Forensic Science International 263 (2016) 107113

Contents lists available at ScienceDirect

Forensic Science International

journal homepage: www.elsevier.com/locate/forsciint

Interpretation of postmortem vitreous concentrations of sodium and

chloride
B. Zilg a,*, K. Alkass a, S. Berg b, H. Druid a
a
Forensic Medicine Laboratory, Department of Oncology-Pathology, Karolinska Institutet, Retzius v. 3, SE-171 77 Stockholm, Sweden
b
Department of Medicine and Health, Linkoping University, University Hospital, SE-581 85 Linkoping, Sweden

A R T I C L E I N F O A B S T R A C T

Article history: Vitreous uid can be used to analyze sodium and chloride levels in deceased persons, but it remains
Received 14 December 2015 unclear to what extent such results can be used to diagnose antemortem sodium or chloride imbalances.
Accepted 4 April 2016 In this study we present vitreous sodium and chloride levels from more than 3000 cases. We show that
Available online 11 April 2016
vitreous sodium and chloride levels both decrease with approximately 2.2 mmol/L per day after death.
Since potassium is a well-established marker for postmortem interval (PMI) and easily can be analyzed
Keywords: along with sodium and chloride, we have correlated sodium and chloride levels with the potassium
Postmortem
levels and present postmortem reference ranges relative the potassium levels. We found that virtually all
Vitreous
cases outside the reference range show signs of antemortem hypo- or hypernatremia. Vitreous sodium or
Sodium
Chloride chloride levels can be the only means to diagnose cases of water or salt intoxication, beer potomania or
Hypernatremia dehydration. We further show that postmortem vitreous sodium and chloride strongly correlate and in
Hyponatremia practice can be used interchangeably if analysis of one of the ions fails.
It has been suggested that vitreous sodium and chloride levels can be used to diagnose drowning or to
distinguish saltwater from freshwater drowning. Our results show that in cases of freshwater drowning,
vitreous sodium levels are decreased, but that this mainly is an effect of postmortem diffusion between
the eye and surrounding water rather than due to the drowning process, since the decrease in sodium
levels correlates with immersion time.
2016 Elsevier Ireland Ltd. All rights reserved.

1. Introduction In this communication we focus on the diagnosis of deranged

The postmortem diagnosis of various electrolyte imbalances is
often difcult to make due to the degradation of the many cells in
the blood, which makes blood or serum unsuitable for such
analyses. Other body uids that can be used for postmortem
chemistry are vitreous, cerebrospinal, pericardial and synovial
uid. Among these, vitreous uid has become the matrix of choice
in forensic pathology, not only because it is much easier to obtain,
but also because of its isolated position, which makes it less
affected by postmortem contamination and putrefaction [1
3]. Further, vitreous uid is almost devoid of cells, and the
changes occurring after death are almost exclusively an effect of
postmortem exchange of ions and other endogenous compounds
between the vitreous and the surrounding cells. Hence, a large
number of postmortem biochemistry studies have been conducted
on vitreous uid, aiming to provide keys to various diagnoses.
* Corresponding author. Tel.: +46 704930844.
E-mail address: brita.zilg@gmail.com (B. Zilg).
sodium and chloride levels. These can arise from a variety of
illnesses, such as kidney failure, liver failure, cancer or diarrhea. In
a forensic setting cases may, for example, concern:
 dehydration may be very important to diagnose properly,
especially in cases of neglect of children or the elderly [1,4];
 salt intoxication mostly seen in children who have been force-
fed salt as a punishment or due to a Munchhausen by proxy
syndrome [4,5];
 water intoxication a condition called psychogenic polydypsia,
where psychiatric patients compulsively drink large amounts of
water [6]. This can lead to brain swelling, seizures and death.
Water intoxications can also occur during MDMA (3,4-methy-
lenedioxy-methamphetamine, ecstasy) use [7];
 beer potomania a condition similar to water intoxication. Large
ingestion of beer, together with poor food intake, can lead to
severe hyponatremia [8];
 drowning hypothetically, aspiration of water during drowning
leads to hyponatremia in freshwater and to hypernatremia in
saltwater. Some authors have proposed that vitreous sodium/
chloride levels can be used to distinguish saltwater from

http://dx.doi.org/10.1016/j.forsciint.2016.04.006
0379-0738/ 2016 Elsevier Ireland Ltd. All rights reserved.
108 B. Zilg et al. / Forensic Science International 263 (2016) 107113

Table 1 (n = 3065). Fourteen cases were excluded due to technical

Postmortem changes in vitreous sodium and chloride according to different
analytical problems; otherwise, no cases were excluded. In order
authors.
to study the changes of sodium and chloride concentrations with
Author (year) Sodium Chloride postmortem time, 462 cases with a known postmortem interval
Coe (1969) [22] # # were selected from this period, and 62 cases from 2012 (altogether
Blumenfeld (1979) [23] ! ! n = 524). When possible changes in sodium and chloride con-
Balasooriya (1984) [17] # centrations due to different medical conditions were studied, the
Farmer (1985) [24] #
concentrations were related either to the true postmortem interval
Madea (2001) [25] ! !
Tao (2006) [26] # # (PMI), or to the potassium concentration as a proxy for the PMI. The
Jashnani (2010) [20] ! ! latter approach provided a much larger material for comparison.
Tumram (2011) [21] ! # For all cases, information was extracted from the Swedish
Chandrakanth (2013) [27] # #
national forensic medicine database [17] and linked to the results
Mitchell (2013) [18] # #
Siddamsetty (2014) [19] # !
obtained by analysis of vitreous uid with the blood gas instrument
before being anonymized. The original forensic pathology case les
were perused to select cases with a certied time of death, to search
freshwater drownings [9] or drownings from non-drownings
for antemortem analytical results whenever copies of medical
[10].
records were available and to more closely study the circumstances
surrounding death and the autopsy ndings.
Sodium is the major cation and chloride the major anion in the
Approximately 0.2 mL vitreous uid was aspirated from the
extracellular uid, with normal levels being 135145 mmol/L and
center of each eye using a 1 mL syringe equipped with an 18-gauge
95110 mmol/L, respectively. Sodium is the major osmotically
needle. The samples from each decedent were collected in the
active ion, so the serum sodium concentration normally reects
same syringe and were not pre-treated, and hence not subjected to
the osmolarity of the extracellular uid and total body sodium
dilution, centrifugation, or sonication. The samples were instead
content determines the volume of the extracellular uid compart-
directly injected into an ABL 625 blood gas instrument (ABL 625,
ment. Water balance and plasma osmolarity, and thereby serum
Radiometer Copenhagen, Brnshj, Denmark). Samples that were
sodium levels, are tightly regulated by the kidneys, which
macroscopically opaque or dirty were also directly analyzed.
continuously make adjustments to the concentration of urine in
However, if there was a massive bleeding in the eye, no analysis of
order to maintain constant plasma osmolarity. Extreme sodium
vitreous potassium analysis was performed. In addition, samples
and chloride imbalances can cause the cells in the body to swell or
from severely decomposed bodies and infants were not consis-
shrink and can be fatal [11].
tently included, since toxicological analyses were prioritized for
There are no reliable published data on human vitreous
several of these cases if the vitreous uid volume was very small.
electrolyte levels in vivo. Studies on various animals, however,
have shown that vitreous sodium levels are 95% of those in blood
2.2. Analytical principles
and vitreous chloride levels are 110% of those in blood [1215].
The postmortem changes in vitreous sodium and chloride levels
All analyses were made with a blood gas instrument (ABL 625,
have been studied extensively. Table 1 summarizes ndings by
Radiometer Copenhagen, Brnshj, Denmark). With this instru-
different authors. There seems to be no real consensus about the
ment it is possible to determine the levels of electrolytes, pH, PCO2,
postmortem changes of these electrolytes, although it should be
PO2, glucose, lactate and Hb. For the determination of the
mentioned that the postmortem changes that were found by some
concentrations of the electrolytes, a voltmeter is used for
authors were marginal.
measuring the difference in potential generated across an ion-
In an article from 2005 on the postmortem diagnosis of
specic membrane between the sample and a reference solution
hypertonic dehydration by Madea and Lachenmeier [16] it is stated
for the specic analyte.
that there are several conceptual problems surrounding postmor-
tem vitreous sodium values that still have to be resolved, for
2.3. Ethical aspects
example the distribution of postmortem vitreous values in
comparison to serum values in vivo and the postmortem changes.
All analyses were performed as a part of the forensic medicine
The aims of this study were:
investigation and were reported to the responsible forensic
pathologist. Therefore, ethical permission was not required for
 To evaluate the postmortem changes of vitreous sodium and
the sampling and analyses conducted according to Swedish
chloride.
regulations. However, ethical approval was obtained for the
 To investigate whether postmortem vitreous sodium levels
formation of a database for these data, and for the perusal of
reect antemortem serum sodium levels and thus can be used to
forensic medicine les, including police reports and medical records
identify antemortem imbalances.
(Regional Ethical Review Board, Stockholm 2008/231-31/3).
 To correlate sodium and chloride levels to specic causes of
death.
2.4. Statistical methods
 To establish postmortem vitreous reference concentrations of
sodium and chloride.
Correlation between variables was assessed by linear regres-
sion, and differences between groups were assessed with the
2. Materials and methods nonparametric KruskalWallis analysis of variance (Statistica v10,
StatSoft Inc., Tulsa, OK, Excel v14.5.5, Microsoft and StatPlus v5,
2.1. Study design AnalystSoft Inc.).

Between January 2003 and June 2006, vitreous uid samples 3. Results
were consistently collected from all deceased subjects admitted to
the Department of Forensic Medicine (Stockholm, Sweden) as soon In Figs. 1 and 2, postmortem vitreous sodium and chloride
as possible following the arrival of the bodies at the morgue concentrations are plotted against potassium concentrations. Since
 B. Zilg et al. / Forensic Science International 263 (2016) 107113 109

190

170

150
Sodium mmol/L

130

110

90
y = -0.8986x + 145.09
r = 0.33117
n=3069
70

50
0 5 10 15 20 25 30 35 40
Potassium mmol/L

Fig. 1. Sodium vs potassium. Black lines indicate mean and the 95% CI, suggesting Fig. 4. Chloride vs PMI. n = 524.
cut-off levels for hypo- and hypernatremia, considering matching potassium levels.

180 analysis. All cases outside the 95% CI for sodium were reviewed
160 regarding the cause of death and are presented in Tables S1 and S2.
Virtually all cases outside the 95% CI had a plausible explanation
140 for their deviating sodium and chloride levels, e.g. water
intoxication and freshwater drowning.
Chloride mmol/L

120
We found that sodium and chloride levels strongly correlated
100 (r2 = 0.82; p = 0.000, Fig. 5). Hypernatremia together with hypo-
80 chloremia or vice versa are extremely rare. This means that in most
cases it is sufcient to only analyze sodium. In cases where sodium
60 y = -0.9694x + 127.34 cannot be analyzed (for example when sodium uoride has been
r = 0.29805

40
n=3069
20
0 5 10 15 20 25 30
Potassium mmol/L
Fig. 2. Chloride vs potassium. Black lines indicate mean and the 95% CI.
potassium is a well-established marker of the postmortem interval
(PMI) [3,1820], we chose to use potassium as a proxy for PMI. The
advantage of using potassium instead of PMI is that the PMI often is
unknown or uncertain and thus not available for all cases, whereas
the vitreous potassium concentration can be included in the same
analytical run as other analytes. Figs. 1 and 2 show that sodium and
chloride levels slowly decrease with increasing potassium levels
(p = 0.000). In 524 cases an exact time of death was available
sodium and chloride levels vs true PMI are shown in Figs. 3 and
4. Sodium and chloride levels both decrease with approximately
2.2 mmol/L per day.
The dashed lines in Figs. 1 and 2 represent the mean and 95% CI
in relation to the potassium levels, based on linear regression
added to the test tube to prevent bacterial growth), chloride can be
analyzed instead.
In 46 cases, serum sodium values that were obtained shortly
35 40
before death were available. Antemortem serum sodium values
correlate very well with postmortem vitreous sodium values,
especially in cases with low potassium values, i.e. short PMIs. For
longer PMIs, postmortem vitreous sodium values are gradually
lower, as expected (Fig. 6). In these cases, we calculated the
expected sodium values at time of death, through extrapolation of
postmortem vitreous sodium by using the vitreous potassium
values and the regression equation. We then compared them to the
actual serum sodium values that were obtained shortly before
death (Fig. 7).
Table 2 and Fig. 8 show statistical data for sodium and chloride
in 658 sudden, violent death cases for different potassium ranges.
These cases represent a presumed healthy normal population
without any antemortem electrolyte imbalances and thus can be
used as a reference range. Fig. 9 shows these cases compared to the
whole material. There are virtually no cases of sudden, violent
death with sodium values outside the 95% condence intervals.

Fig. 3. Postmortem vitreous sodium concentration vs PMI. n = 524. Fig. 5. Sodium levels vs chloride levels. n = 3071.
110 B. Zilg et al. / Forensic Science International 263 (2016) 107113

Figure 8. Vitreous sodium levels in cases of sudden, violent death. n = 658.

Fig. 6. Comparison antemortem hospital serum sodium levels with postmortem

vitreous. n = 46. The black line indicates identity.

Fig. 9. Vitreous sodium levels in cases of sudden, violent death compared to other
cases.

Vitreous sodium levels in cases of drownings were examined.

Fig. 10 shows that victims of freshwater drownings have
signicantly lower sodium levels (median 129 mmol/L, inter-
quartile range 14 mmol/L) than victims drowning in the brackish
water of the Baltic Sea (median 134 mmol/L, IQR 9 mmol/L)
(p < 0.01) and non-drowning persons (median 135 mmol/L, IQR
11 mmol/L) (p < 0.001). Victims of drowning in the Baltic Sea have
similar sodium levels as non-drownings (p = 1.0). We found that

Fig. 7. Comparison antemortem hospital serum sodium levels with calculated

sodium values at time of death. n = 46. The solid black line indicates identity.

Table 2
Statistical data from cases of sudden, violent death. All values are in mmol/L.

Sodium Chloride

All cases (n = 658)

Min 109 91
Mean  SD 135  6.1 119  6.6
Max 148 136
Low potassium (49 mmol/L) (n = 339)
Min 118 101
Mean  SD 138  3.9 121  4.5
Max 148 136
Medium potassium (1019 mmol/L) (n = 261)
Min 115 92
Mean  SD 134  5.1 118  6.1
Max 148 134
High potassium (2035 mmol/L) (n = 58)
Fig. 10. Sodium levels in fresh water drownings, Baltic Sea drownings and non-
Min 109 91
drownings. The inside gure of each box represents the median. The lower and
Mean  SD 124  6.3 108  8.0
upper edges of the boxes represent the 25th and 75th percentiles, respectively.
Max 148 134
Upper and lower lines outside the boxes represent minimum and maximum values.
 B. Zilg et al. / Forensic Science International 263 (2016) 107113
200
180
160
Sodium, mmol/L
140
120
100
80
60
00 05 10 15 20 25 30 35
Potassium, mmol/L
Fig. 11. Sodium levels in fresh water drownings with different immersion times.
the low sodium values in freshwater drownings strongly correlate
with immersion time (Fig. 11).
4. Discussion
4.1. Postmortem changes
Our data show that vitreous sodium and chloride levels both
slowly decrease with approximately 2.2 mmol/L per day after
death. As can be seen in Table 1, previous studies have been
discordant regarding the postmortem stability of vitreous sodium
and chloride levels. This may be due to smaller sample size and a
narrower PMI bracket with a concentration of data points among
short PMIs.
The reason for the postmortem decrease in vitreous sodium and
chloride levels is most likely the diffusion of these electrolytes into
the surrounding cells of the retina and choroid. Intracellular uid
has low sodium and chloride concentrations and high potassium
concentrations and hence the same mechanism, diffusion but in
the opposite direction, is responsible for the postmortem increase
in vitreous potassium. In fact, when the decrease in sodium and
chloride levels is compared with the increase in potassium, the
sum at any PMI remains virtually unchanged (compare [21]). Since
vitreous uid is gradually reduced postmortem, it could be
assumed that the electrolytes would become more concentrated;
however this is contradicted by the constant sum of electrolyte
concentrations. Having said that, it is possible that such
concentration might occur at extreme PMIs, when only a very
small amount of uid remains. There seems to be a attening of
both the chloride and the sodium curve at the very end of the
interval, see Figs. 3 and 4, but this may just as well represent the
nal equilibration concentrations, just like the potassium curve
becomes asymptotic with no tendency to increase further at very
long PMIs [21].
A negative correlation between sodium and PMI means of
course that sodium and chloride levels should be evaluated
together with either the PMI (when known) or with the potassium
concentration. We think however that it often is more practically
feasible to use the potassium concentration, since an exact time of
death often is uncertain or unavailable, whereas the potassium
concentration can be easily obtained from the same analytical run.
Table 2 and Fig. 8 show postmortem vitreous levels for sodium
and chloride in sudden, violent deaths at different potassium
ranges. There is no reason to believe that these subjects had any
major electrolyte imbalances, since they were all active immedi-
ately prior to death. Hence, these concentration ranges and the
regression equation can serve as a guidance for postmortem
normal levels of sodium and chloride.
111
Sodium and chloride levels strongly correlate (Fig. 5). This
means that in practical casework, they can be used interchange-
ably. Even if there are conditions, such as extreme chronic
vomiting that may change sodium or chloride unevenly, we did not
nd any such cases in our material.
4.2. Hyponatremia
Table S1 presents all cases with sodium levels below the 95%
condence intervals (2 SD below the mean), which represent
presumed hyponatremia (n = 96). In virtually all of these cases, the
low sodium levels could be explained by illnesses or circum-
stances. There were for example 17 cases where the cause of death
40 45
was freshwater drowning; these are discussed in detail below.
There were 13 cases where the cause of death was pneumonia.
Hyponatremia is common in patients with community-acquired
pneumonia at admission, and is related to the severity of the illness
and mortality. The hyponatremia is often associated with
increased levels of ADH (antidiuretic hormone) [22,23].
There was one case of water intoxication due to massive
ingestion of water. Two other interesting cases of suspected water
intoxication stand out:
Case 1 (number 26 in Table S1): A 41 year old male with paranoid
schizophrenia was admitted to a psychiatric rehab clinic, where he
suddenly had seizures and vomited watery content, fell on his back
and died despite CPR. He had not eaten anything for two days but
had been drinking water. The staff had noted a swollen face and
stomach. The autopsy revealed an enlarged heart and liver. Blood
toxicology showed fairly low levels of haloperidol, alimemazine and
zuclopentixol. Cause of death was determined as an acute
complication to heart disease with possible contribution of side
effects of medication.
Case 2 (number 15 in Table S1): A 54 year old male with epilepsy,
social anxiety disorder and depression, living in a sheltered
housing, was known to occasionally drink large amounts of water.
One day he started suddenly to have seizures. The nurse who
performed CPR reported that water came out of the mouth during
compressions until there was a small pool on the oor. Autopsy
revealed an enlarged heart and general atherosclerosis. Blood
toxicology showed 14 mg/g caffeine and fairly low levels of
sertraline, phenytoin and paracetamol. Cause of death was
determined as intoxication with caffeine.
In both cases vitreous sodium/chloride levels were available
and showed 104/88 mmol/L and 102/84 mmol/L respectively. Due
to lack of previous studies on vitreous sodium and chloride
imbalances at the time (cases are from 2006 and 2003), it seems
that the postmortem vitreous analyses were not taken into
account. In retrospect however, a more appropriate cause of death
would probably have been water intoxication due to psychogenic
polydipsia. These cases illustrate the usefulness of postmortem
vitreous chemistry in certain situations.
The remainder of the cases in the hyponatremia group were
persons with chronic alcoholism, liver disease, pancreatitis or
heart disease, all conditions which are known to cause hypona-
tremia [2426].
4.3. Hypernatremia
In the hypernatremia group (Table S2, n = 82), there are also quite
a few cases where the cause of death was established as pneumonia
(n = 17). Hypernatremia has also been described to be associated
with community-acquired pneumonia, although much less common
than hyponatremia. Both conditions, however, carry an increased
mortality risk [23]. The cause of hypernatremia could be dehydration
112 B. Zilg et al. / Forensic Science International 263 (2016) 107113
due to fever with uid loss and lack of proper uid intake prior to
death. In some of the cases, the cause of death may even have been
dehydration in combination with illness, rather than the illness
alone. This may be true not only for cases of pneumonia but also for
other cases where the subject was found dead home alone due to, for
example, intoxications (n = 6) or heart disease (n = 12).
There were 5 cases of diabetic coma, a condition which may
cause hypernatremia due to dehydration caused by polyuria [27].
Twenty-two cases in the hypernatremia group were victims of
severe trauma who did not die instantly but became unconscious
and died some time later at an intensive care unit. Most of these
cases involved head trauma. A plausible mechanism for the high
sodium levels in this group is that the brain injury may cause
cerebral edema and subsequent impairment of secretion of ADH
from the hypothalamus, in turn causing massive diuresis and
hypernatremia a condition known as central neurogenic diabetes
insipidus [28]. The same may be true for 11 subjects with brain
injuries, strokes or spontaneous subdural hemorrhages who were
found dead at home. There were also two cases of carbon
monoxide intoxication, a condition that also can cause diabetes
insipidus through damage to the neurohypophysis [29].
In summary, it should be noted that there were virtually no
cases of sudden, violent death of presumed healthy persons in the
hypo- and hypernatremia groups, suggesting that vitreous
sodium and chloride levels corrected for PMI or potassium level
is a robust means for predicting the antemortem levels. This can
be seen in Fig. 9, where cases of sudden, violent death (i.e. cases
where one would not expect any sodium imbalances) are
compared to the whole material. There are only 3 cases of
sudden, violent death (out of 658) with sodium values outside the
95% condence intervals.
Madea and Lachenmeier [16] stated that elevated postmortem
vitreous sodium and chloride values can occur in other causes of
death than dehydration. This may be due to analytical/methodo-
logical issues. The problem with variations in vitreous sodium and
chloride levels, when determined by different analytical methods,
was already pointed out by Coe [1]. We consider the method of
analysing vitreous samples with a blood gas instrument to be very
robust. This is supported by the data in Fig. 6 showing cases in
which antemortem serum sodium values were available from
samples taken shortly prior to death. In cases with a short PMI
(=low vitreous potassium), the values are almost identical. In cases
with longer PMIs, vitreous sodium levels are lower, as expected.
Using the slope in Fig. 8, we calculated the postmortem vitrous
sodium levels back to a potassium level of 4.4 and compared these
values to the actual serum sodium levels obtained shortly before
death at a hospital (Fig. 7). The values correlated fairly well.
4.4. Drowning
Previous studies have shown that vitreous sodium and chloride
levels are increased in cases of saltwater drownings and decreased
in freshwater drownings. However, there is no consensus
regarding whether this is a result of the drowning process with
concentration or dilution of the body water following inhalation/
swallowing of water or if it is due to postmortem diffusion
between the vitreous and surrounding water. Cala et al. [10] have
studied saltwater drownings and found signicantly higher
vitreous sodium and chloride levels in saltwater drownings
compared to bodies retrieved from saltwater that had died of
other causes. The immersion time of these bodies ranged from
several minutes up to 8 h in one case. Byard and Summersides [9]
compared saltwater drownings with freshwater drownings and
found higher vitreous sodium levels in the saltwater group than in
the freshwater group. They concluded, however, that it is unclear
whether this was an effect of perimortem water aspiration or
postmortem diffusion. Sturner et al. [30] studied electrolytes in
bovine eyeballs immersed in saltwater and found a steady increase
of electrolytes with immersion time. Farmer et al. [31] concluded
that changes in vitreous electrolytes in drowning cases rather is an
effect of postmortem diffusion and proposed that these changes
might be used to estimate immersion time.
There are no saltwater drownings in our material, only
drownings in the brackish water of the Baltic Sea in the Stockholm
area (salinity 0.65%, sodium 90 mmol/L) and freshwater. Our
study conrms that vitreous sodium and chloride levels are lower
in freshwater drownings than in non-drownings. The sodium and
chloride levels in drownings in Baltic Sea water are virtually the
same as in non-drownings, which seems reasonable since the salt
level in the Baltic Sea is similar to human serum levels. In Fig. 11 it
can be seen that long immersion time was associated with lower
sodium levels, suggesting that this is an effect of postmortem
diffusion. It can also be seen that in cases with a very short
immersion time, there is no decrease in sodium levels, suggesting
that there is no effect on vitreous sodium concentration of
hemodilution due to water inhalation. Even if electrolyte
imbalances do occur in blood, it is unlikely that these imbalances
would show in the vitreous, since it takes some time for the sodium
blood levels to equilibrate with the vitreous levels [32].
5. Conclusion
Our study has shown that postmortem vitreous sodium and
chloride levels slowly decrease with the postmortem interval.
Postmortem vitreous sodium levels correlate very well with
antemortem serum sodium levels, when interpreted together with
postmortem vitreous potassium levels or PMI. Sodium levels may
frequently help to establish the cause of death, or at least the
antemortem condition of the subject prior to death, and analysis
should be considered routinely at autopsy.
Acknowledgements
We thank the staff at the Department of Forensic Medicine in
Stockholm. These studies were supported in part by the Swedish
National Board of Forensic Medicine and the Swedish Medical
Society.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.forsciint.2016.04.
006.
References
[1] J.I. Coe, Postmortem chemistry update. Emphasis on forensic application, Am. J.
Forensic Med. Pathol. 14 (1993) 91117.
[2] C. Palmiere, P. Mangin, Postmortem chemistry update part I, Int. J. Legal Med. 126
(2012) 187198.
[3] B. Madea, Is there recent progress in the estimation of the postmortem interval by
means of thanatochemistry, Forensic Sci. Int. 151 (2005) 139149.
[4] R.E. Zumwalt, C.S. Hirsch, Subtle fatal child abuse, Hum. Pathol. 11 (1980)
167174.
[5] R. Meadow, Non-accidental salt poisoning, Arch. Dis. Childhood 68 (1993)
448452.
[6] W.V. Vieweg, J.J. David, W.T. Rowe, G.J. Wampler, W.J. Burns, W.W. Spradlin, Death
from self-induced water intoxication among patients with schizophrenic disor-
ders, J. Nerv. Mental Dis. 173 (1985) 161165.
[7] C.M. Milroy, Ecstasy associated deaths: what is a fatal concentration? Analysis of a
case series, Forensic Sci. Med. Pathol. 7 (2011) 248252.
[8] A.Z. Fenves, S. Thomas, J.P. Knochel, Beer potomania: two cases and review of the
literature, Clin. Nephrol. 45 (1996) 6164.
[9] R.W. Byard, G. Summersides, Vitreous humor sodium levels in immersion deaths,
J. Forensic Sci. 56 (2011) 643644.
 B. Zilg et al. / Forensic Science International 263 (2016) 107113
[10] A.D. Cala, R. Vilain, R. Tse, Elevated postmortem vitreous sodium and chloride
levels distinguish saltwater drowning (SWD) deaths from immersion deaths not
related to drowning but recovered from saltwater (DNRD), Am. J. Forensic Med.
Pathol. 34 (2013) 133138.
[11] B. David, M.H.S. Mount, K. Ajay, Singh Core Concepts in the Disorders of Fluid,
Electrolytes and AcidBase Balance, Springer, 2013.
[12] D.V. Reddy, V.E. Kinsey, Composition of the vitreous humor in relation to that of
plasma and aqueous humors, Arch. Ophthalmol. 63 (1960) 715720.
[13] I.W. Wilkie, J.E. Bellamy, Estimation of antemortem serum electrolytes and urea
concentrations from vitreous humor collected postmortem, Can. J. Comp. Med.
(Revue canadienne de medecine comparee), 46 (1982) 146149.
[14] P.S. McLaughlin, B.G. McLaughlin, Chemical analysis of bovine and
porcine vitreous humors: correlation of normal values with serum chemical
values and changes with time and temperature, Am. J. Vet. Res. 48 (1987)
467473.
[15] B.G. McLaughlin, P.S. McLaughlin, Equine vitreous humor chemical concentra-
tions: correlation with serum concentrations, and postmortem changes with time
and temperature, Can. J. Vet. Res. (Revue canadienne de recherche veterinaire), 52
(1988) 476480.
[16] B. Madea, D.W. Lachenmeier, Postmortem diagnosis of hypertonic dehydration,
Forensic Sci. Int. 155 (2005) 16.
[17] H. Druid, P. Holmgren, P. Lowenhielm, Computer-assisted systems for forensic
pathology and forensic toxicology, J. Forensic Sci. 41 (1996) 830836.
[18] L. Adelson, I. Sunshine, N.B. Rushforth, M. Mankoff, Vitreous potassium concen-
tration as an indicator of the postmortem interval, J. Forensic Sci. 8 (1963)
503514.
[19] W.Q. Sturner, G.E. Gantner Jr., The postmortem interval. A study of potassium in
the vitreous humor, Am. J. Clin. Pathol. 42 (1964) 137144.
[20] L. Hansson, U. Uotila, R. Lindfors, K. Laiho, Potassium content of the vitreous body
as an aid in determining the time of death, J. Forensic Sci. 11 (1966) 390394.
113
[21] B. Zilg, S. Bernard, K. Alkass, S. Berg, H. Druid, A new model for the estimation of
time of death from vitreous potassium levels corrected for age and temperature,
Forensic Sci. Int. 254 (2015) 158166.
[22] Z.V. Edmonds, Hyponatremia in pneumonia, J. Hosp. Med. 7 (Suppl. 4) (2012)
S11S13.
[23] S. Kruger, S. Ewig, S. Giersdorf, O. Hartmann, D. Frechen, G. Rohde, et al.,
Dysnatremia, vasopressin, atrial natriuretic peptide and mortality in patients
with community-acquired pneumonia: results from the German competence
network CAPNETZ, Respir. Med. 108 (2014) 16961705.
[24] G.L. Liamis, H.J. Milionis, E.C. Rizos, K.C. Siamopoulos, M.S. Elisaf, Mechanisms of
hyponatraemia in alcohol patients, Alcohol Alcohol. 35 (2000) 612616.
[25] T.D. Filippatos, M.S. Elisaf, Hyponatremia in patients with heart failure, World J.
Cardiol. 5 (2013) 317328.
[26] J.G. Verbalis, Disorders of body water homeostasis, Best Pract. Res. Clin. Endocrin.
Metab. 17 (2003) 471503.
[27] G. Liamis, E. Liberopoulos, F. Barkas, M. Elisaf, Diabetes mellitus and electrolyte
disorders, World J. Clin. Cases 2 (2014) 488496.
[28] P. Hadjizacharia, E.O. Beale, K. Inaba, L.S. Chan, D. Demetriades, Acute diabetes
insipidus in severe head injury: a prospective study, J. Am. Coll. Surg. 207 (2008)
477484.
[29] M.Y. Chang, J.L. Lin, Central diabetes insipidus following carbon monoxide poi-
soning, Am. J. Nephrol. 21 (2001) 145149.
[30] W.Q. Sturner, A. Balko, J. Sullivan, Magnesium and other electrolytes in
bovine eyeballs immersed in sea water and other uids, Forensic Sci. 8 (1976)
139150.
[31] J.G. Farmer, F. Benomran, A.A. Watson, W.A. Harland, Magnesium, potassium,
sodium and calcium in post-mortem vitreous humour from humans, Forensic Sci.
Int. 27 (1985) 113.
[32] Y. Honda, A. Negi, S. Kawano, Mode of ion movements into vitreous. Equilibration
after vitrectomy, Arch. Ophthalmol. 101 (1983) 105111.