Using 94 flow-directed pulmonary artery catheters in 53 ICU patients, we obtained prospectively 282 pulmonary artery wedge pressure (WP) measurements from 286 attempts. After catheter manipulation, 96% of these WP measurements were ultimately free of technical problems, and 84% were confirmed by aspiration of pulmonary capillary blood from the wedge position. The 95% confidence interval for repeated measurements of WP in stable ICU patients was 4 mm Hg. The WP measurement error was defined as the difference between unconfirmed and confirmed WP pairs obtained from stable patients. The probability of encountering a WP measurement error of at least 4 mm Hg was 33% for the 93 WP measurements with technical problems, 5% for the 189 WP measurements without technical problems, and 14% for the entire set of 286 WP measurement attempts. Quality control procedures, including dynamic response testing, easily identified most errors.