Abstract
Purpose :
To determine if universal masking instituted after Coronavirus Disease 2019 (COVID-19) altered rate and outcomes of endophthalmitis after intravitreal injection (IVI) at a single tertiary center.
Methods :
Retrospective chart review of endophthalmitis cases within 4 weeks of IVI (IVIE) occurring at or referred to University of Michigan from 7/1/18 to 11/15/22. Patients (n=37) were treated with either IVI antibiotics with vitreous and/or anterior chamber tap (T/I) or pars plana vitrectomy (PPV). Records were excluded for poor documentation (n=7) or course consistent with non-infectious inflammation (n=6). Rate of IVIE, referral patterns, microbial spectra, management, and visual acuity (LogMAR VA) were investigated. Data is reported as mean±SD with unpaired two-tailed t-test. Proportions were compared with Fisher exact test. Comparisons are pre- v post-masking unless otherwise indicated.
Results :
17 cases presented with varying masking protocols compared to 20 cases after universal masking began 03/2020. Of the 66,449 injections performed by our facility, the internal IVIE rate measured 0.025% versus 0.024% post-masking. No significant difference in demographics, baseline VA, external referral rate, and time to initial treatment was noted pre- or post-masking. Cases had similar presenting VA (2.18±0.37 v 2.35±0.40, p=0.23) and rate of light perception (LP) vision (18.8 v 31.6%, p=0.46) post-masking. Initial treatment with T/I remained consistent (100 v 95.0%). Positive-culture rate was not statistically different (60.0 v 73.7%, p=0.47) with similar rates of antibiotic-drug resistance (ADR) (62.5 v 64.3%). Frequency of streptococcal isolates (45.5 v 35.7%) and coagulase-negative staphylococci (CoNS) (45.5 v 64.3%) were similar post-masking. Isolated CoNS organisms were more likely ADR (60.0 v 77.8%) compared to oral flora (40 v 20%) regardless of masking protocol. Second treatment with T/I v PPV were similar pre-masking (23.5 v 23.5%), however secondary PPV (11.1 v 27.8%) was more common post-masking. For those with ≥30 days follow-up (627±424 v 356±244 days, p=0.04), final VA was similar (1.42±1.03 v 1.43±0.98, p=0.98) with no difference in assessed long-term complications.
Conclusions :
Universal masking had no effect on the rate of IVIE, presenting VA, or isolated microbial spectrum. Initial management with T/I did not differ between groups and resulted in similar visual outcomes.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.