Background:  Efavirenz is a mainstay of first-line HAART in South Africa and elsewhere. Due to concerns about teratogenicity, the drug is contraindicated in women with intentions to become pregnant; these women are typically initiated on nevirapine. Despite this, women who initiate HAART with efavirenz become pregnant during follow-up.

Methods:  We analyzed prospectively data from the Themba Lethu Clinic, an urban HAART clinic in Johannesburg. We used Cox proportional hazards regression to examine risk factors for pregnancy after HAART initiation. We clinically investigated a subset of pregnancies in women receiving efavirenz for birth outcomes.

Results:  Between April 1, 2004 and March 31, 2007, 5011 women initiated HAART in the clinic; of those, 351 (7%) became pregnant, a rate of 4.1 pregnancies per 100 woman-years (95%CI 3.7 to 4.5). Rate of incident pregnancy during follow-up was lower among those aged >35 years, with a lower CD4 count, or employed. As expected, women who initiated HAART with efavirenz were less likely than nevirapine initiators to become pregnant (HR = 0.6, 95%CI 0.4 to 0.8). However, the majority of pregnancies (68%, n = 238) occurred in efavirenz initiators. Of these 238 pregnancies, 136 were investigated for pregnancy outcomes. There were 3 maternal deaths, 39 women were lost to follow-up or could not be contacted, 1 woman refused interview, and 12 women were still pregnant at time of interview. Of the remaining 81 pregnancies, 8 elected voluntary termination of the pregnancy, and 13 miscarried. There were 60 lives births, of which 41 were examined using the Denver Developmental Screening Test. The Denver scale classified 30 infants as “within normal limits” and 11 as “suspect” for neurodevelopment delay. No congenital abnormalities were found.

Conclusions:  Pregnancy after initiation of efavirenz-based HAART is relatively common in our cohort. We found no evidence of congenital defects, but we found that >25% of examined infants were suspect for neurodevelopmental delay. In addition, we found a relatively high risk of miscarriage. While further study including a formal comparison group is required to ascertain whether neurodevelopmental delays or miscarriages are attributable to efavirenz use, rather than to HIV disease or to HAART generally, these results suggest that the risk of efavirenz in pregnancy may be less catastrophic than feared.