Table 2.
Effect of allocation to dimethyl fumarate on key study outcomes
| Dimethyl fumarate (n = 356) | Usual care (n = 357) | Treatment effect (95% CI) | p value | |
|---|---|---|---|---|
| Primary outcome | ||||
| Ordinal scale at day 5a | ||||
| 7 vs 1–6 | 13 (4%) | 10 (3%) | ||
| 6–7 vs 1–5 | 23 (6%) | 22 (6%) | ||
| 5–7 vs 1–4 | 77 (22%) | 73 (20%) | ||
| 4–7 vs 1–3 | 204 (57%) | 193 (54%) | ||
| 3–7 vs 1–2 | 255 (72%) | 249 (70%) | ||
| 2–7 vs 1 | 269 (76%) | 261 (73%) | ||
| Common odds ratio | 1.12 (0.86–1.47) | 0.40 | ||
| Secondary clinical outcomes | ||||
| Sustained improvement in ordinal category within 10 daysb | 246 (69%) | 258 (72%) | 0.96 (0.80–1.16) | 0.70 |
| Improvement in clinical status at day 10c | 246 (69%) | 259 (73%) | 0.95 (0.87–1.05) | 0.31 |
| Baseline-adjusted day 5 S/F94d | 3.57 (0.06) | 3.64 (0.06) | −0.07 (−0.23 to 0.09) | 0.38 |
| Baseline-adjusted day 5 CRPe | 14.4 (1.2) | 14.0 (1.2) | 2% (−18 to 29%) | 0.84 |
| Median duration of hospitalisation, days | 8 | 8 | ||
| Discharged from hospital alive within 28 daysb | 274 (77%) | 281 (79%) | 0.95 (0.80–1.13) | 0.59 |
| Subsidiary clinical outcomes | ||||
| Use of ventilationc,f | 58/284 (20%) | 60/291 (21%) | 0.99 (0.72–1.37) | 0.95 |
| Non-invasive ventilation | 56/284 (20%) | 56/291 (19%) | 1.02 (0.73–1.43) | 0.89 |
| Invasive mechanical ventilation | 14/284 (5%) | 12/291 (4%) | 1.20 (0.56–2.54) | 0.64 |
| Successful cessation of invasive mechanical ventilationg | 0/3 (0%) | 0/0 | - | - |
| Renal dialysis or haemofiltrationc,h | 7/356 (2%) | 6/355 (2%) | 1.16 (0.39–3.43) | 0.78 |
| Safety outcomes | ||||
| Flushingc | ||||
| Some | 23 (7%) | 11 (3%) | 2.08 (1.03–4.21) | 0.04 |
| Severe | 8 (2%) | 0 (0%) | - | - |
| Subtotal: Any flushing | 31 (9%) | 11 (3%) | 2.81 (1.44–5.50) | 0.0026 |
| Gastrointestinal symptomsc | ||||
| Some | 34 (10%) | 18 (5%) | 1.88 (1.08–3.27) | 0.02 |
| Severe | 4 (1%) | 1 (0%) | 3.99 (0.45–35.51) | 0.21 |
| Subtotal: Any gastrointestinal symptoms | 38 (11%) | 19 (5%) | 1.99 (1.17–3.39) | 0.01 |
| Transaminitisc | 57 (19%) | 56 (18%) | 1.05 (0.75–1.46) | 0.78 |
| Acute kidney injuryc | 9 (3%) | 12 (4%) | 0.75 (0.32–1.75) | 0.51 |
| Non-coronavirus infectionc | ||||
| Pneumonia | 18 (5%) | 20 (6%) | 0.90 (0.49–1.68) | 0.75 |
| Urinary tract | 1 (0%) | 4 (1%) | 0.25 (0.03–2.23) | 0.21 |
| Biliary | 0 (0%) | 0 (0%) | - | - |
| Other intra-abdominal | 1 (<0.5%) | 1 (<0.5%) | - | - |
| Blood stream | 4 (1%) | 1 (0%) | 4.01 (0.45–35.71) | 0.21 |
| Skin | 1 (<0.5%) | 1 (<0.5%) | - | - |
| Other | 6 (2%) | 6 (2%) | 1.00 (0.33–3.08) | 1.00 |
| Subtotal: Any non-coronavirus infection | 27 (8%) | 31 (9%) | 0.87 (0.53–1.43) | 0.59 |
All p values are 2-sided.
aNumber of patients with a 'bad' outcome given. Treatment effects are odds ratios for 'bad' vs 'good' outcome. Common odds ratio estimated using a proportional odds model adjusted for ordinal scale at randomisation. For the 22 patients with missing data on ordinal scale at day 5, the median possible category was imputed (rounded up when there are an even number of possibilities).
bTreatment effect is a rate ratio estimated using logrank methods.
cTreatment effect is a risk ratio.
dTreatment effect is difference in mean S/F94 estimated using ANCOVA with adjustment for baseline S/F94. For patients discharged alive by day 5, a value of 4.76 was imputed. All 135 (18.9%) other missing values at day 5 were imputed using multiple imputation.
eANCOVA analyses of log transformed CRP with adjustment for randomisation value were conducted. 276 (38.7%) missing values at day 5 imputed using multiple imputation. Geometric means and approximate standard errors are presented and treatment effect is percentage change in CRP.
fAnalyses include only those on no ventilation support at randomisation.
gAnalyses restricted to those on invasive mechanical ventilation at randomisation.
hAnalyses exclude those on haemodialysis or haemofiltration at randomisation.