Pulmonary artery catheterization. Can there be an integrated strategy for guideline development and research...

To review the evidence regarding indications for pulmonary artery catheterization (PAC) in critically ill patients, and to propose a guideline-generating process that would encourage randomized controlled trials of PAC.

Computerized and manual search for randomized trials involving PAC as an integral part of the protocol, published as of October 1992. Manual search for guidelines endorsed by specialty societies and/or proposed by expert panels following an explicit group process.

Four trials suggest benefit from PAC for pre- or perioperative management of high-risk surgical patients. Two others indicate that PAC for low-risk aortic aneurysm surgery confers no advantages. Use of PAC to guide therapy aimed at improved mixed venous oxygen saturation was beneficial compared with usual care with PAC in two small trials among patients with septic shock and severe trauma. Other PAC indications are either untested or inadequately tested by small trials. Large trials are needed, but trials to date have been impeded by clinicians' uncertainty about PAC and unwillingness to randomize critically ill patients. No published guidelines for PAC have used a formal group process and/or a hierarchical review of evidence to demarcate proven from unproven indications.

We propose a research-promoting expert panel on indications for PAC. Aided by a critical literature review, experts would rate case scenarios on the need for routine PAC with or without a specified intervention strategy. Future trials should test indications where there is either consensus about the uncertainty of need for PAC, or interpanelist disagreement owing to inconclusive evidence. This process could facilitate practice guideline development, utilization management, and large trials of PAC and related interventions.