Mediator-dependent secondary injury after unilateral blunt thoracic trauma.

The pathophysiologic sequence leading to respiratory failure after chest trauma can be an inevitable consequence of the primary injury or a secondary, mediator-driven inflammatory process. To distinguish between these alternatives, a simple cross-transfusion experiment was performed. A captive bolt gun injured the chest of anesthetized pigs that were mechanically ventilated with FiO2 = .21, .50, or .50 plus indomethacin (5 mg/kg intravenous; 15 min before injury). Tube thoracostomy immediately followed. After 30 min, blood from these injured donors was transfused into three matched groups of naive recipients (n = 8, 6, and 4, respectively) for a 33% exchange transfusion. Two control groups received blood from uninjured donors with tube thoracostomies only (FiO2 = .21, n = 7; FiO2 = .50, n = 10). Within 15-30 min after transfusion, in recipients from injured donors versus controls, lung compliance was decreased 20%, stroke volume and cardiac output were decreased 50%, and pulmonary vascular resistance was increased >300% (all p < .05). These changes recovered to baseline within 60-90 min. The stable metabolite of thromboxane A2, thromboxane B2, increased >500% in plasma within 15 min and remained elevated for >120 min. All responses were similar at 21 % or 50% O2, which suggests that hypoxia per se is not a cause of mediator production. All responses were eliminated by indomethacin. By 24 h, histologic changes included atelectasis in 3/3 recipients from injured donors versus 0/3 recipients from uninjured donors. We conclude that 1) blunt chest trauma releases blood borne mediators, including prostanoids; 2) these mediators can cause secondary cardiopulmonary changes in naive recipients similar to those produced by chest trauma; 3) the progression to trauma-induced respiratory failure is multifactorial; 4) early pharmacologic intervention, rather than supportive care alone, may benefit some victims of severe chest trauma.

Acadesine during fluid resuscitation from shock and abdominal sepsis.

OBJECTIVE: To determine properties of acadesine, the prototype adenosine regulating agent, in an experimental model in which abdominal sepsis is superimposed onto hemorrhagic shock. DESIGN: Randomized, blinded animal study. SETTING: University-based animal research facility. SUBJECTS: Twenty-eight anesthetized mongrel pigs (35.5 +/- 1.1 kg). INTERVENTIONS: The cecum was ligated and punctured to produce abdominal sepsis. To produce hemorrhagic shock, 45% to 47% of the estimated blood volume was withdrawn. After 1 hr, shed blood plus supplemental crystalloid (twice the shed blood volume) plus either acadesine (5 mg/kg bolus + 1 mg/kg x 60 min, n = 10) or its vehicle (n = 10) was administered. All animals were awakened and observed for 48 hrs. At 48 hrs, cardiac function, bacterial cultures from the septic focus, and inflammatory changes in the abdomen were quantified. MEASUREMENTS AND MAIN RESULTS: After resuscitation with acadesine vs. vehicle, we observed the following: a) arterial blood pressure and cardiac filling pressures were similar but cardiac index, systemic oxygen delivery, and systemic oxygen consumption were increased; b) plasma lactate was higher, systemic vascular resistance was lower, but ileal mucosal blood flow was not measurably altered; c) lipopolysaccharide-evoked tumor necrosis factor production in whole blood ex vivo was reduced; d) in those animals that survived 48 hrs (10/10 vs. 8/10), sepsis-induced cardiac depression, amount of free intraperitoneal fluid, extra abscess inflammatory reaction, abscess wall formation, abscess bacterial counts, and peritoneal bacterial counts, were all similar, but blood bacterial counts were higher. CONCLUSIONS: Fluid resuscitation with acadesine produced no adverse hemodynamic consequences and probably improved washout of metabolites from the reperfused microcirculation in sites other than the small intestine or heart. Taken together, these observations suggest that adenosine regulating agents might have therapeutic potential during fluid resuscitation from trauma. However, at least in these extreme conditions, the acute salutary effects of acadesine were probably overwhelmed by polymicrobial sepsis. Further studies must determine whether supplemental adjuvants to boost host defense during recovery from trauma will optimize adenosine-based resuscitation solutions.

Prognostic value of blood lactate, base deficit, and oxygen-derived variables in an LD50 model of penetrating trauma.

OBJECTIVE: To determine whether blood lactate, base deficit, or oxygen-derived hemodynamic variables correlate with morbidity and mortality rates in a clinically-relevant LD50 model of penetrating trauma. DESIGN: Prospective, controlled study. SETTING: University research laboratory. SUBJECTS: Anesthetized, mechanically-ventilated mongrel pigs (30+/-2 kg, n = 29). INTERVENTIONS: A captive bolt gun delivered a penetrating injury to the thigh, followed immediately by a 40% to 60% hemorrhage. After 1 hr, shed blood and supplemental crystalloid were administered for resuscitation. MEASUREMENTS AND MAIN RESULTS: After penetrating injury, 50.7+/-0.3% hemorrhage (range 50% to 52.5%), and a 1-hr shock period, seven of 14 animals died, compared with six of six animals after 55% to 60% hemorrhage, and 0 of nine animals after < or =47.5% hemorrhage. Only two of 13 deaths occurred during fluid resuscitation. At the LD50 hemorrhage, peak lactate concentration and base deficit were 11.2+/-0.8 mM and 9.3+/-1.5 mmol/L, respectively, and minimum mixed venous oxygen saturation, systemic oxygen delivery, and systemic oxygen consumption were 33+/-5%, 380+/-83 mL/min/kg, and 177+/-35 mL/min/kg, respectively. For comparison, baseline preinjury values were 1.6+/-0.1 mM, -6.7+/-0.6 mmol/L, 71+/-3%, 2189+/-198 mL/min/kg, and 628+/-102 mL/min/kg, respectively. Of all the variables, only lactate was significantly related to blood loss before and after fluid resuscitation in the 16 survivors. However, r2 values were relatively low (.20 to .50), which indicates that only a small fraction of the hyperiactacidemia was directly related to tissue hypoperfusion. In the whole population of survivors and nonsurvivors, both lactate and base deficit (but none of the oxygen-derived variables) correlated with blood loss. CONCLUSIONS: Arterial lactate is a stronger index of blood loss after penetrating trauma than base deficit or oxygen-derived hemodynamic variables. The reliability of arterial lactate depends on several factors, such as the time after injury, the proportion of survivors and nonsurvivors in the study population, and on factors other than tissue hypoxia.

Determinants of myocardial performance after blunt chest trauma.

OBJECTIVE: This study investigates whether factors that determine myocardial performance (preload, afterload, heart rate, and contractility) are altered after isolated unilateral pulmonary contusion. METHODS: Catheters were placed in the carotid arteries, left ventricles, and pulmonary arteries of anesthetized, ventilated (FiO2=0.5) pigs (31.2+/-0.6 kg; n=26). A unilateral, blunt injury to the right chest was delivered with a captive bolt gun (n=17) followed by tube thoracostomy. To control for anesthesia and instrumentation at FiO2 of 0.5, one group received tube thoracostomy only (sham injury; n=6). To control for effects of hypoxia without chest injury, an additional sham-injury group (n=3) was ventilated with FiO2 of 0.12. To generate cardiac function (i.e., Starling) curves, lactated Ringer's solution was administered in three bolus infusions at serial time points; the slope of stroke index versus ventricular filling pressure defines cardiac contractility. RESULTS: By 4 hours after pulmonary contusion, pulmonary vascular resistance, airway resistance, and dead space ventilation were increased, whereas PaO2 (72+/-6 mm Hg at FiO2=0.5) and dynamic compliance were decreased (all p < 0.05). Despite profound lung injury, arterial blood pressure, heart rate, cardiac filling pressures, and output remained within the normal range, which is inconsistent with direct myocardial contusion. The slope of pulmonary capillary wedge pressure versus left ventricular end-diastolic pressure (LVEDP) regression was reduced by more than 50% from baseline (p < 0.05), but there was no significant change in the slope of the central venous pressure versus LVEDP regression. By 4 hours after contusion, the slope of the stroke index versus LVEDP curve was reduced by more than 80% from baseline (p < 0.05). By the same time after sham injury with FiO2 of 0.12 (PaO2 < 50 mm Hg), the regression had decayed a similar amount, but there was no change in the slope after sham injury with FiO2 of 0.5 (PaO2 > 200 mm Hg). CONCLUSION: After right-side pulmonary contusion, the most often used estimate of cardiac preload (pulmonary capillary wedge pressure) does not accurately estimate LVEDP, probably because of changes in the pulmonary circulation or mechanics. Central venous pressure is a better estimate of filling pressure, at least in these conditions, probably because it is not directly influenced by the pulmonary dysfunction. Also, ventricular performance can be impaired by depressed myocardial contractility and increased right ventricular afterload even with normal left ventricular afterload and preload. It is thus conceivable that occult myocardial dysfunction after pulmonary contusion could have a role in the progression to cardiorespiratory failure even without direct cardiac contusion.

Cardiopulmonary function after pulmonary contusion and partial liquid ventilation.

PURPOSE: To compare the effects of mechanical ventilation with either positive end-expiratory pressure (PEEP) or partial liquid ventilation (PLV) on cardiopulmonary function after severe pulmonary contusion. METHODS: Mongrel pigs (32 +/- 1 kg) were anesthetized, paralyzed, and mechanically ventilated (8-10 mL/kg tidal volume; 12 breaths/min; FiO2 = 0.5). Systemic hemodynamics and pulmonary function were measured for 7 hours after a captive bolt gun delivered a blunt injury to the right chest. After 5 hours, FiO2 was increased to 1.0 and either PEEP (n = 7) in titrated increments to 25 cm H2O or PLV with perflubron (LiquiVent, 30 mL/kg, endotracheal) and no PEEP (n = 7) was administered for 2 hours. Two control groups received injury without treatment (n = 6) or no injury with PLV (n = 3). Fluids were liberalized with PEEP versus PLV (27 +/- 3 vs. 18 +/- 2 mL.kg-1.h-1) to maintain cardiac filling pressures. RESULTS: Before treatment at 5 hours after injury, physiologic dead space fraction (30 +/- 4%), pulmonary vascular resistance (224 +/- 20% of baseline), and airway resistance (437 +/- 110% of baseline) were all increased (p < 0.05). In addition, PaO2/FiO2 had decreased to 112 +/- 18 mm Hg, compliance was depressed to 11 +/- 1 mL/cm H2O (36 +/- 3% of baseline), and shunt fraction was increased to 22 +/- 4% (all p < 0.05). Blood pressure and cardiac index remained stable relative to baseline, but stroke index and systemic oxygen delivery were depressed by 15 to 30% (both p < 0.05). After 2 hours of treatment with PEEP versus PLV, PO2/FiO2 was higher (427 +/- 20 vs. 263 +/- 37) and dead space ventilation was lower (4 +/- 3 vs. 28 +/- 7%) (both p < 0.05), whereas compliance tended to be higher (26 +/- 2 vs. 20 +/- 2) and shunt fraction tended to be lower (0 +/- 0 vs. 7 +/- 4). With PEEP versus PLV, however, cardiac index, stroke index, and systemic oxygen delivery were 30 to 60% lower (all p < 0.05). Furthermore, although contused lungs showed similar damage with either treatment, the secondary injury in the contralateral lung (as manifested by intra-alveolar hemorrhage) was more severe with PEEP than with PLV. CONCLUSIONS: Both PEEP and PLV improved pulmonary function after severe unilateral pulmonary contusion, but negative hemodynamic and histologic changes were associated with PEEP and not with PLV. These data suggest that PLV is a promising novel ventilatory strategy for unilateral pulmonary contusion that might ameliorate secondary injury in the contralateral uninjured lung.