INTRODUCTION — Decisions regarding antiretroviral therapy (ART) in the pregnant female are complex and, beyond the typical considerations of virologic potency, side effects, and formulation, need to also take into consideration other factors, such as possible changes in pharmacokinetics due to physiologic changes, potential toxicities that may be magnified during pregnancy, and potential toxicity to the fetus and child.

This topic will address the clinical data on the safety and pharmacology of the more commonly used antiretroviral medications during pregnancy. Antiretrovirals that are not commonly used or used only in limited situations are not discussed in this topic.

Guidelines for ART of pregnant women with HIV in the United States are developed by the Panel on Treatment of HIV-Infected Pregnant Women and Prevention of Perinatal Transmission; these guidelines include detailed information on the safety and pharmacokinetics of antiretroviral drugs in pregnancy [1]. Guidelines for ART of pregnant women residing in resource-limited settings, where antiretroviral drug access may be more limited and preferred drug choices may differ, are developed by the World Health Organization (WHO) and may differ from recommendations in the United States; WHO guideline updates can be found on its website.

Antiretroviral selection and other management issues for the pregnant woman with HIV are discussed in detail elsewhere. (See "Prenatal evaluation of women with HIV in resource-rich settings" and "Antiretroviral selection and management in pregnant women with HIV in resource-rich settings" and "Intrapartum management of pregnant women with HIV and infant prophylaxis in resource-rich settings" and "Prevention of mother-to-child HIV transmission in resource-limited settings".)

GENERAL PRINCIPLES — Antiretroviral therapy (ART) is recommended for all individuals with HIV, including pregnant women, regardless of immune, clinical, or viral status [1]. ART reduces HIV-related morbidity and mortality, even in individuals with high CD4 cell counts [2,3]. An additional goal of ART in pregnant women is to decrease the risk of perinatal transmission of HIV infection.

In pregnant women, antiretroviral regimen selection should take into account the resistance profile of the virus, the safety and efficacy of the drugs in the mother and fetus, the convenience and adherence potential of the regimen, the potential for drug interactions with other medications, and pharmacokinetic data in pregnancy. In resource-rich settings, certain antiretroviral agents that are recommended in the general population with HIV are not preferred agents for pregnant women because of limited experience during pregnancy. On the other hand, certain agents that are not preferred in the general population may be preferred for pregnant women because of extensive clinical experience during pregnancy. In general, treatment-naïve women should be treated with a regimen made up of preferred agents, if the resistance profile of the virus allows. Treatment-experienced women on a suppressive antiretroviral regimen can generally continue it even if the agents are not specifically preferred during pregnancy. Preferred antiretroviral agents in pregnancy and regimen selection for pregnant women are discussed in detail elsewhere. (See "Antiretroviral selection and management in pregnant women with HIV in resource-rich settings" and "Prevention of mother-to-child HIV transmission in resource-limited settings".)

The benefits of ART generally outweigh the risk of adverse effects to the pregnant woman and her infant. With rare exception, the short-term data are reassuring. However, long-term data on the safety of in utero drug exposure in humans are not available for any antiretroviral drug, and it is possible that in utero exposure to antiretroviral agents could have delayed effects, as were ultimately identified with diethylstilbestrol (DES). Information on in utero antiretroviral exposure should be maintained in a child's medical record in case problems that could be long-term complications are identified in the future.

Clinicians are encouraged to contact the Antiretroviral Pregnancy Registry to prospectively report antiretroviral exposures in pregnancy to assist in monitoring pregnancy outcomes of pregnant women exposed to antiretroviral medications (1-800-258-4263 or www.APRegistry.com). The registry is a collaborative project of pharmaceutical manufacturers with an advisory committee of obstetric and pediatric practitioners. The registry is anonymous. Information is compiled from case reports and updated every six months for public review.

PREGNANCY OUTCOMES WITH COMBINATION ART EXPOSURE — Overall, the known benefits of combination antiretroviral therapy (ART) for pregnant women outweigh the known and potential risks (see "Antiretroviral selection and management in pregnant women with HIV in resource-rich settings", section on 'Rationale for maternal ART'). However, clinicians should be aware of and discuss with patients the potential risk for adverse pregnancy outcome with ART. Some, but not all, studies have suggested a small increase in the risk for preterm birth, pregnancy loss, and impaired fetal growth. Of note, untreated maternal HIV infection itself has been associated with adverse pregnancy outcomes [4,5]. Studies on teratogenicity of specific antiretroviral agents have been overall reassuring but cannot rule out small increased risk in rare defects. For all these outcomes, further longitudinal data are warranted to clarify the risk.

Fetal growth — Studies evaluating the risk of combination antiretroviral agents during pregnancy on poor fetal growth and low birth weight have demonstrated conflicting results [6-14]. Similarly, studies evaluating maternal estradiol and progesterone levels, which can impact fetal growth and birth weight, and their association with antiretroviral agents have been conflicting. Two studies have reported an increase in estradiol in pregnant women receiving protease inhibitor-based therapy [15,16]. One of these studies reported a differential impact of drug class, with an increase of estradiol with lopinavir-ritonavir and a decrease with efavirenz-based therapy, and no effect on progesterone [15]. Other studies have reported decreased progesterone levels in women receiving protease inhibitor-based therapy [17,18]. Nevertheless, whether poor growth is due more to underlying maternal characteristics or to the medications themselves, some assessment of fetal growth is important in pregnant females with HIV. While some experts would rely on appropriate growth of uterine fundal height, others recommend ultrasound assessment in the third trimester to ensure adequate fetal growth. (See "Prenatal evaluation of women with HIV in resource-rich settings", section on 'Prenatal fetal monitoring' and "Fetal growth restriction: Screening and diagnosis".)

Studies of fetal growth effects associated with specific agents are discussed below.

Preterm birth — There may be an increased risk of preterm delivery prior to 37 weeks gestation with combination antiretroviral regimens, particularly with those that contain protease inhibitors; data from prospective cohort studies are conflicting, while a large randomized trial conducted in resource-limited settings did find an increased risk [4,19-30]. Even in those studies that demonstrate an increased risk with protease inhibitors, the increases are modest, and the effects on infant morbidity and mortality are unknown. Thus, given the clear benefits to maternal health and a reduction in perinatal transmission of HIV, certain protease inhibitors remain preferred antiretroviral agents for pregnant women with HIV in resource-rich settings. (See "Antiretroviral selection and management in pregnant women with HIV in resource-rich settings", section on 'Approach by patient population'.)

In a meta-analysis of studies evaluating pregnancy outcomes in women with HIV, use of any antiretroviral regimen was not associated with an increased risk for preterm birth [19]. Compared with no antiretroviral use, however, a protease inhibitor-based regimen was associated with a small, non-significant trend towards increased preterm birth, which reached statistical significance when compared with other combination regimens that did not contain a protease inhibitor (OR 1.35, 95% CI 1.08-1.70). In a subsequent trial comparing a protease inhibitor-containing combination ART regimen with a triple-nucleoside regimen, preterm (<37 weeks) births were more frequent among women in the protease inhibitor group (24 versus 11 percent) [21]. However, there were no differences in infant hospitalizations or mortality rates through six months of age.

Similar results were reported in a subsequent randomized trial conducted in resource-limited settings that compared two protease inhibitor-containing combination ART regimens (lopinavir-ritonavir plus either zidovudine-lamivudine [zidovudine-ART arm] or tenofovir-emtricitabine [tenofovir-ART] arm) with zidovudine plus single-dose nevirapine (zidovudine-alone arm) among nearly 3500 HIV-infected pregnant women with CD4 cell counts >350 cells/microL [28]. Because of initial uncertainties about the safety of in utero exposure to tenofovir, during the first period of the trial (representing approximately 65 percent of enrollment), participants were randomly assigned only to zidovudine-ART or zidovudine alone unless they had HBV coinfection; during the second period of the trial, participants were randomly assigned to any one of the three arms, with approximately 350 women in each arm; comparisons among the three arms were limited to this second period of concurrent enrollment. Rates of preterm birth at <37 weeks were higher with the combination regimens (20.5 percent with zidovudine-ART versus 13.1 percent with zidovudine during the combined first and second period of the trial, and 18.5 percent with tenofovir-ART versus 13.5 percent with zidovudine alone during the second period). Rates of very preterm birth at <34 weeks and neonatal deaths, most of which occurred in very preterm infants, were not statistically higher with each of the combination ART regimens compared with zidovudine alone. However, during the second period of the trial, rates of very preterm delivery and neonatal deaths were both higher with tenofovir-ART (6.0 and 4.4 percent) compared with zidovudine-ART (2.6 and 0.6 percent). The vast majority of the neonatal deaths in the zidovudine-ART study arm occurred during the first period of the study, suggesting that these outcomes may have been unusually lower in the zidovudine-ART arm during the second period rather than abnormally high in the tenofovir-ART arm [31]. The risk of very preterm birth and neonatal death with tenofovir is discussed in detail elsewhere. (See 'Very preterm birth/neonatal mortality' below.)

Pregnancy loss — The impact of combination ART on the risk of pregnancy loss is uncertain and more difficult to study, as pregnancy losses may not be reported as reliably as premature birth. Some studies suggest a possible increased risk with ART exposure at conception, but certain limitations reduce confidence in the findings.

In an observational study from Botswana, the rate of still birth among women with HIV infection who had started various ART regimens prior to conception was 1.4-fold higher than among women without HIV infection (3.4 versus 2.1 percent) [14].

In a post-hoc analysis of a trial of women with CD4 cell counts >400 cells/microL, rates of spontaneous abortion and stillbirth were higher (23.6 versus 11.9 percent, relative risk 2.0, 95% CI 1.1-3.5) among women who had been randomly assigned to continue ART following an earlier pregnancy (and thus were more likely to have been on ART at the time of conception of the index pregnancy) than among those assigned to discontinue ART following an earlier pregnancy (and thus were more likely not to have been on ART at the time of conception) [32]. However, approximately 10 percent of women in the continue ART arm were not taking ART at conception, and 13 percent in the discontinue ART arm were taking ART at conception, and in the as-treated analysis, the association between ART at conception and pregnancy loss was of a lower magnitude and no longer statistically significant.

In contrast, in the United States Women's Interagency HIV study, among women living with HIV, lower rates of miscarriage (spontaneous loss of pregnancy at <29 weeks gestation) were associated with low viral load (which likely reflects ART use; adjusted odds ratio [OR] 0.45, 95% CI 0.24-0.84) and use of protease inhibitor therapy compared with non-use of combination ART (adjusted OR 0.40, 95% CI 0.21-0.79) [33]. Although the proportion of women living with HIV who experienced a miscarriage (37 percent) was higher than the general population (10 to 20 percent of United States women), the rate of miscarriage was similar compared with women without HIV with similar socioeconomic status (39 percent).

Teratogenicity — As of the end of July 2020, the prevalence of birth defects reported to the Antiretroviral Pregnancy Registry among infants born to women with exposure to any antiretroviral agent during pregnancy was 580 birth defects of 20,437 live births, or 2.8 percent (95% CI 2.6-3.1) [34]. The prevalence was not different between women with first trimester exposure and those with initial antiretroviral exposure after the first trimester. These birth defect rates are comparable to the prevalence of birth defects reported in the United States (2.72 percent, according to surveillance by the Centers for Disease Control and Prevention). Other studies have suggested that overall, first-trimester antiretroviral exposure is not associated with an increased risk of birth defects [35,36]. Data on specific antiretroviral exposures are discussed in the sections dedicated to specific agents below.

NUCLEOSIDE REVERSE TRANSCRIPTASE INHIBITORS

Class effects

Mitochondrial toxicity — Mitochondrial toxicity is a major adverse effect of nucleoside reverse transcriptase inhibitor (NRTI) treatment and can lead to myopathy, peripheral neuropathy, and hepatic steatosis with lactic acidosis, the last of which may have a female preponderance and can be life-threatening [37,38]. Clinicians should be vigilant for symptoms of mitochondrial toxicity in pregnant women taking NRTIs, which overlap with commonly-associated symptoms of pregnancy (eg, nausea, vomiting, abdominal bloating, fatigue). The hepatic lactic acidosis syndrome has similarities to the rare, but life-threatening, HELLP syndrome (hemolysis, elevated liver enzymes, and low platelets), which occurs in association with fatty liver during the third trimester. (See "HELLP syndrome (hemolysis, elevated liver enzymes, and low platelets)".)

Mitochondrial toxicity is a consequence of NRTI binding to mitochondrial DNA polymerase gamma, which leads to mitochondrial DNA depletion and dysfunction. The various NRTI agents inhibit mitochondrial DNA polymerase gamma to varying degrees in vitro. The greatest inhibition among commonly used NRTIs is with zidovudine. Tenofovir, emtricitabine, lamivudine, and abacavir inhibit the enzyme to a much lesser extent.

There is also the risk of mitochondrial toxicity among infants whose mothers used NRTIs during pregnancy. Hematologic abnormalities consistent with mitochondrial depletion (eg, anemia and neutropenia) have been reported in studies of infants with in utero NRTI exposure from both the United States and Europe [39-41]. Additionally, mitochondrial dysfunction should be considered as a potential etiology of neurologic abnormalities in children with past ART exposure, although the association is still not well defined. Early reports from France had suggested that exposure to zidovudine with or without lamivudine in utero was associated with significant mitochondrial dysfunction in infancy (with moderate to severe neurologic disease or mitochondrial toxicity associated with mortality) [42,43]. A large cohort of 4392 uninfected children in France estimated that 0.3 to 0.5 percent of infants with in utero NRTI exposure developed severe neurological symptoms thought to be associated with mitochondrial toxicity [43,44]. However, subsequent studies from the United States and Europe have not duplicated these findings [45-50].

Malignancy risk — Research to date has been reassuring regarding the risk of malignancy among HIV-exposed uninfected children with perinatal antiretroviral drug exposure.

Early studies from the United States and United Kingdom, with median follow-up ranging between 14.5 months to 4.2 years, reported no cases of malignancy among children with antiretroviral exposure [51-53]. Additionally, in studies of French children exposed to nucleoside reverse transcriptase inhibitors in utero, the overall incidence of cancer did not differ significantly from that of the general population (21 cases observed among over 15,000 children) [54,55].

In another study of 3087 HIV-exposed but uninfected children followed in New Jersey for a median of 9.8 years, there were four diagnoses of cancer; cancer incidence among HIV-exposed children without perinatal antiretroviral exposure did not differ significantly from that among HIV-exposed children with perinatal exposure (22.5 versus 14.3 per 100,000 person-years) [56]. Additionally, the number of cases in children with perinatal antiretroviral exposure did not significantly differ from the number of cases expected based on state and national reference rates for children aged <19 years.

Abacavir — Abacavir is a preferred NRTI for use in pregnancy in the United States [57].

●Fetal safety – Short-term data on the use of abacavir in pregnancy do not suggest major concerns about fetal safety [30,34]. In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to abacavir are available to exclude a 1.5-fold or more increase in birth defects; the prevalence of birth defects reported to the registry was 3.1 percent (95% CI 2.3-4.2) [34].

In a separate analysis of data from the Antiretroviral Pregnancy Registry, there was no excess risk of spontaneous abortions, stillbirths, preterm births, and low birth weight with in utero exposure to an abacavir-containing antiretroviral regimen compared with other regimens [58].

●Maternal safety – Serious hypersensitivity reactions have been associated with abacavir therapy in nonpregnant adults and have rarely been fatal; symptoms include fever, skin rash, fatigue, and gastrointestinal symptoms such as nausea, vomiting, diarrhea, or abdominal pain. Abacavir should only be used in individuals, pregnant or otherwise, with a negative HLA-B*5701 test and without a history of a possible abacavir hypersensitivity reaction. (See "Abacavir hypersensitivity reaction".)

●Dosing – No dose adjustments are warranted for abacavir during pregnancy [59,60]. In a pharmacokinetic study of 25 women with HIV, pregnancy did not affect abacavir exposure during the third trimester or at 6 to 12 weeks postpartum [59].

Emtricitabine — Emtricitabine is a preferred NRTI for use in pregnancy in the United States [57]. It is an NRTI option in the World Health Organization (WHO) recommended regimen for use in individuals with HIV, including pregnant women, in resource-limited settings [61].

●Fetal safety – Short-term data on the use of emtricitabine in pregnancy do not suggest major concerns about fetal safety. In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to emtricitabine are available to exclude a 1.5-fold or more increase in birth defects; the prevalence of birth defects reported to the registry was 2.6 percent (95% CI 2.1-3.2) [34]. Data from animal studies have also shown no evidence of teratogenicity.

●Dosing – No dose adjustments are warranted for emtricitabine during pregnancy. In a study of 31 patients who received emtricitabine, the pharmacokinetic exposure to emtricitabine was approximately 25 percent lower during the third trimester of pregnancy compared with the post-partum period [62]. However, this was not associated with virologic failure or mother-to-child transmission of HIV.

Lamivudine — Lamivudine is a preferred NRTI for use in pregnancy in the United States [57]. It is an NRTI option in the WHO recommended regimen for use in individuals with HIV, including pregnant women, in resource-limited settings [61].

●Fetal safety – Short-term data on the use of lamivudine in pregnancy do not suggest major concerns about fetal safety. In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to lamivudine are available to exclude a 1.5-fold or more increase in birth defects; the prevalence of birth defects in the registry was 3.1 percent (95% CI 2.7-3.6) [34].

●Dosing – No dose adjustments are warranted for lamivudine during pregnancy. In a pharmacokinetic study of 114 pregnant and 47 nonpregnant women on long-term antiretroviral therapy that included lamivudine, rate of clearance of lamivudine was 22 percent higher during pregnancy [63]. Accordingly, plasma lamivudine levels were lower in pregnant women but comparable to levels historically observed in other nonpregnant adults.

Tenofovir

Tenofovir disoproxil fumarate — Tenofovir disoproxil fumarate is a preferred NRTI for use in pregnancy in the United States [57]. It is part of the WHO-recommended regimen for use in individuals with HIV, including pregnant women, in resource-limited settings [61].

Very preterm birth/neonatal mortality — Concerns have been raised regarding a possible increased risk of very preterm birth and neonatal death with tenofovir disoproxil fumarate (TDF) exposure during pregnancy [64], but we do not feel that the available evidence is clear enough to discontinue use of TDF as a preferred NRTI in pregnancy. However, due to concerns about a potential increase in TDF levels with concomitant lopinavir-ritonavir, we avoid initiating TDF with lopinavir-ritonavir during pregnancy.

One meta-analysis reported an increased incidence of neonatal death with TDF-based treatment [65,66], however, these point estimates were based solely on a single trial (the PROMISE trial). This trial was designed to evaluate the safety and efficacy of protease inhibitor-based ART compared with zidovudine alone in pregnant women with CD4 cell count >350 cells/microL [28]. Women were initially randomly assigned to receive zidovudine plus lamivudine and lopinavir-ritonavir (zidovudine-based ART) or zidovudine alone, but during the second period of the trial (representing 35 percent of study enrollment), women were also randomly assigned to a third arm of TDF plus emtricitabine and lopinavir-ritonavir (TDF-based ART). Among women enrolled during this second period, there was an unexpectedly higher rate of very preterm delivery (<34 weeks) and neonatal death in the TDF-based ART group compared with the zidovudine-based ART group. (See 'Preterm birth' above.)

This meta-analysis was accompanied by a British Medical Journal clinical practice guideline that suggested zidovudine-based ART rather than TDF-based ART for pregnant women based on these findings [64,65]. However, because of several uncertainties, we do not believe these results should be interpreted to indicate that TDF itself increases very preterm birth and neonatal death:

●In the PROMISE trial, TDF-based ART and zidovudine-based ART both increased the risk of preterm delivery (<37 weeks) compared with zidovudine alone. Although rates of very preterm delivery (<34 weeks) and neonatal death (mostly associated with prematurity) were higher with TDF-based ART compared with zidovudine-based ART, rates of these outcomes were not statistically different when comparing TDF-based ART and zidovudine alone. Furthermore, the rates of very preterm delivery and neonatal death with zidovudine-based ART were disproportionately low during the second part of the trial, when comparisons were made with TDF-based ART, compared with the first part of the trial, when zidovudine-based ART was compared only with zidovudine alone (15 of 17 neonatal deaths in the zidovudine-ART group occurred during the first part of the trial, when two-thirds of study patients were enrolled) [31]. Taken together, these observations suggest the possibility that the neonatal mortality rate was unusually low in the zidovudine-based ART group during that part of the trial rather than abnormally high with tenofovir-based ART.

●Furthermore, doses of lopinavir-ritonavir were increased during the third trimester during the trial, and it is possible that coadministration resulted in excess levels of TDF, as studies have previously demonstrated that standard lopinavir-ritonavir doses increase intracellular and plasma TDF levels [67,68]. Thus, it may have been this particular combination, rather than TDF itself, that was associated with very preterm birth. Studies are currently underway to further evaluate this possibility.

●Finally, several observational studies have not suggested an excess of very preterm birth or neonatal mortality with TDF-based regimens [14,69-71]. As an example, in a study of data from United States pregnancy cohorts, in utero exposure to TDF plus emtricitabine and lopinavir-ritonavir was associated with a similar rate of very preterm birth compared with zidovudine plus lamivudine and lopinavir-ritonavir (risk ratio 0.85, 95% CI 0.34-2.13), although the number of events was small [71]. Additionally, in a study of nearly 12,000 HIV-exposed infants in Botswana between 2014 and 2016, in utero exposure to TDF plus emtricitabine and efavirenz was associated with lower rates of very preterm birth and neonatal death compared with zidovudine-based ART (with lamivudine and either nevirapine or lopinavir-ritonavir) [14]. Although observational data are more susceptible to bias than randomized controlled trial results and are thus considered of lower quality, the consistent findings regarding the safety of TDF across observational studies from various locations decreases confidence in the contradictory findings from a single trial.

The meta-analysis noted above [65] also combined the outcomes of neonatal death and stillbirth from the PROMISE trial to conclude that TDF therapy increases the risk of both, but in the PROMISE trial, these outcomes were not combined, and the incidence of stillbirth was not different with TDF-based ART compared with zidovudine-based ART [28].

After review of these data, the British HIV Association and the United States Department of Human Health and Services Panel on Treatment of Pregnant Women Living with HIV and Prevention of Perinatal Transmission continue to recommend use of TDF-based ART during pregnancy [57,72].

Fetal growth and development — Short-term data on the effect of TDF during pregnancy on fetal growth and development are generally reassuring. Some but not all studies have suggested adverse effects on certain growth parameters and bone mineral density in early infancy, although the clinical and long-term significance of these findings are uncertain.

In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to TDF are available to exclude a 1.5-fold or more increase in birth defects; the prevalence of birth defects reported was 2.4 percent (95% CI 2.0-2.9) [34]. TDF has not demonstrated teratogenicity in rodents, monkeys, or rhesus macaques [73].

Studies evaluating the fetal effects of intrauterine TDF exposure have also been overall reassuring [74-76]. In a systematic review of studies evaluating outcomes among pregnant or lactating women with HIV (26 studies) and without HIV (7 studies) who received TDF, the drug appeared safe in pregnancy [74]. No statistically significant differences in most pregnancy, maternal, and infant adverse outcomes between TDF- and non-TDF-containing regimens were identified; these outcomes included pregnancy incidence, stillbirth or pregnancy loss, preterm delivery <37 week, low birth weight <2500 or <1500 grams, small for gestational age, birth defects, and infant (>14 days) or maternal mortality. The impact of TDF-containing ART on very preterm delivery (<34 weeks) and neonatal mortality are discussed elsewhere. (See 'Very preterm birth/neonatal mortality' above.)

Most studies in the systematic review reported normal infant linear growth with TDF exposure. One reported slightly lower [77] and one reported higher [78] length-for-age z-scores at one year of age in TDF-exposed infants. Results from a large randomized trial also suggested no difference in longitudinal growth (height, weight, and head circumference) at 24 weeks of age between infants exposed to maternal TDF-based ART (with a protease inhibitor) during breastfeeding and those without TDF exposure (who had instead received infant nevirapine prophylaxis) [79]. In four studies, no differences in laboratory or bone marker abnormalities were reported [74]. However, in one study of infants born to mothers with HIV, those exposed to TDF in utero (n = 74) had lower bone mineral content, as measured by dual-energy X-ray absorptiometry (DXA) scan, compared to those with no TDF exposure (n = 69) [80], potentially consistent with a rhesus macaque study that found decreased fetal bone porosity with very high-dose TDF use [81]. However, in a subsequent randomized trial, non-TDF-containing ART regimens resulted in similarly lower bone mineral content, suggesting that ART in general, rather than TDF specifically, may be associated with a decrement in neonatal bone mineral content [82]. In a randomized trial of TDF to prevent mother-to-child hepatitis B transmission, there was no significant effect of maternal TDF use compared with placebo on maternal or infant bone mineral density one year after delivery/birth [76]. Several longer-term studies of TDF-exposed infants of women with HIV followed through age 18 months in Vietnam and Malawi found that TDF use during pregnancy was not associated with bone radiologic of biomarker abnormalities, growth impairment, or renal dysfunction [83,84]. Similarly, long-term follow-up of children of hepatitis B-infected mothers who did or did not receive TDF during late pregnancy found comparable long-term growth, renal function, and bone development up to six to seven years after delivery [85].

Maternal safety — The potential adverse effects of tenofovir disoproxil fumarate have been well described in nonpregnant individuals with HIV and mainly include renal toxicity and modest bone density loss (see "Overview of antiretroviral agents used to treat HIV", section on 'Tenofovir disoproxil fumarate'). Similarly, in an African trial of breastfeeding women with HIV and CD4 counts >350 cells/microL, declines in spine and hip bone mineral density from 14 days to 74 weeks postpartum were greater among those who were randomly assigned to receive TDF-based ART (with a protease inhibitor) postpartum compared with those assigned to receive no ART (their infants received nevirapine instead) for postpartum prevention of transmission [86]. Whether this decline reverses after the cessation of breastfeeding merits further study.

Dosing — No dose adjustments are routinely recommended for TDF during pregnancy. However, because of moderate decreases in drug levels during the third trimester, special attention to virologic monitoring to ensure suppression should be conducted for women on a TDF-containing regimen. Several pharmacokinetic studies have demonstrated lower levels of TDF (area under the curve concentrations and/or trough levels) during the third trimester of pregnancy compared with postpartum or nonpregnant women [62,87,88]. However, decreases in TDF levels were not associated with virologic failure or mother-to-child transmission of HIV [62,88]. In one study, increased weight (>90 kg) was associated with TDF levels below target during pregnancy.

We avoid initiating TDF with lopinavir-ritonavir during pregnancy. (See 'Very preterm birth/neonatal mortality' above.)

Tenofovir alafenamide — Tenofovir alafenamide (TAF) is a preferred NRTI for use in pregnancy in the United States [57].

Data on the use of TAF during pregnancy are limited but accumulating. In a randomized trial that included 643 women with HIV who were initiating an antiretroviral regimen during pregnancy, those who received a TAF-containing regimen had a lower rate of preterm delivery compared with those who received an equivalent TDF-containing regimen (6 versus 9 percent) [89]. Maternal weight gain was greater with TAF than TDF (0.38 versus 0.32 kg per week), although weight gain with both was less than that recommended for the second/third trimester (0.42 kg per week). There was no detected difference in maternal or neonatal grade 3 or higher adverse events with TAF versus TDF.

In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to TAF are available to exclude a twofold or more increase in birth defects; the prevalence of birth defects reported was 4.4 percent (95% CI 2.7-6.8) [34].

No dose adjustments are routinely recommended for TAF during pregnancy. In one study of 43 pregnant women given TAF with or without cobicistat, TAF levels were lower during the second and third trimesters compared with postpartum, but were comparable to levels in nonpregnant adults [90].

Zidovudine — Zidovudine is an alternative NRTI for use in pregnancy in the United States [57]. It is used frequently in resource-limited settings, particularly as part of second-line regimens.

Fetal safety — Short-term data on the safety of zidovudine during pregnancy are generally reassuring.

There has been no evidence of an increased incidence of congenital abnormalities in infants born to women with antepartum zidovudine exposure over the general population [34,45,91,92]. The prevalence of birth defects among infants with first trimester exposure to zidovudine in the Antiretroviral Pregnancy Registry was 3.2 percent (95% CI 2.7-3.8) [34]. Other cohort studies in the United States and Europe have also not identified an association between first trimester zidovudine exposure and birth defects [93,94].

However, data from the French Perinatal Cohort have suggested an association between first trimester in utero zidovudine exposure and congenital heart disease (predominantly ventricular septal defects [VSD]) [95,96]. Among 3262 infants with first trimester zidovudine exposure, the rate of congenital heart disease was 1.5 percent compared with 0.7 percent among 9626 infants without zidovudine exposure (adjusted OR 2.2, 95% CI 1.5-3.2) [95]. Most of the defects observed were minor. This was in contrast to the Antiretroviral Pregnancy Registry data, which included 36 reported cases of VSD among over 15,000 live births with any antiretroviral exposure, without an excess of cases associated with zidovudine exposure [34,97].

Studies evaluating other outcomes have also been reassuring [45,98,99]. After nearly six years of follow-up of infants of trial participants, immunologic, neurologic, and growth parameters were similar between children who had in utero zidovudine exposure and those who were exposed to placebo [45,98].

No evidence of teratogenicity or toxicity has been observed in animal models. However, in vitro and animal study evidence has suggested carcinogenic potential. The relevance of these data to humans is unknown. No tumors have been observed in 727 children with in utero zidovudine exposure followed for over 1100 person-years [51]. While these data are reassuring, follow-up is still limited and needs to be continued into adulthood before it can be concluded that there is no carcinogenic risk.

Dosing — No dose adjustments are warranted for zidovudine during pregnancy. The pharmacokinetics of zidovudine do not appear to be changed during pregnancy [100].

NON-NUCLEOSIDE REVERSE TRANSCRIPTASE INHIBITORS

Efavirenz — Efavirenz is a non-nucleoside reverse transcriptase inhibitor (NNRTI) that is an alternative agent for use in pregnancy in the United States [57]. The World Health Organization recommends it as an alternative first-line regimen for use in individuals with HIV, including pregnant women, in resource-limited settings [61].

Fetal safety — In early studies, rare reports of birth defects in human case reports that were potentially consistent with malformations identified in animal studies had raised initial concerns about the potential for teratogenicity, but increasing data from clinical studies and reports of efavirenz use in pregnancy have been overall reassuring.

In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to efavirenz are available to exclude a 1.5-fold or more increase in birth defects; the prevalence of birth defects reported was 2.4 percent (95% CI 1.6-3.5) [34]. Additionally, a meta-analysis of 21 studies evaluating outcomes of 2026 births with first trimester in utero efavirenz exposure, including the Antiretroviral Pregnancy Registry, did not find an increased relative risk of overall birth defects comparing infants born to women receiving efavirenz-based versus non-efavirenz–based regimens (relative risk 0.78, 95% CI 0.56-1.08) [101]. Subsequent studies have also not suggested an increased risk of congenital anomalies with efavirenz exposure [35,96,102]. Although studies from two large pediatric AIDS cohorts (PACTG 219 and P1025) had reported an increased risk of birth defects with first trimester efavirenz exposure compared with exposure to other antiretroviral drugs, no specific pattern of birth defects was observed, and there was overlap in reported defects between the two papers [103,104].

Data also demonstrate that periconception efavirenz exposure is not associated with neural tube defects. In primate studies, significant central nervous system malformations (anencephaly and unilateral anophthalmia in one and microphthalmia in another) and a cleft palate were observed in 3 of 20 infant cynomolgus monkeys born to mothers who received efavirenz from gestational day 20 to 150 in doses comparable to systemic human therapeutic exposure [105]. The Antiretroviral Pregnancy Registry includes one neural tube defect (meningomyelocele) and one report of anophthalmia with severe facial clefts and amniotic banding out of 1037 first trimester exposures (giving a neural tube defect prevalence of 0.10 percent, compared with an overall prevalence of 0.06 percent in the United States) [106,107]. In a large observational study in Botswana, the prevalence of neural tube defects was 0.04 percent among 7959 infants with efavirenz exposure, similar to that among 90,000 women without HIV (0.08 percent) [108].

Two cohort studies have suggested that in utero efavirenz exposure may be associated with a higher incidence of microcephaly and lower scores on neurodevelopmental assessments among HIV-exposed children, although other studies have not reported this association [109-111]. The risk of preterm birth with efavirenz is discussed below. (See 'Maternal safety' below.)

Expert recommendations from the World Health Organization, United States, and United Kingdom do not restrict the use of efavirenz during pregnancy [57,61,112]. In many resource-limited settings, efavirenz is an alternative first-line regimen.

Selection of ART during pregnancy is discussed elsewhere. (See "Antiretroviral selection and management in pregnant women with HIV in resource-rich settings", section on 'ART selection and management'.)

Maternal safety — Important potential side effects with efavirenz include central nervous system toxicity, rash, hyperlipidemia, and elevated hepatic transaminases. However, data from randomized clinical trials suggest that the rate of adverse events with efavirenz-containing ART is similar to that with integrase inhibitor-based regimens.

As an example, in the IMPAACT 2010 VESTED trial of pregnant women starting ART at >14 weeks gestation, there was no difference in maternal or infant grade 3 or higher adverse events between efavirenz- and dolutegravir-based regimens (with either TDF or TAF) [89]. Preterm delivery was more common with efavirenz than dolutegravir-based regimens (12 percent versus 6 to 9 percent) in that trial. However, in the DolPHIN-2 trial of pregnant women starting ART at >28 weeks gestation, efavirenz was associated fewer reported serious adverse effects than dolutegravir (11 versus 22 percent), and the rate of preterm delivery was not statistically different between the two groups [113]. Finally, in a randomized clinical trial comparing initial therapy with efavirenz and raltegravir-based regimens in pregnant women starting ART at >20 weeks gestation, the rates of adverse events and preterm delivery were similar with the two regimens [114]. (See "Overview of antiretroviral agents used to treat HIV", section on 'Efavirenz'.)

Dosing — No dose adjustments are warranted for efavirenz during pregnancy. The pharmacokinetics of efavirenz during pregnancy do not appear to be significantly different from that in nonpregnant adults. In 25 pregnant women taking efavirenz as part of combination antiretroviral therapy (ART), the clearance of efavirenz was higher and the concentration at 24 hours was lower during the third trimester than several weeks postpartum [115]. However, the area under the curve remained similar at both time points and was comparable to that reported in nonpregnant adults.

Rilpivirine — Oral rilpivirine is an alternative NNRTI for use in pregnancy in the United States [57]. It is reserved for those with a viral load <100,000 copies/mL and a CD4 count ≥200 cells/microL.

Data on the use of oral rilpivirine during pregnancy are relatively limited. In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to oral rilpivirine are available to exclude a twofold or more increase in birth defects; the prevalence of birth defects reported to the registry was 1.3 percent (95% CI 0.5-2.7) [34]. Studies in rats and rabbits have revealed no evidence of fetal harm at systemic exposures substantially higher than that achieved with the standard dose.

No dose adjustments are indicated for oral rilpivirine during pregnancy. In pharmacokinetic studies, oral rilpivirine levels were approximately 30 to 40 percent lower during pregnancy [116,117]. Nevertheless, exposure levels exceeded the target during pregnancy in most participants, and there was no loss of virologic suppression.

There are limited data on the pharmacokinetics and safety of injectable rilpivirine during pregnancy. Injectable rilpivirine is not recommended for initiation or continuation of therapy in those who are pregnant [57].

Nevirapine — Nevirapine is not recommended for ART in pregnant women in the United States because of potential for maternal skin and liver toxicity [57]. It is uncommonly used in resource-rich settings and is being phased out in certain resource-limited settings where it has been in use.

Short-term data on the use of nevirapine in pregnancy do not suggest major concerns about fetal safety. In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to nevirapine are available to exclude a 1.5-fold or more increase in birth defects; the prevalence of birth defects reported to the registry was 3 percent (95% CI 2.1-4.1) [34]. Teratogenic effects have not been observed in reproductive studies with rats and rabbits.

The main toxicity concerns with nevirapine in general are a hypersensitivity reaction with severe skin rash and hepatotoxicity; the risk of these is higher with higher CD4 cell counts. Severe nevirapine-associated skin rash, hepatotoxicity, and liver failure have also been reported in pregnant women [118-123], although pregnancy is not clearly associated with a higher risk for these reactions [124]. A pregnant woman who enters care on a nevirapine-containing regimen and is tolerating ART well can continue her initial regimen, regardless of CD4 cell count.

No dose adjustments are warranted for nevirapine during pregnancy. Pharmacokinetic data on chronic antenatal nevirapine dosing in pregnant women during the third trimester have demonstrated that pharmacokinetic parameters in the pregnant woman are similar to those in nonpregnant adults [125]. Serum nevirapine elimination in the infants was accelerated compared to newborns whose mothers received only a single intrapartum dose.

PROTEASE INHIBITORS

Class effects — There are several side effects that appear to be PI class effects, whereas others are agent specific. Some of the class side effects are insulin resistance, hyperglycemia, diabetes, hyperlipidemia, lipodystrophy, hepatotoxicity, and side effects caused by interactions with other hepatically metabolized drugs.

Gestational diabetes — Although protease inhibitors are associated with insulin resistance and impaired glucose tolerance in general [126-128], most studies in pregnant woman do not indicate an increased rate of gestational diabetes with their use [129-132]. In a multicenter prospective study of pregnant women with HIV in the United States, there was no difference between the 76 women taking and the 73 not taking a protease inhibitor in the incidence of abnormal glucose tolerance following oral glucose challenge (33 versus 26 percent), gestational diabetes (8 versus 10 percent), and macrosomia >4 kg (5 versus 6 percent) [129]. The rate of glucose intolerance and gestational diabetes in this study was higher than historically reported in the general population, possibly because of the relatively high body mass index in both groups. However, this may also reflect the finding that use of antiretrovirals of any kind may increase the risk of gestational diabetes in women with HIV [132]. Pregnant women with HIV who are on combination antiretroviral therapy should undergo glucose screening consistent with standard pregnancy recommendations. However, some experts consider protease inhibitor exposure a risk factor for glucose intolerance and therefore perform earlier testing for women maintained on protease inhibitor-containing regimens. (See "Gestational diabetes mellitus: Screening, diagnosis, and prevention", section on 'Laboratory issues'.)

Preterm birth — Although conflicting, several studies that have suggested a small increased risk of preterm birth with combination ART have implicated protease inhibitors. This is discussed elsewhere. (See 'Preterm birth' above.)

Atazanavir — Ritonavir-boosted atazanavir is a preferred protease inhibitor for use in pregnancy in the United States [57].

Fetal safety — Short-term data on the use of atazanavir in pregnancy do not suggest major concerns about fetal safety. In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to atazanavir are available to exclude a 1.5-fold or more increase in birth defects; the prevalence of birth defects reported to the registry was 2.3 percent (95% CI 1.5-3.2) [34]. One study of 2580 children born to mothers with HIV suggested an increased risk of congenital anomalies, predominantly musculoskeletal and skin defects, with first-trimester atazanavir exposure [35], but this finding has not been replicated. Atazanavir did not produce teratogenic effects in rabbits or rats at systemic drug exposure achieved in humans at the recommended therapeutic dose.

Elevation in indirect (unconjugated) bilirubin occurs frequently with atazanavir use in pregnant and nonpregnant adults and children. This is attributable to atazanavir-related inhibition of hepatic UDP glucuronyl transferase enzyme. There are few data as to whether atazanavir therapy would exacerbate physiologic hyperbilirubinemia in the neonate [133].

Dosing — Dosing increases of atazanavir to 400 mg daily, boosted with low-dose 100 mg ritonavir, may be warranted for some women during pregnancy, particularly antiretroviral-experienced women. Although some experts recommend this increased atazanavir dosing in all women during the second and third trimesters, use of the standard dose (300 mg atazanavir plus 100 mg ritonavir daily) with careful virologic monitoring during the second and third trimester can also be considered. The package insert recommends increased atazanavir dosing only for antiretroviral-experienced pregnant women in the second and third trimesters also receiving either tenofovir or an H₂-receptor antagonist or antiretroviral-naïve pregnant women also receiving efavirenz. Atazanavir should not be used in patients receiving both tenofovir and H₂ receptor antagonists or in antiretroviral-experienced patients also taking efavirenz.

Several studies have observed that the area under the plasma concentration-time curves (AUC) for atazanavir are lower in pregnant than nonpregnant women during use of atazanavir-ritonavir 300 mg/100 mg once daily, and in some studies, are lower than in the same woman postpartum [133-136]. Nevertheless, most pregnant women in these studies are able to achieve an undetectable viral load by the time of delivery. Co-administration with tenofovir further lowers the AUC for atazanavir by 25 percent compared with postpartum levels in the same patient and by 50 percent compared with postpartum levels in women who did not receive tenofovir [135]. Use of an increased dose of atazanavir-ritonavir 400 mg/100 mg during pregnancy results in an AUC for atazanavir equivalent to historic nonpregnant women with HIV receiving standard-dose atazanavir [134,135]. Similarly, pregnant women receiving the increased atazanavir dose with tenofovir had an AUC equivalent to that seen in nonpregnant patients with HIV receiving standard-dose atazanavir and tenofovir [137].

Darunavir — Ritonavir-boosted darunavir is a preferred protease inhibitor for use in pregnancy in the United States [57]. Darunavir boosted with cobicistat is not recommended.

Fetal safety — Limited data on the use of darunavir in pregnancy do not suggest major concerns about fetal safety. In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to darunavir are available to exclude a twofold or more increase in birth defects; the prevalence of birth defects reported was 3.5 percent (95% CI 2.2-5.3) [34]. Infants in case reports of maternal darunavir use have described no congenital or major metabolic anomalies [138-140].

No embryotoxicity or teratogenicity has been seen in mice or rats at drug levels approximately 50 percent of that achieved in humans, nor in rabbits at drug levels 5 percent of that achieved in humans at the recommended therapeutic dose.

Dosing — Because of low trough levels with once-daily dosing in pregnancy, 600 mg/100 mg twice-daily dosing of darunavir is recommended during pregnancy [57]. In studies of darunavir-ritonavir administered as 600 mg/100 mg twice daily or 800 mg/100 mg once daily during pregnancy, darunavir levels were substantially reduced during the third trimester compared with postpartum, particularly with once-daily dosing [141-143]. However, one study noted that there was no clinically relevant decrease in levels of the active form of darunavir during pregnancy, and several case reports have described successful maternal viral suppression and prevention of transmission with typical twice-daily dosing [138-140].

Cobicistat-boosted darunavir is not recommended during pregnancy. (See 'Cobicistat' below.)

Lopinavir-ritonavir — Lopinavir-ritonavir is not recommended for use in pregnancy in the United States except in special circumstances [57]. It is used frequently in resource-limited settings for second-line therapy.

Fetal safety — Short-term data on the use of lopinavir-ritonavir in pregnancy do not suggest major concerns about fetal safety. In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to lopinavir-ritonavir are available to exclude a 1.5-fold or more increase in birth defects; the prevalence of birth defects reported was 2.1 percent (95% CI 1.4-3.0) [34]. There is no evidence of teratogenicity with administration of lopinavir-ritonavir to pregnant rats or rabbits.

Dosing — Many experts increase lopinavir-ritonavir dosing to 600 mg/150 mg twice daily during the second and third trimesters to maintain adequate serum drug levels, particularly in protease-inhibitor-experienced patients, and then return to standard dosing (400 mg/100 mg twice daily) immediately postpartum [57]. If the standard doses are used during pregnancy, virologic response and, if available, lopinavir drug concentrations should be carefully monitored. Once-daily dosing, as studied in adults, is not recommended during pregnancy. Pharmacokinetic data demonstrate a reduction in lopinavir levels during the second half of pregnancy with administration of standard doses of lopinavir-ritonavir. We also avoid initiating lopinavir-ritonavir with tenofovir disoproxil fumarate during pregnancy because of concerns for a potential increase in tenofovir disoproxil fumarate levels with concomitant lopinavir-ritonavir. (See 'Very preterm birth/neonatal mortality' above.)

Several studies have suggested that lopinavir concentrations are decreased late in pregnancy compared with postpartum or levels in nonpregnant individuals [144,145]. Reports of clinical experience suggest that most, if not all, pregnant women receiving standard lopinavir-ritonavir tablet dosing during pregnancy will have trough lopinavir levels that exceed 1.0 mcg/mL, which is the usual trough concentration recommended for monitoring antiretroviral-naïve subjects, but not the higher trough concentrations recommended for protease inhibitor-experienced patients [146,147]. A prospective study of 33 women with HIV evaluated the pharmacokinetics with the following dosing strategy: two lopinavir-ritonavir tablets (400 mg/100 mg) twice daily during the second trimester, three tablets (600 mg/150 mg) twice daily during the third trimester, and two tablets (400 mg/100 mg) twice daily post-delivery through two weeks postpartum [148]. Drug serum levels were stable throughout and comparable to those levels seen in nonpregnant adults taking standard dosing of lopinavir-ritonavir (eg, two tablets twice daily).

INTEGRASE INHIBITORS

Dolutegravir — Dolutegravir is a preferred integrase inhibitor for use in pregnancy in the United States [57]. The World Health Organization (WHO) also recommends dolutegravir as part of a first-line regimen in all populations of patients with HIV, including pregnant women and women of childbearing potential [149].

Fetal safety — Dolutegravir use at the time of conception appears to be associated with a small risk of neural tube defects, but it is not statistically different than the risk with non-dolutegravir-based ART. The risk is limited to the first few weeks of pregnancy. Use of dolutegravir in nonpregnant women of childbearing potential is discussed elsewhere. (See "HIV and women", section on 'Individuals of childbearing potential'.)

In a report of an ongoing birth-outcomes surveillance study in Botswana that included data through March 2021, neural tube defects were documented in 9 of 5860 infants (0.15 percent) born to mothers taking dolutegravir at conception compared with 22 of 22,475 infants (0.10 percent) born to mothers taking non-dolutegravir ART at conception; this difference was not statistically significant [150]. There were three neural tube defects among 5535 infants (0.05 percent) born to mothers who started dolutegravir during pregnancy and 97 among 144,967 infants (0.07 percent) born to women without HIV infection; these rates remain significantly lower than those observed with dolutegravir at conception. The initial report from this study had previously suggested a higher risk of neural tube defects with dolutegravir use at the time of conception (observed in 4 of 426 infants, or 0.94 percent); the estimated magnitude of the risk markedly decreased after analysis of more exposures [151-153]. Another observational study from a different region of Botswana had documented one neural tube defect among 152 infants born to mothers taking dolutegravir at conception compared with none among 381 infants born to mothers taking non-dolutegravir ART at conception, but these numbers are too small to make an accurate risk assessment [154].

Other observational studies had suggested that rates of stillbirth, neonatal death, small for gestational age, and preterm birth with dolutegravir-based regimens (started before or during pregnancy) are comparable with other antiretroviral regimens in women with HIV [108,155,156]. In the Antiretroviral Pregnancy Registry prospective reports of pregnancy outcomes, the calculated rate of overall birth defects after first-trimester dolutegravir exposure was 3.3 percent (95% CI 2.0-5.3 percent); one neural tube defect has been reported with preconception dolutegravir exposure in the Antiretroviral Pregnancy Registry, but there were only 357 preconception dolutegravir exposures at the time of that report [34].

Dosing — No dose adjustments are indicated for dolutegravir during pregnancy [157,158]. Third trimester and postpartum dolutegravir pharmacokinetics were evaluated in the context of a randomized trial comparing dolutegravir versus efavirenz-based ART among 60 Ugandan and South African ART-naïve women with HIV [157]. Dolutegravir exposures were lower in the third trimester and postpartum period than those previously reported in nonpregnant individuals; however, only one had dolutegravir levels below the protein-adjusted IC90.

Dolutegravir should not be administered within two hours of iron- or calcium-containing preparations, including prenatal vitamins.

Raltegravir — Raltegravir is a preferred integrase inhibitor for use in pregnancy in the United States [57].

Maternal safety — In a randomized clinical trial (NICHD P1081) comparing raltegravir and efavirenz-based regimens in pregnant women starting ART at >20 weeks gestation, raltegravir was associated with a more rapid decline in viral load than efavirenz; rates of grade 3 or higher maternal adverse events, preterm delivery, and other adverse pregnancy outcomes were similar with the two regimens [114].

Fetal safety — Short-term data on the use of raltegravir in pregnancy do not suggest major concerns about fetal safety. In a randomized trial of pregnant women starting ART at >20 weeks gestation (NICHD P1081), the rates of grade 3 or higher adverse events among infants born to mothers receiving raltegravir versus efavirenz were similar; the frequency of congenital defects was low and similar with the two regimens, and no neural tube defects were observed [114]. In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to raltegravir are available to exclude a twofold or more increase in birth defects; the prevalence of birth defects reported was 3.1 percent (95% CI 1.7-5.1) [34]. Extranumerary ribs were found in rats at dose exposure threefold higher than human.

In an industry-sponsored review of women in resource-rich settings who used raltegravir during pregnancy, there were no neural tube defects documented among the 1991 prospectively reported pregnancies (including 456 with raltegravir use at the time of conception) [159]. While reassuring, the number of periconception exposures is insufficient to determine a lack of association with neural tube defects. Among the 435 pregnancies that were retrospectively reported, there were three neural tube defects, one of which occurred with raltegravir exposure at the time of conception.

Dosing — Raltegravir should only be dosed twice daily in pregnancy; otherwise, no dose adjustments are warranted. In studies of pregnant women who received the standard dose of 400 mg twice daily as part of their antiretroviral regimen, there was great interindividual variability in serum levels but no clinically significant reductions in third trimester concentrations compared with postpartum [160,161]. Raltegravir should not be administered within two hours of iron or calcium containing preparations, including prenatal vitamins.

Elvitegravir — Elvitegravir (combined with cobicistat) is not recommended for initiation in pregnancy until more data are available [57]. In a pharmacokinetic study of 30 women taking an elvitegravir-cobicistat-containing regimen during pregnancy, the elvitegravir area under the curve was reduced by 24 and 44 percent in the second and third trimesters, respectively, compared with postpartum levels, and trough levels were 81 and 89 percent lower in the second and third trimesters; viral suppression at delivery was observed in 76 percent [162]. These data raise concern that subtherapeutic drug levels may occur among some women receiving elvitegravir during pregnancy, with viral rebound leading to potential increased risk of perinatal transmission. A subsequent retrospective study of 134 pregnant women on an elvitegravir-cobicistat-containing regimen reported rates of viral suppression of 81 percent at delivery overall and 88 percent of those who began elvitegravir prior to conception, with perinatal transmission observed in one of 134 infants (0.8 percent) [163]. Additional data are needed to confirm the efficacy of elvitegravir-cobicistat during pregnancy before recommending its use.

In the retrospective study mentioned above, there were no cases of neural tube defects among the 82 women who were taking elvitegravir at the time of conception; however, the number of periconception exposures is insufficient to determine a lack of association with neural tube defects [163]. In the Antiretroviral Pregnancy Registry, sufficient data on first-trimester exposures to elvitegravir are available to exclude a twofold or more increase in birth defects; the prevalence of birth defects reported was 3.1 percent (95% CI 1.5-5.4) [34]. Studies in animals that have received elvitegravir have not demonstrated evidence of teratogenicity.

Bictegravir — Data on bictegravir in pregnancy are extremely limited. Given the lack of data on pharmacokinetics and safety in pregnant women, this drug is not recommended for initial therapy in antiretroviral-naïve pregnant women.

Cabotegravir — Data on cabotegravir in pregnancy are extremely limited. Given the lack of data on pharmacokinetics and safety during pregnancy, this drug is not recommended for initial therapy or continuation of therapy in those who are pregnant [57].

PHARMACOLOGIC BOOSTERS

Ritonavir — Short-term data on the use of ritonavir in pregnancy do not suggest major concerns about fetal safety. In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to ritonavir are available to exclude a 1.5-fold or more increase in birth defects; the prevalence of birth defects reported was 2.3 percent (95% CI 1.8-2.9) [34].

Cobicistat — Cobicistat is not recommended for initiation in pregnancy until more data are available [57]. Emerging data have suggested that second- and third-trimester levels of elvitegravir-cobicistat are reduced by approximately 50 percent and are associated with loss of virologic suppression at the time of delivery [162]. Similar decreases in second- and third-trimester drug levels and associated virologic failure were observed with darunavir-cobicistat [164].

Studies in animals that have received cobicistat have not demonstrated evidence of teratogenicity. In the Antiretroviral Pregnancy Registry, sufficient data on first trimester exposures to cobicistat are available to exclude a twofold or more increase in birth defects; the prevalence of birth defects reported was 3.5 percent (95% CI 2.0-5.7) [34].

OTHER AGENTS — Antiretrovirals that are not commonly used or used only in limited situations are not discussed in this topic. Detailed information on the safety and pharmacokinetics of these antiretrovirals can be found in the United States Department of Health and Human Services guidelines on reducing perinatal HIV transmission [57].

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: HIV treatment in pregnant women".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here is the patient education article that is relevant to this topic. We encourage you to print or e-mail this topic to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Beyond the Basics topic (see "Patient education: HIV and pregnancy (Beyond the Basics)")

SUMMARY

●Antiretroviral therapy (ART) is recommended for all people with HIV, including pregnant women. An additional goal is to decrease the risk of perinatal transmission of HIV infection. Overall, the known benefits of combination ART for pregnant women outweigh the known and potential risks. Nevertheless, decisions regarding ART in pregnant women are complex and need to take into consideration changes in pharmacokinetics and potential toxicities to the mother and fetus. (See 'Introduction' above and 'General principles' above.)

●Long-term data on the safety of in utero drug exposure in humans are not available for any antiretroviral drug; however, with rare exceptions, the short-term data are reassuring. Health professionals are encouraged to contact the Antiretroviral Pregnancy Registry to prospectively report antiretroviral exposures in pregnancy to assist in monitoring pregnancy outcomes of pregnant women exposed to antiretroviral medications. Information on in utero antiretroviral exposure should be maintained in a child's medical record in case problems that could be long-term complications are identified in the future. (See 'General principles' above and 'Pregnancy outcomes with combination ART exposure' above.)

●Uncertain evidence from a single clinical trial that suggested an increase in preterm birth and neonatal mortality with tenofovir disoproxil fumarate (TDF)-based therapy compared with zidovudine-based therapy (each combined with lopinavir-ritonavir) warrants further study. We avoid initiating TDF with lopinavir-ritonavir during pregnancy. (See 'Very preterm birth/neonatal mortality' above.)

●Other short-term data on the safety of commonly used nucleoside reverse transcriptase inhibitors (NRTIs) during pregnancy do not suggest major concerns. Asymptomatic anemia and neutropenia have been reported in infants with in utero NRTI exposure. Some studies have suggested adverse effects on certain growth parameters and bone mineral density in early infancy with in utero exposure to tenofovir, but the long-term significance of these is unclear. Data on an association between in utero zidovudine exposure and congenital heart defects are inconsistent and also of unclear long-term significance. (See 'Nucleoside reverse transcriptase inhibitors' above.)

●Increasing data from clinical studies and reports of efavirenz use during pregnancy have been reassuring despite initial concerns about teratogenicity potential. Recommendations from expert groups do not restrict efavirenz use during pregnancy. Data on the use of rilpivirine during pregnancy are limited. (See 'Non-nucleoside reverse transcriptase inhibitors' above.)

●Short-term data on the safety of commonly used protease inhibitors in pregnancy do not suggest major concerns. Although protease inhibitors are associated with insulin resistance and impaired glucose tolerance in general, most studies in pregnant woman do not indicate an increased rate of gestational diabetes with their use. Several studies that have suggested a small increased risk of preterm birth with combination ART have implicated protease inhibitors, but data are conflicting. Plasma levels of many protease inhibitors are reduced during pregnancy and dosing adjustments may be warranted in certain situations. (See 'Protease inhibitors' above.)

●There appears to be a very small risk of neural tube defects with dolutegravir exposure at the time of conception, but it is no longer statistically different from the risk with non-dolutegravir ART exposure at the time of conception; such risk is limited to use during the first few weeks of pregnancy. There are only limited data on periconception exposures to other integrase inhibitors. Lower levels of cobicistat-boosted agents during pregnancy have been associated with loss of virologic suppression at delivery; hence, initiation of cobicistat-containing regimens during pregnancy is not recommended. (See 'Integrase inhibitors' above and 'Cobicistat' above.)

●More detailed information on the safety and pharmacokinetics of these and other antiretrovirals can be found in the United States Department of Health and Human Services guidelines on reducing perinatal HIV transmission. The selection of antiretroviral regimens for use during pregnancy in resource-rich and resource-limited settings is discussed in detail elsewhere. (See "Antiretroviral selection and management in pregnant women with HIV in resource-rich settings", section on 'ART selection and management' and "Prevention of mother-to-child HIV transmission in resource-limited settings", section on 'Maternal antiretroviral use'.)