orig

inal

artic

le

Diabetes, Obesity and Metabolism 12: 204–209, 2010.© 2010 Blackwell Publishing Ltdoriginal article

A low carbohydrate Mediterranean diet improvescardiovascular risk factors and diabetes control amongoverweight patients with type 2 diabetes mellitus: a 1-yearprospective randomized intervention studyA. Elhayany1,2,5, A. Lustman2,3, R. Abel2, J. Attal-Singer4,5 & S. Vinker2,3

1CEO, Meir Medical Center, Kfar Saba, Israel2Department of Family Medicine, Central District Clalit Health Services, Rishon Le Zion, Israel3Department of Family Medicine, Tel Aviv University, Tel Aviv, Israel4Endocrinology Institute, Rabin Medical Center, Beilinson Campus, Petah Tikva, Israel5Department of Family Medicine, The Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel

Background: The appropriate dietary intervention for overweight persons with type 2 diabetes mellitus (DM2) is unclear. Trials comparingthe effectiveness of diets are frequently limited by short follow-up times and high dropout rates.Aim: The effects of a low carbohydrate Mediterranean (LCM), a traditional Mediterranean (TM), and the 2003 American Diabetic Association(ADA) diet were compared, on health parameters during a 12-month period.Methods: In this 12-month trial, 259 overweight diabetic patients (mean age 55 years, mean body mass index 31.4 kg/m2) were randomlyassigned to one of the three diets. The primary end-points were reduction of fasting plasma glucose, HbA1c and triglyceride (TG) levels.Results: 194 patients out of 259 (74.9%) completed follow-up. After 12 months, the mean weight loss for all patients was 8.3 kg: 7.7 kg forADA, 7.4 kg for TM and 10.1 kg for LCM diets. The reduction in HbA1c was significantly greater in the LCM diet than in the ADA diet (−2.0 and−1.6%, respectively, p < 0.022). HDL cholesterol increased (0.1 mmol/l ± 0.02) only on the LCM (p < 0.002). The reduction in serum TG wasgreater in the LCM (−1.3 mmol/l) and TM (−1.5 mmol/l) than in the ADA (−0.7 mmol/l), p = 0.001.Conclusions: An intensive 12-month dietary intervention in a community-based setting was effective in improving most modifiablecardiovascular risk factors in all the dietary groups. Only the LCM improved HDL levels and was superior to both the ADA and TM in improvingglycaemic control.Keywords: cardiovascular risk, diabetes mellitus, dietary intervention, Mediterranean diet

Date submitted 16 April 2009; date of first decision 29 July 2009; date of final acceptance 07 August 2009

IntroductionThe prevalence of diabetes among the Israeli population in2000 was 3.4% [1]. According to the Centers for DiseaseControl and Prevention (CDC), the prevalence of diabetesin the USA in 2004 was 5.1% [2]. Weight reduction inoverweight patients with type 2 diabetes mellitus (DM2)has been proven to be effective in diabetes treatment [3,4].Proper dietary intervention can restore insulin sensitivity andprevent disease progression and complications [5,6]. However,recommendations differ as to which diet is more effectivein achieving weight loss and glycaemic control and reducingcardiovascular risk factors.

In 2004, the American Diabetic Association (ADA)recommended that carbohydrates and monounsaturated fats(MUFA) together should provide 60–70% of energy intake [7].

Correspondence to: Asher Elhayany, Meir Medical Center, 59 Tchernichovsky Street, KfarSaba 44821, Israel.E-mail: elasher@clalit.org.il

The recommendations focus on the quantity of carbohydratesconsumed, but not on the source or type of carbohydrate.

Diets based on carbohydrates with a low glycaemicindex (LGI) have been found to improve HbA1c levelsand reduce insulin resistance, as well as lower cholesterol,triglyceride (TG) and very low-density lipoprotein cholesterolconcentrations [8]. However, the relative importance ofglycaemic index and of glycaemic load as well as the relativeadvantages of higher fat diets and the type of dietary fat vs.diets with high carbohydrate content are debated [9–11]. Thelack of consensus is largely due to the relatively small numberof participants in some studies and the difficulties with patientfollow-up over long periods particularly in community-basedstudies.

Previous studies [12–15] have suggested that a diet enrichedin MUFA elevates blood levels of high-density lipoproteincholesterol (HDL-C), reduces low-density lipoprotein choles-terol (LDL-C) and TGs in DM2 and has a positive effect oncardiovascular disease risk factors [16,17].

DIABETES, OBESITY AND METABOLISM original articleGiven the unclear data, and the lack of community-based

long-term clinical trials, our group conducted a largecommunity-centered interventional study comparing threetypes of isocaloric diets in overweight DM2 patients.

We compared the 2003 ADA diet (50–55% carbohydrates,30% fats and 20% proteins) with a traditional Mediterranean(TM) diet (50–55% LGI carbohydrates, 30% fats–highin monounsaturated fat content, 15–20% proteins) anda low carbohydrate Mediterranean (LCM) diet (35% LGIcarbohydrates, 45% fats–high in monounsaturated fat content,15–20% proteins). The main outcome measures wereglycaemic control and biomarkers for cardiovascular risk.

MethodsThe study population has been described elsewhere [18].Briefly, 259 eligible patients with DM2 were enrolled betweenMarch 2003 and April 2004. Eligible patients were recruitedfrom 10 urban primary care clinics in Israel’s central regionusing the electronic medical records of Clalit Health Services, aHealth Maintenance Organization.

Inclusion criteria were: (i) age 30–65 years; (ii) DM2diagnosed within 1–10 years; (iii) body mass index (BMI)27–34 kg/m2; (iv) last HbA1c measurement 7–10%; (v) lastplasma TG level 1.8–4.5 mmol/l; (vi) last serum creatinine<123.2 μmol/l; and (vii) no change in diabetes medication forat least 3 months before entering the study.

Exclusion criteria were: (i) proliferative diabetic retinopathy;(ii) current insulin treatment; (iii) active oncologic orpsychiatric disease; and (iv) uncontrolled hypothyroidism orhyperthyroidism.

We identified 2486 patients with DM2, 1063 (42.8%) ofwhom met the inclusion criteria and 644 (60.6%) of those werewilling to participate. Before the study entry, patients wereasked to repeat laboratory tests and to undergo fundoscopy:244 patients did not comply and an additional 141 wereexcluded as a result of new findings, thus 259 patients wereenrolled (figure 1).

Intervention

Prior to randomization, the 259 eligible patients entered a 2-week maintenance period. During this time, the patients wereasked to continue their usual diet and keep a food intake diary.The patients were then randomly assigned to one of threeisocaloric diets: TM, LCM or ADA.

Patients were followed up by the same dietitian every 2weeks for 1 year. All dieticians followed a structured protocolfor the 24 scheduled meetings and treated patients from eachof the three diet groups. All patients were advised to engage in30–45 min of aerobic activity at least 3 days a week.

Dietary Intervention

Table A1 (Appendix) outlines the composition of each diet.The daily recommended intake of calories, protein, sodium(up to 3000 mg), potassium (more than 3000 mg), calcium(approximately 1300 mg) and magnesium (more than 800 mg)were similar in the three diet programs. The TM and LCM

diets included only LGI carbohydrates whereas the ADAdiet included mixed glycaemic index carbohydrates. TMand ADA included the same percentages of carbohydrates(50–55%), fat (30%) and protein (15–20%); LCM included35% carbohydrates and 45% fat (50% MUFA). Participantswere counselled to eat 4–6 meals/day according to theirlifestyle.

Dietary Questionnaires

Participants completed a 24-h food recall questionnaire, avalidated food frequency questionnaire (FFQ) and a physicalactivity questionnaire that included quality of life measures, atbaseline, 3 and 6 months [19,20]. Before the study, the dietitiansparticipated in a training workshop to ensure standardizationin questionnaire administration.

Outcome MeasuresClinical Measurements

Weight, height, waist and hip circumferences were measuredat every visit. Blood pressure was obtained monthly as a meanof three readings.

Laboratory Measurements

All tests were performed in the Clalit Health Services CentralDistrict laboratory. Fasting blood glucose, plasma insulin levels,HbA1c, total cholesterol, HDL-C, TGs, liver enzymes, serumcreatinine and urea were measured at baseline and every3 months.

Statistical AnalysisTo compare the differences between diets and changes frombaseline, t-test, one way ANOVA, chi-square or Fisher’s exacttests were used as appropriate. To evaluate the effect of thediets and examine the changes over time, as well as theinteraction between diet and time, we examined the repeatedmeasures using generalized linear models. Measurements takenat baseline, 3, 6, 9 and 12 months were used in the analyses,and TG levels were examined after log transformation. Posthoc Scheffe test was used to identify the statistically differentdiets. All p values were two-sided; p < 0.05 was consideredstatistically significant. All analyses were conducted using SPSS-PC version 15.

ResultsOf the 259 patients enrolled in the study, 85 were randomlyassigned to the ADA diet, 89 to TM, and 85 to the LCMdiet. There were no significant differences at baseline indemographic, clinical or laboratory measurements amongthe three groups (table 1). Cumulative dropouts were 43,56 and 65 patients at 3, 6 and 9 months respectively. Thediscontinuation rates in the three diet groups were similar(figure 1). The 80 individuals who did not complete the 12-month follow-up had, at baseline, statistically significant higher

Volume 12 No. 3 March 2010 doi:10.1111/j.1463-1326.2009.01151.x 205

original article DIABETES, OBESITY AND METABOLISM

Assessed for eligibility n=2486

Enrollment Not meeting inclusion criteria n=1423 Refused to participate n=419 Other reasons (unable to contact, moved, changed to another HMO) n=385 Total exclusions n=2227

Agreed to participate and randomized

n=259

LCM diet n=85 ADA diet n=85 TM diet n=89

ENROLLMENT

Lost to follow-up Up to 3 months n=14 Up to 6 months n= 6 6 to 12 months n= 4

:ue toD Noncompliance n=13

Changed residence n=3 Domestic problems n=4

Unrelated health problems n=2

Other n=2

Lost to follow-up Up to 3 months n=15 Up to 6 months n= 11 6 to 12 months n= 4

:ue toD Noncompliance n= 10

Changed residence n=2 Domestic problems n=3

Unrelated health problems n=7

Other n=2 Incomplete 12-month

follow up data n=6

Lost to follow-up Up to 3 months n=13 Up to 6 months n= 8 6 to 12 months n= 5

:ue toD Noncompliance n= 11

Unrelated health problems n=5

Other n=1 Incomplete 12-month

follow up data n=9

FOLLOW-UP

n=61 n=55 n=63

Completed 12-month follow-up

Figure 1. Flow chart of the study population.

fasting plasma glucose, total cholesterol and LDL-C levels thanpatients who completed the study. No differences were observedin HbA1c, weight, waist circumference, systolic and diastolicblood pressures, HDL-C and TG levels.

To determine patient adherence to the recommended diets,we evaluated the results of the FFQ administered at 6 months.The mean reported energy intake was similar in the three diets(2221.6 calories ± 1086.6). There was a statistically significanttrend in the percentage of polyunsaturated fat intake of totalenergy, from highest (12.9%) in LCM, to 11.5% in TM, andlowest in ADA 11.2% (p = 0.002). The same significant trendwas observed for monounsaturated fat intake (14.6, 12.8, and12.6% for LCM, TM, and ADA, respectively, p < 0.001). Theopposite trend was observed for the percentage of carbohydrateintake, being highest in the ADA, and lowest in the LCMdiet (45.4, 45.2, 41.9 for ADA, TM and LCM, respectively,p = 0.011).

Clinical Measurements

At the end of the 12-month study period, weight, BMI and waistcircumference were reduced in all the dietary interventions withno significant difference between the groups (table 2). HbA1c,

blood lipids and HOMA decreased in all three groups, whilefasting insulin levels increased over time. Differences betweendiets were found for HbA1c, TGs and HDL-C as well as forLDL-C. The reduction in HbA1c was significantly greater forpatients on the LCM than for patients on the ADA diet. Thedecrease in TGs was greater for the LCM and TM than forthe ADA group (p = 0.001). Patients on the LCM achieved asignificant increase in HDL-C levels compared to the ADA, butdid not differ significantly from those on the TM diet. LDL-Clevels were reduced for all three diets after a year; we found thatthe decrease was greatest in LCM, but not significantly greaterthan the TM diet.

Additional analysis was performed including the 80individuals without 12-month data using the latest available of3-, 6- or 9-month follow-up data. The results were comparableto the analysis of those who completed the study.

DiscussionIn recent years, low carbohydrate diets have been shown tobe effective in improving glycaemic control and body weightin type 2 diabetes [21–24] as well as weight loss in obese

206 Elhayany et al. Volume 12 No. 3 March 2010

DIABETES, OBESITY AND METABOLISM original articleTable 1. Baseline characteristics of the patients in the three diet groups

Diet [data presented as mean (±s.d.)]

ADA TM LCMCharacteristics (n = 55) (n = 63) (n = 61) p-value

Male/female (n) 27/28 35/28 31/30 0.764Age (year) 56.0 (6.1) 57.4 (6.1) 55.5 (6.5) 0.229Years since DM

diagnosis5.1 (2.6) 6.2 (9.9) 5.5 (3.8) 0.229

Weight (kg) 87.9 (13.7) 85.5 (10.6) 86.7 (14.3) 0.544BMI 31.8 (3.3) 31.1 (2.8) 31.4 (2.8) 0.483Waist circumference

(cm)113.4 (10.0) 111.1 (9.1) 112.7 (9.6) 0.401

HbA1c (%) 8.3 (0.8) 8.3 (1.0) 8.3 (1.0) 0.980Fasting plasma

glucose (mmol/l)10.3 (1.7) 10.1 (1.8) 10.5 (2.0) 0.454

Fasting plasmainsulin (μU/ml)

12.7 (6.2) 12.1 (6.5) 13.5 (5.7) 0.476

HOMA 5.8 (3.3) 5.0 (2.9) 5.9 (4.0) 0.549Total cholesterol

(mmol/l)5.4 (0.9) 5.5 (0.8) 5.4 (0.9) 0.848

HDL-C (mmol/l) 1.1 (0.2) 1.1 (0.2) 1.1 (0.2) 0.732LDL-C (mmol/l) 3.0 (0.9) 3.2 (0.8) 3.1 (0.8) 0.769Triglycerides

(mmol/l)3.1 (0.8) 3.0 (0.7) 3.2 (0.8) 0.792

patients [25–27]. Based on this increasing evidence, in 2008the ADA included a low carbohydrate diet as a possibleintervention.

We found that an intensive community-based dietaryintervention reduced cardiovascular risk factors in overweightpatients with DM2 for all three diets. The LCM group hadimproved cardiovascular risk factors compared to either theADA or the TM groups.

The weight loss in the LCM group at 1 year was greaterthan in the other dietary intervention groups, although thedifference was not significant. Similar results have been foundamong moderately obese adults [28]. This may be explained

in part by increased consumption of dietary fat in the LCMgroup [29], however, the non-significant differences in caloricintake between groups may have contributed to this weightloss.

The LCM reduced the LDL levels by an extra 8% whencompared to the ADA and TM, which themselves each reducedLDL by 20% (LCM reduced 24.8%, TM 20.9% and ADA13.8%). This finding is of particular note, as other studies usinghigh MUFA diets have not shown significant reductions in LDL[16]. These studies had a follow-up between 4 and 6 weeks. Itis possible that the reduction of LDL in high MUFA diets takesconsiderably longer to manifest than 4–6 weeks and thereforein these earlier studies the full effect of the MUFA diet was notrecorded.

In our study, LCM was the only diet that led to an increasein HDL-C levels (12%). Although the ADA recommendsreducing fat calories to less than 30% of the total daily caloricintake, several recent studies underlined the superiority of dietscontaining higher amounts of MUFA. Garg et al. [30] founda 7–13% increase in HDL-C levels when comparing a dietrich in MUFA with a diet rich in carbohydrates in diabeticpatients. These findings together with our study suggest that alow carbohydrate, high MUFA diet might be the best methodto increase HDL-C, while still reducing LDL.

Previous studies were unable to clarify the effect of a highMUFA diet on HbA1c. In our study, both TM and LCM dietsinduced a greater decrease in HbA1c than did the ADA. Asboth Mediterranean diets were based on LGI carbohydrates,while the reference diet included mixed glycaemic indexcarbohydrates, this study supports the suggested protectiveeffects of LGI carbohydrates on glycaemic control in patientswith DM2.

Consumption of MUFA is thought to improve insulinsensitivity [31–33], an effect that may explain the favourableeffect of the Mediterranean diet on glucose and HbA1c levels.Furthermore, both the Mediterranean and low carbohydrate

Table 2. Results from generalized linear model analysis describing changes in clinical and laboratory measurements between baseline and 12-monthfollow-up examinations

ADA (n = 55) TM (n = 63) LCM (n = 61)

Baseline 12-month Baseline 12-month Baseline 12-month Change over p valueVariables (± s.d.) (± s.d.) (± s.d.) (± s.d.) (± s.d.) (± s.d.) time p value between diets

Weight (kg) 87.8 (13.7) 80.2 (13.2) 85.5 (10.6) 78.1 (9.9) 86.7 (14.3) 77.8 (13.1) <0.001 0.557BMI 31.8 (3.2) 29.0 (3.3) 31.1 (2.8) 28.5 (2.9) 31.4 (2.8) 28.1 (2.8) <0.001 0.359Waist circumference (cm) 113.1 (9.7) 104.0 (10.2) 110.9 (8.9) 101.6 (8.0) 112.6 (9.7) 102.2 (10.2) <0.001 0.388Fasting plasma glucose (mmol/l) 10.26 (1.69) 7.19 (1.85) 10.07 (1.80) 6.57 (1.34) 10.47 (2.00) 6.18 (0.84) <0.001 0.087HbA1c (%) 8.3 (0.8) 6.7 (0.9) 8.3 (1.0) 6.5 (0.8) 8.3 (1.0) 6.3 (1.4) <0.001 0.021∗,†

Total cholesterol (mmol/l) 5.36 (0.89) 4.50 (0.71) 5.46 (0.85) 4.50 (0.69) 5.38 (0.97) 4.50 (0.85) <0.001 0.204LDL-C (mmol/l) 3.05 (0.87) 2.68 (0.83) 3.18 (0.76) 2.63 (0.67) 3.07 (0.81) 2.46 (0.72) <0.001 0.036∗

HDL-C (mmol/l) 1.05 (0.22) 1.00 (0.19) 1.09 (0.17) 1.09 (0.21) 1.08 (0.23) 1.21 (0.21) <0.001 <0.001∗‡

Triglycerides (mmol/l) 3.14 (0.81) 2.26 (1.12) 3.04 (0.68) 1.58 (0.33) 3.18 (0.78) 1.66 (0.36) <0.001 <0.001∗,†

Fasting insulin (μU/ml) 12.8 (6.7) 13.7 (5.3) 11.2 (6.6) 12.4 (6.2) 12.6 (8.4) 14.9 (5.7) 0.007 0.183HOMA 5.85 (3.32) 4.32 (2.07) 4.96 (2.95) 3.69 (1.90) 5.94 (4.00) 4.20 (1.80) <0.001 0.238

∗LCM different than ADA.†TM different than ADA.‡LCM different than TM.

Volume 12 No. 3 March 2010 doi:10.1111/j.1463-1326.2009.01151.x 207

original article DIABETES, OBESITY AND METABOLISM

diets have been shown to reduce postprandial hyperglycaemia,which likely lowers LDL and TGs [34].

Some limitations in our study must be underlined.Potentially eligible patients refused to repeat fundoscopy andblood tests and therefore, were not enrolled. The dropoutrate was relatively high at 31% (80/259), although this wasin keeping with other nutritional studies [35]. Patients werein independent living conditions without close supervision oftheir food intake. However, food questionnaires showed a goodadherence to the assigned diet and patients were followed upevery 2 weeks in an urban primary care clinic.

In conclusion, we found that nutritional therapy in acommunity-based population of overweight DM2 patientswas effective in reducing BMI, HbA1c, LDL and TG levels forall three of the dietary interventions. A low carbohydrateMediterranean diet also increased HDL-C levels and wassuperior in reducing LDL, TG and HbA1c levels comparedto ADA and standard Mediterranean diets. It would appearthat the low carbohydrate Mediterranean diet should berecommended for overweight diabetic patients.

AcknowledgementsThis study was registered in Clinical Trials.gov ProtocolRegistration System: IDNCT00520182. The authors would liketo thank S. Daniel Abraham International Center for Health andNutrition, Ben Gurion University, for assisting in study designand statistical analysis, Faye Schreiber for editorial assistanceand Nava Yelin for helping in statistical analysis.

References1. The Israeli Ministry of Health Statistics. Selected health for all indicators

2003. Morbidity, disability and hospital discharges. Available from URL:http://www.health.gov.il/Download/pages/Morbidity.pdf. Accessed 27January 2009.

2. Centers for Disease Control and Prevention. Crude and age-adjustedpercentage of civilian, noninstitutionalized population with diagnoseddiabetes, United States, 1980–2006. Available from URL: http://www.cdc.gov/diabetes/statistics/prev/national/figage.htm. Last modifiedOctober 8, 2008. Accessed 27 January 2009.

3. Manco M, Mingrone G. Effects of weight loss and calorie restriction oncarbohydrate metabolism. Curr Opin Clin Nutr Metab Care 2005; 8:431–439.

4. Norris SL, Zhang X, Avenell A et al. Long-term non-pharmacologicalweight loss interventions for adults with type 2 diabetes mellitus. CochraneDatabase Syst Rev 2005; Issue 2, Art. no.: CD004095. DOI: 10.1002/14651858.CD004095.pub2.

5. Tuomilehto J, Lindstrom J, Eriksson JG, Valle TT, Hamalainen H, Ilanne-Parikka P. Prevention of type 2 diabetes mellitus by changes in lifestyleamong subjects with impaired glucose tolerance. N Engl J Med 2001; 344:1343–1350.

6. Pan XR, Li GW, Hu YH, Wang IX, Yang WY, An ZX. Effects of diet andexercise in preventing NIDDM in people with impaired glucose tolerance.The Da Qing IGT and Diabetes Study. Diabetes Care 1997; 26: 537–544.

7. American Diabetes Association. Nutrition principles and recommendationsin diabetes. Diabetes Care 2004; 27 (Suppl. 1): S36–S46.

8. Chandalia M, Garg A, Lutjohann D, von Bergman K, Grundy SM, Brinkley LJ.Beneficial effects of high dietary fiber intake in patients with type 2diabetes mellitus. N Engl J Med 2000; 342: 1392–1398.

9. Hung T, Sievenpiper JL, Marchie A, Kendall CWC, Jenkins DJA. Fat versuscarbohydrate in insulin resistance, obesity, diabetes and cardiovasculardisease. Curr Opin Clin Nutr Metab Care 2003; 6: 165–176.

10. Hu FB, van Dam RM, Liu S. Diet and risk of type II diabetes: the role oftypes of fat and carbohydrates. Diabetologia 2001; 44: 805–817.

11. Sargrad KR, Mozzoli M. Effect of high protein vs high carbohydrate intakeon insulin sensitivity, body weight, hemoglobin A1c, and blood pressurein patients with type 2 diabetes mellitus. J Am Diet Assoc 2005; 105:573–580.

12. Garg A, Bonamone A, Grundy SM, Zhang ZJ, Unger RH. Comparison of ahigh-carbohydrate diet with a high-monounsaturated-fat diet in patientswith non-insulin-dependent diabetes mellitus. N Engl J Med 1988; 319:829–834.

13. Campbell LV, Marmot PM, Dyer JA, Borkman M, Storlien LH. The high-monounsaturated fat as a practical alternative for NIDDM. Diabetes Care1994; 17: 177–182.

14. Toobert DJ, Glasgow RE, Strycker LA et al. Biologic and quality-of-lifeoutcomes from the Mediterranean lifestyle program. Diabetes Care 2003;26: 2288–2293.

15. Garg A. High-monounsaturated-fat diets for patients with diabetesmellitus: a meta-analysis. Am J Clin Nutr 1998; 67: 577S–582S.

16. Rodriguez-Villar C, Perez-Heras A, Mercade I, Casals E, Ros E. Comparisonof a high-carbohydrate and a high- monounsaturated fat, olive oil-richdiet on the susceptibility of LDL to oxidative modification in subjects withtype 2 diabetes mellitus. Diabet Med 2004; 21: 142–149.

17. Esposito K, Marfella R, Ciotola M et al. Effect of a mediterranean-style dieton endothelial dysfunction and markers of vascular inflammation in themetabolic syndrome: a randomized trial. JAMA 2004; 292: 1440–1446.

18. Shahar DR, Abel R, Elhayani A, Vardi H, Fraser D. Does calcium intakeenhance weight loss among overweight diabetic patients? Diabetes Care2007; 30: 485–489.

19. Shahar DR, Shai I, Vardi H, Brener-Azrad A, Fraser D. Development of asemi quantitative food frequency questionnaire (FFQ) to assess dietaryintake of multiethnic populations. Eur J Epidemiol 2003; 18: 855–861.

20. Shai I, Rosner BA, Shahar DR et al. DEARR study: dietary evaluation andattenuation of relative risk: multiple comparisons between blood andurinary biomarkers, food frequency, and 24-hour recall questionnaires:the DEARR study. J Nutr 2005; 135: 573–579.

21. Gannon MC, Nuttall FQ. Effect of a high-protein, low-carbohydrate diet onblood glucose control in people with type 2 diabetes. Diabetes 2004; 53:2375–2382.

22. Boden G, Sargrad K, Homko C, Mozzoli M, Stein TP. Effect of a low-carbohydrate diet on appetite, blood glucose levels, and insulin resistancein obese patients with type 2 diabetes. Ann Intern Med 2005; 142:403–411.

23. Ma Y, Olendzki BC, Hafner AR et al. Low-carbohydrate and high-fat intakeamong adult patients with poorly controlled type 2 diabetes mellitus.Nutrition 2006; 22: 1129–1136 [Epub 2006 Oct 4].

24. Nielsen JV, Joensson E. Low-carbohydrate diet in type 2 diabetes. Stableimprovement of bodyweight and glycemic control during 22 monthsfollow-up. Nutr Metab (Lond) 2006; 3: 22–26.

25. Foster GD, Wyatt HR, Hill JO et al. A randomized trial of a low-carbohydratediet for obesity. N Engl J Med 2003; 348: 2082–2090.

26. Stern L, Iqbal N, Seshadri P et al. The effects of low-carbohydrate versusconventional weight loss diets in severely obese adults: 1-year follow-upof a randomized trial. Ann Intern Med 2004; 140: 778–785.

27. American Diabetes Association. Nutrition recommendations and inter-ventions for diabetes: a position statement of the American DiabetesAssociation. Diabetes Care 2008; 31: S61–S78.

208 Elhayany et al. Volume 12 No. 3 March 2010

DIABETES, OBESITY AND METABOLISM original article28. Shai I, Schwarzfuchs D, Henkin Y et al. Dietary intervention random-

ized controlled trial (DIRECT) group. Weight loss with a low-carbohydrate, Mediterranean, or low-fat diet. N Engl J Med 2008; 17:229–241.

29. Tucker LA, Thomas KS. Increasing total fiber intake reduces risk of weightand fat gains in women. J Nutr 2009; 139: 576–581.

30. Garg A, Bantle JP, Henry RR, et al. Effects of varying carbohydrate contentof diet in patients with NIDDM. JAMA 1994; 271: 1421–1428.

31. Esposito K, Marfella R, Ciotola M et al. Effect of a Mediterranean-stylediet on endothelial dysfunction and markers of vascular inflammationin the metabolic syndrome: a randomized trial. JAMA 2004; 292:1440–1446.

32. Schwenke DC. Insulin resistance, low-fat diets, and low-carbohydratediets: time to test new menus. Curr Opin Lipidol 2005; 16: 55–60.

33. Lara-Castro C, Garvey WT. Diet, insulin resistance, and obesity: zoning inon data for Atkins dieters living in South Beach. J Clin Endocrinol Metab2004; 89: 4197–4205.

34. Gheewala NM. Dietary strategies for improving post-prandial glucose,lipids, inflammation, and cardiovascular health. J Am Coll Cardiol 2008;51: 249–255.

35. Jenkins DJ, Kendall CW, McKeown-Eyssen G et al. Effect of a low-glycemicindex or a high-cereal fiber diet on type 2 diabetes: a randomized trial.JAMA 2008; 300: 2742–2753.

Appendix

Table A1. Composition of diets

Diet ADA TM LCM

Carbohydrates∗ (% of TCI) 50 50 35Fibre (g) 15 30 30

Fat∗ (% of TCI) 30 30 45MUFA (% of fat) 10 10 23PUFA (% of fat) 12 12 15SFA 7 7 7

Protein∗ (% of TCI) 20 20 20Calories/kg body weight 20 20 20

MUFA, monounsaturated fats; PUFA, polyunsaturated fats; SFA, saturatedfats.∗Carbohydrates, fat and proteins are 100% of total caloric intake (TCI).

Volume 12 No. 3 March 2010 doi:10.1111/j.1463-1326.2009.01151.x 209