Study selection
A total of 1,026 articles were originally identified, and 449 were excluded as duplicates. A total of 577 studies were screened, and 26[8, 19-43] of them were eligible and included in this meta-analysis. The literature selection process is detailed in Figure 1. The basic characteristics of the included studies are shown in Table 1. All 20 cross-sectional studies were of medium quality. Within the case-control studies, there were three medium-quality studies and three high-quality studies (Table. 1). There were no cohort studies involved in our analysis.
Overall, the included studies covered 690,322 samples (AE=54,338, CE=635,984), of which 38,358 (AE=1,588, CE=36,770) were positive based on the combined diagnosis of ultrasound and serological methods. The included studies were from the Tibetan Autonomous Region (n=1), Qinghai Province (n=4), Western China (n=2), Yunnan Province (n=1), Gansu Province (n=3), Xinjiang Province (n=5), Ningxia Province (n=5), Sichuan Province (n=3), and the Tibetan Plateau (n=2). Three studies reported only on AE-infected patients, seventeen reported only on CE-infected patients, and six studies reported on both AE- and CE-infected patients.
Ultimately, thirteen potential risk factors reported in more than three studies were included in the meta-analysis, namely, sex, ethnicity, dog ownership, touching fox skin, lack of washing hands before meals, playing with dogs, herder status, feeding viscera to dogs, drinking nonboiled water, presence of stray dogs, number of household dogs, nomadism and eating raw vegetables.
Twenty cross-sectional studies reported the prevalence of echinococcosis until August 2020, and the pooled prevalence of echinococcosis in endemic districts was 5.52% (95% CI: 5.47%-5.58%).
Potential risk factors for AE
Seven risk factors were recognized among the studies including AE, and a meta-analysis was executed on nine cross-sectional studies[8, 25, 31, 32, 35, 39-41, 43]. The results of the meta-analysis and forest plots are summarized in Table 2 and Figure 2.
Four risk factors were statistically significant. According to the strength of correlation, they were ethnicity (Tibetan/Han) (OR=2.93, 95% CI: 1.81-4.75; p<0.001), herder status (OR=2.66, 95% CI: 2.25-3.14; p<0.001), not washing hands before meals (OR=2.40, 95% CI: 1.34-4.28; p=0.003) and sex (female/male) (OR=1.45, 95% CI: 1.26-1.66; p<0.001).
Potential risk factors for CE
Eleven risk factors were recognized among the CE studies, and a meta-analysis was performed on twenty-three[19-30, 32-42] of the included papers. These papers include six case-control studies[20, 22, 33, 34, 37, 38] and seventeen cross-sectional studies[19, 21, 23-30, 32, 35, 36, 39-42]. The results are shown in Table 2 and Figure 3.
All of the risk factors were statistically significant. According to the strength of the correlation, the top three were ethnicity (Tibetan/Han) (OR=3.18, 95% CI: 1.55-6.52; p=0.002), nomadism (OR=2.71, 95% CI: 1.65-4.47; p<0.001) and drinking nonboiled water (OR=2.47, 95% CI=1.37-4.47; p=0.003).
Sensitivity analysis
The sensitivity analysis revealed that for most of the risk factors, the results were stable. However, when we removed two studies (P.M. SCHANTZ 2003 and Xianglin Wu 2010) on AE, the heterogeneity of playing with dogs and dog ownership declined markedly, and their corresponding results became statistically significant. Similarly, when we removed the studies of CE He Ye 2019, AiLuo 2014 and Wenting Wu 2018, the heterogeneity of sex, ethnicity and the presence of stray dogs dropped below 50%. More details are shown in Additional file 1.
Publication bias
The publication bias test was performed for all the risk factors included in this study (see Additional file 1). Based on the results of Egger’s test and the funnel chart, the CE risk factors sex, herder status and feeding viscera to dogs exhibited publication bias. Other risk factors had no bias; for instance, the p value of Egger’s test for the AE risk factor sex was greater than 0.05, and the funnel chart was substantially symmetric (Figure 4).
Subgroup analysis
The study design-specific subgroup analysis was only conducted on CE risk factors because all of the AE articles were cross-sectional studies. The results are shown in Additional file 1. In the case-control studies, two risk factors were identified and were statistically significant: dog ownership (OR=1.35, 95% CI: 1.03-1.83; p=0.029) and feeding viscera to dogs (OR=2.76, 95% CI: 2.00-3.83; p<0.001). In the cross-sectional studies, seven risk factors were identified and were statistically significant. According to the strength of the correlation, the top three were ethnicity (OR=3.71, 95% CI: 1.60-8.59; p=0.002), not washing hands before meals (OR=2.37, 95% CI: 1.40-4.00; p=0.003) and herder status (OR=2.30, 95% CI: 1.74-3.04; p<0.001). The heterogeneity of all the CE risk factors decreased to varying degrees. The results of sensitivity analysis and publication bias are shown in Additional file 1.
In general, the studies had a wide geographical distribution and involved seven provinces in China. The subgroup analysis based on study region was performed for Ningxia, Qinghai and Xinjiang provinces and included 3 risk factors, 2 risk factors and 1 risk factor, respectively. The CE risk factors sex in Xinjiang and drinking nonboiled water in Ningxia had high heterogeneity, and the test of their overall effect was not significant. However, the heterogeneity of other risk factors was considerably low. An additional file shows this in more detail(see Additional file 1). The results of sensitivity analysis and for publication bias are shown in Additional file 1.