Abstract
Purpose :
To assess if specific findings gathered during an examination of patients with a known angle closure diagnosis (probands) could better determine the risk of angle closure in the patient’s sibling.
Methods :
Patients 30 years and older with suspect primary angle closure (PACS) or primary angle closure/primary angle closure glaucoma (PAC/G) seen at the Aravind Eye Hospital in Pondicherry, Tamil Nadu, and a biological sibling above age 30 years, were recruited as ‘probands’ and ‘siblings’ (n=346 pairs). Demographics, ocular history, and an ophthalmic examination with Anterior Segment Optical Coherence Tomography (ASOCT) were obtained. Siblings were classified as having open angles (OA), PACS, or PAC/G per ISGEO criteria. Models were created to analyze the contribution of specific proband factors in predicting sibling angle closure diagnosis.
Results :
When predicting (1) any sibling angle closure (PACS or PAC/G ) vs. open angles, (2) sibling PAC/G vs PACS or OA, or (3) PAC/G vs. PACS amongst siblings with angle closure, models incorporating proband ASOCT data outperformed models including proband diagnosis alone or proband diagnosis plus demographic (age/gender) and exam metrics (gonioscopy, optic nerve exam, visual acuity, and intraocular pressure). For example, in the prediction of PAC/G vs. PACS amongst siblings with angle closure, the model using only proband diagnosis had the lowest explained variability (0.64%). Adding proband demographics and ocular exam metrics improved model accuracy to a limited extent (1.99% of variability explained), while adding ASOCT Metrics created a significant improvement (17% of variability explained, p<.0001 vs other models). The model incorporating ASOCT metrics resulted in the lowest Bayesian Information Criterion (101.9 vs. 123.6 for the diagnosis-only model and 119.7 for the diagnosis + demographics + exam metrics model).
Conclusions :
Utilizing ASOCT information from angle closure patients can help better predict sibling angle closure status, leading to more efficient and cost-effective screening of family members.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.