This multicenter study reports the outcome of ventricular tachycardia (VT) therapy (conversion or acceleration) and the relationship to initial tachycardia cycle length and other clinical variables using an implantable device with the capability of autodecremental or burst pacing, cardioversion, and defibrillation. The device was implanted in 444 patients (mean age 58 +/- 15 years) with 1,240 episodes of VT induced with noninvasive programming and reported in a multicenter database. Only the first sequence attempted for conversion by pacing or cardioversion was assessed, and cardioversion energies were 0.2-5 J. Autodecremental pacing was used to treat 700 induced episodes of VT during titration of pacing therapies (57% converted and 12% accelerated), burst pacing to treat 357 episodes (49% converted under 11% accelerated), and cardioversion to treat 183 episodes (82% converted and 4% accelerated). Cardioversion was the most effective treatment and had the lowest acceleration rate. Shorter VT cycle lengths were more likely to accelerate with burst pacing and longer VT cycle lengths to convert with both burst and autodecremental pacing. Patients with higher ejection fractions were more likely to convert with autodecremental and burst pacing. Use of cardioversion, higher ejection fraction, absence of unrepaired aneurysm, longer VT cycle lengths, coronary artery disease, and use of autodecremental pacing predicted conversion. Lower ejection fraction and VT cycle lengths < or = 300 msec predicted tachycardia acceleration.