the influence of collection site and methods on postmortem

Journal of Analytical Toxicology, Vol. 30, November/December 2006

The Influence of Collection Site and Methods on Postmortem Morphine Concentrations in a Porcine Model

Cameron S. Crandall Department of Emergency Medicine, Center for Injury Prevention, Research and Education, University of New Mexico Health Sciences Center, Albuquerque, New Mexico

Sarah Kerrigan* New Mexico Department of Health, Scientific Laboratory Division, Toxicology, Albuquerque, New Mexico

Roberto L. Aguero Center for Injury Prevention, Research and Education, University of New Mexico Health Sciences Center, Albuquerque, New Mexico

Jonathon LaValley Center for Injury Prevention, Research and Education, University of New Mexico Health Sciences Center, Albuquerque, New Mexico

Patrick E. McKinney ~ Department of Emergency Medicine, University of New Mexico Health Sciences Center, Albuquerque, New Mexico

Abstract ]

This study was to determine the relationship of antemortem to postmortem morphine concentrations in heart and femoral blood in a porcine model following acute intravenous opiate overdose. The study involved 20 swine; each was sacrificed 10 min after injection of 2 mg/kg body weight of morphine. Drug concentrations were assayed from vitreous humor and blood isolated from the femoral vein and artery and left and right ventricles at various times postmortem. Comparisons were made between antemortem and postmortem values to determine agreement and reliability. Both free and total postmortem values varied significantly among animals, sampling sites, and over time. Free postmortem values were generally higher in comparison with antemortem values, whereas postmortem total morphine values were similar to or slightly lower than antemortem values. The effect of time on postmortem values was small. These results demonstrate a significant amount of variability in free and total morphine measurements both over time and within and between sites. Furthermore, a comparison of antemortem to postmortem values demonstrates a lack of consistency relative to the dose of morphine administered. Concentrations of morphine in the femoral vein were typically the lowest observed. This observation is not surprising given the transformation that occurs prior to the

* Current address: College of Criminal Justice, Sam Houston State University, Huntsville, Texas. ~' Deceased.

drug reaching the femoral vein. Values associated with diffuse tissues, relative to femoral veins, demonstrate more stochastic variation.

Introduction

Fatalities due to opioid overdose are an increasingly common occurrence in the United States (1-3). The death rate from unintentional heroin poisoning and more recently, prescription opioid analgesics, is rising not only in the U.S., but in many other parts of the world as well (2,4,5). According to 1999 data from the American Association of Poison Control Centers, over 10,000 cases of opioid exposures were reported to poison centers in that year, and 100 of these cases resulted in fatalities (6). This finding clearly underestimates the problem because many cases of illicit drug toxicity are likely unreported (2). In addition to overdose cases, opioids can be in- cidental findings in cases of death involving medical care and treatment. Opioids detected as part of the postmortem examination are reported as postmortem drug concentrations and add to the circumstances surrounding the death to con- tributing to the determination of cause. Most importantly, postmortem drug concentrations do not necessarily reflect antemortem drug concentrations. Rather, careful sampling,

Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission. 651 Downloaded from https://academic.oup.com/jat/article-abstract/30/9/651/740683by gueston 09 April 2018


review of case information, postmortem interval, and a knowl- edge of the specific properties of the drug facilitate interpre- tation on a case-by-case basis. Comparison data using antemortem and postmortem morphine concentrations in humans are limited, and the existing animal data are scant and offer conflicting results.

Morphine Postmortem morphine can be present from a number of

sources including legitimate prescriptive use or abuse of licit and illicit opioids. Investigation of postmortem factors such as site dependence and redistribution tend to focus on morphine as the target opiate. Reports of morphine-related deaths indi- cate that morphine concentrations (free, bound, and total) in postmortem specimens vary greatly (7-9). Furthermore, blood morphine concentrations in morphine-related deaths (as de- termined by medical the examiner) overlap with those found in non-drug-related deaths, such as homicides and suicides, and compliant therapeutic use (9).

Researchers often use small animal models for convenience of large sample sizes for replication and shorter intergenera- tional periods (10-13). However, despite the difficulties associated with a large animal model (e.g., swine, primate), they more closely simulate the changes that occur in humans than small animal models (e.g., rat, rabbit), thus minimizing any ex- aggeration caused by diffusion or drug redistribution as a result of biological scaling (9,11,13,14). In addition, repeated sampling by moderate volume aspiration (2-3 mL) is less likely to result in significant contamination of blood from regions distal to the sample site in a large animal model. Using paired peripheral arterial and venous sites and left and right heart ventricles facilitates the evaluation of sampling specific sites on postmortem morphine concentrations. Furthermore, these data may provide valuable information regarding the postmortem/antemortem relationship and allow practical evaluation of basic issues of sample collection.

Morphine in vivo Intravenous morphine has relatively rapid distribution,

metabolism and excretion (15,16). After absorption, morphine leaves the circulation rapidly, and the distribution phase half- life for IV morphine is 0.9-2.5 min (17). Morphine readily dis- tributes throughout the body to highly perfused tissues such as lungs, kidney, liver, spleen, and muscle (18,19). Morphine binds to albumin and gamma globulin; binding is independent of dose, but somewhat dependent on protein concentration (20). This hydrophobic nature of the molecule makes it less "bioavailable" and more difficult to predict its reactions in vivo (7,8). Femoral blood is widely accepted as the most reliable postmortem specimen for drug analysis in forensic toxicology (19).

Factors such as age, intercurrent disease, and concomitant drug use also influence morphine pharmokinetics (21,22). Al- though the therapeutic clinical range for morphine is suggested in the literature [2-10 mg oral or intravenous, or approximately 150 ng/kg bodyweight (23)] and the minimum effective concentrations estimated (15-65 ng/mL) (7,8), these values do not represent all clinical scenarios because of the de-

velopment of tolerance to the respiratory and analgesic effects. Although peripheral blood morphine concentrations are reported as closer to the antemortem dose (19), there is also ample concern about the instability of morphine and its metabolites, such that postmortem concentrations make unreliable indicators of antemortem values. Determining the administered dose of morphine from postmortem sampling is not recommended (7,8).

Pig model Morphine is a commonly used analgesic drug in humans,

but it is used less frequently in animals because it causes agi- tation in many species (8). A review of standard references of drug dosing in animals reveals no pharmacokinetic studies defining distribution and elimination half-life, volume of distribution, or pattern of metabolite formation after morphine administration in domestic swine. Consequently, pharmacokinetic parameters of morphine remain understudied in many species commonly used as experimental models.

Pigs are generally considered the closest immunologic and metabolic model to humans, barring primates, and afford cer- tain considerations because of biological scaling not available in rodents (24). Turnblad and colleagues (25), when studying meningitis in swine, noted the administration of intravenous morphine hydrochloride to yield 1 mg/kg with an infusion over 10 min. This amount equates to about half of what we em- ploy in this study. As a first step, we measured basic pharmacokinetic parameters after intravenous injection of a single dose (data not presented) as a prelude to our use of the porcine model to look at postmortem changes in morphine concentrations.

The purpose of this study was to clarify the relationship between ante- and postmortem concentrations of morphine in a controlled study. Understanding this relationship influences how antemortem exposure to morphine is inferred from postmortem toxicology results. The present study measured drug concentrations in postmortem femoral and heart blood samples from swine and compared them with the concentration found in whole-blood samples from the same subjects collected antemortem.

Methods

Sampling animal model An overview of the experimental schema is shown in Figure 1. Twenty New Hampshire swine (Four Daughters Co.)

weighing between 50 and 70 kg were housed in the institutional animal care facility at the University of New Mexico Health Sciences Center. Animals were fed standard pig chow (Purina) and water ad lib. The animals were acclimated to a 12-h diurnal cycle. After a 7-day observation period, animals were fasted for 12 h and then sedated with telazol (4 mg/kg) and xylazine (2 mg/kg) administered by intramuscular injection (IM). After se- dation, a peripheral intravenous (IV) catheter was inserted in an auricular vein. The animals were intubated with an endotra- cheal tube and maintained under general anesthesia with isoflu-

652 Downloaded from https://academic.oup.com/jat/article-abstract/30/9/651/740683by gueston 09 April 2018


rane. The left femoral vein and artery were surgically exposed under general anesthesia. Swine were injected with morphine sulfate, administered through the ear vein at 2 mg/kg, 30 rain prior to sacrifice. A limited thoracotomy was performed to allow visual identification of the left and right ventricle. Ante- mortem specimens were collected approximately 20 min after administration (10 min prior to sacrifice), including blood as- pirated from the femoral vein and artery and the left and right ventricles (all under direct visualization), as well as vitreous humor specimens from the left eye.

Animals were sacrificed by inhatational anesthesia/thoracotomy with death confirmed by cessation of respirations and cardiac activity on visual inspection. Animals were placed in

8

Figure 1. Experimental sampling schematic.

Table I. Number of Samples Collected by Site and Time of Collection

Arterial Venous (Post-Pulmonary) (Pre-Pulmonary)

the supine position for 8 h at room temperature. Following this 8-h interval, cadavers were stored with minimal manipu- lation at 4~ representing morgue storage conditions. Ap- proximately 3-mL blood samples were drawn from the left and right ventricles and the femoral vein and artery at postmortem time points of 5, 30, 60, 120, 240, 480, 720, 1440, 2880, and 5760 rain. Samples were obtained by needle aspiration under direct visualization. Postmortem changes in the conditions of the animals at times inhibited the collection of all samples (e.g., collapse of a blood vessel). When feasible, 3-mL samples were drawn while avoiding the risk of contamination from proximal vessels. In all, 418 specimens were obtained (Table I).

Vitreous samples were obtained at the postmortem times of 1, 4, 24, and 96 h. Because of sample volume limitations, only a single sample was available from each eye, yielding a maximum of 40 samples. Of these, a total of 35 samples pro- vided sufficient volume for quantitative analysis (Table I).

Control

Antemortem

I I I

Total number of samples 119 Grand total

Time Central Peripheral Central Peripheral Thoracic (min) (LV*) (FA) (RV) (FV) Vitreous Pool

-30 19 . . . . .

-20 19 20 6 20 17 -

5 2 2 2 1 - -

30 5 5 5 5 3 -

60 2 2 2 2 1 -

120 7 7 7 7 2 -

240 1 1 1 1 1 -

480 12 6 11 12 3 -

720 1 - 1 1 - -

1440 16 8 16 15 4 -

2880 19 8 20 20 4 -

5760 16 2 17 14 - 17

61 88 98 35 17 418

* Abbreviations: LV, left ventricle; FA, femoral artery; RV, right ventricle; and FV, femoral vein,

Specimen preparation and storage Blood samples were collected in gray-top

Vacutainer TM tubes containing sodium fluo- ride and potassium oxalate. Vitreous fluid was collected in sterile Vacutainer tubes. No preservative was added. Samples were refrig- erated at 4~ upon collection.

Morphine quantification Quantitative free morphine analysis was

conducted using solid-phase extraction and gas chromatography-mass spectrometry (GC--MS). Silanized glassware and deuterated internal standards were used throughout. Morphine-d3 (Cerilliant) was added to 1 mL of sample to give a final concentration of 0.25 mg/L. Blood proteins were precipitated by addition of 2 mL of cold acetonitrile while vortex mixing. Samples were centrifuged for 10 min at 4000 rpm. The supernatant was transferred to a clean glass tube, and I mL of 0.1M HCI was added. Acidified samples were filtered by gravity using Cerex Clin II solid-phase extraction columns (SPEWare) and were washed with 1 mL deionized water, 0.1 M HCt, methanol, and ethyl acetate in that order. Columns were then dried under vacuum for 5 rain. Morphine was eluted from the column using I mL methylene chloride/isopropyl alcohol (80:20) containing 2% concentrated ammonium hydroxide. Extracts were evapo- rated to dryness under nitrogen. Samples were reconstituted using 15 pL ethyl acetate and 15 pL _N-methyl-_N-trimethylsilyltrifluo- roacetamide (MSTFA) followed by derivatiza- tion at 70~ for 15 min.

For total morphine determination, deuterated morphine-3-glucuronide was used as the

6 5 3 Downloaded from https://academic.oup.com/jat/article-abstract/30/9/651/740683by gueston 09 April 2018

internal standard. Following addition of 0.25 mg/L morphine- 3-glucuronide-d3 (Cerilliant), samples were incubated for 3 h at 37~ with 5000 IU of E. coli type IX, ~-glucuronidase (Sigma). After deconjugation and cooling to room temperature, samples were assayed as described.

An HP 5973 MSD with 30-m DB-5 column (0.25~mm i.d., 25- lJm film thickness) was used for separation and identification. A 2-1JL splitless injection was delivered onto the column run according to the following program: 160~ ramped at 30~ to 260~ for 2 rain; 30~ to 290~ for 5 rain. In- jector and interface temperatures were 250 and 280~ respectively. Data were acquired by selective ion monitoring (m/z 429, 324, and 287 for morphine and 432, 327, and 291 for morphine-d3; quantitation ions are shown in bold). Calibration curves were constructed between 0.010 and 1.000 mg/L by fortification of bovine blood with the appropriate drug standard. R 2 values in the linear range were 0.999 or greater. The limits of detection and quantification were 0.005 and 0.010 rag/L; these were defined as the lowest concentration of drug that produced a signal-to-noise ratio of 3:1 or 10:1, respectively, and with ion ratios with acceptable limits (+_ 20%). Accuracy was 93-97% using in-house controls fortified with free morphine between 0.015 and 0.950 mg/L and 89-104% for total morphine in the range 0.067-0.417 mg/L.

Replicate analysis of a commercial whole blood control (Utak Laboratories) indicated 104% accuracy and an intra-assay CV of 6.6% (n = 4) at 0.1 mg/L. A commercial free morphine whole blood control (0.1 rag/L) was included with each run in addition to in-house controls for both free (0.25 rag/L) and total (0.25 mg/L) morphine.

Statistical analysis Data were analyzed using both the raw (untransformed) and

natural log transformed values for both free and total morphine concentrations. The natural log transformation was used as


this transformation is frequently applied to concentration data. The transformation tends to make the distribution of the values symmetric about the mean as well as stabilizing the variance (it amplifies the variance at low values while reducing the variance at high values, reducing heteroscedasticity).

As a measure of agreement between sampling sites, we calculated ratios on untransformed values. As a measure of overall relativity, site values for both free and total morphine concentrations were compared within each animal to their corresponding level at the antemortem measurement (10 rain before euthanasia).

Analyses of data from vitreous samples used both left and right eyes as interchangeable. To test for the effect of time to sampling on morphine concentration, the antemortem to postmortem ratios were regressed against the number of hours to sampling. The slope of the regression line approximated the ap- parent decline of morphine over time.

Inference testing for morphine values was conducted only on the log transformed data. A generalized linear modeling ap- proach was used to test for differences in the mean values of morphine concentrations between the different sampling sites. Tukey's Honestly Significant Difference (HSD) post-hoc pro- cedure was used to adjust for multiple comparisons.

To test for site and time dependent differences, we conducted a repeated measures analysis of variance. Because of the limited samples available at some anatomic sites (e.g., thoracic pool and vitreous samples) and for the incomplete time sampling strategy, we limited our repeated measures analysis of variance to samples collected from the femoral vein, left ventricle and right ventricle at time 0 (antemortem), 8 h, 24 h, and 48 h (postmortem).

For all statistical inference, we used a two-tailed Type I error rate of 5% or calculated 95% confidence intervals about the point estimate.

Data were entered into a Microsoft Excel spreadsheet, then

Table II. Mean Concentration Values of Free and Total Morphine (rag/L) by Sampling Site and Time of Sampling

Free Morphine Total Morphine


Time Central Peripheral Cenlral Peripheral (rain) (LV*) (FA) (RV) (FV) Vitreous


Thoracic Central Peripheral Central Peripheral "[horacic Pool (LV) (FA) (RV) (FV) Vitreous Poor

Control

Antemortem

-30 0.000 - -

-10 0.424 0.407 0.394 0.576 0.138

5 0.911 1.071 0.994 1.060 -

30 0.731 0.861 0.636 1,408 -

60 1.061 1.346 1.071 1.039 0.439

120 0.714 0.676 0.382 0.742 -

240 1.012 1.639 0,982 1,619 0.306

480 0.706 0.461 0.565 0.778 0,198

720 1.746 - 1.865 1.718 -

1440 0.836 0.272 0.687 0.849 0.172

2880 0.790 0.242 0.768 0.910 0.201

5760 0.676 0.686 0.61t 0.592 - 0.947

* Abbreviations: LV, left ventricle; FA, femoral artery; RV, right ventricle; and FV, femoral vein.

0.000 - - - 3.744 3.656 5.584 3.888 0.880 -

6.095 5.841 4.331 4.803 - -

3.794 4.628 4.331 4.954 0.333 -

7.089 4,952 6.060 6.799 1.342 -

4.582 3.633 4.468 4.982 0.479 -

5.671 8,413 5.134 7.266 0.446 -

4.369 2.351 4.718 3.333 0.383 -

3.194 4.516 4.121 - -

4.107 0.819 3.992 2.273 0.586 -

4.321 1.714 3.577 2.924 1.108 -

3.347 3.009 3.322 2.286 - 3.704


Journal of Analyt ical Toxicology, Vol. 30, November /December 2006

transferred to SAS (SAS Institute, version 9.0, Cary, NC) for data management and analysis. Graphical analyses were conducted in S-Plus (Insightful, version 6.2, Seattle, WA).

R e s u l t s

Table I details the number of specimens per site and time of sampling. Table II provides the mean free and total morphine concentrations per site and time of sampling. There was con- siderable variation in values for both free and total morphine regardless of site.

In order to assess the ability of postmortem drug concentration to predict an antemortem value, we compared animal and site-specific antemortem values to corresponding postmortem values. These ratios are presented graphically in Figure 2. The dashed horizontal line presents the expected value if there were a one-to-one correspondence of ante- and

Rat b of pos~nortem free morphine vabes to antermrtem values, by site and ~ me

.10 1440 ~ 57E~

o

=

, ~ , 3 ~ , : , : -o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ; . . . . . . . . . . . . . . . . . .

~ FA; femoral artery =

o FV: ~emoral veh ~ LV: left ven~ic~ =

o ~ o o = RV: r t ~ ventricle o~ ~ ~ ,ll ~ ~

or =~ o o

= , , , , , = , , , , , ,

-~0 ~,~ ~ ~o Sampli"~ tire (rm)

Ratio of pos~'nortemtotal morphine values to a n t e ~ values, by sle and lime

o

o

. . . . . . . . -

o *~, oo

o

=

- : ~- ......... t % ' : ~ o ~ d o

.o. [i ~. ~176

FA: femoral artery FV: femoral veh LV: left ventricle RV: right ventricle VI: vitreous

Samplng the (n'~)

Figure 2. Plots of the ratio of postmortem free and total morphine values (mg/L) to corresponding antemortem values by site and sampling time.

postmortem values. The solid line presents the least squares regression line of the data. In general, postmortem free morphine values were on average higher than antemortem values (the regression line is above the dashed line); whereas postmortem total morphine values were similar to or slightly lower than antemortem values. There was little overall effect of sampling time on the observed postmortem to antemortem ratios, and the effect was not consistent between the different sites. Among the free morphine value ratios, the femoral artery and right ventricle means increased over time; the left ventricle and femoral vein stayed the same; and the vitreous declined slightly (Table II). Among the total morphine value ratios, the right ventricle mean values increased over time, while the other sites decreased (Figure 2). Only the vitreous site declined sub- stantially.

Finally, we conducted analyses to account for the repeated measurements over within the same animal. Free morphine values differed over the sampled time values (p = 0.011) but did not differ between the sampling sites (p = 0.63). Total mor-

phine values did not differ over time (p = 0.62) or between the sampling sites (p = 0.06).

D i s c u s s i o n

Postmortem morphine values demonstrate substantial variation, across both sampling site and time from death. While in population studies it may be possible to conclude that on average postmortem free morphine values are higher (and total morphine values are similar to or are slightly lower) than antemortem values, and that values tend to decline over time, the tremendous variation observed across individual sites and times demonstrates that it is not possible to draw definitive con- clusions about any single measurement in a given individual.

Postmortem examinations of humans, which have attempted to relate postmortem drug concentrations to the antemortem circulating concentration or antemortem dose received, have generally demonstrated that postmortem drug concentration in blood are highly variable, but are often as high as, or higher than, the antemortem circulating plasma drug concentration at the time of death (16,19,26). In several cases within the described study, the difference observed between the two concentrations was significant. Previous studies, citing the wide array of endogenous and exogenous factors that could influence the relationship between antemortem and postmortem morphine concentrations, suggest that this variability precludes the development of a general rule in this re- gard. There are numerous endogenous and

6 5 5 Downloaded from https://academic.oup.com/jat/article-abstract/30/9/651/740683by gueston 09 April 2018

Journal of Analytical Toxicology, Vot, 30, November/December 2006

exogenous factors which could possibly account for the variation seen. Comparing central (C) blood (e.g., heart) to peripheral (P) fluids exhibits another source of variance. Other variables might influence the disparity between central to peripheral concentrations, such as the interval between death and postmortem examination. This factor largely affects site-specific postmortem drug concentrations (27). It is likely that this interval would also affect the postmortem/antemortem drug concentration ratios calculated for these drugs, although drug concentrations in the femoral vein after death appear to be relatively stable with time (27) again making any extrapo- lation using these data unreliable. In an autopsy series, Logan and Smirnow (28) demonstrated that free morphine concentrations from femoral and left ventricular blood were generally in agreement at lower concentrations but varied greatly at higher concentrations. Since untransformed data tend to coa- lesce around zero and demonstrate greater variability at higher concentrations, it is possible that their conclusion is influ- enced by the lack of statistical correction. The finding that drugs exhibiting a high C/P ratio tend to have a high postmortem/antemortem ratio might prove useful in dealing with future cases in which poisoning may be a feature. Despite the variability, the range of quantified morphine we report (approximately 0.5-30.0pM) is consistent with previous authors results (25).

The factors that influence postmortem redistribution are well-established, and a number of authors (29,30) have published useful lists of CIP ratios. It is often desirable to determine whether to attribute the drug concentration found at postmortem examination to therapeutic ingestion versus overdose. The difficulty is mostly attributable to the influences of postmortem change and thereby do not support the use of postmortem/antemortem ratios, or direct inference from a postmortem concentration (19,27). In addition to this point, a significant, although subjective, factor in attribution of a postmortem blood concentration of an opioid to either therapeutic use or overdose is whether the medical examiner can es- tablish whether tolerance existed and, if so, to what degree.

There are a number of human and animal studies regarding forensically significant postmortem redistribution of pharma- ceutical agents (e.g., digoxin, tricyclic antidepressants) (11,12,27,29). Authors propose several theories to explain these changes in drug concentrations after death at various body sites. A drug may be released from tissues with high antemortem drug concentrations and diffuse to adjacent areas after death (30). In cases of oral overdose, a drug may diffuse through the stomach wall into surrounding blood and tissue. Postmortem flow of blood due to pressure changes in the ab- dominal compartment during decomposition has also been described (31,32). Drugs may also repartition as the pH de- creases, as occurs with death and decomposition (32). In addition, drug metabolism due to hydrolysis, continued enzyme activity, or bacterial metabolism may alter drug concentrations (33-35). It is not clear which of these factors, if any, is predominantly responsible for postmortem drug redistribution or if various combinations may predominate depending on the characteristics of the drug and conditions of the body and the postmortem setting. Regardless, it is clear that the majority

of these postulated mechanisms are dependent on the diffusion of a drug with a concentration gradient.

Published data specifically describing postmortem redistribution of morphine, the prototypical opioid, are limited to two categories, a small number of small animal studies and a number of autopsy reviews of heroin (and or morphine) deaths that are severely limited by the absence of antemortem morphine concentrations for comparisons. Published animal data on postmortem morphine distribution include studies using rats, rabbits, guinea pigs, and mice. Sawyer and Forney (10) conducted a series of experiments in rats evaluating postmortem morphine concentrations over time at various body sites while comparing these values to antemortem orbital sinus blood morphine concentrations. Sawyer and Fomey (10) found that postmortem heart-blood morphine concentrations increase when compared to antemortem values, even with postmortem intervals as short as 5 rain. Morphine concentrations continue to rise with increasing postmortem intervals, espe- cially in heart-blood, heart tissue and liver. Authors attribute increasing levels of free morphine over time to conversion or to repartitioning of morphine into areas of lower pH such as the blood. In vitro studies have also suggested that morphine metabolites may undergo postmortem hydrolysis (36). Con- centration changes in postmortem blood were reported during extended storage for 124 days at 4~ However, the authors concluded that the changes were inconsistent and highly variable between samples. Although the extent of possible deconjugation during the postmortem interval is not known it absolute terms, the magnitude of the change is likely small given the study conditions described here.

Koren and colleagues (11,12) examined postmortem morphine distribution in a rat model. Animals received intramuscular morphine prior to sacrifice and free morphine concentrations collected at 0, 24, and 96 h postmortem. Heart- blood morphine concentrations increased 2.7-fold on average by 24 h with no further significant increase at 96 h. Because of a lack of true antemortem concentrations, the differences observed in the study may underestimate the magnitude of change.

Schmidt et al. (37) reported on postmortem morphine redistribution in a rabbit model. In contrast to previous studies, they reported that heart-blood morphine and metabolite concentrations decreased to 70% of antemortem values after a 12-h postmortem interval. Unfortunately, because of the small sample size (n = 6), it is difficult to make statistical comparisons. However, these studies do allude to the limitations of the small animal model. Studies in a rat model demonstrate changes in morphine concentration over time, a finding con- trary to those observed in the rabbit study, possibly reflecting the effect of relative animal size on the degree of drug diffusion. Other limitations of note include requiring large sample vol- umes or repeated samples from a single site. In a small animal, this increases the risk that blood from adjacent vessels may contaminate samples collected. For example, the volume of blood in the femoral vein in a rat is small and large volume as- pirations might potentially draw blood from the iliac vein or in- ferior vena cava. These studies typically refer to sample sites as "femoral vessel" or "heart," with no distinction between heart



chamber, artery, or vein. Drug concentrations can vary according to whether an investigator samples blood from an artery or its paired vein (33,34). In addition, left heart drug concentrations can differ from the right heart drug concentrations (38,39).

Small animal models have limited use for understanding human morphine distribution. Consider two factors relevant to diffusion, distance and the permeability (or thickness) of tissues. In small mammals, these barriers to diffusion may be minimal. Diffuse organs that concentrate opioids, such as the lungs, liver, and stomach, are within millimeters of the heart and great vessels, the traditional postmortem sampling sites. In addition, organ walls and tissue barriers like the diaphragm are relatively thin in small animals, again minimizing diffusion barriers. Because of these factors, small animal models would likely amplify any postmortem distribution effects due to diffusion; however, they are all that are currently available for comparison. Therefore, it seemed logical that studies to resolve issues of postmortem distribution for specific drugs in humans should also occur in large animal models (e.g., swine) to reduce the issues related to biological scaling.

A couple of factors likely contributed to the unexpected variation in these data. We observed that antemortem morphine concentrations, relative to fixed inoculation concentrations (2 mg/kg morphine), were highly variable. This demonstrates that high variance in postmortem values are not a postmortem artifact as alluded to in this and other studies. Even within in the same animal, morphine concentrations were variable among sampling sites relative to time (same time, different sites). These data do support that among the many deaths oc- curring annually from opiate intoxication, attempts to inter- pret antemortem dose from postmortem values are problematic as concentrations detected postmortem vary with the individual, location of sampling, and time after death.

Acknowledgmenis

This work was funded in part by an unrestricted grant from Purdue Pharma L.P. to Dr. Patrick McKinney, who passed away in 2003. We gratefully acknowledge the assistance of Donna Honey at the New Mexico Department of Health, Scientific Laboratory Division, Toxicology Bureau for performing the quantitative analyses.

References

1. Drug Abuse Warning Network (DAWN) data http://dawninfo. samhsa.gov/.

2. E.J. Cone, R.V. Font, J.M. Rohay, Y.H. Caplan, M. Ballina, R.F. Reder, D. Spyker, and J.D. Haddox. Oxycodone involve- ment in drug abuse deaths: a DAWN-based classification scheme applied to an oxycodone postmortem database containing over 1000 cases. J. Anal. Toxicol. 27:57-67 (2003).

3. L.A. Fingerhut and C.S. Cox. Poisoning mortality 1985-1995. Public Health Rep. 113:219-235 (1998).

4. J. Neeleman and M. Farrell. Fatal methadone and heroin over- doses: time trends in England and Wales. J. Epidemiol. Commu- nity Health 51: 435-437 (I 997).

5. W. Hall and S. Darke. Trends in opiate overdose in Australia 1979-I 995. Drug Alcohol Depend. 52:71-77 (1998).

6. T.L. Litovitz, W. Klein-Schwartz, S. White, D.J. Cobaugh, J. Youniss, A. Drab, and B.E. Benson. 1999 Annual report of the American association of poison control centers toxic exposure surveillance system. Am. J. Emerg. Med. 18:517-574 (2000).

7. C.M. Bernards, D.D. Shen, E.S. Sterling, J.E. Adkins, L. Risler, B. Phillips, and W. Ummenhofer. Epidural, cerebrospinal fluid, and plasma pharmacokinetics of epidural opioids (part I): differences among opioids. Anesthesiology 99:455-465 (2003).

8. C.M. Bemards, D.D. Shen, E.S. Sterling, J.E. Adkins, L. Risler, B. Phillips, and W. Ummenhofer. Epidural, cerebrospinal fluid, and plasma pharmacokinetics of epidural opioids (part 2): differences among opioids. Anesthesiology 99:466-475 (2003).

9. F. Moriya and Y. Hashimoto. Distribution of free and conjugated morphine in body fluids and tissues in a fatal heroin overdose: is conjugated morphine stable in postmortem specimens? J. Forensic ScL 42:736-740 (1996).

10. W.R. Sawyer and R.B. Forney. Postmortem disposition of morphine in rats. Forensic Sci. Int. 38:259-273 (1988).

11. G. Koren and J. Klein. Postmortem redistribution of morphine in rats. Then DrugMonit. 14:461-463 (1992).

12. G. Koren and S. MacLeod. Postmortem redistribution of digoxin in rats. J. Forensic ScL 30:92-96 (1985).

13. P. Dechelotte, A. Sabouraud, P. Sandouk, I. Hackbarth, and M. Schwenk. Uptake, 3-, and, 6-glucuronidation of morphine in isolated cells from the stomach, intestine, colon, and liver of the guinea pig. Drug Metab. Dispos. 21 �9 13-17 (1993).

14. E.J. Cone, W.D. Darwin, W.E Buchwald, and C.W. Gorodetzky. Oxymorphone metabolism and urinary excretion in human, rat guinea pig rabbit and dog. Drug Metab: Dispos. 11:446-450 (1983).

15. S.F. Brunk, M. Delle, and W.R. Wilson. Morphine metabolism in man: effect of aspirin. Clin. Pharmacol. Ther. 15:283-290 (1973).

16. B.A. Goldberger, E.J. Cone, T.M. Grant, Y.H. Caplan, B.S. Levine, and .I.E. Smialek. Disposition of heroin and its metabotites in heroin-related deaths. J. Anal Toxicol. 18" 22-28 (1994).

17. K.M. Foley. The treatment of cancer pain. N. Engl. J. Med. 313: 84-95 (1985).

18. O.H. Drummer and J. Gerostamoulos. Postmortem drug analysis: analytical and toxicological aspects. Ther. Drug Monit. 24: 199-~209 (2000).

19. A.L. P61issier-Alicot, J.M. Gaulier, P. Champsaur, and P. Marquet. Mechanisms underlying postmortem redistribution of drugs: a review. J. Anal. Toxicol. 27:533-544 (2003).

20. G.D. Olsen. Morphine binding to human plasma proteins. Clin. Pharmacol. Then 17:31-35 (1975).

21. P.A. Glare and T.D. Walsh. Clinical pharmacokinetics of morphine. Ther. DrugMonit. 13:1-23 (1991).

22. P.A. Glare, T.D. Walsh, and C.E. Pippenger. A simple, rapid method for the simultaneous determination of morphine and its principal metabolites in plasma using high-performance liquid chromatography and fluorometric detection. Ther. Drug Monit. 13:226-232 (1991).

23. J.H. Jaffe and W.R. Martin. Opioid analgesics and antagonists. In The Pharmacological Basis of Therapeutics, 8th ed., A.G. Gilman, T.W. Rail, A.S. Nies, and P. Taylor, Eds. Pergamon Press, New York, NY, 1990, pp 485-521.

24. H.J. Rothkotter, E. Sowa, and R. Pabst. The pig as a model of de- velopmental immunology. Hum. Exp. Toxicol. 21:533-536 (2002).

25. K. Tunblad, P. Ederoth, A. Gardenfors, M. Hammarlund- Udenaes, and C.H. Nordstrom. Altered brain exposure of morphine in experimental meningitis studied with microdialysis. Acta Anaesthesiol Scand. 48:294-301 (2004).

26. M.F. Shepard, K.D. Lake, and M.A. Kamps. Postmortem changes


Journal of Analytical Toxicology, Vo[. 30, November/December 2006

and pharmacokinetics: review of the literature and case report. Ann. Pharmacother. 26" 510-514 (1992).

27. D.S. Cook, R.A. Braithwaite, and K.A. Hale. Estimating antemortem drug concentrations from postmortem blood samples: the influence of postmortem redistribution. J. Clin. Pathol. 53: 282-285 (2000),

28. 8.K. Logan and D. Smirnow. Postmortem distribution and redistribution of morphine in man. J. Forensic Sci. 41:37-46 (1996).

29. G. Koren, D. Beatie, S. Soldin, T.R. Einerson, and S. MacLeod. In- terpretation of elevated postmortem serum concentrations of digoxin in infants and children. Arch. Pathol. Lab. Med. 113: 758-761 (1989).

30. M. Dalpe-Scott, M. Degouffe, D. Garbutt, and M. Drost. A comparison of drug concentrations post mortem cardiac and peripheral blood in 320 cases. Can. Soc. Forensic Sei. J. 28:113-121 (1995).

31. H.F. Gomez, P. McKinney, S. Phillips, D.V. Roberts, J. Brent, and W.A. Watson. Postmortem acetaminophen pharmacokinetics: an experimental study of site and time-dependent concentration changes. J. Forensic Sci. 40:980-982 (1995).

32. J. Gerostamoulos and O.H. Drummer. Postmortem redistribution of morphine and its metabolites. ]. Forensic Sci. 45:843-845 (2O0O).

33. G.R. Jones and D.J. Pounder. Site dependence of drug concen-

trations in postmortem blood--a case study. J. Anal. Toxicol. 11: 186-190 (1987).

34. D.J. Pounder and G.R. Jones. Post-mortem drug redistribution-- a toxicological nightmare. Forensic Sci. Int. 45:253-263 (1990).

35. R.W. Prouty and W.H. Anderson. Implications of site and tem- poral influences on postmortem blood-drug concentrations. J. Forensic ScL 35:243-270 (1990).

36. G. Skopp, L. Potsch, A. Klingmann, and R. Mattern. Stability of morphine, morphine-3-glucuronide, and morphine-6-glucuronide in fresh blood and plasma and postmortem blood samples. J. Anal. Toxicol. 25:2-7 (2001).

37. C. Schmidt, B. Gossage, A. Grinowski, and T. Martinez. Investi- gation of the postmortem redistribution of opiates. Proc. West Pharmacot. Soc. 39:27-28 (1996).

38. F. Moriya and Y. Hashimoto. Redistribution of basic drugs into cardiac blood from surrounding tissues during early-stages postmortem. J. Forensic Sei. 44:10-16 (1999).

39. F. Moriya and Y. Hashimoto. Criteria for judging whether postmortem blood drug concentrations can be used for toxicologic evaluation. Leg. Med. (Tokyo) 2" 143-151 (2000).

Manuscript received November 1, 2005; revision received March 14, 2006.


the influence of collection site and methods on postmortem

Documents