This study investigates the relationship between diabetic retinopathy (DR) and birth weight (BW) in diabetic subjects sampled from the general population.
The Gutenberg Health Study (GHS) is a population-based, observational cohort study in participants aged from 35 to 74 years. Criteria for diabetes diagnosis were HbA1c ≥6.5% at study entry, a doctor-diagnosis of diabetes, or diabetes medication. The presence of DR was determined by evaluating fundus photographs. BW was assessed by self-reports. GHS participants were divided into three different BW groups (low: 4000 g). Logistic regression analysis was conducted as uni- and multivariable analysis with adjustment for age and sex. Effect mediators were separately investigated.
A total of 1,124 GHS participants (7.5% of the cohort) had diabetes at study entry. Of these, 402 subjects (35.8%) had gradable fundus photographs, reported BW data and were included into this study. Overall, 91/402 subjects (23%) had DR. With regard to BW groups, DR was descriptively more frequent in subjects with low (28.1% [95%-CI: 14.4-47.0%; n = 32]) and high BW (30.8% [95%-CI: 19.1-45.3%; n = 52]) compared to normal BW (20.8% [95%-CI: 16.5-25.7%; n = 318]). Both high and low BW were associated with DR in multivariable analysis (high: OR = 1.68, = .037; low: OR = 1.81, = .05). The BW effect was mediated by duration of diabetes in both BW groups and by arterial hypertension in the low BW group.
Low and high BW in persons with diabetes is related to higher risk of diabetic retinopathy. Longer duration of diabetes and higher prevalence of arterial hypertension are factors in these subjects explaining the elevated risk.

Author