Pneumothorax

Background: Placing chest tubes to water seal is superior for patients with an air leak, but when a patient has a pneumothorax and an air leak the best chest tube setting is unknown.

Methods: This is a retrospective analysis of a prospective database on a consecutive series of patients who had a pneumothorax and air leak on the same day. Patients underwent elective pulmonary resection by one surgeon and had their chest tubes placed to water seal on postoperative day 1. Daily chest radiographs were obtained, and the size of the pneumothorax and air leak were measured. Tubes were left on seal unless there was a symptomatic enlarging pneumothorax or subcutaneous emphysema (defined as failing water seal). The primary objective was to evaluate the efficacy of water seal. We also wanted to identify risk factors that predicted failure of water seal.

Results: There were 838 patients [greater than or equal to] 21 years old who underwent elective pulmonary resection, and 86 patients (10%) had an air leak and a concomitant pneumothorax on the same day. Fourteen patients (16%) failed water seal. Multivariate analysis showed that a large air leak (greater than or equal to expiratory 3 in our classification system; odds ratio [OR], 16.5; p < 0.001) and a pneumothorax > 8 cm in size (OR, 4.9; p < 0.005) were predictors of failing water seal.

Conclusions: Keeping chest tubes on water seal is safe for most patients with an air leak and a pneumothorax. However, if the leak or pneumothorax is large, then subcutaneous emphysema or an expanding symptomatic pneumothorax is more likely. A prospective randomized trial is needed to compare water seal to suction in these patients.

Key words: air leak; chest tube management; lung resection; pneumothorax; quantification

Abbreviations: CI = confidence interval; CXR = chest radiograph; DLCO/VA = diffusion capacity of the lung for carbon monoxide corrected for alveolar volume; OR = odds ratio; RDC = Robert David Cerfolio

**********

Until recently, the management of chest tubes after pulmonary resection has not been studied. Most physicians make derisions based on their training and experience as opposed to prospective randomized trials. Over the last few years, there have been several randomized trials (1-4) that have examined chest tube management issues. It has been shown that although the majority of physicians favored placing chest tubes on suction when a patient has an air leak after lung resection, water seal is superior. (1-3,5) Air leaks stop sooner when chest tubes are placed on water seal. Despite these prospective randomized trials, many other questions remain. Since apposition of the parietal to the visceral pleura is thought to be an important factor for air leaks to heal, many favor placing chest tubes on some suction if a patient has a pneumothorax, even if there is an air leak. We decided to examine if water seal, which has been shown in two prospective randomized trials (2,3) to be a superior chest tube setting for patients with air leaks, is efficacious and safe for patients who have an air leak and a pneumothorax. We also wanted to identify risk factors that would predict which patients might fail water seal.

MATERIALS AND METHODS

Inclusion Criteria

The study population consisted of 838 patients [greater than or equal to] 21 years old who underwent elective pulmonary resection at the University of Alabama at Birmingham performed by a general thoracic surgeon (R.J.C.) over a 2-year period. Patients who underwent pneumonectomy, video-assisted thoracoscopic resection, or pulmonary resection performed as a part of a larger procedure (ie, decortication for empyema or Ivor Lewis esophagogastrectomy) were excluded. Data were collected prospectively, and the Institutional Review Board at University of Alabama at Birmingham approved the prospective database and this study. There were 86 patients identified who had an air leak and a concomitant pneumothorax on the same day, and they represent the study cohort.

Intraoperative Management

All patients received standard preoperative and intraoperative management. Incomplete fissures were stapled, and lung sealants were not used. At the completion of the pulmonary resection, the chest was filled with saline solution and the remaining lung was inflated. Air leaks were identified and sewn. Two chest tubes (28F Soft Deknatel; Genzyme Biosurgery; Cambridge, MA) were placed if the patients underwent upper or lower lobectomy, and one chest tube was placed if they underwent segmentectomy, wedge resection, or middle lobectomy. Patients who underwent upper lobectomy had two straight chest tubes placed in the apex of the chest. One was placed posteriorly, and one was placed anteriorly. If the patient had a lower lobectomy performed, one straight chest tube was placed in the anterior apex and a right angle tube was positioned in the posterior base.

Chest Tube Management

Chest tubes were connected to the drainage system and then -20 cm [H.sub.2]O of suction was added on the day of surgery. The chest tubes were then placed on water seal on the morning of postoperative day 1. Chest radiographs (CXRs) were obtained daily. All tubes remained on water seal unless they failed water seal. Failure of water seal was defined as follows: (1) the development of new-onset hypoxia (saturations consistently < 93%, or a fall of 6%, without an increase in the fraction of inspired oxygen), and a repeat CXR (performed at least 4 h after switching to water seal) that showed the pneumothorax was larger; (2) the development of new-onset subcutaneous emphysema or enlarging subcutaneous emphysema, determined by clinical examination and/or radiographs; and/or (3) the development of a change in the pitch of their voice and the presence of subcutaneous emphysema. If any of these events occurred, the patient was defined as "having failed water seal." All other patients were defined as having "tolerated water seal." For those who failed water seal, their chest tubes were then placed to -10 cm [H.sub.2]O of suction. If the problem continued after 24 h the tubes were then placed to -20 cm [H.sub.2]O of suction.

Definitions

Water seal is defined by placing the chest tubes to the underwater seal that is contained in the drainage system that was employed throughout this study (Fig 1). Wall suction is defined by attaching the chest tubes to this drainage system (which contains an underwater seal in it) and then attaching the drainage system to wall suction in the patient's room. Thus, suction is actually added to the underwater seal. A dial on the drainage system 'allows one to vary the amount of suction delivered to the chest tubes from -10, -20, or -40 cm [H.sub.2]O.

[FIGURE 1 OMITTED]

A portable bedside CXR was obtained in the recovery room after the completion of the operation and then daily at 3 AM. A patient was defined as having a pneumothorax if on the daily portable early morning CXR there was noted to be a pneumothorax after review with an attending chest radiologist.

Calculation of Pneumothorax

We wanted to objectify the size of the pneumothorax on a portable CXR. Thus, we modified a previously described technique by Choi et al, (6) which measures three distances between the parietal pleural of the chest wall and the visceral pleural of the collapsed lung. The location of these measurements is the same for all patients and is described in Figure 2. The first measurement is the height of the pneumothorax, which is measured from the most superior aspect of the apex (using a vertical line that is perpendicular to the floor) to the top of the collapsed lung. This is labeled as measurement 1 in Figure 2. The next two distances are the lateral components of the pneumothorax. Some pneumothoraces have a lateral component that goes further down the chest than others, and thus the lateral component is measured in two locations. These two locations are found at the midpoint of the upper and lower hemithoraces. A horizontal line (which is at a 90[degrees] angle to the vertical line used for the first measurement) is drawn from these two locations to the collapsed lung. These two locations are the precise midway point of the two hemithoraces. These two measurements are labeled 2 and 3 in Figure 2. The fourth measurement is a continuation of the same line used for the first measurement. It is the distance between the collapsed lung and the diaphragm. Although one could also add a fifth and sixth measurement to include patients with a medial lateral component of a pneumothorax (a pneumothorax from the midline out laterally), no patient had one in this series, and thus it was not used for this study. The stun of these four numbers represents the "size of the pneumothorax" in centimeters, and all pneumothoraces were measured and reported in this manner.

[FIGURE 2 OMITTED]

Quantification of Air Leak

Air leaks were measured and scored using the previously described Robert David Cerfolio (RDC) classification (Appendix) (2,7,8) and recorded by the same two observers (R.J.C., C.S.B.). A patient was defined as having an air leak on a postoperative day if on morning rounds there was a column of air that consistently crossed into the air leak chamber. Patients were asked to take two deep breaths each morning, and if no leak was seen they were asked to cough twice. A patient was defined as having subcutaneous emphysema if there was air present under the skin that was observable by CXR or physical examination. Any patient with an air leak and a pneumothorax on the same day was entered into the study. The size of the air leak and the size of the pneumothorax on that day were used for statistical analysis. An increase in subcutaneous emphysema was determined by physical examination or by CXR.

Statistical Analysis

The primary objective of this trial was to assess the efficacy of water seal for patients with an air leak and a pneumothorax. We also wanted to identify, predictors for failing water seal. Data for the categorical variables are expressed as the number and percentage of patients and were compared by either the Fisher Exact Test or [chi square] test, as appropriate. For continuous variables, data are reported as means with SDs. Normalization of variables was tested using the Shapiro-Wilk test. Values were compared by Student t test or by Wilcoxon test for those with a nonnormalized distribution.

Multiple logistic regression was performed to identify predictors of water seal failure. All predictor variables that were significant at a two-tailed nominal p value of < 0.10 in univariate analyses were entered into a multivariate logistic model. Stepwise logistic regression was performed, and for each element remaining in the multivariate model, a parameter estimate was calculated from which a p value, odds ratio (OR), and 95% confidence interval (CI) for the variable were derived. A two-sided p value of < 0.05 was considered to indicate statistical significance. The analysis was carried out using statistical software (EpiInfo 2002; CDC; Atlanta, GA; and SAS v.8.2; SAS Institute; Cary, NC).

RESULTS

Table 1 compares the demographics, some previously identified risk factors for air leaks, (5) and the type of surgery for these 86 patients. It also compares the 72 patients who tolerated water seal to the 14 patients who did not.

Variables associated with water seal failure were the diffusion capacity of the lung for carbon monoxide corrected for alveolar volume (DLCO/VA) [reported as percentage of the predicted value] < 60%, an air leak greater than or equal to expiratory 3, and a pneumothorax [greater than or equal to] 8 cm. Of the candidate variables found by univariate analysis, two emerged as possible predictors of water seal failure according to the multivariate logistic regression analysis (Table 2). An air leak greater than or equal to expiratory 3 was the strongest predictor of water seal failure (OR, 14.3; 95% CI, 5.8 to 28.3, p < 0.001). In addition, a pneumothorax [greater than or equal to] 8 cm was also a predictor of failing water seal (OR, 4.9; 95% CI, 1.9 to 13.4; p < 0.005).

Of the 84 patients in this study, 31 had an air leak that was greater than or equal to expiratory 3. Only 11 of those patients (35%) patients failed water seal. There were 16 patients with a pneumothorax > 8 era, and 6 of them (38%) failed water seal. There were seven patients who had both an air leak greater than or equal to expiratory 3 and a pneumothorax > 8 cm, and five patients (71%) failed water seal. We found no association between having a pneumothorax > 8 cm and having an air leak (p = 0.26) in this cohort.

DISCUSSION

The management of drains and tubes after many surgical procedures is highly variable. (9) Few trials have been dedicated to these issues, and thus surgical opinion has become dictum. We have studied some of the various issues concerning chest tube management often using prospective randomized trials. We have tried to address one specific question at a time in a select group of consecutive patients who have undergone similar operations by one surgeon with similar preoperative, intraoperative, and postoperative techniques in an attempt to control as many confounding variables as possible.

In this article, only patients who underwent an elective pulmonary resection and had a concomitant air leak and a pneumothorax on the same postoperative day were studied. The most common day this occurred was postoperative day 2. This specific problem, although not infrequent clinically, has never been studied before. The goal of this study was to evaluate the efficacy and safety of water seal in patients with an air leak and a pneumothorax. Indeed we found that 84% of the patients tolerated water seal well despite having a pneumothorax. From this data, we cannot state that water seal is superior to suction for these patients; only a prospective randomized trial can do that. However, we can state that it works well in most patients and is safe. It does not lead to subcutaneous emphysema or an expanding pneumothorax in most patients, and our previous studies (1,3) (which showed that water seal is superior to suction for stopping air leaks) have already shown the beneficial effect of water seal in patients with an air leak.

The second goal of this study was to identify risk factors for the failure of water seal. Patients who have an air leak that is greater than or equal to expiratory 3 are at increased risk of an enlarging pneumothorax that causes hypoxia or subcutaneous emphysema. Yet only 35% of patients in this series with this large of a leak failed water seal. However, when patients had both a large air leak and a large pneumothorax, 71% failed water seal. We wondered if the large leak was the cause of the pneumothorax but did not find a correlation between the two. In fact, the size of the air leak was a 3.4 times stronger predictor of failing water seal than the size of the pneumothorax (the OR was 16.5, compared to 4.9). It seems that when the chest tubes are on water seal with a large air leak, the volume of evacuated air may not be sufficient. This theory is supported by the fact that when these chest tubes are placed to--10 cm [H.sub.2]O suction, their symptoms of hypoxia and increasing subcutaneous emphysema immediately improve.

Subcutaneous emphysema represents air leaking from the lung that is not controlled by the chest tubes. Although it is not life threatening, it is bothersome for patients. Sometimes, a large amount of air enters the eyelids and patients are unable to open their eyes, and this causes significant anxiety. Air can also dissect into the larynx and cause a high-pitched voice. If this happens, the tubes are not fully evacuating the air that is leaking from the lung. The application of suction alleviates this problem as well, but in our experience it also makes the leak larger and delays its healing. The second identified risk factor of failing water seal is the size of the pneumothorax.

Quantifying the size of a pneumothorax on a portable CXR is difficult to do. (9) The films are often of poor quality, and there may be subcutaneous emphysema or hardware in the way. In order to quantify this area for scientific study, we had to modify a previously described technique. We did not attempt to correlate this measurement with the percent pneumothorax since that is not a verifiable amount nor is it important. Whether our two-dimensional measurement of a three-dimensional space is accurate is irrelevant. The goal of this technique is clear. It is for the busy clinical physician at the bedside to be able to quickly look at a portable CXR, make a few measurements with a ruler on the CXR, and decide if water seal is safe for a particular patient with a pneumothorax and an air leak, or if suction is better. This technique is simple, takes < 1 min to do, and is easy to teach. Although the size of the pneumothorax was a predictor of failure of water seal, it was not as strong a predictor as the size of the leak.

Although we believe that parietal pleura to visceral pleural apposition is an important factor for the healing of air leaks, this study has shown it is not a necessary component. Some patients have fixed pleural space deficits after pulmonary resection. The elasticity and/or compliance of the remaining lung, chest wall, and diaphragm are such that the remaining lung cannot fill the pleural space. This residual space will fill with noninfected fluid over time if the tubes are removed quickly. In this clinical scenario, the addition of suction in these patients does not help and may only prolong air leaks. If the lung does not expand further on tubes that are patent and on suction, the tubes should be left on water seal. However, if the lung does expand further and pleural-pleural apposition is achieved, air leaks may stop sooner. A prospective randomized trial is needed to completely answer this question.

In conclusion, this is the first reported series to evaluate the chest tube management in patients with both an air leak and a pneumothorax. We found that water seal is safe for most patients with this not uncommon problem. Subcutaneous emphysema and/or an enlarging symptomatic pneumothorax can occur if the air leak is larger than or equal to an expiratory 3 and/or if the pneumothorax is [greater than or equal to] 8 cm. Further prospective randomized trials are needed to evaluate if water seal is better than suction at stopping air leaks in patients with both a pneumothorax and leak. This trial may help set up safety standards for such a trial.

APPENDIX

The RDC classification system classifies an air leak based on when it occurs during the respiratory cycle (the qualitative aspect of the classification system) and how big the air leak is (the quantitative aspect of the system). Briefly, the RDC system labels air leaks as one of four types: continuous (occurring during inspiration and expiration), inspiratory, expiratory, or forced expiratory. A forced expiratory leak is present when a patient has no air leak detected with deep inspiration and expiration, but has one only with a forceful cough. Once the air leak is classified as one of these four types, its size is then measured from 1 to 7. This is assessed at the bedside by watching which one of the seven chambers the bubbles from the air leak go into (as shown in Fig 1). Using this classification system, an air leak is completely scored. For example an air leak can be an expiratory 2 or a forced expiratory 3. We have shown this classification system to be reproducible between different observers and useful in the treatment and natural history of air leaks. (2,4,7,8) It has also been shown that forced expiratory leaks are the smallest followed by expiratory, then inspiratory, and finally continuous. (2,3,7,8) The latter two types of leaks are extremely rare. For research purposes, all leaks can be placed on a spectrum, from the smallest air leak (a forced expiratory 1) to the largest air leak (a continuous 7). Leaks are assigned a numeric value for statistical purposes from 1 to 28 so we can compare the size of air leaks and their outcomes. We refer to this as an air leak score.

REFERENCES

(1) Cerfolio RJ, Tummala RP, Holman WL, et al. A prospective algorithm for the management of air leaks after pulmonary resection. Ann Thorac Surg 1998; 66:1726-1731

(2) Marshall MB, Deeb ME, Bleier JIS, et al. Suction vs water seal after pulmonary resection: a randomized prospective study. Chest 2002; 121:831-835

(3) Cerfolio RJ, Bass C, Katholi CR, et al. A prospective randomized trial compares suction versus water seal for air leaks. Ann Thorac Surg 2001; 71:1613-1617

(4) Cerfolio RJ, Bass CS, Pask AH, et al. Predictors and treatment of persistent air leaks. Ann Thorac Surg 2002; 73:1727-1731

(5) Davis JW, Mackersie RC, Hoyt DB, et al. Randomized study of algorithms for discontinuing tube thoracostomy drainage. J Am Coll Surg 1994; 179:553-557

(6) Choi BG, Park SH, Yun EH, et al. Pneumothorax size: correlation of supine anteroposterior with erect posteroanterior chest radiographs. Radiology 1998; 209:567-569

(7) Cerfolio RJ. Chest tube management after pulmonary resection. Chest Surg Clin North Am 2002; 12:507-527

(8) Cerfolio RJ. Advances in thoracostomy tube management. Surg Clin North Am 2002; 82:833-848

(9) Engdahl O, Toft T, Boe J, et al. Chest radiograph: a poor method for determining the size of a pneumothorax. Chest 1993; 103:26-29

Manuscript received July 20, 2004; revision accepted January 31, 2005.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (e-mail: www.chestjournal.org/misc/reprints.shtml).

Correspondence to: Robert J. Cerfolio, MD, FCCP, Chief of Thoracic Surgery, Division of Cardiothoracic Surgery, University of Alabama at Birmingham, 1900 University Blvd, THT 712, Birmingham, AL 35294; e-mail: Robert.cerfolio@ccc.uab.edu

* From the Departments of Epidemiology (Ms. Bryant), Radiology (Dr. Singh), and Biostatistics (Dr. Bartolucci), and Division of Cardiothoracic Surgery (Dr. Cerfolio and Ms. Bass) Department of Surgery, University of Alabama at Birmingham, Birmingham, AL.

COPYRIGHT 2005 American College of Chest Physicians
COPYRIGHT 2005 Gale Group