OBJECTIVE Lithium has occasionally been reported to cause symptomatic sinus node bradyarrhythmias. The prevalence and mechanism of these arrhythmias during long-term treatment are unknown. The aims of this study were (a) to evaluate the systemic effects of lithium treatment on cardiac conduction in individuals who were free from cardiovascular disorders; (b) to assess the prevalence of lithium treatment in a group of patients with pacemakers; and (c) to evaluate the interaction between the parasympathetic limb of the autonomous nervous system and the sinus node cells during long-term lithium treatment. METHODS 45 patients who had been treated with lithium for > 12 months were investigated in a long-term electrocardiography study. Only patients without cardiovascular disease, or concomitant chronotropic medication, or metabolic disorders known to cause rhythm disturbances were included. An age-stratified population was used as a reference group. 21 patients also underwent analysis of carotid sinus pressure and sinus cycle length before and after atropine to clarify whether neural mechanisms were involved. The prevalence of lithium treated patients was determined in 650 patients with pacemakers. RESULTS (a) Signs of moderate sinus node dysfunction (sinus arrest > 1.5 s, minimum heart rate < 50 beats/min) were found in 56% and 78% respectively in the lithium-treated group compared with 30% and 30% respectively in the reference group (p < 0.01). Severe sinus node dysfunction was equally common in both groups. (b) The prevalence of chronic lithium treatment in the pacemaker population was 0.46%. (c) Sinus cycle variations were abnormal in the basal state in three (14%) patients and in 11 (52%) patients after atropine despite signs of intact and normal parasympathetic innervation. CONCLUSIONS Depressed sinus node function was significantly more common in a lithium-treated population than in an age-stratified reference group. Clinically significant dysfunction, however, was uncommon. The effect of lithium on the sinus node seemed to be intrinsic and was not caused by increased parasympathetic tone.