Eur J Trauma Emerg Surg (2012) 38:3–9 DOI 10.1007/s00068-011-0168-4

ORIGINAL ARTICLE

Epidemiology of in-hospital trauma deaths R. Lefering • T. Paffrath • O. Bouamra T. J. Coats • M. Woodford • T. Jenks • A. Wafaisade • U. Nienaber • F. Lecky



Received: 21 October 2011 / Accepted: 23 November 2011 / Published online: 13 December 2011  Springer-Verlag 2011

Abstract Purpose About half of all trauma-related deaths occur after hospital admission. The present study tries to characterize trauma deaths according to the time of death, and, thereby, contributes to the discussion about factors considered as the cause of death. Methods Data from two large European trauma registries (Trauma Registry of the German Society of Trauma Surgery, TR-DGU, and the Trauma Audit and Research Network, TARN) were analyzed in parallel. All hospital deaths with Injury Severity Score (ISS) [ 9 documented between 2000 and 2010 were considered. Patients were categorized into five subgroups according to the time to death (0–6 h; 7–24 h; day 1–6; day 7–30; beyond day 30). Surviving

R. Lefering (&) Institute for Research in Operative Medicine, University Witten/ Herdecke, Ostmerheimer Str. 200, 51109 Cologne, Germany e-mail: [email protected] T. Paffrath  A. Wafaisade Department for Trauma Surgery and Orthopaedics, Cologne Merheim Medical Center, University Witten/Herdecke, Cologne, Germany O. Bouamra  M. Woodford  T. Jenks  F. Lecky Trauma Audit and Research Network, Health Sciences Research Group, School of Community Based Medicine, Manchester Medical Academic Health Sciences Centre, University of Manchester, Salford Royal Hospital, Salford, UK T. J. Coats Emergency Medicine Academic Group, University of Leicester, Leicester, UK U. Nienaber Akademie Unfallchirurgie (AUC) of the German Society for Trauma Surgery (DGU), Cologne, Germany

patients from the same time period served as a control group. Results In total, 6,685 and 6,867 non-survivors were included from the TR-DGU and TARN, respectively. The hospital mortality rate was between 15 and 17%. About half of all deaths occurred within the first 24 h after admission (TR-DGU: 54%; TARN: 45%). The earliest subgroup of trauma deaths showed the highest mean ISS and the highest rate of mass transfusions. Severe head injury was most frequently observed in the subgroup of day 1–6. Late deaths are associated with higher age and more complications (sepsis, multiple organ failure). Conclusions The time to death after severe trauma does not follow a trimodal distribution but shows a constantly decreasing incidence. Keywords Severe injuries  Trauma  Cause of deaths  Mortality  Trauma registry

Introduction About half of the patients who die from severe injuries do not reach a hospital; either they were already dead on arrival of emergency personnel and resuscitation was unsuccessful, or they quickly deteriorate, despite initial attempts to stabilize them. After admission to a hospital, the mortality rate is still high in severely injured patients, especially in the first hours after admission. However, the cause of death is sometimes not easy to determine, since there is usually a combination of factors which lead to death. Injury severity is, of course, the most important prognostic factor, where head injuries are the most critical. Furthermore, if injuries cause massive bleeding, then blood loss is also an

123

4

important contributor to death, despite volume replacement, massive transfusion, and immediate emergency surgery. When the patient survives the initial stabilization phase after trauma, the subsequent phase of intensive care is a real challenge for the patient. Frequently, organ function is compromised, and patients are at risk of developing infection and sepsis. Finally, age is an independent prognostic factor that plays an important role during the whole process of care. Based on the results of two large European trauma registries, the present paper evaluates non-survivors after severe trauma who reached a hospital but then died. Since the cause of death is not available, patients are grouped and analyzed according to their time of death. Differences between early and late deaths will help to understand the physiologic challenge caused by severe injuries.

Materials and methods The present analysis has been performed in parallel in the databases of two large European trauma registries, the Trauma Registry of the German Society of Trauma Surgery (TR-DGU; Deutsche Gesellschaft fu¨r Unfallchirurgie) and the Trauma Audit and Research Network (TARN) from the UK. The TR-DGU was established in 1993 as an instrument for the external comparison of the type and outcome of care for severely injured patients. Cases with potential need for intensive care who were admitted over the shock room were eligible for documentation. Patients who died in the prehospital area as well as those from burns, drowning, poisoning, or hanging were excluded. The documentation of cases consist of about 120 data points per patient covering the prehospital phase, the early in-hospital phase, intensive care, and treatment until discharge. All injuries are coded using the Abbreviated Injury Scale (AIS) 2005 version. Codes from the previous version (AIS-98) were matched appropriately. Organ failure was assessed using the Sequential Organ Failure Assessment (SOFA) score, where organ failure for each organ was defined if the SOFA score was 3 or 4 points [1]. Multiple organ failure was given if two or more organs failed. Pre-existing diseases were derived from a list of 16 different diseases (for details, see [2]). Since 2002, data are collected by an online documentation system with multiple plausibility and completeness checks. Participation has been voluntary until it recently became the obligatory tool for quality assessment in regional trauma networks certified by the DGU. Participating hospitals receive extensive annual audit reports. Actually, the TR-DGU collects data from about 400 hospitals, mostly from Germany, with an annual increase of 15,000 cases per

123

R. Lefering et al.

year [3]. The basis for the present analysis was the period from 2000 to 2010. The TARN cohort contains prospectively recorded patient data from its large European multicenter trauma register using a web-based data collection and reporting system. Eligible for inclusion in the TARN are all patients with trauma presenting to a participating hospital who either require hospital admission for 72 h or more; are transferred to a participating hospital for specialist care; require high-dependency care or intensive care unit treatment; or who die as a result of their injuries within 93 days of admission. The TARN does not include patients dead on arrival at the hospital, those transferred for rehabilitation only, patients with brain injury unrelated to trauma (e.g., spontaneous subarachnoid hemorrhage), simple skin lacerations, contusions or abrasions and minor penetrating injuries, patients with single uncomplicated limb injuries, or patients over 65 years of age with isolated fracture of the femoral neck or pubic ramus. Variables collected by the TARN include age, gender, systolic blood pressure (SBP), heart rate, respiratory rate, Glasgow Coma Scale (GCS) score, injury mechanism, and mortality at 30 days. Injury details are coded using the AIS, as are all patient intervention details during the acute phase. The TARN publishes quarterly reports of the observed and expected mortality rates for each center on the web-based system to allow hospitals to compare their performance with other trauma centers. For the present analysis, the following criteria were used to identify the patients in both databases: • • • •

Time period 2000–2010, Died in the acute care hospital, Date of admission and date of death available, Injury Severity Score (ISS) [ 9.

The day of death was defined as 0 if the patient died within 24 h after admission. The day of death was 1 if death occurred between 24 and 48 h after admission. If the time of admission and/or the time of death were missing, the day of death was defined as the difference between the date of admission and the date of death. Patients were grouped into one of the following subgroups based on the time of death: • • • • •

Died within 6 h after admission (0–6 h), Survived the first 6 h but died within 24 h (7–24 h), Survived the first 24 h but died before day 6 (day 1–6), Died between day 7 and day 30 (day 7–30), Died in hospital beyond day 30 (late).

If the time of death was missing, patients were grouped in the first subgroup (0–6 h) if the date of admission and date of death were identical. For reasons of comparisons, the group of surviving patients from the same period of

Epidemiology of in-hospital trauma deaths Fig. 1 Cumulative number of non-survivors for the first 48 h after hospital admission. Only patients from the Trauma Audit and Research Network (TARN) database with available time of admission and time of death are considered (n = 3,584)

5

4000

3500

3000

2500

2000

1500

1000

500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

0

Hour after admission

time with ISS [ 9 from both databases were used (TRDGU: n = 32,408; TARN: n = 37,091). The data were presented as means and percentages. The time to death was also presented as the median for different conditions (Table 5). More detailed information, such as the number of cases or measures of variation (standard deviation [SD], range) are omitted here in order to focus on the point estimates. Statistical testing was avoided, as the large number of cases would make even minor differences statistically significant. It was also not the intention of this paper to compare differences between both registries. A Kaplan–Meier curve was used to describe the time of death during the first 30 days.

Results From the TR-DGU, 6,685 non-survivors were included. In the same period, 32,408 survivors with ISS [ 9 were documented; this equates to a hospital mortality rate of 17.1%. From the TARN database, 6,867 non-survivors were included, 962 of them with missing time of death. Based on a total of 43,958 cases, the hospital mortality rate was 15.6%. The time until death in the first 48 h is presented in Fig. 1, and Fig. 2 presents a survival curve for the first 30 days. Between 37% (TR-DGU) and 38% (TARN) of deaths occurred within the first 6 h after admission (Fig. 3). Another 14–16% occurred within 24 h, so that about half of all fatalities in the hospital occurred within the first 24 h after admission.

40

38 37

Percentage of non-survivors

TR-DGU TARN

28

30

24 20

16

18 17 14

10

4

5

0 0-6 hours

7-24 hours

day 1-6

day 7-30

later

Time to death

Fig. 2 Kaplan–Meier curve for the first 30 days after admission; only data from the Trauma Registry of the German Society of Trauma Surgery (TR-DGU) are included (n = 38,761)

Fig. 3 Percentage of patients in the five time-to-death subgroups, separately presented for both databases. The selected time periods are not equal

123

6

R. Lefering et al.

Table 1 Demographic data from the non-survivors according to the time of death after hospital admission Time of death

Number of cases

Mean age (years)

Age C 60 years (%)

Males (%)

Pre-existing diseases (%)

Up to 6 h

2,536/2,163

50/45

40/30

66/72

19/21

Up to 24 h

1,065/847

54/52

48/43

63/67

26/36

Until day 7

1,622/1,634

54/59

48/54

69/65

32/48

Until day 30

1,211/980

57/67

52/69

70/61

41/59

Beyond day 30

251/281

60/71

60/75

76/56

51/65

Non-survivors

6,685/6,867

53/55

46/47

67/67

28/39

Survivors

19,036/37,091

40/41

22/22

73/73

26/33

The values presented refer to the TR-DGU and TARN, respectively; 962 patients from the TARN had missing time of death

Patients who die later than 1 week after the incident were considerably older and had more pre-existing diseases than those who died earlier (Table 1). While only about one in five surviving patients is older than 60 years of age, the rate in non-survivors is higher in general, and it specifically increased with the time to death. In the small group of very late deaths (beyond 1 month), the majority of cases are aged 60 years or above (Fig. 4). Injury severity is higher in early deaths and consistently decreases with time (Table 2). Also, penetrating trauma is found more frequently in early deaths. In contrast, the proportion of non-surviving patients with severe head injury is highest in the time period from day 1 to 7. Severe bleeding (as represented by low blood pressure and the number of blood units transfused) is mainly seen in patients who died early (within the first 24 h), and is most prevalent in cases who died within the first 6 h after admission (Table 3). Complications like organ failure and sepsis were mainly found in cases who died later in the course of treatment. Since a sepsis needs some time for development, it is not found in those patients who died early. The sepsis rate in

late deaths (TR-DGU 53%; TARN 22%) is found to be five to tenfold higher than in survivors (8 and 2%, respectively) (Table 4). The mean and median time to death in non-survivors with a specific condition is presented in Table 5. The

Table 2 Injury severity and type of trauma of non-survivors according to the time of death after hospital admission Time of death

Mean ISS

Head injury (AIS C 3) (%)

Penetrating trauma (%)

Up to 6 h

42/36

68/59

7/10

Up to 24 h

37/31

84/80

6/5

Until day 7 Until day 30

34/28 33/25

89/79 74/68

5/2 3/1

Beyond day 30

33/23

54/61

5/2

Non-survivors

37/30

76/69

5/5

Survivors

23/19

45/39

4/6

The values presented refer to the TR-DGU and TARN, respectively

Table 3 Blood pressure (BP) and transfusion requirements according to the time of death after hospital admission 100

age 60+

90 mass transf.

Time of death

BP prehospital

BP on admission

Number of pRBCa

80 sepsis

70

Percent

Mass transfusionb (%)

60

head injury

50

Up to 6 h

84/106

81/104

6/9

23/5

Up to 24 h

116/135

113/133

5/13

18/5

Until day 7

118/137

122/140

3/10

10/2

40

Until day 30

119/141

121/141

3/10

12/2

30

Beyond day 30

118/138

115/140

5/5

16/1

20

Non-survivors

106/129

105/129

5/9

16/3

10

Survivors

122/130

126/133

1/6

4/1

0

0-6 h

7-24 h

1-7 d

8-30 d

later

survivor

Fig. 4 Prevalence of risk factors in the five subgroups according to the time of death. The risk factors are age 60 years or above (age 60?), mass transfusion (mass transf.), sepsis, and severe head injury. Data from the TR-DGU are presented as solid lines and data from the TARN as dashed lines

123

The values presented refer to the TR-DGU and TARN, respectively a

Average number of units of packed red blood cells transfused until intensive care unit admission (TR-DGU) or during hospital stay (TARN)

b

TR-DGU: 10 or more units until intensive care unit admission; TARN: not derived from the number of units but assessed as a separate item

Epidemiology of in-hospital trauma deaths

7

Table 4 Complications (%) according to the time of death after hospital admission Time of death

Organ failure (%)

Multiple organ failure (%)

Sepsis (%)

Up to 6 h

38/12

27/–

0/0

Up to 24 h

79/9

54/–

1/0

Until day 7 Until day 30

86/9 86/15

58/– 68/–

8/7 29/18

Beyond day 30

87/17

76/–

53/22

Non-survivors

67/11

48/–

9/6

Survivors

33/2

19/–

8/2

The values presented refer to the TR-DGU and TARN, respectively

Table 5 Time to death for non-survivors with specific conditions Condition

n

Mean days

Median days

Mass transfusion

1,032/202

5.0/2.3

0/0

Head injury

5,092/4,073

4.9/6.5

1/1

Age C 60 years

3,073/2,778

6.5/9.8

1/3

Multiple organ failure

2,729/668

8.2/9.6

3/1

Sepsis

511/351

22.4/20.8

15/11

The values presented refer to the TR-DGU and TARN, respectively

condition leading most quickly to death was massive blood loss, with a median time to death within 24 h.

Discussion Survival is the ultimate initial goal in treating severely injured patients, in the prehospital setting as well as in the acute care hospital. Therefore, the focus on those cases who did not survive may help to understand the mechanisms involved in a positive or negative outcome. About 30 years ago, Trunkey described a trimodal distribution of trauma deaths [4]. The first peak is immediate deaths that occur within 1 h of the accident. About half of all trauma deaths are said to belong to this group; however, the data vary from 30 to 70% based on various factors, such as type of injury, region, or type of emergency medical system (EMS). Evans et al. [5] found that 66% of trauma deaths were prehospital deaths in Australia, Maegele et al. [6] found that 61% of patients with traumatic brain injury (TBI) died outside the hospital, and Demetriades et al. [7] found a 50% rate for immediate deaths in the county of Los Angeles. However, these very early deaths are usually not seen in hospital-based trauma registries, and the present analysis also could not contribute to this aspect.

The second (within 1–4 h after the accident) and third (around 1–2 weeks later) peaks of trauma deaths described by Trunkey has been repeatedly challenged by various authors [7, 8]. Our analysis also did not show such a third peak, but only a continuous and consistent decrease in mortality (Figs. 1 and 2). We also found that about half of all in-hospital trauma deaths occur within the first 24 h after admission. Within the first day, most deaths occur in the first 6 h after admission, 70% in the TR-DGU and 72% in the TARN. This is in the range of previously published observations of 60–80% [8, 9]. The cause of death is not directly analyzed here since it is not yet part of the documentation in the TR-DGU. It was felt that the determination of the cause of death is sometimes difficult, and, in many patients, there are multiple factors where each one may have contributed to the lethal outcome. Rather, we analyzed indirect variables that could suggest the contribution of one or more causative factors. For example, the initial blood pressure as well as the fact that the patient required massive blood transfusion are good indicators for the role of exsanguination. Our findings confirm that significant blood loss is among the very early lethal factors. More than half of all nonsurvivors who received a mass transfusion died within the first 24 h (median time to death is day 0). Another important finding is that the blood pressure was very low in trauma patients who died within the first 6 h, and it did not increase on average between the first pre-clinical assessment and hospital admission. In all other subgroups, the average blood pressure was either normal or could be improved by prehospital volume administration. Another very early lethal factor is the overall amount of tissue damage and injury severity, as expressed, for example, by the ISS score. Patients who died in the first 6 h had a substantially higher injury severity. The rate of penetrating trauma is also highest in this early subgroup. This result, which is more clearly observed in the TARN data, is in accordance with previous findings [10]. Head injury or TBI is also known to be among the important prognostic factors. It is found to be responsible for about half of all trauma deaths [8, 11]. In our data, the subgroup of patients who died until day 7 did show the highest rate of TBI patients, consistently in both registries. The median time to death is also day 1, i.e., beyond 24 h, in both databases. Although severe head injuries could lead to death very quickly, it is not the predominant cause of death within the first few hours. Also, Bansal et al. [9] found that hemorrhage is more prevalent in early deaths than TBI. The two factors that are responsible for late deaths are complications and age. Organ failure and sepsis are critical complications during the intensive care treatment of trauma patients. In both registries, the median and mean time until

123

8

death for patients with sepsis is about 2 weeks. However, this is not the time point when sepsis was first observed— these data are not available. It is the time point when patients died, probably from that complication. The final prognostic factor is age. Age is known to be an independent prognostic factor in all advanced outcome prediction models, such as the Trauma and Injury Severity Score (TRISS), Revised Injury Severity Classification (RISC), or Probability of Survival (PS04) [12–14]. Patients who die beyond day 30 as a consequence of their traumatic injuries are much older than patients who died earlier. Especially in the TARN dataset, these patients are 70 years old, on average, and more than three out of four non-survivors were older than 60 years of age. Injury severity in these cases is still higher than in survivors, but it is lowest in the considered subgroups of non-survivors. It could be assumed that older patients have an increasingly reduced capacity to recover. Physical repair mechanisms may be less effective, or need more time, than in younger trauma patients. Besides this, there may also be a contribution from pre-existing diseases, which are also more prevalent in the elderly. These diseases could be considered as an additional burden for the organism and may complicate recovery [2]. Modern intensive care strategies are able to stabilize a patient by the substitution of specific organ function (e.g., ventilation, dialysis), nutrition, medication, and nursing support. While many patients benefit from this support and manage to recover after a long intensive care unit and hospital stay, it may also prolong the time to death in some elderly patients. Limitations This analysis is based on data from two different trauma registries in Europe. Registry data in general have to be considered as less valid than data from prospective clinical trials. However, both registries have existed for many decades and have, independently from each other, developed various strategies to improve data quality. Some variables considered here may differ between the two registries based on their definitions. Pre-existing diseases, for example, are assessed in a similar but not identical way. Organ failure and sepsis also require extensive definitions, which were different in both datasets. Massive transfusion after 24 h of admission is often not described in the TARN. Differences between the data from both registries should, therefore, be interpreted with caution. It is not the aim to demonstrate national differences in trauma care, but, rather, to identify similarities in non-survivors. It is, therefore, the trend within the time-to-death subgroups in both registries which is of the most importance. And if this trend is the same in both data banks, then this is a strong indicator of a general validity.

123

R. Lefering et al.

Conclusions Based on data from two large European trauma registries, we were not able to identify a trimodal distribution of trauma deaths. In-hospital trauma deaths occur with a decreasing incidence after admission, with about half of all fatalities occurring within the first 24 h. Injury severity and massive hemorrhage are the most prevalent among the very early deaths (up to 6 h after admission), while the highest rates of severe head injury are found in patients who died in the first week. Late deaths are frequently associated with complications like organ failure or sepsis, and increasing age. Conflict of interest

None.

References 1. Vincent JL, Moreno R, Takala J, Willatts S, De Mendonc¸a A, Bruining H, Reinhart CK, Suter PM, Thijs LG. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intens Care Med. 1996;22:707–10. 2. Wutzler S, Maegele M, Marzi I, Spanholtz T, Wafaisade A, Lefering R; Trauma Registry of the German Society for Trauma Surgery. Association of preexisting medical conditions with inhospital mortality in multiple-trauma patients. J Am Coll Surg. 2009;209:75–81. 3. Ruchholtz S, Lefering R, Paffrath T, Oestern HJ, Neugebauer E, Nast-Kolb D, Pape HC, Bouillon B. Reduction in mortality of ¨ rztebl Int. severely injured patients in Germany. Dtsch A 2008;105:225–31. 4. Trunkey DD. Trauma. Accidental and intentional injuries account for more years of life lost in the U.S. than cancer and heart disease. Among the prescribed remedies are improved preventive efforts, speedier surgery and further research. Sci Am. 1983;249:28–35. 5. Evans JA, van Wessem KJP, McDougall D, Lee KA, Lyons T, Balogh ZJ. Epidemiology of traumatic deaths: comprehensive population-based assessment. World J Surg. 2010;34:158–63. 6. Maegele M, Engel D, Bouillon B, Lefering R, Fach H, Raum M, Buchheister B, Schaefer U, Klug N, Neugebauer E. Incidence and outcome of traumatic brain injury in an urban area in Western Europe over 10 years. Eur Surg Res. 2007;39:372–9. 7. Demetriades D, Murray J, Charalambides K, Alo K, Velmahos G, Rhee P, Chan L. Trauma fatalities: time and location of hospital deaths. J Am Coll Surg. 2004;198:20–6. 8. de Knegt C, Meylaerts SAG, Leenen LPH. Applicability of the trimodal distribution of trauma deaths in a Level I trauma centre in the Netherlands with a population of mainly blunt trauma. Injury. 2008;39:993–1000. 9. Bansal V, Fortlage D, Lee JG, Constantini T, Potenza B, Coimbra R. Hemorrhage is more prevalent than brain injury in early trauma deaths: the golden six hours. Eur J Trauma Emerg Surg. 2009;35:26–30. 10. Demetriades D, Kimbrell B, Salim A, Velmahos G, Rhee P, Preston C, Gruzinski G, Chan L. Trauma deaths in a mature urban trauma system: is ‘‘trimodal’’ distribution a valid concept? J Am Coll Surg. 2005;201:343–8.

Epidemiology of in-hospital trauma deaths 11. Dutton RP, Stansbury LG, Leone S, Kramer E, Hess JR, Scalea TM. Trauma mortality in mature trauma systems: are we doing better? An analysis of trauma mortality patterns, 1997–2008. J Trauma. 2010;69:620–6. 12. Lefering R. Development and validation of the Revised Injury Severity Classification (RISC) score for severely injured patients. Eur J Trauma Emerg Surg. 2009;35:437–47.

9 13. Champion HR, Copes WS, Sacco WJ, Lawnick MM, Keast SL, Bain LW Jr, Flanagan ME, Frey CF. The Major Trauma Outcome Study: establishing national norms for trauma care. J Trauma. 1990;30:1356–65. 14. Bouamra O, Wrotchford A, Hollis S, Vail A, Woodford M, Lecky F. A new approach to outcome prediction in trauma: a comparison with the TRISS model. J Trauma. 2006;61:701–10.

123

Epidemiology of in-hospital trauma deaths.

About half of all trauma-related deaths occur after hospital admission. The present study tries to characterize trauma deaths according to the time of...
314KB Sizes 1 Downloads 11 Views