You Are Here:
Systematic Review to Update the 2005 U.S. Preventive Services Task Force Recommendation
Release Date: November 2012
By Roger Chou, MD; Amy G. Cantor, MD, MHS; Bernadette Zakher, MBBS; and Christina Bougatsos, MPH.
The information in this article is intended to help clinicians, employers, policymakers, and others make informed decisions about the provision of health care services. This article is intended as a reference and not as a substitute for clinical judgment.
This article may be used, in whole or in part, as the basis for the development of clinical practice guidelines and other quality enhancement tools, or as a basis for reimbursement and coverage policies. AHRQ or U.S. Department of Health and Human Services endorsement of such derivative products may not be stated or implied.
This article was first published in Annals of Internal Medicine on November 20, 2012 (Ann Intern Med 2012;157:719-728; http://www.annals.org).
Background: A 2005 U.S. Preventive Services Task Force (USPSTF) review found good evidence that prenatal HIV screening is accurate and can lead to interventions that reduce the risk for mother-to-child transmission.
Purpose: To update the 2005 USPSTF review, focusing on previously identified research gaps and new evidence on treatment.
Data Sources: MEDLINE (2004 to June 2012) and the Cochrane Library (through the second quarter of 2012).
Study Selection: Randomized trials and cohort studies of pregnant women on risk for mother-to-child transmission or harms associated with prenatal HIV screening or antiretroviral therapy during pregnancy.
Data Extraction: 2 reviewers abstracted and confirmed study details and quality by using predefined criteria.
Data Synthesis: No studies directly evaluated effects of prenatal HIV screening on risk for mother-to-child transmission or maternal or infant clinical outcomes. One fair-quality, large cohort study (HIV prevalence, 0.7%) found that rapid testing during labor was associated with a positive predictive value of 90%. New cohort studies of nonbreastfeeding women in the United States and Europe confirm that full-course combination antiretroviral therapy reduces rates of mother-to-child transmission (<1% to 2.4% vs. 9% to 22% with no antiretroviral therapy). New cohort studies found antiretroviral therapy during pregnancy to be associated with increased risk for preterm delivery (<37 weeks’ gestation); there were no clear associations with low birthweight, congenital abnormalities, or infant neurodevelopment. Evidence on long-term maternal harms after short-term antiretroviral therapy exposure during pregnancy remains sparse.
Limitations: Only English-language articles were included. Studies conducted in resource-poor settings may be of limited applicability to screening in the United States.
Conclusion: Antiretroviral therapy in combination with avoidance of breastfeeding and elective cesarean section in women with viremia reduces risk for mother-to-child transmission. Use of certain antiretroviral therapy regimens during pregnancy may increase risk for preterm delivery.
Primary Funding Source: Agency for Healthcare Research and Quality.
Between 6000 and 7000 HIV-positive women give birth each year in the United States (1), and approximately 30% of women are unaware of their HIV-positive status before pregnancy (2). Mother-to-child transmission is responsible for more than 90% of pediatric HIV infections in the United States (3, 4). The number of cases of perinatal HIV infections in the United States peaked at about 1650 in 1992 but has since decreased dramatically, with the widespread adoption of routine prenatal screening coupled with the use of more effective therapies for preventing mother-to-child transmission; the number of cases was estimated at 215 to 370 in 2005 (5).
Current U.S. recommendations are for opt-out HIV screening at the initial prenatal visit as part of standard prenatal testing (6, 7). “Opt-out screening” refers to screening that is performed unless the woman specifically declines. The Centers for Disease Control and Prevention recommend that clinicians consider repeated testing in the third trimester in all women who test negative initially, and they recommend repeated testing for women who continue to practice high-risk behaviors or are in a high-incidence setting.
The current standard of care to prevent perinatal transmission of HIV infection in the United States is a 3-drug antiretroviral regimen started at the beginning of the second trimester of pregnancy or earlier (followed by treatment of the infant in the postnatal period) in all HIV-infected women, regardless of viral load or CD4 cell count; elective cesarean delivery before labor or rupture of membranes in women with HIV RNA levels greater than 1000 copies/mL near delivery; and avoidance of breastfeeding in all women (8, 9). Women who are identified as HIV-positive during pregnancy may also benefit from other interventions that would be considered in nonpregnant women with HIV infection, including long-term antiretroviral therapy, prophylaxis against opportunistic infections, immunizations, and counseling to reduce high-risk behaviors for horizontal transmission.
The U.S. Preventive Services Task Force (USPSTF) last reviewed the evidence on prenatal screening for asymptomatic HIV in 2005 (10) and issued a recommendation to screen all pregnant women (grade A recommendation) (7). The USPSTF did not address repeated prenatal screening. This report updates the previous USPSTF review on benefits and harms of prenatal HIV screening, with an emphasis on research gaps identified in that review and new evidence on benefits and harms of antiretroviral medications. Because perinatal practices and interventions related to prevention of HIV infection are substantially affected by the availability of resources, the report will emphasize evidence that is more applicable to typical practice in the United States.
Scope of the Review
We followed a standardized protocol and developed an analytic framework (Figure) that focused on the following key questions:
1. What are the benefits of HIV screening versus no screening in asymptomatic pregnant women on maternal or child morbidity, mortality, or quality of life or rates of mother-to-child transmission?
2a. What is the yield (number of new diagnoses) of repeat HIV screening in asymptomatic pregnant women?
2b. What are the adverse effects (including false-positive tests and anxiety) of rapid versus standard HIV testing in asymptomatic pregnant women?
3a. What is the effectiveness of newer antiretroviral regimens for reducing mother-to-child transmission?
3b. What are the effects of antiretroviral regimens in pregnant, HIV-positive women on long-term maternal morbidity, mortality, or quality of life?
3c. What are the harms (including longer-term harms) to the mother or child associated with antiretroviral therapy during pregnancy?
The full report (11) provides detailed methods and data for the review, including search strategies and multiple tables with quality ratings of individual studies. Laboratory or imaging effects of antiretroviral therapy for children with uncertain clinical implications, such as mitochondrial dysfunction, echocardiographic abnormalities, and hematologic abnormalities, are also reviewed in the full report but are not presented in this article.
This update focuses on research gaps identified in the previous review, such as harms (including false-positive results and anxiety) of rapid versus standard testing and the yield of repeated screening. The diagnostic accuracy of HIV testing and the effectiveness of breastfeeding avoidance and elective cesarean delivery in selected women are well-established (10, 12) and were not rereviewed. Rather, this update focuses on new evidence on the effectiveness of combination antiretroviral regimens on perinatal transmission, as well as evidence on long-term clinical outcomes of prenatal exposure to antiretroviral therapy in the mother and harms to the mother or infant.
Data Sources and Searches
We searched Ovid MEDLINE from 2004 to June 2012 and the Cochrane Library through the second quarter of 2012 and reviewed reference lists to identify relevant articles published in English.
At least 2 reviewers independently evaluated each study to determine eligibility for inclusion. Articles were selected for full review if they were about HIV infection in pregnancy, were relevant to a key question, and met the predefined inclusion criteria (Appendix Table 1). Outcomes were mother-to-child transmission, morbidity, mortality, quality of life, and harms from antiretroviral therapy (such as adverse pregnancy outcomes; adverse congenital, neurodevelopmental, cardiovascular, metabolic, or hematologic outcomes in exposed children; and adverse clinical outcomes in mothers), including long-term outcomes (those occurring ≥1 year after birth for women and ≥2 years after birth for children). We included randomized, controlled trials and cohort studies for all key questions.
For key questions related to harms and other long-term maternal and infant outcomes, we also included case–control studies and intervention series if randomized trials and cohort studies were unavailable or lacking. For some key questions, we included studies from resource-poor settings that evaluated short-course antiretroviral regimens or breastfeeding populations, because these may provide some information about the effectiveness of antiretroviral therapies in U.S. women who present late in pregnancy or about the general effectiveness of combination antiretroviral therapy.
Data Extraction and Quality Assessment
One investigator abstracted details on the study design, patient population, setting, screening method, treatment regimen, analysis, follow-up, and results. A second investigator reviewed data abstraction for accuracy. Two investigators independently applied criteria developed by the USPSTF (13) to rate the quality of each study as good, fair, or poor. Discrepancies were resolved by consensus.
We assessed the aggregate internal validity (quality) of the body of evidence for each key question as good, fair, or poor by using methods developed by the USPSTF, based on the number, quality and size of studies, consistency of results between studies, and directness of evidence (13). Meta-analysis was not attempted because the data could not be pooled, owing to differences across studies in design, interventions, populations, and other factors.
Role of the Funding Source
This research was funded by the Agency for Healthcare Research and Quality (AHRQ) under a contract to support the work of the USPSTF. Investigators worked with USPSTF members and AHRQ staff at key points to develop and refine the scope, analytic framework, and key questions; resolve issues arising during the project; and finalize the report. AHRQ staff provided project oversight, reviewed the draft report, and distributed the draft for peer review, including by representatives of professional societies and federal agencies. In addition, AHRQ performed a final review of the manuscript to ensure that the analysis met methodological standards. AHRQ had no role in study selection, quality assessment, synthesis, or development of conclusions. The investigators are solely responsible for the content and the decision to submit the manuscript for publication.
The Appendix Figure shows the results of the search and study selection process.
Longer-Term Harms Associated With ART
Key Question 1
What are the benefits of HIV screening versus no screening in asymptomatic pregnant women on maternal or child morbidity, mortality, or quality of life or rates of mother-to-child transmission?
No randomized trial or observational study compared clinical outcomes (including risk for perinatal transmission) between pregnant women who were screened and not screened for HIV infection.
Key Question 2a
What is the yield (number of new diagnoses) of repeat screening in asymptomatic pregnant women?
No randomized trial or observational study evaluated the yield of repeated prenatal HIV screening compared with 1-time screening or compared the yield of different strategies for repeated screening (such as risk-based repeated screening versus a routinely repeated test).
Key Question 2b
What are the adverse effects (including false-positive tests and anxiety) of rapid versus standard HIV testing in asymptomatic pregnant women?
The large (7753 participants), prospective, fair-quality MIRIAD (Mother-Infant Rapid Intervention At Delivery) study provides the strongest evidence on the diagnostic accuracy of the rapid OraQuick test (OraSure Technologies, Bethlehem, Pennsylvania) compared with standard enzyme immunoassay HIV testing (14, 15). MIRIAD specifically enrolled women in labor with unknown HIV status (HIV prevalence, 0.7%), for whom immediate test results are needed to help guide treatment decisions.
Initial (2-year) results from MIRIAD (15) were included in the previous USPSTF review (10). Final (40-month) results (14) found that compared with Western blot (the reference standard), sensitivity was 100% for both tests and specificity was 99.9% and 99.8% for the rapid and standard tests, respectively. On the basis of an HIV prevalence of 0.7% (52 of 7753 persons), the positive predictive value was higher for the rapid test (90% [52 of 58 persons]) than for the standard test (74% [52 of 70 persons]). In clinical practice, a positive result from a standard test would not be available in time to inform interventions during labor and delivery and would require Western blot confirmation.
A study (16) of 910 pregnant women, about 90% of whom were Hispanic, at any gestational age (HIV prevalence, 0.5%) found a positive predictive value of 100% (5 of 5) for the OraQuick rapid test and a value of 36% (5 of 14) for standard enzyme immunoassay (before confirmation).
No study compared psychological or other harms associated with rapid versus standard tests or adverse clinical consequences of interventions given as a result of initial false-positive rapid test results.
Key Question 3a
What is the effectiveness of newer antiretroviral regimens for reducing mother-to-child transmission?
We identified no new randomized trials since the previous review on full-course (started at or before the beginning of the second trimester) combination antiretroviral therapy in non–resource-poor, nonbreastfeeding settings. Consistent with the prior USPSTF review, 3 U.S. and European cohort studies (involving 489 to 7344 participants) published since 2005 found perinatal, full-course, triple antiretroviral therapy to be associated with rates of mother-to-child transmission ranging from less than 1% to 2.4%, compared with 9% to 22% with no antiretroviral therapy (17–19). The largest cohort study (involving 7344 participants), based on U.S. surveillance data from 1999 to 2001, found full-course, single- or multidrug antiretroviral therapy to be associated with a rate of mother-to-child transmission of 2.4%, compared with 22% for no antiretroviral therapy (adjusted odds ratio [OR], 0.09 [95% CI, 0.06 to 0.12]) (18). In women who received antiretroviral therapy, combination regimens with zidovudine plus other drugs were about twice as effective as zidovudine alone for reducing risk for mother-to-child transmission (adjusted ORs, 0.4 to 0.5). Two smaller European cohort studies (17, 19) also reported lower mother-to-child transmission rates with combination antiretroviral therapy (0.6% and 1.0%, respectively) than with no therapy (18% and 9%, respectively).
A fourth study, which analyzed European surveillance data for 7573 participants over 9 years and included 1 of these cohorts, found transmission rates of less than 1% with either zidovudine-sparing or zidovudine-containing regimens of 3 or more drugs (20). Appendix Table 2 provides details on these 4 studies.
One good-quality (21) and 5 fair-quality (22–26) randomized trials published since the 2005 USPSTF review evaluated shorter-course prenatal antiretroviral regimens in primarily breastfeeding African women (Appendix Table 3). In the United States, these studies are most applicable to HIV-infected women identified later in pregnancy, who cannot receive full-course regimens.
In general, these studies reported lower transmission rates with antiretroviral therapy than expected without treatment. Studies that evaluated longer courses of treatment and regimens that included at least 3 drugs reported the lowest transmission rates; 1 trial (709 participants) of various 3-drug regimens started at 18 to 34 weeks' gestation (median, 26 to 27 weeks) reported an HIV transmission rate of 1.1% at 6 months (24), which was similar to the rates observed in U.S. and European cohort studies (17–20) of full-course, triple-drug regimens.
Transmission rates in studies that evaluated antiretroviral regimens initiated later in pregnancy or with fewer than 3 drugs reported rates of mother-to-child transmission ranging from 4% to 12% (21–23, 25), although rates were still lower than expected without treatment (about 25%) (27). One trial (609 participants) found high rates of mother-to-child transmission with ultrashort-course zidovudine (during labor and given to the infant for 72 hours after birth) plus single-dose maternal and infant nevirapine as well as single-dose nevirapine alone (14% vs. 17%), and a high rate of infant mortality (7% at 6 weeks) (26).
Key Question 3b
What are the effects of antiretroviral regimens in pregnant, HIV-positive women on long-term maternal morbidity, mortality, or quality of life?
No study published since the prior USPSTF review evaluated the effects of antiretroviral therapy administered during pregnancy and then discontinued on long-term maternal clinical outcomes. The prior USPSTF review included 1 study of 226 U.S. women that found no difference in risk for AIDS-defining events or death after a mean of 4.1 years between women randomly assigned to receive zidovudine during pregnancy and those assigned to receive placebo (28). A study included in the prior USPSTF review found that women still benefit from subsequent highly active antiretroviral therapy after receiving antiretroviral treatment during pregnancy (29).
Key Question 3c
What are the harms (including longer-term harms) to the mother or child associated with antiretroviral therapy during pregnancy?
New evidence (27 studies [20, 30–55]) on infant and maternal harms associated with perinatal exposure to antiretroviral therapy was generally consistent with the evidence included in the 2005 USPSTF review (10, 12).
Preterm Birth and Other Birth Outcomes
One randomized trial (40) and 10 cohort studies (30–39) published since the prior USPSTF review reported risk for prematurity, low birthweight, and other birth outcomes after in utero exposure to antiretroviral therapy (Appendix Table 4). Sample sizes ranged from 57 to 8793 participants. Eight studies were rated as fair-quality (30, 32, 34, 35, 37–40), and 3 were poor-quality (31, 33, 36). Methodological shortcomings included differences between groups in baseline characteristics and poor reporting of attrition. Six studies reported risk estimates adjusted for important confounders, such as maternal age, CD4 count, and viral load (30, 32, 34, 36–38).
The randomized trial (530 participants) found that protease inhibitor–based antiretroviral therapy was associated with greater risk for preterm delivery than nonnucleoside reverse transcriptase–based antiretroviral therapy (OR, 2.0 [CI, 1.3 to 3.3]) (40). Three prospective cohort studies (183 to 8793 participants) found maternal exposure to combination antiretroviral therapy with a protease inhibitor to be associated with increased risk for preterm delivery (<37 weeks) compared with combination antiretroviral therapy without a protease inhibitor (adjusted OR, 1.8 [CI, 1.1 to 3.0]) (32), dual therapy (adjusted OR, 1.2 [CI, 1.0 to 1.4]) (37), or monotherapy (adjusted OR, 3.4 [CI, 1.1 to 10]) (34). None found exposure to combination therapy without a protease inhibitor to be associated with increased risk for preterm delivery. However, a large cohort study (4939 participants) found combination therapy to be associated with increased risk for preterm delivery (<37 weeks; adjusted OR, 1.4 [CI, 1.1 to 1.8]; P = 0.02) and very preterm delivery (<32 weeks; OR, 2.6 [CI, 1.3 to 5.3]; P = 0.007) compared with monotherapy or dual therapy; risk did not differ according to whether the antiretroviral regimen included a protease inhibitor or not (38). Of 4 studies that did not adjust for confounders, 1 found an association between prenatal antiretroviral therapy and preterm delivery (39) and 3 found no clear association (31, 33, 35).
Seven cohort studies (352 to 8192 participants) published since the 2005 USPSTF review found no clear association between maternal use of antiretroviral therapy and low birthweight or intrauterine growth restriction (30–33, 35, 37, 38).
Three fair-quality cohort studies (1414, 3740, and 8576 participants) published since the 2005 USPSTF review found no association between perinatal exposure to antiretroviral therapy and congenital abnormalities (41–43). Follow-up ranged from 6 months to 17 years. One large study (7573 participants) of European surveillance data over a 9-year period found no difference in the risk for infant congenital abnormalities with maternal use of zidovudine-sparing versus zidovudine-containing antiretroviral therapy (20).
Two cohort studies published since the 2005 USPSTF review (10, 12) found no clear differences in neurodevelopmental outcomes between children exposed to antiretroviral therapy in utero and postnatally compared with unexposed controls at 18 to 36 months of follow-up (44, 45). Both studies used the Bayley Scales of Infant Development II, which include a mental development index and psychomotor development index.
We identified 1 large (2543 participants), fair-quality U.S. cohort study published since the 2005 USPSTF review that found antiretroviral use to be associated with increased risk for maternal anemia compared with nonuse (adjusted OR, 1.6 [CI, 1.1 to 2.4]) (46). It also found late use of antiretroviral therapy (started between 25 and 32 weeks' gestation) to be associated with increased risk for gestational diabetes compared with nonuse (adjusted OR, 3.5 [CI, 1.2 to 10]); however, causality was unclear, because screening for gestational diabetes is typically performed at 24 to 28 weeks' gestation and women may have received a diagnosis before initiation of antiretroviral therapy.
A smaller (167 participants) fair-quality cohort study found exposure to combination therapy to be associated with a trend toward increased risk for gestational diabetes compared with exposure to monotherapy with zidovudine or no antiretroviral therapy, but the difference was not statistically significant (12% vs. 0%; unadjusted relative risk, 0.11 [CI, 0.01 to 1.7]) (47).
As in the 2005 USPSTF review (10, 12), we found no direct evidence on effects of prenatal screening for HIV infection versus no screening on risk for mother-to-child transmission or maternal or infant clinical outcomes. Other evidence reviewed in this update is summarized in the Table.
The 2005 USPSTF review (10, 12) found that HIV tests are accurate. The strongest evidence on potential harms associated with rapid testing is from the fair-quality MIRIAD study, which found a lower positive predictive value for standard enzyme immunoassay than for a rapid test (74% and 90%, respectively) in a population of women presenting in labor among whom the prevalence of undiagnosed HIV infection was 0.7%. This could result in unnecessary maternal and fetal exposure to antiretroviral therapy (14). The positive predictive value would be expected to be lower in lower-prevalence populations, potentially resulting in more unnecessary antiretroviral exposure.
No study has evaluated the clinical consequences of unnecessary exposure to antiretroviral therapy as a result of an initially positive false-positive rapid HIV test, although any such harms must be weighed against the potential benefits of prenatal identification and treatment of undiagnosed HIV infection. As in the 2005 USPSTF review, no study has evaluated the yield of repeated HIV screening during pregnancy, which depends on the incidence of new HIV infection.
New cohort studies of antiretroviral therapy in nonbreastfeeding women in the United States and Europe confirm the finding from the 2005 USPSTF review that full-course combination antiretroviral therapy is effective at reducing the rate of mother-to-child transmission (<1% to 2.4% vs. 9% to 22% with no antiretroviral therapy) (17–19). Randomized trials also found low risk for transmission with combination therapy regimens started around the end of the second trimester in breastfeeding African women (21, 24). Shorter courses of antiretroviral therapy evaluated in randomized trials were not as effective as full-course regimens, but they reduced risk for mother-to-child transmission compared with historical transmission rates without antiretroviral therapy and are relevant for women in the United States who might begin therapy late, owing to delayed diagnosis or treatment (22, 23, 25).
Evidence on harms of prenatal antiretroviral therapy was also largely consistent with the 2005 USPSTF review. Current evidence continues to suggest that the long-term harms associated with antiretroviral exposure are relatively small. New cohort studies found that perinatal antiretroviral therapy was associated with increased risk for preterm delivery (31–40), but there was no clear association with low birthweight (30, 32, 33, 35, 37, 38), congenital abnormalities (20, 41–43), or impaired infant neurodevelopment (44, 45). Although other studies (reviewed in the full report ) found an association between in utero exposure to antiretroviral therapy and echocardiographic abnormalities (48), hematologic abnormalities (49–51), or markers of mitochondrial dysfunction (52–54), the clinical significance of these findings remains unclear. Evidence on long-term maternal harms associated with short-term exposure to antiretroviral therapy during pregnancy, or antiretroviral therapy started during pregnancy and continued after pregnancy, remains sparse.
Antiretroviral therapy during pregnancy is associated with the nonobstetric adverse events typically associated with the specific drugs and regimens, but these often resolve after treatment with the offending drug or drug combination is stopped, and effective alternatives are usually available (8). Antiretroviral therapy regimens for use during pregnancy and indications for initiating long-term antiretroviral therapy continue to evolve, and guidelines on selection of antiretroviral therapy for pregnant women are regularly updated (8).
Our study has limitations. We excluded non–English-language articles, which could result in language bias, although we identified no non–English-language studies that would have met our inclusion criteria. We could not formally assess for publication bias with graphical or statistical methods because of small numbers of studies and differences in the study designs, populations, and outcomes assessed. We included observational studies, which are more susceptible to bias and confounding than well-conducted randomized trials, although we focused on results from studies that performed statistical adjustment for potential confounding. We also included studies conducted in resource-poor and high-prevalence settings, which could limit applicability to U.S. practice.
More research is needed on the long-term maternal effects of transient exposure to antiretroviral therapy during pregnancy or use of less intense antiretroviral regimens during pregnancy. Children exposed to antiretroviral therapy in utero should continue to be followed to help identify unexpected or emerging long-term harms from combination regimens. More research is also needed to understand the clinical significance of the hematologic abnormalities, echocardiographic abnormalities, and markers of mitochondrial dysfunction observed in some children exposed to antiretroviral therapy.
In summary, prenatal HIV screening is accurate and antiretroviral therapy in combination with avoidance of breastfeeding and cesarean section in women with HIV RNA levels greater than 1000 copies/mL near the time of delivery is effective at reducing risk for mother-to-child transmission. Use of certain antiretroviral therapy regimens during pregnancy may be associated with increased risk for preterm delivery, but more evidence is needed to fully understand short- and long-term maternal and infant effects.