Evaluation Of Ocular Trauma Score Accuracy In Predicting The Surgical Outcomes Of Traumatic Corneal Perforation Patients In The United Kingdom: A Single-Centre Retrospective Study

Ocular Trauma Score (OTS) is a well-defined and widely-used scoring system to estimate the likelihood of visual acuity (VA) improvement following ocular trauma. However, there is a lack of research in evaluating the accuracy of OTS in predicting VA improvement in traumatic corneal perforation cases in the United Kingdom (UK), especially in the Midlands. This study aims to evaluate the predictive accuracy of OTS in the British population by comparing the OTS in patients with traumatic corneal perforation against their actual post-operative VA.

This study was conducted at Birmingham and Midland Eye Centre, Sandwell and West Birmingham NHS Trust (BMEC), a tertiary eye care centre, in Birmingham, Midlands UK. It mainly serves patients from the West Midlands region, and is among the top three largest eye hospitals in the UK, receiving a diverse range of cases, including ocular trauma cases, with its urgent care services.

A single-centre retrospective study was conducted on patients who visited BMEC eye emergency services, were diagnosed with full-thickness traumatic corneal perforation between January 2022 and July 2024 and had been operated on their injuries. OTS at the time of presentation was calculated based on the best-corrected visual acuity at presentation (initial BCVA), whether there was initial globe rupture, endophthalmitis, perforating injury, retinal detachment, and/or relative afferent pupillary defect. To evaluate the predictability of OTS in our patients, we compared the OTS against their true, final post-operative (post-op) VA.

23 cases (White, 39.1%; Asian, 34.5%; not specified, 21.7%) were recorded. Mean age was 42.4 years (SD=17.8). Males (86.9%) had a higher rate than females, which aligns with the literature. Mean follow-up duration was 7.85 months. At presentation, mean BCVA was 2.24 logMAR and mean OTS graded score was 2.22. For graded OTS 1 cases (n=6), the predicted % of patients with NPL & PL VA were 73% and 28% while true % recorded in our patients were each 50%. Graded OTS 2 (n=7) predicted 28% of NPL & 26% PL/HM VA while both true values were reported as 14.3%. Graded OTS 3 (n=9) predicted 44% VA >20/40 & 28% VA 20/200 to 20/50 while true were 22.2% and 44.4%, respectively. Graded OTS 4 (n=1) predicted 74% VA >20/40 while the true recording was 100%. 

For all graded OTS categories, the predicted % of patients with the 2 likeliest VA categories differed markedly from the actual true post-operative VA % in our cohort. Notably, graded OTS 2 and 4 had over-estimated their 2 likeliest VA % distribution, which in fact indicates a better post-operative VA improvement than predicted. Our results show that in the UK population, particularly in the Midlands, OTS proves in general to be limited in its ability to predict long-term post-operative VA. Further studies are needed to verify with a larger sample size, especially with graded OTS 4 patients, in order to investigate better predictive alternatives for such patients.