Metformin, glibenclamide and their combination were compared in a randomized, double-blind trial in patients with non-insulin-dependent diabetes mellitus (NIDDM) using a parallel group design. The study was performed in primary health care, and the purpose was to assess possible synergistic effects of combination therapy with the two drugs as primary treatment versus conventional oral therapy, starting with one drug and adding the other, if necessary. Lipids and lipoproteins were measured in the study, and preliminary results are reported for one hundred sixteen patients concluding 6 months maintenance therapy. Comparison of mean differences showed that patients randomized to combination therapy (n = 60) demonstrated a greater decrease in total- and LDL-cholesterol levels after 4 and 6 months treatment than patients randomized to conventional therapy starting with either metformin (n = 28) or glibenclamide (n = 28). For LDL-cholesterol a significant difference was also found between patients solely on monotherapy with lower values after treatment with metformin. Triglycerides did not change significantly, and only minor fluctuations were seen in HDL-cholesterol, independent of treatment. Obese patients had significantly higher triglyceride concentrations than the non-obese group, both at baseline and after treatment, as well as significantly lower HDL-cholesterol levels. The mean triglyceride concentration (+/- SD) after 6 months treatment was 2.32 +/- 1.38 mmol/l in the obese group (n = 69) and 1.54 +/- 0.84 mmol/l in non-obese (n = 47). For HDL-cholesterol the corresponding values were 0.83 +/- 0.23 mmol/l in obese and 0.93 +/- 0.29 mmol/l in non-obese patients.(ABSTRACT TRUNCATED AT 250 WORDS)