INTRODUCTION: Basal insulin dose requirements in patients with type 1 diabetes may be derived from the course of glucose concentrations in the fasting state; i. e. by skipping meals. The present study examined whether fasting tests could be replaced by carbohydrate-free meals. MATERIAL AND METHODS: 16 adult patients with type 1 diabetes (10 male) on intensive insulin therapy participated in this prospective intervention study. Mean age (+/-SD) was 44+/-12 years, BMI 24+/-3 kg/m (2), mean HbA1c was 7.5+/-0.6% and duration of diabetes 15+/-12 years. All participants skipped dinner and plasma glucose concentrations were hourly monitored from 7 p.m. to 11 p.m. This blood glucose profile was compared with three test meals given at 7 p.m. at day 2-4, consumed either in the hospital (meal 1) or at home (meal 2 and 3). No insulin injection (except to basal insulin) was allowed. Test meals consisted of 2.5 g carbohydrate, 32.4 g protein, 52.0 g fat (according to 612 kcal). RESULTS: During 16 fasting tests plasma glucose concentration remained stable between 7.2+/-2.4 mmol/l at 7 p.m. and 6.8+/-2.8 mmol/l at 11 p.m. (p=0.461). Following the intake of near carbohydrate-free meals (48 tests), plasma glucose concentrations rose within 4 h from 6.7+/-2.0 at 7 p.m. to 9.8+/-3.4 mmol/l at 11 p.m. (p<0.0001). The increase in plasma glucose concentrations was similar in all three different meals tested. CONCLUSION: Plasma glucose concentrations significantly increase in patients with type 1 diabetes following the intake of carbohydrate-free meals. Carbohydrate-free meal-tests cannot replace skipping meal tests to determine the basal insulin requirement in patients with type 1 diabetes.