Significant hypothermia occurs when the body core temperature falls from its usual 37 °C to below 35 °C. The findings at autopsy can be quite subtle and there is a lack of accord on the necessary features required to make the diagnosis [1,2,3,4]. It is even less clear to what degree hypothermia may contribute to lethal outcomes in the presence of significant underlying organic illnesses. For example, the relationship between hypothermia and the metabolic derangements that occur in elderly individuals with frailty syndrome has only recently been explored [5]. The following discussion looks at the complexities of this diagnosis and points out the gaps in our knowledge which could result in under diagnosis.

Diagnostic issues with hypothermia were summarized quite succinctly and presciently by Casper in the nineteenth century. Over a century and a half ago he commented that “even in regard to the appearances found on dissection, there is not one which can with any certainty justify the assumption of death from cold” [6]. Hypothermia still carries a significant risk of death with >70% mortality with core temperature of 30 °C, increasing to 90% at 26 °C [7].

Environmental factors clearly play a significant role in deaths attributed to hypothermia with a common scenario being individuals who have been lost in areas with sub-zero temperatures. The finding of a non-traumatized body with no significant underlying disease and a high alcohol level in a snow bank is strong supportive evidence for death from hypothermia. Cases become more complex, however, where there has been an elderly individual found at their home address having collapsed from a cardiovascular, orthopedic, neurological or endocrine event [8]. Exacerbating factors include cold flooring, damp conditions, wet or inadequate clothing, low muscle mass, immobility, certain drugs, and some psychiatric conditions [7, 9]. Poor home insulation and heating are recently recognized factors that have produced the counter-intuitive finding of higher death rates from hypothermia in South Australia with its Mediterranean climate than in Sweden [10, 11]. This has also been reported in other temperate areas [12].

Individuals die of hypothermia because of failure of hypothalamic counter-regulatory mechanisms that initiate and maintain chemical thermogenesis and vasoconstriction [13, 14]. Death likely results from a combination of events such as ventricular fibrillation or asystole, initiated or exacerbated by hypoxia, myocardial ischemia, increased circulating catecholamines and electrolyte derangements [1, 9, 15].

The major problems for forensic pathologists are firstly how to accurately identify cases, and then how to make a credible diagnosis. As with cases of hyperthermia, the core temperature of a decedent at the time of death is not usually available, and so evaluation of potential cases at autopsy requires a clear understanding of the circumstances of the death (ideally with a recording of the ambient temperature) and the exclusion of other possibilities [3, 8, 16]. Issues occur with differences in diagnostic practices arising from the non-specific and subtle nature of findings at autopsy, and the possibility of inflicted injury due to the sometimes bizarre nature of behavior that may occur before death [17,18,19]. Biochemical markers are not currently routinely assessed in these cases [2, 20].

The peri-mortem behavior of individuals who have succumbed to hypothermia that may create concerns of a suspicious death include so-called paradoxical undressing where, most typically, an individual has sequentially removed clothes as he or she has staggered through snow. This has been called “hypothermic confusion” and results from terminal failure of thermoregulatory control [4, 8]. Socially isolated individuals may also exhibit features of the “hide and die” syndrome or “terminal burrowing” where they may be found hidden under a pile of newspapers or rubbish [17, 21]. This is most likely in hoarding disorders as a manifestation of Diogenes, Havisham, or so-called senile squalor, syndrome [22, 23].

At autopsy in those exposed to the cold there are may be a series of characteristic but not diagnostic findings that occur with variable frequency. Externally these features include pink discoloration of the skin over the extensor surfaces of large joints. On internal examination there may be superficial gastric erosions, or Wischnewski spots, vacuolization of renal tubular cells, acute pancreatic inflammation with fat necrosis, fatty change in cells of the heart, liver and kidneys, and skeletal muscle hemorrhage [1, 4].

Wischnewski spots in the gastric mucosa are considered the most reliable marker of hypothermia at post mortem [1, 3, 24,25,26]. There is, however, considerable variation in their reported incidence which may be related to the length of exposure to low temperatures, or to idiosyncratic general and gastric mucosal responses to cold stress [3, 27,28,29].

Spots may occur because of disturbances in microcirculation with ischemic-reperfusion injury and hemoglobin breakdown [3, 24, 27, 30]. The role of stress in their etiology has been debated, with a recent low-stress animal model failing to generate Wischnewski spots in anesthetized rats exposed to low temperatures [19].

Hypothermia is also related to a wide variety of disparate organic conditions that include diabetes mellitus, stroke, sepsis, hypothyroidism and cardiac, renal, and hepatic failure [11, 31,32,33]. These conditions may alter thermoregulatory responses leading to abnormal temperature control and regulation [31]. Determining the exact cause of death in cases with multiple medical conditions may be difficult as the potential role of hypothermia in the deaths may be obscured [12]. In addition, a number of psychiatric and neurodegenerative conditions such as bipolar disease, schizophrenia, Alzheimer’s disease, and dementia increase the risk for lethal hypothermia [12, 34].

In conclusion, it appears that although the phenomenon of lethal hypothermia is well recognized in forensic pathology, the diagnostic criteria and evaluation of cases remain problematic. The percentage of cases with Wischnewski spots exemplifies this situation with the frequency of these lesions varying from 40 to 91% of cases in different series [3, 4, 26]. While this may be a reflection of the pathogenesis and time course of these spots, it may also be a marker of diagnostic practices, with hypothermia being a more likely diagnosis in some centers only if the spots are present.

The conclusion that hypothermia is the cause of death often relies not only on the quality of the death scene examination by police investigators, but also on the documentation of key features such as the ambient temperature, and the adequacy of insulation and heating for indoor deaths. In addition, the detection and correct interpretation of subtle autopsy markers is sometimes not easy. Finally, the role of hypothermia in triggering natural deaths from organic illness is yet another area in forensic practice to be explored.

Thus, it would appear not unreasonable to suggest that some cases of hypothermia might not be being recognized in contemporary forensic facilities. This would, of course come as no surprise to Professor Casper, who also commented in 1862 in Berlin, that “physiology has not yet determined, and probably never will determine, how it happens…” [6].