Introduction

The ongoing mpox outbreaks in DRC, followed by the spread of mpox virus (MPXV) clade Ib from South Kivu, DRC, to neighboring countries, were reasons to call for coordinated actions to stop MPXV spread1,2,3. To this end, the World Health Organization (WHO) declared a Public Health Emergency of International Concern (PHEIC) in August 20244. The Africa Centers for Disease Control and Prevention (Africa CDC), in collaboration with the WHO, developed a continental response plan that describes the steps needed to combat mpox5. This included the need for developing laboratory capacity, public health capacity, and vaccines. A pledge for vaccines enabled a first round of vaccination with modified vaccinia Ankara – Bavarian Nordic (MVA-BN) in DRC, initially targeting healthcare workers and risk groups6,7. To target vaccination to specific risk groups to interrupt further transmission, it is important to understand the extent of MPXV circulation. This is difficult to ascertain as the focal point of the outbreak is in regions with little health system capacity.

Between January 2024 and mid-April 2025, over 92,000 suspected and confirmed cases were reported in DRC, of which over 25,000 were reported in South Kivu (Supplementary Fig. 1)8. MPXV clade Ib was first detected in Kamituga, South Kivu, DRC, in September 20239, where it was associated with a high exposure risk through professional sex workers (PSW)10,11. Rapid person-to-person transmission appears to have driven the outbreak, including cross-border transmissions12. As the outbreak in South Kivu is almost exclusively driven by this novel clade, which is distinct from MPXV historically found in the region9, extrapolation to the situation in the rest of DRC may be difficult. Initial case descriptions suggest that the associated disease is less severe than disease caused by MPXV clade Ia, but the pattern of spread suggests increased rates of human-to-human transmission, corroborated by an enrichment of APOBEC3-type mutations indicative thereof 3,10,13. Better understanding of secondary attack rates is a key knowledge gap. Serological studies can help to understand both the size of the disease pyramid and the true extent of MPXV clade Ib spread. These serological studies are also the cornerstone for understanding and evaluating the immunogenicity of the MVA-BN vaccine, currently being administered in the region to individuals with or without prior exposures through infection and/or vaccination.

Since the spread of MPXV clade Ib and the declaration of the PHEIC, serological studies in Central Africa have been lacking2. Here, we describe a serological study among individuals potentially exposed to an index case in bars with PSW or households, and a pre-outbreak control cohort. Sera from the potentially exposed cohort were collected after the beginning of the clade Ib outbreak in South Kivu, DRC, between September 2023 and May 2024. For the control cohort, archived serum samples were used, which were collected at least two years prior to the ongoing MPXV clade Ib outbreak in an area that has not reported mpox before. These studies were performed to (1) determine the level of MPXV exposure in PSW, (2) determine the level of exposure through non-sexual household transmission, and (3) assess the possibility to perform serological assays at local laboratories.

Results

Study design and cohort description

Sera were obtained from a total of 276 potentially exposed individuals in DRC. Sera collected from 51 individuals pre-outbreak served as negative controls. Most of the sera from the potentially exposed cohort were obtained from individuals working in or visiting bars (n = 168), collected either in November and December 2023 (n = 120), or May 2024 (n = 48) (Table 1). These individuals include bar personnel and mining workers, but most sera were collected from PSW (n = 113). Additionally, sera were obtained from household contacts of confirmed mpox cases (n = 108) between September 2023 and May 2024 (Table 1). Out of the 276 sera obtained from these potentially exposed exposure cohorts, 21 (7.6%) were human immunodeficiency virus-1 (HIV-1) seropositive. According to the UNAIDS, the national seroprevalence in the general population of DRC was 0.7% in 2023; however, in PSW specifically, the reported HIV seroprevalence was 7.5%, similar to what we find in this cohort14.

Table 1 Cohort characteristics
Full size table

Seroprevalence of VACV-binding antibodies among potentially exposed individuals

As an initial screening, we measured levels of vaccinia virus (VACV)-binding antibodies in all sera by ELISA, as described previously15, to assess potential orthopoxvirus exposure. Among all potentially exposed individuals, 62 out of 276 (23%) had detectable VACV-binding antibodies (Table 2). In sera collected from bars in November and December 2023, or May 2024, VACV-binding antibodies were measured in 25/120 (21%) and 10/48 (21%) samples, respectively. In sera collected from households, the ELISA-positivity rate was 27/108 (25%) (Fig. 1a). Since most of the samples in the potentially exposed cohort were obtained from individuals younger than 44 years (267 out of 276, 97%), and smallpox vaccination in DRC stopped around 198016, most of the individuals were expected to be serologically naive for VACV-binding antibodies. Therefore, the high seropositivity rate was likely caused by MPXV infection, although exposure to other (animal) poxviruses cannot be ruled out. Antibody levels were comparable among male and female individuals, and across age groups (Supplementary Fig. 2a). VACV-binding antibodies were detected in 6 out of 51 (12%) of the pre-outbreak controls (Fig. 1a). Four of those individuals were in the age range where they could have received historic smallpox vaccination.

Fig. 1: VACV-binding and MPXV-neutralizing antibodies in sera from potentially exposed (high-risk) and unexposed (pre-outbreak) individuals in DRC.
figure 1

a VACV-binding antibodies in sera obtained from bars in 2023 (n = 120), bars in 2024 (n = 48), households (n = 108) (all potentially exposed, orange), controls (unexposed, teal), and validation sera (gray and red). Fraction of seropositive sera is indicated per group. b MPXV-neutralizing antibodies in sera shown in (a). c Correlation between VACV-binding binding and MPXV-neutralizing antibodies in potentially exposed individuals (orange), controls (teal), and validation sera from vaccinated individuals (red). Colored bubbles indicate manually defined clusters. d Percentage of bars with seropositive professional sex workers (PSW). e Serostatus from all individuals in bars; index cases were not sampled. Every row represents a bar, every square a serum. Sera are categorized as mpox-seropositive (red; positive for VACV-binding and MPXV-neutralizing antibodies) or mpox-seronegative (gray; negative for VACV-binding and MPXV-neutralizing antibodies). Sera testing positive for either VACV-binding or MPXV-neutralizing antibodies are shown as orange. f Percentage of households with seropositive members. g Serostatus from all individuals in households; index cases were not sampled. Every row represents a bar, every symbol a serum from an adult (square; ≥18 years old) or a minor (circle; <18 years old). Color-coding is identical to panel (e). DRC Democratic Republic of the Congo, VACV vaccinia virus, MPXV mpox virus, ELISA enzyme-linked immunosorbent assay, PRNT plaque reduction neutralization test.

Full size image
Table 2 Detection of VACV-binding and MPXV-neutralizing antibodies and associated mpox seropositivity
Full size table

Seroprevalence of MPXV clade Ib-neutralizing antibodies among potentially exposed individuals

Next, we developed a neutralization assay using an MPXV clade Ib isolate obtained from an mpox patient in DRC. To this end, we inoculated swab material on Vero cells, propagated the isolate to passage 2, and grew virus stocks on Calu-3 cells for neutralization assays (see Methods for details). The input material and all virus passages were sequenced to confirm virus identity and the absence of cell culture adaptations (GISAID accession number: EPI_ISL_19621388; Supplementary Fig. 3). Since MPXV clade Ib neutralization assays have not yet been described, we performed neutralization assays on 10 validation sera available at Erasmus MC (Fig. 1b, red and gray circles). From these 10 control sera, 5 did not have VACV-binding antibodies (obtained from individuals known to be unvaccinated and unexposed) and 5 did (obtained from individuals known to have received historic smallpox and recent MVA-BN vaccination) (Fig. 1a, red and gray circles). Indeed, the in-house developed neutralization assay with MPXV clade Ib demonstrated the presence of cross-neutralizing antibodies in the vaccinated individuals, but not the seronegative controls. Based on the negative control sera, we defined the cut-off as a titer threefold over background (>35).

We tested all available sera from the potentially exposed group and the pre-outbreak control group for the presence of clade Ib MPXV-neutralizing antibodies (Fig. 1b). Positivity rates among the two bar groups, collected in November and December 2023, and May 2024, were 51% (61/120) and 42% (20/48), respectively. Among the sera collected from households, 44% (48/108) tested positive (Table 2). Using the combined presence of detectable binding antibodies by ELISA (>10) and confirmation by neutralization (>35) as the definition for mpox seropositivity, the overall seropositivities for the three groups were 18%, 17%, and 23%, respectively (Table 2). Similar to the VACV-binding antibodies, comparable MPXV-neutralizing antibody levels were detected among male and female individuals, and individuals from different age groups (Supplementary Fig. 2b).

While no clear correlation between VACV-binding and clade Ib MPXV-neutralizing antibody levels was observed, the samples from the different cohorts (seropositive potentially exposed, pre-outbreak, and validation sera) clustered together (Fig. 1c). Most of the individuals we defined as mpox-seropositive had high levels of MPXV-neutralizing antibodies, indicative of recent MPXV infection as described previously for MPXV clade IIb15. The seropositive validation sera obtained from individuals with historic smallpox and recent MVA-BN vaccination had high levels of VACV-binding antibodies and intermediate levels of clade Ib MPXV-neutralizing antibodies (Fig. 1c). Notably, many sera tested positive for the presence of MPXV-neutralizing antibodies in the absence of measurable titers by the VACV ELISA. This suggests that mpox seropositivity percentage we report here could be an underestimation of the actual seroprevalence.

Developing local capacity to perform serological assays to detect MPXV infections

As part of capacity building, we performed ELISA training at the University Teaching Hospital of Butare (CHUB; Huye, Rwanda). In total, 146 sera were tested at both sites (Supplementary Fig. 4) to allow comparison of local laboratory readiness for serological screening assays. Overall agreement between test results was 94% (137 out of 146). However, the testing at CHUB yielded 5 positives that could not be confirmed by ELISA at Erasmus MC, all of which did test positive in neutralization assays. Reversely, 4 samples tested positive at Erasmus MC but negative at CHUB, all of which were again neutralization-positive. This, combined with the detection of MPXV-neutralizing antibodies in the absence of VACV-binding antibodies, shows that further work is needed to do standalone serology based on ELISA.

PSW play an important role in the local transmission of MPXV clade Ib

To determine the role of PSW in MPXV clade Ib transmission in South Kivu, DRC, we then focused on the analysis of sera obtained from PSW and visitors of 25 bars (n = 168) (Tables 1, 2; Fig. 1d). In 16 out of 25 bars (64%), there was serological evidence for MPXV circulation, defined as presence of both VACV-binding and MPXV-neutralizing antibodies in at least one serum sample obtained at that bar (Fig. 1e). In 14 out of 16 bars with serological evidence for MPXV circulation, the serum sample(s) testing mpox-seropositive were obtained from PSW (88%). Mpox seropositivity among the PSW and non-PSW individuals was not significantly different (19.5% [22/113] and 14.5% [8/55], respectively; Fisher’s exact test, two-sided; p = 0.52). This serological evidence supports the hypothesis that sexual transmission is an important transmission route for MPXV clade Ib, which was earlier suggested by epidemiological and genomic surveillance of PSW10,11. Especially in Kamituga, South Kivu, a mining region with a high number of PSW, this could have fueled the rapid spread of MPXV clade Ib.

Evidence for close-contact household transmission of MPXV clade Ib

Similarly, we used serological evidence to determine whether household transmission played a role in the rapid spread of MPXV clade Ib. Serum samples were obtained from 34 households with an index case within the household. The index cases themselves were not included in sampling. In 17 out of 34 households (50%), at least one serum sample obtained from contact cases tested mpox-seropositive, defined as both the presence of VACV-binding and MPXV-neutralizing antibodies (Fig. 1f). In five of these households, there were at least two seropositive individuals (households 1, 4, 6, 20, and 43), of which four included at least one minor (<18 years; households 1, 4, 6, and 43) (Fig. 1g). In five additional households (households 3, 7, 9, 18, and 19), we detected serological evidence of MPXV clade Ib infection in minors, often below the age of 10, suggesting that close-contact transmission via another route than sexual transmission had occurred. Secondary transmission of MPXV clade Ib to household members and close contacts has recently been documented for a number of imported travel-related cases in Germany, Belgium, France, the UK, and China17,18.

Discussion

Here, we show that sero-epidemiological studies are vital to understand the true extent of MPXV clade Ib circulation in the African region. The seropositivity rates in three different potentially exposed groups ranged between 17% and 23%, using stringent definitions for seropositivity by combining the VACV-binding antibody ELISA as a screening assay and the newly developed MPXV clade Ib neutralization assay as confirmation. Positivity rates for MPXV clade Ib-neutralizing antibodies were higher than those for VACV-binding antibodies (Table 2), possibly indicating a reduced sensitivity of the VACV ELISA in the context of MPXV clade Ib infection. When measuring antibodies induced by MVA vaccination of non-primed individuals, the VACV ELISA has a sensitivity and specificity of 100% and 98.8%, respectively (Supplementary Fig. 5)19. Consequently, we cannot exclude that the mpox seropositivity we report here might be an underestimation of the number of true positives. High seroprevalence among PSW in bars confirmed that sexual transmission was a likely cause of the rapid MPXV clade Ib spread. Seropositivity in young children—as part of a study investigating follow-up of the introduction of mpox in a household—confirmed the occurrence of household transmission, likely via a close-contact non-sexual transmission route. As a next step, setting up local laboratories that perform serological assays is essential to monitor vaccine immunogenicity. Here, we used an ELISA based screening assay with confirmation by neutralization as evidence for seropositivity, but it remains to be determined whether the ELISA alone is sufficient for sero-epidemiological studies. Conclusively, serological studies could play a crucial role in determining vaccination policies and could guide decision makers towards relevant risk groups to be vaccinated to interrupt transmission.

Methods

Ethics statement

For the samples obtained in DRC, ethical clearance was obtained from the ethical review committee of the Catholic University of Bukavu (UCB/CIES/NC/022/2023) and a tripartite Memorandum of Understanding signed between the Provincial Health Division (Division Provinciale de la Santé, DPS), Stansile and the Erasmus MC. Sex or gender were not considered in study design. All participants provided informed consent. In case the participant was a minor, parental consent was obtained. The 10 validation sera were collected from laboratory workers in the Netherlands who received MVA-BN vaccination for safety reasons as employees of a Biosafety Level 3 (BSL-3) laboratory under the Erasmus Medical Center vaccination cohort (COVA) biobanking study protocol (MEC-2014-398). The protocol was approved by the Erasmus MC Medical Ethics Review Committee. Written informed consent was obtained from all study participants.

Participants

Samples were collected from three potentially exposed cohorts (bars and households with a suspected, probable, or confirmed mpox index case) from South Kivu, DRC, and a pre-outbreak control cohort. Bars or households without a suspected, probable, or confirmed index case were not sampled. Additional validation samples were included.

Potentially exposed bar cohorts. For this part of the study, serum samples were collected in November and December 2023 (n = 120), and in May 2024 (n = 48) from bars that reported at least one suspected, probable, or confirmed mpox case identified in September 2023. The index cases themselves were not included in this study. We cannot exclude that individuals contracted mpox from a source other than the identified index cases. The amount of contact between index cases and sampled participants in bars is not known. Out of the 25 total bars sampled across both time points, 22 bars were sampled either in November/December 2023 or May 2024. Three bars were sampled at both time points; however, there was no re-sampling of previously sampled individuals. None of the participants were symptomatic at the time of sampling. Demographic data from all participants was collected. A suspected mpox case was defined as an individual with acute illness who had fever, intense headache, myalgia, and back pain, followed by one to three days of a progressively developing rash, often starting on the face and spreading to the rest of the body. A probable mpox case was defined by an individual meeting the clinical definitions described for suspected cases, who had an epidemiological link to a confirmed or suspected case but was not laboratory-confirmed. A confirmed mpox case was defined as an individual with a laboratory-confirmed MPXV infection as tested by the National Institute for Biomedical for Research (INRB).

Potentially exposed household cohort. In addition to samples collected in bars, we performed contact tracing for confirmed mpox index cases identified in September 2023 with a link to a household and collected samples from the family members (contact cases). A contact case was defined to have had contact (non-physical, physical, sexual, respiratory, fomites, or any other form) with the confirmed mpox case. It was not possible to sample every member of the households of index cases. The respective index cases were not part of this study. We cannot exclude that individuals contracted mpox from a source other than the identified index cases. None of the participants were symptomatic at the time of sampling. Household samples were collected between September 2023 and May 2024 (n = 108). Demographic data from all participants was collected.

Pre-outbreak cohort and validation sera. For the pre-outbreak cohort, randomly selected archived serum samples collected at least two years prior to the ongoing MPXV clade Ib outbreak, in an area that has not reported mpox before, were used (n = 51). Samples were collected at CHUB (Huye, Rwanda), a hospital with an outpatient clinic that, among other things, distributes anti-retroviral therapy for HIV. For the validation sera, 10 sera available at Erasmus MC from laboratory workers who received MVA-BN vaccination for safety reasons as employees of a Biosafety Level 3 (BSL-3) laboratory were used. Of these 10 sera, 5 were used as negative samples (obtained from individuals known to be unexposed and before MVA-BN vaccination) and 5 were used as positive samples (obtained from individuals known to have received historic smallpox and recent MVA-BN vaccination).

Sampling

Five ml of blood was collected in dry tubes from which 2 ml of serum was obtained by centrifugation, and subsequently stored at −20 °C before assays were performed. To perform virus isolations, lesion swabs from the genital area of suspected mpox cases were collected between 12 August and 3 September 2024, and stored in virus transport medium at −20 °C before shipment on dry ice to Erasmus MC.

Isolation and propagation of MPXV clade Ib

MPXV clade Ib was isolated from swab material obtained from an mpox patient in DRC. Presence of MPXV clade Ib was confirmed by real-time PCR (RT-PCR) and sequencing20. The virus transport medium in which the swab was stored was inoculated on Vero cells in Advanced DMEM/F-12 (Gibco, #12634010), supplemented with 10 mM HEPES (Gibco, #15630056), 1×GlutaMAX (Gibco, #35050061), and 1×Primocin (InvivoGen, #ant-pm-05), referred to as AdDF+++, and the isolated virus was propagated to passage 2 on Vero cells. Each passage was harvested upon development of cytopathic effect across most of the monolayer using freeze-thaw cycles to release the intracellular mature virions (IMV) from the cells15. Virus stocks for neutralization assays were grown on Calu-3 cells. To this end, Calu-3 cells were inoculated at an MOI of 0.05 in AdDF+++ and harvested as described above. All experimental work involving infectious MPXV were performed in a class II biosafety cabinet under BSL-3 conditions. MPXV stock titers were determined as described previously15.

MPXV sequencing

Original input material and all virus passages were submitted to whole genome sequencing to confirm virus identity and exclude cell culture adaptations. DNA was randomly amplified using sequence-independent single-primer amplification21. Sequencing libraries were generated using the Native Barcoding Kit 24 V14 (Oxford Nanopore Technologies, #SQK-NBD114.24), sequenced on a PromethION flow cell (R10.4.1) with super-accurate basecalling using Dorado v7.4.12 (Oxford Nanopore Technologies, #FLO-PRO114M). MPXV consensus was generated using Virconsens (https://github.com/dnieuw/Virconsens) with >30x coverage cut-off and uploaded to GISAID (EPI_ISL_19621388). The consensus was aligned with 165 publicly available MPXV clade Ib sequences from GISAID (Supplementary Table 1) using squirrel v1.0.12 (https://github.com/aineniamh/squirrel) followed by phylogenetic analysis using IQ-TREE v2 (https://github.com/iqtree/iqtree2) (Supplementary Fig. 2). Clade assignment and quality checks were performed on Nextclade v3.9.1 (https://github.com/nextstrain/nextclade).

Detection of VACV-binding IgG antibodies

Binding IgG antibody levels against VACV were measured both at Erasmus MC as well as on site during a workshop at the CHUB, Huye, Rwanda, using an in-house developed ELISA as described previously15. Identical antigens were used. The antigens were generated by mock-treating or infecting HeLa cells with VACV Elstree at a multiplicity of infection (MOI) of 1, and harvesting the cell monolayer in 1% Triton X-100 in PBS (supplemented with 1 X cOmplete Mini EDTA-free protease inhibitor [Roche, #11836170001]) once a cytopathic effect was observed across the majority of the culture. Briefly, high-binding 96-well ELISA plates (Corning, #9018) were coated with the cell lysate of either VACV-infected or mock-treated HeLa cells diluted 1:250 in PBS for 1 h at 37 °C. After coating, plates were washed with a washing buffer (PBS + 0.05% Tween20 [Merck, #P1379]) and blocked for 1 h at 37 °C using a blocking buffer (2% skim milk powder [Merck, #70166] in washing buffer). Fivefold dilution series of the heat-inactivated sera (30 min, 56 °C) in blocking buffer starting at a 1:10 dilution were prepared, added to the plates, and incubated overnight at 4 °C. Afterwards, plates were washed with a washing buffer and incubated for 1 h at 37 °C with horseradish peroxidase (HRP)-conjugated rabbit anti-human IgG (1:6000; Dako, #P0214) diluted in blocking buffer. After another washing step, plates were developed with TMB peroxidase substrate (SeraCare, #5120-0047). Absorbance was measured at 450 nm using a microtiter plate reader (Erasmus MC: Anthos 2001 microplate reader; CHUB: BioTek 800 TS) and corrected by subtracting absorbance at 620 nm. A net OD450 response was determined by subtracting the OD450 value from plates coated with mock-treated cell lysate from those coated with VACV-infected cell lysate. A positive reference serum pool was included on every plate and used to calculate 30% endpoint titers by transforming the net OD450 responses per sample to this positive control S-curve. Additional assay validation is included in Supplementary Fig. 5, including a receiver operating characteristic (ROC) curve analysis and correlations of endpoint titers with VACV antigens A27L, A33R, and B5L using MVA vaccination data from non-primed individuals19.

Detection of MPXV clade Ib-neutralizing antibodies

A plaque reduction neutralization test (PRNT) using the newly isolated MPXV clade Ib was developed based on the MPXV clade IIb neutralization assay that we previously described15. Essentially, Vero cells were seeded 1 day prior in 96-well cell culture plates. Heat-inactivated sera (30 min, 56 °C) were twofold serially diluted in AdDF+++ before adding an equal volume of medium containing 300 PFU MPXV clade Ib. All measurement were performed in duplicate. The virus-serum mix was incubated for 1 h at 37 °C before being transferred to the cell culture plates. After incubation for 16 h at 37 °C and 5% CO2, the cells were fixed with 4% paraformaldehyde for 30 min and permeabilized with 70% ethanol. Following washing with PBS and blocking in 0.6% BSA and 0.1% Triton X-100 in PBS for 30 min, MPXV clade Ib was visualized by overnight staining with rabbit-anti-VACV-FITC (1:1000 in blocking buffer; Abbexa, #abx023199) at room temperature. Cell nuclei were counterstained with Hoechst33342 (Thermo Fisher Scientific, #62249). Imaging was done using the Opera Phenix spinning disk confocal HCS system (PerkinElmer) equipped with a ×10 air objective (NA 0.3) and 405-nm and 488-nm solid-state lasers. Infected cells were quantified using the Harmony software (version 4.9, PerkinElmer). The dilution that would yield 50% reduction of plaques (PRNT50) compared with the infection control was estimated by determining the proportionate distance between two dilutions from which an endpoint titer was calculated. When no neutralization was measured, the PRNT50 was assigned a value of 10. The cut-off for positivity was defined as three times the background of a set of five negative validation sera (≤35, negative; >35, positive).

HIV serological analysis

Initial testing for the presence of HIV-specific antibodies was done on-site in 178 out of the 327 serum samples using the Determine™ HIV Early Detect kit (Abbott, #7D2847). Additionally, all 327 serum samples were screened for the presence of HIV-specific IgG antibodies at Erasmus MC to validate and complete the dataset using a diagnostic combination of a Liaison XL HIV Ab/Ag immunoassay (DiaSorin, #310260) and a line immunoassay (Inno-Lia HIV I/II; FujiRebio, #80540) on an Auto-LiPA (FujiRebio) automated system22. All assays were performed according to the manufacturer’s instructions. Sera above the cut-off for positivity (S/CO ≥ 1) were assessed by a line immunoassay to confirm the presence of HIV-specific antibodies. Sera with a S/CO < 1 were considered non-reactive. Samples were considered seropositive if they tested positive in the screening assay and presence of HIV-specific antibodies was confirmed by a secondary method (detection of reactivity either using the line immunoassay and/or during on-site testing). Samples were considered seronegative if they tested negative in the screening assay. Out of the total of 327 serum samples, 15 yielded an inconclusive test result and four could not be tested due to insufficient sample volumes.

Multiplexed mpox immunoassay

Samples were tested for the presence of IgG antibodies specific to nine MPXV proteins (A5, A27, A29, A35, B2, B6, E8, H3, and M1) and three VACV proteins (A27, A33, and B5) using the MpoxPlex assay as previously described by Jones et al.23. Briefly, diluted samples were combined 1:1 in a 96-well plate with a mixture of MagPlex microspheres of twelve distinct regions (DiaSorin, #MC100xx), each coupled to one of the MPXV and VACV proteins. Following a 30-min incubation, wells were washed and 100 μL per well phycoerythrin-conjugated mouse anti-human IgG antibody at a 1:500 dilution was added. After a second incubation and wash, wells were read using a Luminex xMAP INTELLIFLEX system (DiaSorin) with a minimum count of 100 microspheres per well per protein.

Statistical analysis

Fisher’s exact test (two-sided) was used to compare mpox seropositivity between PSW and non-PSW. Simple linear regression was performed on log-transformed VACV ELISA titers to determine the correlation between the ELISA performed at Erasmus MC (Netherlands) and CHUB (Rwanda) by Spearman’s r.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.