Late-Onset Sepsis among Extremely Preterm Infants of 24-28 Weeks Gestation: An International Comparison in 10 High-Income Countries

Abstract

Introduction: Despite advances in neonatal care, late-onset sepsis remains an important cause of preventable morbidity and mortality. Neonatal late-onset sepsis rates have decreased in some countries, while in others they have not. Our objective was to compare trends in late-onset sepsis rates in 9 population-based networks from 10 countries and to assess the associated mortality within 7 days of late-onset sepsis.

Methods: We performed a retrospective population-based cohort study. Infants born at 24-28 weeks' gestation between 2007 and 2019 were eligible for inclusion. Late-onset sepsis was defined as a positive blood or cerebrospinal fluid culture. Late-onset sepsis rates were calculated for 3 epochs (2007-11, 2012-15, and 2016-19). Adjusted risk ratios (aRRs) for late-onset sepsis were calculated for each network.

Results: Of a total of 82,850 infants, 16,914 (20.4%) had late-onset sepsis, with Japan having the lowest rate (7.1%) and Spain the highest (44.6%). Late-onset sepsis rates decreased in most networks and remained unchanged in a few. Israel, Sweden, and Finland showed the largest decrease in late-onset sepsis rates. The aRRs for late-onset sepsis showed wide variations between networks. The rate of mortality temporally related to late-onset sepsis was 10.9%. The adjusted mean length of stay for infants with late-onset sepsis was increased by 5-18 days compared to infants with no late-onset sepsis.

Conclusions: One in 5 neonates of 24-28 weeks' gestation develops late-onset sepsis. Wide variability in late-onset sepsis rates exists between networks with most networks exhibiting improvement. Late-onset sepsis was associated with increased mortality and length of stay.

Keywords: Extremely preterm infants; Late-onset sepsis; Mortality; Trends.

Fig. 1.

Pairwise comparison of late-onset sepsis between iNeo networks. The table compares pairs of countries. Countries on the left side of the table are compared to countries listed above by aRRs (adjusted for GA, gender, multiple births, and BW z-score). Dark gray indicates the corresponding country in the row has better results than the corresponding country in the column. Light gray indicates vice versa. No color indicates no statistically significant difference. ANZNN, Australian and New Zealand Neonatal Network; CNN, Canadian Neonatal Network; FinMBR, Finnish Medical Birth Register; INN, Israel Neonatal Network; NRNJ, Neonatal Research Network Japan; SEN1500, Spanish Neonatal Network; SNQ, Swedish Neonatal Quality Register; SNN, Swiss Neonatal Network; Tuscan NN, Tuscany Neonatal Network.

Fig. 2.

Late-onset sepsis rates over 3 epochs by iNeo networks. Mean late-onset sepsis rates for each network are shown for 3 time epochs with standard deviations. ANZNN, Australian and New Zealand Neonatal Network; CNN, Canadian Neonatal Network; FinMBR, Finnish Medical Birth Register; INN, Israel Neonatal Network; NRNJ, Neonatal Research Network Japan; SEN1500, Spanish Neonatal Network; SNQ, Swedish Neonatal Quality Register; SNN, Swiss Neonatal Network; Tuscan NN, Tuscany Neonatal Network.

Fig. 3.

Standardized ratios (SR) comparing the Late-onset sepsis of each network to all other networks combined. Vertical bars are the estimated 99% confidence intervals of the SR. The dotted curves represent the 99% confidence limits expected under the null hypothesis of similar outcome rates (SR = 1). ANZNN, Australia and New Zealand Neonatal Network; CNN, Canadian Neonatal Network; FinMBR, Finland Medical Birth Register; INN, Israel Neonatal Network; NRNJ, Neonatal Research Network of Japan; SEN1500, Spanish Neonatal Network; SNQ, Swedish Neonatal Quality register; SwissNeoNet, Swiss Neonatal Network; TuscanNN, Tuscany Neonatal Network.