Baroreflex sensitivity (BRS) conveys useful prognostic information in patients with heart disease, yet methods for its quantification suffer from poor reproducibility and test failure in some patients with heart failure. We set out to compare the short-term reproducibility and success rate of four different methods of assessing BRS in normal subjects and patients with chronic heart failure (CHF). A total of 31 patients with CHF and 18 normal controls underwent BRS testing using four techniques: (1) bolus phenylephrine (BRS(Phe)), (2) alpha-index in both low- and high-frequency bands (BRS(alphaLF) and BRS(alphaHF) respectively), (3) the sequence method (BRS(Seq)), and (4) a new 0.1 Hz controlled-breathing, time-domain analysis method (BRS(Cbr)). Each subject underwent two test episodes with each method on the same day. The average values for BRS in patients and controls respectively were: BRS(Phe), 4.4 (+/-4.4) ms/mmHg and 19.8 (+/-11.5) ms/mmHg; BRS(alphaLF), 5.6 (+/-4.1) ms/mmHg and 15.4 (+/-5.0) ms/mmHg; BRS(alphaHF), 7.1 (+/-7.0) ms/mmHg and 25.1 (+/-8.3) ms/mmHg; BRS(Seq), 7.7 (+/-6.3) ms/mmHg and 22.5 (+/-8.4) ms/mmHg; BRS(Cbr), 6.6 (+/-5.9) ms/mmHg and 22.8 (+/-10.8) ms/mmHg. The coefficients of variation (S.D. of the difference in repeated values divided by mean) in patients and controls respectively were: BRS(Phe), 85.6% and 52.2%; BRS(alphaLF), 65.9% and 33.7%; BRS(alphaHF), 99.7% and 52. 1%; BRS(Seq), 30.7% and 40.4%; BRS(Cbr), 30.7% and 19.6%. The numbers of test failures in patients were: BRS(Phen), 15; BRS(alphaLF), 7; BRS(alphaHF), 5; BRS(Seq), 14; BRS(Cbr), 1. Of the four techniques assessed for measuring BRS, the controlled breathing time-domain method yielded the best reproducibility and lowest failure rate in controls and in patients with CHF.