Two Doses, One Misleading Question
This series examines a specific question about two-dose JYNNEOS/MVA-BN vaccination: what does it mean when someone still develops mpox after completing the recommended series? The useful answer is not that the vaccine “worked” or “failed.” It is to understand which kind of protection is being measured, what a breakthrough infection reveals about the disease, and what the evidence can—and cannot—say about an individual’s future risk.
JYNNEOS is a replication-deficient orthopoxvirus vaccine approved as a two-dose series for preventing smallpox and mpox. A breakthrough infection is mpox diagnosed after a person has received the qualifying doses and enough time has passed for the series to count in surveillance. The existence of such cases is compatible with a vaccine that provides substantial protection. It becomes confusing only when protection is imagined as a switch: vaccinated or immune, exposed or safe, infected or uninfected. Real-world evidence asks a more useful question: how much does vaccination move the odds and consequences of an exposure?
Breakthrough infection is not the same as vaccine failure
The first reason two doses do not mean zero risk is simple: the virus can still reach a vaccinated person. Mpox continues to circulate, and vaccination does not remove every close or intimate exposure from a person’s life. A dose changes the likelihood that an exposure becomes infection; it does not create a physical barrier around the person or guarantee that every encounter with the virus will be blocked. That distinction is especially important when the disease is uncommon. A rare event can still occur, and a rare event can feel like a contradiction when people mistake “unlikely” for “impossible.”
The U.S. Centers for Disease Control and Prevention’s analysis of reported cases from May 2022 through May 2024 illustrates the difference. Among 32,819 probable or confirmed cases included in the comparison, 271 occurred in people who met the report’s definition of fully vaccinated—two JYNNEOS doses, with the second dose at least 14 days before illness onset. Those cases represented 0.8% of the analyzed cases. In a separate subset of jurisdictions with usable vaccination-coverage data, 187 infections were reported among 188,907 fully vaccinated people, an estimated 0.1% infection rate for that subset and period.
Those figures are not interchangeable. The first asks, “What share of analyzed cases occurred in fully vaccinated people?” The second asks, “How many infections were observed among a vaccinated population?” Neither is a personal forecast. The numbers depend on which cases were detected, whether vaccination status was recorded, how the population at risk was estimated, and how much exposure people experienced. They do, however, establish the central tension: breakthrough infection was uncommon in the available vaccinated denominator, but it was not absent.
The distinction also explains why a small number of breakthrough reports cannot be read in isolation. A jurisdiction with more complete records may appear to have more vaccinated cases simply because status is being captured more reliably. A jurisdiction with missing vaccination fields may conceal both vaccinated and unvaccinated cases. Likewise, a vaccinated population with more people protected can still produce some breakthrough reports because the population is larger and because exposure continues. The CDC analysis had missing vaccination status for thousands of reported cases, so the observed count is neither a complete inventory of breakthroughs nor evidence that the missing cases all belonged to one group. A careful reader keeps the numerator, the denominator, and the missingness beside one another.
That tension is more informative than a headline claiming that the vaccine either stops mpox completely or does not work. If a vaccinated person becomes infected, the relevant comparison is not only whether infection happened. It is also what the infection does afterward. A vaccine can be valuable before the virus arrives and still be valuable after it gets past the first line of defense.
Protection has three different endpoints
“Risk” is too broad a word to carry this discussion by itself. For two-dose JYNNEOS/MVA-BN vaccination, at least three endpoints need to be separated.
The first is the chance of infection: after an exposure, how likely is a person to develop mpox? This is the endpoint people usually mean when they say a vaccine prevents disease. It is also the endpoint most likely to produce an overly simple expectation of zero. The U.S. surveillance data show that infections after a completed series were rare, but the existence of reported cases shows that prevention is not absolute.
The second is severity conditional on infection: if a vaccinated person does develop mpox, how many lesions appear, how widely do they spread, which mucosal sites are involved, and how much pain or treatment is required? The 2024 CDC report found that fully vaccinated patients with mpox generally had less systemic illness and a median of one rash location compared with four among unvaccinated patients. The Swedish clinical case series gives the same idea a more detailed shape. Among 30 patients vaccinated with MVA-BN only, the median mpox Severity Score was 5.0, compared with 8.0 among 32 unvaccinated patients. The score combines lesion burden, anatomical spread, mucosal involvement, bacterial superinfection, level of care, and analgesia needs.
The third is the risk of a severe outcome, such as hospitalization or death. In the U.S. case comparison, hospitalization was reported for 1.4% of fully vaccinated patients with known outcome data and 8.4% of unvaccinated patients with known outcome data. No deaths were reported among fully vaccinated cases with available death information, while deaths were reported among unvaccinated cases. A newer U.S. surveillance analysis covering 2024–2025 also found a substantially higher adjusted odds of hospitalization among unvaccinated than fully vaccinated patients in a complete-data subset, after accounting for age group and HIV status.
These endpoints are connected, but they are not the same. Lower infection risk does not mean every infection is prevented. Lower average severity does not mean every breakthrough is mild. Lower hospitalization odds do not mean hospitalization is impossible. The benefit can be real at all three stages while each stage still contains exceptions.
This is why a breakthrough case should not be read as a verdict on the vaccine. It is a case in the part of the distribution that remains after risk has been reduced. The person may still have painful lesions or need clinical care, but the average pattern can be less extensive, less complicated, or less likely to require hospitalization than it would have been without vaccination. “Less likely” and “less severe” are meaningful outcomes even when the word “never” is unavailable.
The distribution can also change in ways that are not uniformly reassuring or uniform across symptoms. In the U.S. report, genital rash was reported more often among fully vaccinated cases even though the median number of rash locations was lower. That does not contradict a lower overall burden: a disease can be more concentrated in one location while being less widespread. The point is not that every symptom moves in the same direction, but that the clinically important pattern is multidimensional. Lesion count, anatomical spread, systemic illness, pain, complications, care level, and death each capture a different part of what protection may change. A single breakthrough symptom therefore cannot stand in for the whole outcome.
This layered view also protects against a common misunderstanding of the word “effective.” Prevention and mitigation are not consolation prizes for a vaccine that missed its target. They are different ways an intervention can reduce harm. Avoiding infection is usually the most visible benefit, but preventing a severe course among the infections that remain can matter to the patient, the need for antibiotics or pain treatment, and the capacity of hospitals. The evidence must be described at the endpoint it actually measured.
Three kinds of evidence answer one practical question
The sources do not all measure protection in the same way, and that is a strength rather than a flaw. National CDC surveillance is broad. It can show how many cases were reported, how vaccination status was distributed, what symptoms and outcomes were recorded, and how patterns changed over time. Its weakness is that reporting is incomplete: vaccination status, exposure history, symptoms, and outcomes may be missing, and mild infections that never reach testing can go unseen.
The Swedish study is narrower but more clinically detailed. It followed laboratory-confirmed cases at a specialized HIV and sexual-health center in Stockholm. Two clinicians who did not know the patients’ vaccination status assigned a validated severity score. This makes “milder disease” more concrete than a single case count. It also limits the conclusion. The study included 79 patients, all MVA-BN doses were given intradermally, and there was no direct comparison with subcutaneous vaccination. Its results describe meaningful attenuation in an observed clinical setting; they do not provide a universal prediction for every vaccinated person.
The newer U.S. surveillance report and its supplementary hospitalization table add a further layer. They place breakthrough cases in an environment where low-level transmission persisted through 2024 and 2025, then compare hospitalization by vaccination status while adjusting for age and HIV status. That analysis is useful because hospitalization matters to patients and health systems. It is still an association from surveillance data, not a randomized experiment. Only cases with complete information entered the model, and the confidence interval around the adjusted estimate was wide.
The kinds of evidence that would change the judgment are therefore clear. A direct, prospective comparison of intradermal and subcutaneous schedules could clarify route questions. More complete vaccination, exposure, HIV, and hospitalization records could test whether the observed patterns are distorted by who was documented or sought care. Longer follow-up with immune measurements could distinguish a true change in protection from the timing of an outbreak or vaccination campaign. Until those comparisons exist, the strongest claim is the one shared by the sources: the two-dose series provides meaningful protection, while the remaining cases and the gaps in the data keep the conclusion conditional.
Taken together, the evidence supports a measured conclusion. Two doses of JYNNEOS/MVA-BN are not a promise that a vaccinated person can ignore exposure or symptoms. They are a major risk-reducing intervention whose value appears in more than one place: fewer infections in the vaccinated population, less severe disease among many people with breakthrough infection, and lower observed hospitalization risk. The size and certainty of each benefit depend on the population, the route of vaccination, the time window, the host factors measured, and the quality of the records.
The practical mistake is to ask a binary question of a layered intervention. “Did the vaccine guarantee that this person would never become infected?” is a question with a no-like answer for almost any real-world vaccine. “Did two doses lower the chance of infection and improve the expected course if infection occurred?” is the question these sources can address. To read the numbers responsibly, the next step is to define exactly what the reports mean by a completed two-dose series—and which denominator makes each claim meaningful.