The incubation periods of Monkeypox virus clade Ib

Study setting and population

This analysis is based on surveillance data from the Mpox Treatment Center in the Uvira health zone in South Kivu province, eastern DRC. Uvira is a predominantly urban area with approximately 460,000 inhabitants, 65% of whom reside in the city of Uvira. Located on the northwestern shore of Lake Tanganyika, it shares a land border with Burundi and is connected to Tanzania via the lake. The city spans 12 km north to south and 2 km inland, with a high population density. Uvira is historically prone to cholera outbreaks, experiencing 1–2 per year due to inadequate water and sanitation infrastructure, frequent flooding, and population displacement caused by both floods and armed conflicts. The Uvira health zone is one of DRC’s mpox clade Ib hotspots, prioritized for interventions such as vaccination. It reported its first mpox case on May 2, 2024, eight months after the outbreak was declared in Kamituga, a remote mining town 360 km away. Initially, suspected cases, mostly adults, were managed in outpatient or community settings. However, as cases increased, an Mpox Treatment Center (MTC) was established on June 8, 2024, within the Uvira general hospital. With support from Médecins Sans Frontières, mpox care, including essential medicines and meals for patients and caregivers, became free at the Uvira MTC on June 17.

Case definitions and data collection

The official DRC Ministry of Health case definition for mpox includes three criteria: (1) sudden high fever followed by a vesiculo-pustular rash, primarily on the face, palms, and soles; (2) the presence of at least five smallpox-type scars; or (3) fever >38.3°C, severe headache, lymphadenopathy, back pain, myalgia, and severe asthenia, followed within 1–3 days by a progressive rash spreading from the face to other body parts, including the palms and soles.¹² The community case definition is broader, classifying any person with fever and a skin rash as a suspected case. At the Uvira MTC, following outbreak confirmation, a simplified case definition was applied: anyone presenting with a skin rash or papulo-vesicular or pustular eruptions was considered a suspected mpox case.

Starting from June 2024, the clinical staff at the Uvira General Hospital’s Mpox Treatment Centre (MTC) conducted structured interviews with suspected cases. Data were collected on demographics, exposure history, and clinical manifestations, using forms on the KoboCollect platform. A suspected case contact was defined as someone with a rash, lesions and/or similar symptoms to the patient or a confirmed mpox case, contacted in the three weeks prior to symptom onset or diagnosis of the clinical case.

MTC staff collected skin lesion secretions with a dry swab, whereas oro-pharyngeal secretions were taken with a swab that was soaked into a viral transport medium. While lesion swabs were meant to be collected and tested from all suspected cases, only a subset had test results due to stockouts of materials and reagents during this phase of the outbreak. The subset of suspected cases that were tested had dry lesion swabs tested by qPCR with GeneXpert (Xpert® Mpox; Cepheid, Sunnyvale, USA) or the Radi^® Fast Mpox Kits (KH Medical, Pyeongtaek-si, Gyeonggi-do, Republic of Korea). A subset of samples was tested with Tibmolbiol^® qPCR kits (TIB Molbiol, Eresburgstrasse, Berlin, Germany) for sub-clade confirmation.

Statistical inference of the incubation period

The aim of the statistical modeling framework was to infer the distribution of the incubation periods of MPXV clade Ib for different definitions of symptom onset (fever, rash, any symptom), and its variability across key factors, including demographic characteristics and transmission routes. To do so, we used Bayesian methods to obtain a robust estimate of the incubation period distribution accounting for the limitations of the data and known sources of bias in the estimation of epidemiologic delay distributions.¹³ A full model description is given in Supplementary Methods Section S2.

For each reported contact with a suspected case, we had information on whether a single or multiple exposures occurred, the type of exposure, and the date of most recent exposure. Our modeling framework accounted for multiple reported contacts per case, censored times from exposure to symptom onset as only time to most recent contact was reported, and right-truncation of the time from exposure to symptom onset as only contacts in the previous 21 days were recorded). We compared across multiple distributional assumptions typically used for incubation period estimation (Log-normal, Gamma and Weibull), and across different assumptions on the probability of unreported exposures in the community based on the observed epidemic curve (no additional community exposures vs. exposures as a function of the epidemic curve).

We produced separate incubation period estimates for different symptom types: fever, rash, and any symptom. In addition, we produced stratified estimates by demographic characteristics (age [below 5 vs. 5 and above], sex), exposure type (sexual vs. non-sexual, see definitions in Supplementary Material Section S1.1), epidemic period (before vs. after to Sep 1st, 2024, Supplementary Figure S2), and hospitalization status (inpatient vs. outpatient).

Inference was drawn using a Hamiltonian Monte Carlo sampler as implemented in the Stan programming language. Model comparison was performed using estimated leave-one-out cross-validation log-predictive density.¹⁴

Clinical surveillance and reported exposures of mpox cases

Of the 973 cases recorded between June and October 2024, 35.6% (346/973) of suspected mpox cases reported having contact with a suspected mpox case, with 70% (243/346) of them reporting temporal information of these contacts. Among the cases with complete contact information and valid laboratory results (123/243), 72% were PCR-positive (92/123) and 25% were negative (31/123). The distribution of cases by age, sex and self-reported exposure types were similar for all clinical cases and those positive by PCR (Table 1).

The 243 clinical cases with complete contact information reported contact with 320 suspected mpox cases. 79% (192/243) reported contact with a single suspected case and 21% (51/243) to more than one suspected case (Table 1). The majority of cases reported non-sexual physical contact with a case (95%) and/or respiratory contact (84%), with most exposures reported to have occurred within the household. Only 8% of cases reported sexual exposures.

Incubation period estimation

We estimated the incubation period based on data from 92 PCR positive cases with complete data on reported exposure to suspected cases, and performed a sensitivity analysis on 243 clinical cases, including those with negative PCR results as the diagnostic sensitivity of the qPCR tests for MPXV Clade Ib is not well characterized and those who were not tested. Among all cases, the median time from most recent contact (before symptoms) to rash onset was 9 days (Interquartile Range, IQR: 3–14), 8 days (IQR: 2–13) to fever onset, and 7 days (IQR: 2– 14) to any symptoms.

We estimate that the median time from exposure to rash onset is 18.6 days (95% CrI: 15.2-22.6) (Figure 1, Table 2), with 5% of infections estimated to develop rash within 5.4 days of exposure (95% CrI; 3.4-7.6), and 95% estimated to develop rash within 37.6 days (95% CrI: 30.3-47.0). Estimates of the median time to onset of fever (17.0, 95% CrI: 13.5-21.1) and of any symptom (16.1, 95% CrI: 12.8-20.1) were slightly shorter (Table 2).

• Download figure
• Open in new tab

Figure 1. Estimates of Mpox clade Ib rash incubation period.

a) Posterior draws of the incubation period cumulative distribution function (lines, 100 draws) and estimates of quantiles of interest (points: mean, bars: 95% CrI) by diagnostic test result. Estimates are shown for the log-normal distribution assumption, and the main model (no community infections). Dots indicate the median incubation period estimate. b) Estimates of the median incubation period (dots: mean, bars: 95% CrI) by strata for the overall estimates (first facet) and across different case grouping categories against the overall estimate (horizontal dashed line).

We found evidence of differences in the incubation period by transmission route (Figure 1b, Table 3). The median rash incubation period was shorter for putative sexual vs. non-sexual transmission (median 9.5, 95% CrI: 3.7-16.4, vs. 18.9, 95% CrI: 15.3-23.0) (Supplementary Material Section S1.2). We estimate a shorter rash incubation period in the early phase of the epidemic (prior to Sep 1st, median 11.7, 95% CrI: 9.0-15.1) versus the later phase (after Sep 1st, median 24.3, 95 % CrI: 19.5-29.7), consistent with previous evidence of a shift from sexual to non-sexual transmission.⁵ While not statistically significant, the median incubation period was also slightly shorter for PCR-positive cases fifteen years and older compared to others, and for inpatients (hospitalized) vs. outpatients (Supplementary Tables S3-S4). We estimate similar differences in stratified estimates for time to fever and any symptom onset (Figure S4, Supplementary Tables S3-S4).

Estimates of the incubation period using all clinical cases were similar to those for the PCR-positive cases (Figure 1, Table 2). Estimates of the median and lower quantiles of the incubation period were robust to distributional assumptions (Table S1); however, estimates of the time to 95% of cases developing symptoms were longer if a log-normal distribution was chosen.