Introduction
Acquisition of data quantifying the exposure of second-line TB drugs in pregnant woman treated for rifampicin-resistant TB (RR-TB) is a priority. Until recently, pregnant and breastfeeding women have typically been excluded from clinical trials of new drugs, including TB treatment. (1) The World Health Organization (WHO) currently recommends individualised treatment regimens with drugs with a preferred safety profile for pregnant women with RR-TB, (2) although there are limited human data guiding these recommendations.
Bedaquiline is a group A drug, recommended for inclusion in all RR-TB treatment regimens, and can be used in pregnant women, although safety data is lacking. The design of pharmacokinetic studies to explore the effect of pregnancy on long half-life drugs like bedaquiline is challenging, as cumulative drug concentrations could mask pregnancy-related effects on drug exposure. Physiological changes in pregnancy result in decreased concentrations of many drugs, particularly in the third trimester. Pharmacokinetics in pregnancy may be complex: one of the reasons for reduced drug concentrations in pregnancy is the reduction in plasma concentrations of the two key drug-binding proteins: albumin and α1-acid glycoprotein. (3) Reduction in these binding protein concentrations reduces the total (bound and unbound) concentrations of drugs, but the unbound fraction typically increases, resulting in unbound drug concentrations that are similar to non-pregnant women. As only the unbound drug is pharmacologically active, recommendations to increase the dose of drugs in pregnancy to approximate total concentrations in non-pregnant women could therefore increase the risk of toxicity. However, pregnancy also increases several drug clearance mechanisms, which could reduce unbound drug concentrations. Although measurement of unbound bedaquiline concentrations is preferable to rationally optimize dosing in pregnancy (bedaquiline is >99% protein-bound) (4), an understanding of the effect of pregnancy on total bedaquiline concentrations would provide a much-needed foundation to understand the effect of pregnancy on the unbound fraction.
Data on the secretion of key drugs for RR-TB into human milk are scarce. The studies describing RR-TB drug exposure in human milk are small with little or no infant plasma pharmacokinetic data available - the study designs are also unclear or unstated. Linezolid (5), levofloxacin (6) and cycloserine (7) penetrate poorly into human milk and exposure to breastfed infants is therefore likely to be low. Clofazimine, in contrast, demonstrates effective human milk penetration with skin discoloration observed in the infants of breastfeeding mothers treated with clofazimine for leprosy;.(8,9) clofazimine exposure in human milk in the context of mothers treated for TB is unfortunately lacking. Animal studies have shown that bedaquiline is concentrated in rat milk with concentrations 6- to 12- fold higher than the concentration in maternal plasma (10), but there are currently no human data available. Information on clinically relevant infant exposure to RR-TB drugs through breastfeeding with mother-infant pairs has not been done, and is an important knowledge gap.
An international consensus panel on the inclusion of pregnant and postpartum women in TB drug trials, convened by the NIH, identified the safety, tolerability and pharmacokinetics of novel agents and regimens for treatment of RR-TB as research priorities (11). It is ethically imperative to study drug dosing and safety in populations where drugs are used - this has not been done satisfactorily for RR-TB. (12) We therefore conducted an observational study of bedaquiline exposure in pregnant and breastfeeding women with RR-TB.