Risk factors for progression to proliferative diabetic retinopathy in the EURODIAB Prospective Complications Study

Abstract

Aims/hypothesis: Proliferative diabetic retinopathy (PDR), a leading cause of blindness, cannot be totally prevented by optimizing metabolic and blood pressure control and responds to no specific treatment other than partially destructive retinal photocoagulation. Recognizing risk factors using large-scale epidemiological studies could help identify targets for treatment. The EURODIAB Prospective Complications Study (PCS) includes the largest cohort so far of patients with Type I (insulin-dependent) diabetes mellitus.

Methods: Baseline data were collected between 1989 and 1991 on 3250 patients who were recalled for follow-up. Physical examination, biochemical tests and assessment of complications were done on both occasions. In particular, 1249 patients had retinal photographs taken both basally and after an average of 7.3 years.

Results: Proliferative retinopathy had developed in 157 patients (cumulative incidence 17.3/1000 patient-years; 95%-CI: 13.6-21.1). HbA(1c) (standardized regression estimate--SRE = 3.03, CI 2.49-3.69), diabetes duration (1.71, 1.42-2.06), age at diagnosis < 12 (1.66, 1.11-2.50), diastolic blood pressure less than or equal to 83 (1.50, 1.03-2.20) and waist-to-hip ratio (1.50, 1.03-2.20) were all independent predictors for progression to PDR when entered simultaneously into a logistic regression model. Including retinopathy at baseline maintained the effects of metabolic control and pre-pubertal onset only. Including the albumin excretion rate maintained the effect of control but reduced SRE for pre-pubertal onset to 1.49 (0.94-2.33). There was no evidence for a threshold effect for HbA(1c)concentrations at baseline and progression to proliferative retinopathy.

Conclusion/hypothesis: Metabolic control and duration of diabetes are strong indicators of progression to proliferative retinopathy. Onset of diabetes before puberty could be an additional independent risk factor.