The objective of this study was to evaluate our thoracostomy tube management protocol following a trauma-induced event. We decided to retrospectively collect data among all the trauma patients who we have consistently used this algorithm on in hopes to not only improve patient care but also to share the algorithm due to its simplicity and effectiveness. In our practice, the mechanism of injury does not alter the management of tube thoracostomy, however, complication rates may vary depending on the severity of injury.
In our analysis, 15 of the 313 patients (4.8%) were found to have a thoracostomy tube complication. This was low compared to other studies. Menger et al. conducted a retrospective chart review in 154 patients with a 22.1% thoracostomy tube complication rate following thoracic trauma (Menger et al. 2012). They concluded that the severity of injury (measured by the abbreviated injury score) should be incorporated into the development of thoracostomy tube management guidelines. Patients in our study had a numerically higher ISS score than patients without complication, although this difference did not achieve statistical significance.
In 1995, Etoch et al. conducted a retrospective institutional review and displayed a 21% complication rate associated with thoracostomy tube management in 379 trauma patients (Etoch et al. 1995). However, the primary outcome of this study was a comparison regarding the complication rate after insertion by a surgeon (6%), an emergency medicine physician (13%), or prior to the transfer of the patient (38%).
In 2000, a retrospective case series determining the complication of tube thoracostomy in trauma patients was 30%. The objective was to determine if the rate was high enough to support a selective reduction in the indications for tube thoracostomy. The conclusion of the study revealed no persuasive evidence to support a selective reduction and the need for a larger study to confirm or refute their findings (Bailey 2000). In 1997, Chan et al. and Collop et al. reported complication rates of 11% and 18.2%, respectively (Chan et al. 1997; Collop et al. 1997). An algorithm regarding thoracostomy tube management with the above portrayed complication rates were not described in any of the above studies.
Based upon the available medical literature and clinical expertise, the Department of Surgical Education at the Orlando Regional Medical Center presented an algorithm similar to what we use in our institution (Cheatham 2009). However, this was based upon literature from other studies rather than upon results from their institution. It was not published as a peer-reviewed document, nor was it used in patients who were not involved in a trauma. In this evaluation, they depicted data based on a defined level of published data. Their recommendations regarding level-one evidence suggested that thoracostomy tube drainage should be ≤ 2 mL/kg/day or ≤ 200 mL/day before removal. Our algorithm suggests that thoracostomy tube output should be ≤ 200 mL/day to advance from wall suction to water seal or water seal to removal. These data were based on randomized studies evaluating the timing of thoracostomy tube removal in regards to the daily drainage volume (Younes et al. 2002; Hessami et al. 2009). Thus, both Hessami et al. and Younes et al. conducted prospective randomized investigations of 138 and 139 trauma patients and documented that a thoracostomy tube output of < 200 mL/day was just as safe as an output of <150 mL/day.
In regards to placing a thoracostomy tube on continuous suction after insertion, Davis and colleagues randomized 80 patients to wall suction versus water seal and observed a similar incidence of recurrent PTX (2.5%) in both groups (Davis et al. 1994). However, they concluded that the suction algorithm could help reduce the length of stay by reducing the total thoracostomy tube time (72.2 hours versus 92.5 hours, P = 0.013), as well as removal time (25.2 hours versus 35.6 hours, P = 0.034).
Interestingly, Martino et al. published a prospective randomized study in 205 trauma patients in which their thoracostomy tubes were removed either on water seal or on wall suction (Martino et al. 1999). Following thoracostomy tube removal, a recurrent PTX was seen in 13 patients in the water seal group with only one patient requiring tube replacement and a recurrent PTX in nine patients from the suction group with seven patients requiring a tube replacement. The study concluded that the water seal group was more likely to have recurrent PTX after thoracostomy tube removal but less likely to need a replacement.
Per our algorithm, we do not remove thoracostomy tubes on wall suction. Martino et al. concluded that a trial of water seal appears to allow occult air leaks to become clinically apparent, thus potentially reducing the need for another thoracostomy tube (Martino et al. 1999). Furthermore, regarding wall suction versus water seal, the literature has suggested that suction compared to water seal does not reduce air leaks in patients whom have had a pulmonary resection. However, it could decrease the occurrence of postoperative PTX from early air leak (Deng et al. 2010). Further, this study did not involve trauma patients. We do not have data to compare the recurrence of PTX after a thoracostomy tube is removed on water seal versus wall suction, which is a limitation of our study.
Schulman et al. prospectively evaluated the time interval for identifying a PTX after placing 119 thoracostomy tubes on water seal for three hours then obtaining a chest x-ray (CXR) (Schulman et al. 2005). They concluded that 31 patients had a PTX on follow-up CXR, 22 were identified early and nine were late. Of the 22 patients identified early, three had a clinically significant increase in size of a PTX. This may suggest that three hours of water seal time may not be safe. Our algorithm is based on approximately 24 hours of water seal time before removal with only 2/313 patients (1.3%) having a recurrent PTX.
In regards to removing a thoracostomy tube, we obtain a thoracostomy tube CXR ≥ 4 hours after removal. From our data, two of the patients (1.3%) developed a recurrent PTX after removal and these patients were observed without an intervention. Bell et al. revealed that up to 24% of patients might have a small apical PTX after thoracostomy tube removal that does not require a repeat thoracostomy tube (Bell et al. 2001). Interestingly, this prospective randomized study also compared the removal of 102 thoracostomy tubes in 69 trauma patients, either at the end of inspiration or at the end of expiration, and found no significant difference in the recurrence of a PTX. In 2000, Pacanowski et al. conducted a retrospective review of 105 patients with 113 thoracostomy tubes removed with a protocol CXR performed eight to 22 hours after removal (Pacanowski et al. 2000). The authors advocated obtaining a CXR 24 hours after thoracostomy tube removal.
The median chest tube drainage time for patients in the study of Younes et al. was approximately three days (Younes et al. 2002). However, there were three treatment arms in this study, and chest tubes were withdrawn depending on the amount of pleural fluid that was drained (<200 mL/day, <150 mL/day, and <100 mL/day), whereas in our study all patients were removed from suction once their drainage was below 200 mL/day. Even though we had an average drainage time of 5.9 days, it is likely that the 313 patients in our study had experienced more serious injuries than the 37 patients who were in the <200-mL/day-drainage group in the study of Younes et al., which would explain the increased chest tube drainage time.
The mean length of hospital stay in our study was 10.4 days. Although Hessami et al. reported a mean hospital stay of 4.1 days, it is possible that this difference can be attributed to more severely injured patients in our study who required longer hospital care (Hessami et al. 2009). Hessami had 138 patients who required chest tube placement due to malignancy and trauma, while all of our patients had trauma injuries, which could contribute to longer hospital duration.
The policy in our department with regard to thoracostomy tube size is in line with the Advanced Trauma Life Support® (ATLS) recommendations, such that a 32–40 French (F) drain should be used for a trauma-induced PTX or hemothorax. Collop et al. found a 36% complication rate associated with a tube thoracostomy of 14 F or less compared to a 9% rate for a standard tube (Collop et al. 1997). In general, the use of a large bore 32 F or greater is likely to reduce the complication associated with a drain becoming kinked or clotted. In our study, one patient with a 32 F thoracostomy tube clotted, requiring a thoracostomy and replacement with two 32 F thoracostomy tubes. Interestingly, Inaba et al. conducted a prospective analysis and compared the efficacy of small (28–32 F) versus large (36–40 F) thoracostomy tubes in 293 patients with thoracic traumas and concluded no differences in retained hemothoraces, the need for additional tube insertion and/or pain level (Inaba et al. 2012).
Occult PTX is a controversial situation that may arise in rare circumstances. By definition, an occult PTX is a PTX identified by computed tomography (CT) scan but not by CXR. We have included information in our algorithm as to how we manage occult pneumothoraces. However, a large prospective randomized study is needed in trauma patients to guide the management. We did not evaluate the number of patients who were admitted with a trauma-induced occult PTX that did not require a thoracostomy tube, which was a limitation of this study.