Neonatal bacterial meningitis (NBM), defined as bacterial inflammation of the meninges and subarachnoid space occurring within 28 days of birth, remains a significant cause of neonatal mortality and long-term neurological impairment [1]. This condition accounts for approximately 190,000 deaths annually worldwide [2]. Among survivors, 25–50 % develop persistent neurological sequelae [3], including cerebral palsy (9 %), epilepsy (5 %), sensorineural hearing loss (3 %), and learning disabilities (8 %) by 5 years of age [4].

Acute neurological complications (ANCs) of NBM encompass a spectrum of conditions including subdural effusion/empyema, ventriculitis, hydrocephalus, brain abscess, cerebral infarction, encephalitis, cerebral venous thrombosis, and arterial stroke [5]. Early identification of neonates at high risk for ANCs is critical for implementing targeted interventions and potentially mitigating permanent neurological deficits. Nevertheless, studies investigating risk factors for ANCs in neonates with bacterial meningitis remain scarce.

Several studies have explored predictors of specific complications or mortality in infant meningitis. A retrospective study in China identified multidrug-resistant infection and C-reactive protein levels > 50 mg/L as risk factors for brain abscess in neonatal meningitis [6]. A prospective study in Angola found that elevated heart rate and the presence of seizures were independent predictors of mortality in infants under 90 days of age with bacterial meningitis [7]. Similarly, a retrospective study in Taiwan reported that seizures at onset, early-onset sepsis, and neurological complications requiring surgical intervention were independent predictors of moderate-to-severe disability or death in infants under 90 days with Gram-negative bacterial meningitis [8]. While these studies provide valuable insights, they have largely focused on either specific complications (e.g., brain abscess) or long-term/mortality outcomes, rather than comprehensively examining risk factors for the full spectrum of ANCs in the neonatal population.

Therefore, this study aimed to identify risk factors for the development of ANCs in neonatal bacterial meningitis.

This study demonstrates a high incidence of ANCs (29.41 %) in neonatal bacterial meningitis and identifies seizures and GBS infection as independent risk factors for their development. A predictive model combining these factors effectively identified NBM patients at risk for acute neurological complications with good discriminatory capacity (AUC = 0.814). Dexamethasone administration did not reduce the incidence of ANCs in our cohort. These findings provide a foundation for early risk stratification and underscore the need for heightened clinical vigilance in neonates presenting with seizures or GBS infection.

The incidence of specific acute neurological complications observed in our cohort—subdural effusion/empyema (11.76 %), ventriculitis (4.41 %), hydrocephalus (10.29 %), brain abscess (5.88 %), and encephalomalacia (17.65 %) — generally aligns with ranges reported in previous literature. The subdural effusion incidence approximates the 11 % reported for neonatal bacterial meningitis and the 7–13 % range for Gram-negative meningitis [13,14]. While ventriculitis has been reported in 5–10 % of neonatal meningitis cases [15], hydrocephalus in approximately 25 % [13,15], and brain abscess in 10 % of neonatal meningitis patients[16] or 10–20 % in Gram-negative meningitis [15,16], our observed rates for these specific complications were somewhat lower. These variations may reflect differences in inclusion criteria, age distribution, pathogen profiles, or treatment protocols across studies.

Our finding of a significantly higher incidence of seizures in the ANC group is consistent with extensive prior research linking seizures to adverse outcomes in NBM. Multiple studies have identified seizures as a predictor of mortality or neurodevelopmental disability in neonatal bacterial meningitis[17,18]. Specifically, seizures have been demonstrated to be an independent risk factor for adverse outcomes in neonates with Gram-negative bacterial meningitis [8], and studies across diverse populations confirm their independent association with poor outcomes in bacterial meningitis overall [19]. While seizures were identified as an independent predictor of ANCs in our cohort, it is important to consider the nature of this association. This relationship may reflect, at least in part, seizures serving as an early clinical manifestation of evolving neurological injury rather than as a truly independent causal factor. Given the observational design of this study, definitive conclusions regarding causality are precluded. Accordingly, seizures should be interpreted primarily as an early indicator of disease severity that may facilitate risk stratification, rather than as a modifiable target for intervention.

This study provides evidence that GBS infection is significantly associated with an increased risk of acute neurological complications in NBM. This finding aligns with the well-documented severe neurological consequences of GBS meningitis. Prospective studies following neonatal GBS meningitis survivors have documented high rates of neurological dysfunction (49 % at 5 years) [20], neurodevelopmental impairment (25 % moderate/mild, 19 % severe) [21], and disabilities persisting beyond 18 months (32 %, with nearly one-fifth being moderate or severe) [22]. Our results suggest that this propensity for neurological damage manifests acutely as a higher risk of the complications studied.

The authors found no evidence supporting the use of dexamethasone for ANC prevention in NBM. The role of adjunctive dexamethasone in neonates remains controversial and inadequately studied compared to older populations. While randomized controlled trials and meta-analyses demonstrate benefits (reduced hearing loss and neurological sequelae) in pediatric and adult bacterial meningitis [23–25], and cohort studies suggest improved outcomes in certain settings [26–29], concerns exist regarding potential adverse effects, such as reduced survival in neuro-listeriosis [30]. Neonatal randomized controlled trials remain limited and inconclusive: one found no difference in outcomes, including complications [31], while another reported reduced mortality and inflammation [32]. A meta-analysis indicated reduced hearing loss but no benefit for mortality or severe neurological outcomes [33]. Our data contribute specifically to the ANC context, indicating no protective effect of dexamethasone administration.

ANCs profoundly impact the long-term prognosis of NBM. The absence of simple predictive models in the literature underscores the utility of our tool, which integrates readily assessable clinical (seizures) and microbiological (GBS) features. Early identification of high-risk neonates could prompt intensified antibiotic therapy, vigilant neuroimaging surveillance, and aggressive complication management, potentially attenuating long-term neurological deficits. Nevertheless, while the model exhibits reasonable discriminatory ability, its relatively low sensitivity (55.0 %) indicates that it would fail to identify nearly half of the neonates who ultimately develop ANCs, limiting its utility as a standalone screening tool in clinical practice. The modest sample size and limited number of predictors available for model development likely contributed to this limitation and may have introduced optimism into the performance estimates. Given its high specificity (93.8 %), the model may be best employed as an adjunct to clinical judgment — helping to identify a subset of patients at particularly high risk — while risk assessment in screen-negative individuals should continue to rely on comprehensive clinical evaluation. Accordingly, external validation in larger, independent cohorts is essential before this model can be considered for clinical application.

This study has several limitations. First, its retrospective design introduces the potential for selection bias. Additionally, reliable data on antibiotic administration prior to lumbar puncture were unavailable, precluding assessment of this potential confounder, which may influence CSF culture positivity rates. Second, the modest sample size limits statistical power for subgroup analyses and detection of smaller effect sizes, and increases the risk of overfitting in our multivariate model. Third, external validation of the predictive model in independent cohorts is pending. Future prospective multicenter studies with larger cohorts are needed to validate and refine this model and explore additional risk factors.

This study identified an incidence of acute neurological complications in neonatal bacterial meningitis of 29.41 %. Seizures and GBS infection were identified as independent risk factors for these complications. Dexamethasone administration did not reduce the incidence of ANCs in this cohort. These findings may help clinicians identify high-risk neonates who require closer monitoring and more aggressive management strategies. However, the results should be interpreted with caution given the study's limitations, including its retrospective design, modest sample size, and single-center setting. The predominant role of GBS observed in this cohort may reflect regional pathogen prevalence and local clinical practices, potentially limiting generalizability to populations with different epidemiological profiles.

This study was conducted in accordance with the Declaration of Helsinki concerning Ethical Principles for Medical Research Involving Human Subjects. To protect participant privacy, names were not used and data confidentiality was maintained throughout the study. Ethical approval was obtained from the Institutional Review Board of Children's Hospital, Zhejiang University School of Medicine (Approval Number: 2025-IRB-0225-p - 01). The requirement for informed consent was waived due to the retrospective nature of the study.

This work was supported by the Zhejiang Province Health Project (Grant No 2021KY764). The funding organization had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

FZ: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Visualization, Writing - original draft, Writing - review & editing; ZY: Data curation, Formal analysis, Investigation, Methodology, Project administration, Validation, Writing - review; JP: Data curation, Formal analysis, Methodology, Resources, Supervision, Validation, Visualization, Writing - review; TY: Conceptualization, Methodology, Resources, Supervision, Validation, Writing - review.

Data are available on reasonable request (email: yuantianming @zju.edu.cn).