To the Editor:--The data provided by Hopf et al. [1]are interesting and provocative. Their conclusions are in stark contrast to the large number of reports published during the last 10 yr and especially the last 2 yr [2-5]that further specified the methodology needed to use the low-frequency spectral component as a marker of sympathetic modulation. These authors state that, after segmental epidural anesthesia (C6-T6) in 10 patients, low-frequency spectral power did not change significantly, therefore low-frequency spectral power of heart rate variability is unlikely to reflect cardiac sympathetic modulation.
Beginning with the report by Akselrod et al. [6]and continuing to the present, investigators have debated the degree of sympathetic-parasympathetic contribution to the low-frequency component of heart rate variability. The early interpretation that sympathetic and parasympathetic modulations contributed to the origin of the low-frequency component was related mainly to the experimental approach followed by the investigators. [6,7]For example, with parasympathetic blockade by atropine, the low-frequency component in absolute units is reduced, in accordance with the dramatic decrease in total variance, thus leading to the hypothesis of a parasympathetic contribution to this low-frequency oscillation. However, if one considers the low-/high-frequency ratio or the low frequency in normalized units, one would observe the prevalence in these conditions of the low-frequency component accordingly with the sympathetic predominance induced by atropine. [8]Montano et al. [4]reported that the low- and high-frequency components expressed in normalized units (nu) and low-/high-frequency ratio provided the best estimates of graded changes in sympathovagal balance. Similarly, situations characterized by a sympathetic activation, such as tilt, mental stress, and coronary occlusion, are associated with an increase of low-frequency component (nu) or low-/high-frequency ratio. [8,9]
In previous experiments, using a Fast Fourier transform (FFT) algorithm, we observed that the rostral spread of spinal anesthesia was associated with a progressive decrease of total variance in both low-and high-frequency components of the HRV spectra. [10]We recently confirmed those data analyzing, by means of autoregressive spectral techniques, the HRV of 24 women who received spinal anesthesia for postpartum tubal ligation surgery. [11]The progressive cephalad spread of spinal block (from T4 to C7) was associated with a significant reduction in total variance, along with decreases in both low- and high-frequency components in absolute units, whereas low- and high-frequency (nu) as well as low-/high-frequency ratio were not significantly changed. A possible conclusion was that low and high frequency, and obviously total power, originate from an interaction of sympathetic and parasympathetic activities and that an acute subtraction of one component does not necessarily represent the best approach to understand the physiologic significance of cardiovascular rhythms. Similar results on the effects of spinal anesthesia on heart rate variability were obtained by Oberlander et al. [12]in premature infants and by Landry et al. [13]in adults. The latter study included patients with epidural anesthesia, and the results were not different from those of patients receiving spinal anesthesia. The patients in the study by Hopf et al. [1]were scheduled for thoracic and upper abdominal surgery, and although the specific surgical procedures and underlying conditions were not discussed, the state of the patients' health could have contributed to the results. [3,5]Moreover, the norepinephrine levels in their patients did not increase during 40-degree tilt, which is known to activate the sympathetic nervous system, before epidural block. The authors did not provide a reason why they chose the low-frequency band of 0.06-0.15 Hz. Most other researchers report using a low-frequency cutoff of 0.02-0.04 Hz. This frequency range could significantly affect interpretation of the data.
A possible explanation of the differences between the results presented by Hopf et al. [1]and other studies [10-13]can be found in the authors' own introductory sentence, "Spinal sympathetic outflow during epidural anesthesia is attenuated by preganglionic blockade." [1]Therefore, if an attenuation, incomplete blockade, of the cardiac sympathetic outflow results from epidural anesthesia, the phenomenon of central recruitment [14]might occur. In this phenomenon, available autonomic pathways that have survived epidural block can compensate by being recruited to carry the bulk of the remaining neural traffic to and from the heart. This possibility appears to suggest that the reason for the controversial results is that some sympathetic innervation to the heart remained intact to varying degrees. This theory is supported by our own report [11]as well as one that appeared in this journal during the writing of this letter. [15]In our paper, 4 of the 24 patients studied under spinal anesthesia had sensory level blocks to T3-T4 determined by pinprick and yet experienced inadequate surgical analgesia. Interestingly, in these four patients, HRV spectra were little changed from the preanesthetic spectra. In the remaining patients, spinal anesthesia had significantly depressed low-and high-frequency components as noted above.
Robert P. S. Introna, M.D., Department of Anesthesiology, Medical College of Georgia, Augusta, Georgia 30912-2700.
Nicola Montano, M.D., Ph.D., Centro Ricerche Cardiovasculari CNR, Medicina Interna II, Ospedale L. Sacco, Universita di Milano, 20157 Milano, Italy.
Edmund H. Yodlowski, Ph.D., Dan C. Martin, M.D., Jack K. Pruett, Ph.D., Robert S. Crumrine, M.D., Department of Anesthesiology, Medical College of Georgia, Augusta, Georgia 30912-2700.
(Accepted for publication July 18, 1995.)
