Reintubation as an Outcome Predictor in Trauma Patients

Reintubation as an Outcome Predictor in Trauma Patients* Brian]. Daley, MD; Felix Garcia-Perez, MD; and Steven E. Ross, MD, FCCP

Study objective: Determine reintubation rate, identify its cause, and detail adverse outcomes from reintubation. Design: Retrospective review of extubation failures in the trauma ICU. Setting: University hospital and regional trauma center. Patients: Four hundred five patients arriving intubated or requiring intubation during hospitalization after 2,516 traumatic injury admissions over 18 months. Interventions: None. Results: Reintubation incidence was 7% (27 times per 405 patients). Comparative mortality of the reintubated group (2/24=8%) is similar to overall trauma center mortality (224/2516=6.5% ), but less than the cohort of patients admitted to the hospital intubated (63/405=16%). Reintubated patients had an increased frequency of stridor than reported previously (33%), and an increased tracheostomy rate (62% vs 30% ). Stridor was not predictable from injury severity score, Glasgow coma score, age, sex, length of intubation, or place of intubation. Pulmonary complications (atelectasis, tracheobronchitis, pneumonia) developed in half of reintubated patients; stridorous patients did not have an increased rate of pulmonary complications. Conclusion: Reintubation in trauma ICU patients does not predict poor outcome. (CHEST 1996; 110:1577-80) Key words: extubation failure; reintubation; stridor; tracheostomy; trauma Abbreviations: GCS=Giascow coma scale; ISS=injury severity score

The use of mechanical ventilation and the length of time ventilated have been well correlated with increased mortality. The safest and most rapid method of weaning mechanical ventilation is controversial. 1 Those patients who fail extubation and require reinstitution of ventilatory support do so from a variety of causes. Regardless of the cause of extubation failure , an increased number of complications and higher mortality rate are reported when compared to those patients who did not require reintubation. 2 For these reasons, premature extubation is undesirable, and reintubation has become a national filter in critical care quality improvement. 3 To define and address preventable causes and/or to identifY those at risk for extubation failure, we sought to determine both the cause and patient outcome after reinstitution of mechanical ventilation in our trauma ICU population. Critically ill trauma patients are different from their *From the Division of Trauma, Department of Surgery, Cooper Hospital/University Medical Center, University of Medicine and Dentistry of New Jersey, Robert Wood Tohnson Medical School, Camden, NJ (Drs. Daley and H.oss ), and' the Division of Trauma, Jersey Shore Medical Center, Neptune, NJ (Dr. Garcia-Perez). Manuscript received February 27, I996; revision acceptedJune 14. Reprint requests: Dr. Daley, Dept. of Surgery, UTMCK, 924 Alcoa Hwy, Knoxville, TN 37920

medical counterparts with regards to chronicity of disease, end-organ dysfunction, functional reserve, in addition to age and cause for the precipitating respiratory failure or need for mechanical support. The population of surgical patients, ie, trauma and elective surgical patients, is also not homogeneous, and there are divergent data on reintubation in patients in the surgical ICU. Demling et al4 report on 700 patients who were extubated in two surgical ICUs at the same institution. A40% mortality from progressive cardiopulmonary failure is reported in the elective and nontrauma patients, whereas trauma patients had only a 10% mortality after reintubation. Most of the reintubations in the victims of trauma occurred following smoke inhalation injury or in head injury. A more recent report by Kaups et al5 showed that mortality was significantly increased in those surgical ICU patients requiring reintubation. In their series, trauma patients requiring reintubation had a significantly increased mortality. 5 MATERIALS AND METHODS

We performed a retrospective case review of the experience at the Southern New Jersey Regional Trauma Center from October CHEST/110/6/DECEMBER, 1996

1577

Table !-Extubation Criteria Criteria l. Resuscitation complete, sepsis controlled, urgent procedures completed 2. Respiratory rate <30 breaths/min 3. Negative inspiratory force >25 em H20 4. Vital capacity >10 mUkg 5. Tidal volume >5 mUkg 6. Minute ventilation <10 Umin

1992 to March 1994. The Southern New Jersey Regional Trauma Center has a catchment area of nearly 2 million people. Data were obtained from review of the patients' hospital records , the Trauma Registry, and minutes of the Quality Assurance Committee of the Division of Trauma, where reintubation is a mandatory review filter. Patients admitted to the trauma ICU who required reintubation during the course of their hospitalization were studied. Patients were intubated in the field by expe rienced e mergency medical technicians at the paramedic level or flight nurses. In the trauma admitting area or operating room, intubation was performed by experienced certified registered nurse anesthetists or anesthesiologists. Patients were maintained on a regimen of mechanical ventilation until resuscitation was complete, operative therapeutic measures were completed, and issues compromising respiratory function (sepsis, paralysis, mechanical impediments) had been addressed or were resolved. Endotracheal tube taping, suctioning, circuit maintenance, and ventilator setting changes are performed by respiratory therapists dedicated to the trauma ICU . Minimal occlusion pressure techniques were checked twice daily at a minimum to avoid high endotracheal tube cuff pressures. Termination of ventilatory support was determined by the trauma ICU attending surgeon, based on these initial criteria, and the achievement of objective weaning parameters (Table 1). Extubations were characterized as "planned"-meeting the subjective and objective criteria, and extubated with the physician present and on his/her order; "self-extubation"-prior to meeting subjective and objective criteria, without physician orders, by the direct actions of the patient; or "unplanned"-extubation prior to meeting subjective and objective criteria, without a physician order, and as a result of actions of a third party. Extubation failure was determined clinically, with supporting laboratory testing when possible. Reintubation was pe rformed by experienced certified registered nurse anesthetist or anesthesiologists. Cause for extubation failure was assigned based on the notation in the c hart at the time of reintubation. Pulmonary complications were identified at chart review-criteria for a diagnosis of pneumonia included a new or pe rsistent inflltrate on chest radiograph, feve r, leukocytosis, purulent sputum, and!or a positive sputum culture; patients without radiographic findings were defined as having tracheobronchitis. Atelectasis was defined as a radiographic finding of a lobar infiltrate without clinical or laboratory signs of

Table 2-Study Population Admissions (2,516) Intubated (405) Not Intubated (2,111 ) [excluded] Extubated/Died (381 ) Reintubated (24) [27 events] Self-extubated (5) Planned (19), Unplanned (3)

1578

infection. Timing of tracheostomy was detennined b ythe attending surgeon in the trauma ICU. RESULTS

During the study period, 2,516 patients were admitted to the trauma center. Blunt mechanisms (motor vehicle crashes, falls, etc) were most frequent cause for hospital admission. There were 405 patients who were intubated or arrived intubated. Twenty-four patients were reintubated 27 times for 19 planned, 5self-extubations, and 3 unplanned extubations (Table 2). All patients were intubated orotracheally. There were 8 women and 16 men, with a mean age of 37 years (range, 14 to 80 years). Mean Injury Severity Score (ISS ) was 20, and the mean Glasgow Coma Scale (GCS) at hospital admission was 11. Length of intubation ranged from 6 h to 9 days, with a mean of3.6 days. No patient suffered direct laryngeal trauma, and there were no head and neck burns or smoke inhalation injuries. Reasons for initial intubation were altered sensorium (12), airway safety after pain control maneuvers (5), respiratory distress (4), or respiratory arrest (3). Seven were intubated in the field, 14 in the admitting area, 1at a transferring facility, and 2 in the operating theater. Extubation occurred in 19 instances when the specific criteria (sensorium, control of infectious processes, respiratory parameters [negative inspiratory force, spontaneous tidal volumes, FVC, rate]) were achieved in a controlled environment. In five instances, the patient was able to dislodge the tube prior to meeting such criteria, and three times the endotracheal tube was dislodged from extraneous forces (ventilator tubing caught in bed rail when patient rolled for nursing care, dislodged during suctioning and displaced during bed-to-bed transfer). Quality assurance review found only one planned extubation requiring reintubation to be deemed inappropriate due to inadequate parameters; all five self-extubations were unpreventable (all patients were restrained and sedated at the time), and the three unplanned extubations were all preventable. Reasons for intubation were respiratory distress/inability to clear secretions (14 times in 11 patients), stridor (9 times in 8 patients), declining sensorium (3 times in 3 patients), and urgent reoperation (once in 1 patient). One patient was reintubated a second time for stridor after having tolerated breathing around a deflated cuff, and one patient with a low cervical spinal cord injury was reintubated three times for respiratory distress, each time achieving the criteria for extubation, and with hopes of self-sufficient respiration. Overall reintubation rate was 7%. Time to reintubation averaged 11.6 h, with a range from 0.1 to 72 h. Two patients in the reintubated group died, for a Clinical Investigations in Critical Care

mortality rate of 8%, 1 from a withdrawal of care at 7 days after family request, and 1 of pseudomonal sepsis with multiple organ system dysfunction at 8 months postinjury. The mortality rate for the intubated patient cohort was 16% (65/405), and the mortality rate for the trauma center over the study time period was 224 of 2,516, or 6.5%, including those who were dead at the time of admission. Death rate among reintubated patients and all intubated patient cohort was significant at a pvalue <0.1 but greater than p<0.05 (x 2 with Yates' correction). Of the 405 patients intubated prior to ICU admission, 123 (30%) had tracheostomy performed for prolonged ventilatory support or airway protection. Fifteen of 24 patients reintubated underwent tracheostomy (62%). These tracheostomy rates are significantly different by x2 analysis (p
Prediction of outcome has always been a goal of medicine. Recently, economic forces , ethical concerns, and resource allocation have spurred mathematically complex prognostic formulas. Efforts to simplifY prognostic determination have focused on simple characteristics or events that can be detected easily. The need for additional ventilatory support or airway protection after initial discontinuation of such support has been proposed to be a simple identifier of mortality risk in critically ill patients. Current data available on reintubation confirm that the differences in patient characteristics from the medical ICU, surgical ICU, and trauma ICU extend to ventilatory mechanics. A group of severely ill medical ICU patients (overall survival, 38%) had a significant relationship between reintubation and an increased

Table 3-Data Comparison of Reintubated Stridor vs Nonstridor Patients

No. Age, yr Sex, male/female ISS GCS Time intubated, d Time to reintubation, h Tube size Pulmonary complications

Stridor

Nonstridor

8 33.8 5/3 15.5 11.8 4.5 0.3 7.6 4

16 38.2 ll/5 20.7 10.8 3.6 18.1 7.9 10

p Value NS 0.324 NS 0.072 0.002* NS NS 1

*Mann-Whitney test. 1x2 with Yates' correction.

rate of complications. Mortality in the reintubated group was not reported; pulmonary compromise was the most frequent cause for reintubation (32.5%), whereas airway obstruction accounted for only 12.5% of extubation failures, with stridor occurring in 5 of 6 patients, accounting for 18% of the study population, but only 4 of whom were reintubated. 2 The surgical ICU patients described by Demling et al4 had a remarkably low reintubation rate of 4%, but this group had a 40% mortality. Reintubation in this group occurred at 48 h, without airway compromise, and indicated continued progressive cardiopulmonary decline. As in the surgical ICU patients, the trauma group had low reintubation rate of 3%. Mortality was one of ten patients, and no patients had stridor reported; air leak and laryngoscopy were employed in the group with inhalation injury to avoid stridor. Our experience with reintubation in critically ill trauma patients found that reintubation was not a predictor for death. Trauma patients are typically younger, without chronic cardiopulmonary dysfunction, and have acute and limiting physiologic disturbances. The patients in this study, although severely injured by accepted scoring systems, still underwent aggressive weaning and termination of ventilatory support in about 3 days. The 7% reintubation rate appears acceptable by literature standards, and certainly the lack of adverse outcome suggests that perhaps even more aggressive weaning is possible to avoid complications associated with prolonged mechanical ventilation, such as nosocomial pneumonia. This study is subject to the limitations inherent in any retrospective data analysis, although the patients were prospectively identified and followed up at the time of reintubation in quality assurance data. A prospective study may also cause alterations in mechanical ventilation weaning or termination practices if reintubation, a nonblindable event, is defined as the marker of outcome. There were more upper airway problems than previous reports. These were treated generally by traCHEST I 11 0 I 6 I DECEMBER, 1996

1579

cheostomy after reintubation and carried little or no threat to life. Additionally, because of the acute nature of traumatic injury and!or multiple organ dysfunction syndrome, with relatively short-term effects on ventilatory physiology, it was possible to reextubate several of these patients. Those requiring prolonged ventilatory support had tracheostomy performed. Thus, reintubation was not a predictor for mortality, but of need for longer ventilatory support or airway protection in our trauma patient population. We do not consider tracheostomy a complication, but rather an effective respiratory therapeutic maneuver, which is effected at the earliest convenience. Tracheostomy facilitates pulmonary toilet, improves patient comfort, and reduces dead space ventilation and airway resistance to allow weaning from mechanical ventilation. 6 Stridor was increased in our patients for reasons that remain unclear. Stridorous patients were intubated, on average, 21 h longer than those who did not develop stridor, although this was not statistically significant. Laryngeal edema occurs in up to 15% of patients intubated orotracheally, 7 secondary to local trauma at the area, either from intubation or the presence of the endotracheal tube, inciting an inflammatory response and edema, which compromise the airway. Steroids delivered preextubation do not reduce the incidence of laryngeal edema, 8 and in this study, laryngeal edema was more frequent in female patients and patients intubated greater than 36 h. In pediatric patients, postextubation stridor is increased in patients with facial bums, and in those patients in whom no air leak was present with cuff deflation immediately prior to extubation. 9 Similarly, in the report of Demling et al, 4 most extubation failure occurred in patients with smoke inhalation injury. We did not routinely assess preextubation air leak since there were no thermal injuries. There was no difference in sex distribution between the groups of patients with and without stridor. Unfortunately, the numbers of patients in these groups are small, and significance may be achievable only in larger populations. We were unable to ascertain traumatic intubations, either in the field or elsewhere, from the records available. The development of nosocomial pneumonia leads to increased costs and mortality, 10 and is probably the most frequent cause of death in the intubated patient.11 Because patients with underlying severe diseases are more likely to develop, and to die from pneumonia, 12 it is understandable that medical ICU patients would have a higher mortality from pulmonary

1580

compromise, and reintubation would act as a marker for severity of disease, rather than a provocative insult. Similarly, nontraumatic surgical ICU patients have more underlying disease, more pneumonia, and greater mortality, making extubation failure a marker of this. The younger trauma patient, with fewer chronic diseases, has other causes for respiratory compromise such as stridor. The reintubated patients in this study did not have increased incidences of pneumonia or pulmonary complications. CoNCLUSION

Reintubation in the traumatized ICU patient is not a predictor of mortality; reintubation implies the need for prolonged airway protection and/or ventilatory support and leads to tracheostomy. Aggressive termination of ventilatory support would be possible in the trauma patient as opposed to the medical ICU or surgical ICU patient, as there is little adverse outcome from reintubation when required. REFERENCES

1 Estaban A, Frutos F, Tobin MJ, et al. A comparison of four methods of weaning patients from mechanical ventilation. N Eng! J Med 1995; 332:345-50 2 Rashkin MC, Davis TL. Acute complications of endotracheal intubation: relationship to reintubation, route, urgency, and duration. Chest 1986; 89:165-67 3 Committee on Trauma. Quality improvement. In: Resources for the optimal care of the injured patient. Chicago: American College of Surgeons, 1993; 80 4 Demling RH, Truman R, Lind LJ, et al. Incidence of morbidity of extubation failure in surgical intensive care unit patients. Crit Care Med 1988; 16:573-77 5 Kaups KL, Shinozaki T, Deane RS, et al. Reintubation as a predictor of mortality [abstract]. J Trauma 1993; 34:181 6 Heffner JE. Timing of tracheostomy in mechanically ventilated patients. Am Rev Respir Dis 1993; 147:768-71 7 Bums HP, Dayal VS, Scott A, et al. Laryngotracheal trauma: observations on its pathogenesis and its prevention following prolonged OT intubation in the adult. Laryngoscope 1979; 89:1316-25 8 Darmon J-Y, Rauss A, Dreyfuss D, et al. Evaluation of risk factor for laryngeal edema after tracheal extubation in adults and its prevention by dexamethasone. Anesthesiology 1992; 77:245-51 9 Kemper KJ, Benson MS, Bishop MJ. Predictorsofpostextubation stridor in pediatric trauma patients. Crit Care Med 1991; 19:352-55 lO Freeman J, McGowan JE Jr. Risk factors for nosocomial infection. J Infect Dis 1978; 138:321-34 11 Fagon J-Y, Chastre J, Hance AJ, et al. Nosocomial pneumonia in ventilated patients: a cohort study evaluating attributable mortality and hospital stay. Am J Med 1993; 94:281-88 12 Craven DE, Kunches LM, Klinski V, et al. Risk factors for pneumonia and fatality in patients receiving continuous mechanical ventilation. Am Rev Respir Dis 1986; 133:792-96

Clinical Investigations in Critical Care