Vertical transmission of HIV in Rio de Janeiro, Brazil

A cohort of 297 HIV-infected pregnant women was followed from January 1996 to December 2001. The overall transmission rate was 3.57% and remained constant over time. Low birthweight was independently associated with a higher risk of vertical transmission (P= 0.0072), whereas a longer duration of antiretroviral drugs during pregnancy was independently associated with a lower risk of transmission (P= 0.0084). Further decreases in transmission should be obtained by initiating prophylaxis earlier in pregnancy. Data about risk factors for the vertical transmission of HIV in Brazil are scarce. Results from a recent study in which HIV testing was offered at delivery to pregnant women demonstrated seroprevalence rates of 0.4–7.2% in various Brazilian cities [1]. The municipality of Rio de Janeiro ranks second in the number of reported cases of AIDS. The highest incidence (36 cases per 100 000 inhabitants) is in the central area, where Hospital dos Servidores do Estado (HSE) is the largest public hospital. It is a referral hospital for primary care units throughout the metropolitan area. We analysed risk factors for mother- to-child transmission (MTCT) in a population of HIV-infected pregnant women followed at HSE over 6 years (January 1996–December 2001). HIV infection was defined by two positive serological tests on different samples, plus a Western blot. Routine assessment included: demographic data, risk factors for HIV and other sexually transmitted diseases; HIV-related history; clinical and obstetric history; baseline clinical, gynaecological and obstetric evaluation; ultrasound for gestational age; viral load (nucleic acid sequence-based amplification or nucleic acid sequence-based amplification/Nuclisens), flow cytometry; routine serologies; complete blood count and biochemistry panel; urinalysis and urine culture, stool examination for parasites, purified protein derivative; Pap smear. The evaluation of HIV-exposed children included birthweight, length, head circumference, gestational age, and length of hospitalization. They were followed at least until the confirmation of HIV infection status, with viral load measured at 6 weeks and 4 months and HIV serology at 18 months. All patients were treated according to guidelines at the time of enrolment, zidovudine monotherapy in 1996 [2], mono or dual nucleosides in 1997 [3], and highly active antiretroviral therapy since 1998 [4,5], and received at least zidovudine for prophylaxis according to the ACTG 076 protocol [6]. The mode of delivery was determined according to obstetric indications until 2000. Thereafter, elective caesarean section was recommended whenever the viral load at 34–40 weeks gestation was 1000 copies/ml or greater, or was unknown [5]. Formula was provided to all mothers, and children were not breastfed. Preterm birth was defined as birth at less than 37 weeks. Low birthweight was defined as a birthweight of less than 2500 g. Analyses were undertaken with approval by the local Institutional Review Board. Bivariate analysis was performed to assess the potential associations between maternal and infant characteristics and MTCT. Chi-squared tests were used to test associations for categorical data, and analysis of variance or Kruskal–Wallis tests were used for continuous data. A multivariate logistic regression was used to model variables associated with the outcome of infection in children [7,8]. From January 1996 to December 2001, 297 pregnant women entered the cohort and underwent 319 deliveries, including 22 women with repeat deliveries and three sets of twins. These 319 deliveries resulted in 10 HIV-infected children, 270 uninfected children, and 39 children whose HIV infection status was not determined. Most of the women (66%) learned about their HIV infection status during the index pregnancy. The transmission rate was 3.13% [95% confidence interval (CI) 1.60, 5.51]. When it was calculated assuming the rate of HIV infection was the same for children lost to follow-up, the transmission rate was 3.57% (95% CI 1.82, 5.91) and was relatively constant over time. Most women (N = 128, 39.4%) received zidovudine only, with 90 (27.7%) receiving the combination of zidovudine/lamivudine, 23 (7.1%) receiving two-drug combinations other than zidovudine/lamivudine, and 84 (25.8%) receiving three antiretroviral drugs. The mean duration of antiretroviral use was 8.77 weeks for mothers of infected versus 16.06 for those of uninfected children. Although the number of antiretroviral drugs received during pregnancy was not associated with MTCT, a longer duration of receipt of antiretroviral drugs during pregnancy was associated with a lower risk of transmission (P = 0.04 in the bivariate analysis). The results of multivariate analyses are shown in Table 1. The variables studied were: antepartum viral load nearest delivery, antepartum CD4 cell count nearest delivery, duration of receipt of antiretroviral drugs during pregnancy, number of antiretroviral drugs received during pregnancy, low birthweight, preterm birth, duration of ruptured membranes, and mode of delivery. Preterm birth in our cohort is moderately related to low birthweight (Pearson's correlation coefficient was 0.42). In these adjusted analyses, low birthweight was independently associated with a higher risk of MTCT (P = 0.0072), whereas a longer duration of receipt of antiretroviral drugs during pregnancy was independently associated with a lower risk of transmission (P = 0.0084).Table 1: Results of multivariate logistic regression modeling: factors associated with mother-to-child transmission among HIV-infected pregnant women receiving care at Hospital dos Servidores do Estado, Rio de Janeiro.In one of the few studies of MTCT of HIV in Brazil after ACTG 076, the vertical transmission rate was estimated to be 7% in 1998 on the basis of data from several units including HSE [9]. In a prospective cohort study of 177 pregnant women in Rio de Janeiro from 1996 to 2000 [10] the transmission rate was 2.75%, but analyses were limited as a result of the relatively small sample size. The estimated rates of MTCT of HIV in our study cohort from 1996 to 2001 are under 4%. We found a significant association between low birthweight and the HIV infection status of the infant. Low birthweight has been associated with HIV infection of the mother [11,12]. Our data showed that 60% of low birthweight children were born full-term (≥ 37 weeks). Other authors also found some association between low birthweight and MTCT of HIV [13,14]. The duration of exposure to antiretroviral prophylaxis or treatment during pregnancy was a significant protective factor in our cohort, as also shown in a paper by Lallemant and colleagues [15]. This emphasizes the importance of diagnosing HIV early in pregnancy because almost 70% of women were diagnosed with a mean gestational age of 24 weeks.