key: cord-0705775-tr9659vp
authors: Steuwer, Bastian; Jamrozik, Euzebiusz; Eyal, Nir
title: Prioritizing between second-generation SARS-CoV-2 vaccines through low-dosage challenge studies
date: 2021-02-13
journal: Int J Infect Dis
DOI: 10.1016/j.ijid.2021.02.038
sha: ca6260edf696cbe021c89c3cb932156e79f97b41
doc_id: 705775
cord_uid: tr9659vp

Design of human challenge studies balances scientific validity, efficiency, and study safety. This Perspective explores some advantages and disadvantages of “low-dosage” challenge studies, in the setting of testing second-generation vaccines against COVID-19. Compared to a conventional vaccine challenge, a low-dosage vaccine challenge would be likelier to start, and start earlier. A low-dosage challenge would also be less likely to rule out a vaccine candidate which would have been potentially effective in target usage. A key ethical advantage of a low-dosage challenge over conventional challenge is that both it and its dose escalation process are safer for each participant. Low-dosage studies usually require larger numbers of participants than conventional challenges, but this and other potential disadvantages are less serious than they may initially appear. Overall, low-dosage challenges should be considered for certain roles, such as prioritizing between second-generation vaccines against COVID-19.

In controlled human infection studies (herein, "challenge studies") research participants, randomized into a treatment and a control group, are deliberately infected so that researchers can test interventions or investigate other scientific hypotheses. The UK government has announced its support for challenge studies, to begin in 2021 (Callaway, 2020) . First-generation SARS-CoV-2 vaccines have high efficacy in protecting against disease (Grady, 2020) , but it is not yet known how much these vaccines will block asymptomatic infection and/or transmission, nor how long immunity will last (Grady, 2020) . By enabling control of the timing of infection, challenge studies excel at demonstrating vaccine efficacy against asymptomatic infection and transmission, as well as the duration of vaccine protection, and generating other information for improved control measures (Douglas and Hill, 2020; Levine et al., 2020; . Challenge studies may also help to refine dosing regimens of existing vaccines and determine the degree of protection against particular viral variants.

This perspective explores the ethical and pragmatic advantages of "low-dosage challenge studies".

This study design reduces risk to individual volunteers, and may also permit vaccine challenge studies to begin earlier and with greater likelihood after the commencement of initial viral dose escalation studies.

A low-dosage challenge study here refers to the use of a lower dose of wild-type virus (as opposed to an attenuated strain) than required in conventional challenge studies. In standard challenge study designs, the viral dose used is calculated to cause a sufficient proportion of participants in the J o u r n a l P r e -p r o o f placebo arm to reach one or more endpoints. Such endpoints can be either disease (e.g. determined by the presence of influenza-like illness), often expressed as "clinical attack risk", or, alternatively, infection (determined by a virological endpoint such as viral load). A WHO group on SARS-CoV-2 challenge studies that included one of us deemed "a 70% clinical attack risk for mild upper respiratory illness accompanied by shedding of SARS-CoV-2" as a sufficiently high proportion for an efficient vaccine challenge study (World Health Organization, 2020) .

Instead of a dosage aimed at 70% risk of influenza-like illness (ILI) in the placebo arm, a lowdosage SARS-CoV-2 challenge could aim at, for example, 35% risk of ILI in the placebo arm. Or it could aim for even lower disease rates (a quarter the clinical attack risk, a tenth of it, and so forth). Or it could aim for a lower infection ratio in the placebo arm, regardless of the clinical attack risk. This option appears suited to the role of prioritizing second-generation vaccines in terms of their impact in blocking infection now that first-generation vaccines reportedly protect against disease. By contrast, a SARS-CoV-2 clinical disease model with a 70% ILI risk would require infecting more than 70% of participants since some infected participants would remain asymptomatic.

Importantly, to achieve the same statistical power as a conventional challenge study, a low-dosage challenge study would require a larger number of volunteers. The lower the dosage, the more volunteers would be needed. With adequate numbers of volunteers, a low-dosage challenge would have similar statistical power to a conventional challenge study.

In order to determine what (if any) viral dose meets the required endpoint target, challenge studies are nearly always preceded by a dose escalation study. Escalation begins by exposing a few volunteers to increasing doses of virus starting at a low dosage and ends if either (a) there are serious adverse events or (b) there is a sufficiently high rate of the endpoint (using our illustration, 35% of participants experiencing ILI or comparably many infections) being met. If neither occurs, dose escalation continuesfurther volunteers are exposed to a higher dose.

Others have recently explored different potential scientific applications of low-dosage challenge studies to COVID-19, for a host of non-vaccine related uses (Hausdorff and Flores, 2020) . We explore the advantages and potential disadvantages of low-dosage challenge over conventional challenge designs in the particular role of developing second-generation vaccines: as an initial efficient test of experimental vaccine efficacy that determines which candidates proceed to standard phase 3 testing (herein, 'field trials') (Jamrozik and Selgelid, 2020; Douglas and Hill, 2020; Levine et al., 2020) . Those are summarized in Table 1 below.

One worry concerning challenge studies is that there will be serious adverse events (SAEs) in the preceding dose escalation study before it reaches its desired endpoint. A low-dosage challenge would reduce the probability of this occurring, insofar as SAEs are less likely with a lower clinical endpoint target than that of a conventional challenge.

Even when all goes well in the dose escalation, that preliminary process typically takes many weeks and delays the start of vaccine challenge studies (Cohen, 2020; Deming et al., 2020) . A dose escalation study with an endpoint reachable with fewer rounds of titration could save many weeks.

Even in environments that put individuals at high probability of SARS-CoV-2 infection (e.g.

sharing a berth on a ship, sitting next to an infectious train passenger for many hours), it is rare for ≥ 70% of exposed individuals to get infected and become symptomatic (Liu et al., 2020; Payne et al., 2020; Hu et al., 2020; Prentiss et al., 2020; Popa et al., 2020) . This is even less likely for young and healthy individuals as challenge trial participants would be. So the vast majority of community-acquired infections may well be due to viral doses lower than the dose that would meet WHO's 70% ILI recommendation in a challenge trial's placebo arm. Low-dosage challenges may therefore be less liable to prematurely rule out vaccines with some degree of efficacy (that is, less liable to a type II error). Based on limited evidence, some vaccines seem to prevent infection via small viral exposures yet fail against larger viral exposures (Langwig et al., 2019) . A vaccine that prevented a significant fraction of community-acquired infections would have powerful herd immunity effects if widely administered, even if it were less efficacious at preventing (higherdosage) infection or severe disease. Such a vaccine would be especially important to identify should high levels of herd immunity prove elusive with first-generation vaccines.

One approach to study program design would be a 'branching model', wherein different challenge studies are performed for different purposes with a single preliminary dose escalation study. Once J o u r n a l P r e -p r o o f a relatively low endpoint is met, the dose escalation study identifies the dose for a first, low-dosage challenge. After this first step, dose titration continues for a second higher dosage challenge study.

The results of both challenge trials together would provide more comprehensive information about potential vaccine performance. Unlike a pure low-dosage challenge, further escalation permits testing vaccine efficacy against higher exposures. Unlike a pure high-dosage challenge, vaccines that may be efficacious against some exposures would not be prematurely ruled out. Alternatively, a higher dose might be necessary for challenge studies to determine clinically useful immune correlates of protection. A further advantage of combining low-and high-dosage studies would be to test how seropositive individuals react to vaccination. A branching model could for example test the hypothesis that vaccine boosting after past exposure results in greater overall protection.

The coordination of these different challenge studies in a branching model could therefore increase efficiency and allow a wider array of hypotheses to be tested. A coordinated branching model may also permit the standardization of protocols for multi-site challenge studies. Such standardization could help to improve the external validity of results, increase overall statistical power, and reduce the wastage of scientific resources (Roestenberg et al., 2012) .

The main ethical advantage of a low-dosage challenge over a conventional challenge is the lower risk for each participant, in two ways. First, a lower viral dose is less likely to infect each individual participant. Thus, each person entering the study faces a lower probability for severe outcomes.

Second, there may also be a positive dose-response relation between viral dose and the severity of disease-a subject of ongoing scientific debate. There is strong evidence from measles (Perry and Halsey, 2004) and cholera (Cash et al., 1974) , and weak evidence from SARS-CoV-1 (Chu and et al., 2005) , and SARS-CoV-2 (Imai and et al., 2020; Rubin and Offit, 2020; Gandhi and Rutherford, 2020; Bielicki and et al., 2020) that higher viral dosage causes more severe outcomes. In contrast, an influenza challenge study found no statistically significant dose-severity relationship (Memoli et al., 2015) . Overall, there is no scientific reason to think that lower viral exposure could, upon infection, make COVID-19 more severe, and good reasons to think that it could make it milder. In the face of continuing uncertainty, any reasonable decision-maker should assign some credence to this possibility.

That a low-dosage challenge is safer to each participant than a conventional challenge study addresses concerns that challenge studies must be too risky, even with careful participant selection and provision of all available therapeutics (Deming et al., 2020; Rosenblatt, 2020; Shah et al., 2020; Kahn et al., 2020) . Such concerns are naturally understood to refer to risks to individual participants (Shah et al., 2020) , a classical focus of research ethicists (Rid and Wendler, 2011; Miller and Joffe, 2009; Resnik, 2012; London, 2009 ) and documents on human subjects protection (Rid, 2014 In principle, one could use progressively-lower doses to satisfy lower tolerances for risk to individual participants. Instead of ½ the risk of ILI (e.g., as suggested earlier, 35% instead of 70%), a lower individual dose could potentially aim at 1/10 th that risk (with suitably many more volunteers), or at 1/100 th (ditto), and so forth. Since one could reduce dosage as much as desired, J o u r n a l P r e -p r o o f some low-dosage challenge design should reduce risks such that no critic could dismiss it as too dangerous for individual participants.

In all vaccine challenge studies, efficacy against severe disease is harder to detect than in vaccine field trials-precisely because of challenge studies' careful risk reduction strategies, including the exclusion of higher risk groups (Corey et al., 2020) . This limitation may seem to be accentuated in low-dosage challenge studies, which go to even greater lengths to prevent severe disease in the cohort. However, this limitation does not affect the usage of challenge studies to assess impact on infection rather than severe disease (as discussed above) and some other uses (Hausdorff and Flores, 2020) .

Another seeming disadvantage of low-dosage challenges compared to conventional ones is their potential higher likelihood of selecting experimental vaccines that are insufficiently protective against high-dosage infections-the flipside of the advantage noted earlier and a type I error (Langwig et al., 2019) . However, the doses associated with conventional challenge studies may be rare in target usage (as argued above). Even if higher dose exposures are common, relative inefficacy against high-dosage exposures would not necessarily undermine low-dosage challenge in their role of weeding out unpromising vaccine candidates in advance of field trials. The low-J o u r n a l P r e -p r o o f dosage challenge would help identify, for further study, those vaccine candidates that prevent infection against at least a low dosage. Field trials could then confirm efficacy in target usage.

Hosting a larger number of volunteers for an extended period in isolation, and consequently needing more staff hours or larger facilities, as low-dosage challenge require, would cost more.

However, these economic costs appear to be outweighed by the benefits of enhanced diseased control. Moreover, significant funds are now available for SARS-CoV-2 vaccine development, including challenge studies (Callaway, 2020) .

Finally, some concerns regarding excessive risk in challenge studies relate to something other than the risk to each participant, but the risk that someone in the cohort of challenge participants will have a serious adverse event: "A single death or severe illness in an otherwise healthy volunteer would be unconscionable..." (Deming et al., 2020; Guarino and Johnson, 2020) . Nothing is said here about the probability of death or serious injury for any particular study participant.

First, it is unknown whether risk for the cohort is larger in a low-dosage challenge than in a conventional challenge. In our illustration of low dosage above, the risk of ILI per individual is ½ of what it would be in a conventional challenge, but for statistical reasons at least twice the number of volunteers would be needed. In that respect, the overall risk of SARS-CoV-2 infection in the cohort is higher in a low-dosage challenge. Nonetheless, following any particular infection, the J o u r n a l P r e -p r o o f individual risk of severe COVID may remain significantly smaller at lower viral doses (again, a matter currently unknown). Combined, these two effects may increase or decrease overall cohort risk.

Second, either a low-dosage challenge or a conventional challenge study should be deemed cumulatively safe enough for study cohorts, insofar as their cumulative risk is on a par with that of vaccine field trials, which are widely accepted. In particular, vaccine field trials vaccinate >100 times more participants than conventional challenge studies, and require more viral exposures (Heriot and Jamrozik, 2020) . Further, vaccine field trials include risks of other adverse events from vaccine side-effects and from increased risk behavior. Adding up these risks over the substantially larger cohort (tens of thousands of participants), and considering poorly characterized risks of disease severity enhancement (Peeples, 2020; Corey et al., 2020) , which can take place in thousands of participants long after the study (Eyal and Lipsitch, 2021) , a vaccine field trial may generate a greater cumulative risk of serious adverse events. All in all, a vaccine field trial is not clearly less risky for the cohort than challenge studies.

When using a challenge study to determine which vaccine candidates should be selected for further testing, low-dosage challenges can remove the main inherent barrier to conducting challenge studies, namely, the potential inexistence of a dose that is both safe enough and enables a study to reach useful endpoints. Additionally, low-dosage challenges permit decision-makers to proceed with vaccine challenge studies at an earlier point in dose escalation. They may mimic some target J o u r n a l P r e -p r o o f vaccine indications much better than conventional challenge studies. Low-dosage challenge studies would be less risky for each participant, as would their dose escalation processes.

The disadvantages of low-dosage challenge studies over conventional challenge studies for triaging vaccine candidates may be less significant than they appear. Therefore, low-dosage challenges may be a useful tool in expediting second-generation SARS-CoV-2 vaccine testing, and specifically in providing estimates of candidate vaccine impacts on infection to select promising candidates for further investigation.

J o u r n a l P r e -p r o o f Conflict of interest: EJ previously served on two WHO working groups on SARS-CoV-2 challenge studies. NE serves on the advisory board for challenge trial intended-volunteer group 1DaySooner, an unpaid position.

Conflict of interest: EJ previously served on two WHO working groups on SARS-CoV-2 challenge studies. NE serves on the advisory board for challenge trial intended-volunteer group 1DaySooner, an unpaid position.

Authors' assessment

Shorter time to start ++ Likelihood to start ++ Limiting financial cost and logistics -Less acute, given high investment in COVID vaccine research.

Avoiding the type II errors of ruling out efficacious vaccines ++ Avoiding type I errors of not ruling out inefficacious vaccines 0 / -Can be avoided if study is followed by a field trial, or by testing in a branching model together with high-dosage challenges.

External validity 0 / + High when there is standardization across multiple sites.

Reducing risk to individuals ++ Risk reduction due to reduced infection risk and (possibly) doseseverity relationship. Reducing risk to cohort 0 Unclear whether risk to cohort is higher or lower than conventional challenge trial; commensurate cohort risks with field trial.

Microbial Challenge Studies of Human Volunteers-A Guidance Document from the Academy of Medical Sciences

Social Distancing Alters the Clinical Course of COVID-19 in Young Adults: A Comparative Cohort Study

Dozens to be deliberately infected with coronavirus in UK 'human challenge' trials

Response of Man to Infection with Vibrio cholerae. I. Clinical, Serologic, and Bacteriologic Responses to a Known Inoculum

Viral Load Distribution in SARS Outbreak

Council for International Organizations of Medical Sciences, World Health Organization. International Ethical Guidelines for Health-related Research Involving Humans

Speed coronavirus vaccine testing by deliberately infecting volunteers? Not so fast, some scientists warn

A strategic approach to COVID-19 vaccine R&D

Accelerating Development of SARS-CoV-2 Vaccines -The Role for Controlled Human Infection Models

Immunological considerations for SARS-CoV-2 human challenge studies

Human challenge trials of covid-19 vaccines still have much to teach us

Testing SARS-CoV-2 vaccine efficacy through deliberate natural viral exposure

Facial Masking for COVID-19 -Potential for "Variolation" as We Await a Vaccine

Early Data Show Moderna's Coronavirus Vaccine Is 94.5% Effective

Volunteers sign up to put their lives on the line for a coronavirus vaccine

Low dose exposures to SARS-CoV-2 to understand and prevent severe COVID-19

Not in my backyard? COVID-19 vaccine development requires someone to be infected somewhere

Risk of Coronavirus Disease 2019 Transmission in Train Passengers: an Epidemiological and Modeling Study

Syrian hamsters as a small animal model for SARS-CoV-2 infection and countermeasure development

COVID-19 human challenge studies: ethical issues

Opinion: For now, it's unethical to use human challenge studies for SARS-CoV-2 vaccine development

Limited available evidence supports theoretical predictions of reduced vaccine efficacy at higher exposure dose

Viewpoint of a WHO Advisory Group Tasked to Consider Establishing a Closely-Monitored Challenge Model of COVID-19 in Healthy Volunteers

Secondary attack rate and superspreading events for SARS-CoV-2

Clinical research in a Public Health Crisis: the Integrative Approach to Managing Uncertainty and Mitigating Conflict

Validation of the Wild-type Influenza A Human Challenge Model H1N1pdMIST: An A(H1N1)

Dose-Finding Investigational New Drug Study

Limits to research risks

SARS-CoV-2 Infections and Serologic Responses from a Sample of

Avoiding pitfalls in the pursuit of a COVID-19 vaccine

The Clinical Significance of Measles: A Review

Genomic epidemiology of superspreading events in Austria reveals mutational dynamics and transmission properties of SARS-CoV-2

Superspreading Events Without Superspreaders: Using High Attack Rate Events to Estimate Nº for Airborne Transmission of COVID-19

Limits on Risks for Healthy Volunteers in Biomedical Research

A Framework for Risk-Benefit Evaluations in Biomedical Research

Setting risk thresholds in biomedical research: lessons from the debate about minimal risk

Comparison of Clinical and Parasitological Data from Controlled Human Malaria Infection Trials

Human challenge trials with live coronavirus aren't the answer to a Covid-19 vaccine

We Know Crowding Affects the Spread. It May Affect the Death Rate

Ethics of controlled human infection to study COVID-19

Potential Value and Limitations of Establishing a Closely Monitored Challenge Model of Experimental COVID-19 Infection and Illness in Healthy Young Adult Volunteers