Epidemiology, microbiology and antibiotic treatment of bacterial and fungal meningitis among very preterm infants in China: a cross-sectional study

Objective Neonatal meningitis significantly contributes to neonatal morbidity and mortality, yet large-scale epidemiological data in developing countries, particularly among very preterm infants (VPIs), remain sparse. This study aimed to describe the epidemiology of meningitis among VPIs in China. Design Cross-sectional study using the Chinese Neonatal Network database from 2019 to 2021. Setting 79 tertiary neonatal intensive care units in China. Patients Infants with gestational age <32 weeks or birth weight <1500 g. Main outcome measures Incidence, pathogen distribution, antimicrobial use and outcomes of bacterial and fungal meningitis. Results Of 31 915 VPIs admitted, 122 (0.38%) infants were diagnosed with culture-confirmed meningitis, with 14 (11.5%) being early-onset (≤6 days of age) and 108 (88.5%) being late-onset (>6 days of age). The overall in-hospital mortality was 18.0% (22/122). A total of 127 pathogens were identified, among which 63.8% (81/127) were Gram-negative bacteria, 24.4% (31/127) were Gram-positive bacteria and 11.8% (15/127) were fungi. In terms of empirical therapy (on the day of the first lumbar puncture), the most commonly used antibiotic was meropenem (54.9%, 67/122). For definitive therapy (on the sixth day following the first lumbar puncture, 86 cases with available antibiotic data), meropenem (60.3%, 35/58) and vancomycin (57.1%, 16/28) were the most used antibiotics for Gram-negative and Gram-positive bacterial meningitis, respectively. 44% of infants with Gram-positive bacterial meningitis and 52% with Gram-negative bacterial meningitis received antibiotics for more than 3 weeks. Conclusion 0.38% of VPIs in Chinese neonatal intensive care units were diagnosed with meningitis, experiencing significant mortality and inappropriate antibiotic therapy. Gram-negative bacteria were the predominant pathogens, with fungi emerging as a significant cause. Data are available upon reasonable request. After the publication of the article, the corresponding author may offer de-identified data, but this requires the provision of scientific rationale and sound methods. Requests for data sharing will be processed according to the Chinese Neonatal Network’s data-sharing policy.