287 - Impact of the COVID-19 Pandemic on Late-Onset Sepsis Among Extremely Preterm Infants

Poster Number: 287
Publication Number: 287.226

Sagori Mukhopadhyay, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; David Kaufman, University of Virginia School of Medicine, Charlottesville, VA, United States; Shampa Saha, RTI International, RTP, NC, United States; Dustin D. Flannery, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Karen M. Puopolo, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Eric Eichenwald, Children's Hospital of Philadelphia, Philadelphia, PA, United States; Pablo J. Sanchez, Nationwide Children's Hospital -OSU, Columbus, OH, United States; Rachel G. Greenberg, Duke Clinical Research Institute, Durham, NC, United States; Kristin Weimer, Duke University School of Medicine, Durham, NC, United States; Abbott Laptook, Women & Infants Hospital of Rhode Island, Brown University Warren Alpert Medical School, Providence, RI, United States; Barbara J. Stoll, McGovern Medical School at the University of Texas Health Science Center at Houston, Atlanta, GA, United States; Michael Cotten, Duke University School of Medicine, Durham, NC, United States

Background: In response to the COVID-19 pandemic, neonatal units implemented enhanced infection-prevention practices to mitigate SARS-CoV-2 transmission. Such strategies can also impact bacterial late-onset sepsis (LOS) transmission

Objective: To compare rate of LOS in extremely preterm infants in the NICHD Neonatal Research Network sites before and during the COVID-19 pandemic.

Design/Methods: Retrospective cohort study of infants with birth weight 400-1000 grams or gestational age 22-28 weeks who survived >12 hours and were hospitalized from 1/1/18 to 8/31/21. For all centers, we considered 4/1/20 as a common timepoint for the start of pandemic-associated infection-prevention practices. Pre-pandemic period was defined from 1/1/18-3/31/20 and the pandemic period from 4/1/20-8/31/21. LOS was defined as a bacterial or fungal pathogen isolated from blood or CSF culture obtained >72 hours of age. We calculated incidence rates (monthly LOS/1000 patient-days) and incidence proportion (LOS cases among all admissions). An interrupted time series analysis using generalized linear mixed model with negative binomial regression and center as a random effect was used to analyze incidence rates. A multivariable Poisson regression model adjusted for confounders was used to determine LOS incidence proportion.

Results: Fifteen centers and 6697 infants with 1170 LOS events and 642,842 patient-days were included. There was a significant downward trend in monthly LOS/1000 patient-days during the pandemic period (P=0.003), with no significant trend before, or change in level at 4/1/20 (Figure 1). Varying the timepoint of change by ±1 month did not alter the results. In the pre-post analysis, the incidence proportion of LOS was not different (18.5% vs 17.4%; P= 0.28). After adjusting for patient characteristics (Table 1), time period was not associated with risk of LOS (aRR 0.92, 95% CI 0.80-1.05) or mortality (aRR 1.14, 95% CI 0.99-1.32) (Table 2).Conclusion(s): We found no difference in LOS incidence proportion between the pre-pandemic and the pandemic period. However, there was a reduction in monthly LOS incidence rates in the pandemic period. Incidence proportion and incidence rates measure different aspects of LOS epidemiology where incidence rates account for length of stay. Similarly, time-series analysis may detect early trends not captured in pre-post analysis. Longer study durations are needed to understand if these trends persist and lead to reduced incidence proportions over time.
Figure 1: Monthly rates of LOS per 1000 patient-days across all centersFigure shows the actual (blue) and estimated (red) rates of monthly LOS/1000 patient-days during study period. Vertical line denotes 4/1/2020, the timepoint chosen to identify start of pandemic related infection prevention practice changes. Estimates from the time-series model adjusted for the centers as random effect.

Designed by Cadmium |Technical Support
© Copyright 2024 Cadmium. All Rights Reserved.