Damage control in severely injured trauma patients - A ten-year experience.

BACKGROUND: This study reviews our 10-year institutional experience with damage control management and investigates risk factors for early mortality.
MATERIALS AND METHODS: The trauma registry of our level I trauma centre was utilized to identify all patients from 01/96 through 12/05 who underwent initial damage control procedures. Demographics, clinical and physiological parameters, and outcomes were abstracted. Patients were categorized as either early survivors (surviving the first 72 hours after admission) or early deaths.
RESULTS: During the study period, 319 patients underwent damage control management. Overall, 52 patients (16.3%) died (early deaths) and 267 patients (83.7%) survived the first 72 hours (early survivors). Early deaths showed significantly deranged serum lactate (5.81±0.55 vs. 3.46±0.13 mmol/L; P<0.001), base deficit (10.10±0.95 vs. 4.90±0.28 mmol/L; P<0.001) and pH (7.16±0.03 vs. 7.29±0.01; P<0.001) levels compared to early survivors on hospital admission. An International Normalized Ratio >1.2, base deficit >3 mmol/L, head Abbreviated Injury Scale ≥3, body temperature <35°C, serum lactate >6 mmol/L, and hemoglobin <7 g/dL proved to be independent risk factors for early mortality on hospital admission.
CONCLUSIONS: Several risk factors for early mortality such as severe head injury and the lethal triad (coagulopathy, acidosis and hypothermia) in patients undergoing damage control procedures were identified and should trigger the trauma surgeon to maintain aggressive resuscitation in the intensive care unit.

INTRODUCTION

Damage control (DC) procedures for patients with major truncal or extremity trauma have evolved over the past 20 years and are recognized as major factors in decreasing morbidity and mortality in severely injured patients.[1] Early physiologic criteria, such as hypothermia, coagulopathy, and acidosis are used to select patients that would benefit from DC techniques. Injury severity scoring or the estimated time required to accomplish definitive repair of injuries are further useful parameters in determining the need for DC.[2]

Numerous studies have reported indications to perform DC operations in severely injured trauma patients.[2-6] However, investigations evaluating predictors of poor outcome among the subset of trauma patients undergoing DC procedures are scarce. Thus, the objectives of this study were to review our institutional experience with DC management and to investigate risk factors for early mortality in patients undergoing DC procedures.

MATERIALS AND METHODS

After approval by the Institutional Review Board, the trauma registry of the Division of Trauma Surgery, University Hospital of Zurich, a verified Level I trauma centre, was reviewed to identify all severely injured patients [Injury Severity Score (ISS) ≥17] from January 1, 1996, through December 31, 2005 who underwent initial DC procedures. A DC management was defined by limited operations for control of hemorrhage and contamination, or temporary fixation of extremity injuries. This included management of solid organ injuries by packing, resection of gastrointestinal tract injuries without re-anastomosis, use of temporary closure techniques at a site of surgical exploration, or temporary stabilization of extremity and pelvic fractures.

Demographic and clinical information collected included age, sex, mechanism of injury (blunt vs. penetrating), systolic blood pressure, Glasgow Coma Score (GCS) upon admission, ISS, and Abbreviated Injury Scale (AIS) for each body area (head, chest, abdomen, and extremity). For the analysis, continuous variables were dichotomized using clinically relevant cut-points: GCS on admission (≤8 vs. >8), systolic blood pressure (SBP) on admission (≤90 mmHg vs. >90 mmHg) and ISS (≥25 vs. <25). Laboratory parameters including hemoglobin, hematocrit, lactate, base deficit (BD), and pH at hospital and intensive care unit (ICU) admission were also recorded.

Statistical analysis

Patients were categorized as early survivors (surviving the first 72 hours after admission) or early deaths (death within 72 hours after admission). The demographic and clinical characteristics between early survivors and early deaths were evaluated using bivariate analysis. The P values for categorical variables were derived from the Chi-square or two-sided Fisher's Exact test and continuous variables were evaluated with the Student's t-test or the Mann-Whitney test. Differences were considered significant at P<0.05.

To identify independent risk factors for early mortality, potential risk factors were examined by bivariate analysis. To explore critical thresholds for physiological and laboratory parameters associated with early mortality, a cut-off analysis was performed using repeated multivariate analysis. The highest R2 value defined the best cut-off. All risk factors with P<0.2 were subsequently entered in a stepwise forward logistic regression model.

Values are reported as mean±standard error of the mean (SEM) for continuous variables and as percentages for categorical variables. All analyses were performed using the Statistical Package for Social Sciences (SPSS Windows©), version 12.0 (SPSS Inc., Chicago, IL).

RESULTS

During the 10-year study period, 1,144 patients with an ISS ≥17 were admitted to our trauma centre. Of those, 319 patients (27.9%) underwent initial DC procedures and constituted our study cohort mean age overall was 39.3±1.0 years, 71.8% were male and the mean ISS was 36.6±0.7. On admission, 8.7% (n=26) of the patients were hypotensive (SBP <90 mmHg), and 55.7% (n=176) had a GCS ≤8. Blunt trauma was responsible for the majority of injuries (n=291, 91.2%).

Overall, 74 patients (23.2%) died. Fifty-two patients (16.3%) died within the first 72 hours after admission (early deaths), with seven deaths occurring during the initial DC procedure. The cause of early deaths included hemorrhagic shock (n=18), head injury (n=28), and multiple organ failure (n=6). A total of 267 patients (83.7%) survived the first 72 hours (early survivors). Basic demographics and associated injuries of the two study groups are presented in Table 1. DC procedures performed are summarized in Table 2.

Table 1

Comparison of clinical and demographic characteristics of early survivors and early deaths

Table 2

Damage control procedures performed in 319 patients

Laboratory parameters including serum lactate and BD level, pH, International Normalized Ratio (INR), hemoglobin, hematocrit, and platelet counts at hospital and ICU admission are delineated in Table 3. Early deaths showed significantly deranged serum lactate, BD and pH levels compared to early survivors on hospital and ICU admission. However, while hemoglobin and hematocrit levels were significantly lower in patients dying within 72 hours post admission, these values were similar upon admission to the ICU in both groups.

Table 3

Comparison of laboratory parameters in early survivors and early deaths

Independent hospital admission predictors of early mortality in patients undergoing DC management

Various factors, potentially associated with early mortality in patients undergoing DC procedures, were identified using bivariate analysis. After stepwise logistic regression, an INR >1.2, BD >3 mmol/L, AIS head ≥3, body core temperature <35°C, lactate >6 mmol/L and hemoglobin <7 g/dL proved to be independent risk factors for early mortality (R2=0.453) [Table 4].

Table 4

Independent risk factors at hospital admission for early mortality in patients undergoing damage control management

Independent ICU admission predictors of early mortality in patients undergoing DC management

For those patients admitted to the ICU (n=312), an arterial lactate level >4 mmol/L on admission to ICU and >10 Units PRBC transfused until ICU admission were found to be independent predictors of early mortality (R2=0.373) [Table 5].

Table 5

Independent risk factors at ICU admission for early mortality in patients undergoing damage control management

Outcomes of early survivors

Intensive care unit length of stay for early survivors was 15.7±0.8 days and the mean hospital length of stay was 36.5±1.5 days. Overall, 59.6% (159 of 267) patients developed major in-hospital complications with the most frequent being sepsis (n=95, 35.6%), pneumonia (n=82, 30.7%), wound infection (n=64, 24.0%), intra-abdominal infection (n=17, 6.4%), and acute respiratory distress syndrome (ARDS) (n=12, 4.5%). Twenty-two early survivors (8.2%) died 22.5±5.3 days later due to multiple organ failure (n=16) or severe head injury (n=6).

DISCUSSION

Managing severely injured patients is a diagnostic and therapeutic challenge for every trauma team. During the last decade, several studies have shown that the DC concept for abdominal, thoracic, and orthopedic injuries improves outcome in blunt and penetrating trauma.[7-11] Johnson et al. recently demonstrated that the continued application of DC principles has led to improved survival in patients with penetrating abdominal injury by comparing a historical cohort with a current study population. In their study, overall survival improved from 58% to 90% with equivalent injury severity. They concluded that the early treatment of hypothermia and coagulopathy and the increasing experience with the open abdomen significantly contributed to improved survival.[7] For multiple injured patients with concomitant orthopedic trauma, it has been demonstrated that DC procedures significantly reduce the incidence of general systemic complications such as ARDS and multiple organ failure.[10] Likewise in our study cohort, although 84% of the 267 patients surviving the first 3 days presented with critical injuries (ISS ≥ 25), only 22 patients (8.2%) died in the later in-hospital course due to multiple organ failure or severe traumatic brain injury.

It is difficult to distinguish which patients qualify for DC management. Analysis of our data has identified several hospital and ICU admission parameters that predict early mortality in severely injured patients despite being managed with DC procedures. Among those variables, coagulopathy upon hospital admission was one of the strongest predictor of poor outcome. Early deaths presented with significantly deranged coagulation parameters including an elevated INR and lower platelet counts on hospital and ICU admission. An INR >1.2 on hospital admission represented the most important independent risk factor (adjusted odds ratio, 10.64) associated with early mortality. Likewise, in numerous previous studies, the development of coagulopathy in trauma patients has been associated with increased morbidity and a several-fold increased mortality.[12-19] In the study by MacLeod and colleagues, coagulation abnormalities upon admission, defined by a prothrombin time >14 seconds and a partial thromboplastin time >36 seconds, were independent predictors of in-hospital mortality with an adjusted odds ratio of 1.35 and 4.26, respectively.[13] In the analysis by Mitra et al., the presence of acute coagulopathy was independently associated with death within 24 hours of hospital admission (odds ratio, 8.77).[19] Additionally, coagulopathy is common following severe traumatic brain injury, mediated by the systemic release of tissue factor from the injured brain parenchyma, resulting in an unregulated activation of the extrinsic coagulation pathway.[1720-24] Talving et al. demonstrated that the development of traumatic brain injury-associated coagulopathy, defined by thrombocytopenia and/or elevated INR and/or activated thromboplastin time, is associated with longer ICU lengths of stay and an almost ten-fold increased risk of death.[17] Also, in a prospective study by Olson and colleagues, elevated fibrin degradation products and an increased consumptive coagulopathy at admission were identified as independent predictors of mortality after isolated head injury.[24] In our study, a significantly higher proportion of patients dying within the first 72 hours suffered severe traumatic brain injury compared to early survivors (71.2% vs. 43.1%, P<0.001), which may have further contributed to their impaired admission coagulation status and may explain their early detrimental outcome. In addition, the severity of brain injury (head AIS ≥3) itself remained an independent predictor of early mortality (Odds Ratio (OR) 4.27, 95% CI [1.55-11.76]).

In the present study, early deaths had significantly higher lactate and BD levels at hospital and ICU admission as compared to early survivors. Furthermore, an elevated BD >3 mmol/L, lactate level >6 mmol/L on hospital admission, and the inability to achieve a lactate level <4 mmol/L until ICU admission, indicating a decreased lactate clearance, were significant risk factors for early mortality in the current analysis. These findings are consistent with previously published results.[4525-35] In the study by Smith and co-workers, lactate, BD, and the combination of the two were good predictors of morbidity and mortality in critically ill patients.[35] Furthermore, a recent retrospective analysis of 95 surgical ICU patients documented a stepwise increase of hospital mortality with increasing time to normal lactate levels (≤2.0 mmol/L).[32] In patients with major vascular injuries, initial emergency department acid-base variables such as pH, BD, and lactate were strongly associated with outcome.[29] Thus, serum lactate and BD levels are useful for estimating and monitoring the adequacy of surgical and resuscitative therapy and provide a means of predicting mortality in severely injured trauma patients.

Hypothermia on admission, defined as a core temperature lower than 35°C, has repeatedly been demonstrated to be a predictor of poor outcomes in severely injured trauma patients.[36-45] In the present investigation, early deaths presented with significantly lower admission core temperature compared to early survivors. Moreover, admission hypothermia was independently associated with early mortality (adjusted OR, 3.68). In the study by Wang et al. admission hypothermia was likewise independently associated with an increased odds of death in both the full cohort of trauma patients (adjusted OR, 3.03) and the subset of patients with isolated severe head injury (adjusted OR, 2.21).[38] In another series of 71 severely injured trauma patients, a temperature below 32°C was associated with no survival.[36] Inaba et al. recently demonstrated that not only admission hypothermia, but also postoperative hypothermia after cavitary surgery represents an independent predictor of mortality. Compared to patients with a temperature of >35°C, the relative risk of death for patients with a temperature between 35°C and 33°C was 4.0, and that for patients with a temperature ≤33°C was 7.1.[46]

CONCLUSION

We identified several risk factors present upon hospital and ICU admission for early mortality such as severe head injury and the lethal triad (coagulopathy, acidosis and hypothermia). These predictors of poor outcome should trigger the trauma surgeon to maintain aggressive resuscitation. However, when surviving the first 72 hours after injury, survival in severely injured trauma patients managed with damage control protocol exceeded 90%.