Control of cardiac output in essential hypertension

STUDIES IN HYPERTENSION

Control of Cardiac Output in Essential Hypertension

MICHEL E. SAFAR, MD NGUYEN P. CHAU YVES A. WEISS, MD GiRARD M. LONDON, MD PAUL L. MILLIEZ, MD Paris, France

This work was performed by the Hemodynamic Laboratory (Prof. Ag. Safar) of the Hypertension Research Center, Department of Prof. Milliet, Hospital Broussais, Paris and U.E.R. Saint-Quentin (Prof. Chau), Universite de Picardie, France. This study was supported by a grant-in-aid from the fnsfitut de la Sante et de la Recherche Medicale (INSERT), Paris, France, the Association pour Wtilisation du Rein Artificiel (AURA), Paris. France; and the Delegation Generate a la Recherche Scientifique et Technique (DGRST), Paris, France. Manuscript received December 29, 1975, accepted February 18, 1976. Address for reprints: Prof. Ag. Michel E. Safar, Hopital Broussais, 96 rue Didot, 75674 Paris Cedex 14, France.

332

September 1976

Cardiac and renal hemodynamics and cardiopulmonary and total blood volume were determined in 202 men, 101 with normotenslon and 101 of the same age with chronic essential hypertension, normal renal functffn and balanced sodium intake and urinary output. Cardiac output was identical in the two groups, whereas blood pressure and total peripheral resistance were significantly different. The two groups exhibited strong differences in the correlation study: (1) Correlations of blood pressure with, respectively, heart rate, cardiopulmonary blood volume and total blood volume were significant in the normotensive group but not in the hypertensive group. (2) Correlations of cardiac output with, respectively, heart rate, cardiopulmonary blood volume and total blood volume were significant in both groups. (3) Correlations of renal blood flow with, respectively, cardiac output, blood pressure and total blood volume were significant in the hypertensive group but not in the normotensive group. This study provides evidence that: (1) the volume and neural control of blood pressure are disrupted in hypertension whereas control of cardiac output is maintained; and (2) adaptive mechanisms involving renal function are necessary to the maintenance of normal cardiac output in patients with essential hypertension.

Since total peripheral resistance has not been directly measured, the respective roles of cardiac output and resistance in the mechanism of hypertension have remained controversial. In most types of hypertension, total peripheral resistance is greatly increased whereas cardiac output is nearly normal. These findings have usually been thought to indicate that hypertension is caused by some primary factor that increases resistance without changing cardiac output. More recently, on the basis of animal experiments and computer analysis,l-” it has been shown that a primary increase in cardiac output can cause increased total peripheral resistance and therefore increased blood pressure. This theory assumes that the maintenance of a normal cardiac outnut in established hypertension requires important adaptive mechanisms in the systemic circuiation.4 That regulation of arterial pressure is an essential ingredient in regulation of cardiac output is easy to demonstrate in animal experiments in which precise time-dependent relations of hemodynamic variables can be determined and compared. However, such data are difficult to establish in man. Clinical hemodynamic studies are performed either in the steady state or in unstable conditions involving only short-term control of blood pressure. Therefore, the mechanism of the normal cardiac output in essential hypertension is still unknown. The purpose of this study was to investigate the adaptive mechanisms that explain (or are related to) the normal cardiac output in patients with chronic essential hypertension.

The American Journal of CARDIOLOGY

Volume 36

CARDIAC OUTPUT IN ESSENTIAL HYPERTENSION-SAFAR

Material and Methods

TABLE

Selection of subjects: The study group was composed of 202 men, 101 with normal blood pressure and 101 with hypertension. All subjects were 20 to 40 years old (mean fl standard error of the mean 28.5 f 1 years). The patients were untreated or had discontinued treatment at least 4 weeks before the study. They were hospitalized for 6 days and given a 110 mEq/day sodium diet. During hospitalization, on the basis of weight, sodium intake and urinary output, sodium balance was considered to be in the steady state. Extensive investigations included determination of serum and urinary electrolytes, catecholamines and endogenous creatinine clearance levels and timed intravenous urography with wash-out test or renal arteriography, or both. The 101 hypertensive patients were diagnosed as having essential hypertension. None had renal, cardiac or neurologic involvement. The mean endogenous creatinine clearance of the overall population was 97 f 3 mlemin-l-1.73 rnmz.The protocols were approved by INSERM (Institut National de la Santi et de la Recherche MBdicale). Informed consent was obtained from the patients after a detailed description of the procedure. Hemodynamic determinations: On the 3rd hospital day, hemodynamic studies were performed after the patients had fasted overnight. No premeditation was administered. Under local procaine anesthesia, polyethylene catheters were in-

I

Hemodynamic Characteristics of the Two Groups (mean + standard error of the mean)

coefficients @values) in the two groups. The r values are presented with both unnormalized and weight-normalized data (see Appendix). CO = cardiac output; DAP = diastolic arterial pressure: HR = heat-l rate; MAP = mean arterial pressure; SAP = systolic arterial pressure; SV = stroke volume; TBV = total blood volume; TPR = total peripheral resistance.

Normotensive Group (I 01 patients)

Variable

28 + 71 i 172+ 1385 76+ 97 ? 76+

Age (Yr) Weight (kg) Height (cm) SAP (mm Hg) DAP (mm Hg) MAP (mm Hg) HR (beats/min) co (ml/min) (ml/kg) SV (ml/kg) TPR (arbitrary unit. kg) TBV (ml) (ml/kg)

7173 ? 103? 1.4 r 990 ?

Hypertensive Group (101 patients)

1 1 1 2 1 2 1

29 i 75? 174i 176 t 107+ 132 -r 77 +

165 2 0.03 23

1 I** 1 3*** 2”“” 2”“” 1

7166* 157 97 ? 3 1.3 + 0.02 1453 f 44”“”

5245 ? 132 75 + 1

5355 + 127 72? I**

* = P
SAP

FIGURE 1. Correlation

ET AL.

DAP

MAP

NORMOTENSIVE

GROUP

(n = 101)

HYPERTENSIVE

GROUP

( n:

HR

CO

TPR

TBV

101)

The correlation exisk in the hypertensive group but not in the normotensive group.

September 1976

SV

The American Journal of CARDIOLOGY

The correlotlonexists in both groups.

Volume 36

333

CARDIAC OUTPUT IN ESSENTIAL HYF’ERTENSION-&AFAR

TABLE

ET AL.

I1

Correlations

Between

Cardiopulmona~

Blood Volume and Renal Blood Flow With Other Variables in the Two Populations

Cardiopulmonary

Correlations

With

Age Weight Height SAP DAP MAP z: %R TBV

Normotensive Group (no. = 28) (mean value + SEM: 14+ 11

Blood Volume

(ml/kg)

Renal Blood Flow fmlimin

Hypertensive Group (no. = 33) (mean value + SEM: 15% 1)

Normotensive Group (no. = 31) (mean value f. SEM: 26% 1)

Hypertensive Group (no. = 48) (mean value +_SEM: 241 1)

-0.20 -0.15 0.16 -0.26 -0.35” -0.35” 0.06 0.54** 0.40” -0.54** 0.32

0.07 -0.02 0.00 -0.23 -0.04 -0.10 0.12 -0.1 1 -0.20 0.03 0.00

-0.33” -0.15 0.20 -0.24 -0.41X” -0.37” * 0.06 0.50*** 0.32” -0.56”“’ 0.39””

0.07 -0.33 -0.18 0.42” 0.49** 0.48’* 0.15 0.61*** 0.59** -0.42” 0.32

*P <0.005; **P
traduced into the right bra&ii artery and median antecubital vein and under fluoroscopic control were advanced, respectively, into the aortic arch and the main pulmonary artery. The arterial catheter was then placed in the aortic root immediately distal to the aortic valves for intraarterial pressure measurements, and blood pressure, cardiac output and cardiop~monary blood volume were measured as previously described elsewhere.5 Total peripheral resistance (TPR) was calculated according to the formula: TPR (arbitrary unit) = MAP/CO X 1,000, where MAP is the mean arterial pressure and CO is cardiac output. Cardiopulmonary blood volume (CPBV) was expressed as the volume between the main pulmonary artery (PA) and the tip of the arterial catheter(Ar). It was calculated as follows: CPBV = CO x TI’M, T-l-M being the mean transit time in seconds between the pulmonary artery and the catheter tip.6 Total blood volume was measured before the hemodynamic study by the isotope dilution method, using radioiodinated albumin as previously described.5@ Renal blood flow was estimated with the iodine-131 hippuran clearance technique and determined with use of the single injection method and the one compartment mamillary

system modeL7 Normalization of weight was used to express blood flow, blood volume and stroke volume (see Appendix).

Classification of patients: The main purpose of the study was to investigate control ofcardiac output at different levels of blood pressure. Therefore, patients were classified into two groups according to the level of blood pressure measured during the hemodynamic study on the 3rd hospital day. In the normotensive group (101 patients), diastolic blood pressure was 90 mm Hg or less and systolic pressure was 160 mm Hg or less. In the hypertensive group (101 patients), diastolic pressure was more than 90 mm Hg. Table I lists the main characteristics of the two groups. Age and cardiac output were identical in both groups; blood pressure and total peripheral resistance were significantly different. Statistical studies: Statistical calculation@ were performed for three purposes: (1) To point out the significant correlations between the main hemodynamic variables in the normotensive and hypertensive groups; (2) To compare the correlations between two given variables. The following three results were observed: (a) the correlation was significant in one group, but not in the other; (b) the correlation was significant in both groups, but at different levels; and (c) the 334

September 1976

par kg)

The Amerfcan Journal of CARDIOLDQY

group. Abbreviations

as in Table

I.

correlation was significant in both groups, at the same level. In the latter case, a regression analysis (Y = aX +b) was performed, and the values of a and b in the two groups were compared. (3) To interpret the results by constructing a systematic presentation of factors that interfere directly or indirectly in the regulation of cardiac output. Some differences between the two groups were interpreted as an adaptive mechanism possibly related to the maintenance of a normal cardiac output in the hypertensive patients.

Results normalization for body weight was used in Figure 1 and Tables I (right column) to III. Correlation coefficients (Figure 1 and Table II): The results in Figure 1 can be summarized as follows. By comparison with findings in the normotensive group, the following occurred in the hypertensive group: 1. These correlations were disrupted: (a) blood pressure (systolic, mean) with body weight; (b) blood pressure (systolic, diastolic, mean) with heart rate; (c) blood pressure (systolic) with cardiac output; (d) blood pressure (systolic, diastolic, mean) with blood volume; and (e) total peripheral resistance with heart rate. Table II showed similar results concerning the correlation between blood pressure and cardiop~monary blood volume. 2. These correlations appeared in the hypertensive group: (a) age with several hemodynamic variables (blood pressure and cardiac output, among others). 3. These correlations were maintained; (a) cardiac output (or stroke volume) with blood volume (or body weight); (b) cardiac output (or stroke volume) with heart rate; and (c) total peripheral resistance with blood volume. Table II summarizes the data on renal blood flow. In the normotensive group, there was no correlation between renal blood flow and the other hemodynamic variables. In contrast, in the hypertensive group, renal blood flow was correlated (1) negatively with blood pressure, and (2) positively with cardiac output and blood volume.

Volume 38

CARDIAC OUTPUT IN ESSENTIAL HYPERTENSION-SAFAR

tance-volume) relation.‘l This correlation demonstrates the existence of a volume-pressure adaptive mechanism.7 When blood pressure increases, volume decreases, and this factor contributes to maintenance of pressure within normal ranges. Such results reflect the long-term pressure-body fluid control mechanism described by Guyton and Coleman.3 The increase in pressure causes the kidney to excrete more water and salt, thus decreasing blood volume, cardiac output and, in turn, total peripheral resistance, and finally causing a decrease in pressure level. Disrupted relations in the hypertensive group (Fig. 2): The correlations between volume and neural control of cardiac output were maintained in the hypertensive group. In contrast, the interrelations between cardiac output and blood pressure were disrupted. No correlation existed between heart rate and blood pressure. In addition, the correlation between blood pressure

Regression coefficients (Table III): To get further information on the correlations that were significant in both groups, a regression analysis was made. Table III shows that the regression coefficients of cardiac output versus weight, heart rate and blood volume were similar in the two groups. In contrast, from the normotensive to the hypertensive group, the volume-resistance regression curve had two characteristics: (1) it was reset to the right, and (2) the slope was significantly more shallow. Comments Despite the large differences in blood pressure levels, the normotensive and hypertensive groups had the same cardiac output. These were striking differences in the correlations between hemodynamic variables: (1) Correlations with age existed in the hypertensive group but not in the normotensive group; (2) correlations with blood pressure existed in the normotensive group but not in the hypertensive group; and (3) correlations with cardiac output were maintained in both groups. Age is a well known factor in hypertension. Therefore, points (2) and (3) are the main subjects of our comments. Control of cardiac output in the normotensive group: As shown in Figure 2, the correlations of cardiac output with heart rate, cardiopulmonary and total blood volume are indications of neural and volume control of cardiac output.g However, our study points to subtle interrelations between cardiac output and blood pressure in normal conditions. Sympathetic activity interferes with regulation of both blood pressure and cardiac output, as demonstrated by the previously described correlation between heart rate and blood pressure.lO Additional evidence was provided in this study by the direct correlation between cardiopulmonary blood volume and blood pressure. More important is the volume control, which has been studied by the pressure-volume (or resis-

Normotensive

FIGURE 2. Control of cardiac output in the normotensive and hypertensive groups. CO = cardiac output; CPBV = cardiopulmonary blood volume; DAP = diastolic arterial pressure; HR = heart rate; RBF = renal blood flow; TBV = total blood volume. Solid line indicates a positive correlation: dashed line a negative correlation: wavy line a disrupted correlation.

ET AL.

TABLE III Regression

Y

Data in the Two Groups

X

Group

co

Weight

N H

co

HR

N H

co

TBV

TBV

TPR

Value

a Value i Standard Error up to 5%t

b Value +_ Standard Error up to 5%t

i 0.38 ? 0.42

160 f 27 153 f 31

0.51 0.34

0.87 i 0.29 0.63 k 0.34

36+ 22 48 ? 27

N H

0.30 0.39

0.57 + 0.35 1.07 ? 0.51

60 + 27 20 ? 37

N H

0.46 0.27

-0.39 -0.34

-0.81 -0.75

-0.02 -0.005

f 0.008 i 0.004”

97 k 9 802 6

*P <0.05. t According to Student’s table. H = hypertensive group (no. = 101 1; N = normotensive group (no. = 1011; Y = aX + b regression line; other abbreviations as in Table I.

group

Hypertensive

group

IRBF~

RBF El ---+ AL

B

A is the determinant of B,

A-

B

The determinant

f---+

September 1976

factor

is

unknown and can be a Yd factor.

The American Journal of CARDIOLOGY

Volume 36

335

CARDIAC OUTPUT IN ESSENTIAL HYPERTENSION-SAFAR ET AL.

and cardiopulmonary blood volume was negative. These observations agree with the reduced neurogenic activity previously emphasized in chronic essential hypertension.12 The most significant finding was the loss of the pressure (or resistance)-volume relation that contributed to the control of cardiac output in the normotensive group. The volume-resistance relation of hypertensive patients had two characteristics: (1) The curve was reset, so that, at each level of resistance, the blood volume was greater in hypertensive than in normotensive subjects; and (2) the slope was significantly decreased, indicating a reduced ability of the system to decrease the volume per unit of rise in resistance. Such a curve was described in patients with a steady state sodium balance. Therefore, the blood volume was sufficient to maintain the steady state of sodium balance and the normal cardiac output, but was too great to allow the pressure to return to normal values. The maintenance of such blood volume implied the existence of an underlying renal mechanism.13 Control of cardiac output in the hypertensive group (Fig. 2): That a renal mechanism is involved in the maintenance of the cardiac output level in hypertensive patients was demonstrated by the renal blood

flow correlation study. In the normotensive group, renal blood flow was not correlated with any hemodynamic variable. This finding was in agreement with the classic concept of autoregulation of blood flow in the normal renal circulation. In contrast, hypertensive patients had direct correlations of renal blood flow with blood volume and cardiac output. These observations point to the loss of the blood flow autoregulation and, consequently, to a renal defect. The negative correlation between blood pressure and renal blood flow does not contradict this concept. Tobian et a1.14 have demonstrated that lack of autoregulation is essential to provide a normal rate of flow in hypertensive kidneys perfused at elevated levels of pressure. In conclusion, our study points to the role of a functional renal abnormality in the maintenance of the normal cardiac output level in patients with essential hypertension.

Acknowledgment We thank Mrs. C. Pillet for technical assistance and Mrs. M. J. Eggers for secretarial work. We greatly appreciate having had authorized access to the computers of the Centre de Calcul de Physique Nucleaire (Facult6 des Sciences, Paris, France).

References 1. Ledingham JM, Cohen RD: Changes in the extracellular fluid volume and cardiac output during the development of experimental renal hypertension. Can Med Assoc J 90:292-298, 1964 2. Coleman TG, Guyton AC: Hypertension caused by salt loading in the dog: Ill. Onset transients of cardiac output and other circulatory variables. Circ Res 25:153-158, 1969 3. Guyton AC, Coleman TG: A quantitative analysis of the pathophysiology of hypertension. Circ Res 24:1-6, 1969 4. Pontryagin LS: Ordinary Differential Equations. London, Addison, Wesley Publishing Company, 1962, 200-213 5. Safar ME, Weiss YA, Levenson JA, et al: Hemodynamic study of 85 borderline hypertensive patients. Am J Cardiol 31:315-319, 1973 6. Safar ME, London GM, Weiss YA, et al: Vascular reactivity to norepinephrine and hemodynamic parameters in borderline hypertension. Am Heart J 89:480-486, 1975 7. Safar ME, London GM, Weiss YA, et al: Altered blood volume regulation in sustained essential hypertension: a hemodynamic study. Kidney Int 8:42-47, 1975 8. Croxton FW, Cowden DJ: Applied General Statistics. New York,

Prentice Hall, 1944, p 327, 347, 653 9. Guyton AC: Circulatory Physiology: Cardiac Output and its Regulation. Philadelphia, WB Saunders, 1963, p 322 10. Sive PH, Medalie JH, Kahn HA, et al: Distribution and multiple regression analysis of blood pressure in 10,000 Israeli men. Am J Epidemiol 93:317-327, 1971 11. Dustan HP, Tarazi RC, Bravo EL, et al: Plasma and extracellular fluid volumes in hypertension. Circ Res 32:Suppl 1:1-73-l-81, 1973 12. DeQuattro V, Miura Y: Neurogenic factors in human hypertension. Mechanism or myth? Am J Med 55:363-375, 1973 13. Guyton AC, Coleman TG, Cowley AW, et al: Arterial pressure regulation: overriding dominance of the kidneys in long-term regulation and in hypertension. Am J Med 52:584-594, 1972 14. Tobian L, Johnson MA, Lange J, et al: Effect of varying perfusion pressures on the output of sodium and renin and the vascular resistance in kidneys of rats with “post-salt” hypertension and Kyoto spontaneous hypertension. Circ Res 36, 37:Suppl l:l-162-l-171, 1975

APPENDIX The correlation coefficients (r values) could depend on the normalization procedure. Figure 1 summarizes the results determined by using either unnormalized or weight-normalized data. This figure suggests two comments: Comparison between unnormalized and weight-normalized data: In both cases, the r values were similar, except for the pressure-volume relation; the correlation existed only for normalized blood volume. This finding could be explained by the strong correlation between unnormalized blood volume and body weight. The negative pressure-volume relation became obvious only after elimination of the role of body weight. The pressure-volume relation existed also when data were normalized for body height.”

336

September 1976

The American Journal of CARDIOLOGY

Comparison between weight normalization and other types of normalization: Normalization for body weight, body surface area and height yielded similar r values. In addition, Figure 1 allowed an indirect comparison between the weight and the height normalization data: (1) The correlations of height with the unnormalized hemodynamic data (cardiac output, blood volume) were nut observed after normalization for weight, and (2) several correlations such as the unnormalized blood volume-cardiac output correlation existed even with a constant weight, but not with a constant height (partial correlation coefficient study). These results reflected the strong weight-height correlation and justified the normalization for weight.

Volume 36