We welcome the interest generated by our paper. Snyder and colleagues question our assertion that the standard medical practice of obtaining a posteroanterior chest radiograph after thoracentesis opposes the 1988 American Thoracic Society guideline [1], which states that "a chest film should be performed after therapeutic thoracentesis in most instances." We chose to emphasize that the guideline does not mention diagnostic thoracentesis at all. Further, we interpret the phrase "in most instances" to mean "in most, but not all, cases." We believe that these points were worth highlighting.

Snyder and colleagues are also concerned that the current U.S. legal system would seek liability if radiographic screening had not been done after thoracentesis and an episode of unsuspected pneumothorax resulting in excess morbidity or mortality then occurred. We believe that each physician, not the court system, must define his or her comfort threshold for accepting the risk of failing to diagnose a pneumothorax. As we concluded in our report, however, by combining our data with those from Collins and Sahn [2] and Gerardi and associates [3], we found that no patient with unsuspected pneumothorax had serious clinical consequences. On the basis of the data from this large aggregate patient group, we believe that our legal system would support our recommendations.

We agree with Drs. Brown and Blair that in certain cases, post-thoracentesis chest radiography is done for reasons other than to identify complications. Both pointed out that such cases usually involve therapeutic thoracenteses of large pleural effusions, for which a new "baseline" radiograph is needed for future comparison. Our study, which was designed to evaluate the necessity of the immediate (within 4 hours) postprocedure chest radiograph for identifying complications, did not intend to imply that all post-thoracentesis chest radiographs lacked clinical value.