NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Chou R, Smits AK, Huffman LH, et al. Screening for Human Immunodeficiency Virus in Pregnant Women [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2005 Jul. (Evidence Syntheses, No. 39.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Screening for Human Immunodeficiency Virus in Pregnant Women

Screening for Human Immunodeficiency Virus in Pregnant Women [Internet].

Show details

3Results

Key Question 1. Does Screening for HIV in Asymptomatic Pregnant Women Reduce Mother-to-Child Transmission or Premature Death and Disability?

We identified no randomized trials or observational studies comparing clinical outcomes from screening or not screening pregnant patients in the general population. Although the number of infants with perinatally acquired HIV transmission has markedly declined in the U.S., this is probably due to a combination of increased screening during pregnancy and increased development and acceptance of interventions to prevent transmission, and some HIV-positive women may have been identified before their pregnancy.5, 18 We identified no studies estimating the relative impact of these factors on transmission rates.

Key Question 2. Can Clinical or Demographic Characteristics (Including Persons in Specific Settings) Identify Subgroups of Asymptomatic Pregnant Women at Increased Risk for HIV Infection Compared to the General Population of Pregnant Women?

Risk factors for HIV infection appear to be similar in pregnant and non-pregnant women and are largely unchanged since 1996. The 1996 USPSTF recommendations defined persons at increased risk of HIV infection as those seeking treatment for sexually transmitted diseases; past or present injection drug users; persons who exchange sex for money or drugs and their sex partners; women whose past or present sex partners were HIV-infected, bisexual, or injection drug users; and persons with a history of transfusion between 1978 and 1985.56 Current CDC guidelines also consider unprotected vaginal or anal intercourse with more than one sex partner a high-risk behavior.17 Late or no prenatal care has also been associated with a higher risk for HIV infection.69

A large study of 73,472 women tested at U.S. federally funded prenatal or obstetrics clinics found that 0.6% were positive for HIV.70 Smaller studies of pregnant women reported prevalence rates ranging from 0.13% to 5%.7, 71, 72 In the U.S., there are regional variations in the prevalence of HIV infection, and HIV-positive women are more likely to be African-American or Hispanic.10 Heterosexual transmission has replaced intravenous drug abuse as the most common route of HIV infection among American women. In 30 U.S. areas with confidential name-based reporting, for example, 79% (5,949 of 7,503) women diagnosed with HIV or AIDS in 2002 identified heterosexual contact as their risk factor.10 More than 25% of U.S. women with AIDS reside in smaller cities and rural areas.73

Targeted screening of HIV-positive pregnant women based on the presence of risk factors may miss a substantial proportion of infected persons.74 Not screening HIV-infected women unaware of their status due to lack of reported risk factors could result in missed opportunities for perinatal HIV prevention and other interventions. Observational studies in high- and low-prevalence settings (all published prior to 1996) found that between 8%75 and 57%76 of HIV-infected pregnant women reported identifiable risk factors.7781 Changes in the criteria used to define high-risk behaviors and varying stringency of risk factor assessment, however, complicate interpretation of these results. In two studies, for example, the number of sexual partners was not determined, even though current CDC guidelines consider unprotected intercourse with more than one sexual partner a high-risk behavior.75, 79 In another study, more detailed risk assessment following testing identified substantially more high-risk behaviors than pre-test risk assessment.77

Prior to 1995, HIV screening was routinely recommended by less than 50% of U.S. physicians.82 After zidovudine was shown to be effective in reducing vertical transmission,83 universal prenatal counseling and voluntary HIV testing was recommended in 1995 by the U.S. Public Health Service20 and the American Academy of Pediatrics,73 and appeared to contribute to an increase in the number of HIV diagnoses in pregnant women in the U.S.84 In a seven-state surveillance study, for example, the proportion of HIV-infected women diagnosed before delivery increased from 70% to 80% between 1993 and 1996.85 In one British study, the incidence of known HIV seropositivity at delivery nearly doubled (0.26% to 0.48%) after the implementation of a universal voluntary screening program.75 In another British study, however, over 50% of cases identified by anonymous testing were not detected after a universal counseling and voluntary testing policy was implemented, despite increased uptake rates.86 We identified no U.S. studies since 1995 evaluating the effectiveness of targeted compared to universal counseling and screening. A recent survey of 138 physicians in Alabama who provide prenatal care found that 12% reported that they did not offer universal prenatal HIV counseling and testing despite guidelines.87

Key Question 3. What Are the Test Characteristics of HIV Antibody Test Strategies in Pregnant Women?

The use of repeatedly reactive enzyme-linked immunosorbent assay (EIA) followed by confirmatory Western blot (WB) or immunofluorescent assay (IFA) on an office-based venipuncture specimen remains the standard strategy for diagnosing HIV-1 infection, and is associated with a sensitivity and specificity greater than 99%.88, 89 The diagnostic accuracy of standard HIV testing is thought to be similar for pregnant and non-pregnant persons, though indeterminate results may occur slightly more frequently among parous and pregnant women.90

Rapid HIV antibody tests provide results in 5–40 minutes, compared to one to two weeks for standard testing.6 Such testing provides an opportunity to reduce transmission of HIV from pregnant women who received no prenatal care or who were not tested earlier in pregnancy for other reasons. Such point of care testing in labor may also allow providers to avoid obstetric practices that may increase the risk of transmission, and gives providers the opportunity to counsel the mother against breastfeeding.6, 18 Notification of rapid test results prior to the availability of confirmatory results is recommended in situations in which preliminary test results might benefit tested persons, such as in women with unknown HIV status presenting in active labor.91 However, this could result in unnecessary exposure to antiretroviral or other therapies if the rapid test result is a false positive. Most studies measure the diagnostic accuracy of rapid tests before confirmatory testing, though CDC guidelines recommend routine confirmation of positive rapid tests.92

The Food and Drug Administration (FDA) has approved four rapid HIV tests (Uni-Gold™ Recombigen®, Reveal™ G2, OraQuick® Advance, and Single Use Diagnostic System [SUDS]), but one (SUDS) is not currently being manufactured. In studies of mostly non-pregnant persons, the sensitivities of rapid HIV tests currently available in the U.S. ranged from 96% to 100% and the specificities >99% compared to standard testing.9398 The OraQuick® test performed slightly better than the other rapid tests on blood samples and was calculated to have a positive predictive value near 100% even in low-prevalence settings.6 Though a newer version of the OraQuick® test (OraQuick® ADVANCE Rapid HIV-1/2 Antibody Test) has recently been FDA approved for testing of oral as well as whole blood specimens, no studies of the diagnostic accuracy of rapid oral specimen testing are yet available.

We identified three good-99101 and four fair-quality102105 studies evaluating the diagnostic test characteristics of rapid HIV testing during pregnancy that used standard EIA and confirmatory Western blot as the reference standard (Table 2). Two of these were conducted among pregnant women in the U.S.,99, 100 but only one99 evaluated a rapid HIV test currently available in the U.S. This was a good-quality prospective study that evaluated the test characteristics of the OraQuick® Rapid HIV-1 serum test among 5,744 women presenting in labor in six U.S. cities between 2001 and 2003. Compared to standard testing, the sensitivity was 100% (95% CI, 90%–100%), specificity 99.9% (95% CI, 99.78%–99.98%), positive predictive value 90% (95% CI, 75%–97%), and negative predictive value 100%, with a prevalence of 0.59%. In studies of rapid tests not currently available in the U.S., sensitivity ranged from 95.8% to 100%, specificity ranged from 98% to 100%, and positive predictive values ranged from 33% to 100%.100105 One African study comparing a strategy of confirming one positive rapid test with a second, different rapid test found a sensitivity of 99.6% and specificity 99.9% compared to standard testing, but this strategy is not used in the U.S.106

Table 2. Test Characteristics Of Rapid HIV-1 Antibody Tests Evaluated In Pregnant Women.

Table 2

Test Characteristics Of Rapid HIV-1 Antibody Tests Evaluated In Pregnant Women.

We identified no studies evaluating the diagnostic accuracy of HIV tests in pregnant women based on home-based sampling kits, non-invasive (urine or oral) specimens, or testing of pooled samples with polymerase chain reaction to detect acute infection using standard testing as the reference standard. Although one Indian study found a lower sensitivity with the OraQuick® test on saliva compared to plasma (75.0% vs. 86.4%), it did not use standard EIA plus WB as the reference standard, and may have been related to decreased saliva due to hot local conditions.107

Repeat testing of women who screen HIV-negative during early pregnancy could identify those who are infected after initial testing but before delivery. Whether to repeatedly screen during pregnancy and the optimal timing of repeat testing would depend in part on the frequency of new HIV infections. The incidence of HIV infection among average-risk U.S. women has been estimated at 0.17 per 1,000 person-years,2 and among a high-risk population of pregnant women at 6.2 per 1,000 person-years.108 Using these rates, one model estimated that repeat testing in the third trimester after a negative test in the first trimester would detect 5.3 infections per 100,000 average-risk women tested and 192 infections per 100,000 high-risk women tested.109

Key Question 4. What Are the Harms (Including Labeling and Anxiety) Associated with Screening? Is Screening Acceptable to Pregnant Women?

False-positive diagnoses are rare with standard testing even in low-risk settings.110 Most of the evidence regarding the frequency and harms from false-positive diagnoses in pregnant women is anecdotal, but could include elective pregnancy termination based on incorrect test results, anxiety, discrimination, or altered partner relationships.111 In a recent U.S. study of rapid HIV testing in women with undetermined status presenting in labor, 4 out of 4,849 tested women had a false-positive test and briefly received antiretroviral prophylaxis prior to receiving results of confirmatory testing.99

False-negative tests during pregnancy may occur in recently infected individuals or those with advanced disease, and could give false assurance. False-negative and true-negative tests could encourage continued risky behaviors unless patients are appropriately counseled, but we identified no studies evaluating changes in behaviors in pregnant women after negative HIV tests. Indeterminate test results are likely to cause anxiety while additional testing is performed, but data are limited on rates and consequences of indeterminate tests in pregnant women.112

True-positive tests can result in anxiety, depression, social stigmatization, changes in relationships with sexual partners, and discrimination.73, 74 Most studies on these harms have been performed in non-pregnant populations. One small (N=40) study of U.S. women found that mean anxiety and depression scores were significantly (p<0.05) higher for HIV-positive women compared to matched uninfected controls.113 A potential increase in the risk of intimate partner violence for pregnant women after disclosure of HIV status is especially concerning. A recent good-quality cohort study, however, found that the rate of violence during pregnancy was similar between HIV-infected women and seronegative at-risk pregnant women, and that receiving an HIV diagnosis prenatally did not increase risk.114 Disclosure-related violence occurred, but was rare. There are insufficient data to determine whether diagnosis of HIV during pregnancy is associated with an increased risk of suicide.115 One small study found a nonsignificant trend towards increased partner dissolution in HIV-positive pregnant women compared to matched seronegative controls.113

There remains general consensus that HIV testing should be voluntary and performed after obtaining informed consent.17 Mandatory testing of pregnant women has been debated, but might result in women avoiding prenatal care to avoid testing.116 Uptake of voluntary HIV testing has increased since recommendations regarding universal counseling were issued. A large U.S. telephone survey, for example, found that testing rates in pregnant women had increased from 41% in 1995 to 60% in 1998.117

A good-quality systematic review found that acceptance rates for voluntary HIV antibody testing among more than 174,000 pregnant women in 25 studies published through 1995 ranged from 23% to 100%.118 Recent data from 16 U.S. states and 5 Canadian provinces found a similar range of prenatal HIV testing (25% to 98%) among pregnant women.119 Smaller recent U.S. studies reported testing rates that also fell within those ranges.120, 121

Several patient or provider factors appear to affect testing rates. One randomized trial found that antenatal uptake rates were significantly higher in patients offered HIV testing (35%) than in patients for whom the test was available, but did not receive a direct offer (6%).122 Strong provider endorsement of testing appears to be a predictor of HIV testing acceptance.121, 123 Lack of prenatal care, on the other hand, was a predictor of declining to be tested in a large observational study.124 Other factors associated with increased acceptance of voluntary HIV testing were inconsistent across studies and included specific ethnic or racial groups, age groups, educational level, marital status, and socioeconomic status.120, 121, 125, 126

Policy factors also appear to influence testing rates. For example, some jurisdictions have adopted an ‘opt-out’ (pregnant women are informed that an HIV test is being conducted as a standard part of prenatal care and that they may refuse it) compared to an ‘opt-in’ (pregnant women are required to consent specifically to an HIV test) policy. Testing rates appeared to be higher in states and Canadian provinces that used an ‘opt-out’ policy (71% to 98% vs. 25% to 83% with opt-in testing).119 Other studies from the U.S. and Canada have also reported high (85% to 88%) rates of testing acceptance using an opt-out approach.122, 127, 128 The implementation of a mandatory newborn testing policy with expedited results was associated with increased prenatal testing rates in two states.119 We identified no studies specifically evaluating the effect of name-based reporting on rates of prenatal screening, though in a Canadian study in which an opt-out approach and name-based reporting were introduced near-simultaneously, testing rates increased.128

Newer screening methods such as home sample collection kits, rapid tests, on-site testing, and non-invasive sampling could increase rates of voluntary prenatal HIV testing.6 The recent observational MIRIAD (Mother-Infant Rapid Intervention At Delivery) study of pregnant women (N=5,744) presenting to labor and delivery units with undocumented HIV status found that 84% accepted rapid testing.99 Higher acceptance was associated with younger age, being Black or Hispanic, gestational age less than 32 weeks, and having no prenatal care. Lower acceptance was associated with being admitted between 4 pm and midnight, possibly because of fewer available hospital personnel.99 We identified no studies evaluating the effect of alternative sampling methods (urine or saliva sampling, home-based collection) on the uptake of prenatal HIV testing.

Key Question 5. How Many HIV-Infected Pregnant Women Who Meet Criteria for Interventions Receive Them?

Current guidelines recommend that antiretroviral prophylaxis be offered to all HIV-infected pregnant women in order to reduce the risk of mother-to-child transmission.14 Some women also meet criteria for antiretroviral treatment to improve maternal outcomes, which is determined by the CD4 count and viral load at the time of diagnosis. HAART is recommended for pregnant women according to the same guidelines used in the general population.14 For pregnant women who do not meet guidelines for HAART, the decision to use other less-intense antiretroviral regimens must be balanced against their potential for inducing resistance.129131

We identified one large U.S. cohort study (the Women and Infants Transmission Study) that reported maternal CD4 count and viral load in women enrolled during pregnancy since 1989.34 It found that 13% (70/546) had CD4 counts <200 cells/mm3 and 56% (307/546) had CD4 counts <500 cells/mm3. Ten percent (57 of 551) had viral loads <1,000 copies/ml.

HIV-tested persons may not return for their test results or regular medical care. In the U.S., however, test notification rates among pregnant women appear high. In one large U.S. study, for example, 91% (3,690 of 4,062) of tested pregnant women received their results.132 In settings with low return rates, rapid testing could increase notification rates by providing patients with same-visit results. We identified one African randomized trial that found that rapid HIV testing increased rates of notification of results compared to standard testing (96% versus 65%) among pregnant HIV-positive women not presenting during labor.133

HIV-infected women appear to widely accept and receive antiretroviral drugs during pregnancy. Several recent U.S. studies found that antiretroviral drugs were used by HIV-infected women in more than 90% of pregnancies, with a recent trend towards increased combination (HAART and non-HAART) regimens (58% to 80% in 1998-1999).37, 134137 In the Woman and Infants Transmission Study, approximately 60% of enrolled pregnant women received HAART in 1999.37 An earlier multi-state observational study found that the proportion of HIV-infected women who received prenatal zidovudine increased from 27% to 83% between 1993 and 1996.85 In a recent U.S. observational study, all (n=18) HIV-infected pregnant women diagnosed during labor in time to receive intrapartum zidovudine received it.99 Elective cesarean section rates for HIV-infected pregnant women also are rising. In several recent large U.S. observational studies, scheduled cesarean section rates ranged from 37% to 50%.134, 137, 138

HIV-infected women who are not tested during pregnancy may not be identified until they present with symptomatic illness or immunologically advanced disease. We identified no studies, however, comparing the proportion of women diagnosed late among those who were tested versus those not tested during pregnancy. Studies in the general population of HIV-infected persons suggest that diagnosis at immunologically advanced stages of disease is associated with poorer response to antiretroviral therapy.139144

Key Question 6. What Are the Harms Associated with the Work-up for HIV Infection in Pregnant Women?

We identified no studies estimating potential harms (anxiety, labeling, effects on close relationships, increased risky behaviors) from checking viral loads or CD4 counts in HIV-infected pregnant women.

Key Question 7a. How Effective Are Interventions (Antiretroviral Prophylaxis [to Prevent Mother-to-Child Transmission] or Treatment [to Improve Maternal Outcomes], Avoidance of Breastfeeding, Elective Cesarean Section [in Selected Patients] or Other Labor Management Practices, Counseling on Risky Behaviors, Immunizations, Routine Monitoring and Follow-up or Prophylaxis for Opportunistic Infections) in Reducing Transmission Rates or Improving Clinical Outcomes (Mortality, Functional Status, Quality of Life, Symptoms, or Opportunistic Infections) in Pregnant Women with HIV Infection?

Antiretroviral Drugs

Zidovudine alone has been shown to be efficacious and effective in reducing the risk of mother-to-child transmission of HIV. In the absence of antiretroviral prophylaxis, the risk for transmission of HIV from mother to infant is 14% to 25% in developed countries, and 13% to 42% in countries with high rates of breastfeeding.145 The landmark Pediatric AIDS Clinical Trials Group (PACTG) protocol 076 study found that a three-phase maternal and infant zidovudine regimen starting at 14 to 34 weeks gestation (median 26 weeks) through 6 weeks postpartum in non-breastfeeding women decreased the risk of mother-to-child transmission by nearly 70%, from about 25% to about 8%, compared to placebo.83 We identified a good-quality systematic review of seven randomized controlled trials that found that any zidovudine treatment (including shorter courses and in breastfeeding women) significantly reduced the risk of mother-to-child transmission compared to placebo (OR 0.46, 95% CI 0.35–0.60), with no significant heterogeneity between trials.146 Zidovudine was also associated with decreased risk of infant death within the first year of life (OR 0.57, 95% CI 0.38–0.85) and decreased risk of stillbirth (RR 0.31, 95% CI 0.11–0.90).

In the U.S., treatment of pregnant women infected with HIV has evolved from zidovudine alone to combination antiretroviral regimens.147 We identified one trial of continuous full-course combination antiretrovirals (nelfinavir or nevirapine plus zidovudine) that was discontinued early (after 38 women enrolled) because of a high rate of treatment-limiting or serious side effects in the nevirapine arm.148 Other randomized trials of full-course combination antiretrovirals in pregnant women are not available. We identified four large American or European cohort studies (three good-quality, one fair-quality) evaluating the relative effectiveness of two or more drug antiretroviral regimens versus placebo or full-course (PACTG 076 protocol) zidovudine monotherapy in non-breastfeeding women (Tables 3 and 4).37, 48, 149, 150 In all four studies, regimens with more antiretroviral drugs were superior to regimens with fewer antiretroviral drugs for preventing mother-to-child transmission (Table 5). The only study that specifically evaluated the effectiveness of HAART regimens compared to no antiretrovirals reported an adjusted odds ratio of 0.13 (95% CI, 0.06 to 0.27) for prevention of mother-to-child transmission.149 One study48 calculated an adjusted odds ratio of 0.07 (95% CI 0.02 to 0.23) for two or more drug antiretroviral regimens compared to no antiretrovirals, and two others37, 150 reported adjusted odds ratios of 0.22 (95% CI 0.10 to 0.50) and 0.30 (95% CI 0.09 to 1.02) for two or more drug regimens compared to full-course (three-part PACTG 076 protocol) zidovudine monotherapy. One study evaluated the effectiveness of HAART versus zidovudine monotherapy (adjusted OR 0.27, 95% CI 0.08 to 0.94).37 The proportion of women undergoing cesarean section in these studies ranged from 16% to 44%.

Table 3. Large Observational Cohort Studies Of Combination Antiretroviral Regimens On Risk Of Mother-to-Child Transmission Of HIV Infection.

Table 3

Large Observational Cohort Studies Of Combination Antiretroviral Regimens On Risk Of Mother-to-Child Transmission Of HIV Infection.

Table 4. Evidence Table: Large Observational Cohort Studies Of Combination Antiretroviral Regimens On Risk Of Mother-to-Child Transmission Of HIV Infection.

Table 4

Evidence Table: Large Observational Cohort Studies Of Combination Antiretroviral Regimens On Risk Of Mother-to-Child Transmission Of HIV Infection.

Table 5. Number of Drugs in Full-course Antiretroviral Regimens and Risk of Mother-to-Child Transmission of HIV Infection.

Table 5

Number of Drugs in Full-course Antiretroviral Regimens and Risk of Mother-to-Child Transmission of HIV Infection.

The addition of single-dose intrapartum (maternal) and postnatal (infant) nevirapine to antiretroviral regimens initiated before 34 weeks has been evaluated in two good-quality randomized controlled trials performed in non-breastfeeding settings (Tables 6 and 7).151, 152 The first trial, from Thailand, found that the addition of single doses of intrapartum and postnatal nevirapine to a slightly abbreviated course of zidovudine monotherapy (from 28 weeks gestation to 1 week postnatal) reduced mother-to-child transmission from 6.3% to 1.9%.152 An earlier international randomized clinical trial, on the other hand, found that the addition of single-dose intrapartum and postnatal nevirapine to primarily (77%) combination antiretroviral regimens did not further decrease already low transmission rates (1.4% to 1.6%).151

Table 6. Randomized Controlled Trials ofLonger (Before 34 Weeks Gestation) Antiretroviral Regimens for Reduction of Mother-to-Child Transmission of HIV Infection in Non-breastfeeding Women.

Table 6

Randomized Controlled Trials ofLonger (Before 34 Weeks Gestation) Antiretroviral Regimens for Reduction of Mother-to-Child Transmission of HIV Infection in Non-breastfeeding Women.

Table 7. Evidence Table: Randomized Controlled Trials of Longer (Before 34 Weeks Gestation) Courses of Antiretroviral Therapy for Reduction of Mother-to-Child Transmission of HIV Infection in Non-Breastfeeding Women.

Table 7

Evidence Table: Randomized Controlled Trials of Longer (Before 34 Weeks Gestation) Courses of Antiretroviral Therapy for Reduction of Mother-to-Child Transmission of HIV Infection in Non-Breastfeeding Women.

Shorter courses of antiretroviral prophylaxis have also been developed for use in resource-poor countries. Data from these studies may also help guide management of women in the U.S. who were not diagnosed early enough to receive a full course of antiretroviral prophylaxis. Several clinical trials have evaluated shorter courses of antiretrovirals in women diagnosed after 34 weeks, but before presenting in active labor (Tables 8, 9, and 10). A randomized controlled trial from Thailand (the Perinatal HIV Prevention Trial) found that the risk of transmission using a “short-short” course of zidovudine (from 35 weeks in pregnancy for the mother, intrapartum, and for the newborn until 3 days old) was higher (OR 2.33, 95% CI 1.16–4.68) than the risk using a “long-long” course (from 28 weeks in pregnancy, intrapartum, and for the infant until 6 weeks old).153 However, intermediate courses were similar in efficacy to the full course. An earlier Thai randomized controlled trial found that prophylaxis with zidovudine from 36 weeks and intrapartum without neonatal treatment was associated with a transmission rate similar to that seen in the short-short leg of the Perinatal HIV Prevention Trial (9.4% vs. 10%), suggesting that short courses of neonatal zidovudine added little benefit.154 Both trials were in non-breastfeeding women. Another recent good-quality randomized controlled trial in African breastfeeding women found that short-course zidovudine combined with lamivudine from 36 weeks gestation reduced mother-to-child transmission from 15.3% (in women receiving placebo) to 5.7 % (OR 0.37; 95% CI, 0.21–0.65).155 Shorter courses of zidovudine and lamivudine were less effective. Two other trials of breastfeeding women in Africa found that zidovudine from 36 weeks reduced mother-to-child transmission of HIV from 26.1% to 27.5% in women in the placebo arms compared to 16.5% to 18.0% in the intervention arms.156, 157

Table 8. Randomized Controlled Trials of Short-Course Antiretroviral Regimens for Reduction of Mother-to-Child Transmission of HIV Infection.

Table 8

Randomized Controlled Trials of Short-Course Antiretroviral Regimens for Reduction of Mother-to-Child Transmission of HIV Infection.

Table 9. Evidence Table: Randomized Controlled Trials of Short Courses of Zidovudine (ZDV) Monotherapy for Reduction of Mother-to-Child Transmission of HIV Infection.

Table 9

Evidence Table: Randomized Controlled Trials of Short Courses of Zidovudine (ZDV) Monotherapy for Reduction of Mother-to-Child Transmission of HIV Infection.

Table 10. Evidence Table: Randomized Controlled Trials of Short Courses of Combination Antiretroviral Therapy for Reduction of Mother-to-Child Transmission of HIV Infection.

Table 10

Evidence Table: Randomized Controlled Trials of Short Courses of Combination Antiretroviral Therapy for Reduction of Mother-to-Child Transmission of HIV Infection.

Some HIV-infected pregnant women may not be diagnosed until very late in pregnancy or during labor. We identified four good-quality African randomized controlled trials of breastfeeding women evaluating the effects of very abbreviated regimens for this situation (Tables 8, 9, and 10).158162 Three of these trials evaluated regimens that consisted of antiretroviral prophylaxis administered during labor and postexposure treatment for the infant. One clinical trial found that in this setting, nevirapine was significantly better at reducing vertical transmission (11.8%) than zidovudine (20.0%).160, 162 Another trial found that nevirapine administered during labor and to the infant was associated with a similar rate of vertical transmission (14.1%) compared to the same regimen with zidovudine also administered to the infant (16.3%).161 In the third trial, short-course nevirapine was associated with a 12.3% transmission rate compared to 9.3% with zidovudine plus lamivudine.158 The fourth trial compared regimens of neonatal postexposure prophylaxis without maternal prophylaxis.159 It found that prophylaxis of the newborn alone was associated with higher transmission rates (15.3% for nevirapine plus zidovudine versus 20.9% for nevirapine alone) than seen in clinical trials that included maternal prophylaxis.

A recent U.S. observational study of rapid testing for women with unknown HIV status presenting during labor and who received zidovudine prophylaxis with or without nevirapine found that the transmission rate was 9% (3 of 32).99

We identified no studies evaluating clinical outcomes (clinical progression, death, quality of life, or horizontal transmission) associated with different antiretroviral regimens for HIV-infected women identified during pregnancy. In one study of women who received zidovudine plus single-dose nevirapine intrapartum and subsequently started a nevirapine-based antiretroviral regimen, no harmful effects on clinical outcomes were observed after six months, but longer term follow-up is not yet available.131

Breastfeeding

We identified two meta-analyses of observational studies that found that breastfeeding was associated with an overall increased rate of mother-to-child transmission of HIV of 14% to 16% (Table 11).24, 27 In two other meta-analyses, the cumulative rate of late transmission was 9.3% after 18 months in one meta-analysis of individual patient data from clinical trials that defined late transmission as occurring after four weeks,26 and 9.2% after 18 months in an earlier meta-analysis of observational studies that defined late transmission as occurring after 2.5 months.28 Factors associated with an increased risk of breastfeeding transmission include low maternal CD4 count, detectable virus in breast milk, higher serum viral load, acute HIV infection, nipple lesions, mastitis, oral candidiasis in the infant, longer duration of breastfeeding, younger maternal age, lower parity, and male sex of the infant.23, 26, 163

Table 11. Evidence Table: Randomized Controlled Trials and Large Observational Studies Evaluating the Association between Breastfeeding and Risk of Mother-to-Child Transmission of HIV Infection.

Table 11

Evidence Table: Randomized Controlled Trials and Large Observational Studies Evaluating the Association between Breastfeeding and Risk of Mother-to-Child Transmission of HIV Infection.

We identified no randomized controlled trials evaluating the rate of vertical transmission associated with breastfeeding in the U.S. or in women on antiretroviral therapy. We identified one large prospective Italian cohort study of 3,770 children that found that breastfeeding significantly increased transmission rates when adjusted for other factors including antiretroviral use (adjusted OR 10.20 [2.73–38.11]).48 An African trial of formula versus breastfeeding among women not receiving antiretroviral therapy found that breastfeeding was associated with a probability of vertical transmission of 36.7% (95% CI, 29.4%–44.0%) at 24 months compared to 20.5% (95% CI, 14.0%–27.0%) in the formula feeding arm, and a mortality rate of 24.4% (95% CI, 18.2% to 30.7%) compared to 20.0% (95% CI, 14.4%–25.6%).164 Another African observational study suggested that mixed feeding (both formula and breast) was associated with a higher risk of mother-to-child transmission of HIV than exclusive breastfeeding, though confidence intervals overlapped.165

Pregnancy and Labor Management

Labor management techniques that minimize contact between infected maternal bodily fluids and the fetus could decrease the risk of mother-to-child HIV transmission. Elective cesarean section has been the most extensively studied labor management technique.35, 146, 166172

One good-quality European cohort study evaluated the effectiveness of elective cesarean section in the HAART era.149 It found an odds ratio of 0.33 (95% CI 0.11 to 0.94) for mother-to-child transmission with elective cesarean delivery compared to vaginal delivery when adjusted for antiretroviral therapy, prematurity, and maternal CD4 count and viral load. In the subgroup of women receiving HAART, the odds ratio was 0.64 (95% CI 0.08 to 5.37) for elective cesarean compared to vaginal delivery, and in the subgroup with undetectable viremia, the odds ratio was 0.07 (95% CI 0.02 to 0.31) for elective cesarean compared to vaginal or emergency cesarean delivery.

Other studies evaluating the effectiveness of elective cesarean section were conducted prior to the widespread use of combination antiretroviral regimens. We identified one good-quality randomized clinical trial examining the impact of elective cesarean section on mother to child HIV transmission (Table 12).171 This European study of 370 mother-child pairs found a reduction in vertical transmission from 10.5% in women randomized to vaginal delivery to 1.8% in those randomized to elective cesarean section (p=0.009). Among 119 babies delivered to women who received zidovudine and underwent cesarean section, the rate of HIV infection was 0.8%. We also identified a meta-analysis of individual patient data from 8,533 mother-child pairs in 15 prospective cohort studies that found a 50% reduction in the likelihood of vertical transmission with elective cesarean section compared to other modes of delivery (OR 0.43, 95% CI 0.33–0.56).170 The benefits of elective cesarean section appeared additive with prophylactic zidovudine monotherapy, with the likelihood of transmission reduced by approximately 87% with both elective cesarean section and full-course zidovudine compared to other modes of delivery (non-elective cesarean section or vaginal delivery) and no antiretroviral therapy (adjusted OR 0.13, 95% CI 0.09–0.19). A meta-analysis of 7 European and U.S. prospective cohort studies of 1,202 women with viral loads <1,000 copies/ml also found that cesarean section (elective and non-elective) was independently associated with a lower risk for transmission (adjusted OR 0.30, p=0.22), but the overall transmission rate was low (3.6%) and reduced by antiretroviral therapy alone (primarily zidovudine) to about 1%.173 We identified no studies evaluating the additive effects of elective cesarean section in women receiving multi-drug antiretroviral regimens.

Table 12. Evidence Table: Meta-Analysis and Randomized Controlled Trial of Effects of Elective Cesarean Section on Risk of Mother-to-Child Transmission of HIV Infection.

Table 12

Evidence Table: Meta-Analysis and Randomized Controlled Trial of Effects of Elective Cesarean Section on Risk of Mother-to-Child Transmission of HIV Infection.

We identified one good-quality systematic review that evaluated the risk of invasive procedures during pregnancy and found only one prospective cohort study that met inclusion criteria.174 In that study, amniocentesis was associated with a significantly increased rate of mother-to-child transmission of HIV.44 We also identified one good-quality systematic review that found no association between vaginal disinfection with chlorhexidine and reduced mother-to-child transmission of HIV (OR 0.93, 95% CI 0.63–1.38).175

Counseling on Risky Behaviors

We identified no studies estimating the effects of counseling HIV-infected pregnant women regarding risky behaviors on vertical or horizontal transmission rates.

Immunizations

We identified no clinical trials or observational studies estimating clinical benefits of recommended immunizations in HIV positive pregnant women.

Prophylaxis for Opportunistic Infections

We identified no clinical trials or observational studies estimating clinical benefits of recommended prophylaxis for different opportunistic infections in HIV infected pregnant women.

Routine Monitoring and Follow-up

HIV-infected women identified during pregnancy might benefit from appropriate monitoring of their status (such as following CD4 count and viral load) or regular follow-up to identify early signs of symptomatic illness, in addition to other interventions. We identified no studies estimating the clinical benefits of linking women with health care for routine monitoring after identification of HIV infection during pregnancy.

Key Question 7b. Does Immediate Antiretroviral Treatment in HIV-Infected Pregnant Women Result in Improvements in Clinical Outcomes Compared to Delayed Treatment until Symptomatic?

Some HIV-infected women who choose to use an antiretroviral regimen during the perinatal period to prevent vertical transmission may not receive long-term HAART in the postnatal period because of low viral loads, high CD4 counts, loss to follow-up, or other reasons. We identified no studies estimating the effects of delayed or discontinued versus continuous HAART in HIV-infected women identified during pregnancy.

Withholding antiretrovirals in the first trimester may be an option for women with low viral loads who have a lower risk of transmitting HIV and wish to minimize the risk for congenital anomalies or reduce the likelihood for poor adherence because of pregnancy-related nausea.14 However, we identified no trials examining the effects of delaying antiretroviral prophylaxis or treatment until after the first trimester on mother-to-child transmission rates or other clinical outcomes.

Key Question 7c. How Well Do Interventions Reduce the Rate of Viremia, Improve CD4 Counts, and Reduce Risky Behaviors? How Does Identification of HIV Infection in Pregnant Women Affect Future Reproductive Choices?

In HIV-infected persons in general, HAART is more effective than less intense regimens in achieving sustained virological suppression and improved CD4 counts.176 In pregnant women, HAART appears similarly effective for improving intermediate outcomes.39

We found little evidence on the effect of counseling HIV-positive pregnant women on subsequent changes in risky behaviors that may be associated with increased rates of vertical transmission, such as unprotected intercourse, cigarette smoking, and hard drug use. A small U.S. study found that 40% of 20 HIV-positive and 20% of 20 HIV-negative women reported always using condoms.113 In another study, most HIV-positive women with a history of intravenous drug use who decreased needle sharing changed their behavior before learning their HIV status.177

Counseling HIV-infected women could also lead to behavior changes that might decrease the risk of horizontal transmission, but most studies evaluating the effects of counseling on behavior changes have been performed in non-pregnant persons. In two good-quality systematic reviews, there was mixed evidence regarding the effectiveness of counseling on changing behaviors among HIV-infected women.178, 179 In one small U.S. study, a high proportion of both HIV-positive and HIV-negative pregnant women had unprotected intercourse after testing.113

Knowledge of HIV infection status could affect future reproductive choices, but we identified few studies evaluating the effects of identifying and counseling HIV-infected pregnant women on subsequent contraceptive choices or pregnancy, sterilization, and abortion rates.180182 In two studies, HIV seropositivity was associated with a lower rate of pregnancy182 or trend towards lower rate181 than in uninfected women, but another study found that the rate of pregnancy in HIV-infected women appears to be increasing.183 One U.S. study found that 27% of HIV-infected women chose tubal ligation compared to 15% in uninfected controls, and oral contraceptive use was less likely in seropositive women.181 Two other non-comparative U.S. studies reported rates of tubal ligation among HIV-infected women of 24% and 27%.85, 180 An African study of single session postpartum counseling in HIV-infected women found that the intervention did not appear to influence decisions on condom use or reproductive behavior.184 No differences in pregnancy termination rates between HIV-infected and uninfected women were seen in two U.S. studies.113, 185

Key Question 8. What Are the Harms (Including Adverse Effects from In Utero Exposure) Associated with Antiretroviral Intervention and Elective Cesarean Section?

Harms of Antiretrovirals to Mothers

Receipt of antiretrovirals during pregnancy is associated with significant short-term non-obstetric adverse events, but these often resolve after stopping the offending drug or drug combination, and effective alternatives usually are available.146 Guidelines reviewing adverse events associated with specific antiretroviral drugs, classes, and combinations are regularly updated, and specific antiretroviral combinations associated with serious complications are not recommended.14, 21 Serious or fatal non-obstetric adverse events appear rare on zidovudine monotherapy and currently recommended combination regimens.186

We identified one good-quality meta-analysis that found that zidovudine monotherapy in pregnant women did not cause any deaths or long-term maternal adverse events.146 The largest prospective study examining obstetric adverse events from combination antiretroviral therapy was an international study of 1,407 women that found that gestational diabetes was the only complication associated with antiretroviral therapy, and was most frequent for combination therapy that included a protease inhibitor and was initiated early in the pregnancy.187 Other observational studies have also found an association between elevated serum glucose levels and protease inhibitor therapy in pregnant women.188, 189 One recent clinical trial was discontinued after enrollment of 38 HIV-infected pregnant women because of a high rate of treatment-limiting hepatic or cutaneous toxicity with long-term nevirapine (29%) compared to nelfinavir (5%) in combination with zidovudine, including one death and one case of Stevens-Johnson syndrome.148 Severe reactions to nevirapine were significantly more frequent in women with CD4 counts greater than 250 cells/mm3. Observational studies (N=46–139) have also reported usually reversible hepatitis or abnormal liver function tests (1.1 to 5.0%) associated with long-term nevirapine that was rarely (2 cases) fatal.190192 No laboratory or clinical evidence of liver toxicity with single-dose intrapartum nevirapine, however, has been reported. Three recent randomized controlled trials of a single maternal intrapartum dose of nevirapine with or without other antiretroviral therapy found no differences in liver function tests or hepatitis between the nevirapine prophylaxis and the control group.152, 158, 160

Another potential harm of antiretrovirals initiated during pregnancy is the development of resistance or viral rebound, particularly in women who receive regimens that do not fully suppress viral replication or discontinue antiretrovirals after pregnancy.52 Zidovudine monotherapy during the PACTG 076 trial, for example, was associated with an increased rate of low-level (but not high-level) genotypic zidovudine resistance.130 Studies examining the effect of limited exposure to zidovudine monotherapy, however, did not find a negative impact on subsequent disease progression or response to later therapy.53, 193, 194 In one of these studies, clinical benefits of HAART started in the postpartum period were comparable to those reported in other studies of persons without a recent pregnancy.53 Lamivudine and zidovudine combination therapy was associated with lamivudine resistance mutations in 39% of treated pregnant women.150

Several recent studies examining the effects of single-dose intrapartum nevirapine prophylaxis have found nevirapine resistance mutations in 5%-32% of treated women six weeks postpartum.65, 66, 131, 195 Of these studies, the only one that evaluated the clinical impact of these resistance mutations was a Thai trial that found that women who received single-dose intrapartum nevirapine in addition to standard zidovudine therapy were less likely to have complete virological suppression after six months of postpartum treatment with a nevirapine-containing regimen (49% vs. 68%).131 CD4 cell count response and degree of weight loss, however, was not significantly different between groups receiving and not receiving nevirapine during pregnancy.

Maternal Harms of Elective Cesarean Section

HIV-infected women appear to be at increased risk for cesarean section-related complications than uninfected women. We identified two retrospective cohort studies that found that HIV-positive women had significantly more postoperative fever (OR 2.5–5.7) and minor complications such as urinary tract infections, endometritis, or wound infection (OR 2.7–3.1) compared to HIV-negative women.196, 197 In one study, major adverse events (pneumonia, pleural effusion, transfusion, and sepsis) were reported in 6 of 156 HIV-infected patients.197

Cesarean section is generally associated with an increased risk of complications compared to vaginal delivery. We identified one randomized clinical trial that found that the rate of postpartum fever was 1.1% (2 of 183) in women delivering vaginally and 6.7% (15 of 225) for women delivering by cesarean section, but no serious postpartum complications occurred in either group.171 We also identified two good-quality large prospective cohort studies198, 199 and one retrospective cohort study200 that evaluated the risk for HIV-infected women undergoing elective cesarean delivery versus vaginal delivery. The largest prospective study, with 1,186 HIV-infected women, found that elective cesarean section was associated with increased rates of postpartum fever (14.3%; RR 4.16, 95% CI 1.99–8.70), hemorrhage (7.1%; RR 1.58, 95% CI 0.58–4.26), endometritis (5.4%; RR 2.57, 95% CI 0.78–8.51), urinary tract infection (5.4%; RR 3.64, 95% CI 1.06–12.54), and any postpartum morbidity (26.7%; RR 2.62, 95% CI 1.61–4.20) compared to vaginal delivery.199 Another prospective study (N=497) found that HIV-infected women delivering by cesarean section (elective or emergent) had an increased risk of endometritis (adjusted OR 4.8, 95% CI 2.5–9.3) and hemorrhage requiring blood transfusion (adjusted OR 2.8, 95% CI 1.0–8.4) compared to those delivering vaginally.198 The retrospective study (N=309) found that HIV-infected women who delivered by elective cesarean section had more serious postpartum complications (fever, endometritis, urinary tract infection, pneumonia, wound infection, deep vein thrombosis [DVT], anemia requiring transfusion, transfer to intensive care, or death) than HIV-infected women who delivered vaginally (OR 1.85, 95% CI 1.00–3.39), but elective cesarean section was associated with fewer complications than emergency cesarean delivery.200 An increased rate of postoperative complications was consistently associated with lower maternal CD4 count, and decreased rate with receipt of antiretrovirals and more recent year of delivery.

Harms of In Utero Antiretroviral Exposure to Infants

The Food and Drug Administration currently classifies didanosine, saquinavir, ritonavir, enfuvirtide, and nelfinavir as pregnancy class B (animal studies fail to demonstrate risk to the fetus and no human studies have been conducted).201 Zidovudine, zalcitabine, stavudine, lamivudine, abacavir, indinavir, amprenavir, lopinavir, nevirapine, efavirenz, fosamprenivir and delavirdine are classified as pregnancy class C (safety in human pregnancy has not been determined). Use of efavirenz in early pregnancy is not recommended due to high rates of fetal anomalies in animal studies and case reports of adverse human pregnancy outcomes.18

HIV-seropositivity appears to be associated with an increased risk of perinatal and neonatal complications. A good-quality systematic review of 31 studies found that compared to HIV-negative women, HIV-positive women had significantly higher rates of spontaneous abortion (OR 4.05, 95% CI 2.75–5.96), stillbirth (OR 3.91, 95% CI 2.65–5.77), intrauterine growth retardation (OR 1.7, 95% CI 1.43–2.02), premature delivery (OR 1.83, 95% CI 1.63–2.06), and low birth weight infants (OR 2.09, 95% CI 1.86–2.35), but no significant increases in fetal abnormalities or neonatal mortality.202

We identified one good-quality U.S. meta-analysis of five prospective cohort studies and one good-quality, large European prospective cohort study that found no significant differences in the rates of congenital anomalies, neonatal conditions, or low birth weight between infants exposed to any combination of antiretroviral therapy and unexposed infants (Tables 13 and 14).39, 203 On the other hand, data regarding the association between combination antiretroviral therapy and increased rates of premature delivery are mixed. The meta-analysis found no increase in premature delivery rates for infants exposed to combination therapy with (OR 1.50, 95% CI 0.72–3.01) or without a protease inhibitor (OR 0.95, 95% CI 0.51–1.69) compared to no treatment,203 but a large European prospective cohort study found an increased rate of premature birth associated with combination therapy (adjusted OR 2.60–4.14 for combination therapy with a protease inhibitor and 1.82–2.66 without a protease inhibitor compared to no treatment).204

Table 13. Large Observational Studies Evaluating Adverse Effects from In Utero Exposure to Combination Antiretrovirals.

Table 13

Large Observational Studies Evaluating Adverse Effects from In Utero Exposure to Combination Antiretrovirals.

Table 14. Evidence Table: Large Observational Studies and Meta-Analyses Evaluating Association between In Utero Exposure to Antiretroviral Therapy and Infant Adverse Effects.

Table 14

Evidence Table: Large Observational Studies and Meta-Analyses Evaluating Association between In Utero Exposure to Antiretroviral Therapy and Infant Adverse Effects.

Although molecular evidence of mitochondrial dysfunction has been reported in infants exposed in utero to antiretroviral therapy,205, 206 the clinical impact of such dysfunction is unclear.207, 208 A recent prospective Canadian observational study, for example, found that 92% of uninfected infants exposed to HAART in utero had elevated plasma lactate levels on at least one occasion, but none were clinically ill.209 Large cohort studies have also found no evidence of clinical mitochondrial dysfunction among HIV-negative infants exposed to antiretroviral therapy.39, 210 In population-based mortality studies, no deaths due to mitochondrial dysfunction among exposed, HIV-negative infants have been reported.211213

Studies with longer duration of follow-up (four to six years) are so far available only for zidovudine monotherapy. We identified one good-quality meta-analysis and one good-quality prospective cohort study that found that in utero and postnatal zidovudine did not cause any increase in detectable long-term adverse clinical events or changes in growth or development in exposed infants up to four years of age.146, 214 Zidovudine use also was not found to increase rates of prematurity or low birth weight. No tumors or deaths from cancers were reported among 727 children exposed to zidovudine in utero and followed for six years.215

Key Question 9. Have Improvements in Intermediate Outcomes (CD4 Counts, Viremia, or Risky Behaviors) in HIV-Infected Pregnant Women Been Shown to Improve Clinical Outcomes or Reduce Mother-to-Child Transmission?

Higher maternal viral loads and lower CD4 counts are associated with an increased risk of mother-to-child transmission of HIV.2935, 37, 39, 41, 49, 216218 Observational studies and clinical trials have consistently found that women receiving highly active antiretroviral regimens who had a reduction in HIV RNA to <1,000 copies/ml had very low rates (about 1%) of perinatal transmission.37, 150, 151, 173, 219

Several maternal behaviors (such as unprotected intercourse,46, 220 illicit drug use,46, 221 or cigarette smoking43, 222) may be associated with an increased rate of vertical transmission, but we identified no studies evaluating the association between changes in these behaviors and subsequent rates of mother-to-child transmission.

Views

  • PubReader
  • Print View
  • Cite this Page

In this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...