Randomized Controlled Trial
Study population and setting
This report from the SOLIDARITY study involved 11,266 adults hospitalized with COVID-19 admitted to 405 hospitals in 30 countries between March 22 and October 4, 2020. At each hospital, patients were randomized in equal proportions between basic standard of care and whichever study drugs were locally available, resulting in 4 intervention groups and 4 pairwise control groups. There were no placebos, and the selection of study drugs was open-label. There was partial overlap between the 4 control groups, as randomization to standard of care would place a patient into the control group for all drugs locally available at the hospital site. 2750 patients were allocated remdesivir, 954 hydroxychloroquine, 1411 lopinavir-ritonavir, 2063 interferon, and 4088 received no drug. Exclusion criteria were prior use of study drug, contra-indication to study drug, or anticipated transfer elsewhere within 72 hours. The primary endpoint was 28-day in-hospital mortality, which was further analyzed within subgroups of varying age and disease severity (defined by ventilation at entry). Secondary outcomes included initiation of ventilation and hospital length of stay.
Summary of Main Findings
In intention-to-treat analysis, no study drug had a significant impact on 28-day in-hospital mortality compared to standard of care alone (death rate ratio [RR]: remdesivir RR=0.95, [CI 0.81-1.11]; hydroxychloroquine RR =1.19, [CI 0.89-1.59]; lopinavir-ritonavir RR = 1.00, [CI 0.79-1.25]; interferon RR = 1.16, [CI 0.96-1.39]). No mortality benefits were apparent in any subgroup defined by age or ventilation at entry. In secondary analysis, no drugs were found to reduce initiation of ventilation. Three drugs were found to increase the percentage of patients remaining in hospital at day 7 (remdesivir 69% vs. 54%, hydroxychloroquine 64% vs. 54%, lopinavir-ritonavir 68% vs. 59%). Meta-analyses were conducted to combine the SOLIDARITY remdesivir results with prior trials (ACTT-1, Wuhan, SIMPLE); in this combining of the trial results, remdesivir was found to confer no significant survival benefit (RR 0.91 [CI 0.79-1.05]).
The randomized design large sample size and high patient adherence were strengths. Overall compliance for each drug was high (94%-96%), and crossover was low (2-6%). A diverse patient population was recruited from hospital sites across multiple countries, enhancing the generalizability of findings. In addition, studies set in low- and middle-income countries are critical to advance COVID-19 therapeutics among these populations. Supplementary analysis was performed to adjust results for outcome predictors, including patient characteristics and corticosteroid use. These multivariate Cox regression analyses also found no significant mortality benefits for any drug.
This study was an open-label design without use of placebo arms, meaning the participants and study teams were not blinded to treatment, possibly subjecting the data to reporting bias. In other words, providers may have consciously or unconsciously altered clinical decision-making based on patient’s status in the trial. Consent to enrollment was mostly prospective but could be retrospective where locally approved, which could have led to bias when selecting patients for enrollment; the extent of retrospective enrollment is unclear. Additionally, follow-up for patients was ceased after discharge, precluding analysis of all-cause mortality or long-term mortality effects of drug treatment. An important omission in this pre-print is that the time from symptom onset to enrollment and study drug administration was not clearly stated; there are emerging data supporting that delayed antiviral treatment is less effective in preventing the worst COVID-19 outcomes. Since most patients with early COVID-19 are not admitted to the hospital, this may lead to lower benefit of antivirals. In addition, treatment time for three drugs (remdesivir, hydroxychloroquine, and lopinavir-ritonavir) was expected to last at least 7 days, preventing accurate comparison of hospital length of stay (defined as percentage remaining at hospital at 7 days) between treatment arm and control arm.
In agreement with the Recovery trial, this study provides further evidence that hydroxychloroquine and lopinavir-ritonavir provide little to no survival benefit for hospitalized COVID-19 patients. As the largest trial to date on mortality benefits of remdesivir and interferon, this study suggests that these drugs have little to no effect on in-hospital mortality when used in low- and middle-income settings in a pragmatic approach agnostic to duration of symptoms or severity.
This review was posted on: 13 November 2020