Post-exposure vaccine effectiveness and contact management in the mpox outbreak, Madrid, Spain, May to August 2022

Background Appropriate vaccination strategies have been key to controlling the outbreak of mpox outside endemic areas in 2022, yet few studies have provided information on mpox vaccine effectiveness (VE). Aim To assess VE after one dose of a third-generation smallpox vaccine against mpox when given as post-exposure prophylaxis (PEP) within 14 days. Methods A survival analysis in a prospective cohort of close contacts of laboratory-confirmed mpox cases was conducted from the beginning of the outbreak in the region of Madrid in May 2022. The study included contacts of cases in this region diagnosed between 17 May and 15 August 2022. Follow up was up to 49 days. A multivariate proportional hazard model was used to evaluate VE in the presence of confounding and interaction. Results Information was obtained from 484 close contacts, of which 230 were vaccinated within 14 days of exposure. Of the close contacts, 57 became ill during follow-up, eight vaccinated and 49 unvaccinated. The adjusted effectiveness of the vaccine was 88.8% (95% CI: 76.0–94.7). Among sexual contacts, VE was 93.6% (95% CI: 72.1–98.5) for non-cohabitants and 88.6% (95% CI: 66.1–96.2) for cohabitants. Conclusion Post-exposure prophylaxis of close contacts of mpox cases is an effective measure that can contribute to reducing the number of cases and eventually the symptoms of breakthrough infections. The continued use of PEP together with pre-exposure prophylaxis by vaccination and other population-targeted prevention measures are key factors in controlling an mpox outbreak.


Introduction
Since detection of mpox (formerly monkeypox) virus transmission outside endemic areas in early May 2022, a large multi-country mpox outbreak spread worldwide, with a total of 86,724 laboratory confirmed cases by 27 March 2023 [1]. On 23 July 2022, the World Health Organization (WHO) Director General declared this outbreak a public health emergency of international concern (PHEIC) and issued a set of temporary recommendations [2]. Until the end of July 2022, Europe remained the epicentre of this large and geographically widespread outbreak, with a steady increase of cases and affected countries [1,3]. Spain was the European country reporting most cases and the third worldwide. The capital Madrid region (6,751,251 inhabitants) [4] had the highest mpox rates in Spain with 373 cases per million inhabitants [5,6].
Both post-exposure prophylaxis (PEP) and pre-exposure prophylaxis (PrEP) by vaccination have been included among the main measures to contain the 2022 outbreak and their eligibility criteria have been included in the guidelines of both national and international organisations [7][8][9][10]. For these interventions, Modified Vaccinia Ankara -Bavarian Nordic (MVA-BN), a third-generation live vaccine authorised in 2013 by the European Medicines Agency (EMA) for the prevention of smallpox was used. Taking these recommendations into account, scientists have been trying to launch studies of vaccine efficacy and effectiveness but there are ethical, methodological and logistical challenges that are complicating the design of these studies. Consequently, robust data on the clinical efficacy or effectiveness of the third-generation vaccines against mpox disease have been lacking at the time of writing [11,12].
The aim of this observational study was to assess postexposure vaccine effectiveness (VE) under the conditions of routine clinical practice during the early stages of the 2022 outbreak and the vaccine's effectiveness in preventing or minimising the disease in close contacts.

Study design and definition of close contacts
A dynamic prospective cohort study to assess VE of one vaccine dose given as PEP in close contacts of laboratory-confirmed cases (by specific or generic Orthopoxvirus PCR) was conducted from 17 May to 15 August 2022, starting shortly after the beginning of the mpox outbreak in Madrid in early May 2022.
This study included close contacts that met the following inclusion criteria: having contact with body fluids or lesions, contaminated fomites or clinical samples (without personal protective equipment) from a case since the time of appearance of the first symptoms, or in case of presenting with a rash since the day before first symptoms appeared [13], and whose index case was laboratory confirmed infection with mpox virus. Close contacts were excluded if: the index case could not be identified, the index case could not be laboratory-confirmed, the contact had compatible symptoms with mpox and the aetiology could not be confirmed, the date of onset of risk of the contact could not be established or they had other risk exposures during the follow-up period outside their home.
The maximum follow-up period was 49 days for household members (a maximum of 28 days of symptomatic disease plus a maximum of 21 days of incubation period [13]) and 21 days for non-cohabitants.

Contact identification and post-exposure prophylaxis
Post-exposure prophylaxis was made available for close contacts based on the criteria agreed at national level on 9 June 2022 [10] and its subsequent update [14]. Vaccination priority was given to contacts at risk of severe disease or with high risk of exposure: immunocompromised persons, pregnant women, children of any age and healthcare or laboratory personnel affected by accidental exposure. If vaccines were still available, these were then offered to contacts with high risk of exposure: sexual partners or household members/cohabitants with no possibility of self-isolation. Those with a history of smallpox vaccination were not excluded from PEP. The vaccine was offered up to 14 days post exposure, preferentially in the first 4 days. If the risk was sustained this period could be extended. Vaccination as PEP was launched on 16 June 2022.
The third-generation smallpox vaccines IMVANEX or JYNNEOS (Bavarian Nordic) were delivered subcutaneously with two 0.5 mL doses administered 28 days apart [15,16]. The administration of the second dose, depending on vaccine availability, had not yet been implemented during the study period.
Confirmed cases were asked by a trained epidemiologist about the number of close contacts and the contacts' identifying data. The risk of each contact was assessed individually and vaccination was offered if they met the aforementioned criteria. Post-exposure prophylaxis was administered at a single centre in Madrid. At the vaccination centre, information about the vaccine and the disease was provided and an individual assessment of possible side effects was made. If the contact agreed to the vaccination, they signed an informed consent. Close contacts who received one dose of the vaccine were considered post-exposure vaccinated. Those who were contacts before the launch of PEP, those who declined vaccination, those who did not attend the scheduled appointment even if they accepted the vaccination or those who had contraindications to the vaccine were considered unvaccinated.

What did you want to address in this study?
In 2022, when a large mpox outbreak occurred outside areas usually affected by the disease, little was known about the effect of one dose of third-generation live vaccine vaccines in a real world scenario. We wished to assess the effectiveness of vaccination against mpox, and specifically whether close contacts of confirmed mpox cases have a lower risk of getting the disease if they are vaccinated vs if they are not.
What have we learnt from this study? Vaccination against mpox virus seems to be an effective measure for up to 7 weeks after vaccination. Vaccinated contacts had a lower risk of developing the disease than unvaccinated contacts and, additionally, if they got the disease, they had fewer symptoms.
What are the implications of your findings for public health? Mpox post-exposure prophylaxis vaccination can contribute to the controlling of outbreaks. It is important to encourage the identification of close contacts as well as to promote vaccination so that they can benefit from reduced risk of the disease or its severity.

KEY PUBLIC HEALTH MESSAGE
Vaccination status was confirmed by consulting the vaccination registry. The variables collected from the contacts were: sex (legal sex as recorded in the sanitary registry), age, type of contact, use of HIV-PrEP [17] (last dispensation in the past 3 months), HIV status, date of symptom onset of index case, date of last contact with index case, vaccination date, date of symptom onset for contacts who became infected, symptoms (exanthema, lymphadenopathy, proctitis, mouth ulcers, ocular involvement, bacterial superinfection of lesions, bronchopneumonia and general symptoms such as fever, asthenia, odynophagia, muscle pain, headache, other), polysymptomatology (four or more symptoms described above) and type of confirmation test (specific or generic Orthopoxvirus PCR).

Data analysis
Due to the lack of individualised computerised records of smallpox vaccination, a sensitivity analysis was used to assess whether close contacts had previously been vaccination against smallpox. Here, in addition to using age as a continuous variable, a dichotomous vicarious variable (I v ) and its interaction with postexposure vaccination was introduced. The cut-off point used was age since it was possible to differentiate by age between those who had or had not received a smallpox vaccination. In Spain, smallpox vaccination was mandatory until 26 October 1979 [18]. Therefore, people aged 42.2 years or older in 2022 were likely to have been vaccinated. Given that 51.8% of the confirmed cases were born in Spain and the remaining cases in South and Central American countries where the Pan American Health Organisation [19] specified that "most people aged 50 or over would have been vaccinated against smallpox", the weighted cut-off point was set at 46.2 years.

Statistical analysis
The descriptive analysis of qualitative variables was performed with frequency distribution and the quantitative variables with mean and standard deviation (SD) or median and interquartile range (IQR). To compare characteristics of the contacts (vaccinated and unvaccinated) a chi-squared test, Fisher's exact test or t-test were used depending on the type of variable (categorical or continuous). Log-rank test was used to evaluate the trends. In all analyses, a significance level of 0.05 was used.
Overall VE and its 95% confidence interval (CI) was assessed using a survival analysis between those not vaccinated and those vaccinated with PEP, with the development of an mpox infection as the main outcome.
To measure crude effectiveness of the vaccine, we performed a univariate analysis of the different characteristics of the groups, calculating the rate ratio using the Mantel-Cox method. We applied the Cox proportional hazards model to a multivariate analysis to measure adjusted VE, adjusting for the different covariates and evaluating the presence of effect modifiers. The estimation of VE was given by VE = (1-HR)⋅100, where HR is the hazard ratio. The detailed description of the method is available in the Supplementary material. The presence of general symptoms or polysymptomatology were the events used in the VE analysis for preventing or minimising the disease in close contacts. Uncensored observations were defined as close contacts who developed the main outcome during followup, while censored observations were defined as close contacts free of the main outcome at the end of the follow-up.
To assess the possible incidence bias that would cause cases with early onset not to be vaccinated and become part of the cohort of unvaccinated cases, overestimating VE, a sensitivity analysis by intention to treat was performed. Here, the estimated VE was compared with VE obtained by including the cohort of vaccinated patients who accepted vaccination but were not vaccinated. A debug of the database was carried out in order to find, study and correct lost, illicit and improbable values, as well as inconsistencies between them, especially for the variables referring to dates. The software used for data cleansing were the Pandas, NumPy and Datetime libraries in Python version 3.9.13 and for all statistical analysis Stata version 16.1 (StataCorp, Texas, United States (US)) was used.

Results
Since the mpox outbreak began in May 2022, until 15 August 2022 2,220 cases were reported in Madrid, of which 1,384 (62.3%) were declared close contacts. In total 4,097 close contacts were recorded, 438 of which (10.7%) had identification data. Another 53 close contacts of confirmed cases were self-reported or notified by other regions. Altogether, 491 close contacts were identified for follow-up, with 400 (81.5%) detected when PEP was available on 16 June 2022. Vaccination was offered to 274 close contacts and the first PEP dose was finally administered to 233 contacts -58.3% of the total number of identified contacts and 85.0% of the total number of contacts who were offered the vaccine (Figure and Table 1).

Vaccine effectiveness
Of A comparison of the socio-demographic and clinical characteristics of these cases is shown in Table 2.
The total follow-up time was 11,873 days with a median of 25 days and a range of 2-46 days for censored contacts. Of the contacts, 97.7% were followed for 21 days or more. The median number of follow-up days among the 57 contacts who became ill was 8 days (range: 1-26 days). Table 3 describes the characteristics of the contacts according to their vaccination status. The extreme age group, the type of contact and sex presented statistically significant differences according to vaccination status.
Contact type was associated with becoming ill, presenting differences in survival functions of the Logrank test (χ 2 (5) = 26.97, p<0.001) and the trend test (χ 2 (1) = 17.61, p<0.001). In sexual contacts (cohabiting and non-cohabiting), the observed number of cases who became ill were 67.9%* higher (47 vs 28) than expected (having the same survival function as that of total contacts). On the other hand, non-sexual household member contacts were 69.5% less likely to become ill than expected (7 vs 23). We found a similar situation according to sex that showed statistically significant differences in survival functions (χ 2 (1) = 7.33, p = 0.007) where women presented 64.3%* fewer cases than expected (5 vs 14). In those contacts using HIV-PrEP, differences in survival functions were observed (χ 2 (1) = 3.99, p = 0.046), showing a 97.3% higher risk of becoming ill than non-users. Contacts infected with HIV also presented differences in survival functions (χ 2 (1) = 45.47, p < 0.001), with a 4.02 times higher risk of becoming ill compared to those not infected.
The results of crude (rate ratio) and adjusted VE are shown in Table 4. Some CIs could not be calculated because there were no cases in that subpopulation or in the cohort of vaccinated patients. For close sexual contacts, the adjusted estimated point of VE for non-cohabitants was slightly higher than for cohabitants. This result was also seen among HIVinfected cases vs non-HIV infected cases. As for the influence of age, a 10-year increase in age reduced the risk of contracting the disease among close contacts by 26.8% (95% CI: 3.2-44.6). The interaction of the proxy smallpox vaccination variable (I v ) with the study factor was not statistically significant from the likelihood ratio (χ 2 (3) = 2.32, p = 0.509). Furthermore, age did not modify the effect of PEP as its interaction was not statistically significant (χ 2 (1) = -13.73, p = 1). Additionally, the VE obtained in the subpopulation that would not have received smallpox vaccination in childhood (87.0, 95% CI: 70.4-94.3)* is similar to that estimated in the present study (88.8%, 95% CI: 76.0-94.7). The previously mentioned interaction was not statistically significant in the models for the assessment of VE in the reduction of general symptoms ( χ 2 (1) = 1.33, p = 0.248) nor in polysymptomatic affectation (χ 2 (1) = -0.11, p = 0.742).
For the sensitivity analysis to assess possible incidence bias, 37 cases were included in the cohort of vaccinated patients who accepted vaccination but were not vaccinated, of which six fell ill (on 3, 7, 7, 8, 11 and 12 days post-exposure) before the date of their vaccination appointment. The VE by intention to treat was 85.4% (95% CI: 72.6-92.2), thus not substantially different from the estimated VE, which was 88.8% (95% CI: 76.0-94.7) without this factor. The vaccination delayed onset of symptoms by 5.5 days (95% CI: 1.7-9.4) although a similar delay was seen in the intention-to-treat analysis where the delay was 3.7 days (95% CI: 0.6-6.9).

Discussion
Using regional data from Madrid during the early phase of the mpox outbreak that began in May 2022, we estimate an adjusted VE of 88.8% (95% CI: 76.0-94.7) with one dose of the third-generation smallpox vaccine used in the context of PEP. This high effectiveness was observed in both cohabiting sexual and non-cohabiting sexual contacts and in HIV-infected contacts, although there was a slight decrease in effectiveness with respect to uninfected patients and the effectiveness was independent of the time from exposure to vaccination. The benefits of vaccination were shown in all age groups and in men. In women, given that 68.4% were not vaccinated and considering the low number of events (one in the vaccinated and four in the unvaccinated group), no conclusions can be drawn regarding VE.
We found that breakthrough infections did not exhibit severe symptoms (none required hospitalisation). Furthermore, there were fewer symptoms, a reduction  in general symptoms of at least 18.1% and a reduction in polysymptomatic disease of at least 26.7% in vaccinated cases compared with unvaccinated cases. This has also been described for other infectious diseases, in which post-exposure vaccination can modify the course of the disease and reduce its severity [20].
Robust data on the clinical VE against mpox are still lacking, but some initial assessments are available, with information mainly from observational studies. Most of the published studies on VE in the 2022 mpox outbreak were carried out in the context of preexposure vaccination, with third-generation vaccines and a single dose. A study carried out in England [21] showed VE against symptomatic mpox of 78% (95% CI: 54-89) with a range of 71-85 in the sensitivity analysis. Another study conducted in Israel estimated an adjusted VE of 86% (95% CI: 59-95) [22]. In the US, it was observed that the incidence rates were 14 times higher among unvaccinated men (mixed PEP and PrEP), while in our current study, they were 8.9 times higher in the exclusive context of PEP for people of both sexes [23].
Two observational studies conducted in France have been published, one in Paris with 276 participants [24] and another in Lyon with 108 participants [25], in which the effect of post-exposure vaccination against mpox has been evaluated with a single dose of a third-generation smallpox vaccine. Vaccine effectiveness results were not provided, but 4% and 10% of infections in the Paris and Lyon studies, respectively, were identified after vaccination. The US study [23] identified 2.4% of all infections as breakthrough infections, a similar result to ours, which was 3.5%.
Our study has several limitations derived from its observational design. The VE found may not be generalisable because the population offered the vaccine was prioritised among the highest risk population. Although this limitation may exist for the general population because of different exposure patterns, it does not exist for the target population in the early outbreak. Furthermore, because this population was prioritised for ethical reasons, given the a priori vaccine efficacy and because of availability and high acceptance of vaccination, immunisation was not randomised. Therefore, the nonvaccinated cohort includes close contacts exposed just prior to the start of the vaccination campaign, as a regression discontinuity design would.
The possible non-detection of cases is deemed minimal, with access to emergency medical services or sexually transmitted infection (STI) clinics being free of charge and irrespective of residency status in the community and clinical laboratories in the region reporting mpox-specific test results electronically. Linkage of test results from contacts who progressed to cases depends on identification.
Selection bias may have occurred due to the difficulty in including all close contacts because they were unknown to the index case (group sex, sauna attendance, cruising etc.). Therefore, it was not possible to assess the characteristics of these contacts or determine whether they differed from the analysed contacts. Also, with the possibility of subclinical diagnoses of the close contacts most exposed to risk or contacts who were periodically monitored, a detection bias could have occurred. There could have also been a response bias between those who accepted vaccination and those who declined it, although this would be negligible given vaccination acceptance was 97.1%.
A possible limitation is that it was not possible to carry out a sensitivity analysis with gender because it was not included in the sanitary registry. The register records only binary legal sex.
Another potential limitation would be the possible existence of an incidence bias that would cause contacts with early disease onset to be in the unvaccinated cohort because they failed to arrive in time for vaccination, which would lead to an overestimation of VE. To assess the possible magnitude of such an incidence bias, an intention-to-treat sensitivity analysis was performed, which showed little change in VE (85.4% for intention-to-treat vs 88.8% for estimated VE). Another impact of this bias would be the delay in onset of symptoms in vaccinated patients compared with unvaccinated patients. The mean delay in the vaccinated cohort was 5.5 days (95% CI: 1.7-9.4) and the mean delay in the intention to vaccinate cohort was 3.7 days (95% CI: 0.6-6.9). Since the delay was observed in both groups, this difference could be due to the fact that the mildness of the symptoms and the history of vaccination delayed the diagnosis.
Another limitation is the VE results in different subpopulations with small simple sizes and wide confidence intervals. These should be interpreted with caution and the lower limit of the confidence intervals taken as the most reliable measure.
The proxy smallpox vaccination variable did not modify the effect of PEP vaccination. Age did not modify the effect either and VE in the age group that did not receive the smallpox vaccine (87.0%) was similar to the age group that did (88.8%). This suggests that prior smallpox vaccination would not modify the post-exposure effect of vaccination against mpox.
Calendar time was not included in the analysis as all participants were exposed to a laboratory-confirmed mpox case at the time of inclusion in the study, therefore the data did not depend on the changing incidence of cases over time. Likewise, diagnostic methods and clinical management did not vary throughout the study, given the relatively short duration of this study, which was sufficient to detect new incident cases based on the incubation period described for the disease and short enough to avoid repeated risk exposures outside the cohabitation setting.
The data collected by the trained epidemiologist would largely avoid any observer or interviewer bias and the review of clinical histories and microbiological results would avoid any possible recall bias.
Among the WHO recommendations to reduce the spread of mpox is the collection of data from sexual partners of confirmed cases to facilitate disease detection. The WHO also propose actions to improve notification of close contacts [26]. When following these recommendations, our main challenges were the identification and vaccination of contacts within the first 4 days after exposure. From the moment the index case sought medical care, the time elapsed between confirmation of the diagnosis, notification of the case, conduction of the epidemiological survey, communication with contacts, assessment of risk to offer PEP and making an appointment for the vaccination exceeds 4 days in a high percentage of cases. This makes it difficult to achieve the objective of vaccinating close contacts within 4 days in order to prevent the infection. Given the high VE of PEP, the delay in recruiting contacts and the fact that half of contacts were sexual partners, many of them could have benefited from PrEP without the need for a confirmed case in their personal environment. Our results show a higher acceptability of PEP (97.1%) than a recent online survey conducted in the WHO European Region (82%) [27]. In our study, 86.5% of sexual contacts who were offered the vaccine became vaccinated while the percentage was ca 10% lower for non-sexual household contacts, possibly reflecting the higher risk perceptions in the former group.
Targeted vaccination strategies, along with other measures aimed at reducing risk behaviours [28], risk communication and community engagement, early diagnosis, isolation and effective contact tracing are suggested to have been key to controlling the 2022 mpox outbreak [12].

Conclusions
Post-exposure vaccination of close contacts of mpox cases is an effective measure that can contribute to controlling the spread of the disease and eventually the symptoms of breakthrough infections. The continued use of PEP vaccination together with PrEP and all other population-targeted prevention measures are key actions in controlling large mpox outbreaks.

*Author's correction
At the request of the authors, we corrected several entries in

Ethical statement
This study was conducted in accordance with the Declaration of Helsinki, as revised in 2013. Approval from a Research Ethics Committee was not required, as Organic Law 3/2018, of 5 December, on the Data Protection and Guarantee of Digital Rights provides, with respect to the processing of health data, that the health authorities and public institutions with public health monitoring powers, may carry out scientific research without the data subject's consent in situations of exceptional relevance and seriousness for public health.