Epidemiology of Trauma Deaths: Location, Location, Location!

Abstract

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