Reply from K. L. Ryan, C. A. Rickards, C. Hinojosa‐Laborde, W. H. Cooke and V. A. Convertino

Abstract

We thank Pagani et al. for their interest in our recent paper (Ryan et al. 2011)) and for their kind comments concerning our work. We appreciate the opportunity for further dialogue about the issue of whether low-frequency (LF) blood pressure oscillations can be used as an accurate non-invasive surrogate of direct measurements of sympathetic nerve activity. In the Letter to the Editor, Pagani et al. rightfully point out that there were minor differences in techniques used in our study versus their previous work. While we used fast Fourier transfer (FFT) to perform our spectral analysis, autoregression (AR) was used for spectral analysis in the studies that Pagani et al. cite and which form the primary literature on which our work was based. Both techniques, however, have been used in studies investigating autonomic function, with comparable results (Persson 1997; Badilini et al. 1998)). While we cannot discount the possibility that utilization of a different technique might yield slightly different quantitative values, the results derived using these techniques, if truly reflective of autonomic control, should reliably reflect the underlying physiology and should therefore be qualitatively similar. Concerning normalization of spectral powers, what does this mean physiologically? As cautioned previously, ‘normalizing the oscillations to each other can uncouple their amplitudes from the physiology’ (Taylor & Studinger, 2006)). Clearly, this is dangerous territory for a measure purporting to track autonomic nerve activity. Importantly, it should also be noted that Pagani et al. (1997) reported significant correlations between both absolute low-frequency oscillations of systolic arterial pressure (SAPLF) and normalized SAPLF with muscle sympathetic nerve activity (MSNA). Based on the evidence, it is unlikely that differences in the techniques used to quantify the amplitude of blood pressure oscillations could be res- ponsible for our findings that challenge the use of SAPLF as a non-invasive surrogate of MSNA. We also agree that, like any analytical technique, interpretations of MSNA measurements must be informed by an appreciation of the technical limitations of microneurography. Over the past 45 years, however, microneurography has become a well-accepted technique that has driven entirely new understandings of the control of autonomic function in humans in both health and disease (Vallbo et al. 2004; Fatouleh & Macefield, 2011)). When continuous MSNA measurements are collected in the same subject in the same experimental session without a shift in baseline (conditions met in our work), issues such as electrode tip placement, damage and electronic settings are minimized if not abrogated. Indeed, the concept that SAPLF might serve as a non-invasive means of assessing sympathetic nerve activity was put forward based on results of studies in which MSNA was measured (Pagani et al. 1997; Furlan et al. 2000)), suggesting that we all agree that direct measurement of MSNA has value in answering this research question. Furthermore, as discussed in our paper (Ryan et al. 2011)), results from animal studies measuring renal (Tsai et al. 2009)) or splanchnic (Persson et al. 1992; Stauss et al. 1995)) sympathetic nerve activity have shown that LF oscillations of blood pressure are not a marker of levels of sympathetic nerve activity, indicating that our conclusion is not a function of the measurement of MSNA per se but represents a more global relationship. We also agree with Pagani et al. that, generally speaking, SAPLF increases during sympathetic activation and that it may reflect oscillatory patterns of sympathetic nerve firing; the results of all of the studies to date are in concordance with this concept. While we appreciate the desire for a convenient and non-invasive biomarker of sympathetic nerve activity, our concern is that SAPLF might be used inappropriately to assess autonomic function in individual patients, as our data do not show a direct relationship between SAPLF and MSNA in all individuals. These healthy subjects were drawn from a fairly homogeneous population and studied in a controlled laboratory setting with a well-accepted experimental model that produces progressively increasing sympathetic activation. If such large interindividual variability in the SAPLF–MSNA relationship exists in these healthy subjects under these conditions, how can one define what a ‘normal’ relationship is for the population as a whole? If we cannot reliably define ‘normal’, how then can we define what is abnormal or indicative of dysautonomia in patient populations? While convenience is an attractive feature, it cannot replace efficacy. We would never advocate throwing the baby out with the bath water. But should we not objectively assess whether the baby is able to perform effectively in the situation in which he or she has been placed?