What Two Doses Actually Change
For two-dose JYNNEOS/MVA-BN vaccination, the phrase fully vaccinated is not a synonym for “cannot be infected.” It is a time-qualified category used to decide which cases can be compared with unvaccinated cases. Understanding that category—and the denominator attached to it—is essential for judging what the vaccine changed when mpox occurred anyway.
The distinction may sound technical, but it answers a practical question. If a person received a dose yesterday, a second dose last week, or two doses many months ago, those situations do not carry the same evidentiary meaning. Nor does “271 vaccinated cases” mean the same thing as “271 infections among vaccinated people.” Before looking at a result, we need to know which cases entered the count and what population the count is being measured against.
A completed series is a time-qualified category
In the CDC analysis of U.S. cases from May 2022 through May 2024, a fully vaccinated case was a probable or confirmed mpox case in someone with documented receipt of two JYNNEOS doses. The second dose had to have been given at least 14 days before illness onset, and the vaccination dates had to fall within the period covered by the analysis. A person with only one recorded dose was not put in the same category. Neither was someone whose second dose occurred less than 14 days before illness began.
The 14-day interval is an important part of the definition. CDC treated a dose as reaching maximum immunogenicity 14 days after administration for the purpose of classifying cases. That does not mean immunity switches on at the exact end of the fourteenth day, or that the body’s response is identical in every person. It means the surveillance analysis needed a consistent boundary. Without one, a case diagnosed immediately after vaccination could be misread as evidence about a fully established series when the series had not had the time the analysis considered necessary.
The newer U.S. surveillance report uses a closely related definition, adding that at least 24 days had to separate the first and second doses. Its supplementary hospitalization table repeats that rule and excludes partially vaccinated people from the multivariable estimate because the number of outcomes was too small for a stable comparison. These definitions are similar, but not identical in every detail. A careful reader should preserve the time window used by each report rather than silently combining their case totals as though they came from one identical study.
This is the first classification link in the evidence chain:
- A vaccination record establishes how many doses were received.
- The time between doses and the time from the last dose to illness determine whether the record fits the report’s “fully vaccinated” category.
- Only then can the case be included in a comparison of breakthrough and unvaccinated disease.
That sequence protects the comparison from an obvious distortion. If someone becomes ill before the second dose has had time to take effect, the case may say something about infection during an incomplete or newly completed schedule. It cannot cleanly answer what happens after a completed, time-qualified series. Excluding that case from the fully vaccinated group is not hiding a failure; it is keeping different starting conditions from being treated as the same.
The calendar rule also tells us how to interpret a case near the boundary. A person who becomes ill 10 days after dose two may genuinely have received two doses, but that person does not fit the same analytic starting point as someone whose illness began months later. The rule cannot measure the person’s immune response directly; it creates a consistent comparison by separating “the series was recently completed” from “the series had the time the analysis specified.” This is a familiar problem in observational evidence: researchers need a rule that is simple enough to apply across thousands of records, even when biology changes gradually rather than on one exact day.
It is also possible for two reports to use slightly different rules and still be asking related questions. The U.S. 2022–2024 report and the 2024–2025 report both require two doses and at least 14 days after the second dose, but the newer report also specifies a minimum interval between doses. That added condition does not make the older analysis invalid. It means the reader must keep each report’s inclusion rule attached to its results. Combining results without carrying those rules forward would make the apparent precision of the comparison greater than the evidence warrants.
The category also carries information about what it does not establish. It does not identify the exact exposure that led to infection. It does not prove that the person’s immune response was identical to the response of everyone else in the category. It does not tell us whether the person had unusually intense exposure, an unmeasured health condition, or a different route of administration. It is a useful classification for surveillance, not a biological certificate.
The denominator determines what a number means
Once the category is defined, the next question is: out of what? A case report has a numerator—the cases observed—but the meaning of that numerator changes with the denominator.
Consider two statements from the CDC analysis. Of 32,819 probable or confirmed U.S. cases included in the comparison, 271 occurred in people who met the fully vaccinated definition. That makes fully vaccinated cases 0.8% of the analyzed cases. This is a case-distribution measure. It tells us how the reported cases in that comparison were divided between two vaccination categories.
A different subset of 31 jurisdictions supplied enough vaccination-status and coverage information to estimate infections within the vaccinated population. In that subset, 187 infections were reported among 188,907 fully vaccinated people, an estimated 0.1% infection rate for the relevant period. This is a population-denominator measure. It asks how many infections were observed among people counted as fully vaccinated, rather than what fraction of all reported cases they represented.
The two measures can both be true because they are answering different questions. A small share of all cases can occur in vaccinated people, while a population-based estimate gives a different percentage. Confusing the two can make a vaccine appear either weaker or stronger than the evidence supports. Saying “only 0.8% of cases were vaccinated” does not mean 0.8% of vaccinated people became infected. Saying “0.1% of vaccinated people were reported infected” does not mean every exposure carried a 0.1% risk, or that undetected infections were impossible.
A simple hypothetical makes the distinction clearer. Imagine 10,000 reported infections in a population where 8,000 people are vaccinated and 2,000 are not. If 100 of the infections occurred in vaccinated people, vaccinated cases would make up 1% of reported cases. But the vaccinated-population rate would be 100 out of 8,000, while the unvaccinated-population rate would be 9,900 out of 2,000. Those rates and the case share describe different slices of the same event. The example is not a new estimate of mpox risk; it is a reminder that a numerator becomes meaningful only when the reader knows which people were eligible to appear in the denominator.
Population denominators also depend on how the population at risk was estimated. The people counted as vaccinated are not necessarily identical in age, behavior, location, health status, or number of opportunities for exposure to the people who were not vaccinated. Protection changes exposure outcomes, but people also differ in how often and how intensely they encounter the virus. A population estimate therefore describes the observed period and population; it is not a universal conversion factor that can be applied to any individual.
Case-distribution measures have their own traps. The CDC comparison included 24,507 unvaccinated cases and 271 fully vaccinated cases, but vaccination status was missing for 3,737 cases. A missing vaccination field does not mean “unvaccinated.” It means the record could not be assigned to either comparison group under the analysis. If missingness is concentrated in particular jurisdictions, age groups, or clinical settings, the observed distribution may not perfectly represent the true distribution of all infections.
Missingness can affect more than the size of a group. It can change which kinds of illness are visible within that group. People with mild symptoms may not seek testing, while people with severe symptoms are more likely to enter medical records. Vaccination status may be easier to verify in some health systems than in others. Exposure information may be incomplete because reporting questions changed over time. These are not reasons to assume the result is wrong, but they are reasons to describe it as a finding from reported, classifiable cases rather than a perfect count of every infection after two doses.
Other cases were excluded for reasons such as receiving one dose or receiving the second dose too close to illness onset. Those exclusions are analytically sensible, but they reduce the population being described. The final comparison is not “every person who ever received a JYNNEOS dose.” It is a selected set of reported cases that met the study’s definitions. That selection makes the groups more comparable for the question being asked, while also limiting how broadly the result should be generalized.
The same logic applies to clinical outcomes. If hospitalization status is unknown for some cases, the percentage hospitalized is calculated among cases with known hospitalization information, not among every reported case. If exposure information is recorded for only a subset, a partner-count comparison describes that subset. Precision in the denominator is not a statistical ornament; it is what prevents an observed fraction from being mistaken for a personal guarantee.
Keep dose, route, timing, and outcome attached
A completed dose count is only one part of the case description. The route of administration and the time since vaccination can change which comparison is relevant. The U.S. data distinguish two subcutaneous doses, two intradermal doses, and a mixed route, while the Swedish clinical study uses its own route-defined setting. Those findings can be related, but they cannot be merged into one undifferentiated “vaccinated” mechanism.
Timing must also stay attached to the outcome. A case occurring 14 days after dose two is not the same observation as a case occurring hundreds of days later. The first is close to the classification boundary; the second raises questions about the persistence of protection and the nature of the exposure. But neither timing alone proves how much immunity a particular person had. It only tells us where the case sits in the study’s time structure.
The outcome being measured matters just as much. If the question is infection, the numerator is a diagnosed infection. If the question is severity, the comparison may involve lesions, mucosal involvement, systemic symptoms, pain, or bacterial complications among people who were already infected. If the question is hospitalization, the analysis may use a smaller group with complete records. A vaccinated case can therefore be counted as a breakthrough infection and still provide evidence of lower severity or lower hospitalization risk. The labels are not contradictory because they refer to different stages.
This is also why a completed series should not be treated as an all-purpose explanation for every difference between groups. A route may be associated with a longer interval before a breakthrough case is reported, but that does not establish that the route caused the difference. A vaccinated group may have a different exposure pattern from an unvaccinated group. A case series may detect mild illness more reliably in patients who are already connected to a specialist clinic. The vaccination category begins the comparison; it does not erase the conditions around it.
A useful reading habit is to attach four labels to any breakthrough result: the dose status, the route, the time since the last dose, and the outcome being measured. “Two doses” without the route may conceal a mixed group. “After vaccination” without the time window may combine a recent dose with a dose given long ago. “Lower risk” without the outcome may leave it unclear whether the report means fewer infections, milder symptoms among infected people, or fewer hospitalizations. Keeping these labels together prevents a result about one stage of protection from being used as evidence about another.
The labels also make uncertainty easier to state precisely. A report may show that a fully vaccinated category had fewer diagnosed infections in a defined period, while a clinical study shows that vaccinated people who were infected had lower average severity. Those findings reinforce one another, but they do not become one single number. The first is about who entered the infection count; the second is about what happened after people had already entered it. The distinction is the reason the same person can be both a “breakthrough case” and an example of vaccine benefit.
The diagram’s two paths after classification are deliberate. The case-count path asks how reported cases are distributed. The population path asks how many infections were observed within a vaccinated group. Neither path alone captures the full individual experience, because exposure, testing, reporting, and clinical follow-up influence what becomes visible.
The phrase fully vaccinated is therefore best read as a starting condition for a comparison. It says, “This record met the study’s dose and timing rules.” It does not say, “This person was guaranteed never to become infected,” “this person had no exposure,” or “this person’s illness could not be severe.” Those stronger statements would require evidence beyond the classification itself.
With that foundation, the national breakthrough numbers can be interpreted without two common errors: treating a share of cases as a population risk, and treating a completed schedule as a guarantee. The next question is what the U.S. data show after those definitions are applied—how rare the reported breakthroughs were, what symptoms they produced, and how their clinical pattern compared with unvaccinated mpox.