THE LABORATORY PARAMETERS OF CEREBROSPINAL FLUID IN ASSESSING THE LIKELIHOOD OF BRAIN EDEMA-SWELLING AND SYSTEMIC COMPLICATIONS IN PATIENTS WITH BACTERIAL MENINGITIS

Abstract

Bacterial meningitis remains to be one of the leading causes of death due to infectious diseases, resulting in more than 318,000 fatalities annually. Emergence of a poor outcome in patients with bacterial meningitis is frequently associated with the development of several complications, including brain edema and systemic complications such as systemic inflammatory response syndrome (SIRS), infectious and toxic shock, sepsis, and multiple organ failure. The identification of poor prognostic markers in individuals with bacterial meningitis would allow for timely adjustment of treatment regimens, thereby reducing a risk of a unfavorable outcome. The current study analyzes a predictive importance of cerebrospinal fluid (CSF) laboratory parameters in assessing a risk of brain edema and systemic complications, such as systemic inflammatory response syndrome (SIRS), infectious and toxic shock, sepsis, and multiple organ failure, in adult immunocompetent patients with bacterial meningitis during the early stages of the disease (Ме = 2 days after the onset).

Medical records retrieved from 110 adult patients with laboratory-confirmed bacterial meningitis A were retrospectively analyzed. The study assessed CSF laboratory data obtained upon admission to the hospital, which corresponded to the second day of disease onset: CSF cytosis level (cells/mL) along with cell profiling, total protein concentration (g/L), and glucose level (mmol/L). During a biochemical CSF analysis, the concentration of lactate, chlorides, and D-dimer was determined. Statistical data analysis was performed using clustering, self-organizing Kohonen map construction, logistic regression modeling, and "Decision Tree" techniques in Loginom 7.1.5, Deductor Studio Academic, and Google Colab platforms. It was found that CSF laboratory parameters with the highest prognostic significance for predicting early-stage brain edema-swelling (BES) were presented as an increase in protein and lactate concentration exceeding 3.44 g/L and 12.495 mmol/L, respectively. These two indicators collectively provide sensitivity and specificity levels exceeding 93% and 86%, respectively, for predicting BES. When assessing the risk of systemic complications, an increase in lactate level > 12.425 nmol/L was the most significant predictor in cerebrospinal fluid analysis. The sensitivity of this indicator for predicting systemic complications varies from 73.9% to 97.14%, while the specificity –  from 82.67% to 87.2%, depending on the machine learning model used. Further research is required to fully understand these results.