We evaluated the Baxter/Edwards Critical-Care Swan Ganz IntelliCath continuous cardiac output catheter and Vigilance continuous cardiac output monitor in critically ill adult intensive care patients, and compared cardiac output measurements obtained from this new system with those from a standard bolus thermodilution technique using cold normal saline. Nine Swan Ganz IntelliCath catheters were inserted into patients selected at random, following the decision that pulmonary artery catheter monitoring was required. A total of 100 comparisons were made in nine patients. We found the Swan Ganz IntelliCath catheter to be both easy to insert and position, and simple to use. There were no complications or technical difficulties. The mean cardiac output measured by the continuous system was 8.00 l/min (SD 1.66/min) compared to 8.02 1/min (SD 1.74 l/min) using bolus techniques. Continuous cardiac output values correlated well with bolus measurements (r = 0.87), there was excellent accuracy (bias of 0.02 l/min) and precision was acceptable (limits of agreement of -1.71 to 1.75 l/min). These larger than anticipated limits of agreement may not reflect any inaccuracy in the continuous measurement system, rather they may highlight the extent of well recognised potential errors inherent in the intermittent bolus technique. The availability of a system to measure cardiac output continuously provides a major improvement in intensive care monitoring. This system will permit the rapid and accurate assessment of the response of patients to therapy while providing increased diagnostic facilities and a new research tool.