A population pharmacokinetic model for escitalopram and its major metabolite in depressive patients during the perinatal period: Prediction of infant drug exposure through breast milk

Abstract

Background and objectives: Escitalopram (SCIT) is frequently prescribed to breastfeeding women. Available information on SCIT excretion into breast milk is based on heterogeneous and incomplete data. A population pharmacokinetic model that aimed to better characterize maternal and infant exposure to SCIT and its metabolite was developed.

Methods: The study population was composed of women treated by SCIT or racemic citalopram and enrolled in the multicenter prospective cohort study SSRI-Breast Milk study (ClinicalTrial.gov NCT01796132). A joint structural model was first built for SCIT and S-desmethylcitalopram (SDCIT) in plasma using NONMEM and the milk-to-plasma ratio (MPR) was estimated by adding the drug breast milk concentrations. The effect of different influential covariates was tested and the average drug exposure with variability through breastfeeding was predicted under various conditions by simulation.

Results: The study enrolled 33 patients treated with SCIT or racemic citalopram who provided 80 blood and 104 milk samples. Mean MPR for both parent drug and metabolite was 1.9. Increased milk fat content was significantly associated with an increased drug transfer into breast milk (+28% for SCIT and +18% for SDCIT when fat amount doubles from 3.1 to 6.2 g/100 mL). Simulations suggested that an exclusively breastfed infant would ingest daily through breast milk 3.3% of the weight-adjusted maternal SCIT dose on average.

Conclusion: The moderate between-subject variability in milk concentration of SCIT and the limited exposure to escitalopram through breast milk observed provide reassurance for treated mothers of breastfed healthy infants.

Keywords: breastfeeding; escitalopram; exposure; population pharmacokinetics.

Conflict of interest statement

There are no competing interests to declare.

© 2020 The British Pharmacological Society.

Figures

FIGURE 1

Compartmental structure of the final model and its main pharmacokinetic parameters. VSCIT, volume of distribution of escitalopram; VMR, volume metabolite ratio; kxy, absorption/metabolic/elimination rate constant; MPRD, milk‐to‐plasma ratio of escitalopram; MPRM, milk‐to‐plasma ratio of S‐desmethylcitalopram

FIGURE 2

Goodness‐of‐fit plots of escitalopram with (A) observed concentrations (DV) vs population predictions (PRED), (B) DV vs individual predictions (IPRED), (C) conditional weighted residuals (CWRES) vs PRED and (D) CWRES vs time after dose (TAD) (solid line: unity line)

FIGURE 3

Goodness‐of‐fit plots of S‐desmethylcitalopram with (A) observed concentrations (DV) vs population predictions (PRED), (B) DV vs individual predictions (IPRED), (C) conditional weighted residuals (CWRES) vs PRED and (D) CWRES vs time after dose (TAD) (solid line: unity line)

FIGURE 4

Prediction‐corrected visual predictive checks of the final covariate model for (A) escitalopram in plasma, (B) S‐desmethylcitalopram in plasma, (C) escitalopram in breast milk and (D) S‐desmethylcitalopram in breast milk. Circles, prediction‐corrected drug concentrations; continuous line, population median prediction; dashed lines, 2.5th and 97.5th percentiles; semi‐transparent grey fields, model‐based percentile confidence intervals

FIGURE 5

Relative infant dose (RID) and adult dose equivalent (ADEQ) of escitalopram in 5000 or 10 000 simulated mother–infant pairs under various conditions (median ± range)