- Email: [email protected]
Hypothermia fatalities in a temperate climate: Sydney, Australia
Pathology (January 2008) 40(1), pp. 46–51
ANATOMICAL PATHOLOGY
Hypothermia fatalities in a temperate climate: Sydney, Australia CATHY LIM*
AND
JOHAN DUFLOU{
*Anatomical Pathology, Royal Prince Alfred Hospital, Camperdown, {University of New South Wales, Department of Pathology, University of Sydney—School of Medical Sciences, and Department of Forensic Medicine, Central Sydney Laboratory Service, Glebe, New South Wales, Australia
Summary Aims: Fatal hypothermia is well known to occur in cold climates, with previous case series reflecting this. However, hypothermia can also occur in temperate climates. This case series describes the features and circumstances surrounding hypothermia-related deaths in Sydney, Australia. Methods: The files of hypothermia-related deaths were reviewed at the Department of Forensic Medicine, Glebe between 1 January 2001 and 31 December 2005 via a search of electronic autopsy records. Results: Twenty-four cases of fatal hypothermia were found. Many of the deaths occurred in winter (46%). The mean age was 76 years (range 56–92), with a female predominance (63%). Risk factors for hypothermia were identified in 58%. The mean body mass index (BMI) was 22 (range 15–33). Nineteen cases (79%) were found indoors. Four decedents were found naked, four were dressed in minimal amounts of clothing, and paradoxical undressing was found in seven cases. Pathological findings included gastric erosions (79%), and patchy reddish brown discoloration over large joints (75%). The majority of cases had significant pre-existing natural disease processes. Out of 18 cases where toxicology was performed, alcohol was detected in four cases, while other psychotropic agents were present in four deaths. No illicit drugs were detected. Conclusion: This study shows that fatal hypothermia, a significant public health problem, is not limited to cold climates. Forensic pathologists in Australia need to be aware of this condition, and not dismiss death as due to natural disease processes. Key words: Forensic science, forensic pathology, hypothermia. Received 15 March, revised 13 May, accepted 16 May 2007
INTRODUCTION Hypothermia refers to lowering of the core body temperature to 358C.1 Death due to hypothermia is a well-known significant public health problem in cold climates, affecting predominantly homeless people, drug-dependent persons, and the elderly. Hypothermia is generally considered very rare in more temperate climates, and there is a general lack of awareness of the problem by both the medical fraternity and the general public. Several factors contribute to hypothermia including environmental surroundings
(temperature and amount of clothing), and age. In addition, conditions including mental illness, hypothyroidism, dementia, and cerebrovascular accident can contribute. Previous series have looked at the epidemiology and autopsy findings of hypothermia.2 These have been conducted in cold climates in London,2 and Hokkaido in northern Japan.3 One series has looked at a temperate climate, Alabama in North America.4 However, no such series has been carried out in Sydney, a geographic location generally viewed as having a temperate to hot climate (Fig. 1). Features pertaining to a case series of hypothermiarelated deaths in Sydney, Australia are described. The social circumstances, relevant death scenes, associated morbidities and pathological findings at autopsy are explored. This case series shows that hypothermia as a cause of death must be considered even in temperate climates. It also reveals that underlying risk factors need to be contemplated as being the primary cause of hypothermia, and that associated comorbidities can sometimes complicate the true cause of death.
MATERIALS AND METHODS The files of hypothermia-related deaths were reviewed at the Department of Forensic Medicine, Glebe. All hypothermia-related deaths that occurred during 1 January 2001 and 31 December 2005 were identified via a search of electronic autopsy records, using the keyword ‘hypothermia’. Only those cases where hypothermia was the primary cause of death were included in this series. Those cases where hypothermia was a secondary cause of death or suggested as a possible cause of death were excluded. Consecutive cases were analysed and the following data items collected: demographic data (age, sex); risk factors for hypothermia (relevant medical history including a known psychiatric illness, alcohol abuse or medications); the circumstances in which the deceased was found (amount of clothing, place where found, place of death, season); evidence of injuries; and pathological findings including relevant toxicological examination and neuropathological examination. Any relevant histology or photographs pertaining to each case were reviewed. Fatal hypothermia had been either diagnosed clinically (a suggestive scene of death scenario, or a temperature measurement at hospital if still alive) or by suggestive autopsy findings (patchy reddish brown discoloration over large joints, acute gastric erosions or Wischnewski ulcers, and haemorrhagic pancreatitis). In Australia, the seasons are defined as December, January, and February (Summer); March, April, and May (Autumn); June, July, and August (Winter); and September, October, and November (Spring). Data were presented as frequency, percentages, and ranges. Analysis was conducted using SPSS 11.0 for Windows (SPSS, USA).
Print ISSN 0031-3025/Online ISSN 1465-3931 # 2008 Royal College of Pathologists of Australasia DOI: 10.1080/00313020701716466
FATAL HYPOTHERMIA IN SYDNEY
47
FIG. 1 Monthly climatic temperature for Sydney, 1971–2000. Source: Australian Bureau of Meteorology.5
RESULTS Ninety-five cases were identified using a search of electronic autopsy records, of which 24 cases of fatal hypothermia were included in the study (Table 1). The mean age was 76 years (range 56–92), with a female predominance (63%). Risk factors for hypothermia were identified in 58%, and these included mental illness (schizophrenia, bipolar disorder), alcoholism, and developmental delay (Table 2). All but two decedents lived alone. The mean body mass index (BMI) was 22 (range 15–33), SD 4.7. Nine decedents were underweight (body mass index 520 kg/m2). In this study, the diagnosis of fatal hypothermia was made by a temperature measurement at hospital (17%), suggestive autopsy findings (33%), a suggestive history (4%), or a combination of autopsy findings and a suggestive history (46%). The majority of the deaths occurred in winter (46%) as expected, while equal numbers occurred in the seasons on either side of winter—spring (25%) and autumn (25%). Despite generally hot summer temperatures, one death occurred during summer. This person had become lost in bushland and was found in a state of extreme dehydration. Nineteen cases (79%) were found in a building: a house, apartment, or other premises. Of these, 46% were dead at the scene, while the remainder died either on the way to hospital or in hospital. In the remaining five cases, the decedent was found outside. Within this group, three were dead at the scene, while two died subsequently in hospital. Four decedents were found naked (17%), and four (17%) were dressed in minimal amounts of clothing. The decedent was adequately dressed in three cases. There was evidence of paradoxical undressing in seven cases (29%). In the remaining six cases, the presence or absence of clothing was not given.
TABLE 1 Epidemiological, seasonal, and autopsy findings of accidental hypothermia, n ¼ 24 Findings
Frequency (%)
Sex Female Male
15 (63) 9 (37)
Risk factors No Yes
10 (42) 14 (58)
Clothing None Minimal Not specified Not given Partial undress
4 4 3 6 7
(17) (17) (12) (25) (29)
Place of death Home Hospital Outside
11 (46) 10 (42) 3 (12)
Where found Inside Outside
19 (79) 5 (21)
Season Summer Autumn Winter Spring
1 6 11 6
Evidence of injuries No Yes
6 (25) 18 (75)
Autopsy findings suggestive of hypothermia None Gastric erosions Pancreatic lesions
4 (17) 19 (79) 1 (4)
Diagnosis of hypothermia Temperature measurement (at hospital) Autopsy findings suggestive of History suggestive of Both autopsy findings and history suggest of
4 8 1 11
(4) (25) (46) (25)
(17) (33) (4) (46)
48
LIM AND DUFLOU
TABLE 2 No.
Pathology (2008), 40(1), January
Risk factors and other findings in accidental hypothermia, n ¼ 24
Age/sex Risk factors
1 2 3 4 5 6 7
56/F 75/M 92/F 73/F 74/F 73/M 60/F
Mental illness Alcoholism – – Alcoholism, dementia Alcoholism Alcoholism
8 9 10 11
61/F 71/M 72/M 89/F
Developmental delay – Alcoholism –
12 13 14 15 16 17 18 19 20
90/F 87/M 57/F 85/F 58/M 88/F 69/M 69/M 87/F
– ? Emotional—wife passed away recently Alcoholism ? History of confusion – Dementia Mental illness – –
21 22 23 24
89/F 73/M 85/F 86/F
– Alcoholism Dementia –
Other findings – IHD, ischaemic bowel, pulmonary oedema, splenic infarct, adrenal tumour Valvular heart disease, pulmonary congestion Metastatic bowel cancer Pneumonia, valve abnormalities, CVA IHD, CAL, cellulitis Renal cell carcinoma, cirrhosis, fracture right humerus post-fall, endometrial polyp, uterine leiomyomas Aortic stenosis with cardiac hypertrophy IHD, cirrhosis, BPH Pulmonary oedema, liver haemangiomas, BPH IHD, CAL, metastatic breast carcinoma, angiomyolipoma kidney, nephrosclerosis, cystic lesions in ovary IHD, pneumonia, acute duodenal ulcer, uterine leiomyoma IHD, CAL, acute renal failure, renal cell carcinoma IHD, rib fractures, bronchitis IHD, acute pulmonary oedema Hypertension, hepatic steatosis IHD, mitral stenosis, uterine leiomyoma, benign oesophageal lesion IHD, glioblastoma multiforme in brain Pericarditis IHD, calcific aortic valve, uterine polyp, pulmonary haemorrhage, cardiomegaly, rhabdomyolysis Hypertensive heart disease, CAL, nephrosclerosis IHD, pneumonia, rib fractures Chronic renal failure, Alzheimer’s disease (as suggested by neuropathology) Pneumonia, renal failure
BPH, benign prostatic hyperplasia; CAL, chronic airways limitation; CVA, cerebrovascular accident; IHD, ischaemic heart disease.
Pathological findings in fatal cases of hypothermia are generally considered non-specific. Gastric erosions were found in 79%. There was one case with acute pancreatitis, and a single case with rhabdomyolysis. The characteristic cutaneous lesions in hypothermia, reddening, and abrasion over the large joints of the limbs were present in 16 (75%) cases (Fig. 2). In four cases (17%), there were no autopsy findings to suggest hypothermia, although either hospital or environmental features strongly supported the diagnosis. Other significant autopsy findings included atherosclerotic cardiovascular disease (50%), pneumonia (17%), pulmonary congestion, and oedema (17%), chronic airways limitation (13%), and single cases of glioblastoma multiforme and metastatic bowel cancer. With the exception of two cases where only a raised blood alcohol was detected, all cases had significant pre-existing natural disease processes. Toxicology was performed in 18 (75%) cases—alcohol was detected in four (17%) cases, and other psychotropic agents were present in four (17%) deaths. No illicit drugs were detected. Neuropathological examinations were conducted in six (25%) cases—two incidental meningiomas were found, one glioblastoma multiforme, and three cases of Alzheimer’s disease.
FIG. 2
Reddish brown discolouration seen over both knees.
DISCUSSION The diagnosis of hypothermia-related death depends on several factors.6 Often, previous medical history, examination of the scene of death, and autopsy findings all play a part. Although Sydney is viewed as having a temperate to hot climate, hypothermia-related deaths are still seen. Hypothermia-related deaths have not yet been
reviewed in Sydney. This retrospective series looked at the hypothermia-related deaths at the Department of Forensic Medicine, Glebe, Sydney. Often, it is difficult to determine whether hypothermia is a consequence or determining factor of death. Other significant autopsy findings are often found (such as
FATAL HYPOTHERMIA IN SYDNEY
ischaemic heart disease, pneumonia, pulmonary congestion and oedema, and chronic airways limitation found in this review), which could in themselves cause death at any time. For example, one of the cases in this series had both glioblastoma multiforme and significant ischaemic heart disease, either of which could account for death. As most of the population in this series were elderly, it is not surprising to find coincidental ischaemic heart disease, pulmonary oedema, or chronic airways limitation. Moreover, pulmonary congestion is a non-specific finding that occurs postmortem. Although pulmonary oedema is said to occur more frequently in hypothermia, it is not diagnostic, as it commonly occurs in cardiac-failure deaths.1 However, in the cases included in the study, based on the clinical information and autopsy findings, and taking into account comorbidities, hypothermia was proposed to be the most likely cause of death. Pneumonia was seen in association with hypothermia. Whether this preceded the onset of hypothermia or occurred after is difficult to determine. However, we postulate that pneumonia occurred after the onset of hypothermia. This is because in hypothermia, muscle function deteriorates, and exhaustion and immobilisation take over.6 This, coupled with the occurrence of hypothermia in mainly elderly victims, sets up an environment conducive to the development of pneumonia. In this study, two of the cases with associated pneumonia occurred following injury, which would further contribute to immobilisation. In one of these cases, acute pneumonia and hypothermia followed a chest injury; and in the other case, hypothermia, pneumonia, and renal failure followed a fracture of the left hip. Life-threatening hypothermia does not occur at any one particular temperature, as other factors such as wind movement, clothing, dampness of the environment, and vulnerability of the individual all play a role.1 However, recovery from hypothermia is rare if the rectal temperature is less than 268C.1 In two cases from the study, where core body temperatures were measured as 2886 and 278C, both victims were found outside and, despite admission to hospital with re-warming and resuscitation, did not survive. This indicates that despite re-warming and supportive care, hypothermia past a certain point is irreversible and fatal. Several factors may account for the minimal clothing found on several of the decedents. One of the cases in this series may have been a result of dementia. This occurred in an 88-year-old lady, who was found lying naked on the back step of her house, with her head towards the top of the steps and her feet towards the bottom of the steps. It was believed that she was going to take a shower, and had gone outside for some unknown reason, and slipped on the stairs. This was supported by evidence of associated injury. Other cases raised the possibility of ‘paradoxical undressing’. This occurs due to a paradoxical feeling of warmth due to vasodilation, causing the victim to undress.7 One such case occurred in a 57-year-old woman who was found dead on a mattress on the dining room floor. She was found in an unusual position with her genitalia exposed. Although this position is suggestive of a criminal act, one must be aware of this phenomenon of ‘paradoxical undressing’, and not immediately assume that a death circumstance of partial undress may be due to assault.
49
The reddish brown patchy discoloration seen over large joints is often assumed to be a cause of injury and, at the scene of death, may raise the possibility of assault. The underlying pathogenesis is unknown, but an awareness of its association with hypothermia should be recognised.
FIG. 3 (A) Wichnewski ulcers as seen macroscopically from the exterior of the stomach. (B) Opened stomach showing multiple jet-black lesions. (C) Closeup view of the jet-black lesions.
50
LIM AND DUFLOU
Pathology (2008), 40(1), January
FIG. 4 (A,B) Coarse pigment scattered amongst regenerating gastric glands (H&E, A: 640; B: 6400).
Injuries can also be found in hypothermia-related deaths, and their occurrence could be a result of several factors. Those hypothermia-related deaths that occurred outside may have been preceded by a period of wandering outdoors where injuries may have easily occurred. The ‘hide and die’ syndrome could also be suggested by injuries. This syndrome involves undressing and hiding in small places such as pantries or wardrobes1, with furniture and items pulled over them. The scene of death in the home may initially appear alarming, with upturned furniture suggesting the possibility of a break and entry.8 It is thought to be due to a disregulation in the function of the hypothalamus to maintain thermoregulation, which occurs as the core body temperature drops.8,9 The paucity of pancreatic lesions in the autopsy findings of hypothermia was surprising. Although it has been said to be present in 20–30% of hypothermia-related deaths,10 this study found only one case. Acute pancreatitis or haemorrhage into the gland can occur, with fat necrosis seen macroscopically in the former. However, haemorrhage into the pancreas can also occur with autolysis, so its appearance should be assessed with caution.1 The mechanism is uncertain, and several theories have been proposed, including acute pancreatitis due to shock to the microcirculation from hypothermia, or secondary to alcohol abuse which is a risk factor for hypothermia.10 The history of alcoholism in 29% of cases suggests that this was not contributory in this study. Moreover, gallstones were not present in any of these cases, which excludes a coincidental gallstone pancreatitis with the occurrence of hypothermia. Acute gastric erosions (Wischnewski’s ulcers) were present in the majority (79%), which compares well with their quoted incidence of 80% in the literature.2,11 Although not pathognomonic, they are highly suggestive of hypothermia. Macroscopically, they are seen as multiple well-circumscribed ‘leopard-skin’ ulcers (Fig. 3A–C). The cause of the microscopic appearance has been debated in previous literature. An earlier study which looked at 14 cases of fatal hypothermia did not find any histopathological evidence of erosions or ulcerations in the gastric mucosa.12
Instead, haemorrhages, infarcts, and dilated mucosal capillaries were found.12 The hypothesis was that these gastric lesions were caused by cooling of the body leading to haemorrhages in the gastric mucosa, subsequent autolysis, and destruction of erythrocytes with release of haemoglobin.12 Our cases showed quite the opposite. Histopathological examination of these jet-black mucosal lesions showed erosions and ulcerations, with the accumulation of dark brown-black material at the surface of the gastric glands (Fig. 4A,B). This pigmented material was not reactive with a Perls’ stain. This study highlights a significant public health problem which is not limited to cold climates. There is little awareness of the hazards of hypothermia in the elderly in Australia and probably other temperate parts of the world. Life-threatening hypothermia does not occur at any one particular temperature, as other factors such as wind movement, clothing, dampness of the environment, and vulnerability of the individual all play a role. Forensic pathologists and other death investigators who have the good fortune of living in warm climates need to remain vigilant of this condition, and not misattribute the death to natural disease processes such as atherosclerotic cardiovascular disease or pneumonia. Address for correspondence: Dr J. Duflou, Department of Forensic Medicine, Central Sydney Laboratory Service, PO Box 90, Glebe, NSW 2037, Australia. E-mail: dufl[email protected]
References 1. Saukko P, Knight B. Knight’s Forensic Pathology. 3rd ed. London: Arnold Publication, 2004. 2. Mant AK. Autopsy diagnosis of accidental hypothermia. J Forensic Med 1969; 16: 126–9. 3. Mizukami H, Shimizu K, Shiono H, Uezono T, Sasaki M. Forensic diagnosis of death from cold. Leg Med 1999; 1: 204–9. 4. Taylor AJ, McGwin G Jr, Davis GG, Brissie RM, Holley TD, Rue LW 3rd. Hypothermia deaths in Jefferson County, Alabama. Inj Prev 2001; 7: 141–5. 5. Australian Bureau of Meteorology. Available online at: http://www. bom.gov.au (accessed 13 May 2007). 6. Lifschultz BD, Donoghue ER. Forensic pathology of heat-and coldrelated injuries. Clin Lab Med 1998; 18: 77–90.
FATAL HYPOTHERMIA IN SYDNEY
7. Wedin B, Vanggaard L, Hirvonen J. ‘Paradoxical undressing’ in fatal hypothermia. J Forensic Sci 1979; 24: 543–53. 8. Cala AD, Lawrence CH. Not murder most foul. Med J Aust 2001; 175: 621–2. 9. Ward ME, Cowley AR. Hypothermia: a natural cause of death. Am J Forensic Med Pathol 1999; 20: 383–6. 10. Foulis AK. Morphological study of the relation between accidental hypothermia and acute pancreatitis. J Clin Pathol 1982; 35: 1244–8.
51
11. Takada M, Kusano I, Yamamoto H, Shiraishi T, Yatana R, Haba K. Wischnewski’s gastric lesions in accidental hypothermia. Am J Forensic Med Pathol 1991; 12: 300–5. 12. Tsokos M, Rothschild MA, Madea B, Rie M, Sperhake JP. Histological and immunological study of Wischnewsky spots in fatal hypothermia. Am J Forensic Med Pathol 2006; 27: 70–4.
ANATOMICAL PATHOLOGY
Hypothermia fatalities in a temperate climate: Sydney, Australia CATHY LIM*
AND
JOHAN DUFLOU{
*Anatomical Pathology, Royal Prince Alfred Hospital, Camperdown, {University of New South Wales, Department of Pathology, University of Sydney—School of Medical Sciences, and Department of Forensic Medicine, Central Sydney Laboratory Service, Glebe, New South Wales, Australia
Summary Aims: Fatal hypothermia is well known to occur in cold climates, with previous case series reflecting this. However, hypothermia can also occur in temperate climates. This case series describes the features and circumstances surrounding hypothermia-related deaths in Sydney, Australia. Methods: The files of hypothermia-related deaths were reviewed at the Department of Forensic Medicine, Glebe between 1 January 2001 and 31 December 2005 via a search of electronic autopsy records. Results: Twenty-four cases of fatal hypothermia were found. Many of the deaths occurred in winter (46%). The mean age was 76 years (range 56–92), with a female predominance (63%). Risk factors for hypothermia were identified in 58%. The mean body mass index (BMI) was 22 (range 15–33). Nineteen cases (79%) were found indoors. Four decedents were found naked, four were dressed in minimal amounts of clothing, and paradoxical undressing was found in seven cases. Pathological findings included gastric erosions (79%), and patchy reddish brown discoloration over large joints (75%). The majority of cases had significant pre-existing natural disease processes. Out of 18 cases where toxicology was performed, alcohol was detected in four cases, while other psychotropic agents were present in four deaths. No illicit drugs were detected. Conclusion: This study shows that fatal hypothermia, a significant public health problem, is not limited to cold climates. Forensic pathologists in Australia need to be aware of this condition, and not dismiss death as due to natural disease processes. Key words: Forensic science, forensic pathology, hypothermia. Received 15 March, revised 13 May, accepted 16 May 2007
INTRODUCTION Hypothermia refers to lowering of the core body temperature to 358C.1 Death due to hypothermia is a well-known significant public health problem in cold climates, affecting predominantly homeless people, drug-dependent persons, and the elderly. Hypothermia is generally considered very rare in more temperate climates, and there is a general lack of awareness of the problem by both the medical fraternity and the general public. Several factors contribute to hypothermia including environmental surroundings
(temperature and amount of clothing), and age. In addition, conditions including mental illness, hypothyroidism, dementia, and cerebrovascular accident can contribute. Previous series have looked at the epidemiology and autopsy findings of hypothermia.2 These have been conducted in cold climates in London,2 and Hokkaido in northern Japan.3 One series has looked at a temperate climate, Alabama in North America.4 However, no such series has been carried out in Sydney, a geographic location generally viewed as having a temperate to hot climate (Fig. 1). Features pertaining to a case series of hypothermiarelated deaths in Sydney, Australia are described. The social circumstances, relevant death scenes, associated morbidities and pathological findings at autopsy are explored. This case series shows that hypothermia as a cause of death must be considered even in temperate climates. It also reveals that underlying risk factors need to be contemplated as being the primary cause of hypothermia, and that associated comorbidities can sometimes complicate the true cause of death.
MATERIALS AND METHODS The files of hypothermia-related deaths were reviewed at the Department of Forensic Medicine, Glebe. All hypothermia-related deaths that occurred during 1 January 2001 and 31 December 2005 were identified via a search of electronic autopsy records, using the keyword ‘hypothermia’. Only those cases where hypothermia was the primary cause of death were included in this series. Those cases where hypothermia was a secondary cause of death or suggested as a possible cause of death were excluded. Consecutive cases were analysed and the following data items collected: demographic data (age, sex); risk factors for hypothermia (relevant medical history including a known psychiatric illness, alcohol abuse or medications); the circumstances in which the deceased was found (amount of clothing, place where found, place of death, season); evidence of injuries; and pathological findings including relevant toxicological examination and neuropathological examination. Any relevant histology or photographs pertaining to each case were reviewed. Fatal hypothermia had been either diagnosed clinically (a suggestive scene of death scenario, or a temperature measurement at hospital if still alive) or by suggestive autopsy findings (patchy reddish brown discoloration over large joints, acute gastric erosions or Wischnewski ulcers, and haemorrhagic pancreatitis). In Australia, the seasons are defined as December, January, and February (Summer); March, April, and May (Autumn); June, July, and August (Winter); and September, October, and November (Spring). Data were presented as frequency, percentages, and ranges. Analysis was conducted using SPSS 11.0 for Windows (SPSS, USA).
Print ISSN 0031-3025/Online ISSN 1465-3931 # 2008 Royal College of Pathologists of Australasia DOI: 10.1080/00313020701716466
FATAL HYPOTHERMIA IN SYDNEY
47
FIG. 1 Monthly climatic temperature for Sydney, 1971–2000. Source: Australian Bureau of Meteorology.5
RESULTS Ninety-five cases were identified using a search of electronic autopsy records, of which 24 cases of fatal hypothermia were included in the study (Table 1). The mean age was 76 years (range 56–92), with a female predominance (63%). Risk factors for hypothermia were identified in 58%, and these included mental illness (schizophrenia, bipolar disorder), alcoholism, and developmental delay (Table 2). All but two decedents lived alone. The mean body mass index (BMI) was 22 (range 15–33), SD 4.7. Nine decedents were underweight (body mass index 520 kg/m2). In this study, the diagnosis of fatal hypothermia was made by a temperature measurement at hospital (17%), suggestive autopsy findings (33%), a suggestive history (4%), or a combination of autopsy findings and a suggestive history (46%). The majority of the deaths occurred in winter (46%) as expected, while equal numbers occurred in the seasons on either side of winter—spring (25%) and autumn (25%). Despite generally hot summer temperatures, one death occurred during summer. This person had become lost in bushland and was found in a state of extreme dehydration. Nineteen cases (79%) were found in a building: a house, apartment, or other premises. Of these, 46% were dead at the scene, while the remainder died either on the way to hospital or in hospital. In the remaining five cases, the decedent was found outside. Within this group, three were dead at the scene, while two died subsequently in hospital. Four decedents were found naked (17%), and four (17%) were dressed in minimal amounts of clothing. The decedent was adequately dressed in three cases. There was evidence of paradoxical undressing in seven cases (29%). In the remaining six cases, the presence or absence of clothing was not given.
TABLE 1 Epidemiological, seasonal, and autopsy findings of accidental hypothermia, n ¼ 24 Findings
Frequency (%)
Sex Female Male
15 (63) 9 (37)
Risk factors No Yes
10 (42) 14 (58)
Clothing None Minimal Not specified Not given Partial undress
4 4 3 6 7
(17) (17) (12) (25) (29)
Place of death Home Hospital Outside
11 (46) 10 (42) 3 (12)
Where found Inside Outside
19 (79) 5 (21)
Season Summer Autumn Winter Spring
1 6 11 6
Evidence of injuries No Yes
6 (25) 18 (75)
Autopsy findings suggestive of hypothermia None Gastric erosions Pancreatic lesions
4 (17) 19 (79) 1 (4)
Diagnosis of hypothermia Temperature measurement (at hospital) Autopsy findings suggestive of History suggestive of Both autopsy findings and history suggest of
4 8 1 11
(4) (25) (46) (25)
(17) (33) (4) (46)
48
LIM AND DUFLOU
TABLE 2 No.
Pathology (2008), 40(1), January
Risk factors and other findings in accidental hypothermia, n ¼ 24
Age/sex Risk factors
1 2 3 4 5 6 7
56/F 75/M 92/F 73/F 74/F 73/M 60/F
Mental illness Alcoholism – – Alcoholism, dementia Alcoholism Alcoholism
8 9 10 11
61/F 71/M 72/M 89/F
Developmental delay – Alcoholism –
12 13 14 15 16 17 18 19 20
90/F 87/M 57/F 85/F 58/M 88/F 69/M 69/M 87/F
– ? Emotional—wife passed away recently Alcoholism ? History of confusion – Dementia Mental illness – –
21 22 23 24
89/F 73/M 85/F 86/F
– Alcoholism Dementia –
Other findings – IHD, ischaemic bowel, pulmonary oedema, splenic infarct, adrenal tumour Valvular heart disease, pulmonary congestion Metastatic bowel cancer Pneumonia, valve abnormalities, CVA IHD, CAL, cellulitis Renal cell carcinoma, cirrhosis, fracture right humerus post-fall, endometrial polyp, uterine leiomyomas Aortic stenosis with cardiac hypertrophy IHD, cirrhosis, BPH Pulmonary oedema, liver haemangiomas, BPH IHD, CAL, metastatic breast carcinoma, angiomyolipoma kidney, nephrosclerosis, cystic lesions in ovary IHD, pneumonia, acute duodenal ulcer, uterine leiomyoma IHD, CAL, acute renal failure, renal cell carcinoma IHD, rib fractures, bronchitis IHD, acute pulmonary oedema Hypertension, hepatic steatosis IHD, mitral stenosis, uterine leiomyoma, benign oesophageal lesion IHD, glioblastoma multiforme in brain Pericarditis IHD, calcific aortic valve, uterine polyp, pulmonary haemorrhage, cardiomegaly, rhabdomyolysis Hypertensive heart disease, CAL, nephrosclerosis IHD, pneumonia, rib fractures Chronic renal failure, Alzheimer’s disease (as suggested by neuropathology) Pneumonia, renal failure
BPH, benign prostatic hyperplasia; CAL, chronic airways limitation; CVA, cerebrovascular accident; IHD, ischaemic heart disease.
Pathological findings in fatal cases of hypothermia are generally considered non-specific. Gastric erosions were found in 79%. There was one case with acute pancreatitis, and a single case with rhabdomyolysis. The characteristic cutaneous lesions in hypothermia, reddening, and abrasion over the large joints of the limbs were present in 16 (75%) cases (Fig. 2). In four cases (17%), there were no autopsy findings to suggest hypothermia, although either hospital or environmental features strongly supported the diagnosis. Other significant autopsy findings included atherosclerotic cardiovascular disease (50%), pneumonia (17%), pulmonary congestion, and oedema (17%), chronic airways limitation (13%), and single cases of glioblastoma multiforme and metastatic bowel cancer. With the exception of two cases where only a raised blood alcohol was detected, all cases had significant pre-existing natural disease processes. Toxicology was performed in 18 (75%) cases—alcohol was detected in four (17%) cases, and other psychotropic agents were present in four (17%) deaths. No illicit drugs were detected. Neuropathological examinations were conducted in six (25%) cases—two incidental meningiomas were found, one glioblastoma multiforme, and three cases of Alzheimer’s disease.
FIG. 2
Reddish brown discolouration seen over both knees.
DISCUSSION The diagnosis of hypothermia-related death depends on several factors.6 Often, previous medical history, examination of the scene of death, and autopsy findings all play a part. Although Sydney is viewed as having a temperate to hot climate, hypothermia-related deaths are still seen. Hypothermia-related deaths have not yet been
reviewed in Sydney. This retrospective series looked at the hypothermia-related deaths at the Department of Forensic Medicine, Glebe, Sydney. Often, it is difficult to determine whether hypothermia is a consequence or determining factor of death. Other significant autopsy findings are often found (such as
FATAL HYPOTHERMIA IN SYDNEY
ischaemic heart disease, pneumonia, pulmonary congestion and oedema, and chronic airways limitation found in this review), which could in themselves cause death at any time. For example, one of the cases in this series had both glioblastoma multiforme and significant ischaemic heart disease, either of which could account for death. As most of the population in this series were elderly, it is not surprising to find coincidental ischaemic heart disease, pulmonary oedema, or chronic airways limitation. Moreover, pulmonary congestion is a non-specific finding that occurs postmortem. Although pulmonary oedema is said to occur more frequently in hypothermia, it is not diagnostic, as it commonly occurs in cardiac-failure deaths.1 However, in the cases included in the study, based on the clinical information and autopsy findings, and taking into account comorbidities, hypothermia was proposed to be the most likely cause of death. Pneumonia was seen in association with hypothermia. Whether this preceded the onset of hypothermia or occurred after is difficult to determine. However, we postulate that pneumonia occurred after the onset of hypothermia. This is because in hypothermia, muscle function deteriorates, and exhaustion and immobilisation take over.6 This, coupled with the occurrence of hypothermia in mainly elderly victims, sets up an environment conducive to the development of pneumonia. In this study, two of the cases with associated pneumonia occurred following injury, which would further contribute to immobilisation. In one of these cases, acute pneumonia and hypothermia followed a chest injury; and in the other case, hypothermia, pneumonia, and renal failure followed a fracture of the left hip. Life-threatening hypothermia does not occur at any one particular temperature, as other factors such as wind movement, clothing, dampness of the environment, and vulnerability of the individual all play a role.1 However, recovery from hypothermia is rare if the rectal temperature is less than 268C.1 In two cases from the study, where core body temperatures were measured as 2886 and 278C, both victims were found outside and, despite admission to hospital with re-warming and resuscitation, did not survive. This indicates that despite re-warming and supportive care, hypothermia past a certain point is irreversible and fatal. Several factors may account for the minimal clothing found on several of the decedents. One of the cases in this series may have been a result of dementia. This occurred in an 88-year-old lady, who was found lying naked on the back step of her house, with her head towards the top of the steps and her feet towards the bottom of the steps. It was believed that she was going to take a shower, and had gone outside for some unknown reason, and slipped on the stairs. This was supported by evidence of associated injury. Other cases raised the possibility of ‘paradoxical undressing’. This occurs due to a paradoxical feeling of warmth due to vasodilation, causing the victim to undress.7 One such case occurred in a 57-year-old woman who was found dead on a mattress on the dining room floor. She was found in an unusual position with her genitalia exposed. Although this position is suggestive of a criminal act, one must be aware of this phenomenon of ‘paradoxical undressing’, and not immediately assume that a death circumstance of partial undress may be due to assault.
49
The reddish brown patchy discoloration seen over large joints is often assumed to be a cause of injury and, at the scene of death, may raise the possibility of assault. The underlying pathogenesis is unknown, but an awareness of its association with hypothermia should be recognised.
FIG. 3 (A) Wichnewski ulcers as seen macroscopically from the exterior of the stomach. (B) Opened stomach showing multiple jet-black lesions. (C) Closeup view of the jet-black lesions.
50
LIM AND DUFLOU
Pathology (2008), 40(1), January
FIG. 4 (A,B) Coarse pigment scattered amongst regenerating gastric glands (H&E, A: 640; B: 6400).
Injuries can also be found in hypothermia-related deaths, and their occurrence could be a result of several factors. Those hypothermia-related deaths that occurred outside may have been preceded by a period of wandering outdoors where injuries may have easily occurred. The ‘hide and die’ syndrome could also be suggested by injuries. This syndrome involves undressing and hiding in small places such as pantries or wardrobes1, with furniture and items pulled over them. The scene of death in the home may initially appear alarming, with upturned furniture suggesting the possibility of a break and entry.8 It is thought to be due to a disregulation in the function of the hypothalamus to maintain thermoregulation, which occurs as the core body temperature drops.8,9 The paucity of pancreatic lesions in the autopsy findings of hypothermia was surprising. Although it has been said to be present in 20–30% of hypothermia-related deaths,10 this study found only one case. Acute pancreatitis or haemorrhage into the gland can occur, with fat necrosis seen macroscopically in the former. However, haemorrhage into the pancreas can also occur with autolysis, so its appearance should be assessed with caution.1 The mechanism is uncertain, and several theories have been proposed, including acute pancreatitis due to shock to the microcirculation from hypothermia, or secondary to alcohol abuse which is a risk factor for hypothermia.10 The history of alcoholism in 29% of cases suggests that this was not contributory in this study. Moreover, gallstones were not present in any of these cases, which excludes a coincidental gallstone pancreatitis with the occurrence of hypothermia. Acute gastric erosions (Wischnewski’s ulcers) were present in the majority (79%), which compares well with their quoted incidence of 80% in the literature.2,11 Although not pathognomonic, they are highly suggestive of hypothermia. Macroscopically, they are seen as multiple well-circumscribed ‘leopard-skin’ ulcers (Fig. 3A–C). The cause of the microscopic appearance has been debated in previous literature. An earlier study which looked at 14 cases of fatal hypothermia did not find any histopathological evidence of erosions or ulcerations in the gastric mucosa.12
Instead, haemorrhages, infarcts, and dilated mucosal capillaries were found.12 The hypothesis was that these gastric lesions were caused by cooling of the body leading to haemorrhages in the gastric mucosa, subsequent autolysis, and destruction of erythrocytes with release of haemoglobin.12 Our cases showed quite the opposite. Histopathological examination of these jet-black mucosal lesions showed erosions and ulcerations, with the accumulation of dark brown-black material at the surface of the gastric glands (Fig. 4A,B). This pigmented material was not reactive with a Perls’ stain. This study highlights a significant public health problem which is not limited to cold climates. There is little awareness of the hazards of hypothermia in the elderly in Australia and probably other temperate parts of the world. Life-threatening hypothermia does not occur at any one particular temperature, as other factors such as wind movement, clothing, dampness of the environment, and vulnerability of the individual all play a role. Forensic pathologists and other death investigators who have the good fortune of living in warm climates need to remain vigilant of this condition, and not misattribute the death to natural disease processes such as atherosclerotic cardiovascular disease or pneumonia. Address for correspondence: Dr J. Duflou, Department of Forensic Medicine, Central Sydney Laboratory Service, PO Box 90, Glebe, NSW 2037, Australia. E-mail: dufl[email protected]
References 1. Saukko P, Knight B. Knight’s Forensic Pathology. 3rd ed. London: Arnold Publication, 2004. 2. Mant AK. Autopsy diagnosis of accidental hypothermia. J Forensic Med 1969; 16: 126–9. 3. Mizukami H, Shimizu K, Shiono H, Uezono T, Sasaki M. Forensic diagnosis of death from cold. Leg Med 1999; 1: 204–9. 4. Taylor AJ, McGwin G Jr, Davis GG, Brissie RM, Holley TD, Rue LW 3rd. Hypothermia deaths in Jefferson County, Alabama. Inj Prev 2001; 7: 141–5. 5. Australian Bureau of Meteorology. Available online at: http://www. bom.gov.au (accessed 13 May 2007). 6. Lifschultz BD, Donoghue ER. Forensic pathology of heat-and coldrelated injuries. Clin Lab Med 1998; 18: 77–90.
FATAL HYPOTHERMIA IN SYDNEY
7. Wedin B, Vanggaard L, Hirvonen J. ‘Paradoxical undressing’ in fatal hypothermia. J Forensic Sci 1979; 24: 543–53. 8. Cala AD, Lawrence CH. Not murder most foul. Med J Aust 2001; 175: 621–2. 9. Ward ME, Cowley AR. Hypothermia: a natural cause of death. Am J Forensic Med Pathol 1999; 20: 383–6. 10. Foulis AK. Morphological study of the relation between accidental hypothermia and acute pancreatitis. J Clin Pathol 1982; 35: 1244–8.
51
11. Takada M, Kusano I, Yamamoto H, Shiraishi T, Yatana R, Haba K. Wischnewski’s gastric lesions in accidental hypothermia. Am J Forensic Med Pathol 1991; 12: 300–5. 12. Tsokos M, Rothschild MA, Madea B, Rie M, Sperhake JP. Histological and immunological study of Wischnewsky spots in fatal hypothermia. Am J Forensic Med Pathol 2006; 27: 70–4.