Hyperphosphataemia in chronic kidney disease

DRAFT FOR CONSULTATION

Management of hyperphosphataemia: NICE guideline DRAFT (October 2012) of 248

Hyperphosphataemia in chronic kidney disease: management of

hyperphosphataemia in patients with stage 4 or 5 chronic kidney disease

NICE clinical guideline

Draft for consultation, October 2012

This guideline was developed by the NICE Internal Clinical Guidelines Team

following the short clinical guideline process. This document includes all the

recommendations, details of how they were developed and summaries of the

evidence they were based on.



Contents

Introduction ...................................................................................................... 3 Hyperphosphataemia ................................................................................... 3 Drug recommendations ................................................................................ 5

Who this guideline is for ............................................................................... 5 Patient-centred care ......................................................................................... 6 1 Recommendations .................................................................................... 8

1.1 List of all recommendations ................................................................ 8 2 Care pathway .......................................................................................... 11

3 Evidence review and recommendations .................................................. 12 3.1 Dietary management for people with stage 4 or 5 CKD who are not on dialysis ....................................................................................................... 12 3.2 Dietary management for people with stage 5 CKD who are on dialysis ....................................................................................................... 37 3.3 Patient information strategies ........................................................... 53 3.4 Use of phosphate binders in people with stage 4 or 5 CKD who are not on dialysis ............................................................................................. 79 3.5 Use of phosphate binders in people with stage 5 CKD who are on dialysis ..................................................................................................... 100 3.6 Use of supplements in people with stage 4 or 5 CKD who are not on dialysis ..................................................................................................... 200 3.7 Use of supplements in people with stage 5 CKD who are on dialysis ... ........................................................................................................ 202 3.8 Sequencing of treatments ............................................................... 215

4 Notes on the scope of the guideline ...................................................... 218

5 Implementation ...................................................................................... 218

6 Other versions of this guideline ............................................................. 218 6.1 NICE pathway ................................................................................. 218 6.2 ‘Understanding NICE guidance’ ...................................................... 218

7 Related NICE guidance ......................................................................... 219 8 Updating the guideline ........................................................................... 219 9 References ............................................................................................ 220

10 Glossary and abbreviations ................................................................ 232 Appendix A Contributors and declarations of interests ................................ 238

Appendix B List of all research recommendations ....................................... 240 Appendix C Guideline scope ........................................................................ 244 Appendix D How this guideline was developed ........................................... 245

Appendix E Evidence tables ........................................................................ 246 Appendix F Full health economic report ....................................................... 247

Appendix G Clinical guideline technical assessment unit analysis (phosphate binders) ........................................................................................................ 248

Appendices C, D, E, F and G are in separate files.



Introduction 1

Hyperphosphataemia 2

Chronic kidney disease (CKD) describes abnormal kidney function and/or 3

structure. It is common and often exists together with other conditions, such 4

as cardiovascular disease and diabetes. 5

The ‘National service framework for renal services’ adopted the US ‘National 6

Kidney Foundation kidney disease outcomes quality initiative’ (NKF-KDOQI) 7

classification of CKD. This classification divides CKD into 5 stages according 8

to the extent of a person’s loss of renal function. Stage 4 CKD is defined by a 9

glomerular filtration rate (GFR) of 15–30 ml/min/1.73 m2, and stage 5 by a 10

GFR of less than 15 ml/min/1.73 m2.1 11

CKD progresses to these more advanced stages in a small but significant 12

percentage of people. In 2010, the Health Survey for England reported a 13

prevalence of moderate to severe CKD (stages 3 to 5) of 6% in men and 7% 14

in women. CKD stages 4 and 5 were reported at a prevalence of 1% or less. 15

Although this figure might seem small, it translates to a prevalence of up to 16

520,000 people in England alone. 17

When CKD stage 5 advances to end-stage renal disease (ESRD), some 18

people progress to renal replacement therapy (RRT)2. The UK Renal Registry 19

reported that 49,080 adult patients were receiving RRT in the UK at the end of 20

2009. Of these, 25,796 were receiving RRT in the form of dialysis (a 21

population sometimes classified CKD stage 5D). 22

As kidney dysfunction advances, there is a higher risk of mortality and some 23

comorbidities become more severe. Hyperphosphataemia is one example of 24

this, and is because of insufficient filtering of phosphate from the blood by 25

1 A GFR of over 90 ml/min/1.73 m

2 is considered normal unless there is other evidence of kidney

disease. 2 Note: in this guideline, those who choose not to participate in an active treatment programme for their

ESRD (which would generally include RRT, diet, pain management etc), instead opting for

‘conservative management’, are considered to be a subset of the stage 5 population who are not on

dialysis.

http://www.ic.nhs.uk/webfiles/publications/003_Health_Lifestyles/hse09report/HSE_09_Volume1.pdf





poorly functioning kidneys. This means that a certain amount of the phosphate 26

does not leave the body in the urine, instead remaining in the blood at 27

abnormally elevated levels. 28

High serum phosphate levels can directly and indirectly increase parathyroid 29

hormone secretion, leading to the development of secondary 30

hyperparathyroidism. Left untreated, secondary hyperparathyroidism 31

increases morbidity and mortality and may lead to renal bone disease, with 32

people experiencing bone and muscular pain, increased incidence of fracture, 33

abnormalities of bone and joint morphology, and vascular and soft tissue 34

calcification. 35

For adults with stage 4 or 5 CKD who are not on dialysis, the UK Renal 36

Association guidelines recommend that serum phosphate be maintained at 37

between 0.9 and 1.5 mmol/l. For adults with stage 5 CKD who are on dialysis, 38

it is recommended that serum phosphate levels be maintained at between 1.1 39

and 1.7 mmol/l. Because of the improved removal of phosphate from the 40

blood through dialysis, adults on dialysis have different recommended levels 41

to those with stage 4 or 5 CKD who are not on dialysis. 42

For children and young people with stage 4 CKD, the NKF-KDOQI guidelines 43

and European guidelines on the prevention and treatment of renal 44

osteodystrophy recommend that serum phosphate be maintained within 45

age-appropriate limits. For those with stage 5 CKD, including those on 46

dialysis, it is recommended that serum phosphate levels be maintained at 47

between 1.3 and 1.9 mmol/l for those aged 1–12 years, and between 1.1 and 48

1.8 mmol/l during adolescence. 49

Standard management of hyperphosphataemia involves the use of both 50

pharmacological and non-pharmacological interventions, as well as the 51

provision of education and support. However, there is wide variation between 52

renal centres in the UK in how these interventions are used. At the end of 53

2009, data from the UK Renal Registry showed that only 61% of patients 54

receiving haemodialysis and 70% of patients receiving peritoneal dialysis 55

achieved serum phosphate levels within the recommended range. This, 56



together with a rising prevalence of CKD, led to the development of this 57

clinical guideline on the management of hyperphosphataemia. 58

Drug recommendations 59

The guideline does not make recommendations on drug dosage; prescribers 60

should refer to the ‘British National Formulary’ for this information. The 61

guideline also assumes that prescribers will use a drug’s Summary of Product 62

Characteristics to inform decisions made with individual patients. 63

This guideline recommends some drugs for indications for which they do not 64

have a UK marketing authorisation at the date of publication, if there is good 65

evidence to support that use. Where recommendations have been made for 66

the use of drugs outside their licensed indications (‘off-label use’), these drugs 67

are marked with a footnote in the recommendations. 68

Who this guideline is for 69

This document is for healthcare professionals and other staff who care for 70

people with stage 4 or 5 CKD, including those with stage 5 CKD who are on 71

dialysis. Where it refers to children and young people, this applies to all 72

people younger than 18 years old. Where it refers to adults, this applies to all 73

people 18 years or older. 74

75



Patient-centred care 76

This guideline offers best practice advice on the care of adults, children and 77

young people with stage 4 or 5 CKD who have, or are at risk of, 78

hyperphosphataemia. 79

Treatment and care should take into account patients’ needs and preferences. 80

People with CKD should have the opportunity to make informed decisions 81

about their care and treatment, in partnership with their healthcare 82

professionals. If patients do not have the capacity to make decisions, 83

healthcare professionals should follow the Department of Health’s advice on 84

consent and the code of practice that accompanies the Mental Capacity Act. 85

In Wales, healthcare professionals should follow advice on consent from the 86

Welsh Government. 87

If the patient is under 16, healthcare professionals should follow the guidelines 88

in the Department of Health’s ‘Seeking consent: working with children’. 89

Good communication between healthcare professionals and patients is 90

essential. It should be supported by evidence-based written information 91

tailored to the patient’s needs. Treatment and care, and the information 92

patients are given about it, should be culturally appropriate. It should also be 93

accessible to people with additional needs such as physical, sensory or 94

learning disabilities, and to people who do not speak or read English. 95

If the patient agrees, families and carers should have the opportunity to be 96

involved in decisions about treatment and care. 97

Families and carers should also be given the information and support they 98

need. 99

Care of young people in transition between paediatric and adult services 100

should be planned and managed according to the best practice guidance 101

described in the Department of Health’s ‘Transition: getting it right for young 102

people’. 103

http://www.dh.gov.uk/en/DH_103643

http://www.dh.gov.uk/en/DH_103643

http://www.dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsPolicyAndGuidance/DH_085476

http://www.wales.nhs.uk/consent

http://www.wales.nhs.uk/consent






Adult and paediatric healthcare teams should work jointly to provide 104

assessment and services to young people with stage 4 or 5 CKD. Diagnosis 105

and management should be reviewed throughout the transition process, and 106

there should be clarity about who is the lead clinician to ensure continuity of 107

care. 108

109



1 Recommendations 110

1.1 List of all recommendations 111

Please note: unless an age group is explicitly specified, the recommendation 112

applies to all age groups (children, young people and adults). 113

Dietary management: children, young people and adults 114

1.1.1 A specialist renal dietitian, supported by healthcare professionals 115

with the necessary skills and competencies, should carry out a 116

dietary assessment and give individualised information and advice 117

on dietary phosphate management. 118

1.1.2 Advice on dietary phosphate management should be tailored to 119

individual learning needs and preferences, rather than being 120

provided through a generalised or complex multicomponent 121

programme of delivery. 122

1.1.3 Give information about controlling intake of phosphate-rich foods 123

(in particular foods with a high phosphate content per gram of 124

protein, as well as food and drinks with high levels of phosphate 125

additives) to control serum phosphate, while avoiding malnutrition 126

by maintaining a protein intake at or above the minimum 127

recommended level. For those on dialysis, take into account 128

possible dialysate protein losses. 129

1.1.4 If a nutritional supplement is needed to maintain protein intake in 130

children and young people with hyperphosphataemia, offer a 131

supplement with a lower phosphate content, taking into account 132

patient preference. 133

Phosphate binders: children and young people 134

1.1.5 For children and young people, offer a calcium-based phosphate 135

binder as the first-line phosphate binder to control serum 136

phosphate in addition to dietary management. 137



1.1.6 For children and young people, if a series of serum calcium 138

measurements shows a trend towards the age-adjusted upper limit 139

of normal, consider a calcium-based binder in combination with a 140

non-calcium-based binder, having taken into account other causes 141

of rising calcium levels. 142

1.1.7 For children and young people who remain hyperphosphataemic 143

despite adherence to a calcium-based phosphate binder, and 144

whose serum calcium goes above the age-adjusted upper limit of 145

normal, consider a combination of a calcium-based phosphate 146

binder and sevelamer hydrochloride3, having taken into account 147

other causes of raised calcium. 148

Phosphate binders: adults 149

1.1.8 For adults, offer calcium acetate as the first-line phosphate binder 150

to control serum phosphate in addition to dietary management. 151

1.1.9 For adults, consider calcium carbonate if calcium acetate is not 152

tolerated or patients find it unpalatable. 153

1.1.10 For adults with stage 4 or 5 chronic kidney disease (CKD) who are 154

not on dialysis and who are taking a calcium-based binder: 155

consider switching to a non-calcium-based binder if 156

calcium-based phosphate binders are not tolerated 157

consider either combining with, or switching to, a 158

non-calcium-based binder if hypercalcaemia develops (having 159

taken into account other causes of raised calcium), or if serum 160

parathyroid hormone levels are low. 161

1.1.11 For adults with stage 5 CKD who are on dialysis and remain 162

hyperphosphataemic despite adherence to the maximum 163

3 At the time of consultation (October 2012), sevelamer hydrochloride did not have a UK marketing

authorisation for this indication. The prescriber should follow relevant professional guidance, taking

full responsibility for the decision. The patient should provide informed consent, which should be

documented. See the General Medical Council’s Good practice in prescribing medicines – guidance for

doctors for further information.

http://www.gmc-uk.org/guidance/ethical_guidance/prescriptions_faqs.asp




recommended or tolerated dose of calcium-based phosphate 164

binders, consider either combining with, or switching to, a 165

non-calcium-based binder. 166

1.1.12 For adults with stage 5 CKD who are on dialysis and who are 167

taking a calcium-based binder, if serum phosphate is controlled by 168

the current diet and phosphate binder regimen but: 169

serum calcium goes above the upper limit of normal, or 170

serum parathyroid hormone levels are low, 171

consider switching the patient to either sevelamer hydrochloride or 172

lanthanum carbonate, having taken into account other causes of 173

raised calcium. 174

Phosphate binders: children, young people and adults 175

1.1.13 If a combination of phosphate binders is used, titrate the dosage to 176

achieve control of serum phosphate while taking into account the 177

effect of any calcium-based binders used on serum calcium levels 178

(also see recommendations 1.1.6, 1.1.7 and 1.1.10 to 1.1.12). 179

1.1.14 Use clinical judgement and take into account patient preference 180

when choosing whether to use a non-calcium-based binder as 181

monotherapy or in combination with a calcium-based binder (also 182

see recommendations 1.1.10 to 1.1.12). 183

1.1.15 Advise patients to take phosphate binders with food in order to 184

control serum phosphate. 185

Review of treatments: children, young people and adults 186

1.1.16 If vitamin D or dialysis is introduced, review the patient’s diet and 187

phosphate binder requirement in order to maintain the control of 188

serum phosphate. 189


Management of hyperphosphataemia: NICE clinical guideline DRAFT (October 2012) of 248

2 Summary pathway 190

191

192

193

Person with stage 4 or 5 chronic kidney disease at risk of, or with, hyperphosphataemia

Dietary management

Dietary assessment, information and advice [1.1.1, 1.1.2, 1.1.3, 1.1.4]

Phosphate binder management [1.1.13, 1.1.14, 1.1.15]

First-line phosphate

binder treatment for children and young people [1.1.5, 1.1.6]

First-line phosphate

binder treatment for adults

[1.1.8, 1.1.9]

Second-line phosphate

binder treatment for children and young people

[1.1.7]

Second-line phosphate

binder treatment for adults

[1.1.10, 1.1.11, 1.1.12]

Revie

w o

f tr

eatm

ents

[1

.1.1

6]



3 Evidence review and recommendations 194

For details of how this guideline was developed see appendix D. 195

3.1 Dietary management for people with stage 4 or 5 CKD 196

who are not on dialysis 197

3.1.1 Review question 198

For people with stage 4 or 5 CKD who are not on dialysis, is the dietary 199

management of phosphate effective compared to placebo or other treatments 200

in managing serum phosphate and its associated outcomes? Which dietary 201

methods are most effective? 202

3.1.2 Evidence review 203

This review question focused on the use of dietary interventions in the 204

prevention and treatment of hyperphosphataemia in patients with stage 4 or 5 205

CKD who are not on dialysis. These interventions are based on varying 206

degrees of restriction in the intake of phosphate and/or protein, with or without 207

supplementation with keto and amino acids. 208

For this review question, papers were identified from a number of different 209

databases (Medline, Embase, Medline in Process, the Cochrane Database of 210

Systematic Reviews, the Cochrane Central Register of Controlled Trials and 211

the Centre for Reviews and Dissemination) using a broad search strategy, 212

pulling in all papers relating to the dietary management of 213

hyperphosphataemia in CKD. Only randomised controlled trials (RCTs) that 214

compared a dietary intervention with either a placebo or another comparator 215

in patients with stage 4 or 5 CKD who are not on dialysis were considered for 216

inclusion (appendix E). 217

Trials were excluded if: 218

the population included people with CKD stages 1 to 3 or 219

the population included people on dialysis. 220

221



From a database of 3026 abstracts, 244 full-text articles were ordered 222

(including 107 identified through review of relevant bibliographies) and 223

13 papers describing 11 primary studies met the inclusion criteria 224

(Cianciaruso et al, 2008; Cianciaruso et al, 2008; Di Iorio et al, 2003; 225

European Study Group for the Conservative Management of Chronic Renal 226

Failure, 1992; Feiten et al, 2005; Ihle et al, 1989; Jungers et al, 1987; Klahr et 227

al, 1994; Kopple et al, 1997; Lindenau et al, 1990; Malvy et al, 1999; Mircescu 228

et al, 2007; Snetselaar et al, 1994). No paediatric studies meeting the 229

inclusion criteria were found. Table 1 lists the details of the included studies. 230

In order to define the interventions covered and aid an overarching analysis of 231

the dietary management of hyperphosphataemia, protein levels in 232

interventions that included protein restriction were categorised as follows 233

through discussion with the guideline development group (GDG): 234

less than or equal to 0.4 g of protein per kilogram of bodyweight per day: 235

very-low-protein diet 236

more than 0.4 g to less than or equal to 0.75 g of protein per kilogram of 237

bodyweight per day: low-protein diet 238


bodyweight per day: moderate-protein diet 240

more than 1.2 g of protein per kilogram of bodyweight per day: high-protein 241

diet. 242

243

There was some pooling of studies, although this was limited because of the 244

presence of considerable heterogeneity. A prominent cause of such 245

heterogeneity was the widespread use of a number of concurrent 246

interventions that are known to impact the outcomes of interest. Additionally, 247

many studies were powered for purposes other than the management of 248

hyperphosphataemia, such as the preservation of renal function or patient 249

responsiveness to erythropoietin, further contributing to the heterogeneity 250

observed across the included studies. 251

Many papers did not report adherence, an outcome considered critical to 252

decision-making by the GDG, in a binary manner. Rather than defining a 253



patient as ‘adherent’ or ‘non-adherent’ with the dietary prescriptions, authors 254

often provided mean actual intakes for protein, energy and/or phosphate. In 255

order to use these continuous measures as indicators of adherence that could 256

be compared across studies and interventions, the reviewer converted these 257

mean actual intake levels into a percentage of the prescribed level. For 258

example: 259

Prescribed protein intake

Reported mean actual protein intake

Actual protein intake expressed as a percentage of the prescribed level

0.3 g/kg of body weight/day


140% of prescription

that is, actual intake exceeded prescription by 40%

260

Mean differences (MDs) were calculated for continuous outcomes and odds 261

ratios (ORs) for binary outcomes, as well as the corresponding 95% 262

confidence intervals where sufficient data were available. Where 263

meta-analysis was possible, a forest plot is also presented. 264

265



Table 1 Summary of included studies for dietary management for people with stage 4 or 5 CKD who are not on dialysis 266

Study Population Intervention Control Follow-up

Low-protein diet (+ ad-hoc binders + vitamin D) compared with moderate-protein diet (+ ad-hoc binders + vitamin D)

Cianciaruso et al, 2008

RCT

Naples, Italy

n = 423

(392 analysed)

Basal eGFR ≤ 30 ml/min/1.73 m2

Age 18 or older

note: same population as Cianciaruso et al, 2009

1

Baseline serum phosphate (mean SD):

Intervention = 1.4±0.3 mmol/l

Control = 1.2±0.2 mmol/l

0.55 g protein/kg/day

At least 30 kcal/kg/day, reduced to a minimum of 25 kcal/kg/day in overweight patients, or if hypertension and hyperlipidaemia are present

Calcium supplemented at 1000–1500 mg/day as calcium carbonate

Further binders (calcium carbonate/sevelamer) prescribed where needed to treat hyperphosphataemia

Ad-hoc prescription of vitamin D analogues when required






18 months

Monitored every 3 months


RCT

Naples, Italy

n = 423

(392 analysed)

Basal eGFR ≤ 30 ml/min/1.73 m2

Age 18 or older

note: same population as Cianciaruso et al, 2008

1














48 months

Monitored every 3 months

Low-protein diet (+ ad-hoc binders) compared with ad libitum diet (moderate-protein intake as baseline) (+ ad-hoc binders)

Ihle et al, 1989

RCT

Melbourne, Australia

n = 72

(64 analysed)

Mean Cr-EDTA clearance at baseline was 13.8±2.4 ml/min in the LPD group and 15 (±1.8) ml/min in the control group




See evidence tables in appendix E for full inclusion/exclusion criteria

0.4 g of protein/kg body weight/day provided by foods of 75–80% biologic value

700 mg of phosphorus/day (30–40% less than conventional diet)

35–40 kcal/kg/day

‘When-required' phosphate binder use (calcium carbonate or aluminium hydroxide) to control phosphate levels

Ad libitum food intake with at least 0.75 g of protein/kg body weight/day

35–40 kcal/kg/day

‘When-required' phosphate binder use (calcium carbonate or aluminium hydroxide) to control phosphate levels

18 months

Monitored monthly, although data provided for every 3 months

Supplemented very-low-protein diet compared with low-protein diet

European Study Group for the Conservative Management of Chronic Renal

n = 202

GFR < 20 ml/min

Progressive renal failure during the 3 month run-in period


Keto/amino acid mixture

≥ 35 kcal/kg ideal body weight/day

0.6 g/kg/day of protein

≥ 35 kcal/kg ideal body weight/day

1 year



Failure, 1992 (only those with poor renal function)

RCT


Intervention = no data available

Control = no data available

Supplemented very-low-protein diet (+ Ca supplementation) compared with low-protein diet (+ Ca supplementation)

Snetselaar et al, 1994 (MDRD pilot study B)

RCT

USA

n = 66

(58 analysed)

Patients with advanced renal disease

GFR 7.5–24 ml/min/1.73 m2

Progressive increase in serum creatinine

Aged 18 to 75 years






4–9 mg phosphorus/kg/day

0.28 g keto acid/amino acid mixture (Cetolog)/kg/day or 0.22 g amino acid mixture (Aminess Novum)/kg/day

1500 mg calcium/day

0.575 g protein/kg/day – > 0.35 g/kg/day high biologic value dietary protein

5–10 mg phosphorus/kg/day

1500 mg calcium/day

2 years

Monitored monthly

Supplemented very-low-protein diet (+ vitamin D) compared with low-protein diet (+ Ca-based binders/supplements + vitamin D)

Lindenau et al, 1990

RCT

n = 40

Patients with advanced renal failure

CCr < 15 ml/min




0.4 g protein/kg body weight/day

Mixture of KAs and AAs (Ketosteril)

20,000–40,000 U vitamin D/day

0.6 g protein/kg body weight/day

750 mg calcium (as calcium carbonate or lactate)

20,000–40,000 U vitamin D/day

12 months

Supplemented very-low-protein diet (+ ad-hoc binders) compared with low-protein diet (+ ad-hoc binders)

Feiten et al, 2005

RCT

Sao Paulo, Brasil

n = 24

CCr 25 ml/min/1.73 m2

18 or older

Absence of catabolic illnesses, diabetes mellitus, auto-immune disease and malignant hypertension




0.3 g/kg/day of vegetable origin protein diet

1 tablet/5 kg ideal body weight/day, divided into 3 doses taken during meals, of keto acids and amino acids (Ketosteril)

30–35 kcal/kg ideal body weight/day

‘When-required' phosphate binder use to control phosphate levels

0.6 g/kg/day of protein (50% of high biological value)

30–35 kcal/kg ideal body weight/day

‘When-required' phosphate binder use to control phosphate levels

4 months

All outcomes monitored monthly, except for iPTH which was measured at 2-month intervals

Mircescu et al, 2007

RCT

Bucharest, Romania

n = 53

(47 analysed)

Non-diabetic adult patients with CKD

eGFR <30 ml/min/1.73 m2

Stable renal function at least 12 weeks before enrolment (i.e. a reduction in eGFR < 4 ml/min/year and well-controlled arterial BP)

Good nutritional status (i.e. Subjective

0.3 g/kg/day of vegetable proteins

1 capsule/5 kg ideal body weight/day of ketoanalogues of essential amino acids (Ketosteril®)

Total recommended energy intake of 30 kcal/kg/day

‘When-required' prescription of calcium carbonate to maintain serum calcium and serum phosphate within the recommended

Continued conventional low-protein diet with 0.6 g/kg/day (including high biological value proteins)

Total recommended energy intake of 30 kcal/kg/day

‘When-required' prescription of calcium carbonate to maintain serum calcium and serum phosphate within the recommended range

48 weeks

Monitored monthly



Global Assessment Score A/B and serum albumin >35 g/l)





range

Di Iorio et al, 2003

RCT

Italy

n = 20

(19 completed the study period)

Patients with CCr ≤ 25 ml/min/1.73 m2

Treated with LPD (0.6 g/kg body weight/day) and EPO for a period of 6 to 12 months





0.3 g/kg body weight/day of protein of vegetable origin

Supplemented with a mixture of ketoanalogues and essential amino acids (Alfa Kappa) administered at the dose of 1 tablet/5 kg body weight

35 kcal/kg body weight/day

Phosphate binders were administered to maintain serum phosphate levels ≤ 5.5 mg/dl

0.6 g/kg body weight/day of protein

35 kcal/kg body weight/day

Phosphate binders were administered to maintain serum phosphate levels ≤ 5.5 mg/dl

18 months (although data were also available at 24 months for the sVLPD group only)

Data provided for every 6 months

Klahr et al, 1994 (MDRD study B)

RCT

USA

n = 255

GFR 13-24 ml/min/1.73 m2

Aged 18 to 70 years





Note: same population as Kopple et al, 1997 (MDRD study)

2


0.28 g keto acid/amino acid mixture

4–9 g phosphorus/kg/day

calcium carbonate prescribed for hyperphosphataemia 'as required'

0.58 g protein/kg/day (including 0.35 g/kg/day in essential AAs)



18 to 45 months

Kopple et al, 1997 (MDRD study B)

n = 255

GFR 13-24 ml/min/1.73 m2

Aged 18 to 70 years





Note: same population as Klahr et al, 1994 (MDRD study)

2


0.28 g keto acid/amino acid mixture



0.58 g protein/kg/day (including

0.35 g/kg/day in essential AAs)



18 to 45 months

Information on hospitalisation was obtained ‘routinely’ during study visits or if a study visit was missed

Supplemented very-low-protein diet (+ ad-hoc binders + vitamin D) compared with low-protein diet (+ ad-hoc binders + vitamin D)

Jungers et al, 1987

RCT

Paris, France

n = 19

(15 analysed)

Established advanced chronic renal failure, defined by an SCr > 600 µmol/l in

0.4 g/kg/day of mixed quality proteins

< 600 mg/day of phosphates

1 capsule/6 kg ideal body weight/day of ketoanalogues of essential amino acids

0.6 g/kg/day of mainly high biological value proteins

< 750 mg/day of phosphates

Ad-hoc/'when-required' calcium carbonate

Minimum of 3 months and maximum of 18 months, although data given for



1 Note: Cianciaruso et al, 2008 and Cianciaruso et al, 2009 are 2 reports of the same study. Individual outcomes were included from only 1 of the 2 papers, to avoid double-

counting of the results: serum phosphate, adherence, need for additional phosphate management and serum PTH were extracted from Cianciaruso et al, 2008; malnutrition (adverse event) was extracted from Cianciaruso et al, 2009. 2 Note: Klahr et al, 1994 and Kopple et al, 1997 are 2 reports of the same study (the MDRD study). Individual outcomes were included from only 1 of the 2 papers to avoid

double-counting of the results: adherence was extracted from Klahr et al, 1994; hospitalisation (adverse event) was extracted from Kopple et al, 1997.

Abbreviations: AA, amino acids; CCr, creatinine clearance rate; CKD, chronic kidney disease; CrEDTA, chromium-ethylenediaminetetraacetic acid complex; eGFR, estimated glomerular filtration rate; GFR, glomerular filtration rate; iPTH, intact parathyroid hormone; KA, keto acids; LPD, low-protein diet; PTH, parathyroid hormone; RCT, randomised controlled trial; SCr, serum creatinine; SD, standard deviation; sVLPD, supplemented very low-protein diet.

267

268

males or > 500 µmol/l in females, or a CCr of 5–15 ml/min/1.73 m

2

Slowly progressive rate of decline in renal function for at least 3 months

Good general and nutritional condition

Motivated to accept a low-protein diet and be available for follow-up




(Ketosteril®), divided into 3 doses taken during meals (average actual daily dose was 11.3±1.5 tablets/day)

Vitamin D (at a dosage of 25–50 µg/day) used to maintain plasma phosphate levels below 1.7 mmol/l

Total recommended energy intake of 35–40 kcal/kg/day

‘When-required' binders (aluminium hydroxide) used to maintain plasma phosphate levels below 1.7 mmol/l

supplementation for hypocalcaemia

Vitamin D (at a dosage of 50 µg/day) used to maintain plasma phosphate levels below 1.7 mmol/l

Total recommended energy intake of 35–40 kcal/kg/day

‘When-required' binders (aluminium hydroxide) used to maintain plasma phosphate levels below 1.7 mmol/l

‘the end of the study period’

Monitored monthly

Malvy et al, 1999

RCT

Bordeaux, France

n = 50

(38 analysed)

CCr < 19 ml/min/1.73 m2






Supplement of ketoanalogues and hydroxyanalogues of amino acids (Ketosteril)

Daily supplement of vitamin D3 (25–50 mg) and nicotinic acid (25 mg)

Calcium (1–4 g per day), and aluminium hydroxide were added depending on calcium and phosphate plasma levels


Daily supplement of vitamin D3 (25–50 mg) and nicotinic acid (25 mg)

Calcium (1–4 g per day), and aluminium hydroxide were added depending on calcium and phosphate plasma levels

At least 3 months

Monitored at baseline, once a month during the first 3 months, and then every 3 months thereafter



Summary GRADE profile 1 Low-protein diet (+ ad-hoc binders + vitamin D) compared with moderate-protein diet (+ ad-hoc 269

binders + vitamin D) 270

Outcome Number of Studies

Number of patients Effect Quality

Low protein

(+ ad-hoc binders + vitamin D)

Moderate protein


Serum phosphate

18-month follow-up

1 RCT


200 192 Absolute effect

MD = 0.1 mmol/l higher (95% CI: 0 to 0.2 higher)

Very low

Adherence to protein prescription

(% of patients adherent to dietary prescription)

18-month follow-up

1 RCT


56/212

(26.4%)

103/211

(48.8%)

Relative effect

OR = 0.38 (95%CI: 0.25 to 0.57)

Absolute effect

22 fewer per 100 (14 to 30 fewer)

Moderate


(% of prescription, calculated from urinary urea nitrogen)

18-month follow-up

1 RCT



MD = 21.6% higher (95% CI: 18.4 to 24.8 higher)

Very low

Adverse events – malnutrition

(% of patients reaching pre-defined malnutrition point)

48-month follow-up

1 RCT


2/212

(0.94%)

1/211

(0.47%)

Relative effect

OR = 2 (95%CI: 0.18 to 22.23)

Absolute effect

0 more per 100 (0 fewer to 9 more)

Very low

Need for additional phosphate management

(% of patients who received phosphate binders [serum phosphate > 0.85–1.85 mmol/l])

18-month follow-up

1 RCT


27% 36% Relative effect

OR = 0.66 (95% CI: 0.36 to 1.2)

Absolute effect

9 fewer per 100 (19 fewer to 4 more)

Very low

Serum PTH

18-month follow-up

1 RCT



MD =3.4 pmol/l lower (7.3 lower to 0.5 higher)

Very low

Abbreviations: CI, confidence interval; MD, mean difference; OR, odds ratio; RCT, randomised controlled trial.

271

272

273

274

275



Summary GRADE profile 2 Low-protein low-phosphate diet (+ ad-hoc binders) compared with ad libitum diet (minimum: 276

moderate-protein) (+ ad-hoc binders) 277



Low-protein low-phosphate diet

(+ ad-hoc binders)

Ad libitum diet (minimum: moderate-protein)

(+ ad-hoc binders)

Serum phosphate

18-month follow-up

1 RCT

Ihle et al, 1989


MD = 0.05 mmol/l lower (95% CI: 0.28 lower to 0.18 higher)

Very low

Serum PTH

(C-terminal radioimmunoassay)

18-month follow-up

1 RCT

Ihle et al, 1989


MD = 5.9 pmol/l lower (95% CI: 10.9 to 0.9 lower)

Very low

Abbreviations: CI, confidence interval; MD, mean difference; RCT, randomised controlled trial

278

Summary GRADE profile 3 Supplemented very-low-protein diet compared with low-protein diet 279



Supplemented very-low-protein diet

Low-protein diet



12 months follow-up

1 RCT

European Study Group for the Conservative Management of Chronic Renal Failure, 1992

1


Difference in medians = 23.4% higher

Very low

1Data were only extracted for those with ‘poor renal function’

Abbreviations: RCT, randomised controlled trial.

280

281

282

283

284



Summary GRADE profile 4 Supplemented very-low-protein low-phosphate diet with (+ calcium) compared with low-protein 285

low-phosphate diet (+ calcium) 286



Supplemented very-low-protein low-phosphate diet

(+ Ca supplement)

Low-protein low-phosphate diet

(+ Ca supplement)



12-month follow-up1

1 RCT

Snetselaar et al, 1994

2

363

22 Absolute effect

MD = 64.1% higher (95% CI: 47.2 to 81.0 higher)

Very low


(% of prescription, calculated from 3-day diet diary)

12-month follow-up1

1 RCT


2

363

22 Absolute effect

MD = 29.9% higher

Very low

Adherence to phosphate prescription


12-month follow-up1

1 RCT


2

363

22 Absolute effect

MD = 1% lower

Very low

1 Data provided are the weighted means of follow-up data collected throughout the 12-month study period

2 Data were only extracted for ‘Study B’

3 Study is a 3-arm trial; reviewer combined data for 2 arms (both VLPDs, but 1 supplemented with KAs, the other with essential AAs) into a weighted mean and, where

possible, pooled standard deviation, producing a pair wise comparison of sVLPD versus LPD

Abbreviations: AAs, amino acids; CI, confidence interval; LPD, low-protein diet; MD, mean difference; RCT, randomised controlled trial; sVLPD, supplemented very-low-protein diet; VLPD, very-low-protein diet.

287

288

289

290

291

292

293

294



Summary GRADE profile 5 Supplemented very-low-protein diet (+ vitamin D) compared with low-protein diet (+ calcium + 295

vitamin D) 296




(+ vitamin D)

Low-protein diet

(+ calcium + vitamin D)

Serum phosphate

12-month follow-up

1 RCT




Very low

Serum iPTH

12-month follow-up

1 RCT




Very low

Abbreviations: CI, confidence interval; iPTH, intact parathyroid hormone; MD, mean difference; RCT, randomised controlled trial.

297

Summary GRADE profile 6 Supplemented very-low-protein diet (+ ad-hoc binders) compared with low-protein diet (+ ad-298

hoc binders) 299



Supplemented very-low-protein diet (+ ad-hoc binders)

Low-protein diet (+ ad-hoc binders)

Serum phosphate (pooled)

48.5-week follow-up (mean)

3 RCTs

Feiten et al, 2005




MD = 0.30 mmol/l lower (95% CI: 0.53 to 0.18 lower)

Very low



4-month follow-up

1 RCT

Feiten et al, 2005


MD = 28.3% higher (95% CI: 1.7 lower to 58.3 higher)

Very low



1

4-month follow-up

1 RCT

Feiten et al, 2005


MD = 80/0% higher (95% CI: 56.1 to 103.9 higher)

Very low



1

48-week follow-up

1 RCT




Very low


(% of prescription, calculated from

1 RCT

Klahr et al,


Difference in medians = 27.7% higher

Very low



urinary urea nitrogen)1

36-month follow-up

19942,3

Adherence to energy prescription


1

4-month follow-up

1 RCT

Feiten et al, 2005


MD = 3.7% lower

Very low



1

48-week follow-up

1 RCT




Very low


(% of patients defined as malnourished according to SGA)

48-week follow-up

1 RCT


13%4 10%

4 Relative effect

OR = 1.34 (95% CI: 0.56 to 3.23)

Absolute effect


Very low

Adverse events – hospitalisation

(number of patients to undergo first hospitalisation during study)

34-month follow-up

1 RCT

Kopple et al, 1997

3,5

28/126

(22.2%)

32/129

(24.8%)

Relative effect

OR = 0.87 (95% CI: 0.49 to 1.55)

Absolute effect


Very low

Need for additional phosphate management (pooled)

(number of patients who received phosphate binders)

33-week follow-up (mean)

2 RCTs

Feiten et al, 2005

Mircescu et al

5/39

(12.8%)

25/38

(65.8%)

Relative effect

OR = 0.07 (95% CI: 0.01 to 0.59)

Absolute effect

54 fewer per 100 (13 to 64 fewer)

Very low

Serum PTH (pooled)

(immunofluorometric assay/radioimmunoassay)

11-month follow-up (mean)

2 RCTs

Feiten et al, 2005



MD = 9.88 pmol/l (95% CI: 12.73 to 7.03 lower)

Very low

1 Data available for outcome inappropriate for meta-analysis across studies

2 Same population as Kopple et al, 1997

3 Data only extracted from ‘Study B’

4 Note: values were unchanged from baseline

5 Same population as Klahr et al, 1994

Abbreviations: CI, confidence interval; MD, mean difference; OR, odds ratio; PTH, parathyroid hormone; RCT, randomised controlled trial; SGA, subjective global assessment of nutrition.

300



Summary GRADE profile 7 Supplemented very-low-protein diet (+ ad-hoc binders + vitamin D) compared with low-protein 301

diet (+ ad-hoc binders + vitamin D) 302





Low-protein diet



Follow-up unclear

2 RCTs

Jungers et al, 1987

Malvy et al, 1999


MD = 0.30 mmol/l (95% CI: 0.58 to 0.01 lower)

Very low



Follow-up unclear

1 RCT

Jungers et al, 1987

7 8 Absolute effect


1

Very low

Adverse events - need for additional calcium supplementation

(number of patients who received calcium supplementation for hypocalcaemia)

Follow-up unclear

1 RCT

Jungers et al, 1987

0/10 3/9 Absolute effect

33 fewer per 100

Very low

Need for additional phosphate management

(number of patients who received phosphate binders)

Follow-up unclear

1 RCT

Jungers et al, 1987


33 fewer per 100

Very low

Serum PTH

Follow-up unclear

1 RCT

Malvy et al, 1999



Very low

1 Actual intake in the intervention group falls in the ‘low-protein’ rather than the ‘very-low-protein’ range; actual protein intake in the intervention group was 0.66 g/kg/day and

0.72 g/kg/day in the control group (MD 0.06 lower [95% CI 0.43 lower to 0.31 higher i.e. not statistically significant])

Abbreviations: CI, confidence interval; MD, mean difference; PTH, parathyroid hormone; RCT, randomised controlled trial.

303

See appendix E for the evidence tables and GRADE profiles in full. 304



3.1.3 Evidence statements 305

For details of how the evidence is graded, see ‘The guidelines manual’. 306

Dietary management for people with stage 4 or 5 CKD who are not on 307

dialysis 308

A low-protein diet (+ ad-hoc binders) compared with a moderate-protein 309

diet (+ ad-hoc binders) 310

Critical outcomes 311

3.1.3.1 Very-low-quality evidence from 1 RCT of 392 patients showed a 312

low-protein diet with ad-hoc phosphate binder use to be associated 313

with a mean serum phosphate level 0.1 mmol/l higher (95% 314

confidence interval [CI] 0.0 to 0.2 i.e. statistically significant) than a 315

moderate-protein diet with ad-hoc phosphate binder use at 18 316

months. 317

3.1.3.2 Moderate-quality evidence from 1 RCT of 423 patients showed that 318

a low-protein diet with ad-hoc phosphate binder use was 319

associated with a smaller proportion of patients adherent to protein 320

prescription than a moderate-protein diet with ad-hoc phosphate 321

binder use (odds ratio [OR] 0.38 [95% CI 0.25 to 0.57 i.e. 322

statistically significant]). 323

3.1.3.3 Very low-quality evidence from the RCT of 392 patients showed 324

that, as a percentage of the prescribed protein intake, those on a 325

low-protein diet with ad-hoc phosphate binder use exceeded the 326

relevant prescription to a greater extent than those on a 327

moderate-protein diet with ad-hoc phosphate binder use (mean 328

difference [MD] 21.6% more [95% CI 18.4 to 24.8 i.e. statistically 329

significant]). 330

3.1.3.4 Very-low-quality evidence from 1 RCT of 423 patients did not show 331

a statistically significant difference in the incidence of malnutrition 332

over 48 months between those on a low-protein diet with ad-hoc 333

http://www.nice.org.uk/GuidelinesManual



phosphate binder use and those on a moderate-protein diet with 334

ad-hoc phosphate binder use (OR 2 [95% CI 0.18 to 22.23]). 335

Important outcomes 336


a statistically significant difference in the number of patients that 338

required/were prescribed phosphate binders between those on a 339

low-protein diet and those on a moderate-protein diet (OR 0.66 340

[95% CI 0.36 to1.20]). 341


a statistically significant difference in mean serum PTH level 343

between those on a low-protein diet with ad-hoc phosphate binder 344

use and those on a moderate-protein diet with ad-hoc phosphate 345

binder use at 18 months (mean serum PTH level 3.4 pmol/l lower in 346

the low-protein diet group [95% CI -7.3 to 0.5]). 347

A low-protein diet (+ ad-hoc binders) compared with an ad libitum diet 348

(moderate-protein intake prescribed as minimum) (+ ad-hoc binders) 349


3.1.3.7 Very-low-quality evidence from 1 RCT of 64 patients showed a low-351

protein diet with ad-hoc phosphate binder use to be associated with 352

a mean serum phosphate level 0.05 mmol/l lower (95% CI -0.28 to 353

0.18) than an ad libitum diet with ad-hoc phosphate binder use at 354

18 months, although the effect was not statistically significant. 355


3.1.3.8 Very-low-quality evidence from the RCT of 64 patients showed a 357

low-protein diet with ad-hoc phosphate binder use to be associated 358

with a mean serum PTH level 5.9 pmol/l lower (95% CI -10.9 to -0.9 359

i.e. statistically significant) than an ad libitum diet with ad-hoc 360

phosphate binder use at 18 months. 361



A supplemented very-low-protein diet compared with a low-protein diet 362


3.1.3.9 Very-low-quality evidence from 1 RCT of 202 patients showed that, 364

as a percentage of the prescribed protein intake, those on a very-365

low-protein diet supplemented with a mixture of keto acids and 366

amino acids exceeded the relevant prescription to a greater extent 367

than those on a low-protein diet (difference in medians 23.4% 368

more). 369

A supplemented very-low-protein diet (+ Ca supplementation) compared 370

with a low-protein diet (+ Ca supplementation) 371



as a percentage of the prescribed protein intake, those on a 374

very-low-protein diet supplemented with a mixture of keto acids and 375


than those on a low-protein diet when estimated by urinary urea 377

nitrogen (MD 64.1% more [95% CI 47.2 to 81.0 i.e. statistically 378

significant])4. 379

3.1.3.11 Very-low-quality evidence from the same RCT of 58 patients 380

showed that, as a percentage of the prescribed protein intake, 381

those on a very-low-protein diet supplemented with a mixture of 382

keto acids and amino acids exceeded the relevant prescription to a 383

greater extent than those on a low-protein diet when estimated by 384

3-day diet diary (MD 29.9% more)3. 385

3.1.3.12 Very-low-quality evidence from the RCT of 58 patients comparing a 386


amino acids and calcium and a low-protein diet supplemented with 388

4 Note: actual intake in the intervention group fell into the ‘low-protein’ rather than the ‘very-low-

protein’ range (the difference in actual intake between the groups was, however, statistically

significant).



calcium, showed little difference in the deviation of mean 389

phosphate intake from that prescribed, with both groups 390

demonstrating good mean adherence5. 391

A supplemented very-low-protein diet (+ vitamin D) compared with a 392

low-protein diet (+ Ca-based binders/supplements + vitamin D) 393




amino acids and vitamin D to be associated with a mean serum 397

phosphate level 0.04 mmol/l lower (95% CI -0.25 to 0.17) than a 398

low-protein diet supplemented with vitamin D at 12 months, 399

although the effect was not statistically significant. 400


3.1.3.14 Very-low-quality evidence from the same RCT of 40 patients 402

showed a very-low-protein diet supplemented with a mixture of keto 403

acids and amino acids and vitamin D to be associated with a mean 404

serum PTH level 9.8 pmol/l lower (95% CI -15.8 to -3.8 i.e. 405

statistically significant) than a low-protein diet supplemented with 406

vitamin D at 12 months. 407

A supplemented very-low-protein diet (+ ad-hoc binders) compared with 408

a low-protein diet (+ ad-hoc binders) 409


3.1.3.15 Very-low-quality evidence from 3 RCTs analysing a total of 87 411

patients showed a very-low-protein diet supplemented with a 412

mixture of keto acids and amino acids and ad-hoc phosphate 413

binders to be associated with a mean serum phosphate level 414

0.30 mmol/l lower (95% CI -0.43 to -0.18 i.e. statistically significant) 415



significant).



than a low-protein diet with ad-hoc phosphate binder use (mean 416

follow-up ~48.5 weeks).6 417




amino acids and ad-hoc phosphate binders exceeded the relevant 421

prescription to a greater extent than those on a low-protein diet with 422

ad-hoc phosphate binder use when estimated at 4 months by 423

urinary urea nitrogen (MD 80.0% more [95% CI 56.1 to 103.9 i.e. 424

statistically significant])5. 425

3.1.3.17 Very-low-quality evidence from the same RCT, in this case 426

analysing 22 patients, showed that as a percentage of the 427

prescribed protein intake, those on a very-low-protein diet 428

supplemented with a mixture of keto acids and amino acids and 429

ad-hoc phosphate binders exceeded the relevant prescription to a 430

greater extent than those on a low-protein diet with ad-hoc 431

phosphate binder use when estimated at 4 months by 3-day diet 432

diary, although this was not statistically significant (MD 28.3% more 433

[95% CI -1.7 to 58.3])7. 434


as a percentage of the prescribed protein intake, the extent to 436

which mean protein intake at 48 weeks exceeded the relevant 437

prescription was not statistically significantly different in those on a 438


amino acids and ad-hoc phosphate binders compared to those on a 440

low-protein diet with ad-hoc phosphate binder use (deviation of 441

8.4% more [95% CI -2.4 to 19.2]). 442


protein’ range, and fell into the ‘moderate-protein’ rather than the ‘low-protein’ range in the control

group when measured by urinary urea nitrogen (the difference in actual intake between the groups was,

however, statistically significant). 7 Note: actual intake in the intervention group fell into the ‘low-protein’ rather than the ‘very-low-

protein’ range when estimated at 4 months by 3-day diet diary (the difference in actual intake between

the groups was, however, statistically significant).






amino acids and ad-hoc phosphate binders exceeded the relevant 446

prescription to a greater extent than those on a low-protein diet with 447

ad-hoc phosphate binder use at 36 months (difference in medians 448

27.7% more)8. 449


as a percentage of the prescribed energy intake, those on a 451


amino acids and ad-hoc phosphate binders fell below the relevant 453

prescription to a greater extent than in those on a low-protein diet 454

with ad-hoc phosphate binder use at 4 months (MD 3.7% more). 455

3.1.3.21 Very-low-quality evidence from an RCT of 24 patients showed that, 456

as a percentage of the prescribed energy intake, those on a 457


amino acids and ad-hoc phosphate binders fell below the relevant 459

prescription to a lesser extent than in those on a low-protein diet 460

with ad-hoc phosphate binder use at 48 weeks, although this effect 461

was not statistically significant (MD 2.7% less [95% CI -1.2 to 6.6]). 462

3.1.3.22 Very-low-quality evidence from 1RCT of 45 patients showed that a 463


amino acids and ad-hoc phosphate binders was associated with a 465

larger proportion of patients defined as malnourished than among 466

those on a low-protein diet with ad-hoc phosphate binder use (OR 467

1.34 [95% CI 0.56 to 3.23]), although the difference was not 468

statistically significant and the proportions were the same as those 469

observed at baseline. 470



significant).



3.1.3.23 Very-low-quality evidence from 1 RCT of 255 patients showed that 471

a very-low-protein diet supplemented with a mixture of keto acids 472

and amino acids and ad-hoc phosphate binders was associated 473

with a smaller proportion of patients undergoing first hospitalisation 474

than among those on a low-protein diet with ad-hoc phosphate 475

binder use (OR 0.87 [95% CI 0.49 to 1.55]), although the difference 476

was not statistically significant. 477


3.1.3.24 Very-low-quality evidence from 2 RCTs of 77 patients in total 479

showed that a very-low-protein diet supplemented with a mixture of 480

keto acids and amino acids and ad-hoc phosphate binders was 481

associated with a smaller proportion of patients requiring the 482

prescription of phosphate binders than among those on a 483

low-protein diet with ad-hoc phosphate binder use (OR 0.07 [95% 484

CI 0.01 to 0.59 i.e. statistically significant])9. 485

3.1.3.25 Very-low-quality evidence from 2 RCTs analysing a total of 486

41 patients showed a very-low-protein diet supplemented with a 487

mixture of keto acids and amino acids and ad-hoc phosphate 488

binders to be associated with a mean serum PTH level 9.88 pmol/l 489

lower (95% CI -12.73 to -7.03 i.e. statistically significant) than a 490

low-protein diet with ad-hoc phosphate binder use (mean follow-up 491

11 months)8. 492

A supplemented very-low-protein diet (+ ad-hoc binders + vitamin D) 493

compared with a low-protein diet (+ ad-hoc binders + vitamin D) 494



57 patients showed a very-low-protein diet supplemented with a 497


protein’ range, and fell into the ‘moderate-protein’ rather than the ‘low-protein’ range in the control

group when measured by urinary urea nitrogen (the difference in actual intake between the groups was,

however, statistically significant).



mixture of keto acids and amino acids, vitamin D and ad-hoc 498

phosphate binders to be associated with a mean serum phosphate 499

level 0.30 mmol/l lower (95% CI -0.58 to -0.01 i.e. statistically 500

significant) than a low-protein diet with vitamin D and ad-hoc 501

phosphate binder use10. 502




amino acids, vitamin D and ad-hoc phosphate binders exceeded 506

the relevant prescription to a greater extent than those on a 507

low-protein diet with vitamin D and ad-hoc phosphate binder use, 508

although the difference was not statistically significant (MD 45.0% 509

more [95% CI -25.9 to 115.9])9. 510

3.1.3.28 Very-low-quality evidence from 1 RCT of 19 patients found that 511

fewer patients on a very-low-protein diet supplemented with a 512

mixture of keto acids and amino acids, vitamin D and ad-hoc 513

phosphate binders required/were prescribed calcium 514

supplementation than those on a low-protein diet with vitamin D 515

and ad-hoc phosphate binders (0/10 versus 3/9 respectively) 9. 516



fewer patients on a very-low-protein diet supplemented with a 519

mixture of keto acids and amino acids and vitamin D required/were 520

prescribed phosphate binders than those on a low-protein diet with 521

vitamin D (0/10 versus 3/9 respectively)11. 522

10

Note: actual intake fell into the ‘low-protein’ rather than the ‘very-low-protein’ range; actual protein

intake in the intervention group was 0.66 g/kg/day and 0.72 g/kg/day in the control group (MD 0.06

lower [95% CI -0.43 to 0.31, that is, not statistically significant]). 11


intake in the intervention group was 0.66 g/kg/day and 0.72 g/kg/day in the control group (MD 0.06

lower [95% CI -0.43 to 0.31, that is, not statistically significant]).



3.1.3.30 Very-low-quality evidence from 1 RCT of 42 patients showed that a 523


amino acids, vitamin D and ad-hoc phosphate binders to be 525

associated with a mean serum PTH level 33.9 pmol/l lower (95% CI 526

-43.5 to -24.3 i.e. statistically significant) than a low-protein diet with 527

vitamin D and ad-hoc phosphate binder use. 528

3.1.4 Evidence to recommendations 529

Relative value of different outcomes

The GDG discussed the relative importance of the outcomes and agreed that serum phosphate, adherence with dietary prescription, and adverse events such as malnutrition were critical for decision making. The need for additional phosphate management was considered important for decision making, but not critical.

Following the review of the evidence, malnutrition and adherence with treatment featured prominently in the GDG’s discussions.

In addition to being a surrogate outcome, the GDG noted that there is a lot of variability in PTH measurements, in both the level of PTH in the serum and the ability of laboratory techniques to detect these levels. Therefore, PTH was not deemed to be a sufficiently reliable basis from which to formulate recommendations.

Trade-off between benefits and harms

A very-low-protein diet supplemented with keto and amino acids was associated with the use of fewer concurrent phosphate binders and appeared to be effective in reducing serum phosphate. However, the GDG had concerns over the quality and applicability of the evidence. Many studies were powered for other outcomes, such as the preservation of renal function or patient responsiveness to erythropoietin, and most of the diets examined were accompanied by concurrent treatments, including phosphate binders and vitamin D, which could confound the observed effects. The GDG was unsure whether the beneficial effect observed was because of the diet or the calcium-based keto acids in the supplement, which may have phosphate-binding properties. Consequently, the supplement may overestimate the phosphate-lowering effect of the very-low-protein diet. As a result, the GDG did not have confidence in the findings of the studies in the context of managing hyperphosphataemia.

Despite significant protein restriction, malnutrition did not appear to be a significant problem with any of the diets reviewed. However, the GDG had concerns about the possible lack of sensitivity of the non-standardised definitions used in 1 of the papers, which may underestimate the actual incidence of malnutrition. In addition the GDG noted that adherence with the higher energy prescriptions used in the evidence seemed good, and because of this the long-term impact of protein restriction on nutritional status may be masked, requiring much longer follow-up periods to observe adverse effects relating to malnutrition. The GDG was also concerned that the studies may have had insufficient sample sizes, further reducing their sensitivity to detect malnutrition. The small number of events recorded support this suggestion.

Concerns relating to restricted diets also included ‘sub-clinical’



malnutrition that may easily be missed in patients on low- or very-low-protein diets. It was noted that in current practice clinicians aim to maintain patients' dietary protein intake at or above the recommended minimum for CKD patients, rather than restrict nutritional intake in patients with advanced CKD.

It was noted that in the case of supplemented protein restricted diets the keto and amino acids may simply alleviate the harmful effects of very-low-protein diets, leading to the reduced rates of malnutrition and hospitalisation rates observed. There are a number of possible explanations for this association, such as a beneficial effect on nutritional status through substitution for the restricted protein or a positive impact relating to the alleged phosphate-binding properties of these supplements, although there is currently insufficient evidence available.

If patients are prescribed a very-low-protein diet supplemented with keto and amino acids, the GDG was concerned that non-adherence with the supplements might still lead to malnutrition; the additional pill burden of the supplements was a significant concern. However, no evidence on patient adherence with keto/amino acid supplements was found.

Adherence with protein restrictions was poor in all of the diets reviewed. There was no evidence on patients’ views (for example, quality of life), although the GDG felt the observed low adherence with protein restrictions to be indicative. It was felt that expecting patients to comply with protein restrictions, particularly given the unpalatability of such diets, is potentially unrealistic.

The GDG considered that the risks and disadvantages of a protein-restricted diet, with or without keto and amino acid supplementation, were greater than the benefit of the observed phosphate reduction and therefore did not feel it appropriate to recommend this kind of diet for the management of hyperphosphataemia in adults with advanced CKD. For the reasons outlined above, the GDG did not feel that the evidence was sufficient to recommend restricting protein intake below minimum recommended nutrient intake levels, the accepted standards used for protein intake in adults. Current recommended protein intake levels for adults with CKD stage 4 or 5 is a minimum of 0.75 g/kg of ideal body weight/day. Furthermore, given that very-low-protein diets supplemented with keto and amino acids also have a large pill burden, the GDG felt that phosphate binders would be more clinically appropriate than supplementation with keto and amino acids.

Although there was no evidence on the effectiveness of a low-phosphate diet without protein restriction (for example exchanging foods with a high phosphate to protein ratio for foods with a low phosphate to protein ratio), there was consensus among the GDG that this has been effective in their own clinical experience. The GDG considered that advising patients to reduce their intake of phosphate-rich foods is good clinical practice. The GDG also felt that the same principle could be extended to the nutritional supplements/substitutes currently used in CKD management to maintain protein intake, giving low-phosphate options where possible.

In children, the GDG felt that malnutrition is of much greater concern than hyperphosphataemia. Progressive CKD is often associated with decreased spontaneous dietary protein intake, which is a priority for



treatment given the need to maintain growth and adequate nutritional status. For these reasons, as well as the lack of paediatric evidence available, the GDG could not recommend a diet based on protein restriction for children. Recommended intakes are instead age-specific according to reference nutrient intakes.

Economic considerations

This question was not prioritised for health economic analysis because the majority of resource inputs are outside the NHS and personal social services perspective.

Quality of evidence

No evidence was found for children.

For adults, little significant evidence was found to suggest that low-protein prescriptions are effective in managing serum phosphate, although it was felt that this could be a consequence of the poor quality of the evidence.

No evidence was found regarding the use of phosphate restriction alone demonstrating, for example, the effectiveness of a low-phosphate diet achieved without protein restriction or through the restriction of food and drink containing phosphate additives.

No evidence was found that compared a moderate protein restriction to high protein or ad libitum protein intake.

The GDG questioned the generalisability of the evidence to a UK setting (particularly dietary differences) because none of the studies are UK-based. The GDG also felt that the age of the studies may further reduce the generalisability of the evidence because of changes in practice over time.

The GDG was concerned about the reliability of the non-standardised definition of malnutrition used by 1 of the studies in place of widely recognised and accepted standards such as the Subjective Global Assessment of Nutrition (SGA) and Mini Nutritional Assessment (MNA). Adherence with dietary prescriptions was not regularly reported; instead, papers reported mean actual intakes for each group, which the reviewer then compared to prescriptions, producing a surrogate measure of adherence. The GDG also had concerns regarding the known variability of PTH measurements between laboratories; this variation is particularly concerning in the multicentre trials included, especially if those studies used multiple laboratories to analyse their biochemical outcomes.

It was noted that studies were often designed for purposes other than the management of hyperphosphataemia, such as slowing renal decline. For this reason, outcomes of interest were often secondary and concurrent treatments (such as phosphate binders or calcium/vitamin D supplementation) were often used, and at least 1 paper excluded those with metabolic imbalances such as hyperphosphataemia. Moreover, the ad-hoc phosphate binder use and vitamin D supplementation observed may be driving some of the results, particularly those for serum phosphate and PTH.

Reporting in many of the studies was poor. For example, details of study designs were often unclear and results were not always cited in the text of the papers, requiring the reviewer to read the data off available graphs. Additionally, it was not always clear what dietary advice was provided, in what manner it was provided, or who provided it. Unit-of-analysis errors were also common, with analyses not following the intent-to-treat principle.



Other considerations

Keto and amino acid supplementation is not currently used in regular practice in the UK; the GDG felt that making recommendations about their use without further evidence on their safety and effectiveness would be premature.

Usual practice is to advise a reduction in certain types of food: generally those with a high phosphate to protein ratio, such as some dairy products and nuts, or food and drinks with high levels of phosphate additives, such as cola drinks or processed foods. The emphasis is more on the phosphate content of food and drinks rather than focusing on the restriction of protein.

The definition of ‘moderate-protein diet’ used in the analysis (0.75–1.2 g/kg/day) corresponds to the approximately normal level of protein intake for adults in the UK (although the top end of this range would be considered relatively high for CKD patients who are not on dialysis).

A distinction needs to be noted between a ‘moderate-protein diet’ and a ‘moderate-protein restriction’ because these are not the same intervention.

There is limited guidance regarding recommended nutrient intakes for phosphate in adults with CKD. There is some guidance available for children, but the evidence informing this guidance is limited.

The GDG noted relevant recommendations in ‘Chronic kidney disease’ (NICE clinical guideline 73).

530

http://guidance.nice.org.uk/CG73




3.1.5 Recommendations and research recommendations for 531

dietary management for people with stage 4 or 5 CKD 532


Recommendations 534

Recommendation 1.1.3

Give information about controlling intake of phosphate-rich foods (in particular

foods with a high phosphate content per gram of protein, as well as food and

drinks with high levels of phosphate additives) to control serum phosphate,

while avoiding malnutrition by maintaining a protein intake at or above the

minimum recommended level12.

535

3.2 Dietary management for people with stage 5 CKD 536

who are on dialysis 537


For people with stage 5 CKD who are on dialysis, is the dietary management 539

of phosphate effective in managing serum phosphate and its associated 540

outcomes compared to placebo or other treatments? Which dietary methods 541

are most effective? 542


This review question focused on the use of dietary interventions in the 544

prevention and treatment of hyperphosphataemia in patients with stage 545

5 CKD who are on dialysis. These interventions are based on varying degrees 546

of restriction in the intake of phosphate and/or protein, with or without 547

12

Note: recommendation 1.1.3 in full states “Give information about controlling intake of phosphate-

rich foods (in particular foods with a high phosphate content per gram of protein, as well as food and

drinks with high levels of phosphate additives) to control serum phosphate, while avoiding malnutrition

by maintaining a protein intake at or above the minimum recommended level. For those on dialysis,

take into account possible dialysate protein losses.” The aspects of recommendation 1.1.3 that relate to

those on dialysis (including the highlighted final sentence, which has been removed from section 3.1.5

since this section relates only to people with stage 4 or 5 CKD who are not on dialysis) are covered in

section 3.2.



supplementation with keto and amino acids, or on the exchange of a 548

proportion of dietary protein with a low-phosphate protein substitute. 549



Systematic Reviews, the Cochrane Central Register of Controlled Trials and 552

the Centre for Reviews and Dissemination) using a broad search strategy, 553

pulling in all papers relating to the dietary management of 554

hyperphosphataemia in CKD. Only RCTs that compared a dietary intervention 555

with either a placebo or another comparator in patients with stage 5 CKD who 556

are on dialysis were considered for inclusion. 557



the population included people with a diagnosis of CKD stage 5 who are 560

not on dialysis. 561


(including 107 identified through review of relevant bibliographies) and 3 563

papers describing 3 primary studies were selected (Guida et al, 2011; Jiang et 564

al, 2009; Li et al, 2011). No paediatric studies meeting the inclusion criteria 565

were found. Table 2 lists the details of the included studies. 566

Protein levels in dietary interventions that included protein restriction were 567

categorised as follows: 568

less than or equal to 0.8 g of protein per kilogram of bodyweight per day: 569

very-low-protein diet 570


bodyweight per day: low-protein diet 572


bodyweight per day: moderate-protein diet 574

more than 1.2 g of protein per kilogram of bodyweight per day: high-protein 575

diet. 576



These definitions are different to those used in the analysis of dietary 577

interventions in patients with stage 4 or 5 CKD who are not dialysis because 578

of the different protein requirements of this population. A certain amount of 579

protein tends to be lost from the body through dialysis; therefore a high level 580

of protein restriction poses a greater concern over malnutrition. For this 581

reason, the lowest thresholds for protein intake are set at a higher level in 582

patients with stage 5 CKD who are on dialysis. 583

None of the papers reported adherence, an outcome considered critical to 584

decision-making by the GDG, in a binary manner. Rather than defining a 585

patient as ‘adherent’ or ‘non-adherent’ to the dietary prescriptions, authors 586

provided mean levels of actual protein intake. In order to use these continuous 587

measures as indicators of adherence that could be compared across studies 588

and interventions, the reviewer converted these mean actual intake levels into 589

a percentage of the prescribed level. For example: 590








591



confidence intervals where sufficient data were available. There was no 594

pooling of studies because of the lack of similarity in the interventions used. 595

596

597


Management of Hyperphosphataemia: NICE clinical guideline DRAFT (October 2012) of 248

Table 2 Summary of included studies for dietary management for people with stage 5 CKD who are on dialysis 598


Low-phosphorus protein supplement (+ binders) compared with usual diet (+ binders)

Guida et al, 2011

RCT

Italy

n = 27

Haemodialysis treatment for at least 6 months

No allergy to cow’s milk protein

Hyperphosphataemia (serum phosphate ≥ 6.5 mg/dl (i.e. ≥

2.1 mmol/l))

Stable dialysis dose and modality, dietary intakes, body weight and biochemical markers for at least 3 months

All patients were on thrice weekly 4-hour standard bicarbonate haemodialysis




Patients not asked to change their eating habits or their total protein/energy intakes

Partially replace dietary protein intake with a low-phosphorus, low-potassium whey protein concentrate (PROther), instructed to consume 30–40 g dissolved in liquid in place of usual daily portion sizes of protein-rich foods (including milk consumed at breakfast and meat, fish, eggs or dairy and poultry products consumed at lunch time)

Most patients received both phosphate binders (sevelamer, lanthanum carbonate, calcium carbonate, and aluminium hydroxide) and vitamin D analogues – continued pre-study regimen

Usual diet maintained

Most patients received both phosphate binders (sevelamer, lanthanum carbonate, calcium carbonate, and aluminium hydroxide) and vitamin D analogues – continued pre-study regimen

3 months

Blood samples taken before each dialysis session

Very-low-protein diet compared with moderate-protein diet

Jiang et al, 2009

RCT

Shanghai, China

n = 30

(29 analysed at month 12)

note: as trial had 3 arms, the moderate-protein diet group was split in 2 to give 2 pair-wise comparisons

Stable peritoneal dialysis for at least 1 month

Urine output of 800 ml or eGFR of 2 ml/min/1.73 m2

(calculated as an average of the creatinine and urea clearances by 24 hour urine) (that is, residual renal function)

Age 18–80 years





0.6–0.8 g protein/kg ideal body weight/day 1.0–1.2 g protein/kg ideal body weight/day

12 months

Patients were assessed ‘serially’ for 12 months (data given for baseline, 1 month, 2 months, and then every 2 months)

Supplemented very-low-protein diet compared with moderate-protein diet



1 Full study lasted 16 weeks: after 8 weeks on the intervention diet, the intervention group switched to the control diet and were observed for 8 weeks; the control group were

on the control diet for the duration of the 16-week study period. Data is only extracted from the initial 8-week RCT period, and not for the 8-week observation period following.

Abbreviations: BMI, body mass index; Ca x P product, calcium-phosphorus product; eGFR, estimated glomerular filtration rate; KA, keto acids; Kt/V, quantification of dialysis treatment adequacy; MPD, moderate-protein diet; RCT, randomised controlled trial; SD, standard deviation.

599

Jiang et al, 2009

RCT

Shanghai, China

n = 30

(28 analysed at month 12)

note: as trial had 3 arms, the MPD group has been split in 2 to give 2 pair-wise comparisons

Stable peritoneal dialysis for at least 1 month

Urine output of 800 ml or eGFR of 2 ml/min/1.73 m2

(calculated as an average of the creatinine and urea clearances by 24 hour urine) (that is, residual renal function)

Age 18–80 years





0.6–0.8 g protein/kg ideal body weight/day

0.12 g KA (Ketosteril)/kg ideal body weight/day

1.0–1.2 g protein/kg ideal body weight/day

12 months

Patients were assessed ‘serially’ for 12 months (data given for baseline, 1 month, 2 months, and then every 2 months)

Supplemented very-low-protein diet + phosphate restriction compared with moderate-protein diet

Li et al, 2011

RCT

Shanghai, China

n = 40

Patients on maintenance haemodialysis 3 times/week for more than 3 months with Kt/V above 1.2 and no residual renal function

Patients who also had uncontrolled hyperphosphataemia – serum phosphate > 5.5 mg/dl (i.e. > 1.8mmol/l) – after 3 months of conventional calcium carbonate treatment and low-calcium dialysate (1.25 mEq/l) therapy to maintain normal serum calcium levels





0.8 g of protein/kg ideal body weight/day

500 mg phosphate/day

12 pills of KA (Ketosteril)/day

30–35 kcal/kg/day

1.0–1.2 g protein/kg ideal body weight/day according to the patient’s normal diet

8 weeks

Monitored at baseline, 1 week, 2 weeks, 4 weeks and 8 weeks

1



Summary GRADE profile 8 Low-phosphorus protein supplement compared with no intervention 600



Low-phosphorus protein supplement

No intervention

Serum phosphate

3-month follow-up

1 RCT

Guida et al, 2011


MD = 0.7 mmol/l lower (95% CI:1.0 to 0.4 lower)

Low

Serum iPTH

(immunoradiometric assay)

3-month follow-up

1 RCT

Guida et al, 2011


MD = 28.9 pmol/l lower (95% CI:36.5 to 21.3 lower)

Very low

Abbreviations: CI, confidence interval; iPTH, intact parathyroid hormone; MD, mean difference; RCT, randomised controlled trial.

601

Summary GRADE profile 9 Very-low-protein diet compared with moderate-protein diet 602



Very-low-protein diet Moderate-protein diet

Serum phosphate

12-month follow-up

1 RCT

Jiang et al, 2009

18 91

Absolute effect


Very low



12-month follow-up

1 RCT

Jiang et al, 2009

18 91

Absolute effect

MD = 15.5% higher

Very low


(% of patients malnourished, as defined by SGA)

12-month follow-up

1 RCT

Jiang et al, 2009

18 91

Relative effect

OR = 0.54 (95% CI: 0.25 to 1.17)

Very low


(% of patients hospitalised)

12-month follow-up

1 RCT

Jiang et al, 2009

20 101

Relative effect

OR = 0.46 (95% CI: 0.08 to 2.54)

Very low



Serum iPTH

12-month follow-up

1 RCT

Jiang et al, 2009

18 91

Absolute effect

Difference in medians = 1.8 pmol/l higher

Very low

Study is a 3-arm trial (LPD versus sLPD versus MPD); the reviewer has therefore broken down the data into 2 pair-wise comparisons, with the population of the common comparator (the MPD group) divided in 2 for the analysis: for LPD versus MPD, n (MPD) = 9; for sLPD versus MPD (see GRADE profile below), n (MPD) = 8 (except for ‘adverse events [hospitalisation]’: for LPD versus MPD, n [MPD] = 10; for sLPD versus MPD, n [MPD] = 10)

Abbreviations: CI, confidence interval; iPTH, intact parathyroid hormone; LPD, low-protein diet; MD, mean difference; moderate-protein diet; OR, odds ratio; RCT, randomised controlled trial; SGA, subjective global assessment of nutrition; sLPD, supplemented low-protein diet.

603

Summary GRADE profile 10 Supplemented very-low-protein diet compared with moderate-protein diet 604

Outcome Number of Studies Number of patients Effect Quality


Moderate-protein diet

Serum phosphate

12-month follow-up

1 RCT

Jiang et al, 2009



Very low



12-month follow-up

1 RCT

Jiang et al, 2009


MD = 3% higher

Very low


(% of patients malnourished, as defined by SGA)

12-month follow-up

1 RCT

Jiang et al, 2009

0% 20%1

Absolute effect

20 fewer per 100

Very low


(% of patients hospitalised)

12-month follow-up

1 RCT

Jiang et al, 2009

5/20 7/101

Relative effect

OR = 0.62 (95% CI: 0.12 to 3.21)

Absolute effect


Very low

Serum iPTH 1 RCT 18 81 Absolute effect Very low



12-month follow-up Jiang et al, 2009 Difference in medians = 16.0 pmol/l lower

Study is a 3-arm trial (LPD versus sLPD versus MPD); the reviewer has therefore broken down the data into 2 pair-wise comparisons, with the population of the common comparator (the MPD group) divided in 2 for the analysis: for sLPD versus MPD, n (MPD) = 8; for LPD versus MPD (see GRADE profile above), n (MPD) = 9 (except for ‘adverse events [hospitalisation]’: for LPD versus MPD, n [MPD] = 10; for sLPD versus MPD, n [MPD] = 10)

Abbreviations: CI, confidence interval; iPTH, intact parathyroid hormone; LPD, low-protein diet; MD, mean difference; moderate-protein diet; OR, odds ratio; RCT, randomised controlled trial; SGA, subjective global assessment of nutrition; sLPD, supplemented low-protein diet.

605

Summary GRADE profile 11 Supplemented very-low-protein diet + phosphate restriction compared with moderate-protein diet 606



Supplemented very-low-protein diet + phosphate restriction

Moderate-protein diet

Serum phosphate

8-week follow-up

1 RCT

Li et al, 2011



Very low



8-week follow-up

1 RCT

Li et al, 2011


MD = 7.5% higher

Very low


(% of prescription, calculated from nPCR)

8-week follow-up

1 RCT

Li et al, 2011


MD = 9.6% higher

Very low


(% of patients malnourished, as defined by MNA of < 17)

8-week follow-up

1 RCT

Li et al, 2011


0 fewer per 100

Very low

Abbreviations: CI, confidence interval; MD, mean difference; MNA, mini-nutritional assessment; nPCR, normalised protein catabolic rate; RCT, randomised controlled trial.

607


Management of hyperphosphataemia NICE clinical guideline DRAFT (October 2012) of 248



Dietary management for people with stage 5 CKD who are on dialysis 611

A low-phosphorus protein substitute compared with no intervention 612


3.2.3.1 Low-quality evidence from 1 RCT of 27 patients showed a diet in 614

which normal dietary protein is partially exchanged for a low-615

phosphorus protein substitute to be associated with a mean serum 616

phosphate level 0.7 mmol/l lower (95% CI -1.0 to -0.4 i.e. 617

statistically significant) than in patients receiving no intervention 618

(that is, usual diet) at 3 months. 619



diet in which normal dietary protein is partially exchanged for a low-622

phosphorus protein substitute to be associated with a mean serum 623

PTH level 28.9 pmol/l lower (95% CI -36.5 to -21.3 i.e. statistically 624

significant) than in patients receiving no intervention (that is, usual 625

diet) at 3 months. 626

A very-low-protein diet compared with a moderate-protein diet 627



very-low-protein diet to be associated with a mean serum 630

phosphate level 0.17 mmol/l lower (95% CI -0.52 to 0.18) than in 631

patients on a moderate-protein diet at 12 months, although the 632

difference was not statistically significant13. 633

13


intake in the very-low-protein diet group was 0.90 g/kg/day and 0.97 g/kg/day in the moderate-protein

diet group (MD 0.07 lower [95% CI -0.19 to 0.05]).




3.2.3.4 Very-low-quality evidence from the RCT of 27 patients showed that, 634

as a percentage of the prescribed protein intake, those on a very-635

low-protein diet exceeded the relevant prescription to a greater 636

extent than those on a moderate-protein diet at 12 months 637

(MD 15.5% more)14. 638

3.2.3.5 Very-low-quality evidence from the RCT of 27 patients found that 639

8.2% fewer patients were defined as malnourished on a very-low-640

protein diet compared to a moderate-protein diet (OR 0.54 [95% CI 641

0.25 to 1.17]), although the difference was not statistically 642

significant12. 643


15% fewer patients were hospitalised on a very-low-protein diet 645

compared to a moderate-protein diet (OR 0.46 [95% CI 0.08 to 646

2.54]), although the difference was not statistically significant12. 647



very-low-protein diet to be associated with a median serum PTH 650

level 1.8 pmol/l higher than in patients on a moderate-protein diet at 651

12 months12. 652

A very-low-protein diet supplemented with keto acids compared with a 653

moderate-protein diet 654




amino acids to be associated with a mean serum phosphate level 658

0.22 mmol/l lower (95% CI -0.58 to 0.14) than in patients on a 659

moderate-protein diet at 12 months. 660

14


intake in the very-low-protein diet group was 0.90 g/kg/day and 0.97 g/kg/day in the moderate-protein

diet group (MD 0.07 lower [95% CI -0.19 to 0.05]).







than those on a moderate-protein diet when estimated by 3-day 665

diet diary at 12 months (MD 3% more). 666


20% fewer patients were defined as malnourished on a 668


amino acids compared to a moderate-protein diet (no patients in 670

the supplemented very-low-protein diet group were defined as 671

malnourished). 672


10% fewer patients were hospitalised among those on a 674


amino acids compared to those on a moderate-protein diet 676

(OR 0.62 [95% CI 0.12 to 3.21]), although the difference was not 677

statistically significant. 678




amino acids to be associated with a median serum PTH level 682

16.0 pmol/l lower than in patients on a moderate-protein diet at 683

12 months. 684

A very-low-protein diet supplemented with keto acids + phosphate 685

restriction compared with a moderate-protein diet 686



very-low-protein and low-phosphate diet supplemented with a 689

mixture of keto acids and amino acids to be associated with a 690



mean serum phosphate level 0.5 mmol/l lower (95% CI -0.7 to -0.3 691

i.e. statistically significant) than in patients on a moderate-protein 692

diet at 12 months15,16. 693




mixture of keto acids and amino acids exceeded the relevant 697

prescription to a greater extent than those on a moderate-protein 698

diet when estimated by normalised protein catabolic rate at 699

8 weeks (MD 9.6% more)13. 700

3.2.3.15 Very-low-quality evidence from the same RCT showed that, as a 701

percentage of the prescribed protein intake, those on a 702


mixture of keto acids and amino acids exceeded the relevant 704

prescription to a greater extent than those on a moderate-protein 705

diet when estimated by 3-day diet diary at 8 weeks (MD 7.5% 706

more)17. 707

3.2.3.16 Very-low-quality evidence from 1 RCT of 40 patients found no 708

difference in the number of patients defined as malnourished on a 709

very-low-protein and low phosphate diet supplemented with a 710

mixture of keto acids and amino acids compared to a 711

moderate-protein diet (0 in both groups)13. 712

713

714

15

Note: actual intake in the intervention group fell into the ‘low-protein’ rather than the ‘very-low-

protein’ range, and into the ‘high-protein’ rather than the ‘moderate-protein’ range in the control group

when estimated by the normalised protein catabolic rate (the difference in actual intake between the

groups was, however, statistically significant). 16

Phosphate intake was considerably reduced in the intervention group compared to the control (mean

difference [95% CI] at 8 weeks = -305 mg/day [-376 to -234]). 17

Note: actual intake in the intervention group fell into the ‘low-protein’ rather than the ‘very-low-

protein’ range when estimated by 3-day diet diary (the difference in actual intake between the groups

was, however, statistically significant).





The GDG discussed the relative importance of the outcomes and agreed that those considered critical or important for decision-making were the same as those in the review of dietary management in patients with CKD stage 4 or 5 who are not on dialysis.


Protein restriction (without supplementation with keto and amino acids) had only a marginal positive impact upon serum phosphate and PTH; both increased during the study, only to a lesser extent than among those with a ‘normal’ protein intake. The effect of protein restriction without supplementation on the incidence of malnutrition and hospitalisation was not significantly different from that of people with a ‘normal’ protein intake. The GDG noted that adherence with the prescribed protein restriction was very poor, with no significant difference in the actual intake between the 2 groups. This was the likely cause of the similarity in the results of the 2 groups. The GDG was concerned that the risk of malnutrition on a protein restricted diet without supplementation could not be determined. The GDG was also concerned that the study may have been underpowered in terms of sample size, further reducing the sensitivity to detect malnutrition. The small number of events recorded support the suggestion that this study was underpowered.

The addition of keto and amino acid supplements to protein restricted diets did not have a significantly different effect on serum phosphate levels compared to those with a ‘normal’ protein intake, although it did significantly improve adherence, as well as the incidence of malnutrition and hospitalisation. However, the GDG was again concerned that the study may have been underpowered in terms of sample size, reducing their sensitivity to detect malnutrition.

Concerns relating to restricted diets also included ‘sub-clinical’ malnutrition that may easily be missed in those on low- or very-low-protein diets. It was noted that in current practice, clinicians aim to increase protein, or at least maintain patients’ dietary intakes at reference nutrient intake levels, rather than restrict nutritional intake in patients on dialysis.

The GDG noted that supplementation with keto and amino acids may simply mediate the harmful effects of very-low-protein diets, leading to the reduced rates of malnutrition and hospitalisation rates observed. There are a number of possible explanations for this association, such as a beneficial effect on nutritional status through substitution for the restricted protein or a positive impact relating to the alleged phosphate-binding properties of calcium-based keto acids, although there is currently insufficient evidence available.

If patients are prescribed a very-low-protein diet supplemented with keto and amino acids, the GDG was concerned that non-adherence with the supplements might still lead to malnutrition; the additional pill burden of these supplements was considered a significant concern in this area. However, no data were found on patient adherence with keto and/or amino acid supplements.

The GDG did not feel that the evidence for benefits outweighed the possible risks and disadvantages of significant protein restriction in these patients (whether supplemented by keto and amino acids or not), particularly given the lack of effect upon serum phosphate.



Therefore the GDG did not feel it appropriate to recommend diets based on protein restriction for the management of hyperphosphataemia in adults on dialysis. The recommended nutrient intake for protein in adults on haemodialysis is a minimum of 1.1 g/kg of ideal body weight/day; the recommended nutrient intake for protein in adults on peritoneal dialysis is a minimum of 1.1 to 1.2 g/kg of ideal body weight/day. For the reasons outlined above, the GDG did not feel that the evidence was sufficient to recommend restricting protein intake below these levels. Furthermore, given that very-low-protein diets supplemented with keto and amino acids also have a large pill burden, the GDG felt that phosphate binders would be more clinically appropriate than supplementation with keto and amino acids.

In children, the GDG felt that malnutrition is a greater concern than hyperphosphataemia. Progressive CKD is often associated with decreases in spontaneous dietary protein intake and dialysis with a loss of protein from the body; effects that clinicians feel are a priority for treatment given the particular need to maintain growth and adequate nutritional status. For these reasons, accompanied by the lack of paediatric evidence available, the GDG could not recommend a diet based upon protein restriction for children. In addition, as the GDG felt that, for a variety of reasons, it would never be part of standard practice in the management of hyperphosphataemia to limit a child’s protein intake, it would be inappropriate to make such a recommendation. Recommended protein intakes are instead age-specific according to reference nutrient intakes, plus additional protein to attempt to compensate for the potential of dialytic and other protein losses.

Although observed within a short follow-up period, the GDG felt that, in the only study to examine it, limiting phosphate intake had a large impact in reducing serum phosphate, particularly when compared to the effectiveness of a similar intervention differing only in its lack of prescribed phosphate restriction. This result reflected the observations of the GDG in their own clinical and personal experience. The group considered advising patients to reduce their intake of phosphate-rich foods to reflect good clinical practice.

Exchanging a proportion of dietary protein with a low-phosphate protein substitute seems to be an effective intervention, with significant effects upon serum phosphate and PTH. The GDG was, however, concerned that the follow-up was relatively short, the sample size was small, and the study was not UK-based (Italy). These observations lead to reduced confidence in the results observed and uncertainty over the intervention’s long-term effects (for example, on nutritional status). There was also uncertainty as to the sustainability of the low-phosphate protein supplement as a long-term intervention, particularly since there is no data available on adherence or patients’ views on the intervention (such as quality of life data) and the GDG was unsure of its palatability. Additionally, it was felt that this could constitute further and unwelcome ‘medicalisation’ of the life of patients with advanced CKD.

The group felt that more evidence is required before a recommendation can be made for the use of such low-phosphate protein supplements as an intervention for the management of hyperphosphataemia. However, extending the principle that dietary restriction of food and drink rich in phosphate is desirable, the GDG



felt that low-phosphate options should be considered in instances where patients require nutritional supplements and/or substitutes to maintain protein intake.


This question was not prioritised for health economic analysis as the majority of resource inputs are outside of NHS personal social services perspective.

Quality of evidence

There was considerable variation across the 3 included papers in terms of design and the interventions used, and no evidence was found for children. The GDG noted that the limited evidence provided little for consideration. In particular, little significant or reliable evidence was found to suggest that low-protein prescriptions are effective in managing serum phosphate, although it was felt that this could be a consequence of the poor quality of the evidence and the general failure to meet the high requirements of protein restriction.

The GDG was also concerned with the applicability of the evidence to the NHS given that the papers were not UK-based, in particular the 2 based in China. It was felt that generalisability was further limited by the short follow-up periods in 2 of the papers and the relatively small sample sizes across all 3 papers.

The GDG raised concerns relating to the measures used for 2 of the outcomes. The incidence of adherence with dietary prescription was not regularly reported; instead, papers reported mean actual intakes for each group, which the reviewer then compared to prescriptions, producing a surrogate measure of adherence. The GDG also had reservations regarding the known variability of PTH measurements.

It was noted that 1 study was powered for purposes other than the management of serum phosphate, focusing instead on the preservation of residual renal function. Outcomes of interest in this study were therefore secondary.

In another study, the patients’ pre-study regimens of phosphate binders were continued. Although use was well-balanced at baseline, these concurrent treatments may have influenced the good results observed for serum phosphate and PTH in those on the low-phosphate protein substitute.

Reporting across the studies was poor. For example, details of study designs were often unclear and results were not always cited in the text of the papers, requiring the reviewer to read the data off available graphs. Additionally, it was not always clear what dietary advice was provided, in what manner it was provided, or who provided it. Unit-of-analysis errors were also common, with analyses not following the intent-to-treat principle.


Keto and amino acids supplementation is not currently used in practice in the UK; therefore, it was felt that making recommendations about their use without further evidence on their safety and effectiveness would be premature.

Usual practice is to advise a reduction in certain types of food: generally those with a high phosphate to protein ratio, such as some dairy products and nuts, or food and drinks with high levels of phosphate additives, such as cola drinks or processed foods. The emphasis is more on the phosphate content of food and drinks rather than focusing on the restriction of protein.

There is limited guidance available regarding recommended nutrient



intakes for phosphate in adults with CKD. There is some guidance available for children, but the evidence informing this guidance is limited.


716


dietary management for people with stage 5 CKD who are 718

on dialysis 719

Recommendations 720


Give information about controlling intake of phosphate-rich foods (in particular

foods with a high phosphate content per gram of protein, as well as food and

drinks with high levels of phosphate additives) to control serum phosphate,

while avoiding malnutrition by maintaining a protein intake at or above the

minimum recommended level. For those on dialysis, take into account

possible dialysate protein losses.


If a nutritional supplement is needed to maintain protein intake in children and

young people with hyperphosphataemia, offer a supplement with a lower

phosphate content, taking into account patient preference.

721

3.3 Patient information strategies 722


For people with stage 4 or 5 CKD, both those on dialysis and those who are 724

not, are patient information strategies effective at promoting adherence to 725

phosphate-lowering dietary interventions, or in the management of serum 726

phosphate and its associated outcomes? Which patient information strategies 727

are most effective? 728






This review question focused on the use of patient information and education 730

interventions to promote adherence to phosphate-lowering dietary 731

interventions, in the context of both the prevention and treatment of 732

hyperphosphataemia, in patients with stage 4, 5 or 5D CKD. These 733

interventions ranged from the provision of written educational material, to 734

educational videos, to counselling sessions of varying frequency, delivered 735

one-on-one or to groups of patients. Other interventions consisted of 736

behavioural feedback and patient contracts or combinations of multiple 737

concurrent approaches, including 2 or more of the following: written 738

educational material, educational videos, counselling sessions, self-739

management tools (such as medication charts, individualised menus and food 740

exchange lists) or competitions. 741


databases (Medline, Medline in Process, Embase, the Cochrane Database of 743

Systematic Reviews, the Centre for Reviews and Dissemination’s DARE and 744

HTA databases, and PsycINFO) using a broad search strategy, pulling in all 745

papers relating to the use of patient information and education interventions to 746

promote adherence to phosphate-lowering dietary interventions in patients 747

with CKD. RCTs, non-randomised controlled trials (non-RCTs) and controlled 748

before-and-after studies comparing a patient education intervention with either 749

no intervention or another comparator were considered for inclusion. 750



the trial examined only the information to be covered by education, rather 753

than the strategy by which it should be delivered. 754

755


(including 9 identified through review of relevant bibliographies) and 9 papers 757

(7 RCTs, 1 cluster RCT and 1 non-RCT) describing 9 primary studies were 758

selected (Ashurst & Dobbie, 2003; Baraz et al, 2010; Campbell et al, 2008; 759

Chen et al, 2006; Ford et al, 2004; Morey et al, 2008; Shaw-Stuart & Stuart, 760



2000; Sullivan et al, 2009; Tanner et al, 1998). No paediatric studies meeting 761

the inclusion criteria were found. Table 3 lists the details of the included 762

studies. 763

The reviewer analysed studies together where possible, producing a number 764

of pooled comparisons. These were structured as follows: 765

Patient education versus no intervention (beyond usual care) 766

Three papers were included in this pooled comparison (2 RCTs and 1 cluster 767

RCT). Data from the 2 RCTs could be meta-analysed for knowledge scores 768

and serum phosphate. The cluster RCT data for these outcomes could not be 769

included in the meta-analysis because of the unit-of-analysis error found 770

within the paper (data were analysed at the participant- rather than cluster-771

level, and insufficient information was available for the reviewer to correct 772

this). Therefore, this cluster RCT data, along with the other non-meta-773

analysed data from each paper, was recorded individually. 774

Interventions including counselling-based education versus written material 775

alone 776

Two papers were included in this pooled comparison (both RCTs). Because of 777

the absence of common outcomes between the papers, no meta-analysis was 778

performed; data for each outcome from each paper were recorded 779

individually. 780

Multiple component interventions versus single component interventions 781

Four papers were included in this pooled comparison (all RCTs). Data from all 782

4 papers could be meta-analysed for serum phosphate. The other, non-meta-783

analysed data from each paper were recorded individually. 784

Multiple component interventions versus specific single component 785

interventions 786

These comparisons were designed to further differentiate the more general 787

comparison above into the specific single component comparators of the 5 788

included papers: individual patient counselling alone (2 papers), written 789

material alone (1 paper) and video education alone (1 paper). Data from the 2 790



papers comparing multiple component interventions against individual patient 791

counselling alone could be meta-analysed for serum phosphate. The other, 792

non-meta-analysed data from each paper were recorded individually. 793

Significant heterogeneity was observed across all of the included studies, 794

particularly in relation to the structure and content of the interventions studied. 795

There was also considerable variation in the locations of many of these 796

studies, with few conducted in the UK. 797

Many of the papers did not report adherence, an outcome considered critical 798

to decision-making by the GDG, in a binary manner. Rather than defining a 799

patient as ‘adherent’ or ‘non-adherent’ with the dietary prescriptions, authors 800

often provided mean levels of actual protein intake. In order to use these 801

continuous measures as indicators of adherence that could be compared 802

across studies and interventions, the reviewer converted these mean actual 803

intake levels into a percentage of the prescribed level. For example: 804








805

Additionally, for the purpose of this review, the summary term 'self-806

management tool' has been used to collectively describe the aids used to help 807

patients in managing their dietary intake or serum phosphate levels in 808

response to the education provided. These tools range from a fridge magnet 809

detailing high-phosphate foods, to an individualised tracking chart that used 810

visual goals to engage patients in achieving phosphate control. 811



confidence intervals where sufficient data were available. Where meta-814

analysis was possible, a forest plot is also presented. 815



Table 3 Summary of included studies on patient information strategies 816

Study Population Intervention Control

Follow-up

Regular individual oral counselling + written educational material + self-management tool (serum phosphate control) compared with no intervention (beyond usual care)

Ford et al, 2004

RCT

Louisiana, US

n = 70

(63 completed the study)

Haemodialysis patients

Mean serum phosphate of > 6.0 mg/dl (i.e. > 1.9 mmol/l) for 3 months prior to the study

Baseline serum phosphate

(mean SD):

Intervention = 2.2 0.2 mmol/l

Control = 2.3 0.4 mmol/l


In addition to routine care, 20–30 mins/month of diet education focusing on improving serum phosphate control

Education sessions:

Dietitian stressed the importance of all aspects of phosphate control: prevention of renal bone disease; foods high in phosphate; medications; importance of diet, dialysis and drug therapy

Educational tools (bright and attention-grabbing, and including analogies to which patients could relate) included: posters/flipcharts; picture handouts; puzzles; individualised phosphate tracking tool (self-monitoring phosphate levels using visual goals)

Routine care (no education):

Review of monthly laboratory report by the dietitian during monthly nutrition rounds

Although the dietitian did discuss abnormal phosphate levels with these patients, additional patient education materials were not provided

6 months

Monitored monthly

One-off individual oral counselling + written educational material + self-management tool (serum phosphate control) compared with no intervention (beyond usual care)

Sullivan et al, 2009

Cluster RCT

Ohio, US

n = 279

Long-term haemodialysis for at least 6 months

Most recent serum phosphate level and mean level for the previous 3 months both above 5.5 mg/dl (i.e. above 1.8 mmol/l)

18 years or older


(mean SD):




Face-to-face education:

Coordinator met in person with each intervention patient during a dialysis treatment in the first month of the study

Provided approximately 30 minutes of education regarding phosphorus-containing additives and their effect on the phosphate content of foods

Educational device:

A small magnifier in a plastic case, on which common phosphorus-containing additives were printed

Patients were instructed to use the magnifier and list of additives when purchasing food to avoid any items whose ingredient lists include phosphorus-containing additives

Written material:

Continued to receive pre-study nutritional care from their facility’s registered dietitian, which included nutritional status assessment, monthly laboratory test result review (including serum phosphorus levels), and education regarding the renal diet (including the deleterious effects of hyperphosphataemia, dietary sources of phosphorus and ways to limit phosphorus intake)

A study coordinator telephoned control patients during the second month of the study and asked questions about how often they read nutrition facts labels and ingredient lists, ate meals from fast-food restaurants, and received phosphorus-related recommendations from their facility dietitian

Did not receive any education or feedback from

3 months

Monitoring unclear – at baseline and at 3 months?



Each patient also received a handout for each fast-food restaurant the patient reported eating at more than once a month, tailored to the menus of common fast-food chains in the area. Each handout listed specific menu items to be avoided because they contained phosphorus additives, as well as better choices that were free of phosphorus additives and were compatible with other renal dietary requirements.

Study coordinator telephoned patients during the second month to reinforce the instructions and answer any questions

study coordinators

One-off individual oral counselling + written educational material + self-management tool (serum phosphate control) compared with written material alone

Ashurst & Dobbie, 2003

RCT

London, UK

n = 58

(56 analysed)

Dialysis patients with serum phosphate of at least one value above 1.7 mmol/l during 3-month monitoring

Over 18 years

Clinically stable

Baseline serum phosphate (mean):

Intervention = 1.96 mmol/l

Control = 1.98 mmol/l


Teaching session (approx 40 minutes in length), administered by a single, trained dietitian on an individual basis

Used an education tool to improve the patients’ knowledge of phosphate balance in dialysis patients and to assist patients in controlling their own phosphate level.

The education package, ‘A Patient’s Guide to Keeping Healthy: Managing Your Phosphate’ comprised:

A teaching booklet which included: a cartoon and written description of phosphate and calcium functions, absorption and excretion; information on PTH and vitamin D function, their interaction, ways to control their balance, and the consequences of increased levels in the body; therapeutic approaches to phosphate management and the patient’s role, including diet, dialysis and phosphate binders; emphasized the importance of adherence with treatment and medications

Medication record chart - a sheet given to each subject in both groups - one side had the patient’s personal details and phosphate binder prescription, as well as a table to be filled in with the medication dose taken at each meal, the other side had blood results for phosphate, calcium and Ca x P product and the normal

Written material delivered on the dialysis unit by the renal haemodialysis dietitian, who was not the research dietitian

Renal staff nurses or physicians refer those patients with persistent hyperphosphataemia for dietary advice.

The dietetic consultation involves a diet history to assess the patient’s intake, followed by phosphate restriction advice based on an A4 double-sided diet sheet. This diet sheet briefly explains about hyperphosphataemia and its association with bone disease; there is also a list of high-phosphate foods to be avoided and suitable low-phosphate alternatives.

Patients were given the same medication record chart as the intervention group

Patients in the control group were only told their biochemistry results if they asked or if they were above the recommended levels; they did not receive the education session nor the education booklet

3 months

Monitored monthly



range for each

Refrigerator magnet

Patients were asked to complete the medication chart for 2 consecutive weeks

Gave appropriate individual advice about diet, medication adherence and lifestyle

One-off individual oral counselling + written educational material + self-management tool (dietary management) compared with individual oral counselling

Chen at al, 2006

RCT

Peking, China

n = 70

Patients on peritoneal dialysis for 3 months

Clinically stable


(mean SD):




All patients received intensive training (‘traditional’ method) within 2 weeks of catheter implementation (patients enrolled 3 months after dialysis)

Detailed information about food contents and appropriate weight were taught by an experienced dietitian

Portion-sized food aids were used

Patients educated to maintain a daily protein intake level of 0.8–1.2 g/kg/day

Patients taught to correctly record their 3-day dietary intakes

In addition, intervention group patients received:

Education utilising: individualised menu from the dietitian based on food preferences; an exchange list as a reference – every food in 1 list contains an equivalent amount of protein; patients were taught how to correctly use their menu by referring to the exchange list; portion-sized food aids

All patients received intensive training (‘traditional’ method) within 2 weeks of catheter implementation (patients enrolled 3 months after dialysis)

Detailed information about food contents and appropriate weight were taught by an experienced dietitian

Portion-sized food aids were used

Patients educated to maintain a daily protein intake level of 0.8 to 1.2 g/kg/day

Patients taught to correctly record their 3-day dietary intakes

1 month

Pre-and post-intervention (at 1 month)

Regular individual oral counselling (dietary management) compared with written material alone

Campbell et al, 2008

RCT

Brisbane, Australia

n = 62

(50 analysed)

Adults (older than 18 years)

CKD with eGFR < 30 ml/min/1.73 m

2

Not previously seen by a dietitian for stage 4 CKD

Absence of malnutrition from a

Administered by a single dietitian

Individualised dietary prescription, including 125–146 kj/kg/day (i.e. 30–35 kcal/kg/day) and 0.75–1.0 g protein/kg/day

The patients were provided with an initial individual consultation at baseline of up to 60 minutes duration followed by a telephone consultation, commonly of 15–30 minutes duration, bi-weekly for the first month, then monthly

Patients received generic nutrition information for patients with CKD, as provided in regular practice

12 weeks

Monitoring unclear – pre- and post-intervention?



cause other than CKD

Not expected to require RRT within 6 months


(mean SD):




Structure:

1. Clinical Data

Initial: medical history; dialysis treatment plan; anthropometry; biochemistry

Follow-up: changes in medical treatment, medications etc; changes in biochemistry

2. Interview

Initial: nutrition assessment and appetite; evaluate food record; functional ability; psychosocial issues; readiness to change

Follow-up: 24-hour dietary recall; recall of changes made

3. Determine the treatment plan: discussion on the role and effect of diet on renal disease; nutrition prescription; development of goals and target strategies

4. Self-management training: goal setting; menu planning; education on identifying protein, energy and other nutrients; recipe modification; label reading

5. Expected outcomes: meeting set goals; making appropriate food choice; maintains body weight, muscle and fat stores; biochemistry within range

Regular individual oral counselling + written material (serum phosphate control) compared with one-off individual oral counselling

Morey et al, 2008

RCT

London, UK

n = 67


On maintenance haemodialysis for > 6 months

Mean serum phosphate level persistently above of < 1.8 mmol/l over the 3-month review period

Age older than 18 years


(mean SD):


Individual review by a specialist renal research dietitian – assessed/advised monthly from baseline until the end of the study period

Dietitian assessed subjects’ diets using diet histories, and made an approximation of dietary phosphate content and nutritional adequacy, as well as phosphate binder adherence by self-report against prescription

Subjects were advised and educated about dietary phosphate restriction and adherence with phosphate binder prescription while maintaining nutritional adequacy, and a variety of strategies were employed to encourage dietary modification, including: motivational counselling; negotiation; behaviour modification therapy; reminders; reinforcement; supportive care; both written

Individual review by a specialist renal research dietitian - assessed/advised infrequently, that is, at baseline and at the end of the study period

Dietitian assessed subjects’ diets using diet histories, and made an approximation of dietary phosphate content and nutritional adequacy, as well as phosphate binder adherence by self-report against prescription

Subjects were advised and educated about dietary phosphate restriction and adherence with phosphate binder prescription while maintaining nutritional adequacy

6 months

Monitored monthly





and verbal

The research dietitian individualised strategies to each subject

Patients were educated on how to match phosphate binders to the phosphate content of a meal; the dietitian also liaised with the medical team to adjust phosphate binder prescriptions to better match the needs of the individual patient

Regular individual behavioural feedback and contracting (serum phosphate control) compared with no intervention (beyond usual care)

Tanner et al, 1998

RCT

Alabama, US

n = 40


On haemodialysis for at least 2 months

Age range 26-78 years

A history of non-adherence for at least 1 month (non-adherence was defined as: interdialytic weight gain of 3 kg or greater on weekdays and 4 kg or greater on weekends for 6 of the 12 dialysis sessions and/or monthly serum phosphate levels of > 5.9 mg/dl (i.e. > 1.9 mmol/l))


(mean SD):




Monthly progress reports and behavioural contracts were reviewed each month with subjects by the investigator; copies were given to the subjects

Monthly feedback included:

Posting of subject’s phosphorus level and number of acceptable IDWG on the monthly progress report – ‘smiley’ face stickers were used to represent acceptable values and ‘frown’ stickers for unacceptable values

The reports were used to educate subjects on acceptable and unacceptable phosphate values and IDWG

Provision of rewards, if indicated – subjects were provided with ‘smiley’ face stickers to wear for each criteria met, and an additional reward (stickers/candy) if both criteria were met

Instruction on recommended dietary behaviours

Setting of 1 or 2 monthly goals (increasing in complexity over time) to improve subjects’ phosphate and fluid control, which were written on a new contract; this contract was dated and signed by the investigator and subject

Review of previous month’s contract goals and progress – together, the subject and investigator identified reasons for non-adherence and/or improvement from the previous month

No intervention (usual care) 6 months

Serum phosphate data monitored (and provided for) each month; other outcomes were only monitored pre- and post-intervention

Video + oral counselling + competitive competition (serum phosphate control) compared with regular individual oral counselling + written material



Shaw-Stuart & Stuart, 2000

Non-RCT

North Carolina, US

n = 81

Adult patients with end-stage renal failure

Currently receiving haemodialysis

Patients were not at risk for malnutrition


(mean SD):




Educational program: ‘A Taste for Life’:

An educational, informational, motivational patient adherence program directed at dietary and medical regimens

Included: flip chart overviewing the basics of bone disease; ‘bone disease demonstrator’, which dramatised the progression of renal osteodystrophy without intervention; interactive educational modules; educational booklets; motivational posters; creative games and puzzles; videos

In-house educational materials depicting alternatives to high phosphate foods

In-centre achievement contest: ‘Bone Voyage’:

Group divided into teams: assembled to include a balanced sample of high and low adherence patients

Objective was to foster a competitive spirit and raise awareness of adherence in an effort to facilitate patient self-management

Pitted teams racing against each other in sailboats from a start to finish line on a game poster that hung in the centre

Movement of a team from start to finish depended on respective teams achieving goals set for the contest

Points were added up for each team, and at the end of the third month, the team with the most points won.

Prizes were awarded to winning teams and most improved patients each month and at the end of the 3 months

Patients were followed-up regularly by a staff dietitian and were counselled monthly concerning phosphate levels during the entire 9-month study period

Therapy was on an individual basis and involved:

Nutrition counselling consistent with the American Dietetic Association’s National Renal Diet

Instruction regarding the use of phosphate binders

In-house printed information supplemented verbal instruction

12 months

Monitored monthly

Oral interactive group counselling sessions + written material (general CKD + diet) compared with educational video viewed alone

Baraz et al, 2010

RCT

Tehran, Iran

n = 63

Aged older than 18 years

Receiving haemodialysis routinely 3 times a week

Receiving haemodialysis for at least 6 months

Patients invited to attend a class on the days after their haemodialysis sessions

Two educational sessions of up to 30 minutes

The principal investigator (a renal nurse expert) performed the teaching intervention

The education was didactic and interactive:

Individually approached during 2 consecutive dialysis sessions in a week

A 30-minute educational film was shown to each patient while they were having haemodialysis – each patient was started on haemodialysis, and then 1-2 hours after initiation of haemodialysis and ensuring that the patient was stable and

2 months

Monitored bi-monthly



817

Summary GRADE profile 12 Patient education compared with no intervention (beyond usual care) 818



Patient education No intervention (beyond usual care)

Changes in adherence-promoting behaviours - frequency of reading ingredients lists

(measured on scale from 0–100; 0 = lowest reading behaviour score; 100 = highest possible reading behaviour score)

3-month follow-up

1 cluster RCT



MD = 22 higher (95% CI:15 to 30 higher)

Very low

Changes in adherence-promoting behaviours - frequency of reading nutritional fact labels

(measured on scale from 0–100; 0

1 cluster RCT



MD = 9% higher (95% CI:1 to 17 higher)

Very low

Living in a home setting

Not received any educational intervention in the past


(mean SD):



Patients could ask questions at the time of the class

An explicitly interactive portion of the program was held at the end of the class - in this part of the session, patients were encouraged to offer support to each other

At the end of group sessions, each patient received a teaching booklet (‘A Patient Guide to Controlling Dietary Regimen’) to take home

The 2 interventions had similar content, covering:

General knowledge about ESRD and dietary management for haemodialysis

Identification of restricted/non-restricted food

Fluid restrictions

Possible reasons for adherence and non-adherence

ready, they were invited to watch the film

The 2 interventions had similar content, covering:

General knowledge about ESRD and dietary management for haemodialysis

Identification of restricted/non-restricted food

Fluid restrictions

Reasons for adherence and possible reasons for non-adherence

Abbreviations: BMI, body mass index; Ca x P product, calcium-phosphorus product; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; ESRD, end-stage renal disease; IDWG, inter-dialytic weight gain; PTH, parathyroid hormone; RCT, randomised controlled trial; RRT, renal replacement therapy; SD, standard deviation.



= lowest reading behaviour score; 100 = highest possible reading behaviour score)

3-month follow-up

Knowledge scores (pooled)

6-month follow-up

2 RCTs

Tanner et al, 1998

Ford et al, 2004


MD = 8.50% (95% CI:5.88 lower to 22.88 higher)

Very low

Change in knowledge scores1

3-month follow-up

1 cluster RCT



MD = 3% higher (95% CI: 1 lower to 7 higher)

Very low


6-month follow-up

1 RCT

Tanner et al, 1998


MD = 6.97% higher

Very low


6-month follow-up

1 RCT

Ford et al, 2004


MD = 7.6% higher

Very low

Serum phosphate (pooled)2

6-month follow-up

2 RCTs

Tanner et al, 1998

Ford et al, 2004



Very low

Serum phosphate2

3-month follow-up

1 cluster RCT



MD = 0.2 mmol/l lower (95% CI:0.3 lower to 0 higher)

Very low

Changes in beliefs and attitudes towards health - perceptions of self-efficacy for self-monitoring

(Self-Efficacy Survey scores: 13–14 points = high; 15–26 points = moderate; 27–29 points = low)

6-month follow-up

1 RCT

Tanner et al, 1998


MD = 1.01 points higher3

Very low

Changes in beliefs and attitudes towards health – health beliefs

(Health Beliefs Survey scores: 9–10

1 RCT

Tanner et al, 1998


MD = 1.00 points higher4

Very low



points = high; 11–19 points = moderate; 20–27 points = low)

6-month follow-up 1 Data available for outcome inappropriate for meta-analysis across studies

2 Data extracted from the cluster RCT suffered from a unit-of-analysis error (analysed at the level of the individual patient, not at the level of the cluster); there was insufficient

data available for the reviewer to reduce the size of the trial to its effective sample size, and therefore the data will not be pooled with the other serum phosphate data 3 Scores in both groups, at both baseline and at 6 months, were interpreted as 'moderate'

4 Scores in both groups, at both baseline and at 6 months, were interpreted as 'high'

Abbreviations: CI, confidence interval; MD, mean difference; RCT, randomised controlled trial.

819

Summary GRADE profile 13 Interventions including counselling compared with written material alone 820



Interventions including counselling

Written material alone



12-week follow-up

1 RCT



MD = 0%

Very low



12-week follow-up

1 RCT



MD = 9.4% higher

Very low

Serum phosphate

3-month follow-up

1 RCT




Very low


821



Summary GRADE profile 14 Multiple component education/information interventions compared with single component 822

education/information interventions 823



Multiple component education/information interventions

Single component education/information interventions


(number of patients that reached prescribed targets were considered adherent - dietary protein intake greater than 0.8 g/kg/day and less than 1.2 g/kg/day)

1-month follow-up

1 RCT

Chen et al, 2006

20/35

(57.1%)

8/35

(22.9%)

Relative effect

OR = 4.5 (95% CI 1.6 to 12.66):

Absolute effect

34 more per 100 (9 more to 56 more)

Very low



1-month follow-up

1 RCT

Chen et al, 2006


MD = 0%

Very low

Phosphate binder requirement

(number of patients in whom phosphate binders could be withdrawn)

6-month follow-up

1 RCT

Morey et al, 2008

1/34

(2.9%)

1/33

(3.0%)

Relative effect

OR = 0.97 (95% CI: 0.06 to 16.17)

Absolute effect

0 fewer per 100 (from 3 fewer to 31 more)

Very low



4 RCTs

Chen et al, 2006

Morey et al, 2008


Baraz et al, 2010


MD = 0.09 mmol/l (95% CI: 0.23 lower to 0.04 higher)

Very low



Serum phosphate

(number of patients with serum phosphate within acceptable limits)

2-month follow-up

1 RCT

Baraz et al, 2010

18/32 18/31 Relative effect

OR = 0.93 (95% CI: 0.34 to 2.52)

Absolute effect


Very low


824

Summary GRADE profile 15 Multiple component education/information interventions compared with individual patient 825

counselling alone 826




Individual patient counselling alone


(number of patients that reached prescribed targets were considered adherent - dietary protein intake greater than 0.8 g/kg/day and less than 1.2 g/kg/day)

1-month follow-up

1 RCT

Chen et al, 2006

20/35

(57.1%)

8/35

(22.9%)

Relative effect

OR = 4.5 (95% CI 1.6 to 12.66):

Absolute effect

34 more per 100 (9 more to 56 more)

Very low



1-month follow-up

1 RCT

Chen et al, 2006


MD = 0%

Very low

Phosphate binder requirement

(number of patients in whom phosphate binders could be withdrawn)

6-month follow-up

1 RCT

Morey et al, 2008

1/34

(2.9%)

1/33

(3.0%)

Relative effect

OR = 0.97 (95% CI: 0.06 to 16.17)

Absolute effect


Very low



1 RCT

Chen et al, 2006


MD = 0.01 mmol/l (95% CI: 0.1 lower to

Very low



Morey et al, 2008

0.08 higher)


827

Summary GRADE profile 16 Multiple component education/information interventions compared with written material alone 828




Written material alone

Serum phosphate

3-month follow-up

1 RCT




Very low


829

Summary GRADE profile 17 Multiple component education/information interventions compared with video education 830

alone 831




Video education alone

Serum phosphate

2-month follow-up

1 RCT

Baraz et al, 2010



Very low

Serum phosphate

(number of patients with serum phosphate within acceptable limits)

2-month follow-up

1 RCT

Baraz et al, 2010


OR = 0.93 (95% CI: 0.34 to 2.52)

Absolute effect


Very low




832

Summary GRADE profile 18 Oral counselling + video + competitive competition compared with regular individual oral 833

counselling + written material 834



Oral counselling + video + competitive competition

Regular individual oral counselling + written material

Serum phosphate

12-month follow-up

1 observational study

Shaw-Stuart & Stuart, 2000



Very low

Abbreviations: CI, confidence interval; MD, mean difference.

835






Patient information strategies 839

Patient education compared with no intervention (beyond usual care) 840


3.3.3.1 Very-low-quality evidence from 1 cluster RCT of 279 patients 842

showed a greater increase in adherence-promoting behaviours in 843

those who had received a patient education intervention compared 844

to those who had not at 3 months. The frequency of reading 845

ingredients lists increased by 22% (95% CI 15 to 30 i.e. statistically 846

significant) and the frequency of reading nutritional fact labels 847

increased by 9% (95% CI 1 to 17 i.e. statistically significant) in the 848

patient education group compared to those who received usual 849

care only. 850



101 patients showed that at 6 months, mean knowledge scores 853

were 8.50% higher (95% CI 5.88 to 22.88 i.e. statistically 854

significant) in those who received a patient education intervention 855

compared to those who did not. Knowledge scores in the patient 856

education groups improved by 6.97% and 7.6% than in the 2 857

groups receiving usual care only. 858

3.3.3.3 Very-low-quality evidence from the cluster RCT of 279 patients 859

found that at 3 months, knowledge scores in the patient education 860

group improved by 3% (95% CI -1 to 7) than in those who received 861

usual care only, although the difference was not statistically 862

significant. 863

3.3.3.4 Very-low-quality evidence from the 2 RCTs of 100 patients showed 864

patient education to be associated with a mean serum phosphate 865




level 0.19 mmol/l lower (95% CI -0.87 to 0.49) at 6 months than in 866

patients who received usual care only, although the difference was 867

not statistically significant. 868

3.3.3.5 Very-low-quality evidence from the cluster RCT of 279 patients 869

found patient education to be associated with a mean serum 870

phosphate level 0.2 mmol/l lower (95% CI -0.3 to no difference i.e. 871

statistically significant) at 3 months than in patients who received 872

usual care only. 873

3.3.3.6 Very-low-quality evidence from 1 RCT of 38 patients showed no 874

difference in the changes observed in patients’ beliefs and attitudes 875

towards health, as measured by a ‘self-efficacy survey’ and a 876

‘health beliefs survey’, whether patients had received a patient 877

education intervention or usual care only. 878

Interventions including counselling compared with written material 879

alone 880


3.3.3.7 Very-low-quality evidence from 1 RCT of 50 patients showed that at 882

12 weeks there was no difference in the amount of deviation of the 883

mean protein intake (estimated by 3-day diet diary) from the 884

prescribed level of protein intake between those who received 885

counselling-based interventions and those who received the renal 886

units’ standard written educational material following a diet history, 887

with both means falling within the prescribed limits. 888


as a percentage of the prescribed energy intake, those who 890

received counselling-based interventions fell below the relevant 891

prescription to a greater extent than those who received the renal 892

units’ standard written educational material following a diet history, 893

when estimated by 3-day diet diary at 12 weeks (MD 9.4% less). 894




3.3.3.9 Very low-quality evidence from an RCT of 56 patients showed 896

counselling-based interventions to be associated with a mean 897

serum phosphate level 0.34 mmol/l lower (95% CI -0.58 to -0.10 i.e. 898

statistically significant) at 3 months than in patients who received 899

the renal units’ standard written educational material following a 900

diet history. 901

Multiple component interventions compared with single component 902

interventions 903



after 1 month significantly more patients who received multi-906

component educational interventions were considered to be 907

adherent with dietary protein prescriptions than those who received 908

single-component educational interventions (OR 4.5 [95% CI 1.60 909

to 12.66 i.e. statistically significant]). 910

3.3.3.11 Very-low-quality evidence from the RCT of 70 patients showed that 911

at 1 month there was no difference in the amount of deviation of the 912



multi-component educational interventions and those who received 915

single-component educational interventions. 916


after 6 months there was no statistically significant difference in the 918

number of patients who could have phosphate binders withdrawn 919

from their regimen between those who received multi-component 920

educational interventions and those who received single-921

component educational interventions (OR 0.97 [95% CI 0.06 to 922

16.17]). 923

924





showed no statistically significant difference in mean serum 927

phosphate levels between those who received multi-component 928


component educational interventions (mean serum phosphate level 930

0.09 mmol/l lower [95% CI -0.23 to 0.04; mean follow-up of 931

3 months]). 932



number of patients considered to have serum phosphate within 935

acceptable limits between those who received multi-component 936


component educational interventions (OR 0.93 [95% CI 0.34 to 938

2.52]). 939

Multiple component interventions (one-off individual oral counselling + 940

written educational material + self-management tool; regular individual 941

oral counselling + written educational material) compared with 942

individual patient counselling alone 943



after 1 month significantly more patients who received multi-946

component educational interventions were adherent with dietary 947

protein prescriptions than those who received individual patient 948

counselling alone (OR 4.5 [95% CI 1.60 to 12.66 i.e. statistically 949

significant]). 950

3.3.3.16 Very-low-quality evidence from the RCT of 70 patients showed that 951

at 1 month there was no difference in the amount of deviation of the 952





multi-component educational interventions and those who received 955

individual patient counselling alone. 956



number of patients who could have phosphate binders withdrawn 959

from their regimen between those who received multi-component 960

educational interventions and those who received individual patient 961

counselling alone (OR 0.97 [95% CI 0.06 to 16.17]). 962



showed multi-component educational interventions to be 965

associated with a mean serum phosphate level 0.01 mmol/l lower 966

(95% CI -0.10 to 0.08) than patients who received individual patient 967

counselling alone (mean follow-up of 3 months), although the 968

difference was not statistically significant. 969

Multiple component interventions (one-off individual oral counselling + 970

written educational material + self-management tool) compared with 971

written material alone 972


3.3.3.19 Very-low-quality evidence from 1 RCT of 56 patients showed multi-974

component educational interventions to be associated with a mean 975

serum phosphate level 0.34 mmol/l lower (95% CI -0.58 to -0.10 i.e. 976

statistically significant) at 3 months than patients who received 977

written educational material alone. 978

Multiple component interventions (interactive group counselling + 979

written educational material) compared with video education alone 980






number of patients considered to have serum phosphate within 984

acceptable limits between those who received multi-component 985

educational interventions and those who received video education 986

alone (OR 0.93 [95% CI 0.34 to 2.52]). 987

3.3.3.21 Very-low-quality evidence from the RCT of 63 patients showed 988

multi-component educational interventions to be associated with a 989

mean serum phosphate level 0.05 mmol/l lower (95% CI -0.22 to 990

0.13) than in patients who received individual patient counselling 991

alone at 2 months, although the difference was not statistically 992

significant. 993

Oral counselling + video + competition compared with regular individual 994

oral counselling + written material 995


3.3.3.22 Very-low-quality evidence from 1 non-RCT of 81 patients showed 997

the intervention (oral counselling and a video relating to serum 998

phosphate control, plus a competition) to be associated with a 999


-0.14 i.e. statistically significant) at 12 months than that in patients 1001

who received the control (oral counselling and written material 1002

relating to dietary management and phosphate binder use for 1003

serum phosphate control). 1004



The GDG discussed the possible outcomes and agreed that serum phosphate, changes in adherence-promoting behaviours, and adherence with dietary prescription were critical to their decision-making. Phosphate binder requirement, changes in knowledge and changes in beliefs and attitudes towards health were considered important, though not critical.

The lack of evidence led to limited differentiation in the relative value of these outcomes, although following review of the evidence serum phosphate featured prominently in the GDG’s discussions.

Although important to decision making, it was felt that adherence-promoting behaviours, knowledge and beliefs and attitudes towards health were only meaningful in the management of hyperphosphataemia in the context of an associated change in serum phosphate levels, rather than as endpoints themselves. In isolation,



these outcomes do not necessarily translate directly into a meaningful effect in the management of serum phosphate. Therefore when interpreting data relating to these 3 outcomes the GDG looked for both a change in the outcomes and a concurrent significant difference in serum phosphate between trial arms.

Knowledge about phosphate-lowering dietary interventions was felt to be particularly far removed, or ‘indirect’, in its impact upon serum phosphate control and was therefore downgraded in quality as a surrogate outcome. The GDG felt that improving a person’s knowledge of a subject does not always bring about changes in that person’s behaviour. In this way, improved knowledge relating to the dietary management of hyperphosphataemia alone does not directly necessitate better adherence with dietary prescriptions or improvements in serum phosphate control.


Patient education interventions that exceed ‘usual care’ had a positive effect on adherence, although the evidence was limited and of low quality. Greater improvements in knowledge and beliefs and attitudes towards health following these interventions appeared to be limited when compared to usual care, and serum phosphate levels were not significantly different between the 2 groups; again, the evidence was limited and of low (or very low) quality. The GDG felt that, despite its limitations, the evidence supported the view that usual care is sufficient in improving a patient’s serum phosphate control, and that interventions over and above this may not be necessary in routine practice.

In defining what constitutes good ‘usual’ practice, the GDG noted that counselling-based interventions were more effective than written material alone, although the evidence was of very low quality and there was some uncertainty over the statistical methods employed. The GDG felt that, in their experience and in conjunction with this evidence, advisory counselling sessions with patients represent good clinical practice. It was felt that counselling-based educational sessions represent an effective opportunity to explore different dietary management options with patients, as well as to monitor their suitability and progress. The GDG considered individualised approaches to be particularly important since different patients will have different diets, needs and preferences. These should be evaluated through dietary assessment, from which a treatment regimen can be developed. Additionally, people learn and respond to interventions in different ways and different approaches to dietary education and management may therefore need to be explored over the course of a patient’s treatment. However, it was felt that such exploration would not routinely require the use of an educational programme using multiple delivery methods (for example, including counselling sessions and comprehensive written material and videos/DVDs and self-management tools). Such interventions were not found to deliver a significant additional benefit over the effect seen following simpler, single-approach interventions, although again the quality of evidence was low or very low.

Given the broad, in-depth knowledge required in formulating effective, individualised therapeutic options, the GDG felt that a specialist renal dietitian would be the most appropriate person to conduct a patient’s dietary assessment and offer them individualised advice. It was also felt that early contact with a dietitian is important as a means of



preventing patient misinformation, for example what constitutes phosphate-rich food and drink. The risks of misinformation can also be mediated by appropriately trained, multidisciplinary healthcare professionals/teams, who also play an important role in a patient’s ongoing dietary education, reinforcing nutritional advice and providing support on a more day-to-day basis. It is important for patients, especially those in the early stages of CKD, to understand the need to manage their health and minimise the risks they face. Education empowers many patients to take steps to limit such risks and, in this instance, to minimise the impact of high phosphate levels on their bones and vasculature.

The GDG highlighted that, although no evidence was found for this population; children need to be considered separately because parents and/or carers provide food and drink, and because their dietary needs change as they get older. Therefore parents and/or carers should also be educated, and the monitoring and revision of advice will need to be more frequent depending on the age of the child and their nutritional and clinical status. Because of the specific nature of children’s dietary needs and habits, the GDG felt that a specialist renal dietitian, specifically a paediatric specialist renal dietitian, would be the most appropriate person to conduct a child’s dietary assessment and offer the individualised advice. The multidisciplinary health professionals and teams that support a child’s ongoing dietary management should be similarly aware of the specific requirements of children with CKD, and the ways in which these change over time.

The GDG also felt that it is important to include information relating to phosphate binder use, giving the specific example of the need to take binders with high-phosphate snacks and not simply with meals, as well as the need to match binder dose with the phosphate load in the snack or meal.


Early contact with a dietitian will require adequate availability.

It is not just dietitians who play a role in patient education; nurses, doctors and psychologists also play an important role, and these healthcare professionals should be appropriately trained. Dietitians will have a role in educating and supporting them.

The GDG noted that substantial resources and costs would be incurred in the development of complex, non-individualised programmes of multiple, concurrent educational approaches, with little evidence for additional benefit.

Quality of evidence

No studies compared the intervention of interest against a true ‘no intervention’ comparator, only usual care.

Limited evidence was found for those not on dialysis (only 1 study), and no evidence on education interventions in children was found.

Definitions of usual care varied across the studies, and were not always clear or explicit.

A significant amount of heterogeneity/diversity was observed across the interventions examined; it was difficult to pool studies to produce overarching interpretations of effectiveness. Content, in addition to the structure of the interventions, varied widely across the studies; for example, some interventions focused on serum phosphate control through diet, some on serum phosphate control more generally (for example, the use of binders), and some on more general dietary



management in CKD (for example, information on fluid intake). There were differences between the papers in the populations studied. Most notably, some studies included only those expected to be adherent, some only those expected to be non-adherent.

The GDG questioned how applicable the evidence is to a UK setting, particularly since many of the studies were not conducted in the UK.

Adherence with dietary prescription was not regularly reported; instead, papers reported mean actual intakes for each group, which the reviewer then compared to prescriptions, producing a surrogate measure of adherence.

Reporting in many of the studies was poor: details of study designs were often unclear, results were not always cited in the text of the papers, requiring the reviewer to read the data off available graphs, and 1 paper did not contain a statistics section leaving uncertainty as to the measure of variance used. Unit-of-analysis errors were also common, with analyses not following the intent-to-treat principle, and 2 studies had follow-up periods that the GDG considered too short (less than 3 months).


In the context of the evidence reviewed, the ‘standard care’ provided seems to be sufficient for educating patients, although what constitutes ‘standard care’ is likely to vary across the UK. Concerns were raised that some patients receive a level of care that is below the standard of the studies reviewed, particularly in the long periods of time they wait to receive dietary advice.

Nurses play a significant role in the education process. They may have the most contact and often the greatest rapport with patients, and are important members of the multidisciplinary teams in reinforcing and supporting implementation of the nutritional advice developed by the dietitian.


1006






patient information strategies 1008

Recommendations 1009


A specialist renal dietitian, supported by healthcare professionals with the

necessary skills and competencies, should carry out a dietary assessment

and give individualised information and advice on dietary phosphate

management.


Advice on dietary phosphate management should be tailored to individual

learning needs and preferences, rather than being provided through a

generalised or complex multicomponent programme of delivery.

1010

3.4 Use of phosphate binders in people with stage 4 or 5 1011

CKD who are not on dialysis 1012


For people with stage 4 or 5 CKD who are not on dialysis, are phosphate 1014

binders effective compared with placebo or other treatments in managing 1015

serum phosphate and its associated outcomes? Which is the most effective 1016

phosphate binder? 1017


This review question focused on the use of phosphate-binding medications to 1019

prevent and treat hyperphosphataemia in patients with stage 4 or 5 CKD who 1020

are not on dialysis. These pharmacological interventions bind dietary 1021

phosphate when taken with food (that is, not on an empty stomach), and can 1022

broadly be defined as: 1023

calcium-based: calcium carbonate or calcium acetate, and 1024



non-calcium-based: sevelamer hydrochloride, sevelamer carbonate, 1025

lanthanum carbonate, aluminium hydroxide or magnesium carbonate. 1026

1027



Systematic Reviews and the Centre for Reviews and Dissemination) using a 1030

broad search strategy, pulling in all papers relating to the use of phosphate 1031

binders in the management of hyperphosphataemia in CKD. Only RCTs that 1032

compared a phosphate binder with either a placebo or another comparator in 1033

patients with stage 4 or 5 CKD who are not on dialysis were considered for 1034

inclusion. The first phase from crossover studies was considered for inclusion 1035

but only if the data were extractable from the overall results. 1036


the population included people with stages 1 to 3 CKD; or 1038

the population included people on dialysis; or 1039

data were obtained through an open-label extension of an RCT. 1040

1041

From a database of 1480 abstracts, 300 full-text articles were ordered and 4 1042

papers were selected (Caglar et al, 2008; Qunibi et al, 2011; Russo et al, 1043

2007; Sprague et al, 2009). No paediatric studies meeting the inclusion 1044

criteria were found. Table 4 lists the details of the included studies. 1045

There was some pooling of studies, although there was considerable 1046

heterogeneity. One prominent cause was the widespread use of a range of 1047

concurrent interventions that are known to impact the outcomes of interest. 1048

Additionally, the time periods for which data were available varied widely 1049

between the studies. 1050

Mean differences (MDs) were calculated for continuous outcomes and relative 1051

risks (RRs) for binary outcomes, as well as the corresponding 95% confidence 1052

intervals where sufficient data were available. Meta-analyses were conducted 1053

either through the pooling of studies into single, direct pair-wise comparisons, 1054

or through the pooling of 3 or more interventions into a network meta-analysis. 1055



These interventions could then be compared, both directly and indirectly, 1056

against each other in multiple treatment comparisons (MTCs). 1057

1058

1059



Table 4 Summary of included studies for the use of phosphate binders in people with stage 4 or 5 CKD who are not on 1060

dialysis 1061

Study (Country)

Interventions Follow-up Target serum PO4/Ca (mmol/l)

Baseline PO4/Ca (mmol/l ± SD)

Outcomes Conclusions

Caglar et al, 2008

RCT

(Turkey)

Sevelamer hydrochloride

versus

Calcium acetate

8 weeks PO4

<1.78

Sevelamer

PO4: 2.52 ± 0.19

Ca: 2 ± 0.1

Calcium acetate

PO4: 2.52 ± 0.23

Ca: 2.2 ± 0.1

Serum phosphate

Serum calcium

No significant differences in serum Ca or PO4 levels as a result of binder assignment, although the non-significant serum Ca levels and Ca x PO4 levels were lower in the sevelamer group.

Qunibi et al, 2011

RCT

(USA)

Calcium acetate

versus

Placebo

12 weeks PO4

1.45 to 0.87

Ca

2.54 to 2.12

Calcium acetate

PO4: 1.65 ± 0.4

Ca: 2.27 ± 0.17

Placebo

PO4: 1.65 ± 1.36

Ca: 2.2 ± 0.2

Serum phosphate

Proportion with controlled serum phosphate

Serum calcium

Hypercalcaemia

Adherence

Calcium acetate was effective in reducing serum PO4 over a 12 week period.

Serum Ca was significantly higher in the calcium acetate group compared to the placebo group.

Russo et al, 2007

RCT

(Italy)

Control diet only

versus

Calcium carbonate

versus


104 weeks None given Calcium carbonate

PO4: 1.48 ± 0.48

Ca: 2.24 ± 0.17

Sevelamer

PO4: 1.55 ± 0.55

Ca: 2.3 ± 0.05

Serum phosphate

Serum calcium

Coronary artery calcification

Progression of coronary artery calcification was slowed in those receiving calcium carbonate but that there was no progression in those treated with sevelamer.

Sprague et al, 2009

RCT

(USA)

Lanthanum carbonate

versus

Placebo

8 weeks PO4

<1.49

Lanthanum

PO4: 1.71 ± 0.22

Ca: 2.22 ± 0.15

Placebo

PO4: 1.74 ± 0.23

Ca: 2.24 ± 0.12

Serum phosphate

Proportion with controlled serum phosphate

Serum calcium

Adverse events

Lanthanum was effective in controlling serum PO4 compared to placebo.

There was a slight increase in serum Ca in the lanthanum group.

The safety profile was similar between the treatments.

Abbreviations: Ca, calcium; Ca x PO4, calcium-phosphorus product; PO4, phosphate; RCT, randomised controlled trial; SD, standard deviation.



1062

Summary GRADE profile 19 Calcium acetate compared with placebo 1063



Calcium acetate Placebo

Serum phosphate

(number of patients in whom control of serum phosphate achieved)

12-week follow-up

1 RCT

Qunibi et al, 2011

22/37

(59.5%)

15/41

(36.6%)

Relative effect

RR = 1.63 (95% CI:1.00 to 2.63)

Absolute effect

230 more per 1000 (0 to 596 more)

Very low

Adherence

(% of prescribed pills taken)

12-week follow-up

1 RCT

Qunibi et al, 2011


MD = 0.70 lower (95% CI: 7.16 lower to 5.76 higher)

Low

Serum calcium

(number of patients with hypercalcaemia)

12-week follow-up

1 RCT

Qunibi et al, 2011

5/37

(13.5%

0/41

(0%)

Relative effect

RR = 12.16 (95% CI: 0.7 to 212.64)

Very low

Abbreviations: CI, confidence interval; MD, mean difference; RCT, randomised controlled trial; RR, relative risk.

1064

Summary GRADE profile 20 Calcium carbonate compared with sevelamer hydrochloride 1065



Calcium carbonate Sevelamer

Serum phosphate

2-year follow-up

1 RCT

Russo et al, 2007



Very low

Serum calcium

2-year follow-up

1 RCT

Russo et al, 2007


MD = 0.02 mmol/l higher (95% CI: 0.06 lower to 0.10 higher)

Very low

Coronary calcification scores 1 RCT

Russo et al,


MD = 20 higher (95% CI: 262.96 lower

Very low



2-year follow-up 2007 to 302.96 higher)

Annualised change in coronary calcification scores

2-year follow-up

1 RCT

Russo et al, 2007


MD = 146.66 higher (95% CI: 35.77 to 275.55 higher)

Very low


1066

Summary GRADE profile 21 Lanthanum compared with placebo 1067



Lanthanum Placebo

Serum phosphate

(number of patients in whom control of serum phosphate achieved)

8-week follow-up

1 RCT

Sprague et al, 2009

25/56

(44.6%)

9/34

(26.5%)

Relative effect

RR = 1.69 (95% CI: 0.9 to 3.17)

Absolute effect


Very low

Adverse events – nausea and vomiting

(number of patients to experience adverse event)

8-week follow-up

1 RCT

Sprague et al, 2009

9/56

(16.1%)

4/34

(11.8%)

Relative effect

RR = 1.37 (95% CI: 0.46 to 4.10)

Absolute effect


Very low

Note: a small number of patients were CKD stage 3 in each group (Lanthanum 25%, Placebo 20.6%)

Abbreviations: CI, confidence interval; RCT, randomised controlled trial; RR, relative risk.

1068

Summary GRADE profile 22 Lanthanum compared with calcium acetate 1069



Lanthanum Calcium acetate

Serum phosphate

(number of patients in whom control

2 RCTs

Qunibi et al,

25/56 22/37 Relative effect Very low



of serum phosphate achieved)

10-week follow-up (mean)

2011

Sprague et al, 2009

(44.6%) (59.5%) RR = 1.14 (95% CI: 0.31 to 4.15)

Absolute effect

83 more per 1000 (410 fewer to 1000 more

Note: indirect comparison calculated by reviewer


1070


1072

Multiple treatment comparisons (serum phosphate and serum calcium) 1073

As there was a lack of evidence allowing a direct comparison between all of the possible treatments, an MTC was carried out to aid 1074

the GDG decision making process (for further information on MTCs, please see the glossary on page 235). A total of 3 studies were 1075

included within the network allowing 4 treatments to be assessed against each other. The study by Russo et al was excluded from 1076

the model as the follow up time (104 weeks) was significantly longer than the other 3 studies within the network (mean follow up 1077

time, 9 weeks). The model used was based upon preliminary work carried out by NICE's technical support unit (appendix G). In 1078

brief, a standard network meta-analysis model in WinBUGS was used, utilising a previously published code (Dias et al, 201118). 1079

18

Dias, S et al (2011) NICE DSU Technical support Document 2: a generalised linear modelling framework for pair wise and network meta-analysis NICE Decision Support

Unit available from http://www.nicedsu.org.uk

http://www.nicedsu.org.uk/



The results below are those generated from a fixed effect analysis, which was recommended in the initial work carried out by the 1080

Technical Support Unit. 1081

Serum phosphate network map CKD patients stage 4 or 5 1082

SH CA P LC

1 Study 1 Study1 Study

1083

Abbreviations: CA, calcium acetate; LC, lanthanum carbonate; P, placebo; SH, sevelamer hydrochloride. 1084

1085

Summary GRADE profile 23 Multiple treatment comparison- serum phosphate 1086

Outcome Number of Studies Limitations Inconsistency Indirectness Imprecision Quality

Serum phosphate (mean follow up 9 weeks)

3 RCTs1 very serious

2 serious

5 serious

3 serious

4 very low

1 Caglar et al, 2008; Qunibi et al, 2011; Sprague et al, 2009

2 Lack of detail on the randomisation, allocation and blinding within the studies

3 Surrogate marker used

4 GRADE rule of thumb sample size <400

5 Inconsistency could not be appraised


1087



Mean difference matrix: serum phosphate 1088

Placebo Lanthanum Sevelamer hydrochloride Calcium acetate

Placebo - -0.120 mmol/l (95%CI: -0.218 to -0.022)

-0.259 mmol/l (95%CI: -0.581 to 0.062)

-0.229 mmol/l (95%CI: -0.488 to 0.030)

Lanthanum - - -0.140 mmol/l (95%CI: -0.475 to 0.196)

-0.109 mmol/l (95%CI: -0.387 to 0.168)

Sevelamer hydrochloride - - - 0.030 mmol/l (95%CI: -0.161 to 0.222)

Calcium acetate - - - -

Abbreviations: CI, credibility interval.

1089

Probability rank: serum phosphate 1090

Intervention Probability best binder Median rank (95% CI)

Sevelamer hydrochloride 0.560 1 (1, 4)

Calcium acetate 0.296 2 (1, 4)

Lanthanum carbonate 0.144 3 (1, 3)

Placebo 0.000 4 (3, 4)


1091



Relative effectiveness compared to placebo: serum phosphate 1092

1093

1094

Multiple treatment comparison: serum calcium 1095

SH CA P LC

1 Study 1 Study1 Study

1096

Abbreviations: CA, calcium acetate; LC, lanthanum carbonate; P, placebo; SH, sevelamer hydrochloride. 1097 1098

Summary GRADE profile 24 Multiple treatment comparison - serum calcium 1099


Serum calcium (mean follow up 9 weeks)

3 RCTs1 very serious2 serious5 serious3 serious4 very low

1 Caglar et al, 2008; Qunibi et al, 2011; Sprague et al, 2009 2 Lack of detail on the randomisation, allocation and blinding within the studies 3 Surrogate marker used 4 GRADE rule of thumb sample size <400 5

Inconsistency could not be appraised


-1 -0.5 0 0.5 1

Lanthanum carbonate

Calcium acetate

MD (mmol/l) v. placebo



1100

Mean difference matrix: serum calcium 1101

Placebo Lanthanum Sevelamer hydrochloride Calcium acetate

Placebo - 0.050 mmol/l (95%CI: 0.006 to 0.094) 0.080 mmol/l (95%CI: -0.017 to 0.177) 0.170 mmol/l (95%CI: 0.090 to 0.250)

Lanthanum - - 0.030 mmol/l (95%CI: -0.076 to 0.136) 0.120 mmol/l (95%CI: 0.029 to 0.211)

Sevelamer hydrochloride - - - 0.090 mmol/l (95%CI: 0.035 to 0.145)

Calcium acetate - - - -

1102

Probability rank: serum calcium 1103

Intervention Probability best Median rank (95%CI)

Placebo 0.936 1 (1, 2)




1104



Relative effectiveness compared to placebo: serum calcium 1105

1106

1107

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

Lanthanum carbonate


Calcium acetate

MD (mmol/l) versus placebo





Use of phosphate binders in people with stage 4 or 5 CKD who are not 1110

on dialysis 1111

Calcium acetate compared with placebo 1112


3.4.3.1 Very-low-quality evidence from 1 RCT of 78 patients showed 1114

calcium acetate to be associated with a greater proportion of 1115

participants that had that achieved serum phosphate control than 1116

that associated with placebo at 12 weeks (RR 1.63 [95% CI 1.00 to 1117

2.63 i.e. statistically significant]). 1118

3.4.3.2 Low-quality evidence from the RCT of 78 patients showed no 1119

statistically significant difference in the degree of adherence 1120

between those that received calcium acetate and those that 1121

received placebo (MD 0.70% of prescription lower [95% CI –7.16 to 1122

5.76]). 1123


3.4.3.3 Very-low-quality evidence from the RCT of 78 patients showed no 1125

statistically significant difference in the proportion of participants 1126

with hypercalcaemia at 12 weeks between those that received 1127

calcium acetate and those that received placebo (RR 12.16 [95% 1128

CI 0.7 to 212.64]). 1129

Calcium carbonate compared with sevelamer hydrochloride 1130



statistically significant difference in mean serum phosphate 1133

between those that received calcium carbonate and those that 1134




received sevelamer hydrochloride at 2 years (MD 0.03 mmol/l 1135

higher [95% CI –0.24 to 0.18]). 1136



statistically significant difference in mean serum calcium between 1139

those that received calcium carbonate and those that received 1140

sevelamer at 2 years (MD 0.02 mmol/l higher [95% CI –0.06 to 1141

0.10]). 1142


statistically significant difference in mean coronary artery 1144

calcification scores between those that received calcium carbonate 1145

and those that received sevelamer at 2 years (MD 20 higher [95% 1146

CI –262.96 to 302.96]) 1147


calcium carbonate to be associated with a mean annualised 1149

change in coronary artery calcification scores 146.66 greater (95% 1150

CI 35.77 to 257.55 i.e. statistically significant) than that associated 1151

with sevelamer. 1152

Lanthanum compared with placebo 1153



statistically significant difference in the number of participants that 1156

achieved serum phosphate control during the 8-week study period 1157

between those that received lanthanum and those that received 1158

placebo (RR 1.69 higher [95% CI 0.90 to 3.17]). 1159



statistically significant difference in the number of participants who 1162

experienced nausea and vomiting during the 8-week study period 1163




placebo (RR 1.37 higher [95% CI 0.46 to 4.10]). 1165

Lanthanum compared with calcium acetate 1166


3.4.3.10 Very-low-quality evidence from 1 indirect comparison of 93 patients 1168

from 2 RCTs showed no statistically significant difference in the 1169

number of participants that achieved serum phosphate control 1170

during a mean 10-week study period between those that received 1171

lanthanum and those that received calcium acetate (RR 1.14 1172

[95% CI 0.31 to 4.15]). 1173

Multi treatment comparison: serum phosphate 1174

Critical outcome 1175

3.4.3.11 Evidence from 1 MTC, containing 3 studies of very low quality, 1176

suggested that sevelamer hydrochloride had a 56% probability 1177

(median rank 1 [95% CI 1, 4]) of being the best treatment to control 1178

serum phosphate at 9 weeks, in patients with stage 4 or 5 CKD, 1179

with calcium acetate having a 29.6% probability (median rank 2 1180

[95% CI 1, 4]). 1181

Multi treatment comparison: serum calcium 1182

Important outcome 1183

3.4.3.12 Evidence from 1 MTC, containing 3 studies of very low quality, 1184

suggested that placebo had a 93.6% probability (median rank 1 1185

[95%CI 1, 2]) of being the best treatment to control serum calcium 1186

at 9 weeks in patients with stage 4 or 5 CKD, with sevelamer 1187

hydrochloride acetate having a 5.2% probability (median rank 3 1188

[95% CI 1, 3]) and lanthanum carbonate having a 1.2% probability 1189

(median rank 2 [95% CI 2, 3]). 1190

1191





The GDG discussed the relative importance of the outcomes and agreed that all-cause mortality, cardiovascular-related mortality, serum phosphate and adherence were critical for decision making.

Adverse events (including nausea and vomiting, constipation, diarrhoea, abdominal distension and upper abdominal pain), serum calcium and cardiovascular calcification scores were also considered important for decision making, though not critical.

Serum PTH was raised by some as potentially important, though the GDG also noted that there is a lot of variability in PTH measurements, both in the sense of the actual levels of PTH in the serum and in the ability of laboratory techniques to detect these levels. Therefore, serum PTH was not deemed to be a sufficiently reliable basis from which to formulate recommendations.

Serum phosphate, serum calcium and cardiovascular calcification scores were considered surrogate measures.


Early intervention to prevent or manage high phosphate levels was considered key to preventing downstream complications resulting from the poor management of serum calcium and PTH. The GDG therefore emphasised the importance of starting phosphate binder therapy early, and stressed that this should be in the context of concurrent dietary management of serum phosphate.

The evidence showed that the phosphate binders examined were all effective in lowering serum phosphate in adults compared to placebo, with a 56% probability that sevelamer hydrochloride is better than calcium acetate or lanthanum. Although no significant difference was observed between lanthanum and calcium acetate in terms of achieving serum phosphate control, the Multiple Treatment Comparison (MTC) suggested calcium acetate to be more likely to have the best mean serum phosphate level (30% versus 14% probability of being the best). It was noted, however, that the median ranking of these binders had wide credibility intervals that reduced the GDG's confidence in the significance of these results.

Sevelamer hydrochloride, lanthanum and calcium acetate were all associated with a raise in serum calcium in comparison to placebo, with calcium acetate raising it the most.

There was no significant difference between sevelamer hydrochloride and calcium carbonate in their effectiveness at controlling serum phosphate, although sevelamer hydrochloride is more effective than calcium carbonate in limiting coronary artery calcification.

The GDG felt that the large body of evidence found for the use of phosphate binders in patients with CKD stage 5 but who are also on dialysis (see section 3.5) was a stronger foundation from which to make recommendations than the small, limited evidence base found during this review. Additionally, the GDG felt that continuity of care was important to both patient well-being and therapeutic success. Because of this, they felt it would be inappropriate to use the limited evidence identified in this review to recommend a different phosphate binder for pre-dialysis patients than that recommended for those on dialysis, which was supported by a more substantial evidence base. Calcium acetate's consistently good performance in the review of phosphate binder use in dialysis patients led the GDG to recommend



it as the first-line phosphate binder for that population. The GDG felt it was appropriate to extrapolate these results to patients with CKD stage 4 or 5, and this, combined with calcium acetate's relatively good performance in the serum phosphate MTC outlined above, led the GDG to recommend calcium acetate as the first-line choice of phosphate binder in both pre-dialysis and dialysis patients.

However, in patients that cannot tolerate calcium acetate (for example, those who find it difficult to swallow tablets or experience side effects), the GDG felt that calcium carbonate would be an effective alternative first-line phosphate binder.

The GDG noted that metabolites and toxins are less likely to accumulate in patients with CKD stages 4 and 5 than in patients who require dialysis since residual renal function allows a greater capacity for their elimination from the body. This means that the clinical concerns associated with calcium-based phosphate binders in terms of higher serum calcium levels were lower in this population, and longer-term use of calcium-based phosphate binders was felt to be less of an issue.

Despite this, the GDG noted that a non-calcium-based binder could still be of use in reducing the daily intake of elemental calcium in those who are hypercalcaemic, have low serum PTH or in whom calcium-based binders are not tolerated. The lack of evidence in this area led the GDG to conclude that the choice of non-calcium-based binder should be determined through clinical judgement and patient preference. Because of the residual renal function discussed above, the clinical concerns associated with aluminium hydroxide and magnesium carbonate19 in terms of the potential build-up and deposition of metabolites and toxins in the body are less in pre-dialysis patients. Therefore, the GDG felt comfortable allowing their inclusion as non-calcium-based options alongside sevelamer hydrochloride and lanthanum carbonate. However, some members of the GDG also felt that regular monitoring of serum aluminium may be beneficial in those receiving aluminium hydroxide.

Although no evidence was found concerning the effectiveness of phosphate binders in children, the GDG felt that a calcium-based binder would be desirable as the first-line phosphate binder used in children. This is because children require additional calcium for their growing bones, but also to avoid the effects of secondary hyperparathyroidism that can arise in young patients with chronically low serum calcium levels. However, in children with high serum calcium or at risk from hypercalcaemia, a combination of a calcium-based and a non-calcium-based binder should be used as the first-line binder regimen. In this way, serum phosphate can be controlled to the desired level without further raising the serum calcium, but also without allowing calcium to decrease to levels that lead to the adverse effects outlined above.

In some children taking a calcium-based binder, serum phosphate can still remain above the recommended level and serum calcium may reach the age-adjusted upper limit of normal. In these patients it was

19

Note: magnesium carbonate is only licensed as a combination with calcium acetate; therefore it

cannot currently be used as monotherapy in the UK.



felt that no further increases should be made to the dose of calcium-based binders. Instead, a non-calcium-based binder could be added to the regimen. As with first-line combination phosphate binder therapy, this second-line combination aims to raise phosphate control to the desired level without further raising the serum calcium.

Currently, the only non-calcium-based binder licensed for use in children is aluminium hydroxide. However, although searched for, no evidence meeting the defined inclusion criteria was found for the use of aluminium hydroxide in children with stage 4 or 5 CKD who are not on dialysis. There was also a lack of evidence found for the use of aluminium hydroxide in those who are on dialysis that might have been extrapolated to the population of interest. The only paediatric trial found examined the effectiveness of sevelamer against calcium carbonate over an 8-month period, and showed sevelamer to be as effective at lowering serum phosphate and associated with lower serum calcium level. Although this trial was conducted in children on dialysis (see section 3.5), the GDG felt it appropriate to extrapolate the evidence to those who are pre-dialysis, particularly since no evidence was found for non-calcium-based binders in this population. Furthermore, while the GDG had concerns over the toxicity of aluminium, it was also noted that the licence is not for longer-term indications. For these reasons, the GDG felt that recommending sevelamer hydrochloride over aluminium hydroxide was appropriate, despite its use being off-label in children. The GDG also felt that the total replacement of calcium-based binders with a non-calcium-based binder would be inappropriate in children because of their higher calcium requirements.

If considering the use of non-calcium-based binders because of high serum calcium levels in patients on calcium-based binders, the GDG felt it important to emphasise the necessity of reviewing other sources of calcium, such as vitamin D, calcium supplements or dietary calcium, before making any changes. They noted that in some cases it might be easier to make small changes to these sources of elemental calcium than changing (and ensuring adherence to) the phosphate binder regimen. For example, it may be the case that a patient is on a high dose of vitamin D, and a reduction in this may be the most appropriate course of action.


There was insufficient evidence to provide a worthwhile model for adults with CKD stage 4 or 5 who are not on dialysis, and for children.

Quality of evidence

No data were found for age groups other than adults, although for this population only 4 papers examining 4 comparisons were identified.

Following the application of GRADE, overall evidence quality was low or very low across the outcomes.

Three of the studies had similar follow-up times and could be combined in a meta-analysis. The fourth study, however, was not considered appropriate for inclusion because of its significantly longer length.

A range of co-treatments were used across the studies, including calcium supplements, vitamin D and low-protein diets. The GDG noted their potential influence as confounders.

Reporting in many of the studies was poor. For example, details of study designs were often unclear and results were not always cited in the text of the papers, requiring the reviewer to read the data off



available graphs. Unit-of-analysis errors were also common, with analyses not following the intent-to-treat principle.


It was noted that dietary management is particularly important in these patients, and should continue to be a key component of regimens designed to manage hyperphosphataemia in CKD stages 4 and 5.

1193




use of phosphate binders in people with stage 4 or 5 CKD 1195




For children and young people, offer a calcium-based phosphate binder as the

first-line phosphate binder to control serum phosphate in addition to dietary

management.


For children and young people, if a series of serum calcium measurements

shows a trend towards the age-adjusted upper limit of normal, consider a

calcium-based binder in combination with a non-calcium-based binder, having

taken into account other causes of rising calcium levels.


For children and young people who remain hyperphosphataemic despite

adherence with a calcium-based phosphate binder, and whose serum calcium

goes above the age-adjusted upper limit of normal, consider a combination of

a calcium-based phosphate binder and sevelamer hydrochloride20, having

taken into account other causes of raised calcium.


For adults, offer calcium acetate as the first-line phosphate binder to control

serum phosphate in addition to dietary management.


For adults, consider calcium carbonate if calcium acetate is not tolerated or

20

At the time of consultation (October 2012), sevelamer hydrochloride did not have a UK marketing









patients find it unpalatable.


For adults with stage 4 or 5 chronic kidney disease (CKD) who are not on

dialysis and who are taking a calcium-based binder:

consider switching to a non-calcium-based binder if calcium-based

phosphate binders are not tolerated

consider either combining with, or switching to, a non-calcium-based binder

if hypercalcaemia develops (having taken into account other causes of

raised calcium), or if serum parathyroid hormone levels are low.


If a combination of phosphate binders is used, titrate the dosage to achieve

control of serum phosphate while taking into account the effect of any

calcium-based binders used on serum calcium levels (also see

recommendations 1.1.6, 1.1.7 and 1.1.10 to 1.1.12).


Use clinical judgement and take into account patient preference when

choosing whether to use a non-calcium-based binder as monotherapy or in

combination with a calcium-based binder (also see recommendations 1.1.10

to 1.1.12).

1198

Research recommendations 1199

See appendix B for full details of research recommendations. 1200



Research recommendation B1

Which binders are most effective in controlling serum phosphate in adults with

stage 4 or 5 CKD who are not on dialysis?


In adults with stage 4 or 5 CKD, including those on dialysis, what is the long-

term effectiveness and safety of aluminium hydroxide in controlling serum

phosphate?


In adults and children with stage 4 or 5 CKD, including those on dialysis, what

is the long-term effectiveness and safety of magnesium carbonate in

controlling serum phosphate?


Which binders are most effective in controlling serum phosphate in children

with stage 4 or 5 CKD who are not on dialysis?


For adults with stage 4 or 5 CKD, including those on dialysis, what is the most

effective sequence or combination of phosphate binders to control serum

phosphate?

1201

3.5 Use of phosphate binders in people with stage 5 CKD 1202

who are on dialysis 1203


For people with stage 5 CKD who are on dialysis, are phosphate binders 1205

effective compared to placebo or other treatments in managing serum 1206

phosphate and its associated outcomes? Which is the most effective 1207

phosphate binder? 1208




This review question focused on the use of phosphate-binding medications to 1210

prevent and treat hyperphosphataemia in patients with stage 5 CKD who are 1211

on dialysis. These pharmacological interventions bind dietary phosphate when 1212

taken with food, that is, not on an empty stomach), and can broadly be 1213

defined as: 1214

calcium-based: calcium carbonate or calcium acetate; and 1215

non-calcium-based: sevelamer hydrochloride, sevelamer carbonate, 1216

lanthanum carbonate, aluminium hydroxide or magnesium carbonate. 1217

1218




broad search strategy, pulling in all papers relating to the use of phosphate 1222

binders in the management of hyperphosphataemia in CKD. Only RCTs that 1223

compared a phosphate binder with either a placebo or another comparator in 1224

patients with stage 5 CKD who are on dialysis were considered for inclusion. 1225



the population included people with a diagnosis of CKD stage 5 who are 1228

not on dialysis or 1229

data were obtained through an open-label extension of an RCT. 1230

1231

From a database of 1480 abstracts, 300 full-text articles were ordered and 1232

46 papers were included in the review (Al-Baaj et al, 2005; Asmus et al, 2005; 1233

Barbarykin et al 2004; Barreto et al, 2008; Block et al, 2005; Braun et al, 2004; 1234

Chertow et al, 1997; Chertow et al, 2002; Chertow et al, 2003; Chiang et al, 1235

2005; Chow et al, 2007; de Francisco et al, 2010; de Santo et al, 2006; 1236

Emmett et al, 1991; Evenepoel et al, 2009; Ferreira et al, 2008; Finn et al, 1237

2004; Finn & the Lanthanum Study Group, 2006; Fishbane et al, 2010; 1238

Freemont et al, 2005; Galassi et al, 2006; Hervas et al, 2003; Hutchison et al, 1239



2005; Iwasaki et al, 2005; Janssen et al, 1995; Janssen et al, 1996; Joy et al, 1240

2003; Kakuta et al, 2011; Katopodis et al, 2006; Koiwa et al, 2005a; Koiwa et 1241

al, 2005b; Lin et al, 2011; Liu et al, 2006; Malluche et al, 2008; Navarro-1242

Gonzalez et al, 2011; Qunibi et al, 2008; Raggi et al, 2004; Ring et al, 1993; 1243

Salusky et al, 2005; Shigematsu & the Lanthanum Carbonate Research 1244

Group, 2008a; Shigematsu & the Lanthanum Carbonate Research Group, 1245

2008b; Spasovski et al, 2006; Spiegel et al, 2007; Suki et al, 2007; Tzanakis 1246

et al, 2008; Wilson et al, 2009). Tables 5 and 6 list the details of the included 1247

studies. 1248

The reviewer was unable to extract the serum phosphate data from 1 study 1249

(Finn & the Lanthanum Study Group, 2006) comparing lanthanum carbonate 1250

against ‘standard therapy’ (the participants’ pre-study binder regimens) due to 1251

its presentation in uninterpretable graphs. It was felt that this study was a 1252

significant part of the evidence base due to its long follow-up of 2 years and 1253

large population of 1359 participants. The reviewer contacted the authors of 1254

the study but was unable to obtain the data required to include this study in 1255

the analyses of serum phosphate data. 1256

Only 1 of the papers located explicitly considered the use of phosphate 1257

binders in the management hyperphosphataemia in children who are on 1258

dialysis. The evidence for this population is presented separately from that for 1259

adults because of the different considerations required in determining a child’s 1260

treatment regimen. Foremost amongst these is the need to support a child’s 1261

development and growth, and the different nutrient and mineral requirements 1262

this entails. 1263

There was some pooling of the 46 studies conducted in adults, although this 1264

suffered from considerable heterogeneity. One prominent cause of this was 1265

the widespread use of a range of concurrent interventions that are known to 1266

impact the outcomes of interest. Additionally, time periods for which data were 1267

available varied widely between the studies, and thresholds for the definition 1268

of dichotomous outcomes, such as the incidence of hypercalcaemia or the 1269

achievement of serum phosphate control, were not always consistent. 1270



Mean differences (MDs) were calculated for continuous outcomes, hazard 1271

ratios (HRs) were calculated for all-cause and cardiovascular-related 1272

mortality, and relative risks (RRs) were calculated for all other binary 1273

outcomes. Corresponding 95% confidence intervals are given where sufficient 1274

data were available. 1275

Meta-analyses were conducted either through the pooling of studies into 1276

single, direct pair-wise comparisons, or through the pooling of 3 or more 1277

interventions into a network meta-analysis. These interventions could then be 1278

compared, both directly and indirectly, against each other in multiple treatment 1279

comparisons (MTCs). Where possible, these MTCs enabled the GDG to 1280

quantitatively rank the phosphate binders according to their impact on the 1281

outcomes of interest. 1282



Table 5 Summary of included studies for the use of phosphate binders in children with stage 5 CKD who are on dialysis 1283

Study (Country)

Interventions Follow-up Target PO4/Ca

(mmol/l)

Baseline PO4/Ca (mmol/l ± SD)


Salusky et al, 2005

RCT

Calcium carbonate

versus


8 months PO4

1.24 to 1.94

Ca

2.1 to 2.54

Calcium carbonate

PO4: 1.91±0.5

Ca: 2.25±0.15

Sevelamer

PO4: 1.81±0.38

Ca: 2.25±0.15

Serum phosphate

Serum calcium

Both treatments controlled serum phosphate and calcium.

However, sevelamer controlled serum calcium closer to the lower end of the normal range.

Abbreviations: Ca, calcium; PO4, phosphate; RCT, randomised controlled trial.

1284

Summary GRADE profile 25 Calcium carbonate compared with sevelamer in children 1285



Calcium carbonate Sevelamer hydrochloride

Serum phosphate

8-month follow-up

1 RCT

Salusky et al, 2005



Very low

Serum calcium

8-month follow-up

1 RCT

Salusky et al, 2005


MD = 0.20 mmol/l higher (95% CI: 0.1 to 0.32 higher)

Very low


1286

See appendix E for the evidence table and GRADE profile in full. 1287

1288



Table 6 Summary of included studies for the use of phosphate binders in adults with stage 5 CKD who are on dialysis 1289

Study (Country) Interventions Follow-up Target PO4 / Ca

(mmol/l)

Baseline PO4 / Ca (mmol/l ± SD)


Al-Baaj et al, 2005

RCT

(UK)

Lanthanum carbonate

versus

Placebo

4 weeks PO4

1.3 to 1.8

Lanthanum

PO4: 1.54 ± 0.29

Placebo

PO4: 1.68 ± 0.27

Achieved phosphate control

Serum phosphate

Adherence

Lanthanum is effective in controlling serum phosphate.

Asmus et al, 2005

RCT

(Germany)


versus

Calcium carbonate

2 years PO4

1 to 1.6

Ca

< 2.6

Sevelamer

PO4: 2.4 ± 0.6

Ca: 2.4 ± 0.1

Calcium carbonate PO4: 2.3 ± 0.2

Ca: 2.2 ± 0.5

Achieved phosphate control

Serum phosphate

Serum calcium


Calcium carbonate was associated with increases in coronary artery calcification and the number of participants suffering hypercalcaemic events.

Both treatments lowered serum phosphate.

Babarykin et al, 2004

RCT

(Latvia)

Calcium carbonate bread

versus

Calcium acetate

8 weeks PO4

1.5 to 1.7

Calcium carbonate bread

PO4: 2.57 ± 0.47

Ca: 2 ± 0.25

Calcium acetate

PO4: 2.1 ± 0.18

Ca: 2.15 ± 0.2

Serum calcium

Serum phosphate

Calcium carbonate bread was able to control serum phosphate and serum calcium as effectively as calcium acetate.

Barreto et al, 2008

RCT

(Brazil)

Calcium acetate

versus


12 months PO4

1.13 to 1.78

Calcium acetate

PO4: 2.3 ± 0.45

Sevelamer

PO4: 2.3 ± 0.7

Serum phosphate


No differences were observed in coronary artery calcification between the sevelamer and calcium acetate groups.

There was no difference between the treatments in terms of controlling serum phosphate.

Block et al, 2005

RCT

(USA)


versus

Calcium-based binders

18 months PO4

< 2.10

Ca

Sevelamer

PO4: 1.68 ± 0.52

Ca: 2.32 ± 0.25

Serum calcium

Serum phosphate

Risk of hypercalcaemia

No difference was observed in those with little evidence of calcification at baseline. However, in those with mild calcification



< 2.54 Calcium-based binders

PO4: 1.74 ± 0.45

Ca: 2.32 ± 0.2

Coronary artery calcification there was evidence of a more rapid progression in those on calcium-based binders.

There were no differences in serum phosphate, but there was a higher risk of hypercalcaemia in those on calcium-based binders.

Braun et al, 2004

RCT

(Germany)


versus

Calcium carbonate

12 months PO4

1 to 1.6

Ca

< 2.6

Sevelamer

PO4: 1.65 ± 0.4

Ca: 2.34 ± 0.15

Calcium carbonate

PO4: 1.65 ± 1.36

Ca: 2.32 ± 0.14

Serum calcium

Serum phosphate


Adherence


Reductions in serum phosphate were similar in both groups.

Calcium carbonate resulted in more instances of hypercalcaemia and was associated with increases in coronary artery calcification.

Chertow et al, 1997

RCT

(USA)


versus

Placebo

2 weeks None given

Sevelamer

PO4: 2.13 ± 0.68

Ca: 2.32 ± 0.22

Placebo

PO4: 2.32 ± 0.77

Ca: 2.4 ± 0.12

Serum calcium

Serum phosphate

Adherence

Adverse events

Sevelamer significantly reduced serum phosphate and was not significantly associated with changes in serum calcium.

There was no difference in the incidence of adverse events between sevelamer and placebo.

Chertow et al, 2002

RCT

(USA)


versus

Calcium based binders

12 months PO4

0.97 to 1.61

Ca

2.12 to 2.62

Sevelamer

PO4: 2.45 ± 0.58

Ca: 2.35 ± 0.17


PO4: 2.39 ± 0.61

Ca: 2.32 ± 0.17

Serum calcium

Serum phosphate

Adherence


Sevelamer was equivalent to calcium in controlling serum phosphate.

However, serum calcium was higher in the calcium group and hypercalcaemia was more common.

Chertow et al, 2003

RCT

(USA)


versus

Calcium acetate

12 months PO4

0.97 to 1.61

Ca

2.12 to 2.62

Sevelamer

PO4: 2.45 ± 0.61

Ca: 2.34 ± 0.17

Calcium acetate

PO4: 2.48 ± 0.17

Ca: 2.34 ± 0.17

Serum calcium

Serum phosphate

Adverse events

Adherence



Calcium acetate was associated with increased hypercalcaemia and calcification.

The reduction in serum phosphate was equivalent in both treatments.



Chiang et al, 2005

RCT

(Taiwan)

Lanthanum

versus

Placebo

4 weeks PO4

0.6 to 1.8

Lanthanum

PO4: 1.77 ± 0.11

Placebo

PO4: 1.83 ± 0.16

Serum phosphate

Proportion achieving phosphate control

Lanthanum was effective in controlling serum phosphate.

Chow et al, 2007

RCT

(China)

Sevelamer hydrochloride high-dose

versus

Sevelamer hydrochloride low-dose

6 months PO4

<1.78

Sevelamer high-dose

PO4: 2.38 ± 0.38

Sevelamer low-dose

PO4: 2.25 ± 0.31

Serum phosphate


Adherence

Serum phosphate was significantly lower in the high dose group. However, the proportion of patients achieving phosphate control was not significantly different.

de Francisco et al, 2010

RCT

(Germany, Poland, Portugal, Romania and Spain)

Calcium acetate + magnesium carbonate

versus


6 months PO4

> 1.78

Ca

2.1 to 2.37


PO4: 2.46 ± 0.40

Ca: 2.14 ± 0.23

Sevelamer

PO4: 2.48 ± 0.47

Ca: 2.19 ± 0.18

Serum calcium

Serum phosphate

There was no difference in the control of serum phosphate.

There were small increases in serum calcium in the calcium acetate + magnesium carbonate group.

de Santo et al, 2006

RCT

(Italy)


versus

Calcium carbonate

6 months PO4

<1.78

Ca

2.12 to 2.62

Sevelamer

PO4: 2.38 ± 0.35

Ca: 2.28 ± 0.19

Calcium carbonate

PO4: 2.42 ± 0.34

Ca: 2.28 ± 0.24

Serum calcium

Serum phosphate

Both treatments were effective in reducing serum phosphate.

There was a slight rise in serum calcium in those receiving calcium carbonate.

Emmett et al, 1991

RCT

(USA)

Calcium acetate

versus

Placebo

2 weeks PO4

1.45 to 1.78

Calcium acetate

PO4: 2.42 ± 0.54

Ca: 2.2 ± 0.18

Placebo

PO4: 2.29 ± 0.45

Ca: 2.25 ± 0.22

Serum calcium

Serum phosphate

Calcium acetate was associated with a decrease in serum phosphate and an increase in serum calcium.

Evenepoel et al, Sevelamer hydrochloride 12 weeks PO4 Sevelamer Serum calcium Both treatments significantly



2009

RCT

(Belgium, Denmark, France, Italy, Spain, The Netherlands and UK)

versus

Calcium acetate

0.97 to 1.78

Ca

2.1 to 2.59

PO4: 2.42 ± 0.45

Ca: 2.38 ± 0.15

Calcium acetate

PO4: 2.4 ± 0.45

Ca: 2.39 ± 0.12

Serum phosphate


reduced serum phosphate.

Hypercalcaemia was experienced by more calcium acetate patients.

Ferreira et al, 2008

RCT

(Portugal)


versus

Calcium-based binder

12 months PO4

1 to 1.6

Ca

< 2.6

Sevelamer

PO4: 1.91 ± 0.6

Ca: 2.33 ± 0.27


PO4: 1.74 ± 0.48

Ca: 2.38 ± 0.27

Serum calcium

Serum phosphate

Both serum phosphate and serum calcium were well controlled in both groups.

Finn et al, 2004

RCT

(USA)

Lanthanum carbonate at various doses

versus

Placebo

6 weeks PO4

< 1.78

No details Serum phosphate


Lanthanum carbonate was effective in lower serum phosphate. Significant reductions were observed in the 1350 and 2250 mg/day groups compared to placebo.

Finn et al, 2006

RCT

(USA, Puerto Rico, Poland and South Africa)

Lanthanum carbonate

versus

Standard therapy

2 years PO4

< 1.9

Lanthanum

Ca: 2.41 ± 0.26

Standard therapy

Ca: 2.37 ± 0.40


Serum calcium

Adverse events

Lanthanum’s efficacy in controlling serum phosphate is similar to that of other therapies.

However, higher serum calcium levels were observed in those using standard therapies.

Fishbane et al, 2010

RCT

(USA)

Sevelamer carbonate powder

versus

Sevelamer hydrochloride tablets

6 months PO4

1.13 to 1.78

Sevelamer carbonate powder

PO4: 2.36 ± 0.41

Ca: 2.25 ± 0.17

Sevelamer tablets

PO4: 2.44 ± 0.41

Ca: 2.25 ± 0.17

Serum phosphate

Serum calcium

Adverse events

Adherence

Once a day sevelamer carbonate powder was not as effective in reducing serum phosphate as thrice daily sevelamer tablets. However, the powder did reduce levels significantly, reaching the KDOQI phosphate targets in most patients.

There were a greater number of



upper GI-related events in those taking the powder.

Freemont et al, 2005

RCT

(UK)

Lanthanum carbonate

versus

Calcium carbonate

12 months None given

Lanthanum

PO4: 1.72 ± 0.4

Ca: 2.24 ± 0.2

Calcium carbonate

PO4: 1.87 ± 0.52

Ca: 2.29 ± 0.32

Serum calcium

Serum phosphate


Adverse events

Both treatments were equally effective in controlling serum phosphate and serum calcium, although there were more instances of hypercalcaemia in the calcium carbonate group.

Galassi et al, 2006

RCT

(USA)


versus


2 years None given

Sevelamer

PO4: 1.60 ± 0.42

Ca: 2.30 ± 0.68


PO4: 2.30 ± 0.15

Ca: 1.7 ± 0.45

Serum phosphate

Serum calcium


Serum phosphate was similar in both groups. However, serum calcium and coronary artery calcification increased in those receiving calcium-based binders.

Hervas et al, 2003

RCT

(Spain)


versus

Calcium acetate

32 weeks None given

Sevelamer

PO4: 2.61 ± 0.52

Ca: 2.45

Calcium acetate

PO4: 2.42 ± 0.48

Ca: 2.45

Serum calcium

Serum phosphate

Sevelamer is effective in lowering serum phosphate.

Hutchison et al, 2005

RCT

(UK, Germany, Belgium, The Netherlands)

Lanthanum carbonate

versus

Calcium carbonate

20 weeks PO4

< 1.8

Lanthanum

PO4: 2.67 ± 0.63

Calcium carbonate

PO4: 2.67 ± 0.63

Serum phosphate



Adverse events

Serum phosphate control was similar in both groups. However, there was a greater incidence of hypercalcaemia in the calcium carbonate group.

Iwasaki et al, 2005

RCT

(Japan)

Sevelamer hydrochloride + calcium carbonate (low-dose)

versus


8 weeks None given

Sevelamer + calcium carbonate (low-dose)

PO4: 2.23 ± 0.26

Ca: 2.23 ± 0.26

Sevelamer + calcium

Serum calcium

Serum phosphate

Adverse events

Serum phosphate significantly decreased in the high dose group only.



+ calcium carbonate (high-dose)

carbonate (high-dose)

PO4: 2.42 ± 0.26

Ca: 2.47 ± 0.15

Janssen et al, 1995

RCT

(Netherlands)

Calcium acetate

versus

Calcium carbonate

12 months PO4

< 1.6

Ca

> 2.2

Calcium acetate

PO4: 2.95 ± 0.86

Ca: 2.30 ± 0.2

Calcium carbonate

PO4: 2.45 ± 0.54

Ca: 2.3 ± 0.18

Serum calcium

Serum phosphate

Serum phosphate did not differ between the groups

Janssen et al, 1996

RCT

(Netherlands)

Calcium acetate

versus

Calcium carbonate

12 months PO4

< 1.6

Ca

2.2 to 3.0

Calcium acetate

PO4: 1.65 ± 0.4

Ca: 2.27 ± 0.17

Placebo

PO4: 1.65 ± 1.36

Ca: 2.2 ± 0.2

Serum phosphate

Serum calcium



There was no difference between the treatments in terms of phosphate-binding capacity.

Calcium carbonate appeared to be associated with more hypercalcaemic episodes than calcium acetate.

Joy et al, 2003

RCT

(USA)

Lanthanum carbonate

versus

Placebo

4 weeks PO4

< 1.91

Lanthanum

PO4: 1.76 ± 0.47

Ca: 2.20 ± 0.16

Placebo

PO4: 1.82 ± 0.53

Ca: 2.17 ± 0.18

Serum phosphate

Serum calcium


Adverse events

Lanthanum was effective in controlling serum phosphate.

The number of drug-related adverse events was similar between the 2 groups.

Kakuta et al, 2011

RCT

(Japan)


versus

Calcium carbonate

12 months PO4

< 2.1

Ca

< 2.54

Sevelamer

PO4: 1.82 ± 0.18

Ca: 2.44 ± 0.2

Calcium carbonate

PO4: 1.86 ± 0.25

Ca: 2.42 ± 0.16

Serum calcium

Serum phosphate

Adverse events


Sevelamer appeared to slow the increase in coronary artery calcification.

The control of biochemical parameters was similar in both groups.

Katopodis et al, 2006

RCT


versus

Aluminium hydroxide

8 weeks None given

Sevelamer

PO4: 2.38 ± 0.43

Ca: 2.3 ± 0.2

Serum phosphate

Serum calcium

Adverse events

Similar reductions in serum phosphate were observed over the course of the study.



(Greece) Aluminium hydroxide PO4: 2.28 ± 0.39

Ca: 2.27 ± 0.2

There were no differences in the control of calcium.

Koiwa et al, 2005a

RCT

(Japan)


versus

Sevelamer +calcium carbonate

versus

Calcium carbonate

4 weeks PO4

< 1.78

Ca

2.1 to 2.37

Sevelamer

PO4: 2.09 ± 0.28

Ca: 2.15 ± 0.22

Sevelamer + calcium carbonate

PO4: 1.92 ± 0.54

Ca: 2.28 ± 0.17

Calcium carbonate

PO4: 2.2 ± 0.39

Ca: 2.28 ± 0.17


Serum phosphate

Serum calcium

Adverse events


There was an additive effect of sevelamer in the treatment of hyperphosphataemia with calcium carbonate.

Koiwa et al, 2005b

RCT

(Japan)

Sevelamer hydrochloride + calcium carbonate

versus

Calcium carbonate

4 weeks None given

Sevelamer + calcium carbonate

PO4: 1.91 ± 0.39

Calcium carbonate

PO4: 2.0 ± 0.29

Serum phosphate Serum phosphate levels were lower in those treated with sevelamer + calcium carbonate.

Lin et al, 2011

RCT

(Taiwan)


versus

Calcium acetate

8 weeks PO4

1.13 to 1.78

Ca

< 2.74

Not applicable Adverse events

Adherence


More hypercalcaemic events were recorded in the calcium acetate group.

Liu et al, 2006

RCT

(Taiwan)


versus

Calcium acetate

8 weeks PO4

1.13 to 1.94

Sevelamer

PO4: 2.62 ± 0.79

Ca: 2.26 ± 0.30

Calcium acetate

PO4: 2.62 ± 0.74

Ca: 2.25 ± 0.37

Serum phosphate

Serum calcium



Serum phosphate did not differ between the 2 treatments.

The number of patients experiencing hypercalcaemic events was significantly higher in the calcium acetate group.

Malluche et al, 2008

RCT

Lanthanum carbonate

versus

Standard therapy

24 months PO4

< 1.91

Lanthanum

PO4: 2.45 ± 0.48

Ca: 2.2 ± 0.24

Serum phosphate

Serum calcium

There was similar phosphate control with both treatments.

However, standard treatment was associated with higher serum



(USA, Puerto Rico, Poland, South Africa)

Standard therapy

PO4: 2.62 ± 0.65

Ca: 2.3 ± 0.28

calcium at most visits.

Navarro-Gonzalez et al, 2011

RCT

(Spain)


versus

Calcium acetate

12 weeks None given

Sevelamer

PO4: 1.74 ± 0.32

Ca: 2.25 ± 0.17

Calcium acetate

PO4: 1.65 ± 0.19

Ca: 2.25 ± 0.12

Serum phosphate

Serum calcium

Both serum phosphate and serum calcium decreased significantly in both groups.

Qunibi et al, 2008

RCT

(USA)

Calcium acetate

versus


12 months PO4

1.13 to 1.78

Calcium acetate

PO4: 2.1 ± 0.61

Ca: 2.2 ± 0.2

Sevelamer

PO4: 2.13 ± 0.48

Ca: 2.2 ± 0.17

Serum phosphate

Serum calcium

Adverse events



Both treatments were effective in controlling serum phosphate and serum calcium.

There did not appear to be any differences it coronary artery calcification progression.

Raggi et al, 2004

RCT

(USA, Germany and Austria)


versus


12 months PO4

0.97 to 1.61

Ca

2.12 to 2.62

Calcium acetate

PO4: 1.65 ± 0.4

Ca: 2.27 ± 0.17

Placebo

PO4: 1.65 ± 1.36

Ca: 2.2 ± 0.2

Coronary artery calcification Patients prescribed sevelamer had lower calcification scores than those on calcium-based binders at 52 weeks.

Ring et al, 1993

RCT

(Denmark)

Calcium acetate

versus

Calcium carbonate

3 weeks None given

Calcium acetate

PO4: 2.09 ± 0.25

Ca: 2.48 ± 0.16

Calcium carbonate

PO4: 2.30 ± 0.59

Ca: 2.40 ± 0.31

Serum calcium

Serum phosphate

Both treatments controlled serum phosphate.

There was no difference between the treatments in terms of serum calcium.

Shigematsu et al, 2008a

RCT

(Japan)

Lanthanum carbonate

versus

Calcium carbonate

8 weeks PO4

1.13 to 1.78

Ca

Lanthanum

PO4: 2.7 ± 0.45

Calcium carbonate

Serum phosphate

Adverse event


Both treatments are effective in controlling serum phosphate, but lanthanum is associated with a lower incidence of



< 2.59 PO4: 2.71 ± 0.46

hypercalcaemia.

Shigematsu et al, 2008b

RCT

(Japan)

Lanthanum carbonate 750 to 3000mg/day

versus

Placebo

6 weeks PO4

1.13 to 1.78

See evidence table Proportion achieving phosphate control

Serum phosphate

Adverse events

Serum phosphate was significantly reduced in a dose-dependent manner.

The incidence of drug-related adverse events was also dose-dependent.

Spasovski et al, 2006

RCT

(Macedonia)

Lanthanum carbonate

versus

Calcium carbonate

12 months PO4

< 1.8

Ca

< 2.6

Lanthanum

PO4: 1.58 ± 0.25

Ca: 2.13 ± 0.2

Calcium carbonate

PO4: 1.76 ± 0.39

Ca: 2.27 ± 0.23

Serum calcium

Serum phosphate


There were no significant differences in serum levels of calcium or phosphate.

Spiegel et al, 2007

RCT

(USA)

Magnesium carbonate

versus

Calcium acetate

12 weeks PO4

< 1.78

Magnesium carbonate

PO4: 2.1 ± 0.27

Ca: 2.06 ± 0.13

Calcium acetate

PO4: 2.13 ± 0.19

Ca: 2.1 ± 0.19


Serum phosphate

Serum calcium

Both treatments equally controlled serum phosphate.

Serum calcium was significantly higher in those taking calcium acetate.

Suki et al, 2007

RCT

(USA)


versus


3.75 years None given

None given Adverse events

Mortality

All-cause and specific mortality rates were not significantly different. Only in those over the age of 65 was there a significant effect of sevelamer in lowering the mortality rate.

Tzanakis et al, 2008

RCT

(Greece)

Magnesium carbonate

versus

Calcium carbonate

6 months PO4

< 1.78

Ca

< 2.62

Magnesium carbonate

PO4: 2.14 ± 0.28

Ca: 2.35 ± 0.13

Calcium carbonate

PO4: 2.12 ± 0.28


Serum phosphate

Serum calcium


Both treatments controlled serum phosphate.

There were fewer instances of hypercalcaemia in those receiving magnesium carbonate.



Ca: 2.28 ± 0.11

Wilson et al, 2009

RCT

(USA, Puerto Rico, Poland and South Africa)

Lanthanum carbonate

versus

Standard treatment

2.6 years PO4

< 1.9

Not applicable Mortality Overall mortality was similar in both groups, but the results suggested a survival benefit for those aged over 65 years.

Abbreviations: Ca, calcium; GI, gastrointestinal; PO4, phosphate; RCT, randomised controlled trial; SD, standard deviation.

1290

Summary GRADE profile 26 Calcium acetate compared with placebo 1291



Calcium acetate Placebo

Serum phosphate

2-week follow-up

1 RCT

Emmett et al, 1991



Very low

Serum calcium

2-week follow-up

1 RCT

Emmett et al, 1991



Very low


1292

Summary GRADE profile 27 Calcium acetate compared with calcium carbonate 1293



Calcium acetate Calcium carbonate

Serum phosphate

3-week follow-up

1 RCT

Ring et al,1993

7 8 Absolute effect


Very low



Serum calcium

3 weeks follow-up

1 RCT

Ring et al,1993

7 8 Absolute effect


Very low

Serum calcium


12 months follow-up

1 RCT

Janssen et al, 1996

9/14 (64.3%)

12/13 (92.3%)

Relative effect

RR = 0.70 (95% CI: 0.46 to 1.06)

Absolute effect


Very low


1294

Summary GRADE profile 28 Calcium carbonate bread compared with calcium acetate 1295



Calcium carbonate bread Calcium acetate

Serum phosphate

8-week follow-up

1 RCT




Very low

Serum calcium

8-week follow-up

1 RCT




Very low


1296

Summary GRADE profile 29 Calcium carbonate compared with sevelamer + calcium carbonate 1297



Calcium carbonate Sevelamer + calcium



carbonate

Serum phosphate

4-week follow-up

1 RCT

Koiwa et al,

2005



Very low

Serum calcium

4-week follow-up

1 RCT

Koiwa et al,

2005



Very low


1298

Summary GRADE profile 30 Sevelamer hydrochloride + calcium carbonate (high-dose) compared with sevelamer 1299

hydrochloride + calcium carbonate (low-dose) 1300



Sevelamer + calcium carbonate (high-dose)

Sevelamer + calcium carbonate (low-dose)

Serum phosphate

8-week follow-up

1 RCT

Iwasaki et al, 2005



Very low

Adverse events - abdominal distension


8-week follow-up

1 RCT

Iwasaki et al, 2005

8/21 (38.1%)

9/30 (30%)

Relative effect

RR = 1.27 (95% CI: 0.59 to 2.75)

Absolute effect

81 more per 1000 (from 123 fewer to 525 more)

Very low

Adverse events - constipation


8-week follow-up

1 RCT

Iwasaki et al, 2005

10/21 (47.6%)

13/30 (43.3%)

Relative effect

RR = 1.10 (95% CI: 0.6 to 2.02)

Absolute effect


Very low

Serum calcium 1 RCT 21 30 Absolute effect Very low



8-week follow-up Iwasaki et al, 2005



1301

Summary GRADE profile 31 Sevelamer hydrochloride compared with placebo 1302



Sevelamer Placebo

Serum phosphate

2-week follow-up

1 RCT

Chertow et al, 1997



Very low

Adherence

(% of prescribed pills taken)

2-week follow-up

1 RCT

Chertow et al, 1997


MD = 4% higher (95% CI: 6.75 lower to 14.75 higher)

Very low

Adverse Events – diarrhoea


2-week follow-up

1 RCT

Chertow et al, 1997

0/24 (0%)

1/12 (8.3%)

Relative effect

RR = 0.17 (95% CI: 0.01 to 3.96)

Absolute effect


Very low

Adverse Events - nausea and vomiting


2-week follow-up

1 RCT

Chertow et al, 1997

1/24 (4.2%)

1/12 (8.3%)

Relative effect

RR = 0.5 (95% CI: 0.03 to 7.32)

Absolute effect


Very low

Adverse events - upper abdominal pain


2-week follow-up

1 RCT

Chertow et al, 1997

0/24 (0%)

1/12 (8.3%)

Relative effect

RR = 0.17 (95% CI: 0.01 to 3.96)

Absolute effect


Very low

Serum calcium

2-week follow-up

1 RCT

Chertow et al,


MD = 0.03 mmol/l lower (95% CI: 013

Very low



1997 lower to 0.07 higher)


1303

Summary GRADE profile 32 Sevelamer hydrochloride compared with calcium-based binders 1304



Sevelamer Calcium-based binders

Mortality

44-week follow-up

1 RCT

Suki et al, 2007

267/1053 (25.4%)

275/1050 (26.2%)

Relative effect

HR = 0.93 (95% CI: 0.79 to 1.1)

Absolute effect


Low

Mortality - in those > 65 years

44-week follow-up

1 RCT

Suki et al, 2007

578 598 Relative effect

HR = 0.77 (95% CI: 0.61 to 0.96)

Very low

Mortality - in those < 65 years

44-week follow-up

1 RCT

Suki et al, 2007


HR = 1.18 (95% CI: 0.91 to 1.53)

Very low

Cardiovascular mortality

44-week follow-up

1 RCT

Suki et al, 2007

142/1053 (13.5%)

147/1050 (14%)

Relative effect

HR = 0.93 (95% CI: 0.74 to 1.17)

Absolute effect


Very low

Cardiovascular mortality - in those > 65 years

44-week follow-up

1 RCT

Suki et al, 2007


HR = 0.78 (95% CI: 0.58 to 1.05)

Very low

Cardiovascular mortality - in those < 65 years

44-week follow-up

1 RCT

Suki et al, 2007


HR = 1.19 (95% CI: 0.82 to 1.73)

Very low

Adverse events - nausea and 1 RCT 1/1053 1/1050 Relative effect Very low



vomiting


3.75-year follow-up

Suki et al, 2007 (0.09%) (0.1%) RR = 1.00 (95% CI: 0.06 to 15.92)

Absolute effect


Serum calcium


8-month follow-up

9 RCTs

Block et al, 2005

Braun et al, 2004

Chertow et al, 2002

Chertow et al, 2003

Evenepoel et al, 2009

Koiwa et al,

2005

Lin et al, 2011

Liu et al, 2006

Qunibi et al, 2008

73/538 (13.6%)

161/497 (32.4%)

Relative effect

RR = 0.40 (95% CI: 0.31 to 0.52)

Absolute effect

194 fewer per 1000 (from 155 fewer to 224 fewer)

Very low

Coronary artery calcification scores

12- to 21-month follow-up

7 RCTs

Asmus et al, 2005

Barreto et al, 2008

Block et al, 2005

Braun et al, 2004

Kakuta et al, 2011

Qunibi et al, 2008


MD = 124.13 lower (95% CI: 205.88 to 42.38 lower)

Low



Raggi et al, 2004

Coronary artery calcification scores – in those with baseline CAC > 30

12- to18-month follow-up

3 RCTs

Bareto et al, 2008

Block et al, 2005

Braun et al, 2004


MD = 210.52 lower (382.36 to 38.69 lower)

Very low

Coronary artery calcification scores – in those with diabetes

2-year follow-up

1 RCT

Galassi et al, 2006


MD = 289.00 lower (651.99 lower to 73.99 higher)

Very low

Coronary artery calcification scores – in those without diabetes

2 years follow-up

1 RCT

Galassi et al, 2006


MD = 100.00 lower (256.33 lower to 56.23 higher)

Very low

Abbreviations: CAC, coronary artery calcification score; CI, confidence interval; MD, mean difference; RCT, randomised controlled trial; RR, relative risk; HR, hazard ratio.

1305

Summary GRADE profile 33 Sevelamer hydrochloride compared with calcium carbonate 1306



Sevelamer Calcium carbonate

Serum phosphate

4-week follow-up

1 RCT

Koiwa et al, 2005



Very low

Adherence

(number of patients considered adherent)

12-month follow-up

1 RCT

Braun et al, 2004

39/46 (84.8%)

30/36 (83.3%)

Relative effect

RR = 1.02 (95% CI: 0.84 to 1.23)

Absolute effect

17 more per 1000 (from 133 fewer to 192 more

Very low





4-week follow-up

1 RCT

Koiwa et al, 2005

5/29 (17.2%)

2/27 (7.4%)

Relative effect

RR = 2.33 (95% CI: 0.49 to 11.01)

Absolute effect


Very low

Adverse events – constipation


4- to 52-week follow-up

2 RCTs

Kakuta et al, 2011

Koiwa et al, 2005

16/120 (13.3%)

4/119 (3.4%)

Relative effect

RR = 3.40 (95% CI: 1.34 to 8.63)

Absolute effect

81 more per 1000 (from 11 more to 256 more)

Very low

Serum calcium

4-week follow-up

1 RCT

Koiwa et al, 2005



Very low


12-month follow-up

3 RCTs

Asmus et al, 2005

Braun et al, 2004

Kakuta et al, 2011


MD = 250.11 lower (95% CI: 480.63 to 19.59 lower)

Very low


1307

Summary GRADE profile 34 Sevelamer hydrochloride compared with calcium acetate 1308



Sevelamer Calcium acetate

Serum phosphate

8-week follow-up

1 RCT

Liu et al, 2006



Very low



Adherence


12-month follow-up

1 RCT

Chertow et al, 2003

42/52 (80.8%)

39/54 (72.2%)

Relative effect

RR = 1.62 (95% CI: 0.65 to 4.02)

Absolute effect


Very low



12-month follow-up

2 RCTs

Chertow et al, 2003

Qunibi et al, 2008

16/154 (10.4%)

14/157 (8.9%)

Relative effect

RR = 1.15 (95% CI: 0.38 to 3.48)

Absolute effect


Low

Adverse events – diarrhoea


12-month follow-up

2 RCTs

Chertow et al, 2003

Qunibi et al, 2008

26/154 (16.9%)

29/157 (18.5%)

Relative effect

RR = 0.91 (95% CI: 0.56 to 1.47)

Absolute effect


Low

Adverse events - nausea and vomiting


12-month follow-up

2 RCTs

Chertow et al, 2003

Qunibi et al, 2008

54/154 (35.1%)

63/157 (40.1%)

Relative effect

RR = 0.86 (95% CI: 0.61 to 1.21)

Absolute effect


Low



12-month follow-up

1 RCT

Qunibi et al, 2008

8/100 (8%)

4/103 (3.9%)

Relative effect

RR = 2.06 (95% CI: 0.64 to 6.63)

Absolute effect


Low

Serum calcium

8-week follow-up

1 RCT

Liu et al, 2006



Very low


12- to 21-month follow-up

2 RCTs

Barreto et al, 2008

Qunibi et al,


MD = 23.04 lower (95% CI: 124.44 lower to 78.36 higher)

Very low



2008


1309

Summary GRADE profile 35 Sevelamer hydrochloride compared with sevelamer hydrochloride + calcium carbonate 1310



Sevelamer Sevelamer + calcium carbonate

Serum phosphate

4-week follow-up

1 RCT

Koiwa et al, 2005



Very low

Serum calcium

4-week follow-up

1 RCT

Koiwa et al, 2005



Very low


1311

Summary GRADE profile 36 Sevelamer compared with aluminium hydroxide 1312



Sevelamer Aluminium hydroxide

Serum phosphate

8-week follow-up

1 RCT




Very low



8-week follow-up

1 RCT


2/15 (13.3%)

0/15 (0%)

Relative effect

RR = 5.00 (95% CI: 0.26 to 96.13)

Very low



Serum calcium

8-week follow-up

1 RCT




Very low


1313

Summary GRADE profile 37 Sevelamer hydrochloride compared with sevelamer carbonate 1314



Sevelamer hydrochloride Sevelamer carbonate

Adherence


24-week follow-up

1 RCT


127/148 (85.8%)

66/72 (91.7%)

Relative effect

RR = 0.94 (95% CI: 0.85 to 1.03)

Absolute effect


Very low



24-week follow-up

1 RCT


4/72 (5.6%)

1/141 (0.71%)

Relative effect

RR = 7.83 (95% CI: 0.89 to 68.8)

Absolute effect


Very low



24-week follow-up

1 RCT


4/72 (5.6%)

12/141 (8.5%)

Relative effect

RR = 0.65 (95% CI: 0.22 to 1.95)

Absolute effect


Very low



24-week follow-up

1 RCT


3/72 (4.2%)

22/141 (15.6%)

Relative effect

RR = 0.27 (95% CI: 0.08 to 0.86)

Absolute effect


Very low




1315

Summary GRADE profile 38 Sevelamer hydrochloride high-dose compared with sevelamer hydrochloride low-dose 1316



Sevelamer high-dose Sevelamer low-dose

Adherence


6-week follow-up

1 RCT

Chow et al, 2007

8/9 (88.9%)

16/18 (88.9%)

Relative effect

RR = 1 (95% CI: 0.75 to 1.33)

Absolute effect


Very low


1317

Summary GRADE profile 39 Lanthanum compared with placebo 1318



Lanthanum Placebo

Serum phosphate


4 RCTs

Al Baaj et al, 2005

Chiang et al, 2005

Joy et al, 2003

Shigematsu et al, 2008



Very low

Adherence


4-week follow-up

1 RCT

Al-Baaj et al, 2005

16/17 (94.1%)

18/19 (94.7%)

Relative effect

RR = 0.99 (95% CI: 0.85 to 1.16)

Absolute effect


Very low

Adverse event – constipation 1 RCT 24/111 1/31 Relative effect Very low




6-week follow-up


(21.6%) (3.2%) RR = 6.70 (95% CI: 0.94 to 47.6)

Absolute effect


Adverse event – diarrhoea



3 RCTs


Finn et al, 2004

Joy et al, 2002

12/273 (4.4%)

13/107 (12.1%)

Relative effect

RR = 0.32 (95% CI: 0.15 to 0.66)

Absolute effect


Very low

Adverse event - nausea and

vomiting



3 RCTs


Finn et al, 2004

Joy et al, 2002

71/273 (26%)

15/107 (14%)

Relative effect

RR = 2.12 (95% CI: 0.27 to 16.53)

Absolute effect


Very low

Adverse event - upper abdominal pain



2 RCTs

Finn et al, 2004 Shigematsu et al, 2008

9/224 (4%)

1/63 (1.6%)

Relative effect

RR = 1.48 (95% CI: 0.26 to 8.57)

Absolute effect


Very low

Serum calcium

8-week follow-up

1 RCT

Al-Baaj et al, 2005



Very low


1319

Summary GRADE profile 40 Lanthanum compared with any other binder 1320



Lanthanum Any other binder

Mortality

40-month follow-up

1 RCT

Wilson et al,

135/680 (19.9%)

157/674 (23.3%)

Relative effect

HR = 0.86 (95% CI: 0.68 to 1.08)

Low



2009 Absolute effect


Mortality - in those > 65 years

40-month follow-up

1 RCT

Wilson et al, 2009

44/163 (27%)

68/173 (39.3%)

Relative effect

HR = 0.68 (95% CI: 0.46 to 0.99)

Absolute effect


Very low

Mortality - in those < 65 years

40-month follow-up

1 RCT

Wilson et al, 2009

91/507 (17.9%)

89/501 (17.8%)

Relative effect

HR = 1.00 (95% CI: 0.75 to 1.34)

Absolute effect


Very low

Adverse events - diarrhoea


2-year follow-up

1 RCT

Finn et al, 2006

164/682 (24%)

217/677 (32.1%)

Relative effect

RR = 0.75 (95% CI: 0.63 to 0.9)

Absolute effect


Low



2-year follow-up

1 RCT

Finn et al, 2006

434/682 (63.6%)

470/677 (69.4%)

Relative effect

RR = 0.92 (95% CI: 0.85 to 0.99)

Absolute effect


Low



2-year follow-up

1 RCT

Finn et al, 2006

119/682 (17.4%)

161/677 (23.8%)

Relative effect

RR = 0.68 (95% CI: 0.52 to 0.88)

Absolute effect


Very low

Abbreviations: CI, confidence interval; RCT, randomised controlled trial; RR, relative risk; HR, Hazard ratio.

1321



Summary GRADE profile 41 Lanthanum compared with calcium carbonate 1322

Outcome Number of Studies Number of patients Effect Quality

Lanthanum Calcium carbonate

Serum phosphate

8-week follow-up

1 RCT




Very low



8-week follow-up

1 RCT


3/126 (2.4%)

5/132 (3.8%)

Relative effect

RR = 0.63 (95% CI: 0.15 to 2.58)

Absolute effect


Very low




3 RCTs




40/708 (5.6%)

33/448 (7.4%)

Relative effect

RR = 0.77 (95% CI: 0.48 to 1.21)

Absolute effect


Very low




2 RCTs



71/582 (12.2%)

30/316 (9.5%)

Relative effect

RR = 1.26 (95% CI: 0.84 to 1.89)

Absolute effect


Very low




3 RCTs




221/708 (31.2%)

76/448 (17%)

Relative effect

RR = 2.17 (95% CI: 1.02 to 4.6)

Absolute effect

198 more per 1000 (from 3 more to 611 more)

Very low



8-week follow-up

1 RCT


4/126 (3.2%)

7/132 (5.3%)

Relative effect

RR = 0.60 (95% CI: 0.18 to 2)

Absolute effect

21 fewer per 1000 (from 43 fewer to 53

Very low



more)

Serum calcium



4 RCTs




Spasovski et al, 2006

12/715 (1.7%)

122/456 (26.8%)

Relative effect

RR = 0.09 (95% CI: 0.03 to 0.27)

Absolute effect


Very low


1323

Summary GRADE profile 43 Magnesium carbonate compared with calcium carbonate 1324



Magnesium carbonate Calcium carbonate

Serum calcium


6-month follow-up

1 RCT

Tzanakis et al, 2008

9/14 (64.3%)

12/13 (92.3%)

Relative effect

RR = 0.37 (95% CI: 0.17 to 0.76)

Absolute effect


Very low


1325



Multiple treatment comparisons (serum phosphate and serum calcium) 1326

As there was a lack of evidence allowing a direct comparison between all of the possible treatments, a series of MTCs were carried 1327

out to aid the GDG decision-making process process (for further information on MTCs, please see glossary on page 235). Due to 1328

the use of continuous measures and the wide range of follow-up times presented within the evidence base, it was not possible to 1329

develop a single network which assessed all of the treatments at the various different time-points. Therefore, a series of network 1330

analyses were carried out at 3 months (90 days), 6 months (180 days) and 12 months (360 days). These analyses utilised the data 1331

available at each of these time points, and as a result, the interventions under consideration at each of these time points varied. In 1332

order to facilitate the construction of complete networks, it was decided that the comparators, “Standard Therapy” and “Calcium 1333

Based Binder”, should be analysed as a meta-class, “Any/Calcium based binders”. This assumption was made on the basis that 1334

within the relevant trials21 over 75% of the binders taken within the standard therapy arms were actually calcium based. MTC 1335

analyses were also carried out on the dichotomous outcomes of proportion of people with phosphate controlled and the risk of 1336

hypercalcaemia. 1337

The models used were based upon preliminary work carried out by NICE's technical support unit (appendix G). In brief, a standard 1338

network meta-analysis model in WinBUGS was used, utilising a previously published code (Dias et al 201122). The results 1339

presented below are those generated from a fixed effect analysis, which was recommended in the initial work carried out by the 1340

Technical Support Unit. 1341

21

Finn 2006, Malluche 2008 22

Dias, S et al (2011) NICE DSU Technical support Document 2: a generalised linear modelling framework for pair wise and network meta-analysis NICE Decision Support Unit available from http://www.nicedsu.org.uk

http://www.nicedsu.org.uk/



1342

Network binder map serum phosphate at 90 days patients CKD stage 5 on dialysis 1343

A total of 13 studies were included within the network allowing 8 treatments to be assessed against each other. 1344

SH

CC

LC

CA MG

CAMG

Any/CB

SC

1 study

1 study

1 study

1 study

1 study

1 study

2 studies 3 studies

2 studies

1345

Abbreviations: Any/CB, Any/calcium-based binder; CA, calcium acetate; CAMG, calcium acetate + magnesium carbonate; CC, calcium carbonate; LC, lanthanum carbonate; 1346 MG, magnesium carbonate; SC, sevelamer carbonate; SH, sevelamer hydrochloride 1347

1348



Summary GRADE profile 44 Quality of comparisons within the network 1349


Serum phosphate (follow up 90 days)


2,3 serious

5 serious

4 no serious very low

1 Barreto et al, 2008; Block et al, 2005; Braun et al, 2004; De Francisco et al, 2010; De Santo et al, 2006; Evenepoel et al, 2009; Ferreira et al, 2008; Fishbane et al, 2010;

Hutchison et al, 2005; Janssen et al, 1996; Malluche et al, 2008; Navarro-Gonzalez et al, 2011; Spiegel et al, 2007

2 Lack of detail on the randomisation

3 Studies were either not blinded or details were not provided




1350

1351



Mean difference matrix: serum phosphate at 90 days 1352

Calcium Carbonate

Any / calcium-based binders

Lanthanum carbonate


Calcium acetate


Sevelamer carbonate

Magnesium carbonate

Calcium Carbonate

- -0.129 mmol/l (95%CI: -0.281, 0.023)

0.068 mmol/l (95%CI: -0.034, 0.170)

-0.158 mmol/l (95%CI: -0.292, -0.023)

-0.154 mmol/l (95%CI: -0.315, 0.007)

-0.341 mmol/l (95%CI: -0.525, -0.157)

-0.648 mmol/l (95%CI: -0.809, -0.486)

-0.225 mmol/l (95%CI: -0.592, 0.141)


- - 0.197 mmol/l (95%CI: 0.048, 0.346)

-0.028 mmol/l (95%CI: -0.153, 0.096)

-0.025 mmol/l (95%CI: -0.179, 0.129)

-0.212 mmol/l (95%CI: -0.389, -0.035)

-0.519 mmol/l (95%CI: -0.672, -0.365)

-0.096 mmol/l (95%CI: -0.458, 0.267)

Lanthanum carbonate

- - - -0.226 mmol/l (95%CI: -0.374, -0.077)

-0.222 mmol/l (95%CI: -0.396, -0.050)

-0.409 mmol/l (95%CI: -0.604, -0.214)

-0.716 mmol/l (95%CI: -0.890, -0.542)

-0.293 mmol/l (95%CI: -0.665, 0.078)


- - - - 0.003 mmol/l (95%CI: -0.088, 0.094)

-0.184 mmol/l (95%CI: -0.310, -0.059)

-0.490 mmol/l (95%CI: -0.580, -0.401)

-0.068 mmol/l (95%CI: -0.407, 0.273)

Calcium acetate

- - - - - -0.187 mmol/l (95%CI: -0.343, -0.032)

-0.494 mmol/l (95%CI: -0.621, -0.366)

-0.071 mmol/l (95%CI: -0.397, 0.256)


- - - - - - -0.306 mmol/l (95%CI: -0.460, -0.153)

0.116 mmol/l (95%CI: -0.245, 0.481)

Sevelamer carbonate

- - - - - - - 0.423 mmol/l (95%CI: 0.072, 0.774)

Magnesium carbonate

- - - - - - - -

1353



Probability rank: serum phosphate at 90 days 1354

Intervention Probability best binder

Median rank (95%CI)

Sevelamer carbonate 0.991 1 (1, 1)

Calcium acetate+magnesium carbonate 0.000 2 (2, 3)

Magnesium carbonate 0.009 3 (2, 8)



Any / calcium-based binders 0.000 5 (3, 7)

Calcium carbonate 0.000 7 (5, 8)



1355

Relative effectiveness compared to calcium carbonate: serum phosphate at 90 days 1356

1357



1358

Network binder map serum Phosphate at 180 days patients CKD stage 5 on dialysis 1359


SH

CC

LC

CA

MG

CAMG

Any/CB

SC

1 study 1 study

1 study

1 study

1 study

2 studies

2 studies

1 study

2 studies

1361

Abbreviations: Any/CB, Any/calcium-based binder; CA, calcium acetate; CAMG, calcium acetate + magnesium carbonate; CC, calcium carbonate; LC, lanthanum carbonate; 1362 MG, magnesium carbonate; SC, sevelamer carbonate; SH, sevelamer hydrochloride 1363

1364







2,3 serious

5 serious


1 Barreto, et al 2008; Block, et al 2005; Braun, et al 2004; De Francisco, et al 2010; De Santo, et al 2006; Fishbane, et al 2010; Ferreira, et al 2008; Hutchison, et al 2005;

Janssen, et al 1995; Janssen, et al 1996; Malluche, et al 2008; Tzanakis, et al 2008. 2

Lack of detail on the randomisation 3 Studies were either not blinded or details were not provided




1366




Calcium carbonate


Lanthanum carbonate


Calcium acetate

Calcium acetate+ magnesium carbonate

Sevelamer carbonate

Magnesium carbonate

Calcium carbonate

- -0.065 mmol/l (95%CI: -0.262, 0.132)

0.006 mmol/l (95%CI: -0.132, 0.145)

-0.098 mmol/l (95%CI: -0.263, 0.066)

-0.097 mmol/l (95%CI: -0.389, 0.194)

-0.148 mmol/l (95%CI: -0.378, 0.081)

-0.088 mmol/l (95%CI: -0.298, 0.120)

-0.070 mmol/l (95%CI: -0.253, 0.113)


- - 0.072 mmol/l (95%CI: -0.127, 0.269)

-0.033 mmol/l (95%CI: -0.192, 0.126)

-0.032 mmol/l (95%CI: -0.339, 0.273)

-0.083 mmol/l (95%CI: -0.308, 0.142)

-0.023 mmol/l (95%CI: -0.228, 0.181)

-0.005 mmol/l (95%CI: -0.275, 0.264)

Lanthanum carbonate

- - - -0.104 mmol/l (95%CI: -0.297, 0.087)

-0.103 mmol/l (95%CI: -0.415, 0.207)

-0.154 mmol/l (95%CI: -0.404, 0.096)

-0.094 mmol/l (95%CI: -0.327, 0.136)

-0.076 mmol/l (95%CI: -0.306, 0.153)


- - - - 0.001 mmol/l (95%CI: -0.268, 0.269)

-0.050 mmol/l (95%CI: -0.209, 0.110)

0.010 mmol/l (95%CI: -0.119, 0.138)

0.028 mmol/l (95%CI: -0.218, 0.275)

Calcium acetate - - - - - -0.051 mmol/l(95%CI: -0.364, 0.262)

0.009 mmol/l (95%CI: -0.289, 0.307)

0.027 mmol/l (95%CI: -0.318, 0.372)


- - - - - - 0.060 mmol/l (95%CI: -0.145, 0.264)

0.078 mmol/l (95%CI: -0.216, 0.373)

Sevelamer carbonate

- - - - - - - 0.018 mmol/l (95%CI: -0.259, 0.297)

Magnesium carbonate

- - - - - - - -

1368

1369





Median rank (95%CI)










1371


1373



1374

Multiple treatment comparison: serum phosphate at 360 days 1375


SH

CA

CCLC

Any/CB

2 studies

3 studies

2 studies

3 studies

1 study

2 studies

1377 Abbreviations: Any/CB, Any/calcium-based binder; CA, calcium acetate; CC, calcium carbonate; LC, lanthanum carbonate; SH, sevelamer hydrochloride 1378

1379





2,3 serious

5 serious


1 Barreto et al, 2008; Block et al, 2005; Braun et al, 2004; Chertow et al, 2002; Chertow et al, 2003; Ferreira et al, 2008; Freemont et al, 2005; Kakuta et al, 2011; Janssen et

al, 1995; Janssen et al, 1996; Malluche et al, 2008; Qunibi et al, 2008; Spasovski et al, 2006 2 Lack of detail on the randomisation





1381




Calcium carbonate Any / calcium-based binders

Lanthanum carbonate Sevelamer hydrochloride Calcium acetate

Calcium carbonate - 0.073 mmol/l (95%CI: -0.058, 0.203)

0.191 mmol/l (95%CI: 0.030, 0.353)

0.031 mmol/l (95%CI: -0.047, 0.109)

-0.029 mmol/l (95%CI: -0.182, 0.124)


- - 0.118 mmol/l (95%CI: -0.063, 0.300)

-0.042 mmol/l (95%CI: -0.153, 0.069)

-0.102 mmol/l (95%CI: -0.276, 0.073)

Lanthanum carbonate - - - -0.160 mmol/l (95%CI: -0.331, 0.010)

-0.220 mmol/l (95%CI: -0.436, -0.005)


- - - - -0.060 mmol/l (95%CI: -0.194, 0.075)

Calcium acetate - - - - -

1383



Median rank (95%CI)







1385




1387 1388

1389

-1 -0.5 0 0.5 1


Lanthanum carbonate


Calcium acetate

MD (mmol/l) v. calcium carbonate



Multiple treatment comparison: serum calcium at 90 days 1390


SH

CC

CAMG

CA MGAny/CBLC3 studies 1 study1 study2 studies

2 studies

1 study

1392 1393 Abbreviations: Any/CB, Any/calcium-based binder ; CA, calcium acetate; CAMG, calcium acetate + magnesium carbonate; CC, calcium carbonate; LC, lanthanum carbonate; 1394 MG, magnesium carbonate; SH, sevelamer hydrochloride 1395

1396

1397





Serum calcium (follow up 90 days)


2,3 serious

5 serious


1 Barreto et al, 2008; Braun et al, 2004; De Francisco et al, 2010; De Santo et al, 2006; Evenepoel et al, 2009; Ferreria et al, 2006; Finn et al, 2006; Malluche et al, 2008;

Navarro-Gonzalez et al, 2011; Spiegel et al, 2007 2





. 1399



Mean difference matrix: serum calcium at 90 days 1400

Calcium carbonate


Lanthanum carbonate


Calcium acetate Calcium acetate+ magnesium carbonate

Magnesium carbonate

Calcium carbonate - -0.080 mmol/l (95%CI: -0.200, 0.040)

-0.170 mmol/l (95%CI: -0.293, -0.048)

-0.120 mmol/l (95%CI: -0.181, -0.058)

-0.050 mmol/l (95%CI: -0.127, 0.028)

-0.037 mmol/l (95%CI: -0.114, 0.040)

-0.060 mmol/l (95%CI: -0.215, 0.095)


- - -0.090 mmol/l (95%CI: -0.116, -0.064)

-0.040 mmol/l (95%CI: -0.143, 0.063)

0.030 mmol/l (95%CI: -0.083, 0.143)

0.042 mmol/l (95%CI:-0.070, 0.156)

0.020 mmol/l (95%CI: -0.156, 0.195)

Lanthanum carbonate

- - - 0.050 mmol/l (95%CI: -0.056, 0.157)

0.120 mmol/l (95%CI: 0.004, 0.236)

0.133 mmol/l (95%CI: 0.018, 0.249)

0.110 mmol/l (95%CI: -0.068, 0.287)


- - - - 0.070 mmol/l (95%CI: 0.023, 0.117)

0.083 mmol/l (95%CI: 0.036, 0.130)

0.060 mmol/l (95%CI: -0.082, 0.202)

Calcium acetate - - - - - 0.013 mmol/l (95%CI: -0.054, 0.079)

-0.010 mmol/l (95%CI: -0.145, 0.124)


- - - - - - -0.023 mmol/l (95%CI: -0.173, 0.127)

Magnesium carbonate

- - - - - - -

1401

1402



Probability rank: serum calcium at 90 days 1403


Median rank (95%CI)









1404

Relative effectiveness compared to calcium carbonate: serum calcium at 90 days 1405

1406 1407





SH

CC

CAMG

CA

MG

Any/CBLC

1 study

1 study2 studies

1 study

SC

1 study

1 study

1 study

1 study

1410 1411 Abbreviations: Any/CB, any/calcium-based binder ;CA, calcium acetate; CAMG, calcium acetate + magnesium carbonate;; CC, calcium carbonate; LC, lanthanum carbonate; 1412 MG, magnesium carbonate; SH, sevelamer hydrochloride; SC, sevelamer carbonate 1413

1414






9RCTs1 very serious

2,3 serious

5 serious


1 Barreto et al, 2008; De Francisco et al, 2010; De Santo et al, 2006; Ferreria et al, 2006; Fishbane et al, 2010; Finn et al, 2006; Janssen et al, 1995; Malluche et al, 2008;

Tzankis et al, 2008 2 Lack of detail on the randomisation





1416




Calcium carbonate


Lanthanum carbonate


Calcium acetate


Sevelamer carbonate

Magnesium carbonate

Calcium carbonate - -0.155 mmol/l (95%CI: -0.284, -0.027)

-0.259 mmol/l (95%CI: -0.392, -0.129)

-0.115 mmol/l (95%CI: -0.183, -0.048)

-0.179 mmol/l (95%CI: -0.315, -0.043)

-0.048 mmol/l (95%CI: -0.128, 0.031)

-0.135 mmol/l (95%CI: -0.222, -0.049)

-0.260 mmol/l (95%CI: -0.328, -0.192)


- - -0.104 mmol/l (95%CI: -0.131, -0.078)

0.040 mmol/l (95%CI: -0.069, 0.149)

-0.024 mmol/l (95%CI: -0.204, 0.155)

0.107 mmol/l (95%CI: -0.010, 0.224)

0.020 mmol/l (95%CI: -0.102, 0.140)

-0.105 mmol/l (95%CI: -0.250, 0.041)

Lanthanum carbonate

- - - 0.144 mmol/l (95%CI: 0.032, 0.256)

0.080 mmol/l (95%CI: -0.101, 0.261)

0.211 mmol/l (95%CI: 0.092, 0.331)

0.124 mmol/l (95%CI: 0.000, 0.248)

-0.001 mmol/l (95%CI: -0.148, 0.148)


- - - - -0.064 mmol/l (95%CI: -0.206, 0.078)

0.067 mmol/l (95%CI: 0.025, 0.109)

-0.020 mmol/l (95%CI: -0.074, 0.034)

-0.145 mmol/l (95%CI: -0.241, -0.048)

Calcium acetate - - - - - 0.131 mmol/l (95%CI: -0.017, 0.279)

0.044 mmol/l (95%CI: -0.108, 0.196)

-0.081 mmol/l (95%CI: -0.233, 0.071)


- - - - - - -0.087 mmol/l (95%CI: -0.155, -0.019)

-0.212 mmol/l (95%CI: -0.316, -0.107)

Sevelamer carbonate

- - - - - - - -0.125 mmol/l (95%CI: -0.235, -0.014)

Magnesium carbonate

- - - - - - - -

1418

1419





Median rank (95%CI)










1421


1423





SH

CA

CCLC

Any/CB

2 studies

3 studies

1 study

2 studies

2 studies

2 studies

1426 1427 Abbreviations: Any/CB, Any/Calcium Based binder; CA, calcium acetate; CC, calcium carbonate; LC, lanthanum carbonate; SH, sevelamer hydrochloride 1428

1429




12RCTs1 very serious

2,3 serious

5 serious


1 Barreto et al, 2008; Braun et al, 2004; Chertow et al, 2002; Chertow et al, 2003; Ferreria et al, 2006; Finn et al, 2006; Freemont et al, 2005; Janssen et al, 1995; Kakuta et al,

2011; Malluche et al, 2008; Qunibi et al, 2008; Spasovski et al, 2006 2 Lack of detail on the randomisation





1431




Calcium carbonate


Lanthanum carbonate Sevelamer hydrochloride Calcium acetate

Calcium carbonate - -0.016 mmol/l (95%CI: -0.062, 0.030)

-0.083 mmol/l (95%CI: -0.131, -0.034)

-0.099 mmol/l (95%CI: -0.133, -0.065)

-0.049 mmol/l (95%CI: -0.100, 0.002)


- - -0.066 mmol/l (95%CI: -0.090, -0.042)

-0.083 mmol/l (95%CI: -0.119, -0.046)

-0.033 mmol/l (95%CI: -0.086, 0.020)

Lanthanum carbonate - - - -0.016 mmol/l (95%CI: -0.058, 0.025)

0.033 mmol/l (95%CI: -0.023, 0.090)


- - - - 0.050 mmol/l (95%CI: 0.011, 0.089)

Calcium acetate - - - - -

1433



Median rank (95%CI)







1435




1437 1438

Multiple treatment comparison: proportion of patients achieving phosphate control 1439

A total of 15 studies were included within the network allowing 9 treatments to be assessed against each other. One of the trials 1440

included within the analysis was concerned with different dosages of sevelamer hydrochloride and this data was included within 1441

that one node. 1442

1443



LC P5 studies

Any-B

1 study

CC1 study

SH

SC

SH+CCCA

MG

1 study

1 study

1 study

1 study

1 study

2 studies

1444 1445 Abbreviations: Any-B, any binder; CA, calcium acetate; CC, calcium carbonate; LC, lanthanum carbonate; MG, magnesium carbonate; P, placebo; SH, sevelamer 1446 hydrochloride; SH+CC, sevelamer hydrochloride + calcium carbonate; SC, Sevelamer carbonate 1447

1448





Proportion of patients achieving phosphate control

15RCTs1 very serious

2,3 serious

5 serious


1 Al-Baaj et al, 2005; Chiang et al, 2005; Chow 2007; Evenepoel 2009; Finn et al, 2004; Finn et al, 2006; Fishbane et al, 2010; Hutchison et al, 2005; Janssen et al, 1996; Joy

et al, 2003; Koiwa et al, 2005; Liu et al, 2006; Shigematsu et al, 2008; Spiegel et al, 2007; Tzanakis et al, 2008. 2





1450



Hazard ratio matrix: proportion achieving phosphate control 1451

Placebo Calcium carbonate

Any binder Lanthanum carbonate


Calcium acetate

Sevelamer hydrochloride+ calcium carbonate

Sevelamer carbonate

Magnesium carbonate

Placebo - 0.274 (95%CI: 0.194, 0.387)

0.238 (95%CI: 0.176, 0.324)

0.284 (95%CI: 0.210, 0.384)

0.338 (95%CI: 0.155, 0.739)

0.317 (95%CI: 0.143, 0.707)

0.133 (95%CI: 0.051, 0.330)

0.470 (95%CI: 0.190, 1.173)

0.209 (95%CI: 0.077, 0.566)

Calcium carbonate

- - 0.868 (95%CI: 0.726, 1.040)

1.034 (95%CI: 0.873, 1.227)

1.230 (95%CI: 0.613, 2.486)

1.154 (95%CI: 0.562, 2.379)

0.485 (95%CI: 0.198, 1.123)

1.715 (95%CI: 0.745, 3.987)

0.761 (95%CI: 0.300, 1.933)

Any binder - - - 1.191 (95%CI: 1.124, 1.261)

1.417 (95%CI: 0.691, 2.928)

1.330 (95%CI: 0.634, 2.797)

0.558 (95%CI: 0.224, 1.319)

1.973 (95%CI: 0.842, 4.678)

0.876 (95%CI: 0.340, 2.262)

Lanthanum carbonate

- - - - 1.190 (95%CI: 0.581, 2.452)

1.116 (95%CI: 0.534, 2.343)

0.469 (95%CI: 0.188, 1.105)

1.656 (95%CI: 0.709, 3.920)

0.736 (95%CI: 0.286, 1.899)


- - - - - 0.941 (95%CI: 0.609, 1.431)

0.393 (95%CI: 0.157, 0.939)

1.390 (95%CI: 0.887, 2.242)

0.618 (95%CI: 0.223, 1.736)

Calcium acetate - - - - - - 0.419 (95%CI: 0.158, 1.052)

1.483 (95%CI: 0.797, 2.816)

0.658 (95%CI: 0.245, 1.810)


- - - - - - - 3.553 (95%CI: 1.326, 9.923)

1.574 (95%CI: 0.474, 5.426)

Sevelamer carbonate

- - - - - - - - 0.443 (95%CI: 0.144, 1.368)

Magnesium carbonate

- - - - - - - - -

1452



Probability rank: proportion achieving phosphate control 1453


Median rank (95%CI)

Sevelamer hydrochloride + calcium carbonate 0.732 1 (1, 5)


Any binder 0.042 3 (1, 6)






Placebo 0.000 9 (8, 9)


1454



Relative effectiveness compared to placebo: proportion achieving phosphate control 1455

1456 1457



Multiple treatment comparison: risk of hypercalcaemia 1458


SH

CA

CC

LC

C-Based

SH+CC

1 study

5 studies

4 studies

MG

2 studies

1 study

2 studies

1 study

1460

Abbreviations: C-Based, calcium based; CA, calcium acetate; CC, calcium carbonate; LC, lanthanum carbonate; MG, magnesium carbonate; SH, sevelamer hydrochloride; 1461 SH+CC, sevelamer hydrochloride + calcium carbonate 1462

1463





Risk of hypercalcaemia 16RCTs1 very serious

2,3 serious

5 serious


1 Asmus et al, 2005; Block et al, 2005; Braun et al, 2006; Chertow et al, 2002; Chertow et al, 2003; Evenepoel et al, 2009; Freemont et al,2005; Hutchison et al, 2005; Janssen

et al, 1996; Koiwa et al, 2005; Lin et al, 2011; Liu et al, 2006; Qunibi et al, 2008; Shigematsu et al, 2008; Spasovski et al, 2006; Tzanakis et al, 2008 2 Lack of detail on the randomisation





1465

Hazard ratio matrix: risk of hypercalcaemia 1466

Calcium carbonate


Lanthanum carbonate


Calcium acetate Sevelamer hydrochloride + calcium carbonate

Magnesium carbonate

Calcium carbonate - 1.412 (95%CI: 0.704, 2.915)

0.059 (95%CI: 0.031, 0.104)

0.436 (95%CI: 0.282, 0.663)

1.031 (95%CI: 0.612, 1.742)

0.386 (95%CI: 0.177, 0.795)

0.231 (95%CI: 0.080, 0.593)

Calcium-based binders - - 0.042 (95%CI: 0.016, 0.103)

0.309 (95%CI: 0.171, 0.531)

0.732 (95%CI: 0.365, 1.413)

0.272 (95%CI: 0.098, 0.715)

0.163 (95%CI: 0.045, 0.532)

Lanthanum carbonate - - - 7.347 (95%CI: 3.600, 15.910)

17.440 (95%CI: 8.062, 39.900)

6.521 (95%CI: 2.488, 17.180)

3.912 (95%CI: 1.164, 12.250)


- - - - 2.366 (95%CI: 1.638, 3.466)

0.885 (95%CI: 0.384, 1.957)

0.531 (95%CI: 0.169, 1.497)

Calcium acetate - - - - - 0.373 (95%CI: 0.152, 0.874)

0.223 (95%CI: 0.068, 0.660)


- - - - - - 0.600 (95%CI: 0.166, 2.036)

Magnesium carbonate - - - - - - -

1467



Probability rank: risk of hypercalcaemia 1468


Median rank (95%CI)



Sevelamer hydrochloride+calcium carbonate 0.000 3 (2, 4)




Calcium-based binders 0.000 7 (5, 7)

Abbreviations: CI, credibility interval

1469

Relative effectiveness compared to calcium carbonate: risk of hypercalcaemia 1470

1471





Use of phosphate binders in children with stage 5 CKD who are on 1475

dialysis 1476

Calcium carbonate compared with sevelamer 1477



statistically significant difference in mean serum phosphate level 1480

between those that received calcium carbonate and those that 1481

received sevelamer at 8 months (MD 0.16 mmol/l higher ([95% CI -1482

0.16 to 0.48]). 1483



calcium carbonate to be associated with a mean serum calcium 1486

level 0.20 mmol/l higher (95% CI 0.10 to 0.32 i.e. statistically 1487

significant) than that associated with sevelamer at 8 months. 1488

Use of phosphate binders in adults with stage 5 CKD who are on dialysis 1489

Calcium acetate compared with placebo 1490



calcium acetate to be associated with mean serum phosphate 1493

levels 0.62 mmol/l lower (95% CI -0.90 to -0.34 i.e. statistically 1494

significant) at 2 weeks than that associated with placebo. 1495



calcium acetate to be associated with mean serum calcium levels 1498




0.15 mmol/l higher (95% CI 0.05 to 0.25 i.e. statistically significant) 1499

at 2 weeks than that associated with placebo. 1500

Calcium acetate compared with calcium carbonate 1501



statistically significant difference in mean serum phosphate levels 1504

at 3 weeks between those that received calcium acetate and those 1505

that received calcium carbonate (MD 0.09 mmol/l lower [95% CI -1506

0.45 to 0.27]). 1507



statistically significant difference in mean serum calcium levels at 1510

3 weeks between those that received calcium acetate and those 1511

that received calcium carbonate (MD 0.09 mmol/l higher [95% CI -1512

0.13 to 0.31]). 1513


statistically significant difference in the risk of hypercalcaemia 1515

between those that received calcium acetate and those that 1516

received calcium carbonate over 12 months (RR 0.70 [95% CI 0.46 1517

to 1.06]). 1518

Calcium carbonate bread compared with calcium acetate 1519



calcium carbonate bread to be associated with mean serum 1522

phosphate levels 0.35 mmol/l lower (95% CI -0.40 to -0.30 i.e. 1523

statistically significant) at 8 weeks than that associated with calcium 1524

acetate. 1525

1526





calcium carbonate bread to be associated with mean serum 1529

calcium levels 0.65 mmol/l lower (95% CI -0.75 to -0.55 i.e. 1530

statistically significant) at 8 weeks than that associated with calcium 1531

acetate. 1532

Calcium carbonate compared with sevelamer + calcium carbonate 1533



calcium carbonate alone to be associated with mean serum 1536

phosphate levels 0.31 mmol/l higher (95% CI 0.05 to 0.57 i.e. 1537

statistically significant) at 4 weeks than that associated with 1538

sevelamer and calcium carbonate. 1539




4 weeks between those that received calcium carbonate alone and 1543

those that received sevelamer and calcium carbonate (MD 1544

0.02 mmol/l higher [95% CI -0.12 to 0.16 ]). 1545

Sevelamer + calcium carbonate (high-dose) compared with sevelamer + 1546

calcium carbonate (low-dose) 1547




at 8 weeks between those that received a high dose of sevelamer 1551

and calcium carbonate and those that received a low dose of 1552

sevelamer and calcium carbonate (MD 0.10 mmol/l lower [95% CI -1553

0.34 to 0.14]). 1554

1555





statistically significant difference in the risk of experiencing 1558

abdominal distension between those that received a high dose of 1559

sevelamer and calcium carbonate and those that received a low 1560

dose of sevelamer and calcium carbonate over 8 weeks (RR 1.27 1561

[95% CI 0.59 to 2.75]). 1562



constipation between those that received a high dose of sevelamer 1565

and calcium carbonate and those that received a low dose of 1566

sevelamer and calcium carbonate over 8 weeks (RR 1.10 [95% CI 1567

0.60 to 2.02]). 1568



8 weeks between those that received a high dose of sevelamer and 1571

calcium carbonate and those that received a low dose of sevelamer 1572

and calcium carbonate (MD 0.02 mmol/l higher [95% CI -0.34 to 1573

0.14]). 1574

Sevelamer compared with placebo 1575



sevelamer to be associated with a mean serum phosphate level 1578




statistically significant difference in degree of adherence to 1582

prescription over 2 weeks between those that received sevelamer 1583

and those that received placebo (MD 4.00% above prescription 1584

[95% CI -6.75 to 14.75]). 1585






diarrhoea over 2 weeks between those that received sevelamer 1589

and those that received placebo (RR 0.17 [95% CI 0.01 to 3.96]). 1590


statistically significant difference in the risk of experiencing upper 1592

abdominal pain over 2 weeks between those that received 1593

sevelamer and those that received placebo (RR 0.17 [95% CI 0.01 1594

to 3.96]). 1595


statistically significant difference in the risk of experiencing nausea 1597

and vomiting over 2 weeks between those that received sevelamer 1598




2 weeks between those that received sevelamer and those that 1602

received placebo (MD 0.03 mmol/l lower [95% CI -0.13 to 0.07]). 1603

Sevelamer compared with calcium-based binders 1604


3.5.3.22 Low-quality evidence from 1 RCT of 2103 patients showed no 1606

statistically significant difference in mortality over 44 months 1607

between those that received sevelamer and those that received 1608

calcium-based binders (HR 0.93 [95% CI 0.79 to 1.10]). 1609


no statistically significant difference in cardiovascular mortality over 1611

44 months between those that received sevelamer and those that 1612

received calcium-based binders (HR 0.93 [95% CI 0.74 to 1.17]). 1613



3.5.3.24 Very-low-quality evidence from a sub-group of 1176 patients aged 1614

over 65 in the RCT of 2103 patients showed sevelamer to be 1615

associated with a smaller risk of mortality over 44 months than that 1616

associated with calcium-based binders (HR 0.77 [95% CI 0.61 to 1617



over 65 in the RCT of 2103 patients showed no statistically 1620

significant difference in cardiovascular mortality over 44 months 1621




under 65 in the RCT of 2103 patients showed no statistically 1625

significant difference in mortality over 44 months between those 1626

that received sevelamer and those that received calcium-based 1627

binders (HR 1.18 [95% CI 0.91 to 1.53]). 1628



significant difference in cardiovascular mortality over 44 months 1631





no statistically significant difference in the risk of experiencing 1636

nausea and vomiting over 3.75 years between those that received 1637

sevelamer and those that received calcium-based binders (RR 1.00 1638

[95% CI0.06 to 15.92]). 1639


showed sevelamer to be associated with a smaller risk of 1641

hypercalcaemia over a mean of 8 months than that associated with 1642



calcium-based binders (RR 0.40 [0.31 to 0.52 i.e. statistically 1643

significant]). 1644

3.5.3.30 Low-quality evidence from 7 RCTs of 751 patients in total showed 1645

sevelamer to be associated with mean coronary artery calcification 1646

scores 124.13 lower (95% CI -205.88 to -42.38 i.e. statistically 1647

significant) than that associated with calcium-based binders when 1648

followed-up for 12 to 21 months. 1649

3.5.3.31 Very-low-quality evidence from a sub-group of 158 patients from 1650

3 RCTs who had baseline coronary artery calcification scores 1651

below 30 showed sevelamer to be associated with mean coronary 1652

artery calcification scores 210.52 lower (95% CI -382.36 to -38.69 1653

i.e. statistically significant) than that associated with calcium-based 1654

binders when followed-up for 12 to 18 months. 1655

3.5.3.32 Very-low-quality evidence from a sub-group of 64 patients with 1656

diabetes from 1 RCT showed no statistically significant difference in 1657

mean coronary artery calcification scores at 2 years between those 1658


binders (MD 289.00 lower [95% CI -651.99 to 73.99]). 1660

3.5.3.33 Very-low-quality evidence from a sub-group of 45 patients without 1661

diabetes from 1 RCT showed no statistically significant difference in 1662

mean coronary artery calcification scores at 2 years between those 1663


binders (MD 100.00 lower [95% CI -256.33 to 56.23]). 1665

Sevelamer compared with calcium carbonate 1666




at 4 weeks between those that received sevelamer and those that 1670

received calcium carbonate (MD 0.04 mmol/l higher [95% CI -0.20 1671

to 0.28]). 1672




statistically significant difference in adherence over 12 months 1674


calcium carbonate (RR 1.02 [95% CI 0.84 to 1.23]). 1676



statistically significant difference in risk of experiencing abdominal 1679

distension over 4 weeks between those that received sevelamer 1680

and those that received calcium carbonate (RR 2.33 [95% CI 0.49 1681

to 11.01]). 1682

3.5.3.37 Very-low-quality evidence from 2 RCT of 239 patients in total 1683

showed sevelamer to be associated with a greater risk of 1684

experiencing constipation than that associated with calcium 1685

carbonate (RR 3.40 [95% CI 1.34 to 8.63 i.e. statistically 1686

significant]), when followed up for 4 to 52 weeks. 1687


sevelamer to be associated with mean calcium level 0.24 mmol/l 1689

lower (95% CI -0.37 to – 0.11 i.e. statistically significant) at 4 weeks 1690

than that associated with calcium carbonate. 1691


showed sevelamer to be associated with mean coronary artery 1693

calcification scores 250.11 lower (95% CI -480.63 to -19.59 i.e. 1694

statistically significant) than that associated with calcium carbonate 1695

at 12 months. 1696

Sevelamer compared with calcium acetate 1697



statistically significant difference in mean serum phosphate level 1700




calcium acetate (MD 0.23 mmol/l higher [95% CI -0.10 to 0.56]). 1702




calcium acetate (RR 1.62 [95% CI 0.65 to 4.02]). 1706



no statistically significant difference in risk of experiencing 1709

constipation over 12 months between those that received 1710

sevelamer and those that received calcium acetate (RR 1.15 [95% 1711

CI 0.38 to 3.48]). 1712

3.5.3.43 Low-quality evidence from the 2 RCTs of 311 patients in total 1713

showed no statistically significant difference in risk of experiencing 1714

nausea and vomiting over 12 months between those that received 1715


CI 0.61 to 1.21]). 1717

3.5.3.44 Low-quality evidence from the 2 RCTs of 311 patients in total 1718

showed no statistically significant difference in risk of experiencing 1719

diarrhoea over 12 months between those that received sevelamer 1720

and those that received calcium acetate (RR 0.91 [95% CI 0.56 to 1721

1.47]). 1722


statistically significant difference in risk of experiencing upper 1724

abdominal pain over 12 months between those that received 1725


CI 0.64 to 6.63]). 1727


statistically significant difference in mean serum calcium level 1729




calcium acetate (MD 0.15 mmol/l lower [95% CI -0.30 to 0.00]). 1731


showed no statistically significant difference in mean coronary 1733

artery calcification scores when followed-up for 12 to 21 months 1734


calcium acetate (MD 23.04 lower [95% CI -124.44 to 78.36]). 1736

Sevelamer compared with sevelamer + calcium carbonate 1737



sevelamer alone to be associated with mean serum phosphate 1740

levels 0.35 mmol/l higher (95% CI 0.17 to 0.53 i.e. statistically 1741

significant) at 4 weeks than that associated with sevelamer and 1742

calcium carbonate. 1743



sevelamer alone to be associated with mean serum calcium levels 1746


at 4 weeks than that associated with sevelamer and calcium 1748

carbonate. 1749

Sevelamer compared with aluminium hydroxide 1750




at 8 weeks between those that received sevelamer and those that 1754

received aluminium hydroxide (MD 0.12 mmol/l lower [95% CI -0.15 1755

to 0.39]). 1756

1757






constipation between those that received sevelamer and those that 1761

received aluminium hydroxide over 8 weeks (RR 5.00 [95% CI 0.26 1762

to 96.13]). 1763



8 weeks between those that received sevelamer and those that 1766

received aluminium hydroxide (MD 0.01 mmol/l lower [95% CI -0.12 1767

to 0.10]). 1768

Sevelamer hydrochloride compared with sevelamer carbonate 1769



statistically significant difference in adherence over 24 weeks 1772

between those that received sevelamer hydrochloride and those 1773

that received sevelamer carbonate (RR 0.94 [95% CI 0.85 to 1.03]). 1774




diarrhoea over 24 weeks between those that received sevelamer 1778

hydrochloride and those that received sevelamer carbonate (RR 1779

0.65 [95% CI 0.22 to 1.95]). 1780



constipation over 24 weeks between those that received sevelamer 1783

hydrochloride and those that received sevelamer carbonate (RR 1784

7.83 [95% CI 0.89 to 68.80]). 1785




sevelamer hydrochloride to be associated with a smaller risk of 1787

experiencing nausea and vomiting over 24 weeks than that 1788

associated with sevelamer carbonate (RR 0.27 [95% CI 0.08 to 1789


Sevelamer (high-dose) compared with sevelamer (low-dose) 1791




between those that received a high dose of sevelamer and those 1795

that received a low dose of sevelamer (RR 1.00 [95% CI 0.75 to 1796

1.33]). 1797

Lanthanum compared with placebo 1798



showed lanthanum to be associated with a mean serum phosphate 1801

level 0.66 mmol/l lower (95% CI -0.86 to -0.47 i.e. statistically 1802

significant) than that associated with placebo when followed up for 1803

4 to 7 weeks. 1804


statistically significant difference in adherence over 4 weeks 1806


placebo [RR 0.99 (95% CI 0.85 to 1.16]). 1808



showed lanthanum to be associated with a smaller risk of 1811

experiencing diarrhoea than that associated with placebo (RR 0.32 1812

[95% CI 0.15 to 0.66 i.e. statistically significant]), when followed-up 1813

for 4 to 104 weeks. 1814



3.5.3.61 Very-low-quality evidence from the 3 RCTs of 380 patients in total 1815

showed no statistically significant difference in the risk of 1816

experiencing nausea or vomiting between those that received 1817

lanthanum and those that received placebo (RR 2.12 [95% CI 0.27 1818

to 16.53]), when followed-up for 4 to 104 weeks. 1819



experiencing upper abdominal pain between those that received 1822

lanthanum and those that received placebo (RR 1.48 [95% CI 0.26 1823

to 8.57]), when followed-up for 6 to 104 weeks. 1824



constipation over 6 weeks between those that received lanthanum 1827



lanthanum to be associated with a mean serum calcium level 1830

0.09 mmol/l higher (95% CI 0.01 to 0.17 i.e. statistically significant) 1831


Lanthanum compared with any other binder 1833



statistically significant difference in mortality over 40 months 1836


other binders [HR 0.86 (95% CI 0.68 to 1.08]). 1838


over 65 in the RCT of 1354 patients showed lanthanum to be 1840

associated with a smaller risk of mortality over 40 months than that 1841

associated with other binders (HR 0.68 [95% CI 0.46 to 0.99 i.e. 1842






significant difference in mortality over 40 months between those 1846

that received lanthanum and those that received other binders (HR 1847

1.00 [95% CI 0.75 to 1.34]). 1848


3.5.3.68 Low-quality evidence from 1 RCT of 1359 patients showed 1850

lanthanum to be associated with a smaller risk of experiencing 1851

diarrhoea over 2 years than that associated with other binders (RR 1852

0.75 [95% CI 0.63 to 0.90 i.e. statistically significant]). 1853



upper abdominal pain over 2 years than that associated with other 1856

binders (RR 0.68 [95% CI 0.52 to 0.88 i.e. statistically significant]). 1857

3.5.3.70 Low-quality evidence from 1 RCT of 1359 patients showed 1858


nausea and vomiting over 2 years than that associated with other 1860

binders (RR 0.92 [95% CI 0.85 to 0.99 i.e. statistically significant]). 1861

Lanthanum compared with calcium carbonate 1862




at 8 weeks between those that received lanthanum and those that 1866

received calcium carbonate (MD 0.08 mmol/l higher [95% CI -0.03 1867

to 0.19]). 1868



showed lanthanum to be associated with a lower risk of 1871

hypercalcaemia over a mean of 8 months than that associated with 1872



calcium carbonate (RR 0.09 [95% CI 0.03 to 0.27 i.e. statistically 1873

significant]). 1874



experiencing constipation between those that received lanthanum 1877

and those that received calcium carbonate (RR 0.77 [95% CI 0.48 1878

to 1.21]). 1879

3.5.3.74 Very-low-quality evidence from the 3 RCTs of 1156 patients in total 1880

showed lanthanum to be associated with a greater risk of 1881

experiencing nausea and vomiting than that associated with 1882

calcium carbonate (RR 2.17 [95% CI 1.02 to 4.60 i.e. statistically 1883

significant]), when followed up for 8 to 52 weeks. 1884


no statistically significant difference in the risk of experiencing 1886

diarrhoea between those that received lanthanum and those that 1887

received calcium carbonate (RR 1.26 [95% CI 0.84 to 1.89]), when 1888

followed up for 20 to 52 weeks. 1889



abdominal distension over 8 weeks between those that received 1892

lanthanum and those that received calcium carbonate (RR 0.63 1893

[95% CI 0.15 to 2.58]). 1894

3.5.3.77 Low-quality evidence from the RCT of 258 patients showed no 1895

statistically significant difference in the risk of experiencing upper 1896

abdominal pain over 8 weeks between those that received 1897

lanthanum and those that received calcium carbonate (RR 0.60 1898

[95% CI 0.18 to 2.00]). 1899

1900



Magnesium carbonate compared with calcium carbonate 1901



magnesium carbonate to be associated with a lower risk of 1904

hypercalcaemia over 6 months than that associated with calcium 1905

carbonate (RR 0.37 [0.17 to 0.76 i.e. statistically significant]). 1906

Multi treatment comparison: serum phosphate 1907


3.5.3.79 Serum phosphate at 90 days 1909

Evidence from 1 MTC of 8 treatments, containing 13 studies of low 1910

or very low quality, suggested that sevelamer carbonate had a 1911

99.1% probability. 1912



or very low quality suggested that calcium acetate with magnesium 1915

carbonate had a 36.3% probability (median rank 2 [95%CI 1, 8]) of 1916

being the best treatment to control serum phosphorus at 180 days 1917

in patients with CKD stage 5 on dialysis, with calcium acetate 1918

having a 25.5% probability (median rank 4 [95% CI 1, 8]), 1919

magnesium carbonate 16.3% probability (median rank 5 [95%CI 1, 1920

8]). 1921



or very low quality suggested that calcium acetate had a 61.2% 1924

probability (median rank 1 [95%CI 1, 4]) of being the best treatment 1925

to control serum phosphorus at 360 days in patients with CKD 1926

stage 5 on dialysis, with calcium carbonate having a 30.3% 1927

probability (median rank 2 [95% CI 1, 4]). 1928



Multi treatment comparison: achieving phosphate control 1929


3.5.3.82 Evidence from 1 MTC of 9 treatments, containing 15 studies of low 1931

or very low quality suggested that sevelamer hydrochloride plus 1932

calcium carbonate had a 73.2% probability (median rank 1 [95%CI 1933

1, 5]) of being the best treatment to achieve phosphate control in 1934

patients with CKD stage 5 on dialysis, with magnesium carbonate 1935

having a 21.2% probability (median rank 2 [95%CI 1, 8]). 1936

Multi treatment comparison: serum calcium 1937


3.5.3.83 Serum calcium at 90 days 1939


or very low quality suggested that Lanthanum Carbonate had a 1941

77.1% probability (median rank 1 [95%CI 1, 3]) of being the best 1942

treatment to control serum calcium at 90 days in patients with CKD 1943

stage 5 on dialysis, with sevelamer hydrochloride having a 14.1% 1944

probability (median rank 2 [95%CI 1, 4]). 1945



or very low quality suggested that Lanthanum carbonate had a 1948


treatment to control serum calcium at 180 days in patients with 1950

CKD stage 5 on dialysis, with Magnesium Carbonate having a 1951

45.6% probability (median rank 2 [95% CI 1, 4]). 1952



or very low quality suggested that sevelamer hydrochloride had a 1955


treatment to control serum calcium at 360 days in patients with 1957



CKD stage 5 on dialysis, with lanthanum carbonate having a 22% 1958


Multi treatment comparison: risk of hypercalcaemia 1960


3.5.3.86 Evidence from 1 MTC of 7 treatments, containing 16 studies of low 1962

or very low quality suggested that lanthanum carbonate had a 1963

98.6% (median rank 1 [95% CI 1, 1]) probability of being the best 1964

treatment to reduce the chance of hypercalcaemia in patients with 1965

CKD stage 5 on dialysis, with magnesium carbonate having a 1.4% 1966


3.5.4 Health economic modelling 1968

This is a summary of the modelling carried out for this review question. See 1969

appendix F for full details of the modelling carried out for the guideline. 1970

Decision problems 1971

Two questions were addressed, based on the literature that had been 1972

identified in the review of clinical effectiveness evidence: 1973

For adults with stage 5 CKD on dialysis, which phosphate binder(s) provide 1974

cost-effective first-line management of hyperphosphataemia and its 1975

associated outcomes? 1976

For adults with stage 5 CKD on dialysis, what is the cost effectiveness of 1977

various sequences of phosphate binders in the management of 1978

hyperphosphataemia? 1979

Both questions were explored using the same model structure. An initial 1980

attempt was made to model the use of different phosphate binders in people 1981

in CKD stage 4 and people in CKD stage 5 but not on dialysis; however, 1982

insufficient data were available to estimate the effectiveness of different 1983

binders in this population (see section 3.5.2). Similarly, there were insufficient 1984

data to undertake a separate model exploring the cost effectiveness of 1985

different phosphate binders in children. 1986



Therefore, the analysis presented directly applies to adults with CKD stage 5 1987

on dialysis, and (to an extent) was extrapolated to people with CKD stage 4 or 1988

5 who are not on dialysis, and children. 1989

Methods 1990

The model used an individual patient (‘discrete event’) simulation approach, 1991

capturing costs and effects associated with a series of discrete health states. 1992

Discrete event simulation was considered to be the most appropriate structure 1993

for the analysis because of the complex relationships between biochemical 1994

intermediate outcomes (such as blood phosphate and calcium concentrations) 1995

and long-term, patient-relevant outcomes such as cardiovascular events, 1996

fractures and death. Figure 1 presents a simplified representation of the 1997

model structure, which was based on the natural history of CKD 5 and inputs 1998

from the GDG. 1999

Figure 1 Model structure 2000

2001

In simulating the course of an individual patient, a virtual patient is created 2002

and assigned several characteristics, including age, sex, baseline serum 2003

phosphate level and baseline serum calcium level, with these data drawn from 2004

distributions reflecting patients in the UK Renal Registry. 2005

Based on these baseline characteristics, the model estimates the phosphate 2006

and calcium profiles of the simulated individual receiving 1 year's treatment 2007

post-PTx

dead

fracture

CV event

CKD 5

(on dialysis)

parathyroid-

ectomy

post-

transplantation

transplant-

ation

cohort starts here



with calcium carbonate, which is then used as a common baseline with 2008

reference to which the effects of all other treatments are estimated. Treatment 2009

effects relative to calcium carbonate are drawn from 6 network meta-analyses 2010

reported in section 3.4.2 (phosphate and calcium at 3 different time points: 3 2011

months, 6 months and 1 year). The data underpinning these analyses enabled 2012

consideration of 4 different phosphate binders: calcium carbonate, calcium 2013

acetate, sevelamer hydrochloride and lanthanum carbonate. Insufficient data 2014

were available to derive conclusions on the use of sevelamer carbonate, 2015

aluminium hydroxide, magnesium carbonate. 2016

In the absence of evidence of treatment effects beyond 1 year's follow-up, it 2017

was necessary to rely on extrapolation and assumption to extend the 2018

phosphate and calcium profiles of simulated patients beyond 1 year. In the 2019

model's base case, phosphate levels are assumed to remain constant, 2020

whereas calcium levels are assumed to continue to rise (or fall) at the same 2021

rate observed on average in year 1. Alternative assumptions were tested in 2022

sensitivity analysis (for full details, see appendix F). 2023

Having estimated a simulated patient's biochemical profile, the model uses 2024

these measures as surrogate predictors to calculate event probabilities. The 2025

events that were deemed relevant for this analysis were: 2026

all-cause mortality 2027

cardiovascular events 2028

requirement for parathyroidectomy, and 2029

fractures. 2030

These events are shown in red, outlined arrows in Error! Reference source 2031

not found., indicating that the transitions are directly related to biochemical 2032

parameters (and, hence, treatment effects). 2033

The estimates used to calculate the probability of the relevant events were 2034

obtained from a systematic review of observational prognostic studies. We 2035

identified 36 studies in people with CKD (stage 4 or 5) relating serum 2036

phosphate and serum calcium to the events of interest in a single multivariate 2037

model. A meta-analysis of the various measures of effect was not performed 2038



because it would be inappropriate to pool estimates that come from a 2039

heterogeneous collection of multivariable models. Instead, the evidence was 2040

systematically appraised and the most appropriate individual study(s) were 2041

selected. For full details of systematic review of prognostic studies, see 2042

appendix F. 2043

Adverse events associated with the different phosphate binders – a proportion 2044

of which lead to discontinuation due to intolerance – were also simulated, on 2045

the basis of reported event-rates in the pooled literature. 2046

Once the incidence of events (and consequent state-transitions) was 2047

estimated, costs and quality-of-life values were attached to the events and 2048

health-states and aggregated for each simulated individual's lifetime. 2049

The cost of the binders themselves was calculated with reference to the 2050

average dose at which each was delivered to achieve the clinical effect 2051

observed in the evidence base. Calculated costs as used in the model are 2052

shown in Table 7. 2053

Table 7 Cost of phosphate binders used in model 2054

Binder Average daily dose (g) Cost per g (£) Cost per day (£)

Calcium carbonate 3.66 0.07 0.27

Calcium acetate 4.82 0.11 0.52

Sevelamer hydrochloride 5.22 0.82 5.22

Lanthanum carbonate 1.93 2.25 4.36

2055

In the analyses presented here, cohorts of 250,000 virtual patients were 2056

simulated for each treatment arm. Finally, the average cost and quality-of-life 2057

values for each cohort were calculated. 2058

The model was implemented in Visual Basic for Applications, using Microsoft 2059

Excel as a ‘front-end’, in which parameters are specified and results collected 2060

and analysed. Costs and benefits were discounted at 3.5% per annum each. 2061



Treatments simulated 2062

First-line binders 2063

To provide a cost–utility estimate for different phosphate binders used as 2064

first-line agents, a simplified scenario was assumed in which patient cohorts 2065

were assigned to a single binder. Aside from dropout due to adverse events, 2066

no switching or addition of different binders was simulated, and serum 2067

phosphate and calcium levels were allowed to change based on the observed 2068

effect in the evidence base without additional intervention. Importantly, this 2069

means that the calcium level of simulated patients receiving calcium-based 2070

binders was allowed to rise indefinitely (which is likely to be at odds with 2071

current practice). 2072

Sequential use of binders 2073

As well as estimating the costs and effects of first-line treatment with various 2074

phosphate binders, the model was configured to simulate cohorts receiving 2075

predetermined sequences of binders, with patients switching between them as 2076

time progresses. Because the GDG believed that the main reason for 2077

switching in practice is hypercalcaemia associated with the use of 2078

calcium-based binders, the scenario of greatest interest was one in which 2079

people switched from a calcium-based to a non-calcium-based binder when 2080

simulated serum calcium levels exceeded 2.54 mmol/l (based on GDG 2081

advice). In this scenario, the sequences modelled are: 2082

Calcium carbonate sevelamer hydrochloride 2083

Calcium carbonate lanthanum carbonate 2084

Calcium acetate sevelamer hydrochloride 2085

Calcium acetate lanthanum carbonate 2086

Calcium carbonate alone (no switch) 2087

Calcium acetate alone (no switch) 2088

Sevelamer hydrochloride alone (no switch) 2089

Lanthanum carbonate alone (no switch) 2090

The latter 4 strategies were included in the decision space to estimate the 2091

potential opportunity costs of switching treatment at all. 2092



Sequences were modelled on basis of generic evidence (based on the 2093

network meta-analyses) because of lack of primary evidence on sequential 2094

use of binders. 2095

Types of analysis 2096

At the time of reporting, detailed exploration of parameter uncertainty 2097

(probabilistic sensitivity analysis) has not been feasible because of the 2098

additional computational burden introduced by the discrete event simulation 2099

approach. However, it should be emphasised that, because first-order 2100

uncertainty is properly captured in this model, the impact of inter-individual 2101

variability in parameter values is reflected in the base-case analysis. 2102

Results: first line 2103

In its base case (Table 8), the economic model suggested that calcium-based 2104

phosphate binders are likely to be preferred to either sevelamer hydrochloride 2105

or lanthanum carbonate in the first-line management of hyperphosphataemia. 2106

Table 8 Cost–utility results for first-line use 2107

Binder

Discounted Incremental

Costs (£)

Effects (QALYs)

Costs (£)

Effects (QALYs)

ICER (£ / QALY)

Calcium carbonate £14,151 3.597

Calcium acetate £15,171 3.721 £1020 0.124 £8197

Lanthanum carbonate £23,615 3.731 extendedly dominateda

Sevelamer hydrochloride £25,823 3.842 £10,652 0.121 £87,916 a extendedly dominated: indicates that the specified strategy has a higher ICER than at least one other 2108

option, when we estimate the additional value each provides over and above that which can be 2109 achieved with a common (cheaper, less effective) comparator 2110

Calcium acetate appears to provide good value for money compared with 2111

calcium carbonate, providing an average health gain of around one-eighth of a 2112

QALY at an additional cost of just over £1000, equating to an incremental cost 2113

effectiveness ratio (ICER) of approximately £8000 per QALY gained. 2114

Sevelamer hydrochloride was found to be the most effective treatment; 2115

however, the additional health gains predicted, when compared with calcium 2116

acetate, come at a cost of almost £90,000 per QALY. The simulated cohort 2117

receiving lanthanum carbonate accrued very similar health gains to that 2118

receiving calcium acetate, but at much higher cost. It is said to be extendedly 2119



dominated, in this analysis, which indicates that – regardless of the assumed 2120

maximum acceptable ICER threshold – better value can be achieved by using 2121

calcium acetate and/or sevelamer hydrochloride. 2122

Figure 2 illustrates these results on the cost–utility plane. 2123

Figure 2 Cost–utility plane for first-line use 2124

2125

In one-way sensitivity analysis, calcium acetate remained good value for 2126

money compared with calcium carbonate, except when the difference in 2127

serum calcium at 12 months was varied so that calcium acetate was 2128

associated with higher levels than calcium carbonate (mean difference 2129

+0.002 mmol/l, compared with a base-case point-estimate of −0.049 mmol/l). 2130

Independently varying all other parameters within plausible ranges had no 2131

effect on the implied decision. 2132

The finding that sevelamer hydrochloride and lanthanum carbonate are 2133

associated with high opportunity costs is entirely robust to individual 2134

parameter variations: none had a qualitative impact on model results, with all 2135

ICERs remaining far greater than £30,000 per QALY gained. 2136

£5.0K

£10.0K

£15.0K

£20.0K

£25.0K

£30.0K

3.55 3.65 3.75 3.85

Co

sts

QALYs

CC CA

SH LC

£20K/QALY £30K/QALY

Frontier



Results: sequential use 2137

Base-case cost–utility results for the sequential treatment scenarios are 2138

tabulated in Table 9. Figure 3 illustrates these results on the cost–utility plane. 2139

Table 9 Cost–utility results for sequential use 2140

Binder sequence

Discounted Incremental

Costs (£)

Effects (QALYs)

Costs (£)

Effects (QALYs)

ICER (£ / QALY)

Calcium carbonate (no switch) £14,151 3.597

Calcium acetate (no switch) £15,171 3.721 £1020 0.124 £8197

Calcium acetate lanthanum carbonate £18,965 3.810 extendedly dominateda

Calcium carbonate lanthanum carbonate £19,124 3.765 dominatedb

Calcium acetate sevelamer hydrochloride £19,856 3.844 £4684 0.123 £38,078

Calcium carbonate sevelamer hydrochloride £20,132 3.798 dominatedb

Lanthanum carbonate (no switch) £23,615 3.731 dominatedb

Sevelamer hydrochloride (no switch) £25,823 3.842 dominatedb

a extendedly dominated: indicates that the specified strategy has a higher ICER than at least one other 2141

option, when we estimate the additional value each provides over and above that which can be 2142 achieved with a common (cheaper, less effective) comparator 2143

b dominated: indicates that one or more treatment options are both cheaper and more effective than the 2144

specified strategy 2145

Figure 3 Cost–utility plane for sequential use 2146

2147

As in the first-line-only setting, reliance on calcium acetate as an initial binder 2148

appears to provide good value for money: such strategies tend to provide 2149

conspicuously greater QALY gains than those based on calcium carbonate at 2150

a net cost that is either modestly increased or reduced. 2151

£5.0K

£10.0K

£15.0K

£20.0K

£25.0K

£30.0K

3.55 3.65 3.75 3.85

Co

sts

QALYs

CC CA SHLC CC -> SH CC -> LCCA -> SH CA -> LC £20K/QALY£30K/QALY Frontier



If remaining on calcium acetate indefinitely is considered a clinically 2152

appropriate option, it may be hard to justify switching to a non-calcium-based 2153

binder, instead, even in people with raised serum calcium. This is because, 2154

although switching strategies are associated with QALY gains, the extra 2155

acquisition costs of the non-calcium-based binders make the opportunity cost 2156

of adopting them represent an ineffective use of NHS resources (the ICER for 2157

best-value alternative – switching from calcium acetate to sevelamer 2158

hydrochloride – approaches £40,000 per QALY gained). 2159

For patients who are unable to tolerate calcium acetate, the cost effectiveness 2160

of switching to non-calcium-based binders should be judged against the only 2161

plausible alternative, which is indefinite treatment with calcium carbonate. 2162

Because calcium carbonate is a less effective choice than calcium acetate, 2163

this comparison is more favourable to the calcium-free alternatives, with 2164

ICERs in the range of £20,000–£30,000 per QALY. 2165

If calcium-based binders are considered to be fundamentally contraindicated 2166

in any individual case, the only remaining options in the decision-space are 2167

sevelamer hydrochloride and lanthanum carbonate. The relative cost-2168

effectiveness of strategies switching from calcium-based binders to either of 2169

these alternatives is difficult to distinguish. Although the model estimates an 2170

ICER of £26,090 for the comparison between the two, this is based on 2171

extremely small differences in costs and QALYs, and it is notable that the two 2172

options have similar ICERs compared with a common baseline of indefinite 2173

calcium acetate treatment (£38,078 per QALY gained for switching to 2174

sevelamer hydrochloride and £42,683 per QALY gained for switching to 2175

lanthanum carbonate). Accordingly, if either treatment is considered 2176

acceptable value for money, it is very likely that the other would have to be 2177

judged similarly. 2178

In one-way sensitivity analysis, very few alterations to individual model 2179

parameters affected the apparent superiority of indefinite calcium acetate 2180

therapy when compared with strategies that transferred people to a non-2181

calcium-based binder when their serum calcium reached a given threshold. 2182

Analyses in which that threshold was raised from its base-case value of 2183



2.54 mmol/l were associated with some improvement in estimated value for 2184

money for switching strategies; however, even when the threshold was as 2185

high as 3 mmol/l, the ICER for switching to either drug remained above 2186

£30,000 per QALY gained. Conversely, switching people to non-calcium-2187

based binders at a lower threshold made switching strategies appear less cost 2188

effective. 2189

Three other parameters had a potentially important impact on findings. 2190

Switching strategies were associated with ICERs between £20,000 and 2191

£30,000 per QALY gained when: 2192

sevelamer hydrochloride or lanthanum carbonate were at the lower (most 2193

effective) bounds of their credibility intervals for impact on serum calcium 2194

after one year; or 2195

calcium acetate was at the higher (least effective) bound of its credibility 2196

interval for the same parameter; or 2197

low calcium levels were associated with a very substantially increased 2198

mortality risk compared with people with normal serum calcium. 2199

Details are provided in appendix F. 2200

For the comparison of strategies switching to sevelamer hydrochloride and 2201

those switching to lanthanum carbonate, plausible alterations to a large 2202

number of parameters result in the apparent superiority of one or other option 2203

over the other. This reinforces the view that it is difficult to distinguish between 2204

these options, from a health-economic perspective; additional research on the 2205

relative effectiveness of the treatments would be necessary, if it is judged 2206

worthwhile to establish which is superior. 2207

Discussion 2208

Principal findings 2209

The base-case economic model suggests that calcium acetate, when 2210

compared with calcium carbonate, sevelamer hydrochloride and lanthanum 2211

carbonate, is likely to be the preferred first-line phosphate binder for the 2212



management of hyperphosphataemia in people with CKD stage 5 who are on 2213

dialysis. 2214

When second-line treatment options are taken into account, the most effective 2215

strategies are those commencing with calcium acetate but switching to 2216

sevelamer hydrochloride or lanthanum carbonate if hypercalcaemia develops. 2217

However, remaining on calcium acetate remains a viable option from a health-2218

economic perspective and, in comparison, the switching approaches are 2219

associated with substantially increased costs. Therefore, if all simulated 2220

strategies are seen as realistic options, it is unlikely that the health gains 2221

provided by the more expensive binders are sufficient to counterbalance the 2222

extra expense they entail (unless it can be assumed that society’s maximum 2223

acceptable ICER threshold is approximately £40,000 per QALY). There may 2224

be an exception to this conclusion for people who are unable to tolerate 2225

calcium acetate; in this circumstance, switching to sevelamer hydrochloride or 2226

lanthanum carbonate in those with hypercalcaemia could be seen as an 2227

effective use of resources when compared with the alternative of indefinite 2228

treatment with calcium carbonate. 2229

In patients for whom calcium-based-binders are considered to be 2230

fundamentally contraindicated, the only remaining options in the decision-2231

space are sevelamer hydrochloride and lanthanum carbonate. Under this 2232

circumstance, either option is likely to be judged acceptable, from a health 2233

economic perspective. 2234

Limitations of the analysis 2235

The model has good validity over its first year, accurately reflecting 2236

biochemical measures seen in the trials underpinning it. It also makes a 2237

relatively good prediction of observed survival with the treatments of interest 2238

over the first 3 years of treatment. However, beyond the first year of the 2239

model, biochemical profiles are estimated on the basis of extrapolation and 2240

simplification and, while this approach had its face validity endorsed by the 2241

GDG, it is impossible to tell how well the model fits the (unknown) reality. 2242



Similarly, it is arguable that, as they extend into the future, the biochemical 2243

profiles of occasional simulated patients become implausible (especially as 2244

regards modelled serum calcium, which may rise very high in a few 2245

instances). However, this is only a problem if it results in implausible effects: if 2246

the level of risk that is predicted is realistic, then it does not matter that the 2247

value of the predictor may be unlikely. For this reason, it should be 2248

remembered that the sole purpose of biochemical parameters in the model is 2249

to estimate the risks faced by patients. 2250

The use of serum phosphate and serum calcium alone as determinants of 2251

treatment effect is an acknowledged simplification of a highly complex 2252

biological interaction. Moreover, it is well known that serum calcium is a 2253

suboptimal index of calcium balance in humans, perhaps especially those with 2254

advanced kidney disease (Houillier et al. 2006). If people who are exposed to 2255

excess calcium intake in their phosphate binding regimen are subject to 2256

greater risks than can be inferred from their serum calcium levels, the model 2257

will underestimate the benefit of switching such people to calcium-free 2258

binders. 2259

When simulating second-line treatment for people experiencing 2260

hypercalcaemia, the model is necessarily reliant on evidence of the 2261

effectiveness of treatments in a broader population. If people with 2262

hypercalcaemia respond differently to treatment than those without, it follows 2263

that the model over- or underestimates the treatment effect that can be 2264

achieved in reality; it is possible that different cost–utility conclusions would be 2265

reached if more specific evidence were available. 2266

It is a significant weakness of this analysis that it has not proved 2267

computationally feasible to undertake full probabilistic sensitivity analysis, to 2268

explore the implications of parameter uncertainty for decision-making. A wide 2269

range of one-way sensitivity analyses was undertaken; this enables a fair 2270

degree of inference on the impact of such ‘second-order’ uncertainty and, in 2271

the light of these analyses, it is possible to state with some confidence that the 2272

options identified as optimal in the base-case analysis would be associated 2273

with the highest probability of cost effectiveness in a fully probabilistic 2274



analysis. However, it is not possible to quantify these probabilities in the 2275

absence of such an analysis. 2276

It is possible that the model somewhat underestimates the effectiveness of 2277

lanthanum carbonate in both first- and second-line settings, since it was not 2278

possible to include data on phosphate control from what is by far the largest 2279

and longest trial of the drug in the evidence-base (Finn & the Lanthanum 2280

Study Group, 2006; see 3.5.2 above). It is speculated that, were these data 2281

available to incorporate into analysis, apparent differences between 2282

treatments – perhaps especially sevelamer hydrochloride and lanthanum 2283

carbonate in the first-line setting – would be somewhat attenuated. However, 2284

the use of lanthanum carbonate would still be associated with a very high 2285

ICER, compared with calcium acetate: an additional gain of at least 0.3 2286

QALYs would be necessary to counterbalance the costs estimated here, 2287

assuming conventional cost–utility thresholds, and there is no evidence that 2288

an underestimate of such magnitude is plausible. 2289

It is regrettable that there was insufficient evidence to provide a worthwhile 2290

model for people in CKD stages 4 and 5 who have not yet commenced 2291

dialysis, or for children at any stage of disease. There was also insufficient 2292

evidence to perform any meaningful modelling to support a recommendation 2293

on aluminium hydroxide, magnesium carbonate or sevelamer carbonate. 2294



The GDG discussed the relative importance of the outcomes and agreed that those considered critical or important for decision-making were the same as those in the review of phosphate binder effectiveness in patients with CKD stage 4 or 5 who are not on dialysis. However, the GDG also felt that as the cut off points that defined "phosphate control" or "hypercalcaemia" varied between the studies it was more appropriate to focus upon the continuous measures of serum phosphate and calcium as these would be more objective.


The evidence showed that the phosphate binders examined were all effective in lowering serum phosphate compared with placebo. According to the results of the MTC, no binder was clearly the most effective in terms of impact on serum phosphate levels at the time points considered, although calcium acetate consistently did well. The GDG noted that at 90 days calcium acetate appeared to not be as effective as other binders. However, the GDG considered 90 days to be the minimum follow up time from which decisions could be made,



and therefore felt that more weight should be given to the longer term analyses. A significant number of patients were considered to have achieved serum phosphate control when treated with sevelamer in conjunction with calcium carbonate. There was also a significant number who were considered to have achieved serum phosphate control amongst those treated with magnesium carbonate.

Both sevelamer and lanthanum appear to reduce the risk of death in those older than 65 years.

Non-calcium-based binders were associated with lower serum calcium levels, with sevelamer and magnesium carbonate performing particularly well. Calcium carbonate consistently performed least well. Calcium carbonate was also associated with a greater risk of hypercalcaemia compared with the other calcium-based phosphate binder, calcium acetate. Although the study that demonstrated this was small in size, the GDG believed this to have biological plausibility because absorption of calcium is lower from calcium acetate than from calcium carbonate because of its lower elemental calcium content. Lanthanum appeared to be associated with a lower risk of hypercalcaemia than sevelamer and magnesium carbonate, although these non-calcium-based binders were better than the calcium-based binders.

Sevelamer was better than calcium-based binders in controlling coronary calcification scores. However, when the effectiveness of the 2 calcium-based binders was considered separately, sevelamer was only better than calcium carbonate; there was no statistically significant difference in coronary calcification scores between sevelamer and calcium acetate.

Little evidence showed statistically significant differences relating to the incidence of adverse events. However, when compared against other binders, lanthanum had a reduced risk of experiencing a number of the adverse events of interest: diarrhoea, upper abdominal pain, and nausea and vomiting. Additionally, sevelamer carbonate was associated with a greater risk of nausea and vomiting when compared with sevelamer hydrochloride. There were also concerns regarding the long-term effects of lanthanum, such as its possible deposition in bone, although no evidence was found for this since it was not an agreed outcome of interest.

Because of its effectiveness in lowering serum phosphate, it was felt that calcium acetate would be an effective choice of first-line treatment in adults. However, in patients that cannot tolerate calcium acetate (for example, those who find it difficult to swallow tablets or experience side effects), the GDG felt that calcium carbonate would be an effective alternative first-line phosphate binder.

The GDG also noted that there may be some subgroups that would be at particular risk of the adverse effects of hypercalcaemia and cardiovascular calcification that can result from the use of calcium-based phosphate binders. Using experience and knowledge gained from their own clinical practice, it was noted that these at-risk patients would likely be defined by the presence of vascular calcification, high serum calcium and/or low serum PTH levels. For some of these patients, non-calcium-based binders could be considered, although it was noted that defining thresholds for these biochemical parameters is difficult. In the case of serum calcium and PTH, it was felt that there



is not sufficient evidence or consensus in practice to categorically define what constitutes too high or too low. Clinicians should use their clinical judgement to determine acceptable levels, using indicators such as trends observed across a series of measurements, as well as relative levels of other biochemical markers that are known to impact serum calcium or PTH levels. In the case of PTH, current guidance from the Renal Association states that the target range for dialysis patients is between 2 and 9 times the upper limit of normal for the intact PTH assay used. In terms of vascular calcification, the rare use of tests to determine its extent (particularly for the sole reason of determining vascular calcification in CKD patients being treated for hyperphosphataemia) and the absence of cheap, simple and reliable screening tests for vascular calcification precluded the GDG from including this parameter in their recommendations.

Additionally, although no significant difference in adherence was observed with any of the binders examined, the GDG was concerned that the large size of the calcium acetate tablets used in the UK, together with the unpalatability of calcium acetate and calcium carbonate, may reduce adherence to these phosphate binders in particular, and therefore further limit their suitability for some patients.

In some patients, serum phosphate can remain above the recommended level despite adherence with calcium-based binders. If they have reached the maximum recommended (or tolerated) daily dose of calcium-based binders23 it was felt that no further increases in the dose of calcium-based binder should be made. Instead, a non-calcium-based binder may need to be added to the regimen, producing a combination. The aim would be for the added phosphate-binding capacity to raise phosphate control to the desired level without exceeding the recommended daily intake for elemental calcium.

For patients taking calcium-based binders in whom hypercalcaemia has developed, it was felt that the calcium-based binders should be stopped, and non-calcium-based binders started in their place. Although no direct evidence was found for this ‘sequencing’ of binders, the clinical experience of the GDG and the results of the health economic model constructed for this guideline supported this course of action. The health economic model demonstrated that in patients for whom calcium-based binders are not an option, both sevelamer hydrochloride and lanthanum carbonate would be cost-effective candidates for this switch in adults, although there was insufficient evidence to include other non-calcium-based binders (such as aluminium hydroxide) in the evaluation.

Alternatively, it may be considered appropriate to replace a proportion of the calcium-based binder dose with non-calcium-based binders, producing a lower calcium combination without impairing the control of serum phosphate.

23

The GDG felt it important to explicitly define the upper threshold of the calcium-based binder dose

by both the maximum recommended dose in the BNF and, alternatively, by the actual dose that could

be tolerated by the patient. This resulted from the divergence noted between the BNF's maximum

recommended daily dose of calcium acetate (12 tablets) and the dose usually tolerated by patients in

their own clinical experience (4–6 tablets). This disparity raised concerns among the GDG over the

impact on patients of aiming to dose calcium acetate up to the BNF's upper limit before considering

alternative phosphate binder combinations.



Evidence for the effectiveness of different combinations of binders was limited, addressing only 2 combinations: sevelamer hydrochloride with calcium carbonate (3 studies of 4 to 8 weeks duration), and calcium acetate with magnesium carbonate (1 study of 6 months duration). Both of these combinations seemed to be effective at controlling serum phosphate in patients with normal serum calcium levels, and the binders involved consistently performed well across the effectiveness analyses. The limited nature of the evidence, combined with the lack of evidence for combinations in hypercalcaemic patients, prevented the GDG from selecting one combination over the other, preferring instead to leave the decision to patient preference and clinical judgement.

In children, the one comparison for which evidence was found showed that sevelamer and calcium carbonate are both effective in controlling serum phosphate. Calcium carbonate was also associated with increased serum calcium levels. Children need additional calcium for growing bones and to avoid the effects of secondary hyperparathyroidism that can arise in young patients with chronically low serum calcium. For these reasons, a calcium-based binder would also be desirable as a first-line treatment in children. However, in children who have had a series of calcium measurements that show a trend towards hypercalcaemia, a non-calcium-based binder could help to avoid the adverse effects of prolonged hypercalcaemia.

Additionally, it was felt that no further increases in the dose of calcium-based binder should be made if children taking calcium-based binders remain hyperphosphataemic and have also become hypercalcaemic. Instead, the addition of a non-calcium-based binder to the regimen should be considered in order to achieve phosphate control at the desired level without exceeding the recommended levels for calcium.

Currently, the only non-calcium-based binder licensed for use in children is aluminium hydroxide. However, despite its use being off-label in children, the GDG felt that recommending sevelamer hydrochloride over aluminium hydroxide was appropriate, because they had no evidence for the use of aluminium hydroxide in children. The only paediatric trial found examined the effectiveness of sevelamer against calcium carbonate over an 8-month period, and showed sevelamer to be as effective at lowering serum phosphate and associated with a lower serum calcium level. Furthermore, the GDG had concerns over the toxicity of aluminium, as well as the fact that its licence is not for longer-term indications. The GDG felt that the total replacement of calcium-based binders with a non-calcium-based binder would be inappropriate in children because of their higher calcium requirements.

If considering the use of non-calcium-based binders because of high serum calcium levels in patients on calcium-based binders, the GDG felt it important to emphasise the necessity of reviewing possible causes of high calcium, such as dialysate calcium content, vitamin D, calcium supplements or dietary calcium, before making any changes to the choice of binder. They noted that in some cases it might be easier to make small changes to these sources of elemental calcium than changing (and ensuring adherence to) the phosphate binder regimen. For example, it may be the case that a patient has a high calcium concentration in their dialysate or be on a high dose of vitamin D, and a reduction in one of these may be the most appropriate



course of action.


First-line use of phosphate binders:

A cost–utility model was built based on effectiveness data from a network meta-analysis comparing various phosphate binders. The model suggests that calcium acetate is likely to be the preferred option, as it provides appreciable benefit, when compared with calcium carbonate, at a relatively modest additional cost (ICER of approximately £8000 per QALY gained). While the use of non-calcium-based binders may be associated with some extension of quality-adjusted life expectation compared with first-line calcium acetate, this gain is insufficient to justify the additional costs of the proprietary binders, when judged according to conventional standards (very high ICERs of around £90,000 per QALY gained or greater were estimated).

There is some uncertainty about the validity of the model’s extrapolations beyond the first year of follow-up (it was only possible to inform the first year’s treatment effect directly from the literature). However, the model is expected to be robust to inter-individual variability in parameter values because first-order uncertainty is properly captured in the approach that was adopted.

Second-line use of phosphate binders (sequencing following first-line use of calcium acetate):

The cost–utility model was also used to compare the use of different sequences of phosphate binders, with patients starting on calcium acetate. When their serum calcium level rose above a given level (2.54 mmol/l, in the base case), simulated patients either stayed on calcium acetate or were moved to sevelamer hydrochloride or lanthanum carbonate. The results suggest that, although switching to a non-calcium-based binder in this circumstance is likely to result in additional quality-adjusted life-expectancy, this comes at an additional cost that is unlikely to be judged an acceptable use of NHS resources for patients for whom remaining on calcium acetate is clinically viable (ICERs of approximately £40,000 per QALY or greater).

Therefore, the GDG felt that the model supported the use of non-calcium-based binders only in patients for whom the increased calcium intake associated with calcium-based binder use is considered particularly harmful. In patients with hypercalcaemia or low serum PTH, or in patients unable to tolerate calcium-based binders, the GDG felt it was appropriate to remove the calcium-based binders from the decision space – that is, to regard the use of binders that increase calcium load fundamentally contraindicated. The remaining second-line phosphate binders for consideration within the model – sevelamer hydrochloride and lanthanum carbonate – both appear to be cost-effective in patients for whom calcium-based binders are contraindicated.

Although the GDG noted that combinations of different phosphate binders may be used in practice, the model was not able to explore the cost-effectiveness of this approach, due to an absence of evidence on the effectiveness of such combinations.

Quality of evidence

Following the application of GRADE, overall evidence quality was low or very low across the outcomes.

Only one study examining one comparison (sevelamer compared with



calcium carbonate) was found for age groups other than adults.

Although some evidence for a small number of combinations was found, the evidence did not reflect the wide range of combinations commonly used in UK practice. Furthermore, the combination of sevelamer and calcium carbonate (examined in a number of included studies) would not generally be used.

In making recommendations on the use of combinations of calcium-based and non-calcium-based binders in patients who are above the upper range of normal through calcium-based binder monotherapy, the GDG noted the lack of evidence for combinations in hypercalcaemic patients.

Although magnesium carbonate appeared to perform well in the control of serum phosphate and calcium, the GDG raised concerns about the possible adverse effects – such as diarrhoea – that might be associated with it. The possible adverse effects were not extracted from the single trial examining the effectiveness of magnesium carbonate. This, in conjunction with the small sample size of the trial found, meant that the GDG did not feel the evidence was sufficient to make any recommendations for the use of magnesium carbonate.

No evidence was included on the toxicity of lanthanum, despite concerns, such as its possible deposition in the bones.

Only one paper was found that examined the effectiveness of aluminium hydroxide (used as a comparator for sevelamer). The paper reported data for serum phosphate, the incidence of constipation as an adverse event and serum calcium, although the short 8-week follow-up period and the small sample size meant that the GDG did not feel confident in making recommendations on its use.

A number of different co-treatments were used across the studies. These included ‘rescue’ binders (used when serum phosphate levels became too high), vitamin D, cinacalcet, and phosphate- and/or protein-restricted diets. These concurrent treatments were also considered potential confounders. Additionally, the GDG noted that there were variations, where reported, in the length of time patients were on dialysis prior to starting the trials.

Concerns over the timings of follow-up measurements were also raised. Firstly, the time points used in the MTCs – 90 days, 180 days and 360 days – suffered from approximately 10 days’ variation above or below the stated time point. This primarily resulted from differences in the measures used to define time in each study (days versus weeks versus months). Furthermore, the studies that formed the MTC networks often entailed different follow-up intervals. This meant that the networks changed over time, both in terms of the comparisons covered and the studies involved. It was felt that this may have contributed to the variations in binder ranking over the different time periods analysed.

For serum calcium, lanthanum could not be included in the MTC until 360 days because of the absence of a ‘connector’. When it does enter the network, it was felt that lanthanum did not perform as well as expected. It was suggested that this might have resulted from its connection to the network by ‘any other binder’, which performs particularly poorly, potentially pulling down lanthanum’s relative performance to the other interventions. Additionally, the reviewer was unable to extract serum phosphate data from 1 study comparing



lanthanum carbonate against ‘standard therapy’ due to its presentation in uninterpretable graphs. It was felt that this study was a significant part of the evidence base due to its long follow-up of 2 years and large population of 1359 participants. However, despite contacting the authors of the study, the reviewer was unable to obtain the data required to include this study in the MTCs for serum phosphate or in the health economic model.

A variety of thresholds were used to define dichotomous outcomes, although the thresholds were sufficiently similar for relative comparisons to still be considered useful. Significant variation was observed in the definitions used for hypercalcaemia in particular, although the thresholds used to define a patient’s achievement of phosphate control also differed in a number of papers.

The GDG felt that the sub-group analysis conducted on patients with a baseline coronary calcification score of more than 30 might not have been a clinically relevant distinction.

Reporting in many of the studies was poor. For example, details of study designs were often unclear and results were not always cited in the text of the papers, requiring the reviewer to read the data off available graphs. Unit-of-analysis errors were also common, with analyses not following the intent-to-treat principle.


Patients on nocturnal, long-hours dialysis may have sufficient phosphate removal to preclude the need for intensive phosphate management interventions. However, their phosphate status should still be closely monitored. Some patients may need interventions to increase their phosphate levels if hypophosphataemia occurs.




use of phosphate binders in adults with stage 5 CKD who 2297

are on dialysis 2298



For children and young people, offer a calcium-based phosphate binder as the

first-line phosphate binder to control serum phosphate in addition to dietary

management.


For children and young people, if a series of serum calcium measurements

shows a trend towards the age-adjusted upper limit of normal, consider a

calcium-based binder in combination with a non-calcium-based binder, having

taken into account other causes of rising calcium levels.


For children and young people who remain hyperphosphataemic despite

adherence with a calcium-based phosphate binder, and whose serum calcium

goes above the age-adjusted upper limit of normal, consider a combination of

a calcium-based phosphate binder and sevelamer hydrochloride24, having

taken into account other causes of raised calcium.


For adults, offer calcium acetate as the first-line phosphate binder to control

serum phosphate in addition to dietary management.


For adults, consider calcium carbonate if calcium acetate is not tolerated or

24

At the time of consultation (October 2012), sevelamer hydrochloride did not have a UK marketing









patients find it unpalatable.


For adults with stage 5 CKD who are on dialysis and remain

hyperphosphataemic despite adherence to the maximum recommended or

tolerated dose of calcium-based phosphate binders, consider either combining

with, or switching to, a non-calcium-based binder.


For adults with stage 5 CKD who are on dialysis and who are taking a

calcium-based binder, if serum phosphate is controlled by the current diet and

phosphate binder regimen but:

serum calcium goes above the upper limit of normal, or

serum parathyroid hormone levels are low,

consider switching the patient to either sevelamer hydrochloride or lanthanum

carbonate, having taken into account other causes of raised calcium.


If a combination of phosphate binders is used, titrate the dosage to achieve

control of serum phosphate while taking into account the effect of any

calcium-based binders used on serum calcium levels (also see

recommendations 1.1.6, 1.1.7 and 1.1.10 to 1.1.12).


Use clinical judgement and take into account patient preference when

choosing whether to use a non-calcium-based binder as monotherapy or in

combination with a calcium-based binder (also see recommendations 1.1.10

to 1.1.12).

2300






In adults with stage 4 or 5 CKD, including those on dialysis, what is the long-

term effectiveness and safety of aluminium hydroxide in controlling serum

phosphate?


In adults and children with stage 4 or 5 CKD, including those on dialysis, what

is the long-term effectiveness and safety of magnesium carbonate in

controlling serum phosphate?




phosphate?

2303

3.6 Use of supplements in people with stage 4 or 5 CKD 2304



For people with stage 4 or 5 CKD who are not on dialysis, are prescribed 2307

supplements, alone or in conjunction with other interventions, effective 2308

compared to placebo or other treatments in managing serum phosphate and 2309

its associated outcomes? Which are the most effective supplements? 2310


This review question focused on the use of supplements in the prevention and 2312

treatment of hyperphosphataemia in patients with stage 4 or 5 CKD who are 2313

not on dialysis. These supplements could consist of a variety of vitamins and 2314

minerals. 2315






broad search strategy, pulling in all papers relating to the management of 2319

hyperphosphataemia in CKD using supplements . Only RCTs that compared a 2320

supplementation intervention with either a placebo or another comparator in 2321

patients with stage 4 or 5 CKD who are not on dialysis were considered for 2322

inclusion. 2323



the population included people on dialysis or 2326

supplementation was intended for any reason other than for the 2327

management of hyperphosphataemia. 2328

2329

From a database of 2020 abstracts, 20 full-text articles were ordered, though 2330

no papers were found to be suitable for inclusion in the analysis. 2331

Supplementation using vitamin D and its metabolites was also not considered 2332

for this review and was excluded at the scoping stage. It was felt that their 2333

effectiveness is not disputed and they are already an accepted and cost-2334

effective part of standard clinical practice. 2335


3.6.3.1 No evidence on the clinical effectiveness of supplements in the 2337

management of hyperphosphataemia in people with CKD stages 4 2338

or 5 who are not on dialysis was identified. 2339



The GDG discussed the relative importance of the outcomes and agreed that those considered critical or important for decision-making were the same as those in the reviews of phosphate binder effectiveness.


No evidence was found to examine the benefits and harms of using supplements in the management of hyperphosphataemia in people with stage 4 or 5 CKD who are not on dialysis.

However, the GDG felt that the evidence found in patients with CKD stage 5 who are on dialysis could be extrapolated to this population because it is likely that the efficacy of the intervention would be similar, regardless of whether the person was on dialysis or not. See ‘3.7 Use of supplements in people with stage 5 CKD who are on



dialysis’ for the review of this evidence.


Because the GDG did not feel that the available evidence supported the use of supplements, it was not necessary to conduct a cost-effectiveness analysis for this question.

Quality of evidence

No evidence was identified for the use of supplements in the management of hyperphosphataemia in people with stage 4 or 5 CKD who are not on dialysis.


2341

3.7 Use of supplements in people with stage 5 CKD who 2342

are on dialysis 2343


For people with stage 5 CKD who are on dialysis, are prescribed 2345

supplements, alone or in conjunction with other interventions, effective 2346

compared to placebo or other treatments in managing serum phosphate and 2347

its associated outcomes? Which is the most effective prescribed supplement? 2348


This review question focused on the use of supplements in the prevention and 2350

treatment of hyperphosphataemia in patients with stage 5 CKD who are on 2351

dialysis. These supplements could consist of a variety of vitamins and 2352

minerals, though in the evidence located included: nicotinamide, calcium and 2353

L-carnitine supplementation. 2354




broad search strategy, pulling in all papers relating to the management of 2358

hyperphosphataemia in CKD using supplements. Only RCTs that compared a 2359

supplement with either a placebo or another comparator in patients with stage 2360

5 CKD who are on dialysis were considered for inclusion. 2361





the population included people with CKD stage 5 who were not on dialysis 2364

or 2365

supplementation was intended for any reason other than for the 2366

management of hyperphosphataemia. 2367

2368

From a database of 2020 abstracts, 20 full-text articles were ordered and 2369

5 papers describing 5 primary studies were selected (Young et al, 2009; 2370

Shahbazian et al, 2011; Rudnicki et al, 1993; Chertow et al, 1999; Cibulka et 2371

al, 2007). No paediatric studies meeting the inclusion criteria were found. 2372

Table 9 lists the details of the included studies. 2373

Supplementation using vitamin D and its metabolites was also not considered 2374

for this review and was excluded at the scoping stage. It was felt that their 2375

effectiveness is not disputed and they are already an accepted and cost-2376

effective part of standard clinical practice. 2377

There was pooling of the 2 studies comparing nicotinamide supplementation 2378

against placebo, although there was some heterogeneity present. This 2379

heterogeneity resulted from the use of different concurrent interventions that 2380

are known to impact the outcomes of interest, but also the different types of 2381

dialysis used. 2382

Where meta-analysis was possible, a forest plot is also presented. Mean 2383

differences (MDs) were calculated for continuous outcomes and odds ratios 2384

(ORs) for binary outcomes, as well as the corresponding 95% confidence 2385

intervals where sufficient data were available. 2386



Table 9 Summary of included studies for the use of supplements in people with stage 5 CKD who are on dialysis 2387


Nicotinamide compared with placebo

Young et al, 2009

RCT

Missouri, US

Randomised = 17 (intervention = 8; control = 9)

Analysed = 14 (intervention = 7; control = 7)

Age > 18 years

On peritoneal dialysis for > 3 months

Dose of phosphate binder(s) stable over the previous 2 weeks

Plasma phosphate > 4.9 mg/dl (i.e. 1.6 mmol/l) based on the most recent laboratory data within 1 month of enrolment (note: this threshold is within recommended phosphate levels, though at the upper end)

Patients with phosphate values > 3.9 mg/dl meeting all other inclusion criteria were eligible, if consenting, for a reduction in the current phosphate binder dose followed by repeat screening within 2–4 weeks; they were then eligible for continued participation if the repeat phosphate value exceeded 4.9 mg/dl





Nicotinamide (250 mg per capsule) and placebo were packaged as identically appearing capsules

Dosing increased throughout study period: study medication or placebo was started at 250 mg twice daily, increased to 500 mg twice daily after 2 weeks and to 75 0 mg twice daily after 4 weeks

Changes in phosphate binder dose were not allowed except if phosphate values exceeded 6.5 mg/dl (i.e. 2.1 mmol/l) or fell below 3 mg/dl (i.e. 1.0 mmol/l)

Active vitamin D and cinacalcet doses were required to remain stable - i.e. followed pre-study protocol

Placebo packaged as identically appearing capsules

Dosing increased throughout study period: study medication or placebo was started at 250 mg twice daily, increased to 500 mg twice daily after 2 weeks and to 750 mg twice daily after 4 weeks

Changes in phosphate binder dose were not allowed except if phosphate values exceeded 6.5 mg/dl (i.e. 2.1 mmol/l) or fell below 3 mg/dl (i.e. 1.0 mmol/l)

Active vitamin D and cinacalcet doses were required to remain stable - i.e. followed pre-study protocol

8 weeks

Serum phosphate and corrected calcium measured every 2 weeks

Incidence of adverse events (diarrhoea) recorded for the 8-week study period

Shahbazian et al, 2011

RCT

Ahvaz, Iran


Analysed = 48

Age ≥ 18 years

Fasting serum phosphate ≥ 5 mg/dl (i.e. ≥ 1.6 mmol/l) (note: this threshold is within recommended phosphate levels, though at the very upper end)

Maintaining on haemodialysis for more than 2 months

Constant dosage of phosphate binders during past 2

Immediate release nicotinamide tablets (500 mg)

During the first 4 weeks, nicotinamide was administered 500 mg/day, and then it was administered 1,000 mg/day in weeks 5 to 8.

In case of adverse effects, phosphate level ≤ 3.5 (i.e. 1.1 mmol/l) or > 8 mg/dl (i.e. 2.6 mmol/l), thrombocytopenia

Placebo tablets

Usual doses of calcium carbonate - i.e. followed pre-study protocol (note: unclear if calcium carbonate taken as a supplement or a phosphate binder)

8 weeks

Serum phosphate and corrected calcium measured every 4 weeks

Incidence of



weeks

Patients had residual renal function of < 10 ml/min

Dialysate concentration of calcium was similar for all patients (in these centres, only 1 kind of dialysate is used)





(platelet counts less than 150,000/mm

3), or clinical findings of

low platelet count, the research committee decided on adjusting the dose or other necessary measures

Usual doses of calcium carbonate - i.e. followed pre-study protocol (note: unclear if calcium carbonate taken as a supplement or a phosphate binder)

adverse events (diarrhoea) recorded for the 8-week study period

Sevelamer hydrochloride + calcium supplementation compared with sevelamer hydrochloride alone

Chertow et al, 1999

RCT

US


Analysed = 71

Age ≥ 18 years

Thrice weekly haemodialysis for at least 3 months

Regular calcium- and/or aluminium-based phosphate binders, with or without vitamin D metabolite replacement therapy at stable doses for at least 1 month before screening; all pre-study phosphate binders discontinued 2 weeks before treatment phase

Participants whose serum phosphate concentration rose to 6.0 mg/dl (i.e. 1.9 mmol/l) or above after 2-week phosphate binder washout period were entered into treatment phase; participants whose serum phosphate rose to above 12 mg/dl (i.e. 3.9 mmol/l) were entered immediately into the treatment phase without necessarily completing the 2-week washout





900 mg of elemental calcium taken in the form of calcium carbonate once nightly on a on an empty stomach

Initial dose of sevelamer hydrochloride (RenaGel) was determined by the highest level of serum phosphate during the 2-week phosphate binder washout period: 2 x 465 mg capsules 3 times daily with meals for serum phosphate concentrations > 6.0 mg/dl (i.e. 1.9 mmol/l) and < 7.5 mg/dl (i.e. 2.4 mmol/l); 3 x 465 mg capsules 3 times daily with meals for serum phosphate concentrations > 7.5 mg/dl (i.e. 2.4 mmol/l) and < 9.0 mg/dl (i.e. 2.9 mmol/l); 4 x 465 mg capsules 3 times daily with meals for serum phosphate concentrations > 9.0 mg/dl (i.e. 2.9 mmol/l)

Sevelamer hydrochloride doses were titrated up and down by 3 capsules per day every 3 weeks (at 3, 6 and 9 weeks after treatment commencement), with the goal of achieving serum phosphate concentrations between 2.5 (i.e. 0.6 mmol/l) and 5.5 mg/dl (i.e.

Initial dose of sevelamer hydrochloride (RenaGel) was determined by the highest level of serum phosphate during the 2-week phosphate binder washout period: 2 x 465 mg capsules 3 times daily with meals for serum phosphate concentrations > 6.0 mg/dl (i.e. 1.9 mmol/l) and < 7.5 mg/dl (i.e. 2.4 mmol/l); 3 x 465 mg capsules 3 times daily with meals for serum phosphate concentrations > 7.5 mg/dl (i.e. 2.4 mmol/l) and < 9.0 mg/dl (i.e. 2.9 mmol/l); 4 x 465 mg capsules 3 times daily with meals for serum phosphate concentrations > 9.0 mg/dl (i.e. 2.9 mmol/l)

Sevelamer hydrochloride doses were titrated up and down by 3 capsules per day every 3 weeks (at 3, 6 and 9 weeks after treatment commencement), with the goal of achieving serum phosphate concentrations between 2.5 (i.e. 0.6 mmol/l) and 5.5 mg/dl (i.e. 1.8 mmol/l)

Participants were asked to maintain their usual eating habits without any intentional changes in dietary

12 weeks

Incidence of mortality and adverse events (vomiting, nausea, constipation, and diarrhoea) recorded for the 12-week study period

Serum phosphate and calcium measured weekly



2388

1.8 mmol/l)

Participants were asked to maintain their usual eating habits without any intentional changes in dietary restrictions

Participants were not permitted to take any aluminium-based products, or any additional calcium-based products

Vitamin D metabolite doses were maintained at baseline levels except in the event of significant hypo- or hypercalcaemia

restrictions

Participants were not permitted to take any calcium- or aluminium-based products

Vitamin D metabolite doses were maintained at baseline levels except in the event of significant hypo- or hypercalcaemia

L-carnitine supplementation compared with placebo

Cibulka et al, 2007

RCT

Czech Republic

Randomised = 112

Analysed = 83 (intervention = 44; control = 39)

Criteria for randomisation:

Regular haemodialysis – 4 hours 3 times weekly

All patients had a GFR < 0.2 ml/sec (i.e. < 12 ml/min)

Baseline serum phosphate (median (range)):

Intervention = 1.74 mmol/l (0.71 – 3.71)

Control = 1.71 mmol/l (1.02 – 3.50)Exclusion criteria (for those who were randomised but who were not included in the analysis): kidney transplant; non-adherence; change of residence; restitution of kidney function; change in vitamin D dose during trial

15 mg of L-carnitine/kg of body weight in a short intravenous infusion after each haemodialysis session (i.e. 3 times weekly)

All patients received calcium carbonate phosphate binder therapy

Patients continued pre-study vitamin D treatment

Isotonic solution of sodium chloride in a short intravenous infusion after each haemodialysis session (i.e. 3 times weekly)

All patients received calcium carbonate phosphate binder therapy

Patients continued pre-study vitamin D treatment

6 months

Incidence of mortality recorded for the 6-month study period

Serum phosphate and calcium measured every 3 months

Abbreviations: Ca, calcium ions; GFR, glomerular filtration rate; HIV, human immunodeficiency virus; RCT, randomised controlled trial; SD, standard deviation.



Summary GRADE profile 52 Niacinamide compared with placebo 2389



Niacinamide1

Placebo1


8-week follow-up

2 RCTs


Young et al, 2009


MD = 0.30 mmol/l lower (95% CI:0.42 to 0.17 lower)

Very low

Change in serum phosphate

8-week follow-up

1 RCT

Young et al, 2009

7 7 Absolute effect


Very low

Serum phosphate

(number of patients to experience an increase in serum phosphate from baseline to the end of the study period)

8-week follow-up

1 RCT

Young et al, 2009

1/7

(14.3%)

5/7

(71.4%)

Relative effect

OR = 0.07 (95% CI: 0.01 to 0.97)

Absolute effect

57 fewer per 100 (1 fewer to 69 fewer)

Very low


(number of patients to experience diarrhoea)

8-week follow-up

2 RCTs


Young et al, 2009

5/32

(15.6%)

0/33

(0%)

Relative effect

OR = 6.78 (95% CI: 0.73 to 62.69)

Very low

Serum calcium

8-week follow-up

1 RCT

Young et al, 2009

7 7 Absolute effect


Very low

1 The 2 studies had different co-treatments, received by both the intervention and control groups:

Shahbazian et al, 2011: pre-study calcium carbonate regimen (note: unclear if calcium carbonate was used as a binder or as a supplement)

Young et al, 2009: pre-study phosphate binder, active vitamin D and cinacalcet regimens


2390



Summary GRADE profile 53 Sevelamer hydrochloride + calcium supplementation (+ pre-study vitamin D regimen) 2391

compared with sevelamer hydrochloride (+ pre-study vitamin D regimen) 2392



Sevelamer hydrochloride + calcium supplementation

(+ pre-study vitamin D regimen)


(+ pre-study vitamin D regimen)

Mortality

(number of deaths among patients hospitalised during study)

12-week follow-up

1 RCT

Chertow et al, 1999

2/36

(5.6%)

0/35

(0%)

Relative effect

OR = 0.21 (95% CI: 0.01 to 4.44)

Very low

Serum phosphate

12-week follow-up

1 RCT

Chertow et al, 1999



Very low

Change in serum phosphate

12-week follow-up

1 RCT

Chertow et al, 1999



Very low

Serum phosphate

(number of patients to achieve a therapeutic response - serum phosphate at or below washout level or <1.78mmol/l)

12-week follow-up

1 RCT

Chertow et al, 1999

35/36

(97.2%)

33/35

(94.3%)

Relative effect

OR = 1.03 (95% CI: 0.14 to 7.75)

Absolute effect


Very low

Change in serum calcium

12-week follow-up

1 RCT

Chertow et al, 1999


MD = 0.08 mmol/l higher (95% CI:0.01 lower to 0.16 higher)

Very low

Serum calcium

(% of patients to experience hypercalcaemia)

12-week follow-up

1 RCT

Chertow et al, 1999

21.1% 2.4% Relative effect

OR = 10.88 (95% CI: 2.77 to 42.70)

Absolute effect

19 more per 100 ( 4 to 49 more)

Very low

Note: the 2 groups were not comparable in terms of their serum phosphate at the start of the treatment period (the mean difference was statistically significant at -0.24 mmol/l (95% CI -0.29 to -0.19)), nor their serum calcium x phosphorus product (both were higher in the control group). It also appears that there was a greater prevalence of



hypercalcaemia in the intervention group at the start of treatment. Additionally, vitamin D use different at baseline, but the difference was not statistically significant (intervention group = 60% use vitamin D versus 42% in control group [p = 0.16])


2393

Summary GRADE profile 54 L-carnitine supplementation (+ calcium carbonate + pre-study vitamin D regimen) compared 2394

with placebo (+ calcium carbonate + pre-study vitamin D regimen) 2395



L-carnitine supplementation

(+ calcium carbonate + pre-study vitamin D regimen)

Placebo

(+ calcium carbonate + pre-study vitamin D regimen)

Mortality

(number of deaths among patients)

6-month follow-up

1 RCT

Cibulka et al, 2007


OR = 0.63 (95% CI: 0.21 to 1.89)

Absolute effect


Very low

Serum phosphate

6-month follow-up

1 RCT

Cibulka et al, 2007


Difference in medians = 0.08 mmol/l lower

Very low

Serum calcium

6-month follow-up

1 RCT

Cibulka et al, 2007


Difference in medians = 0.04 mmol/l lower

Very low

Abbreviations: CI, confidence interval; OR, odds ratio; RCT, randomised controlled trial.

2396






Use of supplements in people with stage 5 CKD who are on dialysis 2400

Niacinamide compared with placebo 2401


3.7.3.1 Very-low-quality evidence from 2 RCTs of 62 participants in total 2403

showed nicotinamide supplementation to be associated with a 2404


-0.17 i.e. statistically significant) than that associated with placebo 2406

at 8 weeks. 2407

3.7.3.2 Very-low-quality evidence from 1 RCT of 14 participants showed 2408

that at 8 weeks’ nicotinamide supplementation was associated with 2409

a mean decrease in serum phosphate levels of 0.23 mmol/l (SD 2410

0.29) and placebo with a mean increase in serum phosphate levels 2411

of 0.13 mmol/l (SD 0.26) (MD 0.36 mmol/l lower [95% CI -0.64 to 2412

-0.07 i.e. statistically significant]). 2413


that a smaller proportion of participants that received nicotinamide 2415

supplementation experienced an increase in serum phosphate from 2416

baseline to the end of the 8-week study period than among those 2417

that received placebo (OR 0.07 [95% CI 0.01 to 0.97 i.e. 2418



3.7.3.4 Very-low-quality evidence from 2 RCTs of 65 participants showed 2421

no statistically significant difference in the proportion of participants 2422

that experienced diarrhoea during the 8-week study period, 2423

between those that received nicotinamide supplementation and 2424

those that received placebo (OR 6.78 [95% CI 0.73 to 62.69]). 2425




3.7.3.5 Very-low-quality evidence from 1 RCT of 14 participants showed no 2426

statistically significant difference in mean serum corrected calcium 2427

level between those that received nicotinamide supplementation 2428

and those that received placebo at 8 weeks (MD 0.05 mmol/l lower 2429

[95% CI -0.21 to 0.11]). 2430

Sevelamer hydrochloride + calcium supplementation compared with 2431

sevelamer hydrochloride 2432



statistically significant difference in the number of deaths during the 2435

12-week study period between those that received sevelamer 2436

hydrochloride and calcium supplementation and those that received 2437

sevelamer hydrochloride alone (OR 0.21 [95% CI 0.01 to 4.44]). 2438


sevelamer hydrochloride and calcium supplementation to be 2440

associated with a mean serum phosphate level 0.31 mmol/l lower 2441

(95% CI -0.59 to -0.04 i.e. statistically significant) than that 2442

associated with sevelamer hydrochloride alone at 6 months25. 2443


statistically significant difference in the mean change in serum 2445

phosphate levels during the 6-month study period between those 2446

that received sevelamer hydrochloride and calcium 2447

supplementation and those that received sevelamer hydrochloride 2448

alone (MD 0.03 mmol/l lower [95% CI -1.83 to 1.77]). 2449


statistically significant difference in the number of patients to 2451

achieve a therapeutic response (defined as serum phosphate at or 2452

25

Note: the difference in mean serum phosphate at treatment initiation was statistically significant

between the 2 groups, with a mean in the group taking sevelamer hydrochloride with supplemental

calcium of 2.51 mmol/l and of 2.75 mmol/l in the group taking sevelamer hydrochloride alone (MD -

0.24 mmol/l (95% CI -0.29 to -0.19)).



below the participant’s pre-binder washout level or < 1.78 mmol/l) 2453

during the 12-week study period between those that received 2454

sevelamer hydrochloride and calcium supplementation and those 2455

that received sevelamer hydrochloride alone (OR 1.03 [95% CI 2456

0.14 to 7.75]). 2457



statistically significant difference in the mean change in serum 2460

calcium levels during the 6-month study period between those that 2461

received sevelamer hydrochloride and calcium supplementation 2462

and those that received sevelamer hydrochloride alone (MD 2463

0.08 mmol/l higher [95% CI -0.01 to 0.16]). 2464


that a greater proportion of participants that received sevelamer 2466

hydrochloride and calcium supplementation experienced 2467

hypercalcaemia during the 12-week study period than among those 2468

that received sevelamer hydrochloride alone (OR 10.88 [95% CI 2469

2.77 to 42.70 i.e. statistically significant]). 2470

L-carnitine supplementation compared with placebo 2471



statistically significant difference in the number of deaths during the 2474

6-month study period between those that received L-carnitine 2475

supplementation and those that received placebo (OR 0.63 [95% 2476

CI 0.21 to 1.89]). 2477


received L-carnitine supplementation to be associated with a 2479

median serum phosphate level 0.08 mmol/l lower than that 2480

associated with placebo. 2481



2482



L-carnitine supplementation to be associated with a median serum 2485

calcium level 0.04 mmol/l lower than that associated with placebo. 2486





From the 2 studies identified, the GDG noted that nicotinamide appears to be effective in lowering serum phosphate levels. Fewer patients experienced an increase in these levels when taking nicotinamide, instead achieving a greater mean decrease in serum phosphate over the study period and lower mean levels at the endpoint. It also appeared that the incidence of diarrhoea was not significantly different between those taking the supplement and those taking placebo. However, the GDG felt that the study periods involved (both studies lasted for 8 weeks) and the sizes of the samples used (14 and 48 patients) were not sufficient to detect a significant difference in the risk of experiencing side effects. The GDG was concerned that the side effects of nicotinamide could be comparable to those associated with nicotinic acid, including: nausea, vomiting, abdominal pain and dyspepsia; flushing; pruritus and rash. No evidence was found on these adverse effects. Additionally, no data were found concerning the mortality, adherence or quality of life associated with this supplement.

One study was found to examine the effectiveness of calcium supplementation in controlling serum phosphate. The comparison was made between sevelamer hydrochloride supplemented with calcium carbonate and sevelamer hydrochloride alone. The GDG distinguished calcium carbonate’s use as a supplement from its use as a binder by the timing of its consumption; calcium carbonate supplementation is taken on an empty stomach and calcium carbonate binders are taken with meals.

The addition of supplemental calcium to sevelamer hydrochloride was associated with a significantly lower mean serum phosphate level at the study’s end compared to sevelamer hydrochloride alone. However, the GDG concluded that this was not due to the intervention’s greater effectiveness in controlling serum phosphate because the intervention group had a significantly lower mean serum phosphate level at the start of treatment. This view was further supported by the finding that there was not a significant difference between the mean changes in serum phosphate of the 2 groups over the course of the study.

The GDG noted that the use of additional calcium supplementation did, however, result in an increase in hypercalcaemic events and an



almost statistically significant increase in mean serum calcium. The GDG felt that there were considerable concerns over when calcium carbonate should and should not be given, and how this might link to the binders that an individual may or may not be given. The available evidence only supports the use of calcium carbonate as a phosphate binder (reviewed in section 3.5), taken with food not on an empty stomach. The GDG noted that the need to take any phosphate binder with a meal, as opposed to as a supplement on an empty stomach, was an important step in ensuring therapeutic success.

The GDG felt that the evidence outlined above could be extrapolated to those with CKD stages 4 and 5 who are not on dialysis and to children.


Because the GDG did not feel that the available evidence supported the use of supplements, it was not necessary to conduct a cost-effectiveness analysis for this question.

Quality of evidence

No data were available for age groups other than adults.

No evidence was found relating to cardiovascular calcification scores, adherence and quality of life, and the GDG also noted the short length of follow-up in the trials that studied mortality.

A range of co-treatments were used across the studies, and concerns were noted over their potential influence as confounders. In particular, heterogeneity was noted for the two studies pooled in order to assess the effectiveness of nicotinamide. The co-treatments in one paper consisted of the patients’ pre-study phosphate binder, vitamin D and cinacalcet regimens, and the patients’ pre-study calcium carbonate regimen in the other. Additionally, the types of dialysis used were different in the two papers: peritoneal dialysis and maintenance haemodialysis, respectively.

Reporting in many of the studies was poor. For example, details of study designs were often unclear and results were not always cited in the text of the papers, requiring the reviewer to read the data off available graphs. Unit-of-analysis errors were also common, with analyses not following the intent-to-treat principle.


The use of supplementation in the management of hyperphosphataemia could add further to the treatment burden already experienced by patients on a demanding regimen.

2488




use of supplements in people with stage 5 CKD who are 2490

on dialysis 2491



Advise patients to take phosphate binders with food in order to control serum

phosphate.

2493

3.8 Sequencing of treatments 2494


In the management of hyperphosphataemia in people with stage 4 or 5 CKD, 2496

in what order should available treatments be considered? What are the clinical 2497

indications for commencing each? 2498


This review question focused on the sequencing of interventions in the 2500

prevention and treatment of hyperphosphataemia in patients with stage 4 or 5 2501

CKD, including those on dialysis. These interventions could include dietary 2502

interventions, phosphate binders, supplements including vitamin D, and 2503

dialysis. 2504



Systematic Reviews, the Health Technology Assessment Database (HTA) 2507

and the Database of Abstracts of Reviews of Effects (DARE). A broad search 2508

strategy was used, pulling in all papers relating to the sequencing of 2509

interventions in the management of hyperphosphataemia in patients with 2510

CKD. RCTs, quasi-RCTs, systematic reviews, non-randomised controlled 2511

trials, cohort studies, cross-sectional studies and case-control studies were 2512

eligible for inclusion. 2513




the population included people with stages 1 to 3 CKD or 2515

they did not compare a sequence of interventions with another sequence of 2516

interventions. 2517

2518

From a database of 1868 abstracts, 16 full-text articles were ordered, 2519

although no papers were found to be suitable for inclusion in the analysis. 2520

2521


3.8.3.1 No evidence was identified on the sequencing of interventions in 2523

order to maximise the management of hyperphosphataemia in 2524

people with stage 4 or 5 CKD, nor in people with stage 5 CKD who 2525

are on dialysis. 2526

3.8.4 Health economic modelling 2527

Health economic analysis within this guideline focused on evaluation of 2528

phosphate binders for people with stage 5 CKD who are on dialysis. These 2529

results assisted the GDG in making recommendations regarding the 2530

sequencing of phosphate binders. See section 3.5.4 and appendix F for full 2531

details on the health economic analysis and modelling carried out for the 2532

guideline. 2533





Although no evidence was found concerning the relative effectiveness of different treatment sequences, the GDG reached a consensus regarding best practice, using their clinical knowledge and experience. The GDG felt strongly that the two key interventions for the management of hyperphosphataemia in people with advanced CKD are, as first-line, dietary management and, as second-line, phosphate binders. They also emphasised the cyclical nature of this sequence, stressing that clinicians should continue to periodically review the dietary and binder regimens throughout the treatment pathway.

The GDG noted that dialysis efficacy and vitamin D and its analogues



are known to influence serum phosphate control. However, it was felt that these are not part of the primary treatment sequence of hyperphosphataemia, rather they fall in parallel. By this the GDG meant that, while they impact serum phosphate, these interventions would be brought in to the treatment pathway for reasons other than hyperphosphataemia: for example, vitamin D for management of serum calcium, serum PTH and/or hyperparathyroidism, and dialysis for severe decline in renal function. For this reason, the GDG did not feel it appropriate to specify the clinical conditions in which vitamin D or dialysis should be initiated within the treatment of hyperphosphataemia. However, acknowledging their impact upon serum phosphate (and other relevant biochemical parameters, such as serum calcium), they noted that diet and binder prescriptions should be reviewed when vitamin D or dialysis regimens are initiated or adjusted.

GDG discussions regarding the sequencing of phosphate binders were included in earlier reviews (see sections 3.4 and 3.5).


No health economic evidence was available to inform the GDG’s consideration of the optimal sequence of different modes of treatment. However, the de novo health economic model developed to examine the cost–utility of phosphate binders provided evidence on the cost-effectiveness of different sequences of binders (see section 3.5.4). It should be noted that, in the GDG’s experience, different phosphate binders may be used in combination, rather than wholesale switching from one to another. The model was not able to explore the cost-effectiveness of different combinations of phosphate binders due to a complete absence of evidence on the effectiveness of such combinations.

Quality of evidence

No evidence was found to examine the effectiveness of different treatment sequences for the management of hyperphosphataemia in people with stage 4 or 5 CKD, including those on dialysis.


2535


sequencing of treatments 2537



If vitamin D or dialysis is introduced, review the patient’s diet and phosphate

binder requirement in order to maintain the control of serum phosphate.

2539








phosphate?

2542

4 Notes on the scope of the guideline 2543

NICE guidelines are developed in accordance with a scope that defines what 2544

the guideline will and will not cover. The scope of this guideline is given in 2545

appendix C. 2546

5 Implementation 2547

NICE has developed tools to help organisations implement this guidance. 2548

Note: these details will apply when the guideline is published. 2549

6 Other versions of this guideline 2550

6.1 NICE pathway 2551

The recommendations from this guideline have been incorporated into a NICE 2552

pathway. Note: these details will apply when the guideline is published. 2553

6.2 ‘Understanding NICE guidance’ 2554

A summary for patients and carers (‘Understanding NICE guidance’) is 2555

available. 2556

For printed copies, phone NICE publications on 0845 003 7783 or email 2557

[email protected] (quote reference number N[xxxx]). Note: these 2558

details will apply when the guideline is published. 2559

http://guidance.nice.org.uk/CGnn

http://pathways.nice.org.uk/pathways/xxx

http://pathways.nice.org.uk/pathways/xxx

http://guidance.nice.org.uk/CGnn/PublicInfo



We encourage NHS and third sector, including voluntary, organisations to use 2560

text from this booklet in their own information about stage 4 or 5 CKD. 2561

7 Related NICE guidance 2562

Published 2563

Peritoneal dialysis. NICE clinical guideline 125 (2011). 2564

Anaemia management in people with chronic kidney disease. NICE clinical 2565

guideline 114 (2011). 2566

Medicines adherence: Involving patients in decisions about prescribed 2567

medicines and supporting adherence. NICE clinical guideline 76 (2009). 2568

Chronic kidney disease. NICE clinical guideline 73 (2008). 2569

Cinacalcet for the treatment of secondary hyperparathyroidism in patients 2570

with end-stage renal disease on maintenance dialysis therapy. NICE 2571

technology appraisal guidance 117 (2007). 2572

Under development 2573

NICE is developing the following guidance (details available from the NICE 2574

website): 2575

Chronic kidney disease (update). NICE clinical guideline. Publication date 2576

to be confirmed. 2577

Acute kidney injury. NICE clinical guideline. Publication date August 2013. 2578

8 Updating the guideline 2579

NICE clinical guidelines are updated so that recommendations take into 2580

account important new information. New evidence is checked 3 years after 2581

publication, and healthcare professionals and patients are asked for their 2582

views; we use this information to decide whether all or part of a guideline 2583

needs updating. If important new evidence is published at other times, we 2584

may decide to do a more rapid update of some recommendations. Please see 2585

our website for information about updating the guideline. 2586

http://www.guidance.nice.org.uk/CG125


http://www.nice.org.uk/CG76

http://www.nice.org.uk/CG76


http://guidance.nice.org.uk/TA117

http://guidance.nice.org.uk/TA117

http://www.nice.org.uk/

http://www.nice.org.uk/



8.1 References 2587

Dietary management for people with stage 4 or 5 CKD who are not on 2588

dialysis 2589

Cianciaruso B; Pota A; Pisani A; Torraca S; Annecchini R; Lombardi P; 2590

Capuano A; Nazzaro P; Bellizzi V; Sabbatini M. (2008) Metabolic effects of 2591

two low protein diets in chronic kidney disease stage 4-5--a randomized 2592

controlled trial. Nephrology Dialysis Transplantation 23(2): 636-644 [ref ID: 2593

555] 2594

Cianciaruso B; Pota A; Bellizzi V; Di Giuseppe D; Di Micco L; Minutolo R; 2595

Pisani A; Sabbatini M; Ravani P. (2009) Effect of a low- versus moderate-2596

protein diet on progression of CKD: follow-up of a randomized controlled trial. 2597

American Journal of Kidney Diseases 54(6): 1052-1061 [ref ID: 556] 2598

Di Iorio BR; Minutolo R; De Nicola L; Bellizzi V; Catapano F; Iodice C; 2599

Rubino R. (2003) Supplemented very low protein diet ameliorates 2600

responsiveness to erythropoietin in chronic renal failure. Kidney International 2601

64(5): 1822-1828 [ref ID: 9475] 2602

European Study Group for the Conservative Management of Chronic Renal 2603

Failure. (1992) Dietary compliance in the trial of the European Study Group. 2604

An interim analysis. Guarnieri G, Panzetta G, Toigo (eds): Metabolic and 2605

Nutritional Abnormalities in Kidney Disease. Contributions to Nephrology 98: 2606

133-141 [ref ID: 5] 2607

Feiten SF; Draibe SA; Watanabe R; Duenhas MR; Baxmann AC; Nerbass FB; 2608

Cuppari L. (2005) Short-term effects of a very-low-protein diet supplemented 2609

with ketoacids in nondialyzed chronic kidney disease patients. European 2610

Journal of Clinical Nutrition 59: 129-136 [ref ID: 814] 2611

Ihle BU; Becker GJ; Whitworth JA; Charlwood RA; Kincaid-Smith PS. (1989) 2612

The effect of protein restriction on the progression of renal insufficiency. New 2613

England Journal of Medicine 321(26): 1773-1777 [ref ID: 1162] 2614

Jungers P; Chauveau P; Ployard F; Lebkiri B; Ciancioni C; Man NK. (1987) 2615

Comparison of ketoacids and low protein diet on advanced chronic renal 2616



failure progression. Kidney International Supplement 22: S67-S71 [ref ID: 2617

1242] 2618

Klahr S; Levey AS; Beck GJ; Caggiula AW; Hunsicker L; Kusek JW; Striker G. 2619

(1994) The effects of dietary protein restriction and blood-pressure control on 2620

the progression of chronic renal disease. Modification of Diet in Renal Disease 2621

Study Group. New England Journal of Medicine 330(13): 877-884 [ref ID: 2622

1339] 2623

Kopple JD; Levey AS; Greene T; Chumlea WC; Gassman JJ; Hollinger DL; 2624

Maroni BJ; Merrill D; Scherch LK; Schulman G; Wang SR; Zimmer GS. (1997) 2625

Effect of dietary protein restriction on nutritional status in the Modification of 2626

Diet in Renal Disease Study. Kidney International 52(3): 778-791 [ref ID: 2627

1380] 2628

Lindenau K, Abendroth K, Kokot F, Vetter K, Rehse C, Fröhling PT. (1990) 2629

Therapeutic effect of keto acids on renal osteodystrophy. A prospective 2630

controlled study. Nephron 55(2): 133-135 [ref ID: 14054] 2631

Malvy D; Maingourd C; Pengloan J; Bagros P; Nivet H. (1999) Effects of 2632

severe protein restriction with ketoanalogues in advanced renal failure. 2633

Journal of the American College of Nutrition 18(5): 481-486 [ref ID: 1650] 2634

Mircescu G; Garneata L; Stancu SH; Capusa C. (2007) Effects of a 2635

supplemented hypoproteic diet in chronic kidney disease. Journal of renal 2636

nutrition 17(3): 179-88 [ref ID: 1794] 2637

Snetselaar L; Dwyer J; Adler S; Petot GJ; Berg R; Gassman JJ; Houser H. 2638

(1994) Reduction of dietary protein and phosphate in the modification of diet 2639

in renal disease feasibility study. Journal of the American Dietetic Association 2640

94(9): 986-992 [ref ID: 2444] 2641

Dietary management for people with stage 5 CKD who are on dialysis 2642

Guida B; Piccoli A; Trio R; Laccetti R; Nastasi A; Paglione A; Memoli A; 2643

Memoli B. (2011) Dietary phosphate restriction in dialysis patients: A new 2644

approach for the treatment of hyperphosphataemia. Nutrition Metabolism & 2645

Cardiovascular Diseases 21: 879-884 [ref ID: 1018] 2646



Jiang N; Qian J; Sun W; Lin A; Cao L; Wang Q; Ni Z; Wan Y; Linholm B; 2647

Axelsson J; Yao Q. (2009) Better preservation of residual renal function in 2648

peritoneal dialysis patients treated with a low-protein diet supplemented with 2649

keto acids: a prospective, randomized trial. Nephrology Dialysis 2650

Transplantation 24(8): 2551-2558 [ref ID: 1214] 2651

Li H; Long Q; Shao C; Fan H; Yuan L; Huang B; Gu Y; Lin S; Hao C; Chen J. 2652

(2011) Effect of short-term low-protein diet supplemented with keto acids on 2653

hyperphosphataemia in maintenance haemodialysis patients. Blood 2654

Purification 31(1-3): 33-40 [ref ID: 1512] 2655

Patient information strategies 2656

Ashurst IB; Dobbie H. A (2003) randomized controlled trial of an educational 2657

intervention to improve phosphate levels in hemodialysis patients. Journal of 2658

Renal Nutrition 13(4): 267-74 [ref ID: 7057] 2659

Baraz S, Parvardeh S, Mohammadi E, Broumand B. (2010) Dietary and fluid 2660

compliance: an educational intervention for patients having haemodialysis. 2661

Journal of Advanced Nursing 66(1): 60-68 [ref ID: 15007] 2662

Campbell KL; Ash S; Davies PS; Bauer JD. (2008) Randomized controlled 2663

trial of nutritional counselling on body composition and dietary intake in severe 2664

CKD. American Journal of Kidney Diseases 51(5): 748-58 [ref ID: 439] 2665

Chen W; Lu XH; Wang T. (2006) Menu suggestion: an effective way to 2666

improve dietary compliance in peritoneal dialysis patients. Journal of Renal 2667

Nutrition 16(2): 132-6 [ref ID: 7186] 2668

Ford JC; Pope JF; Hunt AE; Gerald B. (2004) The effect of diet education on 2669

the laboratory values and knowledge of hemodialysis patients with 2670

hyperphosphatemia. Journal of Renal Nutrition 14(1): 36-44 [ref ID: 15000] 2671

Morey B; Walker R; Davenport A. (2008) More dietetic time, better outcome? 2672

A randomized prospective study investigating the effect of more dietetic time 2673

on phosphate control in end-stage kidney failure haemodialysis patients. 2674

Nephron - Clinical Practice 109(3): c173-80 [ref ID: 7708] 2675



Shaw-Stuart NJ; Stuart A. (2000) The effect of an educational patient 2676

compliance program on serum phosphate levels in patients receiving 2677

haemodialysis. Journal of Renal Nutrition 10(2): 80-4 [ref ID: 7911] 2678

Sullivan C; Sayre SS; Leon JB; Machekano R; Love TE; Porter D; Marbury M; 2679

Sehgal AR. (2009) Effect of food additives on hyperphosphatemia among 2680

patients with end-stage renal disease: a randomized controlled trial. JAMA 2681

301(6): 629-635 [ref ID: 7958] 2682

Tanner JL; Craig CB; Bartolucci AA; Allon M; Fox LM; Geiger BF; Wilson NP. 2683

(1998) The effect of a self-monitoring tool on self-efficacy, health beliefs, and 2684

adherence in patients receiving haemodialysis. Journal of Renal Nutrition 8(4): 2685

203-11 [ref ID: 7981] 2686

Use of phosphate binders in people with stage 4 or 5 CKD who are not 2687

on dialysis 2688

Caglar K; Yilmaz MI; Saglam M; Cakir E; Acikel C; Eyileten T; Yenicesu M; 2689

Oguz Y; Vural A; Carrero JJ; Axelsson J; Lindholm B; Stenvinkel P. (2008) 2690

Short-term treatment with sevelamer increases serum fetuin-a concentration 2691

and improves endothelial dysfunction in chronic kidney disease stage 4 2692

patients. Clinical Journal of The American Society of Nephrology 3: 61-8 [ref 2693

ID: 4173] 2694

Qunibi W; Winkelmayer WC; Solomon R; Moustafa M; Kessler P; Ho CH; 2695

Greenberg J; Diaz-Buxo JA. (2011) A randomized, double-blind, placebo-2696

controlled trial of calcium acetate on serum phosphorus concentrations in 2697

patients with advanced non-dialysis-dependent chronic kidney disease. BMC 2698

Nephrology 12: 9 [ref ID: 4966] 2699

Russo D; Miranda I; Ruocco C; Battaglia Y; Buonanno E; Manzi S; Russo L; 2700

Scafarto A; Andreucci VE. (2007) The progression of coronary artery 2701

calcification in predialysis patients on calcium carbonate or sevelamer. Kidney 2702

International 72: 1255-61 [ref ID: 5033] 2703

Sprague SM; Abboud H; Qiu P; Dauphin M; Zhang P; Finn W. (2009) 2704

Lanthanum carbonate reduces phosphorus burden in patients with CKD 2705



stages 3 and 4: A randomized trial. Clinical Journal of the American Society of 2706

Nephrology 4 (1): 178-185 [ref ID: 5142] 2707

Use of phosphate binders in children with stage 5 CKD who are on 2708

dialysis 2709

Salusky IB; Goodman WG; Sahney S; Gales B; Perilloux A; Wang HJ; 2710

Elashoff RM; Juppner H. (2005) Sevelamer controls parathyroid hormone-2711

induced bone disease as efficiently as calcium carbonate without increasing 2712

serum calcium levels during therapy with active vitamin D sterols. Journal of 2713

the American Society of Nephrology 16: 2501-8 [ref ID: 5043] 2714

Use of phosphate binders in adults with stage 5 CKD who are on dialysis 2715

Al-Baaj F; Speake M; Hutchison AJ (2005) Control of serum phosphate by 2716

oral lanthanum carbonate in patients undergoing haemodialysis and 2717

continuous ambulatory peritoneal dialysis in a short-term, placebo-controlled 2718

study. Nephrology Dialysis Transplantation 20: 775-82 [ref ID: 4029] 2719

Asmus HG; Braun J; Krause R; Brunkhorst R; Holzer H; Schulz W; Neumayer 2720

HH; Raggi P; Bommer J. (2005) Two year comparison of sevelamer and 2721

calcium carbonate effects on cardiovascular calcification and bone density. 2722

Nephrology Dialysis Transplantation 20: 1653-61 [ref ID: 4062] 2723

Babarykin D; Adamsone I; Amerika D; Spudass A; Moisejev V; Berzina N; 2724

Michule L; Rozental R. (2004) Calcium-enriched bread for treatment of uremic 2725

hyperphosphatemia. Journal of Renal Nutrition 14: 149-56 [ref ID: 4068] 2726

Barreto DV; Barreto FC; de Carvalho AB; Cuppari L; Draibe SA; Dalboni MA; 2727

Moyses RM; Neves KR; Jorgetti V; Miname M; Santos RD; Canziani ME. 2728

(2008) Phosphate binder impact on bone remodeling and coronary 2729

calcification - results from the BRiC study. Nephron 110: c273-c283 [ref ID: 2730

4084] 2731

Block GA., Hulbert-Shearon TE., Levin NW., Port FK. Association of serum 2732

phosphorus and calcium x phosphate product with mortality risk in chronic 2733

hemodialysis patients: a national study. American Journal of Kidney Diseases 2734

1998; 31(4):607-17. 2735



Block GA; Spiegel DM; Ehrlich J; Mehta R; Lindbergh J; Dreisbach A; Raggi 2736

P. (2005) Effects of sevelamer and calcium on coronary artery calcification in 2737

patients new to hemodialysis. Kidney International 68: 1815-24 [ref ID: 4114] 2738

British National Formulary No. 63 March 2012 2739

Braun J; Asmus HG; Holzer H; Brunkhorst R; Krause R; Schulz W; Neumayer 2740

HH; Raggi P; Bommer J. (2004) Long-term comparison of a calcium-free 2741

phosphate binder and calcium carbonate--phosphorus metabolism and 2742

cardiovascular calcification. Clinical Nephrology 62: 104-15 [ref ID: 4142] 2743

Chertow GM; Burke SK; Lazarus JM; Stenzel KH; Wombolt D; Goldberg D; 2744

Bonventre JV; Slatopolsky E. (1997) Poly[allylamine hydrochloride] 2745

(RenaGel): a noncalcemic phosphate binder for the treatment of 2746

hyperphosphatemia in chronic renal failure. American Journal of Kidney 2747

Diseases 29: 66-71[ref ID: 4203] 2748

Chertow GM; Burke SK; Raggi P; Treat to Goal Working Group. (2002) 2749

Sevelamer attenuates the progression of coronary and aortic calcification in 2750

hemodialysis patients. Kidney International 62: 245-52 [ref ID: 4209] 2751

Chertow GM; Raggi P; McCarthy JT; Schulman G; Silberzweig J; Kuhlik A; 2752

Goodman WG; Boulay A; Burke SK; Toto RD. (2003) The effects of 2753

sevelamer and calcium acetate on proxies of atherosclerotic and 2754

arteriosclerotic vascular disease in hemodialysis patients. American Journal of 2755

Nephrology 23: 307-14 [ref ID: 4213] 2756

Chiang SS; Chen JB; Yang WC (2005) Lanthanum carbonate (Fosrenol) 2757

efficacy and tolerability in the treatment of hyperphosphatemic patients with 2758

end-stage renal disease. Clinical Nephrology 63: 461-70 [ref ID: 4217] 2759

Chow KM; Szeto CC; Kwan BC; Leung CB; Li PK. (2007) Sevelamer 2760

treatment strategy in peritoneal dialysis patients: conventional dose does not 2761

make best use of resources. Journal of Nephrology 20: 674-82 [ref ID: 4219] 2762

de Francisco AL; Leidig M; Covic AC; Ketteler M; Benedyk-Lorens E; 2763

Mircescu GM; Scholz C; Ponce P; Passlick-Deetjen J. (2010) Evaluation of 2764



calcium acetate/magnesium carbonate as a phosphate binder compared with 2765

sevelamer hydrochloride in haemodialysis patients: a controlled randomized 2766

study (CALMAG study) assessing efficacy and tolerability. Nephrology 2767

Dialysis Transplantation 25: 3707-17 [ref ID: 4283] 2768

De Santo NG; Frangiosa A; Anastasio P; Marino A; Correale G; Perna A; Di 2769

Stazio E; Stellato D; Santoro D; Di Meglio E; Iacono G; Ciacci C; Savica V; 2770

Cirillo M. (2006) Sevelamer worsens metabolic acidosis in hemodialysis 2771

patients. Journal of Nephrology 19: Suppl-14 [ref ID: 4284] 2772

Emmett M; Sirmon MD; Kirkpatrick WG; Nolan CR; Schmitt GW; Cleveland 2773

MB. (1991) Calcium acetate control of serum phosphorus in hemodialysis 2774

patients. American Journal of Kidney Diseases 17: 544-50 [ref ID: 4328] 2775

Evenepoel P; Selgas R; Caputo F; Foggensteiner L; Heaf JG; Ortiz A; Kelly A; 2776

Chasan-Taber S; Duggal A; Fan S. (2009) Efficacy and safety of sevelamer 2777

hydrochloride and calcium acetate in patients on peritoneal dialysis. 2778


Ferreira A; Frazao JM; Monier-Faugere MC; Gil C; Galvao J; Oliveira C; 2780

Baldaia J; Rodrigues I; Santos C; Ribeiro S; Hoenger RM; Duggal A; Malluche 2781

HH; Sevelamer Study Group. (2008) Effects of sevelamer hydrochloride and 2782

calcium carbonate on renal osteodystrophy in hemodialysis patients. Journal 2783

of the American Society of Nephrology 19: 405-12 [ref ID: 4353] 2784

Finn WF; Joy MS; Hladik G; Lanthanum Study Group. (2004) Efficacy and 2785

safety of lanthanum carbonate for reduction of serum phosphorus in patients 2786

with chronic renal failure receiving hemodialysis. Clinical Nephrology 62: 193-2787

201 [ref ID: 4361] 2788

Finn WF; Lanthanum Study Group (2006) Lanthanum carbonate versus 2789

standard therapy for the treatment of hyperphosphatemia: safety and efficacy 2790

in chronic maintenance hemodialysis patients. Clinical Nephrology 65: 191-2791

202 [ref ID: 4365] 2792



Fishbane S; Delmez J; Suki WN; Hariachar SK; Heaton J; Chasan-Taber S; 2793

Plone MA; Moe S. (2010) A randomized, parallel, open-label study to compare 2794

once-daily sevelamer carbonate powder dosing with thrice-daily sevelamer 2795

hydrochloride tablet dosing in CKD patients on hemodialysis. American 2796

Journal of Kidney Diseases 55: 307-15 [ref ID: 4370] 2797

Freemont AJ; Hoyland JA; Denton J; Lanthanum Carbonate SPD405-303 2798

Study Group. (2005) The effects of lanthanum carbonate and calcium 2799

carbonate on bone abnormalities in patients with end-stage renal disease. 2800

Clinical Nephrology 64: 428-37 [ref ID: 4392] 2801

Galassi A; Spiegel DM; Bellasi A; Block GA; Raggi P. (2006) Accelerated 2802

vascular calcification and relative hypoparathyroidism in incident 2803

haemodialysis diabetic patients receiving calcium binders. Nephrology 2804

Dialysis Transplantation 21: 3215-22 [ref ID: 4410] 2805

Gorodetskaya I, Zenios S, McCulloch CE, Bostrom A, Hsu CY, Bindman AB, 2806

Go AS, Chertow GM. Health-related quality of life and estimates of utility in 2807

chronic kidney disease. Kidney Int. 2005 Dec;68(6):2801-8. 2808

Hervas JG; Prados D; Cerezo S (2003) Treatment of hyperphosphatemia with 2809

sevelamer hydrochloride in hemodialysis patients: a comparison with calcium 2810

acetate. Kidney International - Supplement: S69-S72 [ref ID: 4500] 2811

Holland R., Brooksby I., Lenaghan E., Ashton K., Hay L., Smith R., Shepstone 2812

L., Shepstone L., et al. Effectiveness of visits from community pharmacists for 2813

patients with heart failure: HeartMed randomised controlled trial. BMJ. 2007 2814

May 26; 334(7603): 1098. 2815

Houillier P, Froissart M, Maruani G, Blanchard A. What serum calcium can tell 2816

us and what it can't. Nephrol. Dial. Transplant. 2006;21(1):29–32. 2817

Hutchison AJ; Maes B; Vanwalleghem J; Asmus G; Mohamed E; Schmieder 2818

R; Backs W; Jamar R; Vosskuhler A. (2005) Efficacy, tolerability, and safety of 2819

lanthanum carbonate in hyperphosphatemia: a 6-month, randomized, 2820

comparative trial versus calcium carbonate. Nephron 100: c8-19 [ref ID: 4526] 2821



Iwasaki Y; Takami H; Tani M; Yamaguchi Y; Goto H; Goto Y; Goto Y; 2822

Shigematsu T. (2005) Efficacy of combined sevelamer and calcium carbonate 2823

therapy for hyperphosphatemia in Japanese hemodialysis patients. 2824

Therapeutic Apheresis & Dialysis: Official Peer-Reviewed Journal of the 2825

International Society for Apheresis, the Japanese Society for Apheresis, the 2826

Japanese Society for Dialysis Therapy 9: 347-51 [ref ID: 4559] 2827

Janssen MJ; van der Kuy A; ter Wee PM; van Boven WP. (1995) Calcium 2828

acetate versus calcium carbonate and erythropoietin dosages in 2829

haemodialysis patients. Nephrology Dialysis Transplantation 10: 2321-4 [ref 2830

ID: 4567] 2831

Janssen MJ; van der Kuy A; ter Wee PM; van Boven WP. (1996) Aluminum 2832

hydroxide, calcium carbonate and calcium acetate in chronic intermittent 2833

hemodialysis patients. Clinical Nephrology 45: 111-9 [ref ID: 4568] 2834

Joy MS; Finn WF; LAM-302 Study Group (2003) Randomized, double-blind, 2835

placebo-controlled, dose-titration, phase III study assessing the efficacy and 2836

tolerability of lanthanum carbonate: a new phosphate binder for the treatment 2837

of hyperphosphatemia. American Journal of Kidney Diseases 42: 96-107 [ref 2838

ID: 4585] 2839

Kakuta T; Tanaka R; Hyodo T; Suzuki H; Kanai G; Nagaoka M; Takahashi H; 2840

Hirawa N; Oogushi Y; Miyata T; Kobayashi H; Fukagawa M; Saito A. (2011) 2841

Effect of sevelamer and calcium-based phosphate binders on coronary artery 2842

calcification and accumulation of circulating advanced glycation end products 2843

in hemodialysis patients.[Erratum appears in Am J Kidney Dis. 2011 2844

May;57(5):804]. American Journal of Kidney Diseases 57: 422-31 [ref ID: 2845

4593] 2846

Katopodis KP; Andrikos EK; Gouva CD; Bairaktari ET; Nikolopoulos PM; 2847

Takouli LK; Tzallas CS; Elisaf MS; Pappas MV; Siamopoulos KC. (2006) 2848

Sevelamer hydrochloride versus aluminium hydroxide: effect on serum 2849

phosphorus and lipids in CAPD patients. Peritoneal Dialysis International 26: 2850

320-7 [ref ID: 4605] 2851



Koiwa F; Onoda N; Kato H; Tokumoto A; Okada T; Fukagawa M; Shigematsu 2852

T; ROD 21 Clinical Research Group. (2005a) Prospective randomized 2853

multicenter trial of sevelamer hydrochloride and calcium carbonate for the 2854

treatment of hyperphosphatemia in hemodialysis patients in Japan. 2855

Therapeutic Apheresis & Dialysis: Official Peer-Reviewed Journal of the 2856

International Society for Apheresis, the Japanese Society for Apheresis, the 2857

Japanese Society for Dialysis Therapy 9: 340-6 [ref ID: 4634] 2858

Koiwa F; Kazama JJ; Tokumoto A; Onoda N; Kato H; Okada T; Nii-Kono T; 2859

Fukagawa M; Shigematsu T; ROD 21 Clinical Research Group. (2005b) 2860

Sevelamer hydrochloride and calcium bicarbonate reduce serum fibroblast 2861

growth factor 23 levels in dialysis patients. Therapeutic Apheresis & Dialysis: 2862

Official Peer-Reviewed Journal of the International Society for Apheresis, the 2863

Japanese Society for Apheresis, the Japanese Society for Dialysis Therapy 9: 2864

336-9 [ref ID: 4635] 2865

Liem YS, Bosch JL, Hunink MG. Preference-based quality of life of patients on 2866

renal replacement therapy: a systematic review and meta-analysis. Value 2867

Health. 2008 Jul-Aug;11(4):733-41. Epub 2008 Jan 8. 2868

Lin YF; Chien CT; Kan WC; Chen YM; Chu TS; Hung KY; Tsai TJ; Wu KD; 2869

Wu MS. (2011) Pleiotropic effects of sevelamer beyond phosphate binding in 2870

end-stage renal disease patients: a randomized, open-label, parallel-group 2871

study. Clinical Drug Investigation 31: 257-67 [ref ID: 4687] 2872

Liu YL; Lin HH; Yu CC; Kuo HL; Yang YF; Chou CY; Lin PW; Liu JH; Liao PY; 2873

Huang CC. (2006) A comparison of sevelamer hydrochloride with calcium 2874

acetate on biomarkers of bone turnover in hemodialysis patients. Renal 2875

Failure 28: 701-7 [ref ID: 4695] 2876

Malluche HH; Siami GA; Swanepoel C; Wang GH; Mawad H; Confer S; Smith 2877

M; Pratt RD; Monier-Faugere MC; Lanthanum Carbonate Study Group. (2008) 2878

Improvements in renal osteodystrophy in patients treated with lanthanum 2879

carbonate for two years. Clinical Nephrology 70: 284-95 [ref ID: 4744] 2880



Navarro-Gonzalez JF; Mora-Fernandez C; Muros de Fuentes M; Donate-2881

Correa J; Cazana-Perez V; Garcia-Perez J. (2011) Effect of phosphate 2882

binders on serum inflammatory profile, soluble CD14, and endotoxin levels in 2883

hemodialysis patients. Clinical Journal of The American Society of 2884

Nephrology: CJASN 6: 2272-9 [ref ID: 4858] 2885

Peasgood T., Herrmann K., Kanis JA., Brazier JE. An updated systematic 2886

review of Health State Utility Values for osteoporosis related conditions. 2887

Osteoporos Int (2009) 20:853–868 2888

Qunibi W; Moustafa M; Muenz LR; He DY; Kessler PD; Diaz-Buxo JA; Budoff 2889

M; CARE-2 Investigators. (2008) A 1-year randomized trial of calcium acetate 2890

versus sevelamer on progression of coronary artery calcification in 2891

hemodialysis patients with comparable lipid control: the Calcium Acetate 2892

Renagel Evaluation-2 (CARE-2) study. American Journal of Kidney Diseases 2893

51: 952-65 [ref ID: 4965] 2894

Raggi P; Bommer J; Chertow GM (2004) Valvular calcification in hemodialysis 2895

patients randomized to calcium-based phosphorus binders or sevelamer. 2896

Journal of Heart Valve Disease 13: 134-41 [ref ID: 4978] 2897

Ring T; Nielsen C; Andersen SP; Behrens JK; Sodemann B; Kornerup HJ. 2898

(1993) Calcium acetate versus calcium carbonate as phosphorus binders in 2899

patients on chronic haemodialysis: a controlled study. Nephrology Dialysis 2900

Transplantation 8: 341-6 [ref ID: 5004] 2901

Shigematsu T; Lanthanum Carbonate Group (2008a) Multicenter prospective 2902

randomized, double-blind comparative study between lanthanum carbonate 2903

and calcium carbonate as phosphate binders in Japanese hemodialysis 2904

patients with hyperphosphatemia. Clinical Nephrology 70: 404-10 [ref ID: 2905

5095] 2906

Shigematsu T; Lanthanum Carbonate Research Group (2008b) Lanthanum 2907

carbonate effectively controls serum phosphate without affecting serum 2908

calcium levels in patients undergoing hemodialysis. Therapeutic Apheresis & 2909

Dialysis: Official Peer-Reviewed Journal of the International Society for 2910



Apheresis, the Japanese Society for Apheresis, the Japanese Society for 2911

Dialysis Therapy 12: 55-61 [ref ID: 5096] 2912

Spasovski GB; Sikole A; Gelev S; Masin-Spasovska J; Freemont T; Webster 2913

I; Gill M; Jones C; De Broe ME; D'Haese PC. (2006) Evolution of bone and 2914

plasma concentration of lanthanum in dialysis patients before, during 1 year of 2915

treatment with lanthanum carbonate and after 2 years of follow-up. 2916


Spiegel DM; Farmer B; Smits G; Chonchol M. (2007) Magnesium carbonate is 2918

an effective phosphate binder for chronic hemodialysis patients: a pilot study. 2919

Journal of Renal Nutrition 17: 416-22 [ref ID: 5132] 2920

Suki WN; Zabaneh R; Cangiano JL; Reed J; Fischer D; Garrett L; Ling BN; 2921

Chasan-Taber S; Dillon MA; Blair AT; Burke SK. (2007) Effects of sevelamer 2922

and calcium-based phosphate binders on mortality in hemodialysis patients. 2923

Kidney International 72: 1130-7 [ref ID: 5163] 2924

Tzanakis IP; Papadaki AN; Wei M; Kagia S; Spadidakis VV; Kallivretakis NE; 2925

Oreopoulos DG. (2008) Magnesium carbonate for phosphate control in 2926

patients on hemodialysis. A randomized controlled trial. International Urology 2927

& Nephrology 40: 193-201 [ref ID: 5221] 2928

UK Renal Registry. thirteenth annual report. 2010. 2929

http://www.renalreg.com/Reports/2010.html 2930

2931

Wilson R; Zhang P; Smyth M; Pratt R. (2009) Assessment of survival in a 2-2932

year comparative study of lanthanum carbonate versus standard therapy. 2933

Current Medical Research & Opinion 25: 3021-8 [ref ID: 5278] 2934

Use of supplements in people with stage 4 or 5 CKD who are not on 2935

dialysis 2936

No studies found 2937

Use of supplements in people with stage 5 CKD who are on dialysis 2938

Chertow GM; Dillon M; Burke SK; Steg M; Bleyer AJ; Garrett BN; Domoto DT; 2939

Wilkes BM; Wombolt DG; Slatopolsky E. A randomised trial of sevelamer 2940

http://www.renalreg.com/Reports/2010.html



hydrochloride (RenaGel) with and without supplemental calcium. Clinical 2941

Nephrology 51(1): 18-26 (1999) [ref ID: 9336] 2942

Cibulka R; Racek J; Pikner R; Rajdl D; Trefil L; Vesala E; Studenovska M; 2943

Siroka R. Effect of L-carnitine supplementation on secondary 2944

hyperparathyroidism and bone metabolism in hemodialyzed patients. Calcified 2945

Tissue International 81: 99-106 (2007) [ref ID: 9357] 2946

Shahbazian H; Zafar Mohtashami A; Ghorbani A; Abbaspour MR; Belladi 2947

Musavi SS; Hayati F; Lashkarara GR. Oral niacinamide reduces serum 2948

phosphate, increases HDL, and induces thrombocytopenia in hemodialysis 2949

patients: a double-blind racndomised controlled trial. Nefrologia 31(1): 58-65 2950

(2011) [ref ID: 10763] 2951

Young DO; Cheng SC; Delmez JA; Coyne DW. The effect of oral niacinamide 2952

on plasma phosphate levels in peritoneal dialysis patients. Peritoneal Dialysis 2953

International 29: 562-7 (2009) [ref ID: 11182] 2954

Sequencing of treatments 2955

No studies found 2956

2957

9 Glossary and abbreviations 2958

Glossary 2959

Chronic kidney disease 2960

The term ‘chronic kidney disease’ describes abnormal kidney function and/or 2961

structure. It is defined as ‘chronic’ as the condition is long-lasting and often 2962

progresses over time. 2963

The US ‘National Kidney Foundation kidney disease outcomes quality 2964

initiative’ (NKF-KDOQI) classification divides CKD into 5 stages according to 2965

the extent of a person’s loss of renal function. Stages 4 and 5, the stages 2966

covered in this guideline, are the most advanced stages of the condition. 2967

Stage 4 is defined by a glomerular filtration rate (GFR) of 15–2968



30 ml/min/1.73 m2, and stage 5 by a GFR of less than 15 ml/min/1.73 m2. A 2969

GFR of over 90 ml/min/1.73 m2 is considered normal unless there is other 2970

evidence of kidney disease. 2971

A decline in or absence of kidney function leads to a range of adverse effects, 2972

including: fluid retention, anaemia, abnormal levels of lipids in the blood, 2973

protein–energy malnutrition and disturbances in bone and mineral 2974

metabolism. Additionally, chronic kidney disease often exists together with 2975

other conditions, such as cardiovascular disease and diabetes. 2976

Dietary management 2977

Dietary management is, in general, a non-pharmacological approach to 2978

managing hyperphosphataemia involving the adjustment of a patient’s diet to 2979

reduce the amount of phosphate they consume. This can be achieved, for 2980

example, by reducing the intake of food and drinks with a high phosphate to 2981

protein ratio, such as some dairy products and nuts, or a high level of 2982

phosphate additives, such as cola drinks or processed foods. However, it is 2983

also important that a patient’s nutritional status is maintained at a healthy 2984

level. 2985

Dialysis 2986

Dialysis is an artificial process for filtering waste and excess water from the 2987

blood. It is a form of renal replacement therapy and is often initiated when 2988

chronic kidney disease advances to end-stage renal disease, in which there is 2989

a complete or almost complete loss of renal function. 2990

Extended dominance 2991

In health economic modelling, each intervention is compared to the next most 2992

effective alternative by calculating the incremental cost-effectiveness ratio 2993

(ICER). Extended dominance occurs when the ICER for a given treatment 2994

alternative is higher than that of the next, more effective, alternative. A 2995

treatment option that is extendedly dominated, that is has a higher ICER and 2996

lower effectiveness, will almost always be rejected in favour of a treatment 2997

option with a lower ICER and greater effectiveness. 2998



Hypercalcaemia 2999

Hypercalcaemia is the abnormal elevation of calcium in the blood. The 3000

condition arises because of insufficient filtering of calcium from the blood by 3001

poorly functioning kidneys. This, in turn, means that a certain amount of the 3002

calcium does not leave the body in the urine, instead remaining in the blood. 3003

Other causes of hypercalcaemia include a high intake of calcium, for example 3004

through the diet or medications, and high levels of vitamin D or parathyroid 3005

hormone, which increase the absorption of calcium in the intestine and its 3006

release from bone. Calcium is also obtained from dialysate in haemodialysis. 3007

High serum calcium levels can lead to increases in morbidity and mortality 3008

through, for example, abnormalities of bone and joint morphology, increased 3009

vascular and soft tissue calcification, and cardiovascular disease. 3010

Hyperphosphataemia 3011

Hyperphosphataemia is an abnormal elevation of phosphate in the blood. The 3012

condition arises because of insufficient filtering of phosphate from the blood 3013

by poorly functioning kidneys. This, in turn, means that a certain amount of the 3014

phosphate does not leave the body in the urine, instead remaining in the 3015

blood. 3016

Dietary phosphate intake can further contribute to hyperphosphataemia; food 3017

and drink with a large phosphate to protein ratio, such as dairy products, or a 3018

large amount of phosphate additives are of particular concern in patients with 3019

chronic kidney disease. 3020

High serum phosphate levels can directly and indirectly increase parathyroid 3021

hormone secretion, leading to the development of secondary 3022

hyperparathyroidism and increases in morbidity and mortality. 3023

Hyperphosphataemia can contribute to an increased incidence of fracture, 3024

abnormalities of bone and joint morphology, vascular and soft tissue 3025

calcification, and cardiovascular disease. 3026



Multiple treatment comparison 3027

Well-conducted randomised controlled trials are the ‘gold standard’ for directly 3028

comparing the effectiveness of different treatments. However, for some 3029

conditions there are many different treatments available, many of which have 3030

not been directly compared in head-to-head trials. 3031

If there is a lack of evidence from direct comparison trials, or the available 3032

evidence is limited, results from different trials can be used to indirectly 3033

estimate the effects of a treatment relative to each of the other treatments. In 3034

this way, a multiple treatment comparison uses a network of the available 3035

direct and indirect comparisons to provide an overall picture of the available 3036

treatments. Indirect comparisons can also be used to strengthen the 3037

estimates of effect gained from direct comparisons. 3038

Phosphate binders 3039

Phosphate binders are pharmacological interventions that bind to dietary 3040

phosphate while it is in the stomach, preventing its absorption into the blood. 3041

Instead, the bound phosphate is passed out of the body in the faeces. 3042

A range of phosphate binders are licensed in the UK, though they can broadly 3043

be defined as calcium-based and non-calcium-based. They are available as 3044

pills, chewable tablets, capsules and powders. 3045

Because the phosphate-binding action occurs in the stomach, it is important 3046

that these medications are taken with food and not on an empty stomach. 3047

Self-management tool 3048

For the purpose of this guideline, the summary term 'self-management tool' 3049

was used to collectively describe the aids used to help patients to manage 3050

their dietary intake or serum phosphate levels alongside education or 3051

counselling. 3052

These tools ranged from a fridge magnet detailing high-phosphate foods, to 3053

an individualised tracking chart that uses visual goals to engage patients in 3054

achieving phosphate control. 3055



Supplements 3056

Supplements are a potentially wide range of substances, including vitamins 3057

and minerals. A supplement is often defined as a preparation intended to 3058

provide nutrients that may be missing from a person’s diet or may not be 3059

consumed in sufficient quantities. However, in this guideline supplements 3060

were defined as vitamins, minerals and other substances that control serum 3061

phosphate through effects other than phosphate-binding. 3062

It should be noted that in the case of calcium, its classification as a phosphate 3063

binder was distinguished from that as a supplement through the timing of its 3064

administration. As defined in this guideline, calcium-based phosphate binders 3065

are taken with food while calcium supplements are taken on an empty 3066

stomach when the phosphate-binding effect will be limited. 3067

Please see the NICE glossary for an explanation of terms not described 3068

above. 3069

Abbreviations 3070

Abbreviation Term

AA Amino acid

ACR Albumin to creatinine ratio

Any-B Any binder

C-based Calcium-based binder

CA Calcium acetate

CAMG Calcium acetate with magnesium carbonate

CC Calcium carbonate

CCr Creatinine clearance

CHF Chronic heart failure

CI Confidence interval / credibility interval

CKD Chronic kidney disease

CV Cardiovascular

eGFR Estimated glomerular filtration rate

ESRD End-stage renal disease

GDG Guideline development group

GFR Glomerular filtration rate

GI Gastrointestinal

GRADE Grading of Recommendations Assessment, Development and Evaluation

http://www.nice.org.uk/website/glossary/glossary.jsp



HR Hazard ratio

ICER Incremental cost effectiveness ratio

IDWG Interdialytic weight gain

KA Keto acid

LC Lanthanum carbonate

LPD Low-protein diet

MD Mean difference

MG Magnesium carbonate

MNA Multi-nutritional assessment

MPD Moderate-protein diet

MTC Multiple treatment comparison

nPCR Normalised protein catabolic rate

OR Odds ratio

P Placebo

PTH Parathyroid hormone

PTx Parathyroidectomy

QALY Quality-adjusted life year

RCT Randomised controlled trial

RR Relative risk

RRT Renal replacement therapy

SC Sevelamer carbonate

SCr Serum creatinine

SD Standard deviation

SGA Subjective Global Assessment of nutrition

SH Sevelamer hydrochloride

sLPD Supplemented low-protein diet

sVLPD Supplemented very-low-protein diet

Tx Kidney transplant

VLPD Very-low-protein diet

3071

3072



Appendix A Contributors and declarations of interests 3073

The Guideline Development Group 3074

Gary McVeigh (Chair) 3075

Professor of Cardiovascular Medicine, Queen’s University Belfast 3076

David Bennett-Jones 3077

Consultant - Renal Medicine, University Hospitals Coventry & Warwickshire 3078

Shelley Cleghorn 3079

Principal Paediatric Nephrology Dietitian, Great Ormond Street Hospital for 3080

Children 3081

Roy Connell 3082

Clinical Nurse Specialist - Paediatric dialysis, Nottingham University Hospital 3083

Indranil Dasgupta 3084

Consultant Physician and Nephrologist, Birmingham Heartlands Hospital 3085

Sylvia Grace 3086

Renal Dietitian, University Hospitals Coventry & Warwick NHS Trust 3087

Clair Huckerby 3088

Pharmaceutical Adviser - Medicines Management Lead, NHS Dudley 3089

Nora Kerigan 3090

Dialysis Adequacy Practitioner, Lancashire Teaching Hospitals NHS Trust 3091

Fiona Loud 3092

Patient and carer member, The Kidney Alliance 3093

Nicholas Palmer 3094

Patient and carer member, National Kidney Federation 3095

Rukshana Shroff 3096

Consultant in Paediatric Nephrology, Great Ormond Street Hospital for 3097

Children 3098



Internal Clinical Guidelines Technical Team 3099

An Internal Clinical Guidelines Technical team was responsible for this 3100

guideline throughout its development. It prepared information for the Guideline 3101

Development Group, drafted the guideline and responded to consultation 3102

comments. 3103

Emma Banks 3104

Project Manager 3105

Mendwas Dzingina 3106

Technical Analyst (Health Economics) 3107

Sarah Glover 3108

Information Specialist 3109

Michael Heath 3110

Programme Manager 3111

Lucy Hoppe 3112

Assistant Technical Analyst 3113

Dylan Jones 3114

Technical Adviser 3115

Gabriel Rogers 3116

Technical Adviser (Health Economics) 3117

NICE Centre for Clinical Practice 3118

Laura Donegani 3119

Guideline Commissioning Coordinator 3120

Louise Millward 3121

Associate Director 3122

Sarah Palombella 3123

Editor 3124



Rachel Ryle 3125

Guideline Commissioning Manager 3126

Judith Thornton 3127

Technical Lead 3128

Erin Whittingham 3129

Assistant Project Manager, Patient & Public Involvement Programme 3130

Declarations of interests 3131

GDG Member

Interest Declared

Type of Interest Decisions Taken

Gary McVeigh

None

David Bennett Jones

None

Shelley Cleghorn None

Roy Connell None

Indril Dasgupta Chief Investigator in the UK for the Steering study which is an observational study of Osvaren (calmag) in dialysis patients.

Non Personal Pecuniary

Leave the room prior to any decisions and recommendations made

Sylvia Grace None

Clair Huckerby None

Nora Kerigan None

Fiona Loud None

Nicholas Palmer None

Rukshana Shroff None

3132

Appendix B List of all research recommendations 3133

The Guideline Development Group has made the following recommendations 3134

for research, based on its review of evidence, to improve NICE guidance and 3135

patient care in the future. 3136



B1 Phosphate binders in patients with CKD stage 3137

4 or 5 3138

Which binders are most effective in controlling serum phosphate in adults with 3139

stage 4 or 5 CKD who are not on dialysis? 3140

Why this is important 3141

Limited evidence was found on the use of phosphate binders in adults with 3142

stage 4 or 5 CKD. While it is possible in some instances to extrapolate from 3143

the evidence on people with stage 5 CKD who are on dialysis, it is not ideal. 3144

Therefore a series of RCTs should be conducted to examine the comparative 3145

effectiveness of various phosphate binders against each other for the 3146

management of serum phosphate in adults with stage 4 or 5 CKD. These 3147

trials should examine the long-term (ideally 12-month) effects of the various 3148

binders on outcomes such as serum phosphate, serum calcium, adverse 3149

events and the ability of the binders to control serum phosphate and calcium 3150

within the given ranges. 3151

B2 Effectiveness and safety of aluminium 3152

hydroxide in adults 3153

In adults with stage 4 or 5 CKD, including those on dialysis, what is the long-3154

term effectiveness and safety of aluminium hydroxide in controlling serum 3155

phosphate? 3156


Limited evidence was found on the efficacy of aluminium hydroxide in adults 3158

and no evidence was found on the long-term efficacy and safety of aluminium 3159

hydroxide. A series of RCTs should be conducted separately in adults with 3160

stages 4 or 5 CKD, including those on dialysis. These trials should be run for 3161

a minimum of 12 months and should examine the effect of aluminium 3162

hydroxide on outcomes such as serum phosphate, serum calcium, adverse 3163


within the given ranges. In addition, specific data should be collected on 3165

aspects relating to aluminium toxicity. 3166



B3 Effectiveness and safety of magnesium 3167

carbonate 3168

In adults and children with stage 4 or 5 CKD, including those on dialysis, what 3169

is the long-term effectiveness and safety of magnesium carbonate in 3170

controlling serum phosphate? 3171


Limited evidence was found on the use of magnesium carbonate to control 3173

serum phosphate. However, the evidence that was assessed suggested that 3174

magnesium carbonate could be very effective in controlling serum phosphate. 3175

A series of RCTs should be conducted separately in adults and children with 3176

stages 4 or 5 CKD, including those on dialysis. These trials should be run for 3177

a minimum of 12 months and should examine the effect of magnesium 3178

carbonate on outcomes such as serum phosphate, serum calcium, adverse 3179


within the given ranges. In addition, specific data should be collected on 3181

aspects relating to magnesium toxicity. 3182

B4 Phosphate binders in children 3183

Which binders are most effective in controlling serum phosphate in children 3184

with stage 4 or 5 CKD who are not on dialysis? 3185


Limited evidence was found on the use of phosphate binders in children with 3187

stage 5 CKD who are on dialysis, and none was found for those with stage 4 3188

or 5 CKD who are not on dialysis. Therefore a series of RCTs should be 3189

conducted that examine the comparative effectiveness of various phosphate 3190

binders against each other for the management of serum phosphate in 3191

children with stage 4 or 5 CKD. These trials should examine the long-term 3192

(ideally 12-month) effects of the various binders on outcomes such as serum 3193

phosphate, serum calcium, adverse events and the ability of the binders to 3194

control serum phosphate and calcium within the given ranges. 3195



B5 Sequencing and combining of phosphate 3196

binders 3197

For adults with stage 4 or 5 CKD, including those on dialysis, what is the most 3198

effective sequence or combination of phosphate binders to control serum 3199

phosphate? 3200


It is thought that the longer people remain on calcium-based binders, the 3202

greater their risk of developing hypercalcaemia. However, no evidence was 3203

found on the most appropriate sequence or combination of phosphate binders 3204

a person should receive to control serum phosphate and serum calcium. A 3205

series of RCTs should be conducted separately in adults with stages 4 or 5 3206

CKD, including those on dialysis. These trials should be run for a minimum of 3207

12 months and should examine comparative effectiveness of various 3208

sequences and combinations of available phosphate binders on outcomes 3209

such as serum phosphate, serum calcium, adverse events and the ability of 3210

the binders to control serum phosphate and calcium within the given ranges. 3211

3212



Appendix C Guideline scope 3213

See separate file. 3214

3215



Appendix D How this guideline was developed 3216


3218



Appendix E Evidence tables 3219


3221



Appendix F Full health economic report 3222


3224



Appendix G Clinical guideline technical assessment 3225

unit analysis (phosphate binders) 3226


3228

Hyperphosphataemia in chronic kidney disease

Documents

Transcript of Hyperphosphataemia in chronic kidney disease