Women accounted for 26% of newly diagnosed AIDS cases in the United States in 2005 and represented 30% of newly reported HIV infections.¹  Worldwide, nearly half of the 37 million adults estimated to be living with HIV infection in 2006 were women.²  Since 80% of women infected with HIV are of reproductive age, coexistent HIV and pregnancy are relatively common.³

The obstetrician has key roles in educating women to prevent acquisition of HIV infection and in providing state of the art care to HIV-infected pregnant women to optimize their health and prevent HIV transmission to the infant. Treatment regimens for HIV and strategies to prevent perinatal transmission have become more complex. The provider caring for an HIV-infected pregnant woman must understand the interactive effects of HIV and pregnancy and be aware of the most current data on interventions for therapy and prevention of transmission. This chapter is designed to assist the practicing obstetrician in caring for HIV-infected pregnant women, and includes information for obtaining updated information as current treatment guidelines are revised. Providers are urged to consult these online guidelines for the most current information.

The recommendations for HIV testing in pregnancy have evolved with the epidemic. Initially HIV infection was believed to be confined to identifiable high risk groups. As the rate of heterosexual transmission of HIV increased, it became clear that testing pregnant women with self-identified risk factors only detected approximately half of HIV-infected pregnant women.^{4, 5, 6} With the development of interventions including zidovudine to reduce perinatal transmission, the importance of identifying HIV-infected women as early as possible in pregnancy became apparent and voluntary testing was recommended.⁷  With broader uptake of testing and use of zidovudine for HIV-infected pregnant women, the rate of perinatally acquired AIDS cases dropped by 67% from 1992 to 1997, with a minimal decline in the number of HIV-infected women giving birth in the US.⁸ However, while the proportion of HIV-infected women tested during pregnancy has increased, up to 10% of HIV-infected pregnant women did not receive any zidovudine in 1997, suggesting a need for increased detection and treatment.

In an effort to further increase the rate of testing and identification of HIV-infected women and to destigmatize the HIV testing process, in 1999 the Institute of Medicine (IOM) issued a report recommending a shift in testing policy from one of counseling and voluntary testing to universal HIV testing with patient notification as a routine component of prenatal care.⁹ Universal means that the test applies to all pregnant women regardless of risk factors and prevalence rates in their area. Routine with notification means that the HIV test is part of the battery of standard prenatal tests and that women are informed that the HIV test is being done and that they may refuse the test. Incorporation of HIV testing into the routine prenatal tests prevents stigmatization, reduces the cost of testing, and deals with potential geographic shifts in epidemiology. The IOM recommendations have been endorsed by the American College of Obstetricians and Gynecologists (ACOG) and have been incorporated into the CDC Revised Recommendations for HIV Testing of Adults, Adolescents, and Pregnant Women in Health-Care Settings.^{10, 11} Other recommendations in the IOM report for further reduction in perinatal transmission of HIV include development of comprehensive strategies to ensure access to prenatal care, HIV counseling and testing to prevent infection in women, high-quality, coordinated care to provide interventions to reduce transmission rates, avoidance of breastfeeding by HIV-infected women, and appropriate treatment and services for mothers.

Women who present in labor without prenatal care are at increased risk for HIV seropositivity and should be offered HIV testing. In one study, up to 15% of HIV-infected women had no prenatal care and another 20% had fewer than five visits.¹² Several studies have found a two- to four-fold increased risk of HIV seropositivity among women presenting in labor or late pregnancy for care.^{13, 14, 15} Risk factors for late or no prenatal care such as illicit drug use are also risk factors for HIV seropositivity. Since initiation of antiretroviral therapy during labor or for the infant within 24 hours of birth may decrease the risk of infant HIV infection,^{16, 17} testing should be offered as soon as possible after presentation and test results should be available rapidly. Many ethical and logistical issues are involved in offering rapid HIV testing in labor and the immediate postpartum period, and each institution should develop its own policy.¹⁸ Combining rapid testing programs for pregnant women and those providing testing of source patients from health care worker exposures may increase efficiency.

Testing for women presenting in labor without prior testing can be done by use of a rapid diagnostic test or rapid processing and return of results using standard testing such as enzyme immunoassay (EIA). Rapid tests for detecting HIV antibodies can be performed in 5–40 minutes. A guide and model protocol for implementation of rapid testing during labor is available at http://www.cdc.gov/hiv/topics/testing/resources/guidelines/pdf/Labor&DeliveryRapidTesting.pdf.¹⁹ Given the relatively low seroprevalence of HIV in most perinatal settings in the US, the negative predictive value of a single EIA or rapid test is high and further testing is not required.¹⁹ However, since the positive predictive value of a single rapid test will be low in low prevalence populations, a reactive rapid test must be confirmed by additional testing. Providers may choose to initiate antiretroviral therapy for women with a repeatedly reactive rapid HIV test and their neonates while awaiting confirmatory testing. 

Implementation of universal HIV testing with patient notification does not obviate the need for education of women about the risks and implications of HIV infection and means of protecting themselves. All women’s health care providers should provide education regarding protection from HIV and other sexually transmitted diseases to their sexually active patients. Before HIV testing, providers should provide verbal or written information on HIV, how it is spread, interventions available to prevent perinatal transmission, and services available for HIV prevention and treatment. Documentation of a woman’s consent or refusal of HIV testing in the medical record is recommended. Providers should be aware of regulations in their area that may require more specific counseling and informed consent documentation than those recommended by the US Public Health Service. Although women should not be coerced to be tested, women who decline testing during pregnancy should have the reasons for refusal explored and be offered testing again at subsequent visits. Confidentiality of patient information should always be maintained but providers must also be aware of local requirements for partner notification if women are not able to notify partners.

While false positive tests for HIV antibodies do not appear to be increased in pregnancy, the rate of indeterminate western blots may be increased.²⁰ Indeterminate western blot results can be caused by an evolving antibody profile in response to recent HIV infection, loss of antibody in persons with end-stage HIV infection, or non-specific cross-reacting antibodies not related to HIV. Risk factors for indeterminate results not related to HIV include current or previous pregnancy, autoantibodies such as rheumatoid factor, and recent immunization.²⁰ The rate of seroconversion was low among individuals with indeterminate western blots and was always associated with current high risk behavior such as unprotected sex with an infected partner or ongoing injection drug use.^{20, 21} If a pregnant woman has a repeatedly reactive EIA and an indeterminate western blot, she should be questioned regarding recent risk behaviors and undergo repeat antibody testing. Nearly all HIV-infected individuals will develop complete HIV antibody profiles within 1 month of exposure. Continued indeterminate results are highly unlikely to be related to HIV infection, especially if the patient’s sexual partners are negative and there are no other high risk behaviors. Although HIV DNA and RNA assays are not approved for diagnostic use, they may sometimes be helpful in distinguishing the reason for indeterminate western blot testing in high risk individuals. 

Pregnancy may have an impact on coexistent maternal medical conditions such as congenital heart disease and autoimmune diseases, but its impact on HIV infection is less clear. The CD4+ lymphocyte count tends to drop in pregnancy in both HIV-infected and uninfected women, while the CD4+ lymphocyte percentage tends to be more stable.^{22, 23, 24} This drop in count could allow enhanced viral replication and disease progression, and antigenic stimulation by fetal tissues during pregnancy could lead to T-cell activation and viral replication. Conversely, the relative immunosuppression of pregnancy could lead to decreased viral replication. Studies on the progression of HIV disease from the US and Europe do not suggest an increase in disease progression related to pregnancy.^{22, 25, 26, 27, 28, 29} Several studies have suggested an increased risk of bacterial pneumonia among HIV-infected women during pregnancy and the postpartum period, perhaps related to respiratory changes in pregnancy.^{29, 30} Studies from developing countries are less clear and suggest a potential effect that could be related to nutritional differences or coexistent infections.^{26, 31, 32}

As with most serious medical conditions coexisting with pregnancy, the treatment of the HIV-infected mother should be the same as that used if the woman were not pregnant unless clear contraindications to specific therapies exist.^{33, 34} Experience with use of most antiretroviral agents in pregnancy is limited, and long term follow up of children exposed to these agents in utero is unavailable. HIV-infected pregnant women should be counseled about the potential benefits of therapy for their own health and reduction of perinatal transmission, about currently available data on risks, and about the lack of long term information.  Pre-clinical and clinical data pertinent to use in pregnancy for currently approved antiretrovirals are summarized in Table 1.^{33, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51} The predictive value of in vitro and animal studies for carcinogenicity, mutagenicity, reproductive, and teratogenic effects in humans is unknown. However, use of efavirenz should be avoided during the first trimester because of defects seen in a study of cynologous monkeys and concerning data in humans.³³ Defects including anencephaly, anophthalmia, microophthalmia, and facial clefts were seen in three (15%) of 20 monkeys after first trimester exposure to efavirenz. Several retrospective cases of neural tube defects after first trimester exposure to efavirenz in humans have been reported, although the denominator is unknown so a rate cannot be calculated. An increased rate of hypospadias was detected after first trimester zidovudine exposure in the Women and Infants Transmission Study, but this finding has not been confirmed in other studies.⁵² For more detailed information, readers should consult the Safety and Toxicity of Individual Antiretroviral Drugs in Pregnancy hyperlink available in both the “Guidelines for the Use of Antiretroviral Agents in HIV-infected Adults and Adolescents” and “USPHS Task Force Recommendations for the Use of Antiretroviral Drugs in Pregnant Women Infected with HIV-1 for Maternal Health and for Reducing Perinatal HIV-1 Transmission in the United States” available at www.aidsinfo.nih.gov .^{33, 34} This summary is updated as new drugs are approved or new information becomes available. Recommended treatment during pregnancy based on CD4+ cell counts, viral load, and antiretroviral treatment history is discussed in more detail below.

 Table 1. Preclinical and clinical data relevant to the use of antiretrovirals (ARV) in pregnancy^{33, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,}

*No longer available in the US. FDA pregnancy category: A, adequate and well-controlled studies of pregnant women fail to demonstrate a risk to the fetus during the first trimester of pregnancy (and there is no evidence of risk during the later trimesters); B, animal reproduction studies fail to demonstrate a risk to the fetus and adequate and well-controlled studies of pregnant women have not been conducted; C, safety in human pregnancy has not been determined, animal studies are either positive for fetal risk or have not been conducted, and the drug should not be used unless the potential benefit outweighs the potential risk to the fetus; D, positive evidence of human fetal risk based on adverse reaction data from investigational or marketing experiences, but the potential benefits from the use of the drug in pregnant women may be acceptable despite its potential risks; X, studies in animals or reports of adverse reactions have indicated that the risk associated with the use of the drug for pregnant women clearly outweighs any possible benefit.

Pk, pharmacokinetic, VSD, ventricular septal defect.

Providers treating pregnant women with antiretroviral agents are urged to report their cases to the Antiretroviral Pregnancy Registry as early in pregnancy as possible.  Registry collects observational data on antiretroviral exposure during pregnancy to assess the potential for teratogenicity of the drugs and is a collaborative project of pharmaceutical manufacturers and an advisory committee. Prospectively reported cases (identified before the outcome of pregnancy is known) are used for the calculation of rates of defects while retrospectively reported cases are used only for detecting any specific pattern of defects because of the potential for reporting bias. Patient identifiers are not included in reports to the registry, and registry staff contact the reporting provider to obtain birth outcomes. Reports on findings to date are issued every 6 months. To report cases and obtain the most recent report, contact the Antiretroviral Pregnancy Registry, 1410 Commonwealth Drive, Wilmington, NC  28403, telephone 1-800-258-4263, fax 1-800-800-1052, or www.APRegistry.com.

 
Two other concerns regarding potential long term effects of antiretroviral therapy during pregnancy on the infant and child have been raised – the potential for transplacental carcinogenicity and mitochondrial toxicity. Non-metastisizing vaginal tumors have been seen in adult rodents treated with continuous, high-dose zidovudine, likely related to reflux of urine with high concentrations of unmetabolized zidovudine onto the vaginal mucosa in rodents.⁵³ No increase in tumors in other organs has been seen in adult rodents, and since humans excrete only metabolized zidovudine in the urine, this effect is not expected in humans. Because of the frequent use of zidovudine in HIV-infected pregnant women, two studies of transplacental carcinogenesis in mice have been done for zidovudine.^{54, 55} In one study using 25 and 50 times the human dose in the third trimester of pregnancy, the offspring in the highest dose group showed a significant increase in the rate of tumors of the lung, liver, and reproductive organs compared to untreated controls.⁵⁴ Incorporation of zidovudine into the DNA of tissues of newborn mice was detected, but the incorporation did not correlate with development of tumors. In the second study, doses of zidovudine were approximately three times the human dose, and some of the offspring were also treated. No increase in tumors was observed in the offspring, except for the expected rate of vaginal tumor development in animals who continued to receive zidovudine.⁵⁵ The animal data were reviewed by an expert panel convened by the National Institutes of Health, and the panel recommended that the benefits of zidovudine in reducing perinatal transmission of HIV outweigh the theoretical risks of transplacental carcinogenesis but that infants exposed to zidovudine in utero should have long term follow up to detect any potential adverse effects.⁵⁶ In 727 children exposed to zidovudine in utero with over 1100 person-years of follow up, no tumors have been observed.^57,   Similarly, zidovudine was not found to be a risk factor for malignancy in a case–control study of HIV-infected children.⁵⁸

Nucleoside analogue drugs such as zidovudine may induce mitochondrial dysfunction by binding to mitochondrial gamma DNA polymerase and interfering with mitochondrial replication.⁵⁹ The relative potency for inhibition of mitochondrial gamma DNA polymerase in vitro is zalcitabine (highest), didanosine, stavudine, lamivudine, zidovudine, and abacavir (lowest).⁶⁰ Toxicity such as myopathy related to mitochondrial dysfunction has been reported among HIV-infected persons treated with nucleoside agents, but it has generally been reversible when the drugs are stopped. A French group raised concerns of possible persistent mitochondrial dysfunction among infants with in utero and neonatal nucleoside exposure to either zidovudine or zidovudine/lamivudine when they reported eight cases of HIV-uninfected infants with symptoms possibly related to mitochondrial dysfunction, two of whom had progressive neurologic dysfunction and died.⁶¹ A subsequent publication described 12 children (including the previously reported eight cases) with neurologic symptoms and with deficits in one of the mitochondrial respiratory chain complexes, abnormal muscle biopsy, or both, identified from a total of 4392 children in the French Pediatric Cohort or the France National Registry. All symptomatic infants had been exposed to antiretrovirals in utero or the neonatal period.⁶² Probable mitochondrial toxicity was identified among 12 (0.26%) of 2644 antiretroviral-exposed children and none of the 1748 unexposed children. An additional 14 children with possible mitochondrial toxicity after antiretroviral exposure were identified from the same group. These data are consistent with a small study of 30 infants born to HIV-uninfected women and 20 infants born to HIV-infected women, half of whom were exposed to zidovudine.⁶³ Mitochondrial DNA levels were lower in infants born to HIV-infected women, regardless of antiretroviral exposure, and lowest among those with zidovudine exposure. Differences in mitochondrial DNA levels persisted through 2 years of age. No clinical information on the children was provided.

To evaluate potential nucleoside toxicity, a collaborative effort between several large cohorts in the US identified and reviewed 353 deaths in more than 20,000 children born to HIV-infected women. No deaths similar to those reported in France were identified, although only 6% of the children had been exposed to the combination of zidovudine/lamivudine.⁶⁴ Evaluation of living children in these cohorts for potential evidence of mitochondrial dysfunction is ongoing. No differences in growth, immunologic parameters, or cognitive development were observed between children exposed to zidovudine or placebo in PACTG 076 and followed up to 5.6 years.⁶⁵ No deaths or tumors occurred during follow up. Neurologic adverse events have been reviewed among 1798 children enrolled to the PETRA trial that compared zidovudine/lamivudine to placebo for prevention of HIV transmission. No increased risk of neurologic events was seen among children exposed to zidovudine/lamivudine compared to placebo, regardless of intensity of treatment.⁶⁶ In a review of clinical symptoms among 2414 HIV-exposed, uninfected children in the European Collaborative study, including 1008 with perinatal antiretroviral exposure, no association was seen with symptoms suggestive of mitochondrial dysfunction and antiretroviral exposure.⁶⁷ The potential for in utero and neonatal exposure to nucleosides to cause mitochondrial dysfunction later remains unproven. If the association exists, the development of severe or fatal disease appears to be extremely rare and must be compared to the clear benefit of zidovudine in reducing transmission of an ultimately fatal infection.⁶⁸ However, these concerns underscore the need for long term follow up of antiretroviral exposed children.

When treating HIV-infected pregnant women with antiretroviral agents or drugs active against opportunistic infections, the physiologic changes of pregnancy must be considered.⁶⁹ The 45% increase in plasma volume coupled with only a 20–30% increase in red cell mass leads to dilutional anemia, which may be exacerbated by drugs that cause bone marrow suppression such as zidovudine. Cardiac output and glomerular filtration rate increase by 30–50%, potentially causing increased excretion and decreased levels of drugs with primarily renal clearance. Increased tidal volume and pulmonary blood flow may lead to increased absorption of aerosolized medications such as pentamidine. Changes in metabolic enzyme pathways in the liver may occur as well. Placental transfer of drugs and metabolism by the fetus may also affect maternal drug levels. Ideally, specific pharmacokinetic studies of new antiretroviral agents in pregnancy will be done, but if these data are not available to guide dosing, clinicians must consider the metabolism of the drug and potential impact of pregnancy and monitor efficacy of the chosen regimen by assessing the effect on plasma HIV RNA levels. 

In assessing the impact of maternal HIV infection on pregnancy and infant outcome, it is important to consider other risk factors for adverse pregnancy outcome such as smoking and illicit drug use which may be increased among HIV-infected women. Studies in the US and Europe, reported primarily before the routine use of antiretroviral therapy in pregnancy, have not shown increases in adverse pregnancy outcomes in HIV-infected compared to uninfected women.^{30, 70, 71, 72, 73, 74, 75, 76, 77} Of note, the outcomes among both groups in these studies have tended to have increased proportions of women with low birth weight and preterm infants compared to the general population, most likely related to maternal drug use, lack of prenatal care, and other risk behaviors.⁷⁸ In contrast, the majority of studies from developing countries have shown an increase in low birth weight and preterm birth among HIV-infected women compared to uninfected women.^{79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97} In addition, when examined, outcomes appear to be worse for women with symptomatic compared to asymptomatic HIV infection. These differential outcomes among women in developing countries may be related to nutritional differences, co-infections such as malaria and syphilis, or other factors.

Studies evaluating the effect of infant HIV infection on birth weight and gestational age, comparing infants exposed but HIV uninfected and those infected, have had variable results. The majority have not shown a clear impact of infant HIV infection on birth weight and gestational age at delivery,^{76, 77, 80, 88, 98, 99, 100} most likely because of the high proportion of infections which occur very late in pregnancy or during labor, delivery, or breast feeding. Studies reporting a lower birth weight or gestational age among HIV-infected compared to exposed but uninfected infants had high rates of preterm birth in the cohorts, suggesting that the increased risk of transmission related to prematurity may have contributed to the finding.^{94, 101, 102} 

Studies evaluating risk factors for adverse pregnancy outcome among HIV-infected women have found them to be primarily the same risk factors seen in HIV-uninfected women with a variable contribution of HIV infection. Using low birth weight as the outcome, the most consistent contributors in studies in the US and Europe have been injection drug use during pregnancy,^{72, 98, 99,} smoking,⁷² history of previous adverse outcome,^{103, 104} bleeding during pregnancy,¹⁰³ lack of prenatal care,⁷⁸ CD4 percentage below 14,¹⁰³ pre-eclampsia,¹⁰⁴ and entry HIV culture titer but not viral load, CD4+ lymphocyte count, or delivery culture.¹⁰⁴ Risk factors for preterm birth noted in US and European studies included current injection drug use,⁹⁸ prior preterm birth,^{103, 104} multiple gestation,¹⁰⁴ bleeding during pregnancy,¹⁰³ alcohol use,¹⁰⁴ pre-eclampsia,¹⁰⁴ CD4 percentage below 14,¹⁰³ and lack of zidovudine use.¹⁰⁵ In Africa, risk factors for low birth weight included malaria parasitemia,^{95, 96} trichomoniasis,⁹⁶ HIV infection,^{87, 96} and maternal weight.⁹⁷

The impact of antiretroviral therapy on pregnancy outcome is unclear. The drugs could exert a deleterious effect on fetal growth, conversely the suppression of high levels of HIV replication and immune reconstitution could be beneficial for fetal growth. As noted above, several studies in untreated women have suggested a negative effect of HIV infection or low CD4+ lymphocyte counts on pregnancy outcome, and zidovudine appeared to decrease the rate of prematurity in one analysis.^{103, 104, 105} The impact of combination therapy on pregnancy outcome is also unclear. One small retrospective study reported pregnancy outcome among 37 HIV-infected women, 21 of whom received dual nucleoside reverse transcriptase inhibitor therapy and 16 of whom received dual nucleosides plus one or two protease inhibitors.¹⁰⁶ A possible association of combination therapy with preterm birth was suggested because ten (33%) of 30 babies were born prematurely. A subsequent study, combining data from the European Collaborative Study and the Swiss Mother and Child HIV-1 Cohort Study, evaluated pregnancy outcome among 3920 women. After adjustment for CD4+ lymphocyte count and intravenous drug use, the odds ratio for preterm delivery was 2.6 (95% CI 1.4–4.8) for infants born to women receiving combination antiretroviral therapy, with or without a protease inhibitor compared to those on no treatment.¹⁰⁷ Women on therapy from before pregnancy had a higher risk of preterm delivery that those starting therapy in the third trimester. No increase in preterm birth was seen in the monotherapy group. In a combined analysis of seven US studies, no increase in preterm delivery or low birth weight was seen among women on combination antiretrovirals compared to those on monotherapy or no therapy.¹⁰⁸ An increased risk of very low birth weight was seen among women on combination therapy with a protease inhibitor compared to those on combinations without a protease inhibitor, but this included only seven women with very low birth weight infants in the protease inhibitor group. A more recent, smaller study from the US detected an increased risk of preterm birth among women on protease inhibitor therapy, but this study had high rates of preterm birth in all therapy groups.¹⁰⁹ The benefits of combination therapy for maternal health and potential reduction of perinatal transmission are great, and given current data, women should be continued on combination regimens if indicated with careful monitoring for potential pregnancy complications.

Transmission of HIV from mother to child accounts for over 90% of pediatric AIDS cases reported in the US.¹ Early in the epidemic before the availability of serologic testing, vertical transmission rates over 50% were reported in the US and Europe. These rates were spuriously high because infected women were identified because of symptomatic disease or previous birth of an infected child and many infected women were not identified. Before the widespread use of antiretroviral therapy in pregnancy, vertical transmission rates of 15–33% were reported in the US and Europe,^{8, 73, 99, 100, 101, 102, 110, 111, 112, 113, 114} and rates of 20–48% were reported in developing countries.^{80, 83, 84, 85, 86, 87, 89, 115, 116} Higher rates of transmission in developing countries are thought to be related to breast feeding transmission; higher rates of cofactors such as sexually transmitted diseases, chorioamnionitis and malaria; nutritional deficiencies; viral differences; and differential classification of infants who die before determination of infection status. More recently vertical transmission rates in the US and Europe of 3–6% with maternal zidovudine therapy^{16, 117, 118, 119, 120, 121, 122} and 2% or less with maternal zidovudine therapy and scheduled cesarean delivery or combination antiretroviral therapy with undetectable maternal viral load have been reported.^{40, 123, 124, 125, 126, 127} Optimal management of pregnant women to prevent perinatal transmission is discussed below. While great progress has been made in reducing the rate of vertical transmission of HIV in developing countries, much remains to be done to translate research findings into interventions in developing countries to reduce rates there. In addition, greater efforts must be undertaken throughout the world to prevent infections in women, the ultimate method to prevent perinatal transmission. 

Vertical transmission of HIV may occur during the antepartum, intrapartum, or postpartum period.  Although it is difficult to differentiate antepartum from intrapartum transmission, data suggestive of antepartum, or in utero, transmission include detection of HIV in fetal tissues from abortuses as early as 8 weeks’ gestation,^{128, 129, 130} isolation of HIV from amniotic fluid,^{131, 132} identification of HIV in placentas from infants born to HIV-infected women,^{133, 134, 135} and the ability to infect trophoblast in vitro.^{136, 137} Based on the rate of detection of virus by culture or antigen or nucleic acid detection at birth compared to later,^{138, 139, 140, 141, 142, 143, 144, 145, 146} the early onset of symptoms related to HIV in a subset of perinatally infected infants,^{147, 148} and patterns of IgA response,^{149, 150} it appears that 20–60% of perinatal transmission not related to breast feeding occurs in utero, with the remaining 40–80% occurring in late pregnancy or during labor and delivery. Evidence for intrapartum transmission includes the virologic, clinical, and immunologic data discussed above, the frequent detection of HIV in blood and cervicovaginal secretions of infected women which is reduced by antiretroviral therapies shown to reduce transmission,^{151, 152, 153} significantly higher infection rates among first-born compared to second-born twins,¹⁵⁴ the association of higher transmission rates with increasing duration of  membrane rupture,^{111, 155, 156} and reduced transmission among women delivered by cesarean section before labor and membrane rupture.^{124, 135} Thus, although transmission may occur early in pregnancy, the majority appears to occur near delivery and interventions directed at late pregnancy and delivery and the neonate would be expected to significantly impact transmission.

 
Transmission of HIV during the postpartum period through breast feeding, and rarely through household contacts, has been documented.^{88, 157, 158, 159, 160} Two cases of household transmission from infected children to care givers have been documented, but otherwise household transmission appears to be rare.¹⁶⁰ HIV has been detected in both the cell-free and cell-associated fractions of breast milk.¹⁶¹ The risk of transmission during breast feeding appears to be higher with primary maternal infection during breast feeding such as with postpartum blood transfusion when maternal viremia is at high levels before an immune response.¹⁵⁸ A meta-analysis showed the average rate of transmission with breast feeding during primary infection to be 29%, range 16–42%.¹⁵⁷ The rate of transmission through breast feeding with established infection is harder to determine because of difficulty differentiating intrapartum and postpartum infections, but ranges from 9 to 32%, with a median around 14%.^{157, 159, 162, 163} HIV is detected more frequently from colostrum and breast milk in the first 7 days after delivery compared to mature milk, suggesting that transmission rates may be higher early in breast feeding.¹⁶¹ Epidemiologic data also suggest higher transmission rates earlier compared to later in breast feeding.^{164, 165} Other factors which increase the risk of transmission during breast feeding include low CD4+ lymphocyte count or high plasma or breast milk HIV RNA in the mother, vitamin A deficiency, and mastitis.^{162, 166, 167} Studies have found higher transmission rates in infants receiving both breast milk and other liquids compared to those receiving only breast milk.^{168, 169} Mixed feeding, especially if contamination occurs could produce inflammation in the gastrointestinal tract which facilitates HIV entry.¹⁷⁰   

 
Avoidance of breastfeeding by HIV-infected women in most of the world is not a viable alternative to reduce transmission. In a randomized trial of breast feeding versus formula feeding Kenya, the HIV infection rates at 24 months of age were reduced from 37% in the breast feeding group to 21% in the formula feeding group despite only 70% compliance with not breast feeding among the women assigned to formula feeding.¹⁷¹ The mortality rate in the children through 2 years of age was high in both groups, 24% in the breast feeding and 20% in the formula feeding arms, despite a requirement for access to running water and restriction of enrollment to certain districts within the city and free provision of formula, suggesting concerns regarding applicability to large areas of the developing world. In addition, maternal mortality was increased among breast feeding compared to formula feeding women.¹⁷² A study in Botswana randomized women to formula feed or breastfeed with infant zidovudine prophylaxis.¹⁷³ While the HIV infection rates in the infants at 7 months of age were lower in the formula fed group (5.6%) compared to the breastfed/zidovudine group (9.0%, p = 0.04), the cumulative infant mortality rate, primarily related to diarrhea and pneumonia, at 7 months was significantly higher for the formula-fed group (9.3% versus 4.9%, p = 0.003), underscoring the risks of formula feeding. By 18 months, mortality was similar between the two groups. A study comparing exclusive breastfeeding with early, abrupt weaning at 4 months to breastfeeding duration of the mother’s choice in Zambia showed no difference overall in the rate of HIV-free survival at 24 months between groups, but among infants who were HIV-infected by 4 months (diagnosed in retrospect), mortality was significantly higher by 24 months with early weaning (73.6%) compared to continued breastfeeding (54.8%, p = 0.007).¹⁷⁴ Thus, formula feeding and early weaning did not improve HIV-free survival by 2 years in several African settings.

 
More recently, strategies of infant prophylaxis during breastfeeding have met with more success. In a combined analysis of studies from Uganda, Ethiopia, and India, evaluating 6 weeks of infant nevirapine during breastfeeding compared to peripartum prophylaxis only showed a significant reduction in infant HIV infection at 6 weeks of age with continued nevirapine, but the difference was no longer significant by 6 months with continued breastfeeding.¹⁷⁵ In a study in Malawi, infants were randomized to receive single-dose nevirapine and 1 week of zidovudine or this regimen plus nevirapine for 14 weeks or the control regimen plus nevirapine and zidovudine for 14 weeks.¹⁷⁶ Both extended prophylaxis groups had lower rates of transmission (5.2% nevirapine, 6.4% nevirapine plus zidovudine) compared to the control group (10.6%, p < 0.002) at 9 months, with fewer adverse events in the nevirapine only group. Thus, infant nevirapine prophylaxis during breastfeeding appears to reduce the risk of HIV acquisition while accruing the benefits of breastfeeding. Studies are ongoing evaluating the use of maternal antiretroviral therapy during breastfeeding to reduce transmission to the infant and comparing maternal therapy to infant prophylaxis.

Maternal and obstetrical factors consistently associated with an increased risk of transmission when evaluated in the period before widespread use of antiretroviral therapy in pregnancy included advanced maternal disease as defined by clinical AIDS, maternal p24 antigenemia, low CD4+ lymphocyte count/percentage, or high plasma viral load,^{72, 73, 81, 84, 85, 90, 99, 101, 112, 113, 114, 115} longer duration of membrane rupture during labor,^{111, 155, 156} placental inflammation,^{81, 85, 90, 102, 115, 177, 178, 179} and concomitant sexually transmitted diseases.^{84, 96, 102, 115} Factors associated with increased risk of transmission in some but not all studies when evaluated include preterm birth,^{90, 112, 180, 181} illicit drug use,^{100, 102, 182, 183} maternal vitamin A deficiency,^{184, 185, 186} female gender of the infant,¹²³ and mode of delivery.^{100, 110, 112, 187, 188, 189, 190, 191} The impact of cesarean delivery was difficult to evaluate in early studies because of lack of differentiation of cesarean sections done before labor and membrane rupture compared to those done after labor or rupture, variations in management of labor such as use of scalp electrodes and operative vaginal delivery, and lack of differentiation of antepartum compared to intrapartum transmission. 

Data from both untreated and treated cohorts of HIV-infected pregnant women have shown an association between increased maternal viral load and an increased risk of vertical transmission, although this association is attenuated among treated women.^{113, 117, 177, 181, 183, 192, 200, 201} Treatment lowers transmission even among women with plasma HIV RNA below 1000 copies/mL before treatment.²⁰¹ Although transmission rates are very low among women with HIV RNA levels below 1000 copies/mL, there is no threshold below which transmission can be certain not to occur.

Antiretroviral therapy for reduction of vertical transmission

The most striking finding regarding vertical transmission of HIV was the demonstration in 1994 of the reduction of transmission from antepartum, intrapartum, and neonatal therapy with zidovudine in the PACTG 076 trial.²⁰² Several subsequent observational studies confirmed the benefit of zidovudine in reduction of transmission and actually reported rates of transmission in treated cohorts of 3–6%.^{16, 118, 119, 120, 121, 122, 126, 183, 203, 204} Recently, several studies of shorter courses of zidovudine (ZDV) or nevirapine have shown benefit in reducing perinatal transmission even among breast feeding populations (Table 2).^{17, 40, 66, 173, 200, 202, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217} Intrapartum therapy only, even with ZDV/lamivudine (3TC), was not better than placebo, while intrapartum/postpartum ZDV/3TC was better than similar ZDV alone, and similar to the two-dose nevirapine regimen.^{66, 211} Short course antepartum/intrapartum zidovudine regimens seem to have similar efficacy to the intrapartum/postpartum ZDV/3TC regimen, but are cheaper and do not have the potential added toxicity of combination regimens, if therapy can be begun by 36 weeks of gestation. The nevirapine regimen is the simplest, least expensive, and offers the advantage of both maternal and infant doses being administered in the hospital under direct observation. Development of resistance with even single dose nevirapine has been reported²¹⁸ and is discussed further below. While intrapartum and neonatal nevirapine reduced transmission compared to short-course zidovudine, addition of this regimen to established antiretroviral therapy including at least the PACTG 076 zidovudine regimen did not reduce transmission further.⁴⁰ The observation that short courses of therapy during pregnancy or postpartum can lead to sustained reductions in transmission even up to 24 months in breast feeding populations is encouraging. 

Table 2. Completed randomized trials of antiretroviral drugs to reduce perinatal transmission of HIV

ZDV, zidovudine; NVP, nevirapine; 3TC, lamivudine; IV, intravenous; QID, four times/day; BID, twice daily.

*Results are for the combination of this and the following study.

Despite the success in reducing transmission with maternal and infant single dose nevirapine, resistance remains a concern. Resistance rates of 15–69% among women exposed to one or two doses of peripartum nevirapine and 8–87% in exposed infants have been reported.³³ Factors which increase the risk of resistance include multiple doses, higher HIV RNA levels, viral clade, especially type C, and the timing and type of testing. Several studies have evaluated the response to subsequent non-nucleoside reverse transcriptase inhibitors (NNRTI)-based antiretroviral therapy comparing women who received intrapartum nevirapine to those who did not.^{219, 220} Response to therapy did not differ between women initiating antiretroviral 6 or more months after delivery, but the virologic failure was increased if therapy was started within 6 months of delivery. Women requiring therapy within 6 months of delivery would meet WHO guidelines for treatment with highly active antiretroviral therapy (HAART) regimens during pregnancy and should be offered combination regimens during pregnancy rather than single dose nevirapine regimens. The response to antiretroviral therapy among infants who became infected despite maternal/infant nevirapine regimens is currently being evaluated. 

 
Several regimens have been shown to reduce the risk of development of nevirapine resistance when given to the mother during the immediate postpartum period, including 4–7 days of ZDV/3TC, and a single dose of tenofovir/emtricitabine, each of which reduced the rate of resistance in the mothers by about 50%.^{221, 222} In a small study which used ZDV from 28 weeks with single dose nevirapine and added 1 week of tenofovir/emtricitabine after delivery, no resistance to ZDV, nevirapine, tenofovir, or emtricitabine was seen at 4 weeks' postpartum.²²³ Thus if single dose nevirapine is being used, a dual nucleoside “tail” regimen for several days postpartum for the mother should be added to reduce the risk of resistance.

In addition to the randomized trials of short course therapy in developing countries, two observational studies in the US have suggested benefit from zidovudine initiated in the intrapartum or the newborn period. A study of over 900 infants in New York state observed transmission rates of 6.1% when zidovudine was begun during pregnancy before labor, 10.0% when begun during labor and given to the neonate, 9.3% when neonatal therapy was begun within 48 hours of birth, 18.4% when begun after 48 hours after birth, and 26.6% if no zidovudine was given.¹⁶ Similarly, a study from North Carolina demonstrated a transmission rate of 3.1% with the full PACTG 076 zidovudine regimen, 10.7% with intrapartum and neonatal zidovudine, 26.7% with neonatal therapy only, timing not specified, and 30.9% with no antiretroviral therapy in pregnancy.¹²⁰ These data suggest some efficacy for zidovudine especially if begun in the intrapartum period and show lower transmission rates than the zidovudine arm in the HIVNET 012 study. This difference could be related to breastfeeding in the Ugandan study, 6 weeks of infant therapy in the US compared to 1 week in the HIVNET trial, or to other factors. All of these studies underscore the need for HIV testing early in pregnancy, the possible benefits of rapid HIV testing in labor for untested women, and the need to offer therapy with zidovudine, nevirapine, or the untested combination for HIV-infected women presenting in labor or detected after delivery with no therapy during pregnancy.

Combination antiretroviral therapy has become the standard of care for nonpregnant adults with indications for treatment of their HIV infection because of better viral suppression and decreased chance for development of resistant virus.³⁴ Given the strong association of HIV levels and transmission and the greater reduction of viral load with combination compared to monotherapy, combination therapy during pregnancy would be expected to further reduce perinatal transmission. The most recent data from the US and Europe demonstrate transmission rates of 1.2–1.7% among women receiving highly active antiretroviral regimens during pregnancy and at least intravenous zidovudine intrapartum with zidovudine in the neonate.^{40, 126, 127, 224} Highly active antiretroviral therapy is now recommended in the US for all pregnant women to reduce the risk of perinatal transmission. Pregnant women should be apprised of the potential benefits of combination antiretroviral therapy for their health, the impact on perinatal transmission, as well as the potential risks of the therapy including the side effects of the drugs and unknown short and long term effects on the infant. For women who receive HAART during pregnancy for prevention of transmission but do not have indications for treatment for their own health, they may consider stopping therapy after delivery. The long term risks and benefits of stopping compared to continuing this therapy after delivery are unclear and are under study. 

While routine use of HAART has been implemented for preventing mother to child transmission (PMTCT) in developed countries as the standard of care and PMTCT rates of under 2% have been reported,^{40, 126, 127} in resource-limited settings, the World Health Organization (WHO) currently recommends short-course AZT plus intrapartum single dose nevirapine (sdNVP) plus newborn sdNVP plus 1 week infant AZT (ZDV) for HIV-infected pregnant women who do not yet require HAART for their own care.²²⁵ In studies of short course AZT (in some cases with lamivudine added) with sdNVP, transmission rates of 1.1–3.9% at 6 weeks of age have been reported, even when including all pregnant women regardless of CD4+ lymphocyte count.^{212, 226, 227} In several studies in developing countries where HAART regimens were provided for all pregnant women regardless of CD4+ lymphocyte count, transmission rates at 4–6 weeks of age ranged from 1.2 to 4.1% among breastfeeding infants.^{228, 229, 230} In the Kisumu Breastfeeding Study, the transmission rate at 6 weeks among women with CD4+ lymphocyte counts above 350 cells/μL who received HAART was 3.8% and at 12 months was 5.8%.²³¹ While it is difficult to compare data between studies because of differences in populations, breastfeeding rates and approaches, antiretroviral agents available, and obstetrical management, transmission rates are similar in these studies from resource limited settings with short course AZT/sdNVP and HAART.

While the effect of zidovudine therapy on transmission has been demonstrated in both randomized and observational studies, concerns about the impact of viral resistance on the efficacy of antiretrovirals in pregnancy remain. In the PACTG 076 study, high level zidovudine resistance was not detected among 96 women who received zidovudine.²³² Low-level zidovudine resistance (K70R mutation) developed in one (2.6%) of the 39 women with paired isolates from enrollment and delivery. Zidovudine resistance was not associated with an increased risk of vertical transmission of HIV. Women enrolled to PACTG 076 were required to have CD4+ lymphocyte counts above 200 cells/μL and had relatively low viral loads, thus limiting the chance of development of resistant virus. Among a cohort of women receiving zidovudine for their own health during pregnancy before release of the PACTG 076 results, genotypic resistance to zidovudine was more commonly detected.²³³ At least one resistance mutation was observed in 34 (25%) of 142 isolates and either the codon 215 mutation or more than one mutation (potentially high level resistance) was observed in 14 of these. Perinatal transmission occurred among eight (24%) of 34 with any resistance mutation and 18 (19%) of 96 without any resistance mutation (p = 0.7). On multivariate analysis after adjusting for duration of ruptured membranes and total lymphocyte count, resistance mutations were associated with an increased risk of transmission, although this finding may be related to the increased risk of resistance mutations among women with high viral loads. The Swiss HIV in Pregnancy Study reported codon 215Y/F zidovudine resistance among six (9.6%) of 62 consecutive women, four of these at initial sampling in pregnancy (three of these with prior zidovudine), and two developing during pregnancy.²³⁴ None of the women with this high level zidovudine resistance transmitted virus to their infants despite the majority receiving the 076 zidovudine regimen. A study from New York detected codon 215 mutations in one of 33 women delivering before 1997 but in three (9.7%) of 31 women delivering between 1997 and 1999.²³⁵ All three had previous zidovudine exposure. In the multicenter Perinatal AIDS Collaborative Transmission Study, primary or secondary NRTI mutations were detected among 22% of the subset of women with amplifiable virus and protease mutations among 0.5%.²³⁶ The detection of zidovudine or other resistance mutations was not associated with an increased risk of perinatal transmission in PACTG 076, PACTG 185, the Swiss study, or the PACTS.^{232, 234, 235, 236} Perinatal transmission of drug resistant virus has been reported,^{237, 238} but the bulk of the evidence to date does not suggest viral resistance mutations increase the risk of transmission. Rapid development of lamivudine resistance has been reported among pregnant women receiving dual nucleoside therapy including lamivudine, suggesting that if combination therapy is chosen during pregnancy, triple combinations including a protease inhibitor or non-nucleoside reverse transcriptase inhibitor as recommended for non-pregnant adults should be used to maximize suppression and minimize resistance.²³⁹ Resistance testing should be used for pregnant women as in nonpregnant adults, i.e., those who do not respond to initial highly active therapy, who have persistently detectable virus and a history of multiple therapeutic regimens, or where the prevalence of resistant virus in the community is high.³⁴ 

Despite concerns regarding potential long term effects of antiretroviral exposure during pregnancy and the neonatal period on growth and development, the benefits of antiretroviral therapy both to maternal health and reduction of perinatal transmission appear to outweigh the as yet unproven risks. Current guidelines should be consulted to assure the most up to date information as new drugs are approved and new data become available.³³

Mode of delivery and vertical transmission

As discussed above, at least half of the incidents of vertical transmission of HIV appear to occur during the intrapartum period. Early studies of the role of cesarean section on the rate of transmission were inconclusive because of combining of cesarean sections done before and after labor and membrane rupture.^{112, 187, 188, 189, 190} However, several studies have now demonstrated the benefit of scheduled cesarean section in reducing vertical transmission of HIV among women on no antiretroviral therapy or zidovudine monotherapy. Two prospective cohort studies demonstrated rates of transmission under 2% among women who received zidovudine and underwent cesarean section before labor and membrane rupture (scheduled cesarean section).^{123, 191} Subsequently, an individual patient meta-analysis pooling data from 15 North American and European cohorts demonstrated a significantly lower transmission rate among infants delivered by scheduled cesarean section compared to urgent cesarean or vaginal delivery, with an unadjusted odds ratio of 0.45 (95% CI 0.35–0.58) and an odds ratio of 0.43 (0.33–0.56) adjusted for zidovudine use, advanced maternal disease, and birth weight.¹²⁴ The rate of transmission among women not receiving antiretrovirals was 10% when delivered by scheduled cesarean section and 19% for other modes of delivery, and among women receiving zidovudine (predominantly the 076 schedule), the rate was 2% with scheduled cesarean and 7% with other modes of delivery. Finally, results from an international randomized trial confirmed a reduction in rate of transmission with scheduled cesarean section. In an analysis of the assigned mode of delivery, transmission was 1.8% among those assigned to scheduled cesarean delivery and 10.5% in those assigned to deliver vaginally (OR 0.2, 95% CI 0.1–0.6).¹²⁵ Results were similar in the actual mode of delivery analysis with women undergoing emergent cesarean section having a rate of transmission of 8.8%, similar to those delivering vaginally.  In an analysis including zidovudine use, the transmission rate was 4% with scheduled cesarean and 20% for vaginal delivery without zidovudine use (AOR 0.2, 95% CI 0–0.8), and 1% with scheduled cesarean and 4% with vaginal delivery with zidovudine use (AOR 0.2, 95% CI 0–1.7). Because of the small number of transmissions (six total) in the group receiving zidovudine, the difference in transmission rates by mode of delivery was not significant, although the magnitude of reduction was similar. Taken together with the meta-analysis and observational data, the trial suggests that scheduled cesarean section offers added protection from transmission among women not receiving antiretroviral therapy or receiving only zidovudine monotherapy. More recent data from the UK found no difference in transmission rates between women on HAART delivered by scheduled cesarean section (0.7%, n = 2337) or planned vaginal delivery (0.7%, n = 565).¹²⁷ No transmissions occurred among 467 women with HIV RNA levels below 10,000 copies/mL and treated with antepartum and intrapartum zidovudine and delivered by scheduled cesarean delivery. Thus, there is no clear benefit of cesarean delivery among women on HAART with suppressed viral load, and scheduled cesarean delivery is not recommended for women with HIV RNA below 1000 copies/mL.²⁴⁰ Scheduled cesarean delivery may be of benefit for women with high HIV RNA levels because of late care or inadequate treatment or among those who choose ZDV monotherapy over HAART.

Other interventions to reduce perinatal transmission of HIV

Given the high seroprevalence of HIV among pregnant women in developing countries and the lack of resources to provide routine prenatal care, HIV testing, and antiretroviral therapy, interventions which are easy to implement, inexpensive, and applicable to all pregnant women regardless of HIV status have been sought. Interventions which have been tested thus far and not found to impact significantly on perinatal transmission of HIV include chlorhexidine vaginal washing,^{241, 142} benzalkonium chloride vaginal suppositories in late pregnancy and labor,²⁴³ multivitamin supplementation,²⁴⁴ vitamin A supplementation,^{245, 246} selenium supplementation,²⁴⁷ and short course antibiotics at 20–24 weeks of gestation and during labor to reduce chorioamnionitis.²⁴⁸ A subset analysis of one of the chlorhexidine studies did show a reduction in transmission in the chlorhexidine group among the women with ruptured membranes for more than 4 hours before delivery and also demonstrated reductions in maternal and infant sepsis and hospital admissions, and infant mortality from sepsis.²⁴⁹ Likewise, the multivitamin supplementation trial demonstrated significantly lower rates of fetal death and stillbirth, low birth weight, and preterm delivery despite no reduction in HIV transmission.²⁴⁴ The vitamin A trial did not demonstrate a difference in fetal or infant mortality but did show a reduction in preterm birth, and among the preterm infants, a reduction in HIV transmission with vitamin A supplementation.²⁴⁵ One vitamin A trial detected an increased risk of HIV transmission among infants born to women receiving vitamin A (OR 1.38), but the mechanism for this inconsistent finding is unclear.²⁴⁴ 

Evaluation of the infant born to an HIV-infected mother

Infants born to HIV-infected women should receive the 6-week zidovudine regimen (2 mg/kg orally every 6 hours) as outlined in PACTG 076.²⁰² Anemia is the most frequent toxicity of the s6-week zidovudine regimen, although suppression of other bone marrow elements is also possible. A complete blood count and differential should be done on the neonate as a baseline and repeat hemoglobin should be done at 6 and 12 weeks of age.³⁴ Potential toxicities in the neonate of combination therapy in the mother would be expected to be similar to those in adults, and the infant should be monitored accordingly.

Because of transfer of maternal IgG antibody across the placenta, serologic testing of the infant is not useful for diagnosis of HIV infection in the first several months of life. Standards of care for diagnosis and management of the infant born to an HIV-infected mother should be followed, which currently include testing by DNA PCR or other virologic assay at 14–21 days, 1–2 months, and 4–6 months of age, with an additional test at birth being optimal.²⁵⁰ Positive testing at or before 48 hours of age is felt to be indicative of infection in utero, while negative testing initially followed by positive testing at or beyond 2 weeks of age is felt to indicate transmission in late pregnancy or intrapartum, assuming no exposure through breast feeding. Most infants subsequently proven to be infected will have a positive DNA PCR test by 2 weeks of age, so that testing at this point may allow early intensification of antiretroviral therapy rather than continuing zidovudine prophylaxis.^{250, 251} Diagnostic tests should be repeated after completion of infant antiretroviral prophylaxis in the infant with earlier negative testing. Some experts also recommend confirmation of negative HIV antibody status in the infant at 12–18 months of age after negative virologic tests. Infants born to HIV-infected women should be started on prophylaxis against Pneumocystis carinii pneumonia (PCP) at 6 weeks of age after completion of zidovudine prophylaxis unless HIV infection has been presumptively ruled out with negative testing at 2–3 weeks and 4–6 weeks of age.²⁵⁰ PCP prophylaxis should be initiated in HIV-infected infants at the time of diagnosis. Infants with proven HIV infection should be managed by infectious disease specialists with experience in treating pediatric HIV infection, according to current recommendations.²⁵⁰

Antepartum care

Women entering pregnancy with a history of HIV infection or with a positive antibody test for the first time during pregnancy should have their status confirmed by repeat antibody testing. Evaluation of the HIV-infected pregnant woman is summarized in Table 3. Included in the history should be documentation of previous and current antiretroviral drug therapy and whether or not previous children have been assessed for possible HIV infection. Immunization history should be assessed including hepatitis B and pneumococcal vaccine status. Pregnant women without evidence of immunity to hepatitis B should be offered vaccination, and HIV-infected women who have not received pneumococcal vaccine within the last 5 years should be offered this vaccine. Influenza vaccine should be offered to all pregnant women who will be pregnant during influenza season.²⁵² Because of the transient increase in plasma HIV levels which can be seen after immunization and the theoretical concern that this increase may increase the risk of perinatal transmission, pregnant women should be on adequate antiretroviral therapy before immunizations. Cervical cytology, and if indicated Neisseria gonorrhoeae and Chlamydia trachomatis testing, should be collected with the pelvic examination. Baseline neurologic function and fundoscopic examination should be documented. Although not routinely recommended for HIV-uninfected pregnant women, antibody status to Toxoplasma gondii, cytomegalovirus (CMV), and herpes simplex virus should be documented in the HIV-infected woman if not previously done to assess maternal risk of symptomatic infection and potential fetal risk. In rare instances, transmission of T. gondii or CMV from an HIV-infected mother with previously documented antibody to the specific pathogen has been documented in women with severe immunosuppression.^{253, 254} This transmission should be avoided by early initiation and optimization of maternal antiretroviral therapy. HIV-infected women are at increased risk for genital herpes seropositivity and potentially reactivation.²⁷² Tuberculin testing should be done if not done within the past year.   

Table 3.  Evaluation of the HIV-positive pregnant woman

Once the CD4+ lymphocyte count and viral load results are available, women can be counseled regarding antiretroviral therapy for maternal health and prevention of transmission. Women who are on a stable antiretroviral regimen with suppression to undetectable viral levels at the time of pregnancy diagnosis should continue on this regimen unless drugs with specific concerns for teratogenicity such as efavirenz or hydroxyurea are included. If women are initiating antiretroviral therapy during pregnancy, they may choose to wait until after the first trimester to minimize risk during organogenesis. This decision depends on maternal disease stage as measured by viral load and CD4+ lymphocyte count and patient preference. Women with a CD4+ lymphocyte count below 350 cells/μL should be encouraged to begin a highly active antiretroviral regimen as outlined for non-pregnant individuals to maximize their health and reduce the risk of perinatal transmission.^{33, 34} These regimens usually consist of two or more nucleoside agents with either a protease inhibitor or non-nucleoside reverse transcriptase inhibitor. Stavudine and didanosine should not be used together in pregnancy unless other agents are contraindicated because of potentially enhanced mitochondrial toxicity in pregnancy.³³ Nevirapine should not be initiated in women with CD4+ lymphocyte counts above 250 cells/μL because of increased risk of hepatic and skin toxicity probably related to hypersensitivity.³⁸ Unless contraindicated because of previous toxicity or known viral resistance, zidovudine should be included in the regimen. If stavudine is included in a combination regimen during pregnancy, then zidovudine should be avoided because of the possible antagonism. Oral stavudine should be discontinued during labor and intravenous zidovudine given if not contraindicated. For women with CD4+ lymphocyte counts above 350/mL, highly active antiretroviral therapy should be recommended. Use of highly active therapy to suppress the viral load below 1000 copies/mL, and ideally to undetectable levels, may further reduce the risk of perinatal transmission and obviate the need for cesarean delivery to reduce intrapartum transmission. The indications for and patient preferences for continuing therapy can then be reassessed after delivery in this group. If patients in this category decline combination therapy, then zidovudine monotherapy should be offered. Time-limited use of zidovudine alone for prophylaxis of transmission is controversial, but no increase in progression was seen in two follow up studies of women treated during pregnancy.^{256, 257} Development of viral resistance was unusual among the women enrolled to PACTG 076 who all had CD4+ lymphocyte counts above 200 cells/μL at enrollment.²³² For women with higher viral loads and lower CD4+ cell counts, highly active antiretroviral therapy should be encouraged to minimize development of resistance, maternal disease progression, and perinatal transmission. Response to therapy should be monitored with repeat viral load testing 4 weeks after beginning or changing therapy. Once viral load has been suppressed to undetectable levels, levels can be monitored every 3 months. Current guidelines should be followed for changing therapy if poor response or rebound occurs.³⁴ Women should be monitored on a regular basis for toxicity as indicated for specific drugs. A baseline sonogram should be done at 18–20 weeks of gestation to verify dating and rule out major anomalies. Follow up scans in the third trimester should be considered to monitor fetal growth among women on antiretroviral therapy.

Adherence to the multidrug antiretroviral regimens currently in use may be more difficult in pregnancy, especially in the first trimester and the early postpartum period. If nausea and vomiting lead to the need for temporary discontinuation of medications, all antiretrovirals should be discontinued and re-instituted simultaneously to minimize the chance for development of resistance. To maximize adherence, time should be spent educating the patient regarding the goals of therapy and consequences of good compared to poor adherence.³⁴ The regimen should be as simple as possible and chosen to minimize potential adverse effects. The woman’s concerns regarding potential fetal effects should be addressed before starting therapy. Indications for and patient’s ability to comply with therapy should be re-assessed after delivery since the loss of reduction of perinatal transmission as an impetus and the demands of newborn care may make adherence more difficult. Discontinuation of therapy for a period rather than continuation with poor adherence may decrease the chance for development of resistance and loss of agents for future therapy.

In general, prophylaxis for opportunistic infections in pregnancy should be used as it would be in nonpregnant women.²⁵⁸ Ideally, antiretroviral therapy would be initiated early enough to maintain the CD4+ cell count at levels above which prophylaxis is indicated, but the same CD4+ lymphocyte counts or percentages should be used as indications for therapy in the pregnant woman. Likewise, current adult guidelines should be used to guide decisions regarding discontinuation of prophylaxis after immune restoration on antiretroviral therapy. Primary prophylaxis for P. carinii pneumonia should be offered to women with CD4+ lymphocyte counts below 200 cells/μL or history of oropharyngeal candidiasis. Secondary prophylaxis should be offered for all women with previous P. carinii pneumonia unless immune reconstitution has occurred. Trimethoprim-sulfamethoxazole, one double-strength tablet daily, is first choice for prophylaxis in pregnancy. Alternatives include aerosolized pentamidine, 300 mg monthly via Respirgard II nebulizer, or oral dapsone 100 mg daily. Prophylaxis for Mycobacterium avium complex should be offered for CD4+ lymphocyte counts less than 50 cells/μL.  Azithromycin, 1200 mg once weekly, is the first choice for prophylaxis during pregnancy. Trimethoprim-sulfamethoxazole will also provide prophylaxis against toxoplasma encephalitis in women who are seropositive for antibodies to T. gondii. Since toxoplasma encephalitis is rarely encountered in pregnancy, other primary prophylactic regimens including pyrimethamine are best avoided until after delivery. For women with previous toxoplasma encephalitis, an appropriate prophylaxis regimen should be offered throughout pregnancy unless immune restoration to a CD4+ lymphocyte count above 200 cells/μL for at least 3 months has occurred.²⁵⁸ Women with a positive tuberculin skin test without prior treatment, or contact with active tuberculosis with no evidence of active disease themselves may receive isoniazid or rifampin prophylaxis during pregnancy. For women with no evidence of active tuberculosis but exposure to multidrug resistant tuberculosis, prophylactic therapy may best be deferred until after delivery. For treatment of active tuberculosis during pregnancy especially for multidrug resistant tuberculosis, the regimen should be developed in consultation with obstetric and infectious disease specialists. Primary prophylaxis for other conditions including mucosal candidiasis and other fungal infections is best avoided during pregnancy. Treatment of invasive fungal disease should be provided as it would be for a non-pregnant individual except that amphotericin B is preferred over high dose azole therapy in the first trimester because of concerns of teratogenic potential.²⁵⁸ Likewise, prophylaxis for CMV disease is not recommended during pregnancy because of the potential toxicity of the drugs and limited experience with their use in pregnancy. However, for women with life-threatening or sight-threatening CMV infections during pregnancy, treatment should be provided in consultation with obstetric and infectious disease specialists.

Intrapartum management

Mode of delivery should be determined by the woman in concert with her provider taking into account current therapy, most recent plasma HIV RNA results, and other obstetric concerns. For women who choose not to take antiretroviral therapy or who choose zidovudine monotherapy, scheduled cesarean section appears to offer benefit in reducing intrapartum transmission of HIV. For women who maintain plasma HIV RNA levels over 1000 copies/mL despite antiretroviral therapy, planned cesarean section may also reduce the risk of transmission. This information should be explained to the woman and her decision regarding mode of delivery should be respected as cesarean section clearly involves increased risks to the mother.

In HIV-uninfected women, cesarean section is associated with an increase in morbidity and mortality compared to vaginal delivery.²⁵⁹ Some of these complications are related to the indications for cesarean section but much of the morbidity is related to peripartum infections. Scheduled cesarean section has a lower risk of infectious complications than that performed after labor or rupture of membranes, but HIV may confer an increased risk. Studies evaluating the risk of complications by mode of delivery among HIV-infected women suggest a similar increase in magnitude of complications related to cesarean section as that seen in HIV-uninfected women.^{125, 260, 261, 262} In addition to the studies describing morbidity among HIV-infected women, several studies have compared postoperative morbidity after cesarean delivery between HIV-infected and HIV-uninfected women.^{263, 264, 265, 266, 267, 268, 269, 270, 271} Many of these studies were retrospective. Seven of the nine studies detected an increased risk of one or more complications, most often postoperative fever or antibiotic use or specific infections among the HIV-infected compared to the HIV-uninfected women. An increased risk of postoperative pneumonia among the HIV-infected women was a consistent finding. Where evaluated, the risk of postoperative fever or infection was increased consistently among women with lower CD4+ lymphocyte counts, also the group with the most potential benefit in reduction of transmission from cesarean delivery. Thus, HIV-infected women undergoing cesarean section should be monitored closely for development of pneumonia or other infectious complications, especially those women with low CD4+ lymphocyte counts.

Women should be counseled about the available data on the potential benefits of cesarean section to the infant and the risks to themselves. Antiretroviral therapy should be optimized for both maternal health and reduction of vertical transmission. While the risk of transmission increases with an increasing duration of membrane rupture in the absence of antiretroviral therapy, data are not available to allow assessment of any benefit in reduction of transmission if cesarean section is done after labor or membrane rupture. If a scheduled cesarean delivery is chosen, this procedure may be scheduled at or after 38 completed weeks of gestation to minimize the chance of labor or rupture occurring before the procedure.²⁴⁰ If cesarean delivery is planned, intravenous zidovudine should be begun at least 3 hours before the procedure. Other antiretroviral medications should be continued orally as practical. The use of prophylactic antibiotics to reduce the risk of endometritis and wound infection should be considered although the use of antibiotic prophylaxis in this setting has not been studied.

If vaginal delivery is chosen, intravenous zidovudine should be begun at the onset of labor, giving 2 mg/kg over 1 hour, followed by a 1 mg/kg infusion until delivery. Other antiretroviral medications except for stavudine should be continued during labor. The membranes should be kept intact as long as possible and use of scalp electrodes, forceps, and vacuum extraction should be avoided if possible. Given the increased risk of genital herpes reactivation among HIV-infected women, a careful examination of the perineum and cervix should be performed to detect genital lesions at the onset of labor.²⁵⁵

HIV-infected women presenting in preterm labor should receive standard tocolytic and corticosteroid therapy and be continued on antiretroviral therapy. If delivery is indicated, mode of delivery should be determined by obstetric indications, most recent HIV RNA level, and maternal wishes. For HIV-infected women presenting with preterm rupture of membranes, management must be individualized based on gestational age and clinical status such as presence of contractions and HIV RNA level. If expectant management is indicated based on early gestational age, antiretroviral therapy should be maximized to suppress HIV RNA levels.

To minimize the risk of infant inoculation with HIV, the infant should be thoroughly washed before any injections or invasive tests such as glucose levels are performed. Routine nasal gastric suction should not be performed. 

Postpartum care

As discussed above, transmission of HIV by breastfeeding has been well-documented. In areas where safe alternatives are available, HIV-infected women should not breastfeed. Routine postpartum care should be provided. HIV-infected women, especially those undergoing cesarean section may be at increased risk for pneumonia and should be evaluated carefully if fever develops. The postpartum period is normally one of emotional lability for many women. HIV-infected women have the added emotional burden of the uncertainty of their infants’ infection status and guilt over possible transmission. In addition, adherence to antiretroviral regimens may be especially difficult with the added demands of newborn care. Thus, adequate physical and psychosocial support for the HIV-infected woman and her newborn must be ensured.

Adequate continued medical follow up of the woman with assessment of indications for continued antiretroviral therapy and ongoing HIV and gynecologic care must be provided. Infant follow up for routine pediatric care, assessment of HIV infection status, and long term toxicity monitoring must be assured. Contraceptive plans after delivery should be discussed with the woman during pregnancy, and contraception should be provided before hospital discharge. Barrier methods of contraception are recommended to prevent HIV transmission and acquisition of other sexually transmitted diseases. If hormonal contraception is being considered, potential interaction with antiretroviral and opportunistic infection prophylaxis drugs must be assessed.³⁴ Currently licensed nucleoside reverse transcriptase inhibitors do not appear to have significant interactions with hormonal contraceptives, but ritonavir, nelfinavir, and possibly amprenavir decrease ethinyl estradiol levels significantly. Efavirenz and indinavir increase estradiol levels although the significance of this finding is unclear. Rifampin and rifabutin lower hormonal levels and decrease contraceptive efficacy of oral contraceptives. No clinically significant pharmacokinetic interactions between depot medroxyprogesterone and nelfinavir, efavirenz, or nevirapine were found in a recent study.²⁷² While the risk of pelvic infections and excessive bleeding among HIV-infected women using an intrauterine contraceptive device has been theorized to be increased, data have not supported an increased risk.²⁷³ Intrauterine devices can be offered to carefully selected HIV-infected women without severe immunocompromise as they are to HIV-uninfected women. Sterilization can be offered as for HIV-uninfected women.

HIV-infected pregnant women will continue to present challenges to the obstetrician/gynecologist. While primary prevention of HIV infection in women is the ultimate goal, the obstetrician must also be able to offer state of the art care to maximize the health of the mother and her fetus and infant. Providers should utilize the most current information available to provide the best care possible.