Epidemiology Of Trauma Deaths Research Articles

Epidemiology of Trauma Deaths: Volume, Methodology, and Comparability—Reply to Letter

We are thankful for the insightful comments on our prospective population-based epidemiological study on traumatic deaths in our trauma system (John Hunter Hospital, New South Wales, Australia) [1]. Our study evaluated the first year of our trauma center/trauma system after the establishment of the new trauma service and concluded the challenge of an aging population; the potential relative increase of hemorrhage related deaths in high-energy trauma and warned about the growing problem of lowenergy fall related deaths. We are well aware of the Professor Soreide’s 9-year retrospective evaluation of 260 trauma autopsies from Norway [2]. We strongly agree with the concept that geographical location and different time points of assessment are, not surprisingly, leading to different results. These two reasons influenced our study design. First, we wanted to describe the epidemiology of traumatic deaths in a comprehensive (highand low-energy) prospective fashion, including all inhospital and prehospital trauma deaths, which has not been done in Australia before. Second, considering the rapid change in the delivery of trauma care (especially in a young trauma system), we insisted to a 12-month prospective study similar to the studies that we considered landmark papers [3–5]. We believe more frequent comparison of shorter periods (12 months) are potentially more meaningful than lumping many years of data together from relatively low volume centers. Our study described 103 trauma fatalities in 1 year, whereas the Stavanger University Hospital 9-year experience was 260 fatalities (29 trauma deaths per year). Their severely injured (ISS [15) is 80–85 patients/year, whereas the John Hunter Hospital admits alone 400–420 patients in the same category annually. We do not feel that the two trauma systems are highly comparable. However, we are convinced that the SUH standards of trauma care are very different today compared with 1996 when their 9 years retrospective study started. Hopefully, we will be able to document similar improvements in our trauma system. We reported both low-energy and high-energy traumatic deaths, because our trauma center and system is committed to address the often very distinct requirements of these two patient populations. This did not prevent us from publishing both high-energy (12.3/100,000/year) and low-energy (8.6/100,000/year) incidence separately to allow comparisons with other population based studies. Our study design, detailed in the Methods section, explicitly excluding asphyxia without injuries: ‘‘Deaths related to electrocution, drowning, hanging asphyxiation, strangling, and poisoning were excluded from the data analysis.’’ We also agree with the potential misinterpretation of temporal patterns of the epidemiology of trauma deaths in many communications. To avoid this bias our study design (time frames and categories of deaths) followed the previous studies referred to the trimodal death distribution. Again, our study was a baseline description of the epidemiology of traumatic deaths in our young trauma system. Some of the authors of this study have firsthand experience of the European, North American, and Australian trauma care, which might eliminate some of the uncertainties of comparing different trauma systems. We definitively agree with the letter that longitudinal Z. J. Balogh (&) J. A. Evans Department of Traumatology, John Hunter Hospital, University of Newcastle, Newcastle, Australia e-mail: Zsolt.Balogh@hnehealth.nsw.gov.au

Epidemiology of Trauma Deaths: Location, Location, Location!

To the Editor: I read with interest the recently published article by Evans et al. on the epidemiology of trauma deaths in Newcastle, Australia, comparing their results to ‘‘previous landmark studies’’ from San Francisco, San Diego, Denver, and Los Angeles in the USA [1]. I congratulate the authors for sharing their important data on fatal trauma from their region, and I very much enjoyed reading their work. The authors conclude that injury mechanisms, time frames, and causes in their study are different from those in the earlier, seminal reports from North America. I believe this should not be surprising, given the differences in the trauma system, trauma epidemiology, and the changes that have occurred with time since these reports were first published. Further, many of the results reported by Evans et al. [1] are much in line with a similar study published by our group in Norway [2]. The similarity in these two articles [1, 2] suggests that one should (if at all) compare ‘‘likes with likes’’ and recognize that different geographic locations have different prevalence, patterns, and predominating mechanisms of injury [3]. It may not be suitable to compare North American numbers, of which many are influenced by a totally different prevalence of penetrating trauma, to numbers reported from other locations, such as Scandinavia or Australia [1, 2]. Further, different geographic locations may yield inherent differences in delivery of trauma care and maturity grade of the system—could it be that geography, delivery, and maturity in the Australian trauma system is more comparable to a European context [4] than that of North America? In our study [3], we found an even lower incidence of trauma deaths (about 10 per 100,000/year), but we did not include burns and deaths from isolated hip fractures (‘‘low energy’’) or trauma deaths without associated anatomical injuries (i.e., traumatic asphyxia), which may in part explain the differences. Further, in our article, we demonstrated several ways in which the temporal distribution of trauma deaths can be displayed, and, thus, how any unimodal/bimodal/trimodal temporal pattern of trauma death may be (mis)interpreted [3]. Recently, Pfeiffer and colleagues [5] examined 22 publications included in a review of trauma deaths. They noted a decrease of hemorrhage-induced deaths (25–15%) over the last decade; Evans noted a one third reduction [1], and we reported a one quarter reduction in our article [2]. No significant changes in the incidence and pattern of death were found in the review by Pfeiffer, considering all articles together. However, the predominant cause of death after trauma was found to be central nervous system injury, but with a wide range being reported (22–72%), followed by exsanguination (13–27%). Sepsis was reported in 3– 17% and multi-organ failure in about 2–9%. Several of the reports included in Pfeiffer’s review were from North American institutions, and only two were from Europe (Italy and Norway); one study was from New Zealand [5]. The wide range reported among the studies, even from similar time periods, suggest that care should be taken when comparing outcomes from different geographic regions, as this may amount to comparing apples and oranges. Thus, the results reported by Evans et al. [1] may reflect the global variance in fatal injury patterns rather than differences occurring in the trauma system per se. Also, the number and percentage of prehospital deaths (if included in the studies at all) are very much influenced by the geographic location, transport distances, and the K. Soreide (&) Department of Surgery, Stavanger University Hospital, University of Bergen, Bergen, Norway e-mail: ksoreide@mac.com

Epidemiology and Contemporary Patterns of Trauma Deaths: Changing Place, Similar Pace, Older Face

The epidemiology of trauma deaths in Europe is less than well investigated. Thus, our goal was to study the contemporary patterns of trauma deaths within a defined population with an exceptionally high trauma autopsy rate. This was a retrospective evaluation of 260 consecutive trauma autopsies for which we collected demographic, pre-hospital and in-hospital data. Patients were analyzed for injury severity by standard scoring systems (Abbreviated Injury Scale [AIS], Revised Trauma Score [RTS], and Injury Severity Score [ISS]), and the Trauma and Injury Severity Scale [TRISS] methodology. The fatal trauma incidence was 10.0 per 100,000 inhabitants (17.4 per 100,000 age-adjusted > or = 55 years). Blunt mechanism (87%), male gender (75%), and pre-hospital deaths (52%) predominated. Median ISS was 38 (range: 4-75). Younger patients (<55 years) who died in the hospital were more often hypotensive (SBP < 90 mmHg; p = 0.001), in respiratory distress (RR < 10/min, or > 29/min; p < 0.0001), and had deranged neurology on admission (Glasgow Coma Score [GCS] < or = 8; p < 0.0001), compared to those > or = 55 years. Causes of death were central nervous system (CNS) injuries (67%), exsanguination (25%), and multiorgan failure (8%). The temporal death distribution is model-dependent and can be visualized in unimodal, bimodal, or trimodal patterns. Age increased (r = 0.43) and ISS decreased (r = -0.52) with longer time from injury to death (p < 0.001). Mean age of the trauma patients who died increased by almost a decade during the study period (from mean 41.7 +/- 24.2 years to mean 50.5 +/- 25.4 years; p = 0.04). The pre-hospital:in-hospital death ratio shifted from 1.5 to 0.75 (p < 0.007). While pre-hospital and early deaths still predominate, an increasing proportion succumb after arrival in hospital. Focus on injury prevention is imperative, particularly for brain injuries. Although hemorrhage and multiorgan failure deaths have decreased, they do still occur. Redirected attention and focus on the geriatric trauma population is mandated.