ORIGINAL ARTICLE
JOURNAL
OF
THE ROYAL
SOCIETY OF
MEDICINE
96
May 2003
Hypothermia and acute pancreatitis: myth or reality? Rh E Stiff MB BSc
G J Morris-Stiff MD FRCS
J Torkington MD FRCS
J R Soc Med 2003;96:228–229
SUMMARY Among the rarer causes of acute pancreatitis listed in surgical texts is hypothermia. To assess the evidence for cause and effect, we questioned selected consultants about their experience and examined the case-notes of patients admitted with hypothermia. The 31 consultants who returned our questionnaire (69% response rate; 317 consultant-years’ experience) could recall only 5 cases of pancreatitis associated with hypothermia, in 2 of which other aetiological factors were judged primary. In case-notes for 100 months of emergency admissions at a single hospital we identified 310 patients with hypothermia and 1153 with acute pancreatitis; none had the dual diagnosis. Of the hypothermic patients, none had abdominal pain typical of acute pancreatitis. In 43 serum amylase was measured because the patient was unable to give a full history and in 2 of these the enzyme was slightly raised; both had experienced a cerebrovascular accident, which is a known cause of hyperamylasaemia. Considered alongside the weak evidence from previous studies, these findings offer negligible support for the idea that hypothermia is a clinically relevant risk factor for acute pancreatitis.
INTRODUCTION
In a case of acute pancreatitis the most likely cause in the UK is gallstones or alcohol, but surgical texts list many other possibilities. Among those in the fine print is hypothermia. This observation might raise questions about current practice in pancreatic transplantation, whereby the organ is transported and stored in ice. To assess the evidence for a causal association between hypothermia and pancreatitis we questioned selected consultants and reviewed a sample of hospital case-notes. METHODS
A senior consultant in the specialties of surgery, pathology and accident and emergency medicine was identified in each of Wales’ fifteen hospitals with emergency admission facilities. A questionnaire was sent to these consultants, asking about their experience of managing acute pancreatitis in hypothermic patients. This inquired specifically about the total number of patients they had seen with the combined diagnosis. In addition, we looked at the inpatient case-notes of all individuals with the clinical coding diagnosis of hypothermia seen as emergency admissions at the University Hospital of Wales between April 1990 and Department of Surgery, University Hospital of Wales, Heath Park, Cardiff CF14 4XN, UK Correspondence to: Dr Rhianwen Stiff, 26 Canada Road, Heath, Cardiff CF14 3BW, UK
228
Volume
E-mail:
[email protected]
August 1998. Particular attention was paid to the presence or absence of abdominal pain before, during or after admission and to the finding of hyperamylasaemia (serum amylase 4100 IU/L) either at admission or subsequently. Other risk factors for acute pancreatitis (gallstones, excess alcohol intake) were sought. RESULTS
Postal survey
31 (69%) of 45 questionnaires were returned, reflecting total experience of about 317 consultant-years. Median duration of consultant appointment was 12 years for surgeons, 12.5 years for pathologists and 6.2 years for accident and emergency consultants. From this total clinical experience, only 5 cases of a possible association between hypothermia and pancreatitis were identified, and in 2 of these other aetiological factors were identified as primary. Analysis of medical records
Between April 1990 and August 1998, 210 patients were admitted with a diagnosis of hypothermia and 1153 with a diagnosis of acute pancreatitis; none had both. Of those with hypothermia, the mean age was 62.5 years (range 1– 97), male to female ratio 3:2. None of the hypothermic patients complained of abdominal pain at the time of admission. 43 had their serum amylase measured, the reason in all cases being absence of a clinical history due to a reduced level of consciousness at the time of admission.
JOURNAL
OF THE ROYAL SOCIETY OF
Amylase was above normal in 2, but in neither was it greater than 250 IU/L; both these patients had had cerebrovascular accidents and survived the hypothermic episode; neither, at any stage, had evidence of abdominal pain. DISCUSSION
Our consultant survey can be criticized for its reliance on memory and its low response rate. However, combined with the case-note review it offers little support for the notion that hypothermia causes pancreatitis. How, then, do our negative findings fit in with existing evidence? In 1969 Mant1 reported necropsies in 43 patients who had been hypothermic, and described pancreatic changes varying from occasional foci of fat necrosis to frank haemorrhage in 29 (67%). He suspected that the lesions were due to the hypothermia and seems to have dismissed the possible contribution of other factors: 3 had been sleeping in the open during cold weather in London; 7 had a history of mental illness; 3 had Alzheimer’s disease; 3 had severe chronic debilitating natural disease (unspecified); and 3 had become hypothermic as a result of criminal activity (a neglected infant, a road-accident victim left at the roadside and an elderly woman left bound and gagged in her home). In 1982, Foulis2 tried to clarify the relation with a morphological analysis of 8 cases. From the patterns of pancreatic necrosis he proposed three possible mechanisms: (1) microcirculatory shock of hypothermia causing pancreatic necrosis; (2) both hypothermia and pancreatitis secondary to alcohol abuse, and (3) severe pancreatitis the primary disease with hypothermia developing especially in the socially deprived. Of Foulis’ 8 patients, 2 had gallstones and 3 were chronic alcoholics; 5 had had unrecordable blood pressure on admission, prompting a hypothesis that hypotension rather than hypothermia might be a risk factor for pancreatitis. With so few patients it is difficult to draw conclusions. The only other published evidence comes from two isolated cases. Mahood and Evans3 described hypothermia
MEDICINE
Volume
96
May 2003
and pancreatitis in a man of 65: he was a known alcohol abuser; moreover, he developed disseminated intravascular coagulopathy before any clinical or biochemical evidence of pancreatitis. The second4 came from a series of 45 cases of fatal acute pancreatitis: the pancreatitis was said to have occurred ‘. . . possibly from hypothermia’ but this patient also had unexplained somnolence and the association was not straightforward Both hypothermic patients with hyperamylasaemia in our study had suffered cerebrovascular accidents, and intracranial bleeding is a known cause of raised serum amylase. Bouwman and colleagues found hyperamylasaemia in 6 of 10 patients with isolated head trauma and CT evidence of intracranial bleeding, and concluded that serum amylase is not a reliable marker for pancreatic trauma in a patient with multiple injuries.5 Probably the most telling evidence against a direct role of hypothermia in causing pancreatitis comes from pancreatic transplantation where a procured organ is submerged in ice for up to 24 hours and on rewarming is expected to function normally. Although slight hyperamylasaemia is not uncommon after transplantation, moderate or severe acute pancreatitis is rare. In view of the lack of positive evidence, and indeed a tendency towards negative evidence, we conclude that hypothermia is unlikely to be an important risk factor for the development of acute pancreatitis. REFERENCES
1 Mant K. Autopsy diagnosis of accidental hypothermia. J Forens Med 1969;16:126–30 2 Foulis A. Morphological study of the relation between accidental hypothermia and acute pancreatitis. J Clin Pathol 1982;35:1244–8 3 Mahood JM, Evans A. Accidental hypothermia, disseminated intravascular coagulation and pancreatitis. N Z Med J 1978;87:283–4 4 Lankisch PG, Schirren CA, Kunze E. Undetected fatal acute pancreatitis: why is the disease so frequently overlooked? Am J Gastroenterol 1991;86:322–6 5 Bouwman DL, Altshuler J, Weaver DW. Hyperamylasemia: a result of intracranial bleeding. Surgery 1983;94:318–23
229
