Background
Worldwide, ~50% of women take medication during breastfeeding. Data surrounding the
exposure of the breastfed infant to drugs and any associated risks are sparse. Despite
longstanding recommendations from the US Food and Drug Administration (FDA) for lactation
studies to be performed close to licensing for drugs anticipated to be widely used in
women of childbearing age, such studies are rarely undertaken. Drugs taken by the breast
feeding mother on TB treatment can be passed from the maternal circulation to the milk
and then to the breastfed infant, a concern of effects of anti-tuberculosis drugs on
nursing infants. Most TB drugs are metabolized by the liver, triggering a potential risk
of drug accumulation in infants due to their immature liver function particularly in
premature infants.
Drugs are transferred to milk in small quantities, and many have been used without
obvious infant toxicity for many years hence the large gaps in the data. Pharmacokinetic
(PK) information of anti- TB drugs transfer to breast milk and breastfed infant is
crucial to limit the development of drug resistance and understand the safety of
prolonged exposure through breast milk.
Problem Statement
Whilst data on TB drug penetration into breastmilk is limited, information on clinically
relevant infant exposure to TB drug-sensitive is even more limited and is an important
knowledge gap both for safety, and because therapeutic concentrations could be 1)
protective in exposed infants, obviating the need for TB preventive therapy or 2) sub-
therapeutic concentrations could select for resistance in those infants infected with
Mycobacterium tuberculosis.
Recruitment
Women will be identified as they attend the clinic for TB treatment at the IDI and KCCA
clinics in Uganda. Should a woman express willingness to participate, once eligibility
for enrolment in the study has been determined, informed consent will be obtained.
Pharmacokinetic Study Day
On arrival, an intravenous cannula will be inserted into the antecubital fossa, and
samples taken for trough drug measurement. After a standardized breakfast the participant
will be administered standard doses of the prescribed medication. Blood samples will be
collected at 2, 4, 6, 8 and ideally* 24 hours. Advice to freely breastfeed the baby will
be given. The participant will be asked to provide a 2-5 ml sample of expressed breast
milk pre-dose, and at 2, 4, 6 and 8 hours post dosing. A blood sample from the infant
will be collected at maternal trough (pre-dose) and at a 3-8 hours post maternal dose
(the second time point will be allocated sequentially to ensure spread of datapoints).
The mother will be administered a standard lunch.
*Due to the logistic considerations of sampling a postpartum mother and her infant who
may have travelled a long distance to the clinic, the 24-hour sample may not be collected
in all cases.
Maternal albumin and creatinine will be sampled as they are important for isoniazid
exposure. Maternal questionnaires will be filled on each visit to assess depression and
anxiety; Generalised anxiety disorder questionnaire (GAD-7), Patient health questionnaire
(PHQ-9), and the Beliefs about medicines questionnaire (BMQ). Infant clinical assessment
will include use of the Gross Motor Development (GMD) checklist
Sample Size Determination
This study is exploratory, as no prior study has characterized the exposure of these
drugs in maternal plasma, breastmilk and infant plasma. There are no prior data upon
which to build a sample size calculation, and there is no comparison between groups which
requires statistical analysis with a pre-specified certainty.
Since no information is available about the penetration of these drugs into breastmilk,
the following approach was used, described in detail for rifampicin.
A previously published pop-PK model of rifampicin in plasma was modified, adding a
compartment to describe breastmilk concentrations. This was characterised using an
approach similar to an effect compartment described by a time delay and an accumulation
ratio between breastmilk and plasma. The half-life of the delay was fixed to 1 h and the
accumulation ratio to 1.5, with 30% between-subject variability in both parameters. These
were chosen to mimic a PK profile similar to Waitt et al. It was assumed that there would
be 15 individuals (considering a mother-infant dyad as a single unit) with an intensive
PK sampling at 0, 1, 2, 4, 6 and 8 hours post-dose of paired plasma and breastmilk (30%
error in the breastmilk measurements was assumed) and Stochastic Simulations and
Estimations (SSEs) were performed to evaluate trial design. This design can characterise
all the typical values of the plasma PK parameters with precision of better than 11% RSE,
and all the breastmilk parameters are well characterised with a precision of 1.14% and
0.591% RSE on delay and accumulation ratio, respectively.
Interim analysis after five participants is part of study design
Analysis of Endpoints
Pharmacokinetic data will be analysed using a population pharmacokinetic approach to
estimate pharmacokinetic parameters and produce modelled fits to exposure data.
Inter-individual variability will be quantified in relation to the covariates.
Non-compartmental methods will be used to assess correlations between maternal breast
milk drug concentrations and measures of drug exposure in the infant (eg AUC) and
pharmacodynamic factors.