Cross spectral analysis in assessment of baroreflex gain in patients with coronary artery disease.

Abstract

Listen

Translate article

Interest in determination of baroreflex sensitivity in clinical practice is growing because of its prognostic information in patients with heart disease. The purpose of the present study was to assess the feasibility of cross spectral analysis in the determination of baroreflex gain from spontaneous RR interval and systolic pressure fluctuations, and to compare the results to the traditional pharmacological method in patients with coronary artery disease. Methods. We measured the gain and time lag between RR interval and systolic pressure variabilities in the frequency domain, and compared baroreflex indexes obtained by this technique with standard phenylephrine tests in 32 patients with coronary artery disease. Results. Cross spectral analysis by fast Fourier transform techniques yielded acceptable (> 0.5) coherence between systolic pressure and RR interval in the mid- (0.07-0.15 Hz) and in the respiratory-frequency (0.15-0.40 Hz) band fluctuations in 30 patients (94%), with mean coherences of 0.69 and 0.74. The mean phase difference in the mid-frequency hand was greater than in the respiratory-frequency band (-83 vs -23 degrees, P < 0.001), suggesting that the mid-frequency fluctuations of RR intervals followed nearly 2 seconds after pressure changes, while respiratory-frequency fluctuations of RR intervals occurred nearly concomitantly with systolic pressure. The mean baroreflex slope derived from the bolus phenylephrine technique was 6.2 ms/mmHg (range 1.6-16.0), 5 patients had an abnormally low (<3 ms/mmHg) baroreflex sensitivity. Baroreflex gain determined by cross spectral analysis from the mid-frequency band correlated significantly (r = 0.60, P < 0.001, n = 27) with the baroreflex gain determined by the phenylephrine test, while the correlation in the respiratory-frequency band was not significant (r = 0.35, P = 0.09, n = 26). Conclusions. Baroreflex slopes derived from cross spectral techniques provide reliable (but not perfect) information regarding baroreflex gain derived from the clasic phenylephrine technique, even in patients with depressed baroreflex responses. Cross correlation calculation of spontaneous baroreflex slopes should be limited to data in the mid-frequency range, where the slopes are likely to reflect simple baroreflex physiology.