Vital Surveillances: Changing Patterns of Epidemiological Characteristics and Spatial-Temporal Clusters of Human Brucellosis Based on County Level — China, 2011–2023

Human brucellosis surveillance data from 31 PLADs in China (2011–2023) were obtained from the National Notifiable Disease Reporting System (NNDRS). The system incorporated both clinically diagnosed and confirmed cases based on national health industry standards. Clinically diagnosed cases were defined as presumptive cases (those with epidemiological history and clinical manifestations) with positive screening test results, while confirmed cases were defined as presumptive or clinically diagnosed cases with confirmatory experimental evidence (5). We conducted descriptive analyses following established definitions for annual incidence (AI) and regional comparisons between southern and northern regions (3). Trend analyses employed Joinpoint regression models to calculate the average annual percentage change (AAPC). Statistical analyses and disease data visualization were performed using SAS (version 9.4, SAS Institute Inc., Cary, USA) and ArcGIS Desktop (version 10.6; Esri; Redlands, California, USA).

For spatiotemporal cluster identification, we employed a Poisson retrospective spatiotemporal scan statistic using SaTScan^TM software developed by Kulldorff (5). The analysis utilized a discrete Poisson distribution model to examine high spatial clustering areas across the Chinese mainland. Each scanning window was evaluated using log-likelihood ratio (LLR) and relative risk (RR) calculations to test hypotheses. The window with the maximum LLR was designated as the primary cluster, while other statistically significant windows were classified as secondary clusters (4). The analysis was conducted at the county level with daily temporal precision. The temporal aggregation period was set to 14 days, corresponding to the brucellosis incubation period. Maximum spatial and temporal window sizes were configured to encompass no more than 5% of the at-risk population and 50% of the study duration, respectively.

From 2011 to 2023, the Chinese mainland reported 656,628 cases of human brucellosis across 31 PLADs. Males were disproportionately affected, with a male-to-female ratio of 2.6∶1 (475,706 males versus 180,922 females). The mean age of affected individuals was 47.59±14.41 years, with 89.3% of cases occurring in those aged 25–69 years. A notable demographic shift emerged over the study period: the proportion of cases in individuals over 60 years increased annually, while the proportion in those aged 15–29 years showed a consistent decline, resulting in a net demographic shift of 7%. Throughout the study period, agricultural workers and herders constituted the predominant occupational groups affected, accounting for 82.7% to 89.7% of all cases.

The nationwide incidence of brucellosis demonstrated marked fluctuations throughout the study period. Reported cases increased substantially from 38,151 (2.8/100,000) in 2011 to 70,439 (5.2/100,000) in 2023, with a mean annual incidence of 3.66 per 100,000. Peak incidence occurred in 2023 at 5.0 per 100,000. A significant upward trend was observed from 2018 to 2023 (AAPC=14.9%, P=0.01). Southern PLADs exhibited a particularly steep increase in reported incidence from 2011 to 2023 (AAPC=31.5%, P<0.01). The disease demonstrated clear seasonal patterns, with 54.3% of annual cases occurring between March and July.

From 2011 to 2023, PLADs in northern China accounted for 96.1% of total brucellosis cases. However, PLADs in southern China exhibited a marked increase in reported cases, rising from 0.6% in 2011 to 7.4% in 2023. During this period, the geographic distribution of affected counties (districts) expanded substantially, from 834 (25.4%) in 2011 to 2,290 (76.9%) in 2023. In southern China specifically, the number of affected counties (districts) increased nearly fourfold, from 96 (9.0%) in 2011 to 932 (40.7%) in 2023 (Figure 1). Counties (districts) reporting high incidence rates exceeding 100 per 100,000 population increased from 44 (1.5%) in 2011 to 71 (2.4%) in 2023, predominantly concentrated in northern regions, particularly Inner Mongolia and Xinjiang. Similarly, counties (districts) with incidence rates greater than 10 per 100,000 showed substantial expansion, increasing from 247 (8.4%) in 2011 to 525 (17.6%) in 2023. The geographic spread of brucellosis has demonstrated a consistent southward progression from northern China since 2013 (Figure 2 and Supplementary Table S1)

Figure 1. 

Human brucellosis in northern and southern China, 2011–2023. (A) Number of affected counties; (B) Proportion of reported cases.

Figure 2. 

Geographical distribution of incidence of the reported human brucellosis in 31 PLADs of China, 2011–2023. (A)–(M) From 2011 to 2023.

Abbreviation: PLAD=provincial-level administrative division.

Map approval number: GS京(2025)0527号.

From 2011 to 2023, we identified 227 spatiotemporal clusters of human brucellosis, comprising 13 primary and 214 secondary clustering areas. The primary clusters initially centered in eastern Inner Mongolia before progressively expanding to encompass eight adjacent PLADs (Heilongjiang, Jilin, Liaoning, Hebei, Shanxi, Shaanxi, Ningxia, Gansu) and part regions in Xinjiang. This expansion pattern persisted throughout the study period. Notably, in 2015–2016, novel clusters emerged in southern PLADs, specifically affecting Guangdong, Hubei, and Guangxi (Supplementary Tables S2–S3). By 2017, the disease had spread to 183 counties (districts), predominantly concentrated in northern PLADs including Shanxi, Inner Mongolia, Heilongjiang, Xinjiang, Hebei, Liaoning, Gansu, and Shandong, with Xizang reporting secondary clusters for the first time (Supplementary Tables S2–S3). The period between 2018 and 2021 saw a reduction in affected counties, with clusters reconcentrating in Inner Mongolia and adjacent PLADs. While clusters in Yunnan persisted from 2019 onward, Xinjiang ceased reporting clusters from 2020. Re-emergence of spatiotemporal clusters in Xinjiang in 2023. In 2022−2023, there was a resurgence in affected counties, primarily in the Inner Mongolia and its surrounding eight PLADs, as well as in Shandong, Henan, Qinghai, and Xinjiang PLADs. The temporal evolution of primary clustering showed distinct geographical shifts: from 2011 to 2014, it was predominantly in eastern Inner Mongolia and neighboring PLADs such as Heilongjiang and Jilin; from 2015 to 2017, it shifted to the western Inner Mongolia and expanded to Xinjiang, Ningxia, and other regions; from 2018 to 2021, it returned to eastern Inner Mongolia; from 2022 to 2023, shifted to western Inner Mongolia as well as Shanxi, Shaanxi, Ningxia, Gansu, Qinghai and Xinjiang regions (Figure 3). Among the 227 spatiotemporal clusters identified, primary clusters occurred predominantly from February to August, while secondary clusters spanned January to July (Table 1).

Figure 3. 

Spatial clustering of the reported brucellosis cases in 31 PLADs of China, 2011–2023. (A)–(M) From 2011 to 2023.

Abbreviation: PLAD=provincial-level administrative division.

Map approval number: GS京(2025)0527号.

This study reveals significant shifts in the epidemiological landscape and spatiotemporal distribution of brucellosis across China from 2011 to 2023. Our findings demonstrate the persistent severity of the epidemic, characterized by both a resurgence in reported cases and substantial geographic expansion across counties (districts). The disease pattern has evolved from its initial concentration in northern pastoral provinces to encompass adjacent grassland and agricultural regions, ultimately extending into southern coastal and southwestern areas. Spatiotemporal analysis revealed distinct clustering patterns, with primary hotspots centered in Inner Mongolia and traditional pastoral regions of western and northeastern China. Throughout the 13-year study period, the disease exhibited a clear eastward-to-westward progression within Inner Mongolia, subsequently spreading to adjacent PLADs and part regions in Xinjiang. A particularly noteworthy development since 2019 has been the emergence of new hotspots in the southern PLAD of Yunnan, demanding immediate attention from public health authorities. In 2023, many hotspots appeared in Xinjiang and Qinghai, which may be related to its improved ability to detect brucellosis.

The age distribution analysis reveals a concerning trend: the proportion of infections among individuals over 60 years has shown consistent annual increases. This demographic shift likely reflects the aging population structure among Chinese farmers and herders engaged in livestock management (5). This finding underscores the urgent need for targeted health education initiatives specifically designed for this vulnerable age group.

The rising incidence of human brucellosis in China, particularly in southern regions, can be attributed to several interconnected factors. First, there exists a robust positive correlation between human brucellosis incidence and livestock outbreak frequency (6-7). Despite persistent and intensifying animal outbreaks, preventive measures — including livestock vaccination programs and quarantine protocols — have proven insufficient (4,8). Recent surveillance data indicate elevated seroprevalence rates among dairy cattle and sheep populations (9-10). Second, the surge in livestock product demand has intensified north-to-south animal transportation, exacerbating the epidemic’s spread to southern regions (4,8). Third, inadequate regional quarantine measures and limited stakeholder awareness of quarantine protocols have contributed to outbreaks in southern China, as evidenced by cases linked to sika deer and raw goat milk consumption (11-12). Furthermore, the diversification of livestock production systems — encompassing free-range, grazing, and intensive production methods — has complicated disease prevention and control efforts (10,13).

The spatial-temporal clustering patterns of human brucellosis have aligned with broader epidemic trends, with Inner Mongolia and adjacent regions remaining critical focal points. Molecular epidemiological studies have revealed genetic homology between Brucella strains isolated from northern provinces, including Xinjiang and Inner Mongolia (14), and those found in Shanxi and neighboring provinces (15). This genetic relationship underscores the need for enhanced quarantine measures and stringent management of livestock and dairy product movement between high-incidence regions to curtail disease transmission. In Yunnan, the rapid emergence of brucellosis cases, potentially linked to expanding dairy goat operations, necessitates the implementation of prevention and control strategies specifically tailored to local livestock industry developments.

Despite the implementation of the National Brucellosis Prevention and Control Plan (2016−2020) (NBPCP), brucellosis incidence rates have continued their upward trajectory. In response, brucellosis was designated as a priority disease for prevention and control in 2022. This designation necessitates comprehensive animal-based prevention and control strategies from governmental bodies at all levels, encompassing immunization, quarantine, and culling protocols. Furthermore, there is an urgent need to strengthen occupational safety measures for at-risk populations to reduce both inter-animal transmission and occupational exposure, thereby mitigating the disease’s escalating impact on human populations. Additionally, management approaches should be tailored to regional specificities, with enhanced oversight in high-risk areas. While this study presents a spatial-temporal aggregation analysis at the district-county scale, future research could explore patterns at smaller administrative levels, such as townships, to provide more granular insights for precise prevention and control strategies.

This study has several limitations. First, county codes for a small proportion (<10%) of cases could not be matched with the corresponding map dataset, though this discrepancy is minor and does not significantly impact the data’s representativeness. Additionally, the reported case numbers may underestimate the true disease burden due to inherent limitations in surveillance data.