Of 629 paired samples, 14 contained detectable HEV RNA in plasma and cecal samples. RNA was detected in 22 additional plasma samples and in 93 additional cecal samples. The prevalence of current infection defined by detectable plasma HEV RNA at any level, adjusting for clustering within farms, was 5.7% (95% CI 3.9%–7.6%). Similarly, the prevalence of current infection, defined by detection of HEV RNA in cecal samples at any level, again adjusting for clustering within farms, was 17.0% (95% CI 14.0%–20.0%). Taking detection of HEV RNA in either plasma or cecal samples as evidence of infection in 129 animals, we determined that the prevalence of current infection in pigs at slaughter was 20.5% (95% CI 17.2%–23.8%).

The viral load in the plasma ranged from detectable but below the limit of quantification to a maximum of 10⁶ IU/mL. Similarly, values in cecal content ranged from 40 to 7.4 × 10⁷ IU/mL. Six plasma samples contained in excess of 10² IU/mL HEV RNA. Although high-level viremia in plasma samples was generally reflected by a high level of shedding in the cecal content (Table 1), most HEV RNA signals in either plasma or cecal content were present in only 1 of these paired sample types and then often only at low levels (plasma only mean C_t 40.5, range 34.3 to below levels of quantitation; cecal only mean C_t 37.2, range 24.2–45).

A total of 584 of 629 pig plasma samples were positive by manufacturers’ criteria in the Wantai total antibody assay (BR 1.0), and the pigs from which these samples were taken were considered seropositive. After adjusting for clustering of pigs within farms, we determined a seroprevalence of 92.8% (95% CI 90.7%–95.0%). Of the 584 plasma samples positive for antibodies against HEV, 276 (47.4%) also contained measurable IgM (reactivity >3.3 AU/mL), indicating a relatively recent infection. The 45 (7.2%) plasma samples from the survey that were unreactive by Wantai were tested for HEV IgM and IgG. Fifteen contained HEV IgG, albeit at low levels (mean BR 1.7, range 1.1–3.1), and 3 of the 15 contained HEV IgM (66.1, 33.4, and 18.3 AU/mL). Two of 30 that were unreactive for HEV IgG contained HEV IgM only (11.3 and 14.1 AU/mL).

Most (91%, 117/129) pigs with detectable HEV RNA at any site were seropositive, and half of the virus-positive pigs (48%, 56/117) also had detectable HEV IgM (IgM reactivity ≥3.3 AU/mL; Table 2). The remaining 220 seropositive pigs with detectable IgM did not have detectable HEV RNA. Pigs with high-level cecal content and plasma viremia had current or recent acute infection, and all had HEV IgM (Table 1). Seven of 9 seropositive pigs that were both viremic and shedding virus in the cecal content were seropositive for HEV IgM.

Twelve infections were identified in pigs that were seronegative by manufacturers’ criteria in the Wantai assay in which HEV RNA was present at low level in cecal contents, plasma samples, or both (Table 2). Additional serologic testing of the 4 samples from Wantai seronegative pigs in which HEV RNA was detected in both cecal fluid and plasma (Table 3) showed 1 plasma sample to contain IgG just over the cutoff, another to contain low levels of IgM (3.4 AU/mL), and 2 samples to be reactive but below cutoff for IgG and IgM.

Figure

Figure. Phylogeny of genotype 3 hepatitis E viruses (HEVs) from pigs and patients with acute hepatitis in the United Kingdom. Nucleotide sequences of a 304-nt open reading frame 2 fragment (positions 5994–6297...

Amplification of RNA for sequencing was attempted on all samples containing quantifiable HEV RNA. However, due to low levels of input, virus sequences were only derived from 6 pig plasma and 21 cecal content samples. Where viral RNA could be sequenced from both plasma and cecal content, the animals had identical virus sequences at both sites (n = 4). All of the 23 unique sequences belonged to G3; all but 1 clustered within the UK human group 1 (Figure). A single cecal-derived sequence clustered within group 2 (4). Most viruses from UK patients with acute hepatitis clustered within group 2.

We have demonstrated that HEV G3 RNA was present in plasma, cecal contents, or both in 20.5% of all pigs tested, which is likely to approximate the lower limit for a farm prevalence, given the modal number of pigs per farm sampled. A recent survey in France described a farm-level prevalence of 30% for HEV G3, with at least 1 RNA-positive animal detected in 27 of 90 farms sampled (22). Similar findings in Canada (23) and Italy (24) have been documented for other tissues and organs. In Spain, 11.5% of liver or bile samples contained HEV RNA (25).

Porcine HEV is acquired through ingestion of virus (26), and the extended duration of cecal fluid virus and, therefore, fecally shed virus in this study, exceeding that of the more transient plasma viremia (11), will favor porcine fecal-oral transmission. Extended fecal shedding means that older but immunologically naive animals joining the groups at finisher units may become infected closer to the time of slaughter. High cecal carriage and fecal excretion in some animals indicates the potential existence of superspreaders around the time of slaughter. The resulting environmental contamination may complicate attempts to control onward transmission of HEV in pigs.

High-level viremia was unusual, occurring in only 6 pigs, but although rare, this represents the best candidate source of potential dietary transmissions by meat ingestion (27). Of these 6 pigs, 1 infection was in the early acute seroconversion phase. Two were in the acute phase of the infection, with high IgM levels, and the remaining 3 were later in the acute infection, with low IgM levels. All 6 pigs had detectable plasma IgM (Table 1), which probably indicates recent infections. We postulate that plasma viremia is a good marker for possible dietary transmission by meat products. The reported absence of porcine adenovirus (another virus found in pig feces) in HEV-contaminated sausages (12) also implicates viremia as the source of virus rather than fecal contamination at the abattoir.

We have reported (4) that the viruses causing current cases of G3 hepatitis E in humans fall into 2 phylogenetically and temporally separable groups, 1 and 2. These groups derive from the analysis of a 304-nt fragment of ORF2 with levels of bootstrap support in the region of 70% depending on the number of sequences analyzed. Much stronger support for these 2 groups is obtained when a larger 1,300-nt region of ORF2 is analyzed (data not shown). Most sequences of strains in humans contemporary to this study fall within group 2 (along with reference sequence 3c; Figure). In contrast, most G3 HEV (22 of 23) sequences obtained from UK pigs fall into group 1 (along with reference sequences of 3e, 3f, and 3g; Figure). Notably, the group 1 pig viruses are almost identical to those circulating in UK pig populations a decade ago (data not shown), perhaps demonstrating a longstanding zoonosis that may be reflected in the continuing group 1 cases in humans in England and Wales. The sole group 2 G3 HEV was from a pig from Scotland and falls outside the dominant human clade, sitting among a minor grouping.

In England, as in most Western industrialized countries, HEV infection in humans comprises travel-associated (G1 and G3; potentially G2 and G4) and indigenous (G3) infections. Our findings indicates that, in the United Kingdom, indigenous HEV human-to-human infection will be rare, and nontravel-related hepatitis E results from HEV G3 dietary acquisition, as shown by recent and continuing case−control studies (13).

Our findings suggest that slaughtered UK pigs are unlikely to be the source of most HEV G3 infections in humans in England and Wales. Although one could postulate the coexistence of group 2 viruses circulating in UK pigs, the failure to detect this virus at the time of slaughter in 22 of 23 pigs from whom virus could be sequenced would seem to render unlikely high-level viremia and possible infectivity of group 2 viruses through the contribution of UK pig meat to the food chain. We were not able to sequence most infections identified because of low viral levels in the reactive analytes. Consequent with our current understanding about infections in humans, it is instead plausible that the dominant HEV infections in humans that could be linked to pork consumption (9) derive from imported meat or meat products, although we are unable to establish the precise source. Other routes of transmission from hitherto unidentified animal sources to humans also cannot be excluded.

The GenBank sequences most closely related to the dominant clade of UK human HEV sequences derive from a wild boar from Germany sampled in 2006 or from pigs from Italy and France, all of which were sampled more than a decade ago. Trade of both pigs and pig products is a common practice in Europe, with imports accounting for around half of the pig meat consumed in the United Kingdom (15). Some of the acute hepatitis cases in UK patients associated with HEV G3 infection clustered together with most UK pig HEV isolates and the 3e reference sequence, and it is possible that UK pork is one of the sources of these infections. However, our findings conversely suggest that additional sources of pork may be responsible for further cases of human G3 HEV acute hepatitis clustering outside UK pig sequences.

The timing of HEV infection in pigs is a key consideration for informing future management options for mitigation of risk to public health. Because just over half of the pigs in this study had recently cleared infection, exhibiting antibodies against HEV and detectable IgM in the absence of detectable HEV RNA, they were likely to have been infected in the 2 months before slaughter. In this study, a few pigs remained seronegative at the time of slaughter (7.1%, or 4.5% if the 17 Wantai-negative samples reactive for IgG and IgM are considered seropositive) (Table 3). Our data indicate a lower prevalence of susceptibility (seronegativity) at slaughter in UK pigs than in continental Europe (7–10,21,28).

The demonstration of pigs viremic at time of slaughter explains the detection of HEV in processed food products in the absence of apparent hygiene problems. It seems likely that we are measuring by proxy, through the infection of humans, the spread of group 2 HEV G3 infection in pig populations that supply meat to the United Kingdom. The continuing annual increase in hepatitis E cases in humans in the United Kingdom may reflect changes in trade, processing, or husbandry in other countries, or a societal change in how pig meat is consumed. Consideration has to be given to developing a better understanding of this widespread zoonosis. Given the ubiquity and transmissibility of porcine HEV infection in swine, with the simultaneous absence of clinical signs, economic effects, or regulation, the elimination of HEV in pigs is unlikely in the near future.

Immunization of pigs against HEV is currently only a theoretical option because there no vaccines are on the market for pigs. In addition, one would have to ensure that any intervention is not merely delaying infection and increasing the likelihood of viremia at the time of slaughter (29,30). An alternative approach through using husbandry practices to facilitate natural immunity in early life should also be considered. Further investigations into HEV infection in humans and pigs in the United Kingdom and other countries are required to inform farming management practices to reduce active porcine infection rates at the time of slaughter.