Of the 1114 patients assessed for eligibility, 1062 underwent randomization; 541 were assigned to the remdesivir group and 521 to the placebo group (intention-to-treat population) (Figure 1); 159 (15.0%) were categorized as mild to moderate disease, and 903 (85.0%) were in the severe disease tier. Of those assigned to receive remdesivir, 531 patients (98.2%) received treatment as assigned. In 52 patients, remdesivir had been discontinued before day 10 due to an adverse reaction or serious adverse reaction other than death and 10 withdrew their consent. Of those assigned to receive placebo, 517 patients (99.2%) received placebo as assigned. Seventy patients discontinued placebo before day 10 due to an adverse event or serious adverse event other than death and 14 withdrew their consent.
🎬📺 Free Movies and Free TV Shows! 🎭🎬
A total of 517 patients in the remdesivir group and 508 in the placebo group completed the study through day 29, recovered or died. Fourteen patients who received remdesivir and 9 who received placebo ended the study before Day 29. A total of 54 patients who were randomized to be in the mild to moderate tier were subsequently determined to meet the criteria for severe disease, resulting in 105 patients in the mild to moderate disease layer and 957 in the severe layer. The as-treated population included 1048 patients who received the assigned treatment (532 in the remdesivir group, including one patient who was randomly assigned to placebo and received remdesivir, and 516 in the placebo group).
The mean age of the patients was 58.9 years and 64.4% were male (table 1). Based on the evolving epidemiology of Covid-19 during the trial, 79.8% of patients were enrolled in sites in North America, 15.3% in Europe and 4.9% in Asia (Table S1 in the Additional attachment). Overall, 53.3% of patients were Caucasian, 21.3% were Black, 12.7% were Asian, and 12.7% were classified as other or not reported; 250 (23.5%) were Hispanic or Latino. Most patients had one (25.9%) or two or more (54.5%) of the pre-specified coexisting conditions at enrollment, most commonly hypertension (50.2%), obesity (44.8%) and type 2 diabetes mellitus (30.3%).
The median number of days between symptom onset and randomization was 9 (interquartile range, 6 to 12) (Table S2). A total of 957 patients (90.1%) had severe disease at enrollment; 285 patients (26.8%) met Category 7 criteria on the ordinal scale, 193 (18.2%) Category 6, 435 (41.0%) Category 5, and 138 (13.0%) Category 4. Eleven patients (1.0%) had no ordinal scale data at enrollment; all of these patients discontinued the study before treatment. During the study, 373 patients (35.6% of 1048 patients in the AS-treated population) received hydroxychloroquine and 241 (23.0%) received a glucocorticoid (Table S3).
Patients in the remdesivir group had a shorter recovery time than patients in the placebo group (median, 10 days, compared to 15 days; rate ratio for recovery, 1.29; 95% confidence interval [CI], 1.12 to 1.49; P<0.001) (Figure 2 and table 2). In the severe disease stratum (957 patients), the median time to recovery was 11 days compared to 18 days (rate ratio for recovery, 1.31; 95% CI, 1.12 to 1.52) (Table S4). The recovery rate was greatest in patients with a baseline ordinal score of 5 (recovery rate ratio, 1.45; 95% CI, 1.18 to 1.79); in patients with a baseline score of 4 and those with a baseline score of 6, the estimates of the recovery rate ratio were 1.29 (95% CI, 0.91 to 1.83) and 1.09 (95% CI, 0, respectively .76 to 1.57). For those who received mechanical ventilation or ECMO at enrollment (base ordinal score of 7), the recovery ratio was 0.98 (95% CI, 0.70 to 1.36). Information on treatment interactions with the baseline ordinal score as a continuous variable is provided in Table S11. An analysis correcting for the baseline ordinal score as a covariate was performed to evaluate the overall effect (of the percentage of patients in each category with baseline ordinal scores) on the primary outcome measure. This adjusted analysis yielded a similar estimate of treatment effect (ratio to recovery, 1.26; 95% CI, 1.09 to 1.46). Patients who underwent randomization during the first 10 days after symptom onset had a recovery ratio of 1.37 (95% CI, 1.14 to 1.64), while patients who received more than 10 days after onset of symptoms symptoms randomization had a rate ratio for recovery of 1.20 (95% CI, 0.94 to 1.52) (figure 3). The benefit of remdesivir was greater when given earlier in the disease, although benefit persisted in most analyzes of duration of symptoms (Table S6). Sensitivity analyzes censoring data for earliest reported use of glucocorticoids or hydroxychloroquine still showed efficacy of remdesivir (9.0 days to remdesivir recovery versus 14.0 days to placebo recovery; rate ratio 1.28; 95% CI 1 .09 to 1.50, and 10.0 versus 16.0 days to recovery; rate ratio, 1.32; 95% CI, 1.11 to 1.58), respectively (Table S8).
Main secondary outcome
The probability of improvement in the ordinal scale score was higher in the remdesivir group, as determined by a proportional odds model at the day 15 visit, than in the placebo group (odds ratio for improvement, 1.5; 95% CI, 1.2 to 1 ,9, adjusted for disease severity) (table 2 and Fig. S7).
Kaplan-Meier estimates of mortality at Day 15 were 6.7% in the remdesivir group and 11.9% in the placebo group (hazard ratio 0.55; 95% CI 0.36 to 0.83); the estimates at day 29 were 11.4% and 15.2% in two groups, respectively (hazard ratio 0.73; 95% CI 0.52 to 1.03). The differences in mortality between groups varied significantly depending on baseline severity (table 2), with the largest difference seen in patients with an ordinal baseline score of 5 (hazard ratio 0.30; 95% CI 0.14 to 0.64). Information on treatment interactions with the baseline ordinal score with regard to mortality is provided in Table S11.
Additional secondary results
Patients in the remdesivir group had a shorter time to improvement in one or two categories on the ordinal scale from baseline than patients in the placebo group (improvement in one category: median, 7 vs. 9 days; recovery rate ratio, 1.23; 95 % CI, 1.08 to 1.41; improvement in two categories: median, 11 vs. 14 days; ratio, 1.29; 95% CI, 1.12 to 1.48) (table 3). Patients in the remdesivir group had a shorter time to discharge or a National Early Warning Score of 2 or less than those in the placebo group (median, 8 days vs. 12 days; hazard ratio, 1.27; 95% CI, 1.10 to 1.46). The initial hospital stay was shorter in the remdesivir group than in the placebo group (median, 12 days vs. 17 days); 5% of patients in the remdesivir group were readmitted to the hospital, compared to 3% in the placebo group.
Of the 913 patients who received oxygen at enrollment, those in the remdesivir group continued to receive oxygen for fewer days than patients in the placebo group (median, 13 days vs. 21 days), and the incidence of new oxygen use in patients who did not receive oxygen at enrollment was lower in the remdesivir group than in the placebo group (incidence, 36% [95% CI, 26 to 47] versus 44% [95% CI, 33 to 57]). For the 193 patients who received non-invasive ventilation or high-flow oxygen at enrollment, the median duration of use of these interventions was 6 days in both the remdesivir and placebo groups. Of the 573 patients who were not on non-invasive ventilation, high-flow oxygen, invasive ventilation, or ECMO at baseline, the incidence of new non-invasive ventilation or high-flow oxygen use was lower in the remdesivir group than in the placebo group (17% [95% CI, 13 to 22] versus 24% [95% CI, 19 to 30]). Of the 285 patients who received mechanical ventilation or ECMO at enrollment, patients in the remdesivir group received these interventions for fewer consecutive days than those in the placebo group (median, 17 days vs. 20 days), and the incidence of new mechanical ventilation or ECMO use among the 766 patients who did not receive these interventions at enrollment was lower in the remdesivir group than in the placebo group (13% [95% CI, 10 to 17] versus 23% [95% CI, 19 to 27]) (table 3).
In the as-treated population, serious adverse events occurred in 131 of 532 patients (24.6%) in the remdesivir group and in 163 of 516 patients (31.6%) in the placebo group (Table S17). There were 47 serious adverse events of respiratory failure in the remdesivir group (8.8% of patients), including acute respiratory failure and the need for endotracheal intubation, and 80 in the placebo group (15.5% of patients) (Table S19). No deaths were considered by the investigators to be related to the treatment assignment.
Grade 3 or 4 adverse reactions occurred on or before Day 29 in 273 patients (51.3%) in the remdesivir group and in 295 (57.2%) in the placebo group (Table S18); 41 events were assessed by the investigators as related to remdesivir and 47 events to placebo (Table S17). The most common non-serious adverse reactions occurring in at least 5% of all patients were decreased glomerular filtration rate, decreased hemoglobin, decreased lymphocyte count, respiratory failure, anemia, pyrexia, hyperglycemia, increased blood creatinine and increased blood glucose (Table S20). The incidence of these adverse events was generally similar in the remdesivir and placebo groups.
After the Data and Safety Board of Trustees recommended that the preliminary primary analysis report be provided to the sponsor, data from a total of 51 patients (4.8% of the total study enrollment) — 16 (3.0%) in the remdesivir group and 35 (6.7) %) in the placebo group — were not blinded; 26 (74.3%) of those in the placebo group whose data was not blinded received remdesivir. Sensitivity analyzes evaluating blinding (patients whose treatment assignments were unblinded had their data censored at the time of unblinding) and crossover (patients in the placebo group treated with remdesivir had their data censored at the start of treatment with remdesivir) yielded results similar to those of the primary analysis (Table S9).
🎬📺 Free Movies and Free TV Shows! 🎭🎬