Abstract
Background: Avalanches are a significant cause of winter recreational fatalities in mountain regions. The purpose of this study was to determine the relative contributions of trauma and asphyxia to avalanche deaths.
Methods: We reviewed all avalanche fatalities between 1984 and 2005 that had been investigated by the offices of the British Columbia Coroners Service and the Chief Medical Examiner of Alberta. In addition, we searched the database of the Canadian Avalanche Centre for fatal avalanche details. We calculated injury severity scores for all victims who underwent autopsy.
Results: There were 204 avalanche fatalities with mortality information over the 21-year study period. Of these, 117 victims underwent autopsy, and 87 underwent forensic external examination. Asphyxia caused 154 (75%) deaths. Trauma caused 48 (24%) deaths, with the rate of death from trauma ranging from 9% (4/44) for snowmobilers to 42% (5/12) for ice climbers. In addition, 13% (12/92) of the asphyxia victims who underwent autopsy had major trauma, defined as an injury severity score of greater than 15. Only 48% (23/48) of victims for whom trauma was the primary cause of death had been completely buried.
Interpretation: Asphyxia and severe trauma caused most avalanche fatalities in western Canada. The relative rates differed between snowmobilers and those engaged in other mountain activities. Our findings should guide recommendations for safety devices, safety measures and resuscitation.
Over the past 30 years (1978–2007), avalanches have been responsible for 329 deaths in Canada, with a recent rate of 14 deaths per year.1 Avalanches kill a disproportionate number of young people, with victims most commonly being in their 20s.2–5 Outdoor recreation accounts for 92% of avalanche deaths in Canada, and 88% of deaths occur in the provinces of British Columbia and Alberta.1
Mountain rescue agencies emphasize avalanche avoidance strategies but also make recommendations for safety devices and resuscitation techniques.6,7 One such safety device, an airbag designed to increase flotation and prevent burial, reduced the mortality rate in a European study.8 However, the results of that study may not be generalizable to Canada, as it used Swiss and Austrian data, which reflect the European preponderance of avalanche deaths from asphyxia and a reported rate of trauma death of 5.6%.9 In contrast, a National Research Council study of human and property loss in Canada reported that trauma was the cause of death in 27% of avalanche deaths in this country.10 Snowmobilers have become the fastest growing and probably the largest activity group exposed to avalanches in North America,11–13 yet this activity is not common in Europe. Similarly, so-called helicat skiers, who use helicopters or snowcats (vehicles with an enclosed cab and tracks for travelling on snow) to reach their starting point, are not represented in European statistics.
Our purpose was to determine the relative contributions of trauma and asphyxia to avalanche deaths in western Canada and to determine whether this varies between different activity groups.
Methods
Study population
In this retrospective observational case series, we investigated all snow avalanche deaths in British Columbia and Alberta between Apr. 1, 1984, and Apr. 5, 2005. This date range coincided with entries of avalanche data in the Canadian Avalanche Centre Avalanche Incident Database that were complete upon study inception. We included all victims who died from snow avalanches, as the primary event, in outdoor terrain.
The research protocol for this study was approved by the research ethics review committee of the College of Physicians and Surgeons of Alberta.
Data collection and abstraction
We identified records for avalanche deaths in the Canadian Avalanche Centre database, which is compiled from input fields on a standardized avalanche report form.14 One of the investigators (P.H.), who was blinded to cause of death and other findings as recorded in autopsy and external examination reports, extracted avalanche data from this database.
We also identified avalanche fatalities by searching the databases of the Chief Coroner of British Columbia and the Chief Medical Examiner of Alberta using the term “avalanche” as the underlying cause or otherwise related to the cause of death. A second study investigator (J.B.) extracted mortality data from these files using explicit criteria and a standardized format. He was blinded to the avalanche data associated with each record. Data extracted included all cause-of-death fields, results of investigations conducted at the scene, findings of autopsy and external examination, and results of toxicology testing. Staff of the British Columbia Coroners Service and the Office of the Chief Medical Examiner of Alberta then searched their respective databases for specific victims who were reported in the database of the Canadian Avalanche Centre but had not been found using the term “avalanche” in the initial searches. Files for each case were reconciled using the date and location of the incident and the name and date of birth of the person who died and were merged into a single database for analysis.
For all cases in which an autopsy was conducted, we quantified trauma severity by calculating injury severity scores.15,16 The injury severity score is an index of anatomic injury severity ranging from 1 to 75, with higher scores indicating more severe injury and scores over 15 defined as major trauma.16 This score is based on the Abbreviated Injury Scale,17 which scores each specific injury on an ordinal scale from 1 (minor injury) to 6 (maximal and currently untreatable). The injury severity score is calculated by dividing the body into 6 regions, identifying the 3 regions with the highest scores and summing the squared values of these 3 scores.15,16 The injury severity score was calculated by use of information from full autopsy reports and not reports limited to external examination.
The sole abstractor (J.B.) for mortality data received training in calculating injury severity scores from a full-time trauma data analyst. Before scoring the study autopsies, we assessed the reliability of the calculations by comparing scoring and calculation performance between the sole abstractor and the trauma data analyst. This was done using a random selection of 8 cases, each requiring assessments in 6 injury regions (48 scores total). The 2 abstractors were blinded to each other's results. Reliability, as determined by weighted kappa scores,18 was excellent (weighted kappa score 0.89, 95% confidence interval [CI] 0.81–0.97).
Statistical analysis
We developed the questions for our primary analysis and corresponding outcome measures by consensus before undertaking any aspect of the analysis. We limited our analysis to cases for which we obtained both mortality and avalanche data.
We first generated standard descriptive statistics. We then performed comparative analyses of the cause of death by examination technique (autopsy v. external exam), the cause of death by activity group and the degree of burial by cause of death using 2-tailed χ2 tests. We compared burial time and depth by cause of death using the Wilcoxon rank-sum test with continuity correction. We derived confidence intervals for proportions from exact binomial calculations.19
Results
We identified 224 potentially eligible avalanche deaths in the Canadian Avalanche Centre database and a total of 211 potentially eligible deaths in the 2 provincial coroner databases. We excluded asphyxia deaths from nonavalanche causes; deaths caused by hitting an object in the absence of an avalanche, by mountaineering falls and by ice avalanches; and deaths in lift-serviced ski areas, buildings and vehicles. After these exclusions, mortality data with corresponding avalanche data were available for 204 cases (Figure 1). In 117 cases (57%), an autopsy had been performed. In the remaining 87 cases, a forensic external examination without dissection had been conducted.
The immediate cause of death was asphyxia in 154 cases (75%), trauma in 48 cases (24%) and hypothermia in 2 cases (1%) (Table 1).
Snowmobilers represented 44 (22%) of the fatalities, whereas helicat skiers represented 43 (21%) (Table 2). Definitions of all activities are provided in Table 2.
Overall, 179 (88%) of the victims were male (Table 2, Appendix 1, available at www.cmaj.ca/cgi/content/full/180/5/507/DC2). The median age of the study population was 33 (interquartile range 26–43) years, and the largest single age group was 20–29 years (59/203 or 29%). Helicat skiers were generally older (median 43 years, interquartile range 30–48 years) and out-of-bounds skiers younger (median 20 years, interquartile range 17–30 years) (Table 2).
Trauma was the cause of death for 4 (9%; 95% CI 3%–22%) of the 44 snowmobilers, similar to the overall trauma death rate reported in European avalanche statistics.9 In contrast, 13 (30%; 95% CI 17%–46%) of the 43 deaths involving helicat skiers were caused by trauma, a proportion that differs substantially from European data.
The median injury severity score in trauma deaths was 30 (interquartile range 22–75) (Table 3). In asphyxia deaths, the injury severity score exceeded 15 in 12 (13%) of the 92 cases with an autopsy.
Chest trauma was the most common injury, representing 46% (11/24) of cases with single-system trauma. In these cases, the median Abbreviated Injury Scale score for the thoracic region was 4 (interquartile range 3.5–4), which is ranked as severe in the Abbreviated Injury Scale dictionary.17 Head injury represented 42% (10/24) of the cases with single-system trauma. In these cases, the median Abbreviated Injury Scale score for the head region was 3 (interquartile range 3–4), which is ranked as serious. Neck injury accounted for 8% (2/24) of the cases with single-system trauma. In these cases, the median Abbreviated Injury Scale score for the neck region was 5, which is ranked as critical. Abdominal injury occurred in 4% (1/24) of the cases of single-system trauma, and the Abbreviated Injury Scale score was 3.
In 32 (68%) of the 47 trauma deaths in which the object hit was recorded, the victims hit trees (Appendix 2, available at www.cmaj.ca/cgi/content/full/180/5/507/DC2).
Among the trauma deaths, 23 (48%) of the 48 victims were completely buried, whereas 142 (92%) of the 154 asphyxia deaths involved complete burial (p < 0.001). In these cases, the median duration of burial was 25 minutes (interquartile range 10–56 minutes) for trauma victims and 45 minutes (interquartile range 20–231 minutes) for asphyxia victims (p = 0.009). The depth of burial also differed significantly for trauma and asphyxia victims (median 90 cm [interquartile range 30–120 cm] v. 150 cm [interquartile range 100–200 cm]; p < 0.001).
The proportion of victims who wore a combination electronic transmitter–receiver (avalanche transceiver) ranged from 8% (95% CI 0%–36%) among the 13 ice climbers and 13% (95% CI 2%–40%) among the 15 out-of-bounds skiers for whom transceiver use was recorded to 100% (95% CI 92%–100%) for the 42 helicat skiers for whom transceiver use was recorded (Appendix 3, available at www.cmaj.ca/cgi/content/full/180/5/507/DC2). Data were insufficient to assess the use of helmets or other safety devices.
Cardiopulmonary resuscitation was performed on scene in 96 (51%; 95% CI 44%–59%) of 187 cases for which such data were available. Prehospital advanced life-support was provided in 40 (21%; 95% CI 15%–27%) of 192 cases. Cardiopulmonary resuscitation was not performed in 6 (55%) of the 11 cases in which single-system chest trauma was the cause of death.
Blood ethanol levels exceeded the legal limit of 80 mg/dL (17.4 mmol/L) in 3% (5/160) of cases in which alcohol testing was performed (Appendix 4, available at www.cmaj.ca/cgi/content/full/180/5/507/DC2). Cannabinoids were detected in 13% (7/55) of cases in which drug testing was performed (Appendix 5, available at www.cmaj.ca/cgi/content/full/180/5/507/DC2).
Interpretation
We found that asphyxia was the dominant cause of avalanche-related deaths and that trauma was relatively common as the immediate cause of death. Overall, trauma was the single immediate cause of death in 24% of cases. The proportion of deaths attributable to trauma in our study varied significantly with activity, ranging from 9% for snowmobilers to 42% for ice climbers.
Trauma was also a significant contributing factor in a substantial proportion of asphyxia deaths. We found that 13% (12/92) of asphyxia victims who underwent autopsy had major trauma, defined as an injury severity score greater than 15.16 If the entire population of 204 fatalities is considered, this results in 10% (20/204) of cases with combined trauma and asphyxia, over and above the 24% (48/204) of cases determined to be caused by trauma alone. Thus, in western Canada, during the study period, major trauma contributed to a total of 33% (68/204) of avalanche deaths overall and, for ice climbers, to more than 50% (7/13) of deaths. In our study, trees were identified as the most common objects hit in trauma deaths. This may be explained by the much greater access to forested ski terrain in Canada20 than in Europe. In a recent study from Utah, United States, trauma was the immediate cause of death in 5.4% of cases and contributed to death in an additional 8.9% of 56 avalanche fatalities.21 Variations between studies point to differences in geography, such as mountain topography and distances for rescue flights, as well as demographics, notably different activity types.9,21–23
Experts stress the importance of avalanche survival strategies such as trying to escape, discarding potential anchors such as skis and creating a breathing space around the chest and face.24,25 However, any trauma suffered while caught in an avalanche may impair survival strategies and increase the risk of burial and asphyxia.26 Focusing on a single “immediate cause of death,” as is typical in forensic reports and mortality databases, may obscure the combined effects of trauma, asphyxia and hypothermia in avalanche fatalities. This may be further clouded by the nonspecific and inconclusive postmortem findings in asphyxia27 and hypothermia.28 An avalanche victim who suffers any degree of trauma, even minor musculoskeletal trauma, may cease to struggle and may thus come to rest entrapped in the debris and die from a combination of trauma, asphyxia and hypothermia, rather than from 1 of these causes alone.
Young skiers and snowboarders were over-represented in our series. The Canadian Avalanche Centre should continue to focus on education of this high-risk group,29 and social science research is essential to determine effective education strategies.30
Various safety devices have been developed to reduce morbidity and mortality in an avalanche. Avalanche experts consider 1 such device, the transceiver, as essential.24 Only 8% of ice climbers and 13% of out-of-bounds skiers in our study were wearing transceivers.
Head trauma accounted for 10 (42%) of 24 cases of single-system trauma death in our study, which supports the recommendation that helmets be worn in avalanche terrain.20,26 Our finding of a far greater risk of trauma than reported elsewhere, coupled with a low proportion of cases involving complete burial, should prompt critical analysis of the risk spectrum of each activity group in western Canada's mountain terrain. The results of such an analysis should be used to guide adoption of other devices such as inflatable airbags8,22,31,32 and breathing implements.21,23,33
In our study, 51% of victims received cardiopulmonary resuscitation, and 21% received prehospital advanced life support. We were unable to assess whether these interventions were indicated. Also, our study did not capture information on cases in which resuscitation was successful. In 2001, a valuable resuscitation algorithm was developed.34 In the light of our findings of a significant incidence of trauma, we recommend that this algorithm be revised to take traumatic injury into account.
Limitations
Our study had limitations attributable to its retrospective design. We did not examine potential changes over the 21 years of our study that might have reflected trends in safety measures or in rescue and resuscitation techniques. There may have been differences in how the individual forensic investigators reported injuries, even though they received similar training and followed standard reporting formats.27 Another limitation is the nonspecific and inconclusive postmortem findings typical of asphyxia27 and hypothermia.28
Mortality studies such as ours benefit from the mandatory reporting and investigation of fatalities. However, we could not assess the overall health consequences of avalanche incidents, because we did not analyze survivor data. These data are currently subject to reporting losses, as survivors often fail to report their involvement in avalanches.5 Therefore, future research should involve prospective designs that capture important morbidity and mortality data for every avalanche-related injury or death.
Conclusion
Asphyxia and severe trauma cause most avalanche fatalities in western Canada. The relative rates differ between snowmobilers and those engaged in other mountain activities. Furthermore, these fatality patterns differ from European studies, which have reported trauma death rates of about 5%. The differences are likely due to variations in terrain preferences and mountain topography. Our findings should guide recommendations for safety devices, safety measures and resuscitation.
@@ See related commentary by Brugger, page 491
Footnotes
-
Funding: No external funding was received for this research.
Une version française de ce résumé est disponible à l'adresse www.cmaj.ca/cgi/content/full/180/5/507/DC1
This article has been peer reviewed.
Contributors: The study was conceived and designed by Jeff Boyd, Pascal Haegeli, Riyad Abu-Laban and Michael Shuster. John Butt contributed his expertise as a forensic pathologist in the conceptual design of the study and the analysis and interpretation of autopsy and external examination data. Jeff Boyd and Pascal Haegeli were responsible for data extraction and Pascal Haegeli and Riyad Abu-Laban for statistical analysis. Jeff Boyd initially drafted the article, and all of the authors made critical revisions and gave final approval of the manuscript. Jeff Boyd was responsible for overall supervision of the study.
Acknowledgements: We thank Terry Smith, Chief Coroner, and Tej Sidhu of the British Columbia Coroners Service, and Graeme Dowling, Chief Medical Examiner, and Kim Borden and Barbara Hinman of the Office of the Chief Medical Examiner, Alberta Justice, for providing access to mortality data; Rosemary Armour, Ministry of Health Services of BC, for providing additional data; Christi Findlay, Regional Trauma Services, Calgary, for abstraction training and reliability testing; and Clair Israelson and John Kelly, Canadian Avalanche Centre, for providing access to avalanche data and for data confirmation. We also thank the following individuals for data confirmation: Paul Langevin; Marc Ledwidge, Steve Blake and Lisa Paulson, Parks Canada; George Field, Kananaskis Country, Alberta; Walter Bruns, Jon R. Bezzola and Rob Rohn, Canadian Mountain Holidays Inc. for administrative support; and Bruce Jamieson and Chris Stethem.
Competing interests: None declared.