PURPOSE: The purpose of this study is to determine whether short-term heart rate variability (HRV) can be used successfully to predict inducible ventricular tachycardia (VT). METHODS: A high-speed (300 mm/s) electrocardiographic recording was obtained in 32 patients in the supine position prior to programmed ventricular stimulation. Beat-to-beat RR intervals (in milliseconds) were derived from an 11-beat strip (10 RR intervals). Logistic regression was used to study the relationship between several variables and a dichotomous dependent variable (inducible, clinical, or electrocardiographic evidence of VT). RESULTS: Of 32 patients, 12 had inducible VT (inducible VT group) and 20 had no clinical or electrocardiographic evidence of VT (control group). Mean short-term HRV values were significantly lower in those with inducible VT than in the control group in all patients (25+/-15 ms, n=12 vs 67+/-22 ms, n=20; p<0.0001) and in patients with coronary artery disease or congestive heart failure or both (22+/-13 ms, n=11 vs 63+/-23 ms, n=11; p<0.0001). For the group as a whole, short-term HRV was < or =50 ms in 11 of 12 patients (92%) with inducible VT, but was < or =50 ms in only 3 of 20 control subjects (15%; p<0.001). As a result of a stepwise selection procedure conducted within the logistic regression, only the short-term HRV was found to be predictive of inducible VT (p<0.0001). CONCLUSION: Short-term HRV is significantly lower in subjects with inducible VT than in those without clinical or electrocardiographic evidence of VT. The probability of developing sudden death increases substantially when short-term HRV decreases below 50 ms.