Non-travel related Hepatitis E virus genotype 3 infections in the Netherlands; A case series 2004 – 2006

Human hepatitis E virus (HEV) infections are increasingly recognized as an emerging disease in the industrialized countries [1]. HEV infections can be asymptomatic or symptomatic and cause acute hepatitis which usually is self-limiting but may progress into fulminate and even fatal disease [2]. In pregnant women a case-fatality rate up to 31% has been reported [3, 4]. Persistent HEV infections have been described in two separate case reports [5, 6]. HEV infections are caused by an RNA virus recently classified in the genus Hepevirus of the family Hepeviridae which contains four major recognized genotypes (1–4) infecting humans, with distinct geographical distribution [7]. Viruses belonging to a 5^th genotype have been identified in birds [8]. HEV is transmitted by the fecal-oral route and large waterborne outbreaks with high attack rate among young adults (15–40 years) have been described in developing countries with poor sanitary conditions [2, 3, 9, 10]. In developed countries, HEV infections have primarily been related to travel to high-endemic regions. However, following the discovery of genotype 3 HEV in pigs and humans in the US in the mid 1990s, there is an increased recognition of these non-travel related HEV infections in industrialized countries. The HEV seroprevalence measured ranges from 0.9% in France to 2.6% in Italy [11], and recently a seroprevalence of 7.3% was measured among an adult population in Spain [12]. HEV infection is a non-notifiable disease in most countries and the national prevalence of infection and disease due to HEV is mostly unknown.

The long incubation period of HEV (15 – 60 days) complicates identification of a possible source and for most cases of sporadic infections the cause is unknown. In Japan, an identical viral sequence has been identified from a HEV patient and from leftovers of wild boar meat consumed by this patient [13]. This provided the first evidence for a direct link between animals and human infection and supports the hypothesis of HEV as a zoonotic agent. In addition, domesticated pigs and deer are thought to play a role in the transmission of HEV based on the high degree of sequence similarity between HEV strains from these animals and humans [14–18].

In the Netherlands, the first non-travel related case of HEV was diagnosed in 1990 followed by a 12 year period in which no such cases were identified [19]. In the late 1990s, HEV of genotype 3 was identified in 25/115 (22%) pooled fecal samples from Dutch pigs. The first cluster of non-travel related HEV infections in the Netherlands was described in three elderly patients and published in 2003 [20], followed by description of a fatal case in a 58 years old female with preexisting liver disease in 2004 [21]. No link or common source was identified in these investigations. Among 209 non-A, B, C hepatitis patients with no history of foreign travel, a HEV seroprevalence of 6% was found as opposed to 0.5% in a control group indicating locally acquired HEV infections in the Netherlands [22]. By genetic comparison of 14 pig strains with HEV strains from human cases a homology up to 97% was shown, suggesting that swine may be a reservoir of HEV [14]. A seroprevalence study among patients (n = 1027) with unexplained hepatitis with onset of illness between 1999 and 2003, found 9.8% seropositive (IgM and/or IgG) by ELISA confirmed by immunoblot. Of the patients with recent infection (IgM and IgG antibodies), more than 50% could be confirmed by virus detection in serum [18]. Since 2000, approximately 10 cases per year of non-travel related HEV infection have been diagnosed by the Dutch National Institute for Public Health and the Environment (RIVM) which most likely represents considerable underdiagnosis of HEV in the Netherlands.

We conducted a descriptive case study including environmental source tracing and molecular typing of strains to generate hypotheses about possible risk factors and transmission routes for non-travel related HEV infections in the Netherlands.

We received samples from 33 of the 99 laboratories (33%). From January 1 2004 till August 31 2006 we tested 748 samples from 694 patients for HEV. Two laboratories, A and B, submitted 277 and 193 samples, respectively, accounting for 63% of the total number of samples (Figure 1). The remaining 31 laboratories submitted between one and 68 samples (median: 3 samples) during the study period. A total of 19 cases met the case definition and were included in the study (Figure 1). Of these, 17 had a positive HEV PCR and two were confirmed by immunological method according to the case definition described in the Methods section. From laboratory A and B nine and five of the cases were identified, respectively. The remaining five cases were identified from two other laboratories.

Among the 17 cases (89.5%) for whom the HEV RT-PCR was positive material was available from 13 cases for sequence analysis and phylogenic clustering of the 148 bp ORF2 RNA fragment. From one case (ID5) two HEV strains were available, one from serum and one from feces. All strains were identified as genotype 3. In the phylogenetic analysis of these strains together with 13 human HEV strains previously identified in the Netherlands, a 100% match was found for seven pairs of strains (Figure 2). Identical sequences were seen in the two strains from a serum and a fecal sample from the same patient (ID5) and in two cases with reported date of onset April 2005 (ID6) and April 2006 (ID22) with no identified epidemiological link. Furthermore, for two cases with date of onset December 2004 (ID1) and February 2002 who lived in the North and the South of the country, respectively, the sequences were identical. The case from 2002 was not included in our study and no further information could be obtained. Eight strains, seven from cases and one from a surface water sample, formed a cluster with more than 96% sequence homology (Figure 2). Six of these strains came from recent cases and three of these cases lived in the same Dutch province (Overijssel). No other cases were identified from this province.

We extended the cluster analyses to 81 overlapping sequences, including internationally available genogroup 3 HEV strains from animals and humans [27]. Twelve of the 13 HEV strains from patients in this study segregated into two main sub lineages within genotype 3 HEV (Figure 3). Each of these sub lineages contains strains collected from surveillance in pigs in the Netherlands and strains from symptomatic cases in the Netherlands. Nine of the strains clustered with strains mainly from humans and three strains clustered with strains mainly from pigs (Figure 3B). One strain (ID11) clustered with strains outside the two main sub lineages.

Sporadic contact with dogs was reported by 14/19 (73.7%) cases of which six (31.5%) had their own dog. From the dog of one patient, a fecal sample was collected and analyzed but no positive signal for HEV RNA was detected by RT-PCR. None of the cases had contact with rats although eight (42.1%) had contact with other rodents (rabbits and guinea pigs) (Table 4). From the rabbit of one patient a fecal sample was collected and analyzed but no HEV RNA was detected. Analyses of fecal samples from cows (n = 16), horses (n = 4), ponies (n = 2), an Indonesian pig (n = 1), and one pooled cow/sheep fecal sample did not show the presence of HEV RNA. Pig manure used in the garden by one patient did also not reveal the presence of HEV RNA.

We identified 19 cases with non-travel related HEV infection with onset of illness from January 2004 to April 2006. This supports the hypothesis that HEV is endemic in the Netherlands, although infections through consumption of contaminated imported foods can not be ruled out. The geographical distribution of the cases is striking, with the majority of cases living in rural areas in the North/Eastern part of the country and a cluster of cases in the South. However, these results correspond to the location of the two laboratories submitting the majority of samples and must be interpreted with caution. It might reflect a reporting bias as diagnostics of HEV in non-travelers depend on the alertness of the clinician and the microbiologist in charge. The first cases of non-travel related HEV in the Netherlands were reported in the North [20]. In a study from England and Wales, living in coastal or estuarine areas was reported as a risk factor for acquiring non-travel-associated HEV disease [28].

The high male:female ratio and the relatively high median age is difficult to explain but still consistent with previously published data [1, 2, 28, 29]. More than half of the cases had one or more preexisting diseases for which they received medical treatment. This might have influenced the testing scheme of the clinicians and created a selection bias towards a more vulnerable/susceptible case population. However, recent reports describe association between HEV infection and pre-existing disease, such as hyperthyroidism [30] and chronic liver disease [31]. In this case series none of the cases reported other liver disease but their medication might render the liver more susceptible to infection. Sub-clinical infections of HEV are well known [2] and the immune system of an individual might play an important role in the development from infection to clinical disease. Three cases were immunocompromised due to malignant disease and one due to usage of immunosuppressive drugs. Two of the cases with malignancies had received multiple blood transfusions during the estimated incubation period which might be the source of their infection, as transmission of HEV via blood has previously been described [32–34]. Trace back of donor blood for these two cases was, however, not considered at the time of this investigation because a relationship with blood donation was thought to be unlikely by the treating physician. Blood donors in the Dutch blood bank have not previously been screened for HEV but the potential for blood borne HEV has now been included in the risk analysis of an advisory committee of the Dutch blood bank. In a study by Zaaijer et al in 1995, 0.4% of 1,275 samples from blood donors tested positive for HEV antibodies [35]. In a study of sera from 167 healthy individuals two samples (1.2%) were IgG and IgM positive by ELISA and 2/65 (3.1%) by immunoblot [36] although the seroprevalence of HEV in the general Dutch population remains unknown.

The genotyping results of the HEV strains from the PCR positive cases further support the hypothesis of an indigenous HEV reservoir in the Netherlands although we have no evidence for direct zoonotic transmission. There was one 100% match between a case and a surface water sample taken in his close surroundings, but this likely reflects shedding into the environment from this patient. Remarkably, however, this water sample was taken almost three months after onset of illness in the patient. As sewage from the household septic tank was let out in this ditch, the patient might have contaminated the water with HEV. HEV is known to be stable in environmental samples, particularly in watery environments and has previously been identified in sewage samples [37]. As part of a routine control of water from the large river Maas in the Netherlands HEV was isolated from 2/12 samples analyzed in September 2004 and March 2005 [38].

Our epidemiological data do not directly support the "swine as HEV reservoir hypothesis", as only one of the cases had direct contact with pigs in the estimated incubation period and none of the cases reported eating raw pork or raw organs from pork. Dried sausages might, however, serve as a proxy for consumption of pork that is insufficiently processed for HEV elimination and should be an issue for further investigation. In 2005, HEV was isolated from raw pork liver obtained from a butchers shop in the Netherlands [26], showing that people can be exposed through handling or consumption of raw pork livers also in our country. In general, pork consumption was high among the cases, with 78.9% eating pork (well done) more than once a week. Comparable data for the general population is not available, although pork is estimated to represent approximately 25% of the total amount of meat purchased on household level in the Netherlands.

Contact with dog and rodents other than rats might represent a possible riskfactor to be further investigated. The cases who were owners of dogs reported feeding their dogs only canned and dried feed, no raw meat or organs. Due to the long incubation period and time delay from onset of illness to the interview took place (Table 1) we collected data on food consumption using a food preference list and on contact with animals in general and not directly related in time to the HEV infection. This might create a recall bias and specific incidents with possible importance for the transmission of HEV occurring in the incubation period might be left out.

We performed a limited contact investigation around a selection of the cases depending on the household composition and the consent given by the household members. However, none of the 18 serum samples tested showed any indications of a HEV infection, confirming the previously published findings from high endemic areas that HEV has limited person to person transmission [2].

Our study shows that sporadic cases of non-travel related HEV infections continue to occur in the Netherlands and that this most likely is caused by endemic HEV genotype-3 strains. Despite an extended trawling questionnaire combined with environmental investigations and molecular characterization of strains we were not able to point out any common source for the infections in these 19 cases. A direct zoonotic transmission of HEV from pigs to humans could not be confirmed, although food borne transmission via inappropriately processed pork might have occurred. As for most of the cases various possible sources of HEV were reported, an indirect route of transmission is the most likely explanation for these infections. Individual factors as gender, age and pre-existing disease may play an important role in development of disease after infection with HEV genotype 3. How often HEV infections do occur in the Netherlands should be investigated in a seroprevalence study of the general Dutch population.