Table 2.
COVID-19 hospitalization or death prior to day 28 by treatment group
| Control (N=589) |
CCP (N=592) |
P-value1 | |
|---|---|---|---|
| Number randomized, n | 615 | 610 | |
| Number transfused, n | 589 | 592 | |
| Participants with primary outcome (hospitalization), n | 37 | 17 | 0.004 |
| Unadjusted Relative Risk [95% one-sided CI upper bound] | 0.46 [0.73] | 0.004 | |
| Hospitalization not due to COVID-19 2 , n | 3 | 4 | 0.77 |
| Details of primary outcome severity, n | |||
| Death after hospitalization | 3 | 0 | |
| ICU Hospitalization (no mechanical ventilation) | 4 | 3 | |
| Non-ICU hospitalization due to COVID-19, requiring supplemental oxygen | 26 | 12 | |
| Non-ICU hospitalization due to COVID-19, not requiring supplemental oxygen | 4 | 2 | |
| A stay of >24 hours for observation in an ED, field hospital or other healthcare unit or receipt of oxygen for >24 hours, outside of hospital | 0 | 0 | |
| Expected time free of primary outcome3 (days) | 26.27 | 27.26 | |
| Difference (SE) [95% one-sided CI lower bound] | 0.99 (0.28) [0.54] | 0.0002 | |
| Probability of remaining free of hospitalization 3 | 0.93 | 0.97 | |
| Difference (SE) [95% one-sided CI lower bound] | −0.04 (0.01)[-0.02] |
0.0003 | |
| Primary outcome events by sex 4 | |||
| Female | 21 | 9 | |
| Male | 16 | 8 | |
| Primary outcome events by vaccination status 4,5 | |||
| Unvaccinated | 36 | 17 | |
| Partially vaccinated | 1 | 0 | |
| Fully vaccinated | 0 | 0 |
P-values calculated using one-sided Fisher’s exact test for count data and Cochran-Mantel-Haenszel test.
Number of first hospitalizations that were not due to COVID-19 (i.e.- hallucinations due to pre-existing mental illness, renal colic, constipation or pancreatitis)
Adjusted for age, trial site, BMI, baseline albumin, bicarbonate, c-reactive protein, glucose, potassium, and baseline abnormal head, eyes, ears, nose, and throat physical examination, as specified from a random survival forest analysis of baseline characteristics. Primary analysis data set restricted via principal component analysis, reducing data set to 990 complete case participants with full data (494 transfused with control plasma, 496 transfused with convalescent plasma). Changing the model by decreasing covariates did not change inferences.
Numbers of participants in each category found on Table 1
Fully vaccinated 14 days following final administration
Cox proportional hazards models with only treatment assignment and then adjusted for age; age and site; age site, and BMI; and finally same covariates as TMLE model, showed congruence with TMLE results as hazard ratios were approximately 0.45 across proportional hazards models.