Monkeypox: how will we know if the treatments work?

Clinical trials of treatments are a priority for emerging infectious disease outbreaks. When we design trials, fit-for-purpose endpoints are crucial, as these will inform decisions on case management, regulatory approval, and priority for public health funding and interventions.

Monkeypox highlights the difficulties in designing primary endpoints for emerging diseases. Globally, we have a limited understanding of what typical monkeypox is—the common and most severe symptoms, the symptoms that cause most distress to patients, the duration of infectivity, and potential complications. Furthermore, patterns of disease might vary, both between individuals and between different clades of virus. Descriptions of clade I disease emphasise disseminated rash and describe a mortality of around 10%.¹ Experience of clade IIb or III disease outside Africa suggests a predominance of genitourinary and perianal lesions,² with new complications (such as proctitis),² and there have been no deaths caused by these clades in the current outbreak. Although case ascertainment bias cannot be excluded, causes for milder disease need to be established and might be linked to the way the virus is being transmitted. Our primary motivations for treating monkeypox vary depending on severity and risk of transmission and, therefore, might shift focus between symptom relief, preventing complications, shortening the duration of patient isolation, or preventing spread of disease.

Our understanding of a disease grows with the number of cases and, in the field of emerging infections, by use of standardised clinical characterisation and biological sampling protocols.³ However, waiting for optimal clinical understanding before starting a trial is impractical—many outbreaks are short-lived (especially when working within the geographical borders of regulatory agencies) and we perpetually risk being too late, with the outbreak being declared over before the trial recruits.⁴

The challenging work to find primary outcomes that reflect the diversity of disease and meet the needs of patients, regulators, and public health officials is underway. It is uncertain whether one primary outcome will be feasible across trials to facilitate data sharing and synthesis through meta-analysis, or whether a range of trials with different outcomes might better meet these needs.

There are various proposed outcomes being considered for clinical trials evaluating treatment safety and efficacy for monkeypox (appendix p 1). The PALM 007 randomised controlled trial of tecovirimat in the Democratic Republic of the Congo will use time to monkeypox lesion resolution as its primary outcome. This outcome was determined by analysis of several years' worth of clinical data from patients in the Democratic Republic of the Congo with clade I disease⁵ and is appropriate for that context, but might be difficult to extrapolate to emerging disease phenotypes. In terms of pharmaceutical action, resolution of active (presumed infectious) lesions is a precise measure, but might not be representative for what is increasingly a polymorphic disease with other organ manifestations. Even so, there is no consensus on when a lesion is resolved—for example, whether a scab needs to be merely present, or have fallen off, or whether the underlying skin or mucosa must be fully healed. Complete lesion resolution is a more meaningful outcome for patients, but prolonged lesion presence might represent bacterial superinfection for which an antiviral treatment will not have a direct effect. Lesion assessment is likely to be prone to variation between clinicians reviewing patients.

Time to resolution of viral presence in blood, swab, or throat samples is particularly informative for infection control planning, but there is a paucity of longitudinal biological sampling to inform the use of these. Ordinal scale outcomes for disease severity might be possible, but these might not be precise if most cases are mild. For an affected individual, lesions having healed might be of little consequence if they have symptoms from a persistent ulcer. Some studies include exploratory outcomes to capture this phenomenon, but including resolution of such ulcerating lesions as a primary outcome might be more appropriate in some instances.

The way forward should be two-pronged. There are urgent deliberations at present (including those led by WHO) to focus on what is needed to safely commence recruitment in trials. These require the scientific community to reach a consensus over important definitions that will help to shape future research (such as what constitutes an active lesion, or a severe case, or a complication). Deliberations should harmonise where possible, but also facilitate exploration of the diversity of disease being observed and adapt with our growing understanding. We advocate that these are consolidated in the longer-term using strategies employed for other diseases (such as cutaneous leishmaniasis, which shares with monkeypox the issues of heterogenous skin lesions and definitions of resolution)⁶ that do due process to considerations such as patients' preferences for outcomes,⁷ and make use of further natural history and biological sampling evidence as it accrues.

All authors are investigators on the MOSAIC cohort study for monkeypox and the Expanded Access Programme of tecovirimat in the Central African Republic. JD is an investigator on the PALM 007 trial.