Severely injured patients from explosions usually. Blast Lung Injury From an Explosion on a Civilian Bus*

Transcription

1 Blast Lung Injury From an Explosion on a Civilian Bus* Reuven Pizov, MD; Arieh Oppenheim-Eden, MD; Idit Matot, MD; Yoram G. Weiss, MD; Leonid A. Eidelman, MD; Avraham I. Rivkind, M.D; and Charles L. Sprung, MD, FCCP Objective: To assess clinical signs and management of primary blast lung injury (BLI) from explosions in an enclosed space and to propose a BLI severity scoring system. Design: Retrospective analysis. Patients: Fifteen patients with primary BLI resulting from explosions on two civilian buses in Results: Ten patients were extremely hypoxemic on admission (PaO 2 < 65 mm Hg with oxygen supplementation). Four patients remained severely hypoxemic (PaO 2 /fraction of inspired oxygen (FIO 2 ) ratio of < 60 mm Hg) after mechanical ventilation was established and pneumothoraces were drained. Initial chest radiographs revealed bilateral lung opacities of various sizes in 12 patients (80%). Seven patients (47%) had bilateral pneumothoraces and two patients had a unilateral pneumothorax. Five (33%) had clinically significant bronchopleural fistulae. After clinical and laboratory data were collected, a BLI severity score was defined based on hypoxemia (PaO 2 /FIO 2 ratio), chest radiographic abnormalities, and barotrauma. Severe BLI was defined as apao 2 /FIO 2 ratio of < 60 mm Hg, bilateral lung infiltrates, and bronchopleural fistula; moderate BLI as a PaO 2 /FIO 2 ratio of 60 to 200 mm Hg and diffuse (bilateral/unilateral) lung infiltrates with or without pneumothorax; and mild BLI as a PaO 2 /FIO 2 ratio of > 200, localized lung infiltrates, and no pneumothorax. Five patients developed ARDS with Murray scores > 2.5. Respiratory management included positive pressure ventilation in the majority of the patients and unconventional methods (ie, high-frequency jet ventilation, independent lung ventilation, nitric oxide, and extracorporeal membrane oxygenation) in patients with severe BLI. Of the four patients who had severe BLI, three died. All six patients with moderate BLI survived, and four of five with mild BLI survived (one with head injury died). Conclusions: BLI can cause severe hypoxemia, which can be improved significantly with aggressive treatment.the lung damage may be accurately estimated in the early hours after injury. The BLI severity score may be helpful in determining patient management and prediction of final outcome. (CHEST 1999; 115: ) Key words: ARDS; blast trauma; high-frequency jet ventilation; lung barotrauma; mechanical ventilation; nitric oxide Abbreviations: ALI acute lung injury; BLI blast lung injury; Fio 2 fraction of inspired oxygen; HFJV highfrequency jet ventilation; PCIRV pressure-controlled inverse-ratio ventilation; PEEP positive end-expiratory pressure; PPV positive pressure ventilation *From the Department of Anesthesiology and Critical Care Medicine (Drs. Pizov, Oppenheim-Eden, Matot, Weiss, Eidelman, and Sprung), and Trauma Unit (Dr. Rivkind), Department of Surgery, Hadassah University Medical Center, The Hebrew University of Jerusalem, Israel. Manuscript received February 19, 1998; revision accepted June 25, Correspondence to: Reuven Pizov, MD, Department of Anesthesiology and Critical Care Medicine, Hadassah University Medical Organization, P.O. Box , Jerusalem 91120, Israel Severely injured patients from explosions usually suffer from a combination of blast, blunt, penetrating, and burn injuries. 1,2 Primary blast damage is almost always seen only in gas-containing organs: the ears, respiratory system, and GI tract. A study including 828 victims from explosions reported immediate fatalities with mixed patterns of injuries including head, chest, and secondary missile injuries. 3 In 17% of the victims, however, lung damage was the only postmortem finding. 3 In another study from Northern Ireland, pathologic evidence of primary blast lung injury (BLI) was found in 45% of the victims who died at the site of the explosion. 4 BLI was a major cause of death in patients who survived initial resuscitation, accounting for nine deaths among 23 injured persons, 1 and BLI occurred more often in explosions in a confined space than in explosions that took place in the open. 5 7 The exact mechanism of lung injury resulting from CHEST / 115 / 1/ JANUARY,

2 blast waves is not clear. 2,8,9 The pressure waves propagate through the lung and result in pressure differentials at the interface between media with different densities. 2,9 These pressure differentials tear the alveolar walls and disrupt the alveolar/ capillary interface, resulting in giant emphysematous spaces filled with blood. 5,10 The airway epithelium may be injured or stripped away; together with giant emphysema, this predisposes the patient to air penetration into the pleura and mediastinum. These pathologic changes result in the primary clinical presentation of BLI, which is similar to pulmonary contusion from blunt trauma, but without rib fractures or chest wall injury. 2,11,12 Hemoptysis and barotrauma are characteristic signs of BLI. 1,6,11,13 No clinical data on BLI have been presented since the development of modern intensive care medicine two decades ago and the introduction of new technical and pharmacologic modalities to treat acute lung injury (ALI). The present study describes the presenting symptoms and respiratory management of BLI in 15 patients admitted to our hospital after terrorist bomb explosions on two civilian buses in Based on these data, a clinical classification of the severity of BLI is proposed. Patients Materials and Methods On February 25 and March 3 in 1996, two terrorist bombs exploded on civilian buses in Jerusalem, immediately killing 47 passengers at the site of the explosion. Eighteen survivors were admitted to our hospital. One was dead on arrival. Two patients had no lung injury, and the 15 remaining patients with varying degrees of BLI were included in the study. Data were gathered from the reports of the medical teams who rendered primary care at the scene, the records from the admitting area of the Trauma Unit, anesthesia records, and medical charts. The admitting hospital is the primary trauma referral center for the Jerusalem area. Demographic data included the sex, age, and injury severity score of the patients. All nonintubated patients were transferred to the hospital with oxygen masks. All endotracheal intubations were performed at the scene of the explosion by the rescue teams or in the admitting area of the Trauma Unit. The indications for intubation were unconsciousness, respiratory failure, emergency surgery, or a combination of the above. All patients had a neurologic assessment, including Glasgow Coma Scale and examination of the tympanic membrane. Measurements Initial oxygenation values were determined with a pulse oximeter or from arterial blood gases. Arterial blood gases were repeated at least every 6 h and more frequently in severely injured patients. Chest radiographs were taken on arrival to the Trauma Unit and at least 24 h later. The chest radiographs taken at arrival were used to estimate the percentage of lung area involved in the injury. Oxygenation dysfunction was estimated by calculating the Pao 2 /fraction of inspired oxygen (Fio 2 ) ratio. Four values are presented: the first available oxygenation value, the lowest and highest values during the first 6hintheICU, and the value 24 h after admission. Lung injury was estimated 6 and 24 h after admission by the modified Murray score for estimation of lung injury, 14 and was defined as ALI or ARDS. 15 BLI Severity Score After collecting clinical and laboratory data, a BLI score was defined based on hypoxemia (Pao 2 /Fio 2 ratio), abnormalities seen on chest radiographs, and barotrauma (Table 1). Patients oxygenation was measured in the first 6 h after admission to the hospital, after pleural drainage and mechanical ventilation were established. The highest Pao 2 /Fio 2 value was used in determining the score. The initial chest radiograph, performed shortly after admission to the hospital, was estimated for the lung area involved in injury. The existence of bronchopleural fistulae was used to estimate the degree of barotrauma to the lung. The score defined three levels of injury: mild, moderate, and severe BLI (Table 1). Treatment All surgical procedures were performed under general anesthesia. After primary resuscitation or surgery, all patients were admitted to the ICU. All patients were monitored with invasive BP measurement, central venous pressure and/or pulmonary artery catheter, and were volume resuscitated based on intravascular pressures. Within 1 h after admission to the hospital, patients with severe hypoxemia and significant bronchopleural fistulae were ventilated with pressure-controlled ventilation (Puritan-Bennett 7200E; Puritan-Bennett, Inc; Overland Park, KS). An attempt was made to use pressure-controlled inverse-ratio ventilation (PCIRV) with inspiratory/expiratory ratios of 2:1, and to lower the positive end-expiratory pressure (PEEP) as much as possible. Patients with bronchopleural fistulae who remained severely hypoxemic were treated with nitric oxide and/or high-frequency jet ventilation (HFJV) with superimposed positive pressure ventilation (PPV) (Acutronic High Frequency Jet Ventilator; Acutronic Medical Systems; Rappeswill, Switzerland). 16 In one patient, independent lung ventilation was used because of a huge unilateral bronchopleural fistulae. Patients with less severe lung injury were ventilated with volume-controlled or pressure support ventilation. The data are presented as median (range) when appropriate. Table 1 BLI Severity Score Severe BLI Moderate BLI Mild BLI Pao 2 /Fio to Chest radiograph Massive bilateral lung infiltrates Bilateral or unilateral lung infiltrates Localized lung infiltrates Bronchial pleural fistula Yes Yes/No No 166 Clinical Investigations in Critical Care

3 Results Fifteen patients (11 men) who suffered from different degrees of BLI were admitted to our hospital (Table 2). The median age was 29 years (range, 18 to 53 years). Ten patients presented with significant hypoxemia during the first hours after blast injury (Table 3). All but one of the patients were intubated within 2 h after the explosion (Table 2). The only patient (patient 14) who did not require intubation during the first few hours after explosion had an uneventful follow-up (Table 2). In addition to the primary blast injury, seven patients had burn injury of the upper airway, and one (patient 12) had significant smoke inhalation confirmed by bronchoscopy. Ten patients (67%) presented with varying degrees of hemoptysis. All patients had eardrum perforation. Nine patients (53%) had burns of varying degrees. Other associated injuries were a spleen tear in four patients (24%) and intra-abdominal bleeding from an injured mesenterium in two patients. Two patients had vascular lesions and four had bone fractures. Two patients (Nos. 9 and 13) had head injuries and another (patient 11) developed a hemispheric stroke on the second day after the explosion. The initial chest radiographs taken immediately after arrival to the hospital showed bilateral opacities in 12 patients (Table 4). Of the seven patients with bilateral pneumothoraces, five (Nos. 2, 6, 10, 11, and 15) presented with significant bronchopleural fistulae (Table 4). Except for one patient who died from BLI (patient 6), all patients showed improvement in oxygenation during the first 24 h after injury (Table 3). Based on the injury severity score, five patients (33%) had mild BLI, six (40%) had moderate, and four (27%) had severe BLI. BLI severity scores are compared with the Murray scores at 6 and 24 h (Table 5). Respiratory Management Of the five patients with mild BLI, one (patient 14) received oxygen through a face mask, whereas the others were ventilated for indications other than respiratory failure: two (Nos. 9 and 13) were mechanically ventilated because of head injury, one with extensive burns (patient 4) was intubated for 4 days because of repeated surgical procedures, and another (patient 5) was extubated within 24 h after laparotomy (Tables 5, 6). All patients in this group received volume-controlled or pressure support ventilation with PEEP that did not exceed 5 cm H 2 O (Table 5). All six patients with moderate BLI were mechanically ventilated for more than 1 day (median, 5 days). Two were ventilated with volume-controlled ventilation and four with PCIRV (Table 5). All patients with moderate BLI, except one (patient 12), received PEEP levels of 5cmH 2 O (Table 5). Patients with severe BLI were ventilated with PCIRV and PEEP levels of 10 cm H 2 O. Extreme hypoxemia (Pao 2 /Fio 2 60 mm Hg) together with significant bronchopleural fistulae required unconventional therapies in this group of patients (Tables 3, 5). One patient (patient 11) was ventilated with independent lung ventilation due to severe hypox- Table 2 Age, Sex, Intubation, Severity of Injury, Blood and Fluid Transfusion in 24 h, and Outcomes in Patients with BLI* Transfusion Patient No. Age, yr Sex Intubation Location Intubation Indication ISS GCS Blood, Units Clear Fluid, ml ICU Stay, d Outcome 1 29 F ER Unc, RF ,750 6 Survived 2 53 M ER Unc, RF , Died 3 21 F ER Unc, RF ,800 7 Survived 4 27 F ER Surg, RF ,500 8 Survived 5 20 M ER Surg ,800 4 Survived 6 30 M Field Unc, RF 41 Sedated 1 4,500 2 h Died 7 25 M ER Surg, RF , Survived 8 21 M ER Surg ,300 4 Survived 9 21 M Field Unc 41 Sedated 0 5,000 5 Died M Field Unc, RF , Survived M ER RF ,000 6 Died F ER RF ,000 8 Survived M ER Unc, RF ,000 8 Survived M No ,000 2 Survived M ER Surg, RF , Survived *ISS injury severity score; GCS Glasgow Coma Score; ER emergency room; Unc unconsciousness; RF respiratory failure; Surg emergency surgery; F female; M male. CHEST / 115 / 1/ JANUARY,

4 Table 3 Arterial Blood Gases* Patient No. Pao 2 /Fio 2 at Admission Pao 2 /Fio 2 During First 6hinICU Highest Lowest Pao 2 /Fio 2 at 24 h ph Paco 2, mm Hg Base Excess sat: 92% NA sat: 99% sat: 84% sat: 88% Median Range *NA not available because patient died 2 h after arriving at the hospital; sat saturation upon arriving in the emergency room. Values at lowest Pao 2 /Fio 2 within 6hofinjury. emia and a huge unilateral bronchopleural fistula during the first few hours after injury. The lung with the bronchopleural fistula was maintained with continuous positive airway pressure only. The patient s Pao 2 increased from 49 to 77 mm Hg and Paco 2 decreased from 63 to 43 mm Hg after applying independent lung ventilation. In one patient with severe hypoxemia, shock, and massive hemoptysis (patient 6), extracorporeal membrane oxygenation was attempted, but intrapulmonary bleeding increased after heparin administration and the patient died 2 h after admission to the hospital. Two patients (Nos. 2 and 10) were treated with nitric oxide inhalation. One (patient 2) showed a significant improvement in Pao 2 (from 48 to 61 mm Hg) after 3 ppm of nitric oxide; the other did not respond to up to 12 ppm of nitric oxide inhalation. These two patients were subsequently ventilated with HFJV/ PPV because of hypoxemia and massive bronchopleural fistulae. Both showed a significant improvement in oxygenation (from 61 to 242 mm Hg and from 42 to 75 mm Hg, respectively), and there was the impression of decreased flow through the bronchopleural fistulae. Table 4 Chest Radiograph Findings in Patients with BLI Patient No. Opacity on First Radiograph Localization Right Lung Injured Area, % Left Lung Pneumothorax Additional Barotrauma 1 Bilateral Diffuse Bilateral 2 Bilateral Diffuse Bilateral Med, perit* 3 Bilateral Diffuse Right Med 4 Bilateral Subpleural No 5 Bilateral Perihilar No 6 Bilateral Diffuse Bilateral 7 Bilateral Diffuse Bilateral 8 Bilateral Perihilar No 9 Left Peripheral 10 No 10 Bilateral Diffuse Bilateral Med 11 Bilateral Diffuse Bilateral Med, perit 12 Left Basal 20 No 13 Left Basal 10 Left 14 Bilateral Subpleural No 15 Bilateral Diffuse Bilateral *Med mediastinum; perit peritoneum. 168 Clinical Investigations in Critical Care

5 Table 5 Modes of Ventilation and Lung Injury Scores at 6 and 24 h* Patient No. Ventilatory Mode PEEP, cm H 2 O 6-h Scores 24-h Scores BLI Murray Lung Injury Murray Intubation duration, d 1 PPV 7 Moderate 2.7 ALI PCV/HFJV PPV 14 Severe 3.7 ARDS PCV 10 Moderate 3.0 ALI PPV 5 Mild 1.3 No PS 5 Mild 0.7 ALI PCV 10 Severe 3.3 NA NA 2 h 7 PCV 15 Moderate 4.0 ARDS PCV 10 Moderate 2.7 ALI PPV 3 Mild 1.0 No PC/HFJV PPV 14 Severe 3.7 ARDS PCV 10 Severe 3.3 ARDS PPV 5 Moderate 1.0 No PPV 3 Mild 1.0 ALI Spontaneous Mild 1.3 No 0 15 PCV 10 Moderate 2.7 ARDS *PCV pressure-controlled ventilation; PS pressure support; NA not available (patient died 2 h after admission to the hospital). Independent lung ventilation during first hours after injury. Outcome Of the 15 patients with BLI, 11 (73%) survived. Three of the four patients with severe BLI died. Patient 6 died from hypoxemia, severe intrapulmonary hemorrhage, and shock 2 h after admission to the hospital; patient 11 died following a hemispheric stroke after complete clinical recovery from lung injury; and patient 2 died after 58 days from complications of respiratory failure and sepsis (Tables 2, 6). All patients with moderate BLI survived and were discharged from the hospital. One patient with mild BLI (patient 9) died 5 days after the explosion from severe head injury, with no signs of lung damage. ARDS (Murray score 2.5) developed in all patients with severe BLI who survived the first 24 h, and in 33% of patients with moderate BLI (Tables 5, 6). Four patients (two with severe and two with Table 6 Treatments and Outcome by BLI Severity* Severe BLI Moderate BLI Mild BLI No. of patients ARDS Died Respiratory management PCV, PEEP, unconventional therapies PPV, PEEP No PPV, low PEEP *ARDS ARDS 24 h after injury; PCV pressure-controlled ventilation. Unconventional therapies included extracorporeal membrane oxygenation, independent lung ventilation, high-frequency jet ventilation, and nitric oxide. One patient died two hours after admission to the hospital. Patient died from head injury, Two patients in this group were mechanically ventilated for reasons other than respiratory failure. moderate BLI) required prolonged mechanical ventilation and a lengthy stay in the ICU ( 21 days) (Tables 2, 5). Discussion In this retrospective study, the clinical features of BLI in victims who were treated following two explosions on civilian buses in 1996 are presented. To the best of our knowledge, this is the largest group of patients with BLI described in the literature. Previous reports of BLI consist of descriptions of radiographic findings, epidemiologic information, or case reports. 1,13,17 19 The present study focused on analysis of BLI during the first 24 h after injury because this is the most critical and dynamic period. Lung injury developing within 24 h after the explosion was classified as ARDS or ALI, depending on the severity of injury. 15 Although all patients received oxygen during transport, they presented with different degrees of hypoxemia on arrival to the hospital. In 7 of 15 patients with BLI, the first oxygenation measurement showed extreme hypoxemia. Three of these patients had significant improvement over the next few hours, one died 2 h after admission to the hospital, and the other three patients improved later during the first 24 h. Effective pneumothorax drainage and adequate mechanical ventilation resulted in improved oxygenation during the first 24 h in the majority of our patients. Contrary to previous suggestions that the clinical picture of BLI may develop over 24 to 48 h, the present group of patients did not present a delay in the manifestation of lung injury. 4 CHEST / 115 / 1/ JANUARY,

6 In addition, the only nonintubated patient did not show deterioration of lung function during his hospital stay. Radiologic findings in this group were characterized by significant changes in the chest radiograph performed within 2 h after the explosion in all patients. In eight patients, diffuse bilateral injury was seen on the initial chest radiograph, and was accompanied by significant barotrauma in all of these patients. Such an early occurrence of radiologic changes, accompanied by massive hemoptysis, strongly suggests that intrapulmonary hemorrhage, which is characteristic for primary BLI, was the main cause of this finding. 10,20,21 Similar chest radiograph changes were found by Hirsch and Bazini 13 in patients who suffered from underwater BLI. Two patients had only subpleural opacities, which might be explained by rib imprint hemorrhages across the surface of the pleura. 20,22 Another two patients presented with perihilar lung consolidation, which has been previously described in blast-injured patients. 11,20 Early characteristic infiltrates on chest radiographs accompanied by severe barotrauma without rib fractures strongly support the hypothesis that the main lung injury was from the blast wave itself. 2,9 BLI Severity Score Stratification of the severity of lung injury produced by blast may be useful in the treatment and prediction of outcome. This scoring system may be particularly important in assessing mass casualties with blast injury, when triage decisions are crucial. Adequate management of patients with BLI depends on the severity of blast injury. The proposed BLI severity score is based on three objective signs: hypoxemia, chest radiograph findings, and the presence of bronchopleural fistula. Changes in oxygenation and changes seen on chest radiographs are key components of different scoring methods for lung injury and ARDS severity. 14,15,23 The Pao 2 /Fio 2 ratio gives a fairly accurate estimation of alveolar-capillary membrane function and is readily available. Moreover, it is the most frequently used lung injury parameter. 14,15,23,24 APao 2 /Fio 2 ratio of 60 mm Hg for was selected to indicate severe BLI because it is the minimal oxygenation level that provides adequate arterial blood saturation. Patients oxygenation was assessed after a few hours of treatment consisting of airway management, fluid resuscitation, pleural drainage, and mechanical ventilation. Chest radiography, although performed shortly after the explosion, provided an appropriate estimation of the severity of BLI in this group of patients. Pneumothoraces and bronchopleural fistulae create major respiratory problems. They result from the extensive barotrauma and are characteristic signs of BLI. 1,6,11,13 Although the magnitude of bronchopleural fistulae was not measured in our patients, its presence is significant in estimating the extent of primary BLI, since these patients had extreme respiratory failure. Our assessment of the proposed BLI score has its limitations, as the number of patients was relatively small and the study was a retrospective analysis. We did, however, study a homogeneous group of patients. All victims were relatively young, had no preexisting diseases, and were treated in the same hospital. At 24 h after injury, there was a good correlation between the proposed BLI score and modified Murray score. However, in the early stages after injury, the modified Murray score did not differentiate between patients with moderate BLI and those with severe BLI (Table 5), and therefore it would not be useful for triage of blast-injured patients. The BLI severity score will require further validation for confirmation of its utility in the management and outcome of blast patients. Respiratory Management Respiratory management of patients with severe BLI is challenging not only because of the lung injury itself (lung contusion with extensive barotrauma and bronchopleural fistula), but also because it is frequently accompanied by shock and unconsciousness. There have been no studies describing respiratory management of patients with BLI during the last two decades. One case report and several reviews discussed the accentuation of barotrauma and possible augmentation of systemic air embolism by positive-pressure mechanical ventilation. 2,9 The recommendations for treatment of BLI are prevention of intubation and positive-pressure ventilation. 1,2 The severity of the respiratory failure, accompanying unconsciousness, and/or the necessity of emergency surgery in the present group of patients left no alternative other than mechanical ventilation. In addition, the most seriously injured patients had hemodynamic and respiratory instability, which limited neuroradiologic evaluation. Moreover, the unknown neurologic status further limited the use of permissive hypercapnia because of fear of increasing intracranial pressure. 25 The term unconventional therapies in the present study encompassed independent lung ventilation, extracorporeal membrane oxygenation, HFJV/PPV, and nitric oxide. The analysis of the respiratory management received by patients with different severity scores demonstrates that unconventional respiratory therapies were applied only in 170 Clinical Investigations in Critical Care

7 patients with severe BLI (and in all such patients). Although the proposed score was developed retrospectively, the unconventional therapies required for these patients were prospectively applied clinically and did not bias the scoring decision. The score can be helpful to delineate patients with severe respiratory failure who will require more aggressive respiratory care. Although extracorporeal oxygenation has been used for more than two decades in several centers, its use is still under debate. 26,27 This method has been used previously in the late course of severe BLI. 18 Our attempt to use extracorporeal oxygenation 2 h after the explosion, as a last resort in a severely injured patient (patient 6) who suffered from refractory hypoxemia, burns, and shock, resulted in massive intrapulmonary bleeding and death. The early use of extracorporeal oxygenation and heparin in BLI has great risk and therefore must be carefully weighed. Although nitric oxide has become increasingly popular recently for treating severe ARDS, it is not the first choice for treatment of hypoxemia in lung injury. 28,29 Since only two patients were treated with nitric oxide, it is impossible to draw any conclusions. However, a good response in one patient with severe BLI makes it reasonable to try this method early after injury in the course of the treatment. HFJV is recommended for ventilation of patients with bronchopleural fistula. 30 Two patients with severe BLI were successfully treated with HFJV in combination with PPV as previously described. 16 HFJV provided adequate oxygenation and superimposed low-frequency PPV improved ventilation with the lowest possible airway pressures. Outcome In the present study, five patients had mild BLI, six had moderate BLI, and four had severe BLI. Although lung injury developing later in the course of a patient s hospitalization may be secondary to or aggravated by hypovolemic shock, sepsis, fluid transfusion, or aspiration, 1 the severity of the primary BLI had a dominant effect on the development of ARDS in the present group of patients. None of the patients with mild BLI developed any form of lung injury. All three patients with severe BLI who survived the first 24 h developed ARDS, as did 33% of the patients with moderate BLI. Obviously, patients with respiratory failure required prolonged mechanical ventilation and treatment in intensive care. Respiratory failure was the major cause of death in two patients with severe BLI, but in none of those with mild or moderate BLI. In conclusion, the severity of lung dysfunction resulting from blast injury could accurately be estimated in the early hours after injury using the proposed BLI score, but not the modified Murray score. Dramatic improvement occurred within the first 24 h, indicating that extraordinary efforts taken early may be useful. Stratification of BLI severity, as suggested in this study, may be used in future studies of blast-injured patients, may guide appropriate patient management, and may also serve to predict the final outcome in these critically injured patients. ACKNOWLEDGMENT: The authors acknowledge the valuable assistance of Dr. Azriel Perol of the Tel Hashomer Sheba Medical Center in the treatment of these most complicated patients. References 1 Mellor SG. The relationship of blast loading to death and injury from explosion. World J Surg 1992; 16: Phillips YY. Primary blast injuries. Ann Emerg Med 1986; 15: Mellor SG, Cooper GJ. Analysis of 828 servicemen killed or injured by explosion in Northern Ireland : the Hostile Action Casualty System. Br J Surg 1989; 76: Hadden MA, Rutherford WH, Merrett JD. The injuries of terrorist bombing: a study of 1532 consecutive patients. Br J Surg 1978; 65: Cooper GJ, Maynar RL, Cross NL, et al. Casualties from terrorist bombings. J Trauma 1983; 23: Katz E, Ofek B, Adler J, et al. Primary blast injury after a bomb explosion in a civilian bus. Ann Surg 1989; 209: Leibovich D, Gofrit ON, Stein M, et al. Blast injuries: bus vs open-air bombings a comparative study of injuries in survivors of open-air vs confined-space explosions. J Trauma 1996; 41: Cooper GJ. Protection of the lung from blast overpressure by thoracic stress wave decouplers. J Trauma 1996; 40(3 suppl): S105 S110 9 Mellor SG. The pathogenesis of blast injury and its management. Br J Hosp Med 1988; 39: Brown R, Cooper G, Maynard R. The ultrastructure of rat lung following acute primary blast injury. Int J Exp Pathol 1993; 74: Hadfield G, Dunn S. Lung injuries in air raids: a discussion on pathology and diagnosis. Br Med J 1941; August: Brismar BBL. The terrorist bomb explosion in Bologna, Italy, 1980: an analysis of effects and injuries sustained. J Trauma 1982; 22: Hirsch M, Bazini J. Blast injury of the chest. Clin Radiol 1969; 20: Bernard GR, Artigas A, Brigham KL, et al. The American- European Consensus Conference on ARDS: definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994; 149: Marks JD, Marks CB, Luce JM, et al. Plasma tumor necrosis factor in patients with septic shock: mortality rate incidence of adult respiratory distress syndrome, and effects of methylprednisolone administration. Am Rev Respir Dis 1990; 141: Rouby J-J. High frequency ventilation In. Perel A, Stock MC, eds. Handbook of mechanical ventilatory support. Baltimore: Williams & Wilkins, 1992; CHEST / 115 / 1/ JANUARY,

8 17 Hill JF. Blast injury with particular reference to recent terrorist bombing incidents. Ann R Coll Surg Engl 1977; 61: Weiler-Ravell D, Adatto R, Borman JB. Blast injury of the chest: a review of the problem and its treatment. Isr J Med Sci 1975; 11: Uretzky G, Cotev S. The use of continuous positive airway pressure in blast injury of the chest. Crit Care Med 1980; 8: Thomas A. The radiological aspect of the pulmonary changes following exposure to high pressure waves. Br J Radiol 1941; 14: Yelveton J. Pathology scoring system for blast injuries. J Trauma 1996; 40(3 suppl):s111 S Jaffin JH, McKinney L, Kinney RC, et al. A laboratory model for studying blast overpressure injury. J Trauma 1987; 27: Murray JF, Matthay MA, Luce JM, et al. An expanded definition of the adult respiratory distress syndrome. Am Rev Respir Dis 1988; 138: Vincent J-L, Moreno R, Takala J, et al. The SOFA (Sepsisrelated Organ Failure Assessment) score to describe organ dysfunction/failure. Intensive Care Med 1996; 22: Feihl F, Perret C. Permissive hypercapnia: how permissive should we be? Am J Respir Crit Care Med 1994; 150: Gattinoni L, Pesenti A, Macheroni D, et al. Low-frequency positive-pressure ventilation with extracorporeal CO 2 removal in severe acute respiratory failure. JAMA 1986; 256: Morris AH, Wallace CJ, Menlove RL, et al. Randomized clinical trial of pressure-controlled inverse ratio ventilation and extracorporeal CO 2 removal for adult respiratory distress syndrome. Am J Respir Crit Care Med 1994; 149: Rossaint R, Falke K, Lopez F, et al. Inhaled nitric oxide for the adult respiratory distress syndrome. N Engl J Med 1993; 328: Zapol W, Rimar S, Gillis N, et al. Nitric oxide and the lung. Am J Respir Crit Care Med 1994; 149: Slutsky A. Mechanical ventilation: American College of Chest Physicians Consensus Conference. Chest 1993; 104: Clinical Investigations in Critical Care