AJP Heart 2008

Baroreflex responsiveness during ventilatory acclimatization in humans

Brian E. Hunt,1 Renaud Tamisier,2,3 Geoffrey S. Gilmartin,1 Mathew Curley,1 Amit Anand,1

and J. Woodrow Weiss1

1Department of Kinesiology, School of Public Health and Health Sciences, University of Massachusetts, Amherst,Massachusetts; 2Pulmonary and Sleep Research Laboratory, Division of Pulmonary, Critical Care and Sleep Medicine,Beth Israel Deaconess Medical Center, Boston, Massachusetts; and 3Laboratoire du Sommeil, Laboratoire HP2 (HypoxiePathoPhysiologie), Centre Hospitalier Universitaire de Grenoble, Grenoble, France

Submitted 7 February 2008; accepted in final form 25 August 2008

Hunt BE, Tamisier R, Gilmartin GS, Curley M, Anand A, WeissJW. Baroreflex responsiveness during ventilatory acclimatization inhumans. Am J Physiol Heart Circ Physiol 295: H1794–H1801, 2008.First published August 29, 2008; doi:10.1152/ajpheart.131.2008.—Wetested the hypothesis that the decline in muscle sympathetic activityduring and after 8 h of poikilocapnic hypoxia (Hx) was associated witha greater sympathetic baroreflex-mediated responsiveness. In 10 healthymen and women (n � 2), we measured beat-to-beat blood pressure(Portapres), carotid artery distension (ultrasonography), heart period, andmuscle sympathetic nerve activity (SNA; microneurography) during twobaroreflex perturbations using the modified Oxford technique before,during, and after 8 h of hypoxia (84% arterial oxygen saturation). Theintegrated baroreflex response [change of SNA (�SNA)/change of dia-stolic blood pressure (�DBP)], mechanical (�diastolic diameter/�DBP),and neural (�SNA/�diastolic diameter) components were esti-mated at each time point. Sympathetic baroreflex responsivenessdeclined throughout the hypoxic exposure and further declinedupon return to normoxia [pre-Hx, �8.3 � 1.2; 1-h Hx, �7.2 � 1.0;7-h Hx, �4.9 � 1.0; and post-Hx: �4.1 � 0.9 arbitrary integratedunits (AIU) �min�1 �mmHg�1; P � 0.05 vs. previous time point for1-h, 7-h, and post-Hx values]. This blunting of baroreflex-mediatedefferent outflow was not due to a change in the mechanicaltransduction of arterial pressure into barosensory stretch. Rather,the neural component declined in a similar pattern to that of theintegrated reflex response (pre-Hx, �2.70 � 0.53; 1-h Hx,�2.59 � 0.53; 7-h Hx, �1.60 � 0.34; and post-Hx, �1.34 � 0.27AIU �min�1 ��m�1; P � 0.05 vs. pre-Hx for 7-h and post-Hx values).Thus it does not appear as if enhanced baroreflex function is primarilyresponsible for the reduced muscle SNA observed during intermediateduration hypoxia. However, the central transduction of baroreceptorafferent neural activity into efferent neural activity appears to be reducedduring the initial stages of peripheral chemoreceptor acclimatization.

autonomic

AUTONOMIC RESPONSES TO HYPOXIA vary greatly depending on theintensity, duration, and pattern of the stimulus (31). Transientexposure to hypoxia, as little as 20 min, results in an increasein minute ventilation that returns to prehypoxia levels withinseveral minutes after exposure, whereas vascular sympatheticactivity remains elevated for at least an hour after return tonormoxia (26, 48). Prolonged exposure to hypoxia (weeks tomonths) results in an elevated minute ventilation and musclesympathetic activity, and both are sustained for at least severaldays after return to normoxia (2, 6, 15). Although much of thework regarding autonomic regulation in humans has focusedon the effects of either transient or prolonged hypoxia, little isknown about the transitional period between short-term and

prolonged hypoxia. The few data available suggest that whenthe hypoxic exposure is extended to several hours, minuteventilation demonstrates a further increase that persists after areturn to normoxia, whereas sympathetic activity returns topreexposure levels (10, 31).

Transient hypoxia elicits a rapid increase in pulmonaryventilation via direct increases in afferent input from peripheralchemoreceptors. This is followed by a modest decline or “rolloff” in ventilation over the course of 20–30 min, likely causedby reduced sensitivity of arterial chemoreceptors to PO2 (1, 34).Although the mechanisms underlying the rapid and sustainedsympathoexcitation associated with transient hypoxia are notwell understood, the baroreflex regulation of vascular sympa-thetic activity appears well maintained and not a likely candi-date (12). Prolonged hypoxia, several hours to days, results ina time-dependent increase in minute ventilation, termed venti-latory acclimatization to hypoxia (VAH), which outlasts theexposure. Previous data suggest that VAH begins by 4 h ofhypoxic exposure (43). Increased sensitivity of both the arterialchemoreceptor and the central nervous system to PO2 arethought to underlie this chronic response (31). However, themechanisms underlying the return of sympathetic activity toprehypoxic levels during the initial stages of VAH are unclear.

As chemoreceptors are sympathoexcitatory, the dissociationbetween augmented peripheral chemoreceptor sensitivity andreduced sympathetic neural outflow during VAH arguesagainst a direct chemoreceptor-mediated response. However,given the close proximity of carotid chemoreceptor and sym-pathoinhibitory baroreceptor afferent projections into the me-dulla, it is plausible that the increased medullary activityassociated with VAH may augment baroreflex restraint ofmuscle sympathetic activity during this period. Therefore, wetested the hypothesis that sympathetic baroreflex-mediatedresponsiveness would be augmented during and after 8 h ofpoikilocapnic hypoxia.

METHODS

Study Sample

Twelve healthy, nonsmoking, normotensive volunteers, free of vaso-active medications, completed the study (women � 3). Adequate sym-pathetic neurograms were obtained in 10 volunteers. All study volunteersunderwent a routine medical history and physical examination to excludeovert cardiopulmonary or neurological disease before giving writteninformed consent. To eliminate possible confounding hormonal effectson cardiovascular function, the women were studied during the week

Address for reprint requests and other correspondence: B. E. Hunt, AppliedHealth Science, Wheaton College, 501 College Ave., Wheaton, IL 60187(e-mail: [email protected]).

The costs of publication of this article were defrayed in part by the paymentof page charges. The article must therefore be hereby marked “advertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Am J Physiol Heart Circ Physiol 295: H1794–H1801, 2008.First published August 29, 2008; doi:10.1152/ajpheart.131.2008.

0363-6135/08 $8.00 Copyright © 2008 the American Physiological Society http://www.ajpheart.orgH1794

on February 6, 2009

ajpheart.physiology.orgD

ownloaded from

http://ajpheart.physiology.org

following menses and all tested negative for pregnancy via humanchorionic gonadotropin test. This protocol conformed to the provisions ofthe Declaration of Helsinki and was thus approved by the InstitutionalReview Board at the Beth Israel Deaconess Medical Center.

Experimental Protocol

All studies were conducted between 9:00 AM and 7:00 PM after studyvolunteers had fasted overnight. After instrumentation, the volunteersrested in the supine position for 15–30 min until all parameters werestable for at least 5 min. During the following 45 min, continuousbaseline recordings were made of sympathetic activity, arterial pressure,and electrocardiogram (ECG) while the subjects breathed room air. Inaddition, two consecutive baroreflex perturbations using the modifiedOxford technique were performed (8). We then exposed the volunteers tocontinuous poikilocapnic hypoxia (Hx). The volunteers breathed througha leak-free face mask (8940 Series; Hans Rudolph, Kansas City, MO)fitted with a Hans Rudolph 2600 series two-way valve. A three-way valveon the inspiratory tubing allows rapid switching between the 30-literrespiratory bag and room air. The respiratory bag was filled with a gasmixture containing 9% oxygen balanced with nitrogen. Nitrogen or roomair was added to the inspired gas as necessary to maintain arterial oxygensaturation (SaO2) near 85%. During the exposure, the measurements ofblood pressure, ECG, and two baroreflex perturbations were performedafter 1 and 7 h. During the hypoxic exposure study, the volunteers wereallowed to drink water ad libitum and were continuously infused with 5%glucose at a rate of 100 ml/h. After 8 h, the volunteers were returned toroom-air breathing and all measurements were repeated after SaO2 re-turned to baseline, with baroreflex testing beginning at 30 min ofposthypoxia.

Measurements

Cardiovascular and pulmonary variables. Heart rate was derivedfrom the ECG. Clinical arterial pressure was measured in the right armusing an automated arm-cuff sphygmomanometer (Dinamap model,Critikon, Tampa, FL). Continuous beat-by-beat blood pressure measure-ments by photoplethysmography (Portapres TNO-Institute of AppliedPhysics Biomedical Instrumentation, Amsterdam, The Netherlands) wereused in the calculation of baroreflex responsiveness. SaO2 was monitoredwith a pulse oximeter (Biox model 3740; Ohmeda, Louisville, CO), andCO2 fraction was measured continuously using an infrared gas analyzerconnected to the mask (model 17630; Vacumed, Ventura, CA). End-tidalCO2 (ETCO2) was calculated as the product of the Fe CO2 value at the endof the expiratory plateau and barometric pressure. Data from previoussubjects exposed to the same 8 h of poikilocapnic hypoxia (10) demon-strated that ETCO2, fraction of inspired oxygen (FIO2), and SaO2 were notdifferent between the 1- and 7-h time points, suggesting the SaO2 is anadequate index of the hypoxic stress in our healthy study volunteers. Thechange in ETCO2 across time points within an individual was used as agross index of peripheral chemoreflex drive.

Muscle sympathetic nerve activity. We obtained nerve recordingsusing standard tungsten microelectrodes inserted into the peronealnerve posterior in the popliteal area after localization by surfacestimulation. The signals were filtered, amplified, and full-wave recti-fied. The rectified signal was integrated for display on an oscilloscopeand for recording (Nerve Traffic Analyzer, Model 662c-3, Bioengi-neering Dept., University of Iowa, Iowa City, IA). The electrodeposition in the muscle fibers was confirmed by pulse synchronousbursts of activity occurring 1.2–1.4 s after the QRS complex, repro-ducible activation during the second phase of the Valsalva maneuver,elicitation of afferent nerve activity by mild muscle stretching, and theabsence of response to startle. We recorded the sympathetic neuro-gram before hypoxia and continued the recordings for 1 h afterhypoxia was begun. The electrodes were then removed. The micro-electrodes were replaced, and the sympathetic neurogram was re-corded after 7 h of hypoxia and continued recording for 1 h aftervolunteers returned to breathing room air. Sympathetic bursts were

identified by the Hamner-Taylor detection algorithm (14) using Mat-lab software (Mathworks, Natick, MA). Burst amplitude was normal-ized by assigning a value of 1,000 arbitrary units to the largest burstidentified during stable baseline activity before each measurement;this was expressed as both frequency (in bursts/min) and rate (inbursts/100 heart beats). Both metrics are stable and reproducibleindexes and thus are used to compare sympathetic activity betweentime points. The total burst activity was then estimated by calculatingthe area under each sympathetic burst; this was expressed as arbitraryintegrated units (AIU) and used to express the change in sympatheticactivity during baroreflex perturbations. Since the Hamner-Tayloralgorithm has not been independently validated, we compared thesympathetic burst frequency as determined by the semiautomatedHamner-Taylor algorithm to manual burst detection in 14 neurograms(7 steady-state normoxia and 7 steady-state hypoxia). The burstfrequency determined by both methods was highly correlated (r2 �0.79) with the automated detection having a fixed bias of �1.5bursts/min.

Baroreflex engagement. We used the modified Oxford technique,which involves a bolus injection of 100 �g of sodium nitroprussidefollowed in 60 s by a bolus of 150 �g of phenylephrine hydrochloride.This technique generally produces an initial �15 mmHg fall inarterial pressure followed by an �15 mmHg rise in pressure aboveresting supine levels (8, 17). Responses of a representative studyvolunteer are shown in Fig. 1. During baroreflex engagement, thebeat-by-beat arterial pressures from a finger photoplethysmograph(Portapres), a standard three-lead ECG, and muscle sympathetic nerveactivity (MSNA) were recorded continuously at a sampling rate of500 Hz. Longitudinal B-mode images of a common carotid artery justbelow the carotid bulb were also obtained during baroreflex engage-ment using a 7.5-MHz transducer. Commercially available hardware(3155 PCI Frame Grabber, Data Translations, Marlboro, MA) andsoftware (DVI Acquisition, Information Integrity, Stow, MA) wereused to acquire 30-Hz images to a computer triggered from theR-wave of the ECG. Fifteen consecutive carotid images were acquired

Fig. 1. Blood pressure (BP), R-R interval, carotid diameter, and musclesympathetic nerve activity (MSNA) from a single study volunteer duringbaroreflex engagement before hypoxia.

H1795HYPOXIA AND SYMPATHETIC BAROREFLEX FUNCTION

AJP-Heart Circ Physiol • VOL 295 • OCTOBER 2008 • www.ajpheart.org

on February 6, 2009


ownloaded from


with each trigger to approximate at least one third of the cardiac cycle(i.e., 500 ms of a 1,500-ms R-R interval or 40 beats/min heart rate)and therefore encompass both end-diastolic and peak-systolic diam-eters. Each 640 � 480-pixel image provided a resolution of 0.05mm/pixel. We used custom software that employed a verterbi searchalgorithm to estimate carotid diameters. Two trials were performed ateach time point. Each trial was separated by at least 15 min to allowfor the recovery of blood pressure and heart rate to the preinjectionlevels.

For each trial, the diameters were associated to the appropriate bloodpressures, R-R intervals, and sympathetic activity (29, 37). The data werethen averaged across 2-mmHg increments or bins. We excluded bins inthe threshold and saturation regions in estimating the linear responsive-ness of the integrated, mechanical, and neural aspects of the baroreflex.Integrated cardiovagal baroreflex responsiveness was estimated from theslope of the relation between R-R interval and systolic blood pressure(�R-R interval/�systolic blood pressure). The mechanical component ofthe vagal limb was expressed as the slope of the relation between systolicblood pressure and carotid systolic diameter (�systolic blood pressure/�systolic diameter). The neural component of the vagal limb was ex-pressed as the slope of the relation between R-R interval and carotidsystolic diameter (�R-R interval/�systolic diameter). Each aspect of the

sympathetic limb was characterized in a similar fashion: integratedresponsiveness (�MSNA/�diastolic blood pressure), mechanical (�dia-stolic blood pressure/�diastolic diameter), and neural (�MSNA/�dia-stolic diameter). Figure 2 depicts how estimates of baroreflex responsive-ness were calculated from the data shown in Fig. 1.

Because central or peripheral hysteresis can affect the slope ofthese relations, we independently assessed the slopes during thefalling and rising portions of the response. No systematic differenceswere found between slopes during the falling and rising pressures.Therefore, the slopes were estimated from data encompassing the falland rise in blood pressure. Baroreflex responsiveness at each time point isexpressed as the mean of the two trials (average trial-to-trial difference of0.30 ms/mmHg for vagal and 0.20 AIU �burst�1 �mmHg�1 for sympa-thetic gains).

Data Analysis

All comparisons were limited to data sets from the 10 volunteers inwhich adequate pre- and postsympathetic neurograms were obtained.A one-way ANOVA for repeated measures was used to compare alldata across the four time points of the study. Newman-Keuls post hocanalysis was used when ANOVA indicated differences between time

Fig. 2. Vagal baroreflex responses (left) and sympatheticbaroreflex responses (right) derived from data depicted in Fig.1. Top: integrated response. Middle: mechanical component.Bottom: neural component of the integrated baroreflex re-sponse. SBP, systolic BP; DBP, diastolic BP.

H1796 HYPOXIA AND SYMPATHETIC BAROREFLEX FUNCTION


on February 6, 2009


ownloaded from


points. Trends were determined by orthogonal polynomials (11, 16).Data are reported as means � SE in the text, table, and figures.P values � 0.05 were considered statistically significant.

RESULTS

As depicted in Fig. 3, MSNA initially increased before whenVAH has been reported to begin (43), then returned to prehy-poxia levels at 7 h of exposure, and declining further atpostexposure. Along with sympathetic activity, Table 1 dis-plays the hypoxic stimulus as measured by SaO2

and theattendant changes in heart rate, ETCO2

, and mean arterialpressure throughout the study.

There was a linear trend for integrated vagal baroreflexresponsiveness to decline across time (P � 0.07), with thelowest value recorded after volunteers were returned to room-air breathing (P � 0.08; Fig. 4). There was no correspondingsystematic difference in either the mechanical component (pre-Hx, 2.43 � 2.23; 1-h Hx, 2.45 � 2.94; 7-h Hx, 1.83 � 2.43;and post-Hx, 2.06 � 1.22 �m/mmHg) or neural component(pre-Hx, 8.85 � 1.84; 1-h Hx, 7.38 � 1.58; 7-h Hx, 8.93 �1.92; post-Hx, 6.51 � 1.05 ms/�m) of the reflex arc.

Integrated sympathetic baroreflex responsiveness declinedduring the hypoxic exposure, achieving its lowest measuredvalue at 30 to 45 min posthypoxia (Fig. 4). There was nodifference in the mechanical component across time (pre-Hx,3.23 � 0.30; 1-h Hx, 3.28 � 0.30; 7-h Hx, 2.95 � 0.20; andpost-Hx, 2.97 � 0.40 �m/mmHg), as depicted in Fig. 5.However, hypoxia was associated with a decline in the neuralcomponent, with a further decline observed after returning to

Table 1. Cardiovascular and ventilatory variables

Prehypoxia 1-h Hypoxia 7-h Hypoxia Posthypoxia

Arterial O2 Saturation, % 98.8�0.8 83.9�3.8* 83.3�4.2* 98.5�0.8End-tidal CO2, mmHg 44.7�2.7 40.3�4.2* 40.9�3.6* 39.5�3.2*Heart Rate, beats/min 58�13 74�19* 78�18* 62�11MSNA, bursts/100 beats 26�10 32�10* 21�8* 16�6*MAP, mmHg 86�8 87�8 86�10 91�8

Values are means � SE. MSNA, muscle sympathetic nerve activity; MAP,mean arterial pressure measured in the brachial artery. *P � 0.05 vs. prehyp-oxia.

Fig. 3. MSNA and BP responses before andduring baroreflex from a single study volunteerat each time point. Steady-state MSNA (left)and integrated sympathetic baroreflex response(SBR; right) are shown. AU, arbitrary unit.



on February 6, 2009


ownloaded from


room-air breathing (pre-Hx, �2.70 � 0.53; 1-h Hx, �2.59 �0.53; 7-h Hx, �1.60 � 0.34; and post-Hx, �1.34 � 0.27AIU �beat�1 ��m�1; P � 0.05 vs. pre-Hx for 7-h and post-Hxvalues).

Carotid diastolic diameter increased during the first hour ofhypoxia (pre-Hx, 6.00 � 0.15 vs. 6.32 � 0.23 mm, P � 0.05)and continued to be greater after 7 h of hypoxia (6.34 � 0.22mm). Upon the resumption of room-air breathing, the diameterwas not different from prehypoxia (5.93 � 0.17 mm).

Although there was no relation between sympathetic barore-flex responsiveness and sympathetic activity before or after just1 h of hypoxia, the greater sympathetic baroreflex responsive-ness was associated with lower sympathetic bursting rateacross individuals at 7 h of hypoxia (r � �0.89, P � 0.05) andafter returning to room-air breathing (r � �0.79, P � 0.05).However, within individuals, these variables were not consis-tently related across time points, with correlation coefficientsranging from 0.92 to �0.99 (mean � �0.23 � 0.32). Further-more, the baroreflex responsiveness was not related to periph-eral chemoreflex drive, as suggested by ETCO2

, with intra-individual correlation coefficients ranging from 0.26 to �0.57,with all P values 0.23.

DISCUSSION

We hypothesized that the decline in sympathetic activityseen after 8 h of poikilocapnic hypoxia would be associatedwith increased sympathetic baroreflex responsiveness. How-ever, our data fail to support this hypothesis. Rather, theyindicate that baroreflex responsiveness declined during hyp-oxia and was at its lowest after returning to room-air breathing.This reduction does not appear to be related to changes in themechanical transduction of arterial pressure into the barorecep-tor stretch within the carotid artery. Rather, altered neuralproperties within the reflex arc may underlie this unexpectedreduction in responsiveness. We are aware of no previous dataproviding insight into changes in the baroreflex function duringshort-term VAH.

Most data regarding the effects of hypoxia muscle sympa-thetic activity and its regulation are limited to either short-termexposures lasting no more than 30 min or long-term exposureslasting weeks or months. During short-term hypoxia, vascularsympathetic activity is elevated 20–126%, whereas baroreflex

responsiveness is unchanged or slightly augmented (12, 13,48). During long-term exposure to hypoxia, vascular sympa-thetic activity is elevated 200–300% (6, 15). Little is knownregarding baroreflex function after long-term hypoxia. Basedon these data, it is reasonable to assume that the dose-responserelation between hypoxia and MSNA follows a simple thresh-old model. However, the current data, as well as previous datafrom our laboratory, show that MSNA and sympathetic barore-flex responsiveness decline to prehypoxia levels during theinitial stages of VAH. Thus, taking all data together, it wouldappear that the hypoxia and MSNA relation displays hormesis(23). Although it is difficult to explain this finding, it must beconsidered that profound physiological adjustments are occur-ring during this period of time. The sensitivity of arterialchemoreceptors to PO2 initially declines and then is augmentedalong with the sensitivity of the central nervous system (31).Concurrently, there are changes in circulating factors thateffect sympathetic activity, such as a decline in angiotensin II(39, 41, 49).

In humans, both cardiovagal and sympathetic baroreflexresponsiveness are largely unaffected by transient (�25 min)hypoxia (4, 5, 7, 12, 13). However, when the hypoxic stimulusis prolonged to an hour, cardiovagal baroreflex responsivenessdeclines (35). Our data demonstrate that when the hypoxicstimulus is prolonged to 8 h, allowing for short-term VAH,then not only does cardiovagal baroreflex tend to decline, butthe responsiveness of the sympathetic limb declines as well.Similar interactions between autonomic reflex loops have beenreported, even during acute exposure. For example, Somersand coworkers (38) examined the effect of baroreceptor stim-ulation on peripheral chemoreflex responses during transienthypoxia. They show that chemoreceptor-mediated responsesare blunted during baroreflex activation by phenylephrine.

With the addition of ultrasound imaging during reflex per-turbation, we were able to examine two broad components ofthe reflex arc: 1) the mechanical, which is an estimate of theability of barosensory vessels to transduce arterial pressurechanges into vessel stretch; and 2) the neural, which encom-passes afferent, central, and efferent neural function. In ourstudy volunteers, the carotid diameter increased during hyp-oxia. Because the distensibility decreases with increasing ar-terial diameter (27), it is reasonable to presume that the

Fig. 4. Integrated cardiovagal (black bars) and sympathetic (gray bars) barore-flex responsiveness before, during, and after isocapnic hypoxia (Hx). Data arepresented as means � SE. *P � 0.05 vs. previous time point. AIU, arbitraryintegrated unit.

Fig. 5. The mechanical component (black bars) and neural component (graybars) of the sympathetic baroreflex arc before hypoxia, after 1 and 7 h ofpoikilocapnic hypoxia, and after return to room air. Data are presented asmeans (numbers shown in bars) � SE (numbers shown in parentheses). *P �0.05 vs. prehypoxia.



on February 6, 2009


ownloaded from


transduction of arterial pressure into the carotid stretch may bediminished during hypoxia when the carotid diameter isgreater, leading to a reduction in baroreflex responsiveness.However, we did not observe any systematic decline in themechanical component during the hypoxic exposure. Thiscould be due to several factors. First, the dilation observed islikely due to endothelium-dependent (24, 44) or -independent(32) flow-mediated processes. Thus, the increased distensibil-ity associated with vascular smooth muscle relaxation mayoffset the stretch-related decrease in distensibility. Second,estimates of arterial distensibility are often derived from datameasured while arterial pressure and diameter are in a rela-tively steady state. As we have shown before, these pulsatileestimates do not reflect the dynamic pressure-diameter inter-action that occurs during acute baroreflex perturbations (18).Lastly, the fact that we used peripheral blood pressure esti-mates rather than carotid pressure estimates may induce somesystematic error. The use of peripheral pressure estimates indetermining central vascular characteristics has been criticizedin the past (28). However, a direct comparison between thecentral and peripheral pressure-based estimates has shown thatthe use of tonometry-derived carotid pressures makes no dif-ference in the estimate of reflex relations (25). This may bedue, in part, to inherent problems regarding the valid calibra-tion of the tonometric signals (47).

Although the reduction in sympathetic baroreflex respon-siveness was not related to a decline in the mechanical com-ponent in our volunteers, the neural component was reducedduring hypoxia, particularly after acclimatization. This may bedue to a reduction in baroreceptor discharge rate (baroreceptorresetting) in response to the chronic stretch of the barosensoryvessels. However, the receptor resetting is an acute phenome-non in response to the rapid and large changes in arterial bloodpressure (20). A chronic stimulation of receptors does not leadto further reductions in afferent neural activity (20). Therefore,the difference in baroreflex responsiveness at 1 and 7 h ofhypoxia cannot be explained by a resetting of the barorecep-tors. Moreover, the acute reduction in baroreceptor dischargerate is not sustained when the stimulus is removed (20). Thereduced medullary transduction of baroreceptor afferent activ-ity into sympathetic efferent activity could also underlie thedepressed neural component of the sympathetic limb of thebaroreflex. Endogenous nitric oxide production, which canincrease during hypoxia, has been shown to depress the barore-flex afferent responsiveness in animals (22). Although a reduc-tion in afferent responsiveness should lead to a reduction in theneural component for both cardiovagal and sympathetic limbs,our data show a reduction in only the sympathetic limb. Thismay be due in part to the inhibitory effects of nitric oxide onsympathetic efferent outflow (36). In addition, the medullarycardiovagal and sympathetic efferent outflow may be differen-tially regulated as demonstrated by Ma and colleagues (21).Taken together, these previous findings may explain the selec-tive reduction in integrated sympathetic baroreflex responsive-ness seen in our data. However, our data cannot directlyaddress this possibility.

It is plausible that the hyperventilation per se may havereduced baroreflex responsiveness. Using a closed-loop ap-proach to characterize baroreflex function (frequency domainanalysis), Van De Borne and colleagues (45) reported thatisocapnic hyperventilation was associated with a decline in the

-index between systolic blood pressure and MSNA. Thisdecline suggests that hyperventilation alone can reduce barore-flex modulation of vascular sympathetic activity (45). Con-versely, using an open-loop approach (modified Oxford) Hal-liwill et al. (13) demonstrated that hyperventilation had noaffect on baroreflex responsiveness. Therefore, it remains un-clear as to whether hyperventilation per se can reduce barore-flex function, although it is important to note that neither VanDe Borne et al. (45) nor Halliwill et al. (13) reported thathyperventilation affected mean MSNA levels.

Given that baroreflex responsiveness and sympathetic activ-ity were not consistently related within individuals, it is un-likely that the decline in sympathetic activity is primarilymediated through the baroreflex. This appears to be true duringtransient hypoxic exposure as well (12). However, our datacannot address the possibility that, although blunted, thebaroreflex inhibited sympathetic activity to some extent. Theeffect of angiotensin II on MSNA should be considered. Acutehypoxia has been shown to reduce arterial angiotensin IIconcentrations (39, 41, 49). Since angiotensin II stimulatessympathetic outflow from the medulla, the hypoxia-mediateddecline in angiotensin II may play a role in reducing sympa-thetic activity during VAH. However, since we did not collectblood for analysis, our data cannot address this possibility.

It is interesting to note that upon the cessation of the hypoxicstimulus, the forearm vascular resistance remained below pre-exposure levels, whereas the carotid diameters returned tobaseline values. Both peripheral and cerebral blood flow ve-locities increase during poikilocapnic hypoxia. However, fore-arm blood flow remains elevated after hypoxia (42), whereascerebral blood flow returns to preexposure levels after expo-sure (30). With this in mind, the reduced forearm vascularresistance is likely due to reduced vascular sympathetic activityalong with enhanced nitric oxide production causing vasodila-tion in the more muscular peripheral radial and ulnar arteriesand arterioles. In contrast, the blood flow through the carotidartery likely returned to baseline levels, reducing shear stress.Moreover, the less muscular artery is likely less affected by thereduced vascular sympathetic outflow.

Limitations

A consistent subject of debate in the study of autonomicregulation is the proper metric for expressing cardiovagalbaroreflex responsiveness. Because of the inverse, curvelinearrelation between R-R interval and heart rate, a rather largechange in R-R interval can result in a small change in heartrate, potentially obscuring changes in autonomic regulation.Since heart rate is linearly related to cardiac output, and thusblood pressure, this metric may be most appropriate whenexamining the ability of the baroreflex to buffer changes inarterial blood pressure. However, in the current investigation,our experimental question was related to changes in the auto-nomic control of afferent neural activity. Therefore, sincecarotid afferent frequency is linearly related to R-R interval incats and humans (3, 19), this metric seems most appropriate forthis investigation.

Given that the microelectrode was removed after 1 h ofhypoxia and then replaced before the 7-h hypoxia measure-ment, the reproducibility of MSNA is an important issue. First,



on February 6, 2009


ownloaded from


it should be noted that the key dependent variable, sympatheticbaroreflex responsiveness, is based on the change in MSNArather than the absolute levels of MSNA. Nonetheless, intra-individual MSNA is highly reproducible, even when measuredseveral years apart (9). Moreover, MSNA recorded in the leg isstrongly associated with MSNA measured simultaneously inthe arm (33, 40, 46). Taken together, this should provideassurance that our measurements of both steady-state MSNAand sympathetic baroreflex responsiveness, taken in the sameproximity but different site within the peroneal nerve, can becompared quantitatively.

Conclusions

These data provide evidence that baroreflex function isinhibited during the initial stages of VAH. Furthermore, by theaddition of continuous ultrasound images of the carotid arteryduring baroreceptor stimulation, we have been able to providedata that strongly suggest that the reduction in baroreflexresponsiveness is likely due to changes in the medullaryafferent-efferent integration, possibly associated with ventila-tory acclimatization to hypoxia. Although our data do notdirectly address this hypothesis, taken together with previousdata in animals, it is likely. Furthermore, in contrast to previouswork that has only examined the chemoreceptor-baroreceptorinteractions before acclimatization, our work suggests that thistransition period during hypoxia encompasses a unique inte-gration between chemoreceptor, baroreceptor, and end-organfunction.

GRANTS

This research was supported by National Heart, Lung, and Blood InstituteGrant HL-072648-01A1.

REFERENCES

1. Bascom DA, Clement ID, Cunningham DA, Painter R, Robbins PA.Changes in peripheral chemoreflex sensitivity during sustained, isocapnichypoxia. Respir Physiol 82: 161–176, 1990.

2. Bisgard GE. Increase in carotid body sensitivity during sustained hyp-oxia. Biol Signals 4: 292–297, 1995.

3. Borst C, Karemaker JM, Dunning AJ, Bouman LN, Wagner J.Frequency limitation in the human baroreceptor reflex. J Auton Nerv Syst9: 381–397, 1983.

4. Bristow JD, Brown EB Jr, Cunningham DJ, Goode RC, Howson MG,Sleight P. The effects of hypercapnia, hypoxia and ventilation on thebaroreflex regulation of the pulse interval. J Physiol 216: 281–302, 1971.

5. Bristow JD, Brown EB Jr, Cunningham DJ, Goode RC, Howson MG,Sleight P. The influence of ventilation, carbon dioxide and hypoxia on thebaroreceptor reflex in man. J Physiol 198: 102passim–103p, 1968.

6. Calbet JA. Chronic hypoxia increases blood pressure and noradrenalinespillover in healthy humans. J Physiol 551: 379–386, 2003.

7. Cunningham DJ, Petersen ES, Pickering TG, Sleight P. The effects ofhypoxia, hypercapnia, and asphyxia on the baroreceptor-cardiac reflex atrest and during exercise in man. Acta Physiol Scand 86: 456–465, 1972.

8. Ebert TJ, Morgan BJ, Barney JA, Denahan T, Smith JJ. Effects ofaging on baroreflex regulation of sympathetic activity in humans. Am JPhysiol Heart Circ Physiol 263: H798–H803, 1992.

9. Fagius J, Wallin BG. Long-term variability and reproducibility of restinghuman muscle nerve sympathetic activity at rest, as reassessed after adecade. Clin Auton Res 3: 201–205, 1993.

10. Gilmartin G, Tamisier R, Anand A, Cunnington D, Weiss JW. Evi-dence of impaired hypoxic vasodilation after intermediate-duration hy-poxic exposure in humans. Am J Physiol Heart Circ Physiol 291: H2173–H2180, 2006.

11. Glass G, Hopkins K. Statistical Methods in Education and Psychology.Boston: Allyn and Bacon, 1996.

12. Halliwill JR, Minson CT. Effect of hypoxia on arterial baroreflex controlof heart rate and muscle sympathetic nerve activity in humans. J ApplPhysiol 93: 857–864, 2002.

13. Halliwill JR, Morgan BJ, Charkoudian N. Peripheral chemoreflex andbaroreflex interactions in cardiovascular regulation in humans. J Physiol552: 295–302, 2003.

14. Hamner JW, Taylor JA. Automated quantification of sympathetic beat-by-beat activity, independent of signal quality. J Appl Physiol 91: 1199–1206, 2001.

15. Hansen J, Sander M. Sympathetic neural overactivity in healthy humans afterprolonged exposure to hypobaric hypoxia. J Physiol 546: 921–929, 2003.

16. Hubert LJ. The use of orthogonal polynomials for trend analysis. AmEduc Res J 10: 241–244, 1973.

17. Hunt BE, Fahy L, Farquhar WB, Taylor JA. Quantification of mechan-ical and neural components of vagal baroreflex in humans. Hypertension37: 1362–1368, 2001.

18. Hunt BE, Farquhar WB, Taylor JA. Does reduced vascular stiffeningfully explain preserved cardiovagal baroreflex function in older, physicallyactive men? Circulation 103: 2424–2427, 2001.

19. Kunze DL. Reflex discharge patterns of cardiac vagal efferent fibres.J Physiol 222: 1–15, 1972.

20. Landgren S. On the excitation mechanism of the carotid baroceptors. ActaPhysiol Scand 26: 1–34, 1952.

21. Ma X, Abboud FM, Chapleau MW. Analysis of afferent, central, andefferent components of the baroreceptor reflex in mice. Am J PhysiolRegul Integr Comp Physiol 283: R1033–R1040, 2002.

22. Matsuda T, Bates JN, Lewis SJ, Abboud FM, Chapleau MW. Modu-lation of baroreceptor activity by nitric oxide and S-nitrosocysteine. CircRes 76: 426–433, 1995.

23. Mattson MP. Hormesis defined. Ageing Res Rev 7: 1–7, 2008.24. Messina EJ, Sun D, Koller A, Wolin MS, Kaley G. Role of endothelium-

derived prostaglandins in hypoxia-elicited arteriolar dilation in rat skeletalmuscle. Circ Res 71: 790–796, 1992.

25. Monahan KD, Tanaka H, Dinenno FA, Seals DR. Central arterial compli-ance is associated with age- and habitual exercise-related differences incardiovagal baroreflex sensitivity. Circulation 104: 1627–1632, 2001.

26. Morgan BJ, Crabtree DC, Palta M, Skatrud JB. Combined hypoxiaand hypercapnia evokes long-lasting sympathetic activation in humans.J Appl Physiol 79: 205–213, 1995.

27. Nichols WW, O’Rourke MF. McDonald’s Blood Flow in Arteries:Theoretical, Experimental and Clinical Principles (5th ed.). New York:Oxford University Press, 2005.

28. O’Rourke MF, Nichols WW. Does reduced vascular stiffening fullyexplain preserved cardiovagal baroreflex function in older physicallyactive men? Circulation 105: E11, 2002.

29. Pickering TG, Davies J. Estimation of the conduction time of thebaroreceptor-cardiac reflex in man. Cardiovasc Res 7: 213–219, 1973.

30. Poulin MJ, Fatemian M, Tansley JG, O’Connor DF, RobbinsPA. Changes in cerebral blood flow during and after 48 h of bothisocapnic and poikilocapnic hypoxia in humans. Exp Physiol 87:633– 642, 2002.

31. Powell FL, Milsom WK, Mitchell GS. Time domains of the hypoxicventilatory response. Respir Physiol 112: 123–134, 1998.

32. Quan A, Ward ME, Kulandavelu S, Adamson SL, Langille BL.Endothelium-independent flow-induced dilation in the mouse carotid ar-tery. J Vasc Res 43: 383–391, 2006.

33. Rea RF, Wallin BG. Sympathetic nerve activity in arm and leg musclesduring lower body negative pressure in humans. J Appl Physiol 66:2778–2781, 1989.

34. Robbins PA. Hypoxic ventilatory decline: site of action. J Appl Physiol79: 373–374, 1995.

35. Sagawa S, Torii R, Nagaya K, Wada F, Endo Y, Shiraki K. Carotidbaroreflex control of heart rate during acute exposure to simulated alti-tudes of 3,800 m and 4,300 m. Am J Physiol Regul Integr Comp Physiol273: R1219–R1223, 1997.

36. Sakuma I, Togashi H, Yoshioka M, Saito H, Yanagida M, TamuraM, Kobayashi T, Yasuda H, Gross SS, Levi R. NG-methyl-L-argi-nine, an inhibitor of L-arginine-derived nitric oxide synthesis, stimu-lates renal sympathetic nerve activity in vivo. A role for nitric oxide inthe central regulation of sympathetic tone? Circ Res 70: 607– 611,1992.

37. Smith SA, Stallard TJ, Littler WA. Estimation of sinoaortic barore-ceptor heart rate reflex sensitivity and latency in man: a new micro-



on February 6, 2009


ownloaded from


computer assisted method of analysis. Cardiovasc Res 20: 877– 882,1986.

38. Somers VK, Mark AL, Abboud FM. Interaction of baroreceptor andchemoreceptor reflex control of sympathetic nerve activity in normalhumans. J Clin Invest 87: 1953–1957, 1991.

39. Stalcup SA, Lipset JS, Mellins RB. Modulation of converting enzymeactivity by hypoxia and its physiological effects. Ciba Found Symp 78:293–311, 1980.

40. Sundlof G, Wallin BG. The variability of muscle nerve sympatheticactivity in resting recumbent man. J Physiol 272: 383–397, 1977.

41. Szidon P, Bairey N, Oparil S. Effect of acute hypoxia on the pulmonaryconversion of angiotensin I to angiotensin II in dogs. Circ Res 46:221–226, 1980.

42. Tamisier R, Hunt BE, Gilmartin GS, Curley M, Anand A, Weiss JW.Hemodynamics and muscle sympathetic nerve activity after 8 h of sus-tained hypoxia in healthy humans. Am J Physiol Heart Circ Physiol 293:H3027–H3035, 2007.

43. Tansley JG, Clar C, Pedersen ME, Robbins PA. Human ventilatoryresponse to acute hyperoxia during and after 8 h of both isocapnic andpoikilocapnic hypoxia. J Appl Physiol 82: 513–519, 1997.

44. Toporsian M, Govindaraju K, Nagi M, Eidelman D, ThibaultG, Ward ME. Downregulation of endothelial nitric oxide synthase inrat aorta after prolonged hypoxia in vivo. Circ Res 86: 671– 675,2000.

45. Van De Borne P, Mezzetti S, Montano N, Narkiewicz K, Degaute JP,Somers VK. Hyperventilation alters arterial baroreflex control of heartrate and muscle sympathetic nerve activity. Am J Physiol Heart CircPhysiol 279: H536–H541, 2000.

46. Wallin BG, Victor RG, Mark AL. Sympathetic outflow to restingmuscles during static handgrip and postcontraction muscle ischemia. Am JPhysiol Heart Circ Physiol 256: H105–H110, 1989.

47. Wilkinson IB, MacCallum H, Flint L, Cockcroft JR, Newby DE,Webb DJ. The influence of heart rate on augmentation index and centralarterial pressure in humans. J Physiol 525: 263–270, 2000.

48. Xie A, Skatrud JB, Puleo DS, Morgan BJ. Exposure to hypoxiaproduces long-lasting sympathetic activation in humans. J Appl Physiol91: 1555–1562, 2001.

49. Zakheim RM, Molteni A, Mattioli L, Park M. Plasma angiotensin IIlevels in hypoxic and hypovolemic stress in unanesthetized rabbits. J ApplPhysiol 41: 462–465, 1976.



on February 6, 2009


ownloaded from


Documents

AJP Heart 2008