https://orcid.org/0000-0002-8215-1683
TO THE EDITOR:
The risk of venous thromboembolism (VTE) is significantly increased during the postpartum period and remains the leading direct cause of maternal mortality in many Western countries.1 Low molecular weight heparins and vitamin K antagonists are the 2 options currently available for the prophylaxis and treatment of VTE during the postpartum period, because they are known to be safe during lactation.1 However, the parenteral nature of low molecular weight heparin and frequent monitoring required with vitamin K antagonists act as barriers for many women.1-3 In recent years, direct oral anticoagulants are increasingly being prescribed because of the advantages they confer: oral route of administration, a predictable dose-response curve, and no requirement for laboratory monitoring. Therefore, if they could be prescribed to breastfeeding women during the puerperium, they would be a significant improvement on what is currently offered to women.
Apixaban and rivaroxaban are 2 commonly prescribed direct oral anticoagulants. Apixaban has been shown to be extensively secreted into rat milk, with a reported milk:plasma ratio (M:P) of 30,4 but there are no studies in the literature on the extent of transfer into human breast milk. Case reports have shown a relatively small distribution of rivaroxaban into human breast milk, but these case reports lack a plasma concentration-time profile across a 24-hour period along with the respective breast milk concentration.5-7 Currently, apixaban and rivaroxaban are contraindicated in lactating women because of limited data on the extent of their transfer into human breast milk.8 To our knowledge, this is the first clinical trial investigating the distribution of apixaban and rivaroxaban into human breast milk.
We conducted a single-center, open-label trial (EudraCT 2018-003852-19). Breastfeeding women age 18 years or older who were willing to stop breastfeeding during the study period were eligible. Women who had a known contraindication to apixaban or rivaroxaban; treatment with antifungal drugs, investigational study drug(s), or drugs that affect hemostasis; an increased risk of bleeding; or abnormal renal or hepatic function were excluded. As a proof-of-concept trial, the study was designed to recruit 2 volunteers to receive apixaban and 2 volunteers to receive rivaroxaban. All participants provided informed consent, and the trial was approved by the National Research Ethics Committee: London-Fulham (ref: 19/LO/0082).
Home visits were provided to volunteers for breast milk and plasma sample collection to maximize recruitment.9 Participants were allocated to rivaroxaban and apixaban sequentially, in the order of their recruitment; those allocated to rivaroxaban were administered a single oral dose of 20 mg, and those allocated to apixaban were given 2 doses of 5 mg each 12 hours apart. Blood and breast milk samples were collected at t = 0 (before the first dose), 2.5, 6, 10, 12, and 24 hours after the first dose for women given rivaroxaban and at t = 0 (before the first dose), 3.5, 7, 12 hours before the second dose, 16, and 24 hours after the first dose, for women allocated to apixaban. A validated ultra-high-performance liquid chromatography/tandem mass spectrometry method was used to measure apixaban and rivaroxaban concentrations in plasma and breast milk.10 The primary data from the trial were captured in Research Electronic Data Capture (REDCap),11 which were then extracted and analyzed. The area under the concentration-time curve (AUC0-24 hours) was calculated by means of noncompartmental pharmacokinetic analysis using the lin-up/log-down trapezoidal rule in R software (version 3.6.2). Results were reported on the basis of US Food and Drug Administration Guidelines.12
Four women were screened and consented to participate in the trial after seeing an advertisement in February 2020, but 1 volunteer withdrew before sampling. Two women received rivaroxaban (volunteer 1: a 27-year-old woman, 8 months postpartum; volunteer 2: a 42-year-old woman, 8 months postpartum), and 1 woman received apixaban (participant 3: a 44-year-old woman, 23 months postpartum). Both apixaban and rivaroxaban were detectable in breast milk samples after they had been ingested, and the observed concentration-time profiles followed that assayed in plasma (Figure 1). Table 1 provides the calculated pharmacokinetic parameters and breast milk exposure estimates. For apixaban, the M:P ratio was 2.61, and the estimated relative infant dose (RID) was 12.78%. For rivaroxaban, the M:P ratio was 0.27 (± 0.02) and the RID was 1.63% (± 0.01%).
Concentration‐time profiles. (A) Rivaroxaban (n = 2) and (B) apixaban (n = 1) in maternal plasma and breast milk. Data are mean ± standard deviation for 2 volunteers administered rivaroxaban.
Concentration‐time profiles. (A) Rivaroxaban (n = 2) and (B) apixaban (n = 1) in maternal plasma and breast milk. Data are mean ± standard deviation for 2 volunteers administered rivaroxaban.
Calculated apixaban and rivaroxaban exposure parameters
| Parameter | Rivaroxaban 20 mg once per day (± SD) (n = 2) | Apixaban 5 mg twice per day (n = 1) |
|---|---|---|
| AUC0-24 hours, milk, ng × h/mL | 693.77 (23.84) | 2725.33 |
| Cavg0-24 hours, milk, ng/mL | 28.91 (0.99) | 113.56 |
| AUC0-24 hours, maternal plasma, ng × h/mL | 2569.55 (312.73) | 1045.82 |
| Cavg0-24 hours, maternal plasma, ng/mL | 107.06 (13.03) | 43.58 |
| M:P ratio (Cavg milk/Cavg plasma) | 0.27 (0.02) | 2.61 |
| Absolute infant dose, mg/kg per day* | 0.0043 | 0.017 |
| RID (%)† | 1.63 (0.01) | 12.78 |
| Parameter | Rivaroxaban 20 mg once per day (± SD) (n = 2) | Apixaban 5 mg twice per day (n = 1) |
|---|---|---|
| AUC0-24 hours, milk, ng × h/mL | 693.77 (23.84) | 2725.33 |
| Cavg0-24 hours, milk, ng/mL | 28.91 (0.99) | 113.56 |
| AUC0-24 hours, maternal plasma, ng × h/mL | 2569.55 (312.73) | 1045.82 |
| Cavg0-24 hours, maternal plasma, ng/mL | 107.06 (13.03) | 43.58 |
| M:P ratio (Cavg milk/Cavg plasma) | 0.27 (0.02) | 2.61 |
| Absolute infant dose, mg/kg per day* | 0.0043 | 0.017 |
| RID (%)† | 1.63 (0.01) | 12.78 |
AUC, area under concentration-time curve; Cavg, average concentration (AUC0-24 hours/24 hours); RID, relevant infant dose; SD, standard deviation.
Absolute infant dose was calculated as the product from Cavg in milk and an assumed milk intake of 0.15 L/kg per day.
Relative infant dose is the absolute infant dose expressed as a percentage of the weight-adjusted maternal dose, based on an average maternal weight of 75 kg (mg/kg per day).
This study is the first to report the concentration-time profiles for apixaban and rivaroxaban in human plasma and breast milk over a 24-hour period after a single dose. The results suggest that the distribution of apixaban into human breast milk is significantly more compared with that of rivaroxaban. Apixaban and rivaroxaban have similar physiochemical properties. They both exhibit a high degree of protein binding (87% for apixaban and 92% to 95% for rivaroxaban), moderate lipophilicity (logP = 2.23 for apixaban and 1.74 for rivaroxaban), rapid absorption (Tmax = 3-4 hours for apixaban and 2-4 hours for rivaroxaban), and no ionization at physiological pH.13 Given these properties, one might predict that neither of these agents significantly distribute to breast milk.14
This study found that exposure (AUC0-24h) to apixaban was reduced in the plasma of a lactating volunteer compared with that in the plasma of a nonlactating healthy participant,15 because of the extensive distribution of apixaban into the breast milk in this volunteer. This high breast milk distribution might be explained by apixaban being a substrate for the breast cancer resistance protein (BCRP; ABCG2) and being actively secreted into breast milk from the plasma.16 ABCG2 has been shown to be strongly induced in the mammary glands of humans during lactation and is responsible for the active secretion of many drugs into breast milk.17 Although the apixaban results are from only 1 volunteer, the active secretion hypothesis would explain the high M:P ratio reported in animal studies by the manufacturers of apixaban. Further study is clearly required to confirm or refute this.
Despite rivaroxaban also being a substrate of ABCG2, it does not seem to distribute to breast milk extensively. This could be because rivaroxaban is a substrate for both P-glycoprotein and ABCG2,18 and this may have an impact on its level of active secretion into milk, if this is indeed the implicating mechanism.
In our study, the rivaroxaban breast milk concentrations from the 2 volunteers are consistent with those of the few cases published in the literature.5,6 However, 1 case report has calculated a rivaroxaban RID between 4% to 5%, which might be explained by samples being collected at steady state.6 The authors of that report suggest that exposure of rivaroxaban via breast milk, even at 4% to 5%, is not considered clinically relevant (RID <10%) for a breastfed infant.19 The recent EINSTEIN Junior trial determined weight-adjusted rivaroxaban doses for the treatment of acute VTE in pediatrics. The absolute infant dose calculated here (0.0043 mg/kg) is at least 150 times lower than that prescribed for infants in the EINSTEIN Junior trial.20
An important limitation of this study is the small sample size, particularly because only 1 woman received apixaban, and the women who were included in the study were all at the late stages of lactation (>6 months). This study was thus unable to establish interindividual variability and variability of drug secretion during different stages of breastfeeding or to determine a steady-state pharmacokinetic profile in plasma and breast milk, although that was not the aim of this study. Despite the RID of rivaroxaban in our study being similar to that in a previous case in foremilk, the M:P ratio reported here is lower.6 The composition of foremilk differs from that of hindmilk, which could affect passive distribution of drug secretion in breast milk.11 Further studies with women who have recently delivered and been given repeated drug doses are now needed.
This clinical trial found a significant distribution of apixaban into human breast milk, whereas rivaroxaban distribution was significantly less. The results suggest that rivaroxaban may hold promise for breastfeeding women, and further research from women in the early period after delivery is needed, to confirm this.
Data sets and protocols of this study are available by request from Jignesh P. Patel at jig.patel@kcl.ac.uk.
Acknowledgments
The authors thank the volunteer women who took part in the trial and the Health and Social Care (HSCR; Unit of Medical Statistics) at King's College London who provided hosting services for REDCap.
Y.Z. completed this work as part of a personal studentship awarded by the China Scholarship Council.
Authorship
Contribution: Y.Z. contributed to the study design, recruited participants, collected and analyzed the data, and prepared the manuscript; J.P.P. contributed to study design, recruited participants, collected the data, and reviewed and edited the manuscript; R.A. contributed to study design, recruited participants, reviewed and edited the manuscript, and funded the clinical trial; and L.C. contributed to the analysis of the breast milk and plasma samples and analyzed, reviewed, and edited the manuscript.
Conflict-of-interest disclosure: J.P.P. has received an investigator-initiated research grant from Bayer. R.A. received grants from Bayer, personal fees from Bayer, Pfizer, Medtronic, and Sanofi, and nonfinancial support from Bayer, Pfizer, and Sanofi. The remaining authors declare no competing financial interests.
Correspondence: Jignesh P. Patel, Institute of Pharmaceutical Science, King's College London , 5th Floor Franklin-Wilkins Building, 150 Stamford St, London SE1 9NH, United Kingdom; e-mail: jig.patel@kcl.ac.uk.
