Interaction between GPR120 p.R270H loss-of-function variant and dietary fat intake on incident type 2 diabetes risk in the D.E.S.I.R. study.

BACKGROUND AND AIMS: GPR120 (encoded by FFAR4) is a lipid sensor that plays an important role in the control of energy balance. GPR120 is activated by long chain fatty acids (FAs) including omega-3 FAs. In humans, the loss of function p.R270H variant of the gene FFAR4 has been associated with a lower protein activity, an increased risk of obesity and higher fasting plasma glucose levels. The aim of this study was to investigate whether p.R270H interacts with dietary fat intake to modulate the risk of type 2 diabetes (T2D, 198 incident; 368 prevalent cases) and overweight (787 incident and 2891 prevalent cases) in the prospective D.E.S.I.R. study (n = 5,212, 9 years follow-up). METHODS AND RESULTS: The association of p.R270H with dietary fat and total calories was assessed by linear mixed models. The interaction between p.R270H and dietary fat on T2D and overweight was assessed by logistic regression analysis. The p.R270H variant had a minor allele frequency of 1.45% and was not significantly associated with total calories intake, fat intake or the total calories derived from fat (%). However, there was a significant interaction between p.R270H and dietary fat modulating the incidence of T2D (Pinteraction = 0.02) where the H-carriers had a higher risk of T2D than RR homozygotes in the low fat intake category only. The interaction between p.R270H and fat intake modulating the incidence and prevalence of overweight was not significant. CONCLUSION: The p.R270H variant of GPR120 modulates the risk of T2D in interaction with dietary fat intake in the D.E.S.I.R.

Dietary fat intake and polymorphisms at the PPARG locus modulate BMI and type 2 diabetes risk in the D.E.S.I.R. prospective study.

CONTEXT: Fat-rich diets are involved in many disorders such as obesity and type 2 diabetes (T2D). The Pro12Ala variant of peroxisome proliferator-activated receptor-γ (PPARγ) is known to modulate body mass index (BMI) and T2D risk. OBJECTIVE: Our aim was to study the interaction effect between PPARγ gene (PPARG) polymorphisms Pro12Ala and 1431C>T and fat intake on incident T2D and BMI in a 9-year prospective cohort drawn from the French general population, the D.E.S.I.R. (Data from an Epidemiological Study on the Insulin Resistance Syndrome) study (n=4676). METHODS: Nutritional intake was assessed by a food frequency self-questionnaire completed by each participant. Statistical analyses included logistic regression, analysis of covariance and haplotype analysis, with adjustment for confounding variables. RESULTS: A high fat consumption (the third sex-specific tertile of fat intake, as a percentage of energy intake) was associated with an increased T2D risk among ProPro and CC homozygotes (P(interaction)=0.05, odds ratio (OR) (95% confidence interval (95% CI))=1.73 (1.19-2.52) P=0.004 and OR=1.85 (1.27-2.71) P=0.001, respectively) but not in Ala and T carriers. There was a significant interaction effect between Pro12Ala and 1431C>T on BMI (P(interaction)=0.004); Ala was associated with lower BMI in CC homozygotes and with higher BMI in T carriers while the opposite was found for ProPro. There was also an interaction effect between Pro12Ala and dietary fat intake on BMI (P(interaction)=0.02); AlaAla individuals had a higher BMI than Pro carriers among high fat consumers (27.1 ± 1.0 versus 24.9 ± 0.1 for AlaAla and Pro+, respectively). There was no interaction effect between the 1431C>T single-nucleotide polymorphism and fat intake on BMI. CONCLUSION: Our results indicate strong genetic and nutritional interaction effects on BMI and T2D risk at the PPARG locus in a general population.

Decreased insulin secretion and increased risk of type 2 diabetes associated with allelic variations of the WFS1 gene: the Data from Epidemiological Study on the Insulin Resistance Syndrome (DESIR) prospective study.

AIMS/HYPOTHESIS: We investigated associations of allelic variations in the WFS1 gene with insulin secretion and risk of type 2 diabetes in a general population prospective study. METHODS: We studied 5,110 unrelated French men and women who participated in the prospective Data from Epidemiological Study on the Insulin Resistance Syndrome (DESIR) study. Additional cross-sectional analyses were performed on 4,472 French individuals with type 2 diabetes and 3,065 controls. Three single nucleotide polymorphisms (SNPs) were genotyped: rs10010131, rs1801213/rs7672995 and rs734312. RESULTS: We observed statistically significant associations between the major alleles of the three variants and prevalent type 2 diabetes in the DESIR cohort at baseline. Cox analyses showed an association between the G-allele of rs10010131 and incident type 2 diabetes (HR 1.34, 95% CI 1.08-1.70, p = 0.007). Similar results were observed for the G-allele of rs1801213 and the A-allele of rs734312. The GGA haplotype was associated with an increased risk of diabetes as compared with the ACG haplotype (HR 1.26, 95% CI 1.04-1.42, p = 0.02). We also observed statistically significant associations of the three SNPs with plasma glucose, HbA(1c) levels and insulin secretion at baseline and throughout the study in individuals with type 2 diabetes or at risk of developing diabetes. However, no association was observed in those who remained normoglycaemic at the end of the follow-up. Associations between the three variants and type 2 diabetes were replicated in cross-sectional studies of type 2 diabetic patients in comparison with a non-diabetic control group. CONCLUSIONS/INTERPRETATION: The most frequent haplotype at the haplotype block containing the WFS1 gene modulated insulin secretion and was associated with an increased risk of type 2 diabetes.

Relationship between renal capacity to reabsorb glucose and renal status in patients with diabetes.

AIMS: Interindividual variability in capacity to reabsorb glucose at the proximal renal tubule could contribute to risk of diabetic kidney disease. Our present study investigated, in patients with diabetes, the association between fractional reabsorption of glucose (FRGLU) and degree of renal disease as assessed by urinary albumin excretion (UAE) and estimated glomerular filtration rate (eGFR). METHODS: FRGLU [1-(glucose clearance/creatinine clearance)] was assessed in 637 diabetes patients attending our tertiary referral centre, looking for correlations between FRGLU and UAE (normo-, micro-, macro-albuminuria) and Kidney Disease: Improving Global Outcomes (KDIGO) eGFR categories: >90 (G1); 90-60 (G2); 59-30 (G3); and<30-16 (G4) mL/min/1.73 m2. Patients were stratified by admission fasting plasma glucose (FPG) into three groups: low (<6mmol/L); intermediate (6-11mmol/L); and high (>11mmol/L). RESULTS: Median (interquartile range, IQR) FRGLU levels were blood glucose-dependent: 99.90% (0.05) for low (n=106); 99.90% (0.41) for intermediate (n=288); and 96.36% (12.57) for high (n=243) blood glucose categories (P<0.0001). Also, FRGLU increased with renal disease severity in patients in the high FPG group: normoalbuminuria, 93.50% (17.74) (n=135); microalbuminuria, 96.56% (5.94) (n=77); macroalbuminuria, 99.12% (5.44) (n=31; P<0.001); eGFR G1, 94.13% (16.24) (n=111); G2, 96.35% (11.94) (n=72); G3 98.88% (7.59) (n=46); and G4, 99.11% (2.20) (n=14; P<0.01). On multiple regression analyses, FRGLU remained significantly and independently associated with UAE and eGFR in patients in the high blood glucose group. CONCLUSION: High glucose reabsorption capacity in renal proximal tubules is associated with high UAE and low eGFR in patients with diabetes and blood glucose levels>11mmol/L.

Dairy consumption is associated with lower plasma dihydroceramides in women from the D.E.S.I.R. cohort.

AIM: In the D.E.S.I.R. cohort, higher consumption of dairy products was associated with lower incidence of hyperglycaemia, and dihydroceramide concentrations were higher in those who progressed to diabetes. Our aim here was to study the relationships between dairy consumption and concentrations of dihydroceramides and ceramides. METHODS: In the D.E.S.I.R. cohort, men and women aged 30-65 years, volunteers from West-Central France, were included in a 9-year follow-up with examinations every 3 years, including food-frequency questionnaires. Two items concerned dairy products (cheese, other dairy products except cheese). At each examination, dihydroceramides and ceramides were determined by mass spectrometry in a cohort subset; in the present study, the 105 people who did not progress to type 2 diabetes were analyzed, as the disorder per se might be a confounding factor. RESULTS: Higher consumption of dairy products (except cheese) was associated with total plasma dihydroceramides during the follow-up, but only in women (P=0.01 for gender interaction). In fact, dihydroceramide levels were lower in women with high vs low consumption (P=0.03), and were significantly increased during follow-up (P=0.01) in low consumers only. There was also a trend for lower ceramides in women with high dairy (except cheese) intakes (P=0.08). Cheese was associated with dihydroceramide and ceramide changes during follow-up (P=0.04 for both), but no clear trend was evident in either low or high consumers. CONCLUSION: These results show that, in women, there is an inverse association between fresh dairy product consumption and predictive markers (dihydroceramides) of type 2 diabetes.

Alcohol intake modulates the effect of a polymorphism of the cholesteryl ester transfer protein gene on plasma high density lipoprotein and the risk of myocardial infarction.

A polymorphism of the CETP gene (CETP/TaqIB) with two alleles B1 (60%) and B2 (40%) has been investigated in relation to lipid variables and the risk of myocardial infarction in a large case-control study (ECTIM) of men aged 25-64. No association was observed between the polymorphism and LDL or VLDL related lipid variables. Conversely, B2 carriers had reduced levels of plasma CETP (P < 0.0001) and increased levels of HDL cholesterol (P < 0.0001) and of other HDL related lipid variables. The effects of the polymorphism on plasma CETP and HDL cholesterol were independent, suggesting the presence of at least two functional variants linked to B2. A search for these variants on the coding sequence of the CETP gene failed to identify them. The effect of B2 on plasma HDL cholesterol was absent in subjects drinking < 25 grams/d of alcohol but increased commensurably, with higher values of alcohol consumption (interaction: P < 0.0001). A similar interaction was not observed for plasma CETP. The odds-ratio for myocardial infarction of B2 homozygotes decreased from 1.0 in nondrinkers to 0.34 in those drinking 75 grams/d or more. These results provide the first demonstration of a gene-environment interaction affecting HDL cholesterol levels and coronary heart disease risk.

T-cadherin gene variants are associated with type 2 diabetes and the Fatty Liver Index in the French population.

AIM: Adiponectin is an adipocyte-secreted protein associated with insulin sensitivity. T-cadherin is a receptor for high and medium molecular weight adiponectin. In GWAS, T-cadherin gene (CDH13) polymorphisms are associated with circulating adiponectin levels. This study investigated the associations between genetic variants of CDH13 and type 2 diabetes (T2D), and its related parameters, in a Caucasian population. METHODS: Two polymorphisms of CDH13 (rs11646213 and rs3865188) were genotyped in two French cohorts, a general population from the D.E.S.I.R. study (n=5212) and people with T2D in the DIABHYCAR study (n=3123). Baseline adiponectin levels were measured in D.E.S.I.R. participants who were normoglycaemic at baseline, but hyperglycaemic after 3 years (n=230), and in controls who remained normoglycaemic (n=226) throughout. RESULTS: In a cross-sectional analysis, CDH13 genotype distributions differed between those with and without T2D, with T2D odds ratios (OR) of 1.11 (95% CI: 1.04-1.18; P=0.001) and 0.92 (95% CI: 0.87-0.98; P=0.01) for rs11646213 and rs3865188, respectively. The rs11646213 variant, associated with a higher OR for T2D, was also associated with higher BMI (P=0.03) and HbA1c (P=0.006), and lower plasma adiponectin levels (P=0.03) in the D.E.S.I.R. PARTICIPANTS: Conversely, the rs3865188 variant, associated with a lower OR for T2D, was also associated with lower BMI (P=0.03), HbA1c (P=0.02) and Fatty Liver Index (FLI; P≤0.01), and higher plasma adiponectin levels (P=0.002). Associations with HbA1c, FLI and adiponectin levels persisted after adjusting for BMI. CONCLUSION: CDH13 polymorphisms are associated with prevalent T2D in this French population study. The association may be mediated through effects on BMI and/or plasma adiponectin.

Potential risks associated with increased plasma plant-sterol levels.

The consumption of plant sterols is associated with a decrease in LDL cholesterol. However, it is also associated with an increase in plasma plant-sterol (sitosterol, campesterol) levels that may be detrimental. Indeed, the genetic disease sitosterolaemia, which is characterized by elevated plasma levels of plant sterol, is associated with premature atherosclerosis. Yet, although plasma plant-sterol levels are recognized markers of cholesterol absorption, the relationship between such levels and atherosclerosis is not clear. Several studies have analysed the association between plasma plant-sterol levels and cardiovascular disease (CVD), but have found conflicting results. Although the largest prospective trials and genome-wide association studies suggest that high plasma levels of plant sterols are associated with increased CV risk, other studies have reported no such association and even an inverse relationship. Thus, the available data cannot confirm an increased CV risk with plant sterols, but cannot rule it out either. Only a prospective interventional trial to analyse the effects of plant-sterol-enriched food on the occurrence of CV events can exclude a potential CV risk linked with their consumption.

Lowering of HDL2-cholesterol and lipoprotein A-I particle levels by increasing the ratio of polyunsaturated to saturated fatty acids.

The protective role of high-density lipoproteins (HDLs) has been attributed to the subfractions HDL2 (according to the density) and lipoprotein A-I (LpA-I) (according to the composition in apolipoproteins). We investigated the effect of a high ratio of polyunsaturated to saturated fatty acids (P:S) on these subfractions in a homogeneous group of young adult males. Two prescribed diets were consumed successively at the subjects' homes for 3 wk each in a random order; one diet contained 70 g butter (P:S 0.2, diet B), the other contained 70 g sunflower margarine (P:S 1.1, diet M). Total calorie, fat, and cholesterol intakes were similar for the two diets. Cholesterol and apolipoprotein B in serum and in low-density lipoproteins (LDLs) were lower with diet M than with diet B. However, significant decreases in protective subfractions of HDL, HDL2, and LpA-I were observed. This undesirable effect of the diet with a high P:S could cancel the benefits of lowering the LDL-cholesterol concentrations.

n-3 polyunsaturated fatty acids raise low-density lipoproteins, high-density lipoprotein 2, and plasminogen-activator inhibitor in healthy young men.

The effects of a moderate supplementation in n-3 polyunsaturated fatty acids (PUFAs) were investigated in 36 young healthy adult males. Factors investigated were lipoprotein (including HDL subfractions and apolipoproteins) and hemostasis indexes, assessed by platelet aggregation and plasminogen-activator-inhibitor (PAI) activity. Fat-controlled diets were prescribed, one with and one without a fish-oil supplement (control diet), successively during 3 wk in random order. Total calorie, fat, and cholesterol intakes were similar in the two diets. Triglycerides in serum and very-low-density lipoproteins were lower and high-density-lipoprotein 2 cholesterol was higher with the n-3 PUFA-supplemented diet. These effects as well as a significant decrease in platelet aggregation can be considered beneficial in terms of cardiovascular risk. However, significant increases in low-density-lipoprotein cholesterol and PAI activity occurred and were correlated. This latter effect could be detrimental.

Lack of association between dietary alcohol and HDL-cholesterol concentrations in obese women.

The relationships of alcohol intake and corpulence to HDL-cholesterol were studied in 653 women taking medical advice about body weight. The body mass index (BMI) was positively correlated with triglyceride and negatively with HDL-cholesterol. The relation between BMI and HDL-cholesterol was discontinuous. Total cholesterol, triglycerides and diastolic blood pressure were increased for alcohol intakes greater than 10 g/d regardless of body weight. Alcohol intake was associated with higher concentrations of HDL-cholesterol (P = 0.006) in non obese (BMI = 25.2 +/- 1.5 kg/m2) subjects, but not in mildly (27.3 less than or equal to BMI less than 32.3) or massively (BMI greater than or equal to 32.3) obese subjects. The fact that HDL concentrations were not associated with alcohol intake in obese patients suggests that (1) alcohol acts on the HDL pool through one of the pathways which are perturbed in obesity, possibly lipolysis, (2) obesity is one of the reasons for the differences in individual responses of HDL-cholesterol to alcohol, (3) myocardial infarction might not be inversely correlated with alcohol intake in the obese as it is in the non-obese population.

Low high density lipoprotein-2 concentrations in obese male subjects.

In this study we have compared the lipoprotein patterns, in particular HDL subfractions, of 34 obese men to those of 34 normoponderal normolipemic men, matched for age and use of tobacco. Obesity was associated with increased VLDL concentrations in only half the subjects. HDL concentrations in all obese subjects were lower than in matched controls. The decrease was most marked in the HDL2 subfraction in which cholesterol and protein contents were decreased by 50%; it was independent of triglyceride levels and not related to the severity of overweight. Moreover, while HDL2 was negatively correlated with BMI (P less than 0.01) when both populations were considered together, the correlation disappeared when calculated separately within each population, suggesting a threshold effect. The low levels of HDL2 might result from discretely altered lipolysis, not sufficient to cause an elevation in fasting triglyceridemia. In this case, HDL2 should prove to be a sensitive index of lipolytic efficiency.

Lack of association between genetic variations of apo A-I-C-III-A-IV gene cluster and myocardial infarction in a sample of European male: ECTIM study.

The goal of the present study was to compare the allele frequency of four polymorphisms at the apo A-I C-III A-IV cluster gene locus-ApoA-I: XmnI and PstI; ApoC-III: SstI; ApoA-IV: XbaI-between male patients who had had a myocardial infarction (n= 614) and matched controls (n = 764). The association with a number of lipid lipoprotein, apolipoprotein and lipoprotein particle variables was also assessed. Patients and subjects were recruited in Belfast, Lille, Strasbourg and Toulouse in the framework of the ECTIM study. In the control group, the frequencies of the different polymorphic alleles were homogeneous among recruitment centres suggesting the absence of any European North to South gradient for these cluster polymorphisms. There was no evidence for a significant difference in allelic distribution between cases and controls suggesting that apo A-I, C-III, A-IV gene cluster polymorphisms do not explain MI survival in this sample of European men. There was no statistically significant association between apo A-I C-III A-IV cluster gene polymorphisms and lipid, lipoprotein, apolipoprotein, and lipoprotein particle levels. In conclusion, in the ECTIM study, the apo A-I, C-III, A-IV gene cluster polymorphism is associated with neither circulating plasma variables nor MI survival.

A new T-287C polymorphism in the 5' regulatory region of the tissue factor pathway inhibitor gene. Association study of the T-287C and C-399T polymorphisms with coronary artery disease and plasma TFPI levels.

Tissue factor pathway inhibitor (TFPI) is an important regulator of the extrinsic blood coagulation pathway. We screened the untranslated 5' region of the TFPI gene for polymorphisms and investigated their possible involvement in arterial thrombosis. The allele frequencies of a new polymorphism, located 287 base pairs upstream of the transcription start site (T-287C), and that of the previously described C-399T polymorphism, were similar in cases and controls. In controls, the -287C allele was associated with significantly higher levels of total TFPI antigen, arguing for an effect of this polymorphism on TFPI gene expression. In controls, the C-399T polymorphism did not alter TFPI levels. In the cases, however, decreased total and post-heparin free TFPI levels and increased F1+2 levels were significantly associated with the -399T allele. These findings suggest that the T-287C and C-399T polymorphisms are not associated with an increased risk of coronary heart disease, a result which should be confirmed by a larger study. However, their influence on outcome, or a link with subtypes of acute coronary syndromes, cannot be excluded.

Effect of a moderate alcohol intake on the lipoproteins of normotriglyceridemic obese subjects compared with normoponderal controls.

Moderate alcohol intake is frequently associated with an elevated concentration of high-density lipoprotein (HDL), which is one of the potential causes for the relative decrease in cardiovascular risk reported in moderate drinkers. Conversely, low HDL concentrations, particularly HDL2, in obese subjects may be a risk factor. The effect of 30 g alcohol daily (wine) during 14 days following a period of abstinence was studied in seven normolipidemic obese subjects (body mass index [BMI], 30 +/- 1.7 kg/m2) compared with seven normoponderal controls (BMI, 22 +/- 1.2 kg/m2). Alcohol caused apolipoprotein (apo) AI and apo AII concentrations to increase in all controls by 12% and 16% (P less than .05), but not in obese subjects. Lipoprotein (Lp) AI HDL particles (without AII) were initially in the same proportions in the two groups. Their increase in controls only (P less than .03) was not matched by an increase in HDL2 in all subjects. In obese subjects, neither Lp AI nor HDL2 were increased by alcohol, but their HDL-triglyceride (TG) contents, initially elevated, were normalized. Cholesterol ester (CE) transfer activity was not different in controls and obese subjects during abstinence (105.7 +/- 40.8 v 104.8 +/- 34.5 mmol/mg protein/h). It was notably depressed by alcohol in controls (74.2 +/- 27.4, P less than .002), but not in obese subjects.(ABSTRACT TRUNCATED AT 250 WORDS)

Lack of plasmic beta-endorphin response to a gastronomic meal in healthy humans.

In order to study the relationship between the endogenous opiate system and food intake in man, plasma concentrations of beta-endorphin were measured in ten healthy subjects. Time course of beta-endorphinemia was compared under the following conditions: basal (fasting), after an injection of pentagastrin (6 micrograms/kg), or after a gastronomic meal. No changes in plasma beta-endorphin or ACTH concentrations were observed with pentagastrin nor after the meal, despite the combination of very high sensory pleasure with intake of a very large amount of food. It is concluded that blood beta-endorphin concentration is not a sensitive index of the effects of food intake on the endogenous opioid system in man.

Permanent administration of d-fenfluramine in rats: paradoxical effects.

In order to test the hypothesis of Levitzky that d-fenfluramine (d-F) acts by modifying the ponderal set-point, we compared the effects of a permanent infusion of d-F on food intake and body weight (BW). The effect on the weight persisted as long as the infusion; the clear-cut anorectic effect lasted only a few days. This paradox is compatible with the set-point hypothesis. In rats rendered overweight by insulin treatment, the d-F-induced decrease in BW was approximately four times smaller than in controls. In rats rendered overweight by a cafeteria diet, the decrease in BW was twice as large in permanently cafeteria fed rats as in cafeteria, then, ad lib fed rats. In rats rendered underweight by a restricted chow diet and then returned to an ad lib feeding, the final BW depended only on the doses of d-F (0.6 or 12 mg/kg BW/day), whatever the weight at the beginning of infusion. Thus, the underweight paradigm fits well with the set-point hypothesis; the overweight paradigm fits only partially.

Comparative effects of several simple carbohydrates on erythrocyte insulin receptors in obese subjects.

The effects of simple carbohydrates on erythrocyte insulin receptors, plasma insulin and plasma glucose were studied during four hypocaloric, hyperproteic, diets. One diet contained no carbohydrate; the other three contained 36 g of either glucose, galactose or fructose. These diets were given for a 14-day period to groups of moderately obese subjects. The hypocaloric carbohydrate-free diet produced a decrease in plasma insulin and glucose concentrations concomitant with an increase in the number of insulin receptors. A similar increase in insulin receptor number was found when the diet was supplemented with glucose or galactose, but not with fructose. The presence of fructose in the diet prevented any increase in insulin receptor number.

[Polymorphism of aldose reductase gene and susceptibility to retinopathy and nephropathy in Caucasians with type 1 diabetes]. / Polymorphismes du gène de l'aldose réductase et susceptibilité à la rétinopathie et à la néphropathie chez des caucasiens diabétiques de type 1.

Polymorphisms of the -106 mutation and z - 2 or z + 2 microsatellites (-2.1 kb) of the Aldose Reductase (AR) gene have been associated to microangiopathic complications of the diabetes mellitus. The study aimed to establish a relation between the occurrence and progression of the renal and retinal complications and these polymorphisms. The genotypes were realised in 3 populations: DESIR (n = 369), non-diabetic control subjects from the general French population: GENEDIAB (n = 494), type 1 diabetic patients who are suffering from proliferative retinopathy associated with a variable seriousness nephropathy (absent: n = 157; incipient: n = 104; established: n = 126; advanced: n = 107); SURGENE (n = 310), type 1 diabetic patients whom the renal status is prospectively assessed since 1989 in one single center Angers University Hospital. The genotype of the -106 polymorphism was determined using the Molecular Beacons. For the microsatellites analysis, we used an automatized method (GeneScan Abi Prism 3100). There was a strong linkage disequilibrium between the z - 2 allele and the T allele (chi 2 = 120; p = 0.001). The frequency of the C-106T is similar for the DESIR and GENEDIAB cohorts (chi 2 = 3.32; p = 0.19); the Hardy Weinberg law was verified in this group (chi 2 = 0.001, 0.9; p = 1.5 and 0.5 respectively). The law was not verified for the SURGENE cohort (chi 2 = 4.7; p = 0.03) where the frequency of the TT genotype was significantly more important compared to the DESIR population (chi 2 = 6.4; p = 0.01). The z, z - 2 and z + 2 alleles was more frequent compared with other alleles (n = 909, 830 and 349; 39, 38 and 15%). The frequency of the C-106T and microsatellites genotypes did not parallel the nephropathy staging in the GENEDIAB population (chi 2 = 10.9, 2.7, 2.4; p = NS respectively). In the SURGENE population, the survival without renal events did not differ according C-106T and z - 2 or z + 2 microsatellites genotypes (log-rank: 0.6, 3.9, 0.1; p = NS respectively). At the end of the follow-up, we found an effect of the -106 mutation and of the z - 2 microsatellite on the staging of the retinopathy (chi 2 tendency test = 4.61, 0.12; p = 0.031, 0.02; 6 d.f., respectively). The logistic regression multivariable analysis shows that the retinopathy during the final evaluation is independently explained by several factors: diabetes duration (p < 0.0001; OR 21.756; 95% CI: 7.024-67.389), presence of nephropathy (p < 0.0001; OR: 4.086; 95% CI: 2.094-7.973), and genotype TT (p = 0.011; OR: 0.38; 95% CI: 0.18-0.803). In contrast, age of diabetes onset (p = 0.112; OR: 1.556; 95% CI: 0.9-2.692), median HbA1c (p = 0.164; OR: 1.479; 95% CI: 0.85-2.576) and sex (p = 0.156; OR: 1.495; 95% CI: 0.856-2.612) have no independent effect. In conclusion, the association of these AR genetic variants seems absent about the renal risk and slight about the retinal risk associated to the type 1 diabetes mellitus.