- Email: [email protected]
Systolic blood pressure below 110 mmHg is associated with increased mortality in blunt major trauma patients: Multicentre cohort study
Resuscitation 82 (2011) 1202–1207
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical paper
Systolic blood pressure below 110 mmHg is associated with increased mortality in blunt major trauma patients: Multicentre cohort study夽 Rebecca M. Hasler a,∗ , Eveline Nuesch c,d , Peter Jüni c,d , Omar Bouamra a , Aristomenis K. Exadaktylos b , Fiona Lecky a a Trauma Audit and Research Network (TARN), Health Sciences Research Group, School of Community Based Medicine, Manchester Academic Health Sciences Centre, University of Manchester, Salford Royal Hospital, Stott Lane, Salford M6 8HD, UK b Department of Emergency Medicine, University Hospital Bern, Freiburgstr, 3010 Bern, Switzerland c Institute of Social and Preventive Medicine (ISPM), University of Bern, Finkenhubelweg 11, 3012 Bern, Switzerland d Clinical Trials Unit (CTU) Bern, University Hospital Bern, Finkenhubelweg 11, 3012 Bern, Switzerland
a r t i c l e
i n f o
Article history: Received 1 October 2010 Received in revised form 12 April 2011 Accepted 24 April 2011
Keywords: Hypotension Mortality Systolic blood pressure Blunt trauma
a b s t r a c t Introduction: Non-invasive systolic blood pressure (SBP) measurement is often used in triaging trauma patients. Traditionally, SBP < 90 mmHg has represented the threshold for hypotension, but recent studies have suggested redefining hypotension as SBP < 110 mmHg. This study aims to examine the association of SBP with mortality in blunt trauma patients. Methods: This is an analysis of prospectively recorded data from adult (≥16 years) blunt trauma patients. Included patients presented to hospitals belonging to the Trauma Audit and Research Network (TARN) between 2000 and 2009. The primary outcome was the association of SBP and mortality rates at 30 days. Multivariate logistic regression models were used to adjust for the influence of age, gender, Injury Severity Score (ISS) and Glasgow Coma Score (GCS) on mortality. Results: 47,927 eligible patients presented to TARN hospitals during the study period. Sample demographics were: median age: 51.1 years (IQR = 32.8–67.4); male 60% (n = 28,694); median ISS 9 (IQR = 8–10); median GCS 15 (IQR = 15–15); and median SBP 135 mmHg (IQR = 120–152). We identified SBP < 110 mmHg as a cut off for hypotension, where a significant increase in mortality was observed. Mortality rates doubled at <100 mmHg, tripled at <90 mmHg and were 5- to 6-fold at <70 mmHg, irrespective of age. Conclusion: We recommend triaging adult blunt trauma patients with a SBP < 110 mmHg to resuscitation areas within dedicated trauma units for close monitoring and appropriate management. © 2011 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Trauma accounts for 10% of death and 14% of years of life lost worldwide and is the leading cause of death of people aged between 5 and 44 years in developed countries.1,2 Uncontrolled haemorrhage following traumatic injury accounts for 30–40% of trauma deaths,3 and about 25% of patients with multiple trauma and haemorrhage suffer from coagulopathy, acidosis, and hypovolaemic shock.4,5 In addition, excessive red blood cell transfusion is associated with a dose–response development of multiple organ failure,6 a threefold increased intensive care unit
夽 A Spanish translated version of the abstract of this article appears as Appendix in the final online version at doi:10.1016/j.resuscitation.2011.04.021. ∗ Corresponding author. E-mail addresses: [email protected] (R.M. Hasler), [email protected] (E. Nuesch), [email protected] (P. Jüni), [email protected] (O. Bouamra), [email protected] (A.K. Exadaktylos), fi[email protected] (F. Lecky). 0300-9572/$ – see front matter © 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.resuscitation.2011.04.021
admission rate and mortality.7 Non-invasive systolic blood pressure (SBP) measurement is quick and easy to obtain and often used as a triage tool in the pre-hospital and hospital sector to identify patients with major bleeding.8 Hypotensive trauma patients are usually referred to dedicated trauma centres to undergo specific care with maximum support and continued attention provided in dedicated resuscitation areas, whereas normotensive trauma patients might be considered less injured and therefore less observation and care might be provided. SBP levels are part of different guidelines in routine trauma care worldwide.9,10 The American College of Surgeons Committee on Trauma recommends triage of patients with a SBP < 90 mmHg to dedicated trauma centres.9 The multidisciplinary task force for advanced bleeding care in trauma also defines hypotension at a SBP of <90 mmHg10 and an International Consensus Conference in 2007 agreed that SBP < 90 mmHg represents critical hypotension, but recommends that hypotension should not be used to define the state of shock.11 Some authors argue that a SBP < 90 mmHg is a late sign of haemorrhage and rather an indicator for cardiovascular decompensation
R.M. Hasler et al. / Resuscitation 82 (2011) 1202–1207
than an early indicator of hypotension.8,12 Especially in healthy patients, physiologic compensatory mechanisms can initially mask underlying haemorrhage.13,14 It has therefore been suggested to reconsider definitions of critical hypotension using a cut-off for SBP at 110 mmHg.12,15 This analysis of the Trauma Audit and Research Network (TARN) aimed at extending the evidence base on the association of SBP with mortality in blunt trauma patients.
2. Methods 2.1. Data and inclusion criteria This is a cohort study using data from TARN, a European multicentre trauma register with prospective data collection from >140 hospitals using a web based data collecting and reporting system. TARN includes all trauma cases presenting directly to one of its participating hospitals, who either (i) require hospital admission for 72 h or more or (ii) are transferred into these for specialist care, or (iii) require a high dependency or intensive care unit, or (iv) who die as a result of their injuries within 93 days. Excluded are those patients transferred into these hospitals for rehabilitation only, any brain injury unrelated to trauma (e.g. spontaneous subarachnoid haemorrhage), simple skin lacerations, contusions or abrasions and minor penetrating injuries resulting in blood loss <20% (1000 ml). TARN also excludes patients over 65 years with isolated fracture of the femoral neck or pubic ramus and those with single uncomplicated limb injuries.16 Each patient’s injuries are described by the abbreviated injury scale (AIS).17 Each patient record is anonymised and contains details of mechanism of injury, age, gender, presenting physiology (on emergency department (ED) arrival) and the outcome, including mortality, at 30 days. Patients were included in our analysis if they had suffered an injury between January 2000 and December 2009. We excluded patients with penetrating injuries as it has been suggested that it can produce a different cardiovascular response to trauma.18 Patients suffering from concomitant head trauma (AIS 2+) were excluded as there is data suggesting that this as well might be associated with a different cardiovascular response.19 In addition, we excluded patients transferred from a non-TARN hospital as data on their clinical status at admission were missing, and patients referred to a non-TARN hospital as information on mortality was missing. NIGB approval was received for all patients (approval number: PIAG3-04(e)/2006).
2.2. Statistical analysis The primary outcome of this prospectively planned analysis was overall mortality after 30 days. Characteristics were described by median and interquartile ranges and compared using Mann–Whitney-U-Test. We analysed the association between mortality at 30 days with systolic blood pressure, measured in the emergency department, using mixed effects maximum likelihood models with a random intercept for hospitals to account for hierarchical structure in the data, because we know that mortality rates vary among different hospitals.20 Analyses were performed, crude, adjusted for age and adjusted for age, gender, ISS and GCS. Then we used multiple imputation of covariates with age, gender, ISS and hospital as independent variables in the imputation model.21 We used ISS in the following categories: (1) ISS 1–8, (2) ISS 9–15, (3) ISS 16–25 and (4) 26+ and GCS in the following categories: (0) 13–15, (1) 9–12, (2) 6–8, (3) 4–5 and (4) 3. Blood pressure was categorized in 10 mmHg increments, ranging from <70 to ≥200 mmHg. A category for missing SBP was added. Model quality was assessed using Akaike’s information criterion (AIC) and the area under the ROC
1203
Table 1 Injured body regions.
Face Neck Chest Abdomen Spine Upper limbs Lower limbs External
Patients (n)a
Patients (%)a (n = 47,927)
417 72 11,211 2352 4316 615 30,129 611
0.84 0.14 22.55 4.73 8.68 1.24 60.59 1.23
Numbers and percentage of patients with injuries AIS3+ for each body region. Sum total greater than number of patients due to multiply injured patients. a With injury in that body region.
curve (AUC) for prediction of mortality. Analyses were performed in Stata Release 11 (Stata Corp., College Station, TX). 3. Results 161 trauma centres (England: 144; Wales: 12; Ireland: 2; Denmark: 1; Switzerland: 1; Northern Ireland: 1) participated. Hospitals contributed a median of 199 patients (IQR 45–388). 65,330 adult patients, suffering from blunt injuries, other than head injuries, were entered into the TARN data base between January 2000 and December 2009 (Fig. 1). 11,786 (18%) patients transferred in or out of non-TARN hospitals were excluded, because no information on mortality after 30 days or initial physiology was available. 5617 (8%) cases were excluded due to missing data for SBP or GCS. 47,927 (74%) patients matched the inclusion criteria. 3.1. Study population Included patients had a median age of 51.1 years (IQR = 32.8–67.4). 60% (n = 28,694) were male. Median ISS was 9 (IQR = 8–10), median GCS 15 (IQR = 15–15) and median SBP 135 (IQR = 120–152). 1778 patients (3.7%) died during the first 30 days and 46,149 patients (96.3%) were still alive. 269 out of 13,508 patients died with an ISS of 1 to 8 at baseline (2%), whereas 747 out of 28,003 (2.7%), 363 out of 4725 (7.7%) and 399 out of 1691 patients (23.6%) died with an ISS at baseline of 9–15, 16–25, and 26 and above, respectively. The distribution of injured body regions is shown in Table 1. 60.6% of the injuries involved the lower limbs and 22.6% involved the chest, 8.7% the spine and 4.7% were abdominal injuries. Injuries of the face, neck and upper limbs and external injuries were rare (<1.3% for each body region). Characteristics of patients at baseline according to their survival status are presented in Table 2. We observed mortality increases associated with older age, higher ISS, lower GCS and SBP in crude analysis. 7971 patients (16.6%) had an SBP between 130 and 139 mmHg and a mortality rate of 2.4%, whereas 346 (0.7%) had extreme SBP <70 mmHg and 907 (1.9%) had an SBP of ≥200 mmHg. Mortality rates ranged from 7.3% in patients with an SBP between ≥200 mmHg and 33.5% in patients with an SBP of <70 mmHg (p for linear trend <0.001). 296 (0.6%) patients arrived mechanically ventilated. 103 (0.3%) of them had an SBP < 110 mmHg. Fig. 2 shows the association of the mortality after 30 days and the SBP at arrival in the ED. In the univariable analyses, we found an increasing risk of mortality with increasing and with decreasing SBP compared to the baseline category with an SBP of 130–139 mmHg. For example, the odds of death were doubled in patients with an SBP of 100–109 mmHg (OR 1.98, 95% CI 1.62–2.42), 5-fold in patients with an SBP of 80–89 mmHg (OR 5.33, 95% CI 4.21–6.76) and 18-fold in patients with an SBP <70 mmHg (OR 18.46, 95% CI 14.63–23.30) compared to the baseline category. The increase in mortality was
1204
R.M. Hasler et al. / Resuscitation 82 (2011) 1202–1207
Fig. 1. Flow chart showing patient flow according to exclusion criteria applied.
Fig. 2. Univariate and multivariate analysis (adjusted for age only and adjusted for age, gender, ISS and GCS) of categories of SBP levels (mean ± 95% confidence intervals) and its association with the corresponding odds of death, compared to the baseline at 130–139 mmHg.
R.M. Hasler et al. / Resuscitation 82 (2011) 1202–1207
1205
Table 2 Characteristics of patients at baseline. Characteristic at baseline
Patients died
Yes Age (n = 47,927) 16–30 31–45 46–60 61–75 ≥76 Gender (n = 47,927) Males Females SBP (n = 47,927) <70 70–79 80–89 90–99 100–109 110–119 120–129 130–139 140–149 150–159 160–169 170–179 180–189 190–199 ≥200 GCS (n = 47,927) 3 4–5 6–8 9–12 13–15 ISS (n = 47,927) 1–8 9–15 16–25 ≥26
Crude odds ratioa (95% CI)
P-Value
No
149 (1.36) 155 (1.59) 204 (1.89) 322 (4.03) 948 (11.3)
10,795 (98.6) 9616 (98.4) 10,607 (98.1) 7674 (96.0) 7475 (88.7)
1.00 (reference) 1.12 (0.93–1.47) 1.39 (1.12–1.73) 3.04 (2.43–3.81) 9.21 (7.64–11.1)
<0.001
977 (3.40) 801 (4.16)
27,717 (96.6) 18,432 (95.8)
1.00 (reference) 1.23 (1.12–1.36)
<0.001
116 (33.5) 68 (20.8) 113 (14.9) 136 (8.59) 146 (4.58) 178 (3.31) 184 (2.49) 193 (2.42) 153 (2.08) 162 (3.00) 116 (3.56) 52 (2.66) 62 (4.77) 33 (4.04) 66 (7.28)
230 (66.5) 259 (79.2) 648 (85.2) 1448 (91.4) 3045 (95.4) 5202 (96.7) 7209 (97.1) 7778 (97.6) 7194 (97.9) 5232 (97.0) 3142 (96.4) 1901 (97.3) 1237 (95.2) 783 (96.0) 841 (92.7)
20.3 (13.1–31.4) 10.6 (7.71–14.5) 7.03 (5.48–9.02) 3.79 (3.09–4.64) 1.93 (1.49–2.50) 1.38 (1.13–1.68) 1.03 (0.83–1.28) 1.00 (reference) 0.86 (0.68–1.08) 1.25 (1.00–1.56) 1.49 (1.21–1.83) 1.10 (0.81–1.49) 2.02 (1.45–2.82) 1.70 (1.16–2.48) 3.16 (2.26–4.42)
<0.001
222 (45.0) 29 (24.6) 61 (18.9) 108 (13.5) 1,358 (2.94)
271 (55.0) 89 (75.4) 261 (81.1) 691 (86.5) 44,837 (97.1)
27.0 (16.3–44.8) 10.8 (7.29–15.9) 7.72 (5.47–10.9) 5.16 (3.99–6.67) 1.00 (reference)
<0.001
269 (1.99) 747 (2.67) 363 (7.68) 399 (23.6)
3239 (98.0) 27,256 (97.3) 4362 (92.3) 1292 (76.4)
1.00 (reference) 1.35 (1.17–1.56) 4.10 (3.37–4.98) 15.2 (12.4–18.7)
<0.001
SBP, systolic blood pressure (mmHg); GCS, Glasgow Coma Score; ISS, Injury Severity Score. P-Values are from test for linear trend or for association, from maximum likelihood models with random intercept for hospitals to account for clustering of patients within hospitals. a Odds ratios larger than 1 indicate a lower mortality in the reference category.
also prominent, but less pronounced in patients with increasing SBP. The odds of mortality were doubled (OR 2.03, 95% CI 1.51–2.73) in patients with an SBP 180–189 mmHg and tripled in patients with an SBP ≥ 200 mmHg (OR 3.27, 95% CI 2.44–4.39) compared to patients with 130–139 mmHg. When adjusting for age only, we found the associations with mortality robust for the categories of SBP <130 mmHg. In contrast, the odds of death did change substantially with increasing categories of SBP. In the fully adjusted analysis, results were attenuated for SBP categories <130 mmHg. Odds of mortality started to raise at SBP <110 mmHg (OR 1.69, 95% CI 1.35–2.12), doubled at <100 mmHg (OR 2.11, 95% CI 1.65–2.70), tripled at <90 mmHg (OR 2.90, 95% CI 2.20–3.84) and were six-fold at <70 mmHg (OR 5.98, 95% CI 4.42–8.09) compared to the baseline. Results of the fully adjusted analysis were much the same after multiple imputation, except for an attenuation of the association in patients with an SBP <70 mmHg (OR 4.81, 95% CI 3.52–6.58). Model fit was good with an area under the ROC curve of 0.89 and an AIC of 10,811.15 for the fully adjusted model. The Web-Appendix shows a comparison of patients with complete data, patients with incomplete data for SBP and GCS for whom multiple imputation was performed, and patients excluded from the analysis as they were either transferred from a non-TARN hospital or referred to a non-TARN hospital. Patients with incomplete data and necessity of multiple imputation were more likely to be female (45.5% versus 40.1%), remaining differences were clinically irrelevant. Patients excluded were younger (40.9 years versus 51.1 years), more often
male (71.6% versus 59.9%) and had a slightly lower SBP (130 mmHg versus 135 mmHg). Remaining differences were small and clinically irrelevant, albeit statistically significant at the conventional level. 4. Discussion This analysis of prospectively collected data in 47,927 adult cases of blunt trauma suggests that the rate of mortality starts to increase with an SBP < 110 mmHg and continues to rise exponentially with lower blood pressures, irrespective of age. Mortality rates double at <100 mmHg, triple at <90 mmHg and increase 5- to 6-fold at <70 mmHg, compared to the baseline at 130–139 mmHg. Our study therefore supports a cut-off to define hypotension in blunt trauma patients at SBP < 110 mmHg. The apparent association of higher SBP levels with mortality in the crude analysis disappears when age is adjusted for. Interestingly, we were unable to show any impact of age-adjustment on estimated associations of low SBP values with mortality as was previously suggested.15 Below 110 mmHg, the increase of odds of death was much steeper for the univariable and the age adjusted graphs compared to the multivariate analysis. Major strengths include the multicentric nature of the study with a large number of patients with blunt trauma and prospectively collected data. Multivariate analysis using mixed effect logistic regression was applied and adjusted for age, gender, ISS and GCS to overcome their confounding effect on mortality rates.
1206
R.M. Hasler et al. / Resuscitation 82 (2011) 1202–1207
However, we were unable to adjust for a range of further possible confounders, like base excess, lactate levels, body temperature or co-morbidities. Therefore, we cannot exclude residual confounding. Nevertheless, the analysis showed a good model fit. Another limitation of our study is the occurrence of missing data in the covariates. We used two different approaches to deal with missing data (complete case analysis and multiple imputation), which yielded mostly similar results. Imputation for missing data showed, however, slightly lower odds of death for SBPs < 70 mmHg, but was consistent with the multivariable analysis for all other levels of SBP.21 Since characteristics of patients and mortality rates may vary among different hospitals,20 we used multi-level regression models that accounted for the correlation of patients and their mortality risk within hospitals. The patients excluded from the analysis differed from the analysed population in some characteristics and we are unable to assess the effect of their exclusion on the association of SBP with mortality. We excluded patients with traumatic brain injury, as studies in animals support the fact, that brain injury is likely to affect haemodynamic responses in bleeding individuals.22–24 As traumatic brain injury patients can be readily characterised based on clinical grounds at admission, managing them according to separate guidelines should not impose a practical challenge for triage at emergency departments. Previous studies have shown that automated blood pressure may vary from manual measurements, especially in patients with very low or very high SBP.25 In our study, SBP was measured using either automated or manual non-invasive blood pressure measurements, which may have led to random misclassification of some patients and may have biased our results towards underestimation of associations and we were unable to account for this in the analysis. However, oscillometric methods for non-invasive blood pressure measurement prevail in all our centres. Manual sphygomanometres may be used in hypotensive patients in case measurement with oscillometric methods appear unreliable. The low proportion of intubated patients had no impact on our findings. This is due to the fact, that for the majority of TARN centres, ambulances are staffed by paramedics only, who are not allowed to intubate patients. In 2007 Eastridge et al. redefined hypotension using a cut-off of 110 mmHg SBP in a mixed sample of 730,000 American patients with blunt or penetrating trauma by plotting SBPs against mortality rates.15 They postulated that the mortality rate is increased 4.8% for every 10 mmHg decrement between 110 and 60 mmHg peaking in a 26% mortality at 60 mmHg compared to the baseline mortality of <2.5%. Our crude estimates are similar to those reported by Eastridge et al., but we emphasise that statistical adjustment resulted in a considerable decrease in the odds of death associated with low SBP. Bruns et al. performed a similar study in 2008 in the U.S., examining 16,000 blunt and penetrating trauma patients in a pre-hospital setting suggesting a cut off of 110 mmHg as well.12 They found a strong association between falling SBP levels and rising injury severity scores below 110 mmHg, but did not adjust for ISS in their cut off analysis. At SBPs of ≤70 mmHg they observed a 64% mortality rate, which would probably have decreased considerably once ISS had been adjusted for. Parks et al. showed in over 115,000 blunt and penetrating trauma patients that hypotension, defined as an SBP < 90 mmHg, is associated with base deficit levels <−20 and therefore represents a late stage of shock.8 Our data indirectly support this finding. In the recent CRASH-2 trial in adult trauma patients with or at risk of major bleeding,26 68.4% of 20,211 included patients had SBP at inclusion into the trial of ≥90 mmHg. This again supports our notion that 90 mmHg is a late sign of haemorrhage and many bleeding patients are missed by traditional definitions of hypotension at a cut-off of SBP of 90 mmHg. Prior studies defining hypotension examined mixed samples of patients with blunt and penetrating injuries.8,12,14,15 Some were
from single centres,12,14 used pre-hospital systolic blood pressures for their analysis,12 or looked at SBP in combination with less easily obtainable parameters like base excess.8 To the best of our knowledge this is the first multicentre study looking at SBP levels and mortality rates with adjustment for common confounders exclusively in patients with blunt trauma. In our study, only 4775 out of 47,927 patients (10.0%) had an SBP between 90 and <110 mmHg. Using a cut-off of 110 mmHg rather than 90 is therefore unlikely to result in an overflow of resuscitation areas. As this study was performed on SBP measurement on admission to the Emergency Department, recommendations for a cut-off of SBP refer to triage of trauma patients within trauma centres and cannot be necessarily generalised to pre-hospital settings. 5. Conclusion We identified a cut off for definition of hypotension of <110 mmHg. Mortality rates were doubled at <100 mmHg, tripled at <90 mmHg and increased 5- to 6-fold at <70 mmHg, compared to the baseline at 130–139 mmHg. We therefore recommend that all blunt trauma patients with an SBP < 110 mmHg should be triaged to resuscitation areas within trauma centres for close monitoring and optimised clinical management. Conflict of interest statement The authors of this study disclose any financial and personal relationships with other people or organisations that could inappropriately influence (bias) their work. Funding sources The authors of this study declare that they had no sponsors in the study design, in the collection, analysis and interpretation of data; in the writing of the manuscript; and in the decision to submit the manuscript for publication. Acknowledgements We are thankful to staff at the Trauma Audit and Research Network and all participating hospitals. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.resuscitation.2011.04.021. References 1. World Health Organisation (WHO). World Health Statistics 2010: cause specific mortality and morbidity. Avenue Appia 20. 1211 Geneva 27, Switzerland [Accessed 31 August 2010 at http://www.who.int/ whosis/whostat/EN WHS09 Table2.pdf]. 2. Krug EG, Sharma GK, Lozano R. The global burden of injuries. Am J Public Health 2000;90:523–6. 3. Sauaia A, Moore FA, Moore EE, et al. Epidemiology of trauma deaths: a reassessment. J Trauma 1995;38:185–93. 4. Brohi K, Cohen MJ, Davenport RA. Acute coagulopathy of trauma mechanism, identification and effect. Curr Opin Crit Care 2007;13:680–5. 5. Tieu BH, Holcomb JB, Schreiber MA. Coagulopathy: its pathophysiology and treatment in the injured patient. World J Surg 2007;31:1055–64. 6. Moore FA, Moore EE, Sauaia A. Blood transfusion. An independent risk factor for postinjury multiple organ failure. Arch Surg 1997;132:620–4. 7. Malone DL, Dunne J, Tracy JK, Putnam AT, Scalea TM, Napolitano LM. Blood transfusion, independent of shock severity, is associated with worse outcome in trauma. J Trauma 2003;54:898–905. 8. Parks JK, Elliott AC, Gentilello LM, Shafi S. Systemic hypotension is a late marker of shock after trauma: a validation study of advanced trauma life support principles in a large national sample. Am J Surg 2006;192:727–31. 9. American College of Surgeons. Resources for the optimal care of the injured patient. Chicago, IL: American College of Surgeons; 2006.
R.M. Hasler et al. / Resuscitation 82 (2011) 1202–1207 10. Rossaint R, Bouillon B, Cerny V, et al. Management of bleeding following major trauma: an updated European guideline. Crit Care 2010;14:R52. 11. Antonelli M, Levy M, Andrews PJ, et al. Hemodynamic monitoring in shock and implications for management. International Consensus Conference. Intensive Care Med 2007;33:575–90. 12. Bruns B, Gentilello L, Elliott A, Shafi S. Prehospital hypotension redefined. J Trauma 2008;65:1217–21. 13. Shires GT, Carrico CJ, Canizaro PC. In: Dunphy JE, editor. Shock, vol. 8. Philadelphia, PA: WB Saunders Company; 1973. 14. Lipsky AM, Gausche-Hill M, Henneman PL, et al. Prehospital hypotension is a predictor of the need for an emergent, therapeutic operation in trauma patients with normal systolic blood pressure in the emergency department. J Trauma 2006;61:1228–33. 15. Eastridge BJ, Salinas J, McManus JG, et al. Hypotension begins at 110 mm Hg: redefining “hypotension” with data. J Trauma 2007;63:291–7. 16. The Trauma Audit & Research Network. The trauma audit & research network. Procedures Manual. Salford, UK: The Trauma Audit & Research Network; 2009. 17. Committee on Injury Scaling. Association for the advancement of automotive medicine (AAAM). The abbreviated injury scale 2005 revision. Des Plaines, Chicago: AAAM; 2005. 18. Guly HR, Bouamra O, Little R, et al. Testing the validity of the ATLS classification of hypovolaemic shock. Resuscitation 2010;81:1142–7.
1207
19. McMahon CG, Kenny R, Bennett K, et al. The effect of acute traumatic brain injury on the performance of shock index. J Trauma 2010;69:1169–75. 20. The Trauma Audit and Research Network. Performance comparison. Eccles Old Road, Salford: Clinical Sciences Building, Hope Hospital [Accessed 31 August 2010 https://www.tarn.ac.uk/Content.aspx?ca=15]. 21. Sterne JA, White IR, Carlin JB, et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ 2009;338:b2393. 22. Goldstein B, Toweill D, Lai S, Sonnenthal K, Kimberly B. Uncoupling of the autonomic and cardiovascular systems in acute brain injury. Am J Physiol 1998;275:R1287–92. 23. McMahon CG, Kenny R, Bennett K, Kirkman E. Modification of acute cardiovascular homeostatic responses to hemorrhage following mild to moderate traumatic brain injury. Crit Care Med 2008;36:216–24. 24. Law MM, Hovda DA, Cryer HG. Fluid-percussion brain injury adversely affects control of vascular tone during hemorrhagic shock. Shock 1996;6:213–7. 25. Davis JW, Davis IC, Bennink LD, Bilello JF, Kaups KL, Parks SN. Are automated blood pressure measurements accurate in trauma patients? J Trauma 2003;55:860–3. 26. Shakur H, Roberts I, Bautista R, et al. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2): a randomised, placebo-controlled trial. Lancet 2010;376:23–32, 3.
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical paper
Systolic blood pressure below 110 mmHg is associated with increased mortality in blunt major trauma patients: Multicentre cohort study夽 Rebecca M. Hasler a,∗ , Eveline Nuesch c,d , Peter Jüni c,d , Omar Bouamra a , Aristomenis K. Exadaktylos b , Fiona Lecky a a Trauma Audit and Research Network (TARN), Health Sciences Research Group, School of Community Based Medicine, Manchester Academic Health Sciences Centre, University of Manchester, Salford Royal Hospital, Stott Lane, Salford M6 8HD, UK b Department of Emergency Medicine, University Hospital Bern, Freiburgstr, 3010 Bern, Switzerland c Institute of Social and Preventive Medicine (ISPM), University of Bern, Finkenhubelweg 11, 3012 Bern, Switzerland d Clinical Trials Unit (CTU) Bern, University Hospital Bern, Finkenhubelweg 11, 3012 Bern, Switzerland
a r t i c l e
i n f o
Article history: Received 1 October 2010 Received in revised form 12 April 2011 Accepted 24 April 2011
Keywords: Hypotension Mortality Systolic blood pressure Blunt trauma
a b s t r a c t Introduction: Non-invasive systolic blood pressure (SBP) measurement is often used in triaging trauma patients. Traditionally, SBP < 90 mmHg has represented the threshold for hypotension, but recent studies have suggested redefining hypotension as SBP < 110 mmHg. This study aims to examine the association of SBP with mortality in blunt trauma patients. Methods: This is an analysis of prospectively recorded data from adult (≥16 years) blunt trauma patients. Included patients presented to hospitals belonging to the Trauma Audit and Research Network (TARN) between 2000 and 2009. The primary outcome was the association of SBP and mortality rates at 30 days. Multivariate logistic regression models were used to adjust for the influence of age, gender, Injury Severity Score (ISS) and Glasgow Coma Score (GCS) on mortality. Results: 47,927 eligible patients presented to TARN hospitals during the study period. Sample demographics were: median age: 51.1 years (IQR = 32.8–67.4); male 60% (n = 28,694); median ISS 9 (IQR = 8–10); median GCS 15 (IQR = 15–15); and median SBP 135 mmHg (IQR = 120–152). We identified SBP < 110 mmHg as a cut off for hypotension, where a significant increase in mortality was observed. Mortality rates doubled at <100 mmHg, tripled at <90 mmHg and were 5- to 6-fold at <70 mmHg, irrespective of age. Conclusion: We recommend triaging adult blunt trauma patients with a SBP < 110 mmHg to resuscitation areas within dedicated trauma units for close monitoring and appropriate management. © 2011 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Trauma accounts for 10% of death and 14% of years of life lost worldwide and is the leading cause of death of people aged between 5 and 44 years in developed countries.1,2 Uncontrolled haemorrhage following traumatic injury accounts for 30–40% of trauma deaths,3 and about 25% of patients with multiple trauma and haemorrhage suffer from coagulopathy, acidosis, and hypovolaemic shock.4,5 In addition, excessive red blood cell transfusion is associated with a dose–response development of multiple organ failure,6 a threefold increased intensive care unit
夽 A Spanish translated version of the abstract of this article appears as Appendix in the final online version at doi:10.1016/j.resuscitation.2011.04.021. ∗ Corresponding author. E-mail addresses: [email protected] (R.M. Hasler), [email protected] (E. Nuesch), [email protected] (P. Jüni), [email protected] (O. Bouamra), [email protected] (A.K. Exadaktylos), fi[email protected] (F. Lecky). 0300-9572/$ – see front matter © 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.resuscitation.2011.04.021
admission rate and mortality.7 Non-invasive systolic blood pressure (SBP) measurement is quick and easy to obtain and often used as a triage tool in the pre-hospital and hospital sector to identify patients with major bleeding.8 Hypotensive trauma patients are usually referred to dedicated trauma centres to undergo specific care with maximum support and continued attention provided in dedicated resuscitation areas, whereas normotensive trauma patients might be considered less injured and therefore less observation and care might be provided. SBP levels are part of different guidelines in routine trauma care worldwide.9,10 The American College of Surgeons Committee on Trauma recommends triage of patients with a SBP < 90 mmHg to dedicated trauma centres.9 The multidisciplinary task force for advanced bleeding care in trauma also defines hypotension at a SBP of <90 mmHg10 and an International Consensus Conference in 2007 agreed that SBP < 90 mmHg represents critical hypotension, but recommends that hypotension should not be used to define the state of shock.11 Some authors argue that a SBP < 90 mmHg is a late sign of haemorrhage and rather an indicator for cardiovascular decompensation
R.M. Hasler et al. / Resuscitation 82 (2011) 1202–1207
than an early indicator of hypotension.8,12 Especially in healthy patients, physiologic compensatory mechanisms can initially mask underlying haemorrhage.13,14 It has therefore been suggested to reconsider definitions of critical hypotension using a cut-off for SBP at 110 mmHg.12,15 This analysis of the Trauma Audit and Research Network (TARN) aimed at extending the evidence base on the association of SBP with mortality in blunt trauma patients.
2. Methods 2.1. Data and inclusion criteria This is a cohort study using data from TARN, a European multicentre trauma register with prospective data collection from >140 hospitals using a web based data collecting and reporting system. TARN includes all trauma cases presenting directly to one of its participating hospitals, who either (i) require hospital admission for 72 h or more or (ii) are transferred into these for specialist care, or (iii) require a high dependency or intensive care unit, or (iv) who die as a result of their injuries within 93 days. Excluded are those patients transferred into these hospitals for rehabilitation only, any brain injury unrelated to trauma (e.g. spontaneous subarachnoid haemorrhage), simple skin lacerations, contusions or abrasions and minor penetrating injuries resulting in blood loss <20% (1000 ml). TARN also excludes patients over 65 years with isolated fracture of the femoral neck or pubic ramus and those with single uncomplicated limb injuries.16 Each patient’s injuries are described by the abbreviated injury scale (AIS).17 Each patient record is anonymised and contains details of mechanism of injury, age, gender, presenting physiology (on emergency department (ED) arrival) and the outcome, including mortality, at 30 days. Patients were included in our analysis if they had suffered an injury between January 2000 and December 2009. We excluded patients with penetrating injuries as it has been suggested that it can produce a different cardiovascular response to trauma.18 Patients suffering from concomitant head trauma (AIS 2+) were excluded as there is data suggesting that this as well might be associated with a different cardiovascular response.19 In addition, we excluded patients transferred from a non-TARN hospital as data on their clinical status at admission were missing, and patients referred to a non-TARN hospital as information on mortality was missing. NIGB approval was received for all patients (approval number: PIAG3-04(e)/2006).
2.2. Statistical analysis The primary outcome of this prospectively planned analysis was overall mortality after 30 days. Characteristics were described by median and interquartile ranges and compared using Mann–Whitney-U-Test. We analysed the association between mortality at 30 days with systolic blood pressure, measured in the emergency department, using mixed effects maximum likelihood models with a random intercept for hospitals to account for hierarchical structure in the data, because we know that mortality rates vary among different hospitals.20 Analyses were performed, crude, adjusted for age and adjusted for age, gender, ISS and GCS. Then we used multiple imputation of covariates with age, gender, ISS and hospital as independent variables in the imputation model.21 We used ISS in the following categories: (1) ISS 1–8, (2) ISS 9–15, (3) ISS 16–25 and (4) 26+ and GCS in the following categories: (0) 13–15, (1) 9–12, (2) 6–8, (3) 4–5 and (4) 3. Blood pressure was categorized in 10 mmHg increments, ranging from <70 to ≥200 mmHg. A category for missing SBP was added. Model quality was assessed using Akaike’s information criterion (AIC) and the area under the ROC
1203
Table 1 Injured body regions.
Face Neck Chest Abdomen Spine Upper limbs Lower limbs External
Patients (n)a
Patients (%)a (n = 47,927)
417 72 11,211 2352 4316 615 30,129 611
0.84 0.14 22.55 4.73 8.68 1.24 60.59 1.23
Numbers and percentage of patients with injuries AIS3+ for each body region. Sum total greater than number of patients due to multiply injured patients. a With injury in that body region.
curve (AUC) for prediction of mortality. Analyses were performed in Stata Release 11 (Stata Corp., College Station, TX). 3. Results 161 trauma centres (England: 144; Wales: 12; Ireland: 2; Denmark: 1; Switzerland: 1; Northern Ireland: 1) participated. Hospitals contributed a median of 199 patients (IQR 45–388). 65,330 adult patients, suffering from blunt injuries, other than head injuries, were entered into the TARN data base between January 2000 and December 2009 (Fig. 1). 11,786 (18%) patients transferred in or out of non-TARN hospitals were excluded, because no information on mortality after 30 days or initial physiology was available. 5617 (8%) cases were excluded due to missing data for SBP or GCS. 47,927 (74%) patients matched the inclusion criteria. 3.1. Study population Included patients had a median age of 51.1 years (IQR = 32.8–67.4). 60% (n = 28,694) were male. Median ISS was 9 (IQR = 8–10), median GCS 15 (IQR = 15–15) and median SBP 135 (IQR = 120–152). 1778 patients (3.7%) died during the first 30 days and 46,149 patients (96.3%) were still alive. 269 out of 13,508 patients died with an ISS of 1 to 8 at baseline (2%), whereas 747 out of 28,003 (2.7%), 363 out of 4725 (7.7%) and 399 out of 1691 patients (23.6%) died with an ISS at baseline of 9–15, 16–25, and 26 and above, respectively. The distribution of injured body regions is shown in Table 1. 60.6% of the injuries involved the lower limbs and 22.6% involved the chest, 8.7% the spine and 4.7% were abdominal injuries. Injuries of the face, neck and upper limbs and external injuries were rare (<1.3% for each body region). Characteristics of patients at baseline according to their survival status are presented in Table 2. We observed mortality increases associated with older age, higher ISS, lower GCS and SBP in crude analysis. 7971 patients (16.6%) had an SBP between 130 and 139 mmHg and a mortality rate of 2.4%, whereas 346 (0.7%) had extreme SBP <70 mmHg and 907 (1.9%) had an SBP of ≥200 mmHg. Mortality rates ranged from 7.3% in patients with an SBP between ≥200 mmHg and 33.5% in patients with an SBP of <70 mmHg (p for linear trend <0.001). 296 (0.6%) patients arrived mechanically ventilated. 103 (0.3%) of them had an SBP < 110 mmHg. Fig. 2 shows the association of the mortality after 30 days and the SBP at arrival in the ED. In the univariable analyses, we found an increasing risk of mortality with increasing and with decreasing SBP compared to the baseline category with an SBP of 130–139 mmHg. For example, the odds of death were doubled in patients with an SBP of 100–109 mmHg (OR 1.98, 95% CI 1.62–2.42), 5-fold in patients with an SBP of 80–89 mmHg (OR 5.33, 95% CI 4.21–6.76) and 18-fold in patients with an SBP <70 mmHg (OR 18.46, 95% CI 14.63–23.30) compared to the baseline category. The increase in mortality was
1204
R.M. Hasler et al. / Resuscitation 82 (2011) 1202–1207
Fig. 1. Flow chart showing patient flow according to exclusion criteria applied.
Fig. 2. Univariate and multivariate analysis (adjusted for age only and adjusted for age, gender, ISS and GCS) of categories of SBP levels (mean ± 95% confidence intervals) and its association with the corresponding odds of death, compared to the baseline at 130–139 mmHg.
R.M. Hasler et al. / Resuscitation 82 (2011) 1202–1207
1205
Table 2 Characteristics of patients at baseline. Characteristic at baseline
Patients died
Yes Age (n = 47,927) 16–30 31–45 46–60 61–75 ≥76 Gender (n = 47,927) Males Females SBP (n = 47,927) <70 70–79 80–89 90–99 100–109 110–119 120–129 130–139 140–149 150–159 160–169 170–179 180–189 190–199 ≥200 GCS (n = 47,927) 3 4–5 6–8 9–12 13–15 ISS (n = 47,927) 1–8 9–15 16–25 ≥26
Crude odds ratioa (95% CI)
P-Value
No
149 (1.36) 155 (1.59) 204 (1.89) 322 (4.03) 948 (11.3)
10,795 (98.6) 9616 (98.4) 10,607 (98.1) 7674 (96.0) 7475 (88.7)
1.00 (reference) 1.12 (0.93–1.47) 1.39 (1.12–1.73) 3.04 (2.43–3.81) 9.21 (7.64–11.1)
<0.001
977 (3.40) 801 (4.16)
27,717 (96.6) 18,432 (95.8)
1.00 (reference) 1.23 (1.12–1.36)
<0.001
116 (33.5) 68 (20.8) 113 (14.9) 136 (8.59) 146 (4.58) 178 (3.31) 184 (2.49) 193 (2.42) 153 (2.08) 162 (3.00) 116 (3.56) 52 (2.66) 62 (4.77) 33 (4.04) 66 (7.28)
230 (66.5) 259 (79.2) 648 (85.2) 1448 (91.4) 3045 (95.4) 5202 (96.7) 7209 (97.1) 7778 (97.6) 7194 (97.9) 5232 (97.0) 3142 (96.4) 1901 (97.3) 1237 (95.2) 783 (96.0) 841 (92.7)
20.3 (13.1–31.4) 10.6 (7.71–14.5) 7.03 (5.48–9.02) 3.79 (3.09–4.64) 1.93 (1.49–2.50) 1.38 (1.13–1.68) 1.03 (0.83–1.28) 1.00 (reference) 0.86 (0.68–1.08) 1.25 (1.00–1.56) 1.49 (1.21–1.83) 1.10 (0.81–1.49) 2.02 (1.45–2.82) 1.70 (1.16–2.48) 3.16 (2.26–4.42)
<0.001
222 (45.0) 29 (24.6) 61 (18.9) 108 (13.5) 1,358 (2.94)
271 (55.0) 89 (75.4) 261 (81.1) 691 (86.5) 44,837 (97.1)
27.0 (16.3–44.8) 10.8 (7.29–15.9) 7.72 (5.47–10.9) 5.16 (3.99–6.67) 1.00 (reference)
<0.001
269 (1.99) 747 (2.67) 363 (7.68) 399 (23.6)
3239 (98.0) 27,256 (97.3) 4362 (92.3) 1292 (76.4)
1.00 (reference) 1.35 (1.17–1.56) 4.10 (3.37–4.98) 15.2 (12.4–18.7)
<0.001
SBP, systolic blood pressure (mmHg); GCS, Glasgow Coma Score; ISS, Injury Severity Score. P-Values are from test for linear trend or for association, from maximum likelihood models with random intercept for hospitals to account for clustering of patients within hospitals. a Odds ratios larger than 1 indicate a lower mortality in the reference category.
also prominent, but less pronounced in patients with increasing SBP. The odds of mortality were doubled (OR 2.03, 95% CI 1.51–2.73) in patients with an SBP 180–189 mmHg and tripled in patients with an SBP ≥ 200 mmHg (OR 3.27, 95% CI 2.44–4.39) compared to patients with 130–139 mmHg. When adjusting for age only, we found the associations with mortality robust for the categories of SBP <130 mmHg. In contrast, the odds of death did change substantially with increasing categories of SBP. In the fully adjusted analysis, results were attenuated for SBP categories <130 mmHg. Odds of mortality started to raise at SBP <110 mmHg (OR 1.69, 95% CI 1.35–2.12), doubled at <100 mmHg (OR 2.11, 95% CI 1.65–2.70), tripled at <90 mmHg (OR 2.90, 95% CI 2.20–3.84) and were six-fold at <70 mmHg (OR 5.98, 95% CI 4.42–8.09) compared to the baseline. Results of the fully adjusted analysis were much the same after multiple imputation, except for an attenuation of the association in patients with an SBP <70 mmHg (OR 4.81, 95% CI 3.52–6.58). Model fit was good with an area under the ROC curve of 0.89 and an AIC of 10,811.15 for the fully adjusted model. The Web-Appendix shows a comparison of patients with complete data, patients with incomplete data for SBP and GCS for whom multiple imputation was performed, and patients excluded from the analysis as they were either transferred from a non-TARN hospital or referred to a non-TARN hospital. Patients with incomplete data and necessity of multiple imputation were more likely to be female (45.5% versus 40.1%), remaining differences were clinically irrelevant. Patients excluded were younger (40.9 years versus 51.1 years), more often
male (71.6% versus 59.9%) and had a slightly lower SBP (130 mmHg versus 135 mmHg). Remaining differences were small and clinically irrelevant, albeit statistically significant at the conventional level. 4. Discussion This analysis of prospectively collected data in 47,927 adult cases of blunt trauma suggests that the rate of mortality starts to increase with an SBP < 110 mmHg and continues to rise exponentially with lower blood pressures, irrespective of age. Mortality rates double at <100 mmHg, triple at <90 mmHg and increase 5- to 6-fold at <70 mmHg, compared to the baseline at 130–139 mmHg. Our study therefore supports a cut-off to define hypotension in blunt trauma patients at SBP < 110 mmHg. The apparent association of higher SBP levels with mortality in the crude analysis disappears when age is adjusted for. Interestingly, we were unable to show any impact of age-adjustment on estimated associations of low SBP values with mortality as was previously suggested.15 Below 110 mmHg, the increase of odds of death was much steeper for the univariable and the age adjusted graphs compared to the multivariate analysis. Major strengths include the multicentric nature of the study with a large number of patients with blunt trauma and prospectively collected data. Multivariate analysis using mixed effect logistic regression was applied and adjusted for age, gender, ISS and GCS to overcome their confounding effect on mortality rates.
1206
R.M. Hasler et al. / Resuscitation 82 (2011) 1202–1207
However, we were unable to adjust for a range of further possible confounders, like base excess, lactate levels, body temperature or co-morbidities. Therefore, we cannot exclude residual confounding. Nevertheless, the analysis showed a good model fit. Another limitation of our study is the occurrence of missing data in the covariates. We used two different approaches to deal with missing data (complete case analysis and multiple imputation), which yielded mostly similar results. Imputation for missing data showed, however, slightly lower odds of death for SBPs < 70 mmHg, but was consistent with the multivariable analysis for all other levels of SBP.21 Since characteristics of patients and mortality rates may vary among different hospitals,20 we used multi-level regression models that accounted for the correlation of patients and their mortality risk within hospitals. The patients excluded from the analysis differed from the analysed population in some characteristics and we are unable to assess the effect of their exclusion on the association of SBP with mortality. We excluded patients with traumatic brain injury, as studies in animals support the fact, that brain injury is likely to affect haemodynamic responses in bleeding individuals.22–24 As traumatic brain injury patients can be readily characterised based on clinical grounds at admission, managing them according to separate guidelines should not impose a practical challenge for triage at emergency departments. Previous studies have shown that automated blood pressure may vary from manual measurements, especially in patients with very low or very high SBP.25 In our study, SBP was measured using either automated or manual non-invasive blood pressure measurements, which may have led to random misclassification of some patients and may have biased our results towards underestimation of associations and we were unable to account for this in the analysis. However, oscillometric methods for non-invasive blood pressure measurement prevail in all our centres. Manual sphygomanometres may be used in hypotensive patients in case measurement with oscillometric methods appear unreliable. The low proportion of intubated patients had no impact on our findings. This is due to the fact, that for the majority of TARN centres, ambulances are staffed by paramedics only, who are not allowed to intubate patients. In 2007 Eastridge et al. redefined hypotension using a cut-off of 110 mmHg SBP in a mixed sample of 730,000 American patients with blunt or penetrating trauma by plotting SBPs against mortality rates.15 They postulated that the mortality rate is increased 4.8% for every 10 mmHg decrement between 110 and 60 mmHg peaking in a 26% mortality at 60 mmHg compared to the baseline mortality of <2.5%. Our crude estimates are similar to those reported by Eastridge et al., but we emphasise that statistical adjustment resulted in a considerable decrease in the odds of death associated with low SBP. Bruns et al. performed a similar study in 2008 in the U.S., examining 16,000 blunt and penetrating trauma patients in a pre-hospital setting suggesting a cut off of 110 mmHg as well.12 They found a strong association between falling SBP levels and rising injury severity scores below 110 mmHg, but did not adjust for ISS in their cut off analysis. At SBPs of ≤70 mmHg they observed a 64% mortality rate, which would probably have decreased considerably once ISS had been adjusted for. Parks et al. showed in over 115,000 blunt and penetrating trauma patients that hypotension, defined as an SBP < 90 mmHg, is associated with base deficit levels <−20 and therefore represents a late stage of shock.8 Our data indirectly support this finding. In the recent CRASH-2 trial in adult trauma patients with or at risk of major bleeding,26 68.4% of 20,211 included patients had SBP at inclusion into the trial of ≥90 mmHg. This again supports our notion that 90 mmHg is a late sign of haemorrhage and many bleeding patients are missed by traditional definitions of hypotension at a cut-off of SBP of 90 mmHg. Prior studies defining hypotension examined mixed samples of patients with blunt and penetrating injuries.8,12,14,15 Some were
from single centres,12,14 used pre-hospital systolic blood pressures for their analysis,12 or looked at SBP in combination with less easily obtainable parameters like base excess.8 To the best of our knowledge this is the first multicentre study looking at SBP levels and mortality rates with adjustment for common confounders exclusively in patients with blunt trauma. In our study, only 4775 out of 47,927 patients (10.0%) had an SBP between 90 and <110 mmHg. Using a cut-off of 110 mmHg rather than 90 is therefore unlikely to result in an overflow of resuscitation areas. As this study was performed on SBP measurement on admission to the Emergency Department, recommendations for a cut-off of SBP refer to triage of trauma patients within trauma centres and cannot be necessarily generalised to pre-hospital settings. 5. Conclusion We identified a cut off for definition of hypotension of <110 mmHg. Mortality rates were doubled at <100 mmHg, tripled at <90 mmHg and increased 5- to 6-fold at <70 mmHg, compared to the baseline at 130–139 mmHg. We therefore recommend that all blunt trauma patients with an SBP < 110 mmHg should be triaged to resuscitation areas within trauma centres for close monitoring and optimised clinical management. Conflict of interest statement The authors of this study disclose any financial and personal relationships with other people or organisations that could inappropriately influence (bias) their work. Funding sources The authors of this study declare that they had no sponsors in the study design, in the collection, analysis and interpretation of data; in the writing of the manuscript; and in the decision to submit the manuscript for publication. Acknowledgements We are thankful to staff at the Trauma Audit and Research Network and all participating hospitals. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.resuscitation.2011.04.021. References 1. World Health Organisation (WHO). World Health Statistics 2010: cause specific mortality and morbidity. Avenue Appia 20. 1211 Geneva 27, Switzerland [Accessed 31 August 2010 at http://www.who.int/ whosis/whostat/EN WHS09 Table2.pdf]. 2. Krug EG, Sharma GK, Lozano R. The global burden of injuries. Am J Public Health 2000;90:523–6. 3. Sauaia A, Moore FA, Moore EE, et al. Epidemiology of trauma deaths: a reassessment. J Trauma 1995;38:185–93. 4. Brohi K, Cohen MJ, Davenport RA. Acute coagulopathy of trauma mechanism, identification and effect. Curr Opin Crit Care 2007;13:680–5. 5. Tieu BH, Holcomb JB, Schreiber MA. Coagulopathy: its pathophysiology and treatment in the injured patient. World J Surg 2007;31:1055–64. 6. Moore FA, Moore EE, Sauaia A. Blood transfusion. An independent risk factor for postinjury multiple organ failure. Arch Surg 1997;132:620–4. 7. Malone DL, Dunne J, Tracy JK, Putnam AT, Scalea TM, Napolitano LM. Blood transfusion, independent of shock severity, is associated with worse outcome in trauma. J Trauma 2003;54:898–905. 8. Parks JK, Elliott AC, Gentilello LM, Shafi S. Systemic hypotension is a late marker of shock after trauma: a validation study of advanced trauma life support principles in a large national sample. Am J Surg 2006;192:727–31. 9. American College of Surgeons. Resources for the optimal care of the injured patient. Chicago, IL: American College of Surgeons; 2006.
R.M. Hasler et al. / Resuscitation 82 (2011) 1202–1207 10. Rossaint R, Bouillon B, Cerny V, et al. Management of bleeding following major trauma: an updated European guideline. Crit Care 2010;14:R52. 11. Antonelli M, Levy M, Andrews PJ, et al. Hemodynamic monitoring in shock and implications for management. International Consensus Conference. Intensive Care Med 2007;33:575–90. 12. Bruns B, Gentilello L, Elliott A, Shafi S. Prehospital hypotension redefined. J Trauma 2008;65:1217–21. 13. Shires GT, Carrico CJ, Canizaro PC. In: Dunphy JE, editor. Shock, vol. 8. Philadelphia, PA: WB Saunders Company; 1973. 14. Lipsky AM, Gausche-Hill M, Henneman PL, et al. Prehospital hypotension is a predictor of the need for an emergent, therapeutic operation in trauma patients with normal systolic blood pressure in the emergency department. J Trauma 2006;61:1228–33. 15. Eastridge BJ, Salinas J, McManus JG, et al. Hypotension begins at 110 mm Hg: redefining “hypotension” with data. J Trauma 2007;63:291–7. 16. The Trauma Audit & Research Network. The trauma audit & research network. Procedures Manual. Salford, UK: The Trauma Audit & Research Network; 2009. 17. Committee on Injury Scaling. Association for the advancement of automotive medicine (AAAM). The abbreviated injury scale 2005 revision. Des Plaines, Chicago: AAAM; 2005. 18. Guly HR, Bouamra O, Little R, et al. Testing the validity of the ATLS classification of hypovolaemic shock. Resuscitation 2010;81:1142–7.
1207
19. McMahon CG, Kenny R, Bennett K, et al. The effect of acute traumatic brain injury on the performance of shock index. J Trauma 2010;69:1169–75. 20. The Trauma Audit and Research Network. Performance comparison. Eccles Old Road, Salford: Clinical Sciences Building, Hope Hospital [Accessed 31 August 2010 https://www.tarn.ac.uk/Content.aspx?ca=15]. 21. Sterne JA, White IR, Carlin JB, et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ 2009;338:b2393. 22. Goldstein B, Toweill D, Lai S, Sonnenthal K, Kimberly B. Uncoupling of the autonomic and cardiovascular systems in acute brain injury. Am J Physiol 1998;275:R1287–92. 23. McMahon CG, Kenny R, Bennett K, Kirkman E. Modification of acute cardiovascular homeostatic responses to hemorrhage following mild to moderate traumatic brain injury. Crit Care Med 2008;36:216–24. 24. Law MM, Hovda DA, Cryer HG. Fluid-percussion brain injury adversely affects control of vascular tone during hemorrhagic shock. Shock 1996;6:213–7. 25. Davis JW, Davis IC, Bennink LD, Bilello JF, Kaups KL, Parks SN. Are automated blood pressure measurements accurate in trauma patients? J Trauma 2003;55:860–3. 26. Shakur H, Roberts I, Bautista R, et al. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2): a randomised, placebo-controlled trial. Lancet 2010;376:23–32, 3.