Early Antibiotic Use and Retinopathy of Prematurity: A Single-Center Retrospective Cohort Study.

Objective: Retinopathy of prematurity (ROP) has been linked to neonatal sepsis, with antibiotic use suggested as a connection. Given the role of antibiotics in gut dysbiosis and the gut-retina axis, we assessed whether exposure to different antibiotic classes is associated with the incidence of treatment-necessary ROP. Design: Retrospective cohort study. Subjects: Preterm infants born at the University of Chicago Medicine and screened for ROP between January 2012 and December 2023. Methods: Retrospective analysis was performed to compare systemic antibiotic exposure within the first 2 months of life between infants with type 1 ROP (ie, required treatment) and those that did not require treatment. Multivariable adjustment included birth weight (BW), gestational age (GA), bronchopulmonary dysplasia (BPD), intraventricular hemorrhage, necrotizing enterocolitis (NEC), and neonatal sepsis. To reduce potential confounding by indication, propensity score matching was performed. Main Outcome Measures: Type 1 ROP by systemic exposure to antibiotic classes. Results: Seven hundred twenty infants were included, with 50 treated for ROP. The rate of systemic antibiotic use was 97.3% in the control group and 100.0% in infants with type 1 ROP. Infants with type 1 ROP showed higher rates of BPD (P < 0.001) and bacterial infection (P = 0.007). No association was noted with NEC. In multivariable regression, the odds of treatment were significantly greater for infants receiving cephalosporins, carbapenems, and monobactams (odds ratio [OR], 4.25; 95% confidence interval [CI], 1.74-10.4; P = 0.002), and an exploratory analysis suggested a dose-response per day prescribed (OR, 1.07; 95% CI, 1.03-1.11; P = 0.001). When restricted to BW <750 g or GA <27 weeks (N = 259, with 49 treated for ROP), these associations remained significant (OR, 3.87; 95% CI, 1.57-9.51; P = 0.003; per day prescribed: OR, 1.07; 95% CI, 1.03-1.11; P = 0.001). After propensity score matching for BW, GA, BPD, sepsis, and any bacterial infection, the average percent of infants who required treatment was 9.9 percentage points higher among those who received cephalosporins, carbapenems, and monobactams compared to those same infants if they had not received them (95% CI, 2.0-17.8 percentage points; P = 0.014). Conclusions: Early exposure to broad-spectrum antibiotic classes, particularly cephalosporins, carbapenems, and monobactams, may be associated with type 1 ROP. Financial Disclosures: The authors has no/the authors have no proprietary or commercial interest in any materials discussed in this article.