The etiologic agent of hepatitis type E is a single-strand RNA virus temporally classified as Hepevirus, of the Caliciviridae family, non-A, non-B, with a diameter of 27–38 nm. Hepatitis type E is also known as enterically transmitted non-A non-B hepatitis. Hepatitis E virus (HEV) infection causes an acute, self-limited hepatitis associated with high mortality in pregnant women. The infection usually occurs as epidemic outbreaks, is transmitted by the fecal-oral route, and the virus is usually disseminated in contaminated water. Outbreaks affecting hundreds of patients have occurred in developing countries such as India, Indonesia, and Africa. Sporadic clinical cases are also frequent in endemic zones. Therefore, HEV infection is an important cause of morbidity and mortality in several countries.

Most studies describe the prevalence of the infection in patients with hepatitis or in selected groups of subjects. Community-based surveys are scarce and information on HEV infection in populations is needed. In Mexico, two outbreaks affecting patients were reported in 1987, and three seroepidemiologic studies in small groups of subjects reported prevalences of 0.0, 3.9, and 6.3%14. Lisker MM, Dawson G, Sixtos S. Endemicidad de la hepatitis E. Rev Gastroenterol Mex 1994 (Suppl 2);59:143 [Abstract 155]... The aim of this work was to study the prevalence of anti-HEV antibodies in a sample population representative of all regions and all socioeconomic levels in Mexico. The infection usually occurs in young adults; however, as this is a fecal-oral transmitted infection, we considered it important to determine the prevalence during infancy and childhood.

During 1987–1988, the Mexican Ministry of Health performed a national serologic survey with the purpose of creating a National Serum Bank. The collection of serum samples was carried out according to a master sampling-frame, including almost 500,000 households representing the entire country, and was created with information from 1980 general population census data and included carto- and demographic information provided by the National Institute for Statistics, Geography and Information. The survey included all 32 states of Mexico, and was designed to include individuals from 1 to 90 years of age. All socioeconomic levels and geographic zones of the country were represented.

During the survey, 32,200 households were visited, and a blood sample was drawn at each with the consent of the individuals or their parents or guardians. More than 70,000 serum samples were collected, with an overall response rate of 78.4% of the homes surveyed. Serum samples were aliquoted and stored at ?20°C in the National Serum Bank.

Each individual completed a questionnaire detailing personal, socioeconomic, and other demographic data.

To define the socioeconomic level, an index was constructed which included the number of persons per room, the type of floor material used in the house, the availability of municipal water, waste disposal, and the educational level of the head of the household. An urban population was defined as an area with more than 2,500 inhabitants, while a rural population had fewer than 2,500 inhabitants. A metropolitan population included the four largest cities of the country (Mexico City, Guadalajara, Monterrey, and Mérida).

For the purposes of this study, the country was divided into eight regions reflecting the level of development. The level of development was based on socioeconomic, demographic and public health indexes according to the criteria of Kuns et al. The indexes included in this regionalization were as follows: educational level, household characteristics, rate of infant mortality, rate of maternal mortality, rate of mortality due to transmissible and non-transmissible diseases, number of physicians in the population, proportion of the population that is economically active, and urban vs. rural communities, among others.

A subsample from this national survey representing individuals 1 to 29 years of age from all eight regions of the country was selected. The sample size for each region was calculated by means of the Z test, estimating an HEV prevalence of 10%, with a 95% confidence interval (CI) and a relative error of 0.05. The total number of serum samples for study was 3,549.

A multistage stratified design was employed to select serum samples. In stage I, the sampling unit was the region stratified by the type of population (urban/rural). In stage II, the sample units were the household characteristics. In stage III, sample units were the characteristics of the individual (sex, age). In all stages, random selection was proportional to the size of the unit.

Specific anti-HEV antibodies were studied using a commercial ELISA test with a sensitivity of 98.2% and a specificity of 99.5% (HEV EIA, rDNA, Abbott, Wiesbaden-Delkenheim Germany). The antigen is a mixture of two recombinant antigens derived from structural regions ORF-2 and ORF-3 of the Burmese HEV strain. Cut-off was defined with positive and negative control sera that were included in each assay, according to manufacturer instructions. Samples were considered positive if the optical density (OD) value was above the cut-off value and all positive samples were confirmed with a second assay, using the same ELISA test.

Seroprevalence to infection and its association with socioeconomic and demographic characteristics were described as percentages and analyzed using a bivariate model with OR estimates and a CI of 95%. Variables showing a significance of <0.05 were then studied in a multivariate logistic regression model and OR estimates with a CI of 95%.

Antibodies against HEV were detected in 374 (10.5%) of the 3,549 subjects studied. Although no clear association was found between seroprevalence and level of development, a trend was observed, with some regions (regions 4, 7, and showing a higher risk of infection as compared with the region of the highest level of development (Region 1).

The influence of socioeconomic and demographic variables in the frequency of HEV infection with a bivariate analysis. In this model, age, type of community, socioeconomic level, educational level and crowding were found as risk factors for infection. Accordingly, these variables were included in a multivariate model. Seroprevalence increased significantly with age, from 1.1% in children under 5 years of age to 14.2% in individuals 25 to 29 years of age. In this last age group, the risk for being infected was 15.5 times higher than that in children younger than 5 years of age. Individuals living in urban and rural areas had a higher risk of infection as compared with those living in metropolitan areas (which include the four largest cities of the country). Illiterate individuals were at higher risk for infection than those with a basic high school degree. It should be noted that in this model the variables of socioeconomic level and crowding were no longer found to be risk factors for infection.