Chemoprophylaxis trial designs in epidemics: insights from a systematic review of COVID-19 study registrations.
Manoharan L., Olliaro P., Horby PW., Watson CH.
BACKGROUND: Chemoprophylactics against emerging epidemic and pandemic infectious diseases offer potential for prevention but require efficacy and safety analysis before widespread use can be recommended. Chemoprophylaxis with repurposed drugs enables deployment ahead of development of novel vaccines. It may have particular utility as a stopgap ahead of vaccine deployment or when vaccines become less effective on virus variants, in countries where there may be structural inaccessibility to vaccines or in specific risk-groups. Rapid implementation of robust trial designs is a persistent challenge in epidemics. We systematically reviewed SARS-CoV-2 and COVID-19 chemoprophylaxis trial registrations from the first 21 weeks of the pandemic to critically appraise significant design features and alignment of study populations to clinical and public health uses, and describe candidate chemoprophylactic agents. METHODS: We searched online international trial databases from 31 Dec 2019 to 26 May 2020 using keywords "proph*" or "prevention". Trial protocols assessing efficacy of chemoprophylactic agents for COVID-19 were included. Trial components were screened for eligibility and relevant studies extracted. Key trial design features were assessed. RESULTS: We found 76 chemoprophylaxis study registrations, proposing enrolment of 208,367 people with median size of 490 (IQR 262-1710). A randomised design was specified for 63 trials, 61 included a control group and total proposed enrolment size was 197,010, median 600 (IQR 236-1834). Four protocols provided information on effect size sought. We estimate that for a control group attack rate of 10%, 66% of trials would be underpowered to detect a 50% effect size, and 97% of trials would be underpowered to detect a 30% effect size (at the 80% level). We found evidence of adaptive design in one trial registration only. Laboratory-confirmed infection with or without symptoms was the most common primary outcome. Polymerase chain reaction testing alone was used in 46% of trials, serological testing in 6.6% and 14.5% used both testing methods. Healthcare workers were the target population in 52/79 (65.8%) trials: 49 pre-exposure prophylaxis (PrEP) and 3 post-exposure prophylaxis (PEP). Sixteen trials (20.3%) planned PEP in close contacts. Five studies (6.3%) considered chemoprophylaxis in clinical-risk patients. Older adults were the focus of recruitment in only 3 (3.8%) studies (all long-term care facilities). Two (2.5%) studies of PrEP in the general population included older adults. Hydroxychloroquine was the most common candidate agent in 55/79 trials (69.6%), followed by chloroquine (4/79, 5.0%) and lopinavir/ritonavir (3/79, 3.8%). CONCLUSION: Many registered COVID-19 chemoprophylaxis efficacy trials were underpowered to detect clinically meaningful protection at epidemiologically informed attack rates. This, compounded with the time that has taken to organise these trials as compared to the rapid development of COVID-19 vaccines, has rendered these trials of marginal importance. International coordination mechanisms and collaboration is required. Supporting the design of feasible chemoprophylaxis trials, large enough to generate strong evidence, early on in an epidemic using adaptive platform trial designs will allow structured entry and exit of candidate agents and rapid stand-up of trial infrastructure. REVIEW PROTOCOL REGISTRATION: Our protocol is registered at https://www.osf.io/vp56f on May 20, 2020.