Abstract
Purpose :
Automated detection and referral for more-than-mild diabetic retinopathy in primary care offices has the potential to triage care and improve adherence to recommended guidelines. However, this referral policy carries potential risk of falsely reassuring patients with less-than-moderate diabetic retinopathy (ltmDR) who may have other incidental eye disease, disrupting the existing pathway for access to regular exams, screening, and eye care. We used National Health and Nutrition Examination Survey (NHANES) data to determine the prevalence of incidental eye disease findings (IEDF) in participants with diabetes classified as having none or mild diabetic retinopathy. We compared participants with and without IEDF to identify factors associated with IEDF.
Methods :
Our NHANES analysis included participants with diabetes (self-reported or HbA1c > 6.5%) aged > 40 with available fundus photos and classified as none or mild diabetic retinopathy in their worse eye. We classified participants with IEDF as those with at least one of twelve non-diabetic eye disease fundus photography findings of either eye (including AMD, glaucoma, and choroidal nevus). We compared ltmDR participants with IEDF versus ltmDR participants without IEDF across several demographic and clinical characteristics. Considering survey design, t-test was used to compare continuous variables and Rao-Scott Chi-Square test was used for categorical variables.
Results :
Among the 757 (weighted n=11.4M) participants with diabetes and ltmDR, 44.5% of participants (weighted n=5.1 M) had IEDF. Comparatively, 36.3% of 4,053 (weighted n = 90.5M) participants aged greater than 40 without diabetes had IEDF. The most prevalent IEDF finding in participants with ltmDR was epiretinal membrane which was found in 17.2% (weighted n= 1.7M) of participants. IEDF was more prevalent in older participants, participants with elevated blood pressure, and participants who had health insurance (all p < 0.05). There were no other statistically significant differences between participants with ltmDR and IEDF and ltmDR without IEDF.
Conclusions :
Among participants with ltmDR older than forty years old, the prevalence of IEDF was 44.5%. Age, blood pressure, and health insurance status were associated with IEDF in this population. Further study of IEDF prevalence is important to assess the risks of adopting fully automated DR screening from current practice.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.