Glycemic Index, Glycemic Load and Their Association with Glycemic Control Among Patients with Type 2 Diabetes

ORIGINAL ARTICLE Glycemic index, glycemic load and their association with glycemic control among patients with type 2 diabetes

Maryam S Farvid1, F Homayouni2, M Shokoohi1, A Fallah1 and Monir S Farvid3

BACKGROUND/OBJECTIVES: The aim was to investigate the associations of glycemic index (GI), glycemic load (GL), carbohydrate and fiber intakes with hyperglycemia in type 2 diabetic patients. SUBJECTS/METHODS: In a cross-sectional study of 640 type 2 diabetic patients aged 28–75 years, usual dietary intakes were assessed by validated food frequency questionnaire. We used published international and Iranian tables of GI based on the white bread. Multivariable logistic regression models were used to estimate odds ratios (ORs) and 95% confidence intervals (CI). RESULTS: High-GL diet was associated with higher risk of hyperglycemia in type 2 diabetic patients after controlling for potential confounders. In multivariable model, OR (95% CI) for the highest vs the lowest quartile of GL was 2.58 (1.08–6.15) for elevated fasting serum glucose (FSG) (4130 mg/dl) (Ptrend ¼ 0.02) and was 3.05 (1.33–7.03) for elevated HbA1c (48.6%) (Ptrend ¼ 0.008). After additional adjusting for dietary fiber and protein intakes, the relation of GL with elevated FSG and HbA1c was stable. GI was not significantly associated with either elevated FSG or HbA1c. In multivariable model, OR (95% CI) for the highest vs lowest quartile of the substitution of dietary carbohydrate for fat intake was 2.32 (1.37–3.92) for elevated HbA1c (Ptrend ¼ 0.001). Higher intake of dietary fiber was associated with lower risk of elevated FSG (highest vs lowest quartile: OR, 0.53; 95% CI: 0.28–0.99; Ptrend ¼ 0.04), but not with lower risk of elevated HbA1c. CONCLUSIONS: GL and carbohydrate intake were positively associated with the risk of hyperglycemia in type 2 diabetic patients; but the benefit in pursuing a low-GI diet without considering carbohydrate and energy intakes in these patients should be further investigated.

European Journal of Clinical Nutrition (2014) 68, 459–463; doi:10.1038/ejcn.2013.288; published online 19 February 2014 Keywords: glycemic index; glycemic load; carbohydrate; ﬁber; hyperglycemia

INTRODUCTION counting, the use of GI and GL may provide a modest additional 3 Diabetes is a chronic metabolic disease associated with long-term benefit in achieving blood glucose goals. On the other hand, the complications resulting from chronic hyperglycemia.1 Nutrition concept of the GI is considered as an important issue in the therapy is an integral part of diabetes care, and carbohydrate guidelines suggested by the European Association for the Study of intake has the greatest impact on improving glycemic control.2 Diabetes that recommends the substitution of low-GI foods with 15 On the basis of American Diabetes Association’s (ADA) high-GI foods. Therefore, the aim of this study was to examine recommendations,3 there is no ideal percentage of energy from the potential association between dietary GI, GL, carbohydrate carbohydrate and other macronutrients for all diabetic patients. and fiber intakes, and the risk of hyperglycemia in men and ADA proposes diabetic patients to monitor total carbohydrate women with type 2 diabetes. intake via carbohydrate counting, exchanges or estimation.2 Foods containing equal amounts of carbohydrate induce a MATERIALS AND METHODS 4,5 different effect on the postprandial blood glucose. The We conducted a cross-sectional study of 751 patients with type 2 diabetes glycemic index (GI), which quantifies the postprandial blood who were randomly recruited via phone call from registered patients with glucose and insulin responses to carbohydrate composition of diabetes at three major diabetes clinics located in Tehran: the Charity diet,6 may have beneficial effects in addition to carbohydrate Foundation for Special Diseases, the Institute of Endocrinology and counting. On the other hand, the concept of glycemic load (GL), Metabolism, and the Iran Diabetes Association. Eligible participants were which represents both the quality and quantity of carbohydrate 25 years old and above with physician-diagnosed type 2 diabetes at least intake,7 has been developed to better represent overall glycemic 1 year before data collection. We did not measure islet cell autoantibodies. effects of a particular food item.8 The effectiveness of low GI Therefore, anyone with diabetes diagnosed before the age of 25 and taking only insulin therapy was considered to be having type 1 diabetes and GL diets in glycemic control has been examined in 9–14 and was excluded from our sample. epidemiological, clinical trials and meta-analysis. However, in Data on medical history, smoking addiction and medication were obtained some studies, diets with low GI or GL had no benefits in the from personal interview. We excluded those who did not meet the following 13,14 management of diabetes. On the basis of the current criteria: taking insulin, altering medication regimen or dietary intake during 3 evidence, the ADA states that combined with carbohydrate months before the study and having abnormal hepatic tests (n ¼ 30). We also

1Department of Community Nutrition, National Nutrition and Food Technology Research Institute, Faculty of Nutrition Sciences and Food Technology, Shahid Beheshti University of Medical Sciences, Tehran, Iran; 2Paramedical School, Jundishapur University of Medical Sciences and Health Services, Ahvaz, Iran and 3Taleghani Hospital, Shahid Beheshti University of Medical Sciences, Tehran, Iran. Correspondence: Dr Maryam S Farvid, Department of Community Nutrition, National Nutrition and Food Technology Research Institute, Faculty of Nutrition Sciences and Food Technology, Shahid Beheshti University of Medical Sciences, Tehran 19395-4741, Iran. E-mail: [email protected] Received 20 May 2013; revised 28 October 2013; accepted 14 November 2013; published online 19 February 2014 GI and GL and glycemic control Maryam S Farvid et al 460 excluded patients who did not live in Tehran (n ¼ 4), did not complete the GL quartiles. Variables were compared across quartile categories of GI and food frequency questionnaire (FFQ) (n ¼ 48), those (n ¼ 11) who reported a GL by one-way analysis of variance with Tukey’s post hoc comparisons total daily energy intake outside the range of 800–4200 kcal and those for quantitative variables and w2-tests for qualitative variables. Logistic (n ¼ 18) who had missing data for confounding variables. After these regression models were used to estimate odds ratio (OR) and 95% exclusions, 640 type 2 diabetic patients aged 28–75 years remained. confidence intervals (CI) for each category using the lowest quartile of Written informed consent was obtained from all participants. The intake as the reference category, while controlling for potential confound- research protocol was approved by the Ethics Committee of National ing variables. FSG o130 mg/dl is recommended as a target of glycemic Nutrition and Food Technology Research Institute. control by ADA2 and the expected HbA1c levels for normal glucose values We collected 20 ml blood samples from each participant between 0800 are 6–8.6% by the commercial HbA1c kit.22 Therefore, we defined and 1000 hours, before taking any oral hypoglycemic agent(s) and after hyperglycemia with cutoff values of 130 mg/dl for FSG and 8.6% for 12–14-h overnight fasting. Aliquots of serum and red pack cells were HbA1c. All models were adjusted for age, sex and energy intake. transferred to polystyrene tubes that were immediately stored at À 70 1C Multivariable models included additional terms for duration of diabetes, until analysis. Fasting serum glucose (FSG) concentration was measured by smoking, physical activity, BMI, vitamin/mineral supplementation, total the glucose oxidase method. HbA1c was measured by a chromatography hypoglycemic medication, blood pressure-lowering drug and lipid- method using the commercial kit (Globe Diagnostics, Rome, Italy). lowering drug. In additional analyses, we further adjusted for dietary The dietary intake of patients was assessed by interview using a 1-year protein and fiber intakes (multivariable model 2). Multivariable model 1 for validated 168-item semi-quantitative FFQ.16 It consisted of a list of foods total carbohydrate intake was fit without adjusting for protein and fat with standard serving sizes according to Iranian meal patterns, and was intake to simulate the substitution of carbohydrate for the average designed to obtain information on usual food intake during the previous mixture of protein and fat in the study population and with the adjustment year. The reported frequency for each food item was then converted to a for protein (multivariable model 2) to simulate the substitution daily intake. Information on frequency of intake and portion size was of carbohydrate with total fat intake.23 To examine whether the converted to the number of grams of each food item consumed on associations between GL and elevated FSG or HbA1c were modified by average per day. To determine the total dietary carbohydrate, protein, fat, other measures of diabetes risk factors, a cross-product term for the level fiber and energy compositions of Iranian foods, we used the United States of each factor and intake of GL expressed as a continuous variable was Department of Agriculture database17 and Iranian food composition tables. included in the multivariable model. P-values for tests for interactions were The validity and reproducibility of dietary GI and GL were similar to those obtained from a likelihood ratio test with 1 degree of freedom. IBM SPSS of nutrients commonly studied in epidemiologic studies with the use of 21 was used for all analyses. All P-values were two-sided. FFQs.18 Of the 168 food and beverage items included in the FFQ, 30 items (17.8%) contain no available carbohydrate. The calculation of dietary GL and GI was thus based on the remaining 138 items with GI values ranging from 10 to 123. We used published international5 and Iranian tables of GI.19 RESULTS We calculated GL values by multiplying the available carbohydrate content The median index across quartile of GL ranged from 108.2 to 246.6 of each food by its GI value and then multiplied the resultant value and across quartile of GI ranged from 50.6 to 66.7. The main with the amount of consumption (divided by 100) and then summed the contributors of GL were bread (31.3%), rice (23.4%), fruits (21.3%) 8 values from all food items. Each unit of dietary GL represents the and sweets (9.8%). Clinical characteristics and nutrient intakes data equivalent of 1 g of carbohydrate from white bread. The overall GI for each relating to the 640 type 2 diabetes patients studied based on GL participant was estimated by dividing the dietary GL by the total amount and GI quartiles are shown in Table 1. The participants with a higher of carbohydrates consumed.8,20 The calculated values of dietary GI, GL and total dietary carbohydrate, protein, fat and fiber were adjusted for total GL were more likely to be younger, to be male, to have higher energy intake using the residual method.21 intakes of energy, carbohydrate, fat, protein and fiber, and to be Height was measured to the nearest 0.1 cm and weight to the nearest smokers than subjects with lower GL. The participants with higher 0.1 kg. Body mass index (BMI) was calculated as weight/height2 (kg/m2). GL also had lower duration of diabetes. In comparison with the Data on physical activity were obtained using modified international subjects in the lowest quartile of the GI, those in the higher physical activity questionnaire (publicly available at http://www.ipaq.ki. quartiles were more likely to be younger, to be male, to be smokers se/ipaq.htm) and expressed as metabolic equivalent h/day (MET-h/day). and to consume more energy, carbohydrate and fat (Table 1). From 640 type 2 diabetes patients, 440 of them had FSG Statistical analysis 4130 mg/dl and 378 of them had HbA1c 48.6%. The ORs of The data were expressed as mean±s.d. or percentages. We divided GI and elevated FSG (4130 mg/dl) and HbA1c (48.6%) according to GL into quartiles and participants were categorized on the basis of GI and quartile categories of GL, GI, total carbohydrate and fiber intakes

Table 1. Characteristics of type 2 diabetic patients according to quartile of glycemic load and glycemic index

Quartile of GL Quartile of GI

1(o129.6) 2 (129.6–171.2) 3 (171.3–212.0) 4 (4212.0) P-valuea 1(o54.1) 2 (54.1–58.7) 3 (58.8–63.5) 4 (463.5) P-valuea

N 160 160 160 160 160 160 160 160 Sex (M/F) 34/126 56/104 82/78 123/37 o0.0001 62/98 68/92 73/87 92/68 0.005 Age (year) 54.6±7.6 54.3±7.7 52.9±7.7 51.1±8.4 o0.0001 53.6±8.1 54.5±7.5 53.4±7.4 51.4±8.6 0.006 Diabetes duration (year) 10.5±6.6 10.1±6.5 9.0±5.9 8.2±5.4 0.004 10.4±6.1 9.7±7.0 9.1±5.5 8.7±5.8 0.06 Current smokers (n) 15 12 21 29 0.02 16 11 19 31 0.005 Metformin (n) 137 122 129 113 0.16 125 132 129 115 0.24 Sulfonylurea (n) 120 111 109 114 0.25 114 107 123 110 0.39 BMI (kg/m2) 28.6±4.1 28.5±4.0 28.4±3.9 28.5±4.5 0.97 28.4±4.1 28.4±3.4 28.6±4.5 28.5±4.5 0.96 FSG (mg/dl) 160±59 156±57 167±58 172±58 0.07 160±57 162±57 166±60 167±58 0.66 HbA1c (%) 9.1±2.2 9.0±1.7 9.1±2.0 9.2±1.9 0.92 9.1±2.2 9.1±1.9 9.1±1.9 9.1±1.9 0.99 Total energy intake (kcal) 1530±280 1930±305 2306±361 2950±528 o0.0001 2049±581 2107±633 2180±627 2382±698 o0.0001 Total carbohydrate intake (g) 197±32 262±31 323±40 422±72 o0.0001 284±92 290±90 298±85 331±106 o0.0001 Total fat intake (g) 66.2±22.1 73.8±27.0 82.7±31.6 103.3±35.9 o0.0001 74.0±29.0 78.3±33.1 81.8±34.6 87.8±33.9 0.002 Protein intake (g) 52.2±14.0 64.7±14.8 77.6±17.3 93.8±21.3 o0.0001 71.7±22.9 71.3±23.8 72.0±22.7 73.4±22.8 0.87 Total ﬁber intake (g) 17.7±6.0 22.2±6.7 26.5±8.7 31.6±10.5 o0.0001 24.3±11.3 25.9±9.4 24.2±8.8 23.7±8.9 0.19 Physical activity (METs-h/day) 32.7±4.6 32.4±4.8 32.7±5.5 32.8±5.1 0.86 33.1±5.2 32.8±4.7 32.7±4.9 32.0±5.1 0.24 Abbreviations: BMI, body mass index; FSG, fasting serum glucose; GI, glycemic index; GL, gycemic load. aOne-way analysis of variance for quantitative variables and w2-test for qualitative variables.

European Journal of Clinical Nutrition (2014) 459 – 463 & 2014 Macmillan Publishers Limited & 04McilnPbihr iie uoenJunlo lnclNtiin(04 5 463 – 459 (2014) Nutrition Clinical of Journal European Limited Publishers Macmillan 2014

Table 2. Multivariate-adjusted ORs (95% CI) of elevated fasting serum glucose (4130 mg/dl) and HbA1c (48.6%) according to quartile categories of glycemic load and glycemic index, carbohydrate and ﬁber intakes

FSG HbA1c

a a Quartile of intake Ptrend Quartile of intake Ptrend

Glycemic load Range of index o129.6 129.6–171.2 171.3–212.0 4212.0 o129.6 129.6–171.2 171.3–212.0 4212.0 No. of cases with hyperglycemia 102 100 119 119 90 97 96 95 Age-, sex- and energy-adjusted modelb 1 1.01 (0.62–1.67) 1.97 (1.07–3.61) 2.20 (0.97–5.01) 0.03 1 1.41 (0.87–2.29) 1.73 (0.99–3.03) 2.29 (1.07–4.89) 0.03 Multivariable-adjusted model 1c 1 1.09 (0.64–1.85) 2.15 (1.13–4.08) 2.58 (1.08–6.15) 0.02 1 1.52 (0.89–2.58) 2.07 (1.12–3.83) 3.05 (1.33–7.03) 0.008 Multivariable-adjusted model 2d 1 1.15 (0.67–1.97) 2.42 (1.26–4.67) 3.00 (1.22–7.33) 0.007 1 1.67 (0.98–2.84) 2.42 (1.29–4.54) 3.94 (1.66–9.31) 0.002

Glycemic index Range of index o54.1 54.1–58.7 58.8–63.5 463.5 o54.1 54.1–58.7 58.8–63.5 463.5 No. of cases with hyperglycemia 105 109 113 113 89 98 96 95 Age-, sex-, and energy-adjusted modelb 1 1.16 (0.72–1.86) 1.25 (0.78–2.02) 1.18 (0.72–1.92) 0.46 1 1.28 (0.81–2.01) 1.27 (0.81–2.00) 1.30 (0.82–2.05) 0.28 c

Multivariable-adjusted model 1 1 1.30 (0.78–2.16) 1.38 (0.83–2.29) 1.41 (0.84–2.38) 0.19 1 1.43 (0.87–2.40) 1.39 (0.85–2.28) 1.42 (0.86–2.35) 0.20 Farvid control S glycemic Maryam and GL and GI Multivariable-adjusted model 2d 1 1.32 (0.79–2.21) 1.34 (0.80–2.25) 1.31 (0.75–2.26) 0.33 1 1.51 (0.91–2.50) 1.45 (0.88–2.39) 1.52 (0.90–2.59) 0.14

Carbohydrate Range of intake (% energy) o50.4 50.4–55.5 55.6–60.3 460.3 o50.4 50.4–55.5 55.6–60.3 460.3 No. of cases with hyperglycemia 106 107 116 111 78 100 100 100 Age-, sex- and energy-adjusted modelb 1 1.11 (0.69–1.78) 1.44 (0.88–2.34) 1.26 (0.78–2.05) 0.23 1 1.76 (1.12–2.76) 1.76 (1.12–2.77) 1.74 (1.10–2.75) 0.02 e Multivariable-adjusted model 1 1 1.12 (0.68–1.87) 1.54 (0.91–2.59) 1.47 (0.86–2.51) 0.09 1 1.94 (1.18–3.18) 2.08 (1.26–3.44) 2.27 (1.35–3.83) 0.002 al et Multivariable-adjusted model 2d 1 1.13 (0.68–1.87) 1.54 (0.91–2.59) 1.47 (0.86–2.51) 0.09 1 1.91 (1.16–3.14) 2.10 (1.27–3.46) 2.32 (1.37–3.92) 0.001

Fiber Range of intake (g) o17.5 17.5–23.0 23.1–30.2 430.2 o17.5 17.5–23.0 23.1–30.2 430.2 No. of cases with hyperglycemia 115 113 101 111 99 106 83 90 Age-, sex- and energy-adjusted modelb 1 0.82 (0.50–1.35) 0.51 (0.30–0.85) 0.53 (0.29–0.96) 0.02 1 1.17 (0.73–1.87) 0.63 (0.39–1.02) 0.72 (0.41–1.25) 0.12 Multivariable-adjusted model 1c 1 0.76 (0.45–1.30) 0.52 (0.30–0.92) 0.53 (0.28–0.99) 0.04 1 1.14 (0.68–1.89) 0.70 (0.41–1.19) 0.81 (0.44–1.50) 0.36 Abbreviation: OR, odds ratio. aCalculated in a separate regression model with median intake in each quartile as a continuous variable and the same group of covariates specified for the corresponding model. bAdjusted for age (o50, 50–55, 55.1–60, 60.1–65, 465), sex, energy intake (kcal). cAge-, sex- and energy-adjusted model further adjusted for duration of diabetes (o5, 5–10, 10.1–15, 415 year), smoking (yes or no), physical activity (METs-h/day), body mass index (p25, 25.1–27, 27.1–30, 30.1–35, X35 kg/m2), vitamin/mineral supplementation (yes or no), total hypoglycemic medication, blood pressure-lowering drug (yes or no) and lipid-lowering drug (yes or no). dMultivariable-adjusted model 1 further adjusted for protein intake (% energy) and fiber intake (g). eSame as (c) in addition to fiber intake (g). 461 GI and GL and glycemic control Maryam S Farvid et al 462 are shown in Table 2. After making adjustment for demographic, Besides epidemiologic studies, an effect of low-GI diet was also anthropometric, medication and lifestyle factors, higher GL was demonstrated by clinical trials. Consumption of low-GI foods associated with elevated FSG risk (the highest vs the lowest instead of traditional or high-GI foods has clinically useful benefits 11 quartile: OR ¼ 2.58, 95% CI: 1.08, 6.15; Ptrend ¼ 0.02) and elevated on glycemic control in diabetic patients. A recent meta-analysis HbA1c risk (the highest vs the lowest quartile: OR ¼ 3.05, 95% CI: of randomized controlled trials conducted in diabetes reported 1.33, 7.03, Ptrend ¼ 0.008) (multivariable model 1, Table 2). To that consuming low-GI diets significantly decreased HbA1c levels explore whether the association of hyperglycemia with the GL is by 0.4% compared with comparison diets.26 Jenkins et al.9 independent of the dietary intakes of fiber and protein, we identified that 6-month treatment with a low-GI diet in type 2 additionally adjusted for dietary fiber and protein intakes (multi- diabetic patients reduced HbA1c levels compared with a high- variable model 2). The odds of having elevated FSG for the highest cereal fiber diet even after controlling for dietary fiber or body quartile vs the lowest quartile of GL were increased by 200% and weight. In contrast, Ma et al.13 compared the effects of a low-GI OR (95% CI) for elevated HbA1c for the highest quartile vs the diet with the standard ADA guidelines on the HbA1c for patients lowest quartile was 3.94 (1.66–9.31, Ptrend ¼ 0.002). The GI was not with type 2 diabetes. They found that the effects of low-GI diet on significantly related to either elevated FSG or elevated HbA1c. HbA1c improvement was comparable with ADA diet. Mayer-Davis Substitution of carbohydrate for average intake of protein or et al. evaluated GI and GL in relation to average fasting glucose, fat (multivariable model 1) was not significantly associated 2 h plasma glucose following 75-g glucose load and glycated with increased risk of elevated FSG, but increased the risk of hemoglobin in adult participants in the insulin resistance elevated HbA1c by 127% in the highest vs the lowest quartile atherosclerosis study. The results did not support the correlation of carbohydrate intake. Because total energy intake, and percent between dietary GI and any measure of glycemia.14 To compare of energy from carbohydrate and protein intakes were included the long-term effects of diets with different GI or carbohydrate simultaneously in the analysis, the OR for carbohydrate can be amount on HbA1c and plasma glucose, type 2 diabetic patients interpreted as the effect of substituting carbohydrate with an received high-carbohydrate and high-GI, high-carbohydrate and equivalent reduction in the percent of energy from total fat. low-GI, or low-carbohydrate and high-monounsaturated fat diets Substitution of carbohydrate for fat intake was significantly for 1 year. Although there has been higher intake of fiber and associated with higher risk of hyperglycemia measured by HbA1c reduction in postprandial glucose by high-carbohydrate and low- (multivariable model 2). Higher dietary fiber intake was related GI diet, HbA1c was not improved by reductions in the GI or with lower risk of elevated FSG (highest vs lowest quartile: OR, carbohydrate intakes.27 0.53; 95% CI: 0.28, 0.99; Ptrend ¼ 0.04). However, higher dietary fiber For GI, we found no association with glycemic control, intake was not associated with lower risk of elevated HbA1c. as assessed by FSG and HbA1c. The present finding of null We also examined whether the associations between GL and association is hardly due to the range of dietary GI. In our study, elevated FSG and HbA1c differed by levels of hyperglycemia risk median index across quartile of GI ranged from 50.6 to 66.7 factors including age, BMI, duration of diabetes and sex. We have and these values did not differ much from the studies that have not found any significant interaction. found a significant association between dietary GI and risk of diabetes.28–31 Energy intake has a positive effect on glycemic control in diabetic patients32 and high intake of dietary fiber is DISCUSSION associated with lower fasting blood glucose and HbA1c.33 However, In this cross-sectional study, we examined the association of after further adjustment for energy or dietary fiber intake, the quality and quantity of dietary carbohydrate intake with risk of results did not change. The association with GL but not GI suggests hyperglycemia in type 2 diabetic patients and found that dietary that the amounts of carbohydrate and GI of foods should be GL in people with type 2 diabetes was positively associated with considered simultaneously, as represented by GL, in recommended risk of hyperglycemia after the adjustment for demographic, diet to effectively manage glycemic control in diabetic patients. anthropometric, lifestyle, medication and dietary factors. These This study had some limitations. We could not infer causality associations were independent of dietary fiber and protein because of the observational nature of our study. We assessed diet intakes. However, no statistically significant association was using a validated FFQ. Measurement errors may be introduced by observed for GI in relation to either FSG or HbA1c; either after the under- or overreporting of the amounts of food groups usually adjustment for potential confounders or after further adjustment eaten per day. GIs were estimated on the basis of the international for dietary factors. In addition, we found that quantity of and Iranian tables of GI. Thus, the effect of varying degree of carbohydrate intake was positively associated with the risk of ripeness, processing and chewing on food GIs is a concern.34 hyperglycemia measured by HbA1c. Higher intake of fiber was Another possible source of measurement error is that we focused also associated with lower FSG, but not HbA1c. only on dietary fiber consumption without regarding the amount of The importance of low GL or GI diets in the management of fiber supplements. However, we would expect these to bias results diabetes is controversial.9–14 A growing body of evidence supports toward the null. In addition, one limitation of the study is the a pivotal role for dietary GI and GL in the prevention of diabetes. possibility of participant’s weight change before including in the In a recent meta-analysis of prospective cohort studies, study and our study only addresses the association controlling for high-GI and/or high-GL diets have been correlated with the risk weight at time of measuring dietary intake and glycemic control of type 2 diabetes (GI: OR ¼ 1.16, 95% CI: 1.06–1.26; GL: OR ¼ 1.20, factors. However, we did not find any interaction between GL and 95% CI: 1.11–1.30).10 In Greenland’s Inuit population, after BMI. This study had several strengths. We measured known potential adjustment for age, sex, BMI, smoking status and energy intake, confounders and were able to control for them in our analyses. GI was positively associated with fasting plasma glucose and In conclusion, our findings suggest that GL which describes HbA1c, although this association was dependent on educational quality and quantity of carbohydrate consumption, and also the level and physical activity and after further adjustment, only the total carbohydrate intake significantly influence glycemic control association with fasting plasma glucose remained statistically in type 2 diabetic patients. But benefit in pursuing a low-GI diet significant. The inverse associations of GL with fasting plasma without considering carbohydrate and energy intakes in patients glucose and HbA1c have been also attenuated after additional with type 2 diabetes should be further investigated. adjustment for BMI, education, smoking status, physical activity and energy intake.24 In contrast, in a Dutch population, neither GI nor GL was associated with fasting glucose and HbA1c after CONFLICT OF INTEREST 25 multivariable adjustment. The authors declare no conflict of interest.

European Journal of Clinical Nutrition (2014) 459 – 463 & 2014 Macmillan Publishers Limited GI and GL and glycemic control Maryam S Farvid et al 463 ACKNOWLEDGEMENTS 17 USDA Nutrient Database for Standard Reference, Release 24: Department of This work was supported by a grant from the National Nutrition and Food Agriculture ARS, 2011. Technology Research Institute, Tehran, Iran. 18 Levitan EB, Westgren CW, Liu S, Wolk A. Reproducibility and validity of dietary glycemic index, dietary glycemic load, and total carbohydrate intake in 141 Swedish men. Am J Clin Nutr 2007; 85: 548–553. REFERENCES 19 Taleban FA, Esmaeili M (eds). Glycemic Index of Iranian Foods. National Nutrition and Food Technology Research Institute Press: Tehran, Iran, 1999. 1 Marcovecchio ML, Lucantoni M, Chiarelli F. Role of chronic and acute hyper- 20 Wolever TM, Nguyen PM, Chiasson JL, Hunt JA, Josse RG, Palmason C et al. glycemia in the development of diabetes complications. Diabetes Technol Ther Determinants of diet glycemic index calculated retrospectively from diet records 2011; 13: 389–394. of 342 individuals with non-insulin-dependent diabetes mellitus. Am J Clin Nutr 2 American Diabetes Association. Standards of medical care in diabetes--2013. 1994; 59: 1265–1269. Diabetes Care 2013; 36: S11–S66. 21 Willett WC, Howe GR, Kushi LH. Adjustment for total energy intake in epide- 3 American Diabetes Association. Standards of medical care in diabetes--2014. miologic studies. Am J Clin Nutr 1997; 65: 1220S–1228S. Diabetes Care 2014; 37: S14–S80. 22 Glycated HbA1c. Chromatographic determination in tube with preweighted resin 4 O’Dea K, Nestel PJ, Antonoff L. Physical factors influencing postprandial glucose of hemoglobin A1c in blood. REF GD541500, GD540500. http://www.gdsrl.com/ and insulin responses to starch. Am J Clin Nutr 1980; 33: 760–765. upload/prodotti/gd5415_00_glycated_hba1c_0.pdf. 5 Foster-Powell K, Holt SH, Brand-Miller JC. International table of glycemic index 23 Hu FB, Stampfer MJ, Manson JE, Rimm E, Colditz GA, Rosner BA et al. Dietary fat and glycemic load values: 2002. Am J Clin Nutr 2002; 76: 5–56. intake and the risk of coronary heart disease in women. N Engl J Med 1997; 337: 6 Jenkins DJ, Wolever TM, Taylor RH, Barker H, Fielden H, Baldwin JM et al. 1491–1499. Glycemic index of foods: a physiological basis for carbohydrate exchange. 24 van Aerde MA, Witte DR, Jeppesen C, Soedamah-Muthu SS, Bjerregaard P, Am J Clin Nutr 1981; 34: 362–366. Jørgensen ME. Glycemic index and glycemic load in relation to glucose intolerance 7 Liu S, Manson JE, Stampfer MJ, Holmes MD, Hu FB, Hankinson SE et al. among Greenland’s Inuit population. Diabetes Res Clin Pract 2012; 97: 298–305. Dietary glycemic load assessed by food-frequency questionnaire in relation to 25 Du H, van der A DL, van Bakel MM, van der Kallen CJ, Blaak EE, van Greevenbroek MM plasma high-density-lipoprotein cholesterol and fasting plasma triacylglycerols in et al. Glycemic index and glycemic load in relation to food and nutrient intake postmenopausal women. Am J Clin Nutr 2001; 73: 560–566. and metabolic risk factors in a Dutch population. Am J Clin Nutr 2008; 87: 8 Willett WC, Manson J, Liu S. Glycemic index, glycemic load, and risk of type 2 655–661. diabetes. Am J Clin Nutr 2002; 76: 274S–280S. 26 Thomas DE, Elliott EJ. The use of low-glycaemic index diets in diabetes control. 9 Jenkins DJ, Kendall CW, McKeown-Eyssen G, Josse RG, Silverberg J, Booth GL et al. Br J Nutr 2010; 104: 797–802. Effect of a low-glycemic index or a high-cereal fiber diet on type 2 diabetes: a 27 Wolever TM, Gibbs AL, Mehling C, Chiasson JL, Connelly PW, Josse RG et al. randomized trial. JAMA 2008; 300: 2742–2753. The Canadian Trial of Carbohydrates in Diabetes (CCD), a 1-y controlled trial of 10 Dong JY, Zhang L, Zhang YH, Qin LQ. Dietary glycaemic index and glycaemic load low-glycemic-index dietary carbohydrate in type 2 diabetes: no effect on glycated in relation to the risk of type 2 diabetes: a meta-analysis of prospective cohort hemoglobin but reduction in C-reactive protein. Am J Clin Nutr 2008; 87: 114–125. studies. Br J Nutr 2011; 106: 1649–1654. 28 Salmeroń J, Ascherio A, Rimm EB, Colditz GA, Spiegelman D, Jenkins DJ et al. 11 Brand-Miller J, Hayne S, Petocz P, Colagiuri S. Low-glycemic index diets in the Dietary fiber, glycemic load, and risk of NIDDM in men. Diabetes Care 1997; 20: management of diabetes: a meta-analysis of randomized controlled trials. 545–550. Diabetes Care 2003; 26: 2261–2267. 29 Krishnan S, Rosenberg L, Singer M, Hu FB, Djousse´ L, Cupples LA et al. 12 Riccardi G, Rivellese AA, Giacco R. Role of glycemic index and glycemic load in the Glycemic index, glycemic load, and cereal fiber intake and risk of type 2 diabetes healthy state, in prediabetes, and in diabetes. Am J Clin Nutr 2008; 87: 269S–274S. in US black women. Arch Intern Med 2007; 167: 2304–2309. 13 Ma Y, Olendzki BC, Merriam PA, Chiriboga DE, Culver AL, Li W et al. A randomized 30 Salmeroń J, Manson JE, Stampfer MJ, Colditz GA, Wing AL, Willett WC. clinical trial comparing low-glycemic index versus ADA dietary education among Dietary fiber, glycemic load, and risk of non-insulin-dependent diabetes mellitus individuals with type 2 diabetes. Nutrition 2008; 24: 45–56. in women. JAMA 1997; 277: 472–477. 14 Mayer-Davis EJ, Dhawan A, Liese AD, Teff K, Schulz M. Towards understanding of 31 Schulze MB, Liu S, Rimm EB, Manson JE, Willett WC, Hu FB. Glycemic index, glycaemic index and glycaemic load in habitual diet: associations with measures glycemic load, and dietary fiber intake and incidence of type 2 diabetes in of glycaemia in the Insulin Resistance Atherosclerosis Study. Br J Nutr 2006; 95: younger and middle-aged women. Am J Clin Nutr 2004; 80: 348–356. 397–405. 32 Heilbronn LK, Noakes M, Clifton PM. Effect of energy restriction, weight loss, and 15 The Diabetes and Nutrition Study Group (DNSG) of the European Association for diet composition on plasma lipids and glucose in patients with type 2 diabetes. the Study of Diabetes. Recommendations for the nutritional management of Diabetes Care 1999; 22: 889–895. patients with diabetes mellitus. Eur J Clin Nutr 2000; 54: 353–355. 33 Post RE, Mainous 3rd AG, King DE, Simpson KN. Dietary fiber for the treatment of 16 Mirmiran P, Esfahani FH, Mehrabi Y, Hedayati M, Azizi F. Reliability and relative type 2 diabetes mellitus: a meta-analysis. J Am Board Fam Med 2012; 25: 16–23. validity of an FFQ for nutrients in the Tehran lipid and glucose study. Public Health 34 Ciok J, Dolna A. The role of glycemic index concept in carbohydrate metabolism. Nutr 2010; 13: 654–662. Przegl Lek 2006; 63: 287–291.

& 2014 Macmillan Publishers Limited European Journal of Clinical Nutrition (2014) 459 – 463