Multivariate Cox's proportional hazards regression model adjusted for injecting drug use, age at conception, ethnicity, baseline CD4 cell count, gestation at initiation, and whether regimen was START or intended to continue showed baseline viral load impacted on the likelihood of achieving a viral load less than 50 and less than 1000 copies/ml by the time of delivery. Baseline viral load was also associated with the likelihood of achieving less than 50 copies/ml by 36 weeks gestation, the time when decisions around mode of delivery are often finalised (Table 3). The significant effect of increasing viral load on the probability of achieving an acceptable threshold remains.
Furthermore, assessment of the data by baseline viral load grouped into clinically defined strata and by gestation at initiation of HAART showed that when baseline viral load is less than 10 000 copies/ml, the gestational age at initiation of HAART is unlikely to affect success up to 26.3 weeks gestation; however, when viral load is more than 10 000 copies/ml, then deferring HAART past 20.4 weeks is significantly associated with a reduced likelihood of reaching the target by delivery (P = 0.011). When baseline viral load was more than 100 000 copies/ml, the likelihood of reaching a viral load of less than 50 copies/ml was low (37%: hazard ratio 0.31) even if HAART was started before 20 weeks gestation (Table 4). Of the 43 women who commenced HAART with a viral load of more than 100 000 copies/ml, those who achieved a viral load less than 50 by delivery initiated HAART at median of 17.7 weeks, compared with those not reaching an undetectable level of 22.5 weeks. When baseline viral load grouped into unbiased quartile grouped strata was analysed, 87% of the women in the lower three quartiles (<32 640 copies/ml) achieved less than 50 copies/ml by the time of delivery compared with only 46% in the upper quartile (P < 0.01).
Finally, multivariate Cox's proportional hazards regression model showed the choice of therapy was significantly associated with the proportion of women reaching a viral load less than 50 copies/ml delivery. Overall boosted protease inhibitor therapy (n = 246) resulted in 80% less than 50 HIV RNA copies/ml compared with 72% of those using NNRTI-based therapy (n = 129). The hazard ratio for NNRTI-based HAART compared with protease inhibitor-based HAART was 0.7 (95% CI 0.52–0.94; P = 0.016). In all but one case, nevirapine was the chosen NNRTI. Lopinavir/ritonavir was the most common boosted protease inhibitor prescribed (n = 171, 70%), followed by saquinavir/ritonavir (n = 67, 27%) and atazanavir/ritonavir (n = 8, 3%). In those women (n = 96) starting HAART in the earliest quartile of gestation (prior to 20.4 weeks), 49 commenced protease inhibitor-based therapy and 47 NNRTI-based therapy.
The results from our study indicate that, for women with a viral load of more than 10 000 copies/ml and especially for those with a viral load of more than 100 000 copies/ml, the probability of achieving targets of either less than 50 or less than 1000 copies/ml by the time of delivery is compromised by delaying initiation of START beyond 20.4 weeks gestation. Women with viral loads less than 10 000 copies/ml should commence START by 26.3 weeks to optimize the probability of successful therapy. Viral load at initiation of HAART was the most important determinant of achieving an undetectable viral load by delivery, and the worst outcomes were seen in the uppermost viral load quartile of more than 32 000 copies/ml. British guidelines currently recommend commencing START between 20 and 28 weeks  and the median starting gestational age in the UK and Ireland cohort 2000–2006 was 25.6 weeks (IQR 22.4–28.9) . Other national and international guidelines provide similar recommendations for the initiation of START. The German–Austrian guidelines 2008 update recommend 28 weeks . The European AIDS Clinical Society recommend week 28 , or earlier if the viral load is high, although how early is not specified. The World Health Organisation recommends commencement of START, where available, by week 28, whilst recognising possible benefits of reduced MTCT with earlier treatment . These data may be particularly relevant when the availability of PLCS is limited. Our study suggests that in terms of achieving an undetectable viral load by delivery, several guidelines may not afford women commencing START enough time to reach this goal. In contrast, the USA DHHS 2010 guidelines recommend initiating START at between 10 and 12 weeks gestation regardless of HIV viral load .
The European Collaborative Study group examined the time to an undetectable HIV viral load in 240 pregnant women starting HAART at median of 23 weeks gestation (IQR 18–27) . Overall 73% of women achieved a viral load less than LLD, but this could have been less than 50 or less than 400 copies/ml depending on the assay used. They also found that baseline viral load was a predictor of the speed of achieving viral suppression by delivery. Up to 50% of those with a baseline viral load more than 100 000 copies/ml and 70% of those with a baseline viral load more than 10 000 copies/ml achieved a viral load less than 400 copies/ml by delivery. Their data did not find a difference between NNRTI and protease inhibitor regimens in terms of successful treatment, but did find that the time to undetectable viral load was shorter in the NNRTI group, particularly if the baseline viral load was more than 10 000 copies/ml. Our data showed slightly more success overall in the protease inhibitor arm. The major differences between the ECS and our data are that all our patients on a protease inhibitor regimen were taking a ritonavir-boosted protease inhibitor, and the LLD of our assays was uniformly 50 copies/ml or less.
A more recent analysis of the WITS cohort examined outcomes from 1998 to 2005 in 630 HIV-1-infected women commencing HAART in pregnancy . Again, the LLD of most assays used was less than 400 copies/ml, and two thirds of the cohort were treated with the unboosted protease inhibitor nelfinavir, which is now difficult to access and not routinely used. Overall 48% of women with detectable virus initiating HAART during pregnancy still had a detectable (>400 copies/ml) viral load at delivery, although this is not stratified for gestational age at initiation of HAART. They found an increase in the likelihood of detectable HIV at delivery with increasing baseline viral load, with an adjusted hazard ratio of 1.35–1.52 for every log10 above 400 copies/ml. An association with the type of HAART regimen was not supported in multivariate analysis.
Limitations of our study are that it is both retrospective and observational, and we were unable to obtain sufficient reliable data on drug adherence and therapeutic drug levels. As with any nonrandomized study, we cannot exclude confounding from factors not considered in our analysis. Furthermore, for some stratifications of viral load and gestation at initiation of HAART, the absolute numbers in some subcategories are relatively small, for example, only 96 patients commenced HAART prior to 20.4 weeks gestation. Nonetheless, this is one of the largest studies investigating the likelihood of achieving a viral load less than LLD by the point of delivery, and has the advantage of using contemporary regimens and viral load assays.
With the increasing trend to advocate for vaginal delivery in HIV-positive pregnant women, regardless of the viral load cut-off used for this recommendation, the data from this study and both the ECS and WITS cohorts suggest that a significant proportion of women do not reach these targets, especially if the baseline viral load is high. Our study further defines these groups and provides information detailing the effect of gestation at initiation of HAART.
HAART commenced during pregnancy, which is most effective as part of widespread HIV antenatal screening at a population level, has markedly reduced the rates of MTCT of HIV when freely available. Increasingly, with wider worldwide access to HAART, many women will be using this as part of a START regimen, but often in settings when a caesarean section is not a viable option. Whether adhering to USA DHHS guidelines , which do not attribute further risk reduction to caesarean section if the viral load on HAART is less than 1000 copies/ml, or the British HIV Association Guidelines , which currently recommend a more stringent cut off of less than 50 copies/ml, it is clear that HAART must be commenced with sufficient time to reach these goals if vaginal delivery is likely or intended.
We conclude that women with a viral load of more than 10 000 copies/ml should commence HAART by 20.4 weeks, and women with a viral load of more than 100 000 should start HAART without delay. In women with a viral load less than 10 000 copies/ml an undetectable viral load should be achieved by delivery provided START is not deferred beyond 26 weeks gestation. However, this should be balanced against the possible increased risk of in-utero transmission with a delayed start, or the risk of transmission should premature rupture of membranes occur prior to the commencement of therapy in a pregnancy where the gestational age of the fetus is compatible with survival. Current UK and other guidelines for when to commence START may therefore limit the chance of vaginal delivery. If the current trend for vaginal delivery is maintained, we suggest guidelines take these data into account when recommending the timing of START.
Conflicts of interest
Comprehensive Biomedical Research Centre support at Imperial College Healthcare NHS Trust (G.P.T.).
P.J.R. has received travel and conference expenses from MSD, BMS, ViiV, and Tibotec. D.H. has received honorarium for presentations, lectures, and travel from BMS and Janssen. Y.G. has received honoraria for lectures and presentations by Gilead, Janssen, MSD, BMS, and ViiV. J.A. has received consultancy fees from Gilead, Abbott, BMS, and Johnson and Johnson. G.P.T. has received consultancy fees from the WHO, and a research grant from Abbott. A.d.R. has received consultancy fees from BMS, and honoraria for lectures from BMS, Janssen, and MSD. C.N. has received conference registration from ViiV. P.K., U.H., and S.M. have no conflicts of interest to declare.
1. Maternal viral load and vertical transmission of HIV-1: an important factor but not the only one. The European Collaborative Study.AIDS
2. Garcia PM, Kalish LA, Pitt J, et al. Maternal levels of plasma human immunodeficiency virus type 1 RNA and the risk of perinatal transmission
. N Engl J Med
3. Contopoulos-Ioannidis DG, Ioannidis JP. Maternal cell-free viremia in the natural history of perinatal HIV-1 transmission: a meta-analysis
. J Acquir Immune Defic Syndr Hum Retrovirol
4. Connor EM, Sperling RS, Gelber R, et al. Reduction of maternal–infant transmission of human immunodeficiency virus type 1 with zidovudine treatment. Pediatric AIDS Clinical Trials Group Protocol 076 Study Group
. N Engl J Med
5. The European Mode of Delivery Collaboration. Elective caesarean-section versus vaginal delivery in prevention of vertical HIV-1 transmission: a randomised clinical trial.Lancet
6. Boer K, England K, Godfried MH, Thorne C. Mode of delivery in HIV-infected pregnant women and prevention of mother-to-child transmission: changing practices in Western Europe
. HIV Med
7. European Collaborative Study. Mother-to-child transmission of HIV infection in the era of highly active antiretroviral therapy
. Clin Infect Dis
8. Townsend CL, Cortina-Borja M, Peckham CS, de Ruiter A, Lyall H, Tookey PA. Low rates of mother-to-child transmission of HIV following effective pregnancy interventions in the United Kingdom and Ireland, 2000–2006
9. Hawkins D, Blott M, Clayden P, et al. Guidelines for the management of HIV infection in pregnant women and the prevention of mother-to-child transmission of HIV
. HIV Med
2005; 6 (Suppl 2):107–148.
10. Lyall E, Blott M, de Ruiter A, et al. Guidelines for the management of HIV infection in pregnant women and the prevention of mother-to-child transmission
. HIV Med
11. Panel on Treatment
-Infected Pregnant Women and Prevention of Perinatal Transmission. Recommendations for Use of Antiretroviral Drugs in Pregnant HIV-1-Infected Women for Maternal Health and Interventions to Reduce Perinatal HIV Transmission in the United States
; 14 September 2011. pp. 1–207. http://aidsinfo.nih.gov/ContentFiles/PerinatalGL.pdf
. [Accessed 12 February 2012]
12. Buchholz B, Beichert M, Marcus U, et al. German-Austrian recommendations for HIV1-therapy in pregnancy and in HIV1-exposed newborn, update 2008
. Eur J Med Res
13. de Ruiter A, Mercey D, Anderson J, et al. British HIV Association and Children's HIV Association guidelines for the management of HIV infection in pregnant women 2008
. HIV Med
16. Antiretroviral Pregnancy
Registry Steering Committee. Antiretroviral Pregnancy
Registry International Interim Report for 1 January 1989 through 31 January 2011. Wilmington, NC: Registry Coordinating Center; 2011. www.APRegistry.com
. [Accessed 1 November 2011]
17. Tubiana R, Chenadec J, Rouzioux C, et al. Factors associated with mother-to-child transmission of HIV-1 despite a maternal viral load <500 copies/ml at delivery: A case-control study nested in the French Perinatal Cohort (EPF-ANRS CO1)
. Clin Infect Dis
18. Lennox JL, DeJesus E, Lazzarin A, et al. Safety and efficacy of raltegravir-based versus efavirenz-based combination therapy in treatment-naive patients with HIV-1 infection: a multicentre, double-blind randomised controlled trial
19. Patel D, Cortina-Borja M, Thorne C, Newell ML. Time to undetectable viral load after highly active antiretroviral therapy initiation among HIV-infected pregnant women
. Clin Infect Dis
20. Katz IT, Shapiro R, Li D, et al. Risk factors for detectable HIV-1 RNA at delivery among women receiving highly active antiretroviral therapy in the women and infants transmission study
. J Acquir Immune Defic Syndr
Keywords:© 2012 Lippincott Williams & Wilkins, Inc.
HAART; HIV; mother-to-child transmission; pregnancy; treatment; viral load