Pulmonary artery catheterization was adopted about 30 years ago and widely disseminated without rigorous evaluation as to whether it benefited critically ill patients. The technique is used to measure cardiac output and pressures in the pulmonary circulation to guide diagnosis and treatment. Clinicians believe these data can improve patients' outcomes, even in the absence of consensus about the specific interpretation of the data. To assess the effect of pulmonary artery catheterization on mortality and cost of care in adult intensive care patients. We searched the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, Issue 2, 2006); MEDLINE (all records to April 2006); EMBASE (all records to April 2006); CINAHL (all records to April 2006) and reference lists of articles. We contacted manufacturers and researchers in the field. We included all randomized controlled trials in adults, comparing management with and without a pulmonary artery catheter (PAC). We screened the titles and abstracts of the electronic search results and obtained the full text of studies of possible relevance for independent review. We determined the final results of the literature search by consensus between the authors. We did not contact study authors for additional information. We identified 12 studies. Mortality was reported as hospital, 28-day, 30-day, or intensive care unit. We considered studies of high-risk surgery patients (eight studies) and general intensive care patients (four studies) separately for the meta-analysis. The pooled odds ratio for the studies of general intensive care patients was 1.05 (95% confidence interval (CI) 0.87 to 1.26) and for the studies of high-risk surgery patients 0.99 (95% CI 0.73 to 1.24). Of the eight studies of high-risk surgery patients, five evaluated the effectiveness of pre-operative optimization but there was no difference in mortality when these studies were examined separately. Pulmonary artery catheterization did not affect intensive care unit (reported by 10 studies) or hospital (reported by nine studies) length of stay. Four studies, conducted in the United States, measured costs based on hospital charges billed to patients, which on average were higher in the PAC groups. To date, there have been two multi-centre trials of the effectiveness of PACs for managing critically ill patients admitted to intensive care, although only one was adequately powered. Efficacy studies are needed to determine optimal management protocols and patient groups who could benefit from management with a PAC.