It should be noted that these variables can have positive and negative values given the relative presence or absence of low transfer magnitude.
The mean and SD of the HR and power spectra for each study epoch were calculated. Coherence weighted transfer function averages were computed as previously described . To compare the course of individual subjects before and during the spinal anesthesia, nonparametric (Wilcoxon signed rank test) statistics were used. A difference was considered significant when P was less than 0.05. For the purposes of display, individual and coherence weighted group average transfer function estimates of both magnitude and phase for preoperative, operative, and postoperative segments were also computed.
Spinal anesthesia was administered to the eight study infants without complications. Sensory block was rapidly achieved to a dermatomal level of C7-T4. Oxygen saturation remained 97%-100% for all subjects. Mean HR, systolic pressure, and diastolic pressure remained stable across all study periods Table 2. Respiratory power (L/m2)2 did not change significantly over the course of the three study periods Table 2; however, a large, nonsignificant reduction in respiratory power was observed in the postoperative period.
Overall, LFP decreased significantly from baseline with spinal anesthesia among all infants (Wilcoxon signed rank test, Z = 2.52; P = 0.01) Table 3. Similarly, HFP decreased in seven of the eight subjects (Z = -2.38; P = 0.02). Importantly, the LFP/HFP ratio did not change significantly with the anesthetic (Z = -0.70; P = 0.48). The sympathetic and parasympathetic gain factors, As and Ap, derived from the transfer function magnitude data, decreased 43.8% and 30%, respectively, from their preoperative levels; however, the differences were not statistically significant [(Z = -0.98; P = 0.33) and (Z = -0.28; P = 0.78), respectively].
As demonstrated in both the HR time series and autospectra from an individual subject Figure 1, B and C, HRV decreased during the operative period, with a small less obvious increase seen in the postoperative period. Although the respiratory rate slowed and became more regular from the preoperative to the operative period, resulting in a clustering of respiratory power at frequencies between 0.4 and 0.5 Hz, total respiratory power was similar. The reduction in HR power combined with stable respiratory activity resulted in decreased transfer magnitude in the operative period, which in this subject is most obvious at low frequencies. In addition, the phase relation at low frequencies shifted from a negative slope, consistent with a sympathetically mediated delay, to a zero slope, consistent with more vagal HR control. As in this subject decreased from 84.3 to 43.9 [(bpm)/(L/m2)], while Ap remained unchanged (3.9 vs 4.0 [(bpm)/(L/m2)], again consistent with diminished sympathetic modulation of HR, but no significant change in vagal modulation.
Finally, the postoperative period demonstrates the importance of the transfer function analysis which accounts for respiratory activity in evaluating the HRV. The transfer magnitudes and Ap actually increased at most frequencies despite the low HRV, because of the marked reduction in respiratory power. The lack of complete consistency of the data from this subject with the group helps demonstrate why the changes in As and Ap between periods were not statistically significant.
The group transfer functions shown in Figure 2 qualitatively demonstrated the changes in As and Ap from Table 3. Low- and high-frequency magnitude both decreased during the operative period with no obvious change in phase, consistent with the small decrease observed in group mean Ap. During the post-operative period, the magnitude increased at all frequencies, while the phase clustered around 0 degrees, consistent with predominant vagal modulation of HR.
In contrast to adults and older children, the former premature infants in this study undergoing spinal anesthesia to high thoracic levels maintained stable levels of mean HR, mean blood pressure, and respiratory activity, even in the absence of prior fluid loading. Overall HRV (both LFP and HFP) decreased with spinal anesthesia, but the balance between them (LFP: HFP ratio) remained stable, suggesting that in response to the sympathetic block due to spinal anesthesia, there was a reflex reduction in parasympathetic modulation of cardiac function. However, when the effect of respiratory activity on HRV was accounted for, the net autonomic effect remained relatively stable.
Using frequency domain analyses to quantify autonomic function, animal and human studies suggest that low-frequency fluctuations of HR at frequencies between 0.02 and 0.15 Hz result primarily from modulation of both cardiac parasympathetic and sympathetic efferent activity by multiple influences, such as low frequency respiratory activity, and arterial and cardiopulmonary baroreceptor feedback . In contrast, higher frequency fluctuations of HR are almost exclusively mediated by modulation of vagal efferent activity, primarily in response to respiratory activity. Based on these observations and others, HFP can be used to provide a qualitative index of cardiac vagal HR control [11-13], and the ratio of the LFP to the HFP, can provide a measure of cardiac sympathetic HR control .
Although the quantitative changes of sympathetic (As) and parasympathetic (Ap) modulation of HR did not reach significance [14-18], the group average transfer functions yielded some qualitative insight into the autonomic changes which accompanied the anesthesia. The group magnitude plots both prior to and during anesthesia demonstrate a low-fequency predominance, while the phase plots at low frequencies show decreasing phase with frequency Figure 2, both findings which suggest a sympathetic predominance at rest and during anesthesia. In addition, both the low- and high-frequency transfer magnitude decreased with anesthesia, suggesting vagal withdrawal consistent with the spectral data. Finally, in the postoperative period, both low- and high-frequency magnitude increased and the low-frequency phase became flat, suggesting an increase in vagal HR modulation postoperatively Figure 2 to levels higher than those observed either pre- or intra-operatively.
Previous reports have indicated that spinal anesthesia which induces sensory block at the C7-T4 level produces effective peripheral vascular sympathetic denervation Figure 3a  and substantial, if not total, cardiac sympathetic denervation  Figure 3b. These effects on vascular sympathetic efferent activity should decrease peripheral arterial resistance and venous tone, and if the patient has not received prior hydration, decrease central venous volume, right atrial pressure Figure 3c and size Figure 3d, and sinoatrial node stretch Figure 3e. Decreased atrial size should presumably also result in a cardiopulmonary-mediated reflex increase in cardiac vagal activity Figure 3f. In the absence of arterial baroreflex-mediated reflex, withdrawal of cardiac vagal activity, HR should drop, decreasing cardiac output, and further decreasing arterial pressure Figure 3g. Such changes have been noted in adults undergoing spinal anesthesia at the T2 level, with the exception that HR actually remains constant, presumably due to arterial baroreflex-mediated vagal withdrawal [3,19]. Because of these observations, one would expect that sympathetic modulation of HR would also decrease, leading to a drop in LFP, LFP/HFP, and As; however, these spectral variables have not been examined in adults who were not prehydrated. Landry et al.  recently examined HR spectral variables in a group of pregnant women undergoing peridural anesthesia (to at least a T-6 level) after prehydration. Despite the absence of a change in mean HR and a small but insignificant drop in arterial pressure, they found a decrease in both LFP and HFP without a change in the LFP/HFP ratio, suggesting isolated withdrawal of vagal modulation of HR which effects both frequency bands, and no significant change in sympathetic modulation which selectively effects low frequencies. Furthermore, if one accounts for the factors which influence HR variations such as respiration, theoretically more specific variables for vagal and sympathetic HR control can be derived, in this case Ap and As [9,10,16-18].
The data from the infants in this study differs from that in adults in two important ways. First, despite the absence of volume loading, neither HR nor arterial pressure changed in these infants. Second, measurement of respiratory activity provided the potential for a more sensitive metric of cardiac vagal and sympathetic modulation than the spectral data alone. The fact that mean arterial pressure remained constant without volume loading and without an increase in HR suggests strongly that there was not a significant decrease in arterial resistance or venous filling induced by the anesthesia Figure 4a. Such a finding could be secondary to diminished capacitance of the peripheral venous bed in the neonate compared to the adult, or less dependence of the neonate on sympathetic vasoconstriction at rest (shaded area, Figure 4c), such that sympathetic block fails to cause a significant fall in arterial pressure. However, the HRV data in the study lend support to the notion that there were autonomic changes from the anesthesia. The observed decrease of both LFP and HFP without a change in the LFP/HFP was similar to that seen in adults , and is consistent with an isolated withdrawal of vagal HR modulation. Interestingly, the measures A (p) and As derived from the transfer functions demonstrated moderate decreases in both variables which were not statistically significant, but the qualitative appearance the transfer function suggested diminished vagal and preserved sympathetic modulation. Together, these data suggest that, as with adult patients, the neonate maintains a constant HR through a small, baroreflex-mediated withdrawal of cardiac vagal activity in response to, and offsetting, a small anesthesia-induced decrease in cardiac sympathetic modulation due to incomplete or variable cardiac sympathetic block Figure 4d. The variable nature of the block may be due to inconsistencies in the level of anesthesia C7-T4) in this study.
The hemodynamic stability we observed in reponse to C7-T4 spinal anesthesia is consistent with revious studies in infants and children less than 6 yr of age [4,2,20-22]. Many of these authors have attributed their findings to age-related differences in neural development, in particular, incomplete development of the sympathetic nervous system and a relative maturity of the parasympathetic system at birth . Recent work, however, has demonstrated a functionally competent autonomic system at birth in humans and animals [5,23]. Finley et al.  demonstrated a small but significant increase in HR among term human infants in response to a tilt-Table maneuver. This shows that lower limb venous capacitance in neonates is at least large enough to sequester a hemodynamically significant volume of blood during tilt, and that neonates have the ability to mount a cardiac autonomic, probable sympathetic, HR response to a change in central volume. In parallel with this cardiac response, Lagercrantz et al.  showed that preterm infants (25-37 wk gestation) were able to significantly increase peripheral resistance in the lower limbs during a tilt procedure, indicating that neonates have at least partially developed sympathetic vascular control. Thus, the possibility that the infants in this study maintain constant HR and arterial pressure through offsetting changes in cardiac sympathetic and vagal control is well supported in the literature.
Several limitations of this study need to be noted. First, the small sample size, in combination with the variability in the level of sensory block, may have reduced the power to detect a significant sympathetic withdrawal during spinal anesthesia. Second, while the impedance measure provided an index of respiratory activity, it was an indirect measure of tidal volume and, despite strict epoch selection criteria, remained sensitive to infant movement artifact, thereby increasing the variance of the signal. Given the limitations of this measure, As and Ap may not have yielded a precise quantification of the relation between respiration and HR. Further, no independent measures of sympathetic activity (e.g., blood flow, vascular resistance, muscle sympathetic nerve activity), vagal activity, central venous pressure, or plasma volume were obtained in this study to validate the As and Ap variables. It is also unclear whether the operation itself influenced the results. The short time interval between the spinal injection and the incision and the operative preparations made it impossible to acquire stable physiologic signals that would have enabled us to compare pre- and postincision autonomic measures to assess a possible confounding influence of the operation. However, given that the post-operative autonomic measures, which had been obtained when the block had regressed to at least T-12, were not significantly different from preoperative baseline values, we were confident that the spinal anesthetic could have been responsible for the autonomic changes observed. Finally, there are significant limitations in quantifying the response of HR to respiratory activity, particularly when the respiratory activity is not broad band , as was demonstrated by low coherence in this study Figure 3. However, not accounting for the effects of respiration at all clearly has even more disadvantages [9,10,17].
In conclusion, in contrast to older children and adults [3,25], former premature infants in this and other studies [2,4,20,22] maintain stable levels of mean HR, blood pressure, and respiratory activity when undergoing spinal anesthesia without prehydration. Using spectral analysis of HR and transfer function analysis of HR and respiratory activity, the spinal anesthetic predominantly had a net effect on parasympathetic modulation of HR. The decrease in parasympathetic activity we observed may reflect a reflex response to an initial bradycardia which follows a partial or complete cardiac sympatholysis from the spinal anesthetic and decreased arterial blood pressure. Specifically, a decrease in HR that may occur with a high spinal anesthetic is offset by the decreased vagal activity. Although such findings may account for the overall cardiovascular stability observed among neonates and infants during spinal anesthesia, the developmental mechanisms which enable infants to maintain such stability remain to be determined.
1. Pascucci RC, Hershenson MB, Sethna NF, et al. Chest wall motion of infants during spinal anesthesia. J Appl Physiol 1990;68:2087-91.
2. Abajian JC, Mellish RW, Browne AF, et al. Spinal anesthesia for surgery in the high-risk infant. Anesth Analg 1984;63:359-62.
3. Bridenbaugh PO, Greene NM. Spinal (subarachnoid) neural blockade. In: Cousins MJ, Bridenbaugh PO, eds. Neural blockade in clinical anesthesia and management of pain. Philadelphia: JB Lippincott, 1988:213-52.
4. Dohi S, Naito H, Takahashi T. Age-related changes in blood pressure and duration of motor block in spinal anesthesia. Anesthesiology 1979;50:319-23.
5. Hanna BD, Stark AR, Cohen RJ, Saul JP. Transfer function analysis of respiratory sinus arrhythmia: developmental effects in premature infants. Circulation 1990;82:IIT-334.
6. Finley JP, Hamilton R, MacKenzie MG. Heart rate response to tilting in newborns in quiet and active sleep. Biol Neonate 1984;45(1):1-10.
7. Prechtl HF. The behavioral states of the newborn infant (a review). Brain Res 1974;76:185-212.
8. Berger RD, Saul JP, Cohen RJ. Transfer function analysis of autonomic regulation. I. Canine atrial rate response. Am J Physiol 1989;256(1 Pt 2):H142-52.
9. Saul JP, Berger RD, Albrecht P, et al. Transfer functional analysis of the circulation: unique insights into cardiovascular regulation. Am J Physiol 1991;261(4 Pt 2):H1231-45.
10. Saul JP. Beat-to-beat variations of HR reflect modulation of cardiac autonomic outflow. News Physiol Sci 1990;5:32-7.
11. Pomeranz B, Macaulay RJ, Caudill MA, et al. Assessment of autonomic function in humans by HR spectral analysis. Am J Physiol 1985;248(1 Pt 2):H151-3.
12. Pagani M, Lombardi F, Guzzetti S, et al. Power spectral analysis of HR and arterial pressure variabilities as a marker of sympatho-vagal interaction in man and conscious dog. Circ Res 1986;59:178-93.
13. Fouad FM, Tarazi RC, Ferrario CM, et al. Assessment of parasympathetic control of HR by a noninvasive method. Am J Physiol 1984;246(6 Pt 2):H838-42.
14. Arai Y, Saul JP, Albrecht P, et al. Modulation of cardiac autonomic activity during and immediately after exercise. Am J Physiol 1989;256(1 Pt 2):H132-41.
15. Saul JP, Rea RF, Eckberg DL, et al. Heart rate and muscle sympathetic nerve variability during reflex changes of autonomic activity. Am J Physiol 1990;258(3 Pt 2):H713-21.
16. Hirsch JA, Bishop B. Respiratory sinus arrhythmia in humans: how breathing pattern modulates HR. Am J Physiol 1981;241:H620-9.
17. Saul JP, Berger RD, Chen MH, Cohen RJ. Transfer function analysis of autonomic regulation. II. Respiratory sinus arrhythmia. Am J Physiol 1989;256(1 Pt 2):H153-61.
18. Angelone A, Coulter NA. Respiratory sinus arrhythmia: a frequency dependent phenomenon. J Appl Physiol 1964;19:479-82.
19. Malmqvist LA, Bengtsson M, Bjornsson G, et al. Sympathetic activity and haemodynamic variables during spinal analgesia in man. Acta Anaesthesiol Scand 1987;31:467-73.
20. Landry DP, Bennett FM, Oriel NE. Analysis of heart rate dynamics as a measure of autonomic tone in obstetrical patients undergoing epidural or spinal anesthesia. Reg Anesth 1994;19:189-95.
21. Blaise GA, Roy WL. Spinal anaesthesia for minor paediatric surgery. Can Anaesth Soc J 1986;33:227-30.
22. Harnik EV, Hoy GR, Potolicchio S, et al. Spinal anesthesia in premature infants recovering from respiratory distress syndrome. Anesthesiology 1986;64:95-9.
23. Gootman PM. Development of central autonomic regulation of cardiovascular function. In: Gootman PM, ed. Developmental neurobiology of the autonomic nervous system. Clifton, NJ: Humana Press, 1986:279-325.
24. Lagercrantz H, Edwards D, Henderson-Smart D, et al. Autonomic reflexes in preterm infants. Acta Paediatr Scand 1990;79:721-8.
© 1995 International Anesthesia Research Society
25. Triedman JK, Cohen RJ, Saul JP. Mild hypovolemic stress alters autonomic modulation of HR. Hypertension 1993;21:236-47.