Azithromycin in Hospitalised Patients with COVID-19 (RECOVERY): a randomised, controlled, open-label, platform trial

Background: Azithromycin has been proposed as a treatment for COVID-19 on the basis of its immunomodulatory actions. We evaluated the efficacy and safety of azithromycin in hospitalised patients with COVID-19. Methods: In this randomised, controlled, open-label, adaptive platform trial, several possible treatments were compared with usual care in patients hospitalised with COVID-19 in the UK. Eligible and consenting patients were randomly allocated to either usual standard of care alone or usual standard of care plus azithromycin 500 mg once daily by mouth or intravenously for 10 days or until discharge (or one of the other treatment arms). Patients were twice as likely to be randomised to usual care as to any of the active treatment groups. The primary outcome was 28-day mortality. Findings: Between 7 April and 27 November 2020, 2582 patients were randomly allocated to receive azithromycin and 5182 patients to receive usual care alone. Overall, 496 (19%) patients allocated to azithromycin and 997 (19%) patients allocated to usual care died within 28 days (rate ratio 1.00; 95% confidence interval [CI] 0.90-1.12; p=0.99). Consistent results were seen in all pre-specified subgroups of patients. There was no difference in duration of hospitalisation (median 12 days vs. 13 days) or the proportion of patients discharged from hospital alive within 28 days (60% vs. 59%; rate ratio 1.03; 95% CI 0.97-1.10; p=0.29). Among those not on invasive mechanical ventilation at baseline, there was no difference in the proportion meeting the composite endpoint of invasive mechanical ventilation or death (21% vs. 22%; risk ratio 0.97; 95% CI 0.89-1.07; p=0.54). Interpretation: In patients hospitalised with COVID-19, azithromycin did not provide any clinical benefit. Azithromycin use in patients hospitalised with COVID-19 should be restricted to patients where there is a clear antimicrobial indication.


INTRODUCTION 62
usual care was twice that of any of the active arms the patient was eligible for (e.g. 2:1 127 in favour of usual care if the patient was eligible for only one active arm, 2:1:1 if the 128 patient was eligible for two active arms). For some patients, azithromycin was 129 unavailable at the hospital at the time of enrolment or if a macrolide antibiotic was 130 considered by the managing physician to be either definitely indicated or definitely 131 contraindicated. These patients were excluded from the randomised comparison 132 between azithromycin and usual care. Patients allocated to azithromycin were to 133 receive azithromycin 500 mg by mouth, nasogastric tube, or intravenous injection once 134 daily for 10 days or until discharge, if sooner. Allocated treatment was prescribed by the 135 managing doctor. Participants and local study staff were not masked to the allocated 136 treatment. The Steering Committee, investigators, and all others involved in the trial 137 were masked to the outcome data during the trial. 138

Procedures 139
A single online follow-up form was completed when participants were discharged, had 140 died or at 28 days after randomisation, whichever occurred earliest (appendix p 29-35). 141 Information was recorded on adherence to allocated study treatment, receipt of other 142 COVID-19 treatments, duration of admission, receipt of respiratory or renal support, and 143 vital status (including cause of death). In addition, routine healthcare and registry data 144 were obtained including information on vital status (with date and cause of death), 145 discharge from hospital, receipt of respiratory support, or renal replacement therapy. 146 Outcomes 147 . 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) preprint The copyright holder for this this version posted December 14, 2020. ; https://doi.org/10.1101/2020.12.10.20245944 doi: medRxiv preprint Outcomes were assessed at 28 days after randomisation, with further analyses 148 specified at 6 months. The primary outcome was all-cause mortality. Secondary 149 outcomes were time to discharge from hospital, and, among patients not on invasive 150 mechanical ventilation at randomisation, invasive mechanical ventilation (including 151 extra-corporal membrane oxygenation) or death. Prespecified subsidiary clinical 152 outcomes were cause-specific mortality, use of haemodialysis or haemofiltration, major 153 cardiac arrhythmia (recorded in a subset), and receipt and duration of ventilation. 154 Among those on invasive mechanical ventilation at randomisation, a subsidiary clinical 155 outcome of successful cessation of invasive mechanical ventilation was defined as 156 cessation within (and survival to) 28 days. Information on suspected serious adverse 157 reactions was collected in an expedited fashion to comply with regulatory requirements. 158

Statistical Analysis 159
An intention-to-treat comparison was conducted between patients randomised to 160 azithromycin and patients randomised to usual care but for whom azithromycin was 161 both available and suitable as a treatment. For the primary outcome of 28-day mortality, 162 the log-rank observed minus expected statistic and its variance were used to both test 163 the null hypothesis of equal survival curves (i.e., the log-rank test) and to calculate the 164 one-step estimate of the average mortality rate ratio. We constructed Kaplan-Meier 165 survival curves to display cumulative mortality over the 28-day period. The 2059 166 patients (27%) who had not been followed for 28 days and were not known to have died 167 by the time of the data cut for this preliminary analysis (30 November 2020) were either 168 censored on 30 November 2020 or, if they had already been discharged alive, were 169 right-censored for mortality at day 29 (that is, in the absence of any information to the 170 . 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) preprint The copyright holder for this this version posted December 14, 2020. ; https://doi.org/10.1101/2020.12.10.20245944 doi: medRxiv preprint contrary they were assumed to have survived 28 days). [Note: This censoring rule will 171 not be necessary for the final report.] We used similar methods to analyse time to 172 hospital discharge and successful cessation of invasive mechanical ventilation, with 173 patients who died in hospital right-censored on day 29. Median time to discharge was 174 derived from Kaplan-Meier estimates. For the pre-specified composite secondary 175 outcome of invasive mechanical ventilation or death within 28 days (among those not 176 receiving invasive mechanical ventilation at randomisation) and the subsidiary clinical 177 outcomes of receipt of ventilation and use of haemodialysis or haemofiltration, the 178 precise dates were not available and so the risk ratio was estimated instead. 179 Prespecified analyses of the primary outcome were performed separately in seven 180 subgroups defined by characteristics at randomisation: age, sex, ethnicity, level of 181 respiratory support, days since symptom onset, use of corticosteroids, and predicted 182 28-day mortality risk (appendix p 26). Observed effects within subgroup categories were 183 compared using a chi-squared test for heterogeneity or trend, in accordance with the 184 prespecified analysis plan. 185 Estimates of rate and risk ratios are shown with 95% confidence intervals. All p-values 186 are 2-sided and are shown without adjustment for multiple testing. The full database is 187 held by the study team which collected the data from study sites and performed the 188 analyses at the Nuffield Department of Population Health, University of Oxford (Oxford, 189

UK). 190
As stated in the protocol, appropriate sample sizes could not be estimated when the trial 191 was being planned at the start of the COVID-19 pandemic (appendix p 26). As the trial 192 . 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) preprint The copyright holder for this this version posted December 14, 2020. ; https://doi.org/10.1101/2020.12.10.20245944 doi: medRxiv preprint progressed, the trial steering committee, whose members were unaware of the results 193 of the trial comparisons, determined that sufficient patients should be enrolled to provide 194 at least 90% power at a two-sided p-value of 0.01 to detect a clinically relevant 195 proportional reduction in the primary outcome of 20% between the two groups.

Role of the funding source 202
The funder of the study had no role in study design, data collection, data analysis, data 203 interpretation, or writing of the report. The corresponding authors had full access to all 204 the data in the study and had final responsibility for the decision to submit for 205 publication. 206

208
Between 7 April 2020 and 27 November 2020, 9434 (57%) of 16443 patients enrolled 209 into the RECOVERY trial were eligible to be randomly allocated to azithromycin (that is 210 azithromycin was available in the hospital at the time and the attending clinician was of 211 the opinion that the patient had no known indication for or contraindication to 212 azithromycin, figure 1; appendix p 38). 2582 patients were randomly allocated to 213 . 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) preprint The copyright holder for this this version posted December 14, 2020. ; https://doi.org/10.1101/2020.12.10.20245944 doi: medRxiv preprint azithromycin and 5182 were randomly allocated to usual care, with the remainder being 214 randomly allocated to one of the other treatment arms. The mean age of study 215 participants in this comparison was 65·3 years (SD 15·7) and the median time since 216 symptom onset was 8 days (IQR 5 to 11 days) (table 1; appendix p 38). (rate ratio 0·99, 95% CI 0·88 to 1·10; p=0·81). 232 Allocation to azithromycin was associated with a similar time until discharge from 233 hospital alive as usual care (median 12 days vs. 13 days) and a similar probability of 234 discharge alive within 28 days (60% vs. 59%, rate ratio 1·03, 95% CI 0·97 to 1·10, 235 . 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) preprint The copyright holder for this this version posted December 14, 2020. ; https://doi.org/10.1101/2020.12.10.20245944 doi: medRxiv preprint p=0·29) (Table 2). Among those not on invasive mechanical ventilation at baseline, the 236 number of patients progressing to the pre-specified composite secondary outcome of 237 invasive mechanical ventilation or death among those allocated to azithromycin was 238 similar to that among those allocated to usual care (21% vs. 22%, risk ratio 0·97, 95% 239 CI 0·89 to 1·07, p=0·54). Allowing for multiple testing in interpretation of the results, 240 there was no evidence that the effect of allocation to azithromycin vs. usual care on time 241 until discharge from hospital alive or on invasive mechanical ventilation or death differed 242 between pre-specified subgroups of patients (appendix p 43-44). 243 We found no significant differences in the prespecified subsidiary clinical outcomes of 244 cause-specific mortality (appendix p 40), use of ventilation, successful cessation of 245 invasive mechanical ventilation, or need for renal dialysis or haemofiltration (Table 2). 246 We observed no significant differences in the frequency of new cardiac arrhythmias 247 (appendix p 41). There was one report of a serious adverse reaction believed related to 248 azithromycin: a case of pseudomembranous colitis from which the patient recovered 249 with standard treatment. 250

DISCUSSION 252
The results of this large randomised trial show that azithromycin is not an effective 253 treatment for patients hospitalised with COVID-19. Allocation to azithromycin was not 254 associated with reductions in mortality, duration of hospitalisation or the risk of being 255 ventilated or dying for those not on ventilation at baseline. These results were 256 consistent across the prespecified subgroups of age, sex, ethnicity, duration of 257 . 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) preprint The copyright holder for this this version posted December 14, 2020. ; https://doi.org/10.1101/2020.12.10.20245944 doi: medRxiv preprint symptoms prior to randomisation, level of respiratory support at randomisation, use of 258 corticosteroids, and baseline predicted risk of death at randomisation. 259 Azithromycin was proposed as a treatment for COVID-19 based on its 260 immunomodulatory activity. 7 Although no major organization or professional society has 261 was not collected, nor was information on radiological or physiological outcomes. This 295 initial report is based on complete follow-up for the primary outcome in 73% of patients 296 (and partial follow-up for the remaining 27%). However, collection of outcome 297 information both through case report forms and linkage to routine NHS records is 298 ongoing and, based on previous reports from this trial, will deliver complete follow-up 299 information for over 99% of patients by early January 2021. However, additional follow-300 up is unlikely to change the conclusion that azithromycin has no meaningful benefit for 301 hospitalised patients with COVID-19. 302 . 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. 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) preprint
The copyright holder for this this version posted December 14, 2020.

Declaration of interests 397
The authors have no conflict of interest or financial relationships relevant to the 398 submitted work to disclose. No form of payment was given to anyone to produce the study materials are available online at www.recoverytrial.net. As described in the 408 protocol, the trial Steering Committee will facilitate the use of the study data and 409 approval will not be unreasonably withheld. Deidentified participant data will be made 410 available to bona fide researchers registered with an appropriate institution within 3 411 months of publication. However, the Steering Committee will need to be satisfied that 412 any proposed publication is of high quality, honours the commitments made to the study 413 participants in the consent documentation and ethical approvals, and is compliant with 414 . 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) preprint
The copyright holder for this this version posted December 14, 2020. ; https://doi.org/10.1101/2020.12.10.20245944 doi: medRxiv preprint relevant legal and regulatory requirements (e.g. relating to data protection and privacy). 415 The Steering Committee will have the right to review and comment on any draft 416 manuscripts prior to publication. Data will be made available in line with the policy and 417 procedures described at: https://www.ndph.ox.ac.uk/data-access. Those  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) preprint The copyright holder for this this version posted December 14, 2020. ; https://doi.org/10.1101/2020.12.10.20245944 doi: medRxiv preprint

583
. 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) preprint The copyright holder for this this version posted December 14, 2020. ; https://doi.org/10.1101/2020.12.10.20245944 doi: medRxiv preprint 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. 585 586 . 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) preprint  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) preprint The copyright holder for this this version posted December 14, 2020. ; https://doi.org/10.1101/2020.12.10.20245944 doi: medRxiv preprint  Figure 2: Effect of allocation to azithromycin 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.  . 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) preprint The copyright holder for this this version posted December 14, 2020. ; https://doi.org/10.1101/2020.12.10.20245944 doi: medRxiv preprint