Despite cerebrospinal fluid (CSF) analysis being the definitive method to diagnose meningitis, obtaining CSF is often challenging in low-resource settings. The authors evaluated whether cranial ultrasound (CUS) could support more rapid identification of neonatal sepsis-associated meningitis (NSM) at the Neonatal Centre of Excellence, University Teaching Hospitals, Lusaka, Zambia. The study tested whether integrating CUS with clinical assessment improves early detection where lumbar puncture or ventricular tap uptake is low.
This was a prospective cohort study enrolling 273 term mother–neonate pairs. Neonates were initially categorized by the admitting clinical team into those with sepsis-meningitis (exposure) and those with sepsis only (non-exposure). The research team later assigned participants into diagnostic subgroups based on clinical diagnosis and CUS findings. Data collected included socio-demographic variables, clinical characteristics, blood work and CSF analysis reports when available, neurologic findings, and mortality outcomes. Both descriptive and inferential statistics were used to compare groups.
Participants were allocated to four groups according to CUS and clinical diagnosis:
These subgroup definitions were used to compare clinical presentation, imaging findings, laboratory results, and outcomes.
The cohort had a median chronological age at admission of 13 days (IQR 7–21). The overall median hospital stay was 11 days (range 2–45; IQR 6–16). Meningitis was suspected in 11% of neonates admitted with clinical sepsis. Late-onset sepsis accounted for 31.9% of NSM cases. Chronological age at admission (p = 0.002) and duration of hospitalization (p = 0.042) differed significantly across the four diagnostic categories. Neonates with meningitis tended to present at older ages and had longer hospital stays than those with sepsis only. Clinical diagnosis of meningitis was more likely in older neonates and was associated with a shorter hospitalization than CUS-diagnosed meningitis.
The most frequently reported abnormal CUS findings were prominent sulci, hyperechoic brain lesions, ventriculitis, and lateral ventriculomegaly. These features comprised the CUS evidence used to classify neonates as CUS-positive for meningitis. The study reports that CUS identified structural and meningeal irregularities but emphasizes that CUS does not replace CSF analysis.
Uptake for lumbar puncture (LP) or ventricular tap (VT) was low, performed in only 12 neonates (4.4%), illustrating practical barriers to CSF-based diagnosis in this setting. Of the 12 CSF reports, nine (75%) were positive for meningitis. Notably, the majority (5 of 9; 55.6%) of positive CSF results occurred in neonates who had been categorized clinically as sepsis only. The low overall LP/VT uptake limits direct comparison between diagnostic methods across the full cohort.
Using CSF analysis as the reference where available, the study reports diagnostic performance metrics for clinical diagnosis and for CUS:
Percent agreement between clinical and CUS diagnosis for NSM was 71.1%. A clinical diagnosis of NSM was associated with greater odds of abnormal CUS at admission (odds ratio 3.94; 95% CI 1.80–8.62; p < 0.001).
Overall mortality in the cohort was 2.2%. Mortality was highest (16.7%) among neonates diagnosed with NSM by both clinical assessment and CUS. The study reports a prevalence of near-miss cases of NSM of at least 27.6%. Neonates with meningitis had longer hospitalizations than those with sepsis only.
The authors report that CUS demonstrated moderate diagnostic effectiveness compared with CSF analysis but performed better than clinical assessment alone on several metrics (higher sensitivity, specificity, and PPV). However, because CSF sampling was infrequent, the comparison is limited by small numbers of CSF-confirmed cases. The presence of positive CSF results among neonates classified clinically as sepsis only highlights the risk of missed meningitis when CSF is not obtained.
The findings support the potential clinical value of combining bedside clinical assessment with CUS to increase detection of NSM in settings where CSF sampling is limited by clinical, logistical, or resource constraints. Common CUS abnormalities described (prominent sulci, hyperechoic lesions, ventriculitis, ventriculomegaly) provide observable targets for sonographic evaluation in suspected cases.
The study concludes that while CSF analysis remains irreplaceable for definitive diagnosis of neonatal meningitis, integrating clinical assessment with cranial ultrasound (CUS) can enhance prompt identification and guide management in resource-limited contexts. Systematic adoption of CUS for neonates with signs of sepsis or possible meningitis, alongside efforts to improve uptake of CSF analysis, could support earlier targeted therapy and may reduce mortality and long-term neurologic sequelae. The authors note that limitations linked to low CSF sampling and the preprint status of the report warrant cautious interpretation.
This report is a preprint and has not undergone peer review. The study received ethical approval from the University of Zambia Biomedical Research Ethics Committee and the National Health Research Authority (reference UNZA-5664/2024). Data are available on reasonable request to the authors. Funders declared include the Gates Foundation. The authors stated no competing interests.