Primary outcomes were death rates at days 30 and 60 of study randomization. Patients in the CCP group received 3 daily doses of CCP (600 mL/d) in addition to standard treatment control patients received standard treatment only. Patients with severe COVID-19 who were within 10 days of initial symptom onset were eligible. Because most of the data was distributed not normally, results are expressed as median and interquartile range (IQR) as dispersion characteristic, and p value less than 0.05 ( p < 0.05) was considered as statistically significant.To assess whether high-dose coronavirus disease (COVID-19) convalescent plasma (CCP) transfusion may benefit patients with severe COVID-19, we conducted a multicenter randomized trial in Brazil. If data were not normally distributed, the comparison of medians between different groups was switched to non-parametric one-way ANOVA on ranks or Kruskal-Wallis test followed by two-stage step-up method of Benjamini, Krieger and Yekutieli as post-hoc test. Homogeneity of variances was tested using Brown-Forsythe and Bartlett’s tests, and in a case of unequal SDs Brown-Forsythe and Welch ANOVA test was applied.

The comparison of means between different groups of numerical variables was performed using one-way ANOVA. Z-value equal or higher than 2, also equal or lower than −2 indicates that variable significantly differs from the control group mean.

Z-transformation of the QST data was performed using the following expression z = ( x − μ)/ σ, where x corresponds to the individual variable from the datasets that must be normalized with z-transformation, while μ and σ are the control group mean and standard deviation (SD), respectively.

To test whether the collected numerical data are normally distributed, the D’Agostino and Pearson, Anderson-Darling and Shapiro-Wilk normality tests were applied. All the graphs, calculations, and statistical analyses were performed using GraphPad Prism software version 8.0 for Mac (GraphPad Software, San Diego, CA, USA).

Graphpad prism version 7