The period effect in the prevalence of proliferative diabetic retinopathy, gross proteinuria, and peripheral neuropathy in type 1 diabetes: A longitudinal cohort study.

Abstract

To investigate whether, for a specific duration of type 1 diabetes, there is a significant change in the prevalence of proliferative diabetic retinopathy, gross proteinuria and peripheral neuropathy in those more recently diagnosed with diabetes (a period effect), in the Wisconsin Epidemiologic Study of Diabetic Retinopathy. Where present, to determine how common risk factors for diabetic complications might be associated with it, and what might be driving it. Longitudinal cohort study with seven examination phases between 1980 and 2014. Multivariate logistic regression models and ordinal parameterization were used to test for and evaluate any period effect. There is a period effect in the prevalence of gross proteinuria and peripheral neuropathy (decreasing), as seen with proliferative diabetic retinopathy (p < 0.001). Adjusting for changing levels of common risk factors attenuates the period effect, particularly for proliferative diabetic retinopathy. For gross proteinuria and peripheral neuropathy, however there is a persistent period effect in spite of adjusting for the major risk factors. There are period effects in the prevalence of proliferative diabetic retinopathy, gross proteinuria and peripheral neuropathy that cannot be fully explained by changes in common risk factors for complications of type 1 diabetes in this cohort. The role of other potential confounders warrants further exploration.