Objective: To determine whether the frequency of pneumothorax is incre
ased after thoracentesis in chronic obstructive pulmonary disease pati
ents. Design: Prospective. Setting: Northeastern community hospital. P
atients: One-hundred-six patients underwent multiple thoracentesis; 36
patients had co-existent chronic obstructive pulmonary disease (mean
age, 68.2 years). Chronic obstructive pulmonary disease was identified
by radiologic findings consistent with chronic obstructive pulmonary
disease and as forced expiratory volume in one second less than 70% of
predicted. Outcome Measures: Identification of pneumothorax on chest
roentgenogram after thoracentesis. Intervention: Patients had diagnost
ic or therapeutic thoracentesis with follow-up chest radiography withi
n 2 hours, or sooner if clinically indicated. Results: Patients with c
hronic obstructive pulmonary disease had a higher incidence of pneumot
horax (15 of 36 patients; 41.7%) than those patients without underlyin
g chronic obstructive pulmonary disease (13 of 70 patients; 18.5%; p =
0.005). The frequency of pneumothorax was no different according to w
ho performed the procedure (house staff or pulmonologist), whether it
was for diagnostic or therapeutic reasons, and whether a small (<500 m
l) or large (>500 ml) amount of fluid is removed. Conclusion: Pneumoth
orax may frequently occur in patients with chronic obstructive pulmona
ry disease undergoing thoracentesis. The reason may be related to the
altered architecture of the lung parenchyma and the change in mechanic
al forces in chronic obstructive pulmonary disease. Sonography-guided
thoracentesis may offer a safer means of performing thoracentesis in p
atients with chronic obstructive pulmonary disease.