Available online at www.sciencedirect.

com J O U R N A L O F

FORENSIC
A N D L E G AL
ME D IC IN E
Journal of Forensic and Legal Medicine 15 (2008) 312–317
www.elsevier.com/jflm

Original Communication

Biochemical blood markers and sampling sites in forensic autopsy

Koichi Uemura MD, PhD a,b,*, Kaori Shintani-Ishida MSc a, Kanju Saka BSc a,
Makoto Nakajima MVet a, Hiroshi Ikegaya MD, PhD c, Yousuke Kikuchi MD a,
Ken-ichi Yoshida MD, PhD a
a
Department of Forensic Medicine, Graduate School of Medicine, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
b
Section of Forensic Medicine, Department of International Health Development, Division of Public Health, Graduate School,
Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan
c
National Research Institute of Police Science, 6-3-1, Kashiwanoha, Kashiwa-city, Chiba 277-0880, Japan

Received 28 September 2006; received in revised form 24 November 2007; accepted 13 December 2007
Available online 10 March 2008

Abstract

Forensic pathologists often hesitate to use biochemical blood markers due to the risk of large postmortem changes and deviations
from healthy subjects. Biochemical analyses of postmortem blood, if possible, may help to evaluate pathological status and determining
the cause of death in forensic diagnosis, for example, in sudden unexpected death without obvious cause, or young adults with no appar-
ent cause of death or antemortem information. Even commercially available biochemical markers were re-evaluated in the blood samples
of 164 forensic autopsy cases. Biochemical markers examined were HbA1c, fructosamine, blood nitrogen urea (BUN), creatinine, total
protein, total bilirubin, c-glutamyl transpeptidase (c-GTP), triglyceride, total cholesterol, C-reactive protein (CRP) and pseudocholine
esterase (pChE). We collected cardiac blood (left cardiac blood and right cardiac blood) and peripheral blood (femoral vein blood) to
clarify the differences in measured values by sampling site. The measured values were analyzed in relation to postmortem interval, eti-
ology of death and sampling sites. Of all eleven markers, HbA1c is the most useful and reliable because of its negligible postmortem
changes and small deviation from healthy subjects. Total bilirubin, BUN, CRP and total cholesterol can be useful if we set appropriate
limit ranges and pay attention to the interpretation. For the evaluation of changes due to postmortem intervals, none of the markers
except for triglyceride showed significant changes up to three days postmortem. As for sampling sites, femoral vein blood is generally
recommended considering postmortem changes, but left cardiac blood was suitable for creatinine, pChE, and total cholesterol. For clin-
ical forensic diagnosis of biochemical blood markers, we must determine the ‘‘forensic abnormal value” after collecting more cases by
known causes with more information about the population.
Ó 2008 Elsevier Ltd and FFLM. All rights reserved.

Keywords: Biochemical markers; Postmortem change; Right cardiac blood; Left cardiac blood; Femoral vein blood; Sampling sites

1. Introduction pathologist must make a decision on the basis of autopsy

findings. In clinical medicine, a lot of information, including
In forensic autopsies, antemortem information such as biochemical markers in the blood, is available and contrib-
present and past illnesses is rarely available, and the forensic utes to the diagnosis of disease, in addition to physical find-
ings. For forensic pathologists, biochemical analysis of the
postmortem blood, if possible, may help in evaluating path-
*
Corresponding author. Address: Section of Forensic Medicine, ological status and determining cause of death in such cases.
Department of International Health Development, Division of Public However, to date, forensic pathologists have often hesitated
Health, Graduate School, Tokyo Medical and Dental University, 1-5-45
Yushima, Bunkyo-ku, Tokyo 113-8519, Japan. Tel.: +81 3 5803 5197; fax:
to use biochemical markers in the blood for forensic diagno-
+81 3 5803 0128. sis due to concern about large postmortem changes and large
E-mail address: kuemura.legm@tmd.ac.jp (K. Uemura). deviations from healthy subjects.

1752-928X/$ - see front matter Ó 2008 Elsevier Ltd and FFLM. All rights reserved.
doi:10.1016/j.jflm.2007.12.003
 K. Uemura et al. / Journal of Forensic and Legal Medicine 15 (2008) 312–317 313

There is a complete review by Coe on autopsy samples, nitrogen urea (BUN) (urease UV method: 6–20 mg/dL)
covering a lot of markers1 and a report by Tsuji et al. on an and creatinine (enzyme method: male 0.61–1.04 mg/dL,
animal study,2 with respect to postmortem changes in the female 0.47–0.79 mg/dL) for renal failure, total protein (Biu-
markers. Additionally, there are several reports on the use- ret method: 6.7–8.3 g/dL) for malnutrition, total bilirubin
fulness of the individual markers in autopsy diagnosis.3–15 (vanadinate oxidation method: 0.2–1.0 mg/dL) and c-glut-
However, there is insufficient information on commercially amyl transpeptidase (c-GTP) (JSCC standardization
available blood markers and differences in sampling sites in method: male < 70 IU/L, 37 °C, female < 30 IU/L, 37 °C)
forensic autopsy cases. We chose to measure eleven clini- for liver function, triglyceride (enzyme method: 50–
cally available biochemical markers in the blood from three 149 mg/dL) and total cholesterol (enzyme method: 150–
sampling sites. The cost for measuring these eleven markers 219 mg/dL) for hyperlipidemia, C-reactive protein (CRP)
by one sample was very low because these markers are rou- (latex aggregation method, <0.3 mg/dL) for inflammation,
tinely measured in clinical medicine. The cost/benefit factor pseudocholine esterase (pChE) (rate assay, male 242–
is very important for low-budget forensic facilities. 495 IU/L, 37 °C, female 200–495 IU/L, 37 °C) for liver func-
Our aim was to re-examine and evaluate commercially tion and organic phosphate poisoning. The sera volume of
available blood markers and their usefulness in forensic required to measure the 11 markers was 2 mL.
diagnosis. We investigated how biochemical markers in The laboratory rejected 2.4% of samples for bilirubin
the blood suffer from postmortem changes and showed dif- measurement, but not those for other markers. However,
ferences due to the etiology of death, while determining we could not measure all the markers in substantial num-
suitable sampling sites. These will be useful for taking post- bers of cases because of a lack of sample volume.
mortem changes into consideration when selecting markers
and interpreting results. 2.3. Statistical analyses

2. Materials and methods Data are expressed as the mean ± SD. Statistical signif-
icance was determined as follows: for postmortem interval,
2.1. Blood samples Spearman’s rank correlation was carried out (Table 2). For
etiology of death, one-way ANOVA was carried out. When
With the permission of the Ethics Committee of Gradu- there was a difference among groups, Scheffe’s posthoc test
ate School of Medicine, The University of Tokyo (No. between multiple groups was performed (Table 3). For
690), blood was obtained from 164 consecutive autopsy regional differences, one-way repeated measures ANOVA
cases in our department from April 2003 to March 2006 was carried out. When there was a difference among
(age 0–98, average age 54.9 ± 21.8, median age 57.0, male groups, a paired t-test for pair-wise comparisons was per-
112, female 52). The postmortem interval of the sampling formed (Table 4). The software for the above-mentioned
of specimens are as following: 0–12 h (25 cases), 13–69 h statistical analysis was Statview Ver. 4.11 (Abacus Con-
(69 cases), 25–48 h (54 cases), 49–72 h (16 cases). Causes cepts Inc., Berkeley, CA). Significant level was 0.05 (5%).
of death were as follows: blunt injury (52 cases), sharp
injury (seven cases), asphyxiation (18 cases), drowning 3. Results
(four cases), fire death (five cases), intoxication (nine cases),
internal death (39 cases) and others (30 cases). Care was We summarized the data for right cardiac blood, which
paid so that the deceased would not be identified from were obtained in almost all cases (Table 1). First, HbA1c
the data. The bodies were preserved refrigerated and foren- showed almost the same mean value as healthy subjects,
sic autopsies were performed within a day after they were and a very low ratio of abnormal values (24.8%), as com-
found. The blood was sampled from the right and left heart pared with fructosamine (77.7%), another marker for
cavities and femoral vein within 72 h postmortem, as far as chronic hyperglycemia. The next group of markers showed
possible. As soon as whole blood was obtained, the sera much higher mean values than the healthy subjects, and a
was separated by centrifugation at 1000g, 30 min, and higher ratio of abnormal values (37.3–95.1%). This group
stored at 20 °C, while the whole blood was stored at included total bilirubin, triglyceride, BUN, CRP, c-GTP,
4 °C as long as a day, until shipping to the laboratory of fructsamine and creatinine. In the third group, pseudocho-
SRL, Co. Ltd. (Tokyo, Japan), where the samples were line esterase (pChE) and total cholesterol showed lower
analyzed within a day. mean values than the healthy subjects and a higher ratio
of abnormal values (64.1%, 72.9%). In the last group, total
2.2. Biochemical analyses protein showed almost the same mean value as healthy sub-
jects, but a wide variability, and therefore, a high ratio of
We selected the biochemical markers on the basis of post- abnormal values (75.3%).
mortem stability reported in previous reports.1–4 The 11 To evaluate changes due to postmortem intervals, we
markers included HbA1c (latex aggregation method, stan- classified postmortem intervals into four groups (0–12 h,
dard range: 4.3–5.8%) and the fructosamine (calorimetry 13–24 h, 25–48 h, 49–72 h) and carried out correlation
method: 205–285 mM) for chronic hyperglycemia, blood analysis between the results obtained for each of the
314 K. Uemura et al. / Journal of Forensic and Legal Medicine 15 (2008) 312–317

Table 1
Differences from clinical standards (right cardiac blood)
Marker Measured Value obtained from Unit n Results outside the reference
value healthy subjects intervals of healthy subjects (%)
HbA1c 5.23 ± 1.23 4.3–5.8 % 149 24.8
t-Bilirubin 1.32 ± 2.44 0.2–1.0 mg/dL 150 37.3
Triglyceride 129.9 ± 107.4 50–149 mg/dL 155 45.2
BUN 39.8 ± 40.6 6–20 mg/dL 162 59.3
CRP 7.54 ± 11.54 <0.3 mg/dL 163 69.3
c-GTP 154.1 ± 173.4 Male < 70, female < 30 IU/L, 37 °C 148 74.3
Fructsamine 325.3 ± 147.1 205–285 mM 148 77.7
Creatinine 3.29 ± 2.35 Male 0.61–1.04, female 0.47–0.79 mg/dL 163 95.1
Pseudocholine esterase 204.1 ± 120.7 Male 242–495, female 200–459 IU/L, 37 °C 153 64.1
t-Cholesterol 142.3 ± 77.3 150–219 mg/dL 155 72.9
t-Protein 7.45 ± 2.09 6.7–8.3 g/dL 154 75.3
Values are expressed as the mean ± SD.

markers and the postmortem intervals. We summarized the 4. Discussion

results in Table 2. The triglyceride value significantly
decreased according to the postmortem interval. Of all eleven markers examined, HbA1c showed the
Next, to evaluate differences of obtained values by etiol- smallest deviation from healthy subjects (24.8%, Table 1),
ogy of death, we classified the cause of deaths into eight negligible postmortem changes (Table 2) and no difference
categories (blunt injury, sharp injury, asphyxiation, drown- due to etiology of death (Table 3). We confirmed that
ing, fire death, intoxication, internal death and others). We HbA1c showed negligible postmortem changes. Glycated
carried out a comparison by etiology of death and summa- albumin,9 glycated hemoglobin,10–14 and fructsamine14
rized the results in Table 3. There were significant differ- were shown to indicate chronic hyperglycemia in autopsy
ences in pseudocholine esterase, total cholesterol and samples. However, in this study fructsamine showed a large
total protein. For total cholesterol, there were differences deviation from healthy subjects. We recommend HbA1c as
in asphyxiation-blunt injury, and asphyxiation-internal a good marker for chronic hyperglycemia in forensic prac-
death. For total protein, there were differences in fire tice because of its commercial availability and reliability.
death-blunt injury, fire death-sharp injury, and fire death- We can interpret the negligible postmortem changes in
internal death. HbA1c as follows: the level of glycation of hemoglobin is
Differences by site of blood sampling are summarized in a cumulative process the rate of which is determined by
Table 4. Triglyceride, BUN and fructsamine showed no the level of prevailing glucose concentration under the life-
difference by sampling site, but eight other markers showed span of erythrocytes. By contrast, the other markers evalu-
significant differences. Creatinine showed the lowest value ated are components of tissues reflecting their destruction,
in the left cardiac blood, but seven other markers showed or components normally eliminated from the blood by
the lowest values in the femoral vein blood. functioning organs.

Table 2
Postmortem interval (right cardiac blood)
Marker Measured value p-valuea Postmortem change
Postmortem time
0–12 h 13–24 h 25–48 h 49–72 h
HbA1c 5.43 ± 1.78 (21) 5.35 ± 1.05 (66) 4.92 ± 1.05 (49) 5.39 ± 1.59 (13) 0.1571 Unchanged
t-Bilirubin 1.25 ± 2.53 (24) 1.26 ± 2.28 (64) 1.34 ± 2.38 (48) 1.66 ± 3.32 (14) 0.9305 Unchanged
Triglyceride 158.8 ± 82.5 (25) 139.4 ± 138.0 (66) 114.1 ± 73.5 (48) 93.4 ± 65.0 (16) 0.0083 Decrease
BUN 29.5 ± 29.2 (25) 39.1 ± 34.1 (69) 48.6 ± 51.8 (52) 36.2 ± 40.8 (16) 0.2752 Unchanged
CRP 7.12 ± 12.54 (25) 9.30 ± 11.55 (69) 6.67 ± 12.22 (53) 3.44 ± 5.42 (16) 0.7540 Unchanged
c-GTP 131.0 ± 113.3 (25) 135.8 ± 162.9 (61) 165.1 ± 172.6 (49) 229.6 ± 282.8 (13) 0.2545 Unchanged
Fructsamine 304.9 ± 122.5 (22) 308.2 ± 107.7 (66) 360.4 ± 201.7 (45) 301.8 ± 149.3 (15) 0.2711 Unchanged
Creatinine 3.14 ± 2.93 (25) 3.21 ± 2.11 (69) 3.53 ± 2.48 (53) 3.06 ± 2.02 (16) 0.3442 Unchanged
Pseudocholine esterase 235.3 ± 105.0 (24) 200.9 ± 133.1 (66) 191.4 ± 114.9 (47) 207.4 ± 106.0 (16) 0.2599 Unchanged
t-Cholesterol 165.9 ± 59.3 (25) 135.0 ± 83.5 (66) 143.6 ± 79.3 (48) 131.8 ± 67.1 (16) 0.2212 Unchanged
t-Protein 7.16 ± 1.92 (25) 7.23 ± 1.89 (66) 7.63 ± 1.96 (47) 8.34 ± 3.15 (16) 0.0998 Unchanged
Values are expressed as the mean ± SD. (n): n is the sample number.
a
Spearman’s rank correlation test.
 K. Uemura et al. / Journal of Forensic and Legal Medicine 15 (2008) 312–317 315

p-valuea

<0.0001
<0.0001
Total bilirubin showed a tendency towards postmortem

0.2808
0.6044

0.5737
0.0094
0.0063

0.2433

0.0039
0.9079

0.7557
increase time-dependently, but not significantly (Table 2),
and a relatively small deviation from healthy subjects
(37.3%, Table 1). This suggests total bilirubin can be used

7.47 ± 11.23 (30)

182.3 ± 179.1 (28)
346.7 ± 149.6 (28)

119.5 ± 66.7b (29)

7.17 ± 2.04c (30)
2.21 ± 3.52 (29)
4.99 ± 1.09 (28)

101.9 ± 77.1 (29)

47.5 ± 42.5 (30)

3.29 ± 2.73 (30)

159.5 ± 98.6 (30)
as a marker. c-GTP and pseudocholine esterase, other
markers of liver function, are not appropriate for forensic
diagnosis. They showed a high portion of deviation from
Others

healthy subjects (74.3%, 64.1%, respectively, Table 1) and

no postmortem increase, although c-GTP showed a ten-
dency to increase time-dependently (Table 2).
176.2 ± 173.3 (32)
366.8 ± 189.1 (36)

183.7 ± 117.4 (36)

130.1 ± 69.4b (36)
7.14 ± 1.55c (35)
1.34 ± 2.63 (36)
5.33 ± 1.34 (36)

117.1 ± 85.8 (36)

58.1 ± 54.1 (39)
7.24 ± 7.67 (39)

3.64 ± 2.28 (39)

For creatinine and BUN, kidney injury markers, we rec-
Internal death

ommend BUN as a good marker for renal injury. The cases

with BUN > 100 mg/dL showed renal diseases or acute
renal failure due to hemorrhagic shock after intestinal
bleeding. We must be careful when considering the back-
ground since high BUN was reported in association with
365.2 ± 268.9 (6)

288.8 ± 185.5 (8)

1.99 ± 3.91 (7)
5.23 ± 1.50 (8)

111.0 ± 61.1 (8)

25.2 ± 15.3 (8)
3.59 ± 9.43 (9)
117.3 ± 59.9 (7)

3.16 ± 1.37 (9)

131.3 ± 60.8 (8)

9.24 ± 1.51 (7)

severe hypoxia or skeletal muscle damage.15 Creatinine

Intoxication

cannot be recommended because of the extremely high

ratio of abnormal values (95.1%), suggesting a postmortem
increase.
The level of CRP showed no significant postmortem
138.3 ± 166.3 (4)
258.0 ± 131.4 (5)

277.6 ± 148.8 (5)

changes (Table 2). Clinical standard CRP is <0.3 mg/dL,

0.70 ± 0.88 (4)
5.80 ± 2.37 (5)

134.8 ± 79.4 (5)

34.0 ± 29.2 (5)
1.15 ± 2.43 (5)

2.69 ± 1.30 (5)

200.2 ± 92.9 (5)

10.76 ± 3.88 (5)

which is almost equal to zero. In our study, a large stan-

Fire death

dard deviation was observed (7.54 ± 11.54), but 30.7% of

measured samples were within the limit of healthy subjects.
Considering these results, it is concluded that CRP is a
good marker for inflammation, which is also affected by
155.0 ± 116.2 (3)

259.0 ± 214.6 (3)

274.0 ± 109.3 (3)

survival time, as reported previously.6

0.40 ± 0.26 (3)
6.05 ± 1.52 (4)

0.29 ± 0.42 (3)

102.3 ± 76.5 (3)

4.09 ± 2.29 (3)

183.7 ± 22.6 (3)

9.35 ± 0.78 (2)
18.4 ± 3.5 (3)

High total cholesterol can be a marker for hyperlipid-

Drowning

emia, since it tends to decrease postmortem (Table 2).

There were significant differences in asphyxiation-blunt
injury, and asphyxiation-internal death (Table 3). Postmor-
tem time intervals were 25.7 ± 13.6 h (asphyxiation),
162.9 ± 265.5 (18)

285.6 ± 151.6 (17)

219.3 ± 108.0 (18)
0.39 ± 0.25 (18)
5.39 ± 1.43 (17)

163.1 ± 94.5 (18)

22.0 ± 20.3 (18)
1.82 ± 3.44 (18)

336.6 ± 55.8 (17)

2.99 ± 3.07 (18)

8.64 ± 1.23 (18)

22.9 ± 15.6 h (blunt injury), and 29.7 ± 16.3 h (internal

Asphyxiation

death), and there was no significance among the three

groups in terms of postmortem interval. We concluded that
differences in asphyxiation-blunt injury and asphyxiation-
Values are expressed as the mean ± SD. (n): n is the sample number.

internal death reflect the different causes of death. By con-

trast, the mean triglyceride value was higher than that of
152.4 ± 132.1 (7)
247.6 ± 115.6 (5)

6.06 ± 1.87c (7)

0.38 ± 0.41 (6)
5.44 ± 1.01 (7)

135.6 ± 46.9 (7)

1.49 ± 3.30 (7)

1.91 ± 1.19 (7)

197.2 ± 96.1 (5)
98.3 ± 32.3 (7)

healthy subjects, but there was a significant time-dependent

10.4 ± 6.9 (7)
Sharp injury

decrease (Table 2). This marker can easily affected by inges-

tion or starvation, which are sometimes identified in
p < 0.05 (Scheffe’s posthoc test; vs asphyxiation).

autopsy cases and cannot be used as postmortem markers.

p < 0.05 (Scheffe’s posthoc test; vs fire death).

Total protein tended to increase time-dependently,

though not significantly (Table 2). Intravascular fluid is
142.5 ± 148.1 (49)

12.35 ± 15.40 (52)

130.3 ± 152.8 (49)
293.9 ± 103.9 (48)

6.96 ± 2.02c (50)

128.9 ± 53.1b(49)
1.25 ± 1.84 (49)
5.05 ± 0.93 (44)

37.5 ± 35.0 (52)

3.35 ± 2.25 (52)

193.0 ± 91.0 (49)
Measured value

affected either by overhydration caused by infusion or post-

Etiology of death (right cardiac blood)

Blunt injury

mortem extravasation. In our data on etiology of death,

there were differences in fire death-blunt injury, and fire
death-sharp injury, fire death-internal death (Table 3). In
all cases, fire death is related to this difference, reflecting
intravascular extravasation due to heat.
Pseudocholine esterase

One way ANOVA.

As for sampling sites, we recommend the following. For

HbA1c, any site can be used. This is because the values of
t-Cholesterol

the three sampling sites were substantially the same,

Fructsamine
Triglyceride
t-Bilirubin

Creatinine

although femoral vein blood showed a significantly lower

t-Protein
Table 3

Marker

HbA1c

c-GTP

value statistically. For triglyceride, BUN, fructsamine,

BUN
CRP

b
c
a

any site can be used because there were no significant

316 K. Uemura et al. / Journal of Forensic and Legal Medicine 15 (2008) 312–317

Table 4
Regional differences in biochemical markers obtained from postmortem blood
Marker Measured value n p-valuea Regional differenceb Recommended
sampling site
Right cardiac blood Left cardiac blood Femoral vein blood
HbA1c 5.26 ± 0.94 5.23 ± 0.97 5.20 ± 0.99 40 0.0452 r > fe (r = l, l = fe) r, l, fe
t-Bilirubin 1.27 ± 2.12 1.35 ± 2.18 1.06 ± 1.84 34 0.0001 r = l > fe fe
Triglyceride 131.4 ± 144.7 147.4 ± 183.1 130.6 ± 104.6 36 ns r, l, fe
BUN 45.3 ± 42.6 44.1 ± 42.3 45.1 ± 42.5 39 ns r, l, fe
CRP 7.94 ± 10.38 8.12 ± 10.74 6.96 ± 9.27 41 0.0032 r = l> fe fe
c-GTP 161.3 ± 192.4 158.4 ± 183.6 128.9 ± 154.6 38 0.0397 l > fe (r = l, r = fe) fe
Fructsamine 355.9 ± 119.3 355.6 ± 127.0 334.8 ± 106.9 29 ns r, l, fe
Creatinine 3.12 ± 2.32 2.95 ± 2.34 3.30 ± 2.38 40 0.0005 l < r = fe l
Pseudocholine esterase 262.0 ± 137.1 277.3 ± 163.4 250.7 ± 134.0 35 0.0079 l > r = fe l
t-Cholesterol 178.6 ± 96.6 195.4 ± 118.5 174.6 ± 94.7 34 0.0057 l > r = fe l
t-Protein 7.69 ± 1.70 7.78 ± 1.84 6.97 ± 1.89 37 <0.0001 r = l > fe fe
Values are expressed as the mean ± SD.
r, right cardiac blood; l, left cardiac blood; fe, femoral vein blood; ns, not significant.
a
One-way repeated measures ANOVA.
b
Paired t-test.

differences among sampling sites. For total bilirubin, CRP, ysis of all cases. Then, we can obtain a genuine forensic
c-GTP, total protein we recommend femoral vein blood standard.
since these markers tend to increase postmortem and fem- In our study we re-examined commercially available
oral vein blood showed a significantly lower value than the blood markers and presented the possibility of forensic
others. For creatinine, we recommend left cardiac blood diagnosis. For clinical forensic diagnosis using biochemical
since it showed a tendency to increase time-dependently, blood markers, we must collect more samples by known
and there was a lower level in left cardiac blood than fem- causes with more information about the population, and
oral vein blood. For pseudocholine esterase and total cho- thereby determine the ‘‘forensic abnormal value”. Situation
lesterol, we recommend left cardiac blood since it showed a that often require biochemical assessment include sudden
tendency to decrease time-dependently, and showed a unexpected death without obvious cause, such as alcoholics
higher level in left cardiac blood than femoral vein blood. with a low post mortem blood alcohol level, young adults
In fact, left cardiac blood was suitable for measuring the with no apparent cause of death and antemortem informa-
value of creatinine, pseudocholine esterase, total choles- tion such as present and past illnesses.
terol, while femoral vein blood was good for measuring
the other eight markers (Table 4). 5. Conclusion
When biochemical blood markers are used in forensic
practice, we need the reliability of the standard values. In In our postmortem biochemistry study of 164 consecu-
clinical medicine, standard values are calculated from the tive autopsy cases, HbA1c was clearly a reliable marker.
samples obtained from healthy adults (usually 20–40 years Total bilirubin, BUN, CRP and total cholesterol would
of age). Needless to say, there is no perfect standard value have been useful if we had set an appropriate limit range
with blood samples obtained from forensic autopsies and been careful in the interpretation. For the evaluation
because the samples are from the deceased with various of changes due to postmortem intervals, the triglyceride
causes of death, agonal states, complications, treatments value decreased according to the postmortem interval, but
and postmortem changes. These factors generally affect other markers did not show significant changes up to three
the level of markers. This is a serious problem. We tenta- days of postmortem. As for the etiology of deaths, the value
tively tried to calculate a forensic standard value of HbA1c, of asphyxiation showed a higher value that of blunt injury
which shows negligible postmortem changes, and obtained and internal death, reflecting the difference in the cause of
a value of 2.77–7.69% (mean ± 2SD), although our data death. Additionally, femoral vein blood is generally a suit-
included abnormal values due to etiology of death. How- able sampling site for measuring blood biochemical mark-
ever, the limitation of this kind of study is derived from ers because of its relatively slight postmortem changes.
the lack of information on death scene circumstances, pres- Postmortem biochemistry is poorly understood, under
ent and past illness, as well as the various backgrounds of used at present because of concerns about postmortem
the victims. Therefore, we must be aware that above-men- changes and large deviations from healthy subjects, while
tioned standard value is only a guideline. A more accurate it has great potential for forensic service work and future
forensic standard value should be determined after collect- research into the sudden unexpected death without obvious
ing more samples and excluding apparently abnormal data cause, or young adults with no apparent cause of death or
due to pathological changes on the basis of complete anal- antemortem information. As the next step for clinical
 K. Uemura et al. / Journal of Forensic and Legal Medicine 15 (2008) 312–317 317

forensic diagnosis on the basis of biochemical blood mark- special reference to the cause of death and survival time. Forensic Sci
ers, we must determine the ‘‘forensic abnormal value” after Int 2002;130:160–6.
7. Tamaki K, Katsumata Y. Enzyme-linked immunoabsorbent assay for
collecting more samples by known causes with more infor- plasma thyroglobulin following compression of the neck. Forensic Sci
mation about the population. Int 1990;44:193–201.
8. Ansari MQ, Zamora JL, Lipscomb MF. Postmortem diagnosis of
Acknowledgements acute anaphylaxis by serum tryptase analysis. Am J Clin Pathol
1993;99:101–3.
9. Akane A, Shiono H, Matsubara K, Tanabe K, Nakamura H,
We thank Mr. Kakuji Kawai and Ms. Mihoko Nakam- Hasegawa M, et al. Analysis of glycated albumin in postmortem
ura for collecting blood samples in autopsies. blood samples as the diagnostic parameters of diabetes mellitus. Jpn J
Legal Med 1992;46:237–43.
10. Kuroda N. Determination of glycated hemoglobins of cadaveric
References blood: a study involving the use of affinity chromatography with m-
amino phenyl boronic acid. Jpn J Legal Med 1990;44:115–25.
1. Coe JI. Postmortem chemistry update. Am J Forensic Med Pathol 11. Khuu HM, Robinson CA, Brissie RM, Konrad RJ. Postmortem
1993;14:91–117. diagnosis of unsuspected diabetes mellitus established by determi-
2. Tsuji A, Ikeda N, Kojimahara M, Takeichi S. Forensic application of nation of decedent’s hemoglobin A1c level. J Forensic Sci
biochemical blood analysis – experimental studies of postmortem 1999;44:643–6.
changes of value in cadaveric blood. Res Pract Forensic Med 12. Winecker RE, Hammett-Stabler CA, Chapman JF, Ropero-Miller JD.
1993;36:97–103 [in Japanese]. HbA1c as a postmortem tool to identify glycemic control. J Forensic
3. Hart AP, Zumwalt RE, Dasgupta A. Postmortem lipid levels for the Sci 2002;47:1373–9.
analysis of risk factors of sudden death. Usefulness of Ektachem and 13. Goullé J-P, Lacroix C, Bouige D. Glycated hemoglobin: a useful post-
Monarch analyzers. Am J Forensic Med Pathol 1997;18:354–9. mortem reference marker in determining diabetes. Forensic Sci Int
4. Tsuji A, Ikeda N, Kojimahara M, Takeichi S. Forensic application of 2002;128:44–9.
biochemical blood analysis – normal range of cadaveric blood. Res 14. Valenzuela A. Postmortem diagnosis of diabetes mellitus. Quantita-
Pract Forensic Med 1992;35:105–8 [in Japanese]. tion of fructosamine and glycated hemoglobin. Forensic Sci Int
5. Uhlin-Hansen L. C-reactive protein (CRP), a comparison of pre- and 1988;38:203–68.
post-mortem blood levels. Forensic Sci Int 2001;124:32–5. 15. Zhu BL, Ishida K, Quan L, Taniguchi M, Oritani S, Li DR, et al.
6. Fujita MQ, Zhu BL, Ishida K, Quan L, Oritani S, Maeda H. Serum Postmortem serum uric acid and creatinine levels in relation to the
C-reactive protein levels in postmortem blood – an analysis with causes of death. Forensic Sci Int 2002;125:59–66.