Pre-exposure prophylaxis with hydroxychloroquine for COVID-19: initial results of a double-blind, placebo-controlled randomized clinical trial

Pre-exposure prophylaxis (PrEP) is a promising strategy to break COVID-19 transmission. Although hydroxychloroquine was evaluated for treatment and post-exposure prophylaxis, it is not evaluated for COVID-19 PrEP yet. The aim of this study was to evaluate ecacy and safety of PrEP with hydroxychloroquine against placebo in healthcare workers at high risk of SARS-CoV-2 infection during an epidemic period. Methods We conducted a double-blind placebo-controlled randomized clinical trial in three hospitals in Barcelona, Spain. From 350 adult healthcare workers screened, we included 269 participants with no active or past SARS-CoV-2 infection (determined by a negative nasopharyngeal SARS-CoV-2 PCR and a negative serology against SARS-CoV-2). Participants allocated in the intervention arm (PrEP) received 400mg of hydroxychloroquine daily the rst four consecutive days and subsequently, 400mg weekly during the study period. Participants in the control group followed the same treatment schedule with placebo tablets. 52.8% of participants were in the hydroxychloroquine arm and 47.2% in the placebo arm. Both groups showed similar proportion of participants experiencing at least one adverse event (AE) (p=0.548). No serious AE were reported. Almost all AE (96.4%) were mild. Only mild gastrointestinal symptoms were signicantly higher in the hydroxychloroquine arm compared to the placebo arm (28.3% vs 16.9%, p=0.044). Given the national epidemic incidence decay, only one participant in each group was COVID-19 diagnosed. Consequently, our study design deemed underpowered to evaluate any benet regarding PrEP ecacy. relationship of high of hydroxychloroquine AE supported by a recent study evaluating high doses of hydroxychloroquine (up to 600 bid during 10 days) for treating COVID-19 This study showed that 15% of participants prolonged QTc interval and two of them presented ventricular tachycardia. In this specic case, the suggested that this outcome could have been inuenced by most of their participants receiving oseltamivir, which also prolongs QT, and the elder age of them.(20) In our case, no cardiovascular events study


Abstract Background
Pre-exposure prophylaxis (PrEP) is a promising strategy to break COVID-19 transmission. Although hydroxychloroquine was evaluated for treatment and post-exposure prophylaxis, it is not evaluated for COVID-19 PrEP yet. The aim of this study was to evaluate e cacy and safety of PrEP with hydroxychloroquine against placebo in healthcare workers at high risk of SARS-CoV-2 infection during an epidemic period.

Methods
We conducted a double-blind placebo-controlled randomized clinical trial in three hospitals in Barcelona, Spain. From 350 adult healthcare workers screened, we included 269 participants with no active or past SARS-CoV-2 infection (determined by a negative nasopharyngeal SARS-CoV-2 PCR and a negative serology against SARS-CoV-2). Participants allocated in the intervention arm (PrEP) received 400mg of hydroxychloroquine daily the rst four consecutive days and subsequently, 400mg weekly during the study period. Participants in the control group followed the same treatment schedule with placebo tablets.
Results 52.8% of participants were in the hydroxychloroquine arm and 47.2% in the placebo arm. Both groups showed similar proportion of participants experiencing at least one adverse event (AE) (p=0.548). No serious AE were reported. Almost all AE (96.4%) were mild. Only mild gastrointestinal symptoms were signi cantly higher in the hydroxychloroquine arm compared to the placebo arm (28.3% vs 16.9%, p=0.044). Given the national epidemic incidence decay, only one participant in each group was COVID-19 diagnosed. Consequently, our study design deemed underpowered to evaluate any bene t regarding PrEP e cacy.

Conclusions
First month follow-up analysis displayed that PrEP with hydroxychloroquine at low doses is safe.

Trial registration
This trial was registered at clinicaltrials.gov (NCT04331834) on April 2nd 2020.

Background
The novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is the causative agent of coronavirus disease 2019 .(1) Currently, the worldwide prevention strategies for SARS-CoV-2 infection are self-protection equipment use, hand washing, case identi cation, case isolation, contact tracing and exposed people quarantine of close contacts. (2)(3)(4) In these circumstances, secondary attack rate estimates of COVID-19 ranges from 3-15% in the community (5)(6)(7), which can reach 26% in healthcare professionals.(8) Prevention of healthcare workers' infection is crucial for protecting the workforce during pandemic management. Pre-exposure prophylaxis (PrEP) is a promising strategy, which proved effective in preventing other infectious diseases.(9) Thus, several trials with repurposed drugs to evaluate PrEP effectiveness in mitigating SARS-CoV-2 transmission are under development. (10) Chloroquine was observed to effectively inhibit SARS-CoV-2 in vitro. (11,12) In fact, its derivate hydroxychloroquine shows a better in vitro antiviral activity and safety pro le. (13,14) Hydroxychloroquine potentially inhibits entry and post-entry stages of SARS-CoV-2. (11,13) While hydroxychloroquine was evaluated for the treatment of SARS-CoV-2 pneumonia and post-exposure prophylaxis, it is not evaluated for PrEP yet. The aim of this study is to compare e cacy and safety of PrEP with hydroxychloroquine against placebo in healthcare workers in reducing the risk of COVID-19 disease during an epidemic period.

Trial design
We conducted a multicentre double-blind, placebo-controlled randomized clinical trial. We allocated participants to one of the two study arms in a 1:1 ratio by simple randomization. Randomization list was generated prior to enrolment.
Screening of candidates was initiated on April 3rd 2020 and rst recruitment on April 4th 2020. We included adult healthcare workers working at least three days a week in a trial hospital with a negative result of SARS-CoV-2 polymerase chain reaction (PCR) assay in nasopharyngeal swab within four days before enrolment. COVID-19 serology was evaluated in all candidates with a rapid diagnostic test (Vivadiag™ COVID-19 IgM/IgG Rapid Test ©, Hangzhou, China) and con rmed with Enzyme-Linked ImmunoSorbent Assays (VITROS Anti-SARS-CoV-2 Total, © Ortho-Clinical Diagnostics, 2020).
Those individuals positive for COVID-19 serology by any method were excluded. We also excluded participants with any of the following conditions: pregnancy, breastfeeding, ongoing antiviral, antiretroviral or corticosteroids treatment, chloroquine or hydroxychloroquine intake the last month or any contraindication to hydroxychloroquine.

Intervention and comparator
Randomization was generated using a computer random number generator. We used sequentially numbered sealed envelopes of identical appearance containing either hydroxychloroquine or placebo, ensuring allocation concealment. Participants allocated to intervention arm (PrEP) received 400 mg of hydroxychloroquine (two tablets of 200 mg) daily the rst four consecutive days, followed by 400 mg weekly during the study period, initially scheduled to be 6 months. Participants in the control group followed the same treatment schedule with placebo tablets that were indistinguishable from hydroxychloroquine tablets.
Participants took the rst two tablets at recruitment visit under direct observation by a physician, who then provided the needed tablets to complete the rst month of treatment.
Participants, investigators assessing participant eligibility and recruitment, assessing outcomes and follow-up, and/or dealing with data management and analysis were all blinded to arm allocation. Only one person unrelated to participant recruitment and follow-up, clinical assistance, data management and analysis had access to this information.

Outcome
The primary outcome was the incidence of COVID-19 con rmed cases (de ned by compatible symptoms with COVID-19 with seroconversion or a positive PCR for SARS-CoV-2) in the hydroxychloroquine arm compared to the placebo arm at any time during the study follow-up.
The secondary outcomes included: i) the SARS-CoV-2 seroconversion in the hydroxychloroquine group compared to the placebo group in both asymptomatic and symptomatic participants; ii) the occurrence of any adverse event (AE) related with hydroxychloroquine treatment; iii) the incidence of SARS-CoV-2 infection and COVID-19 in healthcare workers in the placebo group during the study period; iv) the risk ratio for the different clinical, analytical and microbiological conditions to develop COVID-19.

Participant's follow-up
Passive and active surveillance was conducted to all participants to detect SARS-CoV-2 infections and any AE.
Active surveillance of each participant was conducted monthly by blinded physicians unaware of the trial arm assignments, which completed a standardized case report form (CRF) for each participant. Follow-up visits included: i) assessment of compliance with PrEP, ii) physical examination and detailed evaluation of symptoms to either detect past and current symptoms and signs related to COVID-19, iii) venepuncture for blood determinations, including SARS-CoV-2 serology test (VITROS Anti-SARS-CoV-2 Total, © Ortho-Clinical Diagnostics, 2020), iv) assessment of COVID-19 risk factors such as known close contacts with suspected and/or con rmed COVID-19 cases or the number of weeks during which they were managing COVID-19 patients, v) standardized questions to collect past and current common side effects along with open free text and, vi) electrocardiogram to evaluate possible cardiac rhythm alterations.
During this study period, a medical doctor was available by phone 24hours a day during the study period for passive surveillance.
All participants were provided with this contact number in case of presenting any COVID-19 related symptom or AE. In that instance, a standardized CRF was lled out to collect the information. A nasopharyngeal swab was performed to all those participants presenting with COVID-19 related symptoms to detect SARS-CoV-2 infection by PCR.
Medical assistance was ensured for all participants diagnosed with SARS-CoV-2 infection following hospital guidelines.
Although the protocol was designed to follow participants for 6 months, this manuscript only includes participants' rst month analysis.

Sample size
We estimated sample size assuming an expected incidence of 10% of COVID-19 in healthcare workers in the control group and 2% in the hydroxychloroquine group, with a hazard ratio of 0.2. Thus, we required a total of 440 subjects (220 per group) for a signi cance level of 5%, statistical power of 90% and assuming a rate of lost-to-follow-up of 10%. (15) Interim analysis Interim analyses of the e cacy and safety of hydroxychloroquine were planned monthly, with the option of early stopping the trial for futility. We planned to restimate incidence and lost-to follow-up rate at the rst month, since these data was unknown when we estimated sample size. After the rst interim analysis the trial was halted on the basis of a very low incidence rate among study participants.

Statistical analysis
We conducted an intention-to-treat analysis, with all patients ful lling inclusion criteria and not presenting exclusion criteria.
Categorical variables were expressed as absolute frequency and percentage and were compared with Fisher's exact test. Continuous variables were expressed as mean and standard deviation (SD) or median and interquartile range (IQR). We conducted all analysis with R.(16)

Screening
We assessed 350 health-care workers for eligibility, 269 of them ful lled the study criteria and were recruited after signing the informed consent form. Participants were randomly assigned to the hydroxychloroquine group (n = 142, 52.8%) and to the placebo group (n = 127, 47.2%). Figure 1 describes participants enrolment and randomization.

COVID-19 risk factors at month 1
After a month of follow-up, 39.9% of participants ceased to work in a COVID-19 hospital unit, with similar proportion between both groups. In addition, only seven participants denied having used always PPE when assisting patients.The proportion of participants in contact with a con rmed and/or a suspected COVID-19 case showed to be lower compared with the screening visit (p-value < 0.001 in the hydroxychloroquine group, p-value = 0.002 in the placebo group). Participants from the hydroxychloroquine group had lower contact with con rmed COVID-19 cases without using PPE compared to the placebo (p-value = 0,018). (Table 2) E cacy During the rst month of follow-up, the overall accumulated incidence of COVID-19 among the study participants was 0.8%. Only one participant, from placebo arm, was diagnosed with COVID-19 by SARS-CoV-2 PCR and serology test. The participant presented with fever, respiratory symptoms and headache 6 days after randomization. The participant did not received speci c treatment for COVID-19 or required hospitalization.
The risk ratio of collected risk factors and clinical and analytical conditions for developing COVID-19 could not be calculated due to low COVID-19 incidence.

Safety
A total of 95 participants experienced at least one AE during the rst month of follow-up, posing an overall accumulated prevalence of 35.3%. The proportion of participants experiencing at least one AE was similar in both groups. Eighty-one events (33.0% in the hydroxychloroquine group and 26.0% in the placebo group, p = 0.207) were judged to be related to the study intervention (hydroxychloroquine or placebo). No serious AE were reported. Almost all AE (96.4%) were considered mild. Only four were reported as moderate: prostate adenocarcinoma in the hydroxychloroquine group and dental infection, hypertensive crisis and myalgia in the placebo group.
Four cardiovascular AE (two in each study group) were detected during the rst month of treatment: two pre-excitation syndromes (Wolf-Parkinson-White), one with hypertensive crisis and two participants presenting with heart palpitations. Only hypertensive crisis was moderate, the rest were mild and none of them was considered related to the study drug. None of the participants presented prolonged QTc intervals at rst month (388 ms  in the hydroxychloroquine group and 393 ms in  in the placebo group).
Only one participant in the hydroxychloroquine group presented mild visual disturbances.
Five participants discontinued the prophylaxis due to AE in the hydroxychloroquine group and one in the placebo group. Table 4 shows a detailed description of the AE presented in both study groups.  (Table 3).

Discussion
We conducted a multicentre double-blind placebo-controlled randomized clinical trial to evaluate whether PrEP with hydroxychloroquine was an effective intervention for preventing COVID-19 among healthcare workers during a COVID-19 epidemic period. These initial data showed that prophylaxis with hydroxychloroquine at the study doses had an excellent safety pro le. Nevertheless, the community incidence of SARS-CoV-2 events decreased during the rst month of follow-up as a consequence of country's control and mitigation strategies. Thus, the overall incidence in the cohort was 0.8% and the study design was deemed underpowered to answer the main objective.
In this trial, AE were similar between the hydroxychloroquine and the placebo group. Mild gastrointestinal events were higher in the hydroxychloroquine group than in the placebo group. Our safety data contrast with available data from a recently published placebo-controlled clinical trial post-exposure prophylaxis with hydroxychloroquine to prevent COVID-19, in which the intervention arm had a higher rate of AE (40.1%) compared to the placebo arm (16.8%). (17) The reasons for these discrepancies may include the high loading doses of hydroxychloroquine required for the post-exposure prophylaxis strategies (3200 mg compared to 1200 mg in 4 days in our intervention). Moreover, the loading dose during 4 consecutive days in PrEP strategies could be avoided since the patient has not yet been exposed to the virus, potentially decreasing the number of AE.
Large observational studies evaluating the hydroxychloroquine effect on SARS-CoV-2 pneumonia showed that higher doses of the drug could be associated with QTc interval prolongation and death due to cardiovascular events; especially when administered with other drugs which favor QT interval prolongation. (18, 19) The relationship of high doses of hydroxychloroquine with severe AE was also supported by a recent study evaluating high doses of hydroxychloroquine (up to 600 bid during 10 days) for treating COVID-19 patients. This study showed that 15% of participants prolonged QTc interval and two of them presented ventricular tachycardia. In this speci c case, the authors suggested that this outcome could have been in uenced by most of their participants receiving oseltamivir, which also prolongs QT, and the elder age of them. (20) In our case, no cardiovascular events related to the study drug were observed.
The adequate safety pro le of the preventive (low) doses of hydroxychloroquine found in our study is supported by many studies on hydroxychloroquine short-term use as antimalarial and long-term use for rheumatic diseases demonstrated to be safe as well. (21)(22)(23)(24) As mentioned above, this trial has some limitations. The main one is low power in our study design to assess the e cacy of PrEP with hydroxychloroquine in healthcare workers at the rst interim analysis. The principal reason was the low incidence of COVID-19 during the study follow-up in the study area. The analysis presented in this manuscript was conducted during April, May and June 2020 in Barcelona. During that period, COVID-19 reported cases were already decreasing in Catalonia region: while 1208 cases were reported in April 3rd, 77 cases were reported in June 12th. Accordingly, SARS-CoV-2 attack rate declined from 0.9 at the initiation of the study to 0.77 at the end of the analysis.(25) In addition, we noticed that healthcare workers risk perception decreased as national epidemic vanished. That had an impact not only on participants' recruitment but also on participant's adherence and follow-up.
Moreover, other healthcare workers factors in uenced their participation. Some study candidates had already self-prescribed hydroxychloroquine assuming its role on COVID-19 prevention, so although displaying interest they were excluded to participate. In addition, a fraction of recruited participants dropped out from the study the following day of test results noti cation. Hence, this set of circumstances prompted us to stop trial recruitment by 8th of May even though we had not reached our estimated sample size. Consequently, both study groups were unevenly distributed (52.8% vs 47.2%).

Conclusions
This trial displayed that administering 400 mg of hydroxychloroquine during 4 consecutive days followed by 400 mg of hydroxychloroquine weekly in adults during a month was safe. The question if COVID-19 could be prevented with hydroxychloroquine PrEP was unanswered. Since the epidemiological situation happening in our country is expected to be reproduced in other areas where similar trials are being conducted, our group will make all efforts to share databases with clinical trials with similar design, doses of hydroxychloroquine, and similar study endpoints, in an effort to answer the main question that initially fostered the design of this and similar studies. Preventing healthcare workers from COVID-19 is critical to control the pandemic. Thus, further studies in countries in the head of the pandemic are needed to break the transmission.  Flow diagram of trial participants at screening, recruitment and follow-up.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. CONSORTPrepBMC.doc