Purchase this article with an account.
C. B. Meyerle, C. M. Greven, R. P. Danis, E. Y. Chew, ACCORD Eye Study Group; Thiazolidinediones and Diabetic Macular Edema: Is There an Association?. Invest. Ophthalmol. Vis. Sci. 2007;48(13):1421.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
We studied participants in the eye substudy of Action to Control Cardiovascular Risk Factors in Diabetes (ACCORD) to determine if thiazolidinedione (TZD) therapy is associated with diabetic macular edema (DME) and if this relationship is confounded by age, HbA1c, duration of diabetes, gender, race, and retinopathy.
We examined the relationship between DME and TZD exposure using the baseline visits of the NEI-funded ACCORD-Eye substudy (n=3,537) of the NHLBI-funded ACCORD (n=10,251). ACCORD-Eye participants underwent an ophthalmic exam with stereoscopic fundus photos at baseline. We will also examine these same participants again at 4-years of follow-up. Exclusion criteria included previous laser photocoagulation or vitrectomy. DME and retinopathy severity were assessed by masked graders. DME was scored for each eye as 0=No DME, 1=Questionable DME, 2=Zone of retina thickening ≥1 disk area part ≤ 1 disc diameter from center, or 3=Retinal thickening or adjacent hard exudates ≤ 500um from center. Retinopathy was graded using ETDRS scale. A generalized estimating equation (GEE) model was used to examine relationships between TZD exposure and DME at baseline, controlling for potential confounders
Of the 3468 participants (6865 eyes) analyzed, 217 people (284 eyes) had DME scores of ≥1. 694/3468 (20.0%) had TZD exposure at baseline. In an unadjusted analysis, GEE indicated that TZD exposure was not associated with the presence of DME at baseline (OR=1.03, 95% CI 0.72-1.46, P=0.89). Adjusting for potential confounders yielded similar results (OR=0.97, 95% CI: 0.67-1.40, P=0.85). Retinopathy (4 df, P<0.0001) and age (OR=0.97, 95% CI: 0.952 to 0.997, P=0.0291) were significant predictors of DME. HbA1c (OR=1.15, 95% CI: 1.00-1.32, P=0.06) was marginally significant. Duration of diabetes (OR=1.00, 95% CI 0.98-1.02, P=0.65), female gender (OR=1.06, 95% CI: 0.78-1.43, P=0.71), and race (4 df, P=0.33) were not significant.
No association between TZD exposure and DME was observed in the ACCORD-Eye baseline population. Due to the size of the confidence interval (0.67-1.40), we cannot rule out the possibility of either a modest protective or deleterious effect of TZD exposure on DME. A more definitive answer will be provided from the 4-year follow-up data which will allow us to examine prospectively for the relationship between TZD exposure and DME incidence.
This PDF is available to Subscribers Only