ABSTRACT
Cerumen (earwax) is a complex mix of apocrine and holocrine secretions and skin cells. In recent years, it has emerged as an innovative specimen for the detection of xenobiotics, suggesting that it has both clinical and forensic applications. An LC–MS/MS method was developed for the detection of 33 psychoactive medications in cerumen and urine and applied to samples from 60 psychiatric inpatient volunteers. The history of administration and qualitative results between urine and cerumen were generally consistent, with haloperidol, quetiapine, and nebivolol being detected more frequently in cerumen. In certain cases, drugs were detected both with very recent and without any recent administration, even when they were not in urine, highlighting opportunities for utilization in monitoring both acute and chronic administration and even following the cessation of administration. However, no clear relationship was found between drug concentrations and time of administration or total daily dose in most cases, consistent with challenges of routine collection. Cerumen concentrations were higher among more hydrophilic drugs with lower molecular weight and protein binding and demonstrated the best fit with an inverse regression model. No relationship was found with patient sex. Our results suggest that the disposition of drugs in cerumen is complex, being downstream of insufficiently elucidated pathways through migrating skin cells, ceruminous, and sebaceous glands. The present study is the first to investigate the routine collection of cerumen in a clinical setting and has broad implications for the interpretation of drug testing results with this emerging alternative biological specimen.
Keywords: alternative biological matrix, apocrine sweat, clinical drug testing, earwax, therapeutic drug monitoring
An LC–MS/MS method was developed for the detection of 33 psychoactive medications in cerumen (earwax) and applied to the samples of 60 psychiatric inpatient volunteers. The findings highlighted the utility of cerumen as an innovative biospecimen in clinical toxicological analysis.
1. Introduction
Human cerumen—known as earwax—is a biospecimen that may offer several advantages in forensic and clinical applications, due to its unique properties of accumulation and stability. Cerumen is produced in the outer ear canal when the secretions from ceruminous and sebaceous glands [1, 2, 3] mix with keratinized desquamated epithelial cells [4]. The resulting acidic and waxy product contains 60% keratin, 12%–20% fatty acids, alcohols, squalene, and 6%–9% of cholesterol [5, 6]. Unlike blood or plasma, cerumen provides a noninvasive means of assessing long‐term drug exposure and adherence, rendering it an attractive alternative. In recent years, the development of sensitive analytical technologies has allowed the use of cerumen for the detection of various xenobiotics, and several studies have emerged presenting findings mostly in volunteers. Antiepileptics and antipsychotics have been detected and quantified by an LC–MS/MS method in cerumen from patients who had been prescribed these medications [7]. Meier et al. conducted a pilot study for the identification of the designer drug 4‐fluoroamphetamine following controlled oral administration [8], and Nicolaou et al. and Christodoulou et al. determined the presence of certain cannabinoids in cerumen among volunteers who were frequent cannabis users [9, 10]. Massadeh et al. used HPLC with a fluorescence detector to compare tramadol concentrations in the cerumen and urine samples from 40 postoperative patients [11]. Opiates, amphetamine derivatives, cocaine, methadone, diazepam, and cannabinoids have been detected by LC–MS and GC–MS in the cerumen of 38 deceased persons, and the findings were compared with those in blood, urine, and hair samples of the same cases [12]. Other studies in post‐mortem cases include the detection of opiates, methadone, cocaine, antidepressants, benzodiazepines, antipsychotics, and antiparkinsonian drugs in cerumen, blood, and urine via LC–MS/MS in 25 decedents and a case study of an apparent fentanyl overdose where the concentrations of cerumen and norfentanyl were determined [13, 14].
In addition, cerumen has previously been shown to indicate the accumulation of persistent organic pollutants [15, 16, 17], heavy metals [18, 19], and nicotine and its metabolites [20].
The mechanism of disposition of drugs in cerumen is not elucidated in scientific literature, but since cerumen is a product of primarily apocrine secretion (ceruminous glands) and to a smaller extent by holocrine secretion (sebaceous glands) [21] as well as dead skin cells migrating gradually from the tympanic membrane, it appears that it is involved in drug excretion in a complex manner and with a wide timeframe. Substances have been detected in cerumen even as early as 1 h after administration [8] and as late as up to 2 months after the last administration [7]. Meier et al. have placed the time window of drug detection for several drug classes in cerumen between urine and hair [12].
Little data are available about the excretion of drugs through apocrine sweat glands, as there are very few studies investigating the detection of drugs in apocrine sweat, conducted by radioimmunoassay and gas chromatography–mass spectrometry for the detection of tetrahydrocannabinol, benzodiazepines, cocaine, barbiturates, morphine, and methadone [22, 23], as opposed to eccrine sweat, which originates from sweat glands present in most of the skin covering the human body. Conversely, apocrine sweat glands are primarily present in the axillary, mammary, perineal, and genital regions [24]. Systemic circulation reaches the external ear primarily by the posterior auricular artery. Blood then supplies the dermis, but not the surface of the skin. Drugs may then be excreted there through various pathways, including via sweat glands and the process of desquamation [25]. Lipophilic drugs seem to accumulate in the adipose tissue and the stratum corneum in the subcutaneous tissue below the skin surface [26, 27], and the transit rate from subcutaneous tissue and from sweat glands is likely different [27]. There is, however, little subcutaneous structure to the medial two‐thirds of the skin of the EAC.
In the present study we developed and validated an UHPLC/MS–MS method for the analysis of 33 pharmaceuticals covering various psychiatric and other drugs (antipsychotics; antidepressants including selective serotonin reuptake inhibitors; anxiolytics, sedatives, and hypnotics, including benzodiazepines; beta‐blockers; analgesics; antiepileptics; anti‐dementia anticholinesterase inhibitors; and antiparkinsonian anticholinergic agents) in cerumen and urine samples from 60 clinical cases with the aim to compare the findings from cerumen and urine in terms of the provided information regarding the medication history. Furthermore, we investigated associations with the administration history (total daily dose and the interval since the last administration) and the relationships of drug concentrations and the drugs' pharmacologic and physicochemical parameters.
Overall, comprehensive data on the detection and concentration levels of multiple pharmaceuticals in cerumen samples from 60 clinical cases are presented, highlighting their potential utility in personalized medicine, forensic investigations, and drug adherence assessment. By exploring the correlation of the findings between cerumen drug levels and traditional biological matrices such as urine, this study aims to showcase cerumen as a reliable specimen for drug monitoring. To our knowledge, this is the first study conducted in a clinical setting with routine collection of cerumen.
2. Materials and Methods
2.1. Chemicals and Reagents
A Milli‐Q water purification system was used for the acquisition of HPLC‐grade water. LC–MS grade acetonitrile was purchased from Fisher Scientific International Inc. (Hampton, New Hampshire, USA), and formic acid was purchased from Honeywell Riedel‐de Haën (Charlotte, North Carolina, USA). Standards of the highest available purity were used, and deuterated internal standards for haloperidol‐d4 and fluoxetine‐d6 were purchased from LGC Standards (Teddington, Middlesex, UK). Commercially available consumer‐grade cotton swabs were used for the collection of cerumen samples.
2.2. Calibrators and Quality Control Solutions
Stock solutions were prepared by dissolving the 33 compounds of interest and the two deuterated internal standards in methanol and stored in glass vials at −20°C. Calibration standards and quality controls were prepared by spiking drug‐free pooled cerumen at eight concentration levels (0.1, 0.5, 1, 5, 10, 15, 20, 25 μg/g). Specifically, eight drug‐free cerumen samples of 0.0743‐g total weight were dissolved in 5 mL of methanol and, following vortexing, were split into 10 pooled aliquots of 500 μL. Then, 50 μL of a working mixture solution of analytes at the various concentration levels and 25 μL of each internal standard were added to the aliquots. Finally, vortexing and centrifugation were applied. For urine calibrants and QCs, five drug‐free urine samples were mixed and vortexed to create a pooled urine sample, which was then split into 160‐μL aliquots. Two hundred microliters of a methanolic working mixture solution of the analytes at various concentrations was evaporated under a gentle stream of nitrogen and redissolved in each urine aliquot. Twenty microliters of each internal standard was then added, and the mixture was vortexed and centrifuged. Spiked urine samples derived at seven concentration levels (topiramate, paracetamol; 0.1, 0.5, 2, 10, 25, 50, 75 μg/mL, metoprolol, levetiracetam, clozapine, paliperidone, citalopram, nebivolol, duloxetine, chlorpromazine, fluoxetine, lamotrigine; 0.05, 0.1, 0.5, 2, 10, 25, 50 μg/mL, lorazepam, diazepam, atenolol, zolpidem, olanzapine, risperidone, paroxetine, biperiden, venlafaxine, bromazepam, quetiapine, haloperidol, sertraline, amitriptyline, mirtazapine, aripiprazole, levomepromazine, donepezil, buspirone, bisoprolol; 0.001, 0.01, 0.05, 0.1, 1, 5, 10 μg/mL, pregabalin; 0.01, 0.1, 1, 5, 25, 50, 150 μg/mL).
2.3. Study Participants and Sample Collection
Sixty cerumen and urine samples were collected from volunteer inpatients at the Psychiatric Clinic “Agia Eirini” in Drama, Greece, and the 1st Department of Psychiatry, Aristotle University of Thessaloniki, located in General Hospital “Papageorgiou” in Thessaloniki, Greece, who had already been prescribed and routinely administered psychoactive medications. Urine samples were collected for comparison with qualitative results with cerumen, as it is a minimally invasive specimen, and the analytes can usually be detected for several hours and/or days, and is thus routinely used to verify medication adherence. Urine samples were initially screened by GC–MS. After the study participants were duly informed of the study aims and procedures and informed consent was obtained, the collection of the samples was performed by the study participants themselves, under the direction of the investigators and medical staff in order to avoid discomfort or possible adverse events from the collection process using untreated cotton swabs and urine collection containers with no added preservatives. Cerumen samples were collected from both ears in separate cotton swabs and stored at room temperature, whereas urine samples were stored at −20°C prior to analysis. Anonymized demographic data and medical history were also collected. The study was approved by the 3rd Regional Health Authority of the Hellenic Republic (28459/14‐06‐2023), the Bioethics Committee of the Department of Medicine, Aristotle University of Thessaloniki (2.481/7‐12‐2021), the Scientific Council of the Papageorgiou General Hospital (ES 370/07‐06‐2023), and the legal representative of the Psychiatric Clinic “Agia Eirini” (08‐07‐2022).
2.4. Sample Pretreatment
The cerumen samples were weighed immediately following sampling and prior to pretreatment and stored at room temperature (sample net weights are given in Table S1). Prior to extraction, samples were stored in a dry space, at room temperature (approximately 25°C), and analyzed within 48 h of collection. Left or right ear samples were analyzed for each participant in a random sequence (53.3% of left ear and 46.7% of right ear samples were used). On the day of the analysis, the sample‐containing cotton tip of the swab was carefully removed and added to an Eppendorf tube containing 600 μL of methanol and the internal standards at 0.1 μg/g. The samples were mixed with a vortex mixer for 10 min, ultrasonicated for 30 min, and then centrifuged for 10 min at 10,000 × g. The supernatant was then collected. The urine samples were stored at −20°C. On the day of the analysis, they were centrifuged for 10 min at 10,000 × g, and 200 μL of the supernatant was collected. Six hundred microliters of water:acetonitrile 95:5 were added and the internal standards at 1 μg/mL. The samples were then mixed with a vortex mixer for 10 min and centrifuged again for 10 min at 10,000 × g, and the supernatant was collected. The sample pretreatment protocol for cerumen was an adaptation of previous studies on the subject, using methanol as a universal solvent to ensure the recovery of a wide range of xenobiotics from the lipid‐rich cerumen matrix. Unlike Mastrogianni et al., acidification was omitted to provide a more versatile extraction environment suitable for a broader variety of multiclass analytes [7, 14].
2.5. Instrumentation and Analysis
Ultra‐high‐performance liquid chromatography was performed with an Exion LC System (Waters Corporation, Milford, Massachusetts, USA) equipped with an Acquity C18 column (150 × 2.1 MM i.d., 1.7 mm) protected by an Acquity BEH C18 VanGuard precolumn (5 mm × 2.1 mm i.d., 1.7 mm). Mobile phase A was water, 0.1% formic acid, whereas mobile phase B was acetonitrile, 0.1% formic acid. A gradient elution program was applied as follows: at 0–0.2 min: 95% Phase A, 0.2–11 min: 40% Phase A, 11–12 min: 0% Phase A, 12–16 min: 95% Phase A. The flow rate was maintained at 0.4 mL/min, the injection volume at 5 μL, the autosampler temperature at 10°C, and the column temperature at 50°C. MS/MS detection was performed with the triple quadrupole SCIEX QTRAP 6500+ in electrospray ionization (ESI) switching between positive and negative polarity modes. Two MRM transitions were selected for each compound, one as a quantifier and one as a qualifier. Curtain gas (CUR) was set at 30 psi, collision gas (CAD) low, IonSpray Voltage (IS) at 3500 V, temperature 500°C, ion source gas 1 (GS1) at 60 psi, and ion source gas 2 (GS2) at 30 psi. The declustering potential (DP), collision energy (ce), and collision cell exit potential (CXP) were optimized for each transition as shown in Table S2. Data acquisition and analysis were controlled by the Analyst software, and quantitation was performed via the Skyline software. GC–MS screening of urine samples was performed by an established method of the lab [28].
2.6. Method Validation
The presented method was validated for the assessment of accuracy, precision, linearity, sensitivity, stability, matrix, and carryover effect in accordance with international guidelines for clinical and forensic toxicology [29, 30].
2.6.1. Linearity, Limits of Detection, and Quantification
As described previously, linearity was assessed based on spiked extracts of drug‐free pooled biological matrices, and an 8‐point calibration curve was constructed for compounds in cerumen and a 7‐point calibration curve for compounds in urine. In order to maintain acceptable analytical accuracy across a wide dynamic range, a segmented linear regression approach was utilized for certain substances, splitting the dynamic range into two distinct intervals from 0.1 to 10 μg/g and 10 to 25 μg/g for cerumen and the first calibrator (0.001, 0.01, 0.05, 0.1 μg/mL depending on the substance) to the fourth (0.1, 2, 5, 10 μg/mL) and the fourth to the seventh (10, 50, 75, 150 μg/mL) for urine. The limit of detection (LOD) was estimated to be three times the signal‐to‐noise ratio, and the limit of quantitation (LOQ) was 3.3 times the LOD.
2.6.2. Precision and Accuracy
Precision was estimated with five QC replicates at low, mid, and high concentration levels within a single day for intraday precision and over five consecutive days for interday precision and expressed as relative standard deviation (% RSD), whereas accuracy was expressed as a percentage of the nominal value at low, mid, and high concentrations.
2.6.3. Matrix and Carryover Effects
In order to assess whether endogenous compounds in cerumen or urine could interfere with the analyte signal, the matrix effect was evaluated, expressed as the response ratio of five blank replicates spiked after the extraction and five methanolic solutions of the same concentration. Blank samples were injected after the injection of cerumen and urine spikes at high concentration to study the carryover effect.
2.6.4. Stability
Postextraction quality control analytes at low, middle, and high concentrations were evaluated as to whether they are sufficiently stable for 48 h at the autosampler at 10°C and after being stored for 72 h at −10°C.
2.7. Data Analysis
Univariate statistics (frequencies, medians, and standard errors) were used to summarize the sample variables. In terms of bivariate analysis, the linear relationships between the urine and cerumen drug concentrations and (a) the total daily dose and (b) the elapsed time since the last administration, as well as the correlation between cerumen drug concentrations and urine drug concentrations, were performed by estimating the Pearson correlation coefficient. The available data on the history of administration were also separated into discrete time‐points, and they were compared with the qualitative results of cerumen.
The relationships between the aforementioned variables were also fitted to various regression models (linear, logarithmic, inverse, power, S, logistic, exponential). The same analysis was performed to examine the median urine and cerumen concentrations of each substance (estimated for substances with five detected cases or more in order to avoid outliers) in relation to various physicochemical and pharmacologic factors, such as (a) molecular weight, (b) partition coefficient (logP), (c) water solubility (logS), (d) acid–base dissociation constant (pKa), (e) time to peak drug concentration in plasma (tmax), (f) the fraction of the dose excreted unchanged in urine (%), (g) the volume of distribution (Vd), (h) total clearance (Cl), (i) bioavailability, (j) plasma protein binding (%), (k) time to steady‐state concentration, and (l) elimination half‐life of each substance. In each case, the best‐fitting model with the highest linearity (r 2) was selected. A Mann–Whitney U‐test was performed to compare cerumen and urine drug concentrations across patient sex. Data analysis was performed with IBM SPSS Statistics 27, and the significance level was set at 5%.
3. Results
3.1. Qualitative and Quantitative Findings in Cerumen and Urine
Based on the detections of the drugs in the paired cerumen and urine samples in the studied cases, it was found that the qualitative results in cerumen samples matched those in urine samples across most drug classes and cases. A summary of qualitative results is given in Figure 1 and Table S3, where the count and frequency (%) of administered substances in the patient history and of detected substances in cerumen and urine are given across all cases. In certain instances, an analyte with a positive history of administration was more likely to be detected in cerumen compared with urine, such as with, e.g., haloperidol: 106.7% and 60%, quetiapine: 88.9% and 55.6%, and nebivolol: 33.3% and 16.7% in cerumen and urine samples, respectively. There were, however, also substances which were more frequently detected in urine rather than in cerumen, e.g., biperiden: 57.1% and 82.1%, and pregabalin: 40% and 90% in cerumen and urine samples, respectively.
FIGURE 1.
Summary of qualitative results of (a) cerumen (analysis performed via LC–MS/MS), (b) urine (analysis performed via LC–MS/MS), and (c) urine (analysis performed via GC–MS) across the history of administration.
The developed method was applied to quantify the levels of detected drugs in the studied samples. Based on the validation data, the method enabled accurate and precise determinations of the drugs. In Tables S4–S7, the method validation and stability assessment data are provided. Intraday precision for the method for the determination of cerumen was below 20% in all cases, with the maximum RSD value at 15.83%. Interday precision was similar, with the highest value at 16.39%. For urine, the maximum RSD value was 14.79% for intraday precision and 15.52% for interday precision. The matrix effect for cerumen indicated that the signal of endogenous compounds did not interfere with that of the compounds of interest, ranging from 71.89% to 103.08%, nor was there a significant carryover effect. Accuracy parameters were acceptable, ranging between 75.96% and 121.24% for cerumen and 79.5%–119.71% for urine.
Regarding the levels of the concentrations found, the substances with the highest concentration in cerumen were sertraline (median concentration ± standard error; 228.781 ± 80.229 μg/g), lamotrigine (45.025 ± 36.5 μg/g), and paracetamol (22.803 ± 55.938 μg/g), whereas for urine they were pregabalin (182.321 ± 84.722 μg/mL), atenolol (26.776 ± 0.238 μg/mL), and lamotrigine (26.466 ± 5.077 μg/mL). Summaries of the quantitative results in cerumen and urine are given in Tables 1 and 2, respectively. Based on these, it was found that the urine and cerumen concentrations did not correlate for most of the determined drugs, with the exception of zolpidem (r(3) = 0.957, p = 0.011), with cerumen concentration increasing along with urine concentration in a logarithmic fashion (r = 0.972, p = 0.002).
TABLE 1.
Summary of cerumen quantitative results.
| Substance | N | Median concentration (μg/g) | Standard error |
|---|---|---|---|
| Bisoprolol | 1 | 0.898 | — |
| Chlorpromazine | 1 | 1.250 | — |
| Donepezil | 2 | 7.110 | 3.171 |
| Levomepromazine | 1 | 0.154 | 0.038 |
| Nebivolol | 2 | 0.159 | 0.136 |
| Lamotrigine | 3 | 45.025 | 36.500 |
| Aripiprazole | 5 | 10.728 | 1.945 |
| Mirtazapine | 3 | 4.514 | 2.076 |
| Amitryptiline | 1 | 79.576 | — |
| Sertraline | 5 | 228.781 | 80.229 |
| Citalopram | 12 | 5.497 | 5.205 |
| Haloperidol | 16 | 6.095 | 3.149 |
| Quetiapine | 16 | 1.916 | 0.878 |
| Paliperidone | 3 | 3.870 | 1.379 |
| Paracetamol | 8 | 22.803 | 55.938 |
| Bromazepam | 2 | 0.484 | 0.075 |
| Venlafaxine | 1 | 0.972 | — |
| Biperiden | 15 | 0.778 | 2.006 |
| Paroxetine | 5 | 4.060 | 0.976 |
| Risperidone | 6 | 0.158 | 0.167 |
| Olanzapine | 7 | 0.105 | 0.159 |
| Zolpidem | 4 | 0.158 | 0.350 |
| Pregabalin | 4 | 6.244 | 0.888 |
| Atenolol | 2 | 2.894 | 2.076 |
| Clozapine | 5 | 6.388 | 2.063 |
| Diazepam | 4 | 0.736 | 0.435 |
| Levetiracetam | 1 | 73.961 | — |
| Lorazepam | 8 | 0.786 | 1.736 |
| Metoprolol | 3 | 4.668 | 2.279 |
| Topiramate | 4 | 18.245 | 2.876 |
TABLE 2.
Summary of urine quantitative results.
| Substance | N | Median concentration (μg/mL) | Standard error |
|---|---|---|---|
| Bisoprolol | 1 | 1.035 | — |
| Buspirone | 1 | 0.393 | — |
| Donepezil | 2 | 1.53 | 0.823 |
| Levomepromazine | 4 | 0.019 | 0.018 |
| Lamotrigine | 4 | 26.045 | 2.539 |
| Aripiprazole | 4 | 0.073 | 0.390 |
| Mirtazapine | 4 | 7.646 | 4.987 |
| Amitriptyline | 1 | 0.047 | — |
| Sertraline | 4 | 0.326 | 0.106 |
| Citalopram | 13 | 1.624 | 0.764 |
| Haloperidol | 7 | 0.034 | 0.198 |
| Quetiapine | 10 | 0.759 | 0.371 |
| Paliperidone | 7 | 1.851 | 0.758 |
| Paracetamol | 4 | 12.029 | 7.607 |
| Bromazepam | 3 | 0.066 | 0.002 |
| Venlafaxine | 1 | 18.049 | — |
| Biperiden | 19 | 0.037 | 0.146 |
| Paroxetine | 5 | 0.580 | 0.190 |
| Risperidone | 16 | 0.248 | 0.741 |
| Olanzapine | 14 | 0.512 | 1.139 |
| Zolpidem | 5 | 0.005 | 0.007 |
| Pregabalin | 9 | 158.513 | 28.241 |
| Atenolol | 2 | 26.776 | 0.168 |
| Clozapine | 5 | 4.391 | 1.845 |
| Diazepam | 5 | 0.005 | 0.007 |
| Levetiracetam | 1 | 1.405 | — |
| Lorazepam | 7 | 0.029 | 0.010 |
| Metoprolol | 4 | 9.131 | 3.357 |
| Topiramate | 4 | 79.165 | 13.457 |
3.2. Drug Detection in Relation to Patient Administration History and Sex
No clear pattern emerged from the data on the history of administration, separated into discrete time points, after they were compared to the qualitative results of cerumen. A detailed summary is given in Table S8. However, among cases without recent administration (> 72 h prior to collection), paracetamol was positive in both cerumen and urine in one case (Case 60, 27 h since administration). Paliperidone was positive in urine but not cerumen (Case 51, 27 h). Lorazepam was not positive in either of the specimens in one case (Case 46, 107 h) and positive only in cerumen in another (Case 49, 155 h). Olanzapine was positive only in cerumen in one case (Case 40, 576 h). Biperiden (Case 43, 624 h) and haloperidol (Case 43, 648 h) were not positive in either specimen. Quetiapine was positive only in cerumen in two cases (Case 43, 495 h; Case 57, 543 h), but in none of them positive in another case (Case 52, 181 h). Mirtazapine was positive only in cerumen in two cases (Case 40, 832 h; Case 58, 445 h) but positive in both urine and cerumen in another case (Case 52, 85 h). Among cases with only very recent administration (< 8 h prior to collection) but no prior history of administration, sertraline was positive in cerumen but not in urine in one case (Case 45, 4 h), whereas paliperidone was positive in urine but not in cerumen in another (Case 41, 4 h).
No statistically significant associations were found between patient sex and cerumen concentration of any of the substances (Mann–Whitney U‐test, p > 0.05).
3.3. Concentration Relationships With Pharmacologic and Physicochemical Parameters of Drugs
The median urine concentration was negatively correlated with molecular weight (r(17) = −0.521, p = 0.022) and fraction of protein‐bound drug (r(17) = −0.712, p < 0.001) and positively correlated with water solubility (r(17) = 0.520, p = 0.023) and fraction of drug excreted unchanged in urine (r(17) = 0.852, p < 0.001). Urine concentration decreased with the increase of molecular weight (r 2 = 0.379, p = 0.019), volume of distribution (r 2 = 0.378, p = 0.019), and time to reach steady‐state concentration (r 2 = 0.349, p = 0.043), with the inverse model having the best fit in each case. A linear decrease was also observed with increasing fraction of the drug bound by plasma proteins (r 2 = 0.511, p = 0.004). Urine concentration increased along with water solubility, with the inverse model providing the best fit (r 2 = 0.663, p < 0.001), and also demonstrated a linear increase with the fraction of the dose excreted unchanged in urine (r 2 = 0.736, p < 0.001).
Cerumen concentration appeared to decrease in an inverse fashion with increasing molecular weight (r 2 = 0.510, p < 0.001), partition coefficient (r 2 = 0.978, p = 0.015), and fraction of the drug bound by plasma proteins (r 2 = 0.673, p < 0.001), whereas it increased with increasing water solubility (r 2 = 0.400, p = 0.015). The ratio between median urine concentration and cerumen concentration increased with total clearance (L/h) in an exponential fashion (r 2 = 0.555, p = 0.034). The statistically significant regression models between median urine concentrations, median cerumen concentrations, and physicochemical and pharmacologic parameters are given in Table 3.
TABLE 3.
Statistically significant regression models between (a) median urine concentration, (b) median cerumen concentration of analytes with n ≥ 5 positive results, and physicochemical and pharmacologic parameters.
| Dependent variable | Independent variable | Model | r 2 | df1 | df2 | Significance | Equation |
|---|---|---|---|---|---|---|---|
| Median urine concentration | Molecular weight | Linear | 0.316 | 1 | 12 | 0.036 | y = −0.298x + 106.151 |
| Logarithmic | 0.361 | 1 | 12 | 0.023 | y = −82.162ln(x) + 4181.814 | ||
| Inverse | 0.379 | 1 | 12 | 0.019 | y = (19,804.898/x) − 56.439 | ||
| logS | Linear | 0.408 | 1 | 12 | 0.014 | y = 24.548x + 100.999 | |
| Inverse | 0.663 | 1 | 12 | < 0.001 | y = (−190.372/x) − 49.527 | ||
| Exponential | 0.337 | 1 | 12 | 0.030 | y = 62.816e 1.499x | ||
| Logistic | 0.337 | 1 | 12 | 0.030 | y = 1/(0.016 * 0.223 x ) | ||
| S | 0.374 | 1 | 12 | 0.020 | y = e (−9.625/x) − 4.405 | ||
| Dose unchanged in urine (%) | Linear | 0.736 | 1 | 12 | < 0.001 | y = 1.303x − 9.534 | |
| Exponential | 0.700 | 1 | 12 | < 0.001 | y = 0.067e 0.085x | ||
| Logistic | 0.700 | 1 | 12 | < 0.001 | y = 1/(14.996 * 0.918 x ) | ||
| Volume of distribution | Inverse | 0.378 | 1 | 12 | 0.019 | y = (39.414/x) − 10.727 | |
| Protein binding (%) | Linear | 0.511 | 1 | 12 | 0.004 | y = −1x + 87.331 | |
| Steady state (days) | Inverse | 0.349 | 1 | 10 | 0.043 | y = (134.805/x) − 25.12 | |
| Median cerumen concentration | Molecular weight | Logarithmic | 0.370 | 1 | 12 | 0.021 | y = −14.242ln(x) + 86.730 |
| Inverse | 0.510 | < 0.001 | y = (4134.805/x) − 8.859 | ||||
| logP | Logarithmic | 0.849 | 1 | 12 | < 0.001 | y = −6.657ln(x) + 11.195 | |
| Inverse | 0.978 | < 0.001 | y = (9.840/x) − 0.152 | ||||
| logS | Inverse | 0.400 | 1 | 12 | 0.015 | y = (−35.156/x) − 5.958 | |
| Protein binding | Linear | 0.402 | 1 | 12 | 0.015 | y = −0.187x + 19.552 | |
| Logarithmic | 0.570 | 0.002 | y = −10.565ln(x) + 50.132 | ||||
| Inverse | 0.673 | < 0.001 | y = (416.184/x) − 1.935 | ||||
| Median urine to cerumen concentration ratio | Total clearance (L/h) | Power | 0.502 | 1 | 6 | 0.049 | y = 1657.835 (x −2.068) |
| Exponential | 0.555 | 0.034 | y = 4.427e −0.031x | ||||
| Logistic | 0.555 | 0.034 | y = 1/(0.226 * 1.031 x ) |
3.4. Cerumen and Urine Concentration Relationships With Dose and Time of Administration
In most cases, there were no straightforward relationships between drug concentrations in urine or cerumen and total daily dose or time since last administration, or between cerumen and urine drug concentrations. The concentration of aripiprazole in urine seemed to increase with increasing total daily dose in a linear fashion (r 2 = 0.999, p < 0.001) as was the case with pregabalin with a power model (r 2 = 0.474, p = 0.040), diazepam concentration in urine increasing exponentially (r 2 = 0.827, p = 14.307) and metoprolol with the S‐curve model having the best fit (r 2 = 0.929, p = 0.036). The urine concentrations of bromazepam and metoprolol decreased as more time passed since the last administration in a logarithmic (r 2 = 1.000, p = 0.001) and exponential fashion (r 2 = 0.955, p = 0.023), respectively.
The cerumen concentrations of citalopram (r 2 = 0.335, p = 0.038) and olanzapine (r 2 = 0.270, p = 0.047) decreased with increasing elapsed time since last administration in an inverse fashion. Levomepromazine (r 2 = 0.997, p = 0.034) and quetiapine (r 2 = 0.776, p < 0.001) increased linearly with total daily dose, whereas olanzapine did so following a power equation (r 2 = 0.320, p < 0.028). The statistically significant regression models between urine concentrations, cerumen concentrations, total daily dose, and time since last administration are given in Table 4.
TABLE 4.
Statistically significant regression models between (a) urine concentrations, (b) cerumen concentrations, and total daily dose, as well as time since last administration.
| Dependent variable | Independent variable | Model | r 2 | df1 | df2 | Significance | Equation |
|---|---|---|---|---|---|---|---|
| Aripiprazole urine concentration | Total daily dose (mg) | Linear | 0.999 | 1 | 2 | < 0.001 | y = (0.104x) × 1.501 |
| Logarithmic | 0.999 | < 0.001 | y = (2.25ln(x)) × 6.035 | ||||
| Inverse | 0.999 | < 0.001 | y = (−46.793/x) + 3.178 | ||||
| Pregabalin urine concentration | Total daily dose (mg) | Power | 0.474 | 1 | 7 | 0.040 | y = 29.726 (x 0.326) |
| Diazepam urine concentration | Total daily dose (mg) | Exponential | 0.827 | 1 | 3 | 0.032 | y = 0.003 (e 0.081x ) |
| Logistic | 0.827 | 0.032 | y = 1/(398.301 * 0.922 x ) | ||||
| Metoprolol urine concentration | Total daily dose (mg) | S | 0.929 | 1 | 2 | 0.036 | y = e (−177.140/x) + 5.397 |
| Time since last administration (hours) | Power | 0.955 | 1 | 2 | 0.023 | y = 110.132 (x −2.186) | |
| S | 0.955 | 0.023 | y = e (14.635/x) − 2.578 | ||||
| Exponential | 0.955 | 0.023 | y = 20.743 (e −0.244x ) | ||||
| Logistic | 0.955 | 0.023 | y = 1/(0.048 * 1.276 x ) | ||||
| Levomepromazine cerumen concentration | Total daily dose (mg) | Linear | 0.997 | 1 | 1 | 0.034 | y = (0.002x) + 0.016 |
| Logarithmic | 0.997 | 0.034 | y = (0.064ln(x)) − 0.122 | ||||
| Inverse | 0.997 | 0.034 | y = (−1.714/x) + 0.176 | ||||
| Citalopram cerumen concentration | Time since last administration (hours) | Inverse | 0.335 | 1 | 11 | 0.038 | y = (101.255/x) − 27.802 |
| Quetiapine cerumen concentration | Total daily dose (mg) | Linear | 0.776 | 1 | 12 | < 0.001 | y = (0.012x) − 2.670 |
| Logarithmic | 0.439 | 0.010 | y = (2.950ln(x)) − 14.781 | ||||
| Power | 0.527 | 0.003 | y = 3.522e −5 (x 1.730) | ||||
| Exponential | 0.655 | < 0.001 | y = 0.072 (e 0.006x ) | ||||
| Logistic | 0.655 | < 0.001 | y = 1/(13.945 * 0.994 x ) | ||||
| Olanzapine cerumen concentration | Total daily dose (mg) | Power | 0.320 | 1 | 13 | 0.028 | y = 0.006 (x 1.064) |
| S | 0.268 | 0.048 | y = e (−3.255/x) − 2.269 | ||||
| Exponential | 0.271 | 0.047 | y = 0.012 (e 0.140x ) | ||||
| Logistic | 0.271 | 0.047 | y = 1/(83.006 * 0.870 x ) | ||||
| Time since last administration (hours) | Inverse | 0.270 | 1 | 13 | 0.047 | y = (0.763/x) + 0.045 | |
| Power | 0.264 | 0.049 | y = 0.218 (x −0.556) | ||||
| Zolpidem cerumen concentration | Urine concentration | Linear | 0.916 | 1 | 3 | 0.011 | y = (3.630x) + 0.037 |
| Logarithmic | 0.972 | 0.002 | y = (0.049ln(x)) + 0.329 | ||||
| Power | 0.911 | 0.012 | y = 1.515 (x 0.621) | ||||
| S | 0.900 | 0.014 | y = e (−0.003/x) − 1.944 |
4. Discussion
A UHPLC–MS/MS method for the analysis of cerumen and urine samples was established for 33 medications and was subsequently successfully applied for the detection and quantification of psychoactive medications in paired cerumen and urine samples from 60 psychiatric inpatient volunteers who had been prescribed and were administered these medications.
The timeframe and mechanism of the excretion of drugs into the cerumen are not well elucidated in the literature. The secretory materials of ceruminous glands are either contained in secretory vessels and/or dissolved in the cell cytoplasm. With their gradual accumulation, apical protrusions form until partial loss of the cell cytoplasm occurs (apocrine secretion). A concomitant release via exocytosis may also occur [2, 31]. Apocrine secretion is thought to be slow and continuous [32, 33], requiring hours to a few days for the tubular reservoir to be filled [32]. The lipid‐rich secretory materials of sebaceous glands accumulate into droplets until the cells become distended and disintegrate (holocrine secretion) [3, 6, 34]. The disintegration has a turnover of about 2 weeks, and the flow of sebum is relatively continuous [34], though cocaine has been detected in sebum within 24–48 h after exposure [35]. The rate of epithelial migration of keratinized desquamated epithelial cells from the tympanic membrane is about 0.05 mm per day [36], and the process may take weeks [37]. 4‐Fluoroamphetamine has been detected in cerumen as soon as 1 h after intake in the majority of the subjects of a controlled oral exposure study and could be detected in all subjects after 12 h [8]. Δ9‐THC has been detected for up to 15 days after a single exposure, even when it is not detectable in urine. In that study, the peak concentration of Δ9‐THC, CBG, and CBN in cerumen was detected on the first day postexposure and then declined markedly [10]. Quinolizidine alkaloids of the plant Lupinus leucophyllus in the cerumen of cattle peaked in the first available time point (48 h) after a single exposure and then remained constant for up to a month [38]. Similar results were found in another study in cattle exposed to alkaloids of larkspurs over a shorter timeframe [39]. In general, cerumen seems to have a long timeframe, allowing detection with a single exposure even after no detection can be achieved in urine or saliva.
In our study, no clear relationships could be established between the interval since last administration and total daily dose with cerumen concentration, with the exceptions of citalopram and olanzapine for time since last administration and levomepromazine, quetiapine, and olanzapine for total dose. This is to be expected because, as opposed to a controlled dose study, routine clinical conditions did not allow for a washout period or a set period for patients to clean their ears before collection. Therefore, the accumulation of cerumen mass was highly variable between patients. Shokry et al. have also attributed the high intrasubject variance of cerumen concentrations to interindividual variability [7]. Nevertheless, we managed to detect certain substances in cerumen, even after 1 week since their administration was discontinued, and in one case, even after a month had passed. In one case, the initial administration of a drug, 4 h prior to collection, could already be detected in cerumen. Moreover, the qualitative results of cerumen and urine were consistently compatible. Therefore, the utility of cerumen in a clinical setting would be for the identification of acute, subacute, or even chronic exposure to a substance. To our knowledge, this is the first study conducted in a clinical setting with routine collection of cerumen.
The lack of association between patient sex and cerumen drug concentration was consistent with a study that investigated the concentration of tramadol in cerumen [11]. This is despite the fact that the lipid composition of cerumen is affected by sex, as well as the fact that the rate of sebum production is under androgen control [6, 40, 41].
We found that substances with low molecular weight and a low fraction of drug bound by plasma proteins had higher median cerumen concentrations. Since ceruminous glands are modified sweat glands, and the excretion of drugs into eccrine sweat occurs with the process of passive diffusion as the primary transport mechanism [42, 43, 44], and since passive diffusion into the dermis‐epidermis is one of the mechanisms by which a drug may arrive at the epidermis from systemic circulation, this seems to be consistent with our findings [45]. Molecules with higher free fraction and lower molecular mass may cross more readily. Both molecular weight and protein binding have been indicated as some of the factors influencing the excretion of drugs into sweat [46].
What is more surprising is the association of higher median cerumen concentration with lower partition coefficient (logP) along with higher water solubility (logS), as lipophilic drugs cross cell membranes more readily and are retained in adipose tissue, leading to a more gradual excretion. However, as Ruwe et al. have noted, there is a relatively small window of partition coefficient and charge state that enables partitioning of drugs from plasma into sweat [47]. It could be the case that, as with drugs with high volume distribution, high sequestration of drugs in tissues rather than plasma is inefficient for partition into cerumen. A review by Patzelt et al. that investigated the inverse penetration of systemically administered substances into the skin noted that although lipophilic substances reach the skin surface via the sebum, hydrophilic substances use sweat to reach the skin surface, with higher concentrations of hydrophilic drugs generally found in sweat compared with sebum. This points to sweat glands as potential inverse penetration pathways. As cerumen is composed of 60% of desquamated skin and the secretions of ceruminous glands comprise most of the cerumen as opposed to sebum, this is an important consideration. Furthermore, other pathways may exist for the transit of drugs of low molecular weight, even hydrophilic ones, from plasma to the skin surface, such as an intracellular penetration pathway or through the process of desquamation [25]. Moreover, hydrophilic drugs have also been found to accumulate in skin and the underlying adipose tissue, which cannot be merely explained by the presence of penetrating vessels or sweat within sweat glands, as sweat glands do not present in adipose tissue [35].
As cerumen is acidic, it might also be expected that basic drugs would preferably accumulate in cerumen, as is the case with eccrine sweat [42, 44]; however, no correlation was found between a drug's pKa and its cerumen concentration. As for the relationship between cerumen concentration and a shorter time to steady state concentration, it has been noted that sweat patch tests for drugs with urine detection times of 2–3 days have shown positive results, but that is not the case for drugs with a longer elimination half‐life such as benzodiazepines or cannabinoids [48].
The findings with regard to the relationships found between urine concentration and physicochemical and pharmacologic properties were generally expected, as low molecular weight hydrophilic drugs are excreted more readily in urine as their parent compounds, whereas drugs with high lipophilicity and low water solubility undergo metabolism before being excreted as hydrophilic low molecular weight metabolites [49].
The disposition of drugs in cerumen is not extensively studied, nor is that of its constituent components. Therefore, future studies investigating the pathways by which drugs are excreted in cerumen, apocrine sweat glands, and skin, as well as the effect of the physicochemical and pharmacological factors of drugs, are therefore necessary for clear conclusions to be drawn.
5. Conclusions
A UHPLC–MS/MS method for the determination of 33 analytes in cerumen was developed, validated, and applied to samples of psychiatric patients who had been administered these substances. The results indicated good consistency in the qualitative results between cerumen and urine and suggest that cerumen may be utilized for the detection of drugs that have been administered both recently and in the long term. No clear pattern emerged between cerumen concentrations and the interval of last administration or total dose, indicating challenges with routine clinical testing. High median cerumen concentrations were associated with lower molecular weight, lipophilicity, fraction of drug bound by plasma proteins, and higher water solubility, reflecting a complex mechanism of drug excretion. Thus, further studies are required to elucidate these processes further and support the practical utility of this specimen. This work presents important aspects for the utilization of cerumen as an innovative specimen for drug screening and the interpretation of analytical results in cerumen.
Funding
The authors have nothing to report.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Table S1: Site of sample collection and net weight (g) of cerumen cotton swabs.
Table S2: Analysis parameters for multiple reaction monitoring. Retention time, precursor and product ions, declustering potential, collision energy, collision cell exit potential, and ionization mode.
Table S3: Summary of (a) history of administration and qualitative results of (b) cerumen, (c) urine (analysis performed via LC–MS/MS), and (d) urine (analysis performed via GC–MS) across cases.
Table S4: Linearity, regression formulas, limits of detection, and quantitation for cerumen.
Table S5: Linearity, regression formulas, limits of detection, and quantitation for urine.
Table S6: Intraday and interday precision, matrix effects, and accuracy parameters for cerumen.
Table S7: Intraday and interday precision and accuracy parameters for urine.
Table S8: Summary of positive cerumen samples across the history of administration in discrete time intervals (%).
Acknowledgments
The publication of this article in OA mode was financially supported by HEAL‐Link.
Contributor Information
Dimitrios Phaedon Kevrekidis, Email: dphkevrekidis@gmail.com.
Helen G. Gika, Email: gkikae@auth.gr.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Bojrab D. I., Bruderly T., and Abdulrazzak Y., “Otitis Externa,” Otolaryngologic Clinics of North America 29 (1996): 761–782. [PubMed] [Google Scholar]
- 2. Wang J., He A., Zhu Y., Zhou G., and Zhang T., “Isolation, Culture, and Identification of Ceruminous Gland Cells,” Journal of Molecular Histology 53 (2022): 227–237. [DOI] [PubMed] [Google Scholar]
- 3. Thiboutot D., “Regulation of Human Sebaceous Glands,” Journal of Investigative Dermatology 123 (2004): 1–12. [DOI] [PubMed] [Google Scholar]
- 4. Adhikari T. R. and Jamtsho S., “A Rare Case of Abnormal Epithelial Migration in the External Auditory Canal Secondary to Cotton Bud Abuse,” SAGE Open Medical Case Reports 11 (2023): 2050313X231183584. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Guest J. F., Greener M. J., Robinson A. C., and Smith A. F., “Impacted Cerumen: Composition, Production, Epidemiology and Management,” QJM: An International Journal of Medicine 97 (2004): 477–488. [DOI] [PubMed] [Google Scholar]
- 6. Baker L. B., “Physiology of Sweat Gland Function: The Roles of Sweating and Sweat Composition in Human Health,” Temperature 6 (2019): 211–259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Shokry E., Marques J. G., Ragazzo P. C., Pereira N. Z., and Filho N. R. A., “Earwax as an Alternative Specimen for Forensic Analysis,” Forensic Toxicology 35 (2017): 348–358. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Meier S. I., Petzel‐Witt S., Schubert‐Zsilavecz M., et al., “Analysis of 4‐Fluoroamphetamine in Cerumen After Controlled Oral Application,” Drug Testing and Analysis 12 (2020): 968–974. [DOI] [PubMed] [Google Scholar]
- 9. Nicolaou A. G., Stavrou I. J., Louppis A. P., Constantinou M. S., and Kapnissi‐Christodoulou C., “Application of an Ultra‐Performance Liquid Chromatography‐Tandem Mass Spectrometric Method for the Detection and Quantification of Cannabis in Cerumen Samples,” Journal of Chromatography. A 1642 (2021): 462035. [DOI] [PubMed] [Google Scholar]
- 10. Christodoulou M. C., Constantinou M. S., Louppis A. P., Christou A., Stavrou I. J., and Kapnissi‐Christodoulou C. P., “Determination of Cannabinoids in Human Cerumen Samples by Use of UPLC‐MS/MS as a Potential Biomarker for Drug Use,” Journal of Pharmaceutical and Biomedical Analysis 231 (2023): 115412. [DOI] [PubMed] [Google Scholar]
- 11. Massadeh A. M., Al‐Rawi N. H., Fayyad M. T., Shotar A. M., Al‐Farras M. I., and Massadeh S. A. A., “Simultaneous Determination of Tramadol in Earwax and Urine Samples: Effects of Age, Duration Time and Sex,” Bioanalysis 16 (2024): 363–374. [DOI] [PubMed] [Google Scholar]
- 12. Meier S. I., Koelzer S. C., Schubert‐Zsilavecz M., and Toennes S. W., “Analysis of Drugs of Abuse in Cerumen—Correlation of Postmortem Analysis Results With Those for Blood, Urine and Hair,” Drug Testing and Analysis 9 (2017): 1572–1585. [DOI] [PubMed] [Google Scholar]
- 13. Mastrogianni O., Kevrekidis D. P., Brousa E., Orfanidis A., Zagelidou H., and Raikos N., “Determination of Fentanyl and Norfentanyl in Cerumen in the Setting of Postmortem Investigation,” Journal of Forensic Sciences 67 (2022): 2130–2137. [DOI] [PubMed] [Google Scholar]
- 14. Mastrogianni O., Orfanidis A., Brousa E., Kevrekidis D.‐P., Zagelidou H., and Raikos N., “Development and Validation of an Ultra‐Performance Liquid Chromatography‐Tandem Mass Spectrometric Method for the Determination of 25 Psychoactive Drugs in Cerumen and Its Application to Real Postmortem Samples,” Forensic Toxicology 41 (2023): 94–104. [DOI] [PubMed] [Google Scholar]
- 15. Wang X. Q., Gao P. Y., Lin Y. Z., and Chen C. M., “Studies on Hexachlorocyclohexane and DDT Contents in Human Cerumen and Their Relationships to Cancer Mortality,” Biomed. Environ. Sci. BES 1 (1988): 138–151. [PubMed] [Google Scholar]
- 16. J. To‐Figueras , Barrot C., Rodamilans M., et al., “Accumulation of Hexachlorobenzene in Humans: A Long Standing Risk,” Human & Experimental Toxicology 14 (1995): 20–23. [DOI] [PubMed] [Google Scholar]
- 17. Geusau A., Tschachler E., Meixner M., Päpke O., Stingl G., and McLachlan M., “Cutaneous Elimination of 2,3,7,8‐Tetrachlorodibenzo‐p‐Dioxin,” British Journal of Dermatology 145 (2001): 938–943. [DOI] [PubMed] [Google Scholar]
- 18. Krishnan U. and Que Hee S. S., “Ear Wax: A New Biological Monitoring Medium for Metals?,” Bulletin of Environmental Contamination and Toxicology 48 (1992): 481–486. [DOI] [PubMed] [Google Scholar]
- 19. Janssens T., Brouwers E. E. M., de Vos J. P., Schellens J. H. M., and Beijnen J. H., “Determination of Platinum Originating From Carboplatin in Canine Sebum and Cerumen by Inductively Coupled Plasma Mass Spectrometry,” Journal of Pharmaceutical and Biomedical Analysis 54 (2011): 395–400. [DOI] [PubMed] [Google Scholar]
- 20. Shokry E., de Oliveira A. E., Avelino M. A. G., de Deus M. M., Pereira N. Z., and Filho N. R. A., “Earwax: An Innovative Tool for Assessment of Tobacco Use or Exposure. A Pilot Study in Young Adults,” Forensic Toxicology 35 (2017): 389–398. [Google Scholar]
- 21. Shichijo S., Masuda H., and Takeuchi M., “Carbohydrate Composition of Glycopeptides From the Human Cerumen,” Biochemical Medicine 22 (1979): 256–263. [DOI] [PubMed] [Google Scholar]
- 22. Balabanova S., Schneider E., Bühler G., and Krause H., “The Presence of Methadone in Apocrine Sweat in Humans,” Z. Für Rechtsmed 103 (1990): 323–327. [DOI] [PubMed] [Google Scholar]
- 23. Balabanova S. and Schneider E., “Detection of Drugs in Sweat,” Beiträge zur Gerichtlichen Medizin 48 (1990): 45–49. [PubMed] [Google Scholar]
- 24. Wilke K., Martin A., Terstegen L., and Biel S. S., “A Short History of Sweat Gland Biology,” International Journal of Cosmetic Science 29 (2007): 169–179. [DOI] [PubMed] [Google Scholar]
- 25. Patzelt A., Antoniou C., Sterry W., and Lademann J., “Skin Penetration From the Inside to the Outside: A Review,” Drug Discov. Today Dis. Mech 5 (2008): e229–e235. [Google Scholar]
- 26. Henderson G. L., “Mechanisms of Drug Incorporation Into Hair,” Forensic Science International 63 (1993): 19–29. [DOI] [PubMed] [Google Scholar]
- 27. Levisky J. A., Bowerman D. L., Jenkins W. W., and Karch S. B., “Drug Deposition in Adipose Tissue and Skin: Evidence for an Alternative Source of Positive Sweat Patch Tests,” Forensic Science International 110 (2000): 35–46. [DOI] [PubMed] [Google Scholar]
- 28. Orfanidis A., Krokos A., Mastrogianni O., Gika H., Raikos N., and Theodoridis G., “Development and Validation of a Single Step GC/MS Method for the Determination of 41 Drugs and Drugs of Abuse in Postmortem Blood,” Forensic Science 2 (2022): 473–491. [Google Scholar]
- 29. Peters F. T., Drummer O. H., and Musshoff F., “Validation of New Methods,” Forensic Science International 165 (2007): 216–224. [DOI] [PubMed] [Google Scholar]
- 30. Scientific Working Group for Forensic Toxicology , “Scientific Working Group for Forensic Toxicology (SWGTOX) Standard Practices for Method Validation in Forensic Toxicology,” Journal of Analytical Toxicology 37 (2013): 452–474. [DOI] [PubMed] [Google Scholar]
- 31. Gesase A. P. and Satoh Y., “Apocrine Secretory Mechanism: Recent Findings and Unresolved Problems,” Histology and Histopathology 18 (2003): 597–608. [DOI] [PubMed] [Google Scholar]
- 32. Hurley H. J. and Witkowski J. A., “The Merocrine Component of Apocrine Secretion.*,” Proceedings of the Society for Experimental Biology and Medicine 113 (1963): 578–580. [DOI] [PubMed] [Google Scholar]
- 33. Nagarajan P., “Ceruminous Neoplasms of the Ear,” Head and Neck Pathology 12 (2018): 350–361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Braun‐Falco O., Plewig G., Wolff H. H., and Burgdorf W. H. C., “Diseases of the Sebaceous Glands,” in Dermatology, eds. Braun‐Falco O., Plewig G., Wolff H. H., and Burgdorf W. H. C. (Springer, 2000), 1051–1081. [Google Scholar]
- 35. Yang W., Barnes A. J., Ripple M. G., et al., “Simultaneous Quantification of Methamphetamine, Cocaine, Codeine, and Metabolites in Skin by Positive Chemical Ionization Gas Chromatography‐Mass Spectrometry,” Journal of Chromatography. B, Analytical Technologies in the Biomedical and Life Sciences 833 (2006): 210–218. [DOI] [PubMed] [Google Scholar]
- 36. Litton W. B., “Epithelial Migration Over Tympanic Membrane and External Canal,” Arch. Otolaryngol. Chic. Ill 1960 77 (1963): 254–257. [DOI] [PubMed] [Google Scholar]
- 37. Makino K. and Amatsu M., “Epithelial Migration on the Tympanic Membrane and External Canal,” Archives of Oto‐Rhino‐Laryngology 243 (1986): 39–42. [DOI] [PubMed] [Google Scholar]
- 38. Lee S. T., Stonecipher C. A., Dos Santos F. C., et al., “An Evaluation of Hair, Oral Fluid, Earwax, and Nasal Mucus as Noninvasive Specimens to Determine Livestock Exposure to Teratogenic Lupine Species,” Journal of Agricultural and Food Chemistry 67 (2019): 43–49. [DOI] [PubMed] [Google Scholar]
- 39. Stonecipher C. A., Lee S. T., Green B. T., et al., “Evaluation of Noninvasive Specimens to Diagnose Livestock Exposure to Toxic Larkspur (Delphinium spp.),” Toxicon: An International Journal Devoted to the Biological and Clinical Aspects of Venoms, Poisons of Animal, Plant and Microbial Origin 161 (2019): 33–39. [DOI] [PubMed] [Google Scholar]
- 40. Cipriani C., Taborelli G., Gaddia G., Melagrana A., and Rebora A., “Production Rate and Composition of Cerumen: Influence of Sex and Season,” Laryngoscope 100 (1990): 275–276. [DOI] [PubMed] [Google Scholar]
- 41. Koçer M., Güldür T., Akarçay M., Miman M. C., and Beker G., “Investigation of Age, Sex and Menstrual Stage Variation in Human Cerumen Lipid Composition by High Performance Thin Layer Chromatography,” Journal of Laryngology and Otology 122 (2008): 881–886. [DOI] [PubMed] [Google Scholar]
- 42. Johnson H. L. and Maibach H. I., “Drug Excretion in Human Eccrine Sweat,” Journal of Investigative Dermatology 56 (1971): 182–188. [DOI] [PubMed] [Google Scholar]
- 43. Rouen D., Dolan K., and Kimber J., A Review of Drug Detection Testing and Examination of Urine, Hair, Saliva and Sweat (National Drug and Alcohol Research Centre, University of New South Wales, 2001). [Google Scholar]
- 44. Jadoon S., Karim S., Akram M. R., et al., “Recent Developments in Sweat Analysis and Its Applications,” International Journal of Analytical Chemistry 2015 (2015): 164974. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Faergemann J., Zehender H., Jones T., and Maibach I., “Terbinafine Levels in Serum, Stratum Corneum, Dermis‐Epidermis (Without Stratum Corneum), Hair, Sebum and Eccrine Sweat,” Acta Dermato‐Venereologica 71 (1991): 322–326. [PubMed] [Google Scholar]
- 46. Raju K. S., Taneja I., Singh S. P., and Wahajuddin M., “Utility of Noninvasive Biomatrices in Pharmacokinetic Studies,” Biomed. Chromatogr. BMC 27 (2013): 1354–1366. [DOI] [PubMed] [Google Scholar]
- 47. Ruwe T., White E., Zebertavage A. S., et al., “Diverse Drug Classes Partition Into Human Sweat: Implications for Both Sweat Fundamentals and for Therapeutic Drug Monitoring,” Therapeutic Drug Monitoring 45 (2023): 731–742. [DOI] [PubMed] [Google Scholar]
- 48. Kintz P., Tracqui A., Mangin P., and Edel Y., “Sweat Testing in Opioid Users With a Sweat Patch,” Journal of Analytical Toxicology 20 (1996): 393–397. [DOI] [PubMed] [Google Scholar]
- 49. Loftsson T., “Basic Concepts of Pharmacokinetics,” in Essent. Pharmacokinet, ed. Loftsson T. (Academic Press, 2015), 9–84. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1: Site of sample collection and net weight (g) of cerumen cotton swabs.
Table S2: Analysis parameters for multiple reaction monitoring. Retention time, precursor and product ions, declustering potential, collision energy, collision cell exit potential, and ionization mode.
Table S3: Summary of (a) history of administration and qualitative results of (b) cerumen, (c) urine (analysis performed via LC–MS/MS), and (d) urine (analysis performed via GC–MS) across cases.
Table S4: Linearity, regression formulas, limits of detection, and quantitation for cerumen.
Table S5: Linearity, regression formulas, limits of detection, and quantitation for urine.
Table S6: Intraday and interday precision, matrix effects, and accuracy parameters for cerumen.
Table S7: Intraday and interday precision and accuracy parameters for urine.
Table S8: Summary of positive cerumen samples across the history of administration in discrete time intervals (%).
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.