Risk of Birth Defects and Perinatal Outcomes in HIV-Infected Women Exposed to Integrase Strand Inhibitors During Pregnancy

Jeanne Sibiude; Jérôme Le Chenadec; Laurent Mandelbrot; Catherine Dollfus; Sophie Matheron; Nathalie Lelong; Véronique Avettand-Fenoel; Maud Brossard; Pierre Frange; Véronique Reliquet; Josiane Warszawski; Roland Tubiana


AIDS. 2021;35(2):219-226. 

In This Article


The incidence of birth defects was not significantly higher in pregnancies with exposure to INSTI at conception, in comparison with darunavir-exposed pregnancies. We did actually observe a higher rate of birth defects in children exposed to raltegravir at conception when compared with later exposure to raltegravir and a trend when compared with the darunavir-exposed matched controls. However, due to the small number of women and children included, the absence of significant difference may be due to insufficient statistical power. The number of pregnancies with exposure to dolutegravir as well as to elvitegravir was small, in accordance with French perinatal guidelines, which were to avoid their use in pregnancy.

However, the absence of cluster of malformation by organ and the absence of neural tube defects is reassuring, pleading against a causal class effect. Our study is the first to compare raltegravir-exposed women with matched controls receiving the same NRTIs at the same moment of pregnancy, differing thus only by the raltegravir exposure. The absence of increased birth defects is consistent with published data from the UK, from a comprehensive prospective cohort, with comparable numbers,[10] which compared birth defect rates among women exposed at conception (2.25%) with those exposed later in pregnancy (2.8%). As in our study, birth defects reported were mostly heart or limb abnormalities. Other data on birth defects and raltegravir are from noncomprehensive reports, wherein reporting of pregnancy may be done individually on a voluntary basis and thus not systematically, assembling various databases and data collection. The largest to date has described no increase in the risk of neural tube defects among 456 women exposed at conception but unlike our study, other birth defects were not reported.[17] Finally, a study using the French National Health system database also did not find significant differences in birth defect rates.[18] However, the diagnoses were extracted only from coding at birth, which is propitious to underreporting. Indeed, only three birth defects (1.3%) were reported among raltegravir-exposed women, which excluded all diagnoses made after discharge. In contrast, in our study, children were followed by paediatricians until the age of 2 years, which made accurate and comprehensive diagnoses of birth defects possible.

The birth defect rate among women exposed to darunavir at conception was 2.9%, which is lower than the rate we reported previously in EPF (4.4%).[5] However, in that previous study, most women were exposed to zidovudine containing combinations, found to be associated with congenital defects. Overtime, the zidovudine backbone has been progressively replaced by tenofovir. In the present study, most women were exposed to tenofovir at conception, rather than zidovudine, and the birth defect rate we report here is close to the rate of 3.6% that was reported in women exposed to tenofovir in the first trimester in our previous study. The results are thus consistent in terms of birth defect rates.

Regarding preterm delivery, there was no difference in our population between raltegravir-exposed women and matched women, whatever the timing of initiation of ART. We expected a possible inferior rate of preterm birth in raltegravir-exposed vs. darunavir-exposed pregnancies, because boosted-protease inhibitors have been shown to be associated with preterm birth.[19–21] However, all studies that showed a difference between protease inhibitor based ART and other regimens were comparing lopinavir-based ART to either NRTI monotherapy[21] or three NRTIs,[20] or nonboosted protease inhibitors.[19] It is possible that darunavir might be less associated with preterm birth than lopinavir in a specific drug-effect. The preterm birth rate in women starting raltegravir or darunavir during pregnancy (12.8 and 11.2%, respectively) was lower than the rate reported among women starting lopinavir during pregnancy in our earlier publication on protease inhibitor containing ART regimen started during pregnancy (14.4%).[19] This rate remains higher than the preterm birth rate in the French general population, which is 7%,[22] but this may be due to other nonmeasured factors particular to the population of women living with HIV and not solely to ART. There are no published data to date on the risk of preterm birth associated with raltegravir.

We found no difference in stillbirth rates nor in birth weight, length or head circumference according to INSTI exposure, which is reassuring.

Our study presents many strengths. It is the first to match INSTI-exposed pregnancies according to coexposure to other antiretroviral drugs and timing of ART initiation, as well as other factors potentially associated with birth defects, such as geographic origin and twins. Most studies compare women exposed at conception to women exposed later, but we showed that these groups differ with many characteristics that may be confounding factors. Women on raltegravir at conception usually received this treatment before the wish to be pregnant was known, and then it was decided not to switch an effective treatment during pregnancy; whereas women for whom raltegravir was prescribed during pregnancy were generally either late-presenters or women for whom the first ART was ineffective, as shown by the high rate of detectable viral-load in these two groups. We also matched according to centre and year of delivery to eliminate possible biases related to local differences in diagnosis of malformations or a time-effect. It is a prospective study with high-quality data collection filled out by the clinicians. Our data are comprehensive with over 95% inclusion in the participating centres. The major limitation is the small number of INSTI-exposed pregnancies, in accordance with French national guidelines that favour protease inhibitor based therapy for women wishing to become pregnant, thus limiting the power of our analyses. This highlights the importance on maintaining an active research on the possible side effects of any ART given during pregnancy and especially those that have a high placental transfer ratio such as INSTIs.[23,24] Another limitation is that we cannot exclude possible confounding factors that may have not been taken into account, such as viral load and CD4+ cell count at conception, or preexisting conditions and concurrent drugs potentially associated to birth defects as these were not available in our study.

In conclusion, in our study, we were not able to find a significant difference in the risk of birth defect between women exposed to raltegravir at conception and women matched on ART backbone, and factors that could be associated with birth defects, but this may be due to limited power. There was no cluster of birth defects and no neural tube defects. No other signal was observed for stillbirth, preterm birth or neonatal measurements. French current guidelines state that raltegravir is a possible alternate treatment when the first-intention treatment, which is protease inhibitor based ART, is not suitable. However, these guidelines also state that the data on raltegravir and pregnancy is mostly on late-pregnancy prescription and that the quality of the data available is limited. Our results need to be re-evaluated by larger prospective surveillance data, as these drugs are increasingly prescribed in women living with HIV.