Preplanned Studies: DiarrheagenicEscherichia coliOutbreak Reporting to Foodborne Disease Outbreaks Surveillance System — China, 2011–2022

Escherichia colitypically exists as a commensal organism in the gastrointestinal tract of humans and animals, remaining harmless under normal conditions (1). However, certain pathogenic strains can cause severe clinical manifestations, including acute diarrhea, hemolytic uremic syndrome (HUS), and thrombocytopenic purpura (2). Foodborne diarrheagenicEscherichia coli(DEC) represents one of the predominant etiological agents of acute diarrhea in low- and middle-income countries (3), and has emerged as a significant pathogen in diarrheal cases throughout China (4).

DEC outbreaks persist globally, affecting both developed and developing nations, with potential fatal outcomes and substantial public health and economic implications (5). Many countries have implemented comprehensive national surveillance systems specifically designed to monitor and track DEC outbreaks, such as the Foodborne Disease Outbreak Surveillance System (FDOSS) in the USA and the Rapid Outbreak Assessment system for multi-country foodborne outbreaks in the European Union (6-7). This study analyzes DEC outbreaks reported to the FDOSS in China during 2011–2022 to inform evidence-based food safety policies and interventions for the Chinese population.

The FDOSS, established in 2011, serves as a centralized system for managing and mitigating foodborne disease outbreaks. For each outbreak, the system collects comprehensive data including reporting unit, temporal and geographic information, food preparation setting, case counts (illnesses, hospitalizations, and deaths), confirmed etiology, and implicated food vehicles. All outbreak investigation reports undergwent systematic review by dedicated personnel. A foodborne DEC outbreak is defined as two or more cases presenting with similar illness and/or one or more deaths resulting from consumption of a common food source, with laboratory confirmation ofE. colias the causative pathogen. Outbreaks involving multiple etiologic agents were excluded from this analysis.

Food source attribution was determined based on epidemiological evidence. Food vehicles were categorized using a two-tiered classification system. The first tier comprises five broad categories (including animal-based foods), while the second tier subdivides foods into 14 mutually exclusive categories. Single-ingredient foods or foods containing ingredients from the same category were classified accordingly. Foods containing ingredients from multiple categories were designated as complex or multiple foods.

Statistical analyses were conducted using R software (version 4.1.2, R Foundation for Statistical Computing, Vienna, Austria). Exact binomial tests were employed to calculate 95% confidence intervals (CI) for outbreak rates, illness rates, and hospitalization rates. Temporal trends were assessed using the Mann-Kendall trend test (M-K test). Population-based rates were computed using demographic data from the National Bureau of Statistics of China (8).

Given the seasonal nature of foodborne DEC infections, we employed a Seasonal Autoregressive Integrated Moving Average (SARIMA) model to generate short-term forecasts (2020–2022) based on historical time series data. The SARIMA model, widely applied across various fields, has demonstrated utility in time series analysis and prediction. For example, it was successfully used to model USA oil consumption patterns during the COVID-19 pandemic, enabling assessment of pandemic-related impacts on consumption from January 2020 to March 2021 (9).

During 2011–2022, 413 foodborne DEC outbreaks were documented in China, resulting in 8,127 illnesses, 2,565 hospitalizations, and one death (Figure 1). The annual averages (rates) were 34.4 outbreaks [0.024 (95%CI: 0.024, 0.025) per 100,000], 677.2 illnesses [0.487 (95%CI: 0.484, 0.490) per 100,000], and 213.8 hospitalizations [0.154 (95%CI: 0.152, 0.156) per 100,000]. The annual rate of reported foodborne DEC outbreaks showed a significant increasing trend from 2011 to 2019 (0.013–0.032 per 100,000) (Pfor trend =0.021), followed by a declining trend from 2020 to 2022 (0.038 to 0.012 per 100,000,Pfor trend =0.085).

Figure 1.

Annual incidence and rate of foodborne DEC outbreaks in China, 2011–2022.

Note: Z andPvalues indicate trend analysis results for the entire period 2011–2022; Z* andP* values indicate trend analysis results for the pre-pandemic period 2011–2019.

Abbreviation: DEC=diarrheagenicEscherichia coli;CI=confidence interval.

School canteens were the predominant outbreak setting, accounting for 21.79% of all outbreaks and the highest proportions of outbreak-related illnesses (3,206, 39.45%) and hospitalizations (1,002, 39.06%). In contrast, fast food establishments were associated with the lowest numbers of illnesses (27, 0.33%) and hospitalizations (10, 0.39%). The single reported death occurred in a workplace cafeteria. Food vehicles were identified in 76.76% (317) of outbreaks, with 187 (58.99%) attributed to a single food category, resulting in 6,000 illnesses and 2,220 hospitalizations. Animal-based foods were the most frequently implicated single commodity (113 outbreaks, 27.36%). The sole outbreak-associated death was linked to multiple foods (fried lesser croaker, quail eggs, lamb liver) served in a workplace cafeteria (Table 1).

Using the SARIMA model, we analyzed and predicted DEC outbreak patterns in China. The model forecasted 51, 57, and 57 outbreaks for 2020–2022, while actual reported numbers were 54, 51, and 17, respectively. Similarly, predicted illness counts for 2021 (1,137) and 2022 (1,389) exceeded actual reported cases (1,125 and 189). Hospitalization predictions (252,240,240) also surpassed actual numbers (181,158,8) during 2020–2022 ( Supplementary Table S1 ).

Geographic distribution analysis revealed substantial regional variation in outbreak occurrence. Yunnan province reported the highest number of outbreaks (n=42), while Beijing, Xinjiang Autonomous Region, and Hebei province each reported the lowest (n=2) during the 12-year period. Provincial outbreak rates ranged from 0.027 per 1,000,000 population in Hebei to 0.899 per 1,000,000 in Yunnan. Illness rates showed similar variation, from 9 cases (0.372 per 1,000,000) in Xinjiang to 1,202 cases (25.74 per 1,000,000) in Yunnan (Figure 2). The single reported death occurred in Inner Mongolia Autonomous Region, with no cross-provincial outbreaks documented.

Figure 2.

Geographic distribution of foodborne DEC outbreaks, illnesses, and hospitalizations by PLAD in China, 2011–2022.

Abbreviation: DEC=diarrheagenic Escherichia coli; PLAD=provincial-level administrative division.

Laboratory confirmation of DEC pathotypes was available for 254 outbreaks (60.50%), accounting for 5,070 illnesses (62.38%) and 1,077 hospitalizations (41.99%). EnteroaggregativeE. coli(EAEC) emerged as the predominant etiologic agent, responsible for 124 outbreaks (48.82%), 2,124 illnesses (41.87%), and 235 hospitalizations (21.82%). EnteroinvasiveE. coli(EIEC) was the least common pathotype, associated with 21 outbreaks (8.27%), 217 illnesses (4.28%), and 150 hospitalizations (19.39%) ( Supplementary Table S2 ).