PURPOSE: To determine the risk factors for retinopathy progression in type 1 (insulin-dependent) diabetes mellitus in a prospective cohort study.

SUBJECTS AND METHODS: Subjects were 485 participants in the Sorbinil Retinopathy Trial, a randomized trial of aldose reductase inhibition among patients aged 18 to 56 years with type 1 diabetes mellitus (duration of 1 to 15 years) and no or only mild retinopathy. Retinopathy progression, assessed by seven-field stereoscopic fundus photography, was defined as worsening by two or more levels on a standardized grading scale at the end of follow-up (median, 41 months).

RESULTS: The relative risks for retinopathy progression according to successively greater quintiles of total glycosylated hemoglobin level at baseline, after adjusting for age, diabetes duration, sorbinil assignment, and other variables, were 1.0, 2.0, 1.6, 3.7, and 4.4 (P trend <0.0001). Risk increased with greater baseline diastolic blood pressure: 1.0 for <70 mm Hg, 1.2 for 70 to 79 mm Hg, and 1.8 for ≥80 mm Hg (P for trend = 0.04). Diastolic blood pressure was a significant risk factor for progression in participants with mild baseline retinopathy (P for trend <0.02) but not in those without retinopathy at entry. Systolic blood pressure, by comparison, was not associated with progression. Baseline total cholesterol level was a marginally significant predictor of retinopathy progression when examined as a categorical variable (relative risks for increasing quartiles; 1.0, 1.6, 1.8, 1.9; P for trend = 0.03) but not when it was examined as a continuous variable or when hypercholesterolemic patients were compared with those with normal levels. Furthermore, when cholesterol levels were updated in subsequent visits, it was not a significant predictor of progression, and low density lipoprotein (LDL) cholesterol levels did not predict progression no matter how analyzed. Smoking was not associated with progression of retinopathy.

CONCLUSIONS: Levels of hyperglycemia and diastolic blood pressure predicted progression of retinopathy in type 1 diabetes mellitus. We found only a suggestion of an association between total cholesterol level (but not of LDL cholesterol level) and progression of retinopathy; resolution of this issue will require additional studies with larger sample sizes and longer follow-up.