BACKGROUND: Cardiac baroreceptor sensitivity (BRS) can now be calculated non-invasively with the advent of new beat-to-beat blood pressure monitors and simulatneous surface electrocardiographic recording. However, it is not known whether BRS values calculated using a pulse interval obtained from the pulse wave are different from those using the electrocardiographic trace. This is important with the advent of new portable monitors (e.g. Portapres), which do not record a simultaneous electrocardiographic trace. OBJECTIVE: To determine whether pulse interval derivations from pulse wave data and electrocardiographic data are comparable.METHODS: Twenty healthy volunteers (10 men) aged 48 +/- 17 years with mean blood pressure 134 +/- 13/77 +/- 7 mmHg were examined. Supine beat-to-beat arterial blood pressure (measured using a Finapres 2300 device from Ohmeda) and surface electrocardiographic readings were recorded and spectral analysis, using fast Fourier transformation (FFT), and sequence analysis used to calculate BRS. Pulse interval was marked using both electrocardiographic and blood pressure traces and we compared the resulting variability of pulse interval and BRS values obtained using the two methodologies. RESULTS: There was good agreement between the two methods for pulse interval power in the low-frequency bandwidth and for BRS sequence analysis, especially for 'down' sequences, (limits of agreement -26.3 to 77.1 ms2 and -1.6 to 2.7 ms/mmHg, respectively). However, in the high frequency bandwidth (0.15-0.35 Hz) and for the combined BRS, the agreement was less good (-69 to 153.8 ms2 and -0.3 to 1.5 ms/mmHg). Overall differences between methods were smaller for BRS calculated in the low frequency rather than high-frequency band, but only in the low-frequency band was the difference positively correlated to mean BRS values. CONCLUSION: Cardiac BRS values calculated using pulse intervals derived from the pulse wave are different from those on the surface electrocardiographic trace, the greatest differences occurring in BRS derived from the high-frequency bandwidth. However, these differences are small, especially in the low-frequency band at lower mean BRS values, and should not discourage use of these newer methods of BRS measurement that do not involve an electrocardiographic recording.