Pediatr Blood Cancer 2010;54:983–989

Glomerular Toxicity Persists 10 Years After Ifosfamide Treatment inChildhood and Is Not Predictable by Age or Dose

Roderick Skinner, PhD, FRCPCH,1,2* Annie Parry, RSCN,1 Lisa Price, RSCN,1 Michael Cole, BSc, MSc,3

Alan W. Craft, MD, FRCPCH,2 and Andrew D.J. Pearson, MD, FRCPCH2,4

INTRODUCTION

About 75% of children and adolescents presenting with

malignant disease can now be cured in developed countries where

about 1 in every 715 young adults is a long-term survivor [1]. The

increasing use of chemotherapy has led to the emergence of late

and sometimes irreversible adverse effects of treatment, causing

persistent ill health and increasing the risk of early death [2]. It is

important to document carefully very long-term toxicity in order to

design treatment regimens that will minimize chronic toxicity

whilst still maximizing the chance of cure, and to inform the

development of appropriate follow-up strategies for long-term

survivors [3].

Ifosfamide was introduced into clinical practice in the early

1970s and soon became an important component of combination

chemotherapy regimens for many solid tumors in both adults and

children [4]. Initial descriptions of glomerular and tubular toxicity

in adults [5] were followed by numerous case reports and

subsequent studies of severe nephrotoxicity in children [6–18].

However, 20 years later, there are still no published data about the

very long-term effect of ifosfamide on renal function.

The aims of this prospective longitudinal single center cohort

study were to perform detailed evaluation of renal toxicity at the

completion of ifosfamide treatment in childhood, and again 1 and

10 years later, thereby allowing documentation of changes

with time, and to ascertain risk factors for long-term nephrotoxicity.

METHODS

Patients and Treatment

Twenty-nine children commencing ifosfamide in Newcastle

between 1986 and 1996 survived at least 10 years after completion

of treatment. One patient moved away from the region and three

declined investigation. The other 25 (16 males) were studied. No

patient commencing ifosfamide during this time period died from

nephrotoxicity. The median (range) age at treatment commence-

ment was 6.0 (0.6–14.7) years. Twelve children had rhabdomyo-

sarcoma, six a soft tissue sarcoma, six Ewing’s sarcoma, and one a

soft tissue primitive neuroectodermal tumor. All patients had

normal renal function (as demonstrated by normal serum creatinine,

bicarbonate, and phosphate concentrations) prior to ifosfamide

treatment or during treatment but before the onset of nephrotoxicity,

whilst none had radiological evidence of urinary tract obstruction or

renal involvement at initial diagnosis of malignancy.

The median total ifosfamide dose was 106 (12–153) g/m2,

administered intravenously every three weeks at 2–3 g/m2/day for

2–3 days (6–9 g/m2/course), with equidose mesna and hyper-

hydration. Other chemotherapy included melphalan in two patients

(during autologous marrow rescue procedures), actinomycin D,

cyclophosphamide, doxorubicin, etoposide, and vincristine. Drugs

used for supportive care included aminoglycosides (with therapeu-

tic level monitoring) and amphotericin.

No child received cisplatin or carboplatin, and none underwent

nephrectomy. A small area of kidney was included in the

radiotherapy fields in two children, and one patient received total

body irradiation (TBI) (12 Gy).

Background. This prospective longitudinal single institutioncohort study evaluated the natural history of and risk factors forchronic nephrotoxicity 10 years after ifosfamide treatment inchildhood. Procedure. Twenty-five patients (16 males) treated withifosfamide were investigated at end of treatment (End), 1 and10 years later. Glomerular filtration rate (GFR), serum phosphate(PO4) and bicarbonate (HCO3) and renal tubular threshold forphosphate (Tmp/GFR) were measured, and total nephrotoxicity score(Ns) graded. Results. More patients had a low GFR at 1 (72%) and 10(50%) years than at End (26%) (P¼ 0.006 for End vs. 1 year).Electrolyte supplementation requirements for tubular toxicityresolved by 10 years (0% vs. 32% at End and 24% at 1 year; bothP< 0.05). At 10 years, 17% of patients had moderate overallnephrotoxicity and 13% clinically significant reduction of GFR

(<60 ml/min/1.73 m2). Neither dose nor age at treatment predictedany measure of toxicity at 10 years or reduced GFR at any timepoint.Higher cumulative ifosfamide dose correlated with greater tubularand overall nephrotoxicity at End and/or 1 year (P<0.05 for each ofPO4, HCO3, Tmp/GFR, Ns), but age at treatment did not differbetween patients with normal or abnormal results. Conclusions.Although clinically significant tubular toxicity had resolved by10 years, GFR was <60 ml/min/1.73 m2 in 13% of patients, raisingconcerns about very long-term glomerular function. Higher cumu-lative dose was associated with greater tubular and overall toxicity atEnd and 1 year, but not at 10 years. Age at treatment did not predictnephrotoxicity at any timepoint. Pediatr Blood Cancer 2010;54:983–989. � 2010 Wiley-Liss, Inc.

Key words: child; drug toxicity; ifosfamide; kidney; late sequelae

� 2010 Wiley-Liss, Inc.DOI 10.1002/pbc.22364Published online in Wiley InterScience(www.interscience.wiley.com)

——————1Department of Paediatric and Adolescent Oncology, Newcastle upon

Tyne Hospitals NHS Foundation Trust, Royal Victoria Infirmary,

Newcastle upon Tyne, United Kingdom; 2Sir James Spence Institute of

Child Health, School of Clinical Medical Sciences, University of

Newcastle upon Tyne, Royal Victoria Infirmary, Newcastle upon Tyne,

United Kingdom; 3Northern Institute for Cancer Research, University

of Newcastle upon Tyne, Newcastle upon Tyne, United Kingdom;4Institute of Cancer Research, Royal Marsden Hospital, Sutton, Surrey,

United Kingdom

Conflict of interest: Nothing to declare.

Grant sponsor: Special Trustees of Newcastle Health Authority; Grant

sponsor: North of England Children’s Cancer Research Fund.

*Correspondence to: Roderick Skinner, Department of Paediatric and

Adolescent Oncology, Royal Victoria Infirmary, Newcastle upon Tyne

Hospitals NHS Foundation Trust, Queen Victoria Road, Newcastle

upon Tyne NE1 4LP, United Kingdom.

E-mail: roderick.skinner@ncl.ac.uk

Received 6 August 2009; Accepted 20 October 2009

Assessment of Renal Function

Glomerular filtration rate (GFR) and proximal tubular function

(serum phosphate [PO4] and bicarbonate [HCO3] concentrations,

renal tubular threshold for phosphate [Tmp/GFR]) were assessed

[17], and nephrotoxicity grading calculated [19], using a stand-

ardized protocol [20] at treatment completion (End), 1 and 10 years

later. Age-related reference ranges were used [17,20–22]. A small

number of measurements were not performed successfully at some

of the timepoints. Out of 25 children entered into the study, results

were available for 21–23 (depending on the investigation) at End,

and 24–25 at 1 and 10 years. Clinically significant complications of

chronic nephrotoxicity (GFR< 60 ml/min/1.73 m2, hypophospha-

taemic rickets [HR], renal tubular acidosis [RTA], nephrogenic

diabetes insipidus [NDI], growth impairment, and hypertension)

were recorded using standard definitions [19], and ongoing

treatment with electrolyte supplements noted.

The study protocol was approved by the Joint Ethics Committee

of Newcastle Health Authority and the University of Newcastle.

Informed consent was obtained from the parents and, where

appropriate, patient.

Statistics

For analysis of changes with time, the results of renal function

investigations were expressed both as raw values and categorized as

either normal or abnormal (i.e., below the age-related reference

range). Ns was defined as normal when categorized as ‘‘no’’ or

‘‘mild’’ nephrotoxicity (Ns 0–3) and abnormal when categorized as

‘‘moderate’’ or ‘‘severe’’ nephrotoxicity (Ns� 4). Changes between

End and 1 year, End and 10 years, and 1 and 10 years were analyzed

by the Wilcoxon sign rank test for the paired raw values and the

McNemar’s test for paired proportions for the percentages of normal

and abnormal results, and the presence or absence of a severe

nephrotoxicity score (Ns� 8), clinically significant complications,

or ongoing supplementary treatment. Statements describing per-

centages (and changes in percentages) of normal results apply to

paired data only.

Analysis of the relevance of cumulative ifosfamide dose as a risk

factor for the presence of nephrotoxicity was performed using linear

regression, with GFR, PO4, HCO3, Tmp/GFR, and Ns as outcome

measures. Linear regression was also used to evaluate age as a risk

factor for GFR, since the lower limit of the GFR reference range

does not vary with age. However, since the reference ranges of PO4,

HCO3, and Tmp/GFR are age-related, whilst Tmp/GFR and HCO3

are components of calculated Ns, the mean ages of those survivors

with normal and those with abnormal results for these variables was

compared using unpaired t-tests.

RESULTS

The 1- and 10-year studies were performed at a median (range) of

1.1 (0.9–2.1) and 10.5 (9.3–11.4) years, respectively, after

ifosfamide completion. Table I summarizes the overall results of

renal function investigations from all patients (i.e., not just paired

data) and the percentages of normal results at End, 1 and 10 years.

Table II shows changes with time expressed as the differences in

renal function investigations and in the percentage of normal results

between paired data from End and 1 year, End and 10 years, and 1

and 10 years.

Changes in Renal Function Investigations

Considerable inter-individual variability was observed with

some survivors showing substantial deterioration and others marked

improvement in glomerular, tubular or overall renal function over

10 years (Fig. 1). However, except where stated, changes with time

in the cohort as a whole were not statistically significant.

Glomerular Function

The percentage of survivors with a GFR below the reference

range increased significantly between End (26%) and 1 year (70%)

(P¼ 0.006). At 10 years, 50% of survivors still had a low GFR

although the difference from End was not significant (P¼ 0.22).

Proximal Tubular Function

Although 22% of survivors had low PO4 at End and 1 year,

compared to only 8% by 10 years, there was no significant difference

in the percentage with normal PO4, HCO3, or Tmp/GFR at

each timepoint.

Nephrotoxicity Grading

Moderate nephrotoxicity (Ns 4–7) was observed in 24%, 24%,

and 17%, and severe nephrotoxicity (Ns � 8) present in 10%, 8%,

and 0% of survivors at End, 1 and 10 years, respectively (P> 0.05 in

both cases).

Pediatr Blood Cancer DOI 10.1002/pbc

TABLE I. Renal Function Investigations at end of, 1 and 10 Years After Ifosfamide Treatment

End of treatment

(n¼ 21–23)a

One year after

treatment

(n¼ 24–25)a

Ten years after

treatment

(n¼ 24–25)a

Percentage of normal results

End of treatment

One year after

treatment

Ten years after

treatment

GFR (ml/min/1.73 m2) 101 (65–147) 82 (59–131) 88 (40–151) 74 (52–90) 28 (12–49) 50 (29–71)

Serum PO4 (mmol/L) 1.20 (0.43–1.61) 1.19 (0.57–1.62) 1.07 (0.74–1.58) 78 (56–93) 72 (51–88) 92 (74–99)

Serum HCO3 (mmol/L) 21.9 (18.0–27.0) 21.4 (8.5–28.0) 23.9 (20.0–28.0) 65 (43–84) 64 (43–82) 72 (51–88)

Tmp/GFR (mmol/L) 0.90 (0.19–1.41) 0.85 (0.01–1.46) 0.85 (0.43–1.49) 52 (30–74) 50 (29–71) 38 (19–59)

Nephrotoxicity score 2.4 (0–9) 2.7 (0–9) 1.9 (0–6) 67 (43–85) 68 (46–85) 83 (63–95)

Values of the renal function investigations are expressed as median (range). Percentage of patients with normal renal function investigations are

expressed as percentage (95% confidence limits). an¼ the number of patients with results of investigations from each timepoint (not all

investigations were performed at every timepoint in every patient).

984 Skinner et al.

Clinically Significant Complications and ElectrolyteSupplement Treatment

Clinically significant glomerular toxicity (GFR� 60 ml/min/

1.73 m2) was not observed in any child at End, but was present in 4%

at 1 year and 13% at 10 years. HR was documented in 20%, 16%,

and 0% at End, 1 and 10 years, respectively (End vs. 10 years

P¼ 0.06). RTA was present in 8% children at 1 year, but not

observed at the other timepoints. Neither NDI nor growth impair-

ment were observed. One survivor (4%) was receiving treatment for

stage 2 hypertension at 10 years.

Thirty-two percent of children were receiving electrolyte

supplements at End (PO4 in 28%, potassium in 8%), 24% at 1 year

(PO4 in all 24%, with additional HCO3, potassium, calcium and 1a-

cholecalciferol in one [4%]), but none at 10 years (End vs. 10 years

P¼ 0.008, 1 vs. 10 years P¼ 0.03).

Risk Factors for Nephrotoxicity

No correlation was observed between cumulative ifosfamide

dose or age at treatment and GFR at any timepoint. However, at End,

higher dose was correlated to increased tubular toxicity measured

by lower PO4 (P¼ 0.03) and HCO3 (P¼ 0.002). An increase in

cumulative ifosfamide dose of 36 g/m2 (i.e., four courses of 9 g/m2/

course) was associated with a fall in PO4 of 0.14 (95% confidence

intervals 0.02–0.25) mmol/L, and in HCO3 of 1.18 (0.53–

1.82) mmol/L. A similar overall pattern was seen at 1 year, with

Pediatr Blood Cancer DOI 10.1002/pbc

TA

BL

EII

.C

ha

ng

esin

Ren

al

Fu

nct

ion

Inv

esti

ga

tio

ns

Bet

wee

nen

do

f,1

or

10

Yea

rsA

fter

Ifo

sfa

mid

eT

rea

tmen

t

Ch

ang

efr

om

end

of

trea

tmen

tto

1y

ear

po

st-t

reat

men

t

(n¼

21

–2

3)a

Ch

ang

efr

om

end

of

trea

tmen

tto

10

yea

rs

po

st-t

reat

men

t

(n¼

21

–2

3)a

Ch

ang

efr

om

1y

ear

to1

0y

ears

po

st-t

reat

men

t

(n¼

24

-25

)a

Per

cen

tag

eo

fp

atie

nts

wit

hn

orm

alre

sult

s

En

d1

yea

rP

En

d1

0y

ears

P1

yea

r1

0y

ears

P

GF

R(m

l/m

in/1

.73

m2)

�1

7.5

(�2

4.5

,�

11

.5)

�1

1.5

(�2

1.0

,1

.5)

5.5

(�2

.0,

12

.0)

74

30

0.0

06

73

55

0.2

22

95

00

.13

Ser

um

PO

4(m

mo

l/L

)0

.00

(�0

.09

,0

.09

)�

0.2

0(�

0.3

6,

0.0

0)

�0

.17

(�0

.29

,0

.01

)7

87

81

.00

78

91

0.3

87

29

20

.13

Ser

um

HC

O3

(mm

ol/

L)

0.0

(�2

.0,

1.5

)2

.0(0

.5,

3.5

)2

.0(0

.0,

4.0

)6

56

11

.00

65

74

0.7

36

47

20

.77

Tm

p/G

FR

(mm

ol/

L)

�0

.08

(�0

.20

,0

.09

)�

0.1

7(�

0.3

8,

0.2

0)

�0

.05

(�0

.28

,0

.13

)5

25

21

.00

52

33

0.3

95

03

80

.58

Nep

hro

tox

icit

ysc

ore

0.5

(�1

.0,

1.0

)0

.0(�

1.5

,1

.0)

�0

.5(�

2.0

,0

.5)

67

71

1.0

06

78

10

.45

71

83

0.5

1

Pai

red

dat

ao

nly

.Ch

ang

esin

ren

alfu

nct

ion

inves

tig

atio

ns

are

exp

ress

edas

med

ian

dif

fere

nce

(95

%co

nfi

den

celi

mit

sfo

rd

iffe

ren

ce).

An

egat

ive

sig

nin

dic

ates

are

du

ctio

nin

the

val

ue

of

the

var

iab

le

fro

mth

efi

rst

toth

ese

con

dti

mep

oin

t.an¼

the

nu

mb

ero

fp

atie

nts

wit

hre

sult

so

fin

ves

tig

atio

ns

fro

mea

chti

mep

oin

t(n

ot

all

inves

tig

atio

ns

wer

ep

erfo

rmed

atev

ery

tim

epo

int

inev

ery

pat

ien

t). a

b

0

20

40

60

80

100

120

140

160

10 years1 yearEnd

Time

GFR

(ml/m

in/1

.73m

2 )0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

10 years1 yearEnd

Time

Tmp/

GFR

(mm

ol/l)

Fig. 1. Renal function, measured by (a) GFR and (b) Tmp/GFR, at the

end of, and 1 and 10 years after, ifosfamide treatment. Joined lines show

data from individual patients, and demonstrate considerable inter-

patient variability in the severity of renal toxicity and in changes

with time.

Ten-Year Follow-Up of Ifosfamide Nephrotoxicity in Childhood 985

higher dose correlated to lower PO4 (P¼ 0.02) and Tmp/GFR

(P¼ 0.008), and also increased Ns (P¼ 0.01). In contrast, at

10 years post-treatment, no correlation existed between dose and

any aspect of nephrotoxicity (P¼ 0.85, 0.69 and 0.79, respectively,

for PO4, HCO3, and Tmp/GFR) (Fig. 2). An increase in ifosfamide

dose of 36 g/m2 was associated with much smaller falls in PO4

(0.009 mmol/L) and HCO3 (0.17 mmol/L) with wide 95% CI spread

around zero (PO4 �0.081 to 0.098, HCO3 �0.70 to 1.04).

There was no significant difference between the mean age of

survivors with normal and those with abnormal PO4, HCO3, Tmp/

GFR, and Ns at any timepoint. The patient with the lowest GFR

(40 ml/min/1.73 m2) at 10 years had received radiotherapy to a small

area of his left kidney, and was also the only individual requiring

antihypertensive treatment during follow-up. Another patient with a

GFR of 55 ml/min/1.73 m2 at 10 years had received melphalan.

None of the other patients suffering from clinically significant

nephrotoxicity (i.e., GFR< 60 ml/min/1.73 m2, HR, RTA, or hyper-

tension) received melphalan, nor radiotherapy to a field including

the kidney (including TBI).

DISCUSSION

Ifosfamide may lead to chronic glomerular, proximal or distal

tubular toxicity, with a very wide range of severity [19]. Emergence

Pediatr Blood Cancer DOI 10.1002/pbc

a

0 20 40 60 80 100 120 140 1600.0

0.4

0.8

1.2

1.6

2.0

Ifosfamide dose (g/m2)

PO4

(mm

ol/l)

at E

nd

0 40 80 120 1605

10

15

20

25

30

Ifosfamide dose (g/m2)

HC

O3

(mm

ol/l)

at E

nd

b

0 40 80 120 1600.0

0.4

0.8

1.2

1.6

2.0

Ifosfamide dose (g/m2)

PO4

(mm

ol/l)

at 1

yea

r

0 40 80 120 1605

10

15

20

25

30

Ifosfamide dose (g/m2)

HC

O3

(mm

ol/l)

at 1

yea

r

c

0 40 80 120 1600.0

0.4

0.8

1.2

1.6

2.0

Ifosfamide dose (g/m2)

PO4

(mm

ol/l)

at 1

0 ye

ars

0 40 80 120 1605

10

15

20

25

30

Ifosfamide dose (g/m2)

HC

O3 (

mm

ol/l)

at 1

0 ye

ars

Fig. 2. Scatter plots and regression lines showing relation between cumulative dose of ifosfamide received and serum PO4 and HCO3 at (a) the end

of, (b) 1 and (c) 10 years after, ifosfamide treatment. Higher cumulative ifosfamide dose predicts greater tubular toxicity at End (P¼ 0.03 for PO4,

P¼ 0.002 for HCO3) and 1 year (P¼ 0.02 for PO4), but not at 10 years.

986 Skinner et al.

of the Fanconi syndrome may lead to serious clinical consequences,

notably HR, osteomalacia in adults, RTA, growth impairment, and

(rarely) NDI [16]. Fortunately, although ifosfamide-induced end-

stage renal failure has been reported in both adults and children [10],

it appears to be very rare in children [23]. However, there are no

published data about the frequency of nephrotoxicity 10 years after

treatment with ifosfamide. Therefore, it is not known whether

nephrotoxicity is truly persistent, or alternatively partially or

completely reversible in some or all patients.

A few isolated survivors in three cross-sectional studies and one

case series have demonstrated very long-term ifosfamide-induced

nephrotoxicity more than 5 years after treatment in childhood or

adolescence [24–27], but its prevalence remains unknown. A cross-

sectional study of 13 adults at a median of 11.7 years after

ifosfamide (median dose 72 g/m2) in childhood described no

glomerular toxicity, but subclinical proximal tubular toxicity

occurred in three survivors and hypertension in two, although it

was not stated whether these patients had ever experienced

clinically significant nephrotoxicity [28]. However, nearly all of

the larger studies describing chronic ifosfamide nephrotoxicity have

investigated survivors on one occasion only after treatment

completion [12,13,15,17]. The few studies that have incorporated

serial post-treatment studies in individual survivors have reported

either no or limited reversibility of renal damage over the first

few years. Deterioration in glomerular function after completion of

treatment has been described in up to 15% of survivors [25,26],

whilst a decline in median GFR from 92 to 63 ml/min/1.73 m2 has

been described during median follow-up of 31 months in 75

survivors who had been treated with a combination of ifosfamide

and cisplatin [14].

The results of a preliminary analysis of changes in renal function

with time derived from the first 12 patients in this study have been

outlined briefly in a meeting proceedings report [29]. However, the

conclusions of the study have evolved considerably since this

preliminary analysis, and this is the first description of the

methodology and full results, including the risk factor analysis,

from the completed cohort of 25 survivors.

This prospective longitudinal single institution cohort study

utilized a well-validated investigation protocol [20], based on

accurate evaluation of the most frequent and important aspects of

glomerular and tubular impairment by well-accepted methods used

in many previous studies [6,7,9,16,17,24,25]. The grading score was

developed to quantify clinically relevant nephrotoxicity [19]. It

provides evidence of very long-term persistence of nephrotoxicity,

with moderate toxicity (i.e., Ns� 4) still present in 17% of survivors

10 years post-treatment. Although the frequency and severity of

overall nephrotoxicity did not change significantly (except for an

increase in glomerular toxicity between End and 1 year), consid-

erable inter-individual variability in renal function over 10 years

was observed. For example, significant glomerular toxicity

(GFR< 60 ml/min/1.73 m2) was not seen in any patient at End,

but was present in 13% at 10 years. In contrast, severe

nephrotoxicity (Ns� 8), HR, RTA, and ongoing electrolyte

supplementation, which were present in 8–24% of survivors at

1 year had all resolved completely by 10 years. Overall, these

observations suggest that significant improvement in tubulopathy is

likely during long-term follow-up but that substantial deterioration

in glomerular function may occur in some patients.

Glomerular damage may develop gradually during or after

treatment [25]. The incidence of significant chronic glomerular

toxicity in children studied within 5 years of treatment completion

has ranged from 1.4% [11] to about 30% [7,16]. This variability may

reflect differences in patient groups, treatment (especially cumu-

lative dose), and in methodology and timing of investigation. For

example, directly measured GFR (used in this study) is a more

sensitive indicator of glomerular toxicity in children than either

serum creatinine concentration or GFR derived from creatinine

clearance measurement or calculated from serum creatinine [30].

This study confirmed previous reports that GFR falls in some

individuals after completion of ifosfamide [25]. Indeed, the low

GFRs seen in 50% of survivors at 10 years, and particularly the

clinically significant reductions (�60 ml/min/1.73 m2) in 13%, raise

concerns about the implications for very long-term renal function

after ifosfamide treatment in childhood. One patient appeared to

show a substantial increase in GFR between 1 and 10 years but it is

unclear whether this was due to genuine improvement in glomerular

function or conceivably hyperfiltration due to ongoing glomerular

damage. Further follow-up of this patient and study of a larger

cohort may clarify this uncertainty.

The reported prevalence of chronic proximal tubular toxicity

during the first 5 years after ifosfamide treatment ranges from

subclinical glycosuria in about 90% [17] to a Fanconi syndrome,

HR, and/or proximal RTA in 7–25% [14,16,24]. A few case reports

in children have described either improvement or deterioration in

tubular function after completion of treatment. A large cross-

sectional study reported that about 50% of 59 children with

ifosfamide nephrotoxicity who required phosphate or bicarbonate

supplements were still receiving them 5 years after completion of

chemotherapy [27]. In contrast, the present study demonstrated full

resolution of clinically significant tubular toxicity and the need for

electrolyte supplementation during 10 years follow-up.

Several risk factors have been documented for the presence of

ifosfamide nephrotoxicity during the first few years after treatment

[10], notably higher cumulative ifosfamide dose [12,17], previous

or concurrent cisplatin treatment [11,13], prior nephrectomy [13],

and less consistently young age at treatment [12,15]. Potentially

nephrotoxic supportive treatments (e.g., aminoglycosides, ampho-

tericin) have not been shown to be significant risk factors in terms of

the overall frequency of nephrotoxicity [17], although they may be

relevant in some individual patients. Two small studies have failed

to provide any evidence that the ifosfamide administration schedule

(short or prolonged infusion) [31] or any pharmacokinetic measure

[32] predicts subsequent chronic nephrotoxicity.

An important finding of this study was that risk factors varied

between the early (End and 1 year) and late (10 years) timepoints.

Greater early tubular and overall nephrotoxicity was correlated to

higher cumulative dose, confirming previous observations, but a

surprising and previously unreported finding was that this relation-

ship was no longer present at 10 years. Although it could be

suggested that the cohort size was insufficient to detect a dose

toxicity relationship at 10 years, it was large enough to allow clear

demonstration of such associations at End and 1 year. This is

exemplified by the scale (and 95% CI) of the changes in PO4 and

HCO3 associated with an increase in ifosfamide dose of 36 g/m2.

Further investigation in a larger number of survivors is required to

confirm this finding since it has great potential importance in terms

of informing future treatment protocols. In contrast to results from

some early studies with shorter follow-up [12,15], but in keeping

with larger and more recent investigations [14,17], this study found

no evidence that young age was a risk factor for nephrotoxicity. The

Pediatr Blood Cancer DOI 10.1002/pbc

Ten-Year Follow-Up of Ifosfamide Nephrotoxicity in Childhood 987

validity of this finding is strengthened by the use of age-related

reference ranges.

It is unlikely that the inter-individual variability in nephrotox-

icity seen in this study can be attributed entirely to the well-known

effect of dose, or to the impact of the other potential risk factors such

as age at treatment, other nephrotoxic treatments (e.g., radiotherapy,

melphalan). Furthermore, no recognized risk factor appears to be

capable of predicting nephrotoxicity at 10 years, implying that

other, as yet undefined, risk factors (e.g., genetic polymorphisms

involved in ifosfamide metabolism [33]) may also be important. It is

possible that other patient factors such as puberty may influence the

natural history of nephrotoxicity, but this has not been studied

formally.

Documentation of the very long-term prognosis of chronic

ifosfamide nephrotoxicity is particularly relevant in pediatric

practice given the anticipation of cure in most survivors, who will

subsequently face the physiological challenges of growth, pubertal

development, and possibly pregnancy, followed by the expected

age-related decline in renal function. Information about long-term

renal outcome is vital not only to guide follow-up of current

survivors but also to inform the design of future treatment regimens

that balance likely efficacy against the possibility of future severe

toxicity. Despite increased understanding of the pathogenesis of

ifosfamide nephrotoxicity and recognition of the potential for

pharmacologically derived preventative strategies [34], the intro-

duction of clinically effective nephroprotective strategies is unlikely

in the near future. Therefore, in view of the continued use of

ifosfamide in many children, our results urge caution with respect to

the very long-term prognosis for glomerular function but offer a

degree of reassurance regarding tubular toxicity.

This single-center study featured a longitudinal design, very

long-term follow-up and the use of consistent clinical and laboratory

methodology with logical investigation of relevant aspects of renal

function. However, three potential limitations of this study should

be borne in mind. First, the study’s power to detect changes in

nephrotoxicity was limited by a relatively small sample size and the

considerable inter-individual variability, although the conclusions

are strengthened by the fact that 86% of a single center cohort of 10-

year survivors were studied. However, the existence of such marked

variability implies that it is not feasible currently, given contempo-

rary understanding of risk factors for ifosfamide nephrotoxicity, to

predict long-term outcome accurately in individual patients.

Second, the median ifosfamide dose of 106 g/m2 was relatively

high, being above the threshold of 84 g/m2 suggested as being

predictive for a greater risk of nephrotoxicity [17]. Nevertheless,

this study has shown that ifosfamide dose is not a risk factor for very

long-term nephrotoxicity. Finally, the confounding influence of age-

related variability in reference ranges complicated statistical

analysis, although the study design whereby results where

categorized as normal or abnormal was intended to allow mean-

ingful conclusions.

In conclusion, this report of the first prospective longitudinal

study of very long-term ifosfamide nephrotoxicity has revealed

persistent renal damage at 10 years. Although the overall frequency

and severity of nephrotoxicity did not change significantly, its nature

evolved with an increased prevalence of glomerular impairment at

later timepoints, raising concern about very long-term renal

function. In contrast, all clinically significant tubular toxicity had

resolved by 10 years. A novel and potentially very important finding

was that neither dose nor patient age at treatment predicted long-

term ifosfamide nephrotoxicity. There is a continued important need

for prolonged follow-up of a larger number of survivors to clarify the

very long-term outcome for renal function in patients treated with

ifosfamide in childhood, and to identify more informative risk

factors that will facilitate further efforts to reduce the future

prevalence of ifosfamide-induced chronic renal failure.

ACKNOWLEDGMENT

Dr. R. Skinner was an MRC Training Fellow when this study was

commenced. Additional financial support was provided by the

Special Trustees of Newcastle Health Authority and the North of

England Children’s Cancer Research Fund. We are grateful to the

Departments of Medical Physics (Dr. M. Keir, Mr. D. Rodham) and

Clinical Biochemistry (Mr. I. Gibb, Miss. M. Goldfinch) at the

Royal Victoria Infirmary, Newcastle upon Tyne for their assistance

with the investigations, Dr. H. Lucraft for documenting exposure of

renal tissue to radiotherapy, and to the patients and their families for

their support of this study.

REFERENCES

1. Campbell J, Wallace WHB, Bhatti LA, et al. Childhood Cancer in

Scotland: Trends in incidence, mortality and survival 1975–1999.

Edinburgh: NHS Scotland Information and Statistics Division;

2004.

2. Oeffinger KC, Mertens AC, Sklar CA, et al. Chronic health

conditions in adult survivors of childhood cancer. N Engl J Med

2006;355:1572–1582.

3. Skinner R, Wallace WHB, Levitt GA. Long-term follow-up of

people who have survived cancer during childhood. Lancet Oncol

2006;7:489–498.

4. Pratt CB, Douglass EC, Etcubanas EL, et al. Ifosfamide in pediatric

malignant solid tumors. Cancer Chemother Pharmacol 1989;24:

S24–S27.

5. Bremner DN, McCormick JS, Thomson JWW. Clinical trial of

isophosphamide (NSC-109724)—Results and side effects. Cancer

Chemother Rep 1974;58:889–893.

6. Arndt C, Morgenstern B, Wilson D, et al. Renal function in children

and adolescents following 72 g/m2 of ifosfamide. Cancer Chemo-

ther Pharmacol 1994;34:431–433.

7. Ashraf MS, Brady J, Breatnach F, et al. Ifosfamide nephrotoxicity

in paediatric cancer patients. Eur J Paediatr 1994;153:90–94.

8. Burk CD, Restaino I, Kaplan BS, et al. Ifosfamide-induced renal

tubular dysfunction and rickets in children with Wilms tumor.

J Pediatr 1990;117:331–335.

9. Ho PTC, Zimmerman K, Wexler LH, et al. A prospective

evaluation of ifosfamide-related nephrotoxicity in children and

young adults. Cancer 1995;76:2557–2564.

10. Loebstein R, Atanackovic G, Bishai R, et al. Risk factors for long-

term outcome of ifosfamide-induced nephrotoxicity in children.

J Clin Pharmacol 1999;39:454–461.

11. Pratt CB, Meyer WH, Jenkins JJ, et al. Ifosfamide, Fanconi’s

syndrome, and rickets. J Clin Oncol 1991;9:1495–1499.

12. Raney B, Ensign LG, Foreman J, et al. Renal toxicity of ifosfamide

in pilot regimens of the Intergroup Rhabdomyosarcoma Study for

patients with gross residual disease. Am J Pediatr Hematol Oncol

1994;16:286–295.

13. Rossi R, Godde A, Kleinebrand A, et al. Unilateral nephrectomy

and cisplatin as risk factors of ifosfamide-induced nephrotoxicity:

Analysis of 120 patients. J Clin Oncol 1994;12:159–165.

14. Rossi R, Pleyer J, Schafers P, et al. Development of ifosfamide-

induced nephrotoxicity: Prospective follow-up in 75 patients. Med

Pediatr Oncol 1999;32:177–182.

Pediatr Blood Cancer DOI 10.1002/pbc

988 Skinner et al.

15. Shore R, Greenberg M, Geary D, et al. Iphosphamide-induced

nephrotoxicity in children. Pediatr Nephrol 1992;6:162–

165.

16. Skinner R, Pearson ADJ, English MW, et al. Risk factors for

ifosfamide nephrotoxicity in children. Lancet 1996;348:578–

580.

17. Skinner R, Cotterill S, Stevens MCG. Risk factors for nephrotox-

icity after ifosfamide treatment in children: A UKCCSG Late

Effects Group Study. Br J Cancer 2000;82:1636–1645.

18. Stohr W, Paulides M, Bielack S, et al. Ifosfamide-induced

nephrotoxicity in 593 sarcoma pateints: A report from the Late

Effects Surveillance System. Pediatr Blood Cancer 2007;48:447–

452.

19. Skinner R, Sharkey IM, Pearson ADJ, et al. Ifosfamide, mesna

and nephrotoxicity in children. J Clin Oncol 1993;11:173–

190.

20. Skinner R, Pearson ADJ, Coulthard MG, et al. Assessment of

chemotherapy-associated nephrotoxicity in children with cancer.

Cancer Chemother Pharmacol 1991;28:81–92.

21. Barratt TM. Assessment of renal function in children. In: Apley J,

editor. Modern Trends in paediatrics, Vol. 4. London: Butterworth;

1974. pp. 181–215.

22. Stark H, Eisenstein B, Tieder M, et al. Direct measurement of TP/

GFR: A simple and reliable parameter of renal phosphate handling.

Nephron 1986;44:125–128.

23. Rossi R, Kleta R, Ehrich JHH. Renal involvement in children with

malignancies. Pediatr Nephrol 1999;13:153–162.

24. De Schepper J, Hachimi-Idrissi S, Verboven M, et al. Renal

function abnormalities after ifosfamide treatment in children. Acta

Paediatr 1993;82:373–376.

25. Prasad VK, Lewis IJ, Aparicio SR, et al. Progressive glomerular

toxicity of ifosfamide in children. Med Pediatr Oncol 1996;27:

149–155.

26. Loebstein R. Risk factors for long-term outcome of ifosfamide-

induced nephrotoxicity in children. J Clin Pharmacol 1999;39:

454–461.

27. Stohr W, Patzer L, Paulides M, et al. Growth impairment after

ifosfamide-inducd nephrotoxicity in children. Pediatr Blood

Cancer 2007;48:571–576.

28. Mansky P, Arai A, Stratton P, et al. Treatment late effects in long-

term survivors of pediatric sarcoma. Pediatr Blood Cancer 2007;

48:192–199.

29. Skinner R. Chronic ifosfamide nephrotoxicity in children. Med

Pediatr Oncol 2003;41:190–197.

30. Hjorth L, Wiebe T, Karpman D. Correct evaluation of renal

glomerular filtration rate requires clearance assays. Pediatr

Nephrol 2002;17:847–851.

31. Rossi R, Schafers P, Pleyer J, et al. The influence of short versus

continuous ifosfamide infusion on the development of renal tubular

impairment. Int J Pediatr Hematol Oncol 1997;4:393–399.

32. Boddy AV, English MW, Pearson ADJ, et al. Ifosfamide

nephrotoxicity: Limited influence of metabolism and mode of

administration during repeated therapy in paediatrics. Eur J Cancer

1996;32A:1179–1184.

33. McCune JS, Risler LJ, Phillips BR, et al. Contribution of CYP3A5

to hepatic and renal ifosfamide N-dechlorethylation. Drug Metab

Dispos 2005;33:1074–1081.

34. Koren G, Chen N, Aleksa K. Drug-induced nephrotoxicity in

children: Pharmacologically based prevention of long-term impair-

ment. Paediatr Drugs 2007;9:139–142.

Pediatr Blood Cancer DOI 10.1002/pbc

Ten-Year Follow-Up of Ifosfamide Nephrotoxicity in Childhood 989