Pulmonary artery catheters or central venous catheters for cardiac surgery: the <scp>PUMA</scp> Pilot randomised clinical trial

INTRODUCTION: Pulmonary artery catheters are used widely in cardiac surgery despite observed associations with worse outcomes and guidelines that recommend against their routine use. No adequately powered randomised trials are available. METHODS: The PUMA Pilot was a multicentre, randomised, parallel assignment, open-label, pilot and feasibility trial conducted at three tertiary cardiac surgery centres. Eligible patients were adults undergoing coronary artery bypass grafting, aortic valve replacement or surgery on the aortic root or ascending aorta with or without aortic valve replacement, with a predicted surgical mortality of < 2%. Patients were allocated randomly to receive a pulmonary artery catheter or a central venous catheter inserted immediately before surgery. The primary feasibility outcome was protocol compliance, defined as receiving the assigned intervention without crossover. Secondary feasibility outcomes were eligibility rate; recruitment proportion and rate; data completeness; and rate of clinician refusal. RESULTS: We screened 480 patients and 206 (43%) were eligible; 150/203 (74%) approached provided informed consent. Three of 206 (1%) eligible patients were not included due to clinician refusal. Of 149 patients who were randomised, 76 were assigned to the pulmonary artery catheter group and 73 to the central venous catheter group. For the primary feasibility outcome, 147 patients (99%) received the allocated intervention. Data were complete for 144 (97%) patients. Median (IQR [range]) days alive and at home at 30 days was 23.7 (21.9-24.7 [7.0-26.0]) in the pulmonary artery catheter group and 22.9 (20.8-23.9 [8.6-25.8]) in the central venous catheter group. Acute kidney injury occurred in 26/76 (34%) patients in the pulmonary artery catheter group and 14/73 (19%) in the central venous catheter group. DISCUSSION: A randomised trial of pulmonary artery catheters compared with central venous catheters in low-risk cardiac surgery is feasible. Such a trial would address significant practice variability and inform international guidelines.