HIV and women – overview

17 March 2001. Related: Conference reports, Women's health, Pregnancy, CROI 8th (Retrovirus) 2001.

Polly Clayden, for HIV i-Base

At first glance this conference showed an impressive range of sessions and posters dedicated to women’s HIV care.

Further examination however revealed very few studies actually addressing women-specific research, independent of their role in reproduction. As usual there were challenging issues of semantics – HIV-infected and ‘normal’ women being our conference favourite.

Many studies reported and planned for the developing world demonstrated ever more complex ways of not treating women’s own health – vertical transmission prophylaxis for mothers, vertical transmission prophylaxis for babies, breastfeeding prophylaxis for mothers, breastfeeding prophylaxis for babies. Rarely was this issue highlighted – after an update on the HIV 012 where concerns about nevirapine resistance were raised if a woman ‘presents with another pregnancy’. Although we were pleased to hear a question asked about the implications of NNRTI resistance when HAART is introduced in this population, it didn’t cause much of a stir.

In addition to this, as is customary, two so-called ‘special’ patient groups were grouped together in sessions entitled Maternal-Foetal transmission of HIV-1 Implications for Care of HIV Affected Women and Children and HIV Infection in Women/Pediatrics (score out of 10 for the latter symposium – 7 presentations on children, 2 on pregnancy and one on women alone). Although not quite as puzzling as the recent Glasgow conference which combined women, children and IDVUs (why?) this was enough to provoke the comment from one activist who remarked ‘Why not men and mice?’

HAART use during pregnancy

Mary Glenn Fowler from the CDC (Center for Disease Control) presented an overview of the most recent research concerning perinatal transmission both in the US and internationally entitled Update on Prevention of Mother-to-Child HIV Transmission: US Successes, International Challenges [1]. She reported findings from recent studies showing that appropriate use of HAART for a woman’s own health during pregnancy has dramatically reduced transmission rates to less than 1%. Trends in antiretroviral use have changed greatly as demonstrated in data from PACTG 367 reporting 78% women using complex therapy in pregnancy in the US between January 1998 and May 2000 – including 42% using PI-containing regimens.

Elective C-section increasing

Dr Fowler reported C-section rates for HIV positive women in the US of 40-45% in the US in 1999 and 2000 and 78% in Europe as reported by The European Collaborative Study [2]. Several posters looked at this trend in detail, including the CDC study which showed how the US rate had risen from 19% in 1994 to 44% in 2000 and that it had doubled from 1997 [3]. The extent to which this intervention is used, particularly in Europe was surprising as no data to date have demonstrated any additional benefit to be gained from using elective C-section among women using potent antiretroviral therapy during pregnancy.

One report concluded that ‘Future studies are necessary to determine whether there is an incremental benefit of C-section delivery in women receiving prenatal ART with successful HIV suppression’ [4]. As transmission rates continue to be as low as less than 1%, regardless of mode of delivery, it seems unlikely that these data will be

Some groups still at risk

Despite great advances for the majority in the developed world, a small group of women still slip through the net. Dr Fowler explained that these included – late presenters with no prenatal care, women not perceived to be at risk and therefore not offered testing, women who were non-adherent although prescribed antiretrovirals and women experiencing treatment failure.

Reports from the MIRIAD study showed the feasibility of offering rapid testing at delivery to late presenters and substantial US funds ($10,000 per year) are being directed at programmes targeting states with high HIV prevalence.

Different picture in resource-poor countries

Dr Fowler went on to speak about the ever-growing contrast between north and south and how obstacles in the developing word still remain daunting. Adolescent infection is increasing disproportionally among girls in Africa. For example in Kisumu, Kenya 8% of 15 year old girls are HIV positive, increasing to 30% by age 17 and 33% by 19 years, whereas infection rates for teenage boys in the same region are still comparatively low (8.6% by age 19).

Because of HIV, a projected figure of 200 per 1000 young children in Zimbabwe will die before they are 5 years old. Among breast-feeding populations about one third of transmissions occur through this mode of transmission. Risk factors include – younger maternal age, lower parity, post-natal seroconversion, duration of breast-feeding, maternal CD4, viral load in breast milk, nipple lesions, mastisis and infant oral thrush. There appears to be a continued risk throughout the duration of breast-feeding but there is an indication that most transmission appears early [5,6]. Risk of transmission is about 0.7 per month in the infant’s first year and 0.3 per month in the second.

Findings from the HIVNet 012 trial showed 16% and 24% of infants infected in their first year in the nevirapine and AZT arms respectively [7]. Over twelve studies are currently investigating interventions to reduce breast-feeding transmission, these include infant prophylaxis during breastfeeding, passive immune therapy is being studied in Uganda, exclusive breastfeeding in South Africa (mixed feeding appears to by associated with the highest transmission rates [8]) and abrupt early weaning in Zambia. Despite being reminded of Dr Ruth Ndati’s chilling report in which breastfeeding HIV-positive women in Kenya were found to have a 3-fold increase in their mortality rate [9], and that uninfected children born to HIV positive mothers also have a higher infant mortality rate, no planned studies were reported that focus on the issue of maternal health.

Nevirapine resistance

A further update was presented from the HIVNET 012 trial, in which treatment naive Ugandan women received prophylaxis nevirapine to prevent vertical transmission, followed by a dose to the baby [10]. Dr Susan Eshleman first presented resistance data from this study at the 7th CROI a year ago [11]. Last year’s findings showed the K103N primary NNRTI mutation present in plasma of 20% of women studied, after only one dose of nevirapine. Further observations from this ongoing trial reported 45% of infected babies also having nevirapine resistance. In addition the mother-infant pairs were found to have differing genotypes – the Y181C being the most common mutation among the infected infants – suggesting that the babies’ resistance may be selected independently to the mothers’. Unsurprisingly the presence of resistance was most prevalent in samples from women with high viral load and low CD4 count at the time of labour. Although a question from the audience at this meeting raised the point ‘Are you possibly underestimating resistance, as you are not detecting it in women of low viral load?’

New findings presented also included the worrying report of one case of nevirapine resistant virus being transmitted from mother to baby via breastfeeding. In the context of ongoing and planned trials to prophylax both mothers and babies with nevirapine alone, this news should raise additional concerns. Dr Eshleman’s presentation concluded with the report that ‘nevirapine mutations appear to fade from detection in women and infants over time, but it is still unknown whether nevirapine would be effective when she presents with a subsequent pregnancyÉ’

In addition a poster reported genotypic resistance analysis in women participating in PACTG 316 with viral load above 400 copies/ml [12]. This study evaluated the effect of adding single-dose of 200mg nevirapine to standard antiretroviral therapy/prophylaxis to the mother at time of labour and 2mg/kg to the baby within 72 hours of birth vs.placebo, to further reduce transmission (see below). Plasma samples from 104 women were analysed both from delivery and 6 weeks postpartum. Of the 46 women who had received study drug, 5 (11%) had detectable mutations associated with nevirapine resistance after a single dose. Nucleoside and PI resistance mutations were also detected, mutations were most frequently observed in women with lower CD4 counts using PI-containing regimens.

Results from PACTG 316

Results from the PACTG 316 were presented for the first time as a late breaker at this meeting [13]. Describing the sample size hypothesis, principle investigator Dr Alejandro Dorenbaum explained that when they designed this study they had anticipated a 5% perinatal transmission rate. So to achieve 80% power to detect a 50% reduction (from 5% to 2.5%) they required a sample size of 2009 (1808 plus 10% loss rate). However due to the overall low transmission rate the trial was stopped on recommendation of their DSMB after 1267 mother and infant pairs had been randomised to receive nevirapine or placebo.

Of the participating women, none had used NNRTIs prior to enrolment. 1% received no treatment in addition to the nevirapine; 23% received AZT monotherapy; 28% received AZT and 3TC and 41% PI-containing combinations. 49% of women had viral loads below detection (less than 400copies/ml) at delivery. There was no significant difference between the nevirapine and placebo arms (9 and 8 transmissions respectively) and the overall rate of transmission was low – 1.5%.

This provoked some interesting reactions, including a mainstream journalist overheard in the pressroom calling his news desk with the news that ‘You know that wonder drug nevirapine? Well it doesn’t work!’ Against a background of fully suppressive HAART, any additional benefit to be gained by adding nevirapine at delivery would be unexpected, it is hoped that women in the developed world can expect to receive optimal therapy in this setting.

However with no treatment or mono or dual therapy and with a larger sample size, using nevirapine in this way would certainly be expected to add benefit to transmission rates, resistance not withstanding. Dr Dorenbaun concluded that ‘A woman should be treated in the best possible manner for her own disease (author’s note -a message that deserves emphasising) and additionally pregnant women are strongly recommended to initiate therapy to prevent transmission, independent of their disease process.’

Important factors in transmission

Several posters looked at factors associated with perinatal transmission. These included an additional report from the PACTG 316 looking at PI concentrations in cord blood [14]. Although the use of protease inhibitors is growing during pregnancy few data are available measuring PI concentrations after their use. In this study cord blood plasma samples were collected at delivery from 68 women using PI-containing regimens. 38 women received nelfinavir, 21 indinavir, 8 saquinavir, and 6 ritonavir (4 women received more than one protease inhibitor).

The last maternal dose was administered a median of 12.2 hours (range 0.67-40.9) before delivery for the 26 women for whom this information was available. PI concentrations were undetectable in 53 (78%) of the samples. Of the 15 women with detectable cord blood concentrations, 14 were using nelfinavir and 1 ritonavir. Cord blood PI concentrations were revealed to be extremely low after maternal treatment during pregnancy. The investigators suggested that this could reflect lack of dosing during delivery, poor maternal PI absorption and/or poor placental PI transfer, and that it is unlikely that maternal PI use provides post exposure prophylaxis to the baby immediately after birth. These findings also highlight a notable take home message to those caring for HIV-positive pregnant women ie the importance of remembering to take their meds during labour.

Seventy-five percent of infants (born to untreated mothers) exposed to maternal HIV infection are protected from in utero and perinatal transmission by mechanisms that are so far incompletely understood. When transmission does occur, non-syncytium inducing (NSI), M-tropic HIV-1 is most commonly involved. CCR5, a chemokine receptor, serves as a co-receptor for NSI/M-tropic HHIV-1 and may be a necessary co-receptor for sexual, parenteral and vertical transmission. Unlike adults, full term foetuses express a naive (CD45 RO negative, CCR5 negative) immunophenotype on their PBMC, which is not thought to be susceptible by NSI/M-tropic, CCR5-utilising virus.

Dr Karen Beckerman’s group examined the hypothesis that the bacterial infection, chorioamnionitis and preterm delivery could render a foetus susceptible to HIV infection via immune activation-associated induction of CCR5 expression by foetal PBMC [15]. Paired maternal and placental cord blood samples were collected from 12 pregnancies, obtained after delivery by elective C-section (n=3), vaginal delivery (n=3) or choirioamnionitis (n=6). Cells were then analysed for surface expression of CD4, CD25, CD45RO, CD69, CXCR4 and CCR5. They found that among foetuses experiencing no labour (C-section) or vaginal deliveries CD45 RO and CCR5 expression were considerably lower than in adult controls.

The investigators concluded that chorioamnionitis and preterm delivery are associated with significant activation of a foetal immunophenotype that predicts increased susceptibility to M-tropic/R5-utilising HIV-1. And they speculated that ‘The protection of foetuses experiencing no labour or normal labour, from HIV-1 infection may be associated with the naive, CCR5 negative immunophenotype observed in the peripheral immune compartment’. Since chorio generally affects 10-20% of deliveries Dr Beckerman recommended ‘Very active management of labour’.

Transmission has been reported to occur occasionally even at very low maternal viral loads. One poster reported findings from a large meta-analysis combining data from seven European and US studies [16]. Out of a total of 1,202 women there were 44 cases of vertical transmission where maternal viral load was less than 1000 copies/ml at the time of or close to delivery. Of these the rate was only 8 out of 834 (1%) for women receiving treatment during pregnancy and/or delivery as oppose to 36 out of 368 (9.8%) receiving no treatment. Although this study demonstrates a vast reduction in transmission rates (90%) conferred by the prophylaxis effect of treatment, more importantly it makes a strong case for use of fully suppressive therapy for pregnant women particularly at the time of delivery regardless of viral load.

Impact of pregnancy and menopause on CD4 count

Studies have shown that CD4 counts are higher and viral load levels lower in HIV-positive women than HIV-positive men. This is speculated to be due to levels of reproductive hormones, which differ by gender and change throughout a woman’s life. Dr Danisman and colleagues’ report investigated the impact of pregnancy and menopause on CD4 counts in 382 HIV-infected women with a known interval of seroconversion [17].

They found that CD4 count dropped during pregnancy but increased again to pre-pregnancy levels after delivery. Although their numbers of post menopausal women were small (17 at the time of analysis), this group had lower CD4 counts 3 years after seroconversion than premenopausal women (333 and 399 cells respectively, p=0.09 after controlling for age).

The investigators concluded that this effect ‘Émay be a result of changes in levels of reproductive hormones. Since levels of reproductive hormones also differ between men and women, this may explain gender differences in CD4 counts’.

“Hard drug use”

A WITS (Women and Infants Transmission Study) prospective study examined the effects of “hard” drugs on markers of HIV progression in a large group of 1148 women enrolled between 1989 and 1995 (and therefore pre-HAART) [18]. The women were interviewed about their drug use, which was self – reported and also measured using periodic urine toxicology tests. The correlation between self-report and urine toxicology was good and hard drug use was defined as cocaine, crack, heroin, methadone and any injection drug use. 40% of the group studied reported drug use and differences in viral load and CD4 count were compared to non- users.

No difference in mean CD4% were found at baseline between the two groups, but users were found to have a higher mean viral load. In multivariate, longitudinal analyses, hard drug use was not associated with higher averages of either CD4% or viral load. The investigators concluded that hard drug use ‘…did not appear to affect immunological or virological parameters in this cohort of HIV-infected women and may not exert a “co-factor” effect upon disease progression.’

Genital compartment issues

Dr Sherlock and colleagues from British Columbia evaluated the reproducibility of vaginal and cervical viral load measurements. The investigators wanted to ensure that these measurements are reproducible before using them to draw any conclusions [19]. A group of 14 HIV positive women were enrolled in a study of menstrual variation in viral load and followed over 2 menstrual cycles. Plasma, cervical and vaginal viral load samples were taken at initial and follow-up visits.

Their analysis of paired measurements showed good reproducibility for same site samples. The paired cervix/vagina measurements showed moderate correlation only; this was thought to be likely to be due to the high rate of negatives in those pairs. Standardised sample collection and modification of the Roche RNA extraction method allowed reproducible measurement of genital viral load.

With semen, drug-selected HIV-1 genotypes differing from those that are dominant in the blood plasma have been frequently observed. In women there are limited data on whether the genital tract may also be a compartment for differing drug-selected variants; Dr De Pasquale’s group investigated this phenomenon [20].

7 HIV-positive women on failing antiretroviral therapy and 3 drug naive women were studied; all had detectable viral loads in plasma and vagina or cervix. Resistance mutations selected by the failing drugs were found to be present in each compartment, 5 of the 7 women on failing therapy had different mutations in cervico-vaginal RNA which were not detected in plasma. In contrast sequences from the untreated women did not differ between compartments.

Several other studies investigated HIV compartment issues in women. A study from Puerto Rico conducted by Dr Yamamura and colleagues also revealed patterns suggesting that resistance mutations emerged differently in each compartment [21]. Another by Dr Wahl and colleagues revealed discordance between viral load levels in 3 compartments (oral, genital and peripheral blood) from 77 women with detectable viral load (out of a group of 114) [22]

HPV infections and cervical abnormalities

The only report on non-pregnant women that made it, as a slide session was a report concerning the impact of HAART on cervical intraepithelial neoplasia (CIN) presented by Dr Heard [23]. This French prospective study was initiated in 1993, 168 HIV-positive women with CIN were analysed, of the group 80 were defined as having high grade and 88 as low grade CIN. 86 women received PI-containing and 10 NNRTI-containing HAART.

After a median follow up of 17.7 months regression from low grade to normal was observed in 30 women, from high grade to normal in 6, and from high grade to low grade in 31 (n=67, 39.9%). Among women treated with HAART the regression rate was twice as high as it was for women without HAART. It was concluded that ‘HAART significantly increased reversion to normality or change to lower grade.’ They are currently conducting studies to further evaluate the impact of HAART on HPV infection.

Another report from Dr Luque and colleagues also reported a beneficial effect of antiretrovirals on HPV and that women receiving therapy were less likely to have abnormal Pap smears [24]. These reports describing the protective effect of HAART against pre-cancerous conditions further substantiate data presented last year by Dr Howard Minkoff and colleagues [25]

Finally a report from the WIHS highlighted a concern with regards to the reliability of using different testing methods to predict abnormal cytology [26]. Data were compared from 1370 HIV-positive and 224 HIV negative women who had undergone cervical cytology, colposcopy and biopsy where these were available. Biopsies were performed on 603 (44%) and positive women at least once during 1015 colposcopy visits and 145 (69%) of negative women over 116 visits.

The positive predictive value of abnormal cytology for CIN was 619/854 (72%) for positive women and39/65 (55%) for negative women and the positive predictive value of a positive colposcopy was 512/720 (71%) for positive women and 44/80 (55%) for negative women. The investigators caution that ‘The correlation between either cervical cytology or colposcopic impression and biopsy is poor. Liberal use of biopsy is essential for the proper management of women with abnormal smears’. Obviously this message is especially important for HIV-positive women given that it has been demonstrated that immune suppression by HIV speeds up progression of cervical abnormalities [27].

Conclusion

This is probably less of a conclusion and more of a wish list. In the developed world vertical transmission rates are down to virtually zero and we are able to use words like ‘elimination’ in this context. But in the developing world, women’s health is rarely addressed above and beyond their role in transmission of HIV. We support the comment from one report after the Durban conference, which stated that ‘Women’s advocates repeatedly express concern that women are being viewed as vessels for treatment and transmission, rather than people in need of treatment themselves’ [28]. And we would welcome news of plans for studies that also address maternal health.

In addition, although we applaud the work of women’s cohorts such as the WIHS, much information about the performance of anti-retroviral in women as compared to men is still unknown, and women still make up only 18% (at the higher end of the range) of subjects enrolled in clinical trials. We have often discussed the reasons for this but would like to make a final plea for better representation in trials of women overall. At a community meeting prior to the Chicago conference Julie Davids from ACT-UP Philadelphia reminded us, that this issue was not just about fairness or feminism or reflecting respective numbers in the epidemic but most importantly clinically ‘it’s about statistical significance’

References:

1. Fowler, Mary-Glenn, Update on Prevention of Mother-to-Child HIV Transmission: US Successes, International Challenges, 8th CROI, S12
2. Fiore S et al, Interventions to Reduce Vertical Transmission of HIV in Europe, 8th CROI, Abstract 697
3. Dominguez K et al, Increasing Trends in Cesarean Section in HIV-infected Mothers of Infants in the Pediatric Spectrum of HIV Disease Cohort, 8th CROI, Abstract 702
4. Peters V et al, Trends in Antiretroviral Therapies and Cesarean, Section Use for Perinatal HIV Prevention in New York City, 8th CROI, Abstract 703
5. Nduati R et al. Effect of Breastfeeding and Formula Feeding on Transmission of HIV-1: a randomised clinical trial. JAMA; 283: 1167-1174
6. Sullivan J., Perinatal Transmission: where do we go from here? 8th CROI, S15
7. Guay LA, Musoke P, Fleming T et al, Intrapartum and Neonatal Single-Dose Nevirapine compared with Zidovudine for Prevention of Mother-to-Child-Transmission of HIV-1 in Kampala: HIVNET 012 randomised trial. Lancet 1999;354:795-802
8. Coutsoudis A et al, Method of Feeding and Transmission of HIV-1 from Mothers to Children by 15 Months of Age; prospective cohort study from Durban. XIII International AIDS Conference, Abstract LbOr6
9. Nduati R et al. Impact of Breastfeeding on Maternal Mortality Among HIV-1 Infected Women. XIII International AIDS Conference, Abstract WeOrC495
10. Eshleman SH et al. Selection of Nevirapine Resistance in Ugandan Women and Infants Receiving NVP Prophylaxis to Prevent HIV-1 Vertical Transmission (HIVNET-012). 8th CROI, Abstract 516
11. Eshleman SH et al. Selection of the K103N Nevirapine Resistance Mutation in Ugandan Women Receiving NVP Prophylaxis to Prevent HIV-1 Vertical Transmission (HIVNET-066), 7th CROI , Abstract 658
12. Cunningham et al. Genotypic Resistance Analysis in Women Participating in PACTG 316 with HIV-1 RNA >400 Copies /ml. CK 8th CROI. Abstract 712
13. Dorenbaum A et al. Report of Results of PACTG 316: An International Phase III Trial of Standard Antiretroviral Prophylaxis for Prevention of Perinatal HIV Transmission. 8th CROI, Abstract LB7
14. Mirochnick M et al. Cord Blood Protease Inhibitor Concentrations in Infants Born to Mothers Receiving PIs. 8th CROI, Abstract 710.
15. Beckerman KP et al. Chorioamnionitis Induces Foetal Chemokine Receptor Expression. 8th CROI, Abstract 706.
16. Ioannidis JPA et al. Perinatal Transmission of HIV-1 from Pregnant Women with RNA Viral Load Less than 1000 copies/ml. 8th CROI. Abstract 516.
17. Thorpe L et al. The Impact of Pregnancy and Menopause on CD4 Counts in HIV-infected Women. 8th CROI. Abstract 205
18. Danisman F et al. The Effects of Hard Drug Use on Virological and Immunological Parameters of HIV-Infected Women. 8th CROI. Abstract 204.
19. Sherlock CH. Are Measurements of Genital HIV-1 Viral Load in Women Reproducible? 8th CROI. Abstract 241.
20. De Plasquale MP. Drug Selected HIV-1 Mutations Can Differ in Cervico-Vaginal and Blood Plasma RNA. 8th CROI. Abstract 446.
21. Yamamura Y et al. Resident HIV of Vaginal Compartment are Distinct From HIV in Both Drug Mutation Patterns and Viral Evolution. 8th CROI. Abstract 719.
22. Wahl S et al HIV-1 in the Oral, Genital and Peripheral Blood Compartments.. 8th CROI. Abstract 715.
23. Heard I et al. Impact of Highly Active Antiretroviral Therapy on Cervical Intraepithelial Neoplasia (CIN) in HIV-Seropositive Women. 8th CROI. Abstract 516.
24. Luque AE et al. Effect of Antiretroviral Therapy on Human Papillomavirus Infection and Disease Among HIV-Infected Women. 8th CROI. Abstract 724.
25. Minkoff H et al. Effect of Highly Active Antiretroviral Therapy on Cervical Cytologic Changes Associated with Oncogenic HPV among HIV-infected Women. 7th CROI. Abstract 674.
26. Massad LS et al. Correlating Pap Smear, Colposcopic Impression and Biopsy: Results from the WIHS. 8th CROI. Abstract 722.
27. Massad LS et al. Incidence and Progression of Cervical Squamous Lesions Among Women with HIV: Insights from the First 13,038 Pap Smears of the WIHS. 7th CROI. Abstract 675.
28. Carmen R, A Tapestry of Women’s Experiences. IAPAC, October 2000, Vol 6 No 10 p 293