Abstract
Purpose :
The use of vitrectomy in endogenous Klebsiella pneumoniae endophthalmitis (EKPE) is controversial due to the rarity of EKPE and the lack of trials. We aimed to compare outcomes in EKPE patients with and without vitrectomy by combining existing individual data using meta-analysis.
Methods :
PubMed, Embase, and Web of Science were searched from inception until December 31, 2021. The eligibility criteria were: (1) enrolled patients with EKPE, (2) included either vitrectomy, intravitreal injection (IVI) of antibiotics, or both as interventions, and (3) reported both initial and final visual acuities (VAs). VAs were converted to the logarithm of the minimum angle of resolution. Poor VA was defined as hand motion or worse. Main outcomes included the mean differences (MDs) of the VA change and the odds ratios (ORs) of poor final VA and anatomical failure between eyes that underwent vitrectomy (the vitrectomy group) and those that received IVI antibiotics alone (the non-vitrectomy group). During meta-analysis, studies were classified into within-study comparisons (WSCs) and single-arm studies (SASs); the former underwent analysis with the generalized linear mixed model, and the latter with generalized estimating equations. A secondary meta-analysis integrated effect estimates through a fixed-effect model. Quality of the evidence was accessed with the Grading of Recommendations Assessment, Development and Evaluation.
Results :
From the 68 retrospective studies included, 324 patients with 383 affected eyes were reported. Overall, the mean initial VA was 2.121 ± 0.750 and the mean final VA was 2.223 ± 0.988. The mean VA change was significantly better in the vitrectomy group (MD = -0.27; 95% CI = -0.40 to -0.14; P < 0.01). In preset subgroup analysis based on initial VA, the result stayed robust in eyes with poor initial VA (MD = -0.35; 95% CI = -0.49 to -0.21; P < 0.01). There were no significant differences in risk of poor final VA (OR = 0.84; 95% CI = 0.40 to 1.75; P = 0.64) and risk of anatomical failure (OR = 0.42; 95% CI = 0.15 to 1.13; P = 0.09) between the two groups. The quality of the evidence was very low mainly due to the lack of randomized controlled trials.
Conclusions :
Vitrectomy resulted in greater VA improvement compared to IVI antibiotics alone in eyes with EKPE, especially in those with poor initial VA.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.