INTRODUCTION
Background: Current American and British Pulmonology Guidelines advise immediate interventional management of most large (>2 cm at hilum; >3 cm at apex) primary spontaneous pneumothoraces.Reference MacDuff1,Reference Baumann, Strange and Heffner2
Objectives: Is observation of moderate-to-large primary spontaneous pneumothoraces (> 32% by Collins method) noninferior to immediate interventional management assessed by complete lung re-expansion within 8 weeks.
METHODS
Design
Multicenter randomized, open-label, noninferiority trial, with 9% noninferiority margin to demonstrate conservative management is not inferior to interventional treatment.
Setting
Thirty-nine hospitals in Australia and New Zealand.
Subjects
Patients aged 14–50 years with first-time unilateral primary spontaneous pneumothorax of ≥32%.
Intervention
Conservative (repeat imaging after 4 hours, discharge if no supplemental oxygen required; chest tube inserted if clinically unwell) v. intervention (chest tube (≤12 F) with suction; remove and discharge if resolved on chest x-ray at 4 hours or admission).
Outcomes
The primary noninferiority outcome was complete radiographic resolution of the pneumothorax 8 weeks after randomization. Numerous secondary outcomes.
RESULTS
A total of 316 of the 2,637 screened patients underwent randomization; 154 patients to the intervention group, and 162 in the conservative-management group. The prespecified noninferiority margin was (-9%) for re-expansion by 8 weeks. The authors based this on an expected resolution rate of 99% for the intervention group and believed a failure rate of 1 in 10 would be clinically acceptable. A margin this size may bias toward finding statistical noninferiority where there is clinical inferiority.
In the Complete Case analysis, the resolution rate in the intervention group was 129/131 (98.5%). Five patients were lost to follow-up, and 18 did not have chest radiography data available at 8 weeks. In the conservative group, 118/125 (94.4%) had a resolved pneumothorax on chest radiography at 8 weeks, 3 patients were lost to follow-up, and 34 did not have complete data. The Risk Difference was −4.1%, 95% CI −8.6 to 0.5, p = 0.02.
APPRAISAL
Strengths
• First study to assess conservative v. interventional therapy in moderate-large primary pneumothoraces
• Used both clinician and patient-centered secondary outcomes that can impact a shared decision-making model
• In addition to a complete case analysis, they also performed two sensitivity analyses, including one in which missing data were imputed as treatment failure (i.e. worst-case scenario).
Limitations
• This study was unblinded, which could bias findings (treating physicians were more likely to report radiographic resolution of pneumothorax in the intervention group when compared with blinded radiologists)
• The two sensitivity analyses identify potential statistical fragility in the main findings
• A shorter-term primary outcome might reveal greater disparity between conservative v. treatment groups
• The noninferiority margin of 9% is large
• 15.4% of patients (25) in the conservative treatment group required intervention
• 19% of randomized patients (60) lost to follow-up or had missing outcome data
• Radiographic resolution is not a patient-centered outcome.
CONTEXT
Conservative management has been the standard of care for small primary spontaneous pneumothoraces for many years. The British Thoracic SocietyReference MacDuff1 and the American College of Chest PhysiciansReference Baumann, Strange and Heffner2 both currently recommend that clinically stable patients with large primary spontaneous pneumothorax undergo interventional management with a small-bore catheter or chest tube, followed by hospitalization in many cases. In Canada, these patients are commonly treated with a pigtail catheter and discharged home from the emergency department (ED). This study is a step toward validating an observational approach to large pneumothoraces in a carefully selected population.
BOTTOM LINE
Although this study contains some statistical fragility, there is modest evidence that a conservative approach may be noninferior to intervention in carefully selected patients with first time, moderate to large spontaneous pneumothorax. While many Canadian ED physicians may choose a pigtail catheter, any form of tube thoracostomy has potential adverse consequences. This study may identify an alternative option that can be used in shared decision-making with the appropriately selected patient.
You have
Access
INTRODUCTION
Background: Current American and British Pulmonology Guidelines advise immediate interventional management of most large (>2 cm at hilum; >3 cm at apex) primary spontaneous pneumothoraces.Reference MacDuff1,Reference Baumann, Strange and Heffner2
Objectives: Is observation of moderate-to-large primary spontaneous pneumothoraces (> 32% by Collins method) noninferior to immediate interventional management assessed by complete lung re-expansion within 8 weeks.
METHODS
Design
Multicenter randomized, open-label, noninferiority trial, with 9% noninferiority margin to demonstrate conservative management is not inferior to interventional treatment.
Setting
Thirty-nine hospitals in Australia and New Zealand.
Subjects
Patients aged 14–50 years with first-time unilateral primary spontaneous pneumothorax of ≥32%.
Intervention
Conservative (repeat imaging after 4 hours, discharge if no supplemental oxygen required; chest tube inserted if clinically unwell) v. intervention (chest tube (≤12 F) with suction; remove and discharge if resolved on chest x-ray at 4 hours or admission).
Outcomes
The primary noninferiority outcome was complete radiographic resolution of the pneumothorax 8 weeks after randomization. Numerous secondary outcomes.
RESULTS
A total of 316 of the 2,637 screened patients underwent randomization; 154 patients to the intervention group, and 162 in the conservative-management group. The prespecified noninferiority margin was (-9%) for re-expansion by 8 weeks. The authors based this on an expected resolution rate of 99% for the intervention group and believed a failure rate of 1 in 10 would be clinically acceptable. A margin this size may bias toward finding statistical noninferiority where there is clinical inferiority.
In the Complete Case analysis, the resolution rate in the intervention group was 129/131 (98.5%). Five patients were lost to follow-up, and 18 did not have chest radiography data available at 8 weeks. In the conservative group, 118/125 (94.4%) had a resolved pneumothorax on chest radiography at 8 weeks, 3 patients were lost to follow-up, and 34 did not have complete data. The Risk Difference was −4.1%, 95% CI −8.6 to 0.5, p = 0.02.
APPRAISAL
Strengths
• First study to assess conservative v. interventional therapy in moderate-large primary pneumothoraces
• Used both clinician and patient-centered secondary outcomes that can impact a shared decision-making model
• In addition to a complete case analysis, they also performed two sensitivity analyses, including one in which missing data were imputed as treatment failure (i.e. worst-case scenario).
Limitations
• This study was unblinded, which could bias findings (treating physicians were more likely to report radiographic resolution of pneumothorax in the intervention group when compared with blinded radiologists)
• The two sensitivity analyses identify potential statistical fragility in the main findings
• A shorter-term primary outcome might reveal greater disparity between conservative v. treatment groups
• The noninferiority margin of 9% is large
• 15.4% of patients (25) in the conservative treatment group required intervention
• 19% of randomized patients (60) lost to follow-up or had missing outcome data
• Radiographic resolution is not a patient-centered outcome.
CONTEXT
Conservative management has been the standard of care for small primary spontaneous pneumothoraces for many years. The British Thoracic SocietyReference MacDuff1 and the American College of Chest PhysiciansReference Baumann, Strange and Heffner2 both currently recommend that clinically stable patients with large primary spontaneous pneumothorax undergo interventional management with a small-bore catheter or chest tube, followed by hospitalization in many cases. In Canada, these patients are commonly treated with a pigtail catheter and discharged home from the emergency department (ED). This study is a step toward validating an observational approach to large pneumothoraces in a carefully selected population.
BOTTOM LINE
Although this study contains some statistical fragility, there is modest evidence that a conservative approach may be noninferior to intervention in carefully selected patients with first time, moderate to large spontaneous pneumothorax. While many Canadian ED physicians may choose a pigtail catheter, any form of tube thoracostomy has potential adverse consequences. This study may identify an alternative option that can be used in shared decision-making with the appropriately selected patient.
Competing interests
None declared.