Trial setting
Participants will be non-hospitalized adults recruited from the community via advertisements and using traditional methods and social media (Twitter, Instagram, Facebook). Recruitment will take place throughout the United States (US), and, at least initially, within three provinces in Canada (Quebec, Manitoba, Alberta). Approval is being sought in additional provinces. To limit the spread of disease, no in-person visits with study personnel will be performed. During recruitment and follow-up, most participants will be at home in self-isolation or quarantine; screening, informed consent, randomization, and data collection will occur online through the Research Electronic Data Capture (REDCap), a secure web-based platform. Study medications will be delivered directly to the participants’ place of residence via courier services that respect safety protocols for home deliveries to patients with COVID-19. Participants can contact the local study teams in their region via email or phone for questions or concerns with timely responses (same day during business hours and the following day after 9 pm). Websites have been created that will provide live enrolment updates, responses to frequently asked questions, and trial information for participants (www.covidpep.umn.edu and www.covid-19research.ca).
Inclusion criteria
Post-exposure prophylaxis
-
1)
At-risk exposure to a COVID-19-positive case within the past 4 days
-
2)
Household contact or healthcare worker
Pre-emptive therapy
-
1)
Non-hospitalized community-dwelling symptomatic (fever, cough, or shortness of breath) COVID-19 within four days of symptom onset
-
2)
Coronavirus disease confirmed by reverse transcriptase polymerase chain reaction (RT-PCR) or for healthcare workers, by compatible symptoms following exposure to a person with a known positive RT-PCR result (exposure occurring within the past 14 days).
Enrolment for the PEP trial will be limited to asymptomatic individuals exposed to a COVID-19-confirmed case within ≤ four days, as the mean incubation period for SARS-CoV-2 is five days.12 Household contacts are defined as residing in the same domicile. At-risk exposures in the healthcare setting are defined as greater than ten minutes spent within 2 m of a COVID-19-confirmed case, with insufficient use of personal protective equipment per institutional guidelines. Healthcare workers are self-identified physicians, nurses, respiratory therapists, or other workers providing direct patient contact. This will select a high-risk group for developing clinical infection but within the time period that the intervention could prevent or ameliorate disease. All efforts will be made to deliver study medicines before the fourth day following exposure; however, delays may lead to symptom onset after randomization but prior to starting the intervention. These events will be tracked by online surveys administered upon receipt of the intervention. Because community testing for COVID-19 is limited, participants will be included in the PET group as long as symptoms are compatible with COVID-19 after exposure to a person with confirmed COVID-19.
Exclusion criteria
-
1)
Known allergy to chloroquine or hydroxychloroquine
-
2)
Current use of chloroquine or hydroxychloroquine for any reason
-
3)
Currently hospitalized
-
4)
Known chronic kidney disease, stage 4-5, or dialysis33
-
5)
Active malignancy on systemic chemotherapy
-
6)
Known porphyria
-
7)
Prior retinal eye disease
-
8)
Known glucose-6-phosphate dehydrogenase (G6PD) deficiency
-
9)
Weight < 40 kg
-
10)
Known history of ventricular arrhythmia, prolonged QTc interval (personal or family history), structural or ischemic heart disease, or any known episode of sudden cardiac death
-
11)
Severe diarrhea and/or vomiting at screening that may interfere with drug absorption
-
12)
Significant hepatic impairment defined as known cirrhosis with a history of hepatic encephalopathy or ascites
-
13)
Known pregnancy or breastfeeding
-
14)
Current use of any of the following medications with known drug-drug interactions: amiodarone, amitriptyline, artemether, azithromycin, ciprofloxacin, citalopram, clarithromycin, dapsone, desipramine, digoxin, dofetilide, doxepin, droperidol, erythromycin, escitalopram, flecainide, fluoxetine, haloperidol, imipramine, itraconazole, ketoconazole, levofloxacin, lithium, lumefantrine, mefloquine, methadone, moxifloxacin, procainamide, propafenone, quetiapine, sertraline, sotalol, sumatriptan, thioridazine, venlafaxine, bupropion, ziprasidone, zolmitriptan
-
15)
Current use of any anti-epileptic.
Exclusion criteria 11-14 are specific to participants in Canada at the request of Health Canada. Exclusion criteria 15 is specific to participants in the US. In the US, the FDA has approved the inclusion of pregnant and breastfeeding women citing hydroxychloroquine safety data in pregnant patients with autoimmune conditions including systemic lupus erythematosus and malaria.34
Interventions
Participants will be randomized 1:1 to receive hydroxychloroquine or placebo.
Intervention arm
Hydroxychloroquine sulfate will be administered in 200 mg tablets for a total dose of 3,800 mg (19 tablets) over five consecutive days. Dosing regimens will consist of a loading dose with subsequent maintenance dose as follows:
-
1)
Day 1 (upon receipt of study medicine): hydroxychloroquine 800 mg orally once (four tablets), followed six to eight hours later by 600 mg orally once (three tablets);
-
2)
Days 2 to 5: hydroxychloroquine 600 mg orally once daily (three tablets per day).
Control arm
Placebo tablets will be administered using the same dosing schedule and number of tablets as the interventional arm.
The rationale for the study dose is adapted from a modified malaria dosing schedule for hydroxychloroquine.35 Considerations were made for maximal viral inhibition while balancing potential toxicity and drug supply with anticipated shortages. A loading dose was used, similar to malaria treatment strategy to achieve antiviral activity as quickly as possible. Monte Carlo simulations were conducted with the proposed dose achieving steady-state serum concentrations above the EC50,22 where 50% of viral inhibition would occur.
The most common anticipated side effect from hydroxychloroquine is gastrointestinal upset.27 Participants will be instructed to spread out doses throughout the day (e.g., two tablets with breakfast, one tablet with lunch) to increase tolerability if required. Participants may voluntarily stop taking the study medicine at any time. Adherence will be determined by self-report. Chronic medications will be taken as usual, but these must be taken 4 hours before or after taking the study medicine. No other concomitant COVID-19 therapy or other trial medicines are permitted for the duration of the trial. In the event of hospitalization, unblinding of group assignment can occur at the request of the treating physician, and participants may receive other medications or therapies, including experimental treatments from other trials.
Outcomes
Primary outcome for PEP
-
1)
Incidence of proven COVID-19 disease onset by 14 days.
In the absence of SARS-CoV-2 testing, COVID-19 diagnosis will be made using clinical criteria and labelled as a possible case. For the “confirmed cases” outcome, there may be limitations related to access to testing. We will use an a priori clinical definition using the symptoms most commonly reported in PCR-proven patients from the World Health Organization (WHO).36
Primary outcome for PET
-
1)
Ordinal scale for clinical severity, defined by WHO COVID-19 therapeutic trial synopsis37
1 = Ambulatory
2 = Hospitalized with mild disease
3 = Hospitalized with severe disease
4 = Death (all-cause).
Secondary outcomes for PEP and PET
-
1)
Incidence of hospitalization for COVID-19;
-
2)
Incidence of death from COVID-19;
-
3)
Incidence of confirmed SARS-CoV-2 detection;
-
4)
Incidence of possible COVID-19 symptoms;
-
5)
Incidence of all-cause study medicine discontinuation;
-
6)
Severity of symptoms at day 5 and 14 by visual analogue scale.
The primary and secondary outcomes listed will be analyzed separately for both trials. Outcome measures will be assessed through self-report via online questionnaires. Wherever possible, COVID-19 disease status will be verified from public health records, medical records, or death certificates. Patients with positive RT-PCR results for SARS-CoV-2 in a nasopharyngeal swab will be considered confirmed cases.38 rT-PCR will be performed as per the local institution’s standard procedures.
For patients with COVID-19 who require hospitalization, outcome data will be collected up to day 90. Participants will complete additional online surveys throughout hospitalization to assess for adverse events related to the study medicine and their COVID-19-related outcomes. All participants will complete virtual assessments on days 1, 5, and 14 (Figure). If symptomatic disease develops, participants will be required to complete additional virtual assessments on days 3 and 10. Participants who are pregnant or become pregnant while participating in the trial will be followed to the post-partum period to assess maternal and fetal outcomes. In the event that a participant does not respond to the day 14 survey, we will contact them by telephone, by registered mail, and if they are still not available, all participants have given a third-party telephone contact to verify vital status.
Recruitment and consent
Recruitment and follow-up will be exclusively internet-based. Prospective participants will visit the trial website, read the provided study information, and can click the link to enrol. The trial website will be advertised through communication strategies and media channels, and through internet-based advertisements and social media. In all regions, COVID-19-positive cases are reported to public health officials where positive patients are notified, and contact tracing is initiated. In some jurisdictions, public health officials will provide information on accessing the trial website. Healthcare workers will be notified of the trial website by email communications and through the offices of occupational health and promotion by senior hospital administration. Online screening questionnaires will assess interested participants for eligibility criteria. Eligible participants will provide informed consent by submitting an online form, linked through the trial website. Contact information will be provided for participants to ask questions by email and/or phone.
Randomization sequence generation, allocation concealment, and blinding
Participants will be randomized via computer-generated permuted block randomization and stratified into the PEP trial (for asymptomatic participants) or the PET trial (for symptomatic participants). The randomization sequence will be pre-specified. Randomization will be recorded on an electronic log, which will only be accessed by the research pharmacy, those packaging study medication, and the unblinded statisticians. Study investigators, trial participants, care providers, outcome assessors, and data analysts will be blinded to the allocation sequence. Once study medication is prepared by the onsite pharmacies, packages will be shipped directly to participants via courier.
Participants will be provided masked study medicine and will not be notified whether they are taking hydroxychloroquine or placebo. Instructions for taking the medicine are identical between the groups. Because a variety of sources of hydroxychloroquine are being used in different jurisdictions, the placebo will not be identically matched to the hydroxychloroquine tablet; however, both placebo and hydroxychloroquine will have similar rounded oval shapes, colour, and indistinct markings making unblinding of the trial arm unlikely. Patients will be asked if they believe they received the intervention or control upon trial closure. In the event of a medical emergency, an on-call investigator will be available for code breaking if required. Any code breaking event will be reported to the research pharmacy and principal investigator.
Data collection methods and data management
All data collection, including baseline demographics, clinical and epidemiological characteristics, study medicine tolerability, adherence, and outcomes will be recorded online through self-report captured in internet-based questionnaires (administered in REDCap). All participant-submitted data will be stored and maintained in a secure server. Data from the US will be stored on the secure servers at the University of Minnesota, and data from Canada will be stored on the secure servers at the Research Institute of the McGill University Health Centre. Both the US and Canadian databases will be stored and managed in their respective countries. For interim data safety monitoring board purposes and also upon completion of recruitment, confidential de-identified data will be shared to facilitate analysis through institutional data sharing agreements. No paper documents will be retained or stored.
Enrolment progress reports will be generated for each trial at 25%, 50%, and 75% of enrolment to highlight the number of participants enrolled, on study, completed, or lost to follow-up. The cumulative incidence of COVID-19 (both arms pooled) and cumulative hospitalizations (both arms pooled) will be reported.
Sample size
According to existing research, the attack rate of transmission to household contacts is approximately 10%.18 We used Fisher’s exact test and a two-sided α = 0.05 and test power of β = 0.90 to detect a 50% relative risk reduction in disease incidence. Using 1:1 randomization, we calculated a required sample size of n = 621 per arm. Accounting for a predicted 20% dropout rate with an internet self-study design the sample size was inflated to n = 746 per arm. Therefore, we aimed to recruit a total of 1,500 patients for the PEP trial. For symptomatic patients, the proportion of patients who progress to require hospitalization is approximately 10%.14 For this trial, we assumed that without treatment, 90% of the control arm would not be hospitalized, 8% would be hospitalized without an ICU stay or death, and 2% would be hospitalized with an ICU stay or death. With 1:1 randomization and 1,464 total participants, the PET trial has 90% power to detect a log odds ratio of 0.70. A number of sample size approximations were made accounting for differences in baseline rates and relative effect sizes. These can be found in Appendices 1 and 2.
Statistical methods
Primary and secondary outcomes will be analyzed separately in PEP and PET trials by intention-to-treat.
Primary outcome for PEP
The incidence of COVID-19 by day 14 will be assessed in the intervention vs control arm by the Fisher’s exact test.
Primary outcome for PET
The ordinal scale for disease severity (outpatient, hospitalized without ICU, hospitalized with ICU, death) at day 14, will be assessed in the intervention vs control arms by proportional odds model.
Secondary outcomes (PEP and PET are analyzed separately)
In the PEP group, for the subset of patients who develop symptoms of COVID-19 after taking at least one dose of the study medicine, we will analyze the change in symptom severity in the intervention vs control group. We will use the same approach as for the primary outcome in the PET group. Participants who are randomized in the PEP group but develop symptoms before taking one dose of the study medicine will be described and analyzed separately in a subgroup analysis.
Secondary outcomes for incidence will be reported with the Fisher’s exact test. Continuous valued secondary outcomes will be reported by mean (standard deviation) or median [interquartile range], depending on their normality. Analysis will be conducted by parametric or non-parametric tests as appropriate. Symptom severity scores will be recorded on days 0, 1, 3, 5, 10, and 14 using a 10-cm continuous visual analogue scale, where participants can electronically control a slider indicating symptom severity, with 0 = no symptoms and 10 = severe symptoms. Measurements will be recorded to the nearest 0.1 cm. The severity of symptoms on day 5 will be compared first by categorical analysis (symptoms present yes or no) via Fisher’s exact Chi square, and subsequently via the independent two-sample t test for symptom severity among those who are symptomatic. Non-normally distributed data will be analyzed via the Mann–Whitney U test.
Preplanned subgroup analyses
-
1)
Participants contact had SARS-CoV-2 confirmed by positive molecular test result;
-
2)
Participant had SARS-CoV-2 confirmed by positive molecular test result;
-
3)
Exposure in healthcare worker vs household contact;
-
4)
Number of days from the exposure;
-
5)
Decile of age;
-
6)
Sex as a biological variable;
-
7)
Censored subjects in the PEP trial who become symptomatic before taking one dose of the study medicine.
Data monitoring
A Data Safety Monitoring Board (DSMB) will be formed independent of the sponsor and without competing interests. For each trial (PEP and PET), an interim analysis will be presented to the DSMB once the 14-day follow-up is completed for 25%, 50%, and 75% of enrolled participants. All interim analyses will pool US and Canadian de-identified data to arrive at a generalizable conclusion as early as possible. Only the DSMB and unblinded statisticians will be provided with unblinded interim analysis results. The investigators will have access to only pooled results. A Lan-DeMets spending function analogue of the O’Brien-Fleming boundaries for the primary outcome for each group will be provided to the DSMB. The stopping boundaries will be truncated at α = 0.001 (|Z| > 3.09). At each analysis, data will be reviewed for safety/efficacy and for achieving stopping rules. Should stopping rules for the primary outcome be met, the DSMB will review secondary outcomes for consistency such that a clear answer is achieved. In the event of early termination for efficacy, the trial will immediately convert to an open-label observational cohort study.
At the second DSMB review, a sample size re-estimation will occur based on the disease transmission rate in the control group. The a priori assumption of 10% transmission risk in close contacts is based on limited data in other jurisdictions. A new sample size estimation will take into account the updated transmission rate in this trial and be powered to detect a 50% relative reduction in the primary outcome. Starting with the second DSMB review, the DSMB will be given the conditional power under both the trial design parameters and under the current data. If the conditional power is less than 20%, trial discontinuation may be considered.
Potential harms
Short-term use of hydroxychloroquine is well-tolerated with a safe track record of use for approved indications since 1955. The most commonly reported side effects include: gastrointestinal upset, nausea, vomiting, diarrhea, headache, skin rash, or itching.27 Gastrointestinal side effects are minimized when taken with meals and when dosages are separated throughout the day. Participants will be instructed to use these optional strategies to increase tolerance.
While a recent study showed a risk of harm with very high doses of chloroquine,24 hydroxychloroquine is much less toxic.26 Doses used in this trial (3.8 g total) are similar to those doses used to treat malaria (2 g total), and are significantly less than doses used to treat Coxiella burnetii infection or other autoimmune conditions (400-600 mg daily, indefinitely), all showing a long history of safety.27 A recent trial in COVID-19 using significantly higher doses (1.2 g daily for three days, then 800 mg daily for two to three weeks) showed no significant differences in serious adverse advents, though gastrointestinal side effects were higher at this dose.28
The safety of taking hydroxychloroquine will be enhanced by excluding those with pre-existing retinopathy, allergic reactions, cardiac conditions, and certain other medications. All adverse reactions will be documented via self-report during virtual visits. Participants will be provided with contact information to be guided on how to manage adverse effects. There are no predefined plans for post-trial care above the standard of care offered at each centre. Participants will be provided with updated public health information on the COVID-19 pandemic and instructions for how to proceed in the event of a medical emergency.
Auditing
Regulatory agencies, institutional sponsors, and REBs are authorized to conduct trial audits throughout.
Confidentiality and access to data
Interactions with trial participants will be through internet-based REDCap electronic case report forms, conforming to required privacy and server security standards. No participant identifying information will be disclosed in any publication or in any other activities arising from this trial. Anonymized data will be pooled with international collaborators, subject to inter-institutional agreements. No information concerning this trial will be released to unauthorized third parties without the prior written approval of the participants, except for monitoring by the REBs or public health authorities. Only immediate study personnel will be authorized to access the database.
Dissemination policy
Any protocol amendments will be reviewed with each REB and updated on the trial registry and websites. Trial enrolment and results, when available, will be updated on www.clinicaltrials.gov and will be publicly available on the trial websites (www.covidpep.umn.edu and www.covid-19research.ca). Because of the nature of this pandemic, pooled and anonymized trial results will be published immediately, once available, on open access websites and will be submitted urgently for accelerated peer reviewed publication. Full de-identified data will be made available to qualified researchers upon request.