Incidence of diabetic retinopathy: a systematic review and meta-analysis of prospective studies

Abstract

Purpose : The reported incidence of diabetic retinopathy (DR) varies widely between countries (4.2% to 71%) over a similar follow-up period of 10 years. Incidence also varied by type of diabetes and study population. We undertook a systematic review and meta-analysis of prospective studies to clarify the incidence of DR.

Methods : A literature search was undertaken using Medline and Embase from 1980 to 2016. Population or clinic-based cohort studies reporting the incidence of DR in either type 1 or type 2 diabetes, and two additional studies with individual-level data (Singapore) were included. Studies with unusually high reported DR incidence (>70%) or small sample size at baseline (<250) were excluded. Pooled cumulative incidence and 95% confidence interval (CI) were calculated using a logistic-normal random effects model. Subgroup analyses were performed by study population; region; type and duration of diabetes; year; and method of ascertainment of DR.

Results : 14 prospective studies (9 population- and 5 clinic-based) involving 10746 participants with diabetes at baseline were included. Six studies were from Asia, three each from Europe and North America and two from the Caribbean respectively. Eight studies reported on incidence in type 2, one in type 1, and five including both. Two studies reported incidence over two time-points. We identified 1984 new cases of DR over a mean follow-up of 5.7 years with a pooled incidence of 26% (95% CI: 18-36%). Incidence was higher in general population (29% vs. 21% in clinic), North America (42% vs. 22% in Europe and 19% in Asia), type 1 (45% vs. 26% in type 2), duration of diabetes ≥7.6 years (29% vs. 22% in <7.6 years), diagnosis of DR before year 2000 (31% vs. 22% after 2000) and by fundus photography (29% vs. 20% by ophthalmoscopy), all p<0.05.

Conclusions : Our analyses demonstrated that more than 1 in 4 with diabetes developed DR over a mean follow-up period of 5.7 years. The information provided could be used to monitor future trends in the incidence of DR, in particular in type 2 diabetes. Large prospective studies estimating the incidence of DR in type 1 diabetes are much needed.

This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.