GASTROENTEROLOGY
1992;102:1009-1016
A Randomized Study of Propranolol on Postprandial Portal Hyperemia in Cirrhotic Patients CARLO SABB& GIOVANNA FERRAIOLI, PAOLO BUONAMICO, THOMAS MAHL, KENNETH J, W. TAYLOR, EMANUEL LERNER, OTTAVIO ALBANO, and ROBERTO J. GROSZMANN Hepatic Hemodynamic Laboratory and Research Service, Veterans Administration Medical Center, Yale University, West Haven, Connecticut; Department of Diagnostic Radiology, Yale University School of Medicine, New Haven, Connecticut; and Istituto di Clinica Medica I, Universith degli Studi di Bari, Bari, Italy
Propranolol, a nonselective j3-adrenergic blocker, has been shown to reduce portal pressure and the risk of variceal bleeding. The portal pressure-reducing effect of propranolol is mediated by splanchnic arterial constriction, which decreases portal flow. A double-blind randomized control study (crossover on 2 consecutive days) was designed to compare the effects of propranolol vs. placebo on portal flow in cirrhotic patients during fasting and after a standardized meal. Portal flow was measured with an ATL Ultramark 8 echoDoppler system (Advanced Technological Laboratories, Bethel, WA) in 23 cirrhotic patients. Fasting portal flow and heart rate were obtained at baseline and 2 hours after the administration of propranolol or placebo. A standard test meal was then given, and measurements were repeated 30 minutes later. Thirteen patients (group 1)received placebo on day 1 and propranolol on day 2, whereas 10 patients (group 2) received propranolol on day 1 and placebo on day 2. In group 1 patients, heart rate declined by 20% (P< 0.0001)and portal flow decreased by 12% (P < 0.05) after propranolol administration. Similar reductions were found in heart rate (-2l%, P < 0.0001) and portal flow (-17%, P < 0.001)for group 2 patients. For all 23 patients, 2 hours after propranolol administration, heart rate declined by 21% (P< 0.0001) and portal blood flow was reduced by 14% (P < 0.0001). The 10 patients who received propranolol on day 1 (group 2) showed a carryover effect of propranolol on day 2. On day 2, baseline portal flow and heart rate values were significantly lower than baseline values on day 1.This long-lasting effect of a single dose of propranolol may be caused by the longer half-life of propranolol in cirrhotic patients. The postprandial portal blood flow percentage increase after the meal was similar for both placebo and propranolol. Propranolol did not blunt postprandial hyperemia. However, whereas
the absolute value of blood flow after the meal increased significantly in comparison with baseline this did not in placebo-treated patients (P< O.OOl), occur with propranolol. Furthermore, in propranolol-treated patients the absolute value of blood flow after the meal was lower than in placebo-treated patients. This may constitute a protective effect of propranolol in portal hypertension.
It
has been shown that in cirrhosis, portal pressure states.“’ This inincreases during postprandial crease in portal pressure is presumably caused by an increase in gut blood (portal) flow, a phenomenon known as postprandial hyperemiaa3s4 In portal hypertensive patients, the increase in portal venous flow and pressure following the ingestion of food may, at times, be involved in inducing variceal bleeding. Propranolol, a nonselective P-adrenergic blocker reduces portal pressure by decreasing portal blood flow (F) and, by this mechanism, may prevent variteal hemorrhage.5 Currently, the wedge hepatic venous pressure is the only hemodynamic parameter that has proven to be useful in the long-term prognosis of cirrhotic patients treated with propranoloL6 However, hepatic vein catheterization is costly, time-consuming, and uncomfortable for patients. Echo-Doppler flowmetry is widely used for portal blood flow quantitation. Recently, we investigated the intraobserver and interobserver variability of study, we this technique.7 In an observer-blinded also showed the sensitivity of echo-Doppler in detecting hemodynamic changes caused by postprandial hyperemia and noted significantly different responses between normal subjects and cirrhotic patients.’ 0
1992
bytheAmerican
Gastroenterological 0016-5085/92/$3,00
Association
1010
SABBA ET AL.
Table 1. Baseline
GASTROENTEROLOGY Vol. 102,No.3
in Groups 1 and 2
Measurements
Group 1 Baseline day 1 Baseline day 2
A
V
F
N
(mm*)
(cm/s)
(mL/min)
13 13
189 + 7 186 ItI 7
10.8 f 0.7 11.0 f 0.8
1229 + 84 1225 f 76
77 + 5 78 f 4
134 + 4 131 f 4
10 10
192 + 6 193 + 7
13.0 + 0.8 11.0 + 0.6’
1509 + 69 1267 f 42’
83 f 2 71 + 2a
135 f 7 131 + 5
(b%
SBP (mm Hg)
Group 2 Baseline day 1 Baseline day 2
NOTE. For group 1,there are no statistically significant differences between portal vein parameters, HR, and SBP baseline values over the 2 days. For group 2, significant reductions in velocity, F, and HR are present 24 hours after propranolol administration (baseline day 2). “P < 0.01.
A, area: V, mean velocity: F, flow; HR, heart rate; SBP, systolic blood pressure; group I, placebo day 1, propranolol propranolol
day 2; group 2,
day 1, placebo day 2.
The current double-blind randomized hemodynamic study was designed (a) to evaluate the sensitivity of echo-Doppler in detecting acute changes in F in cirrhotic patients receiving a single dose of propran0101 and (b) to determine whether propranolol has any influence on portal postprandial hyperemia. In addition, because a persistence of the effect of a single dose of propranolol on heart rate (HR) for >24 hours has been reported,g”O this study was designed to explore whether this effect persists also on F.
female patients with a mean t SEM age of 59 + 2, weight of 72 + 3 kg, and height of 168 + 2 cm. Seventeen patients were examined at the Istituto di Clinica Medica I, University of Bari, and the remaining 6 patients at the West Haven Veterans Administration Medical Center, Clinical Campus of the Yale School of Medicine. The diagnosis of cirrhosis was based on liver biopsy findings in 8 of the 23 patients and on clinical data in the remaining patients, The etiology of liver cirrhosis was postnecrotic hepatitis B surface antigen-positive in 12 patients, alcoholic in 8, cryptogenic in 2, and hemochromatotic in 1.Cirrhosis was considered to be alcoholic in origin when alcohol consumption was estimated to be >80 g/day for >6 years preceding admission and no other cause was evident. According to Child-Turcotte criteria,” 8 patients were in class were A, 11 in class B, and 4 in class C. Measurements performed by G.F. at Veterans Administration Medical Center, West Haven, and by P.B. at Bari. The study was approved by the Human Investigation Committees of the West Haven Veterans Administration Medical Center and of Yale University School of
Materials and Methods Patients Twenty-six cirrhotic patients were entered into the study. Three of them were excluded; 2 because of poor portal vein visualization and 1 because bradycardia made the use of propranolol unsuitable. The remaining 23 cirrhotic patients were studied. They included 20 male and 3
90
day,* placebo 0
Z : 0. @
;r
min
Ii0
1
h
._ i Z ‘5’ 2
120
1
1
1300
1100
day
1
i
day2
7
!
0
-L:~
propranolol
placebo 900
a
120
min
0
120
’
Figure 1. (A) HR and (B) F are shown at baseline and 2 hours after placebo and propranolol administration in the 13 patients receiving placebo on day 1 and proprano101 on day 2 (group 1).'P < 0.05; *P < 0.0001 with respect to the same day baseline (paired t test).
PROPRANOLOL
March 1992
Table 2. Portal
Vein Parameters,
Day 1 placebo Baseline day 1 z h after placebo 30 min after meal Day 2 propranolol Baseline day 2 2 h after drug 30 min after meal
ON PORTAL BLOOD FLOW
1011
HR, and SBP in Group 1 HR (bpml
F (mL/min)
N
A (mm’)
v (cm/s)
13 13 13
189+ 7 191f 7 194 + 6
10.8 f 0.7b 10.8 + 0.6",' 13.8 f 1.0".b.'
1229f84b 1246rf-81",k 1602 + 107°,b,'
77 * 5c 75 f 5"J
13 13
186 * 7 187 + 8
11.0 f 0.8" 9.6 k O.GeXi
1225 It76e 1072 + 77e,f,k
78 k 4g 62 + 3g,'
13
189f
11.8 f 1.G'
1388 + 93L'
7
SBP (mm Hgl 134 +4d 130 * 5d.l
131&4h 116 f 3h.'
Statistical significance (paired t test): c.e.i.iP< 0.05;df.h.LP< 0.01; a.b,‘P< 0.001; PP < 0.0001. A, Area; V, mean velocity; F, portal blood flow; HR, heart rate; SBP, systolic blood pressure.
Medicine. Informed written consent was obtained each patient, both in Bari and West Haven.
plex system calculated the area using the following ellipse formula: A = nD,D,/4. Velocity measurements were performed in quiet suspended inspiration and averaged over a few seconds. They were taken on a longitudinal section of the vessel positioning the sample volume cursor within the vessel at a known angle of insonation (angle between the Doppler beam and the long axis of the vessel). This angle was always ~60”. The Doppler sample volume was positioned in the center of the vessel and was manipulated to cover about 90% of the vessel section to detect the maximum Doppler frequency shift (Af), related to the maximum velocity. According to the Doppler equation
from
Equipment Studies were performed at both centers using an ATL Ultramark 8 echo-Doppler duplex system (Advanced Technological Laboratories, Bethel, WA) with a 3.0-MHz sector mechanical probe. Real-time and Doppler control settings were optimized for each patient according to established techniques.
Calculations Portal vein cross-sectional area (A, square millimeters) and mean blood velocity (V, centimeters per second) were measured to calculate blood flow (F, milliliters per minute) according to the following formulas: F = A X V X 60. Area was obtained from a cross-section of the vessel visualized by B-mode after defining the major and minor diameters of the vessel. The computer software of the du-
Af
cos 8 , 2fOv”MX
=
C
f, is the incident Doppler frequency, 8 is Doppler angle of insonation, and c is the velocity of the ultrasound in tissues. The mean velocity was calculated from maximum
1800
day A
1500 Meal placebo
Figure 2. The effects are shown for placebo and propranolol on postprandial F in the 13 patients receiving placebo on the first day and propranolol on the second day (group 1).*P < 0.05between same time interval measurements on 2 consecutive days (paired t test).
\
propranolol
900
’
I
0
120 TIME
bmin)
Meal
150
1
1012
SABBA ET AL.
GASTROENTEROLOGY Vol. 102. No. 3
Table 3. Portal Vein Parameters,
Day 1 propranolol Baseline day 1 2 h after drug 30 min after meal Day 2 placebo Baseline day 2 2 h after placebo 30 min after meal
HR, and SBP in Group 2 SBP
A
V
N
(mm’1
(cm/s)
F (mL/min)
10 10 10
192 + 6 192 + 7 194 k 6
13.0 f 0.8a.h 10.8 f 0.7°.b 12.3 It O.gb
1509 + 69C,h 1243 + 62Csb 1517 + 91b
83 f Zmeh 65 f 4’ -
135 + 7c 119 f 5c
10 10 10
193 f 7 191 + 7d 200 + 7d
11.0 f 0.6h 11.0 f 0.7e 13.5 k l.ge
1267 f 42h 1264 ?I 59” 1566 + 156e
71 + 2bh 67 !I Zf -
131+ 58 119 + 48 -
(mm Hg)
-
< 0.01; cP < 0.001; OP < 0.0001. Statistical significance (paired t test): b.eafP< 0.05; d*g,hP A, Area; V, mean velocity; F, portal blood flow; HR, heart rate: SBP, systolic blood pressure.
unaware of the effect of propranolol or placebo on arterial pressure and HR. If HR had not decreased by 20%, another dose of 40 mg of propranolol was administered [unless HR was ~60 beats per minute (bpm) or SBP < 90 mm Hg]. Two hours after propranolol or placebo administration, echoDoppler measurements of portal vein parameters were repeated. Then a standard mixed liquid meal (355 kcal, Ensure plus; Ross Laboratories, Columbus, OH) was given, and the same measurements were performed 30 minutes later. The criterion for inclusion of a subject in the data analysis was feasibility of portal vein measurements by echoDoppler, i.e., adequate visualization of the vessel in two planes with a Doppler angle of insonation of lO% and nonresponders if it was 10% (range, -10% to -41%), and 5 of them did not respond (range, -3% to f14%). Compared with the responders, the nonresponders to propranolol had a smaller increase in F after the meal both on the propranolol(50% + 10% vs. 10% + 14%; P < 0.05)and placebo day (40% f 7% vs. 16% + 6%; P < 0.05). Group 2 The baseline values of the 10 patients receiving propranolol on the first day and placebo 24 hours later are shown on Table 1. In comparison with the first day, a significantly lower baseline HR and F were detected on the second day (P < 0.01 and P < 0.01, respectively), indicating a probable residual effect of propranolol administered on day 1 (Figure 3). When propranolol was given, HR, SBP, and F decreased by 2l%, 11% and 17% respectively (P < 0.001). After placebo administration, F did not change significantly; however, HR and SBP were significantly decreased (P < 0.05; Table 3 and Figure 3). In terms of percentage changes from baseline, the reductions in HR and in F induced by propranolol were significantly greater than after placebo (P < 0.001 and P < 0.01, respectively). No significant percentage change was observed in SBP. There were no statistically significant percentage differences in the behavior of all three parameters (HR, F, and SBP) when group 1 and group 2 were compared with each other (Figure 4).
GASTROENTEROLOGYVol.102.No.3
Probably because of different baseline values on day 1 and day 2, no statistically significant differences were observed on F and HR between absolute values attained at 120 minutes after either proprano101(day 1) or placebo (day 2) administration (Figure 3; Table 3). The F changes after meal ingestion were similar to those observed in group 1 (Tables 2 and 3). Group 1 Plus Group 2 For all 23 patients, in comparison with baseline, HR declined by 21% (P < O.OOOl),SBP decreased by 11% (P < O.OOOl),and F was reduced by 14% (P < 0.0001) 2 hours after propranolol administration. In comparison with placebo, propranolol decreased HR by 10% (P < 0.01) SBP by 6% (P < O.Ol), and F by 8% (P < 0.05) at 120 minutes. After the meal, there was a similar and significant percentage increase in F; however, the peak blood flow reached in response to the meal in comparison with baseline F was significantly higher in the placebo group (P < 0.0001) but not in the propranolol group. Despite a 20% difference in baseline F between the two groups, the findings between the total (n = 23) and group 1 (n = 13)and 2 (n = 10) were similar in the HR, SBP, and F response to propranolol. The variability for F on the same day between baseline and 120 minutes (n = 13) ranged from -13.1% to18.2%withamean+SEMof1.8%+2.5%. The variability on consecutive days between baselines ranged from -10.8% to 11.9% with a mean of 0.4% f 1.8% (mean + SEM). No correlation was found between change in HR or blood pressure and F. Discussion In this double-blind ultrasonography and patient study, echo-Doppler flowmetry was able to detect changes in portal hemodynamics after propran0101 administration in cirrhotic patients. The hemodynamic effect of a single dose of propranolol on F appeared to persist for >24 hours. Indeed, when the drug was given the first day, decreases in HR, SBP, V, and F were still present on the second day. The impaired liver-clearing capacity of cirrhosis is probably responsible for maintaining the effects of this drug beyond the usual half-life.‘5*‘” Propranolol is metabolized during a first pass through the liver, and it has been shown that the half-life of propranolol and its effect on HR are extended in patients with liver cirrhosis.g.‘o However, it has not been shown previously that propranolol has a prolonged effect on portal blood flow. Changes in flow were mainly related to changes in velocity. Previous studies by this laboratory and by another
PROPRANOLOL
March 1992
center regarding blood flow changes in the portal vein in response to a meal have reported similar reSUltS.8*‘7 In this study, we observed that, in cirrhotic patients, the postprandial hyperemic responses are similar after propranolol or placebo. This indicates that P-adrenergic blockade does not interfere with the basic mechanism that induces the postprandial hyperemic state. Because of the lower fasting F induced by propranolol, the absolute value of postprandial F was significantly lower on propranolol day despite a similar percentage increase. Furthermore, the absolute increase in F from baseline to the postprandial 150-minute measurement was significantly greater in the placebo group but was not significantly increased in the propranolol group (Figure 2; Tables 2 and 3). This may constitute a protective effect of propranolol in portal hypertension. A previous study showed that propranolol does not prevent the postprandial increase in portal pressure in cirrhotic patients.’ Our results are consistent with physiological studies on animals that have shown that (3-adrenergic blockade does not affect the postprandial hyperemia.” Six of the 23 patients (5 in group 1 and 1 in group 2) did not respond to propranolol with a decrease from baseline F > 10%. These patients showed a significantly reduced postprandial hyperemic response with and without propranolol. This finding suggests that the basic mechanism leading to a lack of response to propranolol and to a meal stimulus may be related. It has been shown previously that (j-adrenergic blockers are not effective in reducing portal pressure in all cirrhotic patients.‘4r’g These results coincide with the portal flow response to propranolol in this study. In this as in a previous report,14 there were no significant differences in any baseline clinical or hemodynamic features that will distinguish responders from nonresponders. Previous studies have shown limitations with the use of the echo-Doppler when studying F.7,‘3 In view of our experience with the variability of echoDoppler technique for studying F,7 we used the mean of three measurements at each time interval for each parameter. In the current and in previous studies, it was observed that in portal hypertensive patients, the cross-sectional area of the portal vein showed small variations in response to changes in F.20*21 Therefore, it may be possible to measure changes in blood velocity without the additional errors caused by calculation of portal vein area. In conclusion, this study shows that the absolute value of blood flow after a meal in propranololtreated patients is reduced in comparison with placebo. This, in itself, may be an important component
ON PORTAL BLOOD FLOW
1015
of the (3-adrenergic blocker-protective effect in the prevention of variceal bleeding. In addition, the current study shows that, under controlled conditions, echo-Doppler is able to detect the effect of a single dose of propranolol on portal blood flow. Additional studies are needed to relate hemodynamic findings with clinical events. References 1. Lee S, Handengue
A, Moreau R, Hillon P, Lebrec D. Post-prandial hemodynamic responses in patients with cirrhosis. Hepatology 1988;8:647-651. 2. McCormack PA, Afdhal NH, Hegarty JE. Effect of propranolol on post-prandial increase in portal venous pressure in patients with portal hypertension (abstr). Gastroenterology 1987;92:1755. 3. Chou CC. Splanchnic
and overall cardiovascular hemodynamics during eating and digestion. Fed Proc 1983;42:1658-
1661. 4. Gallavan
RH, Chou CC. Possible mechanisms for the initiation and maintenance of post-prandial intestinal hyperemia. Am J Physiol 1985;249:G301-G308. 5. Pagliaro L, Burroughs AK, Sorensen TIA, Lebrec D, Morabito A, D’Amico G, Fine T. Therapeutic controversies and randomized controlled trials (RCTs): prevention of bleeding and rebleeding in cirrhosis. Gastroenterol Int 1989;2:71-84. 6. Groszmann RJ, Bosch J, Grace N, Conn HO, Garcia-Tsao G, Navasa M, Alberts J, Rodes J, Fischer R, Bermann M, Rofe S, Patrick M, Lerner E. Hemodynamic events in a prospective randomized trial of propranolol vs. placebo in the prevention of the first variceal hemorrhage. Gastroenterology 1990;99: 1401-1407. 7. SabbL C, Weltin GG, Cicchetti
DV, Ferraioli G, Taylor KJW, Nakamura T, Moriyasu F, Groszmann RJ. Observer variability in echo-Doppler measurements of portal flow in cirrhotic patients and normal volunteers. Gastroenterology 1990;98: 1603-1611.
8. Sabba C, Ferraioli
9.
10.
11. 12.
13. 14.
G, Genecin P, Colombato L, Buonamico P, Lerner E, Taylor KJW, Groszmann RJ. Evaluation of postprandial hyperemia in superior mesenteric artery and portal vein in healthy and cirrhotic humans: an operator-blind study. Hepatology 1991;13:714-718, Arthur MJSP, Tanner AR, Plate1 C, Wright R, Renwick AG, George CF. Pharmacology of propranolol in patients with cirrhosis and portal hypertension. Gut 1985;26:14-19. Wood AJJ, Kornhauser DM, Wilkinson GR, Shand DG, Branc RA. The influence of cirrhosis on steady state blood concentrations of unbound propranolol after oral administration. Clin Pharmacokin 1978:3:478-487. Conn HO. A peek at the Child-Turcotte classification. Hepatology 1981;1:673-676. Moriyasu F, Ban N, Nishida 0, Nakamura T, Miyake T, Uchino H, Kanimatsu Y, Koizumu S. Clinical application of an ultrasonic duplex system in the quantitative measurement of portal blood flow. J Clin Ultrasound 1986;14:579-588. Gill RW. Measurement of blood flow by ultrasound: accuracy and source of error. Ultrasound Med Biol 1985;11:625-641. Garcia-Tsao G, Grace ND, Groszmann RJ, Conn HO, Bermann NN, Patrick MJC, Morse SS, Alberts JL. Short-term effects of propranolol on portal venous pressure. Hepatology 1986; 6:101-106.
15. Paterson
JW, Conolly ME, Dollery CT, Hayes A, Cooper RG. The pharmacodynamics and metabolism of propranolol in man. Pharmacol Clin 1976;2:127-133.
1016
SABBA ET AL.
16. Pessaye D, Lebrec D, Descatoire V, Peignaux M, Benhamou JP. Mechanism for reduced drug clearance in patients with cirrhosis. Gastroenterology 1978;74:566-571. 17. Gaiani G, Bolondi L, Li Bassi S, Santi V, Zironi G, Barbara L. Effect of meal on portal hemodynamics in healthy humans and in patients with chronic liver disease. Hepatology 1989;9:815-819. 18. Takagi T, Naruse S, Shionoya SS. Post-prandial celiac and superior mesenteric blood flows in conscious dog. Am J Physiol 1988;255:G522-G528. 19. Vorobioff J, Picabea E, Villavicenzio R, Puccini V, Rossi 0, Bordato J, Audato M. Acute and chronic hemodynamic effects of propranolol in unselected cirrhotic patients. Hepatology 1987;7:648-653. 29. Bolondi L, Gandolfi C, Arienti V, Caletti GC, Corcioni E, Gasbarrini G, Labd G. Ultrasonography in the diagnosis of portal
GASTROENTEROLOGY Vol. 102. No. 3
hypertension: diminished response of portal vessels to respiration. Radiology 1982;142:167-172. 21. Sabba C, Buonamico P, Ferraioli G, Taylor KJW, Groszmann RJ. Measurement of mean velocity to monitor acute portal blood flow changes in portal hypertensive patients (abstr). Gastroenterology 1990;98:A199.
Received November 13, 1990. Accepted August 12,1991. Address requests for reprints to: Roberto J. Groszmann, M.D., Veterans Administration Medical Center, 950 Campbell Avenue, West Haven, Connecticut 06516. Dr. Ferraioli’s present address is Divisione di Medicina, Ospedale Civile di Scafati, Scafati 84018, Italy. The authors gratefully acknowledge Marge Petrucci for her secretarial assistance.
1992;102:1009-1016
A Randomized Study of Propranolol on Postprandial Portal Hyperemia in Cirrhotic Patients CARLO SABB& GIOVANNA FERRAIOLI, PAOLO BUONAMICO, THOMAS MAHL, KENNETH J, W. TAYLOR, EMANUEL LERNER, OTTAVIO ALBANO, and ROBERTO J. GROSZMANN Hepatic Hemodynamic Laboratory and Research Service, Veterans Administration Medical Center, Yale University, West Haven, Connecticut; Department of Diagnostic Radiology, Yale University School of Medicine, New Haven, Connecticut; and Istituto di Clinica Medica I, Universith degli Studi di Bari, Bari, Italy
Propranolol, a nonselective j3-adrenergic blocker, has been shown to reduce portal pressure and the risk of variceal bleeding. The portal pressure-reducing effect of propranolol is mediated by splanchnic arterial constriction, which decreases portal flow. A double-blind randomized control study (crossover on 2 consecutive days) was designed to compare the effects of propranolol vs. placebo on portal flow in cirrhotic patients during fasting and after a standardized meal. Portal flow was measured with an ATL Ultramark 8 echoDoppler system (Advanced Technological Laboratories, Bethel, WA) in 23 cirrhotic patients. Fasting portal flow and heart rate were obtained at baseline and 2 hours after the administration of propranolol or placebo. A standard test meal was then given, and measurements were repeated 30 minutes later. Thirteen patients (group 1)received placebo on day 1 and propranolol on day 2, whereas 10 patients (group 2) received propranolol on day 1 and placebo on day 2. In group 1 patients, heart rate declined by 20% (P< 0.0001)and portal flow decreased by 12% (P < 0.05) after propranolol administration. Similar reductions were found in heart rate (-2l%, P < 0.0001) and portal flow (-17%, P < 0.001)for group 2 patients. For all 23 patients, 2 hours after propranolol administration, heart rate declined by 21% (P< 0.0001) and portal blood flow was reduced by 14% (P < 0.0001). The 10 patients who received propranolol on day 1 (group 2) showed a carryover effect of propranolol on day 2. On day 2, baseline portal flow and heart rate values were significantly lower than baseline values on day 1.This long-lasting effect of a single dose of propranolol may be caused by the longer half-life of propranolol in cirrhotic patients. The postprandial portal blood flow percentage increase after the meal was similar for both placebo and propranolol. Propranolol did not blunt postprandial hyperemia. However, whereas
the absolute value of blood flow after the meal increased significantly in comparison with baseline this did not in placebo-treated patients (P< O.OOl), occur with propranolol. Furthermore, in propranolol-treated patients the absolute value of blood flow after the meal was lower than in placebo-treated patients. This may constitute a protective effect of propranolol in portal hypertension.
It
has been shown that in cirrhosis, portal pressure states.“’ This inincreases during postprandial crease in portal pressure is presumably caused by an increase in gut blood (portal) flow, a phenomenon known as postprandial hyperemiaa3s4 In portal hypertensive patients, the increase in portal venous flow and pressure following the ingestion of food may, at times, be involved in inducing variceal bleeding. Propranolol, a nonselective P-adrenergic blocker reduces portal pressure by decreasing portal blood flow (F) and, by this mechanism, may prevent variteal hemorrhage.5 Currently, the wedge hepatic venous pressure is the only hemodynamic parameter that has proven to be useful in the long-term prognosis of cirrhotic patients treated with propranoloL6 However, hepatic vein catheterization is costly, time-consuming, and uncomfortable for patients. Echo-Doppler flowmetry is widely used for portal blood flow quantitation. Recently, we investigated the intraobserver and interobserver variability of study, we this technique.7 In an observer-blinded also showed the sensitivity of echo-Doppler in detecting hemodynamic changes caused by postprandial hyperemia and noted significantly different responses between normal subjects and cirrhotic patients.’ 0
1992
bytheAmerican
Gastroenterological 0016-5085/92/$3,00
Association
1010
SABBA ET AL.
Table 1. Baseline
GASTROENTEROLOGY Vol. 102,No.3
in Groups 1 and 2
Measurements
Group 1 Baseline day 1 Baseline day 2
A
V
F
N
(mm*)
(cm/s)
(mL/min)
13 13
189 + 7 186 ItI 7
10.8 f 0.7 11.0 f 0.8
1229 + 84 1225 f 76
77 + 5 78 f 4
134 + 4 131 f 4
10 10
192 + 6 193 + 7
13.0 + 0.8 11.0 + 0.6’
1509 + 69 1267 f 42’
83 f 2 71 + 2a
135 f 7 131 + 5
(b%
SBP (mm Hg)
Group 2 Baseline day 1 Baseline day 2
NOTE. For group 1,there are no statistically significant differences between portal vein parameters, HR, and SBP baseline values over the 2 days. For group 2, significant reductions in velocity, F, and HR are present 24 hours after propranolol administration (baseline day 2). “P < 0.01.
A, area: V, mean velocity: F, flow; HR, heart rate; SBP, systolic blood pressure; group I, placebo day 1, propranolol propranolol
day 2; group 2,
day 1, placebo day 2.
The current double-blind randomized hemodynamic study was designed (a) to evaluate the sensitivity of echo-Doppler in detecting acute changes in F in cirrhotic patients receiving a single dose of propran0101 and (b) to determine whether propranolol has any influence on portal postprandial hyperemia. In addition, because a persistence of the effect of a single dose of propranolol on heart rate (HR) for >24 hours has been reported,g”O this study was designed to explore whether this effect persists also on F.
female patients with a mean t SEM age of 59 + 2, weight of 72 + 3 kg, and height of 168 + 2 cm. Seventeen patients were examined at the Istituto di Clinica Medica I, University of Bari, and the remaining 6 patients at the West Haven Veterans Administration Medical Center, Clinical Campus of the Yale School of Medicine. The diagnosis of cirrhosis was based on liver biopsy findings in 8 of the 23 patients and on clinical data in the remaining patients, The etiology of liver cirrhosis was postnecrotic hepatitis B surface antigen-positive in 12 patients, alcoholic in 8, cryptogenic in 2, and hemochromatotic in 1.Cirrhosis was considered to be alcoholic in origin when alcohol consumption was estimated to be >80 g/day for >6 years preceding admission and no other cause was evident. According to Child-Turcotte criteria,” 8 patients were in class were A, 11 in class B, and 4 in class C. Measurements performed by G.F. at Veterans Administration Medical Center, West Haven, and by P.B. at Bari. The study was approved by the Human Investigation Committees of the West Haven Veterans Administration Medical Center and of Yale University School of
Materials and Methods Patients Twenty-six cirrhotic patients were entered into the study. Three of them were excluded; 2 because of poor portal vein visualization and 1 because bradycardia made the use of propranolol unsuitable. The remaining 23 cirrhotic patients were studied. They included 20 male and 3
90
day,* placebo 0
Z : 0. @
;r
min
Ii0
1
h
._ i Z ‘5’ 2
120
1
1
1300
1100
day
1
i
day2
7
!
0
-L:~
propranolol
placebo 900
a
120
min
0
120
’
Figure 1. (A) HR and (B) F are shown at baseline and 2 hours after placebo and propranolol administration in the 13 patients receiving placebo on day 1 and proprano101 on day 2 (group 1).'P < 0.05; *P < 0.0001 with respect to the same day baseline (paired t test).
PROPRANOLOL
March 1992
Table 2. Portal
Vein Parameters,
Day 1 placebo Baseline day 1 z h after placebo 30 min after meal Day 2 propranolol Baseline day 2 2 h after drug 30 min after meal
ON PORTAL BLOOD FLOW
1011
HR, and SBP in Group 1 HR (bpml
F (mL/min)
N
A (mm’)
v (cm/s)
13 13 13
189+ 7 191f 7 194 + 6
10.8 f 0.7b 10.8 + 0.6",' 13.8 f 1.0".b.'
1229f84b 1246rf-81",k 1602 + 107°,b,'
77 * 5c 75 f 5"J
13 13
186 * 7 187 + 8
11.0 f 0.8" 9.6 k O.GeXi
1225 It76e 1072 + 77e,f,k
78 k 4g 62 + 3g,'
13
189f
11.8 f 1.G'
1388 + 93L'
7
SBP (mm Hgl 134 +4d 130 * 5d.l
131&4h 116 f 3h.'
Statistical significance (paired t test): c.e.i.iP< 0.05;df.h.LP< 0.01; a.b,‘P< 0.001; PP < 0.0001. A, Area; V, mean velocity; F, portal blood flow; HR, heart rate; SBP, systolic blood pressure.
Medicine. Informed written consent was obtained each patient, both in Bari and West Haven.
plex system calculated the area using the following ellipse formula: A = nD,D,/4. Velocity measurements were performed in quiet suspended inspiration and averaged over a few seconds. They were taken on a longitudinal section of the vessel positioning the sample volume cursor within the vessel at a known angle of insonation (angle between the Doppler beam and the long axis of the vessel). This angle was always ~60”. The Doppler sample volume was positioned in the center of the vessel and was manipulated to cover about 90% of the vessel section to detect the maximum Doppler frequency shift (Af), related to the maximum velocity. According to the Doppler equation
from
Equipment Studies were performed at both centers using an ATL Ultramark 8 echo-Doppler duplex system (Advanced Technological Laboratories, Bethel, WA) with a 3.0-MHz sector mechanical probe. Real-time and Doppler control settings were optimized for each patient according to established techniques.
Calculations Portal vein cross-sectional area (A, square millimeters) and mean blood velocity (V, centimeters per second) were measured to calculate blood flow (F, milliliters per minute) according to the following formulas: F = A X V X 60. Area was obtained from a cross-section of the vessel visualized by B-mode after defining the major and minor diameters of the vessel. The computer software of the du-
Af
cos 8 , 2fOv”MX
=
C
f, is the incident Doppler frequency, 8 is Doppler angle of insonation, and c is the velocity of the ultrasound in tissues. The mean velocity was calculated from maximum
1800
day A
1500 Meal placebo
Figure 2. The effects are shown for placebo and propranolol on postprandial F in the 13 patients receiving placebo on the first day and propranolol on the second day (group 1).*P < 0.05between same time interval measurements on 2 consecutive days (paired t test).
\
propranolol
900
’
I
0
120 TIME
bmin)
Meal
150
1
1012
SABBA ET AL.
GASTROENTEROLOGY Vol. 102. No. 3
Table 3. Portal Vein Parameters,
Day 1 propranolol Baseline day 1 2 h after drug 30 min after meal Day 2 placebo Baseline day 2 2 h after placebo 30 min after meal
HR, and SBP in Group 2 SBP
A
V
N
(mm’1
(cm/s)
F (mL/min)
10 10 10
192 + 6 192 + 7 194 k 6
13.0 f 0.8a.h 10.8 f 0.7°.b 12.3 It O.gb
1509 + 69C,h 1243 + 62Csb 1517 + 91b
83 f Zmeh 65 f 4’ -
135 + 7c 119 f 5c
10 10 10
193 f 7 191 + 7d 200 + 7d
11.0 f 0.6h 11.0 f 0.7e 13.5 k l.ge
1267 f 42h 1264 ?I 59” 1566 + 156e
71 + 2bh 67 !I Zf -
131+ 58 119 + 48 -
(mm Hg)
-
< 0.01; cP < 0.001; OP < 0.0001. Statistical significance (paired t test): b.eafP< 0.05; d*g,hP A, Area; V, mean velocity; F, portal blood flow; HR, heart rate: SBP, systolic blood pressure.
unaware of the effect of propranolol or placebo on arterial pressure and HR. If HR had not decreased by 20%, another dose of 40 mg of propranolol was administered [unless HR was ~60 beats per minute (bpm) or SBP < 90 mm Hg]. Two hours after propranolol or placebo administration, echoDoppler measurements of portal vein parameters were repeated. Then a standard mixed liquid meal (355 kcal, Ensure plus; Ross Laboratories, Columbus, OH) was given, and the same measurements were performed 30 minutes later. The criterion for inclusion of a subject in the data analysis was feasibility of portal vein measurements by echoDoppler, i.e., adequate visualization of the vessel in two planes with a Doppler angle of insonation of lO% and nonresponders if it was 10% (range, -10% to -41%), and 5 of them did not respond (range, -3% to f14%). Compared with the responders, the nonresponders to propranolol had a smaller increase in F after the meal both on the propranolol(50% + 10% vs. 10% + 14%; P < 0.05)and placebo day (40% f 7% vs. 16% + 6%; P < 0.05). Group 2 The baseline values of the 10 patients receiving propranolol on the first day and placebo 24 hours later are shown on Table 1. In comparison with the first day, a significantly lower baseline HR and F were detected on the second day (P < 0.01 and P < 0.01, respectively), indicating a probable residual effect of propranolol administered on day 1 (Figure 3). When propranolol was given, HR, SBP, and F decreased by 2l%, 11% and 17% respectively (P < 0.001). After placebo administration, F did not change significantly; however, HR and SBP were significantly decreased (P < 0.05; Table 3 and Figure 3). In terms of percentage changes from baseline, the reductions in HR and in F induced by propranolol were significantly greater than after placebo (P < 0.001 and P < 0.01, respectively). No significant percentage change was observed in SBP. There were no statistically significant percentage differences in the behavior of all three parameters (HR, F, and SBP) when group 1 and group 2 were compared with each other (Figure 4).
GASTROENTEROLOGYVol.102.No.3
Probably because of different baseline values on day 1 and day 2, no statistically significant differences were observed on F and HR between absolute values attained at 120 minutes after either proprano101(day 1) or placebo (day 2) administration (Figure 3; Table 3). The F changes after meal ingestion were similar to those observed in group 1 (Tables 2 and 3). Group 1 Plus Group 2 For all 23 patients, in comparison with baseline, HR declined by 21% (P < O.OOOl),SBP decreased by 11% (P < O.OOOl),and F was reduced by 14% (P < 0.0001) 2 hours after propranolol administration. In comparison with placebo, propranolol decreased HR by 10% (P < 0.01) SBP by 6% (P < O.Ol), and F by 8% (P < 0.05) at 120 minutes. After the meal, there was a similar and significant percentage increase in F; however, the peak blood flow reached in response to the meal in comparison with baseline F was significantly higher in the placebo group (P < 0.0001) but not in the propranolol group. Despite a 20% difference in baseline F between the two groups, the findings between the total (n = 23) and group 1 (n = 13)and 2 (n = 10) were similar in the HR, SBP, and F response to propranolol. The variability for F on the same day between baseline and 120 minutes (n = 13) ranged from -13.1% to18.2%withamean+SEMof1.8%+2.5%. The variability on consecutive days between baselines ranged from -10.8% to 11.9% with a mean of 0.4% f 1.8% (mean + SEM). No correlation was found between change in HR or blood pressure and F. Discussion In this double-blind ultrasonography and patient study, echo-Doppler flowmetry was able to detect changes in portal hemodynamics after propran0101 administration in cirrhotic patients. The hemodynamic effect of a single dose of propranolol on F appeared to persist for >24 hours. Indeed, when the drug was given the first day, decreases in HR, SBP, V, and F were still present on the second day. The impaired liver-clearing capacity of cirrhosis is probably responsible for maintaining the effects of this drug beyond the usual half-life.‘5*‘” Propranolol is metabolized during a first pass through the liver, and it has been shown that the half-life of propranolol and its effect on HR are extended in patients with liver cirrhosis.g.‘o However, it has not been shown previously that propranolol has a prolonged effect on portal blood flow. Changes in flow were mainly related to changes in velocity. Previous studies by this laboratory and by another
PROPRANOLOL
March 1992
center regarding blood flow changes in the portal vein in response to a meal have reported similar reSUltS.8*‘7 In this study, we observed that, in cirrhotic patients, the postprandial hyperemic responses are similar after propranolol or placebo. This indicates that P-adrenergic blockade does not interfere with the basic mechanism that induces the postprandial hyperemic state. Because of the lower fasting F induced by propranolol, the absolute value of postprandial F was significantly lower on propranolol day despite a similar percentage increase. Furthermore, the absolute increase in F from baseline to the postprandial 150-minute measurement was significantly greater in the placebo group but was not significantly increased in the propranolol group (Figure 2; Tables 2 and 3). This may constitute a protective effect of propranolol in portal hypertension. A previous study showed that propranolol does not prevent the postprandial increase in portal pressure in cirrhotic patients.’ Our results are consistent with physiological studies on animals that have shown that (3-adrenergic blockade does not affect the postprandial hyperemia.” Six of the 23 patients (5 in group 1 and 1 in group 2) did not respond to propranolol with a decrease from baseline F > 10%. These patients showed a significantly reduced postprandial hyperemic response with and without propranolol. This finding suggests that the basic mechanism leading to a lack of response to propranolol and to a meal stimulus may be related. It has been shown previously that (j-adrenergic blockers are not effective in reducing portal pressure in all cirrhotic patients.‘4r’g These results coincide with the portal flow response to propranolol in this study. In this as in a previous report,14 there were no significant differences in any baseline clinical or hemodynamic features that will distinguish responders from nonresponders. Previous studies have shown limitations with the use of the echo-Doppler when studying F.7,‘3 In view of our experience with the variability of echoDoppler technique for studying F,7 we used the mean of three measurements at each time interval for each parameter. In the current and in previous studies, it was observed that in portal hypertensive patients, the cross-sectional area of the portal vein showed small variations in response to changes in F.20*21 Therefore, it may be possible to measure changes in blood velocity without the additional errors caused by calculation of portal vein area. In conclusion, this study shows that the absolute value of blood flow after a meal in propranololtreated patients is reduced in comparison with placebo. This, in itself, may be an important component
ON PORTAL BLOOD FLOW
1015
of the (3-adrenergic blocker-protective effect in the prevention of variceal bleeding. In addition, the current study shows that, under controlled conditions, echo-Doppler is able to detect the effect of a single dose of propranolol on portal blood flow. Additional studies are needed to relate hemodynamic findings with clinical events. References 1. Lee S, Handengue
A, Moreau R, Hillon P, Lebrec D. Post-prandial hemodynamic responses in patients with cirrhosis. Hepatology 1988;8:647-651. 2. McCormack PA, Afdhal NH, Hegarty JE. Effect of propranolol on post-prandial increase in portal venous pressure in patients with portal hypertension (abstr). Gastroenterology 1987;92:1755. 3. Chou CC. Splanchnic
and overall cardiovascular hemodynamics during eating and digestion. Fed Proc 1983;42:1658-
1661. 4. Gallavan
RH, Chou CC. Possible mechanisms for the initiation and maintenance of post-prandial intestinal hyperemia. Am J Physiol 1985;249:G301-G308. 5. Pagliaro L, Burroughs AK, Sorensen TIA, Lebrec D, Morabito A, D’Amico G, Fine T. Therapeutic controversies and randomized controlled trials (RCTs): prevention of bleeding and rebleeding in cirrhosis. Gastroenterol Int 1989;2:71-84. 6. Groszmann RJ, Bosch J, Grace N, Conn HO, Garcia-Tsao G, Navasa M, Alberts J, Rodes J, Fischer R, Bermann M, Rofe S, Patrick M, Lerner E. Hemodynamic events in a prospective randomized trial of propranolol vs. placebo in the prevention of the first variceal hemorrhage. Gastroenterology 1990;99: 1401-1407. 7. SabbL C, Weltin GG, Cicchetti
DV, Ferraioli G, Taylor KJW, Nakamura T, Moriyasu F, Groszmann RJ. Observer variability in echo-Doppler measurements of portal flow in cirrhotic patients and normal volunteers. Gastroenterology 1990;98: 1603-1611.
8. Sabba C, Ferraioli
9.
10.
11. 12.
13. 14.
G, Genecin P, Colombato L, Buonamico P, Lerner E, Taylor KJW, Groszmann RJ. Evaluation of postprandial hyperemia in superior mesenteric artery and portal vein in healthy and cirrhotic humans: an operator-blind study. Hepatology 1991;13:714-718, Arthur MJSP, Tanner AR, Plate1 C, Wright R, Renwick AG, George CF. Pharmacology of propranolol in patients with cirrhosis and portal hypertension. Gut 1985;26:14-19. Wood AJJ, Kornhauser DM, Wilkinson GR, Shand DG, Branc RA. The influence of cirrhosis on steady state blood concentrations of unbound propranolol after oral administration. Clin Pharmacokin 1978:3:478-487. Conn HO. A peek at the Child-Turcotte classification. Hepatology 1981;1:673-676. Moriyasu F, Ban N, Nishida 0, Nakamura T, Miyake T, Uchino H, Kanimatsu Y, Koizumu S. Clinical application of an ultrasonic duplex system in the quantitative measurement of portal blood flow. J Clin Ultrasound 1986;14:579-588. Gill RW. Measurement of blood flow by ultrasound: accuracy and source of error. Ultrasound Med Biol 1985;11:625-641. Garcia-Tsao G, Grace ND, Groszmann RJ, Conn HO, Bermann NN, Patrick MJC, Morse SS, Alberts JL. Short-term effects of propranolol on portal venous pressure. Hepatology 1986; 6:101-106.
15. Paterson
JW, Conolly ME, Dollery CT, Hayes A, Cooper RG. The pharmacodynamics and metabolism of propranolol in man. Pharmacol Clin 1976;2:127-133.
1016
SABBA ET AL.
16. Pessaye D, Lebrec D, Descatoire V, Peignaux M, Benhamou JP. Mechanism for reduced drug clearance in patients with cirrhosis. Gastroenterology 1978;74:566-571. 17. Gaiani G, Bolondi L, Li Bassi S, Santi V, Zironi G, Barbara L. Effect of meal on portal hemodynamics in healthy humans and in patients with chronic liver disease. Hepatology 1989;9:815-819. 18. Takagi T, Naruse S, Shionoya SS. Post-prandial celiac and superior mesenteric blood flows in conscious dog. Am J Physiol 1988;255:G522-G528. 19. Vorobioff J, Picabea E, Villavicenzio R, Puccini V, Rossi 0, Bordato J, Audato M. Acute and chronic hemodynamic effects of propranolol in unselected cirrhotic patients. Hepatology 1987;7:648-653. 29. Bolondi L, Gandolfi C, Arienti V, Caletti GC, Corcioni E, Gasbarrini G, Labd G. Ultrasonography in the diagnosis of portal
GASTROENTEROLOGY Vol. 102. No. 3
hypertension: diminished response of portal vessels to respiration. Radiology 1982;142:167-172. 21. Sabba C, Buonamico P, Ferraioli G, Taylor KJW, Groszmann RJ. Measurement of mean velocity to monitor acute portal blood flow changes in portal hypertensive patients (abstr). Gastroenterology 1990;98:A199.
Received November 13, 1990. Accepted August 12,1991. Address requests for reprints to: Roberto J. Groszmann, M.D., Veterans Administration Medical Center, 950 Campbell Avenue, West Haven, Connecticut 06516. Dr. Ferraioli’s present address is Divisione di Medicina, Ospedale Civile di Scafati, Scafati 84018, Italy. The authors gratefully acknowledge Marge Petrucci for her secretarial assistance.
