BOSTON – World Health Organization guidelines on the use of isoniazid to prevent tuberculosis in HIV-infected women during pregnancy may need to be reconsidered in light of new evidence that isoniazid preventive therapy (IPT) is associated with a high risk for adverse pregnancy events, investigators say.

Among 156 HIV-infected pregnant women, the rate of adverse pregnancy outcomes was 23% for those randomly assigned to immediate IPT during pregnancy versus 17% for women randomized to IPT that was delayed until 12 weeks postpartum, a significant difference (P = .009), reported Amita Gupta, MD , from Johns Hopkins University in Baltimore.

“These pregnancy outcomes included in utero demise – stillbirth – and low birth weight – less than 2500 grams,” she said at the annual Conference on Retroviruses and Opportunistic Infections.

There were six maternal deaths: two in the immediate therapy arm and four in the delayed arm. Two of the deaths were related to isoniazid-induced liver failure, and two other instances of liver failure were from other causes. The two remaining deaths were deemed unrelated to isoniazid: One woman died from bacterial sepsis, and the other from pneumonia.

The incidence of TB infections was low in both study arms, Dr. Gupta noted.

“So should we really be prioritizing IPT in pregnancy or give women the choice to know what the facts are and to select if they feel that they want to take on IPT during pregnancy or defer it until after they deliver?” she said at a briefing following her presentation of the data in an oral abstract session.

WHO guidelines for management of latent tuberculosis infections state that “[a]s isoniazid and rifampicin, the drugs commonly used in preventive treatment, are safe for use in pregnant women, pregnancy should not disqualify women living with HIV from receiving preventive treatment. Nevertheless, sound clinical judgment is required to determine the best time to provide it.”

However, the quality of evidence to support that statement is weak, primarily because pregnant women were typically excluded from IPT or other tuberculosis-prevention trials. In addition, isoniazid has been associated in retrospective studies with increased hepatotoxicity in women both during pregnancy and in the postpartum period, Dr. Gupta pointed out.

She and her colleagues conducted the phase 4 IMPAACT P1078 trial to test the hypothesis that IPT can be initiated safely during pregnancy. The study was conducted at centers in Botswana, Haiti, India, South Africa, Tanzania, Thailand, Uganda, and Zimbabwe.

HIV-infected pregnant women from 14 through 34 weeks of gestation who live in a high TB burden area (prevalence of 60 cases or more per 100,000 population) but had no evidence of TB infection were randomly assigned on a 1:1 basis to receive either immediate therapy with isoniazid 300 mg daily for 28 weeks, followed by placebo, or to the same dose of isoniazid started 12 weeks postpartum for 28 weeks. All patients also received vitamin B6 and a prenatal multivitamin until 40 weeks postpartum.

The participants were stratified by gestational age (14 to less than 24 weeks and 24 through 34 weeks). Women were excluded if they were suspected of having active TB, reported recent exposure, or had received TB treatment for more than 30 days in the past year. Women with recent acute hepatitis or peripheral neuropathy were excluded.

All women and their infants received the local standard of care for HIV. The investigators performed intensified TB case finding by using the WHO symptoms screening and exam, monitoring of signs and symptoms, conducting liver function tests, and screening for peripheral neuropathy.

“We had higher than expected adverse events in the study, but there was no statistical difference between arms,” she said.

In an intention-to-treat analysis, the rate of first maternal treatment-related grade 3 or greater adverse event or permanent drug discontinuation caused by toxicity (the primary endpoint) was 15.5% in the immediate IPT arm and 15.2% in the delayed IPT arm, a nonsignificant difference. The immediate therapy arm approached but did not quite reach the prespecified boundary of noninferiority, Dr. Gupta noted.

The per-protocol analysis of the primary endpoint was similar, at 17.6% vs. 17.8%, respectively.

In the intention-to-treat analysis, any-cause grade 3 or greater maternal adverse events were seen in 30.5% of women in the immediate arm versus 28.4% in the delayed arm, with an incidence-rate difference of 4.2 per 100 person-years, which did not reach the noninferiority boundary.

The respective rates in the per-protocol analysis were 33% vs. 30.4%, for an incidence-rate difference of 4.3 per 100 person-years.

In both groups, elevated liver enzymes and weight loss were the most common maternal adverse events.

All cause hepatotoxicity occurred in 6% of participants in the immediate arm and 7% in the deferred. Rates of permanent discontinuation because of toxicity were 4% and 6%, respectively, along with the two women in the immediate IPT arm and four in the delayed arm who died during the study. There were no significant differences in these outcomes between the groups.

Dr. Gupta noted that higher liver function test results were seen after delivery, regardless of whether the women were on IPT or what type of antiretroviral therapy they were receiving.

There were no significant differences seen in infant safety by treatment arm, a secondary endpoint.

“What did we learn? We learned that we had higher than expected adverse events that were at least possibly attributed to IPT in both arms. We did not reach our noninferiority margin partly because of the high rates, but we also did not find any major significant differences between the immediate and the deferred arm in terms of maternal safety when it was looked [at] by itself,” she said.

Although there were no significant differences in any maternal hepatotoxicity, grade 3 or greater infant adverse events, or maternal or infant death by treatment arm, “we did find an important distinction in terms of difference by adverse pregnancy outcomes, where immediate IPT was associated with more adverse pregnancy outcomes,” she said,

“I think we now need to reweigh the evidence for pros and cons for IPT in pregnancy,” Dr. Gupta concluded.

The study was sponsored the National Institutes of Health. Dr. Gupta reported having nothing to disclose.

SOURCE: Gupta A et al. CROI 2018, Abstract Number 142LB .