Pregnancy and Neonatal Outcomes Following Prenatal Exposure to Dolutegravir

Vani Vannappagari, MBBS, MPH, PhD; Claire Thorne, MSc, PhD; for APR and EPPICC


J Acquir Immune Defic Syndr. 2019;81(4):371-378. 

In This Article


The 2 separate analyses using data from APR and EPPICC account for many of the prospectively collected, prenatal exposures of dolutegravir to date. For most pregnancies included in these analyses (APR, 65%; EPPICC, 58%), dolutegravir was initiated during or before the first trimester (including 158 outcomes after dolutegravir exposure at conception), thus providing important safety information for use of dolutegravir-based regimens during early pregnancy. More than 90% of pregnancies from both data sets resulted in live births. Induced and spontaneous abortions primarily were reported in women exposed to dolutegravir during their first trimester of pregnancy (as expected, because most abortions occur in the first trimester).[15,16] Rates of both induced and spontaneous abortions among APR (2.3% and 4.2%, respectively) and EPPICC (both 1.2%) cohorts who received dolutegravir are lower than recent estimates among general populations for these outcomes; however, neither study is designed to capture early pregnancy loss or termination prevalence.[17–19]

Of the 231 singleton live births in the APR data set without defects and with prenatal exposure to dolutegravir, 10.4% were delivered at <37 weeks (preterm) and 12.6% had birth weights <2500 g. Of the 79 singleton live births and 1 stillbirth in the EPPICC data set, 13.8% were delivered preterm, 16.9% of neonates had LBW (1500–2499 g; none were <1500 g at birth), and 18.7% were SGA. Congenital abnormalities occurred in 2.7% and 4.9% (95% CI: 1.4 to 12.2) of live births from APR and EPPICC, respectively, with most anomalies being polydactyly, a common birth defect.[20] Congenital upper-extremity abnormalities comprise approximately 10% of all birth defects;[21] by contrast, the birth prevalence rate of radial polydactyly is approximately 25 per 100,000 births.[22] The risk factors associated with polydactyly include race/ethnicity (predominance of postaxial polydactyly in people of African descent),[23] male sex,[24] birth order,[25] and maternal smoking.[26]

The preterm delivery rate of 14% of the 79 pregnancies assessed in women exposed to dolutegravir-containing regimens from EPPICC was similar to that reported among women living with HIV infection who delivered during the same period in the United Kingdom—the country accounting for approximately 60% of the studied women. In a recent analysis of data from the UK and Ireland National Study of HIV in Pregnancy and Childhood regarding pregnant women taking protease inhibitor–based or non-nucleoside reverse transcriptase inhibitor–based ARV treatment regimens who delivered neonates between 2007 and 2015, the preterm delivery rate was 11.5% for women taking ARV therapy at conception with CD4+ counts ≤350 cells/μL.[27] The 10.4% preterm delivery rate in the APR cohort is similar to rates reported in overall populations in North America (10.6%; excluding Mexico)[28] and worldwide (11.1%; approximately 14.9 million preterm deliveries).[29] However, a meta-analysis of prospective and retrospective cohort studies conducted across developed and developing countries found that women with HIV infection were at increased risk of preterm delivery (summary odds ratio 1.56; 95% CI: 1.49 to 1.63) when compared with their uninfected counterparts.[30]

Approximately 10% of the women in the EPPICC cohort were vertically infected themselves, and an additional 9% were coinfected with hepatitis C virus (HCV)/HIV. These proportions are higher than expected considering the larger cohort populations of pregnant women with HIV infection living in Western Europe from which they are drawn.[12,31] They most likely reflect the preferential use of dolutegravir in specific groups of women. Factors such as coinfection with HCV/HIV may also be associated with worse pregnancy outcomes.[31]

The observed LBW prevalence among singleton live births without defects in APR (10.6%) was higher than reported US rates in the general population (8.2%) in 2016[32] but lower than expected compared with a recent analysis of APR birth outcomes through 2011 reporting birth weight <2500 g was 15.9%.[33] Although 5 cases were reported, the prevalence of VLBW (<1500 g) after dolutegravir exposure (2.2%) was similar to earlier APR data (2.1%) but higher than the 2016 prevalence of the general US population (1.4%).[32] Overall, a meta-analysis of reports comparing pregnancy outcomes in women with or without HIV found that those living with HIV infection are at increased risk of delivering a neonate with LBW compared with their uninfected counterparts (summary odds ratio 1.73; 95% CI: 1.64 to 1.82).[30]

Birth defect numbers were comparable across APR (7 of 255) and EPPICC (4 of 81) data sets. CIs relating to EPPICC data were large, reflecting the small sample size, and that considerable heterogeneity was observed across cohorts (eg, age, region, CD4+ count, timing of dolutegravir exposure). When applying the EUROCAT classification of birth defects to EPPICC (which excludes isolated tongue-tie and hyperpigmentation on the back), the rate decreased to 2.5%. Through January 31, 2018, among APR reports with any ARV exposure during pregnancy, 516 birth defects were identified of the 18,660 live births, with a prevalence of 2.8 birth defects per 100 live births (95% CI: 2.5 to 3.0).[34] This proportion was not substantially higher than the 3.0 and 4.6 per 100 live births reported in other similar US population-based databases, ie, MACDP and the Texas Birth Defects Registry, respectively.[35,36] In the APR, ≥200 first-trimester exposures to any individual drug are needed to estimate the overall prevalence of birth defects, a statistical threshold the present assessment of dolutegravir pregnancy exposure does not meet.

There are multiple limitations with these 2 analyses. It is also possible that birth defects may go unrecognized or be differentially reported. Unlike the APR, data relating to birth defects from participating studies in EPPICC are not reviewed by a teratologist. Conclusions based on the proportions of induced and spontaneous abortions were limited because some cohorts included patients first enrolled in the cohort later in the pregnancy than when most abortions occur. First-trimester exposures in both cohorts are not large enough for definitive conclusions. Conclusions about regional differences in pregnancy outcomes between the 2 studies are limited because the APR and EPPICC cohorts included 18 (6.8%) and 61 (60.4%) patients from the United Kingdom/Ireland, respectively, and it is possible that the some of the 18 participants may have been included in both cohorts. Similarly, because the EPPICC cohort included participants from 6 cohorts in different countries, differences in study design could affect overall conclusions of pregnancy outcomes in Western Europe. Further postmarketing surveillance within EPPICC is ongoing.

HIV-positive women of childbearing potential, pregnant women, and their health care professionals should follow recommendations on dolutegravir use.[9,37,38] In the United Kingdom and Ireland, the percentage of pregnant women exposed to dolutegravir-based regimens has increased >10-fold from 0.3% in 2015 to 3.3% in 2016.[39] Even with this analysis of pregnancies with first-trimester exposure to dolutegravir combined from APR and EPPICC, the expanding use of dolutegravir, together with the potential safety signal of neural tube defects reported with dolutegravir exposure at the time of conception from Botswana,[9] highlights the importance for continuing the prospective monitoring of pregnant women and their infants through APR and EPPICC. Following recommendations from global, US, and European agencies,[9,37,38] conceptions on dolutegravir may decline, but continued monitoring is critical. Our assessment of these collective 198 birth outcomes after first-trimester dolutegravir exposure from 2 prospective registries (APR, 156 outcomes; EPPICC, 42 outcomes) and data cited from 3 other studies with 104 pregnancy outcomes identified no birth defect signal, either in defect prevalence or in defect-type clustering.[40–42] In addition, among birth defects reported after dolutegravir exposure during pregnancy or at the time of conception, none involved the neural tube. These findings may provide some reassurance to women whose pregnancies have already been exposed to dolutegravir or who have limited therapeutic options.