Distribution profiles of diphenhydramine and lidocaine in scalp, axillary, and pubic hairs measured by micro-segmental hair analysis: good indicator for discrimination between administration and external contamination of the drugs

OTC pharmaceutical products, Feminina^® (an anti-itch ointment containing DP and LD), Wutt^® (a sedative-hypnotic medicine containing DP), and Stac^® (a common cold medicine containing chlorpheniramine (CP) and methylephedrine (ME)) were purchased from Kobayashi Pharmaceutical Co., Ltd. (Osaka, Japan), Itami Pharmaceutical Co., Ltd. (Takashima, Japan), and SSP Co., Ltd. (Tokyo, Japan), respectively. Reference standards, LD and desethyllidocaine (DLD) were purchased from Sigma-Aldrich Co. LLC (St. Louis, MO, USA). CP-d₆ maleate was purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). DP hydrochloride, desmethyldiphenhydramine (DDP) hydrochloride, and the other reagents were purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). Acetonitrile, methanol, and water were of liquid chromatography/mass spectrometry grade.

Participant A was an Asian man aged in his 40s with black scalp hair, and participant B was an Asian woman in her 30s with black scalp hair that was dyed brown. Both participants applied 1.0 g of Feminina^® ointment containing DP hydrochloride (20 mg) and LD (20 mg) to their axillary or pubic areas. The skin and hair in/around the areas where the ointment was applied were washed about 24 h later. One Wutt^® tablet containing DP hydrochloride (8.3 mg) was orally administered as a single dose to both participants. Both participants had ingested Stac^® common cold medicine containing CP (2.5 mg) and ME (20 mg) as a single dose with intervals of over 2 weeks as internal temporal markers, which marks a timescale in a hair strand [23, 24]. Both had received gingival injection of LD (6–36 mg) for local anesthesia during dental treatments.

A few strands of scalp hair were plucked from the posterior vertex regions of the heads of the participants several weeks after drug administrations. A few strands of axillary and pubic hairs were plucked at the same time as scalp hair. DP and LD were topically applied, orally administered and injected with intervals of at least 1 month. The participants affirmed that they had not taken any other medicines containing DP, LD, CP, and ME during the study.

The detailed sample preparation is described in our previous reports [11, 12]. Briefly, an individual hair strand was weighed and its full length was measured. After each hair strand was washed by sonicating with 1% sodium dodecyl sulfate (SDS) aqueous solution (1.5 mL) for 1 min, it was washed alternately with water (1.5 mL) and methanol (1.5 mL) 3 times for 1 min each. The hair strand was cut in a 0.4-mm length from the end of the hair root side using a tissue slicer equipped with a micrometer scale (Stoelting Co., Wood Dale, IL, USA), and each segment was placed in a 0.1-mL microtube. A mixture of the mobile phase consisting of an aqueous solution of 5 mM ammonium acetate plus 0.05% formic acid (mobile phase A) and acetonitrile (3:1, by vol.) was used as the extraction solution. Then, the extraction solution (100 μL) containing CP-d₆ (4 pg/mL) as the internal standard was added to the tube containing each 0.4-mm segment. The sample tube was then sonicated at 23 kHz for 10 min and then maintained at approximately 22 °C in a dark place for 24 h. The supernatant (35 μL) was transferred to a 96-well plate and then diluted with mobile phase A (35 μL), and the resultant solution (50 μL) was injected into a liquid chromatograph–tandem quadrupole mass spectrometer. The analytical conditions were summarized in supplementary material Table S1.

The analytical method was validated using spiked hair segments according to the method validation guidelines of the Scientific Working Group for Forensic Toxicology (SWGTOX) [25]. The analytes used in this study had already been validated in our previous reports [12, 24, 26]. The validation data were summarized in supplementary material Table S2.

Distribution profiles were generated by plotting drug concentrations per segment (Y-coordinate) versus segment numbers from the root to the tip side of the hair strand (X-coordinate). Values for peak parameters were measured as follows (see Fig. 1). Segment numbers with the maximal drug concentrations at the root (C_R) and tip (C_T) sides of the hair strands were represented as S_R and S_T, respectively. The distance between S_R and S_T was defined as D_R-T. The intersections at the level of Y = C_R/2 with the root curve of the peaks were determined for DP and DDP; twice the distance between an intersection and the line of X = S_R was defined as the width of the root-side peak (W_R). The intersections at the level of Y = C_T/2 with the tip curve of the peaks were determined for LD and DLD; twice the distance between an intersection and the line of X = S_T was defined as the width of the tip-side peak (W_T). When the X-coordinate of the intersection of a line and a curve was between segments m and m + 1, the segment number was always numbered as m + 0.5, to simplify measurements of peak parameters.

Hair strands that were washed as described above were sonicated with acetone (1.5 mL) for 3 min, followed by dichloromethane (1.5 mL) for 3 min to estimate the effects of vigorously washing the surfaces where the ointment was topically applied. Hair strands were also analyzed as described above.

Figure 2 shows distribution curves obtained from a scalp hair (hair ID; APS-31) of participant A, who applied the ointment containing DP and LD to the pubic area. The distribution curves of DP and its metabolite, DDP, comprised a major and a minor peak at the root and tip sides of the hair, respectively. In contrast, the distribution curves of LD and its metabolite, DLD, comprised a minor and a major peak at the root and tip sides, respectively. The participant had ingested CP and ME when he applied the ointment to link the peak positions on the distribution curve with the time of drug administration. The positions of the major peaks for CP and ME were near those for DP, DDP, LD, and DLD at the root and tip sides of the hair, respectively.

We investigated whether drug administration routes could be identified from the profiles of peaks in distribution curves. Both participants orally ingested a single dose of DP. A comparison of the distribution curves between topical and oral administration showed that the profiles of the peak for topical and oral DP and DDP were similar (Fig. 3a–d).

The participants both received a single gingival injection of LD for local anesthesia during dental treatments. We compared distribution profile between LD topically applied to the skin and that gingivally injected, in hairs. Figure 3e–h shows the distribution profiles of LD and DLD with minor root and major tip peaks. The tip-side peaks of LD and DLD were broader for topical application than gingival injection.

To confirm whether peak profiles significantly differed depending on drug administration routes, we analyzed strands of scalp hair from both participants at different sampling times and determined peak parameter values (Tables 1 and 2). The values of C_R and C_T for any analyte significantly differed between individual hair strands, regardless of the administration route. Although variations in W_R, W_T, C_T/C_R, and D_R-T within the same drug administration route tended to be smaller than those of C_R and C_T, the effects of administration routes did not significantly differ. Therefore, it was difficult to find characteristic peak profiles to identify drug administration routes.

Figure 4 shows typical distribution profiles of analytes in strands of axillary and pubic hairs obtained from participant A. Strands of axillary and pubic hairs were collected 2 weeks after the participant applied the ointment containing DP and LD to the pubic area. The participant had orally ingested CP and ME on the same day as the ointment application. Drug peaks were detectable in only 2 of 6 axillary hair strands analyzed. The positional relation between analyte peaks in axillary hair was similar to that in scalp hair. Whereas, high concentrations of DP, LD, and their metabolites were detected along the shafts of pubic hairs growing in the area where the ointment was applied.

Additionally, strands of axillary and pubic hairs were collected from participant A 2 weeks after DP was orally ingested. The participant also orally ingested CP and ME on the day following DP ingestion. Drug peaks were detectable in only 1 pubic hair strand out of 3 axillary and 3 pubic hair strands analyzed. The positional relation between analyte peaks in the pubic hair was also similar to that in scalp hair (see Fig. 4, right panel; Fig. 3c, d).

High concentrations of analytes were detected in hairs growing in the axillary and pubic areas where the ointment was applied. We washed the hair surfaces with aqueous SDS, water, and methanol, then sonicated them in acetone and dichloromethane for 3 min each. The analyte concentrations in these hairs did not significantly decrease (Fig. 5).

Concentration ratios of metabolites to unchanged DP and LD in each pubic hair segment were measured along individual hair strands. The ratios for both DP and LD were essentially constant at < 0.5% and < 3%, respectively, along the hair shaft, at the tip from segment 10 (4 mm from the hair root end). In contrast, ratios of metabolites fluctuated near the roots of the hairs, especially for those of LD (Fig. 5).

We previously showed that CP and ME are located at different positions in strands of scalp hair even when simultaneously ingested [12]. This is because CP is predominantly incorporated into the hair follicle via the bloodstream, whereas ME is predominantly incorporated into hair via the sweat and sebum secreted inside/near the scalp. The uptake routes of DP and LD into scalp hair were the same as those of CP and ME, respectively. The relationship between the uptake route and chemical properties of drugs has not yet been elucidated, although micro-segmental analysis has shown that many drugs were incorporated mainly via the bloodstream at hair follicles [12, 24, 27, 28].

Variations in the C_R were large even when drugs were administered via the same routes (Tables 1 and 2). Peaks for LD and DLD were not found at the root sides of some hair strands, because these peaks were small and obscured at the bottom of the tip-side peaks (Table 2, Figs. 2 and 3). This could be because the amount of drug incorporated into hair follicles depends on the activity of individual hair matrices. Additionally, the magnitude of the relationship between C_R and C_T was reversed in three hair strands [APS-32 for DDP (Table 1), ALS-22 and BIS-41 for LD (Table 2)]. Amounts of sweat and sebum secreted onto the scalp vary among individuals, as well as time, and the scalp region, and might affect the C_T. Characteristic peak profiles on distribution curves are more important to identify drug administration routes than drug concentrations. Although some peak profiles differed (Fig. 3e–h), all peak parameters considerably varied among individual hair strands even when drugs were administered via the same route. Therefore, peak profiles did not significantly differ according to drug administration routes in the present study. However, the other drugs that can be administered via different routes, such as inhalation and intravenous injection, should also be investigated. If the peak profile of a specific drug significantly changes depending on the administration route, the distribution profiles obtained by micro-segmental analysis would be useful to estimate which route had been applied in drug-related crimes.

Axillary and pubic hairs can be used as alternatives when scalp hair is not available for drug tests due to alopecia or heads shaved to hide evidence of drug use. We first applied micro-segmental analysis to axillary and pubic hairs. The distribution profiles in most axillary and pubic hairs, which are located far from areas where the ointment was applied, did not contain any significant analyte peaks. In contrast, the distribution profiles for all plucked scalp hair strands showed drug peaks in specific regions according to the time when the ointment was applied to the pubic area (Fig. 2). The ratios of axillary and pubic hairs in the catagen and telogen phases were higher than that in scalp hair (40–60% vs. ~ 15% [29]). In fact, only 2 of 9 axillary hair strands and only 1 of 3 pubic hair strands analyzed had sufficient amounts of analytes, regardless of the administration routes (Fig. 4). Thus, the risk of selecting individual hair strands in the catagen/telogen phases is higher among those obtained from the axillary and pubic areas than from the scalp. Therefore, individual axillary and pubic hairs are not effective as the alternatives to single scalp hairs for drug tests from the perspective of hair matrix activities and hair cycles. Although conventional hair analysis has detected drugs from axillary and pubic hairs [19‐22], tens of strands (at least 10 mg) are usually combined into one analytical sample. If micro-segmental analysis of axillary or pubic hairs is required under real circumstances, more hair strands must be analyzed to obtain the distribution profiles of hair strands in the anagen phase with high reproducibility.

High concentrations of DP and LD were detected along the shafts of all analyzed hair strands in the area where the ointment was applied because of direct contact with the pubic hair shaft (Fig. 5). However, DP and LD were essentially undetectable near the hair root because it was below the epidermis, or was generated de novo after ointment application. We confirmed that drugs adsorbed onto the hair surface was difficult to remove by harsh treatments such as sonication in various solvents, once hair strands were contaminated by the procedures with large amounts of drugs such as ointment application.

The respective DDP and DLD metabolites of DP and LD were also detected along the shaft of all analyzed hairs in the area where ointment was applied. This was attributed to decomposition of the unchanged form in air. In fact, DDP and DLD were detected in extracts of the ointment that had been exposed to air at 22 °C for 24 h (unpublished observation). Therefore, the presence of metabolites cannot necessarily be a marker to distinguish external contamination from drug administration by hair analysis when the metabolites and the decomposed compounds are the same.

The concentrations of DP and LD detected at the shafts of hairs growing in the area where ointment was applied (Fig. 5) exceeded the maximal concentrations in scalp hair (Tables 1 and 2). Additionally, the metabolite ratios at the hair shaft in the contaminated hair were lower than those at the maximal concentrations in scalp hair. The concentration of the unchanged form, the metabolite ratio, and the hair regions where the unchanged form and metabolites were detected would be important to distinguish external contamination from drug administration.

The slight increase in the metabolite ratio near the hair root in Fig. 5 implies that the detected analytes were generated by uptake via the bloodstream, sweat, and sebum, and not by external contamination. To precisely identify such characteristics on drug distribution profile, it is impossible to use the conventional hair analysis, in which hair surface contamination is evaluated only by comparing the amounts of a drug in solvents used to wash the hairs with that of the extracted drug [30‐32]. Differences in the amounts of drugs between wash solutions and extracts can be insufficient to evaluate contamination on hair surfaces, because the amounts of drugs in individual hair strands where drugs are localized at specific hair regions are averaged in the conventional hair analysis. Micro-segmental analysis can clearly differentiate distribution profiles between external contamination and drug administration (Figs. 1, 2, 3, 4, 5). For instance, whether a detected drug was derived from external contaminants can be determined if concentrations of an unchanged drug remain high and metabolite ratios remain low along the hair shaft (Fig. 5). In contrast, if the distribution profile has narrow peaks in specific regions, the detected drug would be regarded as being administered, because the partial adsorption of drugs onto the hair surface is difficult to create a distribution curve with major and minor peaks artificially (Figs. 2, 3, 4). Distribution profiles obtained using micro-segmental hair analysis would be useful not only to estimate when a drug was administered, as we previously reported [24, 27, 28], but also to identify whether the hair had been in contact with drugs externally.