Evidence for Hepatitis A Virus Endemic Circulation in Israel Despite Universal Toddler Vaccination Since 1999 and Low Clinical Incidence in All Age Groups

Abstract

Background.

Hepatitis A virus (HAV) molecular epidemiology in Israel was studied 13–14 years after introduction of universal toddler vaccination (UTV), which reduced hepatitis A incidence from 50.4 to <1.0 per 100 000.

Methods.

An outbreak in Tel-Aviv with 75 cases in 2012–2013 was investigated. Real-time reverse-transcription polymerase chain reaction and sequencing of the VP1-2A region (1100 bp) was done on (1) acute hepatitis A patients' sera (n= 12/75 in Tel-Aviv and 31 patients hospitalized in 3 other major cities in 2011–2013) and (2) sewage samples (n = 27 from metropolitan Tel-Aviv, 14 from the other 3 cities, and 6 from Gaza).

Results.

The outbreak patients’ mean age was 33.2 years, 4 of 75 (5.3%) had been vaccinated, and 58 of 75 (77.3%) were hospitalized. Among sewage samples 16 of 27 (59.2%) from Tel-Aviv; 4 of 14 (28.6%) collected throughout Israel; and 6of 6 (100%) from Gaza contained HAV. Genotype IB predominated (n = 52/53 sequenced samples) and identical strains were demonstrated in the Israeli and Palestinian populations.

Conclusions.

Despite the UTV success, HAV circulation in the Israeli population is apparently facilitated by its close contacts with the endemic Palestinian population. Reassessment of vaccination policy is recommended.

Hepatitis A virus (HAV) infection is an acute, self-limited illness, associated with fever, malaise, nausea, anorexia, and jaundice. The primary route of transmission is fecal–oral, which is associated with low sanitation standards. Regardless of the clinical manifestations of infection, HAV can be found in patients’ sera and is shed in the feces of infected persons. Hepatitis A is responsible for a substantial number of viral hepatitis cases worldwide, and outbreaks still occur with increased number of symptomatic cases and hospitalizations [1].

In countries in process of improving socioeconomic and sanitary conditions and especially in regions where childhood vaccination was introduced, HAV transmission from young children to susceptible adults declines, the incidence of HAV infections in all age groups decreases, and the age of infection shifts from early childhood toward adulthood [2–4]. Israel is surrounded by highly endemic regions for HAV, where vaccination programs do not exist and surveillance is lacking [5, 6]. Following introduction of a 2-dose universal toddler vaccination (UTV) in 1999, acute hepatitis A incidence in Israel declined by approximately 90% within 2–3 years and >50-fold by 2012: from 50.4 to <1.0 cases per 100000 [7, 8]. Before initiation of UTV, the highest incidence of HAV infection was observed in the 1–4 years and 5–9 years age groups (>250 cases per 100000). After 2002, the overall incidence of acute HAV infection dropped to <2.3 cases per 100000 in all age groups, including in nonvaccinated adults, regardless of ethnicity, although the incidence rate was somewhat higher in the non-Jewish population [7, 8].

In countries with a low or very low incidence of hepatitis A, with high hygienic conditions and/or routine immunization in childhood, hepatitis A cases are generally imported from endemic regions. In such areas, outbreaks may occur as a result of introduction into high-risk groups, such as intravenous drug users (IVDU) or men who have sex with men (MSM), or by consumption of imported food products followed by further spread into the general population [9, 10]. Yet the origin of many cases remains obscure [11]. One way of improving the ability to identify the possible source of hepatitis A cases is by sequencing and molecular analysis of HAV shed by patients [12]. However, examining only clinical cases may not represent the full range of circulating HAV because of the high rate of unrecognized or unreported cases [13].

Environmental surveillance and molecular analysis of sewage samples may overcome this limitation. This approach has become a useful tool for assessment of HAV infection and circulation [14, 15]. This information is of particular importance because nonimmune adolescents and adults are at greater risk for developing clinical symptoms that require hospitalization [4, 11, 16]. Hepatitis A virus has a single antigenic serotype, but it was classified into 7 genotypes with genetic variation of >15% between them: genotypes I, II, III, and VII are of human origin and genotypes IV–VI are of simian origin [17–19]. Each of the most common human genotypes (I and III) encompasses 2 subgenotypes, A and B, with nucleotide variations of up to 7.5% within each subgenotype.

Our study was initiated in an attempt to discover the source of the virus that caused an outbreak in the Tel-Aviv metropolitan region between March 2012 and March 2013. The study was based on environmental and clinical samples from the Tel-Aviv metropolitan region and other cities in an attempt to elucidate possible chains of transmission through molecular surveillance.

MATERIALS AND METHODS

Outbreak Investigation

Viral hepatitis A is an individually reportable disease. On report of the new cases in Tel-Aviv Health District, an epidemiological investigation was conducted. Local laboratories were contacted to identify unnotified cases. District Health offices in other regions were contacted to identify additional cases and to ascertain the scope of the problem at a national level. A case was defined as a report of an individual who had been in the Tel-Aviv Health District (covering the Tel-Aviv metropolitan region) with a clinical presentation of abrupt onset hepatitis (fatigue, nausea, anorexia, abnormal liver function tests) and laboratory confirmed anti-HAV (immunoglobulin M) serology. In 1 case (a young child), serology was inconclusive, and the case was confirmed by molecular testing.

Clinical Samples for Molecular Analysis

Serum samples from patients in Tel-Aviv collected during the late stages of the outbreak were transferred to the Central Virology Laboratory (CVL) at Sheba Medical Center for molecular analysis and typing. Additional serum samples from acute HAV cases in other regions (including samples from Gaza residents hospitalized in Israel), which were kept at −80°C, were transferred to the CVL from laboratories in the Hadassah Hospital Liver Unit in Jerusalem as well as hospitals in Haifa and Beer-Sheva for molecular analysis.

The use of the clinical samples for molecular analysis was approved by the institutional review board of Chaim Sheba Medical Center (IRB ref no. 1851-14-SMC).

Sewage Sampling

Fifteen sites along the sewage lines covering all regions of Tel-Aviv and satellite cities, as well as the central sewage treatment facility covering the entire Tel-Aviv metropolitan region were sampled between October 2012 and May 2013. Portable Sigma SD900 automatic samplers calibrated to collect either 24 × 400-mL or 48 × 200-mL samples per day for a total volume of approximately 10 liters were used. One liter of well-mixed sewage was transferred to the CVL for further analysis. Sewage was treated and concentrated to 20–30 mL as described previously [20]. Additional sewage samples were collected in sewage treatment facilities of 3 other major cities located throughout Israel and from 3 sites in Gaza. These were concentrated as described previously [20]. These samples represent >50% of the Israeli population and were collected as part of poliovirus routine surveillance.

Total Nucleic Acid Extraction From Sewage Samples

Total nucleic acid (NA) was extracted as described previously [20] using the NucliSENS easyMAG system according to the manufacturer’s instructions. Briefly, external lysis was performed on 1 mL of concentrated sewage to inactivate the virus. This was followed by NA extraction using the easyMAG extractor. RNA from 0.2–1-mL human serum samples was also extracted using the NucliSENS easyMAG system. Extracted NA were eluted in 55-μL elution buffer and stored at −80°C pending analysis.

Real-Time Reverse-Transcription Polymerase Chain Reaction

An HAV real-time reverse-transcription polymerase chain reaction (qRT-PCR) kit was used to detect viral RNA in serum or sewage samples according to the kit instructions.

Sequencing

Samples that were positive for HAV at cycle threshold (ct) <30 were amplified by RT-PCR using primers flanking the VP1-2A region (approximately 1100 nucleotides) as published by Costa-Mattioli et al [17] and sequenced.

Phylogenetic Analysis

A phylogenetic tree based on 837 nucleotides, present in all sequenced samples, from the VP1/2A region (nucleotides 2424–3260 of reference strain M14707 IB) was generated comparing all obtained sequences with the reference sequences EF190998 IB, HQ401265 IB, M14707 IB, M20273 IB, AJ43723 IA, AJ437232 IA, AJ437234 IA, M34084 III, M66695 III, and M59286 IV from GenBank using ClustalX and njplot, with bootstrap value of 1000.

All HAV sequences obtained were submitted to GenBank, and their accession numbers are KU230462–KU230511.

RESULTS

Tel-Aviv Outbreak Description

The first 3 reported cases occurred on weeks 11, 13, and 24 of 2012 (dates of onset of symptoms). This was followed by 2 waves of cases: one between week 27 and week 39 of 2012 and the other between week 47 of 2012 and week 10 of 2013, with 2 seemingly sporadic cases in between. The epidemic curve is shown in Figure 1.

Altogether 75 acute cases were confirmed in the Tel-Aviv metropolitan region, of which 58 (77.3%) were hospitalized. The geographical, age, and sex distribution of the cases are shown in Table 1. The mean age was 33.2 years (range = 4 months–65 years), 31.7 years in males (range = 4 months–65 years) and 37.2 years in females (range = 8–64 years). The predominance of male subjects was in all age categories except the age 55–64 age. Four of the cases (5.3%) had been vaccinated (self-reporting). An epidemiological investigation conducted by the Tel-Aviv District Health Office did not reveal any common environmental source (food or water), and the investigation of the clinical cases found no immediately apparent connection between any of the cases other than 1 instance in which a parent and child were both infected.

Hepatitis A Virus Detection in Sewage Samples

Sixteen of 27 samples (59.2%) from 11 of 15 sampling sites (73.3%) covering the Tel-Aviv metropolitan region were found positive for HAV by qRT-PCR (Table 2). Hepatitis A virus was found in sewage lines covering much larger areas than the areas where clinical cases were reported, particularly satellite towns north and east of Tel-Aviv, indicating that the virus was circulating in the entire metropolitan region.

Environmental samples from other major cities were used in search for the virus origin and assessment of its possible country-wide circulation (Figure 2). Six sewage samples from Jerusalem (center of Israel), 6 sewage samples from Beer-Sheva and Rahat (south of Israel), 1 sewage sample from Haifa (north of Israel), and 6 sewage samples from Gaza (Palestine) collected in September, October, and November 2012 were tested for HAV by qRT-PCR. Positive samples were found in all regions of Israel and in Gaza, as detailed in Table 2: 4 of 14 (28.6%) in Israeli cities outside of the Tel-Aviv metropolitan region, and 6 of 6 (100%) in Gaza. These findings indicate that at the time of the outbreak HAV circulated country-wide and not only in the Tel-Aviv metropolitan region.

Hepatitis A Virus Detection in Serum Samples

Forty-three serum samples that were positive by qRT-PCR were suitable for sequencing. Twelve were from the Tel-Aviv metropolitan region collected during the last stages of the outbreak between week 49 of 2012 and week 8 of 2013 (Figure 1). Seventeen were hospitalized in the Jerusalem region in 2011–2014: 11 Arabs from Palestinian communities in East Jerusalem and adjacent towns, 1 Arab from the West Bank in Palestine, and 5 Jews. Four serum samples were from the Haifa region, 4 were from the Beer-Sheva region and southern Israel, and 6 were from residents of Gaza hospitalized in Israel, all from 2013.

Typing and Phylogenetic Analysis of Hepatitis A Virus From Sewage and Serum Samples

The phylogenetic analysis performed showed that all of the sewage and 42 of the 43 serum samples (except 1 from a tourist) were of genotype IB. Figure 3 shows global IB reference strains from GenBank and representative strains from all regions of Israel. Figure 4 shows the detailed phylogenetic analysis of all sequenced HAV strains included in this study and of 2 of the most closely related IB reference strains.

The Israeli isolates were most closely related to IB strains from Hungary (EF190998; 1.32%–3.48% difference over 833 compared bases) and Spain (HQ401265; 1.32%–2.88% difference over 833 compared bases). The divergence among all Israeli strains was somewhat lower (1.07%–2.38% over 837 bases), but this difference is not sufficient to define a specific Israeli lineage. Figure 4 shows the phylogenetic tree with geographical classification of samples. It contains 3 main branches: Branch I includes a cluster of 16 identical sequences collected during the outbreak period between August 2012 and January 2013, which appear to represent the dominant Tel-Aviv outbreak strain. Eleven patients and 2 sewage samples were from Tel-Aviv, whereas 3 patients were from Haifa (n = 1) and Jerusalem (n = 2) but had permanent addresses in the Tel-Aviv region. Two closely related samples were sewage from Gaza collected in June 2013. Another cluster contains 1 Arab patient from East Jerusalem, 1 Palestinian patient from the West Bank, 1 Jewish patient from Jerusalem whose permanent address is in the Tel-Aviv region, and 1 sewage sample from Tel-Aviv collected in January 2013. The overall variability in this branch is 0%–2.5%.

Branch II includes samples from 6 Arab patients from East Jerusalem and adjacent communities, 1 patient from Beer-Sheva, and 1 patient from Haifa, collected between 2011 and 2013. The overall variability of this branch is 0.35%–1.79%.

Branch III contains representatives from all regions of Israel, including 2 samples from Tel-Aviv collected in Nov./2012. Ten additional samples were from Gaza (4 sewage and 6 patients). A cluster with 5 identical sequences includes a patient from Haifa, 2 Arabs from East Jerusalem, 1 Jew from a settlement in the West Bank, and 1 Jew from a city in the center of Israel; all were hospitalized in Jerusalem. The overall variability in this branch is 0.35%–1.91%.

DISCUSSION

Interruption of endemic transmission and prevention of large outbreaks of hepatitis A could have been anticipated as a result of the success of the UTV program implemented in Israel since 1999 [7, 16, 21, 22]. Our molecular epidemiology study of HAV in Israel between 2011 and 2013 including the outbreak in Tel-Aviv shows that HAV endemic circulation occurs in Israel and describes its molecular epidemiology.

Our phylogenetic analysis, which included environmental and clinical samples from all 4 major cities in Israel (Figure 4), showed that the same HAV IB strains are circulating endemically in all regions and populations of Israel and in the Palestinian populations, including Gaza.

Hepatitis A virus is highly endemic in Palestine but is also more common among Arabs living in Israel than in Israeli Jews, particularly in East Jerusalem. In 2003–2005, 165 of 420 cases reported in Israel were from East Jerusalem, apparently due to lower vaccination coverage of the Arab versus the Jewish population in Jerusalem [23]. In our study, all of the cases in Tel-Aviv occurred in the Jewish population, in contrast with Jerusalem where most of the patients were Arabs from East Jerusalem and adjacent communities. This observation is of particular importance in view of the results of a survey conducted in Jerusalem between 1999 and 2004, during introduction of the UTV in Israel. In this survey, 610 serum samples, positive for HAV RNA by PCR were subsequently genotyped. Eighty-one percent of 377 serum samples from the Jewish population belonged to genotype IA and 96% of 233 serum samples from Arab patients from East Jerusalem belonged to genotype IB (D. Shouval and N. Daudi, personal communication). Thus transmission of the genotype most common among Jews (IA) was apparently interrupted between 2004 and 2012, whereas the genotype that was common among Arabs (IB) continues to circulate among Arabs and Jews. The evolution of this common virus reservoir can be explained by the continuous multiple contacts between the populations: Palestinians living in Palestine, Arab residents of Israel (East Jerusalem), Jews living in the West Bank, and Jews living elsewhere in Israel. Such contacts occur in all regions of Israel in daily life, including in hospitals, or through free trading of food, including fruits and vegetable products, between Palestine and Israel.

The epidemiological investigation and the phylogenetic analysis failed to pinpoint the exact source of the outbreak and modes of transmission. The epidemic curve of the outbreak may point at a silent transmission through asymptomatic excretors. Such transmission was demonstrated for Poliovirus in Israel by environmental surveillance [20, 24]. The mean age of patients in Tel-Aviv (33.2 years) is compatible with a recent HAV seroprevalence study conducted by the Israel Center for Disease Control and the Central Virology Laboratory (http://www.health.gov.il/PublicationsFiles/ICDC_357.pdf) that showed a positivity rate of 80%–90% only in children aged 2.5–13 years who were vaccinated as part of the UTV program and in those aged ≥50 years due to natural exposure in childhood. However, seroprevalence in other age groups, mainly infants before the first vaccine dose at age 18 months and in those aged 14–49 years (who were neither vaccinated nor exposed to high endemic circulation) is ≤57%. The fact that 4 patients reported that they had been vaccinated is puzzling. These patients were born before the implementation of the UTV program in 1999. However, their reports could not be verified. The presence of a large susceptible population, which allows sporadic infections and outbreaks, prompts reassessment of vaccination policies. The current Ministry of Health guidelines recommend immunization of high-risk groups, including drug users, but no specific programs are in place for IVDUs or the homeless. Thus it is advised to expand the immunization policies to reach out and vaccinate high-risk groups such as IVDUs to further reduce the chance for HAV circulation and outbreaks as well as severely ill patients who need hospitalization.

Our study has a number of limitations. Regarding the outbreak in Tel-Aviv: the only available data on severity of the cases is the high rate of hospitalizations (77.3%). The data on vaccination were obtained by self-reporting alone and are incomplete and might be inaccurate. The clinical and environmental samples collected and analyzed did not cover the entire outbreak period, and phylogenetic analysis could be achieved only for a subset of the samples. Because the phylogenetic analysis showed that >1 strain has circulated in Tel-Aviv during the outbreak period (Branch III), it is possible that some of the cases (especially at the earlier stages) were caused by different strains.

Regarding the country-wide epidemiology, the study opened a window on a short period of time during the outbreak. Thus, we can only assume that this period represents the situation before or after the outbreak. A prolonged sewage survey and analysis of more clinical samples from all regions of Israel would provide more data and help to inform public health policies.

Notes

Acknowledgments. The authors wish to thank all contributors to the collection of the epidemiological information at the District Health Office in Tel-Aviv, the workers of the Shaf-Dan for help in collection sewage samples in Tel-Aviv, laboratory technicians who participated in the generation of the laboratory data in all four participating laboratories and the department of epidemiology of the Israeli ministry of health for creating the map in Figure 2.

Financial support. The study was financed by the Central Virology Laboratory of the Israel Ministry of Health annual budget resources.

Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References

Author notes