Postmortem Blood Cadmium, Lead, and Mercury Concentrations ...

Journal of Analytical Toxicology, Vol. 34, April 2010

Postmortem Blood Cadmium, Lead, and Mercury Concentrations: Comparisons with Regard to Sampling Location and Reference Ranges for Living Persons Joshua G. Schier1, Michael Heninger2, Amy Wolkin1, Stephanie Kieszak1, Kathleen L. Caldwell3, Geroncio C. Fajardo2, Robert Jones3, Carol Rubin1, Randy Hanzlick2, John D. Osterloh3, and Michael A. McGeehin1 1Division

of Environmental Hazards and Health Effects, National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, Georgia; 2Fulton County Medical Examiner’s Office, Atlanta, Georgia; and 3Division of Laboratory Sciences, National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, Georgia

Abstract This study’s goal was to determine cadmium (Cd), lead (Pb), total mercury (THg), and inorganic mercury (IHg) levels in human cadavers to compare measured levels with established reference ranges for living persons and to determine whether blood levels varied with time from death to sample collection or by body collection site. Subjects (n = 66) recruited from the Fulton County Medical Examiner’s Office in Atlanta, GA, were 20 years of age or older, had no penetrating trauma, no obvious source of environmental contamination of the vasculature, and had whole blood accessible from the femoral (F) site, the cardiac (C) site, or both. Geometric mean results were as follows: 2.59 µg/L F-Cd; 11.81 µg/L C-Cd; 1.03 µg/L F-THg; 2.01 µg/L C-THg; 0.29 µg/L F-IHg; 0.49 µg/L C-IHg; 1.78 µg/dL F-Pb; and 1.87 µg/dL C-Pb. Both F- and C-Cd levels as well as C-THg levels were significantly higher than reference values among living persons (C- and F-Cd, p < 0.0001 and C-THg, p = 0.0001, respectively). Based on regression modeling, as the postmortem interval increased, blood Cd levels increased (p < 0.006). Postmortem blood Cd concentrations were elevated compared to population values and varied with respect to sampling location and postmortem interval.

Introduction For certain analytes, postmortem testing for xenobiotics (i.e., chemical compounds not produced by a living organism) and interpretation of the test results might reveal whether that analyte contributed to a person’s death (1). To ensure proper result interpretation, however, accurate postmortem testing requires both appropriate sampling procedures and, for comparison purposes, established and validated reference ranges. Yet at present, established postmortem reference

ranges are limited or nonexistent for analytes such as cadmium, lead, and mercury. This sometimes results in comparisons using reference ranges for living persons (1), an approach that can lead to misattribution and to costly and unnecessary follow-up investigations. One example involves a relatively recent criminal investigation of a possible homicide in Pittsburgh, PA. The investigation was initiated because of extremely high cadmium concentrations found in postmortem blood testing of a single person (2,3). The subsequent investigation included the costly exhumation of several other bodies and postmortem cadmium level testing. The additional samples also revealed significantly elevated cadmium levels, but as the investigation progressed, investigators ultimately determined that intentional cadmium poisoning was unlikely. The elevated postmortem cadmium levels were more likely to have originated from another source, such as environmental contamination of the biologic specimens or from the normal postmortem redistribution of cadmium between tissues (4). Postmortem redistribution is one of several phenomena that complicate the interpretation of postmortem blood testing results. This example illustrates the shortage of systematically collected information about postmortem tissue levels for metals such as cadmium, mercury, and lead that may redistribute over time after death. We studied postmortem whole blood levels of cadmium, mercury, and lead in selected medical examiner’s death cases to determine whether general postmortem blood levels were different from established reference ranges for living persons and whether blood levels varied with postmortem interval or by collection site. These analytes were chosen because limited or no data were available for them and because of their relevance (e.g., cadmium) to previous public health investigations.

Reproduction (photocopying) of editorial content of this journal is prohibited without publisher’s permission.

129


Experimental Study design

We conducted a cross-sectional study of postmortem patients and compared the analytic results with data from the 3rd National Report on Human Exposure to Environmental Chemicals. We determined that to detect a fivefold difference between blood cadmium levels (the primary analyte of interest) and the postmortem population, we would need approximately 50 cases. During the study period, which ran from March to November 2006, technicians at the Medical Examiners (ME) Office in Atlanta, GA used a convenience sampling to select study subjects. Inclusion criteria for the study were 1. the deceased was age 20 years or older; 2. no resuscitative medications were given and no IV lines were used (source of possible environmental contamination); 3. whole blood was accessible from femoral site, the cardiac site, or both; and 4. absence of penetrating trauma. We included deaths due to poisonings (a possible source of environmental contamination) but retrospectively reviewed medical examiner records to determine the poisoning agent(s).

0.05% Triton X-100. To reduce intrinsic mercury memory effects, gold was added. Rhodium was also added for the internal standardization of cadmium as was bismuth for the internal standardization of mercury and lead (5–7). The samples were prepared with the following ratio: sample/water/diluent = 1:1:48. Total mercury, lead, and cadmium were quantified based on the ratio of analyte signal to that of the internal standard signal in peak hopping mode. The calibration was externally matrix-matched. During analysis of whole blood samples, we analyzed two bench quality control pools together with “blind” quality control pools interspersed among the participant samples. Accuracy was verified by the analysis of standard reference material (SRM 955c) from the National Institute of Standards Technology (NIST). For cadmium, lead, and mercury, the limits of detection (LOD), using the standardization of base blood material, were 0.20 µg/L (n = 276), 0.25 µg/dL (n = 284), and 0.33 µg/L (n = 275), Table I. Basic Demographic Results on the 66 Study Subjects* Meeting Inclusion Criteria Demographic

Result

Data collection

ME technicians collected on the subjects and abstracted other data from the official ME record. Collected information included 1. medical examiner case number; 2. cause of death; 3. estimated date and time of death; 4. date, time, and location (peripheral or cardiac) of sample collection; and 5. demographic data such as age, sex, smoking history, occupation, and co-morbid medical conditions, if known. Blood samples were collected from postmortem subjects who met the inclusion criteria at the time of autopsy, which is typically done within 24 h of body arrival. Laboratory specimen collection, handling, and analysis

Technicians collected postmortem blood samples in metalfree ethylenediaminetetraacetic acid (EDTA) specimen tubes from the femoral (peripheral) and cardiac sites, which were cleaned using a metal-free alcohol swab. Cardiac samples were typically collected from the vena cava near where it enters the heart or from the right atrium. If sample collection from this site failed, blood was obtained from the root of the aorta or pulmonary artery. Study investigators instructed technicians in the ME’s office on sampling techniques and provided necessary equipment. Samples were stored in a refrigerator at ~4°C until transport to the Division of Laboratory Sciences (National Center for Environmental Health, Centers for Disease Control and Prevention, CDC). At the CDC, concentrations for whole blood lead, total mercury, and cadmium were determined using the PerkinElmer Inductively coupled plasma-dynamic reaction cell-mass spectrometer 6100 ELAN series DRC II, ELAN® DRC II ICP-MS (PerkinElmer SCIEX, Concord, ON, Canada) equipped with a Minehart nebulizer and cyclonic spray chamber. In this multi-element analytical technique, blood samples are diluted with ≥ 18 ΩM.cm water and diluent containing 1% v/v tetramethyl/ammonium hydroxide, 0.5% disodium ethylenediamine tetraacetate, 10% ethyl alcohol, and

130

Number of cases

66

Mean age of cases (range)

45.6 years (20–69 years)

Mean weight (range)

185 lbs (110–356 lbs)

Mean height (range)

70 in. (62.5–78.5 in)

Sex (Frequency) Male Female Missing

45 (68.2%) 20 (30.3%) 1 (1.5%)

Ethnicity (Frequency) Caucasian African-American Asian Hispanic

24 (36.4%) 40 (60.6%) 1 (1.5%) 1 (1.5%)

Smoking History (Frequency) Unknown Yes No Missing

61 (92.5%) 3 (4.5%) 1 (1.5%) 1 (1.5%)

Comorbidity (Frequency) Other† Hypertension Unknown Diabetes Cardiac disease

39 (59.1%) 19 (28.8%) 17 (25.8%) 8 (12.1) 4 (6.1%)

Mean time between death and sample collection

13.3 h

* Three cases failed to meet inclusion criteria on retrospective review, and two cases had questionable data due to mislabeling. † Other co-morbidities included conditions like asthma (3), seizures (4), drug use including ethanol (12), human immunodeficiency virus infection (2), dental caries (1), malignancy (1), neuropathy (2), hyperlipidemia (3), coronary artery disease (1), arthritis (2), obesity (4), hypertension (1), depression (3), bipolar disorder (1), cerebrovascular disease (1), anxiety or panic attacks (2), subdural hematoma (1), renal disease (1), Graves’ Disease (1), lupus (1), congestive heart failure (1), multiple sclerosis (1), psoriasis (1), chronic obstructive pulmonary disease or emphysema (2), idiopathic thrombocytopenic purpura (1), back pain (1), psychiatric problems (2), insomnia (2), and cachexia (1).


respectively. The interassay precisions (relative standard deviations) for cadmium, lead, and mercury were 4.3%, 2.6%, and 3.2% at levels of 2.04 µg/L, 2.89 µg/dL, and 5.77 µg/L, respectively. Inorganic mercury was analyzed as described in Albalak et al. (8). The cold vapor atomic absorption spectrophotometric (CVAAS) method used a PerkinElmer FIMS™ 400 atomic absorption (AA) spectrometer, which consisted of a flow injection mercury system (FIMS) and an AS-91 autosampler (PerkinElmer Life and Analytical Sciences, Shelton, CT). The CVAAS was fully controlled from a personal computer using PerkinElmer AA Winlab™ software. Inorganic mercury samples were analyzed using a Maxidigest MX 350 microwave digester (Prolabo, Paris, France), in which the sample flowed into a knitted digestion coil (3D-Digester loop, 10-m long). Inorganic Hg in blood was measured using stannous chloride as a reductant. A 500-µL sample coil was used to inject the diluted blood sample. The LOD in micrograms per liter was calculated to be 0.35 µg/L (n = 793), and the relative standard deviation was 20% at 1 µg/L of mercury (n = 438) and 16% at 2 µg/L (n = 438) in diluted whole blood. Standard reference material (SRM 966) from NIST was used to verify accuracy. Statistical analysis

For each metal concentration, geometric mean values were determined with respect to sampling location and their respective standard deviations and ranges. For persons 20 years of age and older, as reported by the 3rd National Report on Human Exposure to Environmental Chemicals or data collected from the 2003–2006 National Health and Nutrition Examination Survey, geometric mean concentrations for each metal were compared with established reference ranges (9). The referent populations for each analyte were 4207 persons for cadmium/lead (9) and 9034 persons for total mercury/inorganic mercury (unpublished data). For inorganic mercury, no equivalent referent population was available, and consequently,

no comparison for this analyte was performed. A single sample, two-tailed t-test was used to compare study geometric mean values to the appropriate reference values. The relationship between cadmium concentration and time interval from death to sampling was analyzed by linear regression using SAS version 9.0. The cardiac to peripheral blood (C/P) ratio is a measure of the potential for postmortem metal (or analyte) redistribution; the higher the C/P ratio, the more likely the xenobiotic will undergo postmortem redistribution (1). For each metal, the arithmetic C/P ratio was calculated by taking the average value for all individual C/P ratios; the median C/P ratio was also determined. To assess whether result variability was related to sampling location, we used the Wilcoxon signed rank test to investigate whether metal concentrations were significantly different in cardiac and peripheral blood.

Results Seventy-one cases were enrolled in this study. On further review, however, we determined that five either did not meet eligibility criteria or had questionable data due to possible mislabeling. We excluded them from analysis. Of the remaining 66 subjects, the mean age was 45.6 years (range 20–69 years) and the majority were male (n = 45) (68%) and African-American (n = 40) (61%). The mean time between a subject’s death and sample collection was 13.3 h (range 0.5–24.5 h). Of the 66 remaining subjects, 48 subjects had femoral blood specimens suitable for cadmium, lead, and total mercury analysis; 50 subjects had cardiac blood specimens suitable for cadmium, lead, total mercury, and inorganic mercury analysis; and 47 subjects had femoral blood specimens suitable for inorganic mercury analysis. The presence of microclots was the primary reason for specimen unsuitability for cadmium, lead, and mercury analysis.

Table II. Postmortem Metal Geometric Mean Values, Ranges, C/P Ratios, and Comparison with Established Reference Ranges in Living Persons

Metal—Sampling Site

Geometric Mean (sample)

Range (LOD = limit of detection)

NHANES Geometric Mean

Cadmium—femoral (µg/L) Cadmium—cardiac (µg/L) Total mercury—femoral (µg/L) Total mercury—cardiac (µg/L) Inorganic mercury—femoral (µg/L) Inorganic mercury—cardiac (µg/L) Lead—femoral (µg/dL) Lead—cardiac (µg/dL)

2.59 11.81 1.03 2.01 0.29 0.49 1.78 1.87

< LOD–77 < LOD–600 < LOD–9.5 < LOD–45 < LOD–1.5 < LOD–23 < LOD–12 < LOD–17

0.468 ‡ 0.468 ‡ 1.017 § 1.017§ No value available No value available 1.56 ‡ 1.56 ‡

NHANES 95th Percentile* 1.5 5.4§ 0.7§ 4.6

p-Value for NHANES C/P Ratio † Geometric Mean Arithmetic Mean Comparison (median) < 0.0001 < 0.0001 0.9120 0.0001 Not available Not available 0.28 0.16

14.93 (3.50) 8.04 (1.36) 5.19 (1.00) 1.09 (0.98)

* 95th percentile intervals from the Third National Report on Human Exposure to Environmental Chemicals. † The c/p ratio is the ratio of the cardiac blood concentration of a xenobiotic to the peripheral blood concentration. Xenobiotics with a large C/P ratio are believed to undergo postmortem redistribution. Both the arithmetic mean and the median C/P values are reported. † Data from the 1999–2000 dataset from the Third National Report on Human Exposure to Environmental Chemicals were used for cadmium because the proportion of results below the LOD in the 2001–2002 dataset was too high to provide valid results. Reference values for lead were from data collected in 2001–2002. § Reference values obtained from data collected from the 2003–2006 National Health and Nutrition Examination Surveys in all adults age > 20 years old (unpublished data).

131


Microclots interfered with the analytical process and rendered the sample impossible to analyze. Additional descriptive data on the 66 subjects is listed in Table I. The geometric mean was used because the data were not normally distributed on the original scale. Taking the natural log transformation resulted in a normal distribution in most cases. Geometric mean values for all metals were 2.59 µg/L femoral cadmium; 11.81 µg/L cardiac cadmium; 1.03 µg/L femoral total mercury; 2.01 µg/L cardiac total mercury; 0.29 µg/L femoral inorganic mercury; 0.49 µg/L cardiac inorganic mercury; 1.78 µg/dL femoral lead; and 1.87 µg/dL cardiac lead (Table II). With respect to established normal values in living persons, geometric mean values for postmortem femoral and cardiac cadmium levels were elevated 5-fold and 20-fold, respectively (Table II). Cardiac blood cadmium levels in particular were higher than femoral blood cadmium levels. Additional data on range, comparison to geometric mean population values and comparisons to normal concentrations in living persons are listed in Table II. Median values were as follows: 2.3 µg/L femoral cadmium;

10.5 µg/L cardiac cadmium; 1.185 µg/L femoral total mercury; 1.85 µg/L cardiac total mercury; 0.247 µg/L femoral inorganic mercury; 0.247 µg/L cardiac inorganic mercury; 1.85 µg/dL femoral lead; and 1.85 µg/dL cardiac lead. For cadmium in femoral and cardiac blood and for total mercury in cardiac blood, levels were higher than the premortem referent population values, as evidenced by a single sample t-test. Detailed information on the reference levels used for each comparison is listed in Table II. The causes of death among study subjects were cardiovascular (n = 23, 35%), trauma (n = 17, 11%), drug or toxic exposures (n = 12) (18%), chronic diseases (n = 7, 11%), suicide (n = 4, 6%), infectious etiology (n = 1, 1.5%), and drowning (n = 1, 1.5%). One case (1.5%) had no information listed (Table III). All drug- or toxin-associated deaths were due to agents other than cadmium, mercury, lead, or other metals; these deaths were primarily due to drugs of abuse such as cocaine, opiates, and ethanol. We analyzed regression models to determine whether the time between death and sample collection significantly affected the natural log of each metal; the time difference was Table III. Causes of Death Among 66 Cases* calculated in minutes. For the two cadmium collection sites, time had a significant effect (p = 0.0008 for femoral and p = Cause of Death Number of Cases 0.0050 for cardiac). An increase in time from death to sample collection similarly affected cadmium levels. Lead and inorCardiovascular 23 ganic mercury levels also tended to increase, but the increase (e.g., atherosclerotic heart disease, was not significant for lead (p = 0.0739 for peripheral and p = myocardial infarction) 0.2236 for cardiac) or for inorganic mercury (p = 0.0970 for peTrauma 17 ripheral and p = 0.2040 for cardiac). Conversely, total mercury (e.g., non-penetrating trauma, blunt injury) levels tended to decrease, but again, this effect was not signifDrug/Toxic* 12 icant: total mercury (p = 0.5993 for peripheral and p = 0.8736 (e.g., drug overdoses and exposures) for cardiac). The approximate amounts of variation explained by time were as follows: 23% femoral cadmium; 16% cardiac Chronic diseases 7 cadmium; 1% femoral total mercury; < 1% cardiac total mer(e.g., diabetes, malignancy) cury; 6% femoral inorganic mercury; 3% cardiac inorganic Suicide/Hanging 4 mercury; 7% femoral lead; and 3% cardiac lead. We used the Wilcoxon signed rank test (the nonparametric Infectious 1 equivalent of the paired t-test) to investigate whether the meDrowning 1 dian difference was zero (n = 38 for all pairs except inorganic Missing data 1 mercury, where n = 37) between paired samples (femoral and cardiac metal concentrations in the same person). In fact, the * If multiple causes of death were listed, the first was chosen. If the cause of death difference between femoral and cardiac lead was not significant was listed as a condition exacerbated by an acute toxic exposure, the cause of death was listed as Drug/Toxic. All Drug/Toxic deaths were due to agents other (p = 0.5736). All other differences were significant, however, for than cardiac, lead, mercury or any other metal, and the majority of cases were cadmium (p < 0.0001), total mercury (p = 0.0001), and inordue to drugs of abuse such as cocaine, opiates, or ethanol. ganic mercury (p = 0.0016). Mean differences, median differences, their standard deviations, and their ranges are reported in Table IV. The carTable IV. Number of Paired Samples, Mean Difference, Median Difference, diac concentrations of these metals were Standard Deviation of Difference, and Range for Paired Samples (Femoral and also higher. Calculated C/P ratios (both Cardiac) of Each Metal arithmetic and median values) are in Table II.

Metal Cadmium Total mercury Inorganic mercury Lead

132

Number of Paired Samples

Mean Difference

38 38 37 38

–43.92 –1.85 –0.80 –0.24

Median Difference

–5.8 –0.53333 0 0

Standard Deviation

Range

102.27 5.85 3.70 2.29

–590 to 3.36 –35.5 to 2.1 –22.56 to 0.56 –7.50 to 7.20

Discussion With careful interpretation, postmortem analyte testing can be useful in


attributing cause of death. But comorbid conditions (10), collection site location (11,12), postmortem metabolism (13), stability in the postmortem environment (14,15), and postmortem redistribution (16) are factors that complicate interpretation of those results. As is known for some drugs, postmortem redistribution of xenobiotics sequestered in the gastrointestinal tract, liver, and other tissues may affect blood levels, but for environmental chemicals such as metals this effect is less studied (1). One study indicated a small gradient in two deceased persons (difference between pre- and postmortem cadmium levels ranged from 0.28 to 0.6 µg/g). Although cadmium levels in that study were also shown to vary slightly depending on sampling location, the comparison included a sample from only one patient (17,18). By contrast, our study sought to avoid environmental contamination as a source of falsely elevated cadmium levels and, compared with referenced study, to compare data from a substantially larger sample size. In comparison with established reference values in living persons, both postmortem femoral and cardiac cadmium levels were significantly elevated. In this study, some of the highest (≥ 75 µg/L) cardiac blood cadmium levels (n = 9, range 82–600 μg/L) were incompatible with life and could easily be misinterpreted as the cause of death. However, because the medical examiner reported no evidence of clinically compatible illness, this is an unlikely hypothesis. Geometric mean total mercury cardiac blood levels were, when compared with established normal values in living persons, also elevated twofold. This is likely due to an elevated organic mercury fraction (inorganic mercury levels were extremely low) (Table II). Moreover, these elevated levels most likely represent postmortem redistribution of cadmium and total mercury rather than true excessive exposure (e.g., poisoning). In this study, we did not find evidence for postmortem redistribution of lead. We did find that the highest postmortem metal levels were in cardiac blood. In general, xenobiotic concentrations such as pharmaceuticals in cardiac blood are influenced by postmortem diffusion and redistribution processes from surrounding tissues (e.g., lungs, gastrointestinal tract, and myocardium) as well as from medical treatment administered before death (19,20). In particular, drugs sequestered in the liver may be redistributed via the hepatic veins to the inferior vena cava, right cardiac chambers, and pulmonary vessels (16). This is likely to occur with other agents such as metals stored in tissues during life. Because cadmium is sequestered in the liver during life, its redistribution is a likely explanation for elevated whole blood levels. Liver and kidney levels of cadmium may be 1000 to 10,000 times higher, respectively, than blood levels (21). But as the time interval between death and sample collection increased, so did cadmium levels. This is probably because an increased amount of time was available for postmortem metal redistribution in cases with large time intervals. Among 83 drugs for which a C/P ratio was listed, our cadmium C/P value was higher than 80 of them, further supporting the notion that cadmium is likely to undergo postmortem redistribution (1). The C/P ratio may be affected by the interval between death and sample collection. Drug concentrations in the femoral vein tend to be relatively stable over time and are better markers of exposure than is cardiac blood

(12,20,22). The same is likely true for many other xenobiotics such as metals; our data demonstrating higher cardiac levels for total mercury, inorganic mercury, and cadmium support this. Xenobiotics with a high C/P concentration ratio and those with known high distribution volumes are believed to have a greater potential for undergoing postmortem redistribution (1,22). On the other hand, an uneven distribution of cadmium within the blood specimen itself due to an uneven distribution of red cells (secondary to clotting or lividity) would be an unlikely explanation for elevated postmortem levels of whole blood cadmium. Though 90% of whole blood cadmium resides in red cells (23), specimens containing fewer red cells due to clotting or lividity would tend to decrease or dilute the measured whole blood cadmium level. The arithmetic C/P ratio, not the median C/P ratio, for total mercury and inorganic mercury was also elevated although not nearly as much as cadmium, suggesting some degree of postmortem redistribution may also occur for this metal. However, with respect to sampling location for cadmium, total mercury, and inorganic mercury, postmortem metal concentrations were significantly different. This strongly suggests that the choice of sampling location is another variable that can ultimately affect cadmium and mercury results. Forensic researchers and others are cautioned that from death to sample collection, sampling site and time interval will affect postmortem cadmium and total mercury results and possibly even inorganic mercury results. We did not have antemortem blood samples for comparison with postmortem testing results, and this was a major study limitation. Also, because the samples tested were only from one metropolitan area, they may not be representative of the general population. Furthermore, although such a scenario is considered unlikely, we were unable definitively to rule out significant metal premortem exposure to metals (e.g., poisoning or occupational exposures). Finally, because young, sudden, and violent deaths can be overrepresented, medical examiner cases are probably not representative of the general population.

Conclusions We found that postmortem blood cadmium concentrations are elevated in comparison to premortem reference levels in living persons and that the postmortem results vary with respect to sampling location and interval between death and sample collection. These differences are most likely due to the changes that occur after death, such as postmortem diffusion and redistribution from other organs. We also found that total and inorganic mercury postmortem concentrations vary with regard to sampling location; total mercury concentrations in postmortem cardiac blood (not femoral) are significantly different from established reference ranges in living persons. Moreover, cardiac blood is a less suitable medium for postmortem sampling for cadmium or total mercury, and inorganic mercury and testing results should be interpreted carefully. The use of reference ranges established from living persons for

133


comparisons of postmortem cadmium testing results is of limited value, and finally, further work is needed to determine appropriate postmortem reference levels for metals such as cadmium, lead, total mercury, and inorganic mercury.

10. 11.

Acknowledgment

12.

The authors thank the staff of the Fulton County Medical Examiner’s Office.

13. 14.

References 1. J.B. Leiken and W.A. Watson. Post-mortem toxicology: what the dead can and cannot tell us. J. Toxicol. Clin. Toxicol. 41: 47–56 (2003). 2. Cadmium cases may not be criminal. http://pittsburghlive.com/ x/tribune-review/news/s_173174.html (accessed August 2008). 3. FBI investigates cadmium poisoning mystery. http://www.wpxi. com/health/2567510/detail.html (accessed August 2008). 4. 1994 study may hold cadmium mystery’s answer. http:// www.post-gazette.com/localnews/20040104closeuplocal5p5.asp (accessed August 2008). 5. A.R. Date and A.L. Gray. Applications of Inductively Coupled Plasma Mass Spectrometry. Chapman and Hall, New York, NY, 1989. 6. E. Barany, I.A. Bergdahl, A. Schutz, S. Skerfving, and A. Oskarsson. Inductively coupled plasma mass spectrometry for direct multielement analysis of diluted human blood and serum. J. Anal. At. Spectrom. 12: 1005–1009 (1997). 7. D.E. Nixon, M.F. Burritt, and T.P. Moyer. The determination of mercury in whole blood and urine by inductively coupled plasma mass spectrometry. Spectrochim. Acta 54Part B: 1141–1153 (1999). 8. R. Albalak, K.L. Caldwell, R.J. Jones, and G. Miller. Inorganic mercury determination in whole blood using on-line microwave digestion with flow injection mercury system (FIMS). At. Spectrosc. 26: 234–240 (2005). 9. Centers for Disease Control and Prevention. Third National Report on Human Exposure to Environmental Chemicals. Atlanta (GA):

134

15. 16. 17. 18. 19. 20. 21. 22.

23.

CDC, 2005. http://www.cdc.gov/exposurereport (accessed March 2008). Y. Bentur, A. Tsipiniuk, and U. Taitelman. Postmortem digoxin-like immunoreactive substances (DLIS) in patients not treated with digoxin. Hum. Exp. Toxicol. 18: 67–70 (1999). S. Felby and J. Olsen. Comparative studies of postmortem barbiturate and meprobomate in vitreous humor, blood and liver. J. Forensic Sci. 14: 507–514 (1969). T. Hilberg, S. Rogde, and J. Morland. Postmortem drug redistribution—human cases related to results in experimental animals. J. Forensic Sci. 44: 3–9 (1999). S.B. Karch. Introduction to the forensic pathology of cocaine. Am. J. Forensic Med. Pathol. 12: 126–131 (1991). A.H. Battah and K.H. Hadidi. Stability of trihexyphenidyl in stored blood and urine specimens. Int. J. Legal Med. 111: 111–114 (1998). R.J. Dettling, E.J. Briglia, L.A. Dal Cortivo, and J.H. Bidanset. The production of amitryptiline from nortroptyline in formaldehydecontaining solutions. J. Anal. Toxicol. 14: 325–326 (1990). A.L. Pelissier-Alicot, J.M. Gaulier, P. Champsaur, and P. Marquet. Mechanisms underlying postmortem redistribution of drugs: a review. J. Anal. Toxicol. 27: 533–544 (2003). N. Koizumi, F. Hatayama, and K. Sumino. Problems in the analysis of cadmium in autopsied tissues. Environ. Res. 64: 192–198 (1994). K. Sumino, K. Hayakawa, T. Shibata, and S. Kitamura. Heavy metals in normal Japanese tissues. Arch. Environ. Health 30: 487– 494 (1975). F. Moriya and Y. Hashimoto. Redistribution of basic drugs into cardiac blood from surrounding tissues during early stages postmortem. J. Forensic Sci. 44: 10–16 (1999). M.C. Yarema and C.E. Becker. Key concepts in postmortem drug redistribution. Clin. Toxicol. 43: 235–241 (2005). R.C. Baselt. Disposition of Toxic Drugs and Chemicals in Man, 6th ed. Biomedical Publications, Foster City, CA, 2002. D.S. Cook, D.S. 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). R.R. Lauwerys and P. Hoet. Industrial Chemical Exposure: Guidelines for Biological Monitoring, 3rd ed. Lewis Publishers, New York, NY, 2001, pp 54–69. Manuscript received June 27, 2009; revision received November 16, 2009.