BACKGROUND In patients with sick sinus syndrome, choice of pacing mode has been implicated in the development of congestive heart failure. RESULTS A total of 225 consecutive patients with sick sinus syndrome and intact atrioventricular conduction were randomized to either single-chamber atrial pacing (n = 110) or single-chamber ventricular pacing (n = 115). Clinical assessment included New York Heart Association classification, medication, and M-mode echocardiography before pacemaker implantation, after 3 months, and subsequently once every year. At long-term follow-up (mean, 5.5+/-2.4 years), NYHA class was higher in the ventricular group than in the atrial group (NYHA class I/II/III/IV: 65/44/4/0 versus 84/22/2/1 patients, P=.010). Increase in NYHA class during follow-up was observed in 35 of 113 patients in the ventricular group versus 10 of 109 in the atrial group (P<.0005). Increase in dose of diuretics from randomization to last follow-up was significantly higher in the ventricular group than in the atrial group (21+/-49 versus 8+/-42 mg furosemide/d, P=.033). The left ventricular fractional shortening decreased significantly in the ventricular group (from 0.36+/-0.12 to 0.31+/-0.08, P<.0005) but not in the atrial group (from 0.35+/-0.13 to 0.33+/-0.09, P=.087). The left atrial diameter increased significantly in both treatment groups (ventricular group: from 34+/-7 to 41+/-7 mm, P<.0005; atrial group: from 34+/-6 to 37+/-7 mm, P=.002), but the increase was significantly higher in the ventricular group than in the atrial group (P<.0005). CONCLUSIONS During long-term follow-up, ventricular pacing is associated with a higher incidence of congestive heart failure and consumption of diuretics than atrial pacing. This is accompanied by a decrease in left ventricular fractional shortening and an increased dilatation of the left atrium in the ventricular paced patients.