Different Patterns of Insulin Resistance in Relatives of Type 1 Diabetic Patients With Retinopathy or Nephropathy : The Genesis France-Belgium Study

The Genesis France-Belgium Study is a prospective, multicenter, family-based study, the main aim of which is to study the genetic determinants of diabetic nephropathy and retinopathy in type 1 diabetes. Between November 1998 and December 2000, 662 subjects with type 1 diabetes (defined by age at onset ≤35 years, with initial ketosis and requirement for permanent insulin treatment within 1 year of diagnosis) were recruited in 38 diabetes or nephrology clinics in France and Belgium. The present article is limited to the the 275 type 1 diabetic patients for which familial data are available (referred to as probands). Patients were recruited according to their renal status and family structure. Probands classified as positive for diabetic nephropathy had type 1 diabetes for ≥5 years, background or more severe diabetic retinopathy (see classification below), and incipient, established, or advanced nephropathy (see classification below). Probands classified as negative for diabetic nephropathy had type 1 diabetes for ≥20 years, with background or more severe retinopathy but without kidney disease (see classification below). Probands (diabetic nephropathy–positive case subjects and −negative control subjects) were only included if their two parents were alive (trio families) or if one parent and at least one sibling were alive (pseudotrio families). Some complex families were also recruited. They consisted of at least three family members, including one index not considered simply as trio or pseudotrio (e.g., one diabetic sibpair plus both parents). The study design was approved by the ethics committee of Angers University Hospital. All study participants (patients and relatives) gave written informed consent.

Type 1 diabetic patients were examined by their diabetes specialist in their diabetes clinic (or by their nephrologist if they were being treated for renal failure), and their first-degree relatives were examined by their regular physician. History of arterial hypertension, dyslipidemia, myocardial infaction, stroke, heart failure, and end-stage renal disorder was recorded.

Blood pressure was determined using a mercury sphygmomanometer after a 5-min rest in the sitting position. Body weight and height were measured in light clothes without shoes. A resting electrocardiogram was recorded for all type 1 diabetic patients. During medical visits, 10 ml urine was collected for albumin excretion rate (AER) determination and 38 ml blood for serum and plasma storage and DNA extraction.

Nephropathy was staged for probands by an independent adjudication committee, as previously described (29), using the patient’s medical records and the three highest urinary AERs recorded in the patient’s records during the previous 10 years as well as the antihypertensive treatment and the last plasma creatinine measurement. No nephropathy was classified as persistent normoalbuminuria (median of three urinary AER samples <20 mg/l or 20 μg/min or 30 mg/24 h) without antihypertensive treatment, including ACE inhibitors or angiotensin-II blockers. Incipient nephropathy was classified as persistent microalbuminuria (median of three urinary AER samples 20–200 mg/l or 20–200 μg/min or 30–300 mg/24 h). Established nephropathy was classified as persistent proteinuria (median of three urinary AER samples >200 mg/l or 200 μg/min or 300 mg/24 h) with plasma creatinine <150 μmol/l. Advanced nephropathy was classified as increased AER and plasma creatinine ≥150 μmol/l and/or renal replacement therapy. Case subjects had nephropathy (incipient, established, or advanced), and control subjects did not have nephropathy.

Retinopathy was staged by an independent adjudication committee according to Kohner’s (30) classification, using the data obtained by direct funduscopy after pupil dilatation and/or fluorescein angiography. Stage 1 diabetic retinopathy (background retinopathy) was determined by three or more microaneurysms with or without hard exsudates and/or hemorrhages, stage 2 (preproliferative retinopathy) by intraretinal microvascular abnormalities and/or cotton wood exsudates, and stage 3 (proliferative retinopathy) by new vessels and/or past or present laser panphotocoagulation.

Insulin resistance was assessed in relatives according to WHO recommendations (16). Hypertension, personal history of lipid disorders, personal history of type 2 diabetes, and obesity were considered. The insulin resistance score was calculated as follows.

Relatives received 0 points if they did not have a history of hypertension and 1 point if they did have a history of hypertension (systolic [SBP]/diastolic blood pressure >140/80 mmHg at physical examination and/or antihypertensive treatment) (31).

Relatives received 0 points if they did not have a history of dyslipidemia and 1 point if they had a lipid disorder (total cholesterol >6 mmol/l and/or fasting triglycerides >2.3 mmol/l) or were on hypolipidemic treatment.

Relatives received 1 point if they had a personal history of type 2 diabetes (23) (regardless of diabetes treatment) and 0 points otherwise.

Relatives received 0 points if their weight was normal (BMI <25 kg/m²), 0.5 points if they were overweight (BMI 25–29.9 kg/m²), and 1 point if they were obese (BMI >30 kg/m²). Due to missing data, the total was divided by the number of available items. Thus, the insulin resistance score ranged from 0 to 1 and was expressed in arbitrary units.

HbA_1c was measured by high-performance liquid chromatography (Diamat; Biorad, Ivry-sur-Seine, France) (normal values 4–6%). Serum creatinine was measured using a modified version of Jaffe’s method. Urinary albumin concentration was determined using a nephelometric method (32) on sterile urine (verified by dipstick test). Plasma glucose concentration was determined by using an enzymatic peroxidase method. Fasting serum insulin was determined in non–type 1 diabetic first-degree relatives using an immunoradiometric assay method (Bi-Insulin IRMA; Schering, Gif-sur-Yvette, France), allowing for the calculation of homeostasis model assessment of insulin resistance (HOMA-IR) (33).

Data are presented as means ± SD or median (range), if not normally distributed. Categorical variables were compared by using the χ² test and the χ² test for trend (34). ANOVA and Kruskal-Wallis were used to compare continuous variables according to index microvascular stages in probands or in relatives. A nonparametric Spearman test was used for correlation analyses. To take intrafamilial correlations into account, we used the PROC MIXED software for continuous variables and ran generalized estimating equations as implemented in the PROC GENMOD software for categorical variables. Statview 5.1 and SAS 8.2 (SAS Institute, Cary, NC) were used for calculations. Statistical significance was set at 0.05.

A total of 853 subjects were recruited in the Genesis France-Belgium Family Study (578 relatives and 275 probands; 130 probands with and 145 without diabetic nephropathy). Control families included 120 trios, 22 pseudotrios, and 3 complex families (5 of 306 relatives had type 1 diabetes). Case families included 105 trios, 17 pseudotrios, and 8 complex families (there were 15 subjects with type 1 diabetes among 272 relatives). In total, 269 mothers, 239 fathers, and 70 siblings were analyzed. The familial structure (mothers, fathers, siblings) did not differ between case and control subjects (χ² = 0.107, P = 0.9481).

The characteristics of probands according to nephropathy stage are summarized in Table 1. As expected, SBP and the proportion of patients with proliferative retinopathy increased with severity of nephropathy. BMI did not differ according to severity of nephropathy.

The characteristics of probands according to their diabetic retinopathy stage are summarized in Table 2. Systolic blood pressure differed significantly according to retinopathy stage. Diabetes duration was longer in patients with proliferative retinopathy compared with other subjects. BMI did not differ between groups.

Mean age was 70 ± 9, 68 ± 9, and 47 ± 10 years in fathers, mothers, and siblings, respectively. Type 2 diabetes was found in 44 subjects. History of arterial hypertension was found in 200 subjects, and a personal history of dyslipidemia was found in 190 relatives. Histories of diabetes and hypertension were closely related (χ² = 7.139, P < 0.0001), unlike histories of diabetes and dyslipidemia (χ² = 0.774, P = 0.3488). Myocardial infarction was found in 39 subjects and was more frequent in relatives with type 2 diabetes than in nondiabetic relatives (χ² = 7.069, P = 0.0078). In relatives who gave a blood sample in the fasting state (n = 358), who were not different from nonfasting relatives regarding age (P = 0.89), insulin resistance score (P = 0.57), or BMI (P = 0.69), the insulin resistance score was positively correlated with fasting insulinemia (ρ = 0.109, P = 0.0469) and HOMA-IR (ρ = 0.184, P = 0.0028). The insulin resistance score was positively correlated with HbA_1c in relatives (ρ = 0.284, P < 0.0001), even after exclusion of type 2 diabetic relatives. The insulin resistance score was also correlated with urinary albumin excretion (UAE) (ρ = 0.144, P = 0.0009).

UAE was higher in relatives of proband case subjects with incipient, established, or advanced diabetic nephropathy (8 mg/l [2–1,126]) than in relatives of control subjects without diabetic nephropathy (6 mg/l [2–640], P = 0.0176). The UAE in relatives was not associated with retinopathy severity in probands (data not shown).

The clinical and biological characteristics of first-degree relatives according to proband nephropathy stage are summarized in Table 3. Insulin resistance scores were significantly higher in first-degree relatives of nephropathic (0.37 ± 0.26) than in those of nonnephropathic (0.31 ± 0.25, P = 0.0370) probands. The frequency of lipid disorders in relatives increased with the nephropathy stage of the proband. No such relationship was found with regard to history of type 2 diabetes or hypertension when fathers and mothers were analyzed separately. SBP and BMI did not differ in relatives according to the nephropathy stage of the proband (F = 0.956, P = 0.4132 and F = 1.676, P = 0.1710, respectively). However, maternal SBP increased with the severity of diabetic nephropathy in type 1 diabetic offspring (F = 3.836, P = 0.0104). Systolic blood pressure was higher in mothers of nephropathic (137 ± 17 mmHg) than nonnephropathic (132 ± 14 mmHg, P = 0.0071) probands.

The clinical and biological characteristics of relatives according to proband retinopathy stage are summarized in Table 4. The insulin resistance scores of relatives depended on retinopathy severity. The insulin resistance scores of relatives of probands with proliferative retinopathy was significantly higher compared with relatives of probands without proliferative retinopathy (P = 0.0144). Taking nephropathy into account did not substantially modify the relationship (P = 0.0379). The frequency of type 2 diabetes did not depend on the severity of the retinopathy in the probands when fathers and mothers were analyzed separately. However, the frequencies of obesity and hypertension in the relatives increased significantly with the severity of diabetic retinopathy in the probands. Systolic blood pressure was not related to the severity of retinopathy in the probands even when considering mothers or fathers separately (data not shown).

In this large family-based study, we found that the insulin resistance score was higher in first-degree relatives of probands with rather than without nephropathy. Interestingly, it was also higher in first-degree relatives of probands with severe retinopathy, even after controlling for nephropathy. We found that arterial hypertension and obesity segregated with retinopathy, whereas lipid disorders and obesity segregated with nephropathy. These results were particularly relevant in mothers of type 1 probands.

This study does, however, have some limitations. Hypertension in relatives was diagnosed by using a questionnaire and by recording blood pressure on just one occasion. Furthermore, we did not measure serum lipids; instead, we used a structured questionnaire to interview relatives. Moreover, we were not able to measure insulin resistance with the reference euglycemic clamp technique due to a very large number of relatives. We were also unfortunately not able to consider waist and hip circumferences. However, we found that insulin resistance was positively correlated with fasting insulinemia, HOMA-IR, and HbA_1c, strongly suggesting that this score effectively reflects insulin resistance, as recently suggested (35). A relationship was also found between the insulin resistance score and UAE, as proposed in the WHO working definition (16) and as shown in Finnish patients (36).

The characteristics of our study population were in good accordance with previous studies. In probands, blood pressure significantly increased with the severity of nephropathy (29). Diabetes duration was a major risk factor for severe diabetic retinopathy (37,38). The prevalence of hypertension, lipid disorders, and type 2 diabetes in relatives was in good accordance with data from the French general population (39,40). The sample frequency of case and control probands is very similar to what was expected from the general population (36).

We found that insulin resistance was higher in relatives of proband case subjects with nephropathy than in those of control subjects without nephropathy. This is in good accordance with three previous studies (26–28) showing that relatives of microalbuminuric or nephropathic patients are more resistant to insulin than relatives of normoalbuminuric patients, using the euglycemic clamp reference or other surrogate methods.

Our results extend our knowledge, showing that familial insulin resistance, in addition to nephropathy, is also related to retinopathy. This relationship between familial insulin resistance and the presence of microvascular complications in index type 1 diabetic subjects is in good accordance with the data from prospective follow-up studies (24,25), showing that insulin resistance is a risk factor for progression or development of retinopathy or nephropathy in type 1 diabetic patients. This report is the first to demonstrate an association between familial insulin resistance and the severity of retinopathy in diabetic first-degree relatives.

Our data support the hypothesis that insulin resistance has a familial component that affects the risk of microvascular disease. This is in good accordance with the results from studies showing an increased risk of microangiopathy in diabetic patients that carry the genetic variants associated with insulin resistance in peroxisome proliferator–activated receptor-γ (41) or human glycoprotein PC-1 genes (42). However, insulin resistance syndrome is an intermediate phenotype with both genetic and environmental components. We therefore cannot infer from our data that the genetic components of risk for microangiopathy in type 1 diabetic subjects are due to the genetic components of insulin resistance. Instead, we can conclude that discrete components of the insulin resistance syndrome segregate with discrete manifestations of diabetic microangiopathy.

We evidenced an association between history of hypertension in relatives and diabetic nephropathy, as shown in some (8,9) but not all (43) previous reports. We found that mothers of nephropathic probands had higher SBP values than mothers of control subjects. This discrepancy for fathers may be related to the insufficient sensitivity of the office blood pressure measurement apparatus, as suggested by a Finnish study (10). Consistent with the possibility, the relatives of probands with nephropathy (incipient to advanced) displayed higher UAE than those of control subjects. Some candidate genes could explain the association between familial hypertension and retinopathy, such as endothelial nitric oxide synthase (44).

Personal history of dyslipidemia was slightly more frequent among relatives of index case subjects than others. This report confirms the data from a single-center family-based study showing a clustering of dyslipidemia with nephropathy (28). Familial clustering of lipid disturbances may partly account for the reported familial clustering of cardiovascular disease in type 1 diabetic patients with nephropathy (6,7,28,45). Lipid disturbances are well known to be associated with diabetic nephropathy (46,47), and prospective studies of type 1 diabetic patients have shown that an adverse lipid profile is predictive of the worsening of renal disease (18,48) or the least regression from microalbuminuria to normoalbuminuria (49). The adverse role of plasma lipids was recently found regarding renal and retinal complications in a longitudinal study (50). Several candidate genes involved in lipid metabolism could account for the relationship between familial lipid disorders and nephropathy, with apolipoprotein E being already widely studied (47,51,52).

Obesity in relatives was strongly associated with the risk of both severe retinopathy and nephropathy. This is in good accordance with previous reports showing that BMI was higher in type 1 diabetic patients who developed retinopathy (21) or nephropathy (24). Interestingly, the role of obesity in patients is not due to a familial resemblance, as BMI did not differ according to nephropathy or retinopathy stage of the index patients. Candidate genes for obesity are currently extensively studied, and some of them, such as adiponectin (53), could be associated with microvascular complications, as suggested by our results.

In conclusion, first-degree relatives are more insulin resistant when type 1 diabetic probands have more severe diabetic retinopathy or nephropathy. The components of the insulin resistance syndrome support the search for genetic factors involved in arterial hypertension and obesity for retinopathy and in lipids and obesity for nephropathy.

Scientific committee: M.M. (principal investigator), P.J.L. (chairman) and F. Alhenc-Gelas, B.D., Y.G., S.H., P.P., M.R., V.R. Adjudication committee: M.M., B.B., B.D., Y.G., V.R., M.R. Investigators (alphabetical order of the centers; all centers are in France except Liege, which is in Belgium): Diabetology: Amiens (A. Dubreuil, S. Arlot), Angers Coordinating Center (M.M., B. Bouhanick, S.H., G. Guilloteau), Association Française des Diabétiques, Bordeaux-Pessac (H. Gin, V.R.), Corbeil Essones (G. Charpentier, I. Petit, J.P. Riveline), Grenoble (S. Halimi), Liège (P.J.L., A.J. Scheen, L.W.), Lille (P. Fontaine, G. Prevot), Lyon-Antiquaille (F. Berthezene, F. Bonnet, S. Marsot), Lyon E-Herriot (C. Thivolet), Marseilles (P. Vague, M.F. Jannot), Nantes (B. Charbonnel, L. Chaillous), Nimes-Montpellier (M.R.), Paris-Bichat (M.M., E. Larger, A.B.), Paris-Hotel Dieu (G. Slama, J.L. Selam, A. Sola), Paris-Saint Louis (P.P., C. Chenu), Poitiers (R.M., F. Torremocha, S.H.), Saint Brieuc (C. Colmar-Montiel), Saint Mandé (B.B., H. Mayaudon), Strasbourg (M.P. Arpin-Bott), Nancy-Toul (P. Drouin [deceased], P. Mattei, B. Guerci), Toulouse (J.P. Tauber [deceased], H. Hanaire-Broutin, S. Gronier), Valenciennes (O. Verier-Mine). Nephrology: Amiens (A. Fournier, J.M. Achard), Grenoble (D. Cordonnier), Lyon-E Herriot (M. Laville, J.P. Fauvel), Montauban (P. Giraud), Nancy-Vandoeuvre (M. Kessler, E. Renoult), Paris-Necker (J.P. Grunfeld), Poitiers (G. Touchard, F. Bridoux), Reims (J. Chanard, J.P. Melin), Saint Brieuc (C. Charasse). The English of the text was checked by Alex Edelman and Associates.

This work was supported by grants from the Association Française des Diabétiques, INSERM (Institut National de la Santé et de la Recherche Médicale), Société Française d’Hypertension Artérielle, European Union (Euragedic Project; QLG2-CT2001-01 669), and Association Diabète et Risque Vasculaire (Paris, France).

We thank all the type 1 diabetic patients and their relatives who took part in this study. We thank G. Brossard and V. Benoit for technical assistance, C. Iste for her greatly appreciated secretarial assistance, and J. Guilhot (Centre de Recherche Clinique, Poitiers, France) for statistical help.