Tracheostomy Following Lung Transplantation Predictors and Outcomes


* Corresponding author: John V. Conte,jconte@csurg.


The effect of tracheostomy on patients receiving lung transplantation is unknown. We reviewed our experience by performing a retrospective analysis on all lung transplant recipients at our institution. Patients were assigned to each study group based on whether or not they received a tracheostomy in the acute postoperative period. One hundred and fourteen lung transplants were performed, and 16 of those patients received a tracheostomy. In the tracheostomy group, more patients had undergone bilateral-lung transplantation (81% vs. 34%, p = 0.001), more required cardiopulmonary bypass (75% vs. 38%, p = 0.005), more acquired postoperative pneumonia (88% vs. 30%, p < 0.001), had greater reperfusion injury at 48 h (PaO2/FiO2 of 233 vs. 345, p = 0.047), had longer initial periods on the ventilator (21 ± 7 vs. 2 ± 0.5 days, p < 0.001), more required re-intubation (56% vs. 18%, p = 0.001), spent longer times in the intensive care unit (30 ± 7 vs. 5.5 ± 0.9 days, p < 0.001), and had longer lengths of stay (67 ± 10 vs. 22 ± 2 days, p < 0.001). Despite these differences between the two groups, a significant difference in survival at 180 days (75 vs. 81%) did not exist (p = 0.89). Although tracheostomy is more likely in sicker patients, it is not associated with poor long-term outcomes.


Over the past several years, lung transplantation has become a viable therapeutic option for patients with end-stage lung disease. Despite advances in surgical techniques, patient selection, organ preservation, and immunosuppression, mortality continues to be high in the postoperative period as a result of ischemia-reperfusion injury and infection. Tracheostomy has been used in the intensive care unit (ICU) setting because of the need for prolonged ventilatory dependence and as an aide in ventilatory weaning. Despite the common use of tracheostomy, little data exists regarding its impact on patient outcomes and no data has been published regarding the lung transplant population. Additionally, clinical indicators that predispose a patient to receiving a tracheostomy as well as uniform indications for its use have not been well defined.

We reviewed our experience with tracheostomy following lung transplantation over an 8-year period in order to identify clinical predictors for receiving a tracheostomy and to describe the outcomes of lung transplant recipients receiving a tracheostomy.


All patients who underwent lung transplantation between February 1993 and November 2001 were included in the retrospective chart review. The patient sample consisted of single-lung, double-lung, and heart-lung transplants.

The patients were divided into two groups, based on whether or not they received a tracheostomy in the acute postoperative period. Tracheostomy was performed in lung transplant recipients according to accepted guidelines for weaning patients from mechanical ventilation, as outlined by Heffner (1). He concluded: (1) no consensus exists regarding the ideal time to perform a tracheostomy; (2) after 1–3 days of endotracheal intubation, patients should be evaluated daily to assess whether intubation may exceed 21 days; and (3) at 7 days, a decision should be made to proceed with the tracheostomy unless extubation is probable within the next 7 days. In general, a tracheostomy was performed if a patient failed to wean from the ventilator in the absence of reperfusion injury or treatable pneumonia and weaning was not imminent. Proper consent with the description of risks and benefits was obtained for each procedure. Tracheostomy was either performed surgically under general anesthesia in the operating rooms or percutaneously at the bedside for all patients.

Relevant medical records were reviewed in order to collect data retrospectively. This data has been part of a long-standing database at our institution. The database was formed before our institution required Institutional Review Board approval for information databases. Nevertheless, the database has been submitted for IRB approval, with approval expected on April 14, 2003. Data was collected on several parameters, including patient demographics, intraoperative data, ICU data, and longer-term postoperative statistics. A diagnosis of postoperative pneumonia was defined as the presence of an infiltrate on chest radiograph, an elevated temperature, an elevated white blood cell count, and presence of either purulent sputum or a positive respiratory culture. Reperfusion injury was measured by PaO2/FiO2 at 24 and 48 h after transplantation. Length of stay in the hospital was defined as the time from lung transplant until either death or discharge from the hospital to home or a rehabilitation facility. All data were analyzed without knowledge of the patients' study group assignments.

Frequencies were compared by using analysis of Chi-square. Continuous variables were compared using the Student's t-test for normally distributed variables. Results of continuous variables are displayed as median values ± standard error. Logistic regression was performed to determine the impact of several independent dichotomous variables associated with tracheostomy; the dependent dichotomous variable. Kaplan-Meier analysis was performed with multiple logistic regression for survival analysis. All analyses were performed on SPSS software (Chicago, IL).


Between February 1993 and November 2001, a total of 114 patients underwent lung transplantation. All patients were mechanically ventilated beginning immediately after the lung transplant. Of these, 16 patients eventually received a tracheostomy in the postoperative period. Baseline characteristics at time of transplant were similar for both groups (Table 1).

Table 1.  Demographics of the study patients
 TracheostomyNo tracheostomy 
Characteristic(n = 16)(n = 98)p-value
 Age (mean years±SD)42 ± 1649 ± 140.085
 Sex (male)9 (56%)42 (43%)0.318
 White15 (94%)84 (86%)0.633
 Black1 (6%)11 (11%) 
 Other03 (3%) 
 COPD6 (38%)47 (48%)0.076
 CF5 (31%)10 (10%) 
 Primary pulm. hyperten.3 (19%)4 (4%) 
 IPF1 (6%)15 (15%) 
 Eisenmeger's06 (6%) 
 Obliterative bronchiolitis03 (3%) 
 Scleroderma1 (6%)9 (9%) 
 Sarcoid04 (4%) 

Indications for lung transplant were chronic obstructive pulmonary disease, cystic fibrosis, idiopathic pulmonary fibrosis, primary pulmonary hypertension, Eisenmeger's, obliterative bronchiolitis, scleroderma, and sarcoid. Sixty-five patients underwent single-lung transplantation, 36 underwent double-lung transplantation, and three underwent heart-lung transplantation. A greater percentage of the tracheostomy patients had undergone double-lung transplantation (81% vs. 34%) (Table 2).

Table 2.  Transplant characteristics of the study patients
 TracheostomyNo tracheostomy 
Characteristic(n = 16)(n = 98)p-value
  • 1

    Median ± SE.

Type of transplant
 Single2 (13%)63 (64%) 0.001
 Double13 (81%)33 (34%) 
 Heart/lung1 (6%)2 (2%) 
CPB12 (75%)37 (38%) 0.005
CPB time, min1229±31183±15 0.31
Reintubated9 (56%)18 (18%) 0.001
Pneumonia14 (88%)29 (30%)<0.001
Pneumonia before trach12 (75%)  
Days until trach, days118±5  
SLT ischemic time, min1198±63207±8 0.258
DLT ischemic, lung 1, min1210±18195±12 0.981
DLT ischemic, lung 2, min1355±23292±16 0.437
24- h PaO2/FiO2, mmHg1330±58283±19 0.691
48- h PaO2/FiO2, mmHg1233±32345±47 0.047
Vent time, days121±7 2.0±0.5<0.001
ICU, days130±7 5.5±0.9<0.001
LOS, days167±1022±2<0.001

Although the ischemic times of the transplanted lung in the single-lung transplant patients were higher in the tracheostomy patients compared with the nontracheostomy group (Table 2), this difference was not considered significant. A greater proportion of patients in the tracheostomy group required cardiopulmonary bypass (75%) compared with the nontracheostomy patients (38%). Reperfusion injury, represented by PaO2/FiO2, was similar in both groups 24 h after transplantation (330 vs. 283 mmHg), but was significantly lower in the tracheostomy group 48 h after transplantation (233 vs. 345 mmHg).

Tracheostomy patients spent a longer time in the ICU (median, 30 vs. 5.5 days) and in the hospital (median, 67 vs. 22 days). As expected, they were also mechanically ventilated for a longer duration (median, 21 vs. 2 days). The tracheostomy group had a greater proportion of patients who were re-intubated (56% vs. 18%) as a result of initial extubation failure compared with the nontracheostomy group (Table 2).

The majority of patients receiving a tracheostomy were diagnosed with pneumonia compared with the nontracheostomy group (88% vs. 30%). Interestingly, only two patients in the tracheostomy group developed pneumonia after receiving a tracheostomy. The presence of pneumonia made a patient 16.7-fold more likely to receive a tracheostomy (Table 3). Other variables independently associated with receiving a tracheostomy were receiving a double-lung transplant, the use of cardiopulmonary bypass, and requiring re-intubation (Table 3).

Table 3.  Variables independently associated with tracheostomy1
VariableOdds ratio95% confidence interval
  • 1

    Calculated using logistic regression, giving each variable a coefficient of b1, which measures the variable's independent contribution to variations in the dependent variable (tracheostomy). The natural log of the odds ratio is equal to the coefficient b1.

Double-lung transplant12.42.2–29

All patients undergoing tracheostomy who were discharged from the hospital were successfully decannulated before discharge. There were no major complications of tracheostomy (e.g. pneumothorax, site infection, tracheal stenosis, trachea-innominate artery fistula).

We were not able to find a statistically significant difference in mortality between the two study groups after 6 months. There was a trend towards the tracheostomy group doing better in the short term, but survival was lower by 6 months. Survival between the two groups was statistically insignificant at 30 days (100% vs. 88%), 60 days (94% vs. 86%), 90 days (88% vs. 84%), and 180 days (75% vs. 81%) after undergoing transplantation (p = 0.89) (Figure 1). In the tracheostomy group, three of the seven deaths (43%) were primarily attributed to a pulmonary cause. In the nontracheostomy group, 21 of the 42 deaths (50%) were primarily attributed to a pulmonary cause (p = NS).

Figure 1.

Cumulative survival of patients.


In 2000, 1412 lung transplants were performed nationwide (2). The nationwide 1-year survival was approximately 75% (2). The major early (before 180 days) causes of death following lung transplant included respiratory failure and infection (2). The high proportion of deaths resulting from respiratory failure highlight the importance of the need to improve respiratory support. Unfortunately, the best course of action is rarely known when decisions need to be made. Tracheostomy is an elective procedure performed when patients fail to be successfully weaned off the ventilator. Tracheostomy has generally been associated with poor outcomes following lung transplantation. However, no study has been performed to test this theory in a transplant population. In fact, multiple re-intubations are often performed before tracheostomy is considered.

Studies in different populations have shown that tracheostomy has a beneficial impact on patient outcome. In a study of 521 patients in the intensive care unit requiring prolonged ventilation, Kollef et al. demonstrated a decrease in mortality (13.7%) in patients receiving a tracheostomy vs. those who did not (26.4%) (3). However, lung transplant patients were excluded in the analysis. In a study of 90 patients in a medical ICU who underwent tracheostomy, Brook et al. showed that those patients who underwent early tracheostomy (by day 10 of mechanical ventilation) had shorter overall durations of mechanical ventilation and shorter lengths of stay in the intensive care unit compared with patients who received a late tracheostomy (after day 10 of mechanical ventilation) (4). However, mortality did not significantly differ among the two groups. Additionally, there were no lung transplant patients in this study.

There are several clinical advantages of tracheostomy over endotracheal intubation. A tracheostomy allows for more effective deep tracheal suctioning, improved patient comfort including speech and mobility, facilitated nursing care, and allows for possible oral nutrition (5). Finally, studies suggest that tracheostomy results in clinical improvement in respiratory function and earlier weaning from mechanical ventilation. In a study of 20 patients, Davis et al. showed that work of breathing and expiratory airway resistance was reduced after patients were changed from endotracheal intubation to tracheostomy (6). However, the exact mechanisms by which this occurs has not been elucidated. One hypothesis claims that the additional length and the tortuous path of the endotracheal tube lead to an increase in airway resistance (6). Change in shape of the thermolabile endotracheal tube and adherence of secretions to the inner lumen are also possible contributing factors (6).

Subjecting the patient to the potential risks and costs of another surgical procedure represents the main obstacle to tracheostomy. Early complications of tracheostomy include bleeding, pneumothorax, hypoxia, hypercapnia, and infection (5). In an otherwise healthy patient, these risks are potentially treatable. But in an immunocompromized transplant recipient, these complications are potentially devastating. Long-term complications of tracheostomy include tracheal stenosis, trachea-innominate artery fistula, and nonhealing superficial scarring at the tracheostomy site (5). Moreover, debate continues regarding the optimal technique of tracheostomy placement, and whether the tracheostomy should be performed percutaneously at the bedside or by open surgical placement in an operating room. A few limitations of our study exist. As it is a retrospective study, there was no standard regarding the timing of tracheostomy placement. Placement was based on clinical judgment at the time. The number of patients receiving tracheostomy was small, thus interpretation of the outcome results need further evaluation in larger-scale trials. Patient satisfaction regarding tracheostomy was not assessed. Also, patients who received a tracheostomy were probably in a poorer condition than those in the control group. Therefore, we would normally expect survival rates to be lower in the tracheostomy group because of their worsened baseline condition. However, we showed that survival rates are similar. The effect of tracheostomy on the development of obliterative bronchiolitis could not be assessed because there were too few patients followed for a long enough period.

Our study shows several indicators exist in the lung transplant recipient that predispose a patient to receiving a tracheostomy at our center. Pre-transplant predictors include receiving a double-lung transplant (as opposed to a single-lung transplant) and the use of CPB. Prolonged ventilation, prolonged stay in the ICU, and 48-h reperfusion injury are postoperative indicators that may predispose a patient towards receiving a tracheostomy. Acquiring pneumonia postoperatively may predispose a patient towards receiving a tracheostomy. Additionally, being re-intubated may also contribute. This is similar to another study of 405 trauma patients requiring endotracheal intubation, which showed that re-intubation was associated with a 62% tracheostomy rate compared with 30% for non re-intubated patients (7). Recognition of these factors may lead to a more appropriate strategy for the optimal timing of tracheostomy placement.

No statistically significant difference exists between the tracheostomy and control groups in terms of both short-term and long-term survival (Figure 1). Therefore, we conclude that tracheostomy remains a viable therapeutic option to wean patients off the ventilator, can lead to increased patient comfort, and does not increase mortality in lung-transplant recipients. We recommend that tracheostomy be performed more aggressively in the post-transplant period if extubation is not possible.