Review 2: "Optimal Deployment of Limited Vaccine Supplies to Combat Mpox"

RR\ID Evidence Scale rating by reviewer:

• Potentially informative. The main claims made are not strongly justified by the methods and data, but may yield some insight. The results and conclusions of the study may resemble those from the hypothetical ideal study, but there is substantial room for doubt. Decision-makers should consider this evidence only with a thorough understanding of its weaknesses, alongside other evidence and theory. Decision-makers should not consider this actionable, unless the weaknesses are clearly understood and there is other theory and evidence to further support it.

***************************************

Review: Berry et al.'s study "Optimal Deployment of Limited Vaccine Supplies to Combat Mpox" builds upon their earlier study, "Predicting Vaccine Effectiveness for Mpox" (April 2024, Nature Communications). In the earlier work, the authors conducted a meta-analysis of existing literature to model vaccine effectiveness for mpox after the first dose. They compared four vaccination strategies: a single dose, a second dose administered after the recommended four weeks, a second dose administered after two years, and a three-dose regimen (one after four weeks and one after two years). Their findings suggested that delaying the second dose for two years offers the best long-term protection, with the three-dose strategy providing only marginal additional benefit.

In "Optimal Deployment of Limited Vaccine Supplies to Combat Mpox," Berry at al. extend their analysis to address the question of how to distribute limited vaccine supplies in regions with high incidence of mpox. The study is positioned in the context of an ongoing clade I mpox outbreak in the Democratic Republic of Congo (DRC) where progress on vaccine distribution has been slow. Due to the lack of clade I-specific vaccine data, the authors base their analysis on vaccine effectiveness data from clade II outbreaks. They cite a WHO report referencing two studies that suggest cross-protection of the smallpox vaccine against clade II mpox. While it seems reasonable to assume that a clade II vaccine could similarly provide cross-protection against clade I mpox, there is currently no empirical evidence to support this hypothesis. Furthermore, the epidemiological and severity differences between clade I and clade II mpox are substantial—clade II is primarily spread among men who have sex with men, while clade I affects a broader population and carries a significantly higher fatality rate - see Rivers et al. "The Resurgence of Mpox in Africa," JAMA, August 2024. While these assumptions are a practical necessity given the absence of clade I-specific data, epidemiological differences could limit the generalizability of their conclusions in the context of the current outbreak in the DRC.

The authors investigate optimal allocation strategies, considering whether to prioritize unvaccinated individuals or those awaiting a second dose. Their analysis considers factors such as the risk of infection among unvaccinated individuals, the time elapsed since initial doses were administered, and the severity of potential infections. As in their earlier study, the authors conclude that for most combinations of parameters, the ratio of cases averted is higher when the second dose is delayed beyond the recommended four weeks, especially if the additional doses are redirected to unvaccinated individuals. Importantly, however, the authors emphasize that their results should not be interpreted as advice to contradict vaccine manufacturers' recommendations.

Despite its valuable insights, this paper relies on a series of modeling assumptions, raising questions about the robustness of its conclusions. In their prior study, vaccine effectiveness was estimated from a meta-analysis of studies measuring antibody levels after vaccination. Most of these studies measured short-term responses (weeks to months), with only two providing data at 1-year and 2-year intervals. To evaluate longer vaccine delays (e.g., 6 months or 1 year), the present study extrapolates these estimates, despite the limited empirical data for these timeframes. This extrapolation introduces uncertainty into the predicted outcomes, particularly when applied to real-world vaccination strategies.

One key conclusion the authors make is that "deploying limited vaccine stocks as single doses to as many people as possible is always favored.'' However, we believe this should be interpreted as "usually favored," as there are notable exceptions. For instance, when initial vaccine doses have been administered to high-risk individuals, their study found that it is often optimal to provide these individuals with the second dose rather than vaccinating new, lower-risk individuals—especially when the ratio of high-risk to low-risk individuals is heavily skewed. Since no context is provided - and likely none is available - regarding the risk ratio for clade I mpox, these borderline cases become particularly important.

While the reliance on extrapolated estimates introduces uncertainty, it also underscores the importance of modeling efforts in addressing pressing public health questions when empirical data are scarce. This study provides valuable insight into optimizing vaccine distribution and may help guide decision-making in the absence of complete data. Future studies that validate the time-dependent vaccine effectiveness function, especially those focused on clade I mpox, would further enhance the robustness of these findings and better inform global vaccination strategies.