HIV-infected women on antiretroviral treatment have increased mortality during pregnant and postpartum periods : AIDS

Secondary Logo

Journal Logo

Supplement Articles

HIV-infected women on antiretroviral treatment have increased mortality during pregnant and postpartum periods

Matthews, Lynn T.a,b; Kaida, Angelac; Kanters, Stevene; Byakwagamd, Helend,f; Mocello, A. Raind; Muzoora, Conradf; Kembabazi, Annetf; Haberer, Jessica E.g; Martin, Jeffrey N.d,h; Bangsberg, David R.a,f; Hunt, Peter W.h

Author Information
doi: 10.1097/QAD.0000000000000040
  • Free

Abstract

Introduction

HIV-infected women have a higher risk of maternal mortality compared with women without HIV [1–4], with a recent meta-analysis reporting an eight-fold increased risk of death during pregnancy or postpartum periods [5]. In 2011, there were an estimated 56 100 HIV-related maternal deaths, accounting for approximately 20% of maternal deaths worldwide [1]. HIV infection has been principally associated with indirect causes of maternal death such as increased susceptibility to opportunistic infections during pregnancy and the postpartum period, particularly among women without access to ART [2,4,6–10].

Among women living with HIV, several studies have investigated whether pregnancy confers an independent risk of mortality. A meta-analysis of studies conducted among women not taking ART suggested increased odds of death [aOR1.8 (95% CI 0.99, 3.30)] and HIV disease progression [aOR1.41 (95% CI 0.85, 2.33)] among pregnant HIV-infected women compared with nonpregnant HIV-infected women, with higher risks among women in resource-limited countries [11].

Whether pregnancy remains independently associated with an increased risk of death among HIV-infected women on ART is not known. The few studies evaluating crude mortality rates or proportion of deaths among HIV-infected women on ART show no effect of pregnancy on mortality risk [12,13], or in some cases, a protective effect (although this was limited to women with CD4 cell count between 200 and 500 cells/μl; no difference was observed between women with CD4 cell count below 200 cells/μl) [14]. The studies reporting no effect had high (≥20%) losses-to-follow-up, which might have led to underestimation of maternal mortality. In addition, women who are biologically capable of pregnancy may be healthier than women who cannot get pregnant [15,16]. Thus, comparing overall mortality of HIV-infected women with or without pregnancy without rigorously adjusting for disease stage may underestimate pregnancy-related mortality. Moreover, comparing mortality rates without accounting for the time-limited effects of pregnancy may dilute time-specific effects of pregnancy on mortality.

We assessed the impact of being pregnant or up to 1 year postpartum on mortality among HIV-infected Ugandan women initiating ART in a cohort study with a high level of retention and vital status ascertainment. The cohort is limited by sample size but strengthened by detailed follow-up to allow for classification of women as pregnant or postpartum, alive or dead. Understanding whether pregnancy affects mortality risk among HIV-infected women on ART is critical to optimizing HIV treatment and reproductive health programming for women living with HIV, particularly in settings with high baseline maternal mortality.

Methods

Setting

The Mbarara District of Uganda is a primarily rural setting located approximately 265 km southwest of the Ugandan capital city of Kampala. Regional adult HIV prevalence is estimated at 10% [17]. The Mbarara University HIV clinic offers comprehensive HIV care services, including ART, at no cost to patients, provided through the Ugandan Ministry of Health with support from the President's Emergency Plan for AIDS Relief (PEPFAR), the Global Fund, and the Family Treatment Fund [18].

Study participants

Study participants were sampled from the Uganda AIDS Rural Treatment Outcomes (UARTO) cohort study of over 700 HIV-infected adults initiating their first ART regimen at the Mbarara University HIV clinic. Participants have quarterly study visits with structured interviews and real-time laboratory testing (CD4+ T-cell count and plasma HIV-1 RNA level). For this analysis of pregnancy and mortality, we restricted the sample to 18–49 year old women enrolled from June 2005 and followed through September 2011.

Measurements

The primary outcome was death as assessed through active vital status ascertainment. When cohort participants miss a scheduled study visit, study staff attempt to reach the participant via phone or home visit. If contact is not successful within 6 months, study staff attempt to ascertain participant vital status through review of records in the referring HIV clinic and communication with the participant's family. For participants who die, cause of death is investigated through communication with families and caregivers and, when available, review of clinic and hospital records.

The primary predictor variable was pregnancy status, which we classified as being pregnant or up to 1 year postpartum (‘pregnancy-related’), or being neither pregnant nor postpartum (‘nonpregnancy-related’). Periods of pregnancy were defined based on self-report in structured quarterly interviews. Pregnancy start was considered to be the date pregnancy was first reported and pregnancy end was the subsequent date at which women reported no longer being pregnant. For eight women who reported a live birth and for whom the period of pregnancy based on the above calculation was less than 7 months (and in one case >11 months), live birth date was used to back-calculate a start date to account for a 9-month gestation. The postpartum period was defined as the period from end of pregnancy up until 12 months after the pregnancy outcome, including livebirth, termination/stillbirth/miscarriage, or no longer pregnant without further specification of the pregnancy outcome. Women entering the study reported whether they were currently pregnant but did not report on dates of pre-enrollment pregnancy or postpartum status. Thus, no women were counted as postpartum at study entry. Due to quarterly follow-up, there was insufficient date precision to reliably distinguish between deaths occurring within the first 42 postpartum days (early postpartum deaths) and those occurring between 42 and 365 days postpartum (late postpartum deaths) [19]. The postpartum period was therefore defined as the end of pregnancy up until 12 months after pregnancy outcome, consistent with the WHO definition of late maternal mortality [19].

Analysis

Baseline characteristics of women with prevalent or incident pregnancies were compared with women with no pregnancies using Wilcoxon rank sum test for continuous variables and Fisher's exact test for categorical variables. Crude mortality rates were calculated using person-time methods and are reported as the number of deaths per 100 person years (PY) of follow-up. We compared crude mortality rates using confidence intervals for rate ratios constructed using the Poisson distribution and test-based methods [20]. Loss to follow-up was defined as having unknown vital status six or more months after the last time the participant was known to be alive.

We modeled mortality using Cox proportional hazards regression with time-dependent covariates to assess the effect of being pregnant or postpartum compared with nonpregnancy-related follow-up. Time-updated predictor variables included age, CD4+ T-cell count, and viral load suppression (HIV-1 RNA < 400 log10copies/ml). We evaluated time on ART as a potential effect modifier of the relationship between pregnancy and mortality with the hypothesis that the effect of pregnancy-related periods on mortality might be strongest during early ART. An Akaike information criterion approach was used to determine the inclusion of time on ART as an effect modifier in the model. Women who had a gap of at least 6 months in cohort follow-up (i.e. women who left the study for extended periods of time, but returned) were censored to protect against pregnancy status misclassification and allowed back into the analysis upon return to the study.

Sensitivity analyses

There were four deaths occurring more than 6 months after last study visit (range 7–39 months) in women who were neither pregnant nor postpartum at last visit. As pregnancy could not be excluded prior to death in these cases, they were censored from the primary analysis. We completed a sensitivity analysis in which these women contributed to nonpregnancy-related follow-up.

We also completed a sensitivity analysis to evaluate mortality during pregnancy and 90-day postpartum time periods, compared with nonpregnant time periods. The 90-day postpartum time-period allows a narrower window than 1 year postpartum and is feasible given quarterly follow-up of participants [15].

Data were analyzed with SAS version 9.3 (Carey, North Carolina, USA).

Ethics

All procedures were approved by institutional review boards at Mbarara University of Science Technology (Mbarara, Uganda), Massachusetts General Hospital/Partners Healthcare (Boston, USA), Simon Fraser University (Burnaby, Canada), and the University of California, San Francisco.

Results

Among 354 women contributing 1215 person-years of follow-up, pre-ART initiation median age was 33 years (interquartile range (IQR) 27–37), CD4+ T-cell count was 142 (IQR 82–213) cells/uL, plasma HIV-1 RNA was log10 4.96 (IQR 4.50–5.46) copies/ml, and women had three (IQR 2–5) children (Table 1). Over the follow-up period, 109 (31%) of women experienced at least one pregnancy with most pregnancies in the first 3 years (Fig. 1). Women with a pregnancy at or subsequent to ART initiation were younger (median age 29 vs. 35 years, P < 0.001) with a higher median CD4+ T-cell count than those without pregnancy (162 vs.133 cells/μl, P = 0.019). Loss to follow-up among all women was low (3 and 7% at years 1 and 5) and did not differ among women with and without pregnancy at the time of last observation (7 vs. 8%). We applied interval censoring to 28 women (7.9%) who had extended gaps between visits, in which pregnancy status could not be ascertained. Of these, two had two interval censorings and one had three interval censorings [median gap 11(IQR 11–18) months].

T1-12
Table 1:
Baseline characteristics.
F1-12
Fig. 1:
Proportion of women pregnant or postpartum over time.As postpartum status was not captured at enrollment, none of the women entered the study in a postpartum state. For this reason, the postpartum line begins at 0.5 years. The low number of women at risk at 5 years is due to staggered enrollment: loss to follow-up was 7% at 5 years.

There were a total of 21 deaths in the cohort, five during pregnancy or postpartum follow-up periods and 16 during nonpregnancy-related follow-up periods for an overall mortality rate of 1.73/100 person-years. During the first year of ART, crude mortality rates were higher for pregnancy-related follow-up periods (12.57/100 person-years) than for nonpregnancy-related follow-up periods (3.53/100 person-years), representing an unadjusted mortality rate ratio of 3.56 (95% CI: 0.97–11.07). After the first year of ART, mortality rates declined both for pregnancy-related periods (0.82/100 person-years) and for nonpregnancy-related periods (0.82/100 person-years), and there was no evidence for an increase in pregnancy-related mortality in unadjusted analyses (RR: 0.99, 95% CI: 0.04–6.71).

After adjusting for time-updated age, CD4+ T-cell count, and plasma HIV-1 RNA level (< vs. ≥400 copies/ml), women had an increased hazard of death during pregnancy-related periods of follow-up. This risk was modified by time on ART with risk highest at treatment initiation (aHR: 21.48, 95% CI: 3.73–123.51), decreasing to 13.44 (95% CI 3.28–55.11) at 4 months, 8.28 (95% CI 2.38–28.88) at 8 months, 5.18 (95% CI: 1.36–19.71) at 1 year, and 1.25 (0.10–15.58) at 2 years of ART (Fig. 2).

F2-12
Fig. 2:
Adjusted hazards ratio of death associated with pregnancy and the postpartum period modified by time on ART (95% CI shown with dashed red lines).

Timing and cause of death

Although cause of death was investigated for all participants, definitive cause of death is not known. The qualitative details are therefore limited, but may be informative. Among five women who died during pregnancy-related follow-up periods, four were pregnant at ART initiation (Fig. 3). One death occurred during pregnancy. This participant reported pregnancy 3 months before ART initiation and died 6 months after report of pregnancy (after 3 months on ART). At a clinic visit 3 weeks prior to death, she was profoundly anemic (Hb 5 gm/dl) and treated for a presumed urinary tract infection. She had not had a CD4 cell count or HIV-1 RNA testing since initiating ART.

F3-12
Fig. 3:
Timeline of pregnancy, death, and cause of death data.COD, suspected cause of death; EDD, estimated date of delivery; Hb, hemoglobin; KS, Kaposi's sarcoma; TB, mycobacterium tuberculosis.

The remaining four deaths occurred in the postpartum period, between 1 and 5 months after live birth. Two women died during hospitalization and the suspected causes of death were opportunistic infection: one had smear-positive pulmonary tuberculosis and congestive heart failure; one had suspected cryptococcal meningitis. While both of these women continued to have CD4+ T-cell count less than 200 cells/μl prior to death, both had plasma HIV-1 RNA levels less than 400 copies/ml. Of the two additional women who died during postpartum follow-up, cause of death was not known.

Cause of death was unknown for 14 out of 16 women who died during nonpregnancy-related follow-up periods. There was a wide variability in timing of death with median time 8.5 [IQR: 2.6–22.3] months after ART initiation.

Sensitivity analyses

When we included the four deaths among women that occurred more than 6 months after last study visit (range 7–39 months) as deaths occurring during nonpregnancy periods (as they were neither pregnant nor postpartum at last study visit), the adjusted hazard ratio for mortality during pregnancy-related follow-up remained high with aHR at ART initiation 20.43 (95% CI: 3.62, 115.43); 13.11 (95% CI 3.20–53.75) at 4 months, 8.30 (95% CI 2.39–28.86) at 8 months, 5.32 (95% CI: 1.44, 19.70) at 1 year, and 1.38 at 2 years (95% CI: 0.13, 15.09) after ART initiation.

Using 90-days postpartum as the cut-off for pregnancy-related follow-up, we observed a similar association between pregnancy-related follow-up and mortality with aHR 21.00 at treatment initiation (95% CI: 3.56–124.04), 13.46 (95% CI: 3.26–55.56) at 4 months, 8.51 (95% CI: 2.45–29.52) at 8 months, 5.45 (95% CI: 1.43–20.80) at 1 year, and 1.41 (95% CI: 0.11, 18.88) at 2 years after ART initiation.

Discussion

HIV is a major risk factor for maternal mortality in resource-limited settings, but the impact of pregnancy on mortality among HIV-infected women initiating ART has remained unclear. These data suggest that the combined pregnancy and postpartum period is an independent risk factor for death among HIV-infected women initiating ART in southwestern Uganda. The pregnancy-associated risk of mortality was greatest in the first year of ART with a steady decline that was no longer significant after the second year on treatment.

The level of maternal mortality observed in our study of HIV-infected women initiating ART is much higher than would be expected in the general Ugandan population, where baseline maternal mortality is considered high by international standards [19]. We observed five deaths/78 livebirths or 6410 deaths/100 000 livebirths. This is 18 times the estimated maternal mortality rate for Uganda: 352 (215–558) deaths/100 000 livebirths [3]. This is also higher than mortality rates ranging from less than 1 to 2% reported in recent perinatal transmission studies in sub-Saharan Africa where women initiated ART during the third trimester and continued until breastfeeding ceased [21–23]. However, median CD4 cell count at ART initiation in these studies was relatively high (median 336–403 cells/ul), so these women were likely at lower risk for immunodeficiency-related complications.

Several recent studies have evaluated the relationship between pregnancy and mortality in larger cohorts of HIV-infected women on ART without treating pregnancy as a time-dependent predictor. Westreich et al. [12] showed no difference in the hazards of death among women with incident pregnancy compared with women without pregnancy in a Johannesburg cohort of 7534 women over a median follow-up of 2.1 years (aHR 0.84, 95% CI 0.44, 1.6). Kaplan et al.[24] found no difference in crude mortality rates during 5 years of follow-up among 2131 ART naive women referred for ART including 318 women with a prevalent pregnancy. In the IeDEA South African network of nearly 30 000 women, women with pregnancy at ART initiation were less likely to die than women who were not pregnant during the first year of follow-up [13]. Similarly, our data do not reveal a statistically significant difference in overall unadjusted mortality risk among women with and without pregnancy over the follow-up period. Rather, the excess mortality risk is observed when the pregnant and postpartum state is treated as a time-dependent risk factor. Use of the time-dependent analysis acknowledges that time spent during pregnancy induces physiologic changes that may affect HIV disease progression (including relative immunodeficiency and susceptibility to infection) in a time-limited fashion and that women move into and out of pregnancy states and thus contribute to time at risk in both denominators. Another limitation of these larger cohort studies is high loss to follow-up, which may lead to underestimation of pregnancy-associated mortality. Some of these studies used inverse probability of censoring weighted methods to correct mortality estimates for differential losses to follow-up, but these methods may not fully protect against biased inferences [25].

The majority of maternal deaths in our study occurred within 7 months of ART initiation among women whose CD4 cell counts remained below 200 cells/μl. The pregnancy-associated risk of mortality declined with increasing ART exposure. It has long been recognized that mortality rates are particularly high in the first few months after ART initiation among HIV-infected individuals in resource-limited settings, particularly those with low pre-ART CD4 cell counts [26,27]. This effect could be a consequence of persistent immunodeficiency or an increased risk of immune reconstitution inflammatory syndromes (IRIS). Indeed, three of the four postpartum deaths in our study may have been related to immunodeficiency or IRIS: one woman had smear-positive pulmonary TB and congestive heart failure, one had suspected cryptococcal meningitis, and one had Kaposis sarcoma lesions at her last clinic visit. The relative immunodeficiency of pregnancy and the postpartum period (i.e. via mechanisms like postpartum induction of the immunosuppressive enzyme indoleamine 2,3-dioxygensase-1 [28,29]), could potentially explain an increased risk of immunodeficiency-related complications. Alternatively, the postpartum period may accentuate the risk of IRIS as the relative immunosuppression of pregnancy is reversed [30]. Even in the absence of HIV infection, the postpartum period may increase the risk of certain infections like tuberculosis, either through persistent immunodeficiency or IRIS [30,31]. These effects may be more pronounced in HIV-infected women with advanced immunodeficiency at ART initiation. In a recent analysis of data from six African sites with 636, 213 person-years of observation between 1990 and 2012, mortality among HIV-infected women who were not pregnant or postpartum fell during the post-ART compared with the pre-ART era (mortality rate ratio 0.42, P < 0.0001), however there was no significant change in the mortality risk among women who were pregnant or postpartum (mortality rate ratio 0.7, P = .205) [32]. These data are limited by inability to assess which women were on ART, but the findings may be consistent with our data showing persistently high mortality risk during pregnancy and postpartum periods while on ART.

The decline in pregnancy-associated risk of death with increasing duration of ART observed in our study is also noteworthy. As immune function recovers with ART, the relatively subtle immunologic effects of pregnancy may contribute less to mortality. A decline in pregnancy-associated mortality risk with time on ART was recently reported from other cohorts of HIV-infected women on ART in Malawi and Mozambique [33].

Strengths of this study include long duration of observation, high levels of vital status ascertainment, and cause of death data. There are several limitations to this study. First, the association between pregnancy and mortality in our study was driven by just five deaths among pregnant or postpartum women: these results need to be confirmed in larger studies with active vital status ascertainment. Second, we relied on self-report of pregnancy. Given restrictions on pregnancy termination and cultural norms in Uganda, women are often reluctant to report early pregnancy, thus this study likely underestimates pregnancy and therefore biases our results in the direction of the null hypothesis. Postpartum status was not captured at enrollment, thus only women who became postpartum during study follow-up contributed to postpartum follow-up time. This likely underestimates postpartum periods of follow-up and biases our results in the direction of the null hypothesis. In addition, given important differences between women with prevalent pregnancies and those without [12], we would prefer to complete separate analyses for women with prevalent and women with incident pregnancies. Due to the relatively small sample size in this cohort, we were not able to separate these analyses.

While our finding of an increased risk of mortality during pregnancy and the postpartum period among HIV-infected women initiating ART at advanced disease stages needs to be confirmed in other studies, this may represent an important finding. Pregnant HIV-infected women are appropriately initiated on ART for the health of the woman and to reduce perinatal transmission. However, if pregnancy increases women's mortality risk, these women may merit closer monitoring (particularly given high loss to follow-up in this group [13]) and further study to understand the mechanisms increasing risk. In addition, in longitudinal studies, a large proportion of incident pregnancies occur proximal to ART initiation [34,35]. Whether increased pregnancy incidence after ART is a result of biological (e.g. improved fecundity) or behavioral change (e.g. increased sexual drive and/or fertility intentions with restored health) is not well understood but is likely due to a combination of factors [35–39]. Contraception uptake among HIV-infected women is low in Uganda (and other settings) [40,41]. These findings provide further impetus for earlier diagnosis of HIV and initiation of ART for women with pregnancy or risk factors for pregnancy (e.g. fertility intentions, partner fertility intentions, younger age) [34,35,39,42–45], promotion of contraception for HIV-infected women proximate to ART initiation, and careful monitoring during the postpartum period.

Acknowledgements

The authors would like to thank study participants and the research team for their contributions to this study. L.T.M., A.K., S.K., H.B., J.N.M., D.R.B., and P.W.H. contributed to study conception, design, analysis, article drafting, and revision. A.R.M., C.M., A.K., J.E.H. contributed to study design, data acquisition, and article revision. All authors approved the final article.

Conflicts of interest

The authors report no conflicts of interest.

This project was supported by R01 (Bangsberg MH54907), R56 (Hunt AI100765), R21 (Kaida HD069194), NIAID P30 027763 funds, and The Sullivan Family Foundation. Lynn Matthews and Jessica Haberer are supported by K23 awards (NIMH 095655 and 087228). David Bangsberg is supported by a K24 award (NIMH 87227). Peter Hunt is supported by Doris Duke Clinical Scientist Development Award (2008047). The content of this manuscript is the responsibility of the authors and does not necessarily represent the official views of the NIH.

References

1. Lozano R, Wang H, Foreman KJ, Rajaratnam JK, Naghavi M, Marcus JR, et al. Progress towards Millennium Development Goals 4 and 5 on maternal and child mortality: an updated systematic analysis. Lancet 2011; 378:1139–1165.
2. Khan M, Pillay T, Moodley JM, Connolly CA. and Durban Perinatal TB/HIV Study GroupMaternal mortality associated with tuberculosis-HIV-1 co-infection in Durban, South Africa. AIDS 2001; 15:1857–1863.
3. Hogan MC, Foreman KJ, Naghavi M, Ahn SY, Wang M, Makela SM, et al. Maternal mortality for 181 countries 1980–2008: a systematic analysis of progress towards Millennium Development Goal 5. Lancet 2010; 375:1609–1623.
4. Black V, Brooke S, Chersich MF. Effect of human immunodeficiency virus treatment on maternal mortality at a tertiary center in South Africa: a 5-year audit. Obstet Gynecol 2009; 114 (2 Pt 1):292–299.
5. Calvert C, Ronsmans PC. The contribution of HIV to pregnancy-related mortality: a systematic review and meta-analysis. AIDS 2013; [Epub ahead of print].
6. Zvandasara P, Hargrove JW, Ntozini R, Chidawanyika H, Mutasa K, Iliff PJ, et al. Mortality and morbidity among postpartum HIV-positive and HIV-negative women in Zimbabwe: risk factors, causes, and impact of single-dose postpartum vitamin A supplementation. J Acquir Immune Defic Syndr 2006; 43:107–116.
7. Nduati R, Richardson BA, John G, Mbori-Ngacha D, Mwatha A, Ndinya-Achola J, et al. Effect of breastfeeding on mortality among HIV-1 infected women: a randomised trial. Lancet 2001; 357:1651–1655.
8. Garenne M, McCaa R, Nacro K. Maternal mortality in South Africa in 2001: From demographic census to epidemiological investigation. Popul Health Metr 2008; 6:4.
9. Ronsmans C, Graham WJ. and L.M.S.S.S. GroupMaternal mortality: who, when, where, and why. Lancet 2006; 368:1189–1200.
10. Fawcus SR, van Coeverden de Groot HA, Isaacs S. A 50-year audit of maternal mortality in the Peninsula Maternal and Neonatal Service, Cape Town (1953–2002). BJOG 2005; 112:1257–1263.
11. French R, Brocklehurst P. The effect of pregnancy on survival in women infected with HIV: a systematic review of the literature and meta-analysis. Br J Obstet Gynaecol 1998; 105:827–835.
12. Westreich D, Maskew M, Evans D, Firnhaber C, Majuba P, Sanne I. Incident pregnancy and time to death or AIDS among HIV-positive women receiving antiretroviral therapy. PLoS ONE 2013; 8:e58117.
13. Myer L, Cornell M, Fox M, Wood R, Prozesky HW, Ndirangu J, et al. Loss to follow-up and mortality among pregnant and nonpregnant women initiating ART: South Africa. Abstract #22. CROI 2012: Seattle.
14. Tai JH, Udoji MA, Barkanic G, Byrne DW, Rebeiro PF, Byram BR, et al. Pregnancy and HIV disease progression during the era of highly active antiretroviral therapy. J Infect Dis 2007; 196:1044–1052.
15. Ronsmans C, Khlat M, Kodio B, Ba M, De Bernis L, Etard J. Evidence for a ’healthy pregnant woman effect’ in Niakhar, Senegal?. Int J Epidemiol 2001; 30:467–473.
16. Hurt LS, Ronsmans C, Thomas SL. The effect of number of births on women's mortality: systematic review of the evidence for women who have completed their childbearing. Popul Stud (Camb) 2006; 60:55–71.
17. Uganda HIV/AIDS Sero-behavioural survey 2004–2005, 2006, Ministry of Health (Uganda) and ORC Macro: Calverton, MD.
18. Geng EH, Bwana MB, Kabakyenga J, Muyindike W, Emenyonu NI, Musinguzi N, et al. Diminishing availability of publicly funded slots for antiretroviral initiation among HIV-infected ART-eligible patients in Uganda. PLoS ONE 2010; 5:e14098.
19. WHO, UNICEF, UNFPA, The World Bank, Trends in maternal mortality: 1990–2008, World Health Organization, Editor 2010: Geneva.
20. Sahai H, Khurshid A. Statistics in epidemiology: methods, techniques, and applications. New York:CRC Press; 1996.
21. Shapiro RL, Hughes MD, Ogwu A, Kitch D, Lockman S, Moffat C, et al. Antiretroviral regimens in pregnancy and breast-feeding in Botswana. N Engl J Med 2010; 362:2282–2294.
22. Kesho Bora Study Group and I. de VincenziTriple antiretroviral compared with zidovudine and single-dose nevirapine prophylaxis during pregnancy and breastfeeding for prevention of mother-to-child transmission of HIV-1 (Kesho Bora study): a randomised controlled trial. Lancet Infect Dis 2011; 11:171–180.
23. Thomas TK, Masaba R, Borkowf CB, Ndivo R, Zeh C, Misore A, et al. Triple-antiretroviral prophylaxis to prevent mother-to-child HIV transmission through breastfeeding--the Kisumu Breastfeeding Study, Kenya: a clinical trial. PLoS Med 2011; 8:e1001015.
24. Kaplan R, Orrell C, Zwane E, Bekker LG, Wood R. Loss to follow-up and mortality among pregnant women referred to a community clinic for antiretroviral treatment. AIDS 2008; 22:1679–1681.
25. Geng EH, Glidden DV, Bangsberg DR, Bwana MB, Musinguzi N, Nash D, et al. A causal framework for understanding the effect of losses to follow-up on epidemiologic analyses in clinic-based cohorts: the case of HIV-infected patients on antiretroviral therapy in Africa. Am J Epidemiol 2012; 175:1080–1087.
26. Braitstein P, Brinkhof MW, Dabis F, Schechter M, Boulle A, Miotti P, et al. Mortality of HIV-1-infected patients in the first year of antiretroviral therapy: comparison between low-income and high-income countries. Lancet 2006; 367:817–824.
27. Walker AS, Prendergast AJ, Mugyenyi P, Munderi P, Hakim J, Kekitiinwa A, et al. Mortality in the year following antiretroviral therapy initiation in HIV-infected adults and children in Uganda and Zimbabwe. Clin Infect Dis 2012; 55:1707–1718.
28. Munn DH, Zhou M, Attwood JT, Bondarev I, Conway SJ, Marshall B, et al. Prevention of allogeneic fetal rejection by tryptophan catabolism. Science 1998; 281:1191–1193.
29. Schrocksnadel K, Widner B, Bergant A, Neurauter G, Schennach H, Schrocksnadel H, et al. Longitudinal study of tryptophan degradation during and after pregnancy. Life Sci 2003; 72:785–793.
30. Singh N, Perfect JR. Immune reconstitution syndrome and exacerbation of infections after pregnancy. Clin Infect Dis 2007; 45:1192–1199.
31. Cheng VC, Woo PC, Lau SK, Cheung CH, Yung RW, Yam LY, et al. Peripartum tuberculosis as a form of immunorestitution disease. Eur J Clin Microbiol Infect Dis 2003; 22:313–317.
32. Zaba B, Calvert C, Marston M, Isingo R, Nakiyingi-Miiro J, Lutalo T, et al. Effect of HIV infection on pregnancy-related mortality in sub-Saharan Africa: secondary analyses of pooled community-based data from the network for Analysing Longitudinal Population-based HIV/AIDS data on Africa (ALPHA). Lancet 2013; 381:1763–1771.
33. Liotta G, Mancinelli S, Gennaro E, Scarcella P, Nielsen-Saines K, Magid NA, et al. Is highly active antiretroviral therapy (HAART) in pregnancy protective against maternal mortality? Results from a large DREAM cohort in Malawi and Mozambique. Abstract No. TUAB0201. 6th IAS Conference on HIV Pathogenesis and Treatment 2011: Rome.
34. Kaida A, Matthews LT, Kanters S, Mocello R, Kabakyenga J, Muzoora C, et al. Incidence of pregnancy and pregnancy outcomes among a cohort of reproductive-aged women receiving antiretroviral therapy in Mbarara, Uganda. PLoS ONE 2013; 8:e63411.
35. Homsy J, Bunnell R, Moore D, King R, Malamba S, Nakityo R, et al. Reproductive intentions and outcomes among women on antiretroviral therapy in rural Uganda: a prospective cohort study. PLoS ONE 2009; 4:e4149.
36. Cooper D, Harries J, Myer L, Orner P, Bracken H, Zweigenthal V. Life is still going on’: reproductive intentions among HIV-positive women and men in South Africa. Soc Sci Med 2007; 65:274–283.
37. Andia I, Kaida A, Maier M, Guzman D, Emenyonu N, Pepper L, et al. Highly active antiretroviral therapy and increased use of contraceptives among HIV-positive women during expanding access to antiretroviral therapy in Mbarara, Uganda. Am J Public Health 2009; 99:340–347.
38. Linas BS, Minkoff H, Cohen MH, Karim R, Cohan D, Wright RL, et al. Relative time to pregnancy among HIV-infected and uninfected women in the Women's Interagency HIV Study, 2002–2009. AIDS 2011; 25:707–711.
39. Myer L, Carter RJ, Katyal M, Toro P, El-Sadr WM, Abrams EJ. Impact of antiretroviral therapy on incidence of pregnancy among HIV-infected women in sub-Saharan Africa: a cohort study. PLoS Med 2010; 7:e1000229.
40. Muyindike W, Fatch R, Steinfield R, Matthews LT, Musinguzi N, Emenyonu NI, et al. Contraceptive use and associated factors among women enrolling into HIV care in southwestern Uganda. Infect Dis Obstet Gynecol 2012; 2012:340782.
41. Polis CB, Gray RH, Lutalo T, Nalugoda F, Kagaayi J, Kigozi G, et al. Trends and correlates of hormonal contraceptive use among HIV-infected women in Rakai, Uganda, 1994–2006. Contraception 2011; 83:549–555.
42. Guthrie BL, Choi RY, Bosire R, Kiarie JN, Mackelprang RD, Gatuguta A, et al. Predicting pregnancy in HIV-1-discordant couples. AIDS Behav 2010; 14:1066–1071.
43. Makumbi FE, Nakigozi G, Reynolds SJ, Ndyanabo A, Lutalo T, Serwada D, et al. Associations between HIV antiretroviral therapy and the prevalence and incidence of pregnancy in Rakai, Uganda. AIDS Res Treat 2011. 519492.
44. Gibb DM, Kizito H, Russell EC, Chidziva E, Zalwango E, Nalumenya R, et al. Pregnancy and infant outcomes among HIV-infected women taking long-term ART with and without tenofovir in the DART trial. PLoS Med 2012; 9:e1001217.
45. Beyeza-Kashesya J, Ekstrom AM, Kaharuza F, Mirembe F, Neema S, Kulane A. My partner wants a child: a cross-sectional study of the determinants of the desire for children among mutually disclosed sero-discordant couples receiving care in Uganda. BMC Public Health 2010; 10:247.
Keywords:

Africa; antiretroviral therapy; HIV; immune reconstitution; maternal health; maternal mortality; mortality; postpartum; pregnancy; women

© 2013 Lippincott Williams & Wilkins, Inc.