Assessment of baroreflex sensitivity in patients with preserved and impaired left ventricular function by means of the Valsalva manoeuvre and the phenylephrine test

The purpose of the present study was to assess the agreement between measurements of baroreflex sensitivity (BRS) obtained by the Valsalva manoeuvre and by the phenylephrine test in patients with previous myocardial infarction and different degrees of left ventricular dysfunction. Patients with a previous myocardial infarction were enrolled consecutively into two groups according to their left ventricular ejection fraction (LVEF): less than or equal to 40% (n = 52) or,40% (n = 52). All patients underwent BRS assessment by the phenylephrine technique (Phe-BRS) and by the Valsalva manoeuvre, with the latter using both the overshoot part of phase IV (OV-VM-BRS) and the whole of phase IV (IV-VM-BRS). The linear association between methods was assessed by correlation analysis and the agreement was evaluated by computing the bias and the limits of agreement. IV-VM-BRS and Ov-VM-BRS could not be computed in 26% and 39% of patients respectively. For both indices a much higher percentage of non-computable Valsalva manoeuvre slopes was found in the group of patients with LVEF less than or equal to 40%. Phe-BRS was significantly lower in patients with non-computable Valsalva manoeuvre slopes; 55% and 59% of patients with non-computable OV-VM-BRS and IV-VM-BRS respectively had a markedly depressed Phe-BRS (i.e. < 3 ms/mmHg, as defined by the ATRAMI study). On comparing Phe-BRS and Ov-VM-BRS in patients with LVEF 40%, we obtained the following results: r = 0.75 (P < 0.001), bias = 0.5 ms/mmHg (P = 0.5) and limits of agreement from -6.7 to 7.8 ms/mmHg. In patients with LVEF > 40% the results were: r = 0.91 (P < 0.001), bias = 0.1 ms/mmHg (P = 0.84) and limits of agreement from -4.8 to 5 ms/mmHg. When comparing Phe-BRS and IV-VM-BRS, we found r = 0.67 (P = 0.001), bias = -1.5 ms/mmHg (P = 0.06) and limits of agreement from -8.8 to 5.7 ms/mmHg in the group of patients with LVEF 40%; and r = 0.66 (P < 0.001), bias = -3.1 ms/mmHg(P < 0.001) and limits of agreement from -12.1 to 5.9 ms/mmHg in the group of patients with LVEF > 40%. Dichotomizing Ov-VM-BRS, the best cut-off value to identify patients with a Phe-BRS of < 3 ms/mmHg was found to be 7 ms/mmHg, giving 100% sensitivity and 69% specificity. In conclusion, estimation of BRS by the Valsalva manoeuvre in post-myocardial infarction patients is limited by a large number of non-measurable results. When computable, measurements are well correlated with those obtained by Phe-BRS, but, because of large limits of agreement, the two methods cannot be used interchangeably. If used as a screening test for risk stratification, the Valsalva manoeuvre could reduce by about one-third the need for phenylephrine injection.