Objective To review the pulmonary artery catheter (PAG) use, examining mortality, SICU length of stay, and demographics of patients with and without PACs in comparison to those of the recent SUPPORT study on right heart catheterization effectiveness . Design Retrospective analysis of data extracted from a comprehensive database archived from a computerized ICU clinical record. Setting Surgical intensive care unit (SICU) of a tertiary university medical center. Subjects All SICU patients admitted over a 5 month period (June–October 1996). Results During the period under analysis, 554 patients were admitted to the SICU. Fifty-seven percent were male and 43% were female. The average age was 56 (range 16–100). Of the 554 patients, 141 patients received 202 PACs (25% of admissions). Fifty-eight of these patients were trauma patients, most of whom did not have ARDS or MOSF. This is in contrast to SUPPORT trauma patients, in whom the presence of ARDS or MOSF was a requirement for study entry . The average duration of PAC insertion was 2.6 days (range 0–11.7 days). The average SICU length of stay (LOS) was 8.9 ± 11.6 (median = 4) days (range 0–66 days). This compares favorably with both the SUPPORT PAC patients (mean ICU LOS = 14.8 days, 25th percentile = 5 days, median = 9 days) and the SUPPORT non-PAC patients (mean ICU LOS = 13 days, 25th percentile = 4 days, median = 7 days). Forty-seven of the 554 patients died in the SICU (overall mortality 8.5%), but 26 of the 141 SICU PAC patients died (18.4%). Yet this compares quite favorably with not only the SUPPORT PAC patients (37.5%, 30-day mortality, P < 0.00001 by χ2 analysis) but even the SUPPORT non-PAC patients (32.8% 30-day mortality, P < 0.001). Indeed, the difference in mortality between the SICU PAC patients and either SUPPORT group (PAC or non-PAC) was greater than the reported minor difference in mortalities between the PAC and non-PAC SUPPORT patients. Conclusion The recent SUPPORT study of the effectiveness of right heart catheterization purports to demonstrate an adverse outcome in patients in whom PACs were used. However, the original SUPPORT study population consisted of critically ill patients, most of whom had at least one established organ failure, and the use of PACs in acutely injured burn and trauma patients was specifically excluded. The better outcomes in our SICU PAC patients (even better than the SUPPORT non-PAC patients) could result from an intrinsically better prognosis due to lower acuities or from the proactive use of PAGs to forestall the onset of organ failures. A well conducted randomized, controlled clinical trial of PAC use in all ICU patients is clearly warranted. In the interim, we emphasize that it appears that the SUPPORT study does not represent PAC effectiveness in all ICU patients.