Articles
Sotrovimab versus usual care in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial
RECOVERY Collaborative Group
Lancet Infect Dis Available online 28 August 2025
https://doi.org/10.1016/S1473-3099(25)00361-5
Summary
Background
Sotrovimab is a neutralising monoclonal antibody targeting the SARS-CoV-2 spike protein. We aimed to evaluate the efficacy and safety of sotrovimab in the RECOVERY trial, an investigator-initiated, individually randomised, controlled, open-label, adaptive platform trial testing treatments for patients admitted to hospital with COVID-19.
Methods
Patients admitted with COVID-19 pneumonia to 107 UK hospitals were randomly assigned (1:1) to either usual care alone or usual care plus a single 1 g infusion of sotrovimab, using web-based unstratified randomisation. Participants were eligible if they were aged at least 18 years, or aged 12–17 years if weighing at least 40kg, and had confirmed COVID-19 pneumonia with no medical history that would put them at significant risk if they participated in the trial. Participants were retrospectively categorised as having a high antigen level if baseline serum SARS-CoV-2 nucleocapsid antigen was above the median concentration (the prespecified primary efficacy population), otherwise they were categorised as having a low antigen level. The primary outcome was 28-day mortality assessed by intention to treat. Safety outcomes were assessed among all participants, regardless of antigen level. Recruitment closed on March 31, 2024, when funding ended. The trial is registered with ISRCTN (50189673) and ClinicalTrials.gov (NCT04381936).
Findings
From Jan 4, 2022, to March 19, 2024, 1723 patients were enrolled in the RECOVERY sotrovimab comparison. Of these, 828 (48%) were assigned to usual care plus sotrovimab and 895 (52%) were assigned to usual care only. Mean patient age was 70·7 years (SD 14·8) and 1033 (60%) were male. 720 (42%) patients were classified as having a high antigen level, 717 (42%) as having a low antigen level, and 286 (17%) had unknown antigen status. 1389 (81%) patients were vaccinated, 1179 (82%) of 1438 patients with known serostatus had anti-spike antibodies at randomisation, and 1021 (>99%) of 1026 patients with sequenced samples were infected with omicron variants. Among patients with a high antigen level, 82 (23%) of 355 assigned to sotrovimab versus 106 (29%) of 365 assigned usual care died within 28 days (rate ratio 0·75, 95% CI 0·56–0·99; p=0·046). In an analysis of all randomly assigned patients (regardless of antigen status), 177 (21%) of 828 patients assigned to sotrovimab versus 201 (22%) of 895 assigned to usual care died within 28 days (0·95, 0·77–1·16; p=0·60). Infusion reactions were recorded in 12 (2%) of 781 patients receiving sotrovimab. We found no difference between groups in any other safety outcome.
Table 1. Baseline characteristics
| Empty Cell | Empty Cell | Patients with a high antigen level | All patients who were randomly assigned | ||
|---|---|---|---|---|---|
| Empty Cell | Empty Cell | Sotrovimab (n=355) | Usual care (n=365) | Sotrovimab (n=828) | Usual care (n=895) |
| Age, years | 72·5 (13·3) | 72·1 (13·7) | 70·9 (14·2) | 70·4 (15·4) | |
| Age group, years | |||||
| <70 | 123 (35%) | 141 (39%) | 342 (41%) | 369 (41%) | |
| ≥70 to <80 | 123 (35%) | 121 (33%) | 251 (30%) | 272 (30%) | |
| ≥80 | 109 (31%) | 103 (28%) | 235 (28%) | 254 (28%) | |
| Sex | |||||
| Male | 218 (61%) | 226 (62%) | 490 (59%) | 543 (61%) | |
| Female | 137 (39%) | 139 (38%) | 338 (41%) | 352 (39%) | |
| Race or ethnicity | |||||
| White | 301 (85%) | 333 (91%) | 706 (85%) | 779 (87%) | |
| Black, Asian, and minority ethnic | 32 (9%) | 16 (4%) | 64 (8%) | 65 (7%) | |
| Unknown | 22 (6%) | 16 (4%) | 58 (7%) | 51 (6%) | |
| Time since symptom onset, days | 6 (3–11) | 6 (3–12) | 6 (3–11) | 6 (3–11) | |
| Time since admission to hospital, days | 2 (1–5) | 2 (1–5) | 2 (1–5) | 2 (1–5) | |
| Respiratory support received | |||||
| None | 43 (12%) | 54 (15%) | 119 (14%) | 137 (15%) | |
| Simple oxygen | 226 (64%) | 213 (58%) | 512 (62%) | 557 (62%) | |
| Non-invasive ventilation | 71 (20%) | 87 (24%) | 168 (20%) | 169 (19%) | |
| Invasive mechanical ventilation | 15 (4%) | 11 (3%) | 29 (4%) | 32 (4%) | |
| Previous diseases | |||||
| Diabetes | 107 (30%) | 84 (23%) | 249 (30%) | 219 (24%) | |
| Heart disease | 119 (34%) | 113 (31%) | 259 (31%) | 272 (30%) | |
| Chronic lung disease | 123 (35%) | 128 (35%) | 327 (39%) | 325 (36%) | |
| Tuberculosis | 0 | 1 (<1%) | 2 (<1%) | 4 (<1%) | |
| HIV | 3 (1%) | 1 (<1%) | 6 (1%) | 5 (1%) | |
| Severe liver disease* | 6 (2%) | 3 (1%) | 19 (2%) | 16 (2%) | |
| Severe kidney impairment† | 45 (13%) | 41 (11%) | 84 (10%) | 74 (8%) | |
| Any of the above | 242 (68%) | 237 (65%) | 578 (70%) | 602 (67%) | |
| Severely immunocompromised‡ | 112 (32%) | 112 (31%) | 206 (25%) | 208 (23%) | |
| Received a COVID-19 vaccine | 296 (83%) | 292 (80%) | 675 (82%) | 714 (80%) | |
| Use of other treatments | |||||
| Corticosteroids§ | 329 (93%) | 334 (92%) | 755 (91%) | 801 (89%) | |
| Remdesivir | 144 (41%) | 128 (35%) | 315 (38%) | 313 (35%) | |
| Tocilizumab | 66 (19%) | 60 (16%) | 144 (17%) | 137 (15%) | |
| Plan to use tocilizumab within the next 24 h | 28 (8%) | 33 (9%) | 48 (6%) | 67 (7%) | |
| Viral load in baseline nose swab, log viral copies per mL | 6·1 (4·6–7·0) | 6·1 (5·0–7·2) | 5·6 (3·7–6·7) | 5·6 (3·9–6·8) | |
| Antigen status | |||||
| High | 355 (100%) | 365 (100%) | 355 (43%) | 365 (41%) | |
| Low | 0 | 0 | 339 (41%) | 378 (42%) | |
| Unknown | 0 | 0 | 134 (16%) | 152 (17%) | |
| Serostatus (anti-nucleocapsid antibodies) | |||||
| Positive | 62 (17%) | 76 (21%) | 214 (26%) | 240 (27%) | |
| Negative | 293 (83%) | 289 (79%) | 481 (58%) | 504 (56%) | |
| Unknown | 0 | 0 | 133 (16%) | 151 (17%) | |
| Serostatus (anti-spike antibodies) | |||||
| Positive | 252 (71%) | 262 (72%) | 569 (69%) | 610 (68%) | |
| Negative | 103 (29%) | 103 (28%) | 126 (15%) | 133 (15%) | |
| Unknown | 0 | 0 | 133 (16%) | 152 (17%) | |
*Defined as requiring ongoing specialist care.
†Defined as estimated glomerular filtration rate <30 mL/min per 1·73 m2.
‡In the opinion of the managing clinician.
§Including all those who were randomly assigned in the comparison of high-dose versus low-dose steroids.
Table 2. Effect of allocation to sotrovimab on key study outcomes in patients with a high antigen level
| Empty Cell | Empty Cell | Sotrovimab (n=355) | Usual care (n=365) | Rate ratio, risk ratio, or mean difference (95% CI) | p value |
|---|---|---|---|---|---|
| Primary outcome | |||||
| 28-day mortality | 82 (23%) | 106 (29%) | 0·75 (0·56 to 0·99) | 0·046 | |
| Secondary outcomes | |||||
| Median (IQR) time to being discharged alive, days | 13 (7 to >28) | 16 (7 to >28) | .. | .. | |
| Discharged from hospital within 28 days | 236 (66%) | 226 (62%) | 1·12 (0·93 to 1·34) | .. | |
| Receipt of invasive mechanical ventilation or death* | 82/340 (24%) | 102/354 (29%) | 0·82 (0·64 to 1·03) | .. | |
| Invasive mechanical ventilation | 14/340 (4%) | 11/354 (3%) | 1·71 (0·81 to 3·61) | .. | |
| Death | 74/340 (22%) | 100/354 (28%) | 0·74 (0·58 to 0·95) | .. | |
| Subsidiary clinical outcomes | |||||
| Use of ventilation† | 41/269 (15%) | 41/267 (15%) | 0·97 (0·66 to 1·44) | .. | |
| Non-invasive ventilation | 40/269 (15%) | 41/267 (15%) | 0·95 (0·64 to 1·41) | .. | |
| Invasive mechanical ventilation | 6/269 (2%) | 3/267 (1%) | 1·82 (0·47 to 7·11) | .. | |
| Successful cessation of invasive mechanical ventilation‡ | 5/15 (33%) | 3/11 (27%) | 1·07 (0·25 to 4·65) | .. | |
| Use of haemodialysis or haemofiltration§ | 12/347 (3%) | 6/356 (2%) | 1·97 (0·77 to 5·06) | .. | |
| Virological outcomes | |||||
| Baseline-adjusted viral load (log copies per mL) on day 3 | 4·89 (0·10) | 4·94 (0·10) | −0·05 (−0·32 to 0·23) | .. | |
| Baseline-adjusted viral load (log copies per mL) on day 5 | 4·26 (0·11) | 4·35 (0·10) | −0·09 (−0·38 to 0·20) | .. | |
*Excluding patients receiving invasive mechanical ventilation at randomisation.
†Excluding patients receiving invasive or non-invasive ventilation at randomisation.
‡Excluding patients not receiving invasive mechanical ventilation at randomisation.
§Excluding patients receiving renal replacement therapy at randomisation.
Interpretation
In patients admitted to hospital with COVID-19 pneumonia, sotrovimab was associated with reduced mortality in the primary analysis population who had a high serum SARS-CoV-2 antigen concentration at baseline, but not in the overall population. Treatment options for patients admitted to hospital are limited, and mortality in those receiving current standard of care was high. The emergence of high-level resistance to sotrovimab among subsequent SARS-CoV-2 variants restricts its current usefulness, but these results indicate that targeted neutralising antibody therapy could potentially still benefit some patients admitted to hospital who are at high risk of death in an era of widespread vaccination and omicron infection.
Funding
UK Research and Innovation (Medical Research Council) and National Institute for Health and Care Research.
