Epinephrine, vasodilation and hemoconcentration in syncopal, healthy men and women

Autonomic Neuroscience: Basic and Clinical 93 Ž2001. 79–90 www.elsevier.comrlocaterautneu

Epinephrine, vasodilation and hemoconcentration in syncopal, healthy men and women Joyce M. Evans a,) , Fabio M. Leonelli b,c , Michael G. Ziegler b,c , Casey M. McIntosh a , Abhijit R. Patwardhan a , Andrew C. Ertl b,c , Charles S. Kim a , Charles F. Knapp a a

Center for Biomedical Engineering and DiÕision of Cardiology, UniÕersity of Kentucky, Lexington, KY 40506-0070, USA b Department of Medicine, USCD, San Diego, CA 92103-8341, USA c Cardiology, PennsylÕania State College of Medicine, Hershey, PA 17033, USA Received 27 March 2001; received in revised form 3 May 2001; accepted 9 July 2001

Abstract Healthy young people may become syncopal during standing, head up tilt ŽHUT. or lower body negative pressure ŽLBNP.. To evaluate why this happens we measured hormonal indices of autonomic activity along with arterial pressure ŽAP., heart rate ŽHR., stroke volume ŽSV., cardiac output ŽCO., total peripheral resistance ŽTPR. and measures of plasma volume. Three groups of normal volunteers Ž n s 56. were studied supine, before and during increasing levels of orthostatic stress: slow onset, low level, lower body negative pressure ŽLBNP. ŽGroup 1., 708 head up tilt ŽHUT. ŽGroup 2. or rapid onset, high level, LBNP ŽGroup 3.. In all groups, syncopal subjects demonstrated a decline in TPR that paralleled the decline in AP over the last 40 s of orthostatic stress. Ten to twenty seconds after the decline in TPR, HR also started to decline but SV increased, resulting in a net increase of CO during the same period. Plasma volume ŽPV, calculated from change in hematocrit. declined in both syncopal and nonsyncopal subjects to a level commensurate with the stress, i.e. Group 3 ) Group 2 ) Group 1. The rate of decline of PV, calculated from the change in PV divided by the time of stress, was greater Ž p - 0.01. in syncopal than in nonsyncopal subjects. When changes in vasoactive hormones were normalized by time of stress, increases in norepinephrine Ž p - 0.012, Groups 2 and 3. and epinephrine Ž p - 0.025, Group 2. were greater and increases in plasma renin activity were smaller Ž p - 0.05, Group 2. in syncopal than in nonsyncopal subjects. We conclude that the presyncopal decline in blood pressure in otherwise healthy young people resulted from declining peripheral resistance associated with plateauing norepinephrine and plasma renin activity, rising epinephrine and rising blood viscosity. The increased hemoconcentration probably reflects increased rate of venous pooling rather than rate of plasma filtration and, together with cardiovascular effects of imbalances in norepinephrine, epinephrine and plasma renin activity may provide afferent information leading to syncope. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Arterial pressure; Autonomic balance; Plasma filtration; Cardiac output; Catecholamines; Heart rate; Plasma volume; Hematocrit; Peripheral resistance

1. Introduction Despite short- and long-term adjustments to counteract the effects of orthostasis Ži.e. increased heart rate, peripheral resistance, vasoconstrictive and plasma conserving hormones., failure of blood pressure regulation to the point of fainting Žsyncope. is a commonly reported malady. The sequence of passive and regulatory events leading to this

)

Corresponding author. Tel.: q859-257-2685; fax: q859-257-1856.

loss of blood pressure regulation has been the subject of extensive study that is summarized in detail by Rowell Ž1993a.. In particular, patients with minimal orthostatic tolerance have been the focus of studies identifying autonomic dysfunction as a major contributor to syncope ŽRobertson et al., 1986.. In patients with recurrent, but not completely debilitating, orthostatic problems, autonomic withdrawal was also implicated when cessation of muscle sympathetic nerve activity accompanied the rapid fall in blood pressure associated with frank syncope ŽMorillo et al., 1997; Mosqueda-Garcia et al., 1997.. Healthy subjects can also experience orthostatically induced fainting; how-

1566-0702r01r$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 1 5 6 6 - 0 7 0 2 Ž 0 1 . 0 0 3 2 3 - X

80

J.M. EÕans et al.r Autonomic Neuroscience: Basic and Clinical 93 (2001) 79–90

ever, the sequence of events leading to this occurrence Žsyncope of unexplained etiology. is less well defined ŽRowell, 1993a; Mosqueda-Garcia et al., 1997; Sra et al., 1994.. In general, syncope in these otherwise healthy people results from a mismatch between cardiac output and organ flow that can be induced by exercise, heat stress, andror prolonged periods of reduced exposure to gravity stress ŽRowell, 1993a.. The role of plasma filtration in the development of syncope has not been fully explored even though decrements in intravascular volume in response to orthostatic stress have been reported for years, e.g. ŽRowell, 1993b; Hinghofer-Szalkay et al., 1992; Aratow et al., 1993; Lundvall and Bjerkhoel, 1995.. This loss of intravascular volume has been found to be as great as 18% in response to 15 min of head-up tilt ŽHUT. ŽLundvall and Bjerkhoel, 1995.. Hinghofer-Szalkay et al. Ž1992. and Greenleaf et al. Ž1979. report a 6% decrease in blood volume in response to 15 min of standing while Aratow et al. Ž1993. found a 14% decline in plasma volume in response to 4 h of y30 mm Hg lower body negative pressure ŽLBNP.. Recently, it was reported that plasma filtration and the rate of volume change in the leg were greater in a group of patients with recurring syncope than in healthy young volunteers ŽBrown and Hainsworth, 1999.. However, interactions between venous pooling and plasma filtration are not yet established in the development of syncopal symptoms. Therefore, a major goal of the present study is to examine the net effect of orthostatic stress on plasma filtrationrabsorption during and following orthostatic stress. Measurements made during orthostatic stress will provide an indication of the rate of fluid filtrationrabsorption. Measurements made immediately after return to control conditions will provide an indication of net plasma filtration resulting from the magnitude of venous pooling and give an index of the effective rate of volume loss. A second goal is to address the relative roles of vasoactive hormones in the development of syncope. Among patients referred for complaints of syncope, half became syncopal during tilting while 62% experienced syncope when tilt was accompanied by an infusion of the b agonist drug isoproterenol ŽLinzer et al., 1997.. Many normal subjects without complaints of fainting also become syncopal during tilting. Recurrent syncope in otherwise normal subjects often responds to therapy with b blocking drugs ŽPerry and Garson, 1991; Grubb et al., 1991; Slotwiner et al., 1997; Natale et al., 1996; Cox et al., 1995; Abe et al., 1995.. This evidence has led to the common explanation that excess sympathetic stimulation to the heart and lungs activates the Bezold–Jarisch reflex ŽPerry and Garson, 1991; Abi-Samra et al., 1988; Purcell, 1992; Manolis et al., 1990. leading to vasodilation, bradycardia and syncope. However, syncope is often accompanied by low plasma norepinephrine levels ŽPerry and Garson, 1991; Ziegler and Lake, 1985; Ziegler et al., 1986.. Since the b agonist isoproterenol can precipitate syncope ŽLinzer et al.,

1997. and b blocking drugs can prevent syncope ŽPerry and Garson, 1991; Grubb et al., 1991; Slotwiner et al., 1997; Natale et al., 1996; Cox et al., 1995; Abe et al., 1995., the endogenous b agonist epinephrine might play a more important role than norepinephrine in syncope in normal subjects ŽKlingenheben et al., 1996; Bhargava et al., 1996.. Epinephrine blood levels are much lower than norepinephrine levels and are at the limits of sensitivity for many assays. The development of an assay with better sensitivity for plasma epinephrine permits an investigation of the role of this stress hormone in syncope among normal volunteers ŽKennedy and Ziegler, 1990.. The final goal of this study is to present an overall, integrated hemodynamic and neurohumoral profile of the time course of the development of syncopal symptoms in healthy subjects undergoing orthostatic stress.

2. Methods 2.1. Subjects Data were acquired from three studies performed over a 3-year period at the University of Kentucky. Group 1 consisted of 10 young men Ž73.5 " 1.8 kg, 175 " 1.4 cm and 24.9 " 0.8 years. and 12 young women Ž66.2 " 2.5 kg, 162 " 2.1 cm and 25.2 " 0.7 years.. Group 2 consisted of 8 men Ž72.2 " 3.1 kg, 172.8 " 1.2 cm and 28.1 " 1.2 years. and 8 women Ž61.6 " 2.2 kg, 163.9 " 2.5 cm, 32.1 " 1.1 years.. Group 3 consisted of 9 men Ž77.6 "3.4 kg, 179.0 " 2.0 cm, 25.2 " 1.0 years. and 9 women Ž68.4 " 3 kg, 167 " 2.7 cm and 26.5 "1.3 years.. Of the 56 total subjects, 9 women and 12 men did not engage in any form of routine conditioning while the remaining subjects participated in mild aerobic exercise: 3.3 " 0.5 hrweek for women and 4.1 " 0.6 hrweek for men. The ethnic origins of all groups were predominantly white with three men and two women of Chinese origin, two men of Asian Indian origin, four men of Latin origin and three African American women. All subjects were healthy volunteers whose blood pressures at screening were in the second and third quartiles of blood pressure distribution for their age range. All women were naturally cycling and were studied between days 2 and 10 of their menstrual cycle. All subjects gave written consent to this protocol approved both by the local Institutional Review Board and by the Advisory Committee of the General Clinical Research Center. 2.2. Experimental protocol Subjects in Groups 1 and 3 were admitted to the Clinical Research Center ŽCRC. at the University of Kentucky on the evening before each study and were fasted after midnight. At 7:00 AM, a venous catheter ŽJelco, 20

J.M. EÕans et al.r Autonomic Neuroscience: Basic and Clinical 93 (2001) 79–90

gauge. was placed in an antecubetal vein. Subjects were given a no fat, low protein breakfast and were accompanied on a walk for the next 20 min. Subjects returned to the CRC and were supine for 40 to 60 min during placement and calibration of instrumentation. All hemodynamic data were acquired on a beat-to-beat basis during 20 min of supine rest followed by 20 min of LBNP exposure to either slow onset rate of ; 3 mm Hgrs down to y35 mm Hg ŽGroup 1. or a fast onset rate of ; 10 mm Hgrs down to y50 mm Hg ŽGroup 3.. For both LBNP tests, the lower half of the subject’s body was sealed inside the chamber at the level of the iliac crest. Support was provided by a footrest, custom-fitted to each subject. Group 2 subjects were admitted to the CRC on the morning of the study after fasting for 8 h. Following placement of the venous catheter and a short walk to the Syncope Clinic, they were placed supine on the tilt table for 40 to 60 min during instrumentation placement and calibration. Data were acquired during 20 min of supine rest followed by 30 min of 708 HUT Žonset rate ; 48rs.. For all groups, LBNP or tilt was terminated at the appearance of syncopal symptoms or at the request of the subject. Electrical impedance leads were attached to monitor ECG, stroke volume ŽSV., cardiac output ŽCO. and end diastolic volume ŽEDV. wBoMed, Cardiodynamic Monitorx. The beat-to-beat output of the BoMed was averaged over four beats and was therefore delayed by that period of time. This delay was accounted for in the analysis. Beat-to-beat arterial pressure ŽAP. was measured by finger cuff wFinapres, Ohmedax.

81

2.3.2. Hematocrit, hemoglobin and plasma protein concentrations Microhematocrit ŽHCT. and plasma refractometer ŽProtometer, National Instruments. readings were taken from all subjects in order to estimate red cell and plasma protein densities. All HCT determinations were made using a protocol developed by one of the authors ŽJacob et al., 1998.. Plasma volume estimations were calculated: % PV change s 100r Ž100 y HCTcontrol . = 100ŽHCTcontrol y HCTtest .rHCTtest ŽVan Beaumont, 27.. In a subset of Group 3 Ž n s 9., we also analyzed each sample for hemoglobin ŽHb. content and used these values to compute changes in plasma volume using the method of Dill and Costill Ž1974.. Hemoglobin was measured in triplicate using the cyanmethemoglobin method ŽVan Beaumont, 1972.. For Groups 1 and 3, 1-ml samples were drawn each minute during the last 4 min of control, at 30-s intervals for 5 min following the onset and cessation of LBNP and at 2-min intervals during LBNP. For Group 2, 1-ml samples were drawn at the end of control, at 8-min intervals during HUT and at the cessation of HUT. 2.4. Hemodynamic data analysis For each subject, beat to beat hemodynamic data were averaged over 20-min supine rest. During LBNP or tilt, data were averaged over 10-s intervals for the last 2 min. 2.5. Statistical analysis

2.3. Hormonal assays

2.3.1. VasoactiÕe hormones For Groups 1 and 3, 10 ml of blood were drawn during the last 2 min of both supine rest and LBNP. For Group 2, 10 ml of blood were drawn at the end of supine rest and at 8-min intervals throughout HUT. For all subjects, the sample was drawn within 2 min of the cessation of the stress. All samples were immediately spun in a refrigerated centrifuge. Plasma was extracted and frozen until samples were assayed. To provide an index of sympathetic activity, we assayed for plasma norepinephrine ŽNE. and epinephrine ŽE.. The radioenzymatic assay for these catecholamines has a sensitivity of 6 pgrml and an intra-assay coefficient of variation of 13% ŽKennedy et al., 1994.. To provide an independent assay for vagally mediated activity, pancreatic polypeptide ŽPPP. levels were assayed by Dr. George Greeley, Dept. of Surgery, UTMB, Galveston, TX using a radioimmunoassay technique ŽSchwartz, 1983.. To assess plasma conserving activity, we assayed for plasma renin activity ŽPRA., using a radioimmunoassay technique ŽQuest Diagnostics, San Juan Capistrano, CA. with a sensitivity of 0.1 ngrmlrh and an intra-assay coefficient of variability of 6.5%.

A mixed model ANOVA was used to assess the significance of main effects: gender, time Žof tilt or LBNP stress. and condition Žsyncopal vs. nonsyncopal. and their interactions. The mixed model was used since time is a within subject factor while gender and condition are between subjects factors. A two factor ANOVA was used for testing supine control values. When post hoc testing was warranted Žsignificant F ratio for the effect of interest., t-statistics with degrees of freedom determined by Satterwhaite’s approximation, were constructed to compare mean responses. For some variables, the change in the variable normalized by the time of orthostatic stress was computed and comparisons were made between syncopal and nonsyncopal subjects using two-sample t-statistics. In all cases, statistical significance was determined at the 0.05 level. Data presented are mean " SEM.

3. Results 3.1. Resting control Mean values for all groups, taken during resting control, are given in Table 1. This table includes only subjects with

J.M. EÕans et al.r Autonomic Neuroscience: Basic and Clinical 93 (2001) 79–90

82

Table 1 Pre-stress control data for syncopal and nonsyncopal male and female subjects from all groups for whom complete hemodynamic and neurohumoral values were available Hemodynamic variables Žmean " SEM. Arterial pressure Žmm Hg.

Heart rate Žbpm.

Cardiac output Žlrmin.

Peripheral resistance Žmm HgrŽlrmin..

Stroke volume Žml.

Males Nonsyncopal Ž n s 14. Syncopal Ž n s 10.

95.1 " 2.2 91.1 " 4.9

61.9 " 2.6 62.1 " 3.0

7.11 " 0.46 7.99 " 0.53

14.0 " 0.8 12.1 " 1.1) )

117 " 9 130 " 13

Females Nonsyncopal Ž n s 13. Syncopal Ž n s 13.

90.0 " 4.5 83.2 " 2.7

64.9 " 3.5 67.3 " 2.0

6.80 " 0.53 7.51 " 0.50

14.2 " 1.3 11.6 " 0.7 ) )

106 " 8 111 " 7

Hematocrit Ž%.

Epinephrine Žpgrml.

Norepinephrine Žpgrml.

Pancreatic polypeptide Žpgrml.

Plasma renin activity Žngrmlrh.

Males Nonsyncopal Ž n s 14. Syncopal Ž n s 10.

43.0 " 0.6 42.9 " 1.0

17.6 " 3.4 16.3 " 2.9

173.1 " 13.0 178.4 " 16.8

73.2 " 14.1 65.3 " 9.5

1.1 " 0.3 1.6 " 0.5

Females Nonsyncopal Ž n s 13. Syncopal Ž n s 13.

39.0 " 1.0 ) 37.1 " 0.8 )

14.3 " 2.6 10.4 " 1.9

190.8 " 26.3 173.2 " 13.8

82.8 " 24.0 63.8 " 12.3

0.8 " 0.2 0.8 " 0.2

Vasoactive hormones Žmean " SEM.

)

Indicates a significant gender difference Ž p - 0.0001.. Indicates a significant Ž p - 0.04. difference between syncopal and nonsyncopal subjects.

))

complete data sets for all variables. Even though syncopal subjects of both genders tended to have lower resting values of arterial pressure, peripheral resistance, epinephrine, norepinephrine and pancreatic polypeptide

than did their nonsyncopal counterparts, only peripheral resistance was significantly Ž p - 0.04. lower. Women had lower values of HCT Ž p - 0.001. than did men, but no other gender differences were significant.

Fig. 1. Response of syncopal and nonsyncopal subjects to the last 2 min of orthostatic stress. Data from tilt and LBNP tests were combined in order to compare syncopal and nonsyncopal groups of subjects. ANOVA main effects Žsyncopalrnonsyncopal groups and time of stress. and interactions are given for each variable. Data are mean " SEM.

J.M. EÕans et al.r Autonomic Neuroscience: Basic and Clinical 93 (2001) 79–90

3.2. Response to stress 3.2.1. Incidence of syncopal responses Five Žall women. of twenty-two Ž23%. subjects in Group 1 had syncopal symptoms while undergoing 20 min of slow onset, y35 mm Hg LBNP. Nine Žfive women, four men. of sixteen Ž56%. Group 2 subjects had syncopal symptoms during 30 min of 708 HUT while 12 Žfive women, seven men. of 18 Ž67%. Group 3 subjects had syncopal symptoms during 20 min of rapid onset, y50 mm Hg LBNP. 3.2.2. Nonsyncopal plasma Õolume responses For the low level LBNP stress ŽGroup 1., plasma volume declined 6.7 " 0.7% over 20 min. For Group 2, the decline in response to HUT was 11.6 " 1.1% over 30 min. At the higher level LBNP ŽGroup 3., plasma volume declined 17.9 " 0.5.0% over 20 min. Plasma volume decreases were significantly different between each group, p - 0.02. 3.2.3. Hemodynamic means Arterial pressure, peripheral resistance, heart rate and cardiac output for the last 2 min of orthostatic stress ŽHUT and both levels of LBNP. are given in Fig. 1 for 29 nonsyncopal and 26 syncopal subjects from all groups. For the last 2 min of stress, both blood pressure and peripheral resistance were lower, and heart rate was higher, in syncopal subjects. In addition, group-by-time interactions were significant for blood pressure, peripheral resistance and cardiac output. The decline in blood pressure that defined the syncopal episode began ; 40 s before the cessation of LBNP. This decline in blood pressure was most clearly associated with the decline in peripheral vascular resistance. The decline in heart rate, which came about 20 s later, was compensated by increased stroke volume Žnot shown. resulting in a net increase in cardiac output across the last 30 s of LBNP. The rise of cardiac output in syncopal subjects accounted for the significant Ž p - 0.007. interaction in this variable: cardiac output was significantly greater just before the release of stress Ž t s 0. than at any other time during the last 2 min. Arterial pressure, peripheral resistance, heart rate and cardiac output for the last 2 min of each stress were analyzed separately for Groups 1, 2 and 3 Žnot shown.. Main effects and interactions were significant for heart rate and arterial pressure for each group as was the interaction effect for peripheral resistance.

83

200 " 16 pgrml and in Group 3 the increases were 267 " 35 vs. 187 " 31 pgrml Ž p - 0.11. for nonsyncopal and syncopal subjects, respectively. However, syncopal episodes occurred at different times for different subjects. We therefore sought a measure of catecholamine response that could be used to make a more accurate comparison between syncopal and nonsyncopal subjects. For this we calculated changes of norepinephrine normalized by the time of orthostatic stress for each subject Žtop panel, Fig. 2.. In Groups 2 and 3, syncopal subjects exhibited steeper Ž p - 0.012 and 0.009. rates of increase in norepinephrine even though their post-test values were not significantly different. Finally, in Group 2, we found that in both syncopal and nonsyncopal subjects, norepinphrine levels underwent their greatest increase during the first 8 min of tilt, plateauing after that Žnot shown.. 3.3.2. Epinephrine Plasma increases in epinephrine were greater for syncopal Ž40.8 " 23.7 pgrml. than for nonsyncopal Ž12.8 " 5.3 pgrml. Group 1 subjects, p - 0.05. Stress-induced increases in epinephrine normalized by the time of stress for syncopal and nonsyncopal volunteers are given for each group in the middle panel of Fig. 2. Syncopal subjects from Group 2 had greater Ž p - 0.025. stress-induced in-

3.3. VasoactiÕe hormones 3.3.1. Norepinephrine Average plasma norepinephrine rose significantly in response to each stress with no significant difference between nonsyncopal and syncopal subjects: the increase in Group 1 was 119 " 18 Žnonsyncopal. vs. 92 " 48 pgrml Žsyncopal.. For Group 2, the increases were 225 " 31 vs.

Fig. 2. Changes in vasoactive hormones normalized by time of orthostatic stress for three levels of orthostatic stress ŽGroups 1, 2 and 3.. Syncopalr nonsyncopal differences are indicated. Data are mean"SEM.

84

J.M. EÕans et al.r Autonomic Neuroscience: Basic and Clinical 93 (2001) 79–90

Fig. 3. Hematocrit responses to y50 mm Hg for three subjects with varying tolerance for LBNP. Differences between control and post stress peaks Žopen symbols. were used to assess changes in plasma volume per time of stress.

creases in epinephrinertime than did their nonsyncopal counterparts. In Group 3 subjects, there was a significant gender by group interaction, contributing to the larger variability in this group. Syncopal men had significantly greater increases Ž p - 0.01. in epinephrinertime Ž14.4 " 3.6 pgrmlrmin. than did nonsyncopal men Ž2.4 " 0.2 pgrmlrmin., syncopal women Ž2.0 " 0.7 pgrmlrmin. or nonsyncopal women Ž2.3 " 0.4 pgrmlrmin.. In Group 2, epinephrine levels plateaued after the first 8 min in nonsyncopal subjects but continued to rise in syncopal subjects, not shown. 3.3.3. Plasma renin actiÕity Like norepinephrine and epinephrine, plasma renin activity increased more Ž p - 0.01. at higher levels of stress,

ranging from a 0.47 " 0.15-ngrmlrh increase for nonsyncopal subjects of Group 1 to a 3.56 " 1.34-ngrmlrh increase for nonsyncopal subjects of Group 2 Žnot shown.. Syncopal subjects of Groups 2 and 3 had smaller Ž p - 0.05. increases in PRA than did nonsyncopal subjects Žnot shown. but when these changes were normalized by time of stress, only the Group 2 increase remained significantly smaller, p s 0.05 ŽFig. 2, bottom panel..

3.3.4. Pancreatic polypeptide The response of pancreatic polypeptide to LBNP or tilt stress Žnot shown. was not different for syncopal and nonsyncopal subjects in any of the three groups nor was there a consistent response to stress.

Fig. 4. Changes in hematocrit normalized by time for three levels of orthostatic stress ŽGroups 1, 2 and 3.. Significant differences between syncopal and nonsyncopal subjects are indicated for each group. Data are mean " SEM.

J.M. EÕans et al.r Autonomic Neuroscience: Basic and Clinical 93 (2001) 79–90

85

Fig. 5. Percentage changes in plasma volume for three levels of orthostatic stress ŽGroups 1, 2 and 3.. Significant differences between syncopal and nonsyncopal subjects are indicated for each group. Data are mean " SEM.

3.4. Fluid filtrationr absorption 3.4.1. During orthostatic stress Groups 1 and 3: examples of hematocrit vs. time of LBNP are given in Fig. 3 for three subjects from Group 3 who demonstrated differing levels of tolerance for 20 min of y50 mm Hg LBNP. Hematocrit decreased during the first minute of LBNP and then began to increase. The rate of increase following the hemodilution Žabsorptive. phase was maintained throughout the stress, indicating that the rate of fluid filtration remained relatively constant for each subject. This was the case for all subjects of this group and there was no significant difference between syncopal and

nonsyncopal subjects. Group 2: Hematocrits were measured at 8-min intervals andror at the appearance of syncopal symptoms, therefore no assessment of the absorptive phase was possible. The average within-sample variability in hematocrit readings was 0.017 hematocrit points. 3.4.2. Post orthostatic stress Following the release of LBNP, hematocrit Žhemoglobin and total protein also, not shown. rose sharply. The difference in HCT between the average control value and the peak value reached after the release of LBNP Žopen symbols of Fig. 3. was used to assess the effects of filtration on total plasma volume. The measurement was

Fig. 6. Tolerance for LBNP plotted as a function of change in hematocrit Žnormalized by time of stress. in response to 20 min of rapid onset y50 mm Hg LBNP.

86

J.M. EÕans et al.r Autonomic Neuroscience: Basic and Clinical 93 (2001) 79–90

made at the peak value in order to indicate the difference due to mixing of blood from dependent regions with the total circulation. As with catecholamines and PRA, we normalized this hematocrit difference by the length of time of the stress. Results for nonsyncopal and syncopal subjects in all groups are given in Fig. 4. The relative increases in hematocrit for Groups 2 and 3 were greater for syncopal compared to nonsyncopal subjects Ž p - 0.007 and 0.01.. The relative increases in hematocrit for Group 1 syncopal and nonsyncopal subjects were not significantly different. Plasma protein concentration, calculated from refractometer readings, and hemoglobin concentration Žneither shown. indicated that the pattern of plasma density paralleled that of hematocrit. Changes in absolute values of plasma volume calculated from the HCT response to each stress are given for syncopal vs. nonsyncopal subjects in Fig. 5. For both Groups 1 and 2, decreases in absolute values of plasma volume were not different between nonsyncopal and syncopal subjects. At the higher level LBNP ŽGroup 3., nonsyncopal subjects declined 17.9 " 0.5% while syncopal subjects declined by only 11.1 " 0.7%, p - 0.001. This smaller decrease in plasma volume for syncopal subjects appears to be a result of the rapid Ž6.4 " 0.6 min. development of syncopal symptoms, sometimes within the hemodilution phase of the LBNP response. Percentage changes in plasma volume using both hematocrit and hemoglobin were calculated for nine Group 3 subjects’ responses to LBNP ŽDill and Costill, 1974.. These results correlated well Ž R 2 s 0.91. with changes calculated from hematocrit alone ŽVan Beaumont, 1972.. Finally, we looked at the potential for using the change in hematocrit normalized by time of stress to predict Group 3 subjects’ tolerance for LBNP ŽFig. 6.. The relationship between tolerance to LBNP stress and hematocrit slope wŽpost stress peak y control.rtime of stressx was significant Ž R 2 s 0.67. and did not differ between men and women: Tolerance time s Ž 1r0.6 HCT slope . q 0.1 min The functional relationships between tolerance times and HCT slopes for Groups 1 and 2 demonstrated relationships that appeared to be dependent upon the stress Žnot shown..

4. Discussion 4.1. Syncope In the present study, vasodilation, different kinetics of norepinephrine, epinephrine and plasma renin activity and more rapid sequestration of filtered blood characterized syncopal, compared to nonsyncopal, otherwise healthy,

subjects. Vasodilation, or the loss of vasoconstriction, was the overriding cause of syncope, with subsequent falling heart rate. While there was no significant difference in the magnitude of norepinephrine responses between syncopal and nonsyncopal subjects, syncopal subjects did demonstrate greater increases in norepinephrine when the responses were normalized by time of stress. The norepinephrine increases of the present study are actually quite similar to those from a recent study of astronauts ŽFritsch-Yelle et al., 1996.; however, our interpretation of results is different. Astronauts who became syncopal during a post-flight stand test exhibited a smaller increase in norepinephrine than did subjects with no symptoms. Our Groups 2 and 3 syncopal subjects had slightly smaller increases in norepinephrine than did nonsyncopal subjects Ž p - 0.11.. However, when normalized by the time of stress, our syncopal subjects had steeper increases in norepinephrine than did our nonsyncopal subjects Ž p - 0.02., indicating a hyper- rather than hypo-adrenergic response. Whether that would have been the case for the astronaut study is not known, but the two studies are dissimilar enough to supply other, quite reasonable explanations for this difference. Both syncopal and nonsyncopal astronauts demonstrated cardiovascular effects from spaceflight indicating increased sympathetic, as well as plasma conserving, activity: supine values of TPR, NE, EPI and PRA were twice as great as those in supine subjects of the present study. Epinephrine results of the present study also support a hyperadrenergic response in syncopal subjects. For Group 1, the stress-induced increases in epinephrine were greater in syncopal subjects. In Group 2, the rise in epinephrine normalized by time of stress was significantly greater in syncopal subjects and, in addition, we were able to determine that epinephrine in syncopal subjects continued to rise throughout the tilt while plateauing in nonsyncopal subjects. Finally, in Group 3 subjects we determined that when the rate of rise of epinephrine was normalized by the time of stress, syncopal men had significantly greater rate of rise of epinephrine than did nonsyncopal men, syncopal women or nonsyncopal women. A net increase in epinephrinernorepinephrine ratio was previously implicated in the development of syncope in otherwise healthy subjects ŽMosqueda-Garcia et al., 1997.. This imbalance could favor beta adrenergically mediated vasodilation over mediated vasoconstriction ŽKennedy et al., 1994. and thereby contribute to the decline in peripheral resistance associated with syncope. Plasma renin activity of syncopal and nonsyncopal subjects also differed. In Groups 2 and 3, nonsyncopal subjects exhibited greater increases in plasma renin activity than did syncopal subjects. When these increases were normalized by the time of stress, the change of PRArtime in Group 2 remained significantly greater for nonsyncopal subjects. Diminished PRA, particularly in the face of catecholamine imbalance, could contribute to syncope

J.M. EÕans et al.r Autonomic Neuroscience: Basic and Clinical 93 (2001) 79–90

through the loss of its vasoconstrictor as well as plasmaconserving properties. The new finding of the present study was the difference in hemoconcentration between syncopal and nonsyncopal subjects. The larger post-stress increase in hematocrit could indicate that syncopal subjects had a greater rate of plasma filtration than did nonsyncopal subjects and, if that were the case, syncopal subjects should have steeper slopes of hematocrit during the orthostatic stress. However, hematocrit slopes were not different for syncopal and nonsyncopal subjects during the application of stress. The difference became apparent only after sequestered blood was released to mix with the total circulation. It therefore seems more likely that the greater post-stress increase in syncopal subjects’ hematocrit reflects greater venous pooling of filtered blood. Increased venous pooling has been reported previously in the form of greater, tilt-induced, decreases in central venous pressure in syncopal, but otherwise healthy, subjects than in nonsyncopal controls or patients with neurally mediated syncope ŽMosqueda-Garcia et al., 1997.. Similar to the present study, other investigators concluded that the increases in hematocrit and hemoglobin that continued for 2 min after the cessation of tilt resulted from the mixing of stored blood with circulating blood ŽLundvall and Bjerkhoel, 1995.. However, the transit time of blood from lower limb veins to the central circulation may be less than the 2 min observed in both studies. Therefore, we speculate that another component of this continued rise may be the increase in capillary perfusion pressure resulting from the vasodilation that occurs with the cessation of orthostatic stress. Whether indicative of plasma filtration rate or rate of venous pooling, the change of hematocrit normalized by the time of stress was predictive of orthostatic tolerance and could therefore be proposed to provide afferent information in the development of syncopal symptoms. Whether hemoconcentration had its effect locally, via venoarteriolar reflexes or centrally was beyond the scope of this study. Absolute values of plasma volume, plasma renin, epinephrine or norepinephrine were not predictive of orthostatic tolerance. This does not mean, however, that independently or as a group, these variables do not provide afferent information that contributes to the development of syncope. The increased rate of rise of epinephrine in syncopal subjects could stimulate venoconstriction, cardiac contractility ŽBezold–Jarisch type reflex. and peripheral vasodilation, through preferential stimulation of peripheral beta2-receptors. The compromised ability of syncopal subjects of Groups 1 and 2 to raise plasma renin activity could indicate a role for that hormone in the sequence of events that regulate blood pressure in response to orthostatic stresses. The indications from Group 2 that norepinephrine and epinephrine increases plateaued in nonsyncopal subjects while epinephrine continued to rise in syncopal subjects leads us to conclude that, since no single variable has been clearly implicated in triggering the onset of syncope,

87

future studies should focus on the time course of the interaction between all of these mechanisms. 4.2. Differences among stresses Differences in the magnitudes of plasma volume decreases between the three groups probably resulted from differences in the level and length of the stresses applied: 20 min of slow onset y35 mm Hg LBNP, 30 min of 708 HUT or 20 min of rapid onset y50 mm Hg LBNP. The magnitude of the plasma volume decrease in our Group 3 subjects was similar to the 18% decrease evoked by 15 min of 858 HUT in healthy males ŽLundvall and Bjerkhoel, 1995.. Regardless of the mechanisms involved in producing the response, the present study clearly supports the conclusion of Lundvall that large, rapid shifts of plasma between intra- and extravascular spaces are a normal component of the human response to the application and cessation of orthostatic stress. Just as normal are neural- and plasmaconserving responses to those plasma shifts. Except for those Group 3 subjects where LBNP was terminated very early, our syncopal and nonsyncopal subjects demonstrated decreases in plasma volume commensurate with the stress ŽFig. 5.. These data led us to conclude that the magnitude of the decrease in plasma volume was not the trigger that initiated syncope in our subjects. This was particularly striking in the case of the Group 3 subjects who became syncopal during the hemodilution phase of LBNP. In these subjects, changes in hematocritrtime were the highest of any subjects studied ŽFig. 6.. We speculate that the rate of change of plasma volume or the rate of venous pooling could provide additional afferent input that contributes to the onset of syncope. 4.3. Presyncopal response to stress Except for peripheral resistance, syncopal and nonsyncopal subjects had similar cardiovascular characteristics at rest. However, several stress-induced differences became evident prior to the final vasodilation. These differences included higher heart rates and lower values of peripheral resistance and arterial pressure in syncopal subjects, apparent at least 2 min before the end of the stress in all three groups. The Yelle study also noted lower supine peripheral resistance in astronauts who would subsequently be unable to complete a 10-min stand test. The lower resistance was significant not only on landing day but also on pre and postflight days, indicating that resting TPR may be a predictor of subjects’ tolerance for orthostatic stress. In the group Ž2. where we looked at catecholamines in detail, epinephrine continued to rise in syncopal subjects but plateaued in nonsyncopal subjects while norepinephrine levels plateaued in both syncopal and nonsyncopal

88

J.M. EÕans et al.r Autonomic Neuroscience: Basic and Clinical 93 (2001) 79–90

subjects. The pattern of continually rising epinephrine accompanied by a plateauing of norepinephrine was similar to that observed in mildly hypoxic subjects made syncopal by an infusion of atrial natriuretic peptide ŽWestendorp et al., 1997.. Previous investigations, however, indicate that epinephrine is not the sole mediator of syncopal vasodilation ŽRowell, 1993a.. Specifically, interactions of epinephrine with hypoxia ŽBlauw et al., 1995. andror atrial natriuretic peptide ŽWestendorp et al., 1997. have been proposed to explain the phenomenon of human hypoxic syncope. The sequence of stress-induced hemodilution followed by hemoconcentration indicated that the effect of increased peripheral resistance Žto decrease plasma filtration. ŽHinghofer-Szalkay et al., 1992; Aratow et al., 1993. was soon overridden by the effect of venous pooling Žto increase plasma filtration in the dependent portion of the circulation. ŽRowell, 1993b; Greenleaf et al., 1979; Berne and Levy, 1981..

4.4. Vasodilation Õs. loss of Õasoconstriction The question of vasodilation vs. loss of vasoconstriction could not be directly addressed in the present study. Some recent studies, however, indicate that in young, otherwise healthy, adults who faint ŽMosqueda-Garcia et al., 1997; Shoemaker et al., 1999., and in patients with recurring vasovagal syncope ŽMorillo et al., 1997; Jardine et al., 1998., the presyncopal decline in blood pressure begins during the time in which muscle sympathetic activity ŽMSNA. and heart rate remain elevated. Conclusions from one of the patient studies emphasized that the onset of presyncope was not due to withdrawal of sympathetic activity but instead, decreases in MSNA and heart rate followed the drop in blood pressure during the last minute of presyncope ŽJardine et al., 1998.. Since indexes of sympathetic activity remain elevated at the onset of presyncopal symptoms in otherwise healthy subjects, vasodilation, via an as yet to be determined pathway, rather than loss of vasoconstrictor tone may be hypothesized as the source of the presyncopal decline in peripheral vascular resistance.

5. Study limitations We used impedance cardiography to assess changes in stroke volume, cardiac output and peripheral resistance as a function of time within the same subject and between subjects. Previous studies from our laboratory ŽGriffin, 1996., indicated that Group 1 end diastolic volumes, determined from ECG-gated, spin–echo, magnetic resonance images at rest and in response to y35 mm Hg LBNP,

correlated significantly with impedance Ž p - 0.004. and with echocardiographic Ž p - 0.03. determinations of end diastolic volumes. Volume calculations from impedance measurements gave consistently larger, and calculations from echo measurements gave smaller, heart volumes than did the MRI calculations. The large number of subjects Ž56. of the present study provides further evidence that the effects reported were not due to random influences. For these reasons we are confident that the lower peripheral resistance reported for syncopal subjects from each group is reliable, but we also feel this difference should be verified using other modalities of cardiac output measurement. This study involved primarily young subjects, all of whom were healthy. Autonomic influences change with age, exercise and disease so further studies, particularly in older age groups, should be carried out. At the lowest level of stress, Group 1, all five syncopal subjects were women. However, in Groups 2 and 3, the numbers of syncopal subjects were equally distributed between men and women. In Group 3, the syncopal men had a greater increase of epinephrine per unit time compared to nonsyncopal men, syncopal and nonsyncopal women. For Groups 2 and 3, syncopal men had a greater increase in their epinephrine to norepinephrine ratio than did nonsyncopal men, syncopal women or nonsyncopal women. The overnight stay in the CRC appeared to have had an effect on resting levels of both hemodynamic and hormonal variables. When resting values from Groups 1 and 3 were compared to those of Group 2 Žno overnight stay., epinephrine, norepinephrine, peripheral resistance and blood pressure were significantly lower in subjects ŽGroups 1 and 3. whose activity and diet was controlled by being admitted to the study on the night before, rather than the day of study. In addition, Groups 1 and 3 subjects ate a small meal but the low levels of pancreatic polypeptide indicated that the effect of this meal had dissipated prior to data collection. These subjects also carried out very light exercise to simulate ordinary activity, but again, the low levels of catecholamines prior to LBNP indicted they were resting in a non-stressful environment. Keeping these limitations in mind, this study demonstrated that syncope in otherwise healthy young people resulted from vasodilation that was associated with an imbalance in the kinetics of increases in plasma norepinephrine and epinephrine, increased sequestration of hemoconcentrated blood and a compromised ability to increase plasma renin activity, all of which may provide afferent information leading to the onset of syncope. Our study further suggests that epinephrine might play a multifactorial role in the development of syncope: epinephrinestimulated myocardial contractility andror venoconstriction could evoke a reflex Že.g. Bezhold–Jarish. that stimulates syncope, epinephrine could stimulate venoconstriction in an attempt to limit rapidly pooling venous

J.M. EÕans et al.r Autonomic Neuroscience: Basic and Clinical 93 (2001) 79–90

blood and epinephrine-associated vasodilation could increase plasma filtration.

Acknowledgements We would like to acknowledge the contributions of Dr. David C. Randall of the Department of Physiology, Ms. Tina Julian and Mr. Michael Stenger, Center for Biomedical Engineering, Dr. R.J. Kryscio of the Department of Statistics and the staff of the GCRC at the University of Kentucky. This research was supported by NASA EPSCoR WKU 522611 and NIH RR02602 and NIH RR00827.

References Abe, H., Kobayashi, Nakashima, Y., Izumi, F., Kuroiwa, A., 1995. Plasma catecholamines and cyclic AMP response during head-up tilt test in patients with neurocardiogenic Žvasodepressor. syncope. Pacing Clin. Electrophysiol. 18 Ž7., 1419–1426. Abi-Samra, F., Maloney, J.D., Fouad-Tarazi, F.M., Castle, L.W., 1988. The usefulness of head-up tilt testing and hemodynamic investigations in the workup of syncope of unknown origin. PACE 11, 1202–1214. Aratow, M., Fortney, S.M., Watenpaugh, D.E., Crenshaw, A.G., Hargens, A.R., 1993. Transcapillary fluid responses to lower body negative pressure. J. Appl. Physiol. 74 Ž6., 2763–2770. Berne, R.M., Levy, M.N., 1981. The peripheral circulation and its control. In: Berne, R.M., Levy, M.N. ŽEds.., Cardiovascular Physiology. 4th edn. CV Mosby, St. Louis, pp. 132–135. Bhargava, B., Chandra, S., Kacker, V., Gupta, Y.K., Seth, S.D., Wasir, H.S., 1996. Changes in circulatory biogenic amines during head-up tilt testing in neurocardiogenic syncope. Indian Heart J. 48 Ž6., 659–662. Blauw, G.J., Westendorp, R.G.J., Simons, M., Chang, P.C., Frolich, M., Meinders, A.E., 1995. Beta adrenergic receptors contribute to hypoxia induced vasodilation in man. Br. J. Clin. Pharmacol. 40, 453–458. Brown, C.M., Hainsworth, R., 1999. Assessment of capillary fluid shifts during orthostatic stress in normal subjects and subjects with orthostatic intolerance. Clin. Auton. Res. 9, 69–73. Cox, M.M., Perlman, B.A., Mayor, M.R., Silberstein, T.A., Levin, E., Pringle, L., Castellanos, A., Myerburg, R.J., 1995. Acute and longterm beta-adrenergic blockade for patients with neurocardiogenic syncope. J. Am. Coll. Cardiol. 26 Ž5., 1293–1298. Dill, D.B., Costill, D.L., 1974. Calculation of percentage changes in volumes of blood, plasma and red cells in dehydration. J. Appl. Physiol. 37 Ž2., 247–248. Fritsch-Yelle, J.M., Whitson, P.A., Bondar, R.L., Brown, T.E., 1996. Subnormal norepinephrine release relates to presyncope in astronauts after spaceflight. J. Appl. Physiol. 81 Ž5., 2134–2141. Greenleaf, J.E., Convertino, V.A., Mangseth, G.R., 1979. Plasma volume during stress in man: osmolality and red cell volume. J. Appl. Physiol.: Respir. Environ. Exercise Physiol. 47 Ž5., 1031–1038. Griffin, A.K., 1996. Lower body negative pressure-induced changes in the cardiovascular system assessed by magnetic resonance imaging. Master’s Thesis, University of Kentucky, 91 pp. Grubb, B.P., Temesy-Armos, P., Hahn, H., Elliott, L., 1991. Utility of upright tilt-table testing in the evaluation and management of syncope of unknown origin. Am. J. Med. 9, 6–10. Hinghofer-Szalkay, H., Konig, E.M., Sauseng-Fellegger, G., Zambo-Polz, C., 1992. Biphasic blood volume changes with lower body suction in humans. Am. J. Physiol.: Heart Circ. Physiol. 32 Ž263., H1270– H1275.

89

Jacob, G., Ertl, A.C., Shannon, J.R., Furlan, R., Robertson, R.M., Robertson, D., 1998. Effect of standing on neurohormonal responses and plasma volume in healthy subjects. J. Appl. Physiol. 84 Ž3., 914–921. Jardine, D.L., Ikram, H., Frampton, C.M., Frethey, R., Bennett, S.I., Crozier, I.G., 1998. Autonomic control of vasovagal syncope. Am. J. Physiol.: Heart Circ. Physiol. 43 Ž274., H2110–H2115. Kennedy, B., Ziegler, M.G., 1990. A more sensitive and specific radioenzymatic assay for catecholamines. Life Sci. 47, 2143–2153. Kennedy, B., Shannahof-Khalsa, B., Ziegler, M.G., 1994. Plasma norepinephrine variations correlate with peripheral vascular resistance in resting humans. Am. J. Physiol.: Heart Circ. Physiol. 45 Ž266., H435–H439. Klingenheben, T., Kalusche, D., Li, Y.G., Scheopperl, M., Hohnloser, S.H., 1996. Changes in plasma epinephrine concentration and heart rate during head-up tilt testing in patients with neurocardiogenic syncope: correlation with successful therapy with beta-receptor antagonists. J. Cardiovasc. Electrophysiol. 7 Ž9., 802–808. Linzer, M., Yang, E.H., Estes, N.A.M., Wang, P., Vorperian, V.R., Kapoor, W.N., 1997. Diagnosing syncope: Part 2. Unexplained syncope. Ann. Intern. Med. 127, 76–86. Lundvall, J., Bjerkhoel, P., 1995. Pronounced and rapid plasma volume reduction upon quiet standing as revealed by a novel approach to the determination of the intravascular volume change. Acta Physiol. Scand. 154, 131–142. Manolis, A.S., Linzer, M., Salem, D., Estes, N.A.M., 1990. Syncope: current diagnostic evaluation and management. Ann. Intern. Med. 112, 850–863. Morillo, C.A., Eckberg, D.L., Ellenbogen, K.A., Beightol, L.A., Hoag, J.B., Tahvanainen, K.U.O., Kuusela, T.A., Diedrich, A.M., 1997. Vagal and sympathetic mechanisms in patients with orthostatic vasovagal syncope. Circulation 96, 2509–2513. Mosqueda-Garcia, R., Furlan, R., Fernandez-Violante, R., Desai, T., Snell, M., Jarai, Z., Ananthram, V., Robertson, R.M., Robertson, D.L., 1997. Sympathetic and baroreceptor function in neurally mediated syncope evoked by tilt. J. Clin. Invest. 99 Ž11., 2736–2744. Natale, A., Newby, K.H., Dhala, A., Akhtar, M., Sra, J., 1996. Response to beta blockers in patients with neurocardiogenic syncope: how to predict beneficial effects. J. Cardiovasc. Electrophysiol. 7 Ž12., 1154– 1158. Perry, J.C., Garson, A., 1991. The child with recurrent syncope: autonomic function testing and b-adrenergic hypersensitivity. JACC 17 Ž5., 1168–1171. Purcell, J.A., 1992. Provoking vasodepresor syncopy with head-up tilt-table testing. Progr. Cardiovasc. Nursing 7 Ž3., 15–19. Robertson, D., Goldberg, M.R., Onrot, J., Hollister, A.S., Wiley, R., Thompson, J.G., Robertson, R.M., 1986. Isolated failure of autonomic noradrenergic transmission. N. Engl. J. Med. 314, 1494–1497. Rowell, L.B., 1993a. Passive effects of gravity. In: Rowell, L.B. ŽEd.., Human Cardiovascular Control. Oxford Univ. Press, New York, pp. 3–36. Rowell, L.B., 1993b. Orthostatic intolerance. In: Rowell, L.B. ŽEd.., Human Cardiovascular Control. Oxford Univ. Press, New York, pp. 118–161. Schwartz, T.W., 1983. Pancreatic polypeptide: a hormone under vagal control. Gastroenterology 85, 1411–1425. Shoemaker, J.K., Hogeman, C.S., Sinoway, L.I., 1999. Contributions of MSNA and stroke volume to orthostatic intolerance following bed rest. Am. J. Physiol.: Regul., Integr. Comp. Physiol. 46 Ž277., R1084–R1090. Slotwiner, D.J., Stein, K.M., Lippman, N., Markowitz, S.M., Lerman, B.B., 1997. Response of neurocardiac syncope to beta-blocker therapy: interaction between age and parasympathetic tone. Pacing Clin. Electrophysiol. 20 Ž3.2., 810–814. Sra, J.S., Murthy, V., Natale, A., Jazayeri, M.R., Dhala, A., Deshpande, S., Sheth, M., Akhtar, M., 1994. Circulatory and cateolamine changes during head-up tilt testing in neurocardiogenic Žvasovagal. syncope. Am. J. Cardiol. 73, 33–37.

90

J.M. EÕans et al.r Autonomic Neuroscience: Basic and Clinical 93 (2001) 79–90

Van Beaumont, W., 1972. Evaluation of hemoconcentration from hematocrit measurements. J. Appl. Physiol. 32, 712–713. Westendorp, R.G.J., Blauw, G.J., Frolich, M., Simons, R., 1997. Hypoxic syncope. Aviat., Space Environ. Med. 68, 410–414. Ziegler, M.G., Lake, C.R., 1985. Noradrenergic responses to postural hypotension: Implications for therapy. In: Lake, C.R., Ziegler, M.G.

ŽEds.., The Catecholamines in Psychiatric and Neurologic Disorders. Butterworth Publishers, Woburn, pp. 121–136. Ziegler, M.G., Echon, C., Wilner, K.D., Specho, P., Lake, C.R., McCutchen, J.A., 1986. Sympathetic nervous withdrawal in the vasodepressor Žvasovagal. reaction. J. Auton. Nerv. Syst. 17 Ž4., 273–278.