The aim was to compare clinical efficacy and safety of two treatment regimens: biphasic insulin aspart (BIAsp) injected at all three meals plus neutral protamine Hagedorn (NPH) insulin at bedtime vs. a human insulin regimen, premixed human insulin at breakfast and soluble insulin at lunch and dinner and NPH at bedtime. A total of 167 adolescents (80 males and 87 females) with type 1 diabetes was included in the trial (multinational, randomized, open-label, and parallel group). Each subject received either of two treatment regimens for a 4-month period. BIAsp was injected immediately before main meals, human insulin products 30 min before meals, and NPH at night. Glycemic control was monitored by eight-point evaluations (after 6 and 16 wks) and hemoglobin A1c (HbA1c) (after 2, 6, and 16 wks). Safety evaluations included adverse events and incidence of hypoglycemic episodes. HbA1c (mean+/- SD) after 4 months on BIAsp (9.39+/- 0.14) was not significantly different from that with human insulin (9.30+/- 0.15). The average postprandial glucose increment in the BIAsp group was about half the increment in the human insulin group; the difference not statistically significant. The body mass index (BMI) increased in both groups, but significantly (p=0.005) less in the BIAsp group. However, in males on BIAsp, the BMI decreased compared with those on human insulin (p=0.007). No significant group differences were found for the rate of hypoglycemic episodes. We concluded that the BIAsp regimen was associated with similar glycemic control and similar incidence of hypoglycemic episodes as human insulin. However, the BIAsp regimen caused a significantly smaller increase in BMI, particularly in males, compared with the human insulin regimen.