OBJECTIVE To compare results obtained with metformin versus those obtained with DNA-recombinant insulin in obese patients with NIDDM suffering from secondary failure to sulfonylureas. RESEARCH DESIGN AND METHODS We conducted an open, prospective, randomized, and comparative study comprising a total of 60 patients selected and placed in two parallel groups. We had previously confirmed that the subjects had secondary failure to high doses of sulfonylureas. The initial metformin dosage was a single 850 mg tablet, and the dosage was increased to two or three tablets depending on the patient's metabolic changes. The initial dosage of DNA-recombinant insulin was 24 U, subcutaneously administered and divided into two portions: two-thirds at around 8:00 A.M., before breakfast, and the remaining third at 8:00 P.M., before dinner. The dosage was adjusted based on the patient's clinical and metabolic response. RESULTS The initial average glucose value for the metformin group was 269.1 ± 32.2 mg/dl, decreasing by the end of the study to 159.7 ± 30.5 mg/dl. For the insulin group, these figures went from 270.7 ± 24.0 mg/dl at the beginning of the study to 134.8 ± 26.7 mg/dl. This decrease correlates with the reduction in glycosylated hemoglobin from 12.8 to 8.9% for the first group and from 12.3 to 8.2% for the second, as well as with the reduction in triglyceride values from 230.3 to 183.1 mg/dl and from 218.4 to 186.3 mg/dl, respectively. The BMI (27.5–26.4), blood pressure (systolic from 145.7–132.1 mmHg, diastolic from 90.3–84.8 mmHg), and total cholesterol levels (235–202 mg/dl) decreased in only the metformin group. CONCLUSIONS Metformin is an effective, safe, and well-tolerated treatment that improves metabolic control and favorably modifies secondary clinical alterations due to insulin resistance, such as arterial hypertension, overweight, and hyperlipidemia, in obese patients with NIDDM suffering from secondary failure to sulfonylureas.