Objectives: Neurocritical illnesses occur disproportionately in countries with limited health care capacity. Identifying risk factors of reversible neurodeterioration may improve care in resource-limited settings. We investigated whether early neurodeterioration (END) predicted inpatient mortality at a resource-limited Zambian hospital.
Methods: We conducted a prospective observational cohort study from January to August 2023 of neurology inpatients at Zambia's national referral hospital. All inpatients older than 18 years were eligible for inclusion and monitored from admission to discharge in a clinical registry. END was defined as a drop in the Glasgow Coma Scale (GCS) within 3 days of admission. We used linear regression to compare mortality between participants with and without END.
Results: We enrolled 500 participants. Two hundred ninety-five participants had GCS recorded more than once within 3 days of admission. END occurred in 53 participants (18%) and predicted increased mortality (72% vs 11%, odds ratio [OR] 18.98, 95% CI 7.74-46.51, p < 0.001), even after a single-point GCS drop (70%, OR 13.08, 95% CI 3.2-53.47, p < 0.001). Seizures were more frequent in participants with END (27% vs 12.5%, χ2 = 9.9, p = 0.002). The leading mortality drivers were aspiration pneumonia (34%) and sepsis (31%).
Discussion: END was common, associated with unfavorable outcomes, and frequently driven by preventable and/or treatable factors. Protocolized interventions for reversible complications may improve neurologic outcomes in resource-limited settings.