When administered in combination with acetaminophen or ibuprofen, hydrocodone is an effective postpartum and postoperative oral opioid analgesic.1,2 The use of hydrocodone in nursing mothers is common. However, data describing the transmission of hydrocodone to breast milk are limited. Cases of neonatal sedation, possibly attributable to higher-than-average maternal hydrocodone dosages during breastfeeding, also have been described.3,4 The short-term use of hydrocodone-containing analgesics during breastfeeding is thus cautiously sanctioned.5
Hydrocodone is biotransformed by the hepatic CYP2D6 enzyme to the more potent opiate, hydromorphone. Both the parent drug and the metabolite contribute to the efficacy and toxicity of hydrocodone.6,7 A relative infant dosage of hydrocodone of approximately 3% to 4% was found based on milk hydrocodone concentrations in two postpartum women using hydrocodone-acetaminophen for analgesia.8 Although relative infant dosage values less than 10% generally indicate a medication is safe for use during breastfeeding,9 hydromorphone milk levels were not measured in either case and, thus, total opiate exposure could not be assessed. To better-characterize the quantities of hydrocodone and its active metabolite hydromorphone in breast milk, we conducted a study in a larger group of breastfeeding mothers using hydrocodone-acetaminophen for postpartum pain.
MATERIALS AND METHODS
The study was performed at the University of California, San Diego, Medical Center from November 2006 to October 2009 and was approved by the University of California, San Diego, Human Research Protections Program. All study patients gave their written consent to participate.
Eligible patients were women 18–50 years of age who were at least 48 hours postpartum, breastfeeding their newborns, and receiving an oral hydrocodone-acetaminophen combination product for pain. Patients targeted for recruitment were those admitted to the postpartum nursing unit, lactating mothers admitted to nonobstetric care areas of the medical center, and mothers of newborns in the neonatal intensive care unit. A daily report of inpatients with hydrocodone orders from the inpatient pharmacy database was created for the study and used to identify potential study patients. Potential patients also were referred to the investigators by other clinicians. Recruitment and consenting of identified eligible patients were performed by investigators (J.S., S.R., N.N., L.S.) when possible during their daily practice. Use of the patient's lactation consultant or her newborn's pediatric providers was a required element of the investigators' patient recruitment procedure. Specifically, the lactation consultant or pediatricians were required to grant the investigator their approval of eligible study patients before consent and enrollment. Approval was granted if, in the opinion of the patient's lactation consultant or pediatrician, lactogenesis stage II had occurred and there was sufficient breast milk production to permit donation of milk for the study. Patients were not approached for consent if milk supply was deemed to be insufficient. Patients also were not approached if the primary obstetric provider deemed the patient too ill to comply with the study protocol. Because of funding limitations, the target study size was 30–35 women.
The targeted study period duration for each patient was from 24 to 96 hours or until hospital discharge, whichever was sooner. Patients were not removed from analysis if their actual study period was outside this target duration. During each study period, enrolled patients either emptied both breasts using a hospital breast pump (Symphony or Lactina, Medela) and then provided a 2- to 5-mL aliquot of milk from each pumping for the study or manually expressed and provided 2 to 5 mL of milk from each breast immediately before breastfeeding. The selection of method was made to match the method the patient was using at the time to provide milk for her newborn. Patients were asked to provide a milk sample each time they pumped or expressed during the study period. Milk samples were collected in 120-mL sterile collection cups. The date and time of each collection were noted on the cup by the patient and were confirmed by an investigator (J.S., S.R., or N.N.) during the time that milk was being collected. Collected samples were stored at −80°C until assayed. Each patient's hydrocodone bitartrate doses and times of administration were captured from the electronic medical record and recorded at the end of each milk collection period.
Aliquots of breast milk pumped by the mothers were analyzed for hydrocodone and hydromorphone using isotope-dilution liquid chromatography mass spectrometry.10 The deuterated internal standards, d3-hydrocodone and d3-hydromorphone, were added to milk samples before protein removal with acetonitrile, which preceded analysis. The lower limit of quantitation for both analytes was 1 ng/mL. Accuracy and precision of the assay across its dynamic range (1–1,000 ng/mL) were more than 90% and less than 8%, respectively. Any concomitant medications in the patient's milk, even other opioid medications, would not be expected to interfere with the assay's accuracy or precision.
Descriptive, inferential, graphic analyses and normality tests were performed using SAS 9.2. Normality of the outcome measures was performed using Proc Univariate and the Shapiro-Wilk test. Descriptive statistics of hydrocodone and hydromorphone outcome measure means, standard deviations, medians, first quartiles, third quartiles, and their respective 95% confidence intervals (CIs) (both normality-based and distributional-free 95% CIs) were found using the Proc Univariate procedure. Comparisons of hydrocodone and hydromorphone outcome measures for breast-pumped milk compared with breastfeeding and term compared with preterm were performed using the Proc Npar1Way procedure and the Wilcoxon rank-sum test. The significance level was set at .05. Individual plots of hydrocodone and hydromorphone breast milk concentrations and dose compared with time were performed using the Proc Gplot procedure.
The study period of each patient was between the time of the first dose of hydrocodone bitartrate (time 0) and the time of the last milk opiate level analyzed (time t). To calculate the total dosage of hydrocodone bitartrate received by each of the neonates, the area under the milk concentration–time curve (AUC0-t) was estimated for each woman using a linear rectangular method described previously.11 The linear rectangular method of AUC estimation was used rather than the linear/log-linear trapezoidal methods because of the sparse sampling, the inconsistent increase and decrease in the breast milk concentrations, and the wide and irregular intervals between dosing and breast milk collection.12
The average milk concentration over the course of the study period for each woman then was calculated by dividing the AUC by the number of hours in the study period. This value then was multiplied by the accepted average neonatal milk intake value of 150 mL/kg/day to estimate the neonatal dosage in micrograms per kilogram per day.
The total dosage that each mother ingested was divided by the total number of hours that the mother was studied and multiplied by 24 to calculate an average daily dosage. This dosage then was divided by maternal weight to calculate the average maternal hydrocodone bitartrate dosage in micrograms per kilograms per day. Because hydrocodone in commercial drug products is the bitartrate hemipentahydrate (2.5 waters of hydration), the dosage that the mother received was multiplied by the factor of 0.605 to determine the dosage of anhydrous hydrocodone base that she ingested.
The neonate's estimated dosage of hydrocodone in micrograms per kilograms per day was divided by the average maternal dosage in micrograms per kilograms per day. This value was multiplied by 100 to calculate the weight-adjusted percentage of maternal dosage (relative infant dosage) that the neonate received.
Finally, an estimate of total opiate exposure was made. These values were calculated by taking of the percentage of the minimum infant dosages for commercially available hydrocodone biartrate (600 micrograms/kg/day) and hydromorphone hydrochloride (240 micrograms/kg/day) multiplied by the fraction of active opiate in each (Lortab syrup package insert. UCB Pharma. January 2008).13 These values are 363 micrograms/kg/day for hydrocodone and 213 micrograms/kg/day (240 micrograms/kg/day×0.887) for hydromorphone. The resulting value was multiplied by 100 to calculate the percentage. The two resulting values were then added to obtain an estimate of total opiate dosage.
Thirty-two eligible women gave their consent to participate. Two enrolled patients asked to be withdrawn from study participation during the study, one because of physical exhaustion and one because of reduced milk supply. The demographics of the remaining 30 patients are summarized in Table 1. Seven (19%) of these women had preterm births at less than 36 0/7 weeks of gestation, six of the births were singletons and one was twins. Three of the eight preterm neonates had very low birth weight (less than 1,500 g). Twenty-eight (93%) of the 30 study patients had cesarean deliveries. The study period averaged 55.3 hours (range 16.9–84.1 hours).
Breast milk hydrocodone data from 125 milk samples provided by the study patients are presented in Table 2. All variables were tested for normality using the Shapiro-Wilk test. Because the data were not normally distributed, the median and ranges of values are reported in addition to means and standard deviations. The maximum measured milk hydrocodone base concentration was 715 micrograms/L. This was a single outlying value sampled 3 hours after a 6,050-microgram dose of hydrocodone base. This concentration was inconsistent with those measured in five other milk samples provided by the same woman at different times during her study period and was 54-times greater than her next highest value. Assay of this sample was repeated and confirmed the initial measurement. Considering the extreme incongruity of this outlying value, it was not used to calculate the AUC or other results for this patient. No other values in any of the women were omitted. Among all other study participants, the maximum measured milk hydrocodone base concentration at any time point was 126.5 micrograms/L sampled 3 hours after a 6,050-microgram anhydrous hydrocodone dose. The median milk hydrocodone base concentration in the entire study population was 14.2 micrograms/L (95% CI 7.7–30.2). The calculated median neonatal dosage equals 2.1 micrograms/kg/day (95% CI 1.2–4.5), which is 1.6% (95% CI 0.7–2.4) of the weight-adjusted maternal daily dosage.
Breast milk hydromorphone data are presented in Table 3. Only 12 of the 30 women had hydromorphone detectable in their milk. The majority of collected samples (107/125) had hydromorphone levels below the assay limit of quantification. Four of these 107 samples were detectable (range 0.2–0.8 micrograms/L). In the 12 women in the study with detectable hydromorphone levels, the median milk hydromorphone concentration was 1.9 micrograms/L (95% CI 1.7–16.5). The calculated median neonatal dosage of hydromorphone from the breast milk of these women is 0.3 micrograms/kg/day (95% CI 0.3–2.5).
Table 4 shows the percentage of the minimum neonatal dosages of hydrocodone, hydromorphone, and both together (total) that newborns would receive if they were breastfed exclusively. All hydrocodone and hydromorphone results were included in the estimate of total neonatal dosage. Nondetectable values were considered equal to zero.
The combination oral analgesic hydrocodone-acetaminophen has remained the most frequently prescribed generic medication in the United States for several years and is often used as an alternative to codeine in postpartum mothers. Despite its popularity, published experience with hydrocodone during breastfeeding is limited. No clinical trials are available, but two reports of toxicity have been published. In one case, a mother had used unspecified dosages of methadone and hydrocodone 5 mg/acetaminophen 500 mg before nursing her infant. The 5-week-old infant was found cyanotic with little respiratory effort. Intubation was required and administration of naloxone improved the infant's respiratory effort. The relative contributions of methadone and hydrocodone to the infant's respiratory depression could not be determined.4 In the other case, a nursing mother with a Candida mastitis was using hydrocodone bitartrate 20 mg and 1,300 mg acetaminophen in combination every 4 hours. Both she and her 18-day-old neonate were “groggy and sleepy most of the day” until she reduced her dosage to hydrocodone bitartrate 10 mg and 650 mg acetaminophen every 3 hours.3 Neither of these cases represent typical hydrocodone dosages used postpartum, but the latter case demonstrates that large maternal dosages can adversely affect an otherwise healthy breastfed newborn.
We previously estimated a hydrocodone relative infant dosage of 3%–4% of hydrocodone based on milk concentration data in two postpartum women using a hydrocodone-acetaminophen combination product.8 The hydrocodone relative infant dosages in the current study with a larger patient sample are similar, averaging 2.4%, a median of 1.6%, and a range of 0.2% to 9%. Such relative infant dosage values are generally considered safe.9 The highest individual neonatal hydrocodone exposure documented in the current study was 14.9 micrograms/kg/day. This equals an infant hydrocodone bitartrate dosage of 25 micrograms/kg/day, well below the reported minimally effective pediatric dosage of 600 micrograms/kg/day. The combined effect of the hydromorphone metabolite appears to be minimal in most cases. The majority of patients in this study had no detectable hydromorphone levels in their milk. Of those who did, the median milk concentrations and neonatal exposures would not be expected to have clinical effects (Tables 3 and 4).
To apply the results of our study at their extreme margins, if a 70-kg exclusively breastfeeding mother receives 40 mg per day of hydrocodone bitartrate, equivalent to eight tablets per day of regular strength Vicodin, the highest neonatal exposure, based on the maximum relative infant dosage of 9% documented in our study, would be approximately 50 micrograms/kg/day of hydrocodone bitartrate. If the hydromorphone metabolite is present in the patient's milk at half the hydrocodone concentration (the mean concentration value in Table 3 relative to Table 2), then the neonate would consume 25 micrograms/kg/day of hydromorphone in addition to the 50 micrograms/kg/day of hydrocodone. Because hydromorphone is 2.5-times as potent as hydrocodone, the cumulative neonatal exposure of “hydrocodone equivalents” in this theoretical scenario would be 112 micrograms/kg/day. Such a daily neonatal dosage could possibly have a pharmacologic effect in neonates, given that hydrocodone is considered 60%–100% as potent as oral morphine,14 and standard neonatal oral morphine dosages are 200–500 micrograms/kg/day.15 This illustrates how the combination of high maternal hydrocodone dosages and high breast milk excretion of both hydrocodone and hydromorphone can potentially expose fully breastfed neonates to worrisome quantities of hydrocodone and hydromorphone in breast milk.
The highest combined neonatal opiate exposure documented in the milk of the women in our study was equal to only 9.9% of the minimally effective dosage in older infants (Table 4). Although this may sound reassuring, neonates eliminate drugs more slowly than older infants and are much more sensitive to the effects of opiates than older infants.16 Our use of the minimally effective pediatric–infant dosages recommended to determine the percent therapeutic dosage in breastfed newborns thus may overstate the safety of hydrocodone use in mothers breastfeeding neonates. Mother–newborn dyads were not followed for evidence of hypersomnolence, poor feeding, or other potential side effects of opiates during breastfeeding. Although the measured breast milk opiate levels and pharmacokinetic analysis indicate that short-term postpartum hydrocodone use should be safe, validation of this prediction in a prospective clinical trial is needed.
Pharmacogenetic analysis was not performed in this study. It is intriguing to consider the ratio of hydromorphone to hydrocodone in breast milk as a possible surrogate marker of CYP2D6 activity, assuming this metabolic ratio and the activity are positively correlated. The ratio is calculated by dividing the estimated infant dosage of hydromorphone by the estimated infant dosage of hydrocodone. Of the 12 women with detectable hydromorphone, 10 women had hydromorphone-to-hydrocodone ratios between 0 and 1 and two (7% of all participants) had ratios more than 1 (2.8 and 3.1). The value of 7% is in the range of reported frequencies of CYP2D6 ultrarapid metabolizers in various populations.17 It is possible that hydromorphone added substantially to the total opiate activity in the breast milk of these two women because they were ultrarapid metabolizers. Although neither of their newborns received a dosage near the therapeutic infant dosage, the hydrocodone bitartrate dosages that their mothers received were relatively low during the study period (data not shown). High maternal hydrocodone bitartrate dosages thus may result in relatively large quantities of hydromorphone in the breast milk of a small percentage of women.
Since the death of a breastfed infant was attributed to maternal codeine intake,18 there has been much discussion in the literature of the safety of codeine during breastfeeding.19–21 Analysis by the original authors suggested that the mother's CYP2D6× 2×2 ultrarapid genotype and perhaps uridyl glucuronosyltransferase 2B7 genotype caused this and other reported adverse reactions.17,18 Other investigators using computer modeling techniques suggested that all breastfed neonates are susceptible to the adverse effects of codeine and its morphine metabolite with as little as 4 days of continuous maternal codeine use at dosages more than 1 mg/kg/day, even in mothers who are not ultrarapid metabolizers.22 Because of these findings, numerous warnings and recommendations have been published regarding the use of codeine in nursing mothers.19,23,24
Like codeine, hydrocodone is metabolized by CYP2D6 to a more potent active metabolite, hydromorphone, and to glucuronide metabolites.25 High maternal hydrocodone dosages for a prolonged period of time could, like codeine, be dangerous in a small percentage of breastfed neonates. Because a mother's CYP2D6 genotype is usually unknown to the clinician, a safe plan is to avoid prolonged high dosages.
Nineteen women donated study milk by pumping and emptying their breasts, and 11 donated manually expressed milk. Allowing differences in milk-collection methodology was necessary to permit our study participants to continue transferring milk to their newborns in harmony with their clinical-care and breastfeeding plan. Our intent was to not allow study interventions to interfere with the maintenance of lactation and breastfeeding in study participants. Both hydrocodone and hydromorphone are lipid-soluble and, hence, may preferably distribute to the more fat-rich hind milk.26 It is reasonable to assume that milk sampled by manual expression before a feeding will have lower levels of hydrocodone and hydromorphone than if the same milk were sampled from a pumped and emptied breast. A weakness of this study, therefore, is that the manually expressed samples from the 11 women could be contributing falsely low hydrocodone and hydromorphone milk concentration values to the final results, thus underestimating milk transmission and overestimating safety. To examine this further, we performed a comparison of the hydrocodone concentrations from all the samples collected as aliquots from pumped milk to those from expressed milk and found no significant difference between these two groups when compared using the Wilcoxon rank-sum test. The median hydrocodone concentration was 12.3 micrograms/L (95% CI 7.2–28.3) in the samples collected from pumped milk and 30.2 micrograms/L (95% CI 6.1–56.1) in those collected from expressed milk (P=.53). To control for the confounders of weight and dosage, we also compared the hydrocodone dosage in micrograms/kg/day between the two groups but again found no significant difference. The median dosage in the patients who pumped their milk was 141 micrograms/kg/day (95% CI 102–231) and 175 micrograms/kg/day (95% CI 120–279) in those who manually expressed milk (P=.48). All other values when similarly compared also were not significantly different (data not shown).
As a general rule, nonopioid analgesics should be used first when treating postpartum pain. Opioid analgesics then are added if pain control is inadequate, particularly after cesarean delivery or repair of perineal trauma.5 Based on this study, we recommend that hydrocodone dosages should be limited to 30 mg of the bitatrate salt (eg, six tablets of Vicodin) per day. If the mother requires higher daily dosages (eg, 35–40 mg), the breastfed neonate should be observed closely for depressed neurobehavior and suboptimal breastfeeding. The dosage of hydrocodone should be decreased or discontinued if either occurs. Daily dosages more than 40 mg should be avoided.
1. Beaver WT, McMillan D. Methodological considerations in the evaluation of analgesic combinations: acetaminophen (paracetamol) and hydrocodone in postpartum pain. Br J Clin Pharmacol 1980;10(Suppl 2):215S–23S.
2. Palangio M, Wideman GL, Keffer M, Landau CJ, Morris E, Doyle RT Jr, et al. Combination hydrocodone and ibuprofen versus combination oxycodone and acetaminophen in the treatment of postoperative obstetric or gynecologic pain. Clin Ther 2000;22:600–12.
3. Bodley V, Powers D. Long-term treatment of a breastfeeding mother with fluconazole-resolved nipple pain caused by yeast: a case study. J Hum Lact 1997;13:307–11.
4. Meyer D, Tobias JD. Adverse effects following the inadvertent administration of opioids to infants and children. Clin Pediatr (Phila) 2005;44:499–503.
5. Montgomery A, Hale TW, Academy of Breastfeeding Medicine Protocol Committee. ABM clinical protocol #15: analgesia and anesthesia for the breastfeeding mother. Breastfeed Med 2006;1:271–7.
6. Susce MT, Murray-Carmichael E, de Leon J. Response to hydrocodone, codeine and oxycodone in a CYP2D6 poor metabolizer. Prog Neuropsychopharmacol Biol Psychiatry 2006;30:1356–8.
7. Hutchinson MR, Menelaou A, Foster DJ, Coller JK, Somogyi AA. CYP2D6 and CYP3A4 involvement in the primary oxidative metabolism of hydrocodone by human liver microsomes. Br J Clin Pharmacol 2004;57:287–97.
8. Anderson PO, Sauberan JB, Lane JR, Rossi SS. Hydrocodone excretion into breast milk: the first two reported cases. Breastfeed Med 2007;2:10–4.
9. Ito S. Drug therapy for breast-feeding women. N Engl J Med 2000;343:118–26.
10. Coles R, Kushnir MM, Nelson GJ, McMillin GA, Urry FM. Simultaneous determination of codeine, morphine, hydrocodone, hydromorphone, oxycodone, and 6-acetylmorphine in urine, serum, plasma, whole blood, and meconium by LC-MS-MS. J Anal Toxicol 2007;31:1–14.
11. Begg EJ, Duffull SB, Saunders DA, Buttimore RC, Ilett KF, Hackett LP, et al. Paroxetine in human milk. Br J Clin Pharmacol 1999;48:142–7.
12. Begg EJ, Duffull SB, Hackett LP, Ilett KF. Studying drugs in human milk: time to unify the approach. J Hum Lact 2002;18:323–32.
13. Kraemer FW, Rose JB. Pharmacologic management of acute pediatric pain. Anesthesiol Clin 2009;27:241–68.
14. Stanos SP, Tyburski MD. Minor and short-acting analgesics, including opioid combination products. In: Benzon HT, Rathmell JP, Wu CL, Turk DC, Argoff CE, editors. Raj's practical management of pain. 4th ed. Philadelphia (PA): Mosby Elsevier; 2008. p. 613–42.
15. Oei J, Lui K. Management of the newborn infant affected by maternal opiates and other drugs of dependency. J Paediatr Child Health 2007;43:9–18.
16. Berlin CM Jr. Sensitivity of the young infant to drug exposure through human milk. Adv Drug Deliv Rev 2003;55:687–93.
17. Madadi P, Ross CJ, Hayden MR, Carleton BC, Gaedigk A, Leeder JS, et al. Pharmacogenetics of neonatal opioid toxicity following maternal use of codeine during breastfeeding: a case-control study. Clin Pharmacol Ther 2009;85:31–5.
18. Koren G, Cairns J, Chitayat D, Gaedigk A, Leeder SJ. Pharmacogenetics of morphine poisoning in a breastfed neonate of a codeine-prescribed mother. Lancet 2006;368:704.
19. Bateman DN, Eddleston M, Sandilands E. Codeine and breastfeeding. Lancet 2008;372:625.
20. Ferner RE. Did the drug cause death? Codeine and breastfeeding. Lancet 2008;372:606–8.
21. Madadi P, Chitayat D, Koren G. Codeine and breastfeeding - reply. Lancet 2008;372:626.
22. Willmann S, Edginton AN, Coboeken K, Ahr G, Lippert J. Risk to the breast-fed neonate from codeine treatment to the mother: a quantitative mechanistic modeling study. Clin Pharmacol Ther 2009;86:634–43.
23. Madadi P, Moretti M, Djokanovic N, Bozzo P, Nulman I, Ito S, et al. Guidelines for maternal codeine use during breastfeeding. Can Fam Physician. 2009;55:1077–8.
25. Otton SV, Schadel M, Cheung SW, Kaplan HL, Busto UE, Sellers EM. CYP2D6 phenotype determines the metabolic conversion of hydrocodone to hydromorphone. Clin Pharmacol Ther 1993;54:463–72.
26. Fleishaker JC, Desai N, McNamara PJ. Factors affecting the milk-to-plasma drug concentration ratio in lactating women: physical interactions with protein and fat. J Pharm Sci 1987;76:189–93.
Figure. No caption available.