Methodological quality of the review: Low confidence

Author: Ruta LM, Magliano DJ, LeMesurier R, Taylor HR, Zimmet PZ, Shaw JE.

Region: India, Mauritius, China, Mexico, Iran, Fiji, Sri Lanka and South Africa.

Sector: Diabetic retinopathy

Sub-sector: Type 2 diabetes, developed countries, developing countries.

Equity focus: None specified

Review type: Effectiveness review

Quantitative synthesis method: Narrative analysis

Qualitative synthesis methods: Not applicable

Background

As the global prevalence of diabetes increases, so does the numbers of people with diabetic retinopathy (DR). Authors note that, with nearly 80% of all people with diabetes living in low- and middle-income countries, the burden of diabetes is felt most in developing countries. Healthcare and support for patients with diabetes tends to be inadequate or non-existent, which makes it more likely that complications such as DR will occur. Although there have been a number of studies on the prevalence of diabetes, less is known about diabetic complications. It is unclear how many epidemiological studies of DR have covered the global population and what methods have been used in these studies.

Research objectives

‘Our review aims to provide a comprehensive picture of available studies of diabetic retinopathy and how prevalence varies around the developed and developing world’.

Main findings

A total of 72 studies from 33 countries were included in the review, 13 of which were from developing countries. There were six register based, 23 primary care-clinic based and 43 population based studies. All studies addressed the prevalence of DR.

Prevalence estimates varied from as low as 10% to as high as 61% in persons with known diabetes, and from 1.5 to 31% in newly diagnosed diabetes. Across all the studies, the median prevalence of any DR in known diabetes was 27.9% and 10.5% in newly diagnosed diabetes.

The prevalence of DR was higher in developing countries. In 15 of the 23 population based (community derived) studies in developing countries and in ethnic minority groups within developed countries, the prevalence of DR was of 35% or over. In developed countries only two of 16 studies reported prevalence of 35% or over.

Authors concluded from the review that there were significant gaps in population based data from developing countries and indigenous populations on the prevalence of DR. Additionally, due to the diversity of studies, comparisons were very difficult. Authors noted that due to the heterogeneity of the studies included, a true single prevalence estimate of DR was not possible. As such, they recommended more standardized studies in order to establish accurately DR prevalence rates as this would be able to inform prevention and treatment strategies.

Methodology

Studies which fulfilled the following criteria were included in the review: (1) full-text publications; (2) in humans; (3) written in English; (4) between 1985 and 2012; (5) attempted to describe the prevalence of DR in a defined population; (6) population-based, register-based or primary care clinic studies; and (7) contained at least 150 participants.

A search was conducted on PUBMED of studies written in English only and published between 1985 and 2012. Reference lists of selected articles were also reviewed for additional relevant articles. Relevant studies were reviewed by two authors. However, it was not clear if this was conducted independently. The following information was extracted from studies where possible: country, ethnicity, study period, geography (urban vs. rural), study design, mean age and age range (median age is reported if mean was not available), duration of diabetes, sample size, diabetes classification (i.e. Type 1, Type 2, known or newly diagnosed), prevalence and severity of diabetic retinopathy, method used to diagnose diabetic retinopathy, criteria used to classify severity of retinopathy and personnel grading the retinopathy. It is not clear from the review if data extraction was conducted by two reviewers.

Due to the heterogeneity of the included studies, it was clear that pooling data was not possible and a narrative synthesis was included in the review.

Applicability/external validity

Authors noted that due to the heterogeneity of the studies included, a true single prevalence estimate of DR is not possible. The results discussed offered a general trend on the prevalence of diabetic retinopathy but due to a lack of evidence, it may not be applicable to all country contexts.

Geographic focus

Authors noted that in 15 of the 23 population based studies in developing countries and in ethnic minority groups within developed countries, the prevalence of DR was 35% or over.

In developed countries only two of 16 studies reported a prevalence of 35% or over.  Authors note that this disparity could be attributed to poverty, poor nutrition, poor access to healthcare and lack of medications within developing countries. More standardized research needs to be conducted in developing countries to further investigate any disparities.

Quality assessment

A low confidence was attributed in the conclusions about the effects of this study as major limitations were identified. The search strategy was not sufficiently comprehensive as only one database and reference lists in included studies were searched. Therefore, we could not be confident that relevant studies were not omitted in the review. It was unclear if authors used appropriate methods to select studies and extract data of included studies; as such, it was not possible to determine if authors avoided risk of bias. Additionally, authors did not report the criteria used to quality assess included studies. However, it was recognised that due to a lack of comprehensive data and the diversity of the studies, combining the data was not possible and limitations around the heterogeneity of the studies were acknowledged. No strong conclusions were drawn other than a need for more research on the prevalence of DR globally.