Lower risk and higher yield for thoracentesis when performed by experienced operators.

A position paper by the American College of Physicians states, " Thoracentesis is relatively uncomplicated technically, well tolerated, and quite safe. " This conclusion is reflected in current practice patterns in university hospitals, where house officers routinely perform thoracenteses. Prospective studies22 of house officer thoracenteses do not support the concept that the procedure is easy and benign; they have reported major complication rates ranging from 11.6 to 30.3 percent for thoracenteses performed without direct imaging guidance. Most of the studies reporting complication ratess have not specifically addressed the issue of the operator and have looked instead at different methods of performing thoracen-tesis, with the type of needle or catheter used the most frequent variable. Our goal was to evaluate the issue of the operator performing thoracenteses. We looked prospectively at " routine thoracenteses " performed by pulmonary fellows and pulmonologists at our institution and compared our complication rates with those reported in prospective studies of thora-centeses performed by house officers.22 Patients arid Criteria METHODS We studied the results of 50 consecutive thoracenteses. There were three major inclusion criteria: First, a thoracentesis could be included in the study only if a member of the pulmonary division (attending physician or pulmonary fellow) was the only operator; any procedure in which house officers played an operant role was ineligible. Second, each thoracentesis had to be done " blindly " at the bedside; thoracenteses for which an ultrasound or chest computed tomographic study was used to mark an entry site were excluded. Third, a follow-up chest radiograph had to be obtained; the presence or absence of postprocedure pneumothorax was a radiographic, not a clinical, determination. Pleural biopsies were included if they met other criteria. If the same patient had undergone two thoracenteses at different times, each thoracentesis was considered eligible for the study and was included as a separate event provided the criteria had been met. Two pulmonary fellows were involved in the study. Before the study, each had to first watch a pulmonary attending physician perform several thoracenteses, then to perform a minimum of six thoracenteses under direct supervision, and finally to be certified before being allowed to operate independently. For each thoracentesis, the following data were recorded: pain at the time of the procedure, pain a day after the procedure, local hematoma, cough, dyspnea right after and a day after the thoracen-tesis, pneumothorax, any other complications (eg, liver laceration, hemothorax), and any …