*From the Department of Kinesiology, University of Waterloo,

Waterloo, Ontario, Canada (Drs. Yamamoto and Hughson), andDepartment of Sports Science, Waseda University, Tokorozawa,Saitama, Japan (Dr. Nakamura).

206S Autonomic Nervous System Response to Exercise (Yamamoto, Hug/non, Nakamura)

Autonomic Nervous System Responses to Exercise inRelation to Ventilatory Threshold*Yoshiharu Yamanwto, PhD; Richard L. Hughson, Ph.D.; and

Yoshio Nakamura, Ph.D.

We introduce our recent approach to study autonomic

nervous system control of heart rate during exercise by

means ofheart rate variability (HRV) spectral analysis with

special reference to its relationship to ventilatory threshold

(Tvent). The rationale for the study was that HRV has been

shown to reflect (cardiac) parasympathetic and sympathetic

nervous system (PNS and SNS, respectively) activity, to-

gether with the underlying complexity of cerebral auto-

nomic system in terms of fractal dimension (DF) of HRV

time series. The experimental results showed that PNS was

markedly reduced below Tvent, that the rate of change in

sympathoadrenal activity indicators (plasma norepineph-

rine and epinephrine concentrations and SNS indicator)

was enhanced above Tvent, and that these changes in PNS

and SNS indicators were associated with the appearance of

the low-dimensional (low DF) dynamics that might reflect

less complex autonomic activity. These findings have been

considered with respect to implication for clinical cardiol-

ogy.

D uring incremental exercise, the minute ventilation

(VE) begins to increase nonlinearly with respect to

oxygen uptake (�o�) or work rate (WR) at a moderate

exercise intensity, namely the ventilatory threshold (Tvent).

This al)rupt increase in VE was considered to be due to the

abrupt onset on lactacidosis and the resultant activation of

blood bicarbonate buffering system.’2

Since these early reports, many studies have cast doubt

on the specific hypothesisl2 that cellular dysoxia, leading to

metabolic acidosis, resulted in the occurrence ofTvent. The

evidence has included the observations that blood lactate

concentration (LA) rose significantly before Tvent,3’4 that

lactacidosis induced by preliminary strenuous exercise failed

to alter the pattern of ventilatory responses during subse-

quent incremental exercise,56 that an alteration in blood

acid-base status could not induce the concomitant change in

Tvent,7 and that epinephrine infusion could change both LA

and Tvent.8

In the review of Dempsey et al,� the 2 more direct

counterevidences were provided. One was the report of

Hagberg et al’#{176}in which they demonstrated a Tvent in

patients with McArdle’s disease despite their inability to

develop lactacidemia during exercise because of a lack of

the enzyme, myophosphorylase. On the contrary, Dempsey

et al’ reported that highly trained athletes could not

hyperventilate in heavy exercise despite marked acidemia.

This evidence forced the original authors of the anaerobic

threshold (AT) concept to modify the significance of Tvent

and to incline more to gas exchange � That is,

they recently emphasized that an increase in carbon dioxide

output (Vco,) was always commensurate with an increase in

H � due to lactacidosis while Tvent was not always accom-

panied by this stimulus.12 However, such compromise may

raise definite problems both methodologically and physio-

logically.

Methodologically speaking, it can be said that most

noninvasive determinations of AT have been performed

using parameters relating to VE response to exercise, not

necessarily by what is called the “V-slope” method.’� Then

the question arises; should we reinterpret all ofthe clinically

presented results because these were not examined by the

latter method? The answer to this question given by these

authorsl2 was that AT determined by VE responses could be

validated only for the subjects whose ventilatory control

mechanisms responded “appropriately.” However, consid-

ering that they regarded the appropriate responses as such

that YE tracked the increase in Vco2, this is circular logic.

To answer this question, therefore, a study on the physiologic

mechanisms responsible for the onset of hyperventilation

per se at Tvent is necessary.

Several factors other than lactacidosis were shown to

affect YE during exercise at a moderate to high intensity.

These included body temperature, catecholamines,�”� and

more recently arterial concentration of potassium.’5 All of

these factors are listed in the recent review of Cunningham

et al18 to modulate feedback control of ventilation, and they

are thought to be related to the sympathoadrenal ty.’7

In this article, we introduce our recent approachesl321 to

autonomic nervous system control during exercise by means

of heart rate variability (HRV; heartbeat intervals on beat-

to-beat basis) spectral analysis with special reference to its

relationship to Tvent. As the method was not for investigating

the respiratory controller directly, the results ofthese studies

could not delineate the detailed physiologic mechanisms

responsible for Tvent. However, there was at least one

potential implication of the results for clinical cardiology.

This has been discussed in the last section.

HRV PARAMETERS TO EVALUATE AUTONOMIC

CONTROL OF HEART

The study of HRV has evolved with time series spectral

analysis. In the frequency domain, human HRV spectra have

displayed 2 or more major harmonic components.” One is

at frequency >0. 15 Hz, mediated solely by changing levels

of parasympathetic nervous system (PNS) activity.”” The

other components are usually seen at or below 0. 1 Hz and

are coherent with blood pressure variability.” It has been

demonstrated that the latter components at frequencies

<0. 15 Hz might be associated with both sympathetic

nervous system (SNS) and PNS activities.un

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21646/ on 06/25/2017

CHEST I 101 I 5 I MAY, 1992 I Supplement 2075

There are components in the HRV signal other than those

taken to indicate PNS and SNS activities. HRV also consists

of a nonharmonic component characterized by its broad-

band spectrom.�’ This broad-band noise has been suggested

to be “fractal” in nature , ‘�‘#{176}� and can be characterized by

the fractal dimension (DF) which in some sense is indicative

of the complexity of a given time series.” We have recently

developed a technique, coarse graining spectral analysis

(CGSA), that permits simultaneous evaluation of the har-

monic components related to PNS and SNS activities,’8 as

well as computation of DF from nonharmonic components.2#{176}

The rationale to use this method has been detailed by

Yamamoto and Hughson. “ Briefly, the method was based on

the fact that the fractal nonharmonic component retains its

power after it is cross-correlated with its rescaled version,

whereas, the harmonic component does not. A standard

FFT method was used to estimate auto-power spectra of

the original HRV time series (Sxx) and cross-power spectra

between the original and rescaled (coarse grained) time

series (Sxx’). Sxx was shown to represent harmonic plus

nonharmonic (fractal) components, whereas, Sxx’ repre-

sented the nonharmonic (fractal) 18 Thus, we

subtracted Sxx’ form Sxx to obtain the harmonic component.

From the harmonic component, the integrated power in

0.0 to 0. 15 Hz (P,) and 0. 15 to 1.0 Hz (Ph) was calculated.

PNS and SNS activities were evaluated by high frequency

power (Ph) and ratio of low to high frequency power (P/Ph),

respectively.182�� PNS indicator was normalized by total

HRV power with consideration for the effect of change in

magnitude of respiratory sinus arrhythmia.�

The nonharmonic (fractal) component in the HRV spec-

trum has been reported to have power-law characteristics,

as expressed by ft in log frequency (j) vs log power plots

where 13 is the spectral exponent.a� Therefore, the fractal

component was plotted in a log frequency vs log power

plane with �3 estimated as the slope of the linear regression

of this plot. From the value of 3, DF was calculated as

DF=2J(�3-1) for 1 <�3�3. For 3>3 and 0��3�1, DF was

taken as 1 and infinity (oo), respectively.

The value of DF or �3 can have the following physiologic

meanings. The brain consists of a number of (nonlinear)

oscillators and the interactions among them generate very

complex dynamics.�”� DF is thought to be the physical

reflection ofthe number ofthese oscillators.”’� For example,

the dimensionality of human cerebral activity, evaluated by

correlation dimension (the lower limit of DF) of an electro-

encephalographic (EEG) recording, has been reported to

be 8 to -0o in the awake, eye-opened, but quiet state.�’� A

high-dimensional system is also called “information rich:’

and this makes redundant processing of external stimuli

possible. Considering the fact that there are projections that

could transmit the cerebral activity to the autonomic cen-

ters,” it is not surprising to see high DF (�3 close to unity) in

the resting human HRV�#{176}�” with high level ofPNS activity.’#{176}

Babloyantz and Destexhe,� in the first report for the

existence oflow dimensional dynamics in the field of human

physiology, demonstrated that the correlation dimension of

EEG during epileptic seizure had a low mean value of 2.05.

As to the implication of this finding, they speculated that

the agent producing the seizure tended to drive the brain

activity toward a stable periodic motion. In such states,

information processing would be impossible and recovery

would be extremely diffIcult. However, to process reflex

activities, the brain remains on a chaotic attractor, albeit one

of very low � In an analogous way, we have

recently shown that the increase ofdefensive SNS responses

to orthostatic challenges�#{176} and exercise” was associated with

DF of HRV close to 2.0 (13-�- 1 .0), suggesting that DF of HRV

was indicative of a functional state of the autonomic centers.

The reduction in DF has also been observed in various

physiologic and pathologic situations, including susceptibil-

ity to syncope induced by orthostatic challenges,’#{176}�’ aging,3’

psychologic stress,’8 and electrical instability of the

heart.�”'”�

AUTONOMIC RESPONSE AND VENTILATORY

THRESHOLD

In an earlier study,’� we observed the HRV responses to

constant load exercise with intensity set relative to Tvent.

Eight subjects completed six 17-mm submaximal exercise

tests and one resting measurement in the sitting position.

During submaximal tests, WR was increased for the initial

3 mm in a ramp fashion until it reached constant WRs of 20

W, or 30%, 60%, 90%, 100%, and 110% ofthe predetermined

Tvent while ventilatory profile, alveolar gas exchange, and

HRV were measured continuously. From the HRV data, PNS

and SNS indicators were calculated.

The results showed the PNS indicator to decrease dra-

matically compared with rest when the subjects exercised.

It continued to decrease until the intensity reached 60%

Tvent. The SNS indicator, on the contrary, was statistically

unchanged up to 100% Tvent, while it increased significantly

at 110% Tvent.

We have recently studied these autonomic nervous system

responses to exercise for more detail,” by adopting a ramp

work protocol and by observing catecholamines and DF (�3)

responses.

Briefly, 6 healthy male volunteers performed incremental

exercise test on an electrically braked cycle ergometer,

consisting of a 5-mm warm-up period at 50 W, followed by

WR increment in a ramp fashion until exhaustion. The rate

of increase in WR was 2.0 Wmin’. This very slow ramp

protocol allowed us to observe quasi-steady-state responses

of HRV analyzed for successive 10-mm periods of data. The

steady-state nature of heart rate response was particularly

important for HRV spectral analysis. Arterialized venous

blood was sampled every 5 mm from an indwelling catheter

in a superficial vein on the heated dorsum of the hand for

the analyses of LA, plasma norepinephrine (NE), and

epinephrine (E) concentrations.

Overall responses of the physiologic variables are shown

in Figures 1 through 3. As was suggested by Wasserman

et al,” the rate of change in LA during the very slow ramp

exercise was so small that �Tco, increased almost linearly in

relation to exercise intensity (Fig 1). Thus, the determination

ofTvent ofY-slope method’s was impossible for all subjects.

Nevertheless, the profile of end-tidal 0, (PE’rO,) and CO,

(PE’ICO,, Fig 1) pressures showed clear evidence of the

onset of hyperventilation (Tvent) at moderate exercise

intensity. Such a reduction in PE’rCO2 at Tvent is to be

expected for these very slow ramps because so-called

“isocapnic buffering” is an anomaly of the specific 15

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21646/ on 06/25/2017

C0

CC)

>

80

60

40

20

--�1’

�- -

a)

0

0

(ng-mL1)

� 0.2

�-

�a.w

0

0

80

0.�

40.�

C/)zCl)

0

3rCC)C0aw

C)a

Cl)

21-

-40 -20 0 20 40

(L- mm) (mmol’ L’)6i

2085 Autonomic Nervous System Response to Exercise (Yamamoto, Hughson, Nakamura)

(Torr)

C�,l

.�40

w

-40 -20 0 20 40

delta Work Rate from T� (W)

(L’ mm’)2.0

a-I

0c%J00

1.0

FIGURE 1. VE, VCo2, and PE�rCO, responses during very slow (2.0Wmin’) ramp exercise. All variables shown in Figures 1 through 3are ensemble averages for 6 subjects in relation to ventilatory

threshold (Tvent) determined by the PETO, and PEi’CO, profiles,

and plotted against work rate (WR) relative to WR at Tvent. See

text for abbreviations.

Wmin’ protocol.� In Figures 1 through 3, all variables have

been shown as ensemble averages for 6 subjects in relation

to the Tvent determined by the PETO, and PEiCO, profiles.

The statistical comparison of regression slopes below and

above Tvent showed that only PEi’CO, in Figure 1 changed

systematically at Tvent.

Figure 2 shows the responses of LA, NE, and E. Each of

LA, NE, and E increased gradually with an increase in

exercise intensity, with an apparently greater increase in

concentration as intensity exceeded Tvent. A threshold-like

response of the blood catecholamine concentrations was

suggested in the recent study of Mazzeo and Marshall.”

However, the recent study of Savard et al” suggested that

effiux of catecholamines from exercising leg muscles in-

creases progressively with greater exercise intensity.

The responses of HRV-related variables are summarized

in Figure 3. In agreement with the results of our earlier

study,” PNS indicator decreased as WR increased, reaching

very low values at Tvent, The SNS indicator increased

gradually over WR below Tvent. At WR greater than Tvent,

it increased more steeply with a significant (p<0.05) change

in the slopes. More interestingly, the mean value for �3

reached 2.04 (corresponding to DF of 1.92) at Tvent with

the higher values (lower DF) thereafter. A similar finding

was obtained for the heart rate responses to the graded

levels of orthostatic stress,” indicating that the fractal

component of HRV had low dimensional, less complex

dynamics.

In summary, our recent results concerning autonomic

nervous system responses during exercise in relation to

-40 -20 0 20 40

delta Work Rate from T� (W)

(ngmL’)2

0CI..

a0

1Ca

0z

FIGURE 2. LA, NE, and E responses during very slow ramp

exercise. See text for abbreviations.

0.1

i!1�(0Iz I

delta Work Rate from T� (W)

FIGURE 3. PNS indicator, SNS indicator, and the spectral exponent

(13)responses during very slow ramp exercise.

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21646/ on 06/25/2017

CHEST I 101 I 5 I MAY, 1992 I Supplement 2095

Tvent showed that PNS withdrew almost completely below

Tvent, that the rate of changes in sympathoadrenal activity

indicators was enhanced above Tvent, and that these changes

in PNS and SNS indicators were associated with the

appearance of the low-dimensional dynamics that might

reflect less complex autonomic activity.

DIsCussIoN

Seeking the mechanisms responsible for Tvent is consid-

ered to be important not only physiologically, but also

methodologically. Proper interpretation of the accumulated

results for estimation of the so-called “AT” by noninvasive

ventilatory responses to exercise is required. Within this

context, our results showed a possible link between auto-

nomic nervous system status and Tvent, although a deline-

ation of the central and the peripheral controller scheme for

the hyperventilation was beyond the scope of these studies.

In addition, these results could have at least one potential

implication for clinical cardiology.

The concept ofAT was originally developed by Wasserman

and McIlroy� for evaluating cardiac performance of cardiac

patients. Later, several investigators supported this concept

indirectly by showing that the patients with the lower

maximal cardiac performance” and the higher NYHA clas-

sifications offunction”�� had the lower Tvent. However, this

does not necessarily mean that the lower Tvent in these

patients has been caused by their impaired cardiac perform-

ances via the mechanisms originally proposed, ie, lactaci-

demia due to decreased oxygen transport.’�” Indeed, Coyle

et al” reported that well-trained ischemic heart disease

patients could show similar LA responses during graded

exercise as well as similar endurance capacity to age-

matched trained runners despite the reduced Vo2max and

maximal cardiac output of the patients. This observation

suggested oxidative capacity of the skeletal muscle as a

possible determinant of LA response during exercise.”

Indeed, the recent review by Connett et aF’#{176}has convincingly

shown that lactate effiux and muscle redox potential are

interactive functions of muscle enzyme concentrations and

substrate availability. Neither changes in lactate efilux, nor

in redox indicators such as lactate to pyruvate ratio, as

suggested by Wasserman et al,” can indicate intracellular

dysoxia.�#{176} Thus, if one examined the relationship between

lactacidemia and Tvent, the lower Tvent in cardiac patients

might merely reflect their physical inactivity due to the

disease per se, not the impaired cardiac function.

It should be emphasized that the spectral analysis of HRV

evaluated the activity of cardiac autonomic nerves innervat-

ing the sinoatrial node.” The cardiac parasympathetic tone

affects electrical stability ofthe heart. Electrical stimulation

of the vagus nerves has been shown to increase ventricular

fibrillation threshold and decrease the incidence of sponta-

neous ventricular fibrillation during myocardial ischemia,�’

while bilateral vagotomy or atropine has been shown to

increase arrvthmia formation.42 Recently, Biliman and Hos-

kins,�’ using HRV spectral analysis in dogs with healed

anterior myocardial infarctions, reported that the dogs

susceptible to ventricular fibrillation by a coronary occlusion

during submaximal exercise had lower PNS indicators than

the resistant.

On the other hand, sympathetic neural activity also

increases the ventricular vulnerability to fibrillation” In

anesthetized dogs, stellate ganglion stimulation decreased

the ventricular fibrillation threshold by about 60%.� The

same type of response was also observed for extrasystole

threshold�” It is of note that this effect has been shown to

be evident especially without PNS activity.e� In our results,

Tvent was accompanied by the enhancement of SNS with

the simultaneous withdrawal of PNS almost completely.

Taken together, these data suggested that changes in cardiac

autonomic profile at Tvent might increase susceptibility to

a cardiac electrical instability.

Furthermore, Skinner et al” recently published their

preliminary observations on the dimensionality of HRV and

arrhythmogenesis in a cardiac patient wearing a Holter

monitor at the time ofhis death. They stated the correlation

dimension decreased to an integer dimension approximating

1.0 prior to the onset of lethal arrhythmogenesis and that a

fractal process in the heartbeat generator provided protec-

tion from arrythmogenesis. Considering that DF (�3) of HRV

decreased (increased) gradually to 1 .0 (to 3.0) above Tvent

in our study, indicating loss of complexity, the exercise

intensity above Tvent could lead to increased risk of cardiac

vulnerability. The relationship between the change in cardiac

autonomic nervous system status and a possible cardiac

vulnerability during moderate exercise remains to be inves-

tigated clinically as wellas experimentally in future research.

ACKNOWLEDGMENT: This study was supported in part bygrants from the Ministry ofEducation, Japan (02951196), from MeijiLife Foundation of Health and Welfare, Japan, and from the NaturalSciences and Engineering Research Council ofCanada. The authorsare grateful to Isao Muraoka, Professor of Waseda University, andMinort, Shinohara, M.Sc. , at University ofTokyo for their help inconducting part of the experiments in this study.

REFERENCES

1 Wasserman K, Whipp BJ, Koyal SN, Beaver WL. Anaerobic

threshold and respiratory gas exchange during exercise. J ApplPhysiol 1973; 35:236-43

2 Wasserman K, Whipp BJ, Koyal SN, Cleary MG. Effect ofcarotid body resection on ventilatory and acid-base controlduring exercise, J AppI Physiol 1975; 39:354-58

3 Green HJ, Hughson RL, Orr GW, Ranney DA. Anaerobicthreshold, blood lactate, and muscle metabolites in progressive

exercise. J Appl Physiol 1983; 54:1032-384 Hughson RL. Methodologies for measurement of the anaerobic

threshold. Physiologist 1984; 27:304-11

5 Davis HA, Gass GC. The anaerobic threshold as determined

before and during lactic acidosis. Eur J Appl Physiol 1981;

47: 141-49

6 Farrell SW, Ivy JL. Lactate acidosis and the increase of �1s/�o,

during incremental exercise. J Appl Physiol 1987; 62:1551-55

7 Kowalchuk JM, Heigenhauser GJF, Jones NL. Effect ofpH on

metabolic and cardiorespiratory responses during progressive

exercise. J Appl Physiol 1984; 57:1558-638 Scheen A, Lemaire P Abaissement du seuil d’hyperventilation

per perfusion intraveineuse d’adrenaline lors d’un exercise

triangulaire. Arch Int Physiol Biochem 1983; 91:187-96

9 Dempsey JA, Vidruk EH, Mitchell GS. Pulmonary control

systems in exercise: update. Fed Proc 1985; 44:2260-70

10 Hagberg JM, Coyle EF, Carroll JE, Miller JM, Martin WH,

Brooke MH. Exercise hyperventilation in patients with Mc-Ardle’s disease. J Appl Physiol 1982; 52:991-94

11 Dempsey JA, Hanson P. Henderson K. Exercise-induced arterialhypoxemia in healthy humans at sea-level, J Physiol 1984;

355:161-74

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21646/ on 06/25/2017

2105 AUtOnOmiC Nervous System Response to Exercise (Yamamoto, Hughson, Nakamura)

12 Wasserman K, Beaver WL, Whipp BJ. Gas exchange theory and

the lactic acidosis (anaerobic) threshold. Circulation 1990;

81:1114-30

13 Beaver WL, Wasserman K, Whipp BJ. A new method for

detecting the anaerobic threshold by gas exchange. J AppIPhysiol 1986; 60:2020-27

14 Whipp BJ. Ventilatory control during exercise in humans. AnnRev Physiol 1983; 48:393-413

15 Paterson DJ, Friedland JS, Bascom DA, Clement ID, Cunning-ham DA, Painter R, et al. Changes in arterial K + and ventilation

during exercise in normal subjects and subjects with McArdle’s

syndrome. J Physiol 1990; 429:339-48

16 Cunningham DJC, Robbins PA, Wolff CB. Integration of respi-ratory response to changes in alveolar partial pressures of CO,

and 0, and in arterial pH. In: Handbook of physiology: therespiratory system: control of breathing, section 3, chap 15.

Bethesda,Md: American Physiological Society, 1986; 2:475-528

17 Christensen NJ, Ga1bO H. Sympathetic nervous activity during

exercise. Ann Rev Physiol 1983; 45:139-4518 Yamamoto Y, Hughson RL. Coarse graining spectral analysis:

new method for studying heart rate variability. J Appl Physiol

1991; 71:1143-50

19 Yamamoto Y, Hughson EL, Peterson JC. Autonomic control of

heart rate during exercise studied by heart rate variability

spectral analysis. J Appl Physiol 1991; 71:1136-4220 Yamamoto Y, Nakamura Y, Bufler GC, Hughson RL. Fractal

dimension of heart rate variability and physiological stress. AmJ Physiol, submitted for publication

21 Nakamura Y, Yamamoto Y. Autonomic control of heart rate

during physical exercise and fractal dimension of heart rate

variability. J Appl Physiol, submitted for publication

22 Saul JR Beat-to-beat variations of heart rate reflect modulation

ofcardiac autonomic outflow. NIPS 1990; 5:32-723 Pagani M, Lombardi F, Guzzetto 5, Rimoldi 0, Furlan R,

Pizzinelli R, et al. Power spectral analysis of heart rate andarterial pressure variabilities as a marker of sympatho-vagalinteraction in man and conscious dog. Circ Res 1986; 59:178-93

24 Goldberger AL. Fractal electrodynamics of the heartbeat. Ann

NY Acad Sci 1990; 591:402-09

25 Mayer-Kress G, Yates FE, Benton L, Keidel M, T,rsch W, Poppl

SJ, et al. Dimensional analysis of nonlinear oscillations in brain,heart, and muscle. Math Biosci 1988; 90:155-82

26 Mayer-Kress G, Layne SR Dimensionality ofhuman electroen-

cephalogram. Ann NY Acad Sci 1987; 504:62-87

27 Babloyantz A, Destexhe A. Low-dimensional chaos in an instance

ofepilepsy. Proc Natl Acad Sci USA 1986; 83:3513-1728 Skinner JE, Goldberg AL, Mayer-Kress G, Ideker RE. Chaos

in the heart: implications for clinical cardiology Biotechnology

1990; 8:1018-24

29 Saul JP, Albrecht P. Berger RD, Cohen RJ. Analysis oflong term

heart rate variability: methods, 1ff scaling and implications.Comp Cardiol 1988; 14:419-22

30 Butler GC, Yamamoto Y, Xing HC, Northey DR, Hughson RL.

Heart rate variability and fractal dimension during orthostatic

challenges. Am J Physiol, submitted for publication31 Lipsitz LA, Mietus J, Moody GB, Goldberger AL. Spectral

characteristics ofheart rate variability before and during postural

tilt: relations to aging and risk of syncope. Circulation 1990;

81:1803-10

32 Mazzeo RS, Marshall P. Influence of plasma catecholamines onthe lactate threshold during graded exercise. J Appl Physiol

1989; 67:1319-22

33 Savard GK, Richter EA, Strange 5, Kiens B, Christensen NJ,

Saltin B. Norepinephrine spillover from skeletal muscle duringexercise in humans: role of muscle mass. Am J Physiol 1989;

257:H1812-18

34 Wasserman K, Mcllroy MB. Detecting the threshold of anaer-

obic metabolism in cardiac patients during exercise. Am J

Cardiol 1964; 14:844-52

35 Weber ICT, Kinasewitz GT, Janicki JS, Fishman AP Oxygen

utilization and ventilation during exercise in patients with

chronic cardiac failure. Circulation 1982; 65: 1213-23

36 Matsumura N, Nishijima H, Kojima 5, Hashimoto F, Minami

M , Yasuda H . Determination of anaerobic threshold for assess-

ment of functional state in patients with chronic heart failure.

Circulation 1983; 68:360-67

37 Lipkin DP, PerrinsJ, Poole-Wilson PA. Respiratory gas exchange

in the assessment ofpatients with impaired ventricular function.

Br Heart J 1985; 54:321-28

38 Coyle EF, Martin WH, Ehsani AA, Hagberg JM, Bboomfield

SA, Sinacore DR. et al. Blood lactate threshold in some well-trained ischemic heart disease patients. J Appl Physiol 1983;

54:18-2339 Ivy JL, Withers LIT, Van Handel PJ, Elger DH, Costill DL.

Muscle respiratory capacity and fiber type as determinants of

the lactate threshold. J AppI Physiol 1980; 48:523-27

40 Connet RJ, Honig CR, Gayeski TEJ, Brooks GA. Defininghypoxia: a systems view of Vo,, glycolysis, energetics, and

intracellar Po,. J Appl Physiol 1990; 68:833-42

41 Kolman BS , Verrier RL, Lown B. The effect of the vagus nerve

stimulation upon vulnerability of canine ventricle: role of thesympathetic parasympathetic interactions. Circulation 1976;

52:578-85

42 Hess DS, Hanlon T, Scheinman M, Budge R, Desal J. Termi-

nation of ventricular tachycardia by carotid sinus message.Circulation 1982; 65:627-33

43 Billman GE, Hoskins RS. Time series analysis of heart rate

variability during submaximal exercise: evidence for reduced

cardiac vagal tone in animals susceptible to ventricular fibrilla-

tion. Circulation 1989; 80:146-57

44 Levy MN. Autonomic interactions in cardiac control. Ann NY

Acad Sci 1990; 601:209-21

45 Lown B, Verrier RL. Neural activity and ventricular fibrillation.N EngI J Med 1976; 294:1165-70

46 Furey SA III, Levy MN. The interactions among heart rate,

autonomic activity, and arterial pressure upon the multiple

extrasystole threshold in the dog. Am Heart J 1983; 106:1112-

20

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21646/ on 06/25/2017