Colchicine in patients admitted to hospital with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial

Summary Background Colchicine has been proposed as a treatment for COVID-19 based on its anti-inflammatory actions. We aimed to evaluate the efficacy and safety of colchicine in patients admitted to hospital with COVID-19. Methods In this streamlined, randomised, controlled, open-label trial, underway at 177 hospitals in the UK, two hospitals in Indonesia, and two hospitals in Nepal, several possible treatments were compared with usual care in patients hospitalised with COVID-19. Patients were eligible for inclusion in the study if they were admitted to hospital with clinically suspected or laboratory confirmed SARS-CoV-2 infection and had no medical history that might, in the opinion of the attending clinician, put the patient at significant risk if they were to participate in the trial. Eligible and consenting adults were randomly assigned (1:1) to receive either usual standard of care alone (usual care group) or usual standard of care plus colchicine (colchicine group) using web-based simple (unstratified) randomisation with allocation concealment. Participants received colchicine 1 mg after randomisation followed by 500 μg 12 h later and then 500 μg twice a day by mouth or nasogastric tube for 10 days in total or until discharge. Dose frequency was halved for patients receiving a moderate CYP3A4 inhibitor (eg, diltiazem), patients with an estimated glomerular filtration rate of less than 30 mL/min per 1·73m2, and those with an estimated bodyweight of less than 70 kg. The primary outcome was 28-day mortality, secondary endpoints included time to discharge, the proportion of patients discharged from hospital within 28 days, and, in patients not on invasive mechanical ventilation at randomisation, a composite endpoint of invasive mechanical ventilation or death. All analyses were by intention-to-treat. The trial is registered with ISRCTN, 50189673, and ClinicalTrials.gov, NCT04381936. Findings Between Nov 27, 2020, and March 4, 2021, 11 340 (58%) of 19 423 patients enrolled into the RECOVERY trial were eligible to receive colchicine; 5610 (49%) patients were randomly assigned to the colchicine group and 5730 (51%) to the usual care group. Overall, 1173 (21%) patients in the colchicine group and 1190 (21%) patients in the usual care group died within 28 days (rate ratio 1·01 [95% CI 0·93 to 1·10]; p=0·77). Consistent results were seen in all prespecified subgroups of patients. Median time to discharge alive (10 days [IQR 5 to >28]) was the same in both groups, and there was no significant difference in the proportion of patients discharged from hospital alive within 28 days (3901 [70%] patients in the colchicine group and 4032 [70%] usual care group; rate ratio 0·98 [95% CI 0·94 to 1·03]; p=0·44). In those not on invasive mechanical ventilation at baseline, there was no significant difference in the proportion meeting the composite endpoint of invasive mechanical ventilation or death (1344 [25%] in the colchicine group vs 1343 [25%] patients in the usual care group; risk ratio 1·02 [95% CI 0·96 to 1·09]; p=0·47). Interpretation In adults hospitalised with COVID-19, colchicine was not associated with reductions in 28-day mortality, duration of hospital stay, or risk of progressing to invasive mechanical ventilation or death. Funding UK Research and Innovation (Medical Research Council), National Institute of Health Research, and Wellcome Trust.


INTRODUCTION 57
Inflammation is a key feature of severe COVID-19. Markedly raised levels of inflammatory 58 markers such as C-reactive protein (CRP), ferritin, interleukin-6 (IL-6) and other cytokines 59 are observed in severe cases and are associated with poor outcomes. 1-5 Inflammation is 60 statistical analysis plan, and additional information are available on the study website 100 www.recoverytrial.net. 101 Patients admitted to hospital were eligible for the study if they had clinically suspected or 102 laboratory confirmed SARS-CoV-2 infection and no medical history that might, in the 103 opinion of the attending clinician, put the patient at significant risk if they were to 104 participate in the trial. Children and pregnant women were not eligible for randomisation 105 to colchicine. Patients with severe liver impairment, significant cytopaenia, concomitant 106 use of strong CYP3A4 or P-glycoprotein inhibitors, or hypersensitivity to lactose were 107 excluded (further details in appendix p 80). Written informed consent was obtained from 108 all patients, or a legal representative if patients were too unwell or unable to provide 109 consent. 110

Randomisation and masking 111
Baseline data were collected using a web-based case report form that included 112 demographics, level of respiratory support, major comorbidities, suitability of the study 113 treatment for a particular patient, and treatment availability at the study site (appendix pp 114 32-34). Eligible and consenting, non-pregnant adult patients were assigned in a 1:1 ratio 115 to either usual standard of care or usual standard of care plus colchicine or one of the 116 other available RECOVERY treatment arms using web-based simple (unstratified) 117 randomisation with allocation concealed until after randomisation (appendix pp 30-31). 118 For some patients, colchicine was unavailable at the hospital at the time of enrolment or 119 was considered by the managing physician to be either definitely indicated or definitely 120 contraindicated. These patients were excluded from the randomised comparison between 121 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 18, 2021. ; https://doi.org/10.1101/2021.05.18.21257267 doi: medRxiv preprint colchicine and usual care. Patients allocated to colchicine were to receive 1 mg after 122 randomisation followed by 500 mcg 12 hours later and then 500 mcg twice daily by mouth 123 or nasogastric tube for 10 days in total or until discharge, whichever occurred earlier. 124 Dose frequency was halved for patients receiving a moderate CYP3A4 inhibitor or who 125 had renal impairment (estimated glomerular filtration rate <30 ml/min/1.73m 2 ) or 126 estimated body weight <70 kg (appendix p 80). 127 As a platform trial, and in a factorial design, patients could be simultaneously randomised 128 to other treatment groups: i) convalescent plasma versus monoclonal antibody (REGN-129 CoV2) versus usual care, ii) aspirin versus usual care, and iii) baricitinib versus usual care 130 (appendix pp 31). Until 24 January 2021, the trial also allowed a subsequent 131 randomisation for patients with progressive COVID-19 (evidence of hypoxia and a hyper-132 inflammatory state) to tocilizumab versus usual care. Participants and local study staff 133 were not masked to the allocated treatment. The trial steering committee, investigators, 134 and all other individuals involved in the trial were masked to outcome data during the trial. 135 136

Procedures 137
A single online follow-up form was completed when participants were discharged, had 138 died or at 28 days after randomisation, whichever occurred earliest (appendix pp 35-41). 139 Information was recorded on adherence to allocated study treatment, receipt of other 140 COVID-19 treatments, duration of admission, receipt of respiratory or renal support, and 141 vital status (including cause of death). In addition, in the UK, routine healthcare and 142 registry data were obtained including information on vital status (with date and cause of 143 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 18, 2021. ; https://doi.org/10.1101/2021.05.18.21257267 doi: medRxiv preprint death), discharge from hospital, receipt of respiratory support, or renal replacement 144 therapy. 145

Outcomes 146
Outcomes were assessed at 28 days after randomisation, with further analyses specified 147 at 6 months. The primary outcome was all-cause mortality. Secondary outcomes were 148 time to discharge from hospital, and, among patients not on invasive mechanical 149 ventilation at randomisation, invasive mechanical ventilation (including extra-corporal 150 membrane oxygenation) or death. Prespecified subsidiary clinical outcomes were use of 151 non-invasive respiratory support, time to successful cessation of invasive mechanical 152 ventilation (defined as cessation of invasive mechanical ventilation within, and survival to, 153 28 days), use of renal dialysis or haemofiltration, cause-specific mortality, bleeding 154 events, thrombotic events, and major cardiac arrhythmias. Information on suspected 155 serious adverse reactions was collected in an expedited fashion to comply with regulatory 156 requirements. 157

Statistical Analysis 158
The primary analysis for all outcomes was by intention-to-treat comparing patients 159 randomised to colchicine with patients randomised to usual care but for whom colchicine 160 was both available and suitable as a treatment. For the primary outcome of 28-day 161 mortality, the log-rank observed minus expected statistic and its variance were used to 162 both test the null hypothesis of equal survival curves (i.e., the log-rank test) and to 163 calculate the one-step estimate of the average mortality rate ratio. We constructed 164 Kaplan-Meier survival curves to display cumulative mortality over the 28-day period. We 165 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 18, 2021. ; https://doi.org/10.1101/2021.05.18.21257267 doi: medRxiv preprint used the same method to analyse time to hospital discharge and successful cessation of 166 invasive mechanical ventilation, with patients who died in hospital right-censored on day 167 29. Median time to discharge was derived from Kaplan-Meier estimates. For the pre-168 specified composite secondary outcome of progression to invasive mechanical ventilation 169 or death within 28 days (among those not receiving invasive mechanical ventilation at 170 randomisation), and the subsidiary clinical outcomes of receipt of ventilation and use of 171 haemodialysis or haemofiltration, the precise dates were not available and so the risk 172 ratio was estimated instead. 173 Prespecified analyses were performed for the primary outcome using the statistical test 174 of interaction (test for heterogeneity or trend), in accordance with the prespecified 175 analysis plan, defined by characteristics at randomisation: age, sex, ethnicity, level of 176 respiratory support, days since symptom onset, and use of corticosteroids (appendix p 177

114). 178
Estimates of rate and risk ratios are shown with 95% confidence intervals. All p-values 179 are 2-sided and are shown without adjustment for multiple testing. The full database is 180 held by the study team which collected the data from study sites and performed the 181 analyses at the Nuffield Department of Population Health, University of Oxford (Oxford, 182

UK). 183
As stated in the protocol, appropriate sample sizes could not be estimated when the trial 184 was being planned at the start of the COVID-19 pandemic (appendix p 54). As the trial 185 progressed, the trial steering committee, whose members were unaware of the results of 186 the trial comparisons, determined that sufficient patients should be enrolled to provide at 187 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 18, 2021. ; https://doi.org/10.1101/2021.05.18.21257267 doi: medRxiv preprint least 90% power at a two-sided significance level of 0.01 to detect a clinically relevant 188 proportional reduction in 28-day mortality of 12.5% between the two groups.  Analyses were performed using SAS version 9.4 and R version 3.4. The trial is registered 196 with ISRCTN (50189673) and clinicaltrials.gov (NCT04381936). 197

Role of the funding source 198
The funder of the study had no role in study design, data collection, data analysis, data 199 interpretation, or writing of the report. The corresponding authors had full access to all the 200 data in the study and had final responsibility for the decision to submit for publication. into the RECOVERY trial were eligible to be randomly allocated to colchicine (i.e. 205 colchicine was available in the hospital at the time and the attending clinician was of the 206 opinion that the patient had no known indication for or contraindication to colchicine, figure  207 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 18, 2021. ; https://doi.org/10.1101/2021.05.18.21257267 doi: medRxiv preprint 1). 5610 patients were randomly allocated to colchicine and 5730 were randomly 208 allocated to usual care (36 patients were randomised outside of UK). The mean age of 209 study participants in this comparison was 63.4 years (SD 13.8) and the median time since 210 symptom onset was 9 days (IQR 6 to 12 days) (webtable 1). At randomisation, 10603 211 (94%) of patients were receiving corticosteroids. 212 The follow-up form was completed for 5510 (98%) in the colchicine group and 5605 (98%) 213 in the usual care group. Among patients with a completed follow-up form, 5122 (93%) 214 allocated to colchicine received at least one dose (figure 1; webtable 2). The median 215 duration of treatment with colchicine was 6 days (IQR 3-9 days). Use of other treatments 216 for COVID-19 was similar among patients allocated colchicine and among those allocated 217 usual care, with 87% receiving a corticosteroid, about one-quarter receiving remdesivir, 218 and one-eighth receiving tocilizumab (webtable 2). 219 Primary and secondary outcome data are known for >99% of randomly assigned patients. 220 There was no significant difference in the proportion of patients who met the primary 221 outcome of 28-day mortality between the two randomised groups (1173 [21%] patients in 222 the colchicine group vs. 1190 (21%) patients in the usual care group; rate ratio 1·01; 95% 223 confidence interval [CI], 0·93 to 1·10; p=0·77; figure 2). We observed similar results 224 across all pre-specified sub-groups (figure 3). In an exploratory analysis restricted to the 225 11009 (97%) patients with a positive SARS-CoV-2 test result, the result was virtually 226 identical (rate ratio 1·02, 95% CI 0·94 to 1·10; p=0·70). 227 The median time to discharge from hospital alive was 10 days (IQR 5 to >28) in both 228 groups and there was no significant difference in the probability of being discharged alive 229 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 18, 2021. ; https://doi.org/10.1101/2021.05.18.21257267 doi: medRxiv preprint within 28 days (70% vs. 70%, rate ratio 0·98, 95% CI 0·94 to 1·03, p=0·44) (table 2). 230 Among those not on invasive mechanical ventilation at baseline, the number of patients 231 progressing to the pre-specified composite secondary outcome of invasive mechanical 232 ventilation or death was similar in both groups (25% vs. 25%, risk ratio 1·02, 95% CI 0·96 233 to 1·09, p=0·47). Similar results were seen in all pre-specified subgroups of patients 234 (webfigure 1, webfigure 2). 235 We found no significant differences in the prespecified subsidiary clinical outcomes of 236 cause-specific mortality (webtable 3), use of ventilation, successful cessation of invasive 237 mechanical ventilation, or need for renal dialysis or haemofiltration (table 2). The 238 incidence of new cardiac arrhythmias, bleeding events, and thrombotic events was also 239 similar in the two groups (webtable 4). There were two reports of a serious adverse 240 reaction believed related to colchicine: one case of severe acute kidney injury and one 241 case of rhabdomyolysis. 242

DISCUSSION 244
In this large, randomised trial involving over 11,000 patients from 3 countries and over 245 2000 deaths, allocation to colchicine was not associated with reductions in mortality, 246 duration of hospitalisation or the risk of being ventilated or dying for those not on 247 ventilation at baseline. These results were consistent across the prespecified subgroups 248 of age, sex, ethnicity, duration of symptoms prior to randomisation, level of respiratory 249 support at randomisation, and use of corticosteroids. 250 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 18, 2021. ; https://doi.org/10.1101/2021.05.18.21257267 doi: medRxiv preprint The benefit of dexamethasone in COVID-19 patients requiring respiratory support 251 demonstrates the importance of inflammation in this patient group and colchicine was 252 proposed as a treatment for COVID-19 based on its anti-inflammatory activity. 25 The lack 253 of evidence of benefit from colchicine in this large well-powered trial suggests that the 254 anti-inflammatory properties of colchicine are either insufficient to produce a meaningful 255 reduction in mortality risk or are not affecting the relevant inflammatory pathways in 256 moderate to severe COVID-19. Whilst the majority of patients in this study received 257 concomitant corticosteroid therapy, we saw no evidence that colchicine was beneficial in 258 those patients not receiving a corticosteroid. 259 Strengths of this trial included that it was randomised, had a large sample size, broad 260 eligibility criteria, was international and more than 99% of patients were followed up for 261 the primary outcome. However, detailed information on laboratory markers of 262 inflammation and immune response was not collected, nor was information on 263 radiological or physiological outcomes. Although this randomised trial is open label (i.e., 264 participants and local hospital staff are aware of the assigned treatment), the outcomes 265 are unambiguous and were ascertained without bias through linkage to routine health 266

records. 267
Three other randomised controlled trials have assessed the efficacy of colchicine for the 268 treatment of COVID-19 in hospitalised patients. 18-20 A two day shorter duration of 269 hospitalisation was reported in a trial of 100 patients with laboratory confirmed SARS-270 CoV-2 infection and pulmonary involvement on computed tomography who were 271 randomized to either hydroxychloroquine plus colchicine or hydroxychloroquine plus 272 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. the RECOVERY trial, with more than 11,000 participants and more than 2,000 deaths, 280 had excellent power to detect modest treatment benefits; none were observed. 281 The RECOVERY trial only studied patients who had been hospitalised with COVID-19 282 and, therefore, is not able to provide any evidence on the safety and efficacy of colchicine  using the search terms ("SARS-CoV-2.mp" OR "COVID.mp" OR "COVID-19.mp" OR 299 "2019-nCoV.mp" OR "Coronavirus.mp" OR "Coronavirinae/") AND ("colchicine.mp" OR 300 "colchicine/") in any language, using validated filters to select for randomised controlled 301 trials. 302

303
We identified three relevant randomised trials that compared colchicine with usual care 304 or placebo in hospitalised patients with COVID-19 (two at low risk of bias and one with 305 some concerns due to limited information on randomisation process and lack of clarity 306 about blinding in the study). Each trial suggested a potential favourable impact of 307 colchicine on outcome measures of clinical improvement or duration of hospitalisation, 308 The three trials combined included a total of 285 patients and 7 deaths, so even combined 309 were not adequately powered to detect an effect on mortality. 310 311

Added value of this study: 312
The Randomised Evaluation of COVID-19 therapy (RECOVERY) trial is the first large, 313 randomised trial to report results of the effect of colchicine in patients hospitalised with 314 COVID-19. We found no significant effect of colchicine vs. usual care alone on 28-day 315 mortality, the probability of discharge alive within 28 days, or, among patients who were 316 not receiving invasive mechanical ventilation at randomisation, the probability of 317 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Wittes. 352

Declaration of interests 353
The authors have no conflict of interest or financial relationships relevant to the submitted 354 work to disclose. No form of payment was given to anyone to produce the manuscript. All Oxford has a staff policy of not accepting honoraria or consultancy fees directly or 358 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 18, 2021. study materials are available online at www.recoverytrial.net. As described in the protocol, 364 the trial Steering Committee will facilitate the use of the study data and approval will not 365 be unreasonably withheld. Deidentified participant data will be made available to bona 366 fide researchers registered with an appropriate institution within 3 months of publication. 367 However, the Steering Committee will need to be satisfied that any proposed publication 368 is of high quality, honours the commitments made to the study participants in the consent 369 documentation and ethical approvals, and is compliant with relevant legal and regulatory 370 requirements (e.g. relating to data protection and privacy). The Steering Committee will 371 have the right to review and comment on any draft manuscripts prior to publication. Data 372 will be made available in line with the policy and procedures described at: 373 https://www.ndph.ox.ac.uk/data-access. Those  Above all, we would like to thank the thousands of patients who participated in this trial. 380 We would also like to thank the many doctors, nurses, pharmacists, other allied health The views expressed in this publication are those of the authors and not necessarily those 406 of the NHS, the NIHR, or the UK Department of Health and Social Care. 407

Conflicts of interest 408
The authors have no conflict of interest or financial relationships relevant to the submitted 409 work to disclose. No form of payment was given to anyone to produce the manuscript. All CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted May 18, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

514
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)  Data are n (%) or n/N (%), unless otherwise indicated. RR=rate ratio for the outcomes of 28-day mortality, hospital discharge and successful cessation of invasive mechanical ventilation, and risk ratio for other outcomes. * Analyses exclude those on invasive mechanical ventilation at randomisation. † Analyses exclude those on any form of ventilation at randomisation. ‡ Analyses restricted to those on invasive mechanical ventilation at randomisation. § Analyses exclude those on haemodialysis or haemofiltration at randomisation. 516 517 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)  Figure 2: Effect of allocation to colchicine on 28−day mortality . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 18, 2021. ; https://doi.org/10.1101/2021.05.18.21257267 doi: medRxiv preprint