Dengue virus (DENV) infection is a leading vector‑borne public health problem worldwide. The virus exists as four antigenically distinct serotypes (DENV‑1 through DENV‑4). Primary infection confers durable immunity to the infecting serotype but only transient cross‑protection against others, permitting simultaneous circulation of multiple serotypes. Multi‑serotype transmission influences population immunity, epidemic potential, and the geographic structure of dengue risk. Despite this importance, spatiotemporal characterization of serotype co‑circulation at sub‑national resolution remains limited in many endemic settings.
Mexico, with over 130 million inhabitants distributed across 2,471 municipalities and a wide range of ecological zones, exhibits heterogeneous dengue transmission dynamics. The national surveillance system maintained by the General Directorate of Epidemiology includes laboratory confirmation and serotype data for a substantial fraction of reported cases, enabling municipality‑level analyses. The interval from 2020 to 2025 featured substantial serotype turnover and the regional reemergence of DENV‑3, culminating in a highly synchronized 2024–2025 season.
This study aimed to identify and characterize spatiotemporal clusters of dengue incidence in Mexico from January 2020 through December 2025, applying Kulldorff’s space‑time scan statistic independently to each serotype and to all serotypes combined, and to quantify the extent and geographic distribution of serotype co‑circulation at the municipal level.
The analysis used 103,426 PCR‑confirmed, serotyped dengue cases reported to the national surveillance system between January 2020 and December 2025. Cases covered 1,547 of Mexico’s 2,471 municipalities. Kulldorff’s space‑time scan statistic under a discrete Poisson model was applied separately for each serotype and for the combined dataset to detect statistically significant clusters of excess incidence in space and time.
Co‑circulation was operationally defined as the spatiotemporal overlap of significant clusters from at least two different serotypes within the same municipality for one or more epidemiological weeks. Pairwise temporal overlaps between serotype clusters were quantified, and the frequency of active co‑circulation across epidemiological weeks was reported.
Details on parameter settings for the scan statistic, cluster significance thresholds, and additional analytic choices were reported in the source article. The underlying, de‑identified datasets used for these analyses were made publicly available in a Zenodo repository and as supporting information accompanying the publication.
Across the study period, the space‑time scan detected 159 statistically significant clusters when all serotypes were analyzed together. DENV‑3 emerged as the dominant serotype, comprising 64.4% of PCR‑confirmed, serotyped cases and yielding an annual incidence of 59.9 per 100,000 — consistent with the documented regional reappearance of DENV‑3 following a prolonged period of low circulation.
The 2024–2025 dengue season produced a nationally synchronized epidemic that affected geographically distant regions simultaneously. Co‑circulation of at least two serotypes was detected in 1,545 municipalities. Notably, 217 municipalities experienced simultaneous clustering of all four DENV serotypes during the study window.
Active co‑circulation — defined as the presence of overlapping serotype clusters in the same municipality during the same epidemiological week(s) — was present in 275 of 311 study weeks. Mean pairwise temporal overlap between serotype clusters ranged from 8.0 to 12.0 weeks, and maximum pairwise overlaps reached 29–30 weeks in some municipality‑serotype pairs.
Geographically, four‑serotype co‑circulation and other multi‑serotype clustering were concentrated along the Gulf coast, across the Yucatán Peninsula, and in northeastern Mexico. The 217 municipalities with synchronous four‑serotype clustering were highlighted as loci of particular epidemiological interest.
Applying a reproducible space‑time cluster detection method to a large national series of PCR‑confirmed, serotyped dengue cases provided municipality‑level maps of serotype‑specific and multi‑serotype activity across Mexico from 2020 to 2025. The predominance of DENV‑3 in this period aligns with reports of its regional reemergence and with hypothesized reductions in population‑level DENV‑3 immunity after a prolonged absence.
The identification of 159 significant clusters and widespread co‑circulation in nearly all sampled municipalities underscores the complex, heterogeneous nature of dengue transmission at sub‑national scales. The observed national synchronization during 2024–2025 suggests broad ecological or epidemiological drivers that produced concurrent increases in transmission across diverse regions.
Municipalities showing simultaneous clustering of all four serotypes — 217 in total — may warrant prioritized epidemiological investigation and potentially intensified surveillance, given their status as areas where serotype interactions and shifts in population immunity could alter disease risk patterns.
The study demonstrates that municipality‑level, serotype‑specific cluster detection is feasible using routine surveillance data when serotyping is available, and that such an approach can produce operationally relevant information for serotype‑aware monitoring and response.
Between 2020 and 2025, four‑serotype co‑circulation of dengue was documented at municipal resolution across Mexico. DENV‑3 dominated the period, and the 2024–2025 season produced a nationally synchronized epidemic. Co‑circulation was widespread — present in 1,545 municipalities — with 217 municipalities experiencing simultaneous clustering of all four serotypes. Active co‑circulation occurred in most epidemiological weeks studied, with meaningful temporal overlap between serotype clusters.
This municipality‑level, spatiotemporal framework offers a reproducible method to track multi‑serotype activity and can support serotype‑aware dengue surveillance and targeted public health actions at sub‑national scales.
The datasets analyzed in the study are publicly available in a Zenodo repository and as supporting information to the published article. The authors reported receiving no specific funding for the work and declared no competing interests.