Salvage or planned esophagectomy after chemoradiation therapy for locally advanced esophageal cancer--a review.

Definitive chemoradiation (without esophagectomy) and neoadjuvant chemoradiation followed by planned esophagectomy are commonly used treatments for locally advanced esophageal cancer. These two treatment strategies have similar survival outcomes, so the value of planned esophagectomy is debated. However, persistence or recurrence of local disease is not uncommon after definitive chemoradiation. Salvage esophagectomy for isolated local failures of definitive chemoradiation is an option for selected patients. In this article we review the debate over definitive chemoradiation versus neoadjuvant chemoradiation and surgery, and then restate the argument in terms of salvage versus planned esophagectomy. Although both forms of esophagectomy are done in the setting of previous chemoradiation, they are different in several ways. Salvage esophagectomy appears to be a more morbid operation than planned esophagectomy. Surgeons supportive of the salvage esophagectomy strategy face the challenge of reducing its postoperative mortality.

Effect of fibrin glue on air leak and length of hospital stay after pulmonary lobectomy.

AIM: Air leaks are a common cause of morbidity and prolonged hospital stay after pulmonary lobectomy. We reviewed our experience with intraoperative fibrin glue to determine if it reduced air leak and improved patient outcomes. METHODS: Records of patients undergoing pulmonary lobectomy for benign or malignant disease over a 4-year period (1998-2001) were reviewed. Data was collected on age, sex, pulmonary function, pulmonary pathology, use of fibrin glue, duration of chest tube drainage, length of hospital stay, and postoperative complications. RESULTS: Three hundred and sixty patients underwent lobectomy. Fibrin glue was used intraoperatively to seal air leaks in 102 of the 360 patients (study group: 102;control group: 258). Fibrin glue was used at the discretion of the surgeon, with some surgeons using it routinely. The groups did not differ in age (p=0.29), sex (p=0.42), FEV1 (p=0.57), or pathology (p=0.08). There were no differences in outcomes such as operative mortality (study: 2 of 102, control 6 of 258, p=0.85), empyema (study: 0 of 102, control: 3 of 258, p=0.55), prolonged (>7 days) air leaks (study: 10 of 20; control: 20 of 258, p=0.71), or length of hospital stay (study: 6.3+/-2.5 days, control:7.7+/-7.2 days, p=0.83). The use of fibrin glue was associated with a reduction in the duration of chest tube intubation (study: 4.1+/-3.2 days, control: 5.5+/-3.8 days, p=0.001). CONCLUSION: Patients treated intraoperatively with fibrin glue had a significantly shorter duration of chest tube intubation after pulmonary lobectomy than those treated conventionally. However, the use of fibrin glue did not significantly influence more clinically relevant outcomes such as length of hospital stay and incidence of prolonged (>7 days) air leaks.

Pulmonary resection in metastatic uterine and cervical malignancies.

OBJECTIVE: Although thoracotomy for removal of pulmonary metastasis is well documented in a wide variety of solid tumors, data are sparse regarding management of patients with gynecologic malignancies metastatic to the lung. METHODS: We retrospectively reviewed the Roswell Park Cancer Institute experience between 1982 and 1999. Of 82 eligible patients with gynecologic tumors metastatic and confined to the lung, 25 underwent pulmonary resection. RESULTS: There were 60 uterine and 22 cervix cancer patients with pulmonary metastases. Among patients with uterine cancer primaries undergoing pulmonary resection (n = 19) median survival was 26 months. Uterine cancer patients who underwent surgical resection for leiomyosarcomas (n = 11) had a median survival of 25 months compared to 46 months in patients with adenocarcinoma (n = 6, P = 0.02). Median survival in cervix cancer patients undergoing resection for pulmonary metastases (n = 6) was 36 months. CONCLUSIONS: Pulmonary resection may provide a survival advantage for selected patients with uterine and cervical malignancies with metastases isolated to the lung.

Survival after conservative (palliative) management of pleural malignant mesothelioma.

BACKGROUND AND OBJECTIVES: Malignant mesothelioma is a lethal disease. Aggressive multimodality treatment protocols are reportedly associated with improved survival, but the apparent survival benefits may simply reflect patient selection and the variable natural history of this malignancy. Before embarking on our own protocol of experimental treatment for mesothelioma, we sought to identify important prognostic factors and document the survival of patients treated conservatively (with palliative intent only) in our region. METHODS: We performed a retrospective review of all patients with a diagnosis of malignant mesothelioma seen at our center between 1987 and 1999. Since curative intent treatment had not been given, we assumed that measured survival would largely reflect the natural history of the malignancy. RESULTS: There were 101 patients (80 males and 21 females). Mean age was 65 +/- 9.2 years. Symptoms of disease were present for a median time of 5 months before the diagnosis was established. The most common presenting symptoms were dyspnea (46 patients), chest pain (30 patients), and weight loss (22 patients). Sixty-eight patients (68%) had a history of asbestos exposure. Mesothelioma subtypes included epithelial (43 patients), sarcomatous (26 patients), mixed (19 patients), desmoplastic (4 patients), and unspecified (9 patients). All 101 patients were treated with palliative intent. Talc pleurodesis was performed in 70 patients. At the time of analysis, 90 patients had died and 11 remained alive. Median survival was 213 (95% CI 137-289) days. Survival for the three major histological subtypes was significantly different (log rank, P = 0.0016). Histological subtype (epithelial favorable) was the only significant independent prognostic factor (Cox proportional hazard regression, P = 0.0009). CONCLUSIONS: Patients with epithelial mesothelioma survive longer than those with other histological subtypes. Conservatively managed patients with pleural malignant mesothelioma have a median survival of approximately 7 months. These data from conservatively treated patients can serve as baseline information for future studies of experimental treatments.

Does the interponat affect outcome after esophagectomy for cancer?

Clinical decision-making in esophageal cancer surgery is a process of balancing the risks of treatment against potential benefits, such as survival and quality of life. Various options are available for esophageal reconstruction. While these reconstructive options do not directly have an impact on cancer survival, they do affect operative morbidity and long-term quality of life. The affect of various interponats (reconstructive conduits) and routes of reconstruction on operative morbidity and foregut function is reviewed. Gastric interponats are preferred for esophageal reconstruction because of their reliable vascularity and the relative simplicity of the reconstructive operation. Colon interponats supposedly provide better long-term function as an esophageal substitute (unproven), but at the cost of increased operative complexity and morbidity. Colon interposition is therefore reserved for situations in which gastric transposition is not feasible. Both posterior and anterior mediastinal routes of gastric interponat reconstruction are acceptable (meta-analysis of randomized controlled trials). Posterior mediastinal reconstruction is usually preferred when a complete (R0) resection has been accomplished. Anterior mediastinal reconstruction may prevent secondary dysphagia after incomplete (R1, R2) resections.

Self-inflicted pneumothoraces.

Self-inflicted pneumothoraces are rare manifestations of psychiatric illness. Two patients with self-inflicted pneumothoraces are reported, and the typical clinical features of factitious disorders are described. If thoracic surgeons are aware of these conditions, inappropriate surgery- and poor outcomes-can be avoided.

Pericardial catheter sclerosis versus surgical procedures for pericardial effusions in cancer patients.

BACKGROUND: Approximately 21% of patients with advanced malignancies have cardiac or pericardial involvement with tumor. Controversy exists regarding the optimal approach to the pericardial space when hemodynamic compromise due to effusions occurs. METHODS: A six-year retrospective review of 59 cancer patients with pericardial effusions. RESULTS: Thirty-six patients had subxiphoid pericardial window (SXPW) alone (Group A), 5 had pericardial catheter drainage (PCD) followed by a SXPW (Group B), 10 had PCD with sclerosis (Group C), 5 had PCD alone (Group D), 2 had PCD with pericardial-pleural window (Group E), and one had pericardial-peritoneal window (Group F). The method of procedure, complications, number of hospital and ICU days, cytological or pathologic evidence of malignancy, solid versus hematological tumors, and survival were analyzed. The median survival for those patients in group C was one month compared to 4 months for Group A and 6 months for Group B. Essentially, results were similar regardless of method performed with the exception that professional and hospital charges averaged $4830 for SXPW compared to $1625 for PCD. CONCLUSIONS: Pericardial catheter drainage and sclerosis provides a viable option for the treatment of pericardial effusions in selected cancer patients at markedly reduced cost and patient discomfort.

Prophylactic intracavitary (pneumonectomy space) antibiotic instillation: a comparative study.

BACKGROUND AND OBJECTIVES: Postpneumonectomy empyema is a dreaded complication of pneumonectomy. The effectiveness of prophylactic intracavitary antibiotic instillation is not known. We conducted a retrospective review to assess the effect of pneumonectomy space antibiotic instillation on septic complications (empyema and bronchial fistula) of pneumonectomy. METHODS: Ninety-three consecutive patients underwent pneumonectomy at our institution over a three-year period. Their charts were reviewed retrospectively and data was collected on age, gender, diagnosis, intravenous antibiotics, intracavitary (pneumonectomy space) antibiotics, empyemas, bronchial fistulas, length of hospital stay, and operative mortality. RESULTS: All 93 patients received 3 perioperative doses of prophylactic intravenous antibiotics. One group (n=47) of patients also received intraoperative intracavitary instillation of an antibiotic solution (penicillin G: 5 million units, bacitracin: 50,000 units, gentamicin: 60 mg, in 1 litre of saline) while the other group (n=46) did not. Age, gender, diagnosis, and length of stay were not significantly different in the two groups. There were no empyemas or bronchial fistulas in the intracavitary antibiotic group. Postpneumonectomy empyemas occurred in 6 (13%) patients (empyema with bronchial fistula: 5, empyema alone: 1) that had not received intracavitary antibiotics (p=0.012). There were 4 deaths (9%) in each group (p=0.63). CONCLUSIONS: Prophylactic intraoperative intracavitary antibiotic instillation may reduce the incidence of empyemas after pneumonectomy. However, a randomized trial would be needed to prove the effectiveness of this form of prophylactic antibiotic strategy.

Local recurrence after total or subtotal esophagectomy for esophageal cancer.

Cancer recurrence is a common problem after esophagectomy for esophageal cancer. Local recurrence is especially problematic because it often negates the palliative benefit of esophagectomy. We conducted a retrospective review to assess the effect of extent of esophageal resection (subtotal or total esophagectomy) on local cancer recurrence. Seventy-four consecutive patients with esophageal cancer underwent esophagectomy at our institution over a four-year period. Their charts were reviewed retrospectively and data was collected on age, gender, histology, stage, tumor location, operation, resection margin status, anastomotic leaks, operative mortality, adjuvant therapy, cancer survival, and local recurrence. Total esophagectomy was done in 19 patients (transhiatal - 3; McKeown - 16) and subtotal esophagectomy was done in the other 55 patients (Lewis - 25; left thoracoabdominal - 30). The two groups were similar with respect to age, gender, histology, stage, anastomotic leaks, operative mortality, adjuvant therapy, and overall survival. Resection margins were positive for residual tumor in 2 out of 19 (11%) total esophagectomies and 9 out of 55 (16%) subtotal esophagectomies (p=0.42). Local recurrence occurred in 3 of 19 (16%) patients treated with total esophagectomy and 23 out of 55 (42%) patients treated with subtotal esophagectomy (p=0.04). We conclude that total esophagectomy is associated with fewer local cancer recurrences than subtotal esophagectomy. We, therefore, recommend total esophagectomy for the surgical treatment of esophageal cancer.

Anastomotic leaks after esophagectomy for esophageal cancer: a comparison of thoracic and cervical anastomoses.

BACKGROUND AND OBJECTIVES: Esophagogastric anastomotic leaks remain a significant problem after esophagectomy for esophageal cancer. Many investigators have reported that leaks are more frequent after cervical, as opposed to thoracic, esophagogastric anastomoses. We conducted a retrospective review to assess the effect of anastomotic location (thoracic or cervical) on anastomotic leak incidence and severity. METHODS: Seventy-four consecutive patients with esophageal cancer underwent esophagectomy and esophagogastric anastomoses at our institution over a four-year period. Their charts were reviewed retrospectively and data was collected on age, gender, histology, stage, resection margin status, adjuvant therapy, cancer survival, anastomotic location, anastomotic leaks, and operative mortality. RESULTS: Cervical anastomoses were done in 19 patients and thoracic anastomoses were done in the other 55 patients. The two groups were similar with respect to age, gender, histology, stage, adjuvant therapy, and overall survival. Operative mortality for the entire group of 74 patients was 4% (3 patients). Resection margins were positive for residual tumor in 2 of 19 (11%) patients with cervical anastomoses and 9 of 55 (16%) patients with thoracic anastomoses (p=0.42). Leaks complicated 1 of 19 (5%) cervical and 9 of 55 (16%) thoracic esophagogastric anastomoses (p=0.21). Positive resection margins and anastomotic leaks were not significantly related (p=0.54). One of 9 (11%) leaks in the thoracic group proved fatal. CONCLUSIONS: In our experience cervical esophagogastric anastomoses do not have a higher incidence of leaks than thoracic anastomoses.

Open lung biopsy as an outpatient procedure.

BACKGROUND: Lung biopsies are frequently needed to diagnose diffuse interstitial lung diseases. Both limited thoracotomy (open lung biopsy) and thoracoscopy can be used for lung biopsies, but both procedures have traditionally required hospital admission. We report a series of patients that underwent outpatient open lung biopsy to show the safety and effectiveness of this practice. METHODS: We reviewed records of ambulatory, nonoxygen dependent patients with a clinical diagnosis of diffuse interstitial lung disease that underwent outpatient open lung biopsy between January 1997 and December 1999. All procedures were done by a senior surgeon using single lumen endotracheal anesthesia, a small anterolateral thoracotomy without rib spreading, stapled wedge resection, and no chest tube. Patients were discharged the same day. RESULTS: Thirty-two patients with a clinical diagnosis of diffuse interstitial lung disease underwent outpatient open lung biopsy. Mean age was 58 years (range, 21 to 74 years). Preoperative forced expiratory volume in 1 second was 74.3%+/-7.0% of predicted. A pathologic diagnosis was established in all patients: usual interstitial pneumonia, 26 patients; sarcoidosis, 2; metastatic carcinoma, 2; desquamative interstitial pneumonia, 1; and mixed dust pneumoconiosis, 1 patient. No patient required a chest tube, overnight observation, or hospital admission. No complications occurred. CONCLUSIONS: Selected patients with a clinical diagnosis of diffuse interstitial lung disease can safely and effectively undergo diagnostic outpatient open lung biopsy. However, careful patient selection and attention to operative detail are essential.

The effect of cardiopulmonary lymphatic obstruction on heart and lung function.

Many thoracic surgical procedures involve excision or destruction of intrathoracic and mediastinal lymphatics. It is widely assumed that the mediastinal lymphatic system is surgically expendable, and that destruction of mediastinal lymphatics has no significant adverse physiological effect. Cardiac lymphatic obstruction may give rise to cardiac lymphedema and impaired cardiac function. Similarly, obstruction of pulmonary lymphatics may result in pulmonary perivascular lymphedema, endothelial injury, and pulmonary artery hypertension. This review summarizes the possible deleterious effects of intrathoracic lymphatic destruction and the benefits of pharmacological and surgical enhancement of active lymph drainage.

Duration of knowledge in general thoracic surgery.

BACKGROUND: Medical knowledge changes rapidly, so current medical education approaches emphasize the development of life-long learning skills ("teaching the learner to learn") as opposed to the simple acquisition of contemporary medical knowledge. Because there are no data on the rapidity of change of general thoracic surgical knowledge, we do not know whether this trend in medical education is appropriate for thoracic surgical trainees. We undertook a study to assess the duration of knowledge in general thoracic surgery. METHODS: The first general thoracic surgery article from each issue of The Annals of Thoracic Surgery between 1965 and 1997 was abstracted into a summary statement. A form, made up of 360 summary statements in random order, was assessed by 6 general thoracic surgeons. They assessed statement validity on a 5-point scale (1 = statement false; 5 = statement true). Average statement validity scores for 30 time intervals were calculated. The relationship between time of publication and statement validity was analyzed. RESULTS: Average validity scores ranged from 2.24 (represents 1965 to 1966) to 4.32 (represents 1969 to 1970). Validity scores increased with time (y = 3.46 + 0.017x, where y is validity score and x is time), and this was significant (r = 0.40; p = 0.027). However, the absolute change in average validity scores over the 33-year study period was only 0.52 or 13.1% of the "modern" era scores. CONCLUSIONS: The assumption that medical knowledge changes quickly may not be true in general thoracic surgery. Although life-long learning skills are important, general thoracic surgery training programs should continue to emphasize fundamental knowledge in the specialty.

Handsewn or stapled esophagogastric anastomoses after esophagectomy for cancer: meta-analysis of randomized controlled trials.

Gastric transposition with esophagogastric anastomosis is a common method of reconstruction after esophagectomy for cancer. The anastomosis can be fashioned using a handsewn or stapled technique. The choice of anastomotic technique is often debated but there is little evidence to support the use of one method over the other. We performed a meta-analysis of randomized controlled trials (RCTs) to determine the effect of esophagogastric anastomotic method (handsewn or circular stapled) on patient outcomes. Medline and manual searches were done (completed independently and in duplicate) to identify all published RCTs that addressed the issue of handsewn or stapled esophagogastric anastomosis after esophagectomy for cancer. The selection process was inclusive; no trials were excluded. Trial validity assessment was done and a trial quality score was assigned. Major outcomes for quantitative data synthesis included operative mortality, anastomotic leaks, anastomotic strictures, cardiac morbidity, and pulmonary morbidity. A random-effects model was used and relative risk was the principal measure of effect. Systematic qualitative review was used for other outcomes such as duration of operation and time to complete the anastomosis. Data on cancer survival were not available in the RCTs. Five RCTs were selected with quality scores ranging from 2 to 3 (5-point Jadad scale). Selection and validity agreement was strong. Relative risk (95% confidence interval, CI; P-value), expressed as handsewn vs. stapled (treatment vs. control), was 0.45 (0.20, 1.00; P=0.05) for operative mortality, 0.79 (0.44, 1.42; P=0.43) for anastomotic leaks, 0.60 (0.27, 1.33; P=0.21) for anastomotic strictures, 0.99 (0.55, 1.77; P=0.97) for cardiac morbidity, and 0.93 (0.63, 1.37; P=0.72) for pulmonary morbidity. Data synthesized from existing RCTs show that handsewn and circular stapled esophagogastric anastomotic techniques give similar results for anastomotic outcomes, such as leaks and strictures. The stapled anastomotic method appears to increase operative mortality (P=0.05). Although it is difficult to explain this finding, it should not be dismissed. Several hypotheses are discussed.

A meta-analysis of randomized controlled trials of route of reconstruction after esophagectomy for cancer.

BACKGROUND: A gastric conduit is usually used to reconstruct the foregut after esophagectomy for cancer. It can be transposed through a posterior or anterior mediastinal route. The choice of route is often debated but there is little evidence to support the use of one route over the other. We performed a meta-analysis of randomized controlled trials (RCTs) to determine the effect of route of reconstruction on patient outcomes. METHODS: Medline and manual searches were done (completed independently and in duplicate) to identify all published RCTs that addressed the issue of route of gastric conduit reconstruction after esophagectomy for cancer. The selection process was inclusive; no trials were excluded. Trial validity assessment was done and a trial quality score was assigned. Major outcomes for quantitative data synthesis included operative mortality, anastomotic leaks, cardiac morbidity, and pulmonary morbidity. A random-effects model was used and relative risk was the principal measure of effect. Systematic qualitative review was used for other outcomes such as duration of ventilation, length of hospital stay, operative blood loss, duration of surgery, anastomotic strictures, dysphagia, gastric emptying, and quality of life. Data on cancer survival were not available in the RCTs. RESULTS: Six RCTs were selected with quality scores ranging from 1 to 4 (5-point Jadad scale). Selection and validity agreement was strong. Relative risk (95% confidence interval; P value), expressed as posterior versus anterior mediastinal route (treatment versus control), was 0.56 (0.17, 1.82; P = 0.34) for mortality, 1.01 (0.35, 2.94; P = 0.98) for leaks, 0.43 (0.17, 1.12; P = 0.08) for cardiac complications, and 0.67 (0.34, 1.33; P = 0.26) for pulmonary complications. Systematic qualitative review did not suggest any difference in other perioperative outcomes or conduit function for the two routes of reconstruction. CONCLUSIONS: Data synthesized from existing RCTs show that posterior and anterior mediastinal routes of reconstruction are associated with similar outcomes after esophagectomy for cancer. However, a difference in outcomes for the two reconstructive routes remains possible. Further trials with larger numbers of patients are needed.

Effect of pyloric drainage on the healing of esophagogastric anastomoses in rats.

BACKGROUND AND OBJECTIVES: Esophagogastric anastomotic leaks complicate 5% to 20% of esophagectomies for esophageal cancer and are responsible for approximately one-third of perioperative deaths. Poor gastric emptying is a predisposing factor for anastomotic leakage. An animal experiment was used to test the hypothesis that a pyloric drainage procedure (pyloromyotomy) would have a positive effect on esophagogastric anastomotic healing. METHODS: In 40 rats single-layer esophagogastric anastomoses were constructed with interrupted 7-0 polypropylene sutures. A pyloromyotomy was done in the experimental group (20 rats) but not in the control group (20 rats). Rats were sacrificed on the 7th postoperative day and their anastomoses were excised, mounted in a tensiometer, and distracted at 10 mm/min to measure breaking strength. After that, the hydroxyproline concentration (an indicator of wound collagen) of the anastomotic tissue was measured. RESULTS: There were no anastomotic leaks. The mean (and standard deviation) breaking strength of the esophagogastric anastomosis was 3.96 (1.14) N in the pyloromyotomy rats and 4.11 (0.75) N in the control rats (p = 0.64). The mean (and SD) hydroxyproline concentration in esophagogastric anastomotic tissue was 368.6 (31.5) nmol/mg in the pyloromyotomy rats and 376.6 (31.3) nmol/mg in the control rats (p = 0.77). CONCLUSION: Pyloric drainage (pyloromyotomy) did not have any effect on esophagogastric anastomotic wound healing in this rat model.

A randomized, controlled trial comparing thoracoscopy and limited thoracotomy for lung biopsy in interstitial lung disease.

BACKGROUND: Lung biopsies are frequently needed to diagnose diffuse interstitial lung diseases. A prospective randomized, controlled trial comparing limited thoracotomy (open lung biopsy) and thoracoscopy for lung biopsy was done. METHODS: Ambulatory patients with a clinical diagnosis of diffuse interstitial lung disease were randomized to thoracoscopy or limited thoracotomy. Data on postoperative pain, narcotic requirements, operating room time, adequacy of biopsy, duration of chest tube drainage, length of hospital stay, spirometry, and complications were collected. RESULTS: A total of 42 randomized patients underwent lung biopsy (thoracoscopy 20, thoracotomy 22). The two study groups were comparable with respect to age, gender, corticosteroid use, and preoperative spirometry. Visual analog scale pain scores were nearly identical in the two groups (p = 0.397). Total morphine dose was 50.8 +/- 27.3 mg in the thoracoscopy group and 52.5 +/- 25.6 mg in the thoracotomy group (p = 0.86). Spirometry (FEV1) values in the two groups were not significantly different on postoperative days 1, 2, 14, and 28 (p = 0.665). Duration of operation was similar in both groups (thoracoscopy 40 +/- 30 minutes, thoracotomy 37 +/- 15 minutes; p = 0.67). The thoracoscopy and thoracotomy groups had equivalent duration of chest tube drainage (thoracoscopy 38 +/- 28 hours, thoracotomy 31 +/- 26 hours; p = 0.47) and length of hospital stay (thoracoscopy 77 +/- 82 hours, thoracotomy 69 +/- 55 hours; p = 0.72). Definitive pathologic diagnoses were made in all patients. CONCLUSIONS: There is no clinical or statistical difference in outcomes for thoracoscopic and thoracotomy approaches. Both thoracoscopy and thoracotomy are acceptable procedures for diagnostic lung biopsy in diffuse interstitial lung disease.

Comparison of videothoracoscopy and axillary thoracotomy for the treatment of spontaneous pneumothorax.

Surgical treatment of spontaneous pneumothorax can be done through a thoracotomy or a video-thoracoscopic approach. Although the videothoracoscopic technique is currently popular it is not obviously superior to a more traditional axillary thoracotomy approach. We compared our recent experience with both techniques to determine the optimal surgical treatment for spontaneous pneumothoraces. A retrospective review of 79 patients treated surgically (34 thoracotomy and 45 thoracoscopy) for spontaneous pneumothoraces was done. Patients were treated between 1991 and 1997. Patients older than 60 years of age and those with spontaneous pneumothoraces secondary to generalized pulmonary emphysema were excluded. There were no operative deaths. Recurrence rate [thoracotomy, two of 34; thoracoscopy, three of 45 (P < 0.89)], air leak exceeding 7 days [thoracotomy, three of 34; thoracoscopy, three of 45 (P < 0.73)], operating room times [thoracotomy, 54 +/- 26 minutes; thoracoscopy, 53 +/- 16 minutes (P < 0.59)], and postoperative length of stay [thoracotomy, 5.7 +/- 4.3 days; thoracoscopy, 4.7 +/- 4.4 days (P < 0.26)] were not significantly different for the two techniques. We conclude that axillary thoracotomy and videothoracoscopy are equally effective surgical treatments for spontaneous pneumothoraces. A large randomized trial would be needed to determine whether one approach is truly superior to the other.