Table 3. Primary and Secondary Outcomes.
| Variablea | No. (%) | Risk difference (95% CI)b | Effect, β estimate (95% CI) | P value | FDR | |
|---|---|---|---|---|---|---|
| Helmet noninvasive ventilation (n = 159) | Usual respiratory support (n = 161) | |||||
| Primary outcome (28-d mortality)c | ||||||
| Primary analysis | 43 (27.0) | 42 (26.1) | 1.0 (−8.7 to 10.6) | Relative risk, 1.04 (0.72 to 1.49) | .85 | |
| Per-protocol analysis | 41/146 (28.1) | 41/157 (26.1) | 2.0 (−8.1 to 12.0) | Relative risk, 1.08 (0.74 to 1.56) | .70 | |
| Secondary outcomes | ||||||
| ICU mortalityd | 56 (35.2) | 60 (37.3) | −2.0 (−12.6 to 8.5) | Relative risk, 0.95 (0.71 to 1.26) | .70 | 0.91 |
| Hospital mortalityd | 61 (38.4) | 64 (39.8) | −1.4 (−12.1 to 9.3) | Relative risk, 0.97 (0.73 to 1.27) | .80 | 0.91 |
| ICU-free days at day 28, median (IQR)e | 12 (0 to 20) | 8 (0 to 19) | Median difference, 4.0 (−2.8 to 10.8) | 0.08 (−0.32 to 0.48) | .70 | 0.91 |
| Mechanical ventilation–free days at day 28, median (IQR)e | 28 (0 to 28) | 23 (0 to 28) | Median difference, 5.0 (−2.6 to 12.6) | 0.05 (−0.28 to 0.38) | .76 | 0.91 |
| Kidney replacement–free days at day 28, median (IQR)e | 28 (0 to 28) | 28 (0 to 28) | −0.02 (−0.33 to 0.30) | .92 | 0.92 | |
| Vasopressor-free days at day 28, median (IQR)e | 28 (0 to 28) | 25 (0 to 28) | Median difference, 3.0 (−2.8 to 8.8) | 0.05 (−0.27 to 0.37) | .76 | 0.91 |
| Hospital LOS, median (IQR), df | 18 (11 to 26) | 17 (11 to 30) | Median difference, 1.0 (−2.1 to 4.1) | −0.16 (−0.33 to 0.00) | .05 | 0.42 |
| Endotracheal intubation | 75 (47.2) | 81 (50.3) | −3.1 (−14.1 to 7.8) | Relative risk, 0.94 (0.75 to 1.17) | .57 | 0.91 |
| Time to intubation, median (IQR) [No.], d | 5 (4 to 10) [75] | 4 (2 to 8) [81] | ||||
Abbreviations: FDR, false discovery rate; Fio2, fraction of inspired oxygen; ICU, intensive care unit; LOS, length of stay; PEEP, positive end-expiratory pressure.
The number of patients for each variable is the total number of patients in the respective group unless otherwise specified. Categorical outcomes were compared with a χ2 test, and the results were reported as risk difference and relative risk with 95% CIs. Continuous outcomes were compared with generalized linear mixed models and the results were reported as β estimates with 95% CI. Denominator of the percentage is the total number of subjects in each group. False discovery rate was used to adjust for multiple testing for the analyses of the following secondary outcomes: ICU mortality, hospital mortality, ICU-free days, mechanical ventilation–free days, kidney replacement–free days, vasopressor-free days, endotracheal intubation, and hospital LOS (hospital LOS was a post hoc secondary outcome).
Risk differences were expressed as percentages.
Secondary analyses of the primary outcome in the primary analysis population and per-protocol population are reported in eTable 9 in Supplement 2.
ICU and hospital mortality are defined as death in the index ICU admission or hospital admission, censored by day 180.
ICU-free days, mechanical ventilation–free days, kidney replacement–free days, and vasopressor-free days are calculated according to 28-day observation.
Hospital LOS was a post hoc secondary outcome.