Effects of alpha-adrenergic stimulation and beta-adrenergic blockade on azygos blood flow and splanchnic haemodynamics in patients with cirrhosis

The effects of beta-blockade with propranolol and of alpha-adrenergic stimulation with methoxamine, a powerful alpha-agonist, on azygos blood flow and on systemic and hepatic haemodynamics were investigated in 26 cirrhotic patients with portal hypertension. Beta-adrenergic blockade with propranolol (n = 12), evidenced by a significant reduction of heart rate (-17 +/- 1%, P less than 0.001) and cardiac index (-17 +/- 2%, P less than 0.001), caused a mild but significant decrease of hepatic venous pressure gradient (-10 +/- 2%, P less than 0.05) and a marked fall of azygos venous blood flow (-31 +/- 5%, P less than 0.05). Alpha-adrenergic stimulation with methoxamine (n = 14), manifested by a significant increase of mean arterial pressure (19 +/- 2%, P less than 0.001), mimicked the effects of propranolol on hepatic venous pressure gradient (-10 +/- 4%, P less than 0.05) and cardiac index (-11 +/- 2%, P less than 0.001). However, azygos blood flow was not significantly reduced by methoxamine (0.7 +/- 0.1 vs 0.6 +/- 0.1 l/min). On the contrary, hepatic blood flow was significantly reduced by methoxamine (-19 +/- 4%, P less than 0.01) but not by propranolol (-7 +/- 7%, ns). Similarly, in 8 patients who received methoxamine after being beta-blocked by propranolol, azygos blood flow, that was markedly reduced by beta-blockade, did not experience a further reduction but increased slightly by alpha-adrenergic stimulation, while hepatic blood flow, that was not reduced by propranolol, decreased significantly during the subsequent methoxamine infusion.(ABSTRACT TRUNCATED AT 250 WORDS)     

Effect of oral propranolol on circulating catecholamines in cirrhosis: relationship to severity of liver disease and splanchnic haemodynamics

Patients with cirrhosis, especially those with decompensated disease have enhanced sympathetic nervous activity. We have investigated the effect of a single oral dose of 80 mg propranolol on circulating catecholamines and related the effect to splanchnic and systemic haemodynamics in 22 patients with cirrhosis. Plasma noradrenaline (NA) was significantly above normal average (NA: 0.52 vs. 0.23 ng/ml, p less than 0.01) and increased with the severity of the liver disease (p less than 0.01). NA was negatively correlated with liver function as estimated by ICG clearance (r = -0.74, p less than 0.01). Azygos blood flow was increased (0.75 l/min) and positively related to plasma NA (r = 0.57, p = 0.05, n = 12). After propranolol intake, plasma NA increased from 0.52 to 0.59 ng/ml (p less than 0.01). This response was found in all Child-Turcotte classes (A: 0.37 to 0.43; B: 0.49 to 0.56; C: 0.78 to 0.88 ng/ml), and in patients with as well as without ascites. Plasma adrenaline increased in the same way (p less than 0.01). Hepatic blood flow (from 1.10 to 0.93 l/min, p less than 0.01) and azygos blood flow (from 0.75 to 0.55 l/min, n = 9, p less than 0.05) decreased significantly after oral propranolol. A borderline significant correlation was observed between the decrease in azygos blood flow and the increase in NA (r = 0.64, p = 0.06). Our results suggest that besides a relationship to liver function and severity of disease, sympathetic nervous activity, as reflected by circulating NA, will further enhance during beta-adrenergic blockade, probably by a compensatory mechanism.(ABSTRACT TRUNCATED AT 250 WORDS)     

Effect of oral propranolol on splanchnic oxygen uptake and haemodynamics in patients with cirrhosis

In order to elucidate the effect of beta-adrenergic blockade on liver metabolism and haemodynamics, splanchnic oxygen uptake, hepatic removal of indocyanine green (ICG) and splanchnic and systemic haemodynamics were studied in 13 patients with cirrhosis before and 1.5-2 h after an oral dose of 80 mg propranolol. All patients underwent hepatic vein catheterization and had a primed continuous intravenous infusion of ICG. Azygos vein catheterization was performed in six patients. Splanchnic (hepatic-intestinal) oxygen uptake (median control 68 ml/min vs. beta-blockade 56 ml/min, P less than 0.01), azygos venous oxygen saturation (76 vs. 67%, P less than 0.05), ICG clearance (263 vs. 226 ml/min, P less than 0.01), wedged-to-free hepatic vein pressure (16 vs. 13.5 mm Hg, P less than 0.01), hepatic blood flow (1.18 vs. 0.78 l/min, P less than 0.01), cardiac index (3.42 vs. 2.53 l/min . min 2, P less than 0.01), and heart rate (72 vs. 56 beats per min, P less than 0.01) decreased significantly after oral beta-blockade. The hepatic extraction ratio of ICG increased significantly (0.32 vs. 0.45, P less than 0.01), whereas estimated 'intrinsic' ICG clearance (289 vs. 300 ml/min, n.s.), arterial blood pressure, stroke volume, and systemic vascular resistance remained essentially unchanged. The results indicate that besides the well-known cardiovascular effects of propranolol, beta-adrenergic blockade may also reduce hepatic metabolic functions as evidenced by the significantly decreased splanchnic oxygen uptake. The raised hepatic extraction ratio of ICG may be caused by reduction in hepatic blood flow as well as in intrahepatic shunting.     

Selective dopamine DA1 stimulation with fenoldopam in cirrhotic patients with ascites: a systemic, splanchnic and renal hemodynamic study.

We studied the effects of fenoldopam, a selective dopamine DA1 agonist on systemic and splanchnic hemodynamics, renal blood flow and sodium excretion in 12 patients with alcoholic cirrhosis and ascites. Hepatic, azygos and renal veins were catheterized before and after intravenous administration of fenoldopam, 0.05 micrograms/kg/min for 1 hr and increased to 0.1 micrograms/kg/min for another hour. Mean arterial pressure progressively decreased (from 83 +/- 7 to a minimum of 77 +/- 8 mm Hg 100 min after starting the infusion) but returned to baseline level at 120 min. Plasma norepinephrine and renin activity increased (respectively from 567 +/- 297 to 919 +/- 375 pg/ml, p less than 0.05, and from 17 +/- 14 to 23 +/- 15 ng/ml/hr, p less than 0.05). Renal blood flow, urine output or sodium excretion did not change. Sodium output decreased at 1 hr from 6.9 mumol/min to 4.0 mumol/min, p less than 0.05. Both hepatic venous pressure gradient and azygos blood flow significantly increased by 21%. We conclude that the acute administration of fenoldopam did not improve renal hemodynamics or function in patients with cirrhosis and ascites. In addition, dopamine DA1 agonism caused further increases in norepinephrine concentration and plasma renin activity. Portal pressure also increased, probably because of an increase in mesenteric blood flow. These results question the renal benefit and raise concern about the use of dopamine agonists in patients with cirrhosis and ascites.     

Measurement of azygos venous blood flow in the evaluation of portal hypertension in patients with cirrhosis. Clinical and haemodynamic correlations in 100 patients

Blood flow in the azygos vein, an index of blood flow through gastro-oesophageal collaterals, was measured by continuous thermal dilution in 100 patients with cirrhosis. Azygos blood flow was directly related to portal pressure (r = 0.54, P less than 0.001). Patients with portal hypertension had very high azygos blood flow (692 +/- 32 ml/min) in comparison with controls (n = 11, 174 +/- 29 ml/min). Patients with previous oesophageal bleeding had similar azygos blood flow as those without, but azygos blood flow was significantly greater in patients with massive or recurrent bleeding than in those with less severe haemorrhage, suggesting that the magnitude of collateral flow may influence the course of variceal bleeding. Patients with grade III varices had higher azygos blood flow than those with grades II or I. In addition, both oesophageal tamponade and vasopressin infusion, procedures of known value in variceal bleeding, markedly reduced azygos blood flow (-40% and -25%, respectively). Measurement of azygos blood flow allows evaluation of haemodynamic changes in the oesophageal collaterals of patients with portal hypertension, and provides useful information on the effect of therapeutic procedures aimed at arresting or preventing variceal haemorrhage.     

Hyperglucagonism and glucagon resistance in cirrhosis. Paradoxical effect of propranolol on plasma glucagon levels

Propranolol, a non-selective beta-blocker, is known to decrease glucagon release in normal subjects. The present study was aimed at investigating the effects of propranolol on the hyperglucagonism commonly observed in patients with cirrhosis. Eight cirrhotic patients and 6 matched healthy controls were studied. The plasma concentrations of glucagon, insulin, c-peptide and glucose were measured in basal conditions and after stimulating glucagon secretion by an i.v. infusion of arginine (0.4 g/kg/30 min). The study was repeated 24 h later after inducing beta-blockade by the i.v. infusion of propranolol (10 mg). In baseline conditions, patients with cirrhosis, despite normal levels of insulin and glucose, had a marked hyperglucagonism (654 +/- 303 pg/ml vs. 269 +/- 90 in controls, P less than 0.01). Prior to propranolol, arginine infusion caused greater glucagon release in cirrhotics (71 +/- 31 ng.h.ml-1) than in controls (33 +/- 17 ng.h.ml-1, P less than 0.02), but despite a similar insulin secretion (assessed from c-peptide), blood glucose did not increase. After propranolol, glucagon secretion decreased as expected in controls (29 +/- 12 ng.h.ml-1, P less than 0.05) but experienced a paradoxical increase in cirrhotics (113 +/- 64 ng.h.ml-1, P less than 0.05). Again, despite the marked increase in glucagon release, there was no increase in glucose production, providing further evidence of the glucagon resistance that accompanies hyperglucagonism in cirrhosis. Our results suggest that hyperglucagonism with glucagon resistance might be the initial disturbance in carbohydrate metabolism in patients with cirrhosis. Contrary to what could be expected, propranolol does not correct but further accentuates this disturbance.     

Intact hepatocyte theory of impaired drug metabolism in experimental cirrhosis in the rat.

The elimination of propranolol by perfused livers of rats made cirrhotic by chronic carbon tetrachloride inhalation during phenobarbital treatment has been compared with control animals receiving only phenobarbital. Cirrhosis reduced propranolol clearance at a constant flow of 20 ml/min from 1.43 +/- 0.08 to 1.12 +/- 0.08 ml/min/g liver (P less than 0.025). In addition, an increase in intrahepatic shunting of 15-micron microspheres from 0.41 +/- 0.01 to 9.4 +/- 4.1% was found in cirrhotic livers (P less than 0.05). Finally, in cirrhotic livers, reducing blood flow did not produce the normal rise in hepatic extraction ratio, which actually fell from 0.873 +/- 0.021 at 20 ml/min to 0.836 +/- 0.025 at 15 ml/min and 0.823 +/- 0.026 at 10 ml/min. At each flow the observed extraction was significantly lower than that predicted to result from a reduced enzyme activity alone, consistent with the development of functionally significant intrahepatic shunts. An operational model is proposed that explains impaired drug metabolism in cirrhosis on the basis of the development of intrahepatic shunts which perfuse nonfunctioning tissue, while the remaining blood flow is exposed to a reduced mass of hepatocytes with an apparently normal amount of drug metabolizing enzyme (the intact hepatocyte theory).     

Enhancement of portal pressure reduction by the association of isosorbide-5-mononitrate to propranolol administration in patients with cirrhosis

This study investigated whether oral doses of isosorbide-5-mononitrate, a preferential venous dilator that decreases portal pressure, could enhance the effects of propranolol on portal hypertension. Taking part in the study were 28 patients with cirrhosis and portal hypertension. Twenty patients (group 1) had hemodynamic measurements in baseline conditions after beta-blockade by intravenous administration of propranolol and after receiving oral doses of isosorbide-5-mononitrate. The remaining eight patients (group 2) were given oral isosorbide-5-mononitrate while receiving chronic propranolol therapy. In group 1, propranolol significantly reduced portal pressure (estimated as the gradient between wedged and free hepatic venous pressures) from 21.5 +/- 3.9 to 18.6 +/- 4.2 mm Hg (-13.7%, p less than 0.001), azygos blood flow (-38%, p less than 0.001), hepatic blood flow (-12.8%, p less than 0.05), cardiac output (-24.5%, p less than 0.001) and heart rate (-18.4%, p less than 0.001) without significant changes in mean arterial pressure. Addition of oral isosorbide-5-mononitrate caused a further and marked fall in portal pressure (to 15.7 +/- 3.1 mm Hg, p less than 0.001), without additional changes in azygos blood flow but with significant additional reductions in hepatic blood flow (-15.5%, p less than 0.05), cardiac output (-11.5%, p less than 0.001) and mean arterial pressure (-22%, p less than 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)     

Effect on hemodynamics of a liquid meal alone and in combination with propranolol in cirrhosis.

Thirteen patients with alcoholic cirrhosis had splanchnic and systemic hemodynamics assessed before and after ingestion of a standard liquid meal of 700 kcal (consisting of isocaloric proteins, lipids, and carbohydrates). Half of the patients (n = 6) were randomized to a treatment group receiving intravenous infusion of propranolol in combination with the meal. No significant effects were observed on systemic hemodynamics after the meal alone. Heart rate (-14%; P less than 0.01) and cardiac index (-24%; P less than 0.01) decreased after meal in combination with propranolol. The mean hepatic venous pressure gradient increased significantly after ingestion of the meal alone with a maximal effect after 30 minutes (+13%; P less than 0.05) and returned to baseline values after 2 hours. Meal in combination with propranolol had no significant effect on the hepatic venous pressure gradient. Hepatic blood flow increased substantially after the meal alone with a maximal effect after 30 minutes (+28%; P less than 0.01), whereas no significant effect was observed after meal in combination with propranolol. Azygos blood flow increased significantly after the meal alone (+36%; P less than 0.05), whereas this effect was abolished in combination with propranolol. In conclusion, ingestion of a peroral mixed meal in cirrhotic patients has, contrary to what is observed in normal controls, no effects on systemic hemodynamics. Substantial changes in splanchnic hemodynamics were observed, and these effects were all abolished when the meal was administered in combination with propranolol.     

Propranolol decreases portal pressure without changing portocollateral resistance in cirrhotic rats

Propranolol decreases portal pressure by reducing portal blood inflow. Studies in rats with prehepatic portal hypertension due to portal vein stenosis (a model with extensive portosystemic shunting) have shown that propranolol increases the portocollateral resistance, which hinders the fall in portal pressure. The present study examined the effects of propranolol on splanchnic and systemic hemodynamics in rats with portal hypertension due to cirrhosis of the liver, a model which is characterized by mild portosystemic shunting. Two groups of rats with CCl4-induced cirrhosis were studied: the propranolol group (n = 8), which received a propranolol infusion of 2 mg per 15 min, and controls (n = 9), which received a placebo (saline) infusion. Hemodynamic measurements were done using radiolabeled microspheres. Propranolol-treated rats had significantly lower cardiac output (-31%) and heart rate (-26%) than controls (p less than 0.001). Hepatic artery flow was not modified by propranolol. Propranolol caused splanchnic vasoconstriction, manifested by increased splanchnic resistance (+57%) and by a significant fall in portal blood inflow (4.8 +/- 0.4 vs. 6.3 +/- 0.5 ml per min.100 gm in controls, p less than 0.05). In contrast with rats with prehepatic portal hypertension, propranolol did not increase portal resistance in cirrhotic rats [2.0 +/- 0.2 vs. 2.0 +/- 0.1 mmHg per ml per min.100 gm body weight (not significant)]. Hence, the fall in portal pressure (-19%) was expected from the decrease in portal inflow (-24%). These results suggest that increased portal resistance in rats with prehepatic portal hypertension may represent an intrinsic effect of propranolol on the portocollateral vessels, since beta-blockade does not modify portal vascular resistance in cirrhosis.     

Combination of ketanserin and verapamil or propranolol in patients with alcoholic cirrhosis: search for an additive effect

Drugs reported to reduce portal pressure through different mechanisms were combined in the hope of either additive portal hypotensive effects in "responders," or inducing a portal hypotensive effect in "nonresponders" to the initial drug. Seven patients with alcoholic cirrhosis received verapamil, 10 mg i.v., and, 60 min later, ketanserin, 5 mg i.v. Verapamil decreased heart rate and increased free hepatic venous pressure but had no effect on hepatic venous pressure gradient or azygos blood flow. When combined with verapamil, ketanserin significantly diminished wedged hepatic venous pressure and hepatic venous pressure gradient. Ten other patients with alcoholic cirrhosis received propranolol, 15 mg i.v., and 45 min later, ketanserin, 5 mg i.v. In all patients, heart rate, cardiac index and azygos blood flow significantly decreased after propranolol. After propranolol alone, however, wedged hepatic venous pressure decreased in only five patients, responders. In five other patients, defined as nonresponders, propranolol did not decrease this pressure. The addition of ketanserin to propranolol induced further significant reduction in wedged hepatic venous pressure, hepatic venous pressure gradient and azygos blood flow. Among the five nonresponders, three had a reduced wedged hepatic venous pressure after ketanserin was combined. We conclude that verapamil does not reduce portal pressure or collateral blood flow in patients with alcoholic cirrhosis. The splanchnic hemodynamic effects of propranolol and ketanserin appear to be independent and additive, without significant systemic alteration.     

Effects of propranolol and nitroglycerin on cephalad collateral venous flow in patients with cirrhosis: evaluation using transesophageal real-time two-dimensional Doppler echography

We used transesophageal real-time two-dimensional Doppler echography (TE2DD) to assess the effects of propranolol (n = 18, 6 mg each) and nitroglycerin (n = 18, 0.5 mg each) on blood flow in the intercostal veins, azygos vein, thoracic aorta, and esophagogastric varices. The primary disease in all of the patients was liver cirrhosis. Propranolol infusion markedly reduced the flow velocity in the varices, intercostal vein, azygos vein, and thoracic aorta (-24%, -41%, -34%, and -24%, respectively). It also significantly reduced the blood flow volume index (BFVI), defined as mean velocity in cm/sec X the square of the diameter in cm2 of both the azygos vein and the aorta (-34%, -21%, respectively). Nitroglycerin infusion did not cause significant changes in the hemodynamics of the above vessels, because the hemodynamic responses to the drug differed from individual to individual. The BFVI of the azygos vein correlated well with the azygos venous flow measured by the conventional thermodilution technique (r = 0.79, p less than 0.01). TE2DD appears to be a useful method for studying the hemodynamics of ascending collaterals in patients with portal hypertension.     

Total effective vascular compliance in patients with cirrhosis: a study of the response to acute blood volume expansion.

Although arterial vasodilation is a well-known feature in patients with cirrhosis, the venous system remains unexplored. To measure total effective vascular compliance, a reflection of the properties of the venous system, rapid volume expansion (300 ml of a gelatin solution in 3 min) was performed in 23 patients. Eleven patients had compensated cirrhosis (Child-Pugh grade A or B), and eight had decompensated cirrhosis (Child-Pugh grade C). Four control patients had mild chronic hepatitis, normal hepatic venous pressure and normal liver architecture. Cardiac index, hepatic venous pressures, hepatic and azygos blood flow and renal plasma flow were measured before and immediately after volume expansion. Right atrial pressure was recorded during volume expansion. This allowed the calculation of total effective vascular compliance, which was higher in patients with decompensated cirrhosis than in those with compensated cirrhosis (4.65 +/- 4.21 vs. 1.34 +/- 0.63 ml.mm Hg-1.kg-1; p less than 0.05). In response to volume expansion, renal vascular resistance decreased significantly in patients with compensated cirrhosis, but not in those with decompensated cirrhosis (-30% +/- 33% vs. +2% +/- 23%; p less than 0.05). No change was seen in glomerular filtration rate. Systemic oxygen consumption increased in patients with compensated cirrhosis, but not in those patients with decompensated cirrhosis (25% +/- 33% vs. -4% +/- 9%; p less than 0.05). Although in all patients with cirrhosis volume expansion increased central venous pressures, azygos blood flow and the hepatic venous pressure gradient did not change.(ABSTRACT TRUNCATED AT 250 WORDS)     

Propranolol compared with propranolol plus isosorbide-5-mononitrate for portal hypertension in cirrhosis. A randomized controlled study

OBJECTIVE: To investigate whether isosorbide-5-mononitrate (Is-5-Mn) given with propranolol reduces hepatic portal pressure more than does propranolol alone in patients with cirrhosis. DESIGN: A randomized controlled trial. PATIENTS: Fifty patients with cirrhosis and esophageal varices entered and 42 completed the study. INTERVENTION: Twenty-one patients received oral propranolol at increasing doses until their resting heart rate was reduced by 25%, and 21 patients received oral propranolol (on the same schedule) plus oral Is-5-Mn, 40 mg twice a day. MEASUREMENTS: Hepatic vein pressure gradient, liver function, and splanchnic and systemic hemodynamics before and after 3 months of continuous therapy. MAIN RESULTS: At 3 months, the hepatic venous pressure gradient decreased more (P less than 0.01) in patients given propranolol plus Is-5-Mn (19%, from 18.4 +/- 3.9 to 14.9 +/- 3.8 mm Hg; 95% CI, -2.4 to -4.5 mm Hg) than in those given propranolol alone (10%, from 18.2 +/- 3.5 to 16.3 +/- 3.1 mm Hg; CI, -1.1 to -2.7 mm Hg). The hepatic venous pressure gradient decreased by more than 20% of the baseline value in 10% of patients receiving propranolol, but in 50% of patients receiving combined therapy (P less than 0.02). There were statistically significant decreases in hepatic blood flow and the intrinsic clearance of indocyanine green after propranolol therapy, but not after combined therapy. The treatments caused similar reductions in azygos blood flow and cardiac output. CONCLUSIONS: The long-term combined administration of propranolol plus Is-5-Mn reduces portal pressure more than propranolol alone without adverse effects on hepatic perfusion and liver function. Whether this greater hemodynamic effect translates into better clinical efficacy should be determined in randomized controlled trials.     

Des-1-Asp-angiotensin II. Possible intrarenal role in homeostasis in the dog.

The relative effects of des-a-Asp-angiotensin 3I and angiotensin II on renal function, including renin secretion, were investigated in normal and sodium-depleted dogs. Intrarenal arterial infusion of the heptapeptide fragment into normal dogs at a rate which was calculated to increase blood levels by only 7 ng/100 ml decreased renal blood flow from 254 +/- 9 ml/min to 220 +/- 12 and 219 +/- 12 ml/min (P less than 0.01 for both values) after 10 and 30 minutes of infusion, respectively; renin secretion decreased from 502 +/- 214 ng/min to 253 +/- 109 and 180 +/- 53 ng/min (P less than 0.05 for both values). Infusion of angiotensin II at the same rate decreased renal blood flow from 251 +/- 26 ml/min to 224 +/- 22 and 220 +/- 16 ml/min (P less than 0.01 and 0.025, respectively) and decreased renin secretion from 374 +/- 25 ng/min to 166 +/- 76 and 131 +/- 37 ng/min (P less than 0.025 for both values). Neither peptide significantly changed mean arterial blood pressure, creatinine clearance, or excreted sodium in these dogs. Infusion of des-1-Asp-angiotensin II into sodium-depleted dogs decreased renin secretion from 1094 +/- 211 ng/min to 768 +/- 132 and 499 +/- 31 ng/min (P less than 0.025 for both values) after 10 and 30 minutes of infusion. Angiotensin II infusion decreased renin secretion from 1102 +/- 134 to 495 +/- 235 and 502 +/- 129 ng/min in these dogs (P less than 0.05 and 0.025, respectively). Neither peptide significantly altered renal blood flow, arterial blood pressure, creatinine clearance, or excreted sodium in the sodium-depleted dogs. The data demonstrated that these two peptides have similar effects on the renin secretory mechanism and the vascular receptor at the level of the renal arterioles.     

Long-term effects of oral propranolol on splanchnic and systemic haemodynamics in patients with cirrhosis and oesophageal varices

Splanchnic and systemic haemodynamics were measured in 24 patients with cirrhosis and oesophageal varices and no previous bleeding. The patients were randomized either to long-term treatment with propranolol (14 patients) or no active treatment (controls, 10 patients). Catheterization was performed again 1 year after randomization. After 1 year of treatment the hepatic venous pressure gradient had decreased in both the propranolol and control group (-16% versus -24% (NS), respectively). Hepatic blood flow decreased substantially in both groups but significantly more in the propranolol group (-39% versus -17% (p less than 0.05), respectively). Azygos blood flow was significantly reduced after 1 year in the propranolol group (-47%, n = 5 (p less than 0.05)), and no obvious effect was observed in the control group (-2%, n = 4). The cardiac index decreased significantly in the propranolol group but not in the control group (-20% versus -1% (p less than 0.05), respectively). Our results demonstrate that the splanchnic hyperdynamic condition observed in cirrhosis is in some of the patients partly reversible without pharmacologic treatment. No additional effect of propranolol was observed on portal pressure after 1 year of treatment with propranolol, whereas a decrease in azygos blood flow was observed only in the propranolol group. The beneficial effect of propranolol on the risk of bleeding from oesophageal varices may, therefore, mostly be due to a selective decrease in collateral blood flow and thereby variceal blood flow.     

Superior portosystemic collateral circulation estimated by azygos blood flow in patients with cirrhosis. Lack of correlation with oesophageal varices and gastrointestinal bleeding. Effect of propranolol

In patients with cirrhosis, superior portosystemic collateral circulation was evaluated by the continuous thermodilution method in the azygos vein. Azygos blood flow was 5 times higher in a group of patients with cirrhosis (alcoholic in 27, cryptogenic in 8, post-hepatitic in 2 and primary biliary cirrhosis in 1), than in a group of patients without portal hypertension (steatosis in 2, granulomatous hepatitis in 2, persistent chronic hepatitis in 2 and Hodgkin's disease in 1). Azygos blood flow was not different in cirrhotic patients with no visible, in those with small-sized, and in those with large sized oesophageal varices. Azygos blood flow was not different in cirrhotic patients with and without a previous episode of gastrointestinal bleeding. Fifteen min after intravenous administration of 15 mg of propranolol, azygos blood flow significantly decreased whereas azygos blood flow did not change after placebo. The decrease in azygos blood flow was significantly more marked than the reduction in cardiac output. It is concluded that superior portosystemic collateral blood flow is elevated in patients with cirrhosis and that the reduction in this collateral circulation might explain the efficiency of propranolol in the prevention of recurrent gastrointestinal bleeding.     

Reversal of endothelin-1-induced vasoconstriction by nifedipine in human resistance vessels in vivo in healthy subjects.

The influence of blockade of voltage-operated calcium channels by nifedipine on endothelin-1-induced vasoconstriction was investigated in 10 healthy volunteers. Brachial artery infusions of nifedipine (0.25, 0.5, 1 and 3 micrograms/min/100 ml forearm tissue) resulted in dose-dependent increases (mean +/- SD) in forearm blood flow (103 +/- 63% to 833 +/- 426%). Intraarterial infusions of endothelin-1 (50 ng/min/100 ml) resulted in transient increases in forearm blood flow (2.6 +/- 0.9 vs 3.9 +/- 2.0 ml/min/100 ml, p less than 0.01) in the first minute of infusion and subsequent decreases (to 1.0 +/- .5 ml/min/100 ml, p less than 0.01) in the third minute of infusion. Endothelin-1-induced vasoconstriction was reversed by the lowest dose of nifedipine, whereas the higher dosages of nifedipine further increased forearm blood flow to 12.5 +/- 6.4 ml/min/100 ml. The percent increase of forearm blood flow during co-infusion of endothelin-1 and the highest dosage of nifedipine was significantly greater compared with nifedipine alone (1,204 +/- 531% vs 833 +/- 426%, p less than 0.05). The results demonstrate a dual action of luminally applied endothelin-1 in human resistance vessels in vivo (e.g., transient initial vasodilation followed by pronounced vasoconstriction) and suggest that blockade of voltage-operated calcium channels can effectively counteract the vasoconstrictor effects of endothelin-1.     

Measurement of azygos venous blood flow by a continuous thermal dilution technique: an index of blood flow through gastroesophageal collaterals in cirrhosis

A method to quantitate blood flow through the gastroesophageal collaterals in portal hypertensive patients was developed. Since gastroesophageal collaterals drain into the azygos system, it is postulated that measurement of blood flow in the azygos vein should provide a quantitative measurement of gastroesophageal collateral blood flow changes in portal hypertensive patients. Azygos blood flow was measured using a double thermodilution catheter directed under fluoroscopy to the azygos vein. Ten patients with alcoholic cirrhosis were studied. Five of these patients had a history of repeated bleeding from gastroesophageal varices (Group I). The azygos blood flow in these patients was 596 +/- 78 ml per min. The other five patients all had decompressive surgery of the portal system (Group II). In these patients the azygos venous blood flow was 305 +/- 29 ml per min (p less than 0.01). The coefficient of variation of repeated baseline measurements was of 4.4 +/- 0.6%. The azygos venous blood flow measurement is a rapid, simple and sensitive method to evaluate blood flow changes in the vessels involved in gastroesophageal bleeding due to portal hypertension.     

The extent of the collateral circulation influences the postprandial increase in portal pressure in patients with cirrhosis

BACKGROUND: In cirrhosis, repeated flares of portal pressure and collateral blood flow provoked by postprandial hyperaemia may contribute to variceal dilation and rupture. AIM: To examine the effect of the extent of the collateral circulation on the postprandial increase in portal pressure observed in cirrhosis. PATIENTS AND METHODS: The hepatic venous pressure gradient (HVPG), hepatic blood flow and azygos blood flow were measured in 64 patients with cirrhosis before and after a standard liquid meal. RESULTS: Peak increases in HVPG (median+14.9%), hepatic blood flow (median+25.4%), and azygos blood flow (median+32.2%) occurred at 30 min after the meal. Compared with patients with marked postprandial increase in HVPG (above the median, n = 32), those showing mild (<15%, n = 32) increase in HVPG had a higher baseline azygos flow (p<0.01) and underwent a greater postprandial increase in azygos flow (p<0.02). Hepatic blood flow increased similarly in both groups. Postprandial increases in HVPG were inversely correlated (p<0.001) with both baseline azygos flow (r = -0.69) and its postprandial increase (r = -0.72). Food intake increased nitric oxide products in the azygos (p<0.01), but not in the hepatic vein. Large varices (p<0.01) and previous variceal bleeding (p<0.001) were more frequent in patients with mild increase in HVPG. CONCLUSIONS: Postprandial hyperaemia simultaneously increases HVPG and collateral flow. The extent of the collateral circulation determines the HVPG response to food intake. Patients with extensive collateralisation show less pronounced postprandial increases in HVPG, but associated with marked flares in collateral flow. Collateral vessels preserve their ability to dilate in response to increased blood flow.