A large national study was implemented to evaluate the prevalence and molecular characteristics of HEV in pigs across the country. Overall, our anti-HEV IgG prevalence (approximately 58%) was very similar to the other studies in Germany (47%), Italy (50.2%), Japan (56%), Philippines (50.3%) and Laos (51.2%) [33–37]. Some studies showed that a higher seroprevalence of IgG has been reported as pigs grow older [38, 39]. However, we were not able to evaluate it because sera samples were obtained from fattening pigs (6–8 months old and weighing 60–120 kg) at slaughterhouses. Interestingly, TPs of three provinces (Dak Lak, Nghe An and Son La) showed more than 80%, suggesting a high potential disease transmission to humans. Overall, the RNA-positive rate (6–7%) in pooled fecal samples was very similar to other studies in Laos (11.5%), Philippines (7.4%) and Thailand (3–7%) [21, 33, 40, 41].
Previous studies have provided evidence that there was an epidemiological association between consumption of pork meat and HEV cases [42, 43]. In Vietnam, most of the minorities/farmers in rural areas raise pigs in backyards with poor sanitary conditions. They have more opportunities to come in contact with pigs and contaminated wastage premises. In addition, it is very common that the local people used to consume uncooked pork, livers, meats, sausages, unwashed vegetables and insufficiently treated drinking water. Therefore, it might be possible that people who are exposed to pig and pork meats are at an increased risk of HEV transmission [44]. Further studies are needed to evaluate whether this exposure route poses a public health concern as well as to better understand the human behavior for preventing HEV infections.
The MMC host discrete traits tree provided information about the transmission of HEVs between humans, swine and other animals. Humans are the main factor to spread HEV to other species and the environment through everyday life. Pigs are a well-known as a reservoir of HEV that transmit the virus to humans. In addition, the host phylogenetic tree indicated that HEV can be transmitted from camel to humans. This tree showed that HEV genotypes 3 and 4 are considered zoonotic viruses, and the new HEV genotype 7 can be transmitted from camel to humans. Therefore, HEVs are now recognized as an emerging zoonotic agent.
We found that both genotypes HEV 3 and 4 were detected in domestic pigs across the country which was in line with other Asian countries [44–46]. Most of them (15/19) were classified into the existing sub-genotype 3a which is the most commonly detected subtype in Asia (including Japan, South Korea and Philippines) [33, 47, 48]. Exclusively, HEV 4 was only detected in Son La province which is on the borders with China and Laos [21, 44, 49]. Therefore, it might provide evidence that this type is circulating in these areas. In Son La province, the percentage of minorities (e.g. Thai and H’Mong) is relatively high compared to other areas in Vietnam (the majority ethnic group is Kinh, accounting for 85.7% of the population) [50]. In general, ethnic minorities have lower education levels and incomes than the Kinh group, which has led to limited awareness of hygiene and food safety in Vietnam [51]. In addition, other studies have suggested that people with high education had a better perception of food safety in comparison with those with low education [52, 53]. Ethnic minorities may benefit from efforts to enhance public awareness of food safety and disease prevention. Further epidemiological study needs to be conducted to establish a transmission link between pigs and humans in Son La province. Also, one recent study showed that the HEV seroprevalence was higher among individuals occupationally exposed to pig and port meat [25]. Therefore, the human health impact of HEV should be properly defined to establish appropriate interventions.
The main limitation of our study was that our blood and fecal samples may not be representative because blood samples were only collected at slaughterhouses and were disproportionally collected depending on the number of pigs at each farm. Our research team had to consider the reality that large and medium scale farmers were not willing to participate (mainly for biosecurity reasons) in our study. Although, our pig samples were not representativeness of the general population, we think that our findings provide valuable information on the epidemiology of HEV in Vietnam. It could be possible that the TP of HEV may be under/overestimated due to a random error with our measuring facilities. In addition, our prior estimates for the sensitivity and specificity of the ELISA test were obtained from our experiment and therefore, we are not clear which estimates are suitable for the Vietnamese context. We conducted some simulations to explore what happens with N = 80, 70, 60, etc. out of 95 samples. There was a large discrepancy between the AP and TP as the positive cases decreased. The main reason was that our less confident Bayesian model put more weight on the priors.
A Bayesian method provides a chance to integrate prior information with observed data, which is useful for estimating values for both prevalence and diagnostic characteristics of tests [54]. If a diagnostic test has less than 100% sensitivity and specificity, the estimated prevalence would be biased. Therefore, a Bayesian analysis was utilized to estimate the TP from AP. Overall, Bayesian models are very sensitive depending on the priors, so we used a Jeffreys prior to minimize the impact of prior to the posterior distribution as there was no prior information for HEV prevalence in the study sites. This study was the first attempt to estimate the TP for HEV in Vietnam and demonstrates how a Bayesian approach can be utilized to better estimate the prevalence of diseases.