Role of Angiotensin II in Regulation of Basal andSympathetically Stimulated Vascular Tone in Earlyand Advanced Cirrhosis

AHMED HELMY,* RAJIV JALAN,* DAVID ERNEST NEWBY,‡ PETER CLIVE HAYES,*and DAVID JOHN WEBB§

*Liver Unit, Department of Medicine, Royal Infirmary of Edinburgh, Edinburgh; ‡Department of Cardiology, Royal Infirmary of Edinburgh,Edinburgh; and §Clinical Pharmacology Unit and Research Centre, Department of Medical Sciences, Western General Hospital,University of Edinburgh, Edinburgh, Scotland

Background & Aims: The renin-angiotensin and sympa-thetic nervous systems are activated in cirrhosis. Thisstudy aimed to establish the role of angiotensin II (ANGII) in the regulation of basal and sympatheticallystimulated vascular tone in preascitic cirrhotic patientsand patients with diuretic-refractory ascites comparedwith age- and sex-matched healthy controls. Methods:Forearm blood flow (FBF) responses to lower bodynegative pressure (LBNP) and to subsystemic, intrabra-chial infusions of losartan, an angiotensin II type 1 (AT1)receptor antagonist, norepinephrine, and ANG II weremeasured using venous occlusion plethysmography.Results: In all groups, ANG II and norepinephrine causeddose-dependent reductions in FBF (P F 0.001); re-sponses to norepinephrine were similar across the 3groups but those to ANG II were less in both cirrhoticgroups than in controls (P F 0.01). Losartan caused adose-dependent increase in FBF only in patients withrefractory ascites (PF 0.01). LBNP caused less reductionin FBF in refractory ascites patients than in both preasciticpatients and controls (P F 0.01). Conclusions: Despitehyporesponsiveness to exogenous ANG II in both earlyand advanced cirrhosis, endogenous ANG II contrib-utes to the maintenance of basal vascular tone only inadvanced cirrhosis. These findings suggest a role ofANG II in the pathogenesis of ascites. Attenuated LBNPresponses occurred only in advanced cirrhosis, withoutapparent interaction with endogenous ANG II.

Cirrhosis is characterized by hyperdynamic circulatorychanges, including high cardiac output and de-

creased systemic vascular resistance,1–3 that worsen withdisease progression.4 Although compensatory activationof vasopressin5–8 and the renin-angiotensin 9–12 andsympathetic nervous9,13–16 systems occurs, impaired vas-cular reactivity has been proposed to contribute to thehyperdynamic circulation of cirrhosis and to the develop-ment of its complications, such as portal hypertension,ascites, and hepatorenal syndrome.17

Present evidence of this hyporesponsiveness has beenprovided by systemic studies and in vitro experiments.Although contributing to our understanding of thepathophysiology of portal hypertension, systemic studieshave the disadvantage of invoking neurohumoral, counter-regulatory mechanisms and having effects on a widerange of organs, including the heart, brain, and kid-neys.18,19 Also, previous studies investigating the role ofthe renin-angiotensin system6,8 in the maintenance ofvascular tone in patients with cirrhosis were weakened bythe use of antagonists, such as saralasin, with partialagonist activity.20–22 However, losartan, a selective angio-tensin II type 1 (AT1) receptor antagonist that is devoid ofagonist activity, has recently become available for clinicaluse.23 In addition, the combination of subsystemic andlocally active intrabrachial artery infusions with bilateralforearm blood flow (FBF) measurements using venousocclusion plethysmography represents a powerful andreproducible method of assessing in vivo vascular re-sponses in a single circulation without invoking systemiceffects.18,19

Therefore, the aims of this study were to evaluate, inpatients with preascitic cirrhosis and those with diuretic-refractory ascites, the role of endogenous angiotensin II(ANG II) in the maintenance of basal and sympatheti-cally stimulated vascular tone in the forearm circulation;to assess the forearm vascular responses to exogenousnorepinephrine (NE) and ANG II; and to determinewhether changes in reactivity, if any, are related to diseaseseverity.

Abbreviations used in this paper: ANG II, angiotensin II; AT1,angiotensin II type 1; EP, epinephrine; FBF, forearm blood flow;LBNP, lower body negative pressure; MAP, mean arterial pressure;NE, norepinephrine; PRA, plasma renin activity.

r 2000 by the American Gastroenterological Association0016-5085/00/$10.00

GASTROENTEROLOGY 2000;118:565–572

Patients and Methods

SubjectsEighteen patients with biopsy-proven cirrhosis, 8 in

the preascitic stage and 10 in the diuretic-refractory stage ofthe disease,24 were recruited and compared with 8 age- andsex-matched healthy volunteers (age, 49 6 6 years; 3 womenand 5 men). All patients were ambulant and had normal serumcreatinine levels (,100 µmol/L). Patients with diuretic-refractory ascites were studied during a no-added-salt diet (100mmol/day) 1 week after their regular therapeutic paracentesisand did not receive diuretic therapy for at least 2 weeks beforethe study. Patients with alcoholic liver disease were abstinentfrom alcohol for at least 1 month,25 confirmed by history andrandom blood ethanol testing. None of the subjects receivedvasoactive or nonsteroidal anti-inflammatory drugs in the weekbefore each phase of the study. All subjects abstained from food,tobacco, and caffeine-containing drinks for at least 4 hoursbefore each study. To minimize between-subject variability,female control subjects who were premenopausal were studiedbetween day 7 and day 12 of their menstrual cycle.26 All femalecirrhotic patients were amenorrheic. Studies were undertakenwith written informed consent from each subject, with theapproval of the local research ethics committee, and inaccordance with the Declaration of Helsinki (1989) of theWorld Medical Association.

Drugs and Intra-arterial Administration

Cannulation of the brachial artery of the nondominantarm was performed using a 27-standard wire gauge steel needle(Cooper’s Needle Works, Birmingham, England) under 1%lidocaine (Xylocaine; Astra Pharmaceuticals Ltd., Kings Lang-ley, England) local anesthesia. Patency was maintained bysaline infusion at a constant rate of infusion (1 mL/min) via anIVAC P1000 syringe pump (IVAC Ltd., Basingstoke, England).

Losartan (Dupont-Merck, Wilmington, DE) at doses of 30and 90 µg/min; ANG II (Clinalfa, AG Laufelfingen, Switzer-land) at doses of 1, 3, 10, and 30 pmol/min; and NE (Levophed;Sanofi-Winthrop Ltd., Guildford, England) at doses of 20, 60,180, and 540 pmol/min were dissolved in physiological saline(0.9% Baxter Healthcare Ltd., Thetford, England) and admin-istered intra-arterially at these locally active doses.27,28 Toprevent its oxidation, NE was dissolved in saline containing0.1% ascorbic acid (Evans Medical, Langhurst, England).

Measurements

All studies were performed in a quiet, temperature-controlled room maintained at 22°C–24°C. Blood pressure andpulse rate were measured in the control arm using a noninva-sive oscillometric sphygmomanometer (Takeda UA 751; Tak-eda Medical Inc., Tokyo, Japan) at intervals throughout thestudy.29 FBF was measured as described previously.28,29 Briefly,FBF measurements were made in both arms (i.e., the infusedand the noninfused arms) by venous occlusion plethysmogra-phy using mercury-in-Silastic strain gauges applied to thewidest part of the forearm. The hands were excluded frommeasurements by rapid inflation of wrist cuffs to a pressure of220 mm Hg using E20 Rapid Cuff Inflators (D. E. HokansonInc., Washington, DC). Upper-arm cuffs were inflated to 40mm Hg for 10 seconds at every 15 seconds to achieve venousocclusion and obtain blood flow measurements.

Lower Body Negative Pressure

Subjects were rested supine in a plastic-covered steelcage that enclosed the lower body from the waist downward, asdescribed previously.30,31 A constant negative pressure of 15mm Hg was achieved using an industrial-strength vacuum cleanerregulated by a pressure control unit (Medical Physics Department,Edinburgh University, Edinburgh, Scotland). Alterations to andfrom atmospheric pressure were attained within 1–2 seconds. Thisnegative pressure produces selective forearm vasoconstriction throughlow-pressure baroreceptor unloading without altering arterialpressure or heart rate, thereby eliminating the confoundingeffects of systemic hypotension.31

Study Design

Two protocols (Figure 1) were applied on 2 separateoccasions 1 week apart and in random order. Subjects restedsupine throughout each study. Strain gauges and cuffs wereapplied, and the brachial artery of the nondominant arm wascannulated. FBF was measured in the last 3 minutes of eachinfusion period unless otherwise stated. Before participating ineach of the protocols, saline was infused for 30 minutes to allowtime for equilibration, with FBF measurements made every 10minutes, and the final measure taken as the baseline FBF. Inprotocol 2, the rationale for combining lower body negativepressure (LBNP) with each of the saline and losartan infusionswas to assess the contribution of endogenous ANG II to the

Figure 1. Schematic diagram ofthe 2 study protocols 1 weekapart. F, Bilateral FBF measure-ments each for 3 minutes; L,LBNP application each for 3 min-utes. NE doses are in picomolesper minute, ANG II doses inpicomoles per minute, and losar-tan doses in micrograms perminute.

566 HELMY ET AL. GASTROENTEROLOGY Vol. 118, No. 3

forearm vascular responses during low-pressure cardiopulmo-nary baroreceptor unloading.

Blood Assays

After 30 minutes of supine rest, and before any drugswere administered, venous blood was withdrawn from thenoninfused arm. Ten milliliters was admixed with each of thefollowing: 1 mL of 1% disodium EDTA/2% sodium metabisul-fite for measuring epinephrine (EP) and NE concentrations; 0.5mL of 0.45% O-phenanthroline/4.65% disodium EDTA formeasuring ANG II concentration; and 1 mL of 1% disodiumEDTA and 1000 KIU aprotinin (Bayer AG, Leverkusen,Germany) for measuring plasma renin activity (PRA). Thesamples were placed on ice and immediately centrifuged at1500g for 20 minutes. Plasma was frozen and stored at 280°Cuntil assayed. EP and NE concentrations were determined using anelectrochemical method following separation by reverse-phase liq-uid chromatography.32 Plasma ANG II concentrations were mea-sured by radioimmunoassay following extraction using Bond Elutcolumns (Varian, Harber City, CA) as previously described.33 PRAwas measured under standard conditions through the generationof ANG I using radioimmunoassay as previously described.34

Data Analysis and Statistics

Plethysmographic data were extracted from the Chart(AD Instruments, Castle Hill, Australia) data files, and FBFswere calculated for individual venous occlusion cuff inflationsusing a template spreadsheet (Excel v. 5.0; Microsoft). Recordingsfrom the first 60 seconds after wrist-cuff inflation were not includedin the analysis because of the variability in blood flow thisproduces.35 Usually, the last 5 flow recordings in each 3-minutemeasurement period were calculated and averaged for each arm.

To reduce the variability of blood flow data, the ratio of flowsin the 2 arms was calculated for each time point, thereby usingthe noninfused arm as a contemporaneous control.18,19 Thepercentage change in FBF was calculated as follows:

% Change in FBF 5 100 3 [(It/NIT) 2 (Ib/NIb)]/(Ib/NIb),

where It and NIt are the blood flows in the infused andnoninfused forearms, respectively, at a given time point (t), andIb and NIb are the blood flows in the infused and noninfusedforearms, respectively, at baseline (b); time 0.

The changes in FBF during LBNP were calculated in eachforearm as % change from the flow immediately before theapplication of LBNP as follows:

% Change in FBF During LBNP 5 100

3 [(FBFLBNP 2 FBFPRE)/(FBFPRE)],

where FBFPRE is the forearm blood flow immediately beforeLBNP, and FBFLBNP is the forearm blood flow during LBNP.

Mean arterial pressure (MAP) was calculated according tothe formula: MAP 5 Diastolic Arterial Pressure 1 1/3 PulsePressure (Systolic Arterial Pressure 2 Diastolic Arterial Pres-sure). FBF was expressed in mL · 100 mL tissue21 · min21.36

Data are expressed as mean 6 SEM and examined by analysis ofvariance (ANOVA) with multiple comparisons, the Pearson’s corre-

lation coefficient, and 2-tailed paired and unpaired Student t tests asappropriate. Statistical significance was taken at the 5% level.

Results

A summary of the patient characteristics is shownin Table 1. Patients with diuretic-refractory ascites had asignificantly higher Child–Pugh score and prothrombintime and significantly lower serum albumin level thanpreascitic patients (P , 0.01). After participation in thefirst study, 1 patient with refractory ascites underwentliver transplantation and another underwent transjugularintrahepatic portosystemic stent shunt for variceal bleed-ing, and the second study was performed in another 2patients. Therefore, the 16 studies, 8 in each protocol,were completed in 10 patients.

Baseline Hemodynamics

There were no changes in heart rate or MAPthroughout the studies. Mean heart rate was significantlyhigher and MAP significantly lower in patients withdiuretic-refractory ascites than in preascitic patients andcontrols (P , 0.01; Table 2). Heart rate correlatedpositively with PRA and basal ANG II concentrations (r 510.52, P , 0.01; and r 5 10.65, P , 0.01, respec-tively), and MAP correlated negatively with PRA and thebasal ANG II concentrations (r 5 20.62, P , 0.01; andr 5 20.61, P , 0.01, respectively).

Baseline FBF

There were no significant differences between thebaseline FBF in the infused and noninfused arms on each

Table 1. Patient Characteristics

VariableRefractory

ascitesPreasciticcirrhosis

Age ( yr) 56 6 4 48 6 3Sex (F:M) 4:6 3:5Child–Pugh score 11.1 6 0.5a 6.3 6 0.4Child grade

A 0 6B 0 2C 10 0

Etiology of cirrhosisPBC 1 3ALD 7 4ALD 1 HBV 1 —Cryptogenic 1 —AICAH — 1

Serum bilirubin (mmol/L) 52 6 17 35 6 12Serum albumin ( g/L) 27 6 2a 38 6 2Prothrombin time (s) 18 6 1a 13 6 1Blood urea (mmol/L) 4 6 1 4 6 1Serum creatinine (mmol/L) 89 6 9 82 6 4

NOTE. Results are expressed as mean 6 SEM.PBC, primary biliary cirrhosis; ALD, alcoholic liver disease; HBV,hepatitis B virus; AICAH, autoimmune chronic active hepatitis.aP , 0.01 vs. preascitic cirrhotic patients.

March 2000 REGULATION OF VASCULAR TONE IN CIRRHOSIS 567

study day. The baseline FBF was significantly lower inpatients with refractory ascites than in both preasciticpatients and the control group (2.5 6 0.2 vs. 4.8 6 0.9and 4.2 6 0.8 mL · 100 mL tissue21 · min21, respec-tively; P , 0.05). No significant change in blood flow inthe noninfused arm was observed during the protocols,except during LBNP.

Forearm Vascular Responses to ANG IIand NE Infusions

Both ANG II and NE infusions produced signifi-cant dose-dependent reductions in FBF (P , 0.001) in allsubjects studied (Figure 2). The responses to ANG II

were significantly greater in the control group than inpatients with diuretic-refractory ascites (P , 0.001) andpatients with preascitic cirrhosis (P , 0.01). FBFresponses to NE were similar in all groups.

Forearm Vascular Responses to LosartanInfusions and LBNP

Absolute values of FBF (mL · 100 mL tissue21 ·min21) for each arm at baseline and during saline,losartan, and LBNP phases of the study in all groups areshown in Table 3. Losartan caused dose-dependent in-creases in FBF in patients with refractory ascites (P ,0.001) but not in patients with preascitic cirrhosis orcontrols (P , 0.01 by 3-way ANOVA between groups;Figure 3). LBNP produced significantly less reduction inFBF in patients with refractory ascites than in eitherpatients with preascitic cirrhosis or controls (P , 0.01 by2-way ANOVA; Figure 4). In addition, responses toLBNP were similar during saline and losartan infusionswithin each group. Although LBNP responses appearedto be less during saline than during losartan infusions inthe refractory ascites group, this did not reach statisticalsignificance (P 5 0.13 and P 5 0.42 at doses of 30 and90 µg/min, respectively).

Assays

Basal plasma ANG II concentrations were signifi-cantly higher in patients with diuretic-refractory ascitesthan in preascitic cirrhotics (P , 0.05), and both weresignificantly higher than in controls (P , 0.01). Patientswith refractory ascites had significantly higher basal PRA(P , 0.001) and higher basal plasma NE (P , 0.01) andEP concentrations (P , 0.05) than patients with preasciticcirrhosis and controls. Child–Pugh score correlated positivelywith both PRA and plasma NE concentrations (r 5

Table 2. Baseline Systemic Hemodynamics, FBF, andHumoral Mediators

VariableRefractory

ascitesPreasciticcirrhosis Controls

Pulse (bpm)Before LBNP 88.0 6 3.4b 72.4 6 2.1 67.4 6 2.6After LBNP 89.0 6 4.5b 74.7 6 2.6 62.0 6 2.3

MAP (mm Hg)Before LBNP 80.7 6 1.2c 91.8 6 3.2 93.7 6 3.9After LBNP 81.6 6 2.1c 93.5 6 3.2 94.9 6 3.0

Baseline FBF(mL · 100mL21 · min21)

Infused arm 2.5 6 0.2d 4.8 6 0.9 4.2 6 0.8Noninfused arm 2.2 6 0.4d 4.2 6 0.6 3.5 6 0.7

Basal plasmaEP (nmol/mL) 1.2 6 0.3d 0.7 6 0.2 0.4 6 0.2NE (nmol/mL) 2.9 6 0.3c 1.1 6 0.1 1.4 6 0.4ANG II ( pg/mL) 237.7 6 30.8b 57.3 6 17.3a 3.2 6 0.3PRA

(ng · mL21 · h21) 15.4 6 2.4b 3.5 6 1.0 1.7 6 0.8

NOTE. Results are expressed as mean 6 SEM.bpm, beats per minute.aP , 0.01 vs. controls.bP , 0.05 vs. preascitic cirrhotics and controls.cP , 0.01 vs. preascitic cirrhotics and controls.dP , 0.05 vs. preascitic cirrhotics and controls.

Figure 2. Responses in FBF toincremental doses of ANG II(pmol/min) and NE (pmol/min).NS, not significant. Results aremean 6SEM.

568 HELMY ET AL. GASTROENTEROLOGY Vol. 118, No. 3

10.71, P , 0.001; and r 5 10.60, P , 0.01,respectively).

Discussion

In this study we show that although patients withdiuretic-refractory ascites and those with preascitic cirrho-sis are hyporesponsive to exogenous ANG II, endogenousANG II contributes to the maintenance of basal vasculartone, as demonstrated by the response to losartan, only inpatients with diuretic-refractory ascites. In addition,impaired peripheral vasoconstrictor responses to low-pressure baroreceptor unloading occurs in patients withrefractory ascites, although the vascular responses toexogenous NE are similar in all groups.

Previous studies in patients with cirrhosis have pro-duced contradictory results regarding the responsivenessto systemic administration of a-adrenergic agonists andANG II. Many studies showed that responses to NE werenot affected,37–39 but others showed an attenuated re-sponse.40,41 Responses to ANG II were reported to beimpaired in one study,41 whereas normal responses werereported in others.38–40 However, systemic drug adminis-tration leads to concomitant effects on organs, such as the

brain, kidney, and heart, and influences neurohumoralreflexes by affecting systemic hemodynamics. Conse-quently, vascular responses cannot be wholly attributedto a direct effect of the drug.18,19 In contrast, the use ofbilateral FBF measurements, with unilateral brachialartery infusions of vasoactive mediators at subsystemicand locally active doses, is a very powerful and reproduc-ible tool for directly assessing vascular responses invivo.18,19 However, previous studies using the forearmmodel in patients with cirrhosis are few and their resultsare also variable.28,42–45 Because of the confoundinginfluence of diuretic therapy and some methodologicalconcerns, such as use of unilateral blood flow measure-ments and nonstandardized infusion protocols, conflict-ing results have been reported, e.g., either reduced42–44 orunaltered28,45 vasoconstrictor responses to NE.

Using the bilateral forearm model, we have previouslyshown that patients with a combination of cirrhosis anddiuretic-responsive ascites have a significantly reducedvascular reactivity to exogenous ANG II and cardiopulmo-nary baroreceptor unloading.28 However, maintenance ofpatients on regular diuretic therapy may have influencedthese results by producing relative intravascular hypovo-

Figure 3. Responses in FBF to losartan infusions in the 3 studiedgroups. Results are mean 6SEM. *P , 0.01 by 3-way ANOVA;refractory ascites vs. preascitic cirrhosis and controls.

Figure 4. Responses in FBF to LBNP application in the 3 studiedgroups during saline (h), 30 µg/min losartan (Q), and 90 µg/minlosartan (j) infusions. Results are mean 6SEM. *P , 0.01 by 3-wayANOVA; refractory ascites vs. preascitic cirrhosis and controls.

Table 3. Absolute FBF During Baseline, Saline, Losartan, and LBNP Phases of the Study

Refractory ascites Preascitic cirrhosis Controls

Noninfusedarma

Infusedarma

Noninfusedarm

Infusedarm

Noninfusedarm

Infusedarm

Baseline 2.2 6 0.4 2.5 6 0.2 4.2 6 0.6 4.8 6 0.9 3.5 6 0.7 4.2 6 0.8Saline (1) 2.3 6 0.4 2.6 6 0.3 4.2 6 0.6 5.1 6 1.0 3.6 6 0.7 4.3 6 0.8Saline (1) 1 LBNP 2.4 6 0.4 2.5 6 0.3 3.2 6 0.5 3.8 6 0.7 2.8 6 0.6 3.3 6 0.7Losartan 30 2.3 6 0.5 3.1 6 0.4 5.1 6 0.8 6.0 6 1.1 3.7 6 1.0 4.7 6 1.2Losartan 30 1 LBNP 2.4 6 0.4 2.6 6 0.3 3.7 6 0.5 4.1 6 0.5 3.0 6 0.7 3.6 6 0.6Losartan 90 2.2 6 0.4 3.0 6 0.3 5.1 6 0.8 6.2 6 1.1 4.1 6 0.9 5.2 6 1.1Losartan 90 1 LBNP 2.2 6 0.3 2.4 6 0.3 3.7 6 0.6 4.4 6 0.6 3.5 6 0.8 4.1 6 0.9Saline (2) 2.1 6 0.5 2.7 6 0.3 5.3 6 0.9 6.2 6 1.2 4.1 6 0.9 5.2 6 1.1Saline (2) 1 LBNP 2.4 6 0.5 2.6 6 0.3 3.8 6 0.5 4.5 6 0.6 3.3 6 0.9 4.4 6 1.2

NOTE. Results are expressed as mean 6 SEM. Data are reported in mL · 100 mL tissue21 · min21.aP , 0.001 by 3-way ANOVA; refractory ascites vs. preascitic cirrhotics and controls.

March 2000 REGULATION OF VASCULAR TONE IN CIRRHOSIS 569

lemia and associated neurohumoral activation. In thepresent study, we recruited patients with early andadvanced cirrhosis who were not receiving maintenancediuretic therapy. Patients with refractory ascites werestudied 1 week after their regular therapeutic paracentesisand albumin therapy and were hemodynamically stable.

In agreement with previous studies,4,9–12,28 we haveshown that the renin-angiotensin and sympathetic ner-vous systems are activated in patients with cirrhosis andthis neurohumoral activation becomes more marked withdisease progression. The observed reduction in baselineFBF in patients with diuretic-refractory ascites may, inpart, be explained by this neurohumoral activation andsupports the previously reported reduction in brachial, femo-ral, and cerebral arterial blood flow patients with cirrhosis andascites, using duplex-Doppler ultrasonography.46,47

The present study demonstrates that the forearmcirculation is hyporesponsive to ANG II in cirrhoticpatients even before the development of ascites, suggest-ing that this abnormality may contribute to the pathogen-esis of the hyperdynamic circulatory syndrome. Themechanism of this hyporesponsiveness is unclear, but arecent study has suggested that it may be mediatedthrough nitric oxide.48 In addition, we have shown anincrease in plasma ANG II concentrations in patientswith preascitic cirrhosis. The mechanism underlying thisincrease is not clear, although the presence of an activatedrenin-angiotensin system is consistent with a compen-sated vasodilated state. Moreover, the activity of therenin-angiotensin system correlates with portal pressuregradient49; in the present study, all patients with preas-citic cirrhosis had a high portal pressure as indicated bythe presence of esophageal varices. Increased activity ofthe endogenous vasopressor systems could influenceresponses to exogenously administered ANG II and NE,resulting in a right shift of the dose-response curve.However, this is unlikely, at least for responses to NE,which were similar in all groups despite the presence ofan activated sympathetic nervous system only in patientswith diuretic-refractory ascites.

The finding that selective AT1-receptor blockade withlosartan increases FBF in patients with diuretic-refractoryascites but produces no change in FBF in healthy subjectsor patients with preascitic cirrhosis supports the conceptthat ANG II contributes to the maintenance of basalvascular tone only in patients with advanced liver disease,having no role in either healthy controls or early cirrho-sis.27,28,50 A recent finding that oral losartan effectivelyreduced hepatic venous pressure gradient in patients withcirrhosis and portal hypertension,51 together with theobserved increase in plasma ANG II concentrations andvascular dependence on ANG II in severe cirrhosis,

indicates that ANG II contributes to the hemodynamicconsequences of cirrhosis and, therefore, that losartanmay potentially be helpful in the treatment of portalhypertension. Additionally, the effects of the high plasmaANG II concentrations observed in patients with preas-citic cirrhosis on renal arteriolar blood flow and distribu-tion help to explain the subtle Na1 retention detectedbefore the development of ascites52 and suggest that useof AT1-receptor blockers may improve Na1 excretion inpatients with preascitic cirrhosis, and thus prevent ordelay the development of ascites.

LBNP has been used for many years to examine theneurocirculatory reflex response to reduced venous returnto the heart.53 In healthy subjects, LBNP at intensities upto 20 mm Hg selectively unloads low-pressure cardiopul-monary baroreceptors without causing changes in heartrate or blood pressure. Such low-intensity LBNP selec-tively reduces FBF31,54–57 and increases forearm NEspillover without affecting total peripheral vascular resis-tance or total body NE spillover. In contrast, LBNP atintensities greater than 20 mm Hg also unloads arterialbaroreceptors with concomitant increases in total bodyspillover of NE, increases in heart rate, and decreases inblood pressure.57 Previous studies in patients with decom-pensated cirrhosis have shown impaired vasoconstrictorresponses to both LBNP and body tilting.9,28,40 In thepresent study, we document impaired vasoconstrictorresponses to LBNP in patients with diuretic-refractoryascites despite normal responsiveness to exogenous NE.This may be the result of several mechanisms includingbaroreceptor down-regulation or defective sympatheticneurotransmission secondary to the autonomic dysfunc-tion described with advanced cirrhosis.58 In earlier stud-ies by Wong et al.,52,59 central venous pressure waselevated in patients with preascitic cirrhosis. One maytherefore anticipate that higher intensities of LBNP arerequired to achieve similar reflex responses in thesepatients. Nevertheless, these studies have shown that, inresponse to LBNP of 15 mm Hg, patients with preasciticcirrhosis have reductions in central venous pressure andFBF similar to healthy subjects.52,59

The observation that the vasoconstrictor responses toLBNP of 15 mm Hg in the infused forearm are unaffectedby losartan infusions has been reported in both cirrhoticpatients28 and controls,27,60,61 and indicates that endog-enous ANG II has little, if any, role in mediating theseregional responses to low levels of LBNP in eithersituation. These findings also suggest that influences ofendogenous ANG II does not contribute to the impairedresponse to LBNP in patients with refractory ascites.

One of the potential limitations of the present study isthat patients in the diuretic-refractory group were on a

570 HELMY ET AL. GASTROENTEROLOGY Vol. 118, No. 3

no-added-salt diet (100 mmol/day), whereas patientswith preascitic cirrhosis and controls were allowed nor-mal sodium intake (150 mmol/day). Sodium status mayalter the peripheral vascular responses to vasoactivesubstances. However, pressor responses to exogenous NE,for example, in healthy subjects are affected only byextremes of sodium intake, differing from 10–80-fold.52,62,63 In our study, the variation in sodium intake ismodest (1.5-fold) and such small differences around theusual Western diet have not been shown to influence subse-quent vascular responses. Indeed, we found similar reductionsin FBF to NE infusion in patient and control groups.

In conclusion, despite hyporesponsiveness to exog-enous ANG II in both early and advanced cirrhosis,endogenous ANG II contributes to the maintenance ofbasal forearm vascular tone only in advanced cirrhosis.These findings suggest a role for ANG II in thepathogenesis of ascites. The forearm vascular responses toexogenous NE are unimpaired. However, attenuatedLBNP responses occur only in advanced cirrhosis, with-out apparent interaction with endogenous ANG II,suggesting low-pressure baroreceptor down-regulation ordefective neuronal transmission. With advancing liverdisease, further activation of the neurohumoral systemsoccurs, and the peripheral circulation is progressivelyvasoconstricted. AT1-receptor blockade may provide auseful strategy for preventing the development of ascitesin preascitic patients but must be used with caution inpatients with advanced liver disease.64 Further studies arerequired to elucidate the mechanism of the observedhyporesponsiveness to ANG II.

References1. Murray JF, Dawson AM, Sherlock S. Circulatory changes in chronic

liver disease. Am J Med 1958;32:358–367.2. Kontos HA, Shapiro W, Mauck HP, Patterson JL. General and

regional circulatory alterations in cirrhosis of the liver. Am J Med1964;37:526–535.

3. Kowalski HJ, Abelmann WH. The cardiac output at rest inLaennec’s cirrhosis. J Clin Invest 1953;32:1025–1033.

4. Bendtsen F, Henriksen JH, Sorensen TIA, Christensen NJ. Effectof oral propranolol on circulating catecholamines in cirrhosis:relationship to severity of liver disease and splanchnic hemody-namics. J Hepatol 1990;10:198–204.

5. Bichet DG, Szatalowicz V, Chaimovitz C, Schrier RW. Role ofvasopressin in abnormal water excretion in cirrhotic patients. AnnIntern Med 1982;96:413–417.

6. Reznick RK, Langer B, Taylor BR, Seif S, Blendis LM. Hyponatre-mia and arginine vasopressin secretion in patients with refractoryhepatic ascites undergoing peritoneovenous shunting. Gastroen-terology 1983;84:713–718.

7. Bichet DG, Groves BM, Schrier RW. Mechanisms of improvementof water and sodium excretion by immersion in decompensatedcirrhotic patients. Kidney Int 1983;24:788–794.

8. Epstein M, Weitzman RE, Preston S, DeNunzio AG. Relationshipbetween plasma arginine vasopressin and renal water handling indecompensated cirrhosis. Min Electro Metab 1984;10:155–165.

9. Bernardi M, Trevisani F, Santini C, Ligabue A, Capelli M, Gasbar-rini G. Impairment of blood pressure control in patients with liver

cirrhosis during tilting: study on adrenergic and renin-angiotensinsystems. Digestion 1982;25:124–130.

10. Arroyo V, Bosch J, Mauri M, Ribera F, Navarro-Lopez F, Rodes J.Effect of angiotensin-II blockade on systemic and hepatic hemody-namics and on the renin-angiotensin-aldosterone system incirrhosis with ascites. Eur J Clin Invest 1981;11:221–229.

11. Pariente EA, Bataille C, Bercoff E, Lebrec D. Acute effects ofcaptopril on systemic and renal hemodynamics and on renalfunction in cirrhotic patients with ascites. Gastroenterology1985;88:1255–1259.

12. Schroeder ET, Anderson GH, Goldman SH, Streeten DHP. Effect ofblockade of angiotensin II on blood pressure, renin and aldoste-rone in cirrhosis. Kidney Int 1976;9:511–519.

13. Floras JS, Legault L, Morali GA, Hara K, Blendis LM. Increasedsympathetic outflow in cirrhosis and ascites: direct evidence fromintraneural recordings. Ann Intern Med 1991;114:373–380.

14. Henriksen JH, Christensen NJ, Ring-Larsen H. Noradrenaline andadrenaline concentrations in various vascular beds in patients withcirrhosis. Relation to hemodynamics. Clin Physiol 1981;1:293–304.

15. Bichet DG, Van Putten VJ, Schrier RW. Potential role of increasedsympathetic activity in impaired sodium and water excretion incirrhosis. N Engl J Med 1982;307:1552–1557.

16. Ramond M, Comoy E, Lebrec D. Alterations in isoprenalinesensitivity in patients with cirrhosis: evidence of abnormality ofthe sympathetic nervous activity. Br J Clin Pharmacol 1986;21:191–196.

17. Schrier RW, Arroyo V, Bernardi M, Epstein M, Henriksen JH, RodesJ. Peripheral arterial vasodilation hypothesis: a proposal for theinitiation of renal sodium and water retention in cirrhosis.Hepatology 1988;8:1151–1157.

18. Benjamin N, Calver A, Collier J, Robinson B, Vallance P, Webb D.Measuring forearm blood flow and interpreting the responses todrugs and mediators. Hypertension 1995;25:918–923.

19. Webb DJ. The pharmacology of human blood vessels in vivo. JVasc Res 1995;32:2–15.

20. Case DB, Wallace JM, Keim HJ, Sealey JE, Laragh J. Usefulnessand limitations of saralasin, a partial competitive agonist ofangiotensin II for evaluating the renin and sodium factors inhypertensive patients. Am J Med 1976;60:825–836.

21. Anderson GH, Streeten DHP, Dalakos TG. Pressor response to1-Sar-8-Ala-angiotensin II (saralasin) in hypertensive subjects.Circ Res 1977;40:243–250.

22. Anderson GH, Anderson T, Streeten DHP, Schroeder ET. Acuteeffects of saralasin on plasma aldosterone in different pathophysi-ological conditions. J Clin Endocrinol Metab 1980;50:529–536.

23. Bauer JH, Reams GP. The angiotensin II type I antagonists: a newclass of anti-hypertensive drugs. Arch Intern Med 1995;155:1361–1368.

24. Arroyo V, Gines P, Gerbes AL, Dudley FJ, Gentillini P, Laffi G,Reynolds TB, Ring-Larsen H, Scholmerich J. Definition and diagnos-tic-criteria of refractory ascites and hepatorenal syndrome incirrhosis. Hepatology 1996;23:164–176.

25. Howes LG, Reid JL. Decreased vascular responsiveness tonoradrenaline following regular ethanol consumption. Br J ClinPharmacol 1985;20:669–674.

26. Hashimoto M, Akishita M, Eto M. Modulation of endothelium-dependent flow–mediated dilatation of the brachial artery by sexand menstrual cycle. Circulation 1995;92:3431–3435.

27. Newby DE, Masumori S, Johnston NR, Boon NA, Webb DJ.Endogenous angiotensin II contributes to basal peripheral vascu-lar tone in sodium deplete but not sodium replete man. Cardio-vasc Res 1997;36:268–275.

28. Newby DE, Jalan R, Masumori S, Hayes PC, Boon NA, Webb DJ.Peripheral vascular tone in patients with cirrhosis: role of therenin-angiotensin and sympathetic nervous systems. CardiovascRes 1998;38:221–228.

29. Wiinberg N, Walter-Larsen S, Eriksen C, Nielsen PE. An evaluation

March 2000 REGULATION OF VASCULAR TONE IN CIRRHOSIS 571

of semi-automatic blood pressure manometers against intra-arterial blood pressure. J Amb Monit 1988;1:303–309.

30. Brown E, Goei JS, Greenfield ADM, Plassaras GC. Circulatoryresponses to stimulated gravitational shifts of blood in maninduced by exposure of the body below the iliac crests tosub-atmospheric pressure. J Physiol 1966;183:607–627.

31. Seidelin PH, Collier JG, Struthers AD, Webb DJ. Angiotensin IIaugments sympathetically mediated arteriolar constriction inman. Clin Sci 1991;81:261–266.

32. Goldstein DS, Feuerstein G, Izzo JL, Kopin IJ, Keiser HR. Validityand reliability of liquid chromatography with electrochemicaldetection for measuring plasma levels of norepinephrine andepinephrine in man. Life Sci 1981;28:467–475.

33. Morton JJ, Webb DJ. Measurement of plasma angiotensin II. ClinSci 1985;68:483–484.

34. Haber E, Koerner T, Page LB, Kliman P, Purnode A. Application of aradioimmunoassay for angiotensin I to the physiologic measure-ments of plasma renin activity in normal subjects. J Clin Endocri-nol Metab 1969;29:1349–1355.

35. Kerslake DM. The effect of the application of an arterial occlusioncuff to the wrist on the blood flow in the human forearm. J Physiol1949;108:451–457.

36. Whitney RJ. The measurement of volume changes in humanlimbs. J Physiol 1953;121:1–27.

37. Ames RP, Borkowski J, Sickinski AM, Laragh JH. Prolongedinfusions of angiotensin II and noradrenaline and blood pressure,electrolyte balance, and aldosterone and cortisol secretion innormal man and in cirrhosis with ascites. J Clin Invest 1965;44:1171–1186.

38. Lenz K, Hortnagl H, Magometschigg D, Kleinberger G, Druml W,Langner A. Functions of the autonomic nervous system in patientswith hepatic encephalopathy. Hepatology 1985;5:831–836.

39. Laragh JH, Cannon PJ, Bentzel CJ, Sickinski AM, Meltzer JI.Angiotensin II, norepinephrine, and renal transport of electrolytesand water in normal man and in cirrhosis with ascites. J ClinInvest 1963;42:1179–1192.

40. Lunzar MR, Manghani KK, Newman SP, Sherlock SPV, BernardAG, Ginsberg J. Impaired cardiovascular responsiveness in liverdisease. Lancet 1975;2:282–285.

41. MacGilchrist AJ, Sumner D, Reid JL. Impaired pressor reactivity incirrhosis: evidence for a peripheral vascular defect. Hepatology1991;13:689–694.

42. Ryan J, Jennings G, Dudley F, Chin-Dusting J. Smooth muscle–derived nitric oxide is elevated in isolated forearm veins in humanalcoholic cirrhosis. Clin Sci 1996;91:23–28.

43. Campillo B, Chabrier P, Pelle G, Sediame S, Atlan G, Fouet P,Adnot S. Inhibition of nitric oxide synthesis in the forearm arterialbed of patients with advanced cirrhosis. Hepatology 1995;22:1423–1429.

44. Ryan J, Sudhir K, Jennings G, Esler M, Dudley F. Impairedreactivity of the peripheral vasculature to pressor agents inalcoholic cirrhosis. Gastroenterology 1993;105:1167–1172.

45. Calver A, Harris A, Maxwell JD, Vallance P. Effect of local inhibitionof nitric oxide synthesis on forearm blood flow and dorsal hand veinsize in patients with alcoholic cirrhosis. Clin Sci 1994;86:203–208.

46. Fernandez-Seara J, Prieto J, Quiroga J, Zozaya JM, Cobos MA,Rodrigues-Eire JL, Garci-Plaza A, Leal J. Systemic and regionalhemodynamics in patients with liver cirrhosis and ascites withand without renal failure. Gastroenterology 1989;97:1304–1312.

47. Maroto A, Gines P, Arroyo V, Gines A, Salo J, Claria J, Jimenez W, Bru C,Rivera F, Rodes J. Brachial and femoral artery blood flow in cirrhosis:relationship to kidney dysfunction. Hepatology 1993;17:788–793.

48. Castro A, Jimenez W, Claria J, Ros J, Martinez JM, Bosch M,Arroyo V, Piulats J, Rivera F, Rodes J. Impaired responsiveness toangiotensin II in experimental cirrhosis: role of nitric oxide.Hepatology 1993;18:367–372.

49. Bosch J, Arroyo V, Betriu A, Mas A, Carrilho F, Rivera F, Navarro-

Lopez F, Rodes J. Hepatic hemodynamics and the renin-angiotensin-aldosterone system in cirrhosis. Gastroenterology1980;78:92–99.

50. Baan J, Chang PC, Vermeij P, Pfaffendorf M, van Zwieten PA.Effects of losartan on the vasoconstrictor responses to angioten-sin II in the forearm vascular bed. Cardiovasc Res 1996;32:973–979.

51. Schneider AW, Kalk JF, Klein CP. Effect of losartan, an angioten-sin II receptor antagonist, on portal pressure in cirrhosis.Hepatology 1999;29:334–339.

52. Wong F, Logan A, Blendis L. Systemic hemodynamic, forearmvascular, renal, and humoral responses to sustained cardiopulmo-nary baroreceptor deactivation in well-compensated cirrhosis.Hepatology 1995;21:717–724.

53. Greenfield ADM, Brown E, Goei JS. Circulatory responses to abruptrelease of blood accumulated in the legs. Physiologist 1963;6:191.

54. Zoller RP, Mark AL, Abboud FM, Schmid PG, Heistad DD. The roleof low pressure baroreceptors in reflex vasoconstrictor re-sponses in man. J Clin Invest 1972;51:2967–2972.

55. Johnson JM, Loring B, Niederberger Eisman MM. Human splanch-nic and forearm vasoconstrictor responses to reductions of rightatrial and aortic pressures. Circ Res 1974;34:515–524.

56. Hirsch AT, Levenson DJ, Cutler SS, Dzau VJ, Creager MA. Regionalvascular responses to prolonged lower body negative pressure innormal subjects. Am J Physiol 1989;257:H219–H225.

57. Jacobs M, Goldstein DS, Willemsen JJ, Smits P, Thien T, LendersJWM. Differential effects of low and high intensity lower bodynegative pressure on noradrenaline and adrenaline kinetics inhumans. Clin Sci 1996;90:337–343.

58. Oliver MI, Miralles R, RubiesPrat J, Navarro X, Espadaler JM, SolaR, Andreu M. Autonomic dysfunction in patients with non-alcoholic chronic liver disease. J Hepatol 1997;26:1242–1248.

59. Wong F, Sniderman K, Blendis L. The renal sympathetic andrenin-angiotensin response to lower body negative pressure inwell-compensated cirrhosis. Gastroenterology 1998;115:397–405.

60. Duranteau J, Pussard E, Berdeaux A, Giudicelli JF. Effects of theangiotensin type 1 receptor antagonist, losartan, on systemicand regional vascular responses to lower body negative pressurein healthy volunteers. Br J Clin Pharmacol 1995;40:431–438.

61. Bishop VS, Haywood RJ. Hormonal control of cardiovascularreflexes. In: Zucker I, Gilmore JP, eds. Reflex control of circula-tion. Boca Raton, FL: CRC, 1991:253–272.

62. Rankin LI, Luft FC, Henry DP, Gibbs PS, Weinberger MH. Sodiumintake alters the effects of norepinephrine on blood pressure.Hypertension 1981;3:650–656.

63. Stein M, Nelson R, Brown M, Wood M, Wood AJJ. Dietary sodiumintake modulates vasodilation mediated by nitroprusside but notby methacholine in human forearm. Hypertension 1995;25:1220–1223.

64. Binder HJ. Angiotensin receptor antagonists in the pharmacologi-cal therapy of portal hypertension: a caution. Gastroenterology1999;117:740–742.

Received May 28, 1999. Accepted November 2, 1999.Address requests for reprints to: Peter Clive Hayes, M.D., Ph.D.,

Department of Medicine, Royal Infirmary of Edinburgh, LauristonPlace, Edinburgh EH3 9YW, Scotland. e-mail: p.hayes@ed.ac.uk; fax:(44) 131-229-2948.

Supported by a grant from the Sir Stanley and Lady Davidsonfund, by an Egyptian Government Scholarship (to A.H.), and by aResearch Leave Fellowship from the Wellcome Trust (WT 0526330;to D.J.W.).

The authors thank Neil Johnston (Clinical Pharmacology Unit,Western General Hospital, Edinburgh, Scotland) and Rhona Stevens(Department of Clinical Biochemistry, Royal Hospital for SickChildren, Edinburgh, Scotland) for providing technical assistance inperforming the humoral assays.

572 HELMY ET AL. GASTROENTEROLOGY Vol. 118, No. 3