OUP user menu

Peripheral vascular tone in patients with cirrhosis: role of the renin–angiotensin and sympathetic nervous systems

David Ernest Newby, Rajiv Jalan, Satoko Masumori, Peter Clive Hayes, Nicholas Antony Boon, David John Webb
DOI: http://dx.doi.org/10.1016/S0008-6363(98)00008-X 221-228 First published online: 1 April 1998


Objective: The aims of the study were to establish the roles of angiotensin II and of the cardiopulmonary baroreceptor reflex in the regulation of peripheral vascular tone in patients with cirrhosis. Methods: Forearm blood flow responses to subsystemic, locally active intrabrachial infusions were measured in patients with Child's Grade C cirrhosis and matched controls using bilateral venous occlusion plethysmography. Responses were determined to the angiotensin II type I receptor antagonist, losartan, noradrenaline, angiotensin II and the nitric oxide synthase inhibitor, l-NG-monomethyl arginine. Results: Losartan at 30 and 90 μg/min caused no significant change in blood flow in controls, but caused 23±6% and 27±5% increases in patients respectively (p<0.001). Lower body negative pressure caused a mean bilateral reduction in forearm blood flow of 20±4% in controls (p<0.001) but only tended to reduce flow (9±5%; p=0.06) in patients (p<0.001; controls vs. patients). Noradrenaline, angiotensin II and l-NG-monomethyl arginine caused significant vasoconstriction (p<0.001) in both patients and controls although angiotensin II caused significantly less vasoconstriction in patients (p=0.01). Conclusions: We conclude that angiotensin II makes an important contribution to basal peripheral vascular tone in patients with cirrhosis in the face of reduced vascular responses to its local administration. In addition, the vasoconstrictor response to cardiopulmonary baroreceptor unloading is attenuated despite normal vascular responses to noradrenaline. These responses are consistent with chronic activation of the renin–angiotensin and sympathetic nervous systems in patients with advanced cirrhosis.

  • Angiotensin
  • Regional blood flow
  • Renin angiotensin system
  • Adrenergic agonists
  • Autonomic nervous system

Time for primary review 25 days.

1 Introduction

Cirrhotic liver disease is characterised by a hyperdynamic circulation and high cardiac output [1]associated with widespread vasodilatation [2]and low peripheral vascular resistance [3]. Compensatory activation of neurohumoral reflexes involving the sympathetic nervous [4–6]and renin–angiotensin systems [7–9], helps to sustain vascular tone and blood pressure. A failure to compensate fully for the peripheral vasodilatation may be a consequence of hyporesponsiveness to, and desensitisation of, these counter-regulatory vasoconstrictor mechanisms.

Reduced pressor responses to systemic administration of α-adrenoreceptor agonists and angiotensin II occur in patients with severe cirrhotic liver disease [10]. However, when studying in vivo vascular responses in man, systemically administered agents can cause effects on other organ systems, such as the liver, brain, kidney and heart, and influence neurohumoral reflexes through changes in systemic haemodynamics. Because of these confounding influences, local vascular responses cannot be wholly attributed to a direct effect of the agent on the blood vessels. Subsystemic locally active brachial artery infusions combined with bilateral forearm blood flow measurements using venous occlusion plethysmography provide a means of assessing the responses of an isolated vascular bed without invoking counter-regulatory reflexes [11]. This technique has been instrumental in demonstrating the contribution of basal nitric oxide [12]and endothelin-1 production [13]to the maintenance of peripheral resistance vessel tone in man.

The role of the renin–angiotensin system in counteracting the peripheral vasodilatation and hyperdynamic circulation of cirrhotic liver disease has not previously been fully characterised because, until recently, specific and selective pharmacological antagonists were lacking. Losartan, a selective AT1 receptor antagonist that is devoid of agonist activity, is now available for clinical use. Therefore, the aims of the present study in patients with cirrhosis were as follows: first, to establish the role of endogenous angiotensin II in the maintenance of peripheral vascular tone; second, to determine the effect of sympathetically stimulated vasoconstriction in the presence and absence of angiotensin II antagonism; and third, to document the peripheral vascular responses to the vasoconstrictors noradrenaline, angiotensin II and the nitric oxide synthase inhibitor, l-NG-monomethyl arginine (l-NMMA).

2 Methods

Patients with stable Child's Grade C cirrhosis and ascites were recruited together with age- and sex-matched healthy volunteers. Studies were undertaken with the approval of the local research ethics committee and in accordance with the Declaration of Helsinki (1989) of the World Medical Association. The written informed consent of each subject was obtained before the study. None of the subjects received vasoactive or non-steroidal anti-inflammatory drugs in the week before each phase of the study, and all abstained from alcohol for 24 h and from food, caffeine-containing drinks and tobacco for at least 4 h before each study. Patients with alcoholic cirrhosis abstained from alcohol for at least 6 months prior to entry into the study. All patients had been maintained on spironolactone (50 mg bd) therapy, without additional diuretic therapy, for at least 1 month prior to entry into the study. In previous studies, complete withdrawal of spironolactone has resulted in unacceptable clinical deterioration in these patients. Therefore, patients remained on spironolactone therapy but, to avoid acute dosing effects, the morning dose of spironolactone was omitted on the day of attendance. The studies were performed in a quiet, temperature controlled room maintained at 23.5–25.0°C. To minimise between-subject variability, pre-menopausal female control subjects were studied between 7 and 12 days of their menstrual cycle [14]. All female patients had amenorrhea.

2.1 Drugs and intra-arterial administration

Losartan (Dupont-Merck, Wilmington, USA), noradrenaline (Levophed; Sanofi Winthrop, Guildford, UK), angiotensin II (Clinalfa, AG Läufelfingen, Switzerland) and l-NMMA (Clinalfa AG) were dissolved in physiologic saline (0.9%; Baxter Healthcare, Thetford, UK) and administered intra-arterially. To prevent its oxidation, noradrenaline was dissolved in saline containing 0.1% ascorbic acid (Evans Medical, Langhurst, UK). Doses of losartan (30–90 μg/min) were chosen on the basis of previous studies [15]to achieve an effective subsystemic and locally active concentration.

The brachial artery of the non-dominant arm was cannulated with a 27-standard wire gauge steel needle (Cooper's Needle Works, Birmingham, UK) attached to a 16-gauge epidural catheter (Portex, Hythe, UK) under 1% lignocaine (Xylocaine; Astra Pharmaceuticals, Kings Langley, UK) local anaesthesia. Patency was maintained by infusion of saline via an IVAC P1000 syringe pump (IVAC, Basingstoke, UK). The total rate of intra-arterial infusions was maintained constant throughout all studies at 1 ml/min.

2.2 Measurements

Blood flow was measured in both the infused and non-infused forearms by venous occlusion plethysmography using mercury-in-silastic strain gauges applied to the widest part of the forearm [11]. During measurement periods the hands were excluded from the circulation by rapid inflation of the wrist cuffs to a pressure of 220 mmHg using E20 Rapid Cuff Inflators (D.E. Hokanson, Washington, DC, USA). Upper arm cuffs were inflated intermittently to 40 mmHg for 10 s in every 15 s to achieve venous occlusion and obtain plethysmographic recordings. Analogue voltage output from an EC-4 strain gauge Plethysmograph (D.E. Hokanson) was processed by a MacLab analogue-to-digital converter and Chart v3.3.8 software (AD Instruments, Castle Hill, Australia) and recorded onto a MacIntosh Classic II computer (Apple Computers, Cupertino, USA). Calibration was achieved using the internal standard of the plethysmograph.

Blood pressure was monitored in the non-infused arm at intervals throughout each study using a semi-automated non-invasive oscillometric sphygmomanometer [16](Takeda UA 751, Takeda Medical, Tokyo, Japan).

Subjects were rested supine in a plastic covered steel cage enclosing the lower body from the waist, as described previously [17]. Lower body negative pressure (LBNP) was applied using an industrial strength vacuum cleaner regulated by a pressure control unit (Medical Physics Department, Edinburgh, UK) to produce a constant negative pressure of 15 mmHg which does not affect systemic blood pressure or heart rate [17]. Alteration to and from atmospheric pressure was attained within 1–2 s.

Ten minutes before giving losartan by infusion and 10 min after its completion, 30 ml of blood was withdrawn from the non-infused arm and 10 ml admixed with each of 1 ml of 1% disodium EDTA, 0.5 ml of 0.45% O-phenanthroline/4.65% disodium EDTA and 1 ml of 1% disodium EDTA/2% sodium metabisulphite. The samples were placed on ice and immediately centrifuged at 2000 g for 15 min. Plasma was frozen and stored at −80°C prior to assay for plasma angiotensin II, endothelin-1, big endothelin-1, adrenaline and noradrenaline concentrations. Following extraction using Bond Elut® columns (Varian, Harbor City, CA, USA) [18], plasma angiotensin II (Peninsula Laboratories Europe, St Helens, UK), endothelin-1 (Peninsula Laboratories Europe) and big endothelin-1 (Peninsula Laboratories Europe) concentrations were determined by radioimmunoassay as previously described [19, 20]. The intra-assay coefficients of variability were 5.2%, 7.0% and 7.2% respectively and the inter-assay coefficients of variability were 8.6%, 9.0% and 9.3% respectively. The cross reactivities of the endothelin-1 assay were endothelin-1 (100%), endothelin-2 (7%), endothelin-3 (7%), big endothelin-1 (10%), C-terminal fragment (0%), angiotensin I (0%) and angiotensin II (0%). The cross reactivities of the big endothelin-1 assay were endothelin-1 (0%), endothelin-2 (0%), endothelin-3 (0%), big endothelin-1 (100%) and C-terminal fragment (100%). Plasma adrenaline and noradrenaline concentrations were determined by an electrochemical method after separation by reverse phase high performance liquid chromatography [21].

2.3 Study design

Subjects rested recumbent throughout each study. Strain gauges and cuffs were applied and the brachial artery of the non-dominant arm cannulated. Measurements of forearm blood flow were made for the last 3 min of each infusion period unless otherwise stated. Before participating in one of the following protocols, saline was infused for the first 30 min to allow time for equilibration, with forearm blood flow measurements being made every 10 min and baseline blood flow being taken as the final measurement.

In protocol 1, eight patients and eight matched control subjects received, in order, saline, losartan 30 μg/min, losartan 90 μg/min [15]and saline, each for 13 min. Forearm blood flow was measured continuously for the last 6 min of each infusion. Lower body negative pressure was applied for the last 3 min of the forearm blood flow measurement. Following 15 min of further saline infusion, noradrenaline was infused intra-arterially at doses of 20, 60, 180 and 540 pmol/min [8, 15], each for 6 min.

In protocol 2, eight patients and eight matched control subjects received incremental doses of angiotensin II (0.1, 1, 10 and 100 pmol/min) [8, 15], each dose administered into the brachial artery for 6 min, followed by 30 min of saline infusion. Thereafter, l-NMMA at 4 μmol/min [8, 22]was administered for 15 min.

Protocols 1 and 2 were conducted at least one week apart and in random order with five patients and seven control subjects common to both protocols.

2.4 Data analysis and statistics

Plethysmographic data were extracted from the Chart data files and forearm blood flows were calculated for individual venous occlusion cuff inflations by use of a template spreadsheet (Excel v5.0; Microsoft). Recordings from the first 60 s after wrist cuff inflation were not used because of the initial reflex vasoconstriction this procedure causes [23]. Usually, the last five flow recordings in each 3 min measurement period were calculated and averaged for each arm. To reduce the variability of blood flow data, the ratio of flows in the two arms was calculated for each time point: in effect using the non-infused arm as a contemporaneous control for the infused arm [11, 24]. Percentage changes in the infused forearm blood flow were calculated [11, 24] as follows: Embedded Image where Ib and NIb are the infused and non-infused forearm blood flows at baseline (time 0) respectively, and It and NIt are the infused and non-infused forearm blood flows at a given time point respectively.

Data were examined by two factor analysis of variance (ANOVA) with repeated measures, two tailed paired Student's t-test and regression analysis using Excel v5.0 (Microsoft). All results are expressed as means±standard errors of the mean. Statistical significance was taken at the 5% level. Based on the responses, the dose response shifts were calculated for the ED50; the dose producing 50% vasoconstriction from baseline.

3 Results

Baseline patient characteristics and concomitant therapy is detailed in Table 1. Three patients with cirrhosis underwent liver transplantation before completing their second study phase and were replaced. Patients had a significantly lower systolic and diastolic blood pressure, and higher plasma angiotensin II, endothelin-1 and big endothelin-1 concentrations than controls. However, there were no significant differences in resting heart rate, baseline forearm blood flows or plasma adrenaline and noradrenaline concentrations in the infused and non-infused arms between protocols or between patients and controls. Blood pressure, heart rate and non-infused forearm blood flow did not change significantly during infusions of losartan, noradrenaline, l-NMMA or angiotensin II (data on file; Fig. 1).

Fig. 1

Absolute (infused forearm, ■; non-infused forearm, □) blood flow and percentage change (infused/non-infused forearm, ○) in blood flow from time 0, with intermittent application (arrows) of lower body negative pressure (LBNP), to saline, 30 μg/min and 90 μg/min of losartan in patients with cirrhosis (right panels) and healthy age- and sex-matched control subjects (left panels) respectively. NS – not significant; *p<0.001, ANOVA.

View this table:
Table 1

Subject characteristics

Sex4 male: 7 female4 male: 5 female
Child's class C11-
Pugh score10.2±1.5-
Aetiology of cirrhosisPBC6-
Biochemical profileBilirubin128.3±24.9-(mmol/l)
Prothrombin time16.2±0.7-(s)
Drug therapySpironolactone11-
Blood pressureSystolic114±3124±4*(mmHg)
Heart rate73±567±3(/min)
Baseline blood flowInfused3.2±0.54.2±0.7(ml/100ml/min)
Plasma endothelin-11.7±0.21.3±0.04†(fmol/ml)
Plasma big endothelin-117.5±3.86.9±1.0‡(fmol/ml)
Plasma angiotensin IIBasal21.6±5.83.2±0.3§(fmol/ml)
Post losartan32.6±10.73.3±0.4§(fmol/ml)
Plasma noradrenaline2.7±0.71.3±0.4¶(pmol/ml)
Plasma adrenaline0.3±0.10.4±0.2(pmol/ml)
  • PBC – Primary Biliary Cirrhosis; CAH – Chronic Active Hepatitis; PSC – Primary Sclerosing Cholangitis.

    *p=0.04; †p=0.02; ‡p=0.01; §p=0.007; ¶p=0.09.

3.1 Effects of losartan on basal blood flow and responses to lower body negative pressure

Plasma angiotensin II concentrations did not change significantly during the study (Table 1).

In the healthy control subjects, there were no significant changes in blood flow of the infused forearm during infusions of saline or losartan (Fig. 1). The Student's t distribution gives 95% confidence intervals of −3.3 to +7.8% and −2.6 to +9.0% for percentage changes in forearm blood flow with 30 μg/min and 90 μg/min of intra-arterial losartan respectively. During saline infusion, LBNP caused a 23.1±3.7% and 21.8±2.8% reduction in the infused and non-infused forearm blood flow respectively (p<0.001). The vasoconstriction induced by the application of LBNP was not different between the infused and non-infused arms either during saline or losartan infusion periods (Figs. 1 and 2).

Fig. 2

Percentage vasoconstriction to lower body negative pressure in infused (black hatched block, closed block) and non-infused (grey hatched block, open block) forearm during infusion of saline, 30 μg/min and 90 μg/min of losartan in patients with cirrhosis and healthy age and sex matched control subjects respectively. p<0.001; 2-way ANOVA controls vs. patients.

In patients with cirrhosis, blood flow in the infused forearm increased by 26.3% (95% Cl: +12.7 to +39.9%) and 29.5% (+18.6 to +40.4%) with 30 μg/min and 90 μg/min of intra-arterial losartan respectively (p<0.001; Fig. 1). The increase in blood flow at 90 μg/min of losartan correlated inversely with the baseline plasma angiotensin II concentration (r=−0.93; p=0.007). LBNP did not significantly reduce the blood flow in either arm (p=0.06), and was significantly less than that achieved in the control subjects (p<0.001; Fig. 2). The response to LBNP in the infused forearm was also unaffected by losartan (Figs. 1 and 2).

3.2 Effect of angiotensin II, noradrenaline and l-NMMA infusions

Angiotensin II and noradrenaline caused dose-dependent vasoconstriction in both patients and controls (p<0.001 for all). There was no significant difference in the magnitude of the vasoconstriction response to noradrenaline between patients and controls. However, angiotensin II caused significantly less vasoconstriction in patients with cirrhosis (p=0.01 vs. controls) with a ∼2.5-fold right shift in the ED50 from 13 to 30 pmol/min (Fig. 3).

Fig. 3

Percentage change in blood flow (infused/non-infused forearm) during incremental infusions of angiotensin II (●) and noradrenaline (■) in patients with cirrhosis (closed symbols) and healthy age and sex matched control subjects (open symbols). *p=0.01; 2-way ANOVA controls vs. patients.

At the end of the period of l-NMMA infusion, blood flow in the infused forearm was reduced by 33.0±2.9% and 29.9±4.2% in patients and controls respectively (p<0.001 for both). The difference between responses in patients and controls was not significant.

4 Discussion

We have previously shown in healthy volunteers [15]that intra-arterial losartan infusion, at doses of 30–300 μg/min, is locally active only and is able to reverse completely the vasoconstriction caused by angiotensin II infusion at 1 pmol/min. This dose of angiotensin II is sufficient to cause a comparable increase in plasma angiotensin II to that seen in patients with cirrhosis. In the present study, we have shown for the first time that brachial artery infusion of losartan significantly increases forearm blood flow by ∼30% in patients with cirrhosis. We have also shown, in agreement with previous work [15, 25], that intra-arterial losartan has no effect on basal forearm blood flow or vascular resistance in healthy control subjects. These findings indicate that angiotensin II contributes to the maintenance of basal peripheral vascular tone in patients with liver cirrhosis but not healthy control subjects.

We have previously shown [15]that acute sodium depletion in healthy man causes a 2–3-fold increase in plasma angiotensin II concentrations and is associated with an increase in forearm blood flow of ∼70% with losartan infusion. Therefore, given that in the patients with cirrhosis the plasma angiotensin II concentrations were 7-fold greater than controls, the magnitude of the vasodilatation to losartan is perhaps smaller than might be anticipated, consistent with a reduced responsiveness to angiotensin II. This is also supported by two other findings in our studies; a reduced vasoconstriction to angiotensin II infusion, and a negative correlation of plasma angiotensin II concentrations with the vasodilatation in response to losartan. This is in keeping with the action of angiotensin II as a compensatory humoral mediator, the vasoconstriction to which is diminished with chronic activation.

Also, we have shown an impairment of the vasoconstrictor response to LBNP in patients with cirrhosis. This is consistent with previous reports of a reduced vasoconstrictor [26]and pressor [7]response to head-up tilt. It is well known that the sympathetic nervous and renin–angiotensin systems act synergistically to influence vascular tone. Even at doses insufficient to cause vasoconstrictor or pressor responses, angiotensin II augments sympathetically stimulated vasoconstriction [17]via a prejunctional adrenoreceptor mediated mechanism [27]. Therefore, it might be anticipated that during renin–angiotensin system activation in the presence of elevated plasma angiotensin II concentrations the vasoconstrictor response to LBNP would be augmented; an effect which should be blocked by the infusion of losartan. In contrast, we demonstrated a significant impairment of the LBNP response which was unaffected by losartan infusion. The only study to address this issue of sympathetically stimulated forearm vasoconstriction through cardiopulmonary baroreceptor unloading [28], reported normal forearm vasoconstriction to LBNP in patients with cirrhosis. However, these patients had mild alcoholic cirrhosis without ascites or humoral activation and the forearm blood flow measurements appears to have been made for one minute only. Plethysmographic blood flow recordings within the first 60 s after wrist cuff inflation may be unreliable because of the instability and marked variability in blood flow that occurs initially [11, 23, 24]. We have utilised established methodology and have studied patients with advanced cirrhosis who have high plasma angiotensin II concentrations and a trend for higher plasma noradrenaline concentrations. Given our observation of normal peripheral vascular responses to noradrenaline infusion and the known increases in sympathetic nerve activity in cirrhosis [6], the current findings are consistent with an impaired response to baroreceptor unloading which probably reflects desensitisation of the neuronal reflex secondary to chronic stimulation from relative intravascular volume depletion and hypotension.

Our findings of normal noradrenaline responses but impaired angiotensin II induced vasoconstriction contrast with the results of previous studies by Ryan and colleagues [8, 22]which showed abnormal responses to both agents. In these studies, patients had mild alcoholic cirrhosis without significant humoral changes and were assessed using unilateral forearm plethysmography. The infusion periods of these studies were very brief at 2 min and the plethysmographic recordings and analysis were not predetermined or standardised. Moreover, there is inherent variability in forearm blood flow [11, 24]and the effects of unilateral brachial artery infusions are probably better assessed by comparing the infused forearm responses with the contralateral non-infused arm which acts as a contemporaneous control. It is also recognised that ethanol consumption impairs vascular responsiveness to noradrenaline [29], which remains a concern in patients with alcoholic liver cirrhosis. In our study we excluded all patients with alcoholic cirrhosis who had not achieved sustained abstinence for more than 6 months. In addition, we assessed the noradrenaline responses in the presence of ongoing AT1 receptor antagonism since the effect of intra-arterial losartan persists for at least 60 min after cessation of infusion [15]. An interaction with concomitant endogenous angiotensin II mediated vasoconstriction could, therefore, account for the attenuation in the noradrenaline response seen previously. Overall, the differences between the findings of the current study and those of Ryan et al. [8, 22]probably reflect a combination of differing patient characteristics and methodologies.

The humoral characteristics of our patient population are consistent with the previously documented changes in liver cirrhosis [5, 9]with compensatory increases in plasma angiotensin II and endothelin-1 concentrations which counterbalance the pronounced peripheral vasodilatation. Patients with chronic renal failure have isolated increases in plasma endothelin-1 concentrations without a rise in big endothelin-1 because of a reduction in endothelin-1 clearance by the kidney. In contrast, we have found a predominant increase in big endothelin-1 concentrations consistent with an increase in endothelin-1 production [30].

None of the known endogenous mediators are present in sufficient concentrations to account for the pronounced reduction in peripheral vascular resistance found in cirrhosis. In agreement with previous work [22, 31, 32], we have shown that patients with cirrhosis have normal basal forearm nitric oxide production as assessed by the vasoconstrictor response to l-NMMA. It is therefore unlikely that overproduction of nitric oxide mediates the characteristic peripheral vasodilatation of liver cirrhosis [31]. Indeed, it is also apparent from the l-NMMA and noradrenaline responses that the resistance vessel smooth muscle cells have a normal functional ability to contract and regulate peripheral vascular resistance in patients with cirrhosis.

In this study, we have confined ourselves to examining patients who have severe liver cirrhosis and ascites associated with pronounced humoral activation. Therefore, extrapolation of our findings to milder forms of liver cirrhosis can only be made tentatively. Antecedent spironolactone therapy may also have affected our results, although the main effects of aldosterone antagonism are to cause a mild accentuation of the plasma renin activity [33]and angiotensin II concentrations [Allan D. Struthers, personal communication]. Finally, baseline systemic haemodynamics in the patient and control populations were significantly different. Consequently, the geometry and resting tonus of the forearm resistance vessels will be dissimilar and may affect the magnitude of responses to the study agents [24]. However, given the similar basal forearm blood flows and responses to l-NMMA and noradrenaline infusion, we believe that the observed differences in responsiveness to angiotensin II are real. Moreover, the substantial differences in responses to losartan infusion and LBNP cannot be explained by these haemodynamic differences.

In conclusion, we have found that in patients with advanced cirrhotic liver disease angiotensin II makes an important contribution to the maintenance of peripheral vascular tone even though the resistance vessels are less sensitive to its vasoconstrictor actions. In addition, the vasoconstrictor response to cardiopulmonary baroreceptor unloading is attenuated in association with preservation of vascular responsiveness to noradrenaline. It is suggested that these abnormalities may reflect a response to chronic activation of the renin–angiotensin and sympathetic nervous systems in patients with advanced cirrhosis.


DEN is the recipient of a British Heart Foundation Junior Research Fellowship (FS/95009). This work was supported by a grant from the Sir Stanley and Lady Davidson fund. We would like to acknowledge the assistance of Neil Johnston (Clinical Pharmacology Unit, Western General Hospital, Edinburgh) and Rhona Stevens (Department of Clinical Biochemistry, Royal Hospital for Sick Children, Edinburgh) in performing the assays.


  • 1 Present address: Yokohama City University, School of Medicine, Second Department of Internal Medicine, 3-9 Fukuura, Kanazawa-Ku, Yokohama 236, Japan.


View Abstract