Meta-Analysis of Diagnostic Accuracy of Retinopathy for the Detection of Diabetic Kidney Disease in Adults With Type 2 Diabetes.

Abstract

OBJECTIVES\nThe study aimed to explore whether diabetic retinopathy (DR) could distinguish diabetic kidney disease (DKD) from nondiabetic renal diseases (NDRDs) in patients with type 2 diabetes mellitus and renal disease.\n\n\nMETHODS\nWe searched PubMed, Embase, Cochrane, MEDLINE and China National Knowledge Internet for articles that enrolled patients with DKD and NDRD. The results were summarized as sensitivity, specificity and the area under the curve of summary receiver operating characteristic curve with their 95% confidence intervals (CIs).\n\n\nRESULTS\nA total of 51 studies that included 4,990 participants were collected for evaluation. The overall pooled sensitivity, specificity and area under the curve with their 95% CIs were 0.67 (95% CI 0.61, 0.73), 0.77 (95% CI 0.72, 0.81) and 0.78 (95% CI 0.75 to 0.82), respectively. If the test for DR is negative, the probability of DKD would decrease to 10%, but if the test for DR is positive, the probability would increase only to 42%. In addition, although the mean specificity of proliferative DR for detection of DKD was 0.98 (95% CI 0.92 to 1.00), the mean sensitivity was 0.25 (95% CI 0.16, 0.35).\n\n\nCONCLUSIONS\nDR may lack adequate evidence either to verify DKD or to exclude NDRD, and the severity of DR may not parallel the presence of DKD.