Abstract
Purpose :
Many prior studies showing an inverse relationship between myopia and diabetic retinopathy (DR) severity based their results on a single clinical encounter and single HbA1c result; failing to account for prior glycemic control or potential fluctuations between early DR and no DR. We conducted a retrospective chart review to assess the impact of axial length (AL) on DR, using extensive HbA1c data and multiple assessments of DR.
Methods :
Billing records identified patients who had cataract surgery from 2008-2015. Patients were included if they had diabetes mellitus and had at least 2 dilated eye exam assessing DR severity, with one occurring over 1 year prior to surgery. Exclusion criteria included: treatment of other retinal disease, retinal vascular occlusion, or uveitis. Collected data included: most severe DR rating; prior HbA1c, AL from pre-operative ocular biometry, presence of hypertension and hyperlipidemia, previous DR treatment, and presence of macular edema (ME) prior to surgery or development of ME after surgery. Statistical analysis was done via multinomial regression models and receiver operating characteristic curves.
Results :
276 right eyes (OD) and 255 left eyes (OS) were reviewed. For each 1mm increase in AL; OD had 47% (p<.0001) decreased odds of having proliferative DR (PDR) vs no DR, and OS had 38% (p=.004) decreased odds of having PDR vs no DR. Patients without diagnosed ME OD after surgery were 52% (p=.07) less likely to have background DR (BDR) and 62% (p=.01) less likely to have PDR; while OS was 93% less likely to have BDR (p=.001) and 93% (p=.001) less likely to have PDR. There was also a significant difference in median HbA1c values between each DR group (p<.0001). Analysis determining sensitivity and specificity cut offs for predicting macular laser treatment found an OD AL of 24.75mm had a 93% sensitivity and 16% specificity; while OD AL of 22.05 had a 17% sensitivity and 95% specificity. Similarly, OS AL of 25.55mm had a 95% sensitivity and 9% specificity; while OS A L of 22.25 had an 18% sensitivity and 90% specificity.
Conclusions :
Increased AL was associated with less severe DR. Despite this significance, AL failed to be a good predictor of macular laser treatment requirements. Additionally, pre-operative DR severity was a significant predictor of post-operative ME. Our results are consistent with the literature, the use of multiple assessments of DR severity and multiple HbA1cs solidifies these relationships.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.