Transcript
378 Rev Panam Salud Publica/Pan Am J Public Health 6(6), 1999 Hepatitis B seroprevalence in Latin America Thêmis R.Silveira, 1 José Carlos da Fonseca, 2 Luis Rivera, 3 Oscar H.Fay, 4 Roberto Tapia, 5 José I.Santos, 5 Eduardo Urdeneta, 6 and Sue Ann Costa Clemens 7 The seroprevalence of hepatitis B was investigated in over 12 000 subjects in six countries of Latin America: Argentina, Brazil, Chile, the Dominican Republic, Mexico, and Venezuela.Each study population was stratified according to age, gender, and socioeconomic status. Anti-bodies against hepatitis B core antigen (anti-HBc) were measured in order to determine hepa-titis B infection. The highest overall seroprevalence was found in the Dominican Republic(21.4%), followed by Brazil (7.9%), Venezuela (3.2%), Argentina (2.1%), Mexico (1.4%), andChile (0.6%). In all the countries an increase in seroprevalence was found among persons 16 years old and older, suggesting sexual transmission as the major route of infection. In addi-tion, comparatively high seroprevalence levels were seen at an early age in the Dominican Re-public and Brazil, implicating a vertical route of transmission. ABSTRACT Parts of Latin America have previ-ously been described as high-endemic-ity areas for hepatitis B. However,there is a wide variation in seropreva-lence within Latin America, with thehighest rates around the Amazon basin and the lowest in the temperatesouthern portion of South America (1,2). Furthermore, Latin American coun-tries share many traits with other de-veloping countries, such as the migra-tion from rural communities to citiesthat creates urban areas with low in-comes and social deprivation. Undersuch conditions, health problems, es-pecially communicable diseases, are amajor concern and impose a large eco-nomic burden. However, in urban en-vironments, health care programs havethe potential to quickly reach largenumbers of people. All major authorities agree that vac-cination is the most effective means toreduce the health care burden of hep-atitis B. In 1992 the Expanded Pro-gramme on Immunization (EPI) of theWorld Health Organization (WHO)set targets for the incorporation of hepatitis B virus (HBV) vaccinationinto national programs by 1995 forhigh-endemic countries, and for allcountries by 1997 (3). Having an un-derstanding of the epidemiology of the virus is essential in order to opti-mize the use of vaccination as a meansof controlling the infection and dis-ease. In Latin America, analysis of blood banks has previously been themain source of epidemiological data(2). However, the data were not strati-fied by age or social class. In addition, blood donors are almost exclusivelyhealthy adults in urban areas wherethere are sufficient facilities to performroutine serological screening. In orderto determine the seroprevalence of hepatitis B infection in six countries of Latin America and the carrier rate inthree of them, we carried out a cross-sectional multicenter study. METHODOLOGY Study design This was a multicenter cross-sec-tional study with subjects in six coun-tries: Argentina (cities of Buenos Aires 1 Hospital das Clínicas de Porto Alegre, Porto Ale-gre, Rio Grande do Sul, Brazil. 2 Instituto de Medicina Tropical de Manaus, Ma-naus, Amazonas, Brazil. 3 Maternidad Nuestra Señora de Altagracia, Depar-tamento de Perinatología, Santo Domingo, Domi-nican Republic. 4 Centro de Tecnología en Salud Pública, Facultadde Ciencias Bioquímicas, Universidad Nacional deRosario, Rosario, Santa Fe, Argentina. 5 Consejo Nacional de Vacunación, Mexico, D.F.,Mexico. 6 Grupo Médico Otorrinolaringológico, Caracas,Venezuela. 7 SmithKline Beecham Biologicals, Rixensart, Bel-gium. Send correspondence to:Dr. Sue AnneCosta Clemens, SmithKline Beecham Biologicals,Rue de l’Institut, 89, 1330 Rixensart, Belgium. Tele-phone: 32 2 656 8781; fax: 32 2 656 8133; e-mail:
[email protected] Rev Panam Salud Publica/Pan Am J Public Health 6(6), 1999 379 and Córdoba), Brazil (Rio de Janeiroand Porto Alegre, in the southeast;Manaus, in the Amazon basin; and For-taleza, in the northeast), Chile (northSantiago de Chile), the Dominican Re-public (three centers in Santo Do-mingo), Mexico (in cities in the north,center, and south regions of the coun-try), and Venezuela (Caracas). Thestudy was conducted between June1996 and November 1997. Subjects in-cluded both men and women betweenthe ages of 1 and 40 years old. The samestudy population was used for the gen-der, age, and socioeconomic analysesthat were done. Written informed con-sent was given by the subjects or thesubjects’ parents or guardians. Study populations and selection To prevent clusters, in all the studysites there was a limit of one sub- ject per household or family. The re-cruitment procedures for the studypopulations varied from country tocountry. The following paragraphsdescribe the procedures in each of thestudy sites. In Mexico the study used two-stagecluster sampling, with primary sampleunits made up of states, plus the Fed-eral District, randomly selected in eachof the three regions. The secondarysample units were urban areas be-longing to each of the selected states,plus Mexico City. The sample size wasdistributed proportionately through-out the entire population of each regionand also the total population between 1and 40 years old in each of the regions.In Chile subjects were recruitedamong randomly selected householdsfrom the north of Santiago de Chile. Thecharacteristics of the sampled house-holds corresponded very well withthose of the last census of the metro-politan population of that city. The Venezuelan study populationwas recruited from eight differentplaces in Caracas, including pre-schoolnurseries, colleges, public schools, andcompanies randomly selected and will-ing to participate. In Brazil, one part of the study pop-ulation was recruited in various out-patient clinics attached to public andprivate hospitals mainly located inpoor and middle-class areas. Anotherproportion, mainly the higher socio-economic group, was recruited atschools in wealthy neighborhoods. The Argentine study subjects wererecruited among people attending am- bulatory primary health care centersattached to three public hospitals.In the Dominican Republic thestudy population was recruited fromthree health care centers, where the at-tendees were from a low socioeco-nomic background. (The higher pro-portion of females in this studypopulation as compared with theother countries was due to the fact thatone of the recruitment centers was amaternity hospital.) Assessment of criterion Subjects completed a questionnaireand provided such demographic dataas social class, type of community, pre-sent health status, and relevant med-ical background. A history of priorHBV vaccination was an exclusion cri-teria.Socioeconomic groups were de-fined according to local scales, whichwere based on income, tap water ser-vice, sewage service, refrigerator own-ership, electric power access, and thenumber of people per room in thehome. In Argentina and Mexico amodified Bronfman scale (4) was used,and in Brazil a government scale (5)was applied. In Venezuela and Chileall subjects were from the middle so-cioeconomic group while in the Do-minican Republic people were only re-cruited from the lower socioeconomicgroup. All data requested in the studyprotocol were documented on individ-ual case report forms. Assessment of serology A minimum of 5 mL of blood wastaken from each of the subjects, and thesera were stored at 20 °C. Immuno-globulin G (IgG) antibodies againsthepatitis B core antigen (anti-HBc) wereassessed with the Corzyme enzyme-linked immunosorbent assay (ELISA)test (Abbott Laboratories, Abbott Park,Illinois, United States of America) todetermine seroprevalence. The quali-tative determination of hepatitis B sur-face antigen (HBsAg) was carried outusing the Auszyme commercial en-zyme immunoassay (Abbott Laborato-ries, Abbott Park, Illinois, UnitedStates) and was used to determine thecarrier population. For both assays,values greater than 1.0 mIU/mL wereconsidered to be seropositive. Serol-ogy was conducted at the National In-stitute of Epidemiological Diagnosisand Reference, Mexico City, Mexico,for samples from Mexico and Vene-zuela; at the J. J. Aguirre Hospital, Uni-versity of Chile, Santiago, for samplesfrom Chile; at the Public Health Tech-nology Center, National University of Rosario, Rosario, Santa Fe, Argen-tina, for samples from Argentina andthe Dominican Republic; and at theCarlos Lieberenz Laboratory of Clini-cal Analyses, Rio de Janeiro, for sam-ples from Brazil. Statistical analysis The study population was stratifiedaccording to country, gender, age, andsocioeconomic status. Data were en-tered into the Dbase IV computer pro-gram (Boland International Inc., ScottsValley, California, United States), andthe descriptive statistical analysis wasperformed using two other computerprograms, SPSS (SPSS Inc., ChicagoIllinois, United States) and Epi Info6.04 (U.S. Centers for Disease Controland Prevention, Atlanta, Georgia,United States). Odds ratios and 95%confidence intervals (95% CIs) werealso calculated. RESULTS Of the 12 085 subjects who were en-rolled in the study, 328 of them wereeliminated due to an inability to ana-lyze their blood sample and/or miss-ing data on their case report form. Thestudy attrition by country is shown inTable 1. 380 Silveira et al. • Hepatitis B seroprevalence in Latin America Gender and country distribution Table 2 shows the seroprevalence of anti-HBc in the study population bycountry and gender. The highest over-all seroprevalence was observed in theDominican Republic (21.4%), followed by Brazil (7.9%), Venezuela (3.2%),and Argentina (2.1%). The lowest lev-els were found in Mexico (1.4%) andChile (0.6%). In the different regions of Brazil a wide range of seroprevalenceswas found, with a high rate in Manaus(21%), followed by Porto Alegre (7.5%),Rio de Janeiro (5.5%), and Fortaleza(1.2%). The only significant differencein seroprevalence between males andfemales was in Brazil, with a higherseroprevalence for males (odds ratio =1.32, 95% CI: 1.01–1.65, P = 0.02), andin the Dominican Republic, with ahigher rate for females (OR = 2.19, 95%CI: 1.05–3.07, P = 0.01). The hepatitis B carrier populationwas assessed by the detection of HBsAg in sera in Argentina, the Do-minican Republic, and Mexico. Aswith the seroprevalence data, thehighest carrier rate was found in theDominican Republic (1.9%). Compara-tively low rates were found in Argen-tina (0.2%) and Mexico (0.1%). Age distribution The seroprevalence of hepatitis Bamong different age groups in the var-ious countries is detailed in Table 3and shown graphically in Figure 1.The countries exhibited noticeable in-creases in seroprevalence among per-sons aged 16 and older. At one year of age, children foundanti-HBc-positive were 0/10 in Chile,0/7 in the Dominican Republic, 0/41in Mexico, and 0/30 in Venezuela, but4/129 (3.1%) in Brazil and 5/71 (7.0%)in Argentina. Socioeconomic factors Table 4 shows the seroprevalencedata analyzed by socioeconomic level,which was available for Argentina,Brazil, and Mexico. In Argentina andMexico there was an even distributionover the socioeconomic groups. How-ever, there was a clear association of increasing seropositivity with decreas-ing socioeconomic status in Brazil,where there was a significant differ-ence between the low socioeconomicgroup (10.7% seropositive for anti-HBcantibodies, 95% CI: 9.2–12.2) and thehigh/medium one (5.4% seropositive,95% CI: 4.3–6.3) ( P < 0.001). DISCUSSION The present study was performed toassess the current status of hepatitis Bseroprevalence in Latin America, withrespect to age, gender, and socioeco-nomic status. This study showed avariation in seroprevalence rates fromone country to another within LatinAmerica. Possibly influencing the epi-demiology of hepatitis B infection from TABLE 1. Study population attrition, hepatitis B seroprevalence in Latin America, 1996–1997 DominicanArgentina Brazil Chile Republic Mexico Venezuela TotalEnrolled 1 475 3 879 496 478 5 262 495 12 085Less: hemolysis, insufficient volume, or loss of sample 1 49 0 5 43 22 120Available for serology 1 474 3 830 496 473 5 219 473 11 965Less: missing data 20 177 0 0 7 4 208CRF a study population 1 454 3 653 496 473 5 212 469 11 757 a CRF = case report form. TABLE 2. Seroprevalence of anti-HBc antibodies in the study populations, by gender, LatinAmerica, 1996–1997 Total study SeropositiveCountry Gender population Number % 95% CI P valueArgentina Male 631 14 2.2 1.1–3.3Female 823 17 2.1 1.1–3.0Total 1 454 31 2.1 1.4–2.8 0.817Brazil Male 1 773 158 8.9 7.6–10.2Female 1 880 130 6.9 5.8–8.0Total 3 653 288 7.9 7.0–8.7 0.025Chile Male 229 2 0.9 0.0–2.1Female 267 1 0.4 0.0–1.1Total 496 3 0.6 0.0–1.3 0.475Dominican Republic Male 111 14 12.6 6.6–18.6Female 362 81 24.0 19.7–28.3Total 473 101 21.4 17.8–25.0 0.010Mexico Male 2 007 27 1.3 0.8–1.8Female 3 205 48 1.5 1.0–1.9Total 5 212 75 1.4 1.1–1.7 0.064Venezuela Male 205 5 2.4 0.4–4.4Female 264 10 3.8 1.6–6.0Total 469 15 3.2 1.7–4.7 0.409 Rev Panam Salud Publica/Pan Am J Public Health 6(6), 1999 381 country to country are differences ingeography, climate, wealth, degree of urbanization, and ethnic srcin. We found the highest overall sero-prevalence rates in the Dominican Re-public (21.4%) and Brazil (7.9%),which both also showed high sero-prevalence levels in children. Further-more, in Brazil and Argentina morethan 3% of children were anti-HBc-seropositive by the age of 1 year, indi-cating a vertical route of transmission.This study showed a dramatic andhighly significant increase in hepatitisB seroprevalence among those 16 andolder in all the countries studied,which is compatible with the fact thatsexual activity and other adult life behaviors are a major route of trans-mission (6). An association of higherseroprevalence with lower socioeco-nomic status was only found in Brazil.The similar seroprevalence levels wefound among different socioeconomicgroups in Mexico and Argentina mightmean that such differences may only become visible when dealing withlarger numbers of seropositive indi-viduals. Even if there had been dif-ferences in the seroprevalences amongthe socioeconomic groups in Mexicoand Argentina, the overall levels wouldstill be considered low and thereforesuch socioeconomic differences would be of limited significance.The levels of anti-HBc antibodiesfound in this study were consistentlylower than those that have been re-ported in other studies (1, 2, 7, 8). How-ever, the levels we found are compara- ble when just the adult population isconsidered, and the pattern of sero-prevalence throughout Latin Americais similar between this and other stud-ies. Earlier studies were conductedmainly on adult populations usingsamples from blood banks. In contrast,the current study investigated a di-verse cross-section of the population,including infants and children, and ismore likely to be a realistic representa-tion of the population. In the Dominican Republic wefound a high seroprevalence of hepa-titis B, which was in agreement with a previous study (2). However, theDominican Republic is atypical of theCaribbean, which is generally a low-endemic region (2). Typical of high-endemic countries, in the DominicanRepublic a high seroprevalence wasobserved in early life, suggesting avertical or horizontal mode of trans-mission. In addition, there was an in-crease for the adolescent and older agegroups, emphasizing the impact of sexual transmission. Brazil also showed a higher overallseroprevalence than the other coun-tries in the study, except for the Do-minican Republic. Brazil is a largecountry with a previously reportedspectrum of HBV prevalence and car-rier rates, with the highest values inthe western Amazon region and thelowest in the south of the country (9,10). As with other studies (1), we ob-served a high seroprevalence in Ma-naus, in the western Amazon basin. TABLE 3. Age distribution of anti-HBc antibodies (seropositivity) in Latin America, 1996–1997 Age group Argentina Brazil Chile Dominican Republic Mexico Venezuela(years) No./TP a % 95% CI No./TP % 95% CI No./TP % 95% CI No./TP % 95% CI No./TP % 95% CI No./TP % 95% CI1–5 12/306 3.9 1.7–6.0 25/655 3.8 2.3–5.2 0/100 0.0 0.0–0.0 7/71 9.9 2.9–16.7 7/850 0.8 0.2–1.4 2/121 1.7 0.0–3.96–10 8/293 2.7 0.8–4.5 58/726 8.0 6.0–9.9 0/100 0.0 0.0–0.0 5/76 6.6 1.2–12.1 7/866 0.8 0.2–1.4 3/74 4.1 0.0–8.511–15 3/289 1.0 0.0–2.2 30/621 4.8 3.1–6.5 0/97 0.0 0.0–0.0 6/55 10.9 2.6–19.1 7/847 0.8 0.2–1.4 0/120 0.0 0.0–0.016–20 2/295 0.7 0.0–1.6 37/665 5.6 3.8–7.3 1/97 1.0 0.0–3.0 17/65 26.2 15.4–36.8 9/763 1.2 0.4–1.9 3/65 4.6 0.0–9.721–30 1/183 0.5 0.0–1.6 77/625 12.3 9.7–14.8 0/52 0.0 0.0–0.0 33/119 27.7 19.6–35.7 20/1 119 1.8 1.0–2.5 3/53 5.7 0.0–11.831–40 5/88 5.7 0.8–10.5 60/361 16.6 12.7–20.4 2/50 4.0 0.0–9.4 33/87 37.9 27.7–48.1 25/767 3.3 2.0–4.5 4/36 11.1 0.8–21.3 a No./TP = number of seropositive persons/total population in that age group in that country. FIGURE 1. Seroprevalence of hepatitis B in different age groups in countries of Latin Amer-ica, 1996–1997 % s u b j e c t s a n t i - H B c s e r o p o s i t i v e Age group (years) xxxx xx * * * * * * llllll ¶ ¶ ¶ ¶ ¶¶ 1 to 5 6 to 10 11 to 15 16 to 20 21 to 30 31 to 404035302520151050 * ¶ x l ArgentinaBrazilChileDominican RepublicMexicoVenezuela 382 Silveira et al. • Hepatitis B seroprevalence in Latin America Hepatitis delta virus (HDV) is de-pendent on HBV for its replication,therefore HDV can only infect peoplewho are simultaneously infected withHBV (coinfection) or who are alreadycarriers of HBV (superinfection) (11).As a result, HDV seroprevalence tendsto parallel that of HBV. Although HBVinfection is largely perinatal, HDV issexually transmitted in later life (12–16). HDV superinfection has been im-plicated as the etiological agent in anumber of outbreaks of severe hepati-tis, including in the upper Amazon basin in Brazil and Peru, the upperOrinoco basin in Venezuela and Co-lombia, and in western Colombia.Given the severity of hepatitis result-ing from HDV infection, priority has been given to the immunization of allinfants and newborns in hyperen-demic regions, including the Amazonregion (17). The largest such programwas introduced in Brazil in 1989, andsince then two-thirds of all childrenhave completed a three-dose course of HBV vaccination as part of the WHOExpanded Programme on Immuniza-tion (1). HDV has generally not been foundin other regions of South America withHBV prevalence rates similar to thosein the Amazon region of Brazil. Re-cently, however, HDV infection ap-pears to be increasing among high-riskgroups in Buenos Aires and otherlarge South American cities (2). If thistrend continues, protection of adoles-cents against hepatitis B will be in-creasingly important in order to pre-vent HDV transmission.HBV vaccination has been part of the WHO EPI since 1992 (3). Reportsfrom countries outside Latin Americaclearly demonstrate that neonatal vac-cination programs can produce a dra-matic reduction in HBV seropreva-lence in the infant population within just a few years (18–20). In Latin Amer-ica, neonatal vaccination programshave been successfully implementedin areas of high endemicity in Ar-gentina, Brazil, Chile, Colombia, CostaRica, Cuba, Ecuador, Honduras, Mex-ico, Peru, Uruguay, and Venezuela (1,21). The extensive awareness that per-sons in those areas have of the diseasehas contributed to the high levels of coverage. However, in the rest of LatinAmerica, it appears there is a broaddeficiency in knowledge about HBV,with the exception of such high-riskgroups as intravenous drug users (1).Given this lack of awareness of HBV,along with the impact that sexualtransmission has on seroprevalencerates, as shown in this study, the per-sons most at risk in many areas of Latin America are those ranging in agefrom infancy to young adulthood.Studies in Italy have demonstrated the benefits of vaccinating both neonatesand 12-year-olds (22). This dual-targetvaccination strategy will lead to thecoverage of the population betweenthe ages of 0 and 24 years within 12years, after which time only neonateswill need to be vaccinated. This strategy has cost benefits, aswell as a rapid impact on the disease burden. Furthermore, the WHO hasrecommended combining HBV vac-cines with mandatory childhood vac-cines, such as the diphtheria, tetanus,and pertussis (DTP) vaccine (23). DTPvaccination has reached greater than80% coverage globally, so combiningHBV with it could rapidly expand HBVcoverage. The availability of DTPw-HBV and DTPw-HBV-Hib combina-tion vaccines will facilitate the imple-mentation of hepatitis B vaccinationprograms and also significantly reducethe delivery and storage costs associ-ated with childhood vaccination pro-grams (24). Furthermore, having togive fewer injections will reduce thediscomfort to infants, increase accep-tance among parents and health careworkers, and lead to improved compli-ance (23, 24). Acknowledgments. The authors wishto thank Rosanna Lagos, Tania Ace-vedo, Gabriela Fernández, Anna MariaCavalcanti, Miguel Tregnaghi, AngelaGentile, Alberto Manerola, Ricardo Rut-timan, Marcela Potin, and Assad Safaryfor their review of the data and help inpreparing this manuscript. Funding forthe study was provided by SmithKlineBeecham Biologicals. TABLE 4. Seroprevalence of anti-HBc antibodies analyzed for different socioeconomicgroups in Latin America, 1996–1997 Socioeconomic Argentina Mexicolevel No./TP a % 95% CI No./TP % 95% CIHigh 11/382 2.9 1.3–4.5 29/1 962 1.5 1.0–2.0Medium 10/484 2.1 0.9–3.3 18/1 661 1.1 0.8–1.6Low 10/588 1.7 0.7–2.7 27/1 589 1.7 1.1–2.3Brazil b Socioeconomic level No./TP % 95% CIHigh/Medium 106/1 955 5.4 4.3–6.3Low 182/1 698 10.7 9.2–12.2 a No./TP = number of seropositive persons/total population in that socioeconomic group in that country. b Comparison of high/medium group with low group in Brazil, P < 0.001.
