Bacterial vaginosis and illness during pregnancy increase risk of in utero HIV transmission

Bacterial vaginosis (BV) and illnesses that occur during pregnancy increase the likelihood of mother-to-child transmission (MTCT) of HIV in utero according to a study presented at the Sixteenth International AIDS Conference in Toronto. The study, which was conducted in mother/child pairs in Nairobi, also confirmed earlier observations that in utero transmission is associated with higher viral loads and less than three weeks of AZT (zidovudine, Retrovir) prophylaxis for prevention of MTCT (PMTCT).

The findings have important clinical indications in settings where pregnant women access antenatal services, since many common illnesses in women with HIV can prevented with cotrimoxazole prophylaxis and BV can be treated.

In utero HIV transmission

“In utero transmission is responsible for 20% to 30% of vertical transmission events in breastfeeding populations, which translates into approximately 200,000 intra utero infections annually in sub-Saharan Africa,” said Professor Carrie Farquhar of the University of Washington Seattle, and the Seattle Nairobi Research Group. However, currently most PMTCT interventions target transmission during delivery and breastfeeding (although studies have suggested that short-course antiretroviral regimens may decrease some late in utero transmission, it is usually administered too late in the third trimester to prevent all of these infections).

As the existing PMTCT strategies become more widely implemented, in utero transmission will account for a greater proportion of paediatric HIV-1 infections. However, relatively little is known about transmission in utero or how to stop it, because of difficulties conducting research into these events.

“Blood specimens are required at birth [when] women often deliver at home [in resource limited settings]... Second, in utero events may be easily misclassified as due to intra partum or early breastmilk transmission,” said Prof. Farquhar.

Furthermore, late in utero infection may not be detectable for up to a week after birth. Finally, women in resource-limited settings often do not come in for antenatal care until very late in pregnancy — which makes it very difficult to determine whether acute HIV infection or other risk factors occurred earlier in the course of the pregnancy.

To date, a number of maternal and viral factors that have a role in in utero transmission have been identified. These include maternal viral load, antenatal antiretroviral use, infant gender (with females more likely to be infected), low infant birthweight, and ascending infections, which spread up from the cervical-vaginal regions and can lead to chorioamnionitis (inflammation of the membrane covering the foetus).

The study in Nairobi

Prof. Farquhar and colleagues conducted a prospective study in Nairobi to determine whether they could identify any other factors that contribute to in utero transmission.

From 1999-2002, they attempted to enrol HIV-infected women attending Nairobi city clinics before the 32^nd week of pregnancy. At the enrolment visit, histories were obtained with specific questions about illness during pregnancy, and women were screened for sexually transmitted infections, which were then treated, as well as candida and bacterial vaginosis (which was not treated since the samples were not evaluated until after the study’s conclusion).

Participants were given AZT beginning from week 34 to 36 of pregnancy until delivery and encouraged to come into the hospital for delivery, where nurse midwives obtained plasma and cervical specimens for viral load measurements. Infants also had a blood draw at birth which was tested by both HIV RNA and DNA PCR. In utero transmission was defined as either a positive RNA or DNA test by 48 hours. Follow-up and HIV testing of the infants continued until one year after birth.

Approximately 36,000 women were seen at the clinics; 88% consented to being tested for HIV; and 14% were identified as HIV-positive. Of these women, 463 agreed to participate in the study and were followed through delivery.

Out of the 463 pregnancies, there were 88 HIV-infected infants, 77 infant deaths and 48 infants were lost to follow-up.

Of the HIV infected infants, 29 (33%) tested positive at birth and were thus infected in utero. This represented an in utero transmission rate of 6.3% — one third of all HIV infections in this cohort. Thirty-seven (42%) of the infants were uninfected at birth but tested positive by month one and were thus infected during delivery or by early breastmilk exposure; ten infants were infected via breastmilk after month one; and precise timing of infection could not be determined for twelve infants for whom specimens were unavailable at birth.

Infant and maternal characteristics for the 29 infants known to be infected in utero were then compared to those of the 422 who were either not infected or who became infected later.

In univariate analyses, mean plasma viral loads and cervical viral loads were approximately a half log higher in women who transmitted in utero (pin utero were also less likely to have received at least three weeks of AZT (p=0.04).

Lower infant gestational age was significantly associated with greater transmission risk (p=0.01) and there was a trend for female babies to be more likely to be infected (p= 0.06).

An antenatal diagnosis of a sexually transmitted infection (which included gonorrhoea, chlamydia, trichomonas) was not associated with in utero transmission, but it should be noted that these infections received prompt treatment in this study.

However, BV was diagnosed in significantly more women (59%) who transmitted in utero than in those who did not (35%). BV is relatively common in resource limited settings, and associated with altered vaginal flora, specifically a reduction in lactobacilli and a greater proportion of other bacteria.

In addition, women who transmitted were more likely to have had an AIDS-defining illness in the previous year (p=0.01), and a greater proportion of women who transmitted in utero reported an illness such as fever, cough and diarrhoea during pregnancy (p=0.01).

The researchers then conducted a multivariate analysis that adjusted for 32-week viral load, CD4 percent, and duration of AZT use. Viral load (plasma and cervical) remained highly predictive of in utero transmission conferring a 1.9- and 1.5-fold increased (respectively) risk per log difference in viral load. Receipt of more than three weeks of AZT also remained protective, when adjusting for viral load and disease stage.

Finally, BV and illness during pregnancy were also independent predictors of tranmission in this model. Women with BV were 3-fold more likely to transmit, and women with illness during pregnancy 2.6-fold more likely to transmit.

Clinical implications

“Illness during pregnancy may be contributing to transmission by transiently elevating viral load, or by causing immune activation... it may also represent primary HIV,” said Prof. Farquhar. Other studies have documented that women who become HIV infected during pregnancy are known to be at increased risk of transmitting the virus.

BV may be increasing the likelihood of transmission by increasing the risk of chorioamnionitis and placental compromise.

This study’s findings could have several important clinical implications.

“Interventions should target viral load in pregnant women and also maternal health,” Dr Farquhar said.

Antenatal testing and counselling during pregnancy is crucial and should include services to help women who test negative avoid HIV infection during pregnancy (or breastfeeding). Meanwhile, for those who test HIV-positive, cotrimoxazole prophylaxis to prevent other infections could also reduce in utero HIV transmission.

Finally, BV can be treated with a one week course of oral metronidazole or one dose of tinidazole (which has a longer half-life) or with gels or creams containing metronidazole or clindamycin. It is also important to prevent reinfection from sexual partners (either by treating them or avoiding sex).