Predicting severity of Covid-19 using standard laboratory parameters

Buchmueller, T.; Groening, I.; Ihl, R.

Background More than 1.6 million people have already deceased due to a COVID-19 infection making it a major public health concern. A prediction of severe courses can enhance treatment quality and thus lower fatality and morbidity rates. The use of laboratory parameters has recently been established for a prediction. However, laboratory parameters have rarely been used in combination to predict severe outcomes. Method We used a retrospective case-control design to analyze risk factors derived from laboratory parameters. Patients treated for COVID-19 at an hospital in Krefeld, Germany, from March to May 2020 were included (n =42). Patients were classified into two categories based on their outcome (Mild course vs. treatment in intensive care unit). Laboratory parameters were compared across severity categories using non-parametric statistic. Identified laboratory parameters were used in a logistic regression model. The model was replicated using a) clinical standardized parameters b) aggregated factors derived from a factor analysis. Results Patients in intensive care unit showed elevated ALT, CRP and LDH levels, a higher leukocyte and neutrophile count, a higher neutrophile ratio and a lowered lymphocyte ratio. We were able to classify 95.1% of all cases correctly (96.6% of mild and 91.7% of severe cases, p<.001). Conclusion A number of routinely collected laboratory parameters is associated with a severe outcome of COVID-19. The combination of these parameters provides a powerful tool in predicting severity and can enhance treatment effectiveness.