In this analysis of a large insured population of patients with diabetes, the presence of DME is associated with a higher incidence of MI and CVA. Adjusted estimates show a 2.5-fold higher risk of MI and a 2-fold higher risk of CVA in patients with DME compared with diabetic patients without retinal disorders. Incident stroke appeared to be higher in women than men in either group, which is supported by data from other studies [17
Earlier studies have shown that the presence of diabetic retinopathy is a predictor of cardiovascular morbidity and mortality [4
]. The Diabetes Control and Complications Trial reported that tight control of diabetes with insulin reduced the 36-month risk of progression to retinopathy by 76% (95% CI: 62-85%) [18
]. In the case of type 2 diabetes, the UK Prospective Diabetes Study showed that for every 1% decrease in hemoglobin A1c (HbA1c), there was a 35% reduction in the risk of microvascular complications [19
Other studies have reported cardiovascular outcomes in DME that appear to support the findings of this study. The Wisconsin Epidemiologic Study of Diabetic Retinopathy (WESDR) examined the relationship between several diabetic eye diseases and mortality over a 16-year period [10
]. Patient cohorts with type I or II diabetes with disease onset before and after the age of 30
years were separately examined. In "younger-onset" diabetes subjects with DME (n
92), there was a non-significant decrease in likelihood of death from ischemic heart disease compared with similar subjects without DME (adjusted HR 0.84, 95% CI: 0.43-1.66). In "older-onset" DME subjects (n
136), there was a non-significant increase in the likelihood of death from either heart disease or stroke (adjusted HR 1.10, 95% CI: 0.76-1.58 and 1.17, 95% CI: 0.65-2.10, respectively) [10
]. The design and outcome measures of the WESDR study make its results difficult to directly compare with the results of the current study.
Cheung and colleagues examined the relationship between DR and cardiovascular events in two studies using data from The Atherosclerosis Risk in Communities study [6
]. In the first, investigators performed a prospective cohort study on 1617 patients with diabetes to quantify the relationship between DR and ischemic stroke. In subgroup analyses, there was a non-significant increase in the risk of stroke in subjects with DME (n
23, adjusted HR 1.40, 95% CI: 0.54-3.65) [6
]. Separately, Cheung et al. also found that the presence of DR was associated with 2-times the risk of incident coronary heart disease events (adjusted HR 2.07, 95% CI: 1.38-3.11) [11
]. However, in subgroup analyses, there was no change in the risk of coronary events in subjects with DME (n
81, adjusted HR 1.00, 95% CI: 0.56-1.79). With regards to DR, the authors concluded that their findings supported the role of microvascular disease in the pathogenesis of heart disease in diabetes. Unfortunately, neither study was sufficiently powered with DME patients to detect a significant difference.
Several limitations related to all administrative claims studies are worth mentioning. Previous studies have shown that diabetes duration, elevated HbA1c levels and uncontrolled hypertension are common risk factors for DR [20
] and cardiovascular events [21
]. As with most claims studies, the onset date of a chronic disease is difficult to determine. The incidence of DME increases with longer duration of type I diabetes [20
]. Diabetes duration is a predictor for the development of retinopathy and DME [22
] and is not measurable in this study. However, the cohorts in this study were age-matched, which should minimize any differences between the age of diabetes onset and therefore differences in the duration of diabetes. This does not fully account for residual confounding influenced by duration. In this study, HbA1c results were available for 25% of DME patients and 19% of controls with mean values of 7.3% and 7.0%, respectively (a sample of patients without diabetes had a mean value of 5.0%). The American Diabetes Association’s standards of care establish a HbA1c goal of 7%, so the subset of patients in this study appears to be within reasonable glycemic control. If this is the case, the difference in the available data for this study does not appear to be clinically significant, but glycemic control was not measured in all patients and therefore was not adjusted in the regression model. Blood pressure data were not available to evaluate the effect of uncontrolled hypertension, but overall hypertension diagnoses were adjusted in the analysis. In a post-hoc
analysis of untreated hypertension within each cohort, fewer DME patients had hypertension without antihypertensive therapy than did controls (9.3% vs. 12.7%). Thus, uncontrolled hypertension alone is unlikely to account for the difference in cardiovascular events.
Recently, the code combination 362.53 (cystoid macular edema) plus 250.xx (diabetes mellitus) was found to have high sensitivity and specificity for DME in a regional study of 22 ophthalmology practices [24
]. The current study did not use cystoid macular edema as a definition of DME in the inclusion criteria; therefore, additional subjects with DME may have been excluded. It is not known how the inclusion of this additional code would have affected the results. Lastly, administrative claims cannot differentiate between DME and clinically significant DME; any divergence in the rates of cardiovascular events in these clinical subgroups is not known.
One implication of quantifying baseline rates of cardiovascular events in DME is the advent of agents that locally suppress vascular endothelial growth factors (VEGF) in the retina but that may also have systemic effects. Studies have shown that VEGF inhibitors may be effective in treating DME [25
]. However, if VEGF inhibition reaches the systemic circulation, it may play a role in potentiating cardiovascular complications in diabetes [26
]. As they are adopted into clinical practice, additional population research will be needed to evaluate the long-term safety of advanced drugs that target the microvascular complications of diabetes.