Pregnancy in a Perinatally Infected Woman: Preventing Second-Generation Mother-to-Child Transmission of HIV

A 28-year-old African American woman (gravida 1, para 1) presented to clinic with breast tenderness. The result of a urine pregnancy test was positive. She had received a diagnosis of congenital HIV infection in 1990, at age 12, before the onset of sexual activity. Her mother's HIV infection was discovered at the time of the patient's diagnosis. Although the patient frequently had been ill in childhood, she had received little medical follow-up because both parents were active drug users and she had been cared for by her older brothers. None of her 3 siblings were HIV-infected. Both of the patient's parents died of AIDS before she was 16.

Prior to starting antiretroviral therapy, her CD4⁺ cell count nadir was zero, and her peak HIV RNA level was 750,000 copies/mL. HIV-related illnesses included recurrent bacterial pneumonia, shingles, Pneumocystis jiroveci (formerly carinii) pneumonias, and thrush. Disseminated Mycobacterium avium complex (MAC) infection was diagnosed at age 22. Papanicolaou (Pap) smears documented both atypical squamous cells of undetermined significance and human papillomavirus since age 20 (no records of Pap smears before age 20 were available). The patient also had an episode of aseptic meningitis at age 27, after discontinuing highly active antiretroviral therapy, but she quickly recovered and restarted her antiretroviral drug regimen.

Her medical history also was significant for recurrent genital herpes and pelvic inflammatory disease that were treated with antibiotics at age 19. She reported the onset of sexual activity at age 17. There was no history of viral hepatitis, but prior serological testing indicated a response to vaccination. There was no surgical history. The patient smoked 2 or 3 cigarettes a day, smoked marijuana once a week, and very rarely drank alcohol. There was no history of injection or other drug use.

The patient also had a diagnosis of chronic depression, mostly untreated, which had been exacerbated in the postpartum period of her first pregnancy by a breakup with the father of that child. She had responded to counseling and antidepressants but was currently not receiving treatment.

She became pregnant for the first time at age 22. Although she had taken medroxyprogesterone acetate (Depo-Provera) since she became sexually active, she had missed numerous injections. She had stopped her antiretroviral therapy on her own prior to her first pregnancy but restarted treatment as part of her prenatal care. During her first pregnancy, she developed a disseminated MAC infection with fevers and wasting despite receiving antiretroviral therapy. This complication responded to treatment with ethambutol, azithromycin, and rifabutin without interrupting her antiretroviral therapy. She was treated successfully with a regimen of the lopinavir/ritonavir coformulation plus a fixed-dose combination of zidovudine/lamivudine and gave birth to a full-term, HIV-uninfected girl who has remained healthy. The patient discontinued Depo-Provera after the birth because she experienced breakthrough bleeding. A contraceptive patch was then prescribed, but she had difficulty in remembering when to change it.

Because of continuing concerns about medication adherence and the patient's concerns about abdominal fat accumulation, the patient's antiretroviral regimen had been changed about a year before her curent clinical presentation, from lopinavir/ritonavir coformulation to atazanavir and low-dose ritonavir and from a fixed-dose combination of zidovudine/lamivudine to a fixed-dose combination of tenofovir disoproxil fumarate/emtricitabine. Her most recent CD4⁺ cell count was 777/µL, and her HIV RNA level by polymerase chain reaction was undetectable (less than 50 copies/mL). She wanted to continue this unplanned pregnancy. The father of the baby was HIV-negative, and he was very happy about the pregnancy.

The patient's antiretroviral treatment history included zidovudine monotherapy begun shortly after her HIV diagnosis in 1990. Stavudine, lamivudine, and nelfinavir were added successively. She had been intolerant of both efavirenz (because of CNS toxicity) and indinavir (which caused nausea and vomiting), and her overall adherence to her past antiretroviral drug regimens had been poor. Prior to switching from her previous lopinavir/ritonavir-based regimen, genotypic testing showed only minor protease inhibitor (PI)-associated mutations (N88N/D).

Physical examination revealed a petite, normally developed African American woman with prominent abdominal fat. Her weight was 105 lb, height was 5 ft 0 in, and blood pressure was 105/70 mm Hg. Findings from the remainder of the examination were unremarkable. She was allergic to dapsone and erythromycin (both of which caused a rash). Her medicines at the current clinic visit included the following: valacyclovir 500 mg twice daily, tenofovir/emtricitabine 300/200 mg daily, atazanavir 300 mg daily, and ritonavir 100 mg daily.

TRANSITIONING PATIENTS
As most pediatric HIV clinicians are well aware, transitioning perinatally HIV-infected survivors to the adult care setting presents challenges for providers as well as for patients and their families. Only very recently have studies assessing this population begun to appear in the literature. Probably the most comprehensive series reported to date is from the Pediatric AIDS Clinical Trials Group (pACTG) Study 219C,¹ which evaluated 638 perinatally HIV-infected adolescent girls, aged 13 years and older. This pACTG study was designed to enroll and observe HIV-infected and HIV-uninfected adolescent girls to evaluate the long-term complications of perinatal HIV infection and antiretroviral exposure. The girls were enrolled from 75 pediatric infectious disease clinics in 24 states and districts within the United States.

Of these 638 adolescent girls (age range, 13.0 to 24.7 years), 174 (27.3%) were known to be sexually active (by self-report, a history of sexually transmitted disease, or known pregnancy). Sexually active girls were more likely to be older, be of Hispanic origin, live independently, and have a higher HIV RNA level and a lower CD4 percentage. In this cohort, 38 girls became pregnant for the first time, for an overall first-time pregnancy incidence rate of 18.8 per 1000 person-years. For girls aged 15 to 19 years, the first-time pregnancy rate was 33.5 per 1000 person-years. The authors noted that this rate is substantially lower than the first-time pregnancy rate in presumed HIV-uninfected US girls of similar age in 1999 (86.7/1000).

Twenty-eight of these 38 first pregnancies resulted in live births (there were 8 therapeutic abortions and 2 spontaneous abortions), and all girls were treated with combination antiretrovirals during pregnancy. Six girls had a second pregnancy, and 1 girl had a third. Of these pregnancies, 4 resulted in live births. Of the total of 32 live-born infants, 1 was known to be HIV-infected and 2 were of unknown HIV status, for a known mother-to-child HIV transmission rate of 3.1% (95% confidence interval, 0.1 - 18.6). Details of the contraceptive practices of the 174 sexually active girls were not provided in this report of pACTG Study 219C data, although it was noted that a considerable proportion had reported engaging in unprotected intercourse.

A similar analysis was performed in a group of 58 perinatally infected adolescent girls and young women in Manipur, India.² Among 30 pregnancies, 4 resulted in elective abortions, and 26 resulted in healthy infants, with no mother-to-child HIV transmission. During pregnancy, 90% of the infants were treated with combination antiretrovirals, and all infants received zidovudine prophylaxis. There were no occurrences of intrauterine growth retardation or fetal demise, no spontaneous abortions or maternal deaths, and only a single preterm delivery (at 35.2 weeks).

A brief review from Puerto Rico focused on 8 perinatally HIV-infected young women with 5 first pregnancies resulting in live births.³ Of the 5 pregnant women, 4 were treated with combination antiretroviral therapy during pregnancy, and no mother-to-child transmissions occurred.

Case reports illustrating the unplanned nature of teen pregnancies in this population were published recently.⁴ The authors reported that despite participation in weekly HIV education and support groups and in group leadership programs and the ready availability of contraception, HIV education, and assistance with disclosure, risky sexual practices still occurred, which resulted in pregnancies. These authors also noted that perinatally infected teens often come from chaotic and unstable families in which parents and siblings have died of AIDS. Such adolescents may have acquired only limited skills to deal with difficult decisions regarding sexuality, intimacy, and childbearing.

PREGNANCY PREVENTION
Part of preventing second-generation HIV infection in this population is prevention of unintended pregnancies, which is itself a clinical challenge, especially among adolescents. The lower pregnancy rate among HIV-infected girls compared with the pregnancy rate of a similar population of predominantly HIV-uninfected girls reported in the pACTG Study 219C is comparable to the lower rate observed in HIV-infected US women.¹ Many studies have addressed but failed to find a biological association between female infertility and HIV infection. To this author's knowledge, there have been no published studies evaluating biological correlates of fertility in the perinatally infected population.

There is, however, some data available on contraception in HIV-infected women. The Women's Interagency HIV Study (WIHS) recently published a review of contraceptive use among its participants from 1994 to 2005.⁵ Structured questionnaires on contraceptive use were employed every 6 months. Women who were purposely not using contraception in order to conceive or who were known to be unable to conceive during the 6-month interval between questionnaire administrations were not included. The majority of pregnancies reported by HIV-positive women in a previously published WIHS study occurred in women not seeking to conceive.⁶

The authors of the contraception study suggest that many of these pregnancies may occur because women with HIV infection are not using highly effective contraception (hormonal contraceptives or sterilization, with or without barrier methods to prevent HIV transmission or acquisition). Despite intensive safer sex education, fewer than 40% of women in the group reported using barrier methods over every study interval, and barrier use was not higher among women with HIV infection.

The authors cite several reasons why HIV-infected women fail to use the most effective contraception, including preoccupation with more pressing health issues, lack of provider training, risks associated with oral contraceptive use, potential drug interactions with antiretrovirals, and smoking status. In addition, antiretroviral therapy may decrease the efficacy of oral contraceptives.⁷ Only a small proportion of the sample were younger than 20 years, but 27.7% of the HIV-positive women were younger than 30 years, as were 45.1% of the HIV-uninfected women. Age-related analyses showed that a substantially larger proportion of younger women were using hormonal methods compared with older women (Figure).⁵

Figure.Changes in contraception use with age among HIV-infected women at risk for pregnancy. (Reproduced with permission from Massad LS et al. J Womens Health [Larchmt]. 2007.⁵)

TREATMENT OF HIV INFECTION IN PREGNANCY
In this population the use of combination antiretroviral therapy for the health of mother and child is complicated by the duration of underlying disease. The safety of such agents is also a concern, particularly in early pregnancy during organogenesis. The duration of disease and prior illness affect the management of complications and increase the requirements for monitoring throughout pregnancy.

 In perinatally infected women who have been able to survive to reproductive age and conceive, the progression of their early childhood disease may have been slower than that of their deceased contemporaries. Often these patients have a history of sequential monotherapies that changed as drugs gradually became available to pediatric patients. As a result, they may have acquired substantial drug resistance, which would require a more complex drug regimen. Since these women are also younger, they are at higher risk for acquisition of other sexually transmitted diseases during pregnancy, and they require monitoring and testing as well as active case management.

Many concerns have been raised about the safety of highly effective antiretroviral regimens for both mother and infant throughout the course of the pregnancy. The maternal toxicities associated with nevirapine and with the coadministration of stavudine and didanosine have been well documented.⁸ In addition, metabolic complications associated with antiretroviral combinations were addressed recently in a study of 158 pregnant women who were receiving highly active antiretroviral therapy. These women were enrolled at 20 to 34 weeks of gestation (after being on stable antiretroviral regimens for more than 8 weeks) and were observed through pregnancy to 12 weeks postpartum.^9,10 The patients were divided into those who were receiving PI-based therapies (n = 81) and those who were receiving non–PI-based therapies (n = 77). Most patients on PI-based regimens were receiving nelfinavir or the lopinavir/ritonavir coformulation. The most common NRTI combination in both arms was zidovudine plus lamivudine. In the non-PI group, 52% of women were treated with nevirapine-based regimens.

The study investigators found that although rates of gestational diabetes were high overall (10%), they were not higher in PI-treated groups; however, total cholesterol and triglyceride levels (median, 230 mg/dL and 224 mg/dL, respectively) were higher and remained so in the PI-treated women. No difference was seen in lactate levels or rates of GI symptoms between the 2 groups.

Similar results were reported in a review of 6000 hospitalizations among pregnant HIV-infected women in the United States during the mid-1990s-the period when highly active antiretroviral therapy was introduced into clinical practice-which showed increases in the rate of gestational diabetes (not stratified by therapy type [eg, PI-based regimen vs non–PI-based regimen]); however, substantial concomitant decreases in hospitalizations for puerperal sepsis, genitourinary tract infections, and other bacterial infections compared with data from the pre-HAART era.¹¹ Nonetheless, the rate of these maternal complications remained higher than the rate in HIV-uninfected pregnant women hospitalized in the same period. Rates of preeclampsia and antepartum hemorrhage were not significantly different between HIV-infected women and HIV-uninfected women over the period of observation.

Obstetrical and infant outcomes have been carefully evaluated in antiretroviral-treated pregnant women after early reports of low birth weights and preterm delivery among HIV-infected mothers receiving highly active antiretroviral therapy during pregnancy. According to a recently published study conducted in the United Kingdom, preterm delivery (ie, earlier than 37 weeks) occurred in 6% to 8 % of all pregnancies but was significantly more common (14.2%) in a single-center cohort study of 211 deliveries from HIV-infected mothers.¹² Only the initiation of combination antiretroviral treatment during pregnancy was associated with an increased risk of preterm delivery, and this finding was independent of maternal health and the class of antiretroviral drug used. Another surveillance study of over 5000 pregnancies in the United Kingdom and Ireland reported similar findings with highly active antiretroviral therapy, whether or not a PI was included, although not specifically associated with the timing of treatment.¹³

In the United States, a review of PI use during 233 pregnancies in HIV-infected women did not find an association between PI-based antiretroviral regimens and preterm delivery,¹⁴ although timing of the initiation of therapy was not specifically addressed. Preterm deliveries in this study were associated with the usual predictors. Of interest, the US study that looked at metabolic complications of antiretroviral therapy during pregnancy found an association between preterm delivery and elevated triglyceride levels during pregnancy.⁹ As reviewed by Thorne and Newell,¹⁵ the conflicting data on preterm delivery and the use of PI-based regimens is unresolved and may be influenced by the populations studied and study methodology.

In addition to studies focusing on the use of PIs during pregnancy and their effect on maternal and fetal outcomes, data on mitochondrial dysfunction in HIV-uninfected children with in utero exposure to NRTIs have been extensively analyzed. In 1999, French investigators presented the first description of a syndrome ascribed to NRTI-associated mitochondrial dysfunction. They described a range of primarily CNS dysfunctions associated with the syndrome, the most severe of which were fatal.¹⁶ This report prompted widespread reviews of cohort studies in progress, but no evidence of additional cases in US, European, or African children were found. However, the most recent analysis of 1037 HIV-infected children enrolled in pACTG studies found 20 cases of possible mitochondrial dysfunction in association with first exposure to lamivudine or to zidovudine with lamivudine in the third trimester.¹⁷

FOLLOW-UP
Our patient continued taking her boosted atazanavir regimen initially but experienced substantial nausea and was unable to maintain 100% adherence for much of her first trimester and part of her second trimester. During this time, she had a 50% drop in her CD4⁺ cell count to 350/µL and a slight increase in her HIV RNA level from undetectable levels to 75 copies/mL. Her course was complicated by otitis media early in pregnancy and a urinary tract infection later on; both conditions responded to amoxicillin. Her clinic visits were every 2 weeks throughout her pregnancy, primarily with her internist initially and then later with an obstetrician. She decided to restart treatment with lopinavir/ritonavir and lamivudine/zidovudine, which she tolerated well for the remainder of her pregnancy, resuppressing her HIV RNA level below 50 copies/mL. She gave birth to a full-term HIV-uninfected girl, who, as of this writing (August 2007), was 5 months old and appears neurologically intact.

Dr Vogler reports having served on the speakers bureaus for Gilead Sciences, GlaxoSmithKline, and Merck; served on advisory boards for Abbott Laboratories and GlaxoSmithKline; and received ACTG research support from Bristol-Myers Squibb, Gilead Sciences, and Merck. No other potential conflict of interest relevant to this article was reported by Dr Vogler.

References:

References1. Brogly SB, Watts DH, Ylitalo N, et al. Reproductive health of adolescent girls perinatally infected with HIV. Am J Public Health. 2007;97:1047-1052.
2. Chibber R, Khurranna A. Birth outcomes in perinatally HIV-infected adolescents and young adults in Manipur, India: a new frontier. Arch Gynecol Obstet. 2005;271:127-131.
3. Centers for Disease Control and Prevention. Pregnancy in perinatally HIV-infected adolescents and young adults-Puerto Rico, 2002 [published correction appears in MMWR. 2003;52:191]. MMWR. 2003;52:149-151.
4. Levine AB, Aaron E, Foster J. Pregnancy in perinatally HIV-infected adolescents. J Adolesc Health. 2006;38:765-768.
5. Massad LS, Evans CT, Wilson TE, et al. Contraceptive use among U.S. women with HIV. J Womens Health (Larchmt). 2007;16:657-666.
6. Massad LS, Springer G, Jacobson L, et al. Pregnancy rates and predictors of contraception, miscarriage and abortion: results from the Women's Interagency HIV Study. AIDS. 2004;28:281-286.
7. Clark RA, Theall K. Population-based study evaluating association between selected antiretroviral therapies and potential oral contraceptive failure. J Acquir Immune Defic Syndr. 2004;37:1219-1220.
8. Hitti J, Frenkel LM, Stek AM, et al. Maternal toxicity with continuous nevira-pine in pregnancy: results from PACTG 1022. J Acquir Immune Defic Syndr. 2004;36:772-776.
9. Livingston EG, Cohn SE, Yang Y, et al. Lipids and lactate in human immunodeficiency virus-1 infected pregnancies with or without protease inhibitor-based therapy. Obstet Gynecol. 2007;110:391-397.
10. Hitti J, Andersen J, McComsey G, et al. AIDS Clinical Trials Group 5084 Study Team. Protease inhibitor-based antiretroviral therapy and glucose tolerance in pregnancy: AIDS Clinical Trials Group A5084. Am J Obstet Gynecol. 2007;196:331.e1-7.
11. Kourtis AP, Bansil P, McPheeters M, et al. Hospitalizations of pregnant HIV-infected women in the USA prior to and during the era of HAART, 1994-2003. AIDS. 2006;20:1823-1831.
12. Martin F, Taylor GP. Increased rates of preterm delivery are associated with the initiation of highly active antiretroviral therapy during pregnancy: a single-center cohort study. J Infect Dis. 2007;196:558-561.
13. Townsend CL, Cortina-Borja M, Peckham CS, Tookey PA. Antiretroviral therapy and premature delivery in diagnosed HIV-infected women in the United Kingdom and Ireland. AIDS. 2007;21:1019-1026.
14. Morris AB, Dobles AR, Cu-Uvin S, et al. Protease inhibitor use in 233 pregnancies. J Acquir Immune Defic Syndr. 2005;40:30-33.
15. Thorne C, Newell ML. Safety of agents used to prevent mother-to-child transmission of HIV: is there any cause for concern? Drug Saf. 2007;30:203-213.
16. Blanche S, Tardieu M, Rustin P, et al. Persistent mitochondrial dysfunction and perinatal exposure to antiretroviral nucleoside analogues. Lancet. 1999;354:1084-1089.
17. Brogly SB, Ylitalo N, Mofenson LM, et al. In utero nucleoside reverse transcriptase inhibitor exposure and signs of possible mitochondrial dysfunction in HIV-uninfected children. AIDS. 2007;21:929-938.