Prevalence and risk factors for HIV, syphilis, hepatitis B, hepatitis C, and HTLV-I/II infection in low-income postpartum and pregnant women in Greater Metropolitan Vitória, Espírito Santo State, Brazil


Prevalência e fatores de risco para a infecção por HIV, sífilis, hepatite B, hepatite C e HTLV-I/II em parturientes e gestantes de baixa renda atendidas na Região Metropolitana de Vitória, Espírito Santo, Brasil



Lucia Helena Mello de LimaI; Maria Carmen VianaII

IUnidade de Referência Ambulatorial para DST/AIDS, Serra, Brasil
IIEscola de Medicina da Santa Casa de Misericórdia de Vitória, Vitória, Brasil





Sexually transmitted infections during pregnancy pose a major risk to the fetus due to vertical transmission. The study's objective was to determine the prevalence of HIV, syphilis, hepatitis B and C, and HTLV-I/II infection among low-income postpartum and pregnant women treated in Greater Metropolitan Vitória, Espírito Santo State, Brazil, and the risk factors associated with these infections. A cross-sectional study was conducted from February to October 1999 assessing postpartum and pregnant women from the maternity ward of the Vitória Mercy Hospital and the Carapina Outpatient Referral Unit in the Municipality of Serra, respectively. Patients were systematically interviewed and had blood samples drawn for serological tests (HIV 1&2, VDRL, HbsAg, anti-HCV, and HTLV-I/II). A total of 534 patients (332 postpartum and 202 pregnant women) were assessed. Seroprevalence rates for the target infections in postpartum and pregnant women and the overall sample were as follows, respectively: HIV 0.9%, 0%, and 0.6%; syphilis 2.1%, 3.6%, and 2.7%; HBV 1.2%, 1%, and 1.1%; HCV 1.8%, 0.6%, and 1.4%; and HTLV-I/II 1.7%, 0.6%, and 1.3%. Factors associated with the various infections are presented and analyzed in light of other research findings from the literature.

Sexually Transmitted Diseases; Pregnancy; Vertical Transmission Disease


A ocorrência de doenças sexualmente transmissíveis durante a gravidez representa risco aumentado de morbidade e mortalidade para o feto e neonato em virtude da transmissão vertical. O objetivo deste estudo foi identificar a prevalência da infecção por HIV, sífilis, hepatites B e C e por HTLV-I/II em parturientes e gestantes de baixa renda da Região Metropolitana de Vitória, Espírito Santo, Brasil, e avaliar fatores de risco associados a essas infecções. Foi realizado estudo de corte transversal de fevereiro a outubro de 1999, avaliando 534 pacientes (332 parturientes e 202 gestantes). Todas as participantes foram submetidas a entrevista semi-estruturada, após a qual foi coletada uma amostra de sangue para realização de testes sorológicos (HIV 1 e 2, VDRL, HbsAg, anti-HCV e HTLV-I/II). A soroprevalência das infecções estudadas entre as parturientes, as gestantes e no total da amostra foi, respectivamente: HIV 0,9%, 0% e 0,6%; sífilis 2,1%, 3,6% e 2,7%; HBV 1,2%, 1% e 1,1%; HCV 1,8%, 0,6% e 1,4% e HTV-I/II 1,7%, 0,6% e 1,3%. Os fatores associados às infecções avaliadas são descritos e analisados à luz de achados de outros investigadores.

Doenças Sexualmente Transmissíveis; Gravidez; Transmissão Vertical de Doença




Sexually transmitted diseases (STDs) and their complications are among the five main causes involved in the use of health services in developing countries 1. The high incidence of such infections in women using prenatal services and gynecology and family planning clinics is an indicator of the problem's importance. Women are more vulnerable than men to STDs, in both biological and cultural and socioeconomic terms. Most STDs are asymptomatic in women, and their consequences can be serious, sometimes even fatal, both for themselves and their offspring 2. Knowledge of STD prevalence in pregnant women in a given region is necessary for implementing prevention and intervention aimed at containing the spread of the diseases and their vertical (mother-to-child) transmission.

In Brazil, HIV infection has evolved at a rapid pace among women, posing a serious public health problem. The highest percentage of cases is in the 25-34-year bracket, in the prime of childbearing age. One direct consequence of the increased female presence in the AIDS epidemic is the progressive increase in vertical transmission 3. Syphilis in adults and neonates is still a public health problem in Brazil and elsewhere in the world. Although pregnancy does not alter the clinical course of syphilis, early diagnosis during prenatal care is essential to prevent congenital infection and administer adequate treatment to the mother 4. Although hepatitis B (HBV) infection can be prevented by vaccination, it remains endemic in various regions of the world, with areas of low (< 2%), medium (2-7%), and high prevalence (> 7%) 5. Vertical HBV transmission is the main form of acquiring the infection in regions with high prevalence, where more than 90% of infected children become persistent carriers of the virus and can present late chronic liver disease 6. Hepatitis C (HCV) prevalence in pregnant women has been extensively investigated in various parts of the world through HCV testing, varying from 0.7 to 4.4%, with 1 to 2% in pregnant women in Europe, Japan, and the United States 7. HTLV is endemic in various regions of Japan, Africa, Australia, Alaska, South America, and the Caribbean islands 8. In Brazil, the prevalence varies and is highest in the States of Bahia, Pernambuco, and Pará 9.

Since there is limited information on STD prevalence in pregnant women in Brazil and specifically in the State of Espírito Santo, the current study aimed to evaluate the prevalence of HIV, Treponema pallidum, HBV, HCV, and HTLV-I/II infections in low-income pregnant and postpartum women in Greater Metropolitan Vitória (the State capital of Espírito Santo) and identify factors associated with these infections. A further objective was to identify possible risk factors associated with these infections and detect potential differences between women giving birth at public maternity hospitals and those seeking prenatal care.



A cross-sectional study was performed from February to September 1999, evaluating postpartum and pregnant women treated, respectively, at the maternity ward in the teaching department of the Vitória Mercy Hospital (Santa Casa de Misericórdia) and the Carapina Outpatient Referral Unit, with a service specializing in STD/AIDS, the latter located in the Municipality (county) of Serra, Espírito Santo State. During this period, all postpartum women who had been admitted for childbirth in the obstetrics ward of the Mercy Hospital in Vitória were invited to participate, and the following day they answered a standardized semi-structured interview exploring socio-demographic data, information on sexual activity, use of psychoactive substances, behavior, lifestyle, and domestic and sexual violence, besides clinical and obstetric data and information on possible risk behaviors by their current and previous sex partners. After the interview, a blood sample was drawn to test for HIV, T. pallidum, HBV, HCV, and HTLV-I/II infection. During the same period, and following the same procedures, we evaluated all the pregnant women who came to the Carapina Outpatient Referral Unit for HIV testing as part of their prenatal care and received pre-test counseling.

HIV infection was detected using ELISA, and positive or indeterminate cases were confirmed by indirect immunofluorescence (IIF) using kits provided by the Oswaldo Cruz Foundation. Syphilis testing used VDRL (Laborclin, Pinhais, Brazil). HbsAg (hepatitis B surface antigen) was detected using AXSYM (Abbott Laboratories, São Paulo, Brazil), a third-generation microparticle enzyme immunoassay (MEIA). Presence of hepatitis C antibody was investigated with AXSYM-ABBOTT. HTLV-I/II antibodies were detected using the ABBOTT HTLV 1/HTLV 2 EIA kits.

The sample size was calculated as a function of the lowest estimated prevalence of the infections in the target population, namely HIV infection (from 1 to 2 %).

All the information collected was coded and stored in a databank using SPSS-Data Entry (SPSS Inc., Chicago, USA).

The groups were characterized according to the target variables using descriptive statistical analysis methods. Prevalence of the respective infections was defined as the percentage of positive cases for each serology, and the standard error and 95% confidence interval (95%CI) were calculated for each prevalence rate. Associations between HIV, syphilis, HBV, HCV, and HTLV infection and possible risk factors (socio-demographic, sexual, behavioral, and clinical) were tested with the chi-square test with Yates continuity correction. The magnitude of associations was assessed using odds ratios (OR), with respective 95%CI, and logistic regression models were used to control for socio-demographic variables. In all the analyses the tests were considered two-tailed, and p was set at 0.05.

Participation was voluntary, with written consent provided by the postpartum or pregnant woman. The study was approved by the Institutional Review Board of the School of Sciences, Mercy Hospital (Santa Casa de Misericórdia), Vitória. The diseases that were identified and that had available treatment were treated after identification and communication of the diagnosis, and the viral infections were referred for specific follow-up.



A total of 534 patients participated in the study: 332 postpartum women who had been admitted for childbirth at the Mercy Hospital maternity ward in Vitória and 202 pregnant women receiving prenatal care at the Carapina Outpatient Referral Unit in the Municipality of Serra, with 100% response rates in both groups. The institutions treat low-income patients from the general population that live in Greater Metropolitan Vitória and receive care through the Unified National Health System (SUS).

In relation to socio-demographic data (Table 1), there were no significant differences between the groups except in relation to the woman's home municipality (county), as a function of location of the two services, and the proportion of black women, which was higher in the postpartum group. Both groups had low family income and high dropout rates before completing primary school. Lower schooling was associated with teenage pregnancy: only 17.5% of women who became pregnant before their 19th birthday continued studying beyond primary school, as compared to 33.2% of those who became pregnant later (χ2 = 15.3; p = 0.000).

Table 2 shows the prevalence rates for each infection, as well as the overall prevalence. There was no statistically significant difference between the groups.

Overall, the postpartum women presented more health problems than the group of pregnant women (Table 3). As for childbearing history, the first pregnancy occurred at a quite early age, on average 19.6 years (standard deviation - SD = 4.5), with no significant difference between the two groups. There was a frequent history of abortion, both spontaneous (miscarriage) and induced, and abortion was more frequent in the postpartum group (OR = 2.6; 95%CI: 1.64-4.18), who showed much higher rates than those reported in the literature for this population (around 15%). History of STDs was similar to that in low-risk women in developing countries (6-8%) and was higher in the postpartum group (OR = 2.5; 95%CI: 1.02-6.83). Among the STDs reported by patients, genital warts, syphilis, and gonorrhea were the most frequent, although such information can be imprecise, subject to recall bias and depending on the diseases having been diagnosed.

Table 4 shows the variables associated with the various infections, showing possible risk or predictive factors, as well as the odds ratios resulting from bivariate analysis and after controlling for age, schooling, and ethnicity in logistic regression models. Considering the presence of any of the sexually transmitted infections (N = 35; Table 2), predictive factors for infection were low schooling (< 4th grade, OR = 2.7; 95%CI: 1.24-5.74), multiple sex partners (> 3 partners, OR = 2.7; 95%CI: 1.13-5.97), and early age at sexual initiation (< 15 years, OR = 2.9; 95%CI: 1.35-6.25).



The two groups appear to represent the population of low-income pregnant women treated at public health services at different moments in relation to pregnancy. No socio-demographic, lifestyle, or behavioral differences were detected between the groups. The main difference, and one of the targets of comparison in this study, is that the postpartum sample was representative of a population that had sought urgent treatment (i.e., for childbirth), while the other group consisted of pregnant women participating actively in a prenatal care program. This difference may represent greater attention to health care in general, as corroborated by the clinical differences shown in Table 3. However, there were no significant differences between the groups in the prevalence rates for the infections investigated in this study. Thus, our findings can be generalized to this target population, namely that of low-income pregnant women treated in the SUS, taking these differential characteristics into account. The sample size was calculated to obtain estimated prevalence rates for the respective infections and compare the groups; however, it proved insufficient to demonstrate a statistical association between these infections and the majority of the independent variables, in the attempt to identify possible risk or predictive factors for the infections. However, considering the total sample and the overall prevalence, we identified important factors that increased the likelihood of infection: low schooling, early sexual initiation, and multiple sex partners.

Prevalence studies in pregnant women are particularly useful, since they provide estimates that can be extrapolated to the general population of sexually active women, since by definition the population of pregnant women does not include situations involving special risk, as in the case of women treated at STD or emergency clinics 10. The most common STDs affect 6 to 8% of low-risk women seen at prenatal clinics in developing countries 11. In this study, 6.5% of the patients presented any current STD.

Although testing and prophylactic medication are available in Brazil, children continue to become infected through vertical transmission. As a result of factors like low adherence to prenatal care and/or late uptake, many women reach the maternity departments without adequate prenatal follow-up; in addition, overall prenatal HIV screening coverage is well below 40%, and is even lower in the pregnant women most vulnerable to infection 12. In the current study, although most of the postpartum women had received prenatal care (94.1%), beginning in the first trimester and with more than six visits, only 40.4% had been tested for HIV, suggesting that on this point the prenatal services fail to meet the Ministry of Health guidelines. According to the Sentinel Study on Postpartum Women for the year 2004, HIV prevalence in this population in Southeast Brazil (where Vitória is located) was 0.537%, similar to the rate in our study (0.6%) 13.

History of STDs in the current pregnancy was a risk factor for HIV infection. Data from the literature show that the presence of an untreated STD can increase the risk of both acquiring and transmitting HIV. On the other hand, improving the efficiency of STD management can reduce the incidence of HIV infection in the general population by some 40% 14. The association between other STDs and HIV infection has been demonstrated by various studies, thus reinforcing the need for HIV serological testing in patients with other STDs 10. Prevention and treatment of other STDs are thus important components of any HIV prevention strategy. Other factors associated with HIV infection were current partner with a history of blood transfusion and a previous HIV-positive partner. Importantly, statistical evaluation of possible risk factors was compromised by the low infection rate. Miranda et al. 15, evaluating pregnant women in the Municipality of Vitória, observed that the main HIV transmission route was sexual, in women with a single partner, and who reported no other risk behavior. In another study in the Municipality of Vitória, Miranda et al. 16 showed an association between HIV infection and history of injecting drug use and prostitution in pregnant women, and history of incarceration and injecting drug use in their partners, risk factors that have been reported in various other studies. Lima & Tanure 17, evaluating pregnant women in the Municipality of Serra, reported that the only factor statistically associated with HIV infection in this population was history of a drug-using sex partner.

The current study showed a high prevalence of syphilis in both the postpartum group (2.1%) and the pregnant women (3.6%). Considering that 94.1% and 100%, respectively, reported undergoing or having undergone prenatal care, one has to conclude that there was some flaw: syphilis testing was not performed, or was performed inadequately, or treatment was inadequate or incomplete. In many regions of the country, more than 95% of pregnant women attend prenatal care, and 91.5% give birth in maternity hospitals; however, the maternal and perinatal morbidity and mortality related to syphilis remain high, reflecting important deficiencies in care 12. According to studies conducted in the Municipality of Vitória 16, despite campaigns by the Ministry of Health to reduce congenital syphilis, the rates remained stable and high in a five-year period (2.9% in 1995 and 3% in 2000). In the current study, the prevalence rate (2.7%) was consistent with findings from other studies in Brazil. Of the seven VDRL-positive postpartum women in this study, only one had not attended prenatal care, while the other six had four or more prenatal visits and four had their first prenatal visit in the first trimester; even so, only two (28.6%) reported having treated their syphilis with penicillin during the pregnancy. Factors associated with syphilis in this study were: previous history of STD and STD during the current pregnancy, sexual initiation at 16 years of age or younger, and drug use in the postpartum group when the groups were evaluated separately (OR = 12.2; 95%CI: 2.0-76.1) and history of domestic violence. Factors related to the current partner, like his sexual contact with multiple partners, drug use, alcohol abuse, and STDs were also associated with syphilis. Miranda et al. 16, analyzing pregnant women in the Municipality of Vitória, found similar risk factors: previous history of STDs, sexual initiation before 16 years of age, and sex partners with a history of injecting and non-injecting drug use.

Overall HbsAg prevalence in this study (1.2%) was consistent with areas of low endemicity, where fewer than 2% of the population are HBV carriers, with a lifetime risk of infection less than 20%, and where the majority of infections occur in adults from high-risk groups; it was also similar to the rate found by Miranda et al. 16 in pregnant women in the Municipality of Vitória (1.1%). Although various factors showed high odds ratios, no risk factor was statistically associated with HBV infection, due to the small number of HBV seropositive women in the study sample.

HCV prevalence in the postpartum group (1.8%) and pregnant women (0.6%) was similar to findings from other studies. Body tattooing and blood transfusion were factors associated with hepatitis C, in addition to a report of previous partners with a history of blood transfusion. According to the Centers for Disease Control and Prevention (CDC) 18, although some studies have found an association between body tattoos and HCV infection in selected populations, it is not known whether these results can be transposed to the general population. Any percutaneous exposure has the potential to transfer infected blood and transmit bloodstream pathogens. In addition, body tattoos can be associated independently with other risk factors, like use of psychoactive substances and unprotected sex and/or sex with multiple partners. These associations were corroborated by the current study. Alcohol use during pregnancy (OR = 4.0; 95%CI: 1.79-8.87), history of smoking (OR = 6.1; 95%CI: 2.57-14.47), and drug use (OR = 6.1; 95%CI: 1.61-19.22), as well as reported sexual contact with multiple partners (1-4 vs. > 4, OR = 9.4; 95%CI: 3.79-22.50) were statistically associated with the presence of body tattoos.

This study found high HTLV prevalence, especially in the postpartum group (1.7%). However, the real prevalence could not be determined, since the confirmatory test was not performed. Prevalence in the group of pregnant women was 0.6%, close to the rate found in pregnant women in Salvador, Bahia State (0.7-0.88%) 19. Vertically acquired HTLV-I is associated with the occurrence of adult T-cell leukemia/lymphoma in 1-5% of the infected children and myelopathy/tropical spastic paraparesis in 0.25-3% 20. A significant association was observed between history of domestic violence and HTLV-I/II infection, perhaps as a marker of unfavorable socioeconomic conditions. Santos et al. 19, analyzing 1,024 pregnant women, failed to identify risk factors in the sample, since the seropositive women did not report blood transfusions or intravenous drug use. The evidence that HTLV-I-positive women had more than two sex partners as compared to the reference group, with single partners, reinforces the hypothesis of sexual transmission in this population. Bittencourt et al. 21, evaluating 6,754 pregnant women in Salvador, found a 0.84% prevalence rate and observed that the epidemiological data for the HTLV-positive mothers suggested that vertical transmission was the most frequent mode of HTLV-I transmission. A total of 88.5% of the women reported that they had been breastfed, and few women had received blood transfusions or had more than ten lifetime sex partners. These hypotheses could not be tested in the study, since these variables were not assessed; however, none of the six HTLV-positive women showed associations that might indicate the infection route. However, the high prevalence found in the study might be inflated, since it lacked corroboration by the confirmatory test. At any rate, more studies are needed to assess the importance of this infection before specific serology is included in prenatal testing and for the mothers to be counseled and oriented concerning this disease, considering that HTLV can be transmitted vertically to the infant, especially during breastfeeding.


Final remarks

This study's findings allow reaffirming the importance of serological tests for STD diagnosis during prenatal care, allowing preventive and therapeutic measures to be adopted, both to contain the spread of such diseases and to prevent their transmission to the fetuses and newborns. These measures could definitely help reduce the incidence of HIV/AIDS, syphilis, HBV, HCV, and adult T-cell leukemia/lymphoma and HTLV-I-associated myelopathy in Brazil as a whole and specifically in the State of Espírito Santo.



L. H. M. Lima conducted the data collection and participated together with M. C. Viana in the study design, development and review of the database, analysis of the results, and elaboration of the tables, as well as the literature review and preparation of the manuscript.



The authors wish to thank Dr. Fausto E. Lima Pereira and the Nucleus on Infectious Diseases for his valuable contribution to this study, Dr. Kelly Rose Areal for conducting the serological tests, and the Coordinating Body for Graduate Studies (CAPES) for its financial assistance.



1. World Health Organization. Global prevalence and incidence of selected curable sexually transmitted infections: overview and estimates. Geneva: World Health Organization; 2001.         

2. World Health Organization. Women and sexually transmitted infections. (accessed on 22/Oct/2001).         

3. Coordenação Nacional de DST e AIDS, Secretaria de Políticas de Saúde, Ministério da Saúde. Guia de tratamento: recomendações para a profilaxia da transmissão materno-infantil do HIV e terapia anti-retroviral em gestantes. Brasília: Ministério da Saúde; 2001.         

4. Sheffield JS, Wendel Jr. GD. Syphilis in pregnancy. Clin Obstet Gynecol 1999; 42:97-106.         

5. Margolis HS, Alter MJ, Hadler SC. Hepatitis B: evolving epidemiology and implications for control. Semin Liver Dis 1991; 11:84-92.         

6. Newell ML, Thorne C, Pembrey L, Nicoll A, Goldberg D, Peckham C. Antenatal screening for hepatitis B infection and syphilis in the UK. Br J Obstet Gynaecol 1999; 106:66-71.         

7. Zanetti AR, Tanzi E, Newell ML. Mother-to-infant transmission of hepatitis C virus. J Hepatol 1999; 31 Suppl 1:96-100.         

8. Tsuji Y, Doi H, Yamabe T, Ishimaru T, Miyamoto T, Hino S. Prevention of mother-to-child transmission of human T-lymphotropic virus type-I. Pediatrics 1990; 86:11-7.         

9. Gabbai AA, Bordin JO, Vieira-Filho JBP, Kuroda A, Oliveira AS, Cruz MV, et al. Selectivity of human T-lymphotropic vírus type-1 (HTLV-1) and HTLV-2 infection among different populations in Brazil. Am J Trop Med Hyg 1993; 49:664-71.         

10. Ramos MC, Curcio BL, Tessaro M, Meneghetti H, Marques LC, Duarte TP, et al. Prevalência da infecção pelo HIV e da soropositividade do VDRL em gestantes. J Bras Doenças Sex Transm 1999; 11: 25-30.         

11. Sciarra JJ, Sexually transmitted diseases: global importance. Int J Gynaecol Obstet 1997; 58:107-19.         

12. Programa Nacional de DST e AIDS, Ministério da Saúde. Projeto Nascer. Brasília: Ministério da Saúde; 2003. (Série F. Comunicação e Educação em Saúde).         

13. Secretaria de Vigilância em Saúde, Ministério da Saúde. Recomendações para profilaxia da transmissão vertical do HIV e terapia anti-retroviral em gestantes. Brasília: Ministério da Saúde; 2006.         

14. Gilson L, Mkanje R, Grosskurth H, Mosha F, Picar J, Gavyole A, et al. Cost-effectiveness of improved treatment services for sexually transmitted diseases in preventing HIV-1 infection in Mwanza Region, Tanzania. Lancet 1997; 350:1805-9.         

15. Miranda AE, Tonini RC, Friço AR, Alves MC, Foletto SR, Ferreira BMC, et al. Soroprevalência e fatores de risco da infecção pelo HIV em gestantes em Vitória. GO Atual 2000; IX:21-3.         

16. Miranda AE, Alves MC, Neto RL, Areal KR, Gerbase AC. Seroprevalence of HIV, HBV and Syphilis in women at their first visit to public antenatal clinics in Vitória, Brazil. Sex Transm Dis 2001; 28:70-3.         

17. Lima LHM, Tanure LVC. Prevalência e fatores de risco para a infecção pelo Vírus da Imunodeficiência Humana em gestantes atendidas no município da Serra, Espírito Santo, Brasil e o perfil epidemiológico dessa população. In: Memorias Foro 2003 - II Foro en VIF/SIDA/ITS de América Latina y el Caribe. Habana: Grupo de Cooperación Técnica Horizontal de América Latina y el Caribe en VIH/SIDA; 2003. p. 11-2.         

18. National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, Centers for Disease Control and Prevention. CDC's position on tattooing and HCV infection. (accessed on 02/Mar/2002).         

19. Santos JI, Lopes MAA, Deliège-Vasconcelos E, Couto-Fernandez JC, Patel BN, Barreto ML, et al. Seroprevalence of HIV, HTLV-I/II and other perinatally-transmitted pathogens in Salvador, Bahia. Rev Inst Med Trop São Paulo 1995; 37:149-51.         

20. Kaplan JE, Khabbaz RF. The epidemiology of human T lymphotropic virus types I and II. Rev Med Virol 1993; 3:137-48.         

21. Bittencourt AL, Dourado I, Bastos Filho P, Santos M, Valadão E, Alcântara LC, et al. Human T-cell lymphotropic virus type I infection among pregnant women in northeastern Brazil. J Acquir Immune Defic Syndr 2001; 26:490-4.         



L. H. M. Lima
Unidade de Referência Ambulatorial para DST/AIDS
Rua Amélia Tartuce Nasser 1055, apto. 201
Vitória, ES 29065-020, Brasil

Submitted on 19/Dec/2006
Final version resubmitted on 12/Nov/2007
Approved on 10/Jan/2008

Escola Nacional de Saúde Pública Sergio Arouca, Fundação Oswaldo Cruz Rio de Janeiro - RJ - Brazil