This was an investigator-initiated, multicenter, open-label, randomized trial. Adult patients aged 18 years and older and admitted to the hospital from 21 May 2020 for COVID-19, virologically confirmed using RT-PCR, were included. Patients were recruited from nine hospitals in Saudi Arabia (King Abdulaziz Hospital Al Ahsa, AlMadina General Hospital, Al-Qatif Central Hospital, Imam Abdulrahman Al Faisal Hospital Dammam, King Abdulaziz Medical City Jeddah, King Abdulaziz Hospital Makkah, King Abdulaziz Medical City Riyadh, Imam Abdulrahman Alfaisal Hospital Riyadh, and Al-Iman General Hospital Riyadh). Ethical approval was obtained from the Institutional Review Board (IRB) at the Ministry of National Guard-Health Affairs (MNGHA; study no. RC20/174/R) and Ministry of Health (MOH; log no. 20-83E). The trial was overseen by an independent data and safety monitoring board (DSMB). The trial was done according to the Declaration of Helsinki principles and the International Conference on Harmonization-Good Clinical Practice guidelines. The trial protocol and the statistical analysis plan were submitted for publication before the interim analysis .
Eligible patients were at least 18 years of age, male or non-pregnant females, and diagnosed with COVID-19 by RT-PCR-confirmed SARS-CoV-2. The patients were considered eligible if admitted to the hospital with moderate-to-severe COVID-19, defined as oxygen saturation (SaO2) of ≤ 94% while breathing ambient air or significant clinical symptoms with chest x-ray changes. As an eligibility criterion, patients were enrolled within 10 days of disease onset. Patients were excluded if they had a history of myocardial infarction or an irregular rhythm/QTc (corrected QT interval) in the baseline electrocardiogram (ECG) of > 490 ms. The exclusion criteria also included major comorbidities such as hematologic malignancy, advanced (stage 4–5) chronic kidney disease or dialysis therapy, severe liver damage (Child-Pugh score ≥ C, AST > five times the upper limit), or HIV. Written informed consent was obtained from all patients or their legal representatives if they could not provide consent.
Randomization was conducted using SAS v9.4 PROC PLAN to generate random allocation schedules. The randomization schedule was uploaded through an online web-based system in a 1:1 ratio to receive standard care (control group) or standard care plus the favipiravir and hydroxychloroquine combination (treatment group). It was stratified by enrolling center and by being admitted to ICU or on mechanical ventilation. The investigator disclosed the arm after all information regarding patient enrollment was recorded in the online system. The trial was open-label, and the delegated hospital pharmacy dispensed the study medications.
Daily assessment of the admitted patients and eligibility screening was done for possible study participants within the first 72 h of hospital admission. Once eligible, informed consent was obtained, and randomization was conducted. In addition to other standard medical treatments for COVID-19, per local protocols, patients randomized to the combination therapy were started on favipiravir for 10 days and hydroxychloroquine for 5 days. Favipiravir was given at a loading dose of 1800 mg (9 tablets) by mouth twice on day 1 followed by 800 mg (4 tablets) twice daily. Hydroxychloroquine dose was 400 mg twice daily on day 1 and then 200 mg twice daily. Other aspects of COVID-19 management were at the discretion of the treating physicians. The use of glucocorticoids, other immunomodulators, and antibiotic agents was allowed, and all were recorded throughout the hospitalization. As a safety consideration, the trial team assessed the QTc interval every 24 h while the patients were receiving the combination of the trial drugs.
Observation of possible adverse events (AEs) was done while using the investigational medications until day 14, including regular laboratory monitoring. In case of any concerns that might require holding/stopping medication, the principal investigator conducted a causality assessment, including a consultation with an independent clinician. One point contact at 28 and 90 days was planned to assess the vital status.
The primary endpoint was the time to clinical improvement, defined as the time from randomization to an improvement of two points on a seven-category ordinal scale or live discharge from the hospital, whichever came first. The seven-category ordinal scale consists of the following categories: a score of 1 indicated not hospitalized with the resumption of normal activities; 2, not hospitalized, but unable to resume normal activities; 3, hospitalization, not requiring supplemental oxygen; 4, hospitalization, requiring supplemental oxygen; 5, hospitalization, requiring nasal high-flow oxygen therapy and/or non-invasive mechanical ventilation; 6, hospitalization, requiring extracorporeal membrane oxygenation (ECMO) and/or invasive mechanical ventilation; 7, death.
The key secondary efficacy outcomes included the progress in clinical status and the requirement of ICU admission or mechanical ventilation within 14 days. The 28- and 90-day mortality was also monitored. In addition, the RT-PCR test conversion from positive to negative, length of hospital stay, and status at day 14 post-randomization were also collected. SARS-COV-2 RT-PCRs from a naso-/oropharyngeal swab or lower respiratory samples were obtained on days 1, 5, 10, 14, 21, and 28 or until the test was negative, whichever came first. All PCR testings are done in Central Laboratories, where all procedures and kits were approved by the Saudi FDA and CDC. These include: Perkin Elmer chemagic 360 Instrument (chemagic™ Viral DNA/RNA Extraction Kit), Roche MagnaPure96 (SV kit), Co-Diagnostics (Logix-Smart™-COVID-19), and Real Star SARS-CoV-2 RT-PCR Kit.
The safety outcomes included treatment-related adverse events (AEs), serious adverse events (SAEs), and premature discontinuations of the study medications. SAEs, defined as untoward medical events leading to death, a life-threatening experience, prolongation of hospitalization, or persistent or significant disability or incapacity in the judgment of the site investigator, were also reported.
Under the classical two-arm parallel design, the total adequate sample size required was 472 subjects (236 per group). The estimated sample size achieved 85% power to detect a minimum of a 30% reduction in the median time to clinical improvement by at least two points or hospital discharge in the treatment group compared to the control group. The trial's overall sample size was estimated as 520 (260 per group), including 10% drop-out rate. A sample size adjustment was planned to be considered after the interim analysis of 60% of the subjects' recruitment and the DSMB recommendation. The interim analysis was designed to test for early stopping due to futility or efficacy and the sample size re-estimation. The interim analysis and final analysis plan were based on the method of the sum of the stage-wise p-value discussed in Chow and Chang, 2008. We performed several sensitivity analyses to determine the effect of different assumptions on the sample size. (Details are provided in the statistical analysis plan in Online Appendix 1).
The study's primary endpoint was the median time to clinical improvement by at least 2 points or live discharge from the hospital. It was analyzed using the log-rank test to compare median time to clinical improvement through day 14 between the two groups. The hazard ratios and 95% confidence interval (CI) were estimated using the proportional hazards regression model with treatment as the explanatory variable. The decision rule based on the study protocol were (1) stop the trial for early efficacy if the interim analysis p-value < 0.01, (2) stop the trial for futility if the interim analysis p-value ≥ 0.25, or (3) declare the trial significant if the sum of the interim analysis and final stage p-values < 0.1832. Similar analyses were performed for the secondary endpoints, including hospital discharge and mortality based on the log-rank test and the proportional hazards regression. For the hospital discharge, patients who were alive through day 14 were included. For the mortality, we compared the two treatment groups through day 28.
Subgroup analyses were performed for the primary endpoint and secondary endpoints to test the treatment effect across subgroups of age, gender, corticosteroid use, the baseline score of the seven-category ordinal scale at admission, and the time between onset to enrollment. The analyses were based on the proportional hazards regression with interaction terms of the treatment and subgroup variables. The results were reported in terms of hazard ratios, the 95% CI for each subgroup, and the interaction p-value. Further compressions and analyses of the secondary endpoints were applied using the chi-square or the Wilcoxon signed-rank tests whenever appropriate and indicated in result tables.
The analysis of the AEs data was descriptive and excluded patients who did not receive treatment. The proportions of the participants experiencing AEs were compared between the randomized groups using a Fisher's exact test.
No specific statistical handling of missing values was performed. Statistical significance was set at p < 0.05. Statistical analyses were done using SAS V9.4.
Stopping the Trial
The DSMB members voted for an early interim analysis after recruiting > 50% of the sample size based on the trial steering committee's concern about the low recruitment rate due to the decreased numbers of the new COVID-19 cases. On 26 January 2021, the DSMB conducted a closed session to review the interim data analysis and recommended aborting the trial because of futility. The patients' enrollment in the trial was stopped immediately, and the site's principal investigators, IRB, and Saudi FDA were notified of the termination.