Forensic Science International 133 (2003) 190196

Organ weight effects of drowning and asphyxiation

on the lungs, liver, brain, heart, kidneys, and spleen
Jeffrey A. Hadleya,*, David R. Fowlerb
a
National Study Center for Trauma and Emergency Medical Services, School of Medicine,
University of Maryland at Baltimore, 701 West Pratt Street, Room 517, Baltimore, MD 21201, USA
b
Office of The Chief Medical Examiner, Baltimore, MD, USA

Received 26 November 2001; accepted 22 January 2003

Abstract
An examination of the organ weights associated with victims of drowning, asphyxiation and trauma was undertaken to
determine (a) the effects of asphyxiation compared to a trauma group, and in turn, (b) the effects of drowning compared to an
asphyxiation group. Included in the study were 217 drowning deaths, 166 pure asphyxiation deaths and 381 trauma deaths. The
effects of asphyxiation (compared to trauma) resulted in elevated mean organ weights for the lungs, liver, kidneys and spleen
(with mean increases of 17.8, 10.5, 10.3 and 23.4%, respectively). Effects of drowning (compared to asphyxiation) resulted in
elevated mean organ weights only with the lungs and kidneys (with mean increases of 30.0 and 4.4%, respectively). Only the
mean heart and brain weight remained constant across all experimental groups. A picture of drowning is suggested in which
elevated lung and kidney weights are the result of both asphyxiation and the aspiration of water that occurs with drowning,
whereas elevated spleen and liver weights in drowning victims are associated with only the effects of asphyxiation. In addition,
the common autopsy finding of a small, anemic spleen in drowning, rather than caused by some pathophysiological mechanism
of death, is hypothesized to be a postmortem phenomenon.
# 2003 Elsevier Science Ireland Ltd. All rights reserved.
Keywords: Drowning; Postmortem effects; Asphyxiation; Organ weight; Endogenous alcohol

1. Introduction
One of the most common autopsy findings in drowning
cases is a heavy, edematous lung in which cut surfaces exude
large amounts of fluid [1]. This is assumed to occur primarily
as the result of water that is aspirated into the lungs before
death. At the same time, although studies have shown higher
lung weight in drowning than in non-drowning trauma [24],
only Ito et al. [4] provided a means to determine the effects
of drowning on the lungs beyond that due to asphyxiation.
Surprisingly, a comparison of lung weight/body weight
ratios across drowning and asphyxiation groups did not find
any statistically significant effects. This result suggests that
the autopsy finding of an edematous lung is not specific to
drowning per se, but is caused by mechanisms associated
*
Corresponding author. Tel.: 1-410-328-2683.
E-mail address: jhadley@som.umaryland.edu (J.A. Hadley).

with asphyxiation (or other secondary effects). Of the lungs,

liver, brain, kidneys, heart, spleen, and adrenal glands, only
the spleen in Ito et al.s [4] study demonstrated an effect of
drowning on organ weight (beyond that due to asphyxiation).
In this case, the spleen weight/body weight ratio was actually
lower in drowning than asphyxiation, a finding that has been
replicated by Haffner et al. [5] and discussed as consistent
with the autopsy finding of a small, anemic spleen sometimes observed in drowning victims.
The initial objective of the present study was to test the
hypothesis that the above findings by Ito et al. [4] and
Haffner et al. [5] was confounded by a failure to eliminate
effects of submersion time due to decomposition or other
effects, in which a drop in organ weight occurs over time in
the water [2]. Although a 24 h submersion time cutoff was
used for inclusion of cases in Ito et al.s [4] study, no
drowning victims had submersion times that were less than
8 h. Haffner et al. [5] addressed the effects of submersion

0379-0738/03/$ see front matter # 2003 Elsevier Science Ireland Ltd. All rights reserved.
doi:10.1016/S0379-0738(03)00069-0

J.A. Hadley, D.R. Fowler / Forensic Science International 133 (2003) 190196

time by excluding cases with evidence of decomposition.

This practice however depends upon the assumption that a
drop in organ weight does not occur before decomposition
can be noted at autopsy. Indeed, it has been shown that
decompositional processes occur at a faster rate under water
than in air [6]. As a result, these effects of (a) a lighter spleen
in drowning (compared to asphyxiation) and (b) a lack of an
effect of drowning for the lungs as well as for the brain,
heart, and kidneys (compared to asphyxiation) may be an
artifact of this loss of organ weight that occurs after death
(but before examination of the body). In the present study,
only those drowning victims that had been found within 6 h
of death were included in the study, thus providing an
estimate of organ weight that largely eliminates the confounding effects of a postmortem drop in weight prior to
recovery (and refrigeration) of the body.
The question of interest in the present study was whether
effects on organ weight in a drowning death are the result of
asphyxiation, the aspiration of water into the lungs, or both.
To tease apart these effects of drowning on organ weight, a
trauma group served as the baseline for an asphyxiation
group, which in turn served as the baseline for a drowning
group. Because asphyxial processes exist for the drowning as
well as asphyxiation groups, any additional effects presumably due to the aspiration of water was examined in the
drowning group. An attempt was then made to replicate Ito
et al.s [4] findings of (a) an effect of asphyxiation with
heavier brain, lungs, kidneys, and spleen (compared to a
trauma group), and (b) an effect of drowning that was limited
to a lighter spleen weight (compared to an asphyxiation
group). In addition, whereas the latter finding would also
replicate Haffner et al.s [5] identical results (of the spleen
only), a replication of Ito et al.s [4] finding of a heavier
spleen in asphyxiation compared to trauma would be counter
to their suggestion that this finding is consistent with the
autopsy finding of a small, dry spleen in drowning victims.

191

The trauma group was further divided into head-only and

body-only groups. In the head-only group, 250 homicidal
victims suffered trauma to the head (but not the body). In the
body-only group, 131 homicidal victims suffered trauma to
the body (but not the head). Any individual organ that was
shown at autopsy to sustain an injury was excluded from the
study. In the asphyxiation group, 166 people died of accidents, suicides, and homicides, with no other contributing
cause of death other than the direct result of suffocation,
compression, hanging or strangulation. As in the drowning
group, these cases were restricted to (a) premorbidly healthy
males aged 18 through 65, (b) white or black ethnicity, and
(c) recovery of the body within 6 h of death. In addition,
death had to occur within 4 h if they did not die at the scene.
In an attempt to obtain a similar proportion of cases across
whites and blacks, the number of years for which cases were
obtained varied across asphyxiation and trauma groups.
Asphyxiation cases were obtained from 1990 to 1999 for
whites, but also included 1989 for blacks. Trauma cases were
also obtained from 1990 to 1999 for whites, but only for the
years from 1990 to 1995 for blacks.
2.2. Procedures
The following information was abstracted from autopsies
performed by forensic pathologists at the Office of the Chief
Medical Examiner of the State of Maryland: (a) height, (b)
body weight, (c) combined lung weight, (d) liver weight, (e)
combined kidney weight, (f) heart weight, (g) brain weight,
(h) spleen weight, and (i) evidence of trauma and/or premorbid disease involving the organs of interest. With one
exception, all organs were weighed prior to dissection. To
prevent blood clots in the heart from artificially increasing its
weight, any clots were released before weighing by opening
each chamber. The age, gender, and race of the victim, as
well as the time delay between death and recovery of the
body, was obtained from investigations filed at the medical
examiners office.

2. Methods
2.3. Design and analysis
2.1. Case definition
Of the 1605 drownings that occurred in Maryland from
1982 to 1999, 217 met the criteria listed below for inclusion
in the present study. First, only those autopsied drowning
victims that were found in either (a) outdoor, inland bodies
of water (91% of all cases), or (b) the brackish waters of the
Chesapeake Bay and its estuaries, were included in the study.
Inclusion also depended upon (a) male gender, (b) an age
between 18 and 65, (c) white or black ethnicity, (d) no
contributing causes of death such as trauma or epilepsy, and
(e) no evidence of disease. Finally, only those cases that were
known to be recovered from the water within 6 h of the
drowning incident were included in the study.
Two additional experimental groups were required for the
present study: a trauma group and an asphyxiation group.

The analyses for the present study included two main

components with organ weight as the dependent measure.
First, an analysis of covariance (ANCOVA) was conducted in
order to compare the drowning and asphyxiation groups for
each organ type. Height, weight, race and age were the
covariates in these analyses. Second, the same ANCOVA
was conducted in order to compare asphyxiation and trauma
groups. To ensure that an effect of asphyxiation is not
confounded by a phenomenon in which uninjured organs
in the trauma cases are indirectly affected by trauma to nearby
organs, the head-only trauma group was used to examine the
effects of asphyxiation for the abdominal organs, whereas the
body-only trauma group was used to examine these effects for
the brain. The only exception to this would occur when, for
any given organ type, no statistically significant difference is

192

J.A. Hadley, D.R. Fowler / Forensic Science International 133 (2003) 190196

and organ weight in fact vary independently, the organ

weight/body weight ratio used by Ito et al. [4] and Haffner
et al. [5] would create a bias for (a) lower ratios in any groups
that have greater mean body weight and (b) higher ratios in
any groups that have lower mean body weight. In the present
use of an analysis of covariance, a lack of any association
between organ weights and any covariate would result in
adjusted means that are simply equivalent to the observed
means.

found between the body-only and head-only groups, in which

case these two trauma subgroups would be combined for the
purpose of the above analyses.
The present use of an analysis of covariance to control for
the potentially confounding effects of height, weight, race
and age is a departure from previous studies. Past studies that
have compared the organ weights of drowning to other
causes of death either did not attempt to control for these
potential confounders [2,3] or employed an organ weight/
body weight ratio as the only dependent measure [4,5].
Although this ratio measure controlled for any independent
effects of body weight on organ weight, the present use of an
analysis of covariance was also able to protect against any
confounding effects of height, race and age on the betweensubject differences in mean organ weight. By using the
techniques of linear regression, adjusted means are generated from a prediction of what the mean organ weights would
have been if the means associated with these covariates
were actually equal across experimental groups, resulting in
a test of organ weight effects largely unconfounded by these
influences.
Interestingly, however, an autopsy study of a large Japanese sample suggested that an attempt to control for these
covariates is actually unnecessary. For those 20 years of age
and older, no correlation was found between either body
weight, height or body surface area, and the respective
weights of the brain, heart, lung, kidney, spleen, pancreas,
thymus, thyroid gland or adrenal gland [7]. If body weight

3. Results
The mean weights across experimental group and organ
type, as well as the sample size, mean age, mean height,
mean weight, and the white/black ratio are presented in
Table 1. In addition, the distribution of categorized organ
weights for each organ type and experimental group are
presented in Figs. 1 and 2.
In the ANCOVA intended to uncover effects of asphyxiation, the asphyxiation group showed significantly higher
mean weights than the trauma group for every organ except
the brain and heart (Table 2). The asphyxiation group had
significantly higher organ weights than the trauma group for
the lungs [F(1, 352) 23.9, P < 0:001], liver [F(1, 468)
31.2, P < 0:001], kidneys [F(1, 506) 30.6, P < 0:001] and
spleen [F(1, 508) 33.5, P < 0:001]. Because the comparisons of the head-only and body-only trauma groups

Table 1
Mean organ weight (and standard deviation), sample size (after exclusion criteria applied to individual organs), mean age, mean height (cm),
mean weight (kg), and white/black ratio for each organ type and experimental group
Lungs

Liver

Brain

Heart

Kidneys

Spleen

Trauma
Mean weight
Sample size
Mean age
Mean height
Mean weight
White/black ratio

1039 (354)
201
32.6
179.0
80.0
1.22

1683 (403)
310
31.9
178.3
81.4
1.13

1429 (133)
250
31.7
179.2
81.7
1.08

385 (69)
278
31.3
178.5
80.9
1.00

309 (64)
348
31.8
178.2
80.3
1.04

165 (79)
353
31.7
178.5
81.2
1.08

Asphyxiation
Mean weight
Sample size
Mean age
Mean height
Mean weight
White/black ratio

1226 (372)
152
33.1
177.9
81.3
1.70

1879 (434)
159
33.8
178.1
81.9
1.70

1453 (126)
160
33.9
178.0
80.7
1.70

397 (138)
147
32.9
178.0
81.8
1.63

345 (72)
160
33.7
178.1
81.6
1.70

210 (84)
157
33.8
178.1
81.9
1.78

Drowning
Mean weight
Sample size
Mean age
Mean height
Mean weight
White/black ratio

1576 (467)
216
33.2
176.8
81.1
1.56

1893 (467)
204
32.3
176.7
80.3
1.38

1430 (130)
195
33.1
176.9
80.4
1.44

400 (98)
178
31.6
176.7
79.7
1.38

357 (76)
203
33.1
176.7
80.8
1.50

200 (101)
202
33.2
176.9
80.9
1.50

J.A. Hadley, D.R. Fowler / Forensic Science International 133 (2003) 190196

193

Fig. 1. Distribution (%) of cases across categorized groups of organ weights at each death type for the lungs, liver, and brain.

revealed none that approached significance for any organ

type, these two groups were combined (in the above analyses) for all organs except the brain (which was necessarily
restricted to the body-only group). If abdominal trauma had
any indirect effects on uninjured organs in the abdomen, it
did not impact on their weight.
In the ANCOVA intended to show effects of drowning,
only the lungs and kidneys were significantly heavier in the
drowning group than in the asphyxiation group [F(1, 358)
68.4, P < 0:001; F(1, 354) 4.0, P 0:045, respectively]
(Table 2). Of the liver, brain and heart, only the liver showed

an effect of drowning that approached statistical significance

[F(1, 353) 2.6, P 0:11].
The covariates of age, race, height and weight were often
statistically significant, with either one or two resulting in a
P-value of less than 0.05 in each analysis. Only the heart and
lungs however had the same statistically significant covariates in each analysis: body weight for the heart, and, race and
height for the lungs. Because these covariates turned out
to be very well matched across experimental groups (see
Table 1), the adjusted and observed mean organ weights
were essentially equivalent.

194

J.A. Hadley, D.R. Fowler / Forensic Science International 133 (2003) 190196

Fig. 2. Distribution (%) of cases across categorized groups of organ weights at each death type for the heart, kidneys, and spleen.

4. Discussion
Most important to the present study were the results that
revealed a problem with previous research that did not
properly account for the effects of submersion, which produces an artifact of lower organ weights only in the drowning
group. As predicted, whereas the present study replicated

nearly all of Ito et al.s [4] effects of asphyxiation (compared

to trauma), the present effects of drowning (compared to
asphyxiation) were markedly different from their findings.
Only the lack of any effects for the liver and heart were
replicated in the present study. Of the discrepancies, only the
brain effects could not be explained by submersion time
effects in their study that resulted in lower organ weights

J.A. Hadley, D.R. Fowler / Forensic Science International 133 (2003) 190196

195

Table 2
Adjusted mean organ weight (and standard errors) which were generated with height, weight, race, and age as covariates in the assessment of
the effects of asphyxiation (A) and drowning (B)
Aa

Bb

Trauma group
Lungs
Liver
Brain
Heart
Kidneys
Spleen

1040
1690
1429
386
310
167

(25)
(18)
(8)
(5)
(3)
(4)

Asphyxiation group
1225
1867
1452
395
342
206

(29)
(26)
(10)
(7)
(5)
(6)

Change (%)
17.8
10.5
1.6
2.3
10.3
23.4

Asphyxiation group
1218
1861
1452
393
343
209

(34)
(27)
(9)
(9)
(5)*
(7)

Drowning group
1584
1921
1436
405
358
203

(29)
(24)
(8)
(8)
(5)**
(6)

Change (%)
30.0*
3.2
1.1
3.1
4.4*
2.9

Significance of asphyxiation effects relative to trauma group.

Significance of drowning effects relative to asphyxiation group.
*
P < 0:01.
**
P < 0:05.
b

only in the drowning group. That is, the present emergence

of a higher mean kidney and lung weight, and, the disappearance of a lighter mean spleen weight, is consistent with
the above explanation. Indeed, as illustrated in Appendix A,
the lungs, kidneys, and spleen have the most obvious drop in
organ weight that each began within 12 h after death. Finally,
additional analyses restricted to cases across the same range
of submersion times in Ito et al.s [4] study replicated at least
the direction of their findings.1 A 36 g (17.5%) reduction in
the mean spleen weight for the drowning group (relative to
the present cases using a 6 h submersion time cutoff) instead
resulted in a significantly lower spleen weight than in the
drowning group (211 and 168 g for asphyxiation and drowning, respectively). A corresponding 12 g (3%) drop in combined kidney weight caused the present drowning effect for
this organ to instead disappear (346 g for both asphyxiation
and drowning). Despite a corresponding 126 g (8%) drop in
mean lung weight, a statistically significant effect of drowning was actually maintained (1458 and 1579 g for asphyxiation and drowning, respectively). The weight of the liver,
brain and heart were not affected by submersion times as
long as 24 h, thus maintaining the absence of an effect of
drowning (compared to the asphyxiation group) found in the
present study (with the 6 h submersion time cutoff).
Indeed, Del Campo et al. [8] uses rats in comparison of
sudden death by drowning versus trauma and found the same
percent difference in spleen weight as the present study. It is
important to point out that, whereas the use of a trauma
control group to isolate effects of asphyxiation may be
questionable, evidence was also found to suggest that the
trauma group indeed provides a valid control for this purpose. With no significant differences for any organ type
between the head-only and body-only trauma subgroups
(described earlier), there were apparently little or no effects
1
This data set included 85 cases that met the present criteria for
inclusion but were only associated with bodies recovered from the
water between 8 and 24 h of death.

of trauma on the uninjured organs included in the study. If

systemic effects of trauma such as due to internal or external
bleeding had occurred (to otherwise uninjured organs), one
would have expected to observe reduced effects on abdominal organs when only brain trauma had occurred, that is,
compared to when other abdominal organ(s) had sustained
trauma. Contributing to this result was the fact that nearly all
trauma cases used in this study consisted of people who died
very quickly (usually gunshot victims), with little time to
sustain blood loss or shifts of fluids from one compartment to
another.
Perhaps the most significant finding in the present study
was the absence of any evidence for a small, anemic spleen
as a direct consequence of death by drowning. Not only was
there no spleen weight effect of drowning compared to
asphyxiation, but also, a 23% increase in spleen weight in
asphyxiation compared to trauma actually implicates an
enlarged spleen in drowning, presumably caused by the
asphyxial effects of central pooling associated with peripheral vascular contraction.
Whereas the large drop in mean spleen weight that
occurred soon after death confounds an understanding of
how drowning processes affect spleen weight, it also points
to the possibility that the autopsy finding of a small, anemic
spleen is caused by a postmortem process. With a 15% drop
in mean spleen weight within 6 h of death, and by 33%
within 12 h of death, this apparent fluid shift may be
sufficient to produce the small, dry spleen sometimes
observed at autopsy. In addition, the fact that not every
drowning victim has an abnormally low spleen weight is
consistent with the present hypothesis in which this autopsy
finding is dependent on a sufficiently long submersion time
(which varies across drowning victims). The observation
that mean spleen weight had dropped earlier than would be
expected from decompositional processes alone (within 6 h
of death), and earlier than any other organ, is also consistent
with this passive shift of fluids out of the spleen that occurs
very soon after death.

196

J.A. Hadley, D.R. Fowler / Forensic Science International 133 (2003) 190196

whether an organ weight is elevated for a given individual,

that is, without knowledge of premorbid values. Even an
edematous lung, except for possibly the most severe cases of
edema (with combined weight greater than 2700 g), cannot
distinguish between drowning and asphyxiation (e.g. as the
result of suffocation before being dumped in the water).
In conclusion, the present study has shown that previous
findings of a lighter spleen in drowning than asphyxiation
was an artifact of a drop in spleen weight prior to recovery of
the body, in addition to revealing a small but significant
effect on kidney weight that had been missed in previous
research (for the same reason). The result is a picture of
drowning that includes effects of asphyxiation with elevated
weights for the lungs, liver, kidneys and spleen, but with only
the lungs displaying any sizeable increases in weight beyond
that due to asphyxiation. An important role for the aspiration
of water (and resulting hypervolemia) is therefore implicated to explain how death occurs in freshwater drowning,
that is, by rapidly accelerating the occurrence of pulmonary
edema. Lastly, the common autopsy finding of a small,
anemic spleen in drowning, rather than caused by some
pathophysiological mechanism of death, is argued to be a
postmortem phenomenon.

Given the present lung and kidney findings in which mean

organ weights were greater in asphyxiation than trauma,
and in turn, greater in drowning than asphyxiation, edematous
autopsy findings for these organs can be viewed as the
combined effect of the specific effects of both asphyxiation
and drowning. On the other hand, the liver and spleen displayed an effect of asphyxiation without an effect of drowning. As a result, an autopsy finding of an enlarged liver or
spleen cannot be attributed to any specific effect of drowning.
Because the aspiration of water in drowning is the only
phenomenon that differentiates a drowning death from a
purely asphyxial death, the present study confirms a significant role for this phenomenon in causing death by
drowning (that can increase blood volume by as much as
50% in fresh water drownings [9]). Whereas edematous lung
tissue in drowning victims is commonly attributed to aspirated fluids prior to being drawn into the blood stream, the
present evidence suggests that this autopsy finding is also the
consequence of the back pressure on the lungs associated
with hypervolemia. Although death by drowning has been
described as in part due to cardiac failure resulting in the
changes in blood chemistry with haemodilution [1,9], it is
possible that a sudden and large increase in blood volume
may also play a more direct role in the explanation of how
death occurs in drowning, that is, by a more rapidly developing pulmonary edema.
The diagnostic value of these findings for forensic pathologists is less straightforward. Because the range of weights
for a given organ were similar across trauma, asphyxiation,
and drowning groups, there becomes no means to determine

Acknowledgements
This research was supported in part by a grant from the
National Institute of Alcohol Abuse and Alcoholism
(R29AA07700). We wish to thank Gordon Smith for his
review of an earlier version of this paper.

Appendix A
Unadjusted mean organ weights (and standard errors) in a sample of 361 non-winter drowning victims by organ type and
submersion time group (in hours)
01
Lungs
Liver
Brain
Heart
Kidneys
Spleen

1597
1909
1445
402
358
221

1.16
(45)
(45)
(17)
(11)
(8)
(11)

1594
1867
1425
398
349
188

(44)
(43)
(16)
(10)
(8)
(10)

6.112

12.124

24.196

>96

1464
1795
1417
414
347
149

1377
1937
1452
401
344
185

1128
1644
1365
366
333
156

960
1526
1305
365
329
153

References
[1] W.U. Spitz, Spitz and Fishers Medicolegal Investigation of
Death, Charles C Thomas, Springfield, IL, 1993, pp. 498515.
[2] B. Kringsholm, A. Fiskov, K. Kock, Autopsy cases of
drowning in Denmark 19871989, Forensic Sci. Int. 52
(1991) 8592.
[3] A.R. Copeland, An assessment of lung weights in drowning
cases, Am. J. Forensic Med. Pathol. 6 (1985) 301304.
[4] Y. Ito, R. Tsuda, M. Hara, The weight ratio of organs for 750
judicial autopsy cases, Igaku Kenkyu 59 (1989) 6062 (in
Japanese).

(80)
(78)
(31)
(19)
(15)
(19)

(59)
(60)
(22)
(14)
(10)
(13)

(50)
(49)
(20)
(12)
(9)
(11)

(55)
(53)
(20)
(12)
(9)
(12)

[5] H.T. Haffner, M. Graw, J. Erdelkamp, Spleen findings in

drowning, Forensic Sci. Int. 66 (1994) 95104.
[6] K. Tomita, On putrefaction and floatations of dead bodies
under water, Hiroshima J. Med. Sci. 24 (1975) 117152.
[7] Y. Ogiu, Y. Nakamura, I. Ijiri, K. Hiraiwa, T. Ogiu, A statistical
analysis of the internal organ weights of normal Japanese
people, Health Phys. 72 (1997) 368383.
[8] J.L. Del Campo, S.L. Del Campo, El bazo en la asfixia por
sumersion, Rev. Esp. Med. Leg. 2 (1975) 3943.
[9] W. Lawler, Bodies recovered from water: a personal approach
and consideration of difficulties, J. Clin. Pathol. 45 (1992)
654700.