Table 1 Limitations in the RECOVERY trial and the meta-analysis

RECOVERY trial [7]

No stratification between centers

BMI, ethnicity not reported

Location of patient at randomization unknown (ward/ICU)

Age imbalance in the study population

Distribution of the various factors associated with outcome not reported for the different subgroups

For patients receiving mechanical ventilation, PEEP, FiO₂, PaO₂/FiO₂ not reported

Short-term outcome (28-day mortality) reported

Meta-analysis [11]

Selected data only were included

Some of the trials were stopped prematurely and underpowered

Excessive weight (57%) of a single trial (RECOVERY) with its own limitations see above

Imbalance in age and sex in favor of dexamethasone in the CODEX trial, which accounted for 19% of weight

Imbalance in age and BMI in favor of hydrocortisone in the CAPE COVID trial which accounted for 7% of weight

Selected patients from the REMAP CAP trial were included, increasing the mortality difference from 2 to 7%. This trial accounted for 12% of weight

Conflicting results for the subgroups of mechanical ventilation versus no ventilation between meta-analysis and RECOVERY trial

Short-term outcomes only considered (21- to 28-day mortality)