Five thousand five hundred seventy-two newly diagnosed non-insulin-dependent diabetes mellitus (NIDDM) patients (3,225 men and 2,347 women; mean age, 58.5 years) were recruited through the General Practitioners (GPs) network in France. All had persistent hyperglycemia after a preliminary 3-month period with dietary and life-style modification. Gliclazide (80 to 320 mg/d) was then prescribed as diabetic pharmacotherapy for 2 years. Additional therapy for hypertension and dyslipidemia was started if necessary. The aim of the study was mainly to determine the feasibility of a GP-directed protocol for the monitoring and treatment of newly diagnosed NIDDM patients, and to assess the effectiveness of diabetic therapy in this cohort. Diabetes was diagnosed in 78% of the cohort during routine screening. Among the women, 6.5% had a history of gestational diabetes. Eighteen percent of the patients had a parental history of diabetes, and the dominant maternal role in the genesis of NIDDM was confirmed. High blood pressure (Joint National Committee V criteria) was found at inclusion in 38.8% of the whole cohort. Hyperlipidemia was known in 44.6%. A history of stroke was present in 1.6% of the patients, and coronary heart disease (CHD) in 6.3%. These data support the relationship between the atherogenic state and development of NIDDM. Microalbuminuria defined as urinary albumin excretion (UAE) of at least 20 mg/L was found in 29.6% of the patients, and retinopathy in 9.8%. Among the included patients, 23% did not complete the study and were excluded from the efficacy analysis. Of these, 14% (808 patients) had only baseline evaluation data and 9% (499 patients) withdrew later. Comparison of mean baseline and final results in study completers uncovered a significant improvement in fasting blood glucose ([FBG] 182 +/- 48 v 137 +/- 40 mg/dL), post prandial blood glucose ([PPBG] 209 +/- 68 v 162 +/- 52 mg/dL), and hemoglobin A1c ([HbA1c] 8.7% +/- 2.5% v 7.3% +/- 2.0%). A slight improvement in total cholesterol (228 +/- 44 v 222 +/- 41 mg/dL), body mass index ([BMI] 28.5 +/- 4.7 v 27.9 +/- 4.5 kg/m2), and waist to hip ratio (0.99 +/- 0.1 v 0.98 +/- 0.1) was observed. There was a decrease in the percentage of patients with high blood pressure (38.5% v 30.7%). A mild increase in the prevalence of retinopathy (10.2% v 11.8%) was noted during the study, while the incidence of microalbuminuria remained unchanged (30.2% v 29.5%). In conclusion, the data indicate that the GPs involved in this study were able to successfully monitor and manage NIDDM patients in accordance with a standardized protocol. Gliclazide appeared to be an effective and well-tolerated treatment. The high prevalence of chronic diabetic complications at diagnosis emphasizes the delay encountered in reaching the diagnosis of NIDDM and the problems associated with this delay. In addition to the classic risk factors for NIDDM exhibited in this patient cohort, we have identified CHD and a maternal genetic component as further potential predicting factors.