The dawn phenomenon in Type 1 (insulin-dependent) diabetes mellitus: magnitude, frequency, variability, and dependency on glucose counterregulation and insulin sensitivity

Abstract

In 114 subjects with Type 1 (insulin-dependent) diabetes mellitus the nocturnal insulin requirements to maintain euglycaemia were assessed by means of i. v. insulin infusion by a Harvard pump. The insulin requirements decreased after midnight to a nadir of 0.102±0.03 mU·kg-1·min-1 at 02.40 hours. Thereafter, the insulin requirements increased to a peak of 0.135±0.06mU·kg-1·min-1 at 06.40 hours (p<0.05). The dawn phenomenon (increase in insulin requirements by more than 20% after 02.40 hours lasting for at least 90 min) was present in 101 out of the 114 diabetic subjects, and its magnitude (% increase in insulin requirements between 05.00-07.00 hours vs that between 01.00-03.00 hours) was 19.4±0.54% and correlated inversely with the duration of diabetes (r = -0.72, p<0.001), but not with age. The nocturnal insulin requirements and the dawn phenomenon were highly reproducible on three separate nights. In addition, glycaemic control, state of counterregulation to hypoglycaemia and insulin sensitivity all influenced the magnitude of the dawn phenomenon as follows. In a subgroup of 84 subjects with Type 1 diabetes, the multiple correlation analysis showed that not only duration of diabetes (t = -9.76, p<0.0001), but also % HbA1 significantly influenced the magnitude of the dawn phenomenon (t = 2.03, p<0.05). After 5-9 months of intensive therapy, the magnitude of the dawn phenomenon decreased from 24+-2% to 18±2% (p<0.05) in seven Type 1 diabetic subjects with initially poor glycaemic control, whereas it increased from 17±2% to 24±3% (p<0.05) in five Type 1 diabetic subjects in whom glycaemic control had deteriorated for 2 weeks. In 18 Type 1 diabetic subjects the magnitude of the dawn phenomenon correlated with the indices of adequate glucose counterregulation, namely plasma glucose concentration at the hypoglycaemic nadir (r = -0.79) and the rate of plasma glucose recovery from hypoglycaemia (r = -0.74) (both p<0.01). Finally, in 10 diabetic subjects in whom insulin sensitivity was examined by the euglycaemic glucose clamp technique, there was a correlation between the residual rate of hepatic glucose production (r = 0.78, p<0.005) as well as between the rate of peripheral glucose utilization and the magnitude of the dawn phenomenon (r = -0.70, p<0.025). In conclusion, the dawn phenomenon is a very frequent event in Type 1 diabetes; its magnitude (∼20%) is much lower than that indicated by previous Biostator studies; it is highly reproducible from day to day; it is influenced by the duration of diabetes, glycaemic control, state of the counterregulation system to hypoglycaemia and insulin sensitivity. © 1991 Springer-Verlag.