The predictive value of inflammatory biomarkers on 30-day mortality in the neurological critical care setting

Introduction Prognostication is difficult for patients admitted to a neurocritical care unit (NCCU). Can serum biomarkers obtained as part of routine admission lab work help predict outcomes among patients expected to stay in the NCCU for \u3e48 hours? Methods In this prospective cohort study, the following biomarkers were measured at admission: C-reactive protein (CRP), arterial lactate, neuron specific enolase (NSE), lactate dehydrogenase (LDH), albumin, and brain natriuretic peptide (BNP). We collected information about demographics, comorbidities, hospital procedures and complications and 30-day mortality. We compared these serological biomarkers in patients who were alive versus those who had died at 30 days. Results A total of 112 patients were enrolled over 4 months from June to September 2016, 11 of which whom (9.8%) died within 30 days of admission. There were no statistically significant differences in age or gender between the two groups. The 30-day mortality group had a higher mean Charlson Comorbidity Index (CCI) (3.0 vs 1.6, p=0.027) as well as mean NSE (39.1 vs 13.9 ug/L, p=0.008) and BNP levels (590.2 vs 177.3 pg/mL, p=0.003). Mean CRP, lactate, and LDH were also higher in the 30-day mortality group (79.5 vs 51.8 mg/L, 2.3 vs 1.9 mmol/L, and 307.2 vs 274.2 U/L) while mean albumin was lower (3.0 vs 3.3 g/dL), although these differences were not statistically significant (p\u3c0.35). Conclusions CCI and serological biomarkers may have utility in predicting 30-day mortality among patients admitted to the NCCU. As we continue enrollment, we plan to develop a predictive model for 30-day mortality on admission for patients admitted to the NCCU using serological biomarkers, CCI and admission characteristics