In a randomized cross-over study we compared blood glucose control and patient acceptance of a 12-week basal-prandial regimen with short-acting insulin before meals and isophane (NPH) insulin at bedtime (4 injections) with a scheme with a second injection of isophane (NPH) insulin before breakfast (5 injections). Forty-three Type 1 diabetic patients (age 37 +/- 11 (+/- SD) years, duration of diabetes 15 (range 2-48) years, 26 males and 17 females) completed the study. Mean daily blood glucose was 8.6 +/- 2.4 mmol l-1 at baseline, and 8.1 +/- 2.2 mmol l-1 after the four-injection period and 7.9 +/- 2.0 mmol l-1 with five-injections (NS). HbA1c after 12 weeks was not different with the two treatments (6.6 +/- 1.1 vs 6.5 +/- 0.9%), neither was fasting blood glucose (9.6 +/- 4.2 mmol l-1 with 4 injections, and 9.0 +/- 4.4 mmol l-1 with 5 injections). Daily insulin dose did not differ between regimens (55 vs 56 U day-1). No differences in number or severity of hypoglycaemic events were observed. After the study, 13 patients preferred to continue the 5-injection regimen, and 21 patients preferred 4 injections. Treatment satisfaction with either regimen was equally high. It is concluded that dividing the intermediate-acting insulin into a morning and an evening dose did not lead to an improvement in blood glucose control in these moderately-controlled Type 1 diabetic patients.