Predictors of Mortality in 217 COVID-19 Patients in Northwest Ohio, United States: A Retrospective Study

Link to article at PubMed

J Med Virol. 2020 Dec 22. doi: 10.1002/jmv.26750. Online ahead of print.


INTRODUCTION: The Coronavirus disease 2019 (COVID-19) pandemic continues to cause significant morbidity and mortality worldwide. This study aims to identify specific lab markers, complications, and treatments that may be associated with increased mortality in COVID-19 patients.

METHODS: This study is retrospective in nature; it included 217 COVID-19 positive patients who were admitted to a ProMedica Health System hospital in Northwest Ohio, United States, between March 25 - June 16, 2020. We collected various laboratory values, complications, and treatment courses. T-test and Chi-squared analyses were used to predict mortality. COVID-19 test was confirmed via PCR.

RESULTS: Of 217 patients included in the study, the mean age of the population was 63.13 (SD 17.8), of which 194 (89.4%, mean age 61.7 years) survived while 23 (10.6%, mean age 74.6 years) died. Among them, 53% were females and 47% male. Laboratory values that were associated with mortality were low hemoglobin (p-value 0.0046), elevated INR (p-value 0.0005), low platelets (p-value 0.0246) and elevated procalcitonin (p-value 0.0472). Marginally significant laboratory values included elevated troponin (p-value 0.0661), and elevated creatinine (p-value 0.0741). Treatment with either antibiotic, antifungals, antivirals, blood transfusion, steroids, and intubation were all statistically significant for mortality. COVID-19 related complications with either ARDS, myocarditis, elevated INR, septic shock, or age greater than 63 were significant predictors of mortality.

CONCLUSION: Low hemoglobin, elevated INR, Low platelet, elevated procalcitonin, treated with either antibiotic, antifungal, antiviral, blood transfusion, steroids, and intubation are associated with high mortality related to COVID-19 infection. Healthcare professionals must be aware of these predictors This article is protected by copyright. All rights reserved.

PMID:33350488 | DOI:10.1002/jmv.26750

Leave a Reply

Your email address will not be published.