Prediction of prolonged ventilatory support in blunt thoracic trauma patients

Intensive Care Med (2003) 29:1101 1105 DOI 10.1007/s00134-003-1813-0 ORIGINAL Ioanna Dimopoulou Anastasia Anthi Michalis Lignos Efstratios Boukouvalas Evangelos Evangelou Christina Routsi Konstantinos Mandragos Charis Roussos Prediction of prolonged ventilatory support in blunt thoracic trauma patients Received: 27 August 2002 Accepted: 17 April 2003 Published online: 12 June 2003 Springer-Verlag 2003 I. Dimopoulou M. Lignos E. Evangelou C. Routsi C. Roussos Department of Critical Care Medicine, Evangelismos Hospital, Athens, Greece A. Anthi E. Boukouvalas K. Mandragos Intensive Care Unit, Hellenic Red Cross Hospital, Medical School, National and Kapodistrian University of Athens, Athens, Greece I. Dimopoulou ( ) 2 Pesmazoglou Street, 14 561 Kifissia, Athens, Greece e-mail: idimo@otenet.gr Tel.: +32-10-6200663 Fax: +32-10-6202939 Abstract Objective: To identify predictors of prolonged (>7 days) mechanical ventilation (MV) in patients with blunt thoracic trauma. Design: Prospective analysis of consecutive patients. Setting: Adult intensive care unit (ICU) in a teaching, tertiary-care hospital. Patients and participants: Sixty-nine patients (53 men, 16 women) with thoracic trauma having a median age of 35 (range 17 85) years and a median injury severity score (ISS) of 29 (range 14 41) were enrolled in the present study. Associated injuries included head neck (77%), extremities (72%), external (67%), abdomen pelvis (67%), and face (55%). Interventions: Patient surveillance and data collection. Measurements and results: Thirty-three (48%) of the 69 patients required prolonged ventilatory support, ranging in duration from 8 to 38 (median 18) days. Logistic regression analysis revealed that advancing age (odds ratio=1.04, p=0.04), severity of head injury (odds ratio=1.92, p=0.008), and bilateral thoracic injuries (odds ratio=12.80, p<0.0001) were significant and independent predictors of long-lasting MV. In contrast, gender, injuries affecting the other body regions (face, abdomen pelvis, extremities, and external), laparotomy in patients with abdominal injury, or PaO 2 /FIO 2 on admission in the ICU, were unrelated to prolonged MV. Conclusions: In thoracic trauma patients admitted in the ICU, prolonged mechanical ventilation was primarily determined by presence of bilateral chest injuries, age, and degree of neurotrauma. This information may help in planning the long-term care of such patients. Keywords Thoracic trauma Mechanical ventilation Duration Prediction Bilateral chest injuries Severe head trauma Aging Introduction The importance of thoracic injury is well recognized. It is involved in nearly one third of acute admissions to trauma centers and is the second most common cause of death after head injury, accounting for approximately 25% of all trauma-related deaths. In the civilian population, blunt chest trauma largely exceeds penetrating trauma [1]. Mechanical ventilation remains the mainstay of livesaving therapy in patients with severe thoracic trauma. Clinical experience suggests that a subset of these patients require prolonged ventilatory support. Predictive criteria for long-lasting mechanical ventilation (MV) in the polytraumatized critically ill have been extensively investigated [2, 3, 4, 5], but no study has specifically addressed this issue in thoracic trauma patients. Early identification of patients who may stay for longer periods under artificial ventilation can be useful in many ways. Important therapeutic decisions can be made, including earlier placement of tracheostomy, which has been proven beneficial in trauma patients [6,

1102 7]. Some studies have inferred that prolonged intensive care unit (ICU) stay is associated with a substantial inhospital death rate [8, 9] and with poor functional outcomes in many patients who do survive [9]. Furthermore, such patients consume a disproportionate amount of nursing and financial resources [10]. Thus, prediction of patients requiring long-lasting MV early after the initiation of ventilatory support also has important consequences for informing families on a realistic basis about expectations and evaluating ICU resources. Given these considerations, this prospective, descriptive study was designed to isolate risk factors that might predict long-term mechanical ventilation in adult blunt thoracic trauma patients. Materials and methods Patients All consecutive patients with blunt thoracic trauma admitted to the ICU of the Hellenic Red Cross Hospital between January 2000 and December 2001 were prospectively evaluated. Patients meeting the following criteria were included in this study: (1) intubation and placement on mechanical ventilation within the first 24 h of injury; (2) survival for more than 48 h posttrauma. The Institutional Review Board approved the study, and informed consent was obtained from next-of-kin. Study protocol The following data were collected for each patient: age, sex, mechanism of injury, thoracic along with extrathoracic injuries, total time spent on MV, length of ICU stay, surgical therapy including laparotomy in patients with abdominal trauma, ratio of arterial blood oxygen tension to the concentration of inspired oxygen (PaO 2 /FIO 2 ) on admission to the ICU, and mortality. The abbreviated injury scale (AIS) was evaluated as the measure of anatomic injury for six body regions: (1) head neck, (2) face, (3) thorax, (4) abdomen pelvis, (5) extremities, and (6) external [11]. The injury severity score (ISS) was calculated as the sum of the squares of the highest AIS grade in each of the three most severely injured body regions [12]. Brain, neck, thoracic, and abdominal injuries were assessed by computed tomography scanning (CT scan) on admission. Certain complications, including pneumonia and adult respiratory distress syndrome (ARDS), were recorded. Pneumonia was diagnosed by a new and persistent infiltration on chest X-ray, a body-core temperature >38 C, and purulent tracheal secretions with evidence of many neutrophils and bacteria. ARDS was defined according to the American-European Consensus Conference on ARDS as PaO 2 /FIO 2 ratio <200 with bilateral infiltrates on chest X-ray with no evidence of congestive heart failure [13]. Prolonged MV was defined as the need for MV support of more than 7 days [5, 8]. For analysis, patients were subdivided in two groups based solely on duration of MV. Group 1 included patients who remained on the ventilator for up to 7 days, and group 2 comprised all patients who required MV for more than 8 days. Statistical analysis The observed data did not follow the normal distribution (categorized Kolmogorov-Smirnov test). Nonparametric statistical tests were used, and results are presented as medians and ranges. Comparisons between groups were performed using Mann-Whitney ranked sum test, chi square analysis, or Fisher exact test, where appropriate. Logistic regression analysis was carried out to examine the association between prolonged MV and the following possible risk factors: gender, age, PaO 2 /FIO 2 on admission to the ICU, AIS for five anatomic regions of the body (head neck, face, abdomen pelvis, extremities, external), performance of laparotomy in patients with abdominal trauma, and thoracic injuries classified in two groups one group with any unilateral thoracic injury (lung contusion, hemothorax, pneumothorax, or rib fractures), and a second group having any bilateral thoracic injury. All factors were included in the model, and nonsignificant variables were deleted by backward elimination (deletion criterion; p>0.05). The odds ratios and the 95% confidence intervals were calculated. Statistical significance was determined at the p<0.05 level. The SAS statistical package was used for all analyses (SAS Institute Inc., Cary, NC, USA). Results During the study period, 79 patients with thoracic trauma and other associated injuries were admitted to the ICU of the Hellenic Red Cross Hospital. Of these, 15 patients (19%) died in the ICU. Early deaths (within 48 h postinjury) were recorded in ten patients and were related to head trauma (n=6), major thoracic aorta injury (n=2), or massive bleeding (n=2). These patients were excluded from further analysis. Late deaths (4 28 days after injury) occurred in five patients due to sepsis/multiple organ failure (n=3) or severe brain injury (n=2). Thus, 69 patients were included in the current study. All injuries were caused by blunt trauma related to motor vehicle crashes. Demographic and clinical characteristics of the 69 patients are shown in Table 1. No patient had isolated thoracic trauma, and the most common concomitant injuries were brain trauma along with bone fractures, occurring in more than 70% of patients. Sixtyfive patients (94%) were severely injured (ISS >16). The most common thoracic injuries were lung contusions, rib fractures, and hemothorax (in more than 60% of patients), while pneumothorax and flail chest were less common (Table 2). No patient sustained sternal fracture, myocardial contusion, or rupture of the diaphragm, great vessels, or tracheobronchial tree. No patient underwent thoracotomy, whereas in 15 patients with abdominal injury laparotomy was performed. Analgesia was common in all patients and included the continuous intravenous infusion of fentanyl. In the entire patient population, median PaO 2 /FIO 2 on admission in the ICU was 200, ranging from 80 to 460. Median duration of MV and ICU stay were 7 (range 1 40) days and 18 (range 3 50 ) days respectively. Overall, duration of MV lasted from 1 to 7 days in 36 patients (group 1), whereas 33 patients remained on the ventilator for more than 8 (range 8 38) days (group 2). Associated injuries in both groups included: head neck (22 patients versus 31 patients), face (16 patients versus

1103 Table 1 Demographic and clinical characteristics of the 69 thoracic trauma patients Variable Median or number (%) (Range) Age (years) 35 (17 85) Gender M 53 F 16 Type of injury Head/Neck 53 (77) Face 38 (55) Thorax 69 (100) Abdomen pelvis 46 (67) Extremities 50 (72) External 46 (67) ISS 29 (14 41) Distribution of ISS 1 15 4 16 30 44 31 41 21 AIS Head neck 3 Face 2 Thorax 4 Abdomen pelvis 3 Extremities 2.5 External 2 ISS injury severity score AIS abbreviated injury scale Table 2 Details of thoracic injuries in the entire patient population (n=69) Injury Lung contusion 58 (84) Unilateral 22 Bilateral 36 Rib fractures 55 (80) Unilateral 47 Bilateral 8 Hemothorax 44 (64) Unilateral 30 Bilateral 14 Pneumothorax 24 (35) Unilateral 18 Bilateral 6 Flail chest 13 (19) Number (percent) 22 patients), abdomen pelvis (21 patients versus 25 patients), extremities (29 patients versus 21 patients), and external (23 patients versus 23 patients). The two groups were similar with regard to type of thoracic injuries (contusion, rib fractures, or hemo/pneumothorax). Demographic and clinical characteristics of the two groups are shown in Table 3. Table 3 Comparison of demographic and clinical characteristics between group 1 and group 2 patients Group 1 Group 2 P value MV <7 days MV > 8 days (n=36) (n=33) Age (years) 30 37 0.01 Gender M 30 23 0.29 F 6 10 ISS 22 33 <0.001 AIS Head neck 2 3 0.36 Face 2 1 0.17 Thorax 3 4 <0.001 Abdomen pelvis 3 3 0.51 Extremities 3 2 0.69 External 2 2 0.11 Bilateral TI, n 10 28 <0.001 PaO 2 /FIO 2 285 136 <0.001 MV duration, days 4 18 <0.001 ICU duration, days 10 28 <0.001 Pneumonia 2 19 <0.001 ARDS 1 13 <0.001 MV mechanical ventilation ISS injury severity score AIS abbreviated injury scale PaO 2 /FIO 2 ratio of arterial blood oxygen tension to the concentration of inspired oxygen ARDS adult respiratory distress syndrome Table 4 Results of logistic regression analysis of factors predicting prolonged mechanical ventilation (MV) in 69 thoracic trauma patients. (Final model resulting from backward elimination procedure (r 2 =0.35) Variables Odds ratio 95% P value Confidence interval Age 1.04 1.00 1.08 0.04 AIS head 1.92 1.18 3.11 0.008 Bilateral TI 12.80 3.36 48.73 <0.0001 AIS abbreviated injury scale TI thoracic injury Group 2 patients were older, had a higher ISS and AIS for the thorax, and a lower PaO 2 /FIO 2 ratio on admission to the ICU compared to group 1 patients. In group 2, duration of mechanical ventilation and ICU stay were significantly longer. Furthermore, the incidence of bilateral thoracic injuries, pneumonia, and ARDS was higher in group 2. Logistic regression analysis revealed that, among the potential risk factors studied, advancing age and severity of head and bilateral thoracic injuries were significant and independent predictors of prolonged MV (Table 4). More

1104 specifically, it was found that the probability of prolonged MV increased by 4% for a 1-year increase in age and by two times per 1 unit increase in head AIS. Furthermore, the probability of long-lasting MV increased by almost 13 times if bilateral thoracic injuries were present. In contrast, gender, injuries affecting other body regions (face, abdomen pelvis, extremities, and external), performance of laparotomy in patients with abdominal injury, PaO 2 /FIO 2 on admission to the ICU, and any unilateral chest injury, were unrelated to prolonged MV. Discussion The aim of this prospective study was to predict prolonged MV following thoracic injury by objective data or simple derived variables, which can be available early on admission in organized trauma centers. Advancing age and degree of neurotrauma, along with bilateral thoracic injuries, were found to be independently associated with long-lasting artificial ventilation. A plethora of research has been carried out on thoracic trauma patients, with the vast majority of the available studies focusing on factors affecting outcome in terms of morbidity or mortality rates [14, 15, 16, 17, 18, 19, 20, 21]. In contrast, predictors of long-lasting MV in these patients have not been well characterized. No consensus exists on how to define prolonged MV in critical illness, and any time from 7 to 21 days has been occasionally cited as prolonged [3, 4, 5, 8, 22]. On the other hand, it has been stressed that a period longer than 7 days of oral or nasal endotracheal intubation significantly increases the risks for laryngotracheal pathology [23, 24]. Thus, in this report, 7 days of ventilatory support was selected as the cutoff for defining long-lasting assisted ventilation [5, 8], and it was found that about 50% of our patients fell into this category. The impact of increasing age on duration of MV can be attributed to several factors, such as comorbid diseases, medications for chronic illness, decreased cardiopulmonary reserve, poor nutritional status, or preexisting limitations in mobility. All these can diminish the ability of older patients to respond to injury. For instance, it has been demonstrated that the presence of one premorbid illness in trauma patients increased the likelihood of death by 30%, and that two or more preexisting conditions increased the death rate by 60% [25]. Blunt high-energy trauma is frequently associated with extrathoracic injuries [14, 15, 17, 21]. Similarly in our series, 67% of patients sustained abdominal trauma and, in addition, bone fractures or head injury were present in more than 70%. The importance of neurotrauma on outcome in the critically injured has been previously emphasized. In chest trauma, severe brain injury as expressed by the Glasgow Coma Scale (CGS) score seems to be the main cause of death [15]. Similarly in polytrauma a low initial GCS score has been associated with a need for prolonged MV [4] and a higher mortality [20]. Although the GCS score is more objective in assessing head trauma severity, in the current study we selected the anatomical injury severity scoring system [17], because a certain number of our patients were intubated at the site of the accident. They were therefore sedated or even paralyzed upon admission to the hospital. In such patients, head AIS is eventually more accurate in ranking brain injury severity compared to the GCS score. By using this scoring system, our data confirmed the widely reported view in the literature that degree of neurotrauma negatively affects outcome in severe physical injury. In this study, type and distribution of thoracic injuries were assessed on the basis of chest CT scan, which has proven superior to radiograph in visualizing thoracic injuries, including hemothorax, pneumothorax, contusions, and diaphragmatic lesions [26]. Bilateral thoracic injuries were found to be associated with a 13-fold higher probability of prolonged MV. This striking finding is indirectly supported by a recent study showing that, in comparison to unilateral chest injuries, bilateral involvement carries a significantly higher mortality rate in patients with thoracic trauma and associated injuries [27]. Presumably bilateral thoracic injuries induce profound derangements in pulmonary, chest wall, and diaphragmatic mechanics, which delay the weaning process. Pulmonary complications including pneumonia and ARDS may occur following thoracic trauma [17, 19] and, as our data point out, are much more common in patients with long-lasting artificial ventilation. Both conditions further prolong duration of assisted ventilation. The appropriate timing of tracheostomy in the critically injured patient is poorly defined. The consensus conference on artificial airways has concluded that the decision to convert endotracheal intubation to tracheostomy should be made as soon as possible in the course of management to minimize the duration of translaryngeal intubation [22]. Indeed, early tracheostomy has been shown to be beneficial in multiple trauma patients. A randomized clinical trial enrolled 106 trauma patients to early (7 days or fewer) or delayed (>8 days) tracheostomy. Mean duration of MV (12 versus 32 days), ICU (16 versus 37 days) and hospital stay (34 versus 51 days) were significantly lower in the group with early tracheostomy. Furthermore, these patients experienced a lower incidence of pneumonia (78 versus 96%), particularly if tracheostomy was performed within the first 2 days of admission to the ICU [6]. Similarly a retrospective review of patients with blunt multiple injuries showed that tracheostomy within the first 4 days of injury helped in early weaning from the ventilator by reducing the incidence of nosocomial pneumonia [7]. Limitations of the present study should be acknowledged. First, our patients had chest trauma and other associated extrathoracic injuries; therefore, the findings

1105 cannot be extrapolated to patients with isolated thoracic trauma. In addition, this investigation does not have adequate power to examine the impact of oxygenation status in predicting prolonged ventilatory support; to better clarify this issue serial measurements of PaO 2 /FIO 2 would be more appropriate. In summary, we found that in thoracic trauma patients bilateral chest injuries, advancing age, and concomitant severe neurotrauma predict independently long-lasting MV. Patients fulfilling these criteria should be considered for early tracheostomy to prevent the complications of prolonged translaryngeal intubation. References 1. Lewis FR (1982) Thoracic trauma. Surg Clin North Am 62:97 104 2. Troche G, Moine P (1997) Is the duration of mechanical ventilation predictable? Chest 112:745 751 3. Johnson SB, Kearney PA, Barker DE (1992) Early criteria predictive of prolonged mechanical ventilation. J Trauma 33:95 100 4. Ross BJ, Barker DE, Russel WL, Burns RP (1996) Prediction of long term ventilatory support in trauma patients. Am Surg 62:19 25 5. Velmahos GC, Belzberg H, Chan L, Avari S, Cornwell EE, Berne TV, Asensio J, Murray J, Demetriadis D (1997) Factors predicting prolonged mechanical ventilation in critically injured patients: introducing a simplified quantitative risk score. Am Surg 63:811 817 6. Rodriguez JL, Steinberg SM, Luchetti FA, Gibbons KJ, Taheri PA, Flint LM (1990) Early tracheostomy for primary airway management in the surgical critical care setting. Surgery 108:655 659 7. Lesnik I, Rappaport W, Fulginiti J, Witze D (1992) The role of early tracheostomy in blunt, multiple organ trauma. Am Surg 58:346 349 8. Lipsett PA, Swoboda SM, Dickerson J, Ylitalo M, Gordon T, Breslow M, Campbell K, Dorman T, Pronovost P, Rosenfeld B (2000) Survival and functional outcome after prolonged Intensive Care Unit stay. Ann Surg 231:262 268 9. Spicher JE, White DP (1987) Outcome and function following prolonged mechanical ventilation. Arch Intern Med 147:421 425 10. Miller RS, Patton M, Graham RM, Hollins D (2000) Outcomes of trauma patients who survive prolonged lengths of stay in the Intensive Care Unit. J Trauma 48:229 234 11. Greenspan L, McLellan BA, Greig H (1985) Abbreviated Injury Scale and Injury Severity Score: a scoring chart. J Trauma 25:60 64 12. Baker SP, O' Neill B, Haddon W, Long WB (1974) The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 14:187 196 13. The American-European consensus conference on ARDS, part II: ventilatory, pharmacologic, supportive therapy, study design strategies and issues related to recovery and remodelling (1998) Intensive Care Med 24:378 398 14. Pinilla JC (1982) Acute respiratory failure in severe blunt chest trauma. J Trauma 22:221 226 15. Schulpen TM, Doesburg WH, Lemmens WAJ, Gerritsen SM (1986) Epidemiology and prognostic signs of chest injury patients. Injury 17:305 308 16. Johnson JA, Cogbill TH, Winga ER (1986) Determinants of outcome after pulmonary contusion. J Trauma 26:695 697 17. Clark GC, Schecter WP, Trunkey DD (1988) Variables affecting outcome in blunt chest trauma: flail chest vs. pulmonary contusion. J Trauma 28:298 304 18. Gaillard M, Herve C, Mandin L, Raynaud P (1990) Mortality prognostic factors in chest injury. J Trauma 30:93 96 19. Wisner DH (1990) A stepwise logistic regression analysis of factors affecting morbidity and mortality after thoracic trauma: effect of epidural analgesia. J Trauma 30:799 804 20. Stellin G (1991) Survival in trauma victims with pulmonary contusion. Am Surg 57:780 784 21. Rashid MA, Wikstrom T, Ortenwall P (2000) Outcome of lung trauma. Eur J Surg 166:22 28 22. Consensus conference on artificial airways in patients receiving mechanical ventilation (1989) Chest 96:178 180 23. Gaynor EB, Greenberg SB (1985) Untoward sequelae of prolonged intubation. Laryngoscope 95:1461 1467 24. Whited RE (1984) A prospective study of laryngotracheal sequelae in longterm intubation. Laryngoscope 94:367 377 25. MacKenzie EJ, Morris JA, Edelstein SL (1989) Effect of pre-existing disease on length of hospital stay in trauma patients. J Trauma 29:757 764 26. Guerrero-Lopez F, Vazquez-Mata G, Alcazar-Romero PP, Fernandez- Mondejar E, Aguayo-Hoyos E, Linde-Valverde CM (2000) Evaluation of the utility of computed tomography in the initial assessment of the critical care patient with chest trauma. Crit Care Med 28:1370 1375 27. Pape HC, Remmers D, Rice J, Ebisch M, Krettek C, Tscherne H (2000) Appraisal of early evaluation of blunt chest trauma: development of a standardized scoring system for initial clinical decision making. J Trauma 49:496 504

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.