Hepatitis E Clusters in Shanghai Map to Dense Urban Neighborhoods Near Rivers

Hepatitis E has long lived in the shadow of its viral hepatitis cousins, hepatitis A, B, and C, yet it remains one of the most common causes of acute viral liver infection worldwide. A new six-year analysis of surveillance data from Shanghai, one of the most densely populated metropolitan areas on Earth, now offers one of the most granular pictures yet of where and when this underappreciated pathogen strikes in an urban setting. Drawing on nearly five thousand confirmed cases reported between 2017 and 2022, a team of researchers from the Shanghai Municipal Center for Disease Control and Prevention, Shanghai Jiao Tong University, and Fudan University has mapped the disease’s spatial and temporal fingerprints at the level of individual communities, and the results point to a distinctive urban geography of risk.

The study, published in BMC Infectious Diseases, analyzed 4,668 hepatitis E cases drawn from China’s National Notifiable Disease Reporting System, a nationwide passive surveillance network to which clinicians and laboratories must report diagnoses of legally designated infectious diseases. Over the six-year window, Shanghai recorded an average annual notification rate of 3.14 cases per 100,000 population. That figure places hepatitis E firmly on the map of locally relevant infectious diseases in the city, even though it rarely attracts the public attention devoted to respiratory pathogens or foodborne outbreaks. Hepatitis E virus, or HEV, is transmitted primarily through the fecal-oral route, most often via contaminated water, and through zoonotic pathways, particularly the consumption of undercooked pork, game meat, and shellfish from infected animals.

To understand the disease’s rhythm in time, the researchers applied temporal scan statistics, a method that slides a window of variable length across the surveillance timeline and asks whether case counts within any given interval exceed what would be expected by chance. The analysis identified a statistically significant temporal cluster spanning January 1, 2017 to May 31, 2019, with a relative risk of 1.35 and a log likelihood ratio of 52.18, well beyond the threshold of statistical significance. In practical terms, cases accumulated during this early window at a rate roughly a third higher than the six-year baseline. The team also detected a recurring seasonal signature: notifications clustered in the months from December through May, a winter-to-spring peak that echoes patterns reported in other regions of China and is often linked to seasonal dietary customs, including festival-period consumption of raw or undercooked animal products.

The spatial dimension of the analysis proved more surprising. When the researchers ran a spatial-temporal scan across the entire six-year period, they found one statistically significant cluster with a relative risk of 1.73 and a log likelihood ratio of 135.74. Communities with the highest raw notification rates tended to sit in the southeastern parts of the city, but the most likely cluster identified by the scan statistic was not there. Instead, it was located in the urban core. When the team repeated the scan year by year, the same pattern held: the most likely cluster consistently fell in central urban districts, with secondary clusters appearing in suburban towns on the metropolitan periphery. The discrepancy between where cases are most numerous and where the statistical signal of clustering is strongest is itself informative, because scan statistics adjust for underlying population size and expected case counts rather than simply flagging the tallest bars on a map.

That adjustment matters. Dense urban neighborhoods generate large numbers of cases in absolute terms simply because so many people live there, but a cluster statistic asks a subtler question: are there more cases than the local population structure would predict? The fact that the urban core repeatedly emerged as the most likely cluster suggests that something about these districts, beyond sheer headcount, elevates hepatitis E transmission or detection. Possibilities include greater reliance on food purchased from restaurants and markets, more frequent consumption of seafood and pork dishes, higher turnover of food handlers, or simply better access to healthcare and laboratory testing, which would raise the probability that infections are recognized and reported. The study’s design cannot distinguish among these mechanisms, a limitation the authors acknowledge explicitly.

To probe what distinguishes clustered communities from the rest, the researchers turned to binary logistic regression, a statistical technique that estimates the odds of membership in a high-risk cluster as a function of community-level characteristics. Three variables emerged as significant. Population density showed the strongest association: communities with higher density had markedly higher odds of falling into a high-risk cluster, with an odds ratio of 7.367. The count of shopping malls, used as a proxy for commercial activity and food-service intensity, was also positively associated, with an odds ratio of 1.531. Intriguingly, distance to the nearest river showed a negative association, with an odds ratio of 0.742, meaning that communities closer to rivers had higher odds of clustering. This last finding is consistent with the hypothesis that waterways and the aquatic food chains they support, including shellfish and fish harvested or sold in riverside markets, may play a role in HEV transmission, although the ecological nature of the analysis means the link remains speculative.

The authors are careful, and rightly so, about how far these associations can be pushed. Because the analysis operates at the level of communities rather than individuals, it is vulnerable to the ecological fallacy: a community-level correlation does not establish that the people within a cluster acquired their infections through the hypothesized route. A mall-dense neighborhood may have many cases not because its residents eat at malls but because mall density tracks with other unmeasured features of urban life. Similarly, proximity to a river may correlate with historical settlement patterns, sanitation infrastructure, or dietary traditions rather than with any direct waterborne exposure. The researchers describe their findings as ecological and hypothesis-generating, and they emphasize that confirming the actual transmission routes of hepatitis E in Shanghai will require individual-level epidemiological studies, including case-control designs that compare exposures of confirmed cases with those of matched controls.

Even with those caveats, the study carries practical weight for public health planning. Shanghai sits within the Yangtze River Delta, one of the most economically dynamic and densely interconnected regions in the world, and its surveillance system feeds into national and global assessments of viral hepatitis burden. The World Health Assembly has set targets for eliminating viral hepatitis as a public health threat, and hepatitis E, though often self-limiting in healthy adults, can be devastating for pregnant women, who face elevated risks of fulminant liver failure, and for people with chronic liver disease or compromised immune systems. Knowing that risk concentrates in dense, commercially active, riverside urban communities gives health authorities a concrete template for targeting interventions, whether that means food-safety inspections in high-risk districts, health education campaigns timed to the December-to-May seasonal peak, or enhanced testing of at-risk populations such as pregnant women in clustered neighborhoods.

The methodological approach also deserves attention from the wider infectious-disease community. Spatial-temporal scan statistics, originally developed for cancer cluster detection and later adapted for communicable disease surveillance, have become a standard tool for turning routine notification data into actionable geographic intelligence. By combining them with community-level socioeconomic and environmental covariates in a regression framework, the Shanghai team demonstrated a pipeline that other cities with robust notifiable-disease systems could replicate. The work also underscores the value of long surveillance windows: a six-year dataset made it possible to separate a one-off elevated period in 2017 through 2019 from a stable seasonal rhythm and a persistent geographic core, distinctions that shorter studies would blur.

What remains to be resolved is the biology behind the map. Genotyping of viral sequences from urban and suburban cases could reveal whether a single transmission chain or multiple introductions drive the clusters, and molecular epidemiology could test whether zoonotic strains from the pork supply chain dominate in commercial districts while water-associated genotypes predominate near rivers. Seroprevalence surveys could measure how much asymptomatic infection the notification data miss, since hepatitis E is frequently mild or silent in young, healthy adults. Until such studies are done, the Shanghai analysis stands as a carefully constructed hypothesis: that in a modern megacity, hepatitis E risk is written into the urban fabric itself, concentrated where people, commerce, and waterways converge. It is a hypothesis that public health officials in Shanghai, and in riverine megacities across Asia and beyond, now have both the reason and the roadmap to test.

Subject of Research: Spatial-temporal epidemiology and community-level risk factors of hepatitis E virus infection in Shanghai, China

Article Title: Identification of the spatial-temporal cluster and risk factors of hepatitis E from 2017 to 2022 in Shanghai, China

Article References: Zhi-Tao, M., Ling-Xiao, Q., Kai-Yun, C., Xin, S., Di, X., Zhao-He, L., Yi-Han, L., Kang, C., Jing, L., & Hong, R. (2026). Identification of the spatial-temporal cluster and risk factors of hepatitis E from 2017 to 2022 in Shanghai, China. BMC Infectious Diseases. https://doi.org/10.1186/s12879-026-14463-4

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14463-4

Keywords: hepatitis E, hepatitis E virus, spatial-temporal analysis, disease cluster, Shanghai, epidemiology, public health surveillance, risk factors, population density, zoonotic transmission, viral hepatitis, logistic regression