• Hepatitis;
  • seroepidemiology;
  • seroprevalence;
  • zoonosis;
  • HEV


  1. Top of page
  2. Abstract
  3. Transparency Declaration
  4. References

Clin Microbiol Infect 2012; 18: E119–E121


Hepatitis E virus (HEV) is one of the leading agents of acute hepatitis. This study investigated the prevalence and risk factors of HEV infection in the Tunisian adult general population, either in blood donors (n = 687) or in patients hospitalized for acute hepatitis (n = 202). The mode of transmission differed between these two populations: contact with animals and living in a rural habitat were the main risk factors for being in contact with HEV in asymptomatic blood donors, while HEV was contracted through contaminated water in symptomatic cases. HEV seroprevalence in adult blood donors in Tunisia was relatively low (5.4%) and increased with age.

Tunisia is considered as an endemic country for HEV [1]. However, in contrast to other North African countries such as Egypt, Algeria or Morocco [2–4], HEV outbreaks have never been reported in Tunisia [5]. We evaluated the seroprevalence of HEV infection in the Tunisian adult general population and identified the risk factors associated with seropositive cases.

Two distinct adult groups were studied. Participants were included in the study during September 2007 to June 2008. Records about age, sex, living conditions and personal and family medical history were evaluated by reviewing the patient’s medical chart and by interviewing each patient through a structured questionnaire. Subjects were asked about their contact with animals (farming or trade of sheep, cattle, goats, horses, rabbits and poultry, and hunting of wild animals) and about their sanitary conditions (use of municipal water vs. water from other sanitary uncontrolled sources, and linkage to the municipal drainage). Patients’ sera were tested for the presence of IgG and IgM anti-HEV (MP Diagnostics, Singapore) and IgM anti-HAV (ADVIA Centaur, Siemens, France). Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated using the Pearson χ2 test and a multiple unconditional logistic regression model.

First we studied blood donors (n = 687), mean age 32.6 ± 8.6 years, originating from the governorates of Tunis (northern region), Monastir and Mahdia (eastern region). The main advantage of these areas was the ability to reflect the whole spectrum of sociological behaviours and housing conditions, from rural to urban, and from low-income high-density to economically advanced urban areas. The second group (n = 202) consisted of adults (mean age 39.0 ± 16.0) hospitalized for acute hepatitis at Charles-Nicolle Hospital, Tunis. Characteristics of each group and results of the risk factor analyses are summarized in Tables 1 and 2. Results did not differ by multivariate analysis for hospitalized patients (data not shown).

Table 1.   Blood donors: population characteristics and risk factors associated with IgG anti-HEV positivity
 Descriptive analysisBivariate analysisMultivariate analysis
Risk factor parameter%IgG anti-HEV seropositive (%)OR95% CIOR95% CI
  1. OR, odds ratio; CI, confidence interval.

  2. Bold value indicate statistically significant results.

Age > 3054.38.03.841.66–8.873.051.27–7.34
Origin: Northern region54.65.30.980.5–1.91 
 Urban high density, low income8.–17.654.591.31–16.15
 Urban economically advanced51.81.71   
 >5 persons per household35.55.30.9830.49–1.97  
Education level
 Primary school23.–24.851.120.52–2.39
 Secondary school47.35.53.831.11–13.170.350.10–1.26
 High school29.01.51  
Personal history of hepatitis11.–6.502.250.97–5.19
Family history of hepatitis16.28.81.730.79–3.77  
Contact with animals19.416.74.962.52–9.752.901.20–6.98
Municipal + other water18.213.63.351.68–6.661.570.71–3.48
No connection to sewage39.79.23.381.67–6.850.560.13–2.41
Table 2.   Acute hepatitis: population characteristics and bivariate analysis of the risk factors associated with IgM anti-HEV or IgM anti-HAV positivity
  IgM anti-HEV seropositiveIgM anti-HAV seropositive
Risk factor parameter%%OR95% CI%OR95% CI
  1. OR, odds ratio; CI, confidence interval.

  2. Bold value indicate statistically significant results.

Age ≤ 3039.60.00[0–Na]56.316.147.18–36.28
Rural habitat54.03.71.730.31–9.6736.73.271.64–6.51
Personal history of hepatitis23.80.00Na–Inf18.80.560.25–1.25
Family history of hepatitis28.–29.5519.00.550.26–1.16
Contact with animals33.21.50.390.04–3.4113.40.310.14–0.68
Municipal + other water31.77.811.611.33–101.5540.62.691.41–5.15

Anti-HEV IgG were detected in 5.4% of the blood donors. Seroprevalence increased significantly with age, from 2.2% of patients younger than 30 years old to 8.0% of patients older than 30. In hospitalized patients, no IgM anti-HEV was detected in patients younger than 30 years old. Seroprevalence was 26.7% for IgM anti-HAV, with a mean age of 26.3 ± 6.3 and 3.0% for IgM anti-HEV with a mean age of 41.0 ± 5.9. The age difference was statistically significant (p <10−5, Student’s t-test).

In bivariate analysis, the factors associated with anti-HEV positivity in blood donors were age, living in rural or in low-income high-density urban areas, a low level of education, personal history of hepatitis, poor sanitary conditions and contact with animals. No difference in the prevalence of IgG anti-HEV was observed with respect to sex, region of origin, number of persons per household or familial history of hepatitis. In multiple logistic regression analysis, the remaining independent risk factors were age, habitat and contact with animals but not education level or sanitary conditions. In contrast, the only factor associated with symptomatic hepatitis E cases was the consumption of uncontrolled water.

Compared with the two recent age-related studies performed in central Tunisia [6,7], we found a lower prevalence of hepatitis E. This discrepancy might be attributable to the difference in the study sample sizes, the region of origin of the populations studied or the ELISA kits used (Globe Diagnostics SRL, Milano, Italy, for the studies by Hannachi et al.). Indeed the MP Diagnostics’ kit might underestimate the seroprevalence linked to genotype 3 strains [8]. However, the Globe Diagnostics kit was not accurately evaluated on this genotype.

A higher risk of infection was found in high-density urban areas compared with standard-density urban areas. However, neither a large family household nor a familial history of hepatitis infections constitute a risk factor, thus indicating a low risk of person-to-person transmission, as suggested in other studies [9], but in contrast to what is suggested in Hannachi’s study [6]. Population density per se does not seem to constitute a risk factor. Indeed, in Tunisia, similarly good sanitary conditions (tap water and link to drainage network) are found in standard and in densely populated urban areas. The key differences between inhabitants of densely populated urban areas and other urban areas are a generally lower economic status, with a lower level of education and, for the majority of them, a previous rural origin. Therefore, in this population, HEV might have been contracted previously while living in a rural area.

The low seroprevalence and the age distribution of HEV in Tunisia do not reflect the classical epidemiological profile described for an endemic country [10]. These findings are in agreement with the absence of waterborne epidemics described in Tunisia, in contrast to other North African countries. Although more than 90% of the Tunisian population has already been in contact with HAV at the age of 20 [5], more than a quarter of acute hepatitis infections were due to HAV and only 3% to HEV. HAV and other enteroviruses but no HEV could be detected in treated wastewater (A. M. Aouni Mahjoub, unpublished data and [11]), which is likely to explain these observations. A relatively late acquisition of HEV compared with HAV suggests a role of what people are eating rather than contaminated water supplies only, as is observed in industrialized countries [1]. Thus, hepatitis E in Tunisia seemed to be mainly a subclinical, asymptomatic infection caused by contact with animals, as in developed countries [12]. However, symptomatic infections appeared to be mainly caused by contaminated water. Remarkably, waterborne HEVs (genotypes 1 and 2) are known to be more virulent than strains transmitted through animals (genotypes 3 and 4) [1,13]. Clinical presentation and modes of transmission suggest that genotype 1 might not be the only genotype circulating in Tunisia as previously thought [1,13]. Our study highlights the need for genotype characterization of the Tunisian strains and determination of the animal reservoir [14]. Indeed, domestic and wild pigs can be excluded here: pork products are not consumed, live pigs are not raised in the country and the wild boars present in some parts of the territory are not endemic in the two regions studied. Goats in particular are raised in the proximity of Tunisian people and might be a source of contamination.

Transparency Declaration

  1. Top of page
  2. Abstract
  3. Transparency Declaration
  4. References

The authors declare no conflicting interests.


  1. Top of page
  2. Abstract
  3. Transparency Declaration
  4. References