Statistical analysis
Data analysis was performed in IBM SPSS Statistics for Windows version 22.0 (IBM Corp., Armonk, NY) and Medcalc version 16 (MedCalc Software bvba, Ostend, Belgium). Shapiro-Wilk test and Kolmogorov-Smirnov test were used to evaluate the normality of the distribution of numeric data. The descriptive statistics were presented as frequency with percentage for categorical variables, median with interquartile range (IQR) for non-normally distributed numeric data, and mean with standard deviation (SD) for normally distributed data. The laboratory findings were compared among the study groups using Independent-Samples Kruskal-Wallis Test for non-normally distributed numeric data, and One-way ANOVA Test for normally distributed numeric data. Post-hoc pairwise comparisons were done using Dunn’s Post-hoc test with Bonferroni correction for non-normally distributed data. We calculated the difference in median for determining the effect size in post-hoc pairwise comparison of laboratory findings among the study groups. Spearman correlation analysis was used to evaluate the relationships between serum procalcitonin, lactate, aspartate transaminase (AST) and alanine transaminase (ALT) levels. Receiver operating characteristic (ROC) curve analysis was used to estimate the severe disease using serum procalcitonin and lactate levels. Youden index was used to find the best cut-off values of these parameters. The AUCs, sensitivity and specificity with 95% confidence intervals were calculated. A p-value less than 0.05 was considered statistically significant.