Neonatal Exposure to Tramadol through Mother’s Breast Milk

Abstract

Tramadol is an opioid used in the treatment of moderate to moderately severe pain. Tramadol’s use during pregnancy is generally avoided and may cause some reversible withdrawal effects in neonates, and its use during lactation is not licensed by the manufacturer. A small clinical trial reported infants were exposed to <3% of a mother’s tramadol dose through breast milk with no evidence of harmful effects. Presented is a case study of breast milk, neonatal urine, and neonatal oral fluid for the analysis of tramadol and its metabolites, along with the validation of a method for the analysis of tramadol, O-desmethyltramadol, and N-desmethyltramadol in breast milk. Tramadol and its metabolites were extracted by solid-phase extraction after saponification of breast milk to remove lipids. Samples were analyzed by ultra-pressure liquid chromatography-tandem mass spectrometry. To the author’s knowledge, this is the first report of tramadol and its metabolites in neonatal oral fluid. The breast milk concentrations were 63, 22, and 76 ng/mL for the analysis of tramadol, O-desmethyltramadol, and N-desmethyltramadol, respectively, on day of life 12. On day of life 20, the breast milk concentrations were 1,254, 388, and 937 ng/mL for the analysis of tramadol, O-desmethyltramadol, and N-desmethyltramadol, respectively. Oral fluid concentrations were 1,011, 1,499, and 406 ng/mL for the analysis of tramadol, O-desmethyltramadol, and N-desmethyltramadol, respectively, on day of life 20. Oral fluid concentrations were similar to breast milk for tramadol, almost four times higher for O-desmethyltramadol, and less than half for N-desmethyltramadol. The absolute infant dose was calculated to be 10 μg/kg/day and 294 μg/kg/day for tramadol on day of life 12 and 20, respectively.

Introduction

Breastfeeding provides numerous benefits to both the mother and neonate. Neonatal benefits include decreased infection rate, increased immune system development, enhanced cognitive development, higher IQ at age 6–7 years (1), reduced incidence of necrotizing enterocolitis (1, 2), decreased gastrointestinal tract illness (3), lower respiratory tract illness (2, 3), lower risk of enteritis, otitis media, sudden infant death syndrome, respiratory syncytial virus infection, insulin-dependent diabetes mellitus, and allergies (2). Benefits to the nursing mother include decreased postpartum bleeding and quicker uterine involution due to increased levels of oxytocin, increased maternal–infant bonding, increased maternal sense of fulfillment and self-worth, and decreased risk of ovarian and pre-menopausal breast cancer (4). The current recommendation set forth by the American Academy of Pediatrics is that infants are exclusively breastfed for the first 6 months of life (1, 4). With an abundance of benefits to both mother and neonate, the decision to stop (or not start) breastfeeding while taking medications is not straight forward. The benefits of breastfeeding and risks of drug exposure to the neonate must be assessed (5). In general, most drugs are not of concern in breastfeeding (6) even though most licit and illicit drugs can pass into the breast milk (2). The decision is often heavily based on prenatal drug exposure, with a higher propensity to breastfeed when continuing treatment with a drug that was used while the baby was in utero compared to a new pharmacological treatment (1, 5). In order to minimize risks to the neonate, alternative medications with shorter half-lives can be chosen, and the mother can be advised to feed the infant immediately prior to taking the medication (6). Even though breast feeding women and their infants are often excluded from the drug development process (1), causing a gap in safety, only a small proportion of medications are contraindicated for concurrent use with breastfeeding (5).

Almost all drugs are transferred into the breast milk to some extent (6). Drugs enter the breast milk mainly through passive diffusion (1, 2, 4, 6, 7), but some active transport (1, 4) or carrier-mediated transport (2, 4) are also present. Breast milk has a pH slightly lower than plasma, 6.6–7.0 (8). Due to its increased acidity, ion trapping can occur in the breast milk for weak bases (6, 7). Infant exposure to drugs in breast milk is dependent on the milk-to-plasma ratio, milk intake, and infant drug clearance (2). Daily infant dose can be calculated as the average concentration of drug in breast milk times the volume of milk ingested in 24 hours (1, 2, 5) or by the milk-to-plasma ratio times maternal plasma concentration times the volume of milk ingested in 24 hours. In the breastfed infant, the estimated volume of milk is 150 mL/kg/day by day 3–4 of life (1, 2, 4–6, 9, 10). The relative infant dose must take into account the advancing postnatal age of the infant given the change in infant stomach capacity at birth. Thus, the dose can be calculated as the daily infant dose/maternal dose times 100 (9) or by calculating the milk-to-plasma ratio divided by clearance times the daily milk intake times 100 (1). Error often occurs in the estimation of the milk-to-plasma ratio due to infant postnatal age, breast milk volume consumed, and the changes in breast milk composition over time and during feed (7). A relative infant dose <10% is often considered to be safe for use with breastfeeding and is unlikely to cause harmful effects to a full-term infant (11). Several properties of drugs promote low milk concentrations including large volume of distribution, low lipid solubility, ionized in the plasma, large molecular weight, and high protein binding (7), which can minimize harmful effects to the neonate by limiting total drug exposure. Once drugs have entered the breast milk, drugs with short elimination half-lives and low bioavailability have less effect and improved safety in the neonate (7).

Tramadol is an opioid used in the treatment of moderate to moderately severe pain (9–12). It acts on the μ-opioid receptors and is a serotonin–norepinephrine reuptake inhibitor (10). Tramadol is metabolized through the cytochrome P450 isozyme CYP2B6, CYP2D6 and CYP3A4 (10, 13). The major metabolites produced are the O- and N-desmethyltramadol (10). The O-desmethyltramadol metabolite has 300 to 400 times the μ-affinity of the parent drug (12). Tramadol’s use during pregnancy is generally avoided and may cause reversible withdrawal effects in neonates, and use during lactation is not licensed by the manufacturer (11). The literature is contradictive in its safety for use in nursing mothers. One case study strongly advises against it due to the death of an infant exposed to tramadol through the breast milk of a tramadol addict (12). On the other hand, a small clinical trial reported infants were exposed to <3% of a mother’s tramadol dose through breast milk with no evidence of harmful effects and suggested short-term maternal use was compatible with breastfeeding (9). This clinical study reported a milk-to-plasma ratio of 2.2 and 2.8 for tramadol and O-desmethyltramadol, respectively (9). Additionally, they reported a relative infant dose of 2.24 and 0.64% for tramadol and O-desmethyltramadol, respectively (9).

Prior reports indicate the use of umbilical cord tissue/blood (14–19), meconium (14, 15, 18–23), urine (14, 15, 18, 19, 21), blood (14, 19), or hair (14, 15, 19, 21) for the detection of neonatal drug exposure. However, umbilical cord tissue/blood is only useful for in utero exposure, and meconium is only useful for in utero exposure or drug exposure shortly after birth. Urine, blood, and hair can be collected at any time after birth for the determination of drug exposure. While blood will provide the most accurate account of current drug load in the neonate’s system, the collection is invasive, difficult to collect, and limited sample volume is available from newborns (14, 24).

Urine is a commonly used matrix for neonatal drug testing; however, it is difficult to collect from the neonate due to placement and timing of collection bags or cotton balls (14, 15, 25). The window of detection, like blood, is extremely dependent on the elimination half-life of the drug of interest (26) and polar metabolites are the predominate form found in the urine (24).

Neonatal hair can be difficult or nearly impossible to collect due to the limited amount of specimen available (15, 19, 24, 27, 28) and refusal from mother for cosmetic or cultural reasons (15, 27). When specimen volume is too low the collection process can be almost invasive (24). Additionally, hair collected within a few weeks after birth will only detect in utero drug exposure (14).

Due to the invasiveness, challenges, and limitations in collecting specimens and/or analytical difficulties of these matrices, oral fluid has the potential to be an alternate matrix in detecting drug exposure in neonates. Oral fluid collection devices have become a significant tool in workplace and regulated drug testing for adults due to their simplicity and ease of a non-invasive collection of a standardized volume of oral fluid; however, oral fluid has not been traditionally used as a specimen for neonatal drug exposure. Neonatal or infant oral fluid has been used in therapeutic drug monitoring (TDM) as an alternative to blood (29–34). The use of ad-hoc collection devices has been proven effective in several studies, as commercial devices are too large for neonatal or infant mouths (29–31, 33–35). Oral fluid collection is fast, easy, non-invasive, (14) and will detect recent exposure (14, 28) that may correlate to blood concentrations (14).

Presented is an evaluation of breast milk, neonatal urine, and neonatal oral fluid for the analysis of tramadol and its metabolites, along with the validation of a method for the analysis of tramadol, O-desmethyltramadol, and N-desmethyltramadol in breast milk. Tramadol and its metabolites were extracted by solid-phase extraction after saponification of breast milk to remove lipids. Samples were analyzed by ultra-pressure liquid chromatography–tandem mass spectrometry (UPLC–MS-MS).

Case History

A female neonate was born via caesarean delivery with use of an epidural with a gestational age of 38 weeks and growth restriction with a birth weight of 1,920 g. The mother was enrolled in methadone maintenance therapy at the Children’s Hospital of Richmond at Virginia Commonwealth University (VCU). Maternal history included treatment with prescriptions for methadone, ondansetron, alprazolam, citalopram, and tramadol; self-reports of cigarette and opiate use, and a history of obesity, depression, anxiety, migraines, and asthma. The neonate was initially fed by donor milk and transitioned to mixed donor and mother’s breast milk starting on day of life 2. On day of life 4, the neonate began treatment with methadone for neonatal opioid withdrawal symptoms. In accordance with a clinical study, oral fluid samples were collected on day of life 0, 1, 2, 3, 4, 12, and 20; urine samples were collected on day of life 2, 4, 7, 12, and 20 and breast milk samples were collected on day of life 12 and 20 when the infant was exclusively feeding mother’s milk. Methadone treatment was ceased on day of life 22, and the infant was discharged from the hospital on day of life 25.

Materials

All glassware, tubing, and fritted glass dispersion tubes were purchased from Colonial Scientific (Richmond, VA, USA). Acetonitrile, ammonium hydroxide (NH4OH), ammonium formate, ethyl ether, hydrochloric acid (HCl), methanol (MeOH), n-butyl chloride, sulfuric acid, butyl acetate, and water (H₂O) were purchased from Fisher Scientific (Hanover Park, IL, USA). Helium, hydrogen, and nitrogen gases were purchased from Praxair (Richmond, VA, USA). Tramadol, O-desmethyltramadol. N-desmethyltramadol, tramadol-C13-d₃, and O-desmethyltramadol-d₆ reference standards were purchased from Cerilliant (Round Rock, TX, USA). Enfamil^® Premium™ Newborn Infant Formula 0–3 months powder was purchased from a local grocery store. LiquiChek™ Urine Chemistry Control Level 2 and LiquiChek™ TDM Controls Level 3 were purchased from Bio-Rad Laboratories (Hercules, CA).

Experimental

Sample collection and screening analysis

All samples were collected at the Children’s Hospital of Richmond at VCU after obtaining informed written consent under VCU IRB number HM20006070. Breast milk samples, up to 2 mL, were collected at the start of feeding by hand expression or electric pump and stored at −80°C until analysis. Samples were collected on day of life 12 at 15:40, and day of life 20 at 20:00.

Urine samples were collected by placing a cotton ball inside of the neonate’s diaper. Approximately 2 mL of urine was expressed by hand from the cotton ball and stored at −80°C until analysis. Samples were collected on day of life 2 at 02:30, 4 at 04:00, 7 at 05:00, 12 at 16:35, and 20 at 23:00.

Oral fluid samples were collected prior to feeding using a modification of the ad-hoc collection by Gesseck et al. (35). In brief, three sterile foam-tipped swabs were rolled along the infant’s oral mucosal surface of the mouth, inner cheeks, and tongue until saturated. The swabs were rinsed in phosphate-buffered saline (PBS) and discarded. The contents were centrifuged, and the supernatant was stored at −80°C until analysis. Samples were collected on day of life 0 at 06:00, 1 at 13:58, 2 at 08:00, 3 at 16:20, 4 at 09:30, 12 at 15:40, and 20 at 20:00.

Urine samples were screened for basic drugs of abuse by gas chromatography–mass spectrometry (GC–MS) after a liquid–liquid extraction (LLE), using a modified version of a previously validated method for basic drugs of abuse (36). In brief, 100 μL matrix-matched calibrators, controls, or collected urine samples, 50 μL internal standard in MeOH and 100 μL NH₄OH were added to a glass test tube and vortexed. After addition of 2 mL 3:1 n-butyl chloride:ethyl ether, the samples were mixed by mechanical means for 5 minutes, followed by centrifugation for 5 minutes, and the top layer was transferred to a clean glass test tube. One milliliter of 2 N sulfuric acid was added, mixed by mechanical means for 5 minutes, followed by centrifugation for 5 minutes, and the top layer was aspirated to waste. One milliliter of hexane was added, mixed by mechanical means for 5 minutes, followed by centrifugation for 5 minutes, and the top layer was aspirated to waste. Five-hundred microliters of NH₄OH was added, samples were vortexed and transferred to a conical glass tube. One-hundred microliters of butyl acetate was added, mixed by mechanical means for 5 minutes, followed by centrifugation for 5 minutes, and all but 100 μL of the bottom layer was removed. Samples were centrifuged again for 5 minutes, and 50 μL of the top layer was transferred to auto-sampler vials.

Urine samples were analyzed by GC–MS using a modification of a previously validated method for basic drugs of abuse (36). In brief, a 1-μL aliquot was injected into a Shimadzu GC–MS QP2020 (Kyoto, Japan) equipped with an Agilent 30 m x 0.25 mm x 0.25 μm HP-5 MS column (Santa Clara, CA). Helium was used as the carrier gas and set to a flow rate of 2.0 mL/min. The GC oven was programmed to an initial temperature of 70°C with a hold time of 2 min, followed by a 15°C/min ramp to a final temperature of 300°C held for 10 min. The total run time was 27.33 min. The ion source and interface temperatures were 200°C and 280°C, respectively. The MS, equipped with an electron ionization (EI) source, was operated using scan mode from m/z 35 to 550 (scan speed = 3,333, event time = 20 msec). Generated mass spectra were compared to the SWGDRUG MS library version 3.3c.

Confirmatory method

Breast milk, urine, and oral fluid samples were confirmed for tramadol, O-desmethyltramadol, and N-desmethyltramadol by UPLC–MS-MS after SPE. In brief, due to limited access to genuine breast milk, Enfamil^® Premium™ newborn infant formula was used to prepare calibrators (10, 25, 50, 100, 250, 500, and 1,000 ng/mL), controls (10, 30, 150, and 750 ng/mL), and blank samples. Additional calibrators were also prepared in urine at the same concentrations and purchased controls LiquiChek™ Urine Chemistry Control Level 2 and LiquiChek™ TDM Control Level 3 were used. To each calibrator, control, or sample, 10 μL of 100 ng/mL tramadol-C13-d₃ and O-desmethyltramadol-d₆ in MeOH was added to a glass tube. After the addition of 100 μL of 1 N HCl, samples were vortexed for 0.5 minutes to remove lipids by saponification (37). Samples were allowed to equilibrate for 5 minutes, followed by centrifugation for 5 minutes at 5,000 g. A Waters™ Oasis^® MCX 96-well μElution plate (Milford, MA) with 2 mg sorbent per well was conditioned with 200 μL MeOH followed by 200 μL H₂O. Samples were added to the plate and aspirated. Columns were washed consecutively with 200 μL H₂O and 200 μL MeOH. Samples were eluted with 50 μL of 3:2:0.25 ACN:MeOH:NH₄OH (v:v:v). The eluent was diluted with 100 μL H₂O.

A 5-μL sample was injected and analyzed by UPLC–MS-MS on a Xevo TQD LC–MS-MS with an ACQUITY UPLC^® (Waters Corporation; Milford, MA) using a previously validated method for urine. In brief, chromatographic separation was achieved by using a gradient consisting of 20 mM ammonium formate in H₂O (mobile phase A) and 20 mM ammonium formate in MeOH (mobile phase B) at a flow rate of 0.6 mL/min on a UCT Selectra^® DA, C18, 2.1 × 50 mm, 3 μm column (Bristol, PA) at 40°C. The gradient used was: 0.00 to 1.5 minutes at 60% A, 1.5 to 3 minutes at 30% A, 3 to 3.5 minutes at 95% A and then return to 5% B at 3.6 minutes, at a flow rate of 0.6 mL/min. The MS parameters were as follows: source temperature, 150°C; capillary voltage, 3.0 kV; desolvation temperature, 600°C; desolvation gas (nitrogen), 100 L/h and ion source gas, 100 L/h. Samples were acquired using multiple reaction monitoring (MRM) mode, the transition ions (m/z) with their corresponding cone voltage (V) corresponding collision energies (eV) are shown in Table I.

Table I.

MRM Parameters

Analyte	Precursor ion (m/z)	Product ion (m/z)	Cone voltage (V)	Collision energy (eV)
Tramadol	264.3	42.3	26	47
	264.3	58.0	26	15
Tramadol-C13-d3	268.3	42.3	26	49
	268.3	58.0	26	17
O-desmethyltramadol	250.2	42.4	28	49
	250.2	58.0	28	15
	250.2	232.1	28	11
O-desmethyltramadol-d6	256.3	45.4	28	41
	256.3	64.0	28	17
	256.3	238.2	28	9
N-desmethyltramadol	250.2	44.0	24	11
	250.2	232.1	24	9

Method Validation

This method was validated according to Approved American National Standard/American Academy of Forensic Sciences (AAFS) Standards Board (ANSI/ASB) Standards (38) for linearity, limits of detection and quantitation, bias and precision, carryover, selectivity, stability, and interferences. The linearity of the assay was verified from seven-point calibration curves prepared in breast milk with the following concentrations: 10, 25, 50, 100, 250, 500, and 1,000 ng/mL. A linear regression of the ratio of peak area counts of the compound to that of the deuterated internal standard versus the corresponding concentration ratio was used to construct the calibration curve. The limits of detection and quantitation were administratively set to 10 ng/mL.

Bias and precision were determined from quality control samples prepared in infant formula: limit of quantitation control (LOQC) at 10 ng/mL, low control (LQC) at 30 ng/mL, mid control (MQC) at 150 ng/mL, high control (HQC) at 750 ng/mL, and a dilution control (DQC) at 5,000 ng/mL diluted 1:4. Control samples were analyzed in triplicate over 5 days. Bias was calculated as a percentage of deviation (%Dev) and acceptable bias did not exceed ± 20% of the target concentration. Two types of precision were assessed during the validation: intra-day and inter-day. Intra-day precision was calculated for each concentration with six replicates over one day. Inter-day precision was calculated for each concentration in three replicates over the 5 days. Precision was calculated as a coefficient of variation (% CV, standard deviation/mean) and acceptable precision did not exceed 20%.

Sample carryover was evaluated over 5 days using two different procedures. First, immediately following the injection of the highest calibrator a negative control (analyte free) was injected. Second, an injection of HQC was immediately followed by injection of the LQC. Lack of carryover was confirmed if none of the compounds were detected in the negative control, and the LQC samples did not demonstrate quantified bias of more than 20%.

Selectivity was determined by analyzing 10 authentic breast milk samples fortified at the LQC concentration with internal standard, and unfortified with and without internal standard. Acceptable selectivity was if there were no peaks that co-eluted with tramadol, its metabolites, or the internal standards.

Stability was determined under several conditions and time intervals. The experiments were performed using two of the control samples: LQC and HQC. All studies included three replicate analyses of each control sample. The “freeze/thaw” stability of analytes was assessed by putting control samples through three freeze/thaw cycles with the last freeze cycle lasting 24 hours. The “bench-top” stability of the analytes was assessed to evaluate the possible effects of specimen transportation and processing in the laboratory by having the control samples sit at room temperature for 72 hours. Both freeze/thaw and benchtop were prepared and quantitated against freshly prepared calibrators. The “post-preparative” stability of the analytes was evaluated by having extracts sit in the auto-sampler of the UPLC–MS-MS. A batch of the extracted controls was quantitated against a freshly prepared calibration curve. The extracted controls were then allowed to sit in the auto-sampler for 24 hours at 4°C after which they were re-injected and quantitated from the initial calibration. The results of the initial analysis were compared to those of the re-injected samples.

Interferences were evaluated by extracting and analyzing LiquiChek™ Urine Chemistry Level 2 and TDM Level 3 controls in the method described. Interferences were evaluated based on the absence or presence of peaks that co-eluted with tramadol, its metabolites, or the internal standards.

Results

Method validation

The assay was determined to be linear from 10 to 1,000 ng/mL with a r² > 0.995 for all analytes using a linear regression through zero with 1/x weighting. The bias and precision results for the LOQC, LQC, MQC, HQC and DQC can be found in Table II. All results were within ± 20% of the target concentration and within 20% CV. No carryover was observed in the negative control immediately following the injection of the highest calibrator. Additionally, no bias was observed in the LQC following the high control. No endogenous compounds interfered with tramadol, its metabolites, or the internal standards. The results of the stability tests demonstrated that the analytes of interest were stable for three freeze-thaw cycles, 72 hours on the bench-top, and 24 hours post-preparative reinjection.

Table II.

Bias and Precision

		Intra-day (n = 6)			Inter-day (n = 15)
Control	Conc. (ng/mL)	TRAM	ODT	NDT	TRAM	ODT	NDT
LOQC	10	0 (1)	1 (4)	−10 (7)	0 (11)	0 (13)	0 (16)
LQC	30	0 (2)	−3 (3)	−3 (10)	0 (7)	3 (11)	3 (13)
MQC	150	0 (3)	−1 (4)	−7 (10)	3 (3)	2 (4)	0 (10)
HQC	750	5 (3)	2 (3)	−4 (9)	5 (3)	1 (4)	3 (8)
DQC	5,000	8 (3)	4 (3)	12 (8)	6 (4)	0 (6)	12 (7)

Results listed as %Dev. (%CV). TRAM = Tramadol, ODT = O-desmethyltramadol, NDT = N-desmethyltramadol.

No interferences were detected in the analysis of LiquiChek™ Urine Chemistry Level 2 and TDM Level 3 controls.

The analysis of urine samples collected on day of life 2, 4, 7, 12, and 20 by LLE and GC–MS base screen yielded a presumptive positive for tramadol and its metabolites on day of life 20, Figure 1. All samples prior to day of life 20 had no detectable levels of tramadol or its metabolites.

Figure 1.

GC–MS total ion chromatogram of tramadol and its metabolites in urine day of life 20 sample.

The results of the analysis of urine, oral fluid, and breast milk samples collected on day of life 12 and 20 by SPE and UPLC–MS-MS can be seen in Table III. The chromatogram for the day of life 20 breast milk sample can be seen in Figure 2.

Table III.

UPLC–MS-MS Results

		Analyte
Matrix	Day of life	TRAM	ODT	NDT
Breast milk	12	63	22	76
Breast milk	20	1,254	388	937
Oral fluid	12	ND	ND	ND
Oral fluid	20	1,011	1,499	406
Urine	12	ND	ND	ND
Urine	20	14	<10	<10

Results listed as ng/mL. TRAM = Tramadol, ODT = O-desmethyltramadol, NDT = N-desmethyltramadol. ND = None Detected.

Figure 2.

LC–MS-MS total ion chromatogram of tramadol and its metabolites in breast milk day of life 20 sample.

Discussion

The presented UPLC–MS-MS method demonstrated acceptable reliability and reproducibility for the detection and quantitation of tramadol, O-desmethyltramadol, and N-desmethyltramadol in breast milk using SPE. Bias was determined to be within ± 20% of the target concentrations. Precision of the assay were determined not to exceed CVs of >20% over the dynamic range of the assay demonstrating the robustness of the assay. Additionally, the assay was free of significant interference from the matrix and free from analyte carryover.

Due to limited access of genuine drug-free breast milk, Enfamil^® Premium™ newborn infant formula was used to prepare calibrators and controls. While it is best practice to use true matrix-matched controls, previous studies have shown no significant differences between Enfamil^® and human breast milk (37). Additionally, the composition of human breast milk is complex and changes throughout the feed and over time (7). Enfamil^® Premium™ newborn infant formula was used as an alternative matrix due to (i) the lack of drug-free human breast milk and (ii) the composition of the samples being collected solely during the start of feed which is not comparable to donor milk from a breast milk bank. Deuterated internal standards were used to compensate for any extraction issues and/or matrix effects.

The composition of neonatal oral fluid is different than adults with statistically significant differences for calcium, chloride, magnesium, phosphorous, IgA, amylase and pH concentrations and changes throughout the postnatal period (39). Due to the inability to obtain matrix-matched neonatal oral fluid for each day of collection, the authors could not prepare matrix-matched oral fluid calibrators or controls. The oral fluid samples were extracted alongside the breast milk and urine calibrators and controls with the use of deuterated internal standards to compensate for any extraction issues and/or potential matrix effects.

To the authors’ knowledge, this is the first report of tramadol and its metabolites in neonatal oral fluid. Oral fluid concentrations were similar to breast milk for tramadol, almost four times higher for O-desmethyltramadol, and less than half for N-desmethyltramadol. Urine concentrations were much lower than breast milk or oral fluid for tramadol and its metabolites. The low urine concentration may have occurred due to the potential of drug loss due to absorption to the cotton ball and/or hand, or due to the timing of collection compared to the last dose of tramadol which is unknown. The absolute infant dose was estimated to be 10 and 294 μg/kg/day tramadol on day of life 12 and 20, respectively. Previously reported absolute infant doses for tramadol are 112 (102–122) μg/kg/day when given 100 mg tramadol every 6 hours (9). The absolute infant dose was estimated to be 3 and 91 μg/kg/day O-desmethyltramadol on day of life 12 and 20, respectively. Previously reported absolute infant doses for O-desmethyltramadol are 30 (28–32) μg/kg/day when given 100 mg tramadol every 6 hours (9). The absolute infant dose was estimated to be 13 and 220 μg/kg/day N-desmethyltramadol on day of life 12 and 20, respectively. No previous reports of absolute infant doses for N-desmethyltramadol are found in the literature. The mother’s prescription history reported 50 mg tramadol taken orally, no interval was provided; however, tramadol at 50 mg orally is often taken every 4–6 hours as needed (40). The average milk concentration at steady state for 100 mg taken every 6 hours is 748 (681–815) ng/mL tramadol and 203 (188–217) ng/mL O-desmethyltramadol (9). The breast milk concentrations observed on day of life 12 (63 and 22 ng/mL tramadol and O-desmethyltramadol, respectively) were well below the average steady-state concentration. With low concentration in the breast milk samples on day of life 12 and negative neonatal urine samples prior to day of life 12, it is suspected that the mother began taking tramadol on day of life 12. The breast milk concentrations observed on day of life 20 (1,254 and 388 ng/mL tramadol and O-desmethyltramadol, respectively) were well above the average steady-state concentration. It is suspected that the mother may have reduced clearance of tramadol, has slow metabolism, or is taking tramadol more frequently than prescribed, leading to an accumulation of tramadol in the breast milk. Assuming the mother took 50 mg tramadol every 4 hours, the relative infant dose on day of life 12 is 0.3 and 0.1% tramadol and O-desmethyltramadol, respectively, and the relative infant dose on day of life 20 is 9.1 and 2.8% tramadol and O-desmethyltramadol, respectively. Previously reported relative infant doses for tramadol and O-desmethyltramadol are 2.24 (2.04–2.44)% and 0.64 (0.59–0.69)%, respectively (9). The detection of tramadol, O-desmethyltramadol, and N-desmethyltramadol demonstrates proof of concept that the simple collection using a three sterile foam-tipped swabs rinsed in PBS is viable for the collection of oral fluid from neonates for the detection of tramadol and similarly related compounds.

Conclusion

The SPE using Waters™ Oasis^® MCX 96-well μElution plate with a saponification pre-step and UPLC–MS-MS analysis is a robust, reliable method for the detection and quantitation of tramadol, O-desmethyltramadol, and N-desmethyltramadol in breast milk. A significant finding in this case is that the estimations of the absolute infant dose and relative infant dose were greater than previously reported in the literature. Additionally, this study demonstrates the use of three sterile foam-tipped swabs rinsed in PBS as an acceptable alternate to traditional collection devices for the detection of tramadol and its metabolites in neonates. Oral fluid appears to have an increased ability to detect recent neonatal drug exposure compared to urine. The collection of oral fluid from neonates was easier than using a cotton ball within the diaper for urine collection. Oral fluid collection from neonates also has the potential to be used for the evaluation of neonatal pharmacokinetics and breast milk-to-oral fluid ratios. This non-invasive oral fluid collection during wellness checks postpartum may help indicate infant exposure to substances, and/or if a mother has refrained during pregnancy.

Funding

This project was supported in part by the National Institutes of Health Center for Drug Abuse [P30DA033934]. The opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect those of the National Institutes of Health.