Endophthalmitis following pars plana vitrectomy: An 11-year retrospective cohort of 36,179 procedures in London, United Kingdom.

Abstract

Objective: Endophthalmitis is a rare but devastating complication of pars plana vitrectomy (PPV). We assessed the incidence and profile of post-PPV exogenous endophthalmitis and investigated causality and prognostic factors in a large ethnically diverse cohort over an 11-year period.

Design: A retrospective single-site cohort study.

Participants: Adult patients (aged 18 and over) undergoing PPV at Moorfields Eye Hospital.

Methods: PPV procedures between January 1st 2013 and January 1st 2024 were extracted from the electronic health record and cross-referenced with all endophthalmitis cases using clinical documentation, prescription records, and incident reports.

Main outcome measures: The incidence proportion was estimated and stratified by additional procedures during PPV. Odds ratios (OR) with 95% confidence intervals (CI) for the association between endophthalmitis incidence and sociodemographic and procedural factors were estimated using univariable and multivariable logistic regression models.

Results: There were 36,179 procedures from 26,533 patients included in the analysis. The overall incidence of post-PPV endophthalmitis was 0.05% (n=19), of which 63.2% (n=12) were culture-positive. Five cases occurred after 28 days, of which three had coexisting anterior segment pathology including corneal abscesses and loose sutures. Incidence figures varied from 0.05% in PPV with internal limiting membrane peel to 0.65% in those undergoing intraocular lens (IOL) exchange. Higher odds of endophthalmitis were seen with fluid tamponade (adjusted OR [aOR] 5.70, 95% CI: 1.80, 18.03, p = 0.003) and increasingly complex secondary IOL procedures - secondary IOL insertion alone (aOR 5.42, 95% CI: 1.23, 23.97, p=0.026); IOL removal (aOR 9.81, 95% CI: 3.10, 31.07, p=1.0 x 10^-4); and, IOL exchange (aOR 13.64, 95% CI: 4.29, 43.43, p=9.7 x 10^-6). When adjusting for tamponade, IOL removal (aOR: 4.64, 95% CI: 1.14, 18.86, p = 0.032) and IOL exchange (aOR 6.24, 95% CI: 1.43, 27.27, p=0.015) remained significantly associated with endophthalmitis. Endophthalmitis associated with secondary IOL procedures were all culture-positive (n=3 Staphylococcus spp, n=3 Streptococcus spp).

Conclusions: Although post-PPV exogenous endophthalmitis is rare, individuals undergoing PPV with IOL removal and exchange have a considerably increased odds of developing endophthalmitis. Delayed cases of endophthalmitis frequently have coexisting anterior segment pathology.