Prospective Association between Diabetic Retinopathy and Cardiovascular Disease – A Systematic Review andMeta-analysis of Cohort Studies

Methodological quality of the review: Low confidence

Author: Guo VY, Cao B, Wu X, Lee JJW, Zee BC.

Region: Finland, US, Sweden, United Kingdom (UK), Japan, Australia, Italy.

Sector: Diabetic retinopathy

Sub-sector: cardiovascular disease

Equity focus: None specified

Review type: Other review

Quantitative synthesis method: Meta-analysis

Qualitative synthesis method: Not applicable

Background:

The presence of diabetic retinopathy (DR) is associated with increased risk of cardiovascular (CV) disease incidence and CV disease-related mortality in both type 1 diabetes mellitus (T1DM) and type 2 diabetes mellitus (T2DM). Furthermore, there was a positive association between DR severity and risk of CV disease demonstrated in the Action to Control Cardiovascular Risk in Diabetes trial. However, the findings were not consistent across reports

Objectives:

To determine the association between DR and CV disease, and to investigate the factors that influence the association.

Main findings:

A total of 13 cohort studies (5 in T1DM and 8 in T2DM) were included in the analysis. Of these, 2 studies in T1DM and 5 studies in T2DM reported the adjusted association between DR and CV disease events with adjustments of other risk factors and confounders.

Subgroup analyses showed that the pooled risk ratios (RRs) were different between T1DM and T2DM. Univariate meta-regression revealed that diabetes type was the main cause of the heterogeneity between T1DM and T2DM, whereas age, male percentage, and follow-up period did not explain the variance among subgroups. In studies with T1DM patients, we found that systolic blood pressure was the main cause of heterogeneity.

Authors noted that funnel plot showed a certain degree of asymmetry. Begg’s test revealed significant publication bias (P = .024), whereas Egger’s test showed nonsignificant publication bias (P = .253). Based on the meta-analysis authors reported that DR was associated with increased risk of CV disease (relative risk [RR]: 2.42, 95% confidence interval [CI]: 1.77-3.31) in diabetes. Specifically, the RR was 3.59 (95% CI: 1.79-7.20) for type 1 diabetes and 1.81 (95% CI: 1.47-2.23) for type 2 diabetes. Significant heterogeneity was found in studies with type 1 diabetes. Meta-regression analysis showed that baseline systolic blood pressure was a key factor leading to the heterogeneity. Authors concluded that DR is significantly associated with CV disease incidence and CV disease-related mortality in diabetes. Patients with DR may need more intensive management to control future CV disease attacks.

Methodology:

Inclusion criteria consisted of: (1) the study design was a cohort study; (2) the study populations should be patients with either T1DM or T2DM, or the association between DR and CV disease was assessed separately in different diabetes types; (3) the included study should report data on the first attack of fatal and/or nonfatal CV disease, rather than the recurrent CV disease or with previous surgery; (4) the study should be written in English; and (5) the study should not be a letter, review, and commentary article. The quality of the cohort studies was assessed using the Newcastle–Ottawa scales. Articles with a score of 6 or higher were considered of high quality. Authors conducted a search on MEDLINE and EMBASE from their inception to October 2015. Authors restricted to studies that reported the following diseases as an outcome: fatal or nonfatal CV disease, cerebrovascular disease, CHD, and stroke. The following search terms without restrictions: “retinopathy” and (“coronary” or “myocardial” or “stroke” or “cardiovascular disease” or “cerebrovascular diseases”) and “diabetes.” In addition, authors further manually searched the reference list of identified reviews and research articles to include any eligible study.

The data were collected using a form incorporating study details (author name, publication year, etc.) and characteristics of the study population (age, percentage of male patients, etc.), as well as the definition and evaluation criteria for DR and CV disease.

Potential publication bias was visually inspected through Begg’s funnel plot. We further performed Begg’s andEgger’s tests, and a P value less than .05 indicated significant publication bias. RR was used to measure the association between DR and CV disease risk across studies. Subgroup analyses were conducted based on diabetes types. To explore the potential causes of the heterogeneity, authors performed meta-regression analysis. To assess the influence of single study, a “leave-one-out” sensitivity analysis was conducted by omitting 1 study at a time. The meta-analysis and statistical analyses were performed using Review Manager software.

Applicability/external validity:

Authors noted that results of this review indicated that the presence of DR might provide more accurate CV risk stratification in patients with diabetes.

Geographic focus:

Authors did not include studies conducted on low- and middle-income countries and did not discuss how findings may be applicable to these settings.

Summary of quality assessment:

Overall, low confidence was attributed in the conclusions about the effects of this review. Authors did not report if screening of eligible studies and data extraction of included studies were conducted by two reviewers independently, avoiding biases. In addition, authors did not conduct a thorough search of the literature to ensure that all relevant studies were included in the review.