OBJECTIVE: To assess the role of caffeine restriction in the management of patients with symptomatic idiopathic ventricular premature beats. DESIGN: A randomised, double blind, 6 week intervention trial incorporating dietary caffeine restriction, caffeinated coffee, and decaffeinated coffee. SETTING: Cardiac outpatient clinic. PATIENTS: 13 patients with symptomatic frequent idiopathic ventricular premature beats. MAIN OUTCOME MEASURES: Weekly measures of serum caffeine concentration, coffee consumption, visual analogue score of palpitations, and 24 hour ventricular premature beat frequency. RESULTS: The interventions achieved significant alterations in serum caffeine concentrations (P < 0.001) which correlated with coffee consumption (r = 0.70; P < 0.001). Visual analogue palpitation scores showed a small, but significant correlation with ventricular premature beat frequencies (r = 0.34; P = 0.003). However, there were no significant changes in palpitation scores or ventricular premature beat frequencies during the intervention weeks and no significant correlations were found between these variables and serum caffeine concentrations. CONCLUSIONS: Caffeine restriction has no role in the management of patients referred with symptomatic idiopathic ventricular premature beats.