Objective: To determine the necessity of posteroanterior chest roentgenography for the identification of pneumothorax and other complications after thoracentesis.
Design: Prospective cohort study.
Setting: Tertiary care teaching hospital.
Patients: 67 men and 43 women (mean age ±SD, 62.4 ± 13.2 years). Exclusion criteria included age younger than 18 years, concurrent pleural biopsy, ultrasound guidance, and use of mechanical ventilation.
Measurements: 174 thoracenteses done between March 1991 and June 1993.
Results: 2 hemothoraces (1.2%) occurred, and 8 patients had a total of 9 pneumothoraces (5.2%). The roentgenograms obtained immediately after the procedures identified 8 pneumothoraces; the other pneumothorax was seen incidentally on a delayed roentgenogram obtained 3 days later. Pneumothorax was suspected in 5 of the 8 cases, and tube thoracostomy was done in 4 of these 5 cases. Patients with unsuspected pneumothorax identified on the roentgenogram obtained immediately after the procedure did not receive treatment for their pneumothoraces.
Univariate analysis showed that the variables that correlated significantly with pneumothorax were aspiration of air during the procedure (relative risk ratio, 12.3; 95% CI, 3.7 to 41.4), number of passes with the thoracentesis needle (relative risk ratio, 6.1; CI, 1.6 to 23.3), history of thoracic radiation therapy (relative risk ratio, 10.5; CI, 2.5 to 44.4), and operator suspicion of pneumothorax (relative risk ratio, 25.9; CI, 8.6 to 78.5).
Conclusion: Among hospitalized patients with pleural effusions, we identified a subgroup of patients in whom the risk for pneumothorax is low enough (approximately 1%) with sufficiently minimal clinical consequences to justify the avoidance of about 60% of chest roentgenograms obtained after thoracentesis. These patients are clinically stable, have not previously received chest irradiation, had only one pass at thoracentesis attempted without the aspiration of any air, and give no other indication of pneumothorax.