Journal of Analytical Toxicology, 2018;42:265–275

doi: 10.1093/jat/bkx106
Advance Access Publication Date: 28 December 2017
Technical note

Technical note

Detection of Cocaine and Metabolites in Bone

Following Decomposition Using 2D LC–MS-MS
Malorie Mella1, Brendan Schweitzer1, Claude R. Mallet2, Tara Moore3,

Downloaded from https://academic.oup.com/jat/article/42/4/265/4780788 by guest on 16 February 2024

and Sabra Botch-Jones1,*
1
Department of Anatomy and Neurobiology, Biomedical Forensic Sciences, Boston University School of Medicine,
72 E. Concord St, L805C, Boston, MA 02118, USA, 2Separation Technologies, Waters Corporation, 34 Maple Street,
Milford, MA 01757, USA, and 3Department of Anatomy and Neurobiology, Forensic Anthropology, Boston
University School of Medicine, 72 E. Concord St, L805C, Boston, MA 02118, USA
*Author to whom correspondence should be addressed. Email: [email protected]

Abstract
Introduction: In forensic toxicology, challenges exist with quantification analysis of cocaine and
metabolites in postmortem samples following extensive decomposition. Alternative matrices,
such as bone could prove useful when other specimens are not available. Detection and quantifi-
cation of drugs in complex matrices require time-consuming extraction processes. The objective
of this study was to develop a robust extraction and clean-up methodology to efficiently extract
cocaine, and its metabolites, in bone and reach target limits of detection using multidimensional
chromatography.
Materials and methods: Under an Institutional Animal Care and Use Committee protocol, rat spe-
cimens underwent a 10–12 weeks chronic intravenous self-administration of cocaine with average
daily dosages ranging 13–19 mg/kg. This was followed by a 6-week period of abstinence, followed
again by a 3-week period of cocaine self-administration before being euthanized. Fourteen cocaine
positive rats were placed at the Boston University Forensic Anthropology Outdoor Research
Facility (Holliston, MA, USA) for a period of 12 months. Skeletal remains were collected for testing
as well as drug-free control rats. After homogenization of whole bones, the extraction process was
performed using a mixed mode reversed-phase/ion exchange sorbent with an extraction time of
1 h followed by analysis using a 2D liquid chromatography–mass spectrometry-mass spectrome-
try which allowed for direct injection of the eluent without evaporation or reconstitution. The anal-
ysis was performed using 100 μL of the final methanol extracts.
Results: The limit of quantitation for cocaine and benzoylecgonine was measured at 0.05ng/g and
for ecgonine methyl ester it was 0.1ng/g. The analytical method for cocaine gave a linear dynamic
range of 0.05–10ng/g with an R2 = 0.998.
Conclusions: The microextraction protocol combined with a multidimensional chromatography
used in this study decreased sample preparation time without sacrificing the quality seen with cur-
rent single dimension chromatography techniques.

Introduction human body has undergone decomposition and putrefaction, toxi-

In the field of forensic toxicology, several challenges exist with quan- cology screens of soft tissue may be possible, however, correlating
titative analysis of cocaine in postmortem samples, including its concentrations of cocaine and metabolites with drug dosage before
rapid half-life due to hydrolysis and postmortem redistribution. If a death is difficult. The process of decomposition includes autolysis,

© The Author(s) 2017. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected] 265
266 Mella et al.

putrefaction and postmortem redistribution. Whether due to the nat- for extraction of drugs of abuse in bone can take upwards of 70 h to
ural metabolic actions, stability of the drug itself, alkalinity, chemi- complete (2–14). Therefore, the objective of this research was to inves-
cal changes, bacteria and/or enzymatic activity, these variables will tigate whether cocaine and its major metabolites can be detected in
result in complex analysis and interpretation of drug concentrations bone samples after a year of decomposition and a second year of stor-
(1). When a body is encountered fully skeletonized, alternative age in room temperature conditions by incorporating 2D LC–MS-MS
matrices, such as hair, nails and bone could prove useful in detecting technology with a rapid extraction method for an overall goal of sam-
chronic drug use in postmortem cases. Bone and bone marrow are ple preparation and instrument runtime of <1 h.
highly vascularized and well protected from the environment. It sur-
vives intact for a greater length of time than any soft tissue in the
body. There are many routes in which drugs may incorporate itself
Material and Methods
into the bone material, including the highly vascularized periosteum, Instrumentation
surrounding blood vessels or Haversian canals. It is also possible For the purposes of this research, an ACQUITY UPLC® (Waters
that drugs are absorbed into bone surfaces during decomposition Corporation, Milford, MA, USA) was used, with a 2D configuration
when there is total breakdown of soft tissues. Regardless of how for “Trap and Elute” with at-column dilution (Figure 1). The detec-

Downloaded from https://academic.oup.com/jat/article/42/4/265/4780788 by guest on 16 February 2024

drugs enter the bones, detection of drugs of abuse in bone samples is tor used with the chromatography system was a Xevo TQD (Waters
possible (2–13). Studies utilizing rat or porcine bones have identified Corporation, Milford, MA, USA) mass spectrometer. The configura-
compounds such as ketamine, diazepam, citalopram, amitriptyline, col- tion was constructed with two Binary Solvent Managers (BSM)—
chicine, meperidine, tramadol and pentobarbital in bones after decom- one set for gradient elution and one set for “At-column” dilution.
position with drug regiments ranging from one dose to chronic The at-column dilution pump was plumbed to create two streams.
exposure (2–13). Further, a forensic case involving human skeletal The A side was set for loading the extracts from the injection loop
remains found in a riverine environment included toxicology testing on onto a 50 μL mixer, while the B side was set at high-flow rate for
bones, which revealed the presence of diazepam and nordiazepam (14). dilution. The diluted extracts are loaded and retained on a trap col-
Cocaine, also known as methylbenzoylecgonine, is an alkaloid umn before the elution pump performs a backflush elution on the
psychotropic drug derived from the South American Andes trap in order to re-focus the molecules onto the analytical column.
Erythroxylon coca plant. The Drug Enforcement Administration’s The final elution of the extracts flows into the mass spectrometer for
Controlled Substances Act classifies cocaine as Schedule II (15). It is detection. MassLynx© version 4.1 (Waters Corporation, Milford,
commonly administered as a topical anesthetic for ear, nose and MA, USA) software was utilized to view and analyze all chromato-
throat surgery. Symptoms of chronic cocaine use include central ner- grams and spectra referenced throughout this research, as well as
vous system stimulation, psychosis, respiratory dysfunction, rhinitis control all instrumentation. TargetLynx© version 4.1 (Waters
and more (15). After consumption, cocaine is rapidly distributed Corporation, Milford, MA, USA) software was utilized for all quan-
throughout the bloodstream, with ~90% of the compound binding titation performed throughout. Signal intensities were evaluated by
to plasma proteins (16). Cocaine is primarily metabolized to benzoy- measuring the area count of the peaks produced in the chromato-
lecgonine via spontaneous hydrolysis at alkaline pH and ecgonine grams. Each sample was analyzed in three replicate injections.
methyl ester via enzymatic hydrolysis by pseudocholinersterase.
Both metabolites can also be derived via liver carboxylesterases (15).
Reference standards and reagents
Cocaine is an ideal candidate for testing in alternative hard
matrices such as bone, hair and nails as it is commonly abused in a Analytical reference standards in the form of 1 mg/mL solutions dis-
chronic regiment and therefore has more time to become incorpo- solved in methanol were obtained from Cerilliant (Round Rock,
rated into the cellular activity associated with hair and nail growth TX, USA). They include cocaine, benzoylecgonine, ecgonine methyl
(20) and potentially bone remodeling. There is a distinct lack of ester and their respective deuterated internal standards cocaine-d3,
research studying cocaine and metabolites in skeletal remains as well benzoylecgonine-d3 and ecgonine methyl ester-d3. All solvents uti-
as lack of research comparing the effect of repeated dosing and sin- lized throughout this research, including methanol, acetonitrile, ace-
gle dosing in drug concentrations of skeletal remains. This study tone, ammonium hydroxide, hydrochloric acid and formic acid,
investigates the detection of cocaine in rat skeletal remains following were Optima grade and purchased from Fischer Scientific
long-term exposure to cocaine and decomposition. Though rats lack (Waltham, MA, USA). All water consumed in this research was
Haversian remodeling seen in humans, male rats still experience lon-
gitudinal bone growth up to 30 months in age (21). Furthermore,
cortical bone growth in rats occurs in the periosteum—the highly
vascularized connective tissue surrounding the bone (21). Whether
drugs are incorporated overtime or acutely, the age of the rats used
in this study as well as the drug regiment leave the possibility of
either route of drug absorption.
Hyphenated instrumentation, particularly liquid chromatography–
tandem mass spectrometry (LC–MS-MS) are widely used in forensic
and clinical toxicology, particularly due to the need for high specificity
and sensitivity (17). By incorporating sample clean-up methods to iso-
late the analyte of interest, matrix effects, which may result in ion sup-
pression or enhancement, associated with complex samples can be
reduced. This research utilized a 2D LC–MS-MS system that incorpo-
rated a “Trap and Elute with At-Column Dilution” concept (18, 19).
Further, traditional sample preparation methods in the literature used Figure 1. Fluidic pathway of 2D LC system.
Detection of Cocaine and Metabolites in Bone Following Decomposition 267

Millie-Q grade water (Merck/Millipore Co, Darmstadt, Germany, Additionally, loading and dilution pH were optimized by testing pH
EU). A stock solution of cocaine, benzoylecgonine and ecgonine 3, 7 and 10 on the two different sized columns. A low and high pH
methyl ester was prepared at a concentration of 10 μg/mL in 100% loading phase was compared for each combination. An example of
methanol. This stock solution was utilized to prepare serial dilutions the chromatography evaluation scheme is pictured in Figure 2. A
of lower concentration stock solutions throughout this project was 10 ng/mL stock solution was prepared in water, methanol and ace-
prepared fresh each month. tonitrile and run on an automated process testing different para-
meters over 18 h. Elution consisted of two mobile phases, A and B.
Mobile phase A consisted of water with 0.5% formic acid and
Compound optimization mobile phase B consisted of acetonitrile with 0.5% formic acid. It
Compound optimization was performed on all standards. Each com- was determined that all analytes gave the best signal intensity with a
pound was prepared in a 50/50 water/methanol solution at a con- pH 10 loading and dilution on the 80 mg HLB trap column.
centration of 10 μg/mL (10 ppm). The solutions were directly
infused into the mass spectrometer. Precursor ions were determined
by running a scan in MS1 mode. Two product ions were then Sample extraction
selected for each standard in MS/MS mode, while varying the colli- To choose a solid phase extraction sorbent, the pKa of each com-

Downloaded from https://academic.oup.com/jat/article/42/4/265/4780788 by guest on 16 February 2024

sion energy values. The most intense transition was selected for pound was referenced. Cocaine, benzoylecgonine and ecgonine
quantification and the second most intense transition was selected methyl ester have pKa’s of 8.6, 10.82 and 9.57, respectively. All
for confirmation. The collision energy value at which the transition compounds are basic in nature, therefore an Oasis® MCX solid
gave the highest signal was chosen. The mass spectrometer and mul- phase extraction barrel (Waters Corporation, Milford, MA, USA)
tiple reaction monitoring values are listed in Tables I and II. was chosen for testing, as it is a mixed mode sorbent consisting of a
strong cation exchange portion meant to retain ionized basic com-
pounds and a reverse phase portion meant to retain acids and neu-
Analytical and trap columns trals. For preliminary method development, two sorbent packing
The analytical separation column used for all analyses was an sizes of this solid phase extraction sorbent were compared. The mix-
ACQUITY HSS T3, 2.1 × 50 mm2, 1.7 μm (Waters Corporation, ture of compounds was prepared at a concentration of 1 ng/mL
Milford, MA, USA), which is a high strength silica C18 sorbent. (1 ppb) in methanol, acetonitrile, acetone, 50/50 methanol and ace-
The sample injection volume was set at 100 μL. The sample is tonitrile, 95/5 acetonitrile and ethyl acetate and 95/5 acetonitrile
loaded with water at 0.1 mL/min and diluted at 5% with water at a and methyl tertial butyl ether—all with 5% ammonium hydroxide
flow rate of 1.9 mL/min. The mixture was loaded on the trap col- added. A 2 mL of each mixture was dissolved in 50 mL of Millipore
umn, subsequently described, at a flow rate of 2.0 mL/min. Then the water for loading on the 60 mg MCX cartridge and 100 mL of
compounds retained on the trap were refocused onto the analytical water for the 150 mg MCX cartridge (Figures 3 and 4). It was deter-
column during elution, which was completed with a 5-min linear mined that the 150 mg MCX cartridge, using methanol as the elu-
gradient from 5% of organic solvent mobile phase to 95% mobile tion solvent gave the best recoveries for all three compounds and
phase. The total runtime was 10 min. therefore chosen for the final method (Figure 5).
The chromatography conditions were tested on two trapping
chemistries, Oasis® hydrophilic–lipophilic balance (HLB) 40 and
Homogenization and extraction
80 mg trap columns (Waters Corporation, Milford, MA, USA). This
A Precellys Evolution homogenizer (Bertin Technologies, Montigny-
is a strong hydrophilic, high-retention strength sorbent material.
le-Bretonneux, France, EU) was utilized for homogenization of the
bone samples. The homogenization was performed by loading the
Table I. Multiple reaction monitoring table for all compounds
Precellys with 15 mL tubes containing 4 mm ceramic ball bearings
Compound Ion Precursor Cone Product ion CE (Bertin Technologies, Montigny-le-Bretonneux, France, EU), and the
mode ion extraction solvent which were then pressurized and shaken at speed

Cocaine ESI + 304.1 30 182.1 20

30 82.1 30
Benzoylecgonine ESI + 290.1 30 168.1 15
30 105.1 30
Ecgonine methyl ester ESI + 200.1 30 182.1 20
30 82.1 25
Cocaine-D3 ESI + 307.1 30 185.1 20
30 85.1 30
Benzoylecgonine-D3 ESI + 293.1 30 171.1 20
30 105.1 30
Ecgonine methyl ester-D3 ESI + 203.1 30 185.1 20
30 85.1 25

Table II. Mass spectrometer settings

Capillary Cone Source Desolvation Desolvation Cone

voltage voltage temperature temperature gas gas

3.0 kV 30 V 150°C 550°C 1100 L/h 50 L/h

Figure 2. Methods tested for optimization of method.
268 Mella et al.

Downloaded from https://academic.oup.com/jat/article/42/4/265/4780788 by guest on 16 February 2024

Figure 3. Area counts comparison of cocaine MCX 60 vs 150 mg.

Figure 4. Area counts comparison of benzoylecgonine MCX 60 vs 150 mg.

Detection of Cocaine and Metabolites in Bone Following Decomposition 269

of 5,000 RPM for periods of 90 s at a time. It was determined that To choose the best extraction solvent that would give the highest
three rounds of shaking for 90 s at 5000 RPM was required to pul- recovery of the compounds of interest, several solvents were tested:
verize the bone into a fine powder. The ceramic ball bearings were Millie-Q water, methanol, acetonitrile and acetone. Additionally,
washed twice after each use with MeOH + 5% formic acid with each solvent was tested at a different pH—hydrocholoric acid was
15 min of sonication followed by 100% MeOH for 15 min of soni- used for pH 1, ammonium hydroxide for pH 10, and no additive
cation. A comparison between new and used ceramic ball bearings for pH 7. After homogenization, the tubes were centrifuged at 4,000
showed no effect in signal intensity, which indicated that the clean- RPM for 5 min. The final extract was then diluted in 100 mL of
ing method utilized in between each use was efficient. Millie-Q water and underwent solid phase extraction via the method
previously described. Results indicated that a methanol extraction
at pH 7, with no additives, gave the best intensity for cocaine.
However, the recoveries of benzoylecgonine and ecgonine methyl
ester were relatively low. The water extraction at pH 7 gave high-
er intensities for benzoylecgonine and ecgonine methyl ester. Therefore,
it was concluded that two sequential extractions were necessary

Downloaded from https://academic.oup.com/jat/article/42/4/265/4780788 by guest on 16 February 2024

for optimal extraction efficiency. After the methanol extraction,
the supernatant was collected. Water was added to the 15 mL tube and
re-homogenized. The supernatants were then pooled and then diluted
into 100 mL of water for solid phase extraction.
Although it is not possible to know the precise concentration of
target compounds initially present in bone matrices and therefore
not possible to reliably determine overall extraction efficiency, an
estimation of the final method extraction’s capabilities was con-
ducted. Unextracted standards, water extractions, and bone extrac-
tions were compared. The 100 mL water used for dilution pre-solid
phase extraction was adjusted at pH 1, 7 or 10 prior to loading
the extracts. When comparing the bone extractions to the water
extractions, it was determined that a pH 10 load gave the best
overall recovery for all three compounds in comparison to the
water extraction values (Figure 6). With the exception of ecgo-
nine methyl ester, the water extractions for cocaine and benzoy-
Figure 5. Final extraction protocol. lecgonine produced recoveries over 90%. The bone extractions

Figure 6. Solid phase extraction yields of unextracted standards, water extractions and bone extractions at different loading pH.
270 Mella et al.

for cocaine and benzoylecgonine were 50–60%. There was a low Fourteen cocaine positive and four drug-free control rats were
extraction efficiency for ecgonine methyl ester in both water and bone. placed outside and above ground at the Boston University Forensic
Anthropology Outdoor Facility (Holliston, MA, USA) for a period
of 12 months. Rat specimens were completely skeletonized upon col-
Filtration lection. All recoverable skeletal samples were collected, cleaned and
It was apparent during method development that there was a need stored at room temperature for an additional year prior to testing
for a filtration step before the dilution step. The supernatants of the due to time constraints associated with instrument access and
bones after homogenization and centrifugation still contained particles method development. Due to size of skeletal specimens and outdoor
which caused the solid phase extraction cartridges to clog. A Sartorius weather conditions, some samples such as caudal vertebrae, phalan-
Vivaspin 6 with 0.2 μm PES membrane centrifuge spin tube (Sartorius, ges and vertebrae were not recovered. Remaining bone specimens
Goettingen, Germany) containing a 0.45 μm PTFE filter was chosen. were grouped by bone type and each sample averaged ~0.5 g. Femora
After centrifugation of the bones, the supernatant was poured into the (n = 21), tibio-fibula (n = 20), humeri + ulna combinations (n = 19)
spin tubes and then centrifuged again at 4,000 RPM for 5 min. and innominate (n = 17) bones were prepared via the final extraction
method previously described. Additionally, one humerus and ulna com-

Downloaded from https://academic.oup.com/jat/article/42/4/265/4780788 by guest on 16 February 2024

Long-term cocaine exposure to rat specimens bination, two ulna and two radii combination, 0.5 g of vertebrae and
one tibio-fibula taken from a rat immediately after being euthanized
All male Wistar rat specimens, ~6 months in age, used for this study
was analyzed for comparison. After the 1-year period of decomposi-
were donated from the Laboratory of Behavioral Neuroscience at
tion, all recovered bones were washed with water and stored in plastic
Boston University’s Psychology and Brain Sciences department and
containers at room temperature until analysis could be performed. The
fell under an Institutional Animal Care and Use Committee
bones retrieved immediately after death of the rodent specimen were
(IACUC) protocol (21). The rats underwent 10–12-week chronic
extracted from the rodent body and cleaned manually.
intravenous self-administration of cocaine. This was followed by a
6-week period of abstinence, followed again by a 3-week period of
cocaine self-administration before being euthanized and donated for
Results
the purposes of this study. Average daily dosages for each rat fell
within a range of 13–19 mg/kg. As dose–death intervals of these rats Calibration curve
were not measured in the previously mentioned research it was A calibration curve ranging from 0.05 to 10 ng/mL was run in both
therefore unknown to the investigators of this study. water and bone on the same day as the final extractions to assist in

Figure 7. Matrix effects and recover calculations for ecgonine methyl ester.
Detection of Cocaine and Metabolites in Bone Following Decomposition 271

quantitation. The calibrators were used to spike 100 mL water solu- counts of the post-spiked internal standard of the extracted standard.
tions as well as spiked 0.5 g bone samples. A full unextracted curve The percentage difference between the values was the calculated suppres-
containing only neat standards, a water extracted curve, and a bone sion. The suppression effect for the 100 mL Millipore water enrichment
extracted curve were analyzed on the same day as the unknown sam- process is 6.3% for ecgonine methyl ester, 8.8% for benzoylecgonine
ples. Additionally, a blank and a recovery sample at a concentration of and 9.5% for cocaine. The bone matrix incorporation increases the sup-
1 ng/mL were run with the water extraction and the bone extraction. A pression effect to 8.6% for ecgonine methyl ester, 24% for benzoylecgo-
mixture of internal standards containing cocaine-D3, benzoylecgonine- nine and 21% for cocaine. It is expected that bone would cause a larger
D3 and ecgonine methyl ester-D3 was post-spiked to all final elutions at amount of suppression of the compounds because it is a complex matrix
a concentration of 1 ng/mL. The recovery samples were analyzed with containing many potential interferences.
the curve as an unknown in order to calculate how accurate the curve The water extraction recovery was calculated using the area counts
was when analyzing an unknown in the same matrix. A ratio of the and the ion ratio of the water extraction and that of the unextracted
internal standard area count and the calibrator area count of each standard. In the water extraction, ecgonine methyl ester showed a 54%
concentration was used to calculate the calibration curve equation. recovery based on the compound’s intensity, and a 57% recovery based
The final linear dynamic range for cocaine was 0.05–10 ng/mL, on the ion ratio (Figure 7). Benzoylecgonine had a 72% recovery based

Downloaded from https://academic.oup.com/jat/article/42/4/265/4780788 by guest on 16 February 2024

and 0.1–10 ng/mL for benzoylecgonine and ecgonine methyl ester. The on the compound’s intensity, and an 80% recovery based on the ion
unextracted standard calibration curve, the water extraction calibration ratio (Figure 8). Cocaine had an 85% recovery based on the com-
curve, and the bone extraction calibration curve all show excellent lin- pound’s intensity and a 94% recovery based on the ion ratio (Figure 9).
earity with R2 values >0.995. The within curve precision was calculated Overall, cocaine demonstrated the best recoveries overall. High recovery
using the TargetLynx software. All points on all curves demonstrated calculations for cocaine prove that the sample preparation method
<20% deviation. The 1 ng/mL recovery sample for ecgonine methyl described previously was efficient at extracting and isolating cocaine
ester was calculated at 88.7% accuracy in the calibration curve, 80% from both a clean and very complicated matrix, though additional
for benzoylecgonine and 94.3% accurate for cocaine. method develop is required to improve recovery of metabolites overall.

Matrix suppression and recovery calculations Quantitation of unknown samples

Matrix suppression was calculated by comparing the intensity or area The results of the calculated concentrations of the biological specimens
counts of the internal standard of the neat standard and the area are displayed in Tables III–VI, divided by bone type. Quantitation of

Figure 8. Matrix effects and recover calculations for benzoylecgonine.

272 Mella et al.

Downloaded from https://academic.oup.com/jat/article/42/4/265/4780788 by guest on 16 February 2024

Figure 9. Matrix effects and recovery calculations for cocaine.

Table III. Quantitation of compounds in innominate bones (TLD) Table IV. Quantitation of compounds in femora (TLD)

Cocaine Benzoylecgonine Ecgonine methyl Cocaine (ng/g) Benzoylecgonine Ecgonine methyl

(ng/g) (ng/g) ester (ng/g) (ng/g) ester (ng/g)

Innominate 1 0.015 0.104 ND Femur 1 TLD TLD ND

Innominate 2 TLD 0.13 ND Femur 2 TLD 0.052 ND
Innominate 3 0.104 0.1 ND Femur 3 TLD 0.23 ND
Innominate 4 0.01 TLD ND Femur 4 TLD 0.2 ND
Innominate 5 0.002 0.044 ND Femur 5 TLD 0.042 ND
Innominate 6 0.1 0.054 ND Femur 6 TLD 0.108 ND
Innominate 7 0.002 0.066 ND Femur 7 TLD TLD ND
Innominate 8 0.016 0.038 ND Femur 8 0.0004 0.066 ND
Innominate 9 TLD TLD ND Femur 9 0.08 0.048 ND
Innominate 10 0.022 0.006 ND Femur 10 0.24 0.01 ND
Innominate 11 0.0016 TLD ND Femur 11 TLD 0.026 ND
Innominate 12 TLD TLD ND Femur 12 TLD 0.034 ND
Innominate 13 TLD 0.062 ND Femur 13 0.016 0.038 ND
Innominate 14 TLD 0.054 ND Femur 14 0.054 0.216 ND
Innominate 15 0.018 0.064 ND Femur 15 0.026 0.112 ND
Innominate 16 0.008 ND ND Femur 16 0.056 0.038 ND
Innominate 17 0.0004 TLD ND Femur 17 0.042 0.224 ND
Femur 18 0.048 TLD ND
Femur 19 0.274 0.1 ND
each specimen was achieved by using the linear equations generated
Femur 20 0.036 ND ND
from the calibration curve in bone, shown above. The biological speci- Femur 21 0.036 ND ND
mens include 17 innominate bones, 21 femora, 19 humerus + ulna
Detection of Cocaine and Metabolites in Bone Following Decomposition 273

Table V. Quantitation of compounds in humeri and ulan (TLD) of many challenges to accurately quantifying drug levels in bone is
the long extraction and sample preparation times required for analy-
Cocaine Benzoylecgonine Ecgonine methyl
sis. The Precellys Evolution homogenizer used in this study could
(ng/g) (ng/g) ester (ng/g)
completely pulverize the bone samples into a fine powder, with little
Humerus + ulna 1 0.052 TLD ND to no large particulates left, therefore, exposing the extraction sol-
Humerus + ulna 2 TLD TLD ND vent to the greatest possible surface area within the sample. This
Humerus + ulna 3 TLD TLD ND process was also used to spike bone samples for the calibration
Humerus + ulna 4 TLD TLD ND curve. It is necessary to remember that it is impossible to form a
Humerus + ulna 5 TLD 0.028 ND completely homogenous mixture with a spiked bone sample, thus ren-
Humerus + ulna 6 0.038 TLD ND
dering it challenging to make an accurate calibration curve. However,
Humerus + ulna 7 0.124 0.254 ND
the elimination of evaporation to dryness steps seen in traditional sam-
Humerus + ulna 8 TLD TLD ND
Humerus + ulna 9 0.24 0.124 ND ple preparation methods eliminated the potential of evaporative loss.
Humerus + ulna 10 TLD TLD ND These factors give confidence to the quantitation of samples.
Humerus + ulna 11 0.038 TLD ND The next largest challenge to quantifying drugs in bone for

Downloaded from https://academic.oup.com/jat/article/42/4/265/4780788 by guest on 16 February 2024

Humerus + ulna 12 TLD ND ND forensic casework is the inability to directly correlate systemic drug
Humerus + ulna 13 TLD ND ND levels and bone drug levels. Due to the lack of access to rats’ sys-
Humerus + ulna 14 TLD ND ND temic levels of cocaine in vivo, a comparison of drug levels is outside
Humerus + ulna 15 TLD ND ND the scope of this study. Furthermore, because these animals were
Humerus + ulna 16 0.366 0.058 ND exposed to cocaine both chronically and immediately before death,
Humerus + ulna 17 0.094 0.026 ND
it is unknown whether the detection of cocaine was due to accumu-
Humerus + ulna 18 0.016 TLD ND
lation of the drug in the bone tissue overtime or due to the most
Humerus + ulna 19 0.001 ND ND
recent cocaine exposure prior to sacrifice. Finally, an incubation
period was not used nor assessed in this evaluation.
Table VI. Quantitation of compounds in tibio-fibulae (TLD)

Cocaine Benzoylecgonine Ecgonine methyl

(ng/g) (ng/g) ester (ng/g) Conclusions
Tibio-fibula 1 TLD TLD ND
The product of this research was development of a rapid and robust
Tibio-fibula 2 TLD TLD ND method to detect trace levels of cocaine in bone, particularly in the
Tibio-fibula 3 TLD ND ND case of fully skeletonized remains. In rodent bone samples that
Tibio-fibula 4 TLD 0.068 ND underwent decomposition for 1 year, followed by 1 year of storage,
Tibio-fibula 5 TLD 0.01 ND cocaine was detected at a range of 0–0.366 ng/g. Benzoylecgonine
Tibio-fibula 6 TLD TLD ND was detected at a range of 0–0.254 ng/g and ecgonine methyl ester
Tibio-fibula 7 TLD 0.044 ND was not detected in any of the 2-year samples. It was detected in a
Tibio-fibula 8 TLD TLD ND range of 0.09–0.401 ng/g in the time zero bone samples. Due to
Tibio-fibula 9 TLD ND ND
extraction efficiency issues for ecgonine methyl ester during the
Tibio-fibula 10 0.0318 TLD ND
method development phase, it is unclear whether ecgonine methyl
Tibio-fibula 11 TLD TLD ND
Tibio-fibula 12 TLD 0.034 ND
ester was totally absent from the bone samples or if the method was
Tibio-fibula 13 TLD TLD ND unable to extract and detect it. Cocaine has proven to remain detect-
Tibio-fibula 14 0.001 TLD ND able in bone after full decomposition and storage at room tempera-
Tibio-fibula 15 TLD TLD ND ture, in both quantifiable and trace levels. Results that indicate TLD
Tibio-fibula 16 TLD 0.034 ND showcase a sharp, Gaussian peak shape for cocaine with a low base-
Tibio-fibula 17 TLD 0.04 ND line, demonstrating that the extraction and sample clean-up method
Tibio-fibula 18 TLD 0.052 ND was capable of detecting analytes in the range of low picogram or
Tibio-fibula 19 TLD 0.02 ND high femtogram per gram.
Tibio-fibula 20 TLD TLD ND
This study provides a reference for forensic casework and those
seeking to test for drugs in bone samples, since human remains are
often incomplete, badly decomposed or fully skeletonized. The
combinations and 20 tibio-fibulae. Each set of bones ranged from 0.5 ability to test any bone with confidence would be beneficial.
to 1.0 g of material and results are mass normalized. Despite the inability to correlate bone drug levels to an accurate
With the exception of cocaine, the bones analyzed at time zero systemic level, the presence of drugs in bone lends investigators
(T0) (Figure 10), all show the greatest concentrations of all analytes. information about what the deceased may have been exposed to
Ecgonine methyl ester was not detected in any bone analyzed after 2 preceding death.
years (T2 years). Trace level detection (TLD) indicates concentrations This methodology and instrumentation could be easily incorpo-
below 0.0001 ng/g. All drug-free control specimens tested yielded rated into toxicology testing laboratories. LC–MS is already an
negative results as expected. established and court approved method for forensic scientists.
With the addition of a pump and a trap column, sample prepara-
tion times could be reduced 10-fold as well as LOQ’s and LOD’s
Discussion can be potentially improved. For forensic casework, which often
This research presents the development process of a method to has very time sensitive demands, this technology would be an asset
extract and detect drugs in skeletal samples in an hour or less. One to any laboratory.
274 Mella et al.

Downloaded from https://academic.oup.com/jat/article/42/4/265/4780788 by guest on 16 February 2024

Figure 10. Analyte distribution in bone at T0 by bone type.

Acknowledgments 6. Watterson, J.H., Botman, J.E. (2009) Detection of acute diazepam

exposure in bone and marrow: influence of tissue type and the dose-
This study was performed in a collaborative effort between Boston University death interval on sensitivity of detection by ELISA with liquid chroma-
and Waters Corporation. The authors wish to thank Waters Corporation for tography tandem mass spectrometry confirmation. Journal of Forensic
providing training and instrument access for the purposes of this study. We Sciences, 54, 708–714.
also wish to thank the Boston University School of Medicine’s Forensic 7. Watterson, J.H., Desrosiers, N.A. (2011) Examination of the effect of
Anthropology graduate program including Dr Don Siwek, Dr James Pokines, dose-death interval on detection of meperidine exposure in decomposed
Dr Jonathan Bethard and Gary Reinecke for their assistance with anthropo- skeletal tissues using microwave-assisted extraction. Forensic Science
logical training and access to the Outdoor Research Facility in Holliston, International, 207, 40–45.
MA. We wish to also thank the Boston University’s Department of 8. Wiebe, T.R., Watterson, J.H. (2014) Analysis of tramadol and O-
Psychological and Brain Sciences for providing the specimens used in this desmethyltramadol in decomposed skeletal tissues following acute and
study. repeated tramadol exposure by gas chromatography mass spectrometry.
Forensic Science International, 242, 261–265.
9. Cornthwaite, H.M., Watterson, J.H. (2014) Microwave assisted extrac-
tion of ketamine and its metabolites from skeletal tissues. Analytical
References Methods, 6, 1142.
1. Skopp, G. (2004) Preanalytic aspects in postmortem toxicology. Forensic 10. Cornthwaite, H.M., Watterson, J.H. (2014) The influence of body posi-
Science International, 142, 75–100. tion and microclimate on ketamine and metabolite distribution in
2. Watterson, J.H., Desrosiers, N.A., Betit, C.C., Dean, D., Wyman, J.F. decomposed skeletal remains. Journal of Analytical Toxicology, 38,
(2010) Relative distribution of drugs in decomposed skeletal tissue. 548–554.
Journal of Analytical Toxicology, 34, 510–515. 11. Imfeld, A.B., Watterson, J.H. (2015) Influence of dose-death interval on
3. Watterson, J.H., Donohue, J.P. (2011) Relative distribution of ketamine colchicine and metabolite distribution in decomposed skeletal tissues.
and norketamine in skeletal tissues following various periods of decompo- International Journal of Legal Medicine, 130, 371–379.
sition. Journal of Analytical Toxicology, 35, 452–458. 12. Desrosiers, N.A., Watterson, J.H. (2010) The effects of burial on drug
4. Watterson, J.H., Donohue, J.P., Betit, C.C. (2012) Comparison of relative detection in skeletal tissues. Drug Testing and Analysis, 2, 346–356.
distribution of ketamine and norketamine in decomposed skeletal tissues 13. Desrosiers, N.A., Watterson, J.H., Dean, D., Wyman, J.F. (2011)
following single and repeated exposures. Journal of Analytical Toxicology, Detection of amitriptyline, citalopram, and metabolites in porcine bones
36, 429–433. following extended outdoor decomposition*. Journal of Forensic
5. Watterson, J.H., Cornthwaite, H.M. (2013) Discrimination between pat- Sciences, 57, 544–549.
terns of drug exposure by toxicological analysis of decomposed skeletal 14. Delabarde, T., Keyser, C., Tracqui, A., Charabidze, D., Ludes, B. (2013)
tissues. Part II: amitriptyline and citalopram. Journal of Analytical The potential of forensic analysis on human bones found in riverine envi-
Toxicology, 37, 565–572. ronment. Forensic Science International, 228, e1–e5.
Detection of Cocaine and Metabolites in Bone Following Decomposition 275

15. Isenschmid, D.S. Cocaine. In: Levine, B. (ed). Principles of Forensic 19. Mallet, C., Botch-Jones, S. (2016) Illicit drug analysis using two-
Toxicology. AACC Press: Washington, DC, 2013; pp. 293–316. dimension liquid chromatography/tandem mass spectrometry. Journal of
16. Barroso, M., Gallardo, E., Queiroz, J. (2009) Bioanalytical methods for Analytical Toxicology, 40, 617–627.
the determination of cocaine and metabolites in human biological sam- 20. Kronstrand, R., Druid, H. Hair in postmortem toxicology. In: Kintz, P.
ples. Bioanalysis, 1, 977–1000. (ed). Analytical and Practical Aspects of Drug Testing in Hair. Taylor &
17. Maurer, H.H. (2007) Current role of liquid chromatography–mass spec- Francis Group: Boca Raton, FL, 2007; pp. 223–240.
trometry in clinical and forensic toxicology. Analytical and Bioanalytical 21. Lelovas, P.P., Xanthos, T.T., Thoma, S.E., Lyritis, G.P., Dontas, I.A.
Chemistry, 388, 1315–1325. (2008) The laboratory rat as an animal model for osteoporosis research.
18. Mallet, C.R., Murphy, B.P. Multi-Dimensional Chromatography Comparative Medicine, 58, 424–430.
Compendium. Waters Corporation 2015; 1–8.

Downloaded from https://academic.oup.com/jat/article/42/4/265/4780788 by guest on 16 February 2024