An initial improvement in glycemic control is often followed by gradual deterioration of glycemia during insulin treatment of patients with noninsulin-dependent diabetes mellitus (NIDDM). We examined the causes of such worsening in a 12-month follow-up analysis of 100 insulin-treated NIDDM patients in the Finnish Multicenter Insulin Therapy Study who were treated with either combination therapy with insulin or insulin alone. In the entire study group, glycemic control averaged 9.7 +/- 0.2% at 0 months and 8.0 +/- 0.1%, 8.0 +/- 0.1%, 8.2 +/- 0.1%, and 8.5 +/- 0.2% at 3, 6, 9, and 12 months (P < 0.001 for each time point vs. 0 months). Glycemic control at 12 months was significantly worse than that at 3 (P < 0.001), 6 (P < 0.001), and 9 months (P < 0.02). Baseline body mass index was the most significant predictor of deterioration in glycemic control. During 1 yr, hemoglobin A1c decreased almost 3-fold more (by 1.7 +/- 0.2%; P < 0.001 vs. 0 months) in patients whose baseline weight was below the mean baseline body mass index of 28.1 kg/m2 (nonobese patients) than in those whose weight exceeded 28.1 kg/m2 (obese patients; 0.5 +/- 0.2%; P = NS vs. 0 months; P < 0.01 vs. obese patients). Glycemic control improved similarly over 1 yr in the nonobese subjects and deteriorated similarly in the obese patients regardless of their treatment regimen. Insulin doses, per body weight, were similar in the nonobese and obese patients. The nonobese patients consistently gained less weight during 12 months of combination therapy with insulin (3.5 +/- 0.6 kg at 12 months) than during insulin therapy alone (5.1 +/- 0.6 kg; P < 0.05). The treatment regimen did not influence weight gain in the obese group, who gained 4.4 +/- 1.0 kg during combination therapy with insulin and 4.5 +/- 1.1 kg during insulin therapy alone. We reached the following conclusions: 1) after an initial good response, glycemic control deteriorates more in obese than in nonobese patients with NIDDM; 2) in obese patients, weight gain per se cannot explain the poor glycemic response to combination or insulin therapy, but it may induce a disproportionately large increase in insulin requirements because of greater insulin resistance in the obese than in the nonobese; 3) in nonobese patients, glycemic control improves equally during 1 yr with combination therapy with insulin and insulin alone, but combination therapy with insulin is associated with less weight gain than treatment with insulin alone; 4) weight gain appears harmful, as it is associated with increases in blood pressure and low density lipoprotein cholesterol.