Preplanned Studies: Epidemiological Features and Spatial-Temporal Clustering of Visceral Leishmaniasis — China, 2011–2022

Visceral leishmaniasis (VL) is the second deadliest parasitic disease globally, caused by Leishmania spp. and transmitted through the bite of female sandflies (1). The World Health Organization (WHO) classifies VL as a neglected tropical disease and prioritizes its elimination in the Roadmap for the Prevention and Control of Neglected Tropical Diseases 2021–2030 (2). China reports three VL types: anthroponotic visceral leishmaniasis (AVL), mountain-type zoonotic visceral leishmaniasis (MT-ZVL), and desert-type zoonotic visceral leishmaniasis (DT-ZVL), each exhibiting distinct epidemiological characteristics (3). Once hyperendemic across 16 provincial-level administrative divisions (PLADs) north of the Yangtze River in the 1950s, VL was largely eliminated in most endemic areas of China by the 1980s, with few cases reported in Xinjiang, Gansu, Sichuan, Shaanxi, and Shanxi PLADs (3). However, recent environmental changes have led to a resurgence in VL incidence. Understanding the epidemiological features and spatial-temporal clustering of this disease is crucial. This study analyzed VL case data from the National Notifiable Disease Reporting System (NNDRS) between 2011 and 2022. Joinpoint regression and spatial-temporal clustering analysis identified epidemiological features and VL hotspots. Findings indicate a large-scale MT-ZVL resurgence and new incidence hotspots within the Loess Plateau and its extensions. These findings underscore the need for proactive measures, particularly addressing the disease burden in children. This study provides policymakers with valuable insights for formulating appropriate strategies.

Data on VL cases in China from 2011 to 2022 were collected from the NNDRS. Indigenous and imported cases were defined according to epidemiological investigations of individual cases. Joinpoint regression (version 4.3.1; National Cancer Institute, Bethesda, USA) was used to identify turning points and divide the disease incidence trend into statistically significant sections. A t-test was used to determine significant differences in VL incidence trends within specific periods. The long-term trend of each linear segment is depicted according to the best-fit result, and the annual percentage change (APC) was calculated. Getis-Ord Gi* statistical data in ArcGIS (version 10.1; ESRI, California, USA) were used for hotspot analysis. The Getis-Ord Gi* statistic is a spatial autocorrelation index based on a weighted distance matrix. It determines the spatial clustering of locations using high (hotspot) or low (coldspot) values, with statistical significance determined by Z scores and P-values. Spatial-temporal cluster analysis was conducted using retrospective space-time permutation scan statistics. Spatial clustering of VL incidence was detected based on a Poisson model using a flexible spatial scan statistic in SaTScan software (version 10.1.2; Information Management Services, Maryland, USA). All incidence data were processed separately for the 3 years for which data were available, and potential spatial clusters were detected using the restricted log-likelihood ratio (RLLR). A Monte Carlo simulation was used for permutation testing, with statistical significance determined by P-value. A P<0.05 indicated a statistically significant cluster.

From 2011 to 2022, 3,041 VL cases were reported from 400 counties in 24 PLADs. Of these, 2,572 (84.6%) were indigenous cases that occurred in 124 endemic counties. These cases comprised 173 AVL, 788 DT-ZVL, and 1,611 MT-ZVL cases. The remaining 469 (15.4%) cases were imported to 276 non-endemic counties (Table 1).

Figure 1 presents the results of the joinpoint regression, showing the long-term trends of crude incidence of three types of VL between 2011 and 2022. The incidence of MT-ZVL declined during 2011–2015 (APC=–17.98, P<0.05) but increased significantly during 2015–2022 (APC=14.06, P<0.05). Conversely, the incidence of DT-ZVL increased rapidly during 2011–2015 (APC=163.20, P<0.05) and decreased during 2015–2022 (APC=–57.75, P<0.05). Overall, the incidence of AVL showed a slight decline between 2011 and 2022.

Figure 1. 

The joinpoint regression analysis for the trend of visceral leishmaniasis incidence.

Abbreviation: AVL=anthroponotic visceral leishmaniasis; APC=annual percentage change.

*indicate statistically significant in the the Annual Percentage Change(APC).

As shown in Table 1, approximately three-quarters of VL cases were distributed across three provinces: Xinjiang (n=947, 31.1%), Gansu (n=829, 27.3%), and Shanxi (n=481, 15.8%) PLADs. A total of 118 endemic counties were identified in eight PLADs, including Xinjiang (28 counties), Shanxi (29 counties), Gansu (19 counties), Shaanxi (14 counties), Henan (13 counties), Sichuan (7 counties), Hebei (5 counties), and Beijing (2 counties) PLADs. In addition, 7 regimental farms from the Xinjiang Production and Construction Corps (XPCC) were also considered endemic counties. Among these counties, three were responsible for 43.9% (1,130/2,572) of indigenous cases: Jiashi County (637 cases) in Xinjiang, Wudu District (283 cases), and Zhouqu County (210 cases) in Gansu Province. Jiashi County recorded the highest annual incidence (1.4/10,000), followed by Zhouqu County (1.3/10,000); other counties reported less than 1.0/10,000. Sixty-eight counties in seven PLADs were identified as re-emerging areas between 2011 and 2022, including 27 counties in Shanxi, 13 in Henan, 11 in Shaanxi, 9 in Gansu, 5 in Hebei, 2 in Beijing, and 1 in Sichuan. A total of 675 indigenous cases were reported in these re-emerging counties, most of which (671/675) were the MT-ZVL type. The remaining four cases, reported from two counties in Gansu Province, were the DT-ZVL type.

Hotspot analysis and spatiotemporal cluster analysis showed consistent results. Major high-incidence regions were observed in southern Gansu Province and northern Sichuan Province from 2011 to 2019, in southern Xinjiang Uygur Autonomous Region from 2014 to 2016, and in the border areas of Shanxi-Hebei provinces and Shanxi-Shaanxi provinces from 2019 to 2022.

VL cases occurred mainly in preschool children, particularly those aged 0–2 years, who accounted for 42.3% (n=1,285) of all cases (Table 2). VL type exhibited different age distributions. Notably, DT-ZVL most commonly affected young age groups, while AVL and MT-ZVL tended to occur in individuals ≥15 years. Approximately 87.8% of imported cases occurred in individuals ≥15 years. In terms of occupation, farmers were the most affected population, excluding preschool children. Farmers accounted for 26.4% and 56.7% of total cases in endemic and non-endemic areas, respectively.

The peak incidence of DT-ZVL occurred from April to May, while that of MT-ZVL occurred from October to November. No significant peak incidence of AVL was observed, likely due to the small number of cases.