The relation of glycaemia to the risk of development and progression of retinopathy in the Diabetic Control and Complications Trial

Aims/hypothesis. We have assessed the relation between the quarterly capill ary glucose profile and the risk of the development and progression of reti nopathy in the DCCT. Methods. Seven point (preprandial and 90-min postprandial for each meal and bedtime) capillary glucose data were analysed from volunteers whose collec tions were complete in 80%, or more, of quarterly periods and who were in t he study longer than 4 years (n = 296, conventional therapy; n = 269, inten sive therapy). The study cohort differed from excluded patients in having m ore women and lower HbA(1c) at baseline and fewer adolescents, older age an d lower baseline mean blood glucose in the intensive therapy group. Results. Univariate analysis showed significant (p < 0.01) associations to sustained 3-step change in retinopathy of each updated glycaemic parameter: mean blood glucose, mean preprandial glucose, mean postprandial glucose, e ach preprandial, postprandial and bedtime glucose; range glucose, standard deviation glucose; M-value of Schlichtkrull and mean amplitude of glycaemic excursions, albeit with relatively small hazard ratios. Multivariate analy ses showed updated mean blood glucose to be the primary risk factor (p < 0. 001) with a weak contribution of mean amplitude of glycaemic excursions at baseline (p < 0.005); no other variables added significantly to the model. The association between updated mean blood glucose and risk for retinopathy was nonlinear: risk progressively increased above updated mean blood gluco se of 8.3 mmol/l. A gradient of risk could not be determined below this lev el because events were few. Conclusion/interpretation. Within the limitations provided by quarterly 7-p oint capillary glucose measurements as an expression of overall glycaemic b ehaviour, the major risk for progression of retinopathy is conveyed by upda ted mean blood glucose especially above 8.3 mmol/l.