Early prognostic value of repeat CSF parameter measurements in neonatal bacterial meningitis following initial antimicrobial therapy.

Background: Bacterial meningitis (BM) in neonates leads to high mortality rates and long-term complications. Current guidelines recommend repeat cerebrospinal fluid (CSF) examinations to assess treatment efficacy, but their clinical value remains controversial. This study aims to evaluate the clinical utility and prognostic value of repeat CSF parameters within 3-5 days following initial antimicrobial therapy in neonates with BM.

Methods: A retrospective analysis of electronic medical records from January 2017 to December 2022 yielded 241 neonates of diagnosed with neonatal BM upon discharge in our neonatal intensive care unit. Of these, 85 met the stringent inclusion criteria for enrollment. Data on clinical characteristics, laboratory findings, cranial imaging results and other relevant information were collected. The criteria for an adequate response to the initial antimicrobial therapy, established for this study, included a 48-hour afebrile period, negative meningeal irritation signs, a C-reactive protein (CRP) level of ≤ 20 mg/L, and a neutrophil count of ≤ 10,000/mm³. Short-term adverse outcomes were defined as the occurrence of complications during hospitalization or sequelae at discharge.

Results: In the analysis of 54 neonates demonstrating an adequate response to initial antimicrobial therapy, a comparison between 29 neonates with repeated CSF parameter measurements and 25 neonates without such measurements revealed no significant differences in antibiotic duration (p = 0.728), complications during hospitalization (p = 0.517), or sequelae at discharge (p = 1.000). A multiple logistic regression analysis of the group with repeated CSF measurements identified elevated CSF protein levels as the only potential risk factor for complications during hospitalization (p = 0.045; OR = 15.046; 95% CI: 1.062-213.089). However, none of the other repeated CSF measurements, such as total leukocyte count in CSF (CSF-TLC), the absolute neutrophil count to total leukocyte count ratio in CSF [CSF-Neutrophil (%)], and CSF glucose, were significant predictors of complications during hospitalizations (P > 0.05 for all). Furthermore, except for CSF-Neutrophil (%), the other repeated CSF measurements were significantly positively correlated with their initial CSF parameters (P < 0.05 for all). In the full cohort of 85 neonates, significant differences were observed between the 54 neonates with an adequate response and the 31 neonates with an inadequate response to therapy. Significant variables included admission weight (p = 0.047), peak blood CRP levels before antibiotic use (p < 0.001), initial CSF protein levels (p = 0.007) and blood culture results (p = 0.037). Both the risk of complications during hospitalization and sequelae at discharge were also significantly different (p < 0.001). In the cohort of 85 neonates, there were significant differences between the 52 neonates without short-term adverse outcomes and the 31 neonates with short-term adverse outcomes. The variables showing significant differences included the presence of neurological symptoms (p = 0.021), peak blood CRP levels before starting antibiotics (p = 0.022), initial CSF glucose levels (p = 0.023), initial CSF protein levels (p = 0.001), and an inadequate response to initial antibiotic therapy (p < 0.001).

Conclusions: For neonates with BM who have shown an adequate response to initial antimicrobial therapy, conducting repeated CSF examinations within the first 3-5 days of treatment appears to have limited value, as it does not provide additional information that could help physicians shorten the duration of antibiotic therapy or reduce the risk of short-term adverse outcomes, such as risk of complications during hospitalizations or sequelae at discharge. In contrast, for the initial assessment of short-term prognosis in neonates newly diagnosed with bacterial meningitis, the clinical global impression (CGI) at admission, particularly factors such as neonatal weight, neurological symptoms, response to initial antimicrobial therapy, and early laboratory data-including peak blood CRP levels before antibiotic use, initial CSF glucose levels, initial CSF protein levels, and blood culture. results [In Table 4, a comparison between neonates with an adequate response to therapy and those without revealed a significant difference in blood culture results (p = 0.025)] -have shown stronger predictive capabilities and deserve greater attention.