Postintravitreal Injection Endophthalmitis: An Infection Control Investigation and Case–Control Analysis of Risk Factors

Abstract

Background: Intravitreal injection of vascular endothelial growth factor inhibitors with or without steroids is a well-established, effective therapy for several ocular disorders. The expected rate of complications from these injections is low, with meta-analyses reporting 5–6 occurrences of infectious endophthalmitis per 10,000 injections. Through October 2019, our health system observed 8 cases of endophthalmitis among 7,693 injections (10.4 per 10,000 injections), compared to 1 case in 2018. This unusually high rate prompted an infection control investigation and a case control study to examine risk factors for the development of postintravitreal injection endophthalmitis. Methods: Infection control providers performed direct observation of several ophthalmologists performing intravitreal injections on 3 separate occasions to determine points of intervention to prevent infection. To define risk factors for postintravitreal injection endophthalmitis, we conducted a retrospective case-control study of the 8 affected patients. Four control patients were selected per case, matched by clinic location, drug injected, and date of injection (total subjects, N = 40). We extracted patient-level risk factors from medical records; documentation was not sufficient to compare procedure-level factors. We conducted unadjusted univariate Poisson regression and Mantel–Cox method rate ratios to identify significant risk predictors of endophthalmitis. Results: Direct observation yielded variable practice in use of masks, gloves, sterile lid speculum, and the duration of povidone-iodine contact on the ocular surface prior to injection. The location of alcohol hand gel relative to the procedure field was suboptimal. Due to patient volume, there were significant delays between procedure and patient prep and injection time. The mean age was 76 years among cases and 74.1 years among controls; 35% of patients were men. Age-related macular degeneration was the most common indication for injection (55%). Only 10% of injections were bilateral. Although not statistically significant, patients with coronary artery disease had a higher rate of infection than those without coronary artery disease (165.3 vs 16.3 per 10,000 person years; IRR = 3.0; 95% CI, 0.60–14.8; P = .18); current smokers were also at higher risk (86.9 per 10,000; IRR, 3.2; 95% CI, 0.33–30.4; P = .32). Conclusions: Coronary artery disease and smoking were risk factors for the development of postintravitreal injection endophthalmitis in a 2019 cluster of cases in our organization. We are continuing to work with our ophthalmologists to optimize infection prevention in the injection environment, including strict use of gloves, appropriate use of povidone-iodine, and routinely wearing a mask and encouraging a no-talking policy during injections.

Funding: None

Disclosures: None