Mapping prognostic factors for globe survival in panophthalmitis using logistic regression and Cox proportional hazard models: A retrospective study

Purpose

The aim of this study was to report the outcomes of panophthalmitis and to identify factors significantly affecting globe survival in the disease.

Methods

This was a retrospective study on patients with panophthalmitis in a tertiary hospital between January 1, 2017, and December 31, 2019. The demographics, treatment details, culture results, and final outcomes were recorded. Logistic regression and Cox proportional hazards (CPH) were calculated to identify variables associated with globe loss. A P < 0.05 was considered significant.

Results

Eighty-five eyes of 85 patients (31 culture positive) were eligible for review. The mean age of the participants was 55.21 ± 20.17 years with a male-to-female ratio of 2.04:1. Corneal ulcer (38.82%; n = 33) and open globe injuries (OGIs) (38.82%; n = 33) were the most common etiologies. Pseudomonas aeruginosa (n = 10; 11.76%) was the most common isolate. The mean duration of hospital stay was 7.58 ± 2.32 days. Overall, 44 (51.76%) globes could be salvaged. The need for evisceration (P = 0.901) and hospital stays (P = 0.095) were similar for culture-positive and -negative cohorts. The unadjusted logistic regression and CPH models showed that culture sterility did not affect globe survival [OR = 1.210 (0.501-2.950), P = 0.668; HR = 1.176 (0.617-2.243), P = 0.623]. The adjusted logistic regression and the CPH models showed that corneal ulcers [OR = 10.900 (2.460-48.200), P = 0.002; HR = 5.393 (1.603-18.140), P = 0.006] and OGI [OR = 7.360 (1.650-32.700), P = 0.009; HR = 4.548 (1.321-15.660), P = 0.016] were significantly associated with globe loss.

Conclusion

Corneal ulcer or OGI as the primary etiology is detrimental to globe survival in panophthalmitis.