View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by LSHTM Research Online Comment 7 Revello MG, Lazzarotto T, Guerra B, et al. A randomized trial of 10 Leruez­Ville M, Ville Y. Is it time for routine prenatal serological screening hyperimmune globulin to prevent congenital cytomegalovirus. for congenital cytomegalovirus? Prenat Diagn 2020; published online N Engl J Med 2014; 370: 1316–26. May 27. https://doi.org/10.1002/pd.5757. 8 Shahar­Nissan K, Pardo J, Peled O, et al. Valaciclovir to prevent vertical 11 Griffiths PD. Burden of disease associatedwith human cytomegalovirus transmission of cytomegalovirus after maternal primary infection during and prospects for elimination by universal immunisation. pregnancy: a randomised, double­blind, placebo­controlled trial. Lancet Infect Dis 2012; 12: 790–98. Lancet 2020; 396: 779–85. 12 Leruez­Ville M, Ghout I, Bussières L, et al. In utero treatment of congenital 9 Bilavsky E, Pardo J, Attias J, et al. Clinical implications for children born with cytomegalovirus infection with valacyclovir in a multicenter, open­label, congenital cytomegalovirus infection following a negative amniocentesis. phase II study. Am J Obstet Gynecol 2016; 215: 462. Clin Infect Dis 2016; 63: 33–38. COVID-19 vaccine trials should seek worthwhile efficacy Three issues are crucial in planning COVID­19 vaccine even weaker vaccine as being non­inferior (so­called Published Online trials: (1) whether to demand not only proof of some bio­creep).1 August 27, 2020 https://doi.org/10.1016/ vaccine efficacy but also proof of worthwhile efficacy; The criteria used to define a successful vaccine in the S0140­6736(20)31821­3 (2) whether the initial trials of vaccine against placebo initial clinical trials of vaccination versus placebo should should prioritise not only single­vaccine trials but also therefore be strict enough to protect against the risk a multivaccine trial; and (3) whether to assess safety, of a weakly effective vaccine being deployed, especially protection against severe disease, and duration of since there are already many candidate vaccines against protection by continuing blinded follow­up of the COVID­19 to be tested,2 providing many chances to vaccine and placebo groups after definite evidence overestimate efficacy. Hence, the initial trials comparing of short­term efficacy has emerged, but before an COVID­19 vaccines versus placebo should seek reliable effective vaccine has been deployed locally in the general evidence not only of some efficacy but of worthwhile population. efficacy. The world needs efficient, speedy, and reliable evalu­ WHO recommends that successful vaccines should ation of many candidate vaccines against COVID­19. show an estimated risk reduction of at least one­half,3 There is a danger that political and economic pres­ with sufficient precision to conclude that the true vac­ sures for rapid introduction of a COVID­19 vaccine cine efficacy is greater than 30%. This means that the could lead to widespread deployment of a vaccine 95% CI for the trial result should exclude efficacy less that is in reality only weakly effective (eg, reducing than 30%. Current US Food and Drug Administration COVID­19 incidence by only 10–20%), perhaps guidance includes this lower limit of 30% as a because of a misleadingly promising result from criterion for vaccine licensure.4 As an example of an underpowered trial. Deploy ment of a weakly a result that would just satisfy these two criteria, effective vaccine could actually worsen the COVID­19 an evenly randomised trial with 50 cases arising pandemic if authorities wrongly assume it causes a in those vaccinated and 100 cases arising in those substantial reduction in risk, or if vaccinated indi­ given placebo would have a 95% CI that just excludes viduals wrongly believe they are immune, hence 30%, but would suggest 50% short­term efficacy. A reducing implementation of, or compliance with, vaccine that has 50% efficacy could appreciably reduce other COVID­19 control measures. Deployment of incidence of COVID­19 in vaccinated individuals, and a marginally effective vaccine could also interfere might provide useful herd immunity. Hence, although with the evaluation of other vaccines, as subsequent efficacy far greater than 50% would be better, efficacy vaccines would then have to be compared with it of about 50% would represent substantial progress. rather than with a placebo. For a vaccine superior to In comparison with individual trials for each of the the weakly effective vaccine, the increased sample size many different vaccines, a global multivaccine trial required could delay recognition of its efficacy. More with a shared control group could provide more rapid importantly, if the weak vaccine is compared against and reliable results. Additionally, its continuous use an even weaker vaccine, the statistical criteria used to of established clinical trial infrastructure could save analyse non­inferiority trials could well endorse the time and effort, accelerating the needed discovery of www.thelancet.com Vol 396 September 12, 2020 741 Comment Why have an international randomised controlled trial of several candidate vaccines? Evaluates several different Expeditiously enrols participants at Eliminates inefficiency of designing International collaboration and candidate vaccines sites with high rates of COVID-19 and conducting separate trials countries’ commitment Permits selected vaccines to Flexible mix of fixed sites and Each site helps assess several Fosters participation of sites with enter the trial whenever ready mobile (pop-up) sites vaccines in parallel high COVID-19 rates Vaccine selection for trial Sufficient enrolment to assess Shared placebo group increases Any effective vaccines will be tested assessed using a priori criteria efficacy and safety of all vaccines efficiency and attractiveness at many sites All vaccines selected for trial are Adaptive design accommodates If placebo can no longer be used, Paves the way for international eligible for testing at some sites unanticipated circumstances another vaccine becomes comparator deployment of effective vaccines Increases the likelihood of Rapid accumulation of data to Results within 3–6 months after Fosters international deployment finding several effective vaccines support rigorous evaluation each vaccine is ready for inclusion with equity of access Figure: Selected design features of the WHO Solidarity Vaccines Trial The primary outcome is laboratory­confirmed symptoms >14 days after vaccination is completed.Analyses of each vaccine after about 40, 70, and 100 primary outcomes occur in the placebo group will report success if they show ≤10 versus 40, ≤30 versus 70, or ≤50 versus 100 outcomes. The third analysis is reported regardless of its findings. In all cases placebo­controlled follow­up continues until at least month 12 (or local deployment of an effective vaccine) to assess safety, disease severity, and duration of protection. several safe and effective vaccines. High enrolment rates of multiple vaccines,10 helping to ensure that weakly facilitated by flexible trial design and hundreds of study effective vaccines are not deployed. The trial seeks to sites in high­incidence locations could yield results on achieve rapid, reliable results by the simplicity of the short­term efficacy for each vaccine within just a few trial design plus real­time checks on the quality of months of including that vaccine. the limited amount of data sought, facilitating high Reliable evidence is also needed about longer­term recruitment rates. A major challenge with vaccine efficacy, vaccine safety, and protection against severe trials at fixed study sites is that unexpectedly low COVID­19. Trials of sufficient size and duration are attack rates can delay progress. The WHO trial will needed to provide this, and to determine whether mitigate this by geographical diversity, recruiting in the vaccine can make COVID­19 more hazardous many high­incidence countries through fixed and (so­called disease enhancement).5,6 Trials that assess mobile (pop­up) research sites in localities where there only immunological endpoints cannot provide this are substantial COVID­19 attack rates at the time of evidence, and human challenge studies in young, enrolment. otherwise healthy, adult volunteers might not provide For a one­dose or two­dose vaccine that halves risk sufficient evidence of safety or efficacy in other the main result on short­term efficacy should emerge populations. Assessments of safety in multivaccine trials within 3–6 months, unless definite results for a highly can determine directly whether particular vaccines have effective vaccine emerge in interim analyses. Placebo­ adverse effects not shared by other vaccines. Evaluation controlled follow­up then continues until at least of multiple COVID­19 vaccines with standardised month 12, or until an effective vaccine is deployed methodology will facilitate regulatory and deployment locally. This approach increases the reliability of the decisions.7 Unless such decisions are informed by reliable evidence on younger and older adults, duration of randomised evidence, the effect on public acceptance protection, efficacy against severe disease, and any of COVID­19 vaccines could adversely affect COVID­19 disease enhancement. control and the uptake of vaccines against other Funders, vaccine developers, researchers, and gov­ diseases.8 ern ment institutions11 have signed an international The WHO Solidarity Vaccines Trial9 (figure)