SINGLE DOSE STUDIES (TABLE 1)

There are four studies involving 63 cirrhotics using ACE inhibitors.22-25 Captopril was used in 55 patients at doses ranging from 12.5 to 150 mg. In the first three studies captopril caused a significant reduction in glomerular filtration rate (GFR) with a consequent decrease in urinary volume and urinary sodium excretion, in association with significant decreases in mean arterial pressure (MAP).22-24 In the only randomised placebo controlled study, by Gentilini and colleagues,25 patients who had arterial hypotension after a single intravenous dose of captopril were excluded from the analysis so as to elucidate if the deleterious renal effects in the previous studies were secondary to arterial hypotension. However, captopril still induced a significant decrease in GFR and urinary sodium excretion in patients with or without ascites. Therefore, ANG-II contributes to the maintenance of renal haemodynamics in cirrhosis and consequently ACE inhibitors are harmful.

A subsequent study of 10 mg of enalapril in eight cirrhotic patients22 showed increased hourly urine volume and renal sodium excretion during the first two hours, but 24 hour urinary volume was unchanged.

Recently, Girgrah and colleagues26 evaluated the role of ANG-II in sodium homeostasis in nine preascitic cirrhotic patients using losartan, its receptor antagonist. A dose-response study showed that 7.5 mg was the optimal dose to produce natriuresis without systemic hypotension, normalising renal sodium without changes in systemic or renal haemodynamics. In the absence of GFR or renal plasma flow (RPF) changes, one would speculate that ANG-II has a direct sodium retaining effect on the renal tubule. However, baseline systemic ANG-II levels were significantly lower in preascitic cirrhotics compared with controls. The authors suggested that intrarenal production of ANG-II or increased sensitivity of preascitic cirrhotics to the sodium retaining effect of ANG-II would explain this discrepancy. Alternatively, losartan could have reduced portal pressure and thus improved sodium retention, without a direct effect on renal haemodynamics.

MULTIPLE DOSE STUDIES (TABLE 2)

In 1983, Saruta and colleagues27 used the ACE inhibitor SQ14225 in six cirrhotics for seven days. ANG-II and plasma aldosterone levels decreased and PRA increased markedly. Although systemic blood pressure was lowered, urinary sodium excretion increased significantly in five of six patients, and ascites improved but did not disappear. Since then, five studies28-32 comprising 60 patients with ascites have been reported using captopril (9–75 mg/day) given for 3–21 days. In three studies,28-30 captopril caused a significant reduction in body weight, enhancement of urinary sodium excretion, and improvement of fluid balance in the majority of patients. However, Wood and colleagues31 used captopril for three days in four patients with resistant ascites and reported falls in MAP with subsequent falls in GFR, decreases in sodium excretion, and no improvement in ascites. These authors suggested that ACE inhibitors have little place in the treatment of patients with resistant ascites and may be harmful. The only randomised placebo controlled study was performed by Tsai and colleagues32and included 50 patients with ascites. PRA became significantly raised in the captopril group, and this was considered as adequate blockade of ACE. Nevertheless, there was no significant change in urine volume, creatinine clearance, or RPF, and the improvement in urine sodium excretion was not significant.

Enalapril was used in two small studies33 ,34 with 18 patients. In one,33 a significant increase in urinary volume and sodium excretion was seen. In both studies creatinine clearance improved significantly.

Evaluation of the above studies is difficult because most were neither placebo controlled nor randomised. Furthermore, as patient characteristics differed considerably between studies (for example, Child-Pugh class, presence of ascites, salt restriction, use of diuretics, and renal function impairment) it is difficult to make comparisons.

Single doses of captopril worsen renal function22-25 but this reflects only the immediate haemodynamic effects, which may not be the case with long term treatment. Indeed, some neurohormonal changes (for example, decrease in aldosterone levels) are evident only after long term treatment, as highlighted by Ohnishi and colleagues.33 There are five multiple dose studies showing that captopril and enalapril enhance renal sodium excretion.28-30 ,33 ,34 However, it is noteworthy that baseline urinary sodium excretion was more than 70 mmol/day in all of these studies compared with the very low baseline sodium excretion in the studies reporting no improvement.31 ,32 It is possible that only patients with an earlier stage of liver disease, and hence less renal impairment, may benefit from the use of ACE inhibitors.

However, in decompensated cirrhotics there is reduced peripheral resistance and a compensatory increase in plasma volume and cardiac output.35 Under these conditions, activation of RAAS plays a critical role in maintaining arterial pressure and in autoregulation of GFR, which maintains GFR at near constant values despite a reduction in renal perfusion pressure.36 Therefore, administration of ACE inhibitors or ANG-II antagonists in these patients would result in deterioration of renal function and worsening of sodium excretion.

Conversely, in patients with preascitic cirrhosis, in whom a subtle sodium handling abnormality is found, it is still not entirely clear if this is due to compensatory activation of the RAAS counteracting the consequences of hyperdynamic circulation37 or if the RAAS plays a crucial pathogenetic role in abnormal sodium handling.38 Girgrah and colleagues26emphasised the primary role of ANG-II in increased renal sodium reabsorption to explain the positive role of losartan in normalisation of sodium handling. Conversely, in the recent study by Helmy and colleagues39 (despite the increase in plasma ANG-II concentrations in preascitic cirrhotics) there was still hyporesponsiveness of the peripheral vessels after administration of losartan. This supports the concept that ANG-II does not have a role in maintenance of vascular tone in these patients. Therefore, use of ANG-II inhibitors could be beneficial in early cirrhosis to counteract the effects of raised ANG-II levels in renal sodium handling and would not have deleterious effects on systemic haemodynamics.

ACE INHIBITORS

Captopril has been evaluated in four studies (table3).23 ,29 ,32 ,40 In three (22 patients), captopril was given as a single oral dose.23 ,29 ,40 No significant change in HVPG was detected in any study. In addition, there was a small but significant decrease in MAP with deterioration of systemic haemodynamics. Ibarra and colleagues29 evaluated captopril after three weeks of administration, reporting a significant reduction in free hepatic venous pressure while wedged hepatic venous pressure (WHVP) decreased less. Therefore, HVPG was increased, although this was not significant. Finally, in the study of Tsai and colleagues,32 captopril given for 14 days had no effect on the systemic or hepatic circulation.

The efficacy of enalapril in the blockade of RAAS and therefore in reducing portal pressure was evaluated in two small studies (table 3). Chiang and colleagues41 used 2.5 mg of enalaprilat, the active metabolite of enalapril, by intravenous infusion, in 20 patients with cirrhosis associated with hepatitis B. HVPG was reduced significantly within 30 minutes in 13 of 20 patients. Although MAP fell, liver and renal function did not change significantly. Svoboda and colleagues42 reported a significant drop in HVPG after three months of administration of enalapril in 12 cirrhotics undergoing long term sclerotherapy; no reduction in HVPG was found in 13 controls who received placebo. Systemic haemodynamics and liver function did not change.

Only two of the above studies were randomised with the use of placebo32 ,42 but the small number of patients with different clinical characteristics makes it difficult to generalise. However, similar to the absence of renal effects, captopril failed to reduce portal pressure in all four studies.23 ,29 ,32 ,40Interestingly, enalapril proved effective in two small studies (32 patients) although there was a poor response in patients with severe liver dysfunction (Child C). This may be attributed to inadequate dosage to counteract the high RAAS in these patients. However, the significant changes in MAP after enalapril raise the possibility that the potential adverse effects may outweigh the potential benefits in larger doses. These opposite results for captopril compared with the enalapril studies are not easy to understand. In one study, enalapril was more effective in preventing the postural induced increase in aldosterone concentrations compared with captopril.43 This may reflect more complete inhibition of ACE. Whether such differences result in different clinical effects is debatable.

ANG-II INHIBITORS

During long term ACE inhibition, alternative enzymatic pathways to ACE, such as trypsin and cathepsin, can also convert angiotensin I to ANG-II.19 This could explain why ACE inhibitors do not reduce portal pressure in cirrhotics.44 Moreover, ACE participates in the breakdown of bradykinin to inactive peptides. Accumulation of bradykinin after ACE inhibition would worsen the systemic hyperdynamic circulation in cirrhosis and thus outweigh the beneficial effects of ACE inhibitors on portal pressure.

The development of specific antagonists of ANG-II receptors has renewed interest in the role of RAAS activation in portal hypertension and the potential benefit from its inhibition. Early in 1981, Arroyo and colleagues45 reported that intravenous infusion of saralasin (a short acting ANG-II antagonist) induced a significant reduction in WHVP but it correlated significantly with the decrease in MAP.

Recently, Schneider and colleagues20 evaluated seven days of 25 mg of the orally active ANG-II receptor antagonist losartan in 45 cirrhotics (table 4 ): 30 had severe (HVPG ⩾20 mm Hg) and 15 moderate portal hypertension (HVPG <20 mm Hg). Losartan caused a significant (p<0.001) decrease in HVPG in patients with both severe (−46.8 (15.5)%) and moderate (−44.1 (14.7)%) portal hypertension while no significant reduction was found in controls. Losartan caused a slight, albeit statistically significant, decrease in MAP in both groups of patients (−3.1 (5.0) and −3.5 (4.3) mm Hg, respectively). Liver and renal function did not deteriorate, and the authors concluded that losartan was safe and highly effective in the treatment of portal hypertension. It should be noted that only three of 45 cirrhotics had Child-Pugh grade C (25 Child A, 17 Child B). A further study published in abstract form by Debernardi-Venon and colleagues21supported the beneficial effect of ANG-II receptor antagonists in the reduction in portal pressure (table 4): 11 cirrhotics given 300 mg of irbesartan daily for two months had a significant reduction in HVPG (mean −3.7 mm Hg, −20.7%) without renal side effects. Only one patient developed symptomatic hypotension.

Irbesartan does not require hepatic metabolism to an active metabolite such as losartan and can be administered in normal dosage despite hepatic impairment.

Although the results of the study of Schneider and colleagues20 appear impressive, a number of concerns have been raised.46 Firstly, despite the mean reduction in HVPG being approximately 45% (the most effective reduction in portal pressure described for any portal hypotensive agent), there was no effect on systemic pressure. Secondly, the study was neither randomised nor blinded, and both of these methodological considerations are important to internally validate the results. Thirdly, the same dose (25 mg) of losartan in another study caused a decrease in blood pressure of 10–12 mm Hg in patients with arterial hypertension and 7–8 mm Hg in subjects with normal blood pressure.47Surprisingly, the MAP reduction in Schneider's study was only 3 mm Hg.20 One proposed mechanism for the reduction in portal pressure by ANG-II antagonists is a decrease in portal blood flow caused by reflex splanchnic arterial vasoconstriction which in turn is caused by the reduction in blood pressure. This mode of action in cirrhotics, with an already marked peripheral vasodilatation and resultant decrease in effective blood volume, should induce further deterioration in systemic haemodynamics with deleterious effects on renal function.

However, Schneider et al suggested that the reduction in portal pressure with losartan was mostly due to decreased resistance of the intrahepatic vasculature without major effects on systemic haemodynamics.20 However, a study using isolated perfused cirrhotic rat liver48 showed that only 20–30% of intrahepatic resistance in cirrhosis is amenable to pharmacological manipulation and therefore only half of the average 45% reduction in portal pressure observed with losartan theoretically can be attributed to a decrease in intrahepatic resistance.

The above concerns and putative explanations for the reduction in portal pressure surfaced in the first double blind placebo controlled study using the ANG-II antagonist irbesartan by Schepke and colleagues,49 recently published in abstract form (table4). Thirty cirrhotic patients with portal hypertension were randomised to receive either 150 mg of irbesartan once daily or placebo for one week. Irbesartan caused a significant reduction in HVPG (mean reduction 16.7 (23)%) which was less pronounced compared with losartan.20 However, in 21% of patients, treatment had to be discontinued due to arterial hypotension. Moreover, in those who completed the study there was reduced creatinine clearance.

In a further twist to this series of observations, a recently published study by Bataller and colleagues50 showed that activated human stellate cells (HSCs) express ANG-II receptors (AT1 subtype) and that binding of ANG-II to AT1 receptors induces contraction and proliferation. The effects of ANG-II were undetectable in quiescent stellate cells, suggesting that activated stellate cells may be a target for ANG-II. Previous reports on the presence of ANG-II receptors in activated HSCs provided conflicting results16 ,51 but this could be attributed to differences in the presence or number of ANG-II receptors among species or to differences in methodology. Interestingly, in a previous study,52 systemic infusion of ANG-II resulted in significant cardiac and renal fibrosis but no fibrogenic response was detected in the liver. Further investigation is required to elucidate the in vivo effects of ANG-II in liver injury and hepatic fibrogenesis.53 This effect, rather than that on portal hypertension, may be of considerable clinical relevance; for example, in patients with chronic hepatitis who would have no, or at least minimal, activation of the RAAS. If inhibition of the contractile and proliferative properties of HSCs by losartan and other ANG-II receptor antagonists is confirmed in animal models, this would open up a new therapeutic avenue in the precirrhotic patient.