We studied the contribution of iatrogenic illness to cardiac arrest among patients hospitalized in 1981 in a university teaching hospital. During this 1-year period, 28 (14%) of 203 arrests in which resuscitation was attempted followed an iatrogenic complication. Seventeen (61%) of the 28 patients died. The demographic characteristics of patients with iatrogenic arrest did not differ strikingly from those of other patients who arrested. However, patients with iatrogenic arrest were less likely to be in cardiogenic shock or to have suffered an acute myocardial infarction prior to arrest. They were more likely to survive to discharge from the hospital and to be taking digoxin or antiarrhythmic medication prior to arrest. Among the 28 cases of iatrogenic cardiac arrest, 18 (9% of all arrests) might have been prevented by stricter attention to the patient's history, findings on physical examination, and laboratory data. The most common causes of potentially preventable arrest were medication errors and toxic effects (44%) as well as suboptimal response by physicians to clinical signs and symptoms (28%), most frequently dyspnea and tachypnea. Rapid, appropriate response to abnormal drug levels, to electrocardiographic signs of adverse drug effects, and to signs and symptoms of congestive heart failure or toxic effects from digoxin might decrease the incidence of cardiac arrest among hospitalized patients.