Hiv Infection among Pregnant Women Attended in Testing and Counseling Centers for Aids

OBJECTIVE: To estimate HIV prevalence and identify high-risk sexual behavior for infection in pregnant women who were given prenatal assistance. METHODS: Cross-sectional study based on attendance records of 8,002 pregnant women (25% of all municipalities) who lived in 27 municipalities in Southern Brazil in 2003 and had anti-HIV tests done in a testing and counseling center that performed prenatal assistance. Sociodemographic and behavioral data were gathered, as well as syphilis and HIV test results, during the individual counseling sessions registered in the data bank of the Sistema de Informações dos Centros de Testagem e Aconselhamento (Information System on Testing and Counseling Centers). Women who sought the centers for confi rmation of previous serology or were referred to this service due to the presence of AIDS symptoms were excluded from the data base. RESULTS: A total of 0.5% of all the pregnant women analyzed (CI 95%=0.3;0.6) were HIV positive. The only variable associated with HIV seropositivity was schooling. The majority of them were basically exposed through unprotected sexual intercourse with the only partner they had a steady relationship with. Younger pregnant women who were single, unemployed and had lower level of education constituted the group with highest exposure. CONCLUSIONS: The Sistema de Informações dos Centros de Testagem e Aconselhamento turned out to be useful for the epidemiological surveillance of HIV infection and high-risk behavior among pregnant women and could also be useful as regards other populations.


INTRODUCTION
In Brazil, the national HIV/AIDS control policy corresponds to the work that has been performed by governments and civil society for over two decades.The epidemiological surveillance component of this policy has as its general purpose to follow the temporal and spatial tendencies of the occurrence of AIDS and HIV infection, aiming at guiding the epidemic control actions on all levels of administration of the Sistema Único de Saúde (SUS -National Health System).*In order to subsidize this control, the epidemiological surveillance needs to make available relevant, reliable, up-to-date information about the occurrence of AIDS and HIV infection at an opportune time, as well as high risk behavior associated with the epidemic.This, as a result, enables to promote actions and measure changes in the epidemiological situation, with an emphasis on populations that present higher risk of HIV infection, making use of data originating from diverse sources of information.14   During the fi rst 20 years of the epidemic, however, the universal and compulsory notifi cation of AIDS cases was the fi rst strategy of the epidemiological surveillance in Brazil, included on the list of compulsory notifi cation aggravations since December 22 nd of 1986. 4,18,* However, apart from problems related to the completion and quality of the information produced, it must be added that AIDS surveillance refl ects the dynamics of transmission that occurred in the past (infection period). 19The criteria of defi nition of cases have been subject to reviews throughout time and the antiretroviral treatment has been modifying the course of expression of the disease in populations, rendering this information even more diffi cult to be interpreted.The population of pregnant women is especially important to be considered as part of a system of HIV surveillance, not only due to the risk of their having had sexual intercourse with HIV infected men, but also because of the possibility of vertical transmission to their children. 13The rates of HIV infection among pregnant women reported in several studies in Brazil are variable, but low, when compared to other developing countries.6-8,16,18   On the other hand, even though the Ministry of Health recommends that the prenatal services conduct serological tests and HIV/AIDS counseling, in a decentralized system such as the SUS, there are many assistance models in operation in the more than 5,000 Brazilian municipalities.In some of them, pregnant women have used or use the CTAs, usually as a decision made by the prenatal service itself.** The objective of the present study was to estimate the HIV prevalence and identify high risk sexual behavior for infection among pregnant women who were given counseling and were serologically tested for HIV in the CTAs as part of the prenatal assistance routine.

METHODS
Cross-sectional study based on computerized attendance records of 8,002 pregnant women living in 27 municipalities of the Southern Brazil, who were given counseling and had anti-HIV tests done in a CTA as an integral part of their prenatal routine.
The south region was selected as a result of the high proportion of CTAs that already used the SI-CTA in 2003.Among the municipalities that used the SI-CTA to register their attendances in the CTAs, those that performed part of the prenatal service (the STD/AIDS counseling and the test) for at least 25% of the pregnant women in the CTA were selected.The average coverage of pregnant population in these 27 municipalities was 51% (±24%).For the calculation of this coverage, the number of live newborns in 2002 was used, obtained from the Sistema de Informações de Nascidos Vivos (SINASC -Information System on Live Newborns), as a proxy of the pregnant population in each municipality.
The data were collected in interviews during individual counseling sessions (before or after the test), with guaranteed privacy and confi dentiality of information for the participants.The data collection was carried out by professionals of the services and the data were registered on the SI-CTA (during or after the session).Women who sought the CTAs for confi rmation of previous serology and those who were referred to this service due to the presence of suggestive AIDS symptoms were excluded from the data bank.
Sociodemographic variables (age, marital status, level of education and situation in the work market), variables relative to sexual behavior (sexual orientation, number of sexual partners in the previous year, use of condoms with steady or occasional partners throughout life and in the last sexual relation), and the results of serological tests for syphilis (VDRL) and HIV were analyzed.As a norm from the Ministry of Health, two different analyses by ELISA and one confi rmatory test for HIV detection were performed to defi ne an HIV positive result.
Two synthetic variables were also built: "consistent use of condoms", with steady or occasional partners, and "exposure to HIV during sexual intercourse"."Consis-tent use of condoms" was defi ned as the use of condoms in the last sexual relation and "every time" one has sexual intercourse with current partner.*On the other hand, "inconsistent use of condoms" was defi ned by one who has never used a condom with current partner or has used it "occasionally", or one who did not use a condom in the last sexual relation.
Regarding "HIV exposure during sexual intercourse" (Table 1), participants were classifi ed as: "low exposure" -women with steady or casual partner who referred to consistent use of condoms; "intermediate exposure" -women with a single, steady partner who related inconsistent use of condoms or those with more than one partner who related consistent use of condoms with at least one of them; and "high exposure" -those with more than one partner who did not relate consistent use of condoms with any of them.
The statistical analyses were carried out by means of the SPSS software.The statistical associations were assessed by the prevalence ratio, using 5% as the statistical signifi cance level.Differences in proportion were assessed by means of Pearson's Chi-square test.

RESULTS
The age of participants ranged from 12 to 55 years, the average age was 24.9 years (±6.76) and the median, 24 years, of which 58% were within the 12-25 year age group.Approximately 85% of participants related being married or living with a partner, of which 56.1% had eight or more years of formal education.A total of 47.4% of pregnant women related working, though 23.1% of these were unemployed (Table 2).Further-more, a total of 87% of those who did not report working were housewives and 12.6%, students.
Regarding high risk behavior for HIV infection, 11.5% related having had two or more sexual partners in the last 12 months and 13.1% related having sexual intercourse with steady and occasional partners (not mutually exclusive).As for the use of condoms, 2.2% related the consistent use of condoms with a steady partner, and none of these women was infected by HIV.Among those with casual partners, 52% reported consistent use of condoms.The positive VDRL rate was 0.5% (Table 2).
The prevalence of HIV infection in the sample studied was 0.5% (CI 95%: 0.3;0.6%).The only variable associated with the occurrence of HIV was the level of education, nearly three times lower (PR=0.34;CI 95%: 0.14;0.83)among women with eight or more years of formal education (0.4%), when compared to those who have three or fewer years of education (1.1%).Statistical difference between other variables studied and HIV infection was not observed (Table 2).
Concerning HIV exposure during sexual intercourse, 2.1% of participants were classifi ed as "low exposure", 87.1% as "intermediate exposure" and 10.8% as "high exposure" (Table 3).Women considered to be under high exposure were usually younger, single, unemployed, and had fewer years of formal education than those from the average and low exposure groups.Even though more HIV infected women had been classifi ed in the high and average exposure groups and no cases had been observed in the low exposure group, differences were not statistically signifi cant.maternity ward in the Southern Brazil, HIV infection prevalence was estimated at 0.5%. 18In another study conducted in Brazil, a 0.4% prevalence was registered among pregnant women at 13 to 24 years of age.*Both studies are from 1998 and their results fall under the same estimated confi dence interval as the present study

DISCUSSION
HIV prevalence rates comparable to the estimated ones in other Brazilian studies on pregnant women were observed. 7In a study with pregnant women at 15 to 49 years of age, during labor performed in a public  (CI 95%: 0.3;0.6).However, these rates are lower than the ones found in twelve Latin American and Caribbean countries, whose prevalences were equal to or higher than 1%, a prevalence of 2.9% found in the state of Rio Grande do Sul in 1999, and 3.4% in Southern Brazil in 1998. 7 In the present study, the level of education was the variable most strongly associated with HIV prevalence, a fact that corroborates fi ndings from other studies.In Brazil, the level of education is considered the best social status indicator, so that an association with the perception of the risk of getting infected by HIV has been suggested. 5,12,13,,20In fact, a Brazilian study on sexual behavior and HIV/AIDS* perception, conducted in a representative sample of adult urban population, revealed that the educational level is the most relevant variable to characterize the degree of knowledge about HIV/AIDS, indicating that the higher this degree is, the more education one has.Apart from education, high risk sexual behavior was also associated with a low degree of knowledge and unemployment.
In the present study it was observed that younger and single pregnant women are probably more exposed to HIV infection according to the defi nition of sexual risk previously proposed here.The context of sexual intercourse between partners who do not live together appears to be linked to higher exposure to HIV than those of women who are either married or living with a partner.Furthermore, single women are more likely to have multiple sexual partners.*Frequent casual dates can increase the probability of their having intercourse with an HIV infected partner,** even though this may not be the case in Brazil.*** In spite of the variables "consistent use of condoms" and "HIV exposure during sexual intercourse" not permitting the effectiveness of condoms as a protective barrier to be assessed, nor signifi cant statistical associations between exposure patterns and risk of HIV infection to be identifi ed, they proved to be useful to describe high risk behavior and to characterize this populational segment in socioeconomic terms.
The validity of information on referred use of condoms has been questioned, which possibly harms the assessment of the impact of prevention policies that focus on their use.*Some studies validated the consistency of the self-account about use of condoms, relating it to HIV infection. 3,* In the present study, it was not possible to identify signifi cant statistical differences between women who related the consistent use of condoms with casual or steady partner and those who related non-consistent use.Pregnancy enables to infer that these women had unprotected sexual intercourse in the last nine months, which renders the analysis of this variable in the study more diffi cult.
The validation of the data produced has been the issue of national and international debate. 15,17Nonetheless, there is a consensus on the need to inform possible bias concerning selection, information and confusion, in order to enable assessment of contributions. 11In this sense, it is important to emphasize that the pregnant women attended to in the CTAs of the 27 municipalities in the south region of Brazil are not representative of the group constituted by Brazilian pregnant women, so that the results obtained cannot be extrapolated to other municipalities in this country as the inner validation limits compromise the outer validation.
Furthermore, even though the rates of prevalence found among pregnant women and women during delivery are commonly used to represent the population of women at reproductive age, it is important to observe the possibility of selection bias in this type of approach.Apart from the demand of and access to healthcare services being often associated with several factors of HIV infection risk, this can reduce fertility and, thus, underestimate the prevalence of the infection, especially among women at 30 years of age or older. 9,10,16On the other hand, the association of the demand for prenatal care with high risk of infection could be minimized in areas with low HIV infection prevalence and high test coverage, something which would reduce the effect of selection bias.
The bias related to the trust regarding participants' responses to questions that could result in overestimation of the use of condoms or underestimation of less acceptable sexual behavior, for instance, could have been minimized in the present study.The interviews were conducted by qualifi ed counselors, with guaranteed information confi dentiality and privacy during the counseling sessions.
Another limitation of the present study refers to the cross-sectional design, which does not permit a temporal relation between events to be established.
Even though the data of health services cannot be extrapolated to the population as a whole, they are important for the epidemiological surveillance of HIV infection and high risk behavior in priority populations defi ned by the diverse levels of administration of the SUS.These data represent useful supplement for HIV/AIDS surveillance, currently performed by means of the Sistema de Informação de Agravos de Notifi cação (SINAN -Information System of Notifi cation Diseases) and can subsidize the planning of preventive activities and assistance for groups under higher HIV exposure.
The precision of estimates could be improved by regular serological studies based on population and by behavioral studies, especially to monitor temporal tendencies.Furthermore, the use of CTA data as a source of information for epidemiological surveillance is low cost and easy to implement on a local level of healthcare administration.
Among the possibilities and advantages for the SI-CTA to perform this HIV epidemiological surveillance, the possibility to follow the infection stages in population sectors must be pointed out and, simultaneously, high risk behavior for infection among priority groups attended to in the CTAs must be focused on.Behavioral data help explain the epidemic tendencies, permitting a realistic perspective of potential vectors of change throughout time to be prepared.

Table 1 .
Operational model of HIV exposure level per sexual relation.Southern, Brazil, 2003.

Table 2 .
Prevalence rates of HIV infection and prevalence ratio according to sociodemographic and behavioral variables among pregnant women.Southern, Brazil, 2003.(N=8002)*