The general prevalence of cardiac arrhythmias in 2,820 consecutive patients was 78%, ranging from 44% in multiple trauma patients to 90% in primary cardiovascular patients. Patients without recorded arrhythmias (22%, n = 621) were used as control subjects. No clinical group was free from cardiac arrhythmias. Atrial tachyarrhythmias had the highest prevalence in the population as a whole (28%) and in all clinical groups except multiple trauma. Atrial fibrillation was the most common atrial arrhythmia (52%); ventricular arrhythmias followed. Patients with atrial tachyarrhythmias, nodal rhythm ventricular bradyarrhythmias, and ventricular rapid rhythms had significantly (p less than .01) increased mortality rates (40%, 44%, 77%, and 51%, respectively) when compared with patients without arrhythmias (35%). The relative risk of dying (RRD) of these clinical groups was increased by 1.16, 1.27, 2.20, and 1.47, respectively. Patients with cardiorespiratory precipitating disease and any arrhythmia except atrial bradyarrhythmia had a mortality rate between 32% and 74%, significantly (p less than .05) different from that of patients within the same clinical groups without arrhythmias. The RRD was increased by 1.67 to 3.40. Septic patients with atrial tachyarrhythmia or nodal rhythm and neurologic patients with nodal or ventricular arrhythmias also had significantly (p less than .01 and .05, respectively) increased mortality and were at higher RRD (1.53 to 2.81). Our data suggest that severe illness may be present in some clinical groups of critically ill patients with cardiac arrhythmias.