- Email: [email protected]
Variation of sympathetic reflex latency in man
Journal of the Autonomic Nervous System, 21 (1987) 157-165 Elsevier
157
JAN 00774
Variation of sympathetic reflex latency in man Jan Fagius
1,2, G~Sran
S u n d l S f 3 a n d B. G u n n a r W a l l i n 4
Departments of 1 Neurology, 2 Clinical Neurophysiology and s Internal Medicine, Akademiska Sjukhuset, University of Uppsala, Uppsala (Sweden) and 4 Clinical Neurophysiology, Sahlgren Hospital, University of Gothenburg, Gothenburg (Sweden)
(Received 15 June 1987) (Revised version received and accepted 30 September 1987)
Key words: Muscle nerve sympathetic activity; Microneurography; Reflex latency; Baroreflex; Central sympathetic mechanism
Summary Microelectrode recordings of muscle nerve sympathetic activity (MSA) in man have shown a reflex relationship between heart beat and corresponding sympathetic burst, the latency of which is stable at rest and independent of heart rate. In peroneal nerve recordings in 35 healthy subjects this latency was reduced during the Valsalva manoeuvre by 120 ms (mean; range 40-245 ms; P < 0.001) from a mean value at rest of 1300 ms. Slow deep breathing and simulated diving shortened the latency by 60 (P < 0.001) and 80 ms (P < 0.05), respectively. When intrinsic heart rate was induced by i.v. administration of atropine and propranolol, the latency was increased by 70 ms (P < 0.001). A number of other manoeuvres affecting the outflow of MSA did not change the latency. It is suggested that the findings indicate the existence of more than one central pathway involved in the baroreflex regulation of MSA. Alternatively, altered central processing time may follow influence from other receptors in different manoeuvres.
Introduction H u m a n muscle nerve sympathetic activity (MSA), as recorded by microneurography, appears as bursts of impulses, time-locked in the cardiac r h y t h m [17]. This cardiac rhythmicity is due to inhibitory baroreflex influence, whereby the systolic blood pressure peak causes interruption of the outflow of sympathetic activity [10]. The temporal relationship between h a e m o d y n a m i c and related neural event is stable at rest, irrespective of spontaneous fluctuations of heart rate, and hence a reflex latency can be defined f r o m the heart beat to the corresponding burst of M S A [8]. If this latency is due to impulse traffic in one well-de-
Correspondence: J. Fagius, Department of Neurology, Akademiska Sjukhuset, S-751 85 Uppsala, Sweden.
fined reflex circuit, one would expect it to be relatively stable under different circumstances. We have, however, observed a shortening of the reflex latency during the Valsalva manoeuvre. This p r o m p t e d the present work and we have now systematically analyzed the influence of different experimental perturbations on the baroreflex latency in MSA.
Materials and Methods Nerve recordings A n insulated Tungsten microelectrode (tip diameter about 5 # m ) was inserted manually through the intact skin into the underlying peroneal nerve at the fibular head. A low impedance reference electrode was placed s.c. at a distance of 1 - 2 cm. Electrical stimuli derived through the recording
0165-1838/87/$03.50 © 1987 Elsevier Science Publishers B.V. (Biomedical Division)
158
electrode served to localize the nerve. When the nerve was found, an electrode position within a muscle nerve fascicle was identified by muscle twitches evoked by the electrical stimuli and by the appearance of afferent, mechanoreceptive activity elicited by stretching or tapping the appropriate muscle. Thereafter, minor changes of the electrode position were made until the characteristic pattern of multi-unit MSA was recorded. The evidence that the recorded signals derive from sympathetic fibres has been summarized previously [4]. The search for the nerve may cause minor discomfort, but when a suitable electrode position is found, nothing is felt during the recording. The nerve signal was amplified in two steps (total gain 50,000 × ) and fed through a 700-2000 Hz band pass filter and an amplitude discriminator to improve signal-to-noise ratio. A resistancecapacitance integrating network with time constant 0.1 s delivered a mean voltage display of the multi-unit neural activity. ECG was recorded by chest electrodes. Subjects were supine (except in the diving experiments, c.f. below, when they were prone) on a comfortable bed. Room temperature was 2224 ° C. All recorded signals were stored on tape (FM tape recorder Sangamo Sabre VI, Sangamo Weston-Schlumberger, Sarasota, FL) for subsequent analysis. During the experiments the signals were displayed on a storage oscilloscope.
Subjects and manoeuvres Recordings were performed in a large number of subjects (c.f. Table I), with the primary aim to study the effect of different procedures on the outflow of MSA. The studies were approved by the Ethics Committee of the Medical Faculty at the University of Uppsala and informed consent was obtained from each subject. Tapes from such experiments were re-analyzed for reflex latency. The following manoeuvres were studied: (1) The Valsalva manoeuvre. Subjects were asked to inspire deeply and attempt to expire forcibly against the closed glottis for 12-15 s. On a few occasions subjects were instead instructed to blow into a tube connected to a manometer and hold a pressure of 40 m m Hg for 15 s. This
manoeuvre regularly evokes a strong increase of MSA [5]. In addition to healthy subjects we included 3 patients with diabetes mellitus, two of whom exhibited signs of slight polyneuropathy. and 4 patients with acute inflammatory polyradiculoneuropathy without signs of autonomic involvement. (2) Slow deep breathing (0.1 Hz), which causes a grouping of the outflow of MSA in the respiratory rhythm [6]. (3) Simulated diving. This procedure elicits a strong increase of MSA [11]. (4) Apnoea. A brief apnoea period exerts little influence on MSA. whereas a prolonged apnoea with resulting hypoxia is accompanied by an increased outflow [11]. (5) Intravenous injection of atropme, which strongly reduces or abolishes the outflow of MSA (Wallin. B.G.. Blumberg, H. and Sundltif. G.. unpublished observation) naturally, only recordhags in which MSA persisted after the injection could be analyzed. (6) Intravenous injection of the B-blocking agents propranolol, which is followed by an increase of MSA (Wallin, B.G.. Blumberg, H. and SundliSf, G.. unpublished observation). (7) Intravenous injection of atropine and propranolol, giving rise to an intrinsic heart rate due to pharmacological denervation. (8) Application of lower body negative pressure, which brings about an increased outflow of MSA [16]. (9) Cold pressure test, i.e. immersion of one hand into ice water for 1 min. During this procedure MSA increases in most subjects [9,18]. (10) Isometric muscle work, performed as a sustained hand grip at 30% of maximal effort for 2 min. A significant increase of MSA occurs during the second minute of this manoeuvre [13]. (11) Intravenous injection of sodium nitroprusside (3.2 Cg/kg/min), which acts as a peripheral vasodilator, thereby reducing peripheral resistance and blood pressure. As a consequence there is a marked increase in the outflow of MSA [7].
Reflex latency determination The method for refle,x latency determj'nation is illustrated in Fig. 1A. Since the l~ro~-d'mx is in-
159 hibitory, the latency is defined as the delay from the R wave of the ECG to the peak of the corresponding burst of MSA in the mean voltage neurogram, the peak being interpreted as the beginning of inhibition [8]. Measured in this way the latency will include both haemodynamic events and neural conduction time. In peroneal nerve recordings the latency lies in the range of 1.2-1.5 s with a close correlation to the length of the extremity in question. The way of determining what burst of MSA is corresponding to a certain heart beat has been described previously [8]. In brief, by comparing R-wave triggered oscilloscope sweeps of the neurogram at different heart rates it was found that only one burst is time-locked to the triggering heart beat, independent of heart rate. Once the approximate latency to the appropriate burst has been defined, the latency to selected bursts is easy to measure (c.f. Fig. 1). For this purpose, 25 m m / s paper displays of ECG and mean voltage neurograms at rest and during manoeuvres were used. Latencies were measured by use of a digitizing board (Hipad, Houston Instruments, Austin, TX), connected to a computer (Digital Dec 11/40, Digital Equipment, Maynard, MA). For each subject the latency at rest was calculated as the mean of 50 consecutive individual latencies. The reliability of this method was checked in 17 subjects by repeating this procedure for another 50 consecutive bursts of MSA. During manoeuvres the latency was measured as the mean of 30-50 individual latencies or, d u r i n g short-lasting manoeuvres, as the mean of all occurring bursts. With few exceptions, short-lasting manoeuvres were repeated 2-4 times and the mean value was calculated. To estimate the duration of the cardiovascular events included in the reflex latency, simultaneous ECG and carotid pulse wave recordings were made in two subjects. The delay between the R wave in the ECG and the steep upslope of the carotid pulse wave were calculated for 25 individual heart beats at rest and during the Valsalva manoeuvre, and mean values were compared. Statistical analysis
Values are given as mean + S.D. For compari-
son of group mean latencies at rest and during a manoeuvre, two-tailed Student's t-test for paired observations was applied. Due to the varying number of observations for different manoeuvres (c.f. Table I), the latency at rest used for comparison varied from one analysis to the other. In order to simplify the result presentation, only mean change of latency from the resting value in question is given; however, statistical analysis is based on absolute values. Results Fig. 1 illustrates how the reflex latency is defined and its independence of heart rate variation in one subject. Mean reflex latency at rest in 35 subjects was 1300 ms (1297 + 93). When two different rest periods were compared in 17 subjects, the value was 1295 ___95 and 1302 ___91 ms, respectively. The changes of reflex latency from rest to different manoeuvres are summarized in Table I. The Valsalva manoeuvre. A mean reduction in reflex latency of 120 ms was observed in 35 subjects (P < 0.001; Table I). There was, however, considerable variation between subjects. In one subject the latency was unchanged, and the range of reduction in the remaining subjects was 40-245 ms. An example of the distribution of reflex latency values at rest and during the manoeuvre is given in Fig. 2A. There was no stepwise change of latency but a gradual reduction and the temporal pattern shown in Fig. 2B was common (seen in 12 of 18 recordings analyzed in this way). Usually the maximal reduction of latency occurred around the 5th burst of MSA during the manoeuvre and then there was a gradual return toward resting values. However, exceptions from this pattern were seen with a random distribution of shorter than normal latencies during the manoeuvre. As seen in Fig. 2B, individual latency values could reach a very low level, but the range of individual latency values did not increase during the manoeuvre; instead the whole distribution was moved to a lower level (mean difference between longest and shortest latency at rest was 300 + 65 ms and during Valsalva manoeuvre 310 + 73 ms).
160
A
t
B
t ,
~" 1300"
i
-: ~
1200"
MSA !
!
ls
R'R interval (ms) Fig. 1. A: exam#e of an MSA recotdinR (mean voltage neuxogram) with indication of reflex latency from R wave in the ECG to peak of corresponding burst of MSA. B: distPd3ution of reflex latencies at different heart rates in one subject. Each dot represents the mean latency of 2-8 MSA bursts and corresponding R-R interval.
The reduction of reflex latency during the Valsalva manoeuvre was observed also in 3 subjects with diabetes mellitus and 4 patients with acute inflammatory polyradieuloneuropathy, At a follow-up recording after recovery in 3 of the latter patients, the same degree of reduction was found in each subject. A second recording was made also in 7 healthy subjects, all of whom displayed similar reductions on both occasions. Slow deep breathing. A significant shortening of reflex latency (60 ms; P < 0.001, Table I) was observed also during 0.1 Hz deep b r e a ~ . One subject did not reduce the latency during this procedure (but did so when performing the Valsalva manoeuvre), and the range in the remaining 18 subjects was 30-95 ms. Also this shortening of latency was due to a redistribution of individual latency range to a lower level without increase of variation during the manoeuvre; Lndividual latencies were randomly distributed w i ~ this range without any temporal pattern during the respiratory cycle. This manoeuvre was repeated in a second recording in 7 subjects, w h o all exhibited a similar degree of reduction on both ~ o n s . To further elucidate the influence of respiration, reflex ~ ¢kn'ing breathing with positive end expiratory pressure in two subjects. M s procedure is a ~ t e d with an enhanced outflow of M ~ (Wallin, B.G., unpub-
lished observation), but did not cause any change of reflex latency. Diving. Simulated diving entailed a mean reduction of reflex latency of 80 ms ( P < 0.05: Table I) with a wide range (10-190 ms) in the 6 subjects studied. The reduced latency values were randomly distributed during the procedure without any temporal development. During apnoea no change of the reflex latency was observed. This was independent of whether the apnoea was brief or of maximal duration, and in the latter situation no change was found in-the late phase of the breath-holding, when hypoxia presumably was present. Intrinsic heart rate was associated with a significant increase of reflex latency (70 ms: P < 0.001; Table I) with a wide range, 0-90 ms. In 3 subjects a total of 4 Valsalva manoeuvres were performed at intrinsic heart rate, and on each occasion a reduction of latency from the value seen at intrinsic heart rate to a value lower than the one at rest was observed (range 140-200 ms from the intrinsic heart rate value). Propranolol and atropine injected separately did not change the reflex latency. Lower body negatioe pressure, cold pressure test, isometric muscle work, and administration of sodium nitropmsside likewise caused no change of the reflex latency. The mean delay between the R wave in the ECG and the carotid pulse wave was 100 ms m both subjects, in whom such measurements were
161
e~ ~4D N
i.
o~
.{ +Z
Rv~ 4~
4-~
"~ i
~
+Z
v
o
e.O
~z
I
T~ a~
[-
I
162
A
B 1400-
m
A
10 u~
..g 1200 X
,..6 O O
Z
!
100(
12oo
1600
Reflex latency (ms)
1000
i
J
;b
Burst no
Fig. 2. Influence of Valsalva manoeuvre on reflex latency in one subject. A: distribution of individual latency values during Valsalva manoeuvre (open bars; summary of 3 manoeuvres; n = 43; mean latency 1240 ms) and at rest (shaded bars, n = 50; m e ~ latency 1430 ms). B: temporal course, burst by burst, of reflex latency values during Valsalva manoenvre. Averaged values from 3 superimposed manoeuvres.
made. During the Valsalva manoeuvre this delay was increased by 10 ms.
Discussion
Our main findings were (1) a considerable shortening of the reflex latency from the R wave of the ECG to the peak of the corresponding burst of MSA during the Valsalva manoeuvre, deep breathing and simulated diving, and (2) a lengthening of the latency when intrinsic heart rate was induced, whereas (3) a number of other manoeuvres changing the outflow of MSA did not affect the latency.
Characteristics of the measured latency At rest there is a considerable variation between latencies measured for individual bursts at rest (Fig. 2A). Despite this, mean latency values are reproducible from one recording to another [8], and in the present study measurement of mean latencies at rest from different se~aants within the same recording provided identical values. The
change of latency does not seem to evolve from alterations of measurement conditions during the manoeuvres since (1) the R wave of the ECG remained well defined during each p r o c ~ u r e and there is no reason to assume a change of thetime relationship between the mechanical heart beat and the recorded R wave during e.g. the Valsalva manoeuvre; and (2) the bursts of MSA had their normal appearance during all manoeuvres with maintained pulse synchrony and a clearly detectable peak in the mean voltage neurogram. The fact that the change of reflex latency was due to a redistribution of the range of latencies and not to an increased variability also speaks again the possibility of an artefact. Neither is the change of latency due to an error in determination of the 'appropriate burst', since selection of neighbouring bursts would have given values diverging 700-1500 ms from the latency at rest. The reflex latency is not del~ndent on heart rate (c.f. Fig. 113) and the lack of correlation between heart rate and reflex latency during manoeuvres ctmai0ng or not c h ~ the latency (see Table I) further excludes the possibiti~ that
163 the latency should vary with R-R interval changes. This conclusion is corroborated by the lack of influence on reflex latency following administration of propranolol or atropine. Consequently, we postulate that the observed changes of reflex latency represent a physiological effect of the procedures performed. Possible cause of the change A number of hypothetical explanations for the effects observed may be put forward: change of propagation velocity of the arterial pulse wave during a manoeuvre; activation of receptors not involved in the events responsible for the reflex latency at rest; complex central interactions and activation of alternative central reflex pathways during manoeuvres; and change of efferent conduction time. Finally, complex pharmacological influence may occur when intrinsic heart rate was induced. Pulse wave propagation Pulse wave propagation was studied during the Valsalva manoeuvre only. The lack of change during this procedure, which caused the biggest reduction of reflex latency, makes it very unlikely that a faster pressure wave propagation contributes to the observed reflex changes. Moreover, it has been estimated that cardiovascular events account for less than 100-150 ms of the measured reflex latency [8], which is in accordance with the direct measurements performed in the present study. This means that a change in the duration of cardiovascular events alone cannot explain a latency reduction averaging 120 ms. Activation of other receptors As mentioned above, arterial baroreceptors are responsible for the cardiac rhythmicity of MSA. The wide range of manoeuvres applied in this study involves a variety of other receptors of importance for cardiovascular regulation. The Valsalva manoeuvre is complex with involvement of both arterial and cardiopulmonary baroreceptors as well as lung stretch receptors; to some extent deep breathing may have similar effects. From animal studies it has been suggested that aortic inputs may reach the medulla 10-20 ms
before the sinus nerve input [12], which might be of some importance for the present effects in humans. However, application of lower body negative pressure causes unloading of intrathoracic baroreceptors [2], and administration of sodium nitroprusside presumably influences both high-pressure and intrathoracic receptors, and those two procedures did not change the reflex latency. At diving, a composite input from facial receptors seems to be a potent activator of MSA and strong bursts may be seen so early during diving that the corresponding R-R interval must have occurred before diving [11]. This observation indicates that a fast mechanism comes into action, which may result in an altered time relationship between heart beat and MSA. Apnoea plays an important role for the diving response of MSA [11], but apnoea did not influence the reflex latency. Activation of arterial chemoreceptors presumably occurred during prolonged apnoea, without change of reflex latency. The increase of MSA during isometric muscle work is assumed to be caused by chemoreceptors in the muscle [13]. Obviously, this receptor influence did not affect reflex latency. Central mechanisms The time for central processing of baroreceptor inputs in humans has been estimated with different techniques to be 0.25-0.30 s [1,8]. This remarkably long delay may imply that many interneurons are involved in the reflex arc and therefore activation of faster central pathways would be a possible explanation for a reduced reflex latency. McAllen and Spyer [14] found that central vagal motoneurons in anaesthetized cats could be activated by sinus nerve input with two distinct ranges of central delay. The functional significance of their observation is unclear, but it implies that different time relationships exist between sinus nerve input and a given central neuron responsible for evoked efferent activity. Another observation that may be relevant for the present results is that by McCloskey and Potter [15], who described, in rats, a shorter latency from sinus nerve stimulation to the evoked response in a given vagal cardiac efferent fibre during expira-
164
tion than during inspiration. In the present study all manoeuvres associated with reduced latency involved respiratory function and it is possible that some form of respiratory modulation of the baroreceptor input to sympathetic motoneurones may contribute to the baroreceptor input to sympathetic motoneurons may contribute to the effect. In rats there seem to be two subsets of efferent cardiovascular neurones, projecting from the rostral ventrolateral medulla to the intermediolateral cell column of the spinal cord, with mean conduction velocities of 0.64 and 3.83 m/s, respectively [3]. However, if there were two discrete reflex pathways between the high-pressure receptors and sympathetic vasoconstrictor neurons to muscle, one would expect two discrete subsets of latency values. Therefore, the varying reduction of reflex during different manoeuvres (see Table I) may imply that even more pathways exist. Alternatively, the variable reduction may be caused by complex modulation of the baroreceptor input by inflow from other receptors in different manoeuvres. Such as possibility may also explain the times course of reflex latency changes during the Valsalva manoeuvre.
Efferent conduction time Efferent conduction time depends on conduction velocity in pre- and postga.nglionic fibres and ganglionic transmission. Conduction velocity of single fibres should be stable. The possibility that faster fibres, usually not in use, should be activated is remote, since an increased time dispersion of bursts with disturbance of pulse synchrony should follow, and this was not the case. Variation of transmission time due to facilitation in the sympathetic ganglion is not likely to cause any major contribution to the shortening of latency.
Central pharmacological actions Central pharmacological actions must be considered for the increase of reflex latency that occurred when intrinsic heart rate was induced by pharmacological denervation of the heart. The effects of the combined administration of a nonselective fl-blocking agent and atropine on central
sympathetic connections is unclear. Interestingly, however, the reduction of reflex latency was still evoked when a Valsalva manoeuvre was performed at intrinsic heart rate.
Significance of observed effects The functional importance of the reduced reflex latency is difficult to evaluate. From a teleological point of view one might assume that faster pathways would be made use of in emer, gency situations. The Valsalva manoeuvre and diving would fit such a reasoning, but apnoea causing hypoxia and the effect of sodium nitroprusside might be regarded as equivalent emergencies and they did not reduce reflex latency. The appearance of a shorter than normal sympathetic reflex latency is of neurophysiological interest, since it seems to verify in humans the existence of more than one central pathway, or variable central processing time, between baxoreceptors of the high-pressure type and sympathetic vasoconstrictor neurons to muscle vessels.
Acknowledgements Supported by the Swedish Medical Research Council, Grants no. B86-04X-7468,01A and B8614X-3546-15B, and by the Swedish Society of Medical Sciences. The authors thank Dr. Finn Mannting, who assisted in measurements of arterial pulse wave propagation velocity and Mrs. Catarina F~irnstrand, who spent many hours measuring reflex latencies on the digitizing board.
References 1 Borst. C. and Karemaker, J.M., Time delays in the human barorcceptor reflex. J. Auton. Nerv. Syst., 9 (1983) 399-409. 2 Brown, E., Goci, J.S., Greenfield, A.D.M. and Plassaras, G.C.. Circulatory responses to simulated gravitational shifts of blood in man induced by exposure of the bodybelow-the iliac crest to subatmospherie pressure. J. Physiol. (London), 183 (1966) 607-627. 3 Brown, D.L. and Guyenet, P.G., Electrophysiological study of cardiovascular neurons in the rostral ventrolateral medulla in rats, Circ. Res., 56 (t985) 359-369. 4 Delius, W., Hagbarth, K.-E., Hongdl, A. and Wallin, B.G.. Genvral c h a t ~ c s of sympathetic acti~ty in human muscle nerves, Acta Physiol. Scand. 84 (t97'2) 65L81.
165 5 Delius, W., Hagbarth, K.-E., Hongell, A. and Wallin, B.G., Manoeuvres affecting sympathetic outflow in human muscle nerves, Acta Physiol. Scand., 84 (1972) 82-94. 6 Eckberg, D.L., Nerhed, C. and Wallin, B.G., Respiratory modulation of muscle sympathetic and vagal cardiac outflow in man, J. Physiol. (London), 365 (1985) 181-196. 7 Eckberg, D.L., Rea, R.F., Andersson, O.K., Hedner, T., Pernow, J., Lundberg, J. and Wallin, B.G., Baroreflex modulation of sympathetic activity and sympathetic neurotransmitters in man, submitted. 8 Fagius, J. and Wallin, B.G., Sympathetic reflex latencies and conduction velocities in normal man, J. Neurol. Sci., 47 (1980) 433-448. 9 Fagius, J., Muscle nerve sympathetic activity in migraine. Lack of abnormality, Cephalalgia, 5 (1985) 197-203. 10 Fagius, J., Wallin, B.G., Sundli3f, G., Nerhed, L. and Englesson, S., Sympathetic outflow in man after anaesthesia of the glossopharyngeal and vagus nerves, Brain, 108 (1985) 423-438. 11 Fagius, J. and SundliSf, G., The diving response in man: effects on sympathetic activity in muscle and skin nerve fascicles, J. Physiol. (London), 377 (1986) 429-443. 12 Gabriel, M. and Seller, H., Interaction of baroreceptor
13
14
15
16
17
18
afferents from carotid sinus and aorta at the nucleus tractus solitarii, Pfliigers Arch., 318 (1970) 7-20. Mark, A.L., Victor, R.G., Nerhed, C. and Wallin, B.G., Microneurographic studies of the mechanisms of sympathetic nerve responses to static exercise in humans, Circ. Res., 57 (1985) 461-469. McAllen, R.M. and Spyer, K.M., The baroreceptor input to cardiac vagal motoneurones, J. Physiol. (London), 282 (1978) 365-374. McCloskey, D.I. and Potter, E.K., Excitation and inhibition of cardiac vagal motoneurones by electrical stimulation of the carotid sinus nerve, J. Physiol. (London), 316 (1981) 163-175. SundliSf, G. and Wallin, B.G., Effect of lower body negative pressue on human muscle nerve sympathetic activity, J. Physiol. (London), 278 (1978) 525-532. Wallin, B.G. and Fagius, J., The sympathetic nervous system in man. Aspects derived from microelectrode recordings, Trends Neurosci., 9 (1986) 63-67. Victor, R.G., Leimbach, W.N., Seals, D.R., Wallin, B.G. and Mark, A.L., Effects of the cold pressor test on muscle sympathetic nerve activity, Hypertension, 9 (1987) 429-436.
157
JAN 00774
Variation of sympathetic reflex latency in man Jan Fagius
1,2, G~Sran
S u n d l S f 3 a n d B. G u n n a r W a l l i n 4
Departments of 1 Neurology, 2 Clinical Neurophysiology and s Internal Medicine, Akademiska Sjukhuset, University of Uppsala, Uppsala (Sweden) and 4 Clinical Neurophysiology, Sahlgren Hospital, University of Gothenburg, Gothenburg (Sweden)
(Received 15 June 1987) (Revised version received and accepted 30 September 1987)
Key words: Muscle nerve sympathetic activity; Microneurography; Reflex latency; Baroreflex; Central sympathetic mechanism
Summary Microelectrode recordings of muscle nerve sympathetic activity (MSA) in man have shown a reflex relationship between heart beat and corresponding sympathetic burst, the latency of which is stable at rest and independent of heart rate. In peroneal nerve recordings in 35 healthy subjects this latency was reduced during the Valsalva manoeuvre by 120 ms (mean; range 40-245 ms; P < 0.001) from a mean value at rest of 1300 ms. Slow deep breathing and simulated diving shortened the latency by 60 (P < 0.001) and 80 ms (P < 0.05), respectively. When intrinsic heart rate was induced by i.v. administration of atropine and propranolol, the latency was increased by 70 ms (P < 0.001). A number of other manoeuvres affecting the outflow of MSA did not change the latency. It is suggested that the findings indicate the existence of more than one central pathway involved in the baroreflex regulation of MSA. Alternatively, altered central processing time may follow influence from other receptors in different manoeuvres.
Introduction H u m a n muscle nerve sympathetic activity (MSA), as recorded by microneurography, appears as bursts of impulses, time-locked in the cardiac r h y t h m [17]. This cardiac rhythmicity is due to inhibitory baroreflex influence, whereby the systolic blood pressure peak causes interruption of the outflow of sympathetic activity [10]. The temporal relationship between h a e m o d y n a m i c and related neural event is stable at rest, irrespective of spontaneous fluctuations of heart rate, and hence a reflex latency can be defined f r o m the heart beat to the corresponding burst of M S A [8]. If this latency is due to impulse traffic in one well-de-
Correspondence: J. Fagius, Department of Neurology, Akademiska Sjukhuset, S-751 85 Uppsala, Sweden.
fined reflex circuit, one would expect it to be relatively stable under different circumstances. We have, however, observed a shortening of the reflex latency during the Valsalva manoeuvre. This p r o m p t e d the present work and we have now systematically analyzed the influence of different experimental perturbations on the baroreflex latency in MSA.
Materials and Methods Nerve recordings A n insulated Tungsten microelectrode (tip diameter about 5 # m ) was inserted manually through the intact skin into the underlying peroneal nerve at the fibular head. A low impedance reference electrode was placed s.c. at a distance of 1 - 2 cm. Electrical stimuli derived through the recording
0165-1838/87/$03.50 © 1987 Elsevier Science Publishers B.V. (Biomedical Division)
158
electrode served to localize the nerve. When the nerve was found, an electrode position within a muscle nerve fascicle was identified by muscle twitches evoked by the electrical stimuli and by the appearance of afferent, mechanoreceptive activity elicited by stretching or tapping the appropriate muscle. Thereafter, minor changes of the electrode position were made until the characteristic pattern of multi-unit MSA was recorded. The evidence that the recorded signals derive from sympathetic fibres has been summarized previously [4]. The search for the nerve may cause minor discomfort, but when a suitable electrode position is found, nothing is felt during the recording. The nerve signal was amplified in two steps (total gain 50,000 × ) and fed through a 700-2000 Hz band pass filter and an amplitude discriminator to improve signal-to-noise ratio. A resistancecapacitance integrating network with time constant 0.1 s delivered a mean voltage display of the multi-unit neural activity. ECG was recorded by chest electrodes. Subjects were supine (except in the diving experiments, c.f. below, when they were prone) on a comfortable bed. Room temperature was 2224 ° C. All recorded signals were stored on tape (FM tape recorder Sangamo Sabre VI, Sangamo Weston-Schlumberger, Sarasota, FL) for subsequent analysis. During the experiments the signals were displayed on a storage oscilloscope.
Subjects and manoeuvres Recordings were performed in a large number of subjects (c.f. Table I), with the primary aim to study the effect of different procedures on the outflow of MSA. The studies were approved by the Ethics Committee of the Medical Faculty at the University of Uppsala and informed consent was obtained from each subject. Tapes from such experiments were re-analyzed for reflex latency. The following manoeuvres were studied: (1) The Valsalva manoeuvre. Subjects were asked to inspire deeply and attempt to expire forcibly against the closed glottis for 12-15 s. On a few occasions subjects were instead instructed to blow into a tube connected to a manometer and hold a pressure of 40 m m Hg for 15 s. This
manoeuvre regularly evokes a strong increase of MSA [5]. In addition to healthy subjects we included 3 patients with diabetes mellitus, two of whom exhibited signs of slight polyneuropathy. and 4 patients with acute inflammatory polyradiculoneuropathy without signs of autonomic involvement. (2) Slow deep breathing (0.1 Hz), which causes a grouping of the outflow of MSA in the respiratory rhythm [6]. (3) Simulated diving. This procedure elicits a strong increase of MSA [11]. (4) Apnoea. A brief apnoea period exerts little influence on MSA. whereas a prolonged apnoea with resulting hypoxia is accompanied by an increased outflow [11]. (5) Intravenous injection of atropme, which strongly reduces or abolishes the outflow of MSA (Wallin. B.G.. Blumberg, H. and Sundltif. G.. unpublished observation) naturally, only recordhags in which MSA persisted after the injection could be analyzed. (6) Intravenous injection of the B-blocking agents propranolol, which is followed by an increase of MSA (Wallin, B.G.. Blumberg, H. and SundliSf, G.. unpublished observation). (7) Intravenous injection of atropine and propranolol, giving rise to an intrinsic heart rate due to pharmacological denervation. (8) Application of lower body negative pressure, which brings about an increased outflow of MSA [16]. (9) Cold pressure test, i.e. immersion of one hand into ice water for 1 min. During this procedure MSA increases in most subjects [9,18]. (10) Isometric muscle work, performed as a sustained hand grip at 30% of maximal effort for 2 min. A significant increase of MSA occurs during the second minute of this manoeuvre [13]. (11) Intravenous injection of sodium nitroprusside (3.2 Cg/kg/min), which acts as a peripheral vasodilator, thereby reducing peripheral resistance and blood pressure. As a consequence there is a marked increase in the outflow of MSA [7].
Reflex latency determination The method for refle,x latency determj'nation is illustrated in Fig. 1A. Since the l~ro~-d'mx is in-
159 hibitory, the latency is defined as the delay from the R wave of the ECG to the peak of the corresponding burst of MSA in the mean voltage neurogram, the peak being interpreted as the beginning of inhibition [8]. Measured in this way the latency will include both haemodynamic events and neural conduction time. In peroneal nerve recordings the latency lies in the range of 1.2-1.5 s with a close correlation to the length of the extremity in question. The way of determining what burst of MSA is corresponding to a certain heart beat has been described previously [8]. In brief, by comparing R-wave triggered oscilloscope sweeps of the neurogram at different heart rates it was found that only one burst is time-locked to the triggering heart beat, independent of heart rate. Once the approximate latency to the appropriate burst has been defined, the latency to selected bursts is easy to measure (c.f. Fig. 1). For this purpose, 25 m m / s paper displays of ECG and mean voltage neurograms at rest and during manoeuvres were used. Latencies were measured by use of a digitizing board (Hipad, Houston Instruments, Austin, TX), connected to a computer (Digital Dec 11/40, Digital Equipment, Maynard, MA). For each subject the latency at rest was calculated as the mean of 50 consecutive individual latencies. The reliability of this method was checked in 17 subjects by repeating this procedure for another 50 consecutive bursts of MSA. During manoeuvres the latency was measured as the mean of 30-50 individual latencies or, d u r i n g short-lasting manoeuvres, as the mean of all occurring bursts. With few exceptions, short-lasting manoeuvres were repeated 2-4 times and the mean value was calculated. To estimate the duration of the cardiovascular events included in the reflex latency, simultaneous ECG and carotid pulse wave recordings were made in two subjects. The delay between the R wave in the ECG and the steep upslope of the carotid pulse wave were calculated for 25 individual heart beats at rest and during the Valsalva manoeuvre, and mean values were compared. Statistical analysis
Values are given as mean + S.D. For compari-
son of group mean latencies at rest and during a manoeuvre, two-tailed Student's t-test for paired observations was applied. Due to the varying number of observations for different manoeuvres (c.f. Table I), the latency at rest used for comparison varied from one analysis to the other. In order to simplify the result presentation, only mean change of latency from the resting value in question is given; however, statistical analysis is based on absolute values. Results Fig. 1 illustrates how the reflex latency is defined and its independence of heart rate variation in one subject. Mean reflex latency at rest in 35 subjects was 1300 ms (1297 + 93). When two different rest periods were compared in 17 subjects, the value was 1295 ___95 and 1302 ___91 ms, respectively. The changes of reflex latency from rest to different manoeuvres are summarized in Table I. The Valsalva manoeuvre. A mean reduction in reflex latency of 120 ms was observed in 35 subjects (P < 0.001; Table I). There was, however, considerable variation between subjects. In one subject the latency was unchanged, and the range of reduction in the remaining subjects was 40-245 ms. An example of the distribution of reflex latency values at rest and during the manoeuvre is given in Fig. 2A. There was no stepwise change of latency but a gradual reduction and the temporal pattern shown in Fig. 2B was common (seen in 12 of 18 recordings analyzed in this way). Usually the maximal reduction of latency occurred around the 5th burst of MSA during the manoeuvre and then there was a gradual return toward resting values. However, exceptions from this pattern were seen with a random distribution of shorter than normal latencies during the manoeuvre. As seen in Fig. 2B, individual latency values could reach a very low level, but the range of individual latency values did not increase during the manoeuvre; instead the whole distribution was moved to a lower level (mean difference between longest and shortest latency at rest was 300 + 65 ms and during Valsalva manoeuvre 310 + 73 ms).
160
A
t
B
t ,
~" 1300"
i
-: ~
1200"
MSA !
!
ls
R'R interval (ms) Fig. 1. A: exam#e of an MSA recotdinR (mean voltage neuxogram) with indication of reflex latency from R wave in the ECG to peak of corresponding burst of MSA. B: distPd3ution of reflex latencies at different heart rates in one subject. Each dot represents the mean latency of 2-8 MSA bursts and corresponding R-R interval.
The reduction of reflex latency during the Valsalva manoeuvre was observed also in 3 subjects with diabetes mellitus and 4 patients with acute inflammatory polyradieuloneuropathy, At a follow-up recording after recovery in 3 of the latter patients, the same degree of reduction was found in each subject. A second recording was made also in 7 healthy subjects, all of whom displayed similar reductions on both occasions. Slow deep breathing. A significant shortening of reflex latency (60 ms; P < 0.001, Table I) was observed also during 0.1 Hz deep b r e a ~ . One subject did not reduce the latency during this procedure (but did so when performing the Valsalva manoeuvre), and the range in the remaining 18 subjects was 30-95 ms. Also this shortening of latency was due to a redistribution of individual latency range to a lower level without increase of variation during the manoeuvre; Lndividual latencies were randomly distributed w i ~ this range without any temporal pattern during the respiratory cycle. This manoeuvre was repeated in a second recording in 7 subjects, w h o all exhibited a similar degree of reduction on both ~ o n s . To further elucidate the influence of respiration, reflex ~ ¢kn'ing breathing with positive end expiratory pressure in two subjects. M s procedure is a ~ t e d with an enhanced outflow of M ~ (Wallin, B.G., unpub-
lished observation), but did not cause any change of reflex latency. Diving. Simulated diving entailed a mean reduction of reflex latency of 80 ms ( P < 0.05: Table I) with a wide range (10-190 ms) in the 6 subjects studied. The reduced latency values were randomly distributed during the procedure without any temporal development. During apnoea no change of the reflex latency was observed. This was independent of whether the apnoea was brief or of maximal duration, and in the latter situation no change was found in-the late phase of the breath-holding, when hypoxia presumably was present. Intrinsic heart rate was associated with a significant increase of reflex latency (70 ms: P < 0.001; Table I) with a wide range, 0-90 ms. In 3 subjects a total of 4 Valsalva manoeuvres were performed at intrinsic heart rate, and on each occasion a reduction of latency from the value seen at intrinsic heart rate to a value lower than the one at rest was observed (range 140-200 ms from the intrinsic heart rate value). Propranolol and atropine injected separately did not change the reflex latency. Lower body negatioe pressure, cold pressure test, isometric muscle work, and administration of sodium nitropmsside likewise caused no change of the reflex latency. The mean delay between the R wave in the ECG and the carotid pulse wave was 100 ms m both subjects, in whom such measurements were
161
e~ ~4D N
i.
o~
.{ +Z
Rv~ 4~
4-~
"~ i
~
+Z
v
o
e.O
~z
I
T~ a~
[-
I
162
A
B 1400-
m
A
10 u~
..g 1200 X
,..6 O O
Z
!
100(
12oo
1600
Reflex latency (ms)
1000
i
J
;b
Burst no
Fig. 2. Influence of Valsalva manoeuvre on reflex latency in one subject. A: distribution of individual latency values during Valsalva manoeuvre (open bars; summary of 3 manoeuvres; n = 43; mean latency 1240 ms) and at rest (shaded bars, n = 50; m e ~ latency 1430 ms). B: temporal course, burst by burst, of reflex latency values during Valsalva manoenvre. Averaged values from 3 superimposed manoeuvres.
made. During the Valsalva manoeuvre this delay was increased by 10 ms.
Discussion
Our main findings were (1) a considerable shortening of the reflex latency from the R wave of the ECG to the peak of the corresponding burst of MSA during the Valsalva manoeuvre, deep breathing and simulated diving, and (2) a lengthening of the latency when intrinsic heart rate was induced, whereas (3) a number of other manoeuvres changing the outflow of MSA did not affect the latency.
Characteristics of the measured latency At rest there is a considerable variation between latencies measured for individual bursts at rest (Fig. 2A). Despite this, mean latency values are reproducible from one recording to another [8], and in the present study measurement of mean latencies at rest from different se~aants within the same recording provided identical values. The
change of latency does not seem to evolve from alterations of measurement conditions during the manoeuvres since (1) the R wave of the ECG remained well defined during each p r o c ~ u r e and there is no reason to assume a change of thetime relationship between the mechanical heart beat and the recorded R wave during e.g. the Valsalva manoeuvre; and (2) the bursts of MSA had their normal appearance during all manoeuvres with maintained pulse synchrony and a clearly detectable peak in the mean voltage neurogram. The fact that the change of reflex latency was due to a redistribution of the range of latencies and not to an increased variability also speaks again the possibility of an artefact. Neither is the change of latency due to an error in determination of the 'appropriate burst', since selection of neighbouring bursts would have given values diverging 700-1500 ms from the latency at rest. The reflex latency is not del~ndent on heart rate (c.f. Fig. 113) and the lack of correlation between heart rate and reflex latency during manoeuvres ctmai0ng or not c h ~ the latency (see Table I) further excludes the possibiti~ that
163 the latency should vary with R-R interval changes. This conclusion is corroborated by the lack of influence on reflex latency following administration of propranolol or atropine. Consequently, we postulate that the observed changes of reflex latency represent a physiological effect of the procedures performed. Possible cause of the change A number of hypothetical explanations for the effects observed may be put forward: change of propagation velocity of the arterial pulse wave during a manoeuvre; activation of receptors not involved in the events responsible for the reflex latency at rest; complex central interactions and activation of alternative central reflex pathways during manoeuvres; and change of efferent conduction time. Finally, complex pharmacological influence may occur when intrinsic heart rate was induced. Pulse wave propagation Pulse wave propagation was studied during the Valsalva manoeuvre only. The lack of change during this procedure, which caused the biggest reduction of reflex latency, makes it very unlikely that a faster pressure wave propagation contributes to the observed reflex changes. Moreover, it has been estimated that cardiovascular events account for less than 100-150 ms of the measured reflex latency [8], which is in accordance with the direct measurements performed in the present study. This means that a change in the duration of cardiovascular events alone cannot explain a latency reduction averaging 120 ms. Activation of other receptors As mentioned above, arterial baroreceptors are responsible for the cardiac rhythmicity of MSA. The wide range of manoeuvres applied in this study involves a variety of other receptors of importance for cardiovascular regulation. The Valsalva manoeuvre is complex with involvement of both arterial and cardiopulmonary baroreceptors as well as lung stretch receptors; to some extent deep breathing may have similar effects. From animal studies it has been suggested that aortic inputs may reach the medulla 10-20 ms
before the sinus nerve input [12], which might be of some importance for the present effects in humans. However, application of lower body negative pressure causes unloading of intrathoracic baroreceptors [2], and administration of sodium nitroprusside presumably influences both high-pressure and intrathoracic receptors, and those two procedures did not change the reflex latency. At diving, a composite input from facial receptors seems to be a potent activator of MSA and strong bursts may be seen so early during diving that the corresponding R-R interval must have occurred before diving [11]. This observation indicates that a fast mechanism comes into action, which may result in an altered time relationship between heart beat and MSA. Apnoea plays an important role for the diving response of MSA [11], but apnoea did not influence the reflex latency. Activation of arterial chemoreceptors presumably occurred during prolonged apnoea, without change of reflex latency. The increase of MSA during isometric muscle work is assumed to be caused by chemoreceptors in the muscle [13]. Obviously, this receptor influence did not affect reflex latency. Central mechanisms The time for central processing of baroreceptor inputs in humans has been estimated with different techniques to be 0.25-0.30 s [1,8]. This remarkably long delay may imply that many interneurons are involved in the reflex arc and therefore activation of faster central pathways would be a possible explanation for a reduced reflex latency. McAllen and Spyer [14] found that central vagal motoneurons in anaesthetized cats could be activated by sinus nerve input with two distinct ranges of central delay. The functional significance of their observation is unclear, but it implies that different time relationships exist between sinus nerve input and a given central neuron responsible for evoked efferent activity. Another observation that may be relevant for the present results is that by McCloskey and Potter [15], who described, in rats, a shorter latency from sinus nerve stimulation to the evoked response in a given vagal cardiac efferent fibre during expira-
164
tion than during inspiration. In the present study all manoeuvres associated with reduced latency involved respiratory function and it is possible that some form of respiratory modulation of the baroreceptor input to sympathetic motoneurones may contribute to the baroreceptor input to sympathetic motoneurons may contribute to the effect. In rats there seem to be two subsets of efferent cardiovascular neurones, projecting from the rostral ventrolateral medulla to the intermediolateral cell column of the spinal cord, with mean conduction velocities of 0.64 and 3.83 m/s, respectively [3]. However, if there were two discrete reflex pathways between the high-pressure receptors and sympathetic vasoconstrictor neurons to muscle, one would expect two discrete subsets of latency values. Therefore, the varying reduction of reflex during different manoeuvres (see Table I) may imply that even more pathways exist. Alternatively, the variable reduction may be caused by complex modulation of the baroreceptor input by inflow from other receptors in different manoeuvres. Such as possibility may also explain the times course of reflex latency changes during the Valsalva manoeuvre.
Efferent conduction time Efferent conduction time depends on conduction velocity in pre- and postga.nglionic fibres and ganglionic transmission. Conduction velocity of single fibres should be stable. The possibility that faster fibres, usually not in use, should be activated is remote, since an increased time dispersion of bursts with disturbance of pulse synchrony should follow, and this was not the case. Variation of transmission time due to facilitation in the sympathetic ganglion is not likely to cause any major contribution to the shortening of latency.
Central pharmacological actions Central pharmacological actions must be considered for the increase of reflex latency that occurred when intrinsic heart rate was induced by pharmacological denervation of the heart. The effects of the combined administration of a nonselective fl-blocking agent and atropine on central
sympathetic connections is unclear. Interestingly, however, the reduction of reflex latency was still evoked when a Valsalva manoeuvre was performed at intrinsic heart rate.
Significance of observed effects The functional importance of the reduced reflex latency is difficult to evaluate. From a teleological point of view one might assume that faster pathways would be made use of in emer, gency situations. The Valsalva manoeuvre and diving would fit such a reasoning, but apnoea causing hypoxia and the effect of sodium nitroprusside might be regarded as equivalent emergencies and they did not reduce reflex latency. The appearance of a shorter than normal sympathetic reflex latency is of neurophysiological interest, since it seems to verify in humans the existence of more than one central pathway, or variable central processing time, between baxoreceptors of the high-pressure type and sympathetic vasoconstrictor neurons to muscle vessels.
Acknowledgements Supported by the Swedish Medical Research Council, Grants no. B86-04X-7468,01A and B8614X-3546-15B, and by the Swedish Society of Medical Sciences. The authors thank Dr. Finn Mannting, who assisted in measurements of arterial pulse wave propagation velocity and Mrs. Catarina F~irnstrand, who spent many hours measuring reflex latencies on the digitizing board.
References 1 Borst. C. and Karemaker, J.M., Time delays in the human barorcceptor reflex. J. Auton. Nerv. Syst., 9 (1983) 399-409. 2 Brown, E., Goci, J.S., Greenfield, A.D.M. and Plassaras, G.C.. Circulatory responses to simulated gravitational shifts of blood in man induced by exposure of the bodybelow-the iliac crest to subatmospherie pressure. J. Physiol. (London), 183 (1966) 607-627. 3 Brown, D.L. and Guyenet, P.G., Electrophysiological study of cardiovascular neurons in the rostral ventrolateral medulla in rats, Circ. Res., 56 (t985) 359-369. 4 Delius, W., Hagbarth, K.-E., Hongdl, A. and Wallin, B.G.. Genvral c h a t ~ c s of sympathetic acti~ty in human muscle nerves, Acta Physiol. Scand. 84 (t97'2) 65L81.
165 5 Delius, W., Hagbarth, K.-E., Hongell, A. and Wallin, B.G., Manoeuvres affecting sympathetic outflow in human muscle nerves, Acta Physiol. Scand., 84 (1972) 82-94. 6 Eckberg, D.L., Nerhed, C. and Wallin, B.G., Respiratory modulation of muscle sympathetic and vagal cardiac outflow in man, J. Physiol. (London), 365 (1985) 181-196. 7 Eckberg, D.L., Rea, R.F., Andersson, O.K., Hedner, T., Pernow, J., Lundberg, J. and Wallin, B.G., Baroreflex modulation of sympathetic activity and sympathetic neurotransmitters in man, submitted. 8 Fagius, J. and Wallin, B.G., Sympathetic reflex latencies and conduction velocities in normal man, J. Neurol. Sci., 47 (1980) 433-448. 9 Fagius, J., Muscle nerve sympathetic activity in migraine. Lack of abnormality, Cephalalgia, 5 (1985) 197-203. 10 Fagius, J., Wallin, B.G., Sundli3f, G., Nerhed, L. and Englesson, S., Sympathetic outflow in man after anaesthesia of the glossopharyngeal and vagus nerves, Brain, 108 (1985) 423-438. 11 Fagius, J. and SundliSf, G., The diving response in man: effects on sympathetic activity in muscle and skin nerve fascicles, J. Physiol. (London), 377 (1986) 429-443. 12 Gabriel, M. and Seller, H., Interaction of baroreceptor
13
14
15
16
17
18
afferents from carotid sinus and aorta at the nucleus tractus solitarii, Pfliigers Arch., 318 (1970) 7-20. Mark, A.L., Victor, R.G., Nerhed, C. and Wallin, B.G., Microneurographic studies of the mechanisms of sympathetic nerve responses to static exercise in humans, Circ. Res., 57 (1985) 461-469. McAllen, R.M. and Spyer, K.M., The baroreceptor input to cardiac vagal motoneurones, J. Physiol. (London), 282 (1978) 365-374. McCloskey, D.I. and Potter, E.K., Excitation and inhibition of cardiac vagal motoneurones by electrical stimulation of the carotid sinus nerve, J. Physiol. (London), 316 (1981) 163-175. SundliSf, G. and Wallin, B.G., Effect of lower body negative pressue on human muscle nerve sympathetic activity, J. Physiol. (London), 278 (1978) 525-532. Wallin, B.G. and Fagius, J., The sympathetic nervous system in man. Aspects derived from microelectrode recordings, Trends Neurosci., 9 (1986) 63-67. Victor, R.G., Leimbach, W.N., Seals, D.R., Wallin, B.G. and Mark, A.L., Effects of the cold pressor test on muscle sympathetic nerve activity, Hypertension, 9 (1987) 429-436.