- Email: [email protected]
Arterial vasodilation is not the cause of increased cardiac output in cirrhosis
GASTROENTEROLOGY 1992;102:1024-1029
Arterial Vasodilation Is Not the Cause of Increased Cardiac Output in Cirrhosis FREDERICK
W. LEWIS, OLIVIA ADAIR,
and WILLIAM G. RECTOR, Jr.
Divisions of Cardiology and Gastroenterology, Department of Medicine, Denver General Hospital, Denver, Colorado
A pathological state of arterial vasodilation has been postulated to cause the increased cardiac output commonly observed in cirrhosis. Further, suhsequent arterial underfilling has been proposed as the stimulus to sodium retention and ascites formation. Left ventricular size during the cycle of a cardiac contraction is predictably altered by a decrease in afterload. Specifically, increased systolic emptying should he observed. The relationship of left ventricular size during the cardiac cycle to systemic hemodynamic indices and urinary sodium retention was investigated in patients with alcoholic cirrhosis to test these hypotheses. Echocardiographic studies were performed on 24 male patients with alcoholic cirrhosis and compared with the results obtained in 10 age-matched male controls. Patients with cirrhosis had increased cardiac output and heart rate and decreased arterial pressure compared with normal subjects, confirming the presence of a hyperdynamic circulation. Patients with cirrhosis had enlarged left ventricular diameter at bothenddiastole (0.08 fO.Olvs.0.07'+0.007 cm/kg dry body wt; P < 0.001) and end systole (0.08 f 0.01 vs. 0.05+ 0.005cm/kg; P < 0.05). Left ventricular end-diastolic diameter was directly related to blood volume (r = 0.56, P < 0.005). No significant differences in cardiac output, arterial pressure, or systemic resistance were found between patients with and patients without ascites. Increased cardiac output in cirrhosis occurs in conjunction with an enlarged ventricle throughout the cardiac cycle. The increase in left ventricular end-systolic diameter indicates that diminished afterload is not responsible for the increase in cardiac output. As the diameter of the ventricle during diastolic filling correlates with vascular volume, cardiac output in cirrhosis may be primarily determined by an increase in vascular volume.
C
irculatory abnormalities observed in patients with cirrhosis include increased cardiac output and heart rate and diminished systemic resistance.14 A circulating vasodilator has been postulated to initiate this hyperdynamic circulation. In this proposal,
the circulation of patients with cirrhosis is analogous to the state of high cardiac output encountered in subjects with arteriovenous fistulae, severe anemia, Paget’s disease of bone, and beri-beri heart disease.44 Study of cardiac pressure-volume relationships by Braunwald, Sonnenblick, and others established the mechanisms of intrinsic cardiac compensation to changes in arterial resistance (“afterload”) and venous pressure (“preload”).‘,* A decrease in arterial resistance increases systolic ejection, and thus cardiac output, by allowing more complete emptying of the ventricle in systole. Conversely, increased diastolic filling of the left ventricle increases cardiac stroke volume by an increase in left ventricular enddiastolic volume. If the arterial vasodilation hypothesis is correct, patients with cirrhosis should show increased systolic emptying of the left ventricle as a result of diminished afterload. Arterial vasodilation has also been postulated to cause sodium retention in cirrhosis. According to this hypothesis, counterregulatory systems respond to the perceived “underfilling” of the arterial circulation caused by vasodilation5.’ Although the degree of circulatory abnormality in cirrhosis is believed to parallel the severity of liver dysfunction, the relationship between the hyperkinetic systemic circulation of cirrhosis and the occurrence of sodium retention is uncertain.“-13 Therefore, an additional aim of this study was to compare hemodynamic findings and left ventricular dimensions in patients with and without ascites in order to evaluate the validity of the postulated sequence of arterial vasodilation, sodium retention, and ascites formation. Materials and Methods Patients and Protocols The protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the Human Subjects Committee of the University of Colorado Health Sciences Center. All patients gave written in-
formed consent. 0 1992by the American Gastroenterological 0016-5065/92/$3.00
Association
CARDIAC OUTPUT IN CIRRHOSIS 1025
March 1992
This study was conducted as part of a larger investigation designed to evaluate factors potentially related to SOdium retention in patients with alcoholic liver disease. These factors include autonomic nervous system function,14 atria1 size and blood volume,*5 liver metabolic capacity and portal-systemic shunt fraction,16 plasma atria1 natriuretic factorI and glucagon concentration,17 and plasma renin activity. Aspects of this investigation not discussed here or in cited articles will be reported in future communications. The study group included 24 men with alcoholic liver disease. The diagnosis of alcoholic liver disease was made by a history of alcoholism, typical laboratory findings, and liver biopsy where indicated. Patients with a history of cardiac disease or with diminished left ventricular fractional shortening (~20%) on echocardiographic examination were excluded.” Ten male volunteers of similar ages (mean age, 39 + 7 years; range, 25-60 years) without cardiac or hepatic disease served as the control group. Repeated studies were performed after an interval of 3-26 months on 12 patients as a part of this longitudinal investigation. This study included 8 patients who remained without ascites for >6 months, 1 patient in whom ascites persisted for >6 months, 2 patients in whom ascites resolved before the second study, and 1 patient without ascites who developed sodium retention in conjunction with renewed alcohol intake. Each patient was admitted to the Adult Clinical Research Center of the University of Colorado Health Sciences Center for 5 days. Diuretics had been discontinued at least 10 days before admission to eliminate alterations in circulatory features or vascular volume attributable to such treatments. A dietitian interviewed each patient, and average sodium intake at home was determined. The patients were placed on a diet containing 200 mEq of sodium daily, and complete 24-hour urine collections for sodium and creatinine determination were performed on days 2-4. The presence of sodium retention and ascites formation was established by physical examination and by noting weight change and sodium excretion during the study period. Plasma volume was measured by isotope dilution using’ZSI-albumin and blood volume calculated as Plasma Volume/[(l - Hematocrit) (0.82)].” On day 5, supine blood pressure and heart rate were measured using an automated sphygmomanometer (Dinamap-Vital Signs Monitor 8100; Critikon Inc., Tampa, FL) with the patient in the fasting state and after 30 minutes’ rest in a quiet, darkened room. Two-dimensional echocardiography, color-flow Doppler imaging, and directed pulsed-wave Doppler imaging were then performed as described, using a Hewlett Packard Model 77020 AC phasedarray system (Palo Alto, CA), and stored on VHS video tape for subsequent analysis. “JO Images were acquired using a 2.5-mHz or 3.5-mHz transducer while the patient was in a supine or left lateral position. A parasternal long-axis view, with M-mode directed recordings of the aortic valve, aortic root, and left ventricular outflow tract, and left ventricle diastolic and systolic images were obtained. Apical four-chamber and apical long-axis two-chamber views and color-flow-directed pulsed Doppler of the left ventricular outflow tract were recorded with electrocardio-
graphic (ECG) display. Left ventricular dimensions (Mmode), flow rates (pulsed-wave Doppler), ejection time, and heart rate were analyzed using Hewlett Packard online computer software. Fractional shortening was calculated as [lo0 X (Diastolic Diameter - Systolic Diameter)/ Diastolic Diameter]. 20*22 Left ventricular outflow tract area was calculated as z(D/~)~, D being the left ventricular outflow tract diameter. Stroke volume was calculated using the color-flow-directed pulsed-wave Doppler by the formula (Left Ventricular Outflow Tract Area) X (Mean Blood Velocity Across the Aortic Valve) X (Ejection Time). Cardiac output was calculated as (Stroke Volume) X (Heart Rate). We have previously shown that this method of determining cardiac output in cirrhosis gives results similar to those of thermodilution methodology.23 Cirrhosis is associated with abnormalities in body muscle and fat mass that progress with the severity of liver disease, as well as alterations in total body water and sodium content. Consequently, the body surface area calculated from height and weight may not be a valid method of correcting for differences in body size in this population. Therefore, comparisons of cardiac dimension and volume were made by expressing these data per kilogram of dry body mass.23 Dry body mass was taken as actual body mass in patients without ascites and as body mass after subsequent complete diuresis in patients with ascites. The systemic vascular resistance index was used to obtain meaningful units in which to express resistance, and was calculated as (Mean Arterial Pressure X 80)/Cardiac Index. Statistics Data are expressed as mean ? SD. Data were analyzed using the SAS statistical program (Cary, NC). An unpaired t test was used to make group comparisons between patients with cirrhosis and normal subjects and between patients with cirrhosis with and without ascites. Pearson correlation coefficients were determined using a linear regression model. All statistical calculations were corrected for the effect of multiple observations in 12 of the 24 patients as follows: repeat studies on patients who gained or lost ascites (n = 3) were treated as separate observations, with one set in the “ascites” group and one set in the “no ascites” group; studies in each patient undergoing two studies without a change in ascites status (n = 9) were treated as a single observation, and the mean of the two values was used for each calculation. As a result, 27 observations are reported in Tables l-3. For completeness, each individual measurement (n = 36) is represented separately in Figures 1 and 2. A P value of 0.05 was accepted as statistically significant in all calculations.
Results Patients with cirrhosis had hyperkinetic circulatory characteristics as manifested by increased cardiac output and heart rate compared with control subjects (Table 2). The systemic resistance index tended to be lower in patients with cirrhosis, but this difference did not reach statistical significance. En-
1026
LEWIS ET AL.
Table I. Patient
GASTROENTEROLOGY Vol. 102, No. 3
Table 3. Hemodynamic Features, Blood Volume, and Left Ventricular Dimensions
Characteristics No ascites
n 46 34 (2.0 32 (3.2
Age (yr) Serum bilirubin concentration [wmol/L (mg/dL)l Serum albumin concentration [g/L (g/dL)l Aspartate aminotransferase W/L) Alanine aminotransferase
17 + 10 f 38 f 2.2) f 7 f 0.7)
94 f
W/L) Body weight (kg) Dry mass (kg) Sodium intake at home (mEq/doy) Sodium excretion (mEq/72 h)
Ascites
’
45 74 (4.3 22 (2.2
10 f 10 + 92 f 5.4) f 3 f 0.3)
104 rfr64
100
45 * 37 72 f 12 72+12
a
52 + 26 69 f 10 64 + 6
325 + 150 438+ 123
’
182 f 122 71 z!I92
Cirrhosis, no ascites n Cardiac output (mL.min-’ . kg-‘) Heart rate (beats/min) Mean arterial pressure (mm Hg) Systemic resistance index (dyne. s *crnm5. m-‘) Blood volume (mL ‘kg-‘) Left ventricular diameter End diastole (cm akg-‘) End systole (cm -kg-‘) Stroke volume (mL . kg-‘) Fractional shortening (%)
Cirrhosis, ascites
17
10
118 f 57 82 f 12
129 + 43 87 + 12
93 f 8
94 f 6
1990 + 800
1760 + 600 81 + 10
75 + 15 0.08 f 0.01 0.06 f 0.01 1.5 + 0.6 29f6
0.08 + 0.01 0.05 + 0.01 a
1.5 -+ 0.4 36 _+ 10
“P < 0.05.
“P < 0.05.
larged left ventricular diameter, both at end diastole and at end systole, and increased fractional shortening were also seen in patients with cirrhosis. When patients with and without sodium retention were compared, the only significant difference was increased fractional shortening in those with ascites (Table 3). No significant differences were detected in cardiac output, heart rate, arterial pressure, left ventricular dimensions, or blood volume, although trends toward increased cardiac output and enlarged left ventricle in diastole in the group with ascites were present. Left ventricular end-diastolic diameter was directly related to cardiac output (Figure 1A). In addition, a significant linear correlation between blood volume and end-diastolic diameter of the left ventricle (Figure 2) was seen. However, increased cardiac
output was not associated with increased systolic emptying. Indeed, the reverse was observed (Figure 1B). Similar relationships were observed between stroke volume and left ventricular dimensions (Figure 1C and D). Arterial hypotension was not associated with increased ventricular emptying. Rather, inverse linear correlations were detected between mean arterial pressure and both end-diastolic and end-systolic ventricular diameters (Figure 1E and F). No significant relationships were observed between either ventricular dimension and systemic resistance (r = -0.11 andr = -0.27 for end-diastolic diameter and end-systolic diameter, respectively; both P = NS).
Table 2. Hemodynamic Dimensions
n Cardiac output (mL - min-’ -kg-‘) Heart rate (beats/min) Mean arterial pressure (mm Hg) Systemic resistance index (dyne. s. crnm5. m-“) Left ventricular diameter End diastole (cm. kg-‘) End systole (cm - kg-‘) Stroke volume (mL *kg-‘) Fractional shortening (%) OP < 0.05. bP < 0.005.
Features
and Left Ventricular Control
Cirrhosis
10
27
94+25 67 f 7
a b
122 + 51 84 + 12
95 f 8
93 + 7
2140 f 400
1900 f 700
0.07 * 0.007 0.05 f 0.005
b o
1.4 f 0.3
26 f 4
0.08 + 0.01 0.06 + 0.01 1.5 + 0.5
b
32 + 8
Discussion This is the first investigation to examine the relationship of left ventricular size to systemic hemodynamics in cirrhosis, although several previous studies have measured left ventricular dimensions in conjunction with evaluation of cardiac function in cirrhosis or alcoholism. These reports are consistent with our finding of enlarged left ventricular diameter at both end diastole and end systole in patients with cirrhosis.24-27 The principle aim of this study was to test the hypothesis that increased cardiac output in cirrhosis is attributable to arterial underfilling. Our data do not support this mechanism. The observed enlargement of the left ventricle throughout the cardiac cycle suggests that the increase in cardiac output is effected by increased ventricular filling during diastole rather than by increased ventricular emptying during systole. The concept that increased ventricular filling underlies the increase in cardiac output in cirrhosis is supported by several further observa-
CARDIAC
March 1992
OUTPUT
r=.53,
I
A
Figure 1. Direct linear correlations between end-diastolic and end-systolic left ventricular dimensions and cardiac output (A and B) and stroke volume (C and D); inverse correlation between left ventricular diameters and arterial pressure (E and F).
.41,
END SYSTOLE
LEFT VENTRICULAR
Cons. Direct correlations were observed between both cardiac output and stroke volume and left ventricular diameter, both at end diastole and at end systole. Furthermore, an inverse correlation was observed between mean arterial pressure and both left ventricular dimensions, and no relationship was discernable between systemic vascular resistance and either ventricular dimension. The opposite, a direct relationship between ventricular diameter and both systemic resistance and arterial pressure, should have been observed if arterial underfilling were the cause of the observed increase in cardiac output. Finally, the significant linear correlation documented between blood volume and left ventricular end-diastolic diameter, similar to the relationship we have previously reported between left atria1 volume and
PLO05
pco5
p<.Ol
1027
b
A.-
,.
r=.51,
r=-
IN CIRRHOSIS
,*
,
A ’
A r=--40,
pco5 F
END DIASTOLE
DIAMETER
(cm/kg)
blood vo1ume,‘5 is reasonable evidence that ventricular enlargement probably results from increased intravascular volume. Ventricular enlargement can, of course, result from cardiac failure. It has been suggested that some patients with cirrhosis, despite showing normal resting cardiac function, have a cardiomyopathy that results in subnormal responses to cardiac stress.6~24~28-30 It is unlikely that this postulated condition influenced the findings of this study. The patients in the current study did not have any history, signs, or symptoms of cardiac disease. Moreover, there was by definition no echocardiographic evidence of cardiac dysfunction because patients with abnormal left ventricular function had been excluded. Second, previously noted circulatory abnormalities28-30 have not
1028
_,
LEWIS ET AL.
,D1
cn
GASTROENTEROLOGY Vol. 102, No. 3
r=.56, pc.005 v =no ascites
v V
*=ascites
bl
I;!
!
vv 0
60--
V W
W
d
50, 0.050
0.060
0.070
0.080
LEFT VENTRICULAR AT END-DIASTOLE
0.090
0.100
DIAMETER
rial pressure and filling to decrease, initiating renal sodium retention. We have previously shown that hepatic metabolic capacity and portal-systemic shunting are unrelated to the systemic hemodynamic features of patients with cirrhosis.” An additional aim of the present investigation was to relate indices of arterial filling to the presence of sodium retention. We have shown that cardiac output is not an indicator of arterial filling. No significant differences were found between other commonly used indices of arterial filling, including arterial blood pressure and systemic vascular resistance, in patients with and without ascites. These findings do not support the mechanism of sodium retention in cirrhosis proposed by the arterial vasodilation hypothesis.5
(cm/kg)
Figure 2. Blood volume is directly related to left ventricular end-diastolic diameter.
been proved to result from cardiomyopathy. Circulatory or autonomic abnormalities associated with cirrhosis may instead be responsible. Indeed, patients in the current study showed increased cardiac output compared with controls, rather than the decreased output that would be expected if intrinsic cardiac disease were present. Finally, within the patient group, a direct correlation was observed between both cardiac output and stroke volume and left ventricular size. This relationship would be an unlikely result of cardiomyopathy. Increased blood volume in patients with cirrhosis and ascites is a well-recognized feature of this condition.lg In the current study, blood volume and left ventricular size tended to be greater in patients with ascites than in patients without ascites, although the differences did not reach statistical significance. More importantly, our data indicate that volume expansion is probably responsible, through increased diastolic filling, for the increase in cardiac output observed in cirrhosis. As stated, a close, direct relationship was observed between blood volume and the diameter of the left ventricle in diastole. Longitudinal study of an experimental model of cirrhosis in dogs supports the postulate that sodium retention precedes hemodynamic changes and ascites formation.31 According to this mechanism, increased cardiac output is an effect of volume expansion rather than part of its cause. Even if arterial underfillingis not directly responsible for the increase in cardiac output observed in cirrhosis, it may still be the stimulus for sodium retention. According to this explanation, increased systemic availability of a vasodilator as a result of liver dysfunction or portal-systemic shunting causes arte-
References 1. Kowalski
HJ, Abelmann WH. The cardiac output at rest in Laennec’s cirrhosis. J Clin Invest 1953;32:1025-1033. 2. Murray JF, Dawson AM, Sherlock S. Circulatory changes in chronic liver disease. Am J Med 1958;24:358-367. 3. Kontos HA, Shapiro W, Mauck HP, Patterson JL. General and regional circulatory alterations in cirrhosis of the liver. Am J Med 1964;37:526-535. 4. Sherlock S. Vasodilation associated with hepatocellular disease: relation to functional organ failure. Gut 1990;31:365367. 5. Schrier RW. Body fluid volume regulation in health and disease: a unifying hypothesis. Ann Intern Med 1990;113:155159. 6. Lee SS. Cardiac abnormalities in liver cirrhosis. West J Med 1989;151:530-535. 7. Grossman W, Braunwald E. High-cardiac output states. In: Braunwald E, ed. Heart disease. Philadelphia: Saunders, 1988:778-792. 8. Sonnenblick EH, Schlant RC. Normal physiology of the cardiovascular system. In: Hurst JW, ed. The heart. 6th ed. New York: McGraw-Hill, 1986:37-72. 9. Shrier RW, Arroyo V, Bernardi M, Epstein M, Henriksen JH, Rodes J. Peripheral arterial vasodilatation hypothesis: a proposal for the initiation of renal sodium and water retention in cirrhosis. Hepatology 1988;8:1151-1157. 10. Siegel JH, Goldwyn RM, Farrell EJ, Gallin P, Friedman HP. Hyperdynamic states and the physiologic determinants of survival. Arch Surg 1974;108:282-292. 11. Braillon A, Cales P, Valla D, Gaudy D, Geoffery P, Lebrec D. Influence of the degree of liver failure on systemic and splanchnic haemodynamics and on response to propranolol in patients with cirrhosis. Gut 1986;27:1204-1209. 12. Moreau R, Lee SS, Soupison T, Roche-Sicot J. Sicot C. Abnormal tissue oxygenation in patients with cirrhosis and liver failure. J Hepatol 1988;7:98-105. 13. Bosch J, Arroyo V, Betriu A, Mas A, Carrilho F, Rivera F, Navarro-Lopez F, Rodes J. Hepatic hemodynamics and the reninangiotensin-aldosterone system in cirrhosis. Gastroenterol 1980;78:92-99. 14. Lewis FW, Cohen JA, Rector WG. Autonomic dysfunction in alcoholic cirrhosis: relationship to indicators of sympathetic activation and the occurrence of sodium retention. Am J Gastroenterol 1991;86:553-559. 15. Rector WG, Adair 0, Hossack K, Rainquet S. Atria1 volume in cirrhosis: relationship to blood volume and plasma concentra-
CARDIAC OUTPUT IN CIRRHOSIS
March 1992
tion of atria1 natriuretic factor. Gastroenterology 1990;99:766770. 16. Rector WG Jr, Lewis FW, Robertson AD, Everson GT. Renal sodium retention complicating alcoholic liver disease: relation to portal-systemic shunting and liver function. Hepatology 1990;12:455-459. 17. Lewis FW, Adair 0, Hossack KF, Everson GT, Rector WG. Plasma glucagon concentration in cirrhosis is related to liver function but not portal-systemic shunting, systemic hemodynamic features, or sodium retention. J Lab Clin Med 1991;117:67-75. 18. Feigenbaum H. Echocardiography. 4th ed. Philadelphia: Lea & Febiger, 1986. 19. Lieberman FL, Reynolds TB. Plasma volume in cirrhosis of the liver: its relation to portal hypertension, ascites and renal failure. J Clin Invest 1967;46:1297-1307. 20. Lewis JF, Kuo C, Nelson JG, Limacher MC, Quinones MA. Pulsed Doppler echocardiographic determination of stroke volume and cardiac output: clinical validation of two new methods using the apical window. Circulation 1984;70:425431. 21. Himelman RB, Cassidy MM, Landzberg JS, Schiller NB. Reproducibility of quantitative two-dimensional echocardiography. Am Heart J 1988;115:425-431. 22. Sahn DJ, De Maria A, Kisslo J, Weyman A. Recommendations regarding quantitation in M-mode echocardiography. Circulation 1978;58:1072-1083. 23. Rector WG, Hossack KF. Pathogenesis of sodium retention complicating cirrhosis: is there room for diminished “effective” arterial blood volume? Gastroenterol 1988;95:16581663. 24. Kelbaek H, Eriksen J, Brynjolf I, Raboel A, Lund JO, Munck 0, Bonnevie 0, Godtfredsen J. Cardiac performance in patients
25.
26.
27.
28.
29.
30.
31.
1029
with asymptomatic alcoholic cirrhosis of the liver. Am J Cardiol 1984;54:852-855. Keller H, Bezjak V, Stegaru B, Buss J, Holm E, Heene DL. Ventricular function in cirrhosis and portasystemic shunt: a twodimensional echocardiographic study. Hepatology 1988;8: 658-662. Shah A, Easwaran V. Pericardial effusion and left ventricular dysfunction associated with ascites secondary to hepatic cirrhosis. Arch Intern Med 1988;148:585-588. Dancy M, Bland JM, Leech G, Gaitonde MK, Maxwell JD. Preclinical left ventricular abnormalities in alcoholics are independant of nutritional status, cirrhosis and cigarette smoking. Lancet 1985;1:1122-1125. Gould L, Shariff M, Zahir M, Di Lieto M. Cardiac hemodynamics in alcoholic patients with chronic liver disease and a presystolic gallop. J Clin Invest 1969;48:860-868. Limas CJ, Guiha NH, Lekagul 0, Cohn JN. Impaired left ventricular function in alcoholic cirrhosis with ascites. Circulation 1974;44:755-760. Ramond MJ, Comoy E, Lebrec D. Alterations in isoprenaline sensitivity in patients with cirrhosis: evidence of abnormality of the sympathetic nervous activity. Br J Clin Pharmacol 1976;21:191-196. Levy M, Allotey JBK. Temporal relationships between urinary salt retention and altered systemic hemodynamics in dogs with experimental cirrhosis. J Lab Clin Med 1978;92:560-569.
Received October 17, 1990. Accepted August 9, 1991. Address requests for reprints to: Frederick W. Lewis. M.D., 950 East Harvard, no. 540, Denver, Colorado 80210. Supported by Public Health Service grants lR29AA07832 and MOl-RR00051.
Arterial Vasodilation Is Not the Cause of Increased Cardiac Output in Cirrhosis FREDERICK
W. LEWIS, OLIVIA ADAIR,
and WILLIAM G. RECTOR, Jr.
Divisions of Cardiology and Gastroenterology, Department of Medicine, Denver General Hospital, Denver, Colorado
A pathological state of arterial vasodilation has been postulated to cause the increased cardiac output commonly observed in cirrhosis. Further, suhsequent arterial underfilling has been proposed as the stimulus to sodium retention and ascites formation. Left ventricular size during the cycle of a cardiac contraction is predictably altered by a decrease in afterload. Specifically, increased systolic emptying should he observed. The relationship of left ventricular size during the cardiac cycle to systemic hemodynamic indices and urinary sodium retention was investigated in patients with alcoholic cirrhosis to test these hypotheses. Echocardiographic studies were performed on 24 male patients with alcoholic cirrhosis and compared with the results obtained in 10 age-matched male controls. Patients with cirrhosis had increased cardiac output and heart rate and decreased arterial pressure compared with normal subjects, confirming the presence of a hyperdynamic circulation. Patients with cirrhosis had enlarged left ventricular diameter at bothenddiastole (0.08 fO.Olvs.0.07'+0.007 cm/kg dry body wt; P < 0.001) and end systole (0.08 f 0.01 vs. 0.05+ 0.005cm/kg; P < 0.05). Left ventricular end-diastolic diameter was directly related to blood volume (r = 0.56, P < 0.005). No significant differences in cardiac output, arterial pressure, or systemic resistance were found between patients with and patients without ascites. Increased cardiac output in cirrhosis occurs in conjunction with an enlarged ventricle throughout the cardiac cycle. The increase in left ventricular end-systolic diameter indicates that diminished afterload is not responsible for the increase in cardiac output. As the diameter of the ventricle during diastolic filling correlates with vascular volume, cardiac output in cirrhosis may be primarily determined by an increase in vascular volume.
C
irculatory abnormalities observed in patients with cirrhosis include increased cardiac output and heart rate and diminished systemic resistance.14 A circulating vasodilator has been postulated to initiate this hyperdynamic circulation. In this proposal,
the circulation of patients with cirrhosis is analogous to the state of high cardiac output encountered in subjects with arteriovenous fistulae, severe anemia, Paget’s disease of bone, and beri-beri heart disease.44 Study of cardiac pressure-volume relationships by Braunwald, Sonnenblick, and others established the mechanisms of intrinsic cardiac compensation to changes in arterial resistance (“afterload”) and venous pressure (“preload”).‘,* A decrease in arterial resistance increases systolic ejection, and thus cardiac output, by allowing more complete emptying of the ventricle in systole. Conversely, increased diastolic filling of the left ventricle increases cardiac stroke volume by an increase in left ventricular enddiastolic volume. If the arterial vasodilation hypothesis is correct, patients with cirrhosis should show increased systolic emptying of the left ventricle as a result of diminished afterload. Arterial vasodilation has also been postulated to cause sodium retention in cirrhosis. According to this hypothesis, counterregulatory systems respond to the perceived “underfilling” of the arterial circulation caused by vasodilation5.’ Although the degree of circulatory abnormality in cirrhosis is believed to parallel the severity of liver dysfunction, the relationship between the hyperkinetic systemic circulation of cirrhosis and the occurrence of sodium retention is uncertain.“-13 Therefore, an additional aim of this study was to compare hemodynamic findings and left ventricular dimensions in patients with and without ascites in order to evaluate the validity of the postulated sequence of arterial vasodilation, sodium retention, and ascites formation. Materials and Methods Patients and Protocols The protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the Human Subjects Committee of the University of Colorado Health Sciences Center. All patients gave written in-
formed consent. 0 1992by the American Gastroenterological 0016-5065/92/$3.00
Association
CARDIAC OUTPUT IN CIRRHOSIS 1025
March 1992
This study was conducted as part of a larger investigation designed to evaluate factors potentially related to SOdium retention in patients with alcoholic liver disease. These factors include autonomic nervous system function,14 atria1 size and blood volume,*5 liver metabolic capacity and portal-systemic shunt fraction,16 plasma atria1 natriuretic factorI and glucagon concentration,17 and plasma renin activity. Aspects of this investigation not discussed here or in cited articles will be reported in future communications. The study group included 24 men with alcoholic liver disease. The diagnosis of alcoholic liver disease was made by a history of alcoholism, typical laboratory findings, and liver biopsy where indicated. Patients with a history of cardiac disease or with diminished left ventricular fractional shortening (~20%) on echocardiographic examination were excluded.” Ten male volunteers of similar ages (mean age, 39 + 7 years; range, 25-60 years) without cardiac or hepatic disease served as the control group. Repeated studies were performed after an interval of 3-26 months on 12 patients as a part of this longitudinal investigation. This study included 8 patients who remained without ascites for >6 months, 1 patient in whom ascites persisted for >6 months, 2 patients in whom ascites resolved before the second study, and 1 patient without ascites who developed sodium retention in conjunction with renewed alcohol intake. Each patient was admitted to the Adult Clinical Research Center of the University of Colorado Health Sciences Center for 5 days. Diuretics had been discontinued at least 10 days before admission to eliminate alterations in circulatory features or vascular volume attributable to such treatments. A dietitian interviewed each patient, and average sodium intake at home was determined. The patients were placed on a diet containing 200 mEq of sodium daily, and complete 24-hour urine collections for sodium and creatinine determination were performed on days 2-4. The presence of sodium retention and ascites formation was established by physical examination and by noting weight change and sodium excretion during the study period. Plasma volume was measured by isotope dilution using’ZSI-albumin and blood volume calculated as Plasma Volume/[(l - Hematocrit) (0.82)].” On day 5, supine blood pressure and heart rate were measured using an automated sphygmomanometer (Dinamap-Vital Signs Monitor 8100; Critikon Inc., Tampa, FL) with the patient in the fasting state and after 30 minutes’ rest in a quiet, darkened room. Two-dimensional echocardiography, color-flow Doppler imaging, and directed pulsed-wave Doppler imaging were then performed as described, using a Hewlett Packard Model 77020 AC phasedarray system (Palo Alto, CA), and stored on VHS video tape for subsequent analysis. “JO Images were acquired using a 2.5-mHz or 3.5-mHz transducer while the patient was in a supine or left lateral position. A parasternal long-axis view, with M-mode directed recordings of the aortic valve, aortic root, and left ventricular outflow tract, and left ventricle diastolic and systolic images were obtained. Apical four-chamber and apical long-axis two-chamber views and color-flow-directed pulsed Doppler of the left ventricular outflow tract were recorded with electrocardio-
graphic (ECG) display. Left ventricular dimensions (Mmode), flow rates (pulsed-wave Doppler), ejection time, and heart rate were analyzed using Hewlett Packard online computer software. Fractional shortening was calculated as [lo0 X (Diastolic Diameter - Systolic Diameter)/ Diastolic Diameter]. 20*22 Left ventricular outflow tract area was calculated as z(D/~)~, D being the left ventricular outflow tract diameter. Stroke volume was calculated using the color-flow-directed pulsed-wave Doppler by the formula (Left Ventricular Outflow Tract Area) X (Mean Blood Velocity Across the Aortic Valve) X (Ejection Time). Cardiac output was calculated as (Stroke Volume) X (Heart Rate). We have previously shown that this method of determining cardiac output in cirrhosis gives results similar to those of thermodilution methodology.23 Cirrhosis is associated with abnormalities in body muscle and fat mass that progress with the severity of liver disease, as well as alterations in total body water and sodium content. Consequently, the body surface area calculated from height and weight may not be a valid method of correcting for differences in body size in this population. Therefore, comparisons of cardiac dimension and volume were made by expressing these data per kilogram of dry body mass.23 Dry body mass was taken as actual body mass in patients without ascites and as body mass after subsequent complete diuresis in patients with ascites. The systemic vascular resistance index was used to obtain meaningful units in which to express resistance, and was calculated as (Mean Arterial Pressure X 80)/Cardiac Index. Statistics Data are expressed as mean ? SD. Data were analyzed using the SAS statistical program (Cary, NC). An unpaired t test was used to make group comparisons between patients with cirrhosis and normal subjects and between patients with cirrhosis with and without ascites. Pearson correlation coefficients were determined using a linear regression model. All statistical calculations were corrected for the effect of multiple observations in 12 of the 24 patients as follows: repeat studies on patients who gained or lost ascites (n = 3) were treated as separate observations, with one set in the “ascites” group and one set in the “no ascites” group; studies in each patient undergoing two studies without a change in ascites status (n = 9) were treated as a single observation, and the mean of the two values was used for each calculation. As a result, 27 observations are reported in Tables l-3. For completeness, each individual measurement (n = 36) is represented separately in Figures 1 and 2. A P value of 0.05 was accepted as statistically significant in all calculations.
Results Patients with cirrhosis had hyperkinetic circulatory characteristics as manifested by increased cardiac output and heart rate compared with control subjects (Table 2). The systemic resistance index tended to be lower in patients with cirrhosis, but this difference did not reach statistical significance. En-
1026
LEWIS ET AL.
Table I. Patient
GASTROENTEROLOGY Vol. 102, No. 3
Table 3. Hemodynamic Features, Blood Volume, and Left Ventricular Dimensions
Characteristics No ascites
n 46 34 (2.0 32 (3.2
Age (yr) Serum bilirubin concentration [wmol/L (mg/dL)l Serum albumin concentration [g/L (g/dL)l Aspartate aminotransferase W/L) Alanine aminotransferase
17 + 10 f 38 f 2.2) f 7 f 0.7)
94 f
W/L) Body weight (kg) Dry mass (kg) Sodium intake at home (mEq/doy) Sodium excretion (mEq/72 h)
Ascites
’
45 74 (4.3 22 (2.2
10 f 10 + 92 f 5.4) f 3 f 0.3)
104 rfr64
100
45 * 37 72 f 12 72+12
a
52 + 26 69 f 10 64 + 6
325 + 150 438+ 123
’
182 f 122 71 z!I92
Cirrhosis, no ascites n Cardiac output (mL.min-’ . kg-‘) Heart rate (beats/min) Mean arterial pressure (mm Hg) Systemic resistance index (dyne. s *crnm5. m-‘) Blood volume (mL ‘kg-‘) Left ventricular diameter End diastole (cm akg-‘) End systole (cm -kg-‘) Stroke volume (mL . kg-‘) Fractional shortening (%)
Cirrhosis, ascites
17
10
118 f 57 82 f 12
129 + 43 87 + 12
93 f 8
94 f 6
1990 + 800
1760 + 600 81 + 10
75 + 15 0.08 f 0.01 0.06 f 0.01 1.5 + 0.6 29f6
0.08 + 0.01 0.05 + 0.01 a
1.5 -+ 0.4 36 _+ 10
“P < 0.05.
“P < 0.05.
larged left ventricular diameter, both at end diastole and at end systole, and increased fractional shortening were also seen in patients with cirrhosis. When patients with and without sodium retention were compared, the only significant difference was increased fractional shortening in those with ascites (Table 3). No significant differences were detected in cardiac output, heart rate, arterial pressure, left ventricular dimensions, or blood volume, although trends toward increased cardiac output and enlarged left ventricle in diastole in the group with ascites were present. Left ventricular end-diastolic diameter was directly related to cardiac output (Figure 1A). In addition, a significant linear correlation between blood volume and end-diastolic diameter of the left ventricle (Figure 2) was seen. However, increased cardiac
output was not associated with increased systolic emptying. Indeed, the reverse was observed (Figure 1B). Similar relationships were observed between stroke volume and left ventricular dimensions (Figure 1C and D). Arterial hypotension was not associated with increased ventricular emptying. Rather, inverse linear correlations were detected between mean arterial pressure and both end-diastolic and end-systolic ventricular diameters (Figure 1E and F). No significant relationships were observed between either ventricular dimension and systemic resistance (r = -0.11 andr = -0.27 for end-diastolic diameter and end-systolic diameter, respectively; both P = NS).
Table 2. Hemodynamic Dimensions
n Cardiac output (mL - min-’ -kg-‘) Heart rate (beats/min) Mean arterial pressure (mm Hg) Systemic resistance index (dyne. s. crnm5. m-“) Left ventricular diameter End diastole (cm. kg-‘) End systole (cm - kg-‘) Stroke volume (mL *kg-‘) Fractional shortening (%) OP < 0.05. bP < 0.005.
Features
and Left Ventricular Control
Cirrhosis
10
27
94+25 67 f 7
a b
122 + 51 84 + 12
95 f 8
93 + 7
2140 f 400
1900 f 700
0.07 * 0.007 0.05 f 0.005
b o
1.4 f 0.3
26 f 4
0.08 + 0.01 0.06 + 0.01 1.5 + 0.5
b
32 + 8
Discussion This is the first investigation to examine the relationship of left ventricular size to systemic hemodynamics in cirrhosis, although several previous studies have measured left ventricular dimensions in conjunction with evaluation of cardiac function in cirrhosis or alcoholism. These reports are consistent with our finding of enlarged left ventricular diameter at both end diastole and end systole in patients with cirrhosis.24-27 The principle aim of this study was to test the hypothesis that increased cardiac output in cirrhosis is attributable to arterial underfilling. Our data do not support this mechanism. The observed enlargement of the left ventricle throughout the cardiac cycle suggests that the increase in cardiac output is effected by increased ventricular filling during diastole rather than by increased ventricular emptying during systole. The concept that increased ventricular filling underlies the increase in cardiac output in cirrhosis is supported by several further observa-
CARDIAC
March 1992
OUTPUT
r=.53,
I
A
Figure 1. Direct linear correlations between end-diastolic and end-systolic left ventricular dimensions and cardiac output (A and B) and stroke volume (C and D); inverse correlation between left ventricular diameters and arterial pressure (E and F).
.41,
END SYSTOLE
LEFT VENTRICULAR
Cons. Direct correlations were observed between both cardiac output and stroke volume and left ventricular diameter, both at end diastole and at end systole. Furthermore, an inverse correlation was observed between mean arterial pressure and both left ventricular dimensions, and no relationship was discernable between systemic vascular resistance and either ventricular dimension. The opposite, a direct relationship between ventricular diameter and both systemic resistance and arterial pressure, should have been observed if arterial underfilling were the cause of the observed increase in cardiac output. Finally, the significant linear correlation documented between blood volume and left ventricular end-diastolic diameter, similar to the relationship we have previously reported between left atria1 volume and
PLO05
pco5
p<.Ol
1027
b
A.-
,.
r=.51,
r=-
IN CIRRHOSIS
,*
,
A ’
A r=--40,
pco5 F
END DIASTOLE
DIAMETER
(cm/kg)
blood vo1ume,‘5 is reasonable evidence that ventricular enlargement probably results from increased intravascular volume. Ventricular enlargement can, of course, result from cardiac failure. It has been suggested that some patients with cirrhosis, despite showing normal resting cardiac function, have a cardiomyopathy that results in subnormal responses to cardiac stress.6~24~28-30 It is unlikely that this postulated condition influenced the findings of this study. The patients in the current study did not have any history, signs, or symptoms of cardiac disease. Moreover, there was by definition no echocardiographic evidence of cardiac dysfunction because patients with abnormal left ventricular function had been excluded. Second, previously noted circulatory abnormalities28-30 have not
1028
_,
LEWIS ET AL.
,D1
cn
GASTROENTEROLOGY Vol. 102, No. 3
r=.56, pc.005 v =no ascites
v V
*=ascites
bl
I;!
!
vv 0
60--
V W
W
d
50, 0.050
0.060
0.070
0.080
LEFT VENTRICULAR AT END-DIASTOLE
0.090
0.100
DIAMETER
rial pressure and filling to decrease, initiating renal sodium retention. We have previously shown that hepatic metabolic capacity and portal-systemic shunting are unrelated to the systemic hemodynamic features of patients with cirrhosis.” An additional aim of the present investigation was to relate indices of arterial filling to the presence of sodium retention. We have shown that cardiac output is not an indicator of arterial filling. No significant differences were found between other commonly used indices of arterial filling, including arterial blood pressure and systemic vascular resistance, in patients with and without ascites. These findings do not support the mechanism of sodium retention in cirrhosis proposed by the arterial vasodilation hypothesis.5
(cm/kg)
Figure 2. Blood volume is directly related to left ventricular end-diastolic diameter.
been proved to result from cardiomyopathy. Circulatory or autonomic abnormalities associated with cirrhosis may instead be responsible. Indeed, patients in the current study showed increased cardiac output compared with controls, rather than the decreased output that would be expected if intrinsic cardiac disease were present. Finally, within the patient group, a direct correlation was observed between both cardiac output and stroke volume and left ventricular size. This relationship would be an unlikely result of cardiomyopathy. Increased blood volume in patients with cirrhosis and ascites is a well-recognized feature of this condition.lg In the current study, blood volume and left ventricular size tended to be greater in patients with ascites than in patients without ascites, although the differences did not reach statistical significance. More importantly, our data indicate that volume expansion is probably responsible, through increased diastolic filling, for the increase in cardiac output observed in cirrhosis. As stated, a close, direct relationship was observed between blood volume and the diameter of the left ventricle in diastole. Longitudinal study of an experimental model of cirrhosis in dogs supports the postulate that sodium retention precedes hemodynamic changes and ascites formation.31 According to this mechanism, increased cardiac output is an effect of volume expansion rather than part of its cause. Even if arterial underfillingis not directly responsible for the increase in cardiac output observed in cirrhosis, it may still be the stimulus for sodium retention. According to this explanation, increased systemic availability of a vasodilator as a result of liver dysfunction or portal-systemic shunting causes arte-
References 1. Kowalski
HJ, Abelmann WH. The cardiac output at rest in Laennec’s cirrhosis. J Clin Invest 1953;32:1025-1033. 2. Murray JF, Dawson AM, Sherlock S. Circulatory changes in chronic liver disease. Am J Med 1958;24:358-367. 3. Kontos HA, Shapiro W, Mauck HP, Patterson JL. General and regional circulatory alterations in cirrhosis of the liver. Am J Med 1964;37:526-535. 4. Sherlock S. Vasodilation associated with hepatocellular disease: relation to functional organ failure. Gut 1990;31:365367. 5. Schrier RW. Body fluid volume regulation in health and disease: a unifying hypothesis. Ann Intern Med 1990;113:155159. 6. Lee SS. Cardiac abnormalities in liver cirrhosis. West J Med 1989;151:530-535. 7. Grossman W, Braunwald E. High-cardiac output states. In: Braunwald E, ed. Heart disease. Philadelphia: Saunders, 1988:778-792. 8. Sonnenblick EH, Schlant RC. Normal physiology of the cardiovascular system. In: Hurst JW, ed. The heart. 6th ed. New York: McGraw-Hill, 1986:37-72. 9. Shrier RW, Arroyo V, Bernardi M, Epstein M, Henriksen JH, Rodes J. Peripheral arterial vasodilatation hypothesis: a proposal for the initiation of renal sodium and water retention in cirrhosis. Hepatology 1988;8:1151-1157. 10. Siegel JH, Goldwyn RM, Farrell EJ, Gallin P, Friedman HP. Hyperdynamic states and the physiologic determinants of survival. Arch Surg 1974;108:282-292. 11. Braillon A, Cales P, Valla D, Gaudy D, Geoffery P, Lebrec D. Influence of the degree of liver failure on systemic and splanchnic haemodynamics and on response to propranolol in patients with cirrhosis. Gut 1986;27:1204-1209. 12. Moreau R, Lee SS, Soupison T, Roche-Sicot J. Sicot C. Abnormal tissue oxygenation in patients with cirrhosis and liver failure. J Hepatol 1988;7:98-105. 13. Bosch J, Arroyo V, Betriu A, Mas A, Carrilho F, Rivera F, Navarro-Lopez F, Rodes J. Hepatic hemodynamics and the reninangiotensin-aldosterone system in cirrhosis. Gastroenterol 1980;78:92-99. 14. Lewis FW, Cohen JA, Rector WG. Autonomic dysfunction in alcoholic cirrhosis: relationship to indicators of sympathetic activation and the occurrence of sodium retention. Am J Gastroenterol 1991;86:553-559. 15. Rector WG, Adair 0, Hossack K, Rainquet S. Atria1 volume in cirrhosis: relationship to blood volume and plasma concentra-
CARDIAC OUTPUT IN CIRRHOSIS
March 1992
tion of atria1 natriuretic factor. Gastroenterology 1990;99:766770. 16. Rector WG Jr, Lewis FW, Robertson AD, Everson GT. Renal sodium retention complicating alcoholic liver disease: relation to portal-systemic shunting and liver function. Hepatology 1990;12:455-459. 17. Lewis FW, Adair 0, Hossack KF, Everson GT, Rector WG. Plasma glucagon concentration in cirrhosis is related to liver function but not portal-systemic shunting, systemic hemodynamic features, or sodium retention. J Lab Clin Med 1991;117:67-75. 18. Feigenbaum H. Echocardiography. 4th ed. Philadelphia: Lea & Febiger, 1986. 19. Lieberman FL, Reynolds TB. Plasma volume in cirrhosis of the liver: its relation to portal hypertension, ascites and renal failure. J Clin Invest 1967;46:1297-1307. 20. Lewis JF, Kuo C, Nelson JG, Limacher MC, Quinones MA. Pulsed Doppler echocardiographic determination of stroke volume and cardiac output: clinical validation of two new methods using the apical window. Circulation 1984;70:425431. 21. Himelman RB, Cassidy MM, Landzberg JS, Schiller NB. Reproducibility of quantitative two-dimensional echocardiography. Am Heart J 1988;115:425-431. 22. Sahn DJ, De Maria A, Kisslo J, Weyman A. Recommendations regarding quantitation in M-mode echocardiography. Circulation 1978;58:1072-1083. 23. Rector WG, Hossack KF. Pathogenesis of sodium retention complicating cirrhosis: is there room for diminished “effective” arterial blood volume? Gastroenterol 1988;95:16581663. 24. Kelbaek H, Eriksen J, Brynjolf I, Raboel A, Lund JO, Munck 0, Bonnevie 0, Godtfredsen J. Cardiac performance in patients
25.
26.
27.
28.
29.
30.
31.
1029
with asymptomatic alcoholic cirrhosis of the liver. Am J Cardiol 1984;54:852-855. Keller H, Bezjak V, Stegaru B, Buss J, Holm E, Heene DL. Ventricular function in cirrhosis and portasystemic shunt: a twodimensional echocardiographic study. Hepatology 1988;8: 658-662. Shah A, Easwaran V. Pericardial effusion and left ventricular dysfunction associated with ascites secondary to hepatic cirrhosis. Arch Intern Med 1988;148:585-588. Dancy M, Bland JM, Leech G, Gaitonde MK, Maxwell JD. Preclinical left ventricular abnormalities in alcoholics are independant of nutritional status, cirrhosis and cigarette smoking. Lancet 1985;1:1122-1125. Gould L, Shariff M, Zahir M, Di Lieto M. Cardiac hemodynamics in alcoholic patients with chronic liver disease and a presystolic gallop. J Clin Invest 1969;48:860-868. Limas CJ, Guiha NH, Lekagul 0, Cohn JN. Impaired left ventricular function in alcoholic cirrhosis with ascites. Circulation 1974;44:755-760. Ramond MJ, Comoy E, Lebrec D. Alterations in isoprenaline sensitivity in patients with cirrhosis: evidence of abnormality of the sympathetic nervous activity. Br J Clin Pharmacol 1976;21:191-196. Levy M, Allotey JBK. Temporal relationships between urinary salt retention and altered systemic hemodynamics in dogs with experimental cirrhosis. J Lab Clin Med 1978;92:560-569.
Received October 17, 1990. Accepted August 9, 1991. Address requests for reprints to: Frederick W. Lewis. M.D., 950 East Harvard, no. 540, Denver, Colorado 80210. Supported by Public Health Service grants lR29AA07832 and MOl-RR00051.