To investigate a cluster of postoperative bleeding following open heart surgery.A cohort and case/control study.Palo Alto Veterans Administration Medical Center, Palo Alto, California.Six (21.4%) of 28 patients undergoing open heart surgery who developed severe, nonsurgical, postoperative bleeding from July 1 through August 30, 1988 (outbreak period). All case-patients had chest tube drainage of greater than or equal to 1000 ml within 4 hours of surgery but did not have identifiable bleeding vessel(s) on exploration.Upon comparison of the pre-outbreak (January 1986 through June 1988) and the outbreak period, a significant increase was found in the incidence of postoperative nonsurgical bleeding (5/440 versus 6/28, p = .0006), but not of postoperative surgical bleeding (8/440 versus 0/28, p = 1.0). Of all patients undergoing open heart surgery during the outbreak period, case patients were found to be older (67.8 versus 60.6, p = .02) and to have received a larger volume of hetastarch (HES), a synthetic colloidal plasma-volume expander (mean = 19.4 ml/kg versus 14.1 ml/kg, p = .02).We conclude that the use of large volumes of HES during surgery in the elderly open heart surgery patient may increase the risk for severe, nonsurgical postoperative bleeding, probably caused by alterations of the coagulation system. As the incidence of open heart surgery increases among the elderly, surgeons and anesthesiologists should be alert to possible adverse reactions from exposures not associated with adverse reactions in younger patients.