Long-term follow-up of treatment comparisons in RECOVERY: a randomised, open-label, platform trial for patients hospitalised with COVID-19

BackgroundThe Randomised Evaluation of COVID-19 Therapy (RECOVERY) trial evaluated the effects of sixteen potential treatments for patients hospitalised with COVID-19. Dexamethasone (at a dose of 6mg daily), tocilizumab, baricitinib, and the monoclonal antibodies casirivimab-imdevimab and sotrovimab were shown to reduce 28-day mortality in all or specific groups of patients. Here we report the long-term efficacy and safety of all sixteen therapies. MethodsPatients hospitalised with COVID-19 were potentially eligible to join this randomised, controlled, open-label, platform trial. Participants were randomly allocated to receive each trial treatment, or not, on top of usual care. Analyses were by intention to treat comparing each treatment with its own usual care control group. The pre-specified primary long-term follow-up outcome was 6-month all-cause mortality, presented as mortality rate ratios adjusted for baseline age and ventilation status. The key safety outcomes were major non-COVID infection and non-COVID death at 6 months. ISRCTN50189673 and NCT04381936. FindingsBetween 19 March 2020 and 19 March 2024, 48,402 patients were included in RECOVERY COVID-19 treatment comparisons. For each of the treatments previously demonstrated to be effective at 28 days, the early mortality benefit was preserved up to 6 months. Among 6425 patients in the dexamethasone (6mg daily) comparison, 6-month mortality was 34.3% vs 44.4% in the invasive mechanical ventilation group (rate ratio [RR] 0.68; 95% confidence interval [CI] 0.55-0.85; p=0.0006); 27.7% vs 29.2% in the oxygen or non-invasive ventilation group (RR 0.87; 95% CI 0.77-0.99; p=0.034); and 26.1% vs 22.5% in the no oxygen group (RR 1.10; 95% CI 0.89-1.36; p=0.39); test for trend p=0.0024. Among 4116 patients in the tocilizumab comparison, 34.3% vs 38.9% died within 6 months (RR 0.87; 95% CI 0.79-0.96; p=0.0077). Among 8156 patients in the baricitinib comparison, 15.7% vs. 16.6% died (RR 0.89; 95% CI 0.80-0.99; p=0.032). Among 3153 anti-SARS-CoV-2 serum antibody negative patients (the primary analysis population) in the casirivimab-imdevimab comparison, 29.3% vs. 34.7% died (RR 0.87; 95% CI 0.77-0.98; p=0.024). Among 720 patients with high serum nucleocapsid antigen concentration (the primary analysis population) in the sotrovimab comparison, 33.0% vs. 38.6% died (RR 0.78; 95% CI 0.61-1.00; p=0.050). In line with the 28-day results, aspirin, azithromycin, colchicine, convalescent plasma, dimethyl fumarate, empagliflozin, lopinavir-ritonavir, molnupiravir, and nirmatrelvir-ritonavir did not reduce 6-month mortality. Mortality at 6 months was higher with hydroxychloroquine therapy (33.3% vs. 30.2%; RR 1.14; 95% CI 1.02-1.27; p=0.018) and, among hypoxic patients not requiring ventilatory support, with higher dose dexamethasone (initial dose 20mg daily, 22.0% vs. 17.6%, RR 1.34; 95% CI 1.04-1.72; p=0.021). Allocation to dexamethasone 6mg once daily resulted in a small increase in major non-COVID infection within 6 months compared with usual care (21.4% vs. 19.1%; absolute difference 2.2%; 95% CI 0.2-4.5%). We found no evidence that any other treatments increased the risk of major non-COVID infection. InterpretationIn patients hospitalised with COVID-19, dexamethasone (at a dose of 6mg daily in hypoxic patients), tocilizumab (in hypoxic patients with CRP [≥]75 mg/L), baricitinib, casirivimab-imdevimab (in seronegative patients), and sotrovimab (in high antigen patients) reduced 6-month mortality. Dexamethasone at a dose of 6mg daily was associated with an increase in major non-COVID infection but there was no evidence of other later emerging harms. Other treatments tested in RECOVERY did not reduce 6-month mortality. FundingUK Research and Innovation (Medical Research Council) and National Institute for Health and Care Research (Grant ref: MC_PC_19056), and Wellcome Trust (Grant Ref: 222406/Z/20/Z).