We examined prospectively for one year the hearts from 141 consecutive autopsy cases in which a central catheter was present at the time of death. Three deaths were attributable to catheter use, two to perforation. Furthermore, mural thrombi were present in 33 (33%) of 99 patients with pulmonary arterial catheters and in 12 (29%) of 42 patients with central venous catheters. The incidence of pulmonary emboli or bacteremia was no greater in patients with thrombi than in those without. The use of central catheters may thus be complicated by perforation or the development of mural thrombi. Although the thrombi may embolize or may become infected, the incidence and clinical significance appear to be low. The incidence of catheter-related deaths in our autopsy population does not necessarily reflect the incidence in a population of living patients.