I would like to reply to the issues raised by Dr. Fee in his recent letter (July 2001).1I was impressed by his very low rate of pneumothorax. Since pneumothorax can only occur at an integral rate, a rate of only 0.025%, as reported by Dr. Fee in his letter, would imply an occurrence of five pneumothoracesis out of 20,000 procedures. By my calculation, it would require only 54.7 years to accumulate this series if one performed a single thoracentesis every day of the year with no time off for weekends or holidays, leap years not withstanding. If one took a slightly more conservative interpretation of his number, then the rate would be one out of 5,000. This would still require 13.7 years to accomplish at the rate described above. For the more usual 5-day work week, at a rate of one every day (without fail), this would still require 19.2 years. In our study2 at a relative large tertiary center with eight active pulmonologists, there were only slightly > 300 thoracenteses (including all inpatients and outpatients) performed in 2 years. We would have had to continue our study for 16 years to accumulate that series of 5,000. There can be only one conclusion. Dr. Fee does not have an accurate accounting of the number of thoracenteses he has performed. That very fact raises a significant question as to the validity of his estimation of the number of complications that have followed this purported number of thoracenteses. He does not and, in fact, can not provide any reliable data to support the implied safety of using a vacuum bottle for the removal of pleural fluid. Our study, however, clearly demonstrated the increased risk associated with the use of the vacuum bottle.