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


Between 2008 and 2017, we identified 808 women exposed to INSTI-based ART during pregnancy. Among these, 37% (N = 301) were exposed at conception, 23% (N = 183) started INSTI-based ART during pregnancy as first-line ART and 40% (N = 324) switched during pregnancy from a non-INSTI-based ART to a INSTI-based ART. Raltegravir was most often received (N = 703), followed by dolutegravir (N = 57) and elvitegravir (N = 48).

Characteristics of patients differed significantly among the three groups of timing of INSTI exposure (Table 1). Women receiving INSTI at conception were older, more often from metropolitan France and over 40% were living with HIV for over 10 years. They had a lower viral load and a higher CD4+ cell count at delivery than the two other groups (P < 0.01 for all differences).

Perinatal outcomes did not differ significantly among the three groups (Table 2); however, when restricting to raltegravir, there was a slightly higher rate of birth defects in infants exposed at conception vs. infants exposed later in pregnancy (6.7 vs. 2.9% and 2.5%, respectively, P = 0.04, Table 3). The odds ratio (OR) for the association between birth defect and raltegravir exposure at conception vs. started as first-line ART during pregnancy was 2.4 [95% confidence interval (95% CI) 0.8–6.7, P = 0.09]. After adjusting on maternal age, ethnic origin and multiple pregnancy, the magnitude of the effect was decreased: aOR = 1.6 (95% CI 0.5–4.8), P = 0.37.

When compared with matched controls, INSTI exposure at conception was not significantly associated with a higher risk of birth defects (5.7 vs. 2.9%, P = 0.13, Table 4). When restricting to raltegravir exposure, there was a trend towards more birth defects, but it did not reach significance: 6.4 vs. 2.3%, P = 0.08 (Table 5). There was no cluster of birth defect type among raltegravir-exposed children and no neural tube defects were observed in this population (Table 6). In order to evaluate the potential role of other drugs and to allow comparison with the literature, we described NRTI-backbone-exposure: in the group exposed to raltegravir at conception, and their matched counterparts: the backbone included zidovudine in 10 cases (6%). Most NRTI-backbones included tenofovir disoproxil fumarate (102 cases, 60%). The sensitivity analysis conducted by considering 'at least one twin' with a birth defect did not change the numbers.

Other perinatal outcomes, such as preterm birth and stillbirths, did not differ significantly between INSTI-exposed women and their darunavir-exposed matched counterparts, or when restricting the comparison to raltegravir-exposed women, whatever the timing of ART initiation (Table 4). Among stillbirths, two were associated with birth defects: one complex heart defect (exposed to darunavir at conception) and one trisomy 18 (exposed to raltegravir at conception). Other stillbirths were unexplained according to the database.

Preterm birth and stillbirth rates were not different when comparing dolutegravir or elvitegravir-exposed women to matched INSTI unexposed women (data not shown). A small number of women could not be matched exactly (N = 55 Group1, N = 58 Group 2 and N = 19 Group 3). Outcomes did not differ between INSTI-exposed mother-infant pairs that could be matched and those that could not be matched.