Culture positivity is a strong prognostic indicator of in-hospital mortality for COVID-19 patients with sepsis

ABSTRACT

Introduction: Despite numerous clinical scoring systems, outcome modeling for COVID-19 patients with sepsis remains poor. To address this deficit, we assessed the impact of culture positivity on in-hospital mortality for COVID-19 patients with sepsis. We report that culture positive sepsis derived from blood, bronchoalveolar lavage (BAL), or cerebrospinal fluid (CSF) is a stronger prognostic indicator of in-hospital mortality for COVID-19 patients than the Sequential Organ Failure Score (SOFA). These results support inclusion of culture status in future clinical scoring systems. Methods: The cohort was defined by inpatients from 03/20 to 09/21 with a COVID-19 + test (PCR, rapid-antigen, antibody) and septic event (n = 792) as defined by Sepsis-3 guidelines [1]. Each patient's worst SOFA score was computed during their suspected infection window (defined as 24 h prior to and 48 h after the first antibiotic administration or body-fluid culture taken). Study groups included culture positive (n = 478) and culture negative (n = 314) sepsis patients. Charlson comorbidity scores for each patient were calculated prior admission. Positive predictors of in-hospital mortality were assessed with multivariate logistic regression and evaluated for statistical significance using the CAR-ANOVA Type-III test with Bonferroni method. Results: Multivariate logistic regression analysis showed that culture positivity had the greatest adjusted odds ratio (OR 3.19, 95% CI 2.09- 4.98, p < 0.001, corr. p < 0.001), compared to worst SOFA score (OR 1.91, 95% CI 1.61-2.27, p < 0.001, corr. p < 0.001), patient age (OR 1.46, 95% CI 1.20-1.80, p < 0.001, corr. p < 0.001), male sex (OR 1.67, 95% CI 1.15-2.42, p < 0.006, corr. p = NS) and comorbidity score (OR 1.02, 95% CI 0.84-1.22, p = NS, corr. p = NS) (Fig. 1). Conclusions: Culture positivity is a strong prognostic indicator of inhospital mortality for COVID-19 sepsis patients and warrants investigation as a candidate variable for future clinical outcome algorithms.