- Top of page
- PATIENTS AND METHODS
Propranolol, a non-selective beta-adrenoceptor antagonist, is the treatment of choice for primary prophylaxis of variceal bleeding. It also prevents rebleeding in patients with portal hypertension.1, 2 The blockade of both β1- and β2-adrenoceptors leads to a decrease in portal pressure by reducing cardiac output and by splanchnic vasoconstriction due to unopposed α-adrenergic activity.2–6 Prospective studies have shown that a hepatic venous pressure gradient reduction of at least 20% or a drop below 12 mmHg should be the goal in order to achieve a clinically significant degree of reduction in bleeding risk.6–8 Unfortunately, the pharmacological effects of propranolol on the portal haemodynamics vary greatly. At least one-third of patients do not respond adequately.9–11 Numerous studies indicate that neither clinical nor systemic haemodynamic parameters (e.g. severity of liver disease, change in heart rate) can satisfactorily indicate the portal pressure response.4, 6, 9, 10, 12, 13 Thus, the measurement of portal haemodynamic parameters is necessary to assess the portal haemodynamic response to propranolol. Pulsed Doppler ultrasound has been established as a reliable and reproducible method for non-invasively assessing the acute effect of propranolol on portal haemodynamics in patients with liver cirrhosis.12
It is remarkable that total plasma propranolol levels do not correlate with the portal haemodynamic response.10 Propranolol is administered as a racemic mixture. However, the R-enantiomer is haemodynamically inactive.15–17 Thus, a considerable proportion of the total plasma propranolol has no haemodynamic effect and may obscure the relationship between plasma concentrations and haemodynamic response because a linear correlation between the plasma concentrations of the active isomer (S-propranolol) and total propranolol has not been established. The hepatic metabolism of the two isomers is stereospecific in healthy subjects; R-propranolol is metabolized faster than S-propranolol, resulting in higher S-propranolol plasma levels.18 It has been shown for carvedilol that such a metabolic stereoselectivity can be greatly altered in patients with liver cirrhosis as compared to healthy subjects.19 These results have been substantiated in rats that underwent portacaval shunting.20 Certain studies have used R-propranolol as a model drug for investigating reduced hepatic drug clearance in patients with cirrhosis.21, 22 To date, however, no data exist on the stereoselectivity of the propranolol metabolism in patients with chronic liver disease. The bioavailability and the metabolism of the racemate are known to be considerably altered in liver cirrhosis,14 e.g. due to changes in the hepatic glucuronidation and oxidation, changes in protein binding or due to portosystemic shunting. If such changes affect the stereoselectivity of the propranolol metabolism, this could explain both the inter-individual differences in portal haemodynamic response to propranolol and the lack of correlation between total propranolol plasma levels and the effect on portal haemodynamics.
We therefore investigated the plasma concentrations of the haemodynamically active (S-propranolol) and inactive (R-propranolol) propranolol stereoisomers and correlated these levels with the changes in portal blood velocity in patients with liver cirrhosis and portal hypertension.
- Top of page
- PATIENTS AND METHODS
This is the first study to investigate the portal haemodynamic effects of propranolol with respect to the drug’s racemic nature. The potential clinical importance of this issue is illustrated by the fact that R-propranolol, in contrast to S-propranolol, does not affect the smooth musculature and that both isomers exhibit different pharmacokinetics.16, 18
In the present study we found that an oral load with 40 mg propranolol induces a decrease in non-invasively measured PBV which reached a maximum at 2–3 h. Seven of the patients (35%) did not respond adequately (i.e. less than a 20% reduction in PBV). The individual PBV responses were not related to either the S- or R-propranolol plasma levels or to the individual S/R-propranolol ratios.
Propranolol became established in the treatment of portal hypertension more than 15 years ago.24 Its effect on portal haemodynamics is mainly caused by a reduction in cardiac output and splanchnic blood flow. There are numerous clinical and haemodynamic studies which clearly show that propranolol reduces portal hypertension and prevents related upper intestinal bleeding.2–10
In all these studies—even in those that assessed propranolol plasma concentrations10—the fact that propranolol is given as a racemate, with only the S-isomer being haemodynamically active, was not adequately appraised. Following oral administration, propranolol has a high first-pass elimination.25 In healthy subjects this first-pass elimination is more pronounced for the R-isomer, resulting in higher S-propranolol plasma concentrations.18 In this study we could demonstrate that this pharmacokinetic stereoselectivity is preserved in patients with liver cirrhosis, as indicated by the approximately two-fold higher AUC0–4 h for S-propranolol as compared to that for the R-isomer. These results contrast with a study on carvedilol, demonstrating that in patients with liver cirrhosis—when compared to normal subjects—the differences in bioavailability between the two carvedilol enantiomers disappear.19 With respect to propranolol, 2 h after its oral administration, the S-propranolol plasma levels were approximately 30% higher than the R-propranolol concentrations in healthy subjects.18 In contrast, our series of liver cirrhotic patients had S-enantiomer plasma concentrations that on average exceeded the R-propranolol levels by 75% 2 h after oral intake.
The metabolism of both total propranolol14 and its inactive isomer R-propranolol21, 22 has been shown to be severely altered in liver cirrhosis. However, until now no study has assessed the active isomer, S-propranolol, in liver cirrhotics. In our study we found a wide range of plasma S-propranolol concentrations among the patients but a close correlation between these S-propranolol plasma levels and the total propranolol concentrations. This indicates that despite the well-known stereoselectivity of the propranolol metabolism, pharmacokinetic alterations in liver cirrhosis obviously affect both stereoisomers to a similar extent. Thus, total plasma propranolol concentrations, which have commonly been assessed in previous studies,10, 26, 27 can serve as surrogate markers for circulating S-propranolol.
Changes in the propranolol metabolism of liver cirrhotic patients, as described by Arthur et al.,14 are characterized by a prolonged terminal half-life in patients with severely impaired liver function and serum albumin levels below 3 g/dL. In our study population there were 80% Child A and B patients and only 10% had a serum albumin below 3 g/dL. This may explain the lack of any correlation between the severity of liver disease (expressed as Child-Pugh score28) and the S- or R-propranolol AUC0–4 h. The degree of pharmacokinetic stereoselectivity can be expressed as an S/R AUC ratio.29 In our study population, this ratio correlated neither with the severity of the liver disease nor with the portal haemodynamic response.
Among our study population, 35% were propranolol non-responders with respect to the portal haemodynamic effect; a value that agrees well with the literature.9–11 Although numerous studies have addressed the issue of which clinical characteristic predicts the portal haemodynamic response to propranolol, no reliable parameter has so far been identified.6, 9, 10 Similarly, we were not able to identify a clinical parameter (e.g. severity of liver disease, presence of ascites, aetiology of cirrhosis, reduction in heart rate or blood pressure) that allowed us to differentiate between responders and non-responders. Furthermore, none of these parameters correlated with the PBV reduction. As reported by others,12, 30 we observed a trend for more non-responders to have had a previous variceal bleeding (43%) than responders (31%). However, this trend did not reach statistical significance. Furthermore, non-responders had a higher mean arterial blood pressure at baseline.
As reported for total propranolol plasma concentrations,10 our study also showed no correlation between the plasma concentrations of the haemodynamically active S-enantiomer (expressed as AUC0–4 h) and the portal haemodynamic response (PBV reduction). This is conceivable, because there was a linear correlation between S-propranolol and total plasma propranolol concentrations. It might well be that the binding properties of S-propranolol to the receptor or post-receptor mechanisms are the important factors for the haemodynamic effect in liver cirrhosis and not its plasma levels.
In conclusion, we were able to demonstrate that there is a stereoselectivity of propranolol pharmacokinetics in patients with liver cirrhosis and that S-propranolol remains the predominant stereoisomer in plasma. In liver cirrhosis the differences in plasma concentrations between S- and R-propranolol are even more pronounced than previously reported for healthy subjects. Interestingly, among our patients with liver cirrhosis, the S- and R-propranolol clearance was affected to a similar extent. Thus, it is improbable that differences in the splanchnic haemodynamic response to propranolol can be explained by an altered stereoselectivity of the diseased liver.