Isosorbide-5-mononitrate is a long acting organic nitrate. The mononitrate, either as native drug or formed from the denitration of isosorbide dinitrate in the lever,44 is the active component, and undergoes minimal first class pass metabolism unlike isosorbide dinitrate,45 thus assisting in appropriate dosing for patients with liver disease and portal shunting. Isosorbide mononitrate reaches peak concentrations within an hour of oral dosing, and has a half-life of approximately 5 h. Only 1–2% of an orally administered dose is excreted unchanged in the urine, with the remainder being eliminated as inactive metabolites. Its pharmacokinetic properties are unchanged in the elderly and in patients with renal failure, or liver cirrhosis.46 It is therefore preferred to isosorbide dinitrate in such patients. To date it is the only nitrate to be used in large randomized controlled trials for preventing variceal bleeding.
The molecular mechanism of action of nitrates is uncertain. It is thought the vasodilatory actions may be a result of enhanced production of intrahepatic nitric oxide or cyclic-GMP.47
There have been a number of haemodynamic studies using isosorbide-5-mononitrate in patients with portal hypertension (Table 3). All the studies with the exception of one, 49 which included predominantly Childs A patients, demonstrated significant reductions in the HVPG. This appears to have been achieved by a fall in the HVWP. Three studies looking at the chronic effects show that the effect of reduced portal pressure is sustained.51–53 Indeed the portal hypotensive effect seemed to be amplified after rechallenge following chronic dosing, confirming lack of tolerance.
Table 3. The effect of isosorbide-5-mononitrate on portal and systemic haemodynamics
Pronounced effects on other parameters may help to explain the mechanism underlying the fall in HVPG, which is comparable to that of propranolol. Early studies48 noted that the hepatic blood flow fell acutely, and this along with an increase in systemic vascular resistance index (SVRI) suggested that a baroreceptor-mediated splancnic vasoconstriction may be responsible for the fall in portal pressure rather than portal venous dilation. However recent work55 demonstrated a significant fall in the portal pressure gradient without affecting the portal blood flow in patients with a transjugular intrahepatic portosystemic stent shunt (TIPSS). It was clear, therefore, that in the study group of patients reflex baroreceptor-mediated vasoconstriction of the splancnic bed could not be the case and that any vasoconstrictive effect to account for the rise in SVRI was limited to the periphery. The observed findings were attributed to reduced intrahepatic vascular resistance rather than a reduction in the liver blood flow (which would be undesirable). Chronic administration resulted in no change or even an increase in the hepatic blood flow50, 51 and may reflect the buffer response of hepatic artery blood flow to a decrease in portal flow.56
Isosorbide-5-mononitrate also reduces the cardiac preload and hence the cardiac output, at least acutely. Significant correlation between the reduction in portal pressure gradient and cardiac output suggests that this may partly be responsible for the reduced portal blood flow observed in this study.48 In all cases the mean arterial pressure fell acutely. It is interesting to note that chronic administration does not appear to have a significant effect on cardiac output or mean arterial pressure.
Azygous blood flow has already been demonstrated as a useful indicator of variceal blood flow in patients with cirrhosis.57 Azygous blood flow responds in a variable fashion to isosorbide-5-mononitrate. Jones and colleagues53 demonstrated no significant change in the azygous blood flow in response to varying doses of nitrates, both acutely and following chronic dosing. However, a relationship was noted between baseline azygous blood flow and the response to nitrates, with those patients with a high azygous blood flow responding by reducing their flow and vice versa. The dose did not seem to influence the azygous blood flow.
Nitrate tolerance is clearly documented in cardiovascular medicine.58, 59 However, of the studies looking at the chronic effects of isosorbide-5-mononitrate therapy only one has reported partial tolerance in five out of 11 patients,51 with others reporting no tolerance.52, 53 The exact mechanism behind why patients with cirrhosis do not develop nitrate tolerance is unknown. It has been suggested that patients with cirrhosis may not be able to develop compensatory mechanisms that are necessary to bring about nitrate tolerance.53
An important observation with nitrate monotherapy has been the deleterious effect on renal function.60 Activation of the renin-angiotensin system was felt to be a major factor. In particular, patients with ascites suffered from a reduced glomerular filtration rate, sodium excretion and renal plasma flow. It is interesting that combination therapy with other portal hypotensive agents abolished these undesirable renal effects. The combination therapies will be covered later.
Drugs acting on the seretonin S2 receptors
It has been shown in an experimental model that portal hypertensive animals are more sensitive to the vasoconstrictor effects of seretonin on mesenteric veins, and that administration of ketanserin, a 5HT2 receptor blocker with α adrenergic antagonist activity, resulted in significant reductions in the portal venous inflow and portal pressure.73 The reduction in portal pressure caused by ketanserin was due mainly to a decrease in portal venous inflow secondary to a decreased cardiac output, which was only seen in portal hypertensive rats. This would be consistent with venous dilatation and pooling of blood in the portal venous system secondary to 5HT2 receptor blockade. These finding led to human studies investigating the effect of 5HT2 receptor blockade on portal pressure.
Early trials in cirrhotic patients demonstrated a significant reduction of 23% in the HVPG following ketanserin administration, which was accompanied by reductions in the azygous blood flow and mean arterial pressure with the hepatic blood flow remaining unaffected.74 Subsequent studies corroborated these findings.75, 76 The chronic administration of ketanserin was associated with a sustained drop in the portal pressure of 14.6%, a reduction in the cardiac index, and a drop in the mean arterial pressure.76 This study also demonstrated that 50% of patients developed reversible portosystemic encepahlopathy. Hypotension probably results from α receptor blockade.
Combination treatments have also been studied. Ketanserin in combination with propranolol, both administrated intravenously, has been shown to reduce the HVPG in patients who did not initially respond to propranolol.77 Ritanserin, a more selective serotonin S2 blocker, was combined with propranolol in a study investigating the haemodynamic effects of the chronic dosing of these agents.78 An initial reduction in the portal pressure was noted, but this effect was not sustained during follow up.
These agents have not been studied in randomized controlled trials for the prevention of variceal bleeding or for the treatment of variceal bleeding. The high incidence of encepahlopathy observed with monotherapy, and the potential for systemic hypotension may limit their clinical use. Combination therapy with non-selective β-blockers seems more promising.
Drugs affecting plasma volume
The expansion of plasma volume leading to increased cardiac index is believed to play a major role in sustaining portal hypertension.79 Thus diuretics or a low sodium diet may in theory help to reduce portal pressure. Early studies suggested that spironolactone had the potential to be as potent a portal hypotensive agent as propranolol.80 A significant reduction in the portal pressure of between 10 and 15% was shown to be accompanied by reductions in plasma volume, cardiac output, mean arterial pressure and azygous blood flow.81–83 Hepatic blood flow was unaffected. Although there was no significant correlation between the plasma volume and HVPG, a significant inverse relationship between post treatment serum aldosterone levels and the HVPG was noted, thus confirming the mode of action of aldosterone.81 The reductions in the HVPG following spironolactone administration were not affected by dietary sodium content suggesting that a low sodium diet alone is not sufficient to reduce the portal pressure.82, 83 In clinical practice the use of spironolactone, like propranolol, may be limited by its side effects, particularly painful gynaecomastia which was present in 55% of male patients in one series.84 There are no published clinical trials assessing the efficacy of spironolactone in preventing variceal bleeding, although an ongoing study is comparing nadolol monotherapy with nadolol and spironolactone combination therapy in the primary prevention of variceal bleeding.85 Initial results show no differences in the bleeding rate in the two groups.
Combination therapy was first used for the treatment of portal hypertension using nitrates and vasopression.86, 87 Enhancement of the portal hypotensive effect was observed. Studies using propranolol, as already discussed, have revealed that 30% of patients failed to reduce portal pressure.13 This observation and the fact that nitrate monotherapy consistently reduced portal pressure led to studies to investigate the effect of combined nitrate and β-blocker therapy (Table 4), a combination that was first investigated in vitro by Kroeger and Groszmann.88 In general there is an enhanced portal hypotensive effect of the combination therapy using isosorbide-5-mononitrate leading to a further 13–16% fall in the HVPG. This effect is particularly striking in those patients who did not respond to a β-blocker alone.16 The mechanism proposed has been that of a decrease in the outflow resistance.91 It is of note that 1 year after combination therapy with nadolol there did not seem to be any additional effect of the nitrate either in the hepatic or systemic haemodynamics93 despite there being a sustained effect after 3 months of therapy. This may be partly explained by a study by Bendsten and associates who demonstrated that placebo treatment had an equal effect to propranolol after 1 year suggesting that portal hypertension improves spontaneously in some patients.18 Although studies looking at the short term effect of nitrates failed to demonstrate tolerance,52, 53 nitrate tolerance could still be an explanation for the lack of effect due to a longer follow-up period in this study.93
Table 4. The effect of combination therapy on portal and systemic haemodynamics
The hepatic blood flow and liver metabolic activity are unaffected, but the azygous blood flow decreased in most cases with the effect being less pronounced with longer duration of treatment. Mean arterial blood pressure and cardiac output both fell, again with the effect most pronounced following acute administration.
Renal function and ascites formation have been the focus of some of the studies,89, 90, 92 particularly as isosorbide-5-mononitrate has been associated with deteriorating renal function.60 The findings suggest that the combination of isosorbide-5-mononitrate with either propranolol or nadolol had no detrimental effect on renal function in patients with or without ascites, despite significant effects on hepatic haemodynamics. These encouraging findings have led to a number of clinical trials using such combinations.
Other combination therapies have also been studied, but none of them have been studied in large clinical trials. One combination therapy worth noting is that of prazosin and propranolol.92 This study showed a greater portal hypotensive effect of prazosin and propranolol than isosorbide-5-mononitrate and prazosin, with no effect on hepatic or renal function. Undesirable systemic effects unfortunately offset the impressive hepatic haemodynamic results with more patients experiencing symptomatic hypotension in the prazosin arm of the study, thus limiting its use in clinical practice.