Forensic Science International 188 (2009) 23–30

Microwave-assisted extraction in toxicological screening of skeletal tissues

Nathalie A. Desrosiers, Caroline C. Betit, James H. Watterson *

Forensic Toxicology Research Laboratory, Laurentian University, 935 Ramsey Lake Rd., Sudbury, Ontario, Canada P3E 2C6

A R T I C L E I N F O

Article history:

Received 13 October 2008

Received in revised form 23 January 2009

Accepted 8 March 2009

Available online 18 April 2009

Keywords:

Forensic toxicology

Bone

Microwave

Drug

Pentobarbital

Ketamine

Diazepam

Immunoassay

A B S T R A C T

The use of microwave-assisted extraction (MAE) in screening of decomposed bone tissue for model drugs

of abuse is described. Rats received 50 mg/kg (i.p.) pentobarbital (n = 2), 75 mg/kg (i.p.) ketamine (n = 2)

or 16 mg/kg (i.p.) diazepam (n = 1), or remained drug-free (control). Drug-positive animals were

euthanized within 20 min of drug administration. Animal remains were allowed to decompose in a

secure outdoor environment to the point of complete skeletonization. Bones (tibiae, femora, vertebrae,

ribs, pelvi, humeri and scapulae) were collected and pooled (according to drug) in order to minimize

effects due to inter-bone differences in drug distribution. Bones were crushed and cleaned of marrow

and residual soft tissue in alkaline solution or phosphate buffer with ultrasonication. Cleaned bones were

then ground and underwent MAE in phosphate buffer (pH 6), methanol or a methanol:water mixture

(1:1, v/v) at atmospheric pressure in a domestic microwave oven, or passive extraction in methanol.

Bone extracts (control and drug-exposed) containing methanol were evaporated to dryness before

reconstitution in phosphate buffer (pH 6) and subsequent analysis by ELISA, while bone extracts

containing only phosphate buffer were assayed directly by the same ELISA protocol. Measured

absorbance values were expressed as the decrease in absorbance, measured as a percentage, relative to

the corresponding drug-free control bone extract. The semi-quantitative nature of the ELISA assay

allowed examination of the effects of extraction solvent and bone sample mass on the assay response for

each drug examined, and subsequent comparison to assays of extracts obtained through passive

methanolic extraction of various bone tissues. Overall, the time required for maximal extraction varied

with extraction solvent and bone mass for each drug investigated, with significant extraction occurring

with all solvent systems examined. MAE may represent a substantially faster extraction system than

passive extraction, with significant extraction recovery observed within 1 min of exposure for all

samples examined. The implications of these results in the context of the available literature on drug

analysis in skeletal tissues are discussed.

� 2009 Elsevier Ireland Ltd. All rights reserved.

Contents lists available at ScienceDirect

Forensic Science International

journa l homepage: www.e lsevier .com/ locate / forsc i in t

1. Introduction

When skeletonized remains are encountered in death inves-tigations, conventional sample matrices, such as blood, are scant orcompletely unavailable. Therefore, bone tissue may be the onlysource of toxicological information. Although a growing body ofresearch has been performed in this area [1–10], with drugs such asketamine [1,2], tricyclic antidepressants [3,4], benzodiazepines [4–6], antipsychotics [4] and morphine [7] being detected in bone orbone marrow, understanding of the implications and limitations ofdrug measurements in skeletal tissues remains poor. For example,drug and metabolite uptake into bone is poorly understood, as isthe time course of drugs within those tissues. Only a handful ofstudies have considered the distribution of drug within a bone (i.e.,trabecular vs. cortical bone) [1,2,5]; drug distribution between

* Corresponding author. Fax: +1 705 675 1151.

E-mail address: jwatterson@laurentian.ca (J.H. Watterson).

0379-0738/$ – see front matter � 2009 Elsevier Ireland Ltd. All rights reserved.

doi:10.1016/j.forsciint.2009.03.008

different bones (e.g., femur vs. vertebrae) has not been studiedunder controlled conditions.

Drug recovery from mineralized bone may be challenging, andremains poorly characterized. Review of the literature shows thatdrug isolation from bone has been attempted using a number ofdifferent methods, including passive methanolic extraction oftransverse slices from the mid-femoral region [4], bone slivers [8],or ground bone [1,2,5], Soxhlet extraction [9], and acid digestion[7,10]. Typically, the methods reported thus far have been time-consuming, typically requiring 12–72 h of incubation time(Table 1). In laboratory studies of drug distribution betweendifferent tissue types, which typically make use of experimentalanimals, it would not be unreasonable to expect to generate 5–10different samples per animal under a given set of experimentalconditions (e.g., drug, dose, etc.). Clearly then, use of passiveextraction or digestion methods represents a significant limitationto analytical throughput.

One approach to expedite this process is microwave-assistedextraction (MAE). MAE is well established and commonly used in

Table 1Summary of reported extraction methodologies for toxicological analysis of skeletal tissues, with time used for extraction.

Study Sample matrix Extraction solvent Drug(s) detected Extraction time (h)

McIntyre et al. [4] Mid-femoral sections—human Methanol Tricyclic antidepressants,

nontricyclic antidepressants

benzodiazepines, antipsychotics

18

Horak and Jenkins [8] Mid-femoral slivers—human Methanol or water Citalopram 24

Guillot et al. [7] Bone (unspecified anatomic source)—mouse 0.1N HCl Morphine 12

Raikos et al. [10] Femoral fragment—human 3N HNO3 Morphine 24

Watterson et al. [1,2,5] Diaphyseal and epiphyseal femoral sections—rat Methanol Ketamine, diazepam 72

Wohlenberg, et al. [3] Cancellous vertebral Bone—Human Methanol Nortriptyline 15

Gorczynski and Melbye [6] Bone (unspecified anatomic source)—mouse and rats PBS Midazolam, diazepam, lorazepam Unspecified

N.A. Desrosiers et al. / Forensic Science International 188 (2009) 23–3024

environmental analysis, with reported extractions of polyaromatichydrocarbons (PAHs), polychlorinated biphenyls (PCBs), andorganochlorine pesticides (OCPs) [11,12] from a variety ofsubstrates of environmental interest (e.g., sediments). Microwaveheating is also used for digestion of heavy metals [13]. Variablessuch as time, pressure, temperature and moisture content havebeen found to affect extraction efficiency [14,15].

In ordinary passive extraction, heat is transferred to the vessel,which heats the solvent in order to desorb constituents from thematrix, which then diffuse into and become dissolved in the bulksolvent. Microwave heating makes use of the absorption ofmicrowaves by the solvent and/or the test substance. Microwaves,which can range from 300 MHz to 300 GHz (2450 MHz is reservedfor commercial or domestic microwave ovens [15]), are a form ofelectromagnetic radiation composed of oscillating perpendicularmagnetic and electric fields. Interaction of polar molecules withthese fields results in dipolar rotation in order to facilitatemolecular alignment with the applied field. However, due to theoscillatory nature of the field, molecular dipoles ultimatelyundergo rotational oscillations, creating friction which is releasedas heat. Ionic solutions undergo a second heating mechanism thatproceeds by resistive heating of the solution following movementof the dissolved ions in the presence of the oscillating field.Ultimately, the passive (conductive/convective) heating processis by-passed, and heating may be applied in a focused manner andalmost instantaneously to the matrix undergoing extraction. Theresultant molecular motions result in desorption of boundanalyte, and subsequent dissolution in the surrounding solvent[15,16].

The ability to absorb microwave energy is material depen-dent. Energy absorption is related to the dielectric loss factor,which, in turn, is a function of the permittivity of the material.Generally, polar substances have a larger dielectric loss factorthan non-polar substances, and thus tend to absorb microwavesmuch more efficiently. Depending on the choice of extractionsolvent and the nature of the sample, heating can be focuseddirectly and almost exclusively on the sample itself and/or to thesurrounding solvent.

The application of MAE in forensic toxicology has beenextremely rare. Franke et al. discussed the use of MAE to facilitatethe extraction of drugs into an organic solvent mixture [17] andreported substantial increases in extraction recovery in somecases. In this study, we examined the utility of MAE for theextraction of model drugs from ground bone, as part of an effort tostreamline methodologies to improve the efficiency of studies ofdrug disposition in skeletal tissues. The objective of this work wasto examine the effects of target drug, solvent, irradiation time andsample mass on the time required for drug recovery, and tocompare this methodology against a passive methanolic extractionstrategy which has been used in our laboratory previously [1,2,5].This work involved the extraction of three model drugs (ketamine,diazepam and pentobarbital) from pooled, ground bone tissuesobtained from acutely exposed animals following a period of

significant decomposition. These drugs were chosen for theirforensic relevance as well as their varying chemical properties (e.g.,acid–base character). We used an open-vessel/atmosphericpressure configuration and a domestic microwave oven in theextraction process. Tissue extracts were analyzed by automatedELISA. We have shown recently [1,2,5] that ELISA is a very sensitiveand reliable method for screening of skeletal tissues for drugexposure. In this application, ELISA is valuable in that it provides asemi-quantitative response, which facilitates estimation of opti-mal extraction conditions, while allowing for parallel analysis ofsamples.

2. Methods

2.1. Chemicals

Methanol used in drug extraction was HPLC grade and purchased from EMD

chemicals (Gibbstown, NJ) Drug standards (Cerilliant, Round Rock, TX) were obtained

as 1 mg/ml methanolic solutions and diluted as required. All other chemicals were

reagent grade and were obtained from EMD chemicals (Gibbstown, NJ).

2.2. Animals and drug administration

Adult male Wistar were given 16 mg/kg diazepam (n = 1), 75 mg/kg ketamine

(n = 2) or 50 mg/kg pentobarbital (n = 2). Rats were euthanized with CO2 within

20 min. The remains were placed in secure cages, and were left to decompose

outdoors at a rural Ontario site during late summer (August–September), with full

exposure to sunlight and other weather features (precipitation, etc.). The mean

maximum and minimum daily temperatures were 23.8 and 11.8 8C, respectively,

and the total rainfall was 55.8 mm. Decomposition was allowed to proceed

naturally to the point of nearly complete skeletonization (only skin, fur and bones

remaining), and required roughly 3 weeks. Bones, including femora, tibiae,

vertebrae, pelvi, humeri, scapulae and ribs were collected. Drug-free animals were

also prepared in this fashion, to provide negative control tissues.

2.3. Bone preparation

Ketamine-exposed, pentobarbital-exposed and control bones were immersed in

a 0.5 M NaOH:NaCl (50:50) solution and cleaned by ultrasonication until no soft

tissue remained (approximately 1.5 h for femoral, tibial and pelvic bones and 3 h for

vertebrae). Solution was replaced once for leg bones and three times for vertebrae.

Diazepam-exposed bones were cleaned with a 0.5 M phosphate buffer, pH 8.5 (PB8)

solution because of observed degradation of the ELISA response following exposure

of the drug to the alkaline solution with ultrasonication. Bones were immersed in

PB8 and underwent ultrasonication for approximately 2 h. The solution was

replaced twice.

Once clean, the bones were washed twice with distilled water and once with

acetone. Bones were then dried under a gentle stream of argon at 50 8C. The

different bones were then pooled (according to drug) and pooled bones were then

ground into fine particles using a domestic grinder, resulting in a total mass of

approximately 9 g of each of the drug-exposed tissues. Glass screw-cap tubes

(20 ml) were used as extraction vessels. A single 1/8 in. hole was drilled in the top of

each cap and a glass capillary was inserted into the hole and held in place with

Teflon tape, in order to serve as a vent.

2.4. Passive extraction

Samples (0.5 g) of each bone were accurately weighed into screw-cap test tubes,

and 2 ml of methanol was added to each tube. Samples were incubated on a hot

plate at 50 8C (�1–2 8C). For each sample, at successive, defined intervals (1, 6, 12 and

24 h), the solvent was removed and replaced with clean solvent in order to monitor the

rate of extraction. Following each solvent recovery, bones were washed twice with

N.A. Desrosiers et al. / Forensic Science International 188 (2009) 23–30 25

1 ml of methanol and the methanol washes were pooled with the originally recovered

fraction. All recovered methanol fractions were then evaporated under a gentle stream

of argon at 50 8C. The samples were reconstituted in 2 ml of 0.1 M phosphate buffer, pH

6 (PB6).

2.5. Stability of drugs to microwave exposure—ELISA

The solvents examined in the extraction of each drug were PB6, methanol:water

(MeOH:H2O, 1:1, v/v) and methanol (MeOH). The stability of the ELISA response to

ketamine, pentobarbital and diazepam under microwave irradiation in each of

these solutions was examined. Solutions of ketamine (50 ng/ml), pentobarbital

(50 ng/ml) and diazepam (2.5 ng/ml) were prepared in each solvent and transferred

to capillary screw-cap test tubes. A household Danby microwave oven DMW1153W

(1100 W, 2450 MHz) equipped with a turntable was used for the extraction.

Solutions were then irradiated for 1, 2, 3, 4 and 5 cycles, where each cycle consisted

of successive 10 s intervals lasting for a total of 3 min. Between successive 10 s

irradiation intervals, the tubes were swirled to ensure good mixing of solvent, and

to liberate any dissolved gases in order to prevent boiling. MeOH and MeOH:H2O

were evaporated under argon at 50 8C, and then reconstituted in 1 ml PB6 for

analysis.

2.6. Stability of drugs to microwave exposure—gas chromatography/mass spectrometry

The stability of the drugs examined to microwave irradiation in PB6 and

methanol was also examined by gas chromatography/mass spectrometry (GC/

MS), in order to further investigate whether the drugs were decomposing under

microwave irradiation. Three solutions (1 ml) of ketamine (400 ng/ml),

pentobarbital (400 ng/ml) and diazepam (400 ng/ml) were prepared in each

solvent and transferred to capillary screw-cap test tubes. Chlorpheniramine

(200 ng/ml) was used as an internal standard for the ketamine and diazepam

analyses, while secobarbital (200 ng/ml) was used as an internal standard for the

pentobarbital analysis. Each solution was irradiated as described above in

successive 10 s intervals lasting for a total of 0 min (control), 7 min or 15 min.

MeOH was evaporated under argon at 50 8C, and then reconstituted in 1 ml PB6

for analysis.

For ketamine and diazepam, the solution pH was raised to approximately 10

with 4 M NaOH, and 5 ml ethyl acetate:toluene (1:1, v/v) was added. The mixture

was rotated for approximately 1 h. The organic phase was recovered, and

evaporated to dryness under argon (50 8C). The residues were reconstituted in

ethyl acetate (50 ml) and analyzed by GC/MS. For pentobarbital, 5 ml ethyl acetate

was added to the buffer solution and the mixture was rotated for approximately

1 h. The organic phase was recovered, and evaporated to dryness under argon

(50 8C). The residues were reconstituted in 50 ml of 0.2 M trimethylphenylam-

monium hydroxide in methanol (United Chemical Technologies, Bristol, PA),

which was then injected directly into the GC/MS to facilitate flash methylation of

the barbiturates. The GC/MS used was a PerkinElmer Clarus 600 (PerkinElmer LAS,

Shelton, CT), equipped with an Elite 5-MS column (30 m � 0.25 mm) and operated

in the electron impact ionization mode, using helium as the carrier gas at a flow

rate of 1 ml/min. The injection port temperature was programmed, with an initial

temperature of 60 8C, which was held for 3 min after an injection of 5 ml. The

injection port temperature was then increased 270 8C, with the split vent open and

a split flow rate of 50 ml/min. The initial column temperature was 60 8C, which

was held for 2 min, increased directly to 160 8C, and then increased linearly at a

rate of 10 8C/min to a final temperature of 300 8C, where it was held for 3 min. Each

drug was examined using selected ion monitoring (ketamine: tR 10.0 min; m/z

180, 167, 152; diazepam: tR 15.3 min; m/z 283, 256, 221; chlorpheniramine: tR

11.3 min; m/z 203, 167, 152; pentobarbital: tR 8.0 min m/z 184, 169, 112;

secobarbital: tR 8.3 min; m/z 196, 181, 169; ions for area comparison in bold). The

response ratio was determined as the peak area ratio of the appropriate ions from

the analyte and internal standard, where the peaks examined were required to

have tR values within 1% of the expected values. The stability of each drug was

examined by comparison of the response ratios of a given drug after being exposed

to microwave radiation for different periods of time (0, 7 or 15 min). The detection

limits for this method were approximately 15, 20, and 15 ng/ml for ketamine,

diazepam and pentobarbital, respectively. At concentrations of 25 and 100 ng/ml,

the precision (%CV) response ratios was 13% and 3%, respectively, for ketamine,

17% and 6%, respectively, for diazepam, and 12% and 3%, respectively, for

pentobarbital.

2.7. Effect of extraction solvent

Samples (0.5 g) of each bone were accurately weighed into MAE tubes, and 2 ml

of MeOH, MeOH:H2O or PB6 was added to each tube. Samples were irradiated for

10 s intervals, separated by 30 s intervals during which the solvent was allowed to

cool to ensure that the boiling point was not reached. After six 10 s cycles (1 min),

the supernatant solvent was removed and transferred to a regular test tube. Further

(2–4) 1 min cycles were undertaken, for a maximum of 5 min of irradiation. MeOH

and MeOH:H2O were evaporated at 50 8C under a continuous flow of argon. Samples

were then reconstituted in 2 ml PB6 for ELISA analysis. For samples which used PB6

as the extraction solvent, the PBS was analyzed directly with ELISA.

2.8. Effect of sample mass

Bone samples (0.5, 1 and 2 g) derived from control and drug-positive animals

were accurately weighted into capillary screw-cap test tubes. PB6 (3 ml) was added

to each sample. Samples underwent microwave irradiation for 3 min, in cycles of

10 s. Solvent was transferred to a test tube and set aside. Another 3 ml of PB6 was

added to screw-cap tubes and the above irradiation pattern was repeated. The

above process of solvent recovery and replacement was repeated until samples had

undergone a maximum of 15 min of irradiation.

2.9. Enzyme-linked immunosorbent assay (ELISA)

Bone extracts were assayed for the appropriate drug with commercially

available ELISA kits (Immunalysis Corp., Pomona, CA) for barbiturates (group

assay), ketamine or benzodiazepines (group assay), as per the manufacturer’s

instructions. Immunoassay was automated using a ChemWell1 2910 Automated

EIA and Chemistry Analyzer (Awareness Technologies, Palm City, FL) Sample

(10 ml) was mixed with 100 ml of enzyme conjugate in the antibody-coated

microwells, maintained at an operating temperature of 25 8C. The plate was

shaken for 10 s and incubated for 1 h. The solution delivery probe was washed

with 1N HCl to remove any traces of enzyme conjugate. The wells were then

washed six times with 100 ml wash solution (0.025 M phosphate buffer (pH 7),

containing 0.025% Tween 20). After aspiration of the wash solution, 150 ml of

enzyme substrate (3,30 ,5,50-tetramethylbenzidine, TMB) was added to each well

and allowed to incubate for 30 min. The reaction was stopped by the addition of

50 ml of 1N HCl stop solution to each well. The absorbance of each well was

measured at 450 nm.

3. Results

3.1. Performance characteristics of the ELISA method for detection of

pentobarbital, ketamine and diazepam exposure—precision,

concentration dependence of response and cross-reactivity

For the purpose of this work, measured absorbance valueswere considered directly, and also transformed to give thedecrease in absorbance, relative to drug-free controls. Thefollowing formula was used to determine the percent decreasein absorbance

% decrease in absorbance ð%DAÞ ¼ Actrl � A

Actrl� 100% (1)

where A is the absorbance of a given sample and Actrl is theabsorbance of a matrix-matched sample of the same treatment.Consideration of %DA values has been adopted in earlier work[1,2,5] as a means of controlling for the effects of endogenouscross-reactants and differences in the ELISA response to variationsin tissue types. We continue to use it here for both consistency withour experimental approach, and to make the graphical descriptionof data more intuitive.

The precision of replicate analyses (% coefficient of variationof measured absorbance values) of matrix-matched spiked boneextracts ranged from 2 to 11% for the diazepam assay, 0.5–5% forthe pentobarbital assay and 0.9–6% for the ketamine assay.Limits of detection for the assays were approximately 0.25, 2.5and 2.5 ng/ml for diazepam, pentobarbital and ketamine,respectively.

3.2. Concentration dependence of ELISA response

Standard solutions of diazepam (0–10 ng/ml), ketamine (0–200 ng/ml) and pentobarbital (0–200 ng/ml) were prepared in 1 mlsolutions of bone extracts derived from drug-free animals andanalyzed by ELISA. Fig. 1 illustrates the concentration dependenceof all three drug assays.

3.3. Passive extraction in methanol

Drug-positive bone and drug-free control bone samples (0.5 g)underwent passive methanolic extraction for purposes of

Fig. 1. Concentration dependence of ELISA response. Relative decrease in

absorbance (%) measured using ELISA method for (A) pentobarbital, (B)

ketamine, (C) diazepam as a function of drug solution concentration in spiked

extracts of bone.

Fig. 2. Incremental relative decrease in absorbance measured following successive

cycles of passive methanolic extraction of pentobarbital, ketamine and diazepam.

Fig. 3. Stability of ELISA response to (A) pentobarbital, (B) ketamine and (C)

diazepam following drug exposure to microwave irradiation, expressed as a relative

decrease in absorbance over time.

N.A. Desrosiers et al. / Forensic Science International 188 (2009) 23–3026

comparison relative to the microwave-assisted extraction meth-ods used. Fig. 2 presents the relative decrease in absorbance after1, 6, 12 and 24 h of passive extraction for bone samples derivedfrom animals exposed to pentobarbital, ketamine or diazepam,respectively.

3.4. Stability of assay response to drugs exposed to microwave

radiation

The effect of microwave radiation on the stability of the ELISAresponse to each drug was examined for each solvent typeinvestigated. The data are summarized in Fig. 3. The data show that

there was no clear loss in assay response to any of the drugsexamined. Further examination of drug stability to microwaveirradiation through the GC/MS assay showed that the responseratios of each drug after 7 or 15 min total irradiation time did notdiffer significantly from those of the control samples, whichcontained drug at the same concentration but did not undergo

N.A. Desrosiers et al. / Forensic Science International 188 (2009) 23–30 27

irradiation. Overall, these data indicate that the drugs examinedhere are stable under the conditions applied (i.e., choice ofsolvents, irradiation times, microwave power, etc.)

3.5. Effect of extraction solvent: comparison of phosphate buffer,

methanol and methanol:water mixture

The effect of the extracting solvent was examined. Phosphatebuffer (PB6), methanol (MeOH) and a methanol–water mixture(MeOH:H2O, 1:1, v/v) were used to extract 0.5 g of drug-positiveand drug-free control bone. Extraction continued in rounds of1 min until 5 cycles had been completed or the assay signalreached the limit of detection (Actrl—3S.D.) for each solvent. Fordiazepam and ketamine, the extraction proceeded more quicklyusing MeOH and MeOH:H2O, than for PB6. However, signalmagnitude was greater for the PB6 solvent system, which may beassociated with a greater net drug recovery. In the case of

Fig. 4. Incremental relative decrease in absorbance measured following successive

cycles of MAE of (A) pentobarbital, (B) ketamine and (C) diazepam from 0.5 g bone,

in various extraction solvents.

pentobarbital, extraction was similar for all solvents. Extractionprofiles can be seen in Fig. 4.

3.6. Effect of bone mass on number of extraction cycles

Samples of control and drug-positive bone (0.5, 1 and 2 g) wereextracted in 3 ml of PB6. Irradiation was done in cycles of 3 min.Extraction continued until the relative decrease in absorbance of aspecific sample reached the limit of detection for at least one of thesamples (i.e., for the 0.5, 1 or 2 g sample). In the case of diazepam, thedetection limit was not approached for any of the samples after 5rounds (15 min), and so the extraction of diazepam from bonesamples of different mass (0.5, 1 and 2 g) was attempted with MeOH.Fig. 5 illustrates the effect of sample mass on the extraction time.

4. Discussion

Drug screening in skeletal tissues is an inherently complexprocess. The use of these tissues as an analytical matrix wouldtypically only occur in cases where there is little or no otheralternative; when decomposition processes have rendered conven-tional samples (blood, soft tissues) unavailable. Such decompositionprocesses are generally accompanied by a variety of environmentalstresses, such as water exposure, scavenging and temperaturefluctuations. Understanding the effects of these variables istremendously challenging as they are very difficult to control.

In this work, we made use of animal models (rats) in order tofacilitate control over drug exposure history, drug dose andpostmortem environment. One difficulty with the use of smallanimal models is the sensitivity requirements for determination oftrace quantities of drugs in small tissue masses. ELISA is a verysuitable approach for this experimental application, as it providesfor highly sensitive and semi-quantitative parallel analysis ofnumerous samples. It is acknowledged that the primary limitationof ELISA is the potential for interference from cross-reactingsubstances, including drug-metabolites or compounds endogen-ous to the sample matrix, and that a higher degree of analyticalselectivity may be achieved through methods such as GC/MS.However, such methods are less sensitive, and the selectivitylimitation was accounted for somewhat through the expression ofELISA response as the relative decrease in absorbance, measured asa percentage of matrix-matched, drug-free control tissues, asdescribed in equation (1). This corrects, to some extent, for thepresence of cross-reacting endogenous compounds that may bepresent in both the control tissues, and those derived from drug-positive animals. Indeed, we have exploited the sensitivity of ELISAto selectively detect acute drug exposure for each of thecompounds investigated here [1,2,5,9] in fragments of individualbones of acutely exposed rats.

4.1. Considerations for experimental design

As part of efforts to improve analytical sensitivity and samplepreparation time, the goal of this study was to examine the effect ofsome experimental factors (e.g., extraction solvent, bone mass,irradiation time, etc.) in microwave-assisted extraction of groundbone on the analytical response. This requires a homogeneoussample for analysis under these different conditions. As a solidmatrix, bone presents a particular challenge in this regard, sincewe noted in previous work [1,2,5,9] that there may be variation inanalytical response based on the tissue type sampled (i.e., corticalvs. trabecular bone and bone vs. marrow). Consequently, we choseto collect a variety of different bone types, which incorporate bothcortical and trabecular bone, and to pool them for analysis.Recovery of spinal and pelvic bone from freshly euthanizedanimals is difficult due to the significant amount of soft tissue, so

Fig. 5. Incremental relative decrease in absorbance measured following successive cycles of MAE using various masses of bone sample. (A) pentobarbital in PB6, (B) ketamine

in PB6, (C) diazepam in PB6 and (D) diazepam in MeOH.

N.A. Desrosiers et al. / Forensic Science International 188 (2009) 23–3028

allowing the animals to decompose to the point of skeletonizationfacilitated collection of these tissues, and also provided anapplication that was more relevant to forensic casework, sinceskeletal tissues are most likely to undergo toxicological analysiswhen there are no conventional samples (e.g., blood and liver)available. Finally, to facilitate maximal drug recovery and to createas homogeneous a sample as possible, the pooled bones wereground into a fine powder and mixed extensively.

In order to minimize drug contributions from residual marrowand soft tissue, the bones were vigorously cleaned prior to poolingand analysis, as we have described in earlier work [1,2,5,9]. It isacknowledged that this process may lead to some drug loss fromwithin the bone itself, and that this loss cannot be resolved from theremoval of drug-laden soft tissues or marrow. Further, it remainsunclear whether drug recovered from the cleaned bone andsubsequently detected has been absorbed directly into osteocytes,or is situated within the microscopic cannaliculi that exist within themineralized matrix. The process of decomposition, through theliquefaction of soft tissue, may facilitate further deposition of druginto the microporous network of the bone. Regardless of the sourceof the recovered drug, it seems that MAE provides some driving forcefor rapid dissolution of drug retained within this tissue into theextraction solvent for subsequent analysis.

4.2. Stability of drugs to microwave radiation

In order to verify that decreases in the %DA values observedfollowing successive irradiation cycles were due to progressiveextraction of the drug from the bone sample as opposed tobreakdown of the drug induced by the microwave radiation, the

stability of the ELISA response to standard solutions of each drug inPBS, methanol and methanol:water (1:1) was examined (Fig. 3).The data showed that the ELISA response was stable to irradiationfor all drugs and solvent systems examined. However, due to thepotential for cross-reactivity in immunoassay, it is possible thatsome chemical decomposition occurred and the decompositionproducts displayed significant cross-reactivity with the immobi-lized antibodies. Consequently, GC/MS was used to assay standardsolutions of each drug examined, prepared in PBS or in methanol,following microwave irradiation for 7 or 15 min, in 10 s intervals,as well as similar solutions which did not undergo irradiation. Theresults of those experiments showed no significant differences inanalytical response for samples irradiated for different periods oftime, suggesting that no significant breakdown of the drugsoccurred as a result of the exposure to microwave radiationdescribed here.

4.3. Extraction parameters

As can be seen in Fig. 4, there is a noticeable effect of theextraction solvent on the apparent rate of extraction and resultantsignal intensity. In the case of diazepam and ketamine, drugextraction appears to proceed more quickly in the MeOH andMeOH:H2O. For the extraction in PB6, diazepam extraction was stillnot maximal after 5 cycles of irradiation (60 s in duration). As shownin Table 2, it appears that the use of PB6 as an extraction solventprovided greater extraction yield. In the case of pentobarbitalextraction, the solvent effect was not as prominent with respect totime required for maximal extraction, but the use of PB6 appeared toprovide some advantage of in terms of relative recovery.

Table 2Semi-quantitative comparison of extraction recovery: ratio of sum of incremental

%decrease in absorbance from successive rounds of MAE relative to sum of

incremental %decrease in absorbance from successive rounds of passive extraction.

Drug R MAE (PBS)

vs. MAE

(MeOH)

R MAE (PBS) vs.

passive extraction

(MeOH)

R MAE (MeOH) vs.

passive extraction

(MeOH)

Pentobarbital 1.9 1.3 0.69

Ketamine 1.8 2.1 1.2

Diazepam 1.4 2.6 1.8

N.A. Desrosiers et al. / Forensic Science International 188 (2009) 23–30 29

In addition, the effect of mass on the extraction time wasexamined. As can be seen in Fig. 5, the number of cycles needed formaximal extraction increased with sample mass. In the case ofdiazepam, 15 min was not sufficient for maximal extraction from0.5, 1 or 2 g bone. For ketamine extraction, 15 min was notadequate for the maximal extraction from the 2 g sample.Therefore, as would be expected, increased extraction time isnecessary for larger bone samples.

Clearly, the optimal extraction procedure may be expected to bedrug-dependent. Under the conditions examined here, pentobar-bital was extracted from 0.5 g of bone within 3 min of PB6, whilediazepam and ketamine required longer irradiation times formaximal recovery from the 0.5 g sample. Also, when studying theeffect of bone mass, pentobarbital was the only drug that appearedto have been maximally extracted from 2 g of bone within 15 min.Conversely, the observed %DA values in assays of extracts fordiazepam from 2 g bone samples did not seem to display anysensitivity to the number of irradiation cycles initially, which wasattributed to the high sensitivity of the benzodiazepine ELISA andthe non-linear nature of response with respect to diazepamconcentration. Subsequent dilution of the samples resulted in theexpected profile. Another uncharacterized effect is the relative rateof extraction of diazepam metabolites that are known to have ahigh-degree of cross-reactivity with the ELISA used [5]. In thiswork, the aim was to illustrate the effects of various operationalvariables associated with MAE on the overall performance of theextraction, and so resolution of these cross-reacting metabolites isrelatively unimportant in this case, but would be of obviousimportance in the application of this methodology in forensiccasework, where the relative quantities of parent drugs andmetabolites may provide valuable information.

Overall, the variables examined here would affect the approachtaken if applied in forensic casework. Obviously, extractionmethods would require a separate validation for each solventused, and the extraction time would need to be selected accordingto the longest time necessary for maximum drug extraction of allpotential drugs under consideration. The volume of solvent usedshould also be examined as an extraction parameter.

4.4. Extraction recovery: MAE vs. passive extraction

The concentration dependent nature of the ELISA methodologyallows for a semi-quantitative comparison of the recovery of twopreparation methodologies. Examination of Fig. 1 shows that the%DA parameter is approximately proportional to drug concentra-tion over a limited range. For samples in this pseudo-linear range,two different extraction methodologies may be compared in termsof which provides the greater recovery through the ratio of the sumof incremental %DA values:

R ¼P

i%DAiPj%DA j

(2)

Values of R comparing MAE using PB6 or methanol as anextraction solvent to passive extraction were determined for each

drug, after any dilutions necessary to yield sample concentrationsinto the pseudo-linear region, and are summarized in Table 2. Thedata in Table 2 show that there was a tendency towards higherextraction recovery through MAE in PBS than from either MAE orpassive extraction in methanol for all three drugs, with the effectappearing to be more pronounced with ketamine and diazepam.Overall, while it may be that both the optimal extraction solventsystem and the magnitude of the benefit of MAE will be drug-dependent, direct extraction into PB6 or other phosphate buffersshould be considered as they may provide high recoveries withrapid extractions. Again, it is important to remember that thisapproach should be treated as being semi-quantitative only, due tothe potential for cross-reactivity effects.

4.5. Benefits of microwave-assisted extraction

The use of MAE offers a substantial advantage in terms ofextraction time, and may provide improvements in extractionefficiency. The data in Table 1 summarize some of the samplepreparation approaches used for toxicological analysis of bone asdescribed in the literature. Overall, the approaches used to datehave involved lengthy extraction steps, using a Soxhlet or passiveextraction configuration, which typically required many hours (6–72 h). Here, passive extraction of drug from ground bone required6 h or more (Fig. 2). Interestingly, the data in Fig. 2 show anunexpected result wherein the passive extraction of diazepamshowed a substantial increase between the 12 and 24 h samplingperiods. The analysis of this fraction was repeated with similarresults, ensuring this was not a randomly anomalous immunoas-say result. Given that the sample concentration in that sample wastoo low for GC/MS analysis, we cannot comment conclusively onthe source of this effect, although one potential explanation may bethat multiple sources of drug exist within the bone sample; onetype being more loosely adsorbed to the bone surface and morereadily diffusing under the extraction conditions, and anothersource which is more strongly bound within the matrix and whichis slower in diffusing into the extraction solvent. Even in theabsence of a conclusive mechanism, this phenomenon should beborne in mind and extra analysis of second extractions from agiven bone sample may be used to verify that maximal extractionhas in fact been achieved [1,4].

Overall, MAE may reduce the time required for maximalextraction to as little as 5 min, depending on the drug in question(and, most likely, the associated dose), extraction solvent and bonemass. Furthermore, while extraction into phosphate buffer may besub-optimal for a particular drug in terms of extraction time, theuse of buffer as an extraction solvent facilitates immediatesubsequent ELISA or extraction procedures (e.g., solid-phaseextraction, liquid–liquid extraction), eliminating the need for anintermediate solvent evaporation step. Even under sub-optimalconditions where maximal recovery is not achieved, sufficient drugmay be extracted within 1–3 min to give a positive response onELISA. If samples are positive, further extraction can be continuedand samples can be pooled and confirmed. This could lead to areduction in unnecessary extraction of negative samples and asignificant reduction in the use of organic extraction solvents.

4.6. Future work

The data presented here suggests that MAE may constitute avaluable methodology for the efficient toxicological analysis ofskeletal tissue samples. While the work here made use of animalmodels, as is common in toxicological research involving bonetissues, this methodology may find utility in forensic analysis ofhuman bones. Laboratory-grade microwave ovens facilitate the useof larger extraction vessels (e.g., 50 ml) which can accommodate the

N.A. Desrosiers et al. / Forensic Science International 188 (2009) 23–3030

larger sample masses that would be required for human boneanalysis. The use of closed-vessels similar to those used inenvironmental chemistry could prove useful to the extraction ofdrugs from skeletal tissues, as that configuration typically providesfor significantly elevated solvent temperatures, in excess of theirnormal boiling points [12,14–16].

5. Conclusions and summary

Overall, experimental parameters such as temperature, pres-sure and sample moisture content should be optimized for eachdrug. An important consideration for future work is the extractionbehaviour of drugs and their metabolites, as the ratio of parentdrug to a given metabolite may serve as an indicator of the patternof drug use (i.e., recent exposure, acute vs. chronic exposure, etc.).In so doing, MAE should prove to be a useful tool for the extractionof drugs from skeletal tissues, and its utility may be extended to therapid extraction of drugs from other solid matrices of forensicvalue, such as hair.

Acknowledgement

The authors wish to acknowledge the Natural Sciences andEngineering Research Council of Canada for financial support ofthis work.

References

[1] T.C. VandenBoer, S.A. Grummett, J.H. Watterson, Utility of immunoassay in drugscreening in skeletal tissues: sampling considerations in detection of ketamineexposure in femoral bone and bone marrow following acute administration usingELISA, J. Forensic Sci. 53 (2008) 1474–1482.

[2] J.H. Watterson, T.C. VandenBoer, Effects of tissue type and the dose-death intervalon the detection of acute ketamine exposure in bone and marrow with solid-

phase extraction and elisa with LC/MS/MS confirmation, J. Anal. Toxicol. 32 (2008)631–638.

[3] N. Wohlenberg, T. Lindsey, R. Backer, K. Nolte, Nortriptyline in maggots, muscle,hair, skin and bone in skeletonized remains, TIAFT Bull. 22 (1992) 19–22.

[4] I.M. McIntyre, C.V. King, M. Boratto, O.H. Drummer, Post-mortem drug analyses inbone and bone marrow, Ther. Drug Monit. 22 (2000) 79–83.

[5] J.H. Watterson, J.E. Botman, Detection of acute diazepam exposure in bone andmarrow: influence of tissue type and the dose-death interval on sensitivity ofdetection by ELISA with liquid chromatography tandem mass spectrometryconfirmation. J. Forensic Sci., doi:10.1111/j.1556-4029.2009.00998.x.

[6] L.Y. Gorczynski, F.J. Melbye, Detection of benzodiazepines in different tissues,including bone, using a quantitative ELISA assay, J Forensic Sci. 46 (2001) 916–918.

[7] E. Guillot, P. de Mazancourt, M. Durigon, J.C. Alvarez, Morphine and 6-acetylmor-phine concentrations in blood, brain, spinal cord, bone marrow and bone afterlethal acute or chronic diacetylmorphine administration to mice, Forensic Sci. Int.166 (2007) 139–144.

[8] E.L. Horak, A.J. Jenkins, Postmortem tissue distribution of olanzapine and citalo-pram in a drug intoxication, J. Forensic Sci. 50 (2005) 679–681.

[9] J.H. Watterson, M.S. Snow, Detection of Acute Pentobarbital Exposure in TibialBone and Marrow by Solid-phase Extraction, ELISA and Positive Ion ChemicalIonization Gas Chromatography–Mass Spectrometry, Society of Forensic Tox-icologists Conference, Phoenix, AZ, October 2008.

[10] N. Raikos, H. Tsoukali, S.N. Njau, Determination of opiates in postmortem boneand bone marrow, Forensic Sci. Int. 123 (2001) 140–141.

[11] S. Thompson, H. Budzinski, K. LeMenach, M. Letellier, P. Garrigues, Multi-residueanalysis of polycyclic aromatic hydrocarbons, polychlorobiphenyls, and organo-chlorine pesticides in marine sediments, Anal. Bioanal. Chem. 372 (2002) 196–204.

[12] V. Lopez-Avila, J. Benedicto, Microwave-assisted extraction combined with gaschromatography and enzyme-linked immunosorbent assay, Trends Anal. Chem.15 (1996) 334–341.

[13] J. Sastre, A. Sahuquilllo, M. Vidal, G. Rauret, Determination of Cd, Cu Pb and Zn inenvironmental samples: microwave assisted total digestion versus aqua regia andnitric acid extraction, Anal. Chim. Acta. 462 (2002) 59–72.

[14] C. Sparr Eskilsson, E. Bjorklund, Analytical-scale microwave-assisted extraction, J.Chromatogr. A 902 (2000) 227–250.

[15] M. Letellier, H. Budzinski, Microwave assisted extraction of organic compounds,Analysis 27 (1999) 259–271.

[16] V. Mandal, Y. Mohan, S. Hemalatha, Microwave extraction—an innovative andpromising extraction tool for medicinal plant research, Pharmacol. Rev. 1 (2007)7–18.

[17] M. Franke, C.L. Winek, H.M. Kingston, Extraction of selected drugs from serumusing microwave irradiation, Forensic Sci. Int. 81 (1996) 51–59.