Post hoc insights from PAC-Man--the U.K. pulmonary artery catheter trial.

OBJECTIVES: To provide descriptive information on patients considered for management with a pulmonary artery catheter (PAC) in U.K. intensive care units and to generate hypotheses to guide future research by examining subsets of patients included in the PAC-Man Study. DESIGN: Randomized controlled trial. SETTING: U.K. general intensive care units. PATIENTS: Adult critically ill patients deemed to require management with a PAC by the treating clinician. INTERVENTIONS: Management with a PAC. MEASUREMENTS AND MAIN RESULTS: A Cox proportional hazards model was used to estimate interactions between treatment effect and time to randomization, age, surgical status, Sequential Organ Failure Score (SOFA) at randomization, organs supported at randomization, and use of flow measurement devices. Type of hospital and size of unit were tested for an interaction with the treatment effect using multilevel logistic regression modeling. There was no effect (or trend) on hospital survival related to the timing of randomization in relation to intensive care unit admission, type of organ support or SOFA score at randomization, age, type of hospital, or size of intensive care unit. No overall difference in acute hospital outcome was seen between use of a PAC and no flow measurement (p = .748) or between use of an alternative flow measurement device and no flow measurement (p = .395). CONCLUSIONS: Post hoc analyses of the PAC-Man Study data set revealed no benefit associated with being managed with a PAC in critically ill patients. However, such analyses are limited, and adequately powered clinical trials are needed of specific population subsets receiving targeted therapies delivered early in the patient's critical illness to optimize the likelihood of reversing or preventing further organ dysfunction. Furthermore, the utility of other flow measurement devices must be investigated as these have already become integrated into critical care management without adequate evaluation.