key: cord-1021451-cfbj6ize
authors: Ramakrishnan, S.; Nicolau, D. V.; Langford, B.; Mahdi, M.; Jeffers, H.; Mwasuku, C.; Krassowska, K.; Fox, R.; Binnian, I.; Glover, V.; Bright, S.; Butler, C.; Cane, J. L.; Halner, A.; Matthews, P. C.; Donnelly, L. E.; Simpson, J. L.; Baker, J. R.; Fadai, N. T.; Peterson, S.; Bengtsson, T.; Barnes, P. J.; Russell, R. E. K.; Bafadhel, M.
title: Inhaled budesonide in the treatment of early COVID-19 illness: a randomised controlled trial
date: 2021-02-08
journal: nan
DOI: 10.1101/2021.02.04.21251134
sha: 3f084f31018e9fb4824cb6e423105e6dee5938a0
doc_id: 1021451
cord_uid: cfbj6ize

Background Multiple early hospital cohorts of coronavirus disease 2019 (COVID-19) showed that patients with chronic respiratory disease were significantly under-represented. We hypothesised that the widespread use of inhaled glucocorticoids was responsible for this finding and tested if inhaled glucorticoids would be an effective treatment for early COVID-19 illness. Methods We conducted a randomised, open label trial of inhaled budesonide, compared to usual care, in adults within 7 days of the onset of mild Covid-19 symptoms. The primary end point was COVID-19-related urgent care visit, emergency department assessment or hospitalisation. The trial was stopped early after independent statistical review concluded that study outcome would not change with further participant enrolment. Results 146 patients underwent randomisation. For the per protocol population (n=139), the primary outcome occurred in 10 participants and 1 participant in the usual care and budesonide arms respectively (difference in proportion 0.131, p=0.004). The number needed to treat with inhaled budesonide to reduce COVID-19 deterioration was 8. Clinical recovery was 1 day shorter in the budesonide arm compared to the usual care arm (median of 7 days versus 8 days respectively, logrank test p=0.007). Proportion of days with a fever and proportion of participants with at least 1 day of fever was lower in the budesonide arm. Fewer participants randomised to budesonide had persistent symptoms at day 14 and day 28 compared to participants receiving usual care. Conclusion Early administration of inhaled budesonide reduced the likelihood of needing urgent medical care and reduced time to recovery following early COVID-19 infection.

Evidence before this study

The majority of interventions studied for the COVID-19 pandemic are focused on hospitalised patients. Widely available and broadly relevant interventions for mild COVID-19 are urgently needed.

In this open label randomised controlled trial, inhaled budesonide, when given to adults with early COVID-19 illness, reduces the likelihood of requiring urgent care, emergency department consultation or hospitalisation. There was also a quicker resolution of fever, a known poor prognostic marker in COVID-19 and a faster self-reported and questionnaire reported symptom resolution. There were fewer participants with persistent COVID-19 symptoms at 14 and 28 days after budesonide therapy compared to usual care.

The STOIC trial potentially provides the first easily accessible effective intervention in early COVID-19. By assessing health care resource utilisation, the study provides an exciting option to help with the worldwide pressure on health care systems due to the COVID-19 pandemic. Data from this study also suggests a potentially effective treatment to prevent the long term morbidity from persistent COVID-19 symptoms.

Introduction COVID-19 is the most serious pandemic for over 100 years and with increasing mortality and morbidity worldwide. Other than age, obesity, and gender 1,2 , there are no clear predictors to forecast who will deteriorate needing hospital-based care. The onset of COVID-19 illness is almost always mild 3 giving a potential window to intervene prior to the development of severe disease 1,2 .To date, almost all studies have focussed on investigating and treating severe and hospitalised COVID-19 infection 4 . However, there is currently little knowledge on therapeutic targets in early COVID-19 infection to prevent progression and clinical deterioration.

In early reports describing COVID-19 infection from China 1,2 , Italy 5 and the United States 6 , there was a significant under representation of patients with asthma and chronic obstructive pulmonary disease (COPD) in patients hospitalised with COVID-19. We hypothesized that this may be due to the widespread use of inhaled glucocorticoids in these patients 7 . Furthermore, the main indication for the use of inhaled glucocorticoids in patients with asthma and COPD is to reduce exacerbations, which are recognised to be often viral in etiology 8, 9 . In-vitro studies have shown that inhaled glucocorticoids reduce the replication of SARS-CoV-2 in airway epithelial cells 10 ; in addition to the downregulation of expression of angiotensin converting enzyme-2 (ACE2) and transmembrane protease serine 2 (TMPRSS2) genes which are critical for viral cell entry 11 .

Here we present the analysis of the Steroids in COVID-19 (STOIC) trial, a phase 2 trial designed to evaluate the efficacy of inhaled budesonide in individuals with early COVID-19 illness in the community. We examined the effect of inhaled budesonide on likelihood of urgent care or hospitalisation, clinical recovery, temperature and oxygenation. We also evaluated the effect of inhaled budesonide on SARS-CoV-2 viral load. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint 

The STOIC trial is a randomised, open label, parallel group phase 2 clinical trial conducted in the community in Oxfordshire, United Kingdom. Adults over the age of 18, with symptoms suggestive of COVID-19 within 7 days were eligible for inclusion. Participants were excluded if they had a known allergy or contraindication to the interventional medicinal product, or if they had recent use (within 7days) of inhaled or systemic glucocorticoids. Recruitment for the study was via local primary care networks, local COVID-19 testing sites and via multi-channel advertising. Volunteers were able to contact the study staff using the advertised phone numbers or via e-mail and all participant information was publicly available on the study website (www.stoic.ndm.ox.ac.uk). The trial was registered on clinicaltrials.gov (NCT04416399).

Participants who met the inclusion criteria were randomised to usual care (UC) or intervention with budesonide (BUD) dry powder inhaler (Pulmicort Turbuhaler, AstraZeneca) at a dose of 800µg twice a day. Participants were seen at their homes at randomisation (day 0), day 7 and day 14 by a trained respiratory research nurse to obtain written informed consent, provide inhalers and to obtain nasopharyngeal swabs for SARS-CoV-2 reverse transcription polymerase chain reaction (RT-PCR) testing (see supplementary methods for further details). Each participant received a paper symptom diary, calibrated pulse oximeter and thermometer for daily home monitoring. All participants were telephone contacted daily to record oxygen saturations, temperature and assessed for any adverse events by the study team. Participants allocated BUD were asked to stop taking the inhaler when they felt they had recovered (self-reported symptom recovery) or if they hit primary outcome; and all participants ceased daily monitoring (including daily telephone calls) when symptoms had recovered (again self-reported symptom recovery) or if the primary outcome was achieved. Finally, at day 28, all study participants were seen in the trial centre and serum SARS-CoV-2 antibody testing was performed.

The primary outcome was defined as a COVID-19-related Urgent Care visit, Emergency Department assessment or hospitalisation. During the pandemic, members of the public in the United Kingdom, were encouraged to contact a government telephone advice line prior to attending the emergency department and COVID-19 specific general practice hubs were created for patients who were deteriorating at home, for medical treatment including transfer to hospital. Secondary outcomes included clinical recovery as defined by self-reported time to symptom resolution; viral symptoms measured by the Common Cold Questionnaire (CCQ) 12 and the InFLUenza Patient-Reported Outcome (FluPRO®) 13 questionnaire; blood oxygen saturations and body temperature; and SARS-CoV-2 viral load.

The statistical packages R (R Core Team (2020). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/), Gauss ((https://www.aptech.com/), and SAS v9.4 (https://www.sas.com/en_us/home.html) are used. Descriptive statistics are used to describe variables between the groups in the BUD arm and the UC arm. Appropriate parametric or nonparametric statistical tests are performed. For continuous variables, the difference between treatments in the means or medians and the corresponding 95% confidence interval are is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint reported. For continuous variables, ANCOVA models (t-tests) adjusted for stratification factors or Wilcoxon rank sum tests are applied to compare the BUD and UC arm. For categorical variables the number (and percentage) of patients in each category are reported for each treatment group and chi-squared tests will be used for comparing treatment arms. Time to self-reported clinical recovery and FLUPro® symptom recovery are analysed using the Kaplan Meier method and presented as the median time to event. Comparison is performed with a log rank test where participants with primary outcome are censored at Day 28 when not meeting the event. Participants are randomly allocated to UC or BUD, stratified for participant age,(≤40 years/ >40 years) gender, and number of co-morbidities (≤1/ ≥2). The randomisation sequence was created using a random number generation function and allocation to each arm was performed through block randomisation in a 1:1 ratio. All tests are performed at a 5% 2sided significance level and all comparative outcomes are presented as summary statistics with 95% confidence intervals and reported in accordance with the CONSORT Statement. Missing data from study visits and daily monitoring are handled by last observation carried forward (LOCF) for temperature, oxygen saturations and time to FLUPro® symptom resolution. For FLUPro® total score and individual domain time-series plots, missing data was handled by imputation of zero score when self-reported symptom resolution occurred or LOCF when this was not available. Stochastic simulations of a "virtual" trial with the same study design, primary endpoint and duration, and community detection are presented in full in the supplementary materials. P-values are reported to a minimum of 3 decimal places. Further full details are available in the supplementary material.

At study inception in March 2020 and using published data available at the time 1,2 , we assumed that 20% of all COVID-19 illness is severe and would require hospitalisation. Using 80% power at 0.05 level, we required 199 patients in each arm to demonstrate a 50% reduction of urgent care visits or hospitalisations. The primary outcome was analysed for both the perprotocol (PP) and intention to treat (ITT) population. The PP population is defined as the population that received the study treatment and had at least 1 day of study observations; the ITT population is defined as all participants that were randomised to a study arm.

The trial was sponsored by the University of Oxford, and was approved by the Fulham London Research Ethics Committee (20/HRA/2531) and the National Health Research Authority. All participants provided written informed consent. The study team requested an independent statistical monitoring committee review on the 9 th of December 2020 due to reduced recruitment after the second national lockdown in England; implementation of the COVID-19 vaccine; and ethical consideration of the primary outcome. A priori stop criteria were used to determine futility of further recruitment (see statistical analysis plan).

The study was funded by the Oxford NIHR Biomedical Research Centre and AstraZeneca (Gothenburg, Sweden). The funders had no role in study design, data collection, data analysis and decision to publish. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint 

From July to December 2020, 146 participants were randomised, of which 139 were in the PP analysis, with 70 and 69 in the BUD and UC arm respectively (see figure 1) . Participant characteristics were similar between the study arms, as shown in table 1 (see supplementary  table 1 for ITT population) . COVID-19 infection occurred in 94% measured by RT-PCR. The median duration of symptoms prior to randomisation was three days. The median time to symptom resolution was eight days. Inhaled budesonide was taken for a median (IQR) duration of 7 (4 to 10.5) days.

Simulations using bootstrap was performed to determine the conditional power for an evaluation of an early stop, using the a priori decisions described in the supplement at approximately 146 patients. Estimated power was >99% using both the total population (N=124) and, at the time of the simulation, the known subgroup of COVID-19 positive patients (N=78).

In the PP analysis, the primary outcome occurred in 10 participants in the UC arm versus 1 participant in BUD arm (difference in proportions 0.131, 95% CI (0.043, 0.218), p=0.004) indicating a relative risk reduction of 90% for BUD. The number needed to treat with inhaled budesonide to reduce COVID-19 related urgent care/hospitalisation was 8. Sensitivity analysis in participants with confirmed COVID-19 showed that the difference in proportions was 0.125, 95% CI (0.035, 0.216), p=0.007. There was no difference in participants with a primary outcome event compared to those without (see supplementary table 2) . For the ITT population, the primary outcome occurred in 11 participants in the UC arm and 2 participants in the BUD arm (difference in proportion 0.123, p=0.009).

Self-reported clinical recovery was 1 day quicker with BUD compared to UC (median of 7 days versus 8 days, logrank test p=0.007, see figure 2 ). The mean time to recovery, in days was 8 and 11 in the BUD and UC arm respectively. Further sensitivity analysis for clinical recovery in participants with confirmed COVID-19 status showed a similar median times to recovery (7 vs. 8 days, p=0.039) (see supplementary figure 1 ). At day 14, self-reported symptoms were present in 10% (n=7) of participants randomised to BUD compared to 30% (n=21) of participants randomised to UC (difference in proportion 0.204, p=0.003).

The mean proportion of days with a documented fever (≥37.5 C) during the first 14 days, was 2.1% in the BUD and 7.7% in the UC arms (Wilcoxon test, p= 0.051). The percentage of participants with at least 1 day of fever was 11% and 23% in the BUD and UC arm respectively (Fishers test, p=0.076). Violin plots, showing distribution of pooled highest temperatures are presented in figure 3 , with a statistically higher mean in the UC arm (mean difference 0.49, 95%CI 0.32 to 0.66, p<0.001). Temperature plots relative to the day of randomisation showed that temperature fell quicker in the BUD compared to UC arm (see supplementary figure 2 ). The median (IQR) proportion of total days that participants required as needed anti-pyretics (paracetamol, aspirin, ibuprofen) in the BUD and UC arms was 27% (0 to 50) and 50% (15 to 71) respectively (Mann-Whitney test, p=0.025).

Symptom resolution at day 14, as defined by the FLUPro® user manual, occurred in 82% and 72% of the BUD and UC arms respectively (p=0.166); whilst the median time to symptom . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101/2021.02.04.21251134 doi: medRxiv preprint resolution as measured by the FLUPro® was 3 and 4 days in the BUD and UC arms respectively (logrank test p=0.080, see supplementary figure 3 ). The mean change (95%CI) in FLUPro® total score between day 0 and day 14 in the BUD and UC are -0.65 (-0.80 to -0.50) and -0,54 (-0.69 to -0.40) respectively (mean difference of -0.10, 95%CI of the difference -0.21 to -0.00, p=0.044). Figure 4 (panels a-g) shows the mean daily FLUPro® scores for the total symptom burden and individual domains. The mean change of the FLUPro® domains showed that systemic symptoms were significantly greater in BUD compared to UC (supplementary table 3 ). The mean change in CCQ total score between day 0 and day 14 in the BUD and UC was -0.49 (-0.63 to -0.35) and -0.37 (-0.51, -0.24) respectively (mean difference of -0.12 (-0.21 to -0.02), p=0.016). The common cold questionnaire (CCQ) symptom daily mean score is presented as supplementary figure 4.

The proportion of days with oxygen saturations ≤94%, during the first 14 days, was 19% and 22% in the BUD and UC arm respectively (Wilcoxon test p=0.627). During the first 14 days 59% and 58% in the BUD and UC arms had at least one day with oxygen saturations ≤94%. 

Stochastic simulations, in a 'virtual twin' study design, demonstrated that the daily odds ratio of reaching the primary outcome, with BUD reduced by 3000% (see figure 5 and supplementary materials). is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint 

We have demonstrated that the inhaled glucocorticoid, budesonide, given for a short duration, may be an effective treatment of early COVID-19 disease in adults. This effect, with a relative reduction of 90% of clinical deterioration is equivalent to the efficacy seen following the use of COVID-19 vaccines 14, 15 and greater than that reported in any treatments used in hospitalised and severe COVID-19 patients [16] [17] [18] . The broad inclusion criteria make this study intervention relevant to health care systems worldwide. Inhaled budesonide is a simple, safe, well studied, inexpensive and widely available treatment. The number of participants needed to treat to prevent increased health care resource utilization is 8, and combined with the short treatment period required to achieve benefit, makes this potentially an affordable and scalable intervention for early COVID-19. This is especially significant in low-and middle-income countries where the majority of currently approved COVID-19 treatments are unlikely to ever reach patients as a consequence of variable healthcare systems 19 . For example, although dexamethasone is a widely available and cheap medicine, and is effective in reducing mortality in severe and intensive care related COVID-19, this is unfortunately largely irrelevant in countries which have limited intensive care or hospital capacity 20, 21 . Furthermore, in high income countries, inhaled budesonide could work as an adjunct to reduce pressure on health care systems until widespread SARS-CoV-2 vaccination can be achieved. Unlike with vaccines, the efficacy of inhaled budesonide is unlikely to be affected by any emergent SARS-CoV-2 variant 22 .

We selected this treatment intervention due to the unexpected observation of an underrepresentation of patients with asthma and COPD in hospitalised cohorts of COVID-19 23 . This finding from early hospitalised cohorts in Wuhan 1,2 was at odds with prior respiratory viral pandemics, such as H1N1 influenza 24 . The common therapy between these lung diseases is inhaled glucocorticoids, either as a mono-, dual-or triple-constituent. Furthermore, inhaled glucocorticoids are among the most prescribed medicine of any class around the world, listed by the World Health Organisation (WHO) of essential medicines 25 . Moreover, evidence of the utility of inhaled glucocorticoids, in reducing viral exacerbations of asthma have been known for many decades 9 , whilst specifically, inhaled budesonide has shown effect at reducing rhinovirus replication in-vitro 26 . The efficacy of Dexamethasone in RECOVERY 17 , for severe disease also supports our findings, whilst there is plausibility that the immune-modulatory effect of inhaled glucocorticoids 27 , may also apply to any future viral epidemics. In our study, we found that inhaled budesonide also demonstrated benefit in the secondary outcomes, with quicker symptom resolution in those treated with budesonide either measured using a selfreport of symptom recovery, or defined as normalisation of prospectively collected symptom scores measured using the FluPRO® 13 or the CCQ 12 . The positive impact on temperature when used to treat early COVID-19 is further evidence that inhaled budesonide is modifying the disease process. Of note, there was a significantly greater population of participants randomised to budesonide who were free of symptoms at 28 days compared to participants randomized to usual care. In the United Kingdom, up to 20% of patients 28 report persistent symptoms 5 weeks after COVID-19. This interesting finding, also suggests that intervention with an inhaled glucocorticoids might impact on rate of the persistent long-term symptoms in COVID-19 ("long COVID"); and should be investigated further in view of the significant longterm health and economic impact of long COVID.

There are limitations to our study. This is an open-labelled study, which was stopped early due to the impact of the national pandemic control measures and national prioritisation rules for clinical research trials in the UK. Our power calculations were made from the best available predictions in early 2020. Therapeutic randomised clinical trial design and sample size calculations are often dictated by statistical assumptions with treatment effect estimations . CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101/2021.02.04.21251134 doi: medRxiv preprint based on the evidence of best available care. However, in trial design for a new disease, with no known effective treatment, statistical assumptions are thus arbitrary. In our study the event rate in the control arm was half of what was predicted, and this is consistent with subsequent primary care clinical trials performed in COVID-19 [29] [30] [31] . We found that the budesonide treatment effect size, was larger than predicted; and independent statistical assessment concluded that the current sample size and treatment effect had a 99% power to reject the null hypothesis. Furthermore, the positive concordance of temperature and symptoms as secondary outcomes gives us confidence in our results. Our study design involved randomisation at home, with home visits for study assessments and a daily contact until symptom resolution by the study team. To our knowledge, the STOIC trial is first to assess daily physiological measures in early COVID-19. The robust study design meant very good participant retention and completion of symptom diaries. Stopping a study early is unusual and is a decision which is not taken without due diligence 32 However, we ensured that a priori stop decision analysis was performed by an independent statistical team for statistical rigor. During these unprecedented times, our findings favouring budesonide, despite these limitations can not be ignored and requires urgent dissemination and validation; especially in the setting of a treatment that is relatively safe.

In conclusion, budesonide, a safe inhaled glucorticoid, appears to be an effective treatment for early COVID-19 infection which could be applicable to global healthcare systems. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint Data presented as mean (SD) or mean (range) unless otherwise stated; *at randomisation; ¥median (IQR); #median (range); $ mean (95%CI)

. CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101/2021.02.04.21251134 doi: medRxiv preprint (g)

. CC-BY-NC-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101/2021.02.04.21251134 doi: medRxiv preprint Figure 5 .

These results indicate that in order to observe our findings (shown by the dotted line), the daily treatment effect needed represents approximately 3000% (30x) reduction in the daily odds of reaching primary outcome. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted February 8, 2021. ; https://doi.org/10.1101/2021.02.04.21251134 doi: medRxiv preprint

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

Clinical Characteristics of Coronavirus Disease 2019 in China

The Incubation Period of Coronavirus Disease 2019 (COVID-19) From Publicly Reported Confirmed Cases: Estimation and Application

A real-time dashboard of clinical trials for COVID-19

COVID-19 in Severe Asthma Network in Italy (SANI) patients: Clinical features, impact of comorbidities and treatments

Characteristics and outcomes of COVID-19 patients in New York City's public hospital system

Inhaled corticosteroids in virus pandemics: a treatment for COVID-19?

Respiratory viruses in exacerbations of chronic obstructive pulmonary disease requiring hospitalisation: a case-control study

The Relationship of Rhinovirus-Associated Asthma Hospitalizations with Inhaled Corticosteroids and Smoking

The inhaled steroid ciclesonide blocks SARS-CoV-2 RNA replication by targeting the viral replication-transcription complex in cultured cells

COVID-19-related Genes in Sputum Cells in Asthma. Relationship to Demographic Features and Corticosteroids

Validity of the common cold questionnaire (CCQ) in asthma exacerbations

Performance of the inFLUenza Patient-Reported Outcome (FLU-PRO) diary in patients with influenza-like illness (ILI)

Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine

Safety and efficacy of the ChAdOx1 nCoV-19 vaccine (AZD1222) against SARS-CoV-2: an interim analysis of four randomised controlled trials in Brazil

Projected health-care resource needs for an effective response to COVID-19 in 73 low-income and middle-income countries: a modelling study

Availability of critical care resources to treat patients with severe sepsis or septic shock in Africa: a self-reported, continent-wide survey of anaesthesia providers

Intensive Care Unit Capacity in Low-Income Countries: A Systematic Review

Covid-19: New coronavirus variant is identified in UK

Risk factors for hospitalisation and poor outcome with pandemic A/H1N1 influenza: United Kingdom first wave

World Health Organisation. World Health Organization Model List of Essential Medicines, 21st List

Antiviral and anti-inflammatory activity of budesonide against human rhinovirus infection mediated via autophagy activation

Office of National Statistics United Kingdom. The prevalence of long COVID symptoms and COVID-19 complications

Fluvoxamine vs Placebo and Clinical Deterioration in Outpatients With Symptomatic COVID-19: A Randomized Clinical Trial

Stopping guidelines for an effectiveness trial: what should the protocol specify?

De-identified individual participant data and a data dictionary defining each field in the set, can be made available to others upon approval of a written request to the corresponding author. The request will be evaluated by a committee formed by a subset of co-authors to determine the research value. A data sharing agreement will be needed

Ramakrishnan reports grants and non-financial support from Oxford Respiratory NIHR BRC, during the conduct of the study; non-financial support from AstraZeneca, personal fees from Australian Government Research Training Program, outside the submitted work

Mrs Langford has nothing to disclose

Mahdi has nothing to disclose

Mrs Helen Jeffers reports personal fees from AstraZeneca, outside the submitted work

Butler reports grants from National Institute for Health Research (NIHR)

Cane has nothing to disclose

Halner has nothing to disclose

Matthews has nothing to disclose

Donnelly reports grants from AstraZeneca, from Boehringer-Ingelheim, outside the submitted work

Simpson has nothing to disclose

Dr Baker has nothing to disclose

Fadai has nothing to disclose

Peterson reports personal fees from AstraZeneca, outside the submitted work

Bengtsson reports personal fees from AstraZeneca, outside the submitted work

Barnes reports grants and personal fees from AstraZeneca, grants and personal fees from Boehringer Ingelheim, personal fees from Teva, personal fees from Covis, during the conduct of the study

Russell reports grants from AstraZeneca, personal fees from Boehringer Ingelheim, personal fees from Chiesi UK, personal fees from Glaxo-SmithKline, during the conduct of the study

Bafadhel reports grants from AstraZeneca, personal fees from AstraZeneca, Chiesi, GSK, other from Albus Health, ProAxsis, outside the submitted work

We are grteful for the support of Rhys Painton, Sophie Morgan, Nick Thomas, Jennifer Quint, Philip Maini, Helen Byrne, Sunil Lakhani, Ruairidh Battleday and Adriaan Bax. We would also llke to thank Investec PLC for generously supporting the trial with personal protective equipment during times of supply shortages. We thank the Department of Health and Social Care who were instrumental in supporting the promotion of STOIC at Oxfordshire testing sites and the Oxford Respiratory Trials Unit.