- Email: [email protected]
Randomised controlled double-blind trial of the calcium channel antagonist amlodipine in the treatment of acute alcoholic hepatitis
Journal of Hepatology 1998; 28: 194–198 Printed in Denmark ¡ All rights reserved Munksgaard ¡ Copenhagen
Copyright C European Association for the Study of the Liver 1998
Journal of Hepatology ISSN 0168-8278
Randomised controlled double-blind trial of the calcium channel antagonist amlodipine in the treatment of acute alcoholic hepatitis George L. A. Bird1, Andrzej T. Prach2, Alex D. McMahon2, John A. H. Forrest, Peter R. Mills3 and Booth J. Danesh 1
Department of Medicine, Stobhill General Hospital, Glasgow, 2Medicines Monitoring Unit, Department of Clinical Pharmacology, University of Dundee, Dundee, and 3Gastroenterology Unit, Gartnavel General Hospital, Great Western Road, Glasgow, UK
Background/Aims: Calcium channel blockers have a hepatoprotective action in animal models of alcoholinduced liver injury but their effect in alcoholic liver disease in humans has not been previously investigated. We have conducted a randomised, placebo-controlled trial to investigate the possible benefit of the calcium channel blocker amlodipine in terms of 4week survival in hospitalised patients with severe acute alcoholic hepatitis. Methods: Sixty-two patients with acute alcoholic hepatitis were randomised to receive 5–10 mg amlodipine each day for 1 year or an identical capsule containing placebo. In 36 (58%), acute alcoholic hepatitis was confirmed on biopsy and in the remainder on clinical and laboratory criteria. There were no statistically significant differences in clinical characteristics and disease severity in the treated and placebo groups. Results: Of the 32 patients receiving amlodipine, there
were six deaths (19%) in the first 4 weeks compared with seven (23%) of the placebo patients (pΩ0.329). Causes of death were similar in the amlodipine and control groups, with liver failure predominant. Analysis by the Cox proportional hazards model after adjustment for other prognostic factors showed survival was not significantly influenced by active treatment (pΩ0.07). One patient in each group was withdrawn because of the development of hypotension, but this did not recur on reintroduction of the capsules. Conclusions: This study shows that calcium channel blockers are well tolerated with few side effects in advanced alcoholic liver disease, but there is no conclusive evidence from this study that calcium channel blockers are helpful in the treatment of alcoholic hepatitis.
of severe acute alcoholic liver disease within the first 4 weeks of admission to hospital is as high as 65% (1), and many of the short-term survivors require prolonged admission to hospital and medical supervision for recurrent episodes of liver decompensation (2). Management of acute alcoholic hepatitis is currently based upon general supportive care, treatment of the various complications of liver decompensation, nutritional support and withdrawal of alcohol until spontaneous recovery occurs (3). There has recently been interest in corticosteroid therapy as
a specific therapy to improve the short-term survival of patients with a poor prognosis as determined by Maddrey’s Discriminant Function index, but many patients do not meet the criteria for receiving this treatment and the results have not been evaluated beyond the first few weeks of admission to hospital (4–6). A number of studies have suggested that failure of calcium homeostasis arises after exposure to hepatotoxins, and loss of integrity of the mechanisms controlling intracellular calcium compartmentalisation could be an intermediate step in the development of hepatocyte necrosis (7,8). Calcium channel antagonists appear to have a hepatoprotective effect in animal models of toxin-induced liver injury, as measured by both biochemical and histological parameters (9–11). The mechanism of cytoprotection is not understood, but could arise secondary to a decrease in the release of
T
Received 24 October 1996; revised 19 August; accepted 3 September 1997
Correspondence: George L. A. Bird, Department of Gastroenterology, Maidstone Hospital, Hermitage Lane, Maidstone, Kent, ME16 9QQ, UK. Tel: 01622 729000. Fax: 01622 723014.
194
Key words: Alcoholic hepatitis; Alcoholic liver disease; Amlodipine; Calcium channel blocker.
Amlodipine in acute alcoholic hepatitis
intracellular calcium by blocking calcium channels (12), a decrease in hepatocellular ischaemic and reperfusion type injury (13) or vasodilatation with improved oxygen delivery to the centrilobular region of the liver lobule (9). Furthermore, in experimental systems where a high level of circulating endotoxin could be contributing to cell injury through activation of the cytokine cascade (14), addition of a calcium channel blocker has been reported to have a hepatoprotective effect (15). Although the exact mechanism of alcohol-induced liver injury is not understood, a recent observation of particular interest is that chronic ethanol administration up-regulates the number of cell surface calcium channels on a wide range of cells, including Kupffer cells and neuronal tissue, and it has been suggested that this may have specific consequences within the liver (16,17). The possible clinical applications of calcium channel antagonists have led to calls for clinical trials to evaluate their use as hepatoprotective agents (18). The objective of the present study was to assess the hypothesis that the dihydropyridine calcium channel antagonist amlodipine could improve 4-week survival in patients admitted to hospital with acute alcoholic hepatitis. Amlodipine was used because it is a widely available dihydropyridine calcium channel antagonist with a low intersubject bioavailablity and a low incidence of side effects (19).
Materials and Methods Consecutive patients admitted to Stobhill and Gartnavel Hospitals with acute alcoholic hepatitis between January 1992 and October 1994 were considered for inclusion in the study. All patients had histories of alcohol intake greater than 40 g/day in women and 60 g/day in men (20). When possible, the diagnosis was confirmed with a liver biopsy showing characteristic histological features, including liver cell damage, inflammatory cell infiltrate and pericellular fibrosis (21). In those patients in whom percutaneous biopsy was precluded by coagulopathy, the diagnosis of acute alcoholic hepatitis was made on the basis of compatible clinical and laboratory features, including at least two of the following: palpable hepatomegaly, leukocytosis of more than 11¿109 cells/l and typical appearances on liver and spleen isotope scanning, in the absence of any other known or suspected aetiological factor. Maddrey’s Discriminant Function was calculated on the day of randomisation (5). Enrollment in the trial was made within 5 days of admission. Patients were excluded if they had pancreatitis, severe gastrointestinal haemorrhage, malignant disease or were seropositive for HBsAg, anti-HCV or anti-HIV. In addition,
patients with established alcoholic cirrhosis who were admitted primarily for the control of complications arising directly from cirrhosis (bleeding oesophageal or gastric varices, recurrent or chronic hepatic encephalopathy, ascites or oedema) or who had had previous admissions for the complications of portal hypertension were excluded. The trial protocol was approved by the local Research Ethics Committee and patients gave written informed consent in accordance with the Declaration of Helsinki. In patients with hepatic encephalopathy on admission, written informed consent was obtained from the next of kin; where possible, permission was subsequently obtained from the patient. Patients were randomised in a double-blind fashion using a ‘sets of four’ procedure to receive either amlodipine or placebo after stratification on the basis of the severity of illness (bilirubin ∞80 mmol/l, 80–120 mmol/ l, or ±120 mmol/l) to improve the likelihood that the treatment and placebo groups were comparable. The treated patients received amlodipine (Pfizer, Sandwich, UK) 10 mg (two capsules) daily, or 5 mg (one capsule) if the prothrombin time was more than 3 s prolonged. Amlodipine is metabolised in the liver and in patients with cirrhosis the elimination half-life is prolonged (19). However, the manufacturers do not provide specific recommendations with respect to dosage in liver disease, and in preliminary studies we found that the standard 5-mg and 10-mg doses in patients with decompensated liver disease did not give rise to an increased incidence of side effects. The control group received placebo capsules, similar in appearance and taste to the amlodipine. Amlodipine and placebo capsules were administered daily by the nursing staff until discharge from hospital, although study medication was withheld if the systolic blood pressure was below 90 mmHg. In patients who left hospital before completion of 4 weeks of treatment, compliance was evaluated by tablet counts in the returned containers. In subjects with a prolonged prothrombin time on entry to the study, the daily dose of the trial medication was increased to 10 mg (i.e. two capsules) when the prothrombin time improved to only 3 or fewer seconds prolonged. All patients were under the supervision of a dietician and were offered a diet providing a minimum intake of 1800 kcal/day. In those patients unable to maintain this intake, enteral or parenteral nutrition totalling 1800 kcal/day was administered as appropriate. Further medical management was based upon the individual requirements of the patient. On entry to the trial, clinical assessment included hepatic encephalopathy, hepatomegaly, splenomegaly and ascites. Nutritional assessment included mid-arm
195
G. L. A. Bird et al.
circumference (MAC) and triceps skinfold thickness (TSF). Laboratory investigation comprised full blood count, prothrombin time, urea, creatinine, bilirubin, alkaline phosphatase, AST, gamma glutamyl transpeptidase, albumin, gammaglobulins and IgA. Clinical and laboratory reassessment was performed weekly. Patients were counselled with regard to their alcohol intake on enrollment into the trial and, if appropriate, a psychiatric referral was made. Subjects were invariably abstinent from alcohol during hospitalisation and were advised to remain completely abstinent when seen subsequently. Statistical analysis An observation period prior to the commencement of the study indicated an expected 12-month mortality rate of 35% in the placebo group. The study was designed to have a power level of 80% to detect a reduction in mortality from 35% to 19% (i.e. 45% reduction) in the treatment group when compared with the control group. On this basis a minimum of 60 patients were required (22). The statistical analysis was performed according to the intention-to-treat method. Parametric tests (t-test and Fisher’s exact test) and non-parametric tests (Mann-Whitney) were used to compare the characteristics of the patients. Survival percentages were estimated by the Kaplan-Meier method and compared by the log-rank test. Prognostic factors were examined using the same method. To estimate the treatment effect adjusted for prognostic factors, a multidimensional Cox proportional-hazards regression model was used. A p-value of less than or equal to 0.05 was considered statistically significant.
Results Sixty-four patients were eligible for the study and all consented to participate. Two patients were subsequently excluded from the analysis: one in whom a diagnosis of genetic haemochromatosis was made after liver biopsy and a second patient in whom an adenocarcinoma of the large bowel was diagnosed 2 weeks after entering the study. The final analysis therefore included 32 subjects taking amlodipine and 30 receiving placebo. In 36 (58%) of the patients the diagnosis of acute alcoholic hepatitis was confirmed by percutaneous liver biopsy and in 22 of these there was associated cirrhosis. At randomisation, 29 patients had a discriminant function of greater than 32, and 11 of these also had hepatic encephalopathy. Only four patients had spontaneous hepatic encephalopathy with discriminant function values below 32.
196
All subjects were Caucasians, and the two patient groups did not differ significantly with respect to clinical or biochemical characteristics, nor was there any difference with respect to the previous number of episodes of admissions to hospital with alcoholic liver disease (Table 1). Survival Within 4 weeks of entry to the study there were 6 (19%) deaths in the amlodipine group and 7 (23%) in the placebo group, pΩ0.329. Of the 29 patients with a Discriminant Function of 32 or more (15 amlodipine and 14 placebo), there were 5 deaths in each group. The following variables in the proportional hazards model were considered to be potential covariates with prognostic importance for survival: treatment, age, sex, presence of ascites and encephalopathy, prolongation of prothrombin time, peripheral white cell count, serum bilirubin, serum creatinine and albumin. The variables that correlated well with survival during the
TABLE 1 Comparison of clinical and laboratory features on admission in treatment and placebo groups Amlodipine No. of patients No. male/female Age Previous decompensation Spontaneous hepatic encephalopathy (no. patients) Ascites Hepatomegaly Splenomegaly MAC (cm) TSF (mm) Haemoglobin (g/dl) Platelets (¿109/l) Prothrombin time (seconds prolonged) Urea (mmol/l) Creatinine (mmol/l) Serum bilirubin (mmol/l) Serum albumin (g/l) Discriminant function Serum globulins (g/l) IgA (g/l) Alkaline phosphatase (U/l) GGT (U/l) AST (U/l)
Placebo
32 20/12 53.7∫4.3 8 9
30 16/14 51.0∫3.1 12 6
19 20 7* 23.6∫2.1 9.9∫1.6 11.8∫1.7 178∫10.6 5.8∫2.1 4.8∫1.9 90.4∫6.2 140.5∫13.2 29.7∫2.7 31.2∫5.5 34.6∫3.3 8.0∫1.8 428∫20.1 355∫21.0 111∫8.0
15 16 3 21.9∫3.0 9.2∫2.0 11.2∫1.5 155∫8.3 5.8∫2.1 6.0∫2.3 88.2∫6.6 142.8∫13.1 28.2∫2.6 31.7∫5.2 34.9∫2.5 6.5∫1.7 463∫17.2 435∫20.3 108∫9.4
* One patient in the amlodipine group had a past medical history of splenectomy. Results are expressed as means∫SE. MAC: Mid-arm circumference, TSF: Triceps skinfold thickness, IgA: Immunoglobulin A, GGT: Gammaglutamyl transpeptidase, AST: Aspartate aminotransferase. Normal ranges: Haemoglobin: 11.5–16.5 g/dl; platelets: 140– 440¿109/l; urea: 2.5–7.5 mmol/l; creatinine: 60–110 mmol/l; bilirubin: 3–18 mmol/l; albumin: 36–50 g/l; globulins: 24–36 g/l; alkaline phosphatase: 30–120 U/l; GGT: 5–50 U/l; AST: 10–35 U/l.
Amlodipine in acute alcoholic hepatitis TABLE 2 Causes of death in the 13 patients who died during the study
Liver failure (including hepatorenal failure) Gastrointestinal bleeding Respiratory tract infection Vasculitis
Amlodipine (nΩ6)
Placebo (nΩ7)
4
5
1 0 1
1 1 0
study were prolongation of prothrombin time (pΩ 0.002), presence of encephalopathy (pΩ0.007) and serum bilirubin (pΩ0.011). (Correlations of survival with albumin and creatinine were pΩ0.052 and pΩ 0.075, respectively.) The two treatment groups were very similar with regard to cause of death (Table 2). Complications of treatment The trial medication was withdrawn for 2 days in one patient receiving amlodipine, on the grounds that it was causing low blood pressure. Another patient, receiving placebo, developed postural hypotension and the trial medication was withdrawn for a period of 6 days. One patient receiving amlodipine had the trial treatment withdrawn after developing a severe leucocytoclastic vasculitis (13 days after entering the study). The vasculitis progressed and death from vasculitic necrosis of the small bowel occurred 4 days later, but at post mortem the pathologist reported this was unconnected to the amlodipine treatment.
Discussion Although the use of calcium channel blockers as hepatoprotective agents has been suggested previously (18), we believe this is the first controlled clinical trial to evaluate their role as a specific therapy for severe liver disease. Our results show that there is no benefit in short-term survival in acute alcoholic hepatitis for patients taking amlodipine when compared with those on placebo. Three of the criteria of severity (hepatic encephalopathy, prolongation of prothrombin time and serum bilirubin) were found to be significantly predictive of survival when analysed by a multidimensional model, and in this respect our patients were similar to other recent series of patients with severe acute alcoholic hepatitis (23,24). It is unlikely that the negative results of this study are related to a lack of power as the expected mortality rate was similar to that actually observed during the study. Further-
more, the treatment and placebo groups were equally matched for severity of disease and the causes of death were similar. As a potential therapy for acute alcoholic hepatitis, amlodipine was an attractive option on account of its ease of administration and low incidence of side effects. Furthermore, there were no concerns about its use in immunocompromised patients or the possibility that it could promote upper gastrointestinal bleeding. Despite the lack of effect upon survival, this study confirms that amlodipine was well tolerated and not associated with an increased incidence of side effects in advanced liver disease. There were several shortcomings of the study: these included the presence of a confirmatory liver biopsy in only 58% of subjects, raising the possibility that a proportion of the patients could have failed to satisfy the criteria for alcoholic hepatitis. In two other recent studies of therapy for alcoholic hepatitis, the diagnosis has been confirmed on every patient with transjugular biopsies (6,23), but this is not a facility routinely available in our centre and has a higher mortality than percutaneous biopsy (25). However, only one patient (3%) was excluded from the final analysis when the biopsy showed the original diagnosis of alcoholic hepatitis was incorrect, fewer than the exclusions in the other recent trials (8% (6) and 12% (23), respectively). Entry into the trial was not precluded by the absence of a liver biopsy and thus there was very little delay between admission to hospital and initiation of trial medication (mean 2.5 days), allowing the use of therapy before the patient was too ill to respond. High-dose corticosteroids remain the only proven specific therapy for severe acute alcoholic hepatitis, although they should be reserved for patients with a poor short-term prognosis and it is not known whether long-term survival is improved (6). There is still no generally accepted therapy suitable for the bulk of patients with a discriminant function value of less than 32, and in those for whom corticosteroids are contraindicated. Our study suggests that although calcium channel blockers are safe to administer in advanced liver disease and are well tolerated, they do not have a place in the treatment of alcoholic hepatitis.
Acknowledgements We thank Dr George Smith and Professor Roderick MacSween for interpreting the liver biopsies, and Mr Carl Fenelon for administering the randomisation procedure and dispensing the study capsules. Financial assistance was kindly provided by the Greater Glasgow Health Board.
197
G. L. A. Bird et al.
References 1. Hardison WG, Lee FI. Prognosis in the acute liver disease of the alcoholic patient. N Engl J Med 1966; 275: 61–6. 2. Bird G, Lau JYN, Koskinas K, Wicks C, Williams R. Insulin and glucagon infusion in acute alcoholic hepatitis: a prospective randomised controlled trial. Hepatology 1991; 14: 1097– 101. 3. Sherlock S. Diseases of the Liver and Biliary System, 8th ed. Oxford: Blackwell; 1989. 4. Imperiale TF, McCullough AJ. Do corticosteroids reduce mortality from alcoholic hepatitis? A meta-analysis of the randomised trials. Ann Intern Med 1990; 113: 299–307. 5. Carithers RL Jnr, Herlong HF, Diehl AM, Shaw EW, Combes B, Fallon HJ, et al. Methylprednisolone therapy in patients with severe alcoholic hepatitis: a randomised multicentre trial. Ann Intern Med 1989; 110: 685–90. 6. Ramond M-J, Poynard T, Rueff B, Mathurin P, Theodore C, Chaput J-C, et al. A randomised trial of prednisolone in patients with severe alcoholic hepatitis. N Engl J Med 1992; 326: 507–12. 7. Mauger J-P, Claret M. Calcium channels in hepatocytes. J Hepatol 1988; 7: 278–82. 8. Thomas CE, Reed DJ. Current status of calcium in hepatocellular injury. Hepatology 1989; 10: 375– 84. 9. Landon EJ, Naukam RJ, Sastry BVR. Effects of calcium channel blocking agents on calcium and centrilobular necrosis in the liver of rats treated with hepatotoxic agents. Biochem Pharmacol 1986; 35: 697–705. 10. Liang D, Thurman RG. Protective effects of the calcium antagonists diltiazem and TA3090 against hepatic injury due to hypoxia. Biochem Pharmacol 1992; 44: 2207–11. 11. Iimuro Y, Ikejima K, Rose ML, Bradford BU, Thurman RG. Nimodipine, a dihydropyridine-type calcium channel blocker, prevents alcoholic hepatitis caused by chronic intragastric ethanol exposure in the rat. Hepatology 1996; 24: 391–7. 12. Buss WC. Savage DD, Stepanek J, Little SA, McGuffee LJ. Effect of calcium channel antagonists on calcium channel uptake and release by isolated rat cardiac mitochondria. Eur J Pharmacol 1988; 152: 247–53.
198
13. Stein HJ, Oosthuizen MM, Hinder RA. Effect of verapamil on hepatic ischaemic-reperfusion injury in normal and glutathione-depleted rats [abstract]. Gastroenterology 1989; 96: 663. 14. Bird GLA, MacSween RNM. Pathogenesis of alcoholic liver disease: immune mechanisms. In: Hall P, editor. Alcoholic Liver Disease. London: Edward Arnold; 1995. 15. Sayeed MM, Maitra SR. Effect of diltiazem on altered cellular calcium regulation during endotoxic shock [review]. Am J Physiol 1987; 253: 549–54. 16. Hijioka T, Goto M, Lemasters JJ, Thurman RG. Effect of short-term ethanol treatment on voltage-dependent calcium channels in Kupffer cells. Hepatology 1993; 18: 400–5. 17. Littleton J, Little H. Current concepts of ethanol dependence. Addiction 1994; 89: 1397–412. 18. Deakin CD, Fagan E, Williams R. Cytoprotective effects of calcium channel blockers. J Hepatol 1991; 12: 251–5. 19. Istin: Product Monograph. Sandwich, UK: Pfizer; 1990. 20. Royal College of Physicians Working Party on Alcohol. Medical Consequences of Alcohol Abuse: a Great and Growing Evil. London: RCP; 1987. p. 1–30. 21. Baptista A, Bianchi L, de Groote R, Desmet VJ, Gedigk P, Korb G, et al. Alcoholic liver disease: morphological manifestations. Lancet 1981; i: 707–11. 22. Young MJ, Bresnitz, EA, Strom BL. Sample size nomograms for interpreting negative clinical studies. Ann Intern Med 1983; 99: 248–51. 23. Trinchet J-C, Balkau B, Poupon RE, Heintzmann F, Callard P, Gotheil C, et al. Treatment of severe alcoholic hepatitis by infusion of insulin and glucagon: a multicentre sequential trial. Hepatology 1992; 15: 76–81. 24. Bennett WG, Bird GLA, Lau JYN. Prognostic models in acute alcoholic hepatitis (AAH) [abstract]. Hepatology 1994; 20: 320. 25. Vlavianos P, Bird G, Portman B, Westaby D, Williams R. Transjugular liver biopsy: use in a selected high risk population. Eur J Gastroenterol Hepatol 1991; 3: 469–72.
Copyright C European Association for the Study of the Liver 1998
Journal of Hepatology ISSN 0168-8278
Randomised controlled double-blind trial of the calcium channel antagonist amlodipine in the treatment of acute alcoholic hepatitis George L. A. Bird1, Andrzej T. Prach2, Alex D. McMahon2, John A. H. Forrest, Peter R. Mills3 and Booth J. Danesh 1
Department of Medicine, Stobhill General Hospital, Glasgow, 2Medicines Monitoring Unit, Department of Clinical Pharmacology, University of Dundee, Dundee, and 3Gastroenterology Unit, Gartnavel General Hospital, Great Western Road, Glasgow, UK
Background/Aims: Calcium channel blockers have a hepatoprotective action in animal models of alcoholinduced liver injury but their effect in alcoholic liver disease in humans has not been previously investigated. We have conducted a randomised, placebo-controlled trial to investigate the possible benefit of the calcium channel blocker amlodipine in terms of 4week survival in hospitalised patients with severe acute alcoholic hepatitis. Methods: Sixty-two patients with acute alcoholic hepatitis were randomised to receive 5–10 mg amlodipine each day for 1 year or an identical capsule containing placebo. In 36 (58%), acute alcoholic hepatitis was confirmed on biopsy and in the remainder on clinical and laboratory criteria. There were no statistically significant differences in clinical characteristics and disease severity in the treated and placebo groups. Results: Of the 32 patients receiving amlodipine, there
were six deaths (19%) in the first 4 weeks compared with seven (23%) of the placebo patients (pΩ0.329). Causes of death were similar in the amlodipine and control groups, with liver failure predominant. Analysis by the Cox proportional hazards model after adjustment for other prognostic factors showed survival was not significantly influenced by active treatment (pΩ0.07). One patient in each group was withdrawn because of the development of hypotension, but this did not recur on reintroduction of the capsules. Conclusions: This study shows that calcium channel blockers are well tolerated with few side effects in advanced alcoholic liver disease, but there is no conclusive evidence from this study that calcium channel blockers are helpful in the treatment of alcoholic hepatitis.
of severe acute alcoholic liver disease within the first 4 weeks of admission to hospital is as high as 65% (1), and many of the short-term survivors require prolonged admission to hospital and medical supervision for recurrent episodes of liver decompensation (2). Management of acute alcoholic hepatitis is currently based upon general supportive care, treatment of the various complications of liver decompensation, nutritional support and withdrawal of alcohol until spontaneous recovery occurs (3). There has recently been interest in corticosteroid therapy as
a specific therapy to improve the short-term survival of patients with a poor prognosis as determined by Maddrey’s Discriminant Function index, but many patients do not meet the criteria for receiving this treatment and the results have not been evaluated beyond the first few weeks of admission to hospital (4–6). A number of studies have suggested that failure of calcium homeostasis arises after exposure to hepatotoxins, and loss of integrity of the mechanisms controlling intracellular calcium compartmentalisation could be an intermediate step in the development of hepatocyte necrosis (7,8). Calcium channel antagonists appear to have a hepatoprotective effect in animal models of toxin-induced liver injury, as measured by both biochemical and histological parameters (9–11). The mechanism of cytoprotection is not understood, but could arise secondary to a decrease in the release of
T
Received 24 October 1996; revised 19 August; accepted 3 September 1997
Correspondence: George L. A. Bird, Department of Gastroenterology, Maidstone Hospital, Hermitage Lane, Maidstone, Kent, ME16 9QQ, UK. Tel: 01622 729000. Fax: 01622 723014.
194
Key words: Alcoholic hepatitis; Alcoholic liver disease; Amlodipine; Calcium channel blocker.
Amlodipine in acute alcoholic hepatitis
intracellular calcium by blocking calcium channels (12), a decrease in hepatocellular ischaemic and reperfusion type injury (13) or vasodilatation with improved oxygen delivery to the centrilobular region of the liver lobule (9). Furthermore, in experimental systems where a high level of circulating endotoxin could be contributing to cell injury through activation of the cytokine cascade (14), addition of a calcium channel blocker has been reported to have a hepatoprotective effect (15). Although the exact mechanism of alcohol-induced liver injury is not understood, a recent observation of particular interest is that chronic ethanol administration up-regulates the number of cell surface calcium channels on a wide range of cells, including Kupffer cells and neuronal tissue, and it has been suggested that this may have specific consequences within the liver (16,17). The possible clinical applications of calcium channel antagonists have led to calls for clinical trials to evaluate their use as hepatoprotective agents (18). The objective of the present study was to assess the hypothesis that the dihydropyridine calcium channel antagonist amlodipine could improve 4-week survival in patients admitted to hospital with acute alcoholic hepatitis. Amlodipine was used because it is a widely available dihydropyridine calcium channel antagonist with a low intersubject bioavailablity and a low incidence of side effects (19).
Materials and Methods Consecutive patients admitted to Stobhill and Gartnavel Hospitals with acute alcoholic hepatitis between January 1992 and October 1994 were considered for inclusion in the study. All patients had histories of alcohol intake greater than 40 g/day in women and 60 g/day in men (20). When possible, the diagnosis was confirmed with a liver biopsy showing characteristic histological features, including liver cell damage, inflammatory cell infiltrate and pericellular fibrosis (21). In those patients in whom percutaneous biopsy was precluded by coagulopathy, the diagnosis of acute alcoholic hepatitis was made on the basis of compatible clinical and laboratory features, including at least two of the following: palpable hepatomegaly, leukocytosis of more than 11¿109 cells/l and typical appearances on liver and spleen isotope scanning, in the absence of any other known or suspected aetiological factor. Maddrey’s Discriminant Function was calculated on the day of randomisation (5). Enrollment in the trial was made within 5 days of admission. Patients were excluded if they had pancreatitis, severe gastrointestinal haemorrhage, malignant disease or were seropositive for HBsAg, anti-HCV or anti-HIV. In addition,
patients with established alcoholic cirrhosis who were admitted primarily for the control of complications arising directly from cirrhosis (bleeding oesophageal or gastric varices, recurrent or chronic hepatic encephalopathy, ascites or oedema) or who had had previous admissions for the complications of portal hypertension were excluded. The trial protocol was approved by the local Research Ethics Committee and patients gave written informed consent in accordance with the Declaration of Helsinki. In patients with hepatic encephalopathy on admission, written informed consent was obtained from the next of kin; where possible, permission was subsequently obtained from the patient. Patients were randomised in a double-blind fashion using a ‘sets of four’ procedure to receive either amlodipine or placebo after stratification on the basis of the severity of illness (bilirubin ∞80 mmol/l, 80–120 mmol/ l, or ±120 mmol/l) to improve the likelihood that the treatment and placebo groups were comparable. The treated patients received amlodipine (Pfizer, Sandwich, UK) 10 mg (two capsules) daily, or 5 mg (one capsule) if the prothrombin time was more than 3 s prolonged. Amlodipine is metabolised in the liver and in patients with cirrhosis the elimination half-life is prolonged (19). However, the manufacturers do not provide specific recommendations with respect to dosage in liver disease, and in preliminary studies we found that the standard 5-mg and 10-mg doses in patients with decompensated liver disease did not give rise to an increased incidence of side effects. The control group received placebo capsules, similar in appearance and taste to the amlodipine. Amlodipine and placebo capsules were administered daily by the nursing staff until discharge from hospital, although study medication was withheld if the systolic blood pressure was below 90 mmHg. In patients who left hospital before completion of 4 weeks of treatment, compliance was evaluated by tablet counts in the returned containers. In subjects with a prolonged prothrombin time on entry to the study, the daily dose of the trial medication was increased to 10 mg (i.e. two capsules) when the prothrombin time improved to only 3 or fewer seconds prolonged. All patients were under the supervision of a dietician and were offered a diet providing a minimum intake of 1800 kcal/day. In those patients unable to maintain this intake, enteral or parenteral nutrition totalling 1800 kcal/day was administered as appropriate. Further medical management was based upon the individual requirements of the patient. On entry to the trial, clinical assessment included hepatic encephalopathy, hepatomegaly, splenomegaly and ascites. Nutritional assessment included mid-arm
195
G. L. A. Bird et al.
circumference (MAC) and triceps skinfold thickness (TSF). Laboratory investigation comprised full blood count, prothrombin time, urea, creatinine, bilirubin, alkaline phosphatase, AST, gamma glutamyl transpeptidase, albumin, gammaglobulins and IgA. Clinical and laboratory reassessment was performed weekly. Patients were counselled with regard to their alcohol intake on enrollment into the trial and, if appropriate, a psychiatric referral was made. Subjects were invariably abstinent from alcohol during hospitalisation and were advised to remain completely abstinent when seen subsequently. Statistical analysis An observation period prior to the commencement of the study indicated an expected 12-month mortality rate of 35% in the placebo group. The study was designed to have a power level of 80% to detect a reduction in mortality from 35% to 19% (i.e. 45% reduction) in the treatment group when compared with the control group. On this basis a minimum of 60 patients were required (22). The statistical analysis was performed according to the intention-to-treat method. Parametric tests (t-test and Fisher’s exact test) and non-parametric tests (Mann-Whitney) were used to compare the characteristics of the patients. Survival percentages were estimated by the Kaplan-Meier method and compared by the log-rank test. Prognostic factors were examined using the same method. To estimate the treatment effect adjusted for prognostic factors, a multidimensional Cox proportional-hazards regression model was used. A p-value of less than or equal to 0.05 was considered statistically significant.
Results Sixty-four patients were eligible for the study and all consented to participate. Two patients were subsequently excluded from the analysis: one in whom a diagnosis of genetic haemochromatosis was made after liver biopsy and a second patient in whom an adenocarcinoma of the large bowel was diagnosed 2 weeks after entering the study. The final analysis therefore included 32 subjects taking amlodipine and 30 receiving placebo. In 36 (58%) of the patients the diagnosis of acute alcoholic hepatitis was confirmed by percutaneous liver biopsy and in 22 of these there was associated cirrhosis. At randomisation, 29 patients had a discriminant function of greater than 32, and 11 of these also had hepatic encephalopathy. Only four patients had spontaneous hepatic encephalopathy with discriminant function values below 32.
196
All subjects were Caucasians, and the two patient groups did not differ significantly with respect to clinical or biochemical characteristics, nor was there any difference with respect to the previous number of episodes of admissions to hospital with alcoholic liver disease (Table 1). Survival Within 4 weeks of entry to the study there were 6 (19%) deaths in the amlodipine group and 7 (23%) in the placebo group, pΩ0.329. Of the 29 patients with a Discriminant Function of 32 or more (15 amlodipine and 14 placebo), there were 5 deaths in each group. The following variables in the proportional hazards model were considered to be potential covariates with prognostic importance for survival: treatment, age, sex, presence of ascites and encephalopathy, prolongation of prothrombin time, peripheral white cell count, serum bilirubin, serum creatinine and albumin. The variables that correlated well with survival during the
TABLE 1 Comparison of clinical and laboratory features on admission in treatment and placebo groups Amlodipine No. of patients No. male/female Age Previous decompensation Spontaneous hepatic encephalopathy (no. patients) Ascites Hepatomegaly Splenomegaly MAC (cm) TSF (mm) Haemoglobin (g/dl) Platelets (¿109/l) Prothrombin time (seconds prolonged) Urea (mmol/l) Creatinine (mmol/l) Serum bilirubin (mmol/l) Serum albumin (g/l) Discriminant function Serum globulins (g/l) IgA (g/l) Alkaline phosphatase (U/l) GGT (U/l) AST (U/l)
Placebo
32 20/12 53.7∫4.3 8 9
30 16/14 51.0∫3.1 12 6
19 20 7* 23.6∫2.1 9.9∫1.6 11.8∫1.7 178∫10.6 5.8∫2.1 4.8∫1.9 90.4∫6.2 140.5∫13.2 29.7∫2.7 31.2∫5.5 34.6∫3.3 8.0∫1.8 428∫20.1 355∫21.0 111∫8.0
15 16 3 21.9∫3.0 9.2∫2.0 11.2∫1.5 155∫8.3 5.8∫2.1 6.0∫2.3 88.2∫6.6 142.8∫13.1 28.2∫2.6 31.7∫5.2 34.9∫2.5 6.5∫1.7 463∫17.2 435∫20.3 108∫9.4
* One patient in the amlodipine group had a past medical history of splenectomy. Results are expressed as means∫SE. MAC: Mid-arm circumference, TSF: Triceps skinfold thickness, IgA: Immunoglobulin A, GGT: Gammaglutamyl transpeptidase, AST: Aspartate aminotransferase. Normal ranges: Haemoglobin: 11.5–16.5 g/dl; platelets: 140– 440¿109/l; urea: 2.5–7.5 mmol/l; creatinine: 60–110 mmol/l; bilirubin: 3–18 mmol/l; albumin: 36–50 g/l; globulins: 24–36 g/l; alkaline phosphatase: 30–120 U/l; GGT: 5–50 U/l; AST: 10–35 U/l.
Amlodipine in acute alcoholic hepatitis TABLE 2 Causes of death in the 13 patients who died during the study
Liver failure (including hepatorenal failure) Gastrointestinal bleeding Respiratory tract infection Vasculitis
Amlodipine (nΩ6)
Placebo (nΩ7)
4
5
1 0 1
1 1 0
study were prolongation of prothrombin time (pΩ 0.002), presence of encephalopathy (pΩ0.007) and serum bilirubin (pΩ0.011). (Correlations of survival with albumin and creatinine were pΩ0.052 and pΩ 0.075, respectively.) The two treatment groups were very similar with regard to cause of death (Table 2). Complications of treatment The trial medication was withdrawn for 2 days in one patient receiving amlodipine, on the grounds that it was causing low blood pressure. Another patient, receiving placebo, developed postural hypotension and the trial medication was withdrawn for a period of 6 days. One patient receiving amlodipine had the trial treatment withdrawn after developing a severe leucocytoclastic vasculitis (13 days after entering the study). The vasculitis progressed and death from vasculitic necrosis of the small bowel occurred 4 days later, but at post mortem the pathologist reported this was unconnected to the amlodipine treatment.
Discussion Although the use of calcium channel blockers as hepatoprotective agents has been suggested previously (18), we believe this is the first controlled clinical trial to evaluate their role as a specific therapy for severe liver disease. Our results show that there is no benefit in short-term survival in acute alcoholic hepatitis for patients taking amlodipine when compared with those on placebo. Three of the criteria of severity (hepatic encephalopathy, prolongation of prothrombin time and serum bilirubin) were found to be significantly predictive of survival when analysed by a multidimensional model, and in this respect our patients were similar to other recent series of patients with severe acute alcoholic hepatitis (23,24). It is unlikely that the negative results of this study are related to a lack of power as the expected mortality rate was similar to that actually observed during the study. Further-
more, the treatment and placebo groups were equally matched for severity of disease and the causes of death were similar. As a potential therapy for acute alcoholic hepatitis, amlodipine was an attractive option on account of its ease of administration and low incidence of side effects. Furthermore, there were no concerns about its use in immunocompromised patients or the possibility that it could promote upper gastrointestinal bleeding. Despite the lack of effect upon survival, this study confirms that amlodipine was well tolerated and not associated with an increased incidence of side effects in advanced liver disease. There were several shortcomings of the study: these included the presence of a confirmatory liver biopsy in only 58% of subjects, raising the possibility that a proportion of the patients could have failed to satisfy the criteria for alcoholic hepatitis. In two other recent studies of therapy for alcoholic hepatitis, the diagnosis has been confirmed on every patient with transjugular biopsies (6,23), but this is not a facility routinely available in our centre and has a higher mortality than percutaneous biopsy (25). However, only one patient (3%) was excluded from the final analysis when the biopsy showed the original diagnosis of alcoholic hepatitis was incorrect, fewer than the exclusions in the other recent trials (8% (6) and 12% (23), respectively). Entry into the trial was not precluded by the absence of a liver biopsy and thus there was very little delay between admission to hospital and initiation of trial medication (mean 2.5 days), allowing the use of therapy before the patient was too ill to respond. High-dose corticosteroids remain the only proven specific therapy for severe acute alcoholic hepatitis, although they should be reserved for patients with a poor short-term prognosis and it is not known whether long-term survival is improved (6). There is still no generally accepted therapy suitable for the bulk of patients with a discriminant function value of less than 32, and in those for whom corticosteroids are contraindicated. Our study suggests that although calcium channel blockers are safe to administer in advanced liver disease and are well tolerated, they do not have a place in the treatment of alcoholic hepatitis.
Acknowledgements We thank Dr George Smith and Professor Roderick MacSween for interpreting the liver biopsies, and Mr Carl Fenelon for administering the randomisation procedure and dispensing the study capsules. Financial assistance was kindly provided by the Greater Glasgow Health Board.
197
G. L. A. Bird et al.
References 1. Hardison WG, Lee FI. Prognosis in the acute liver disease of the alcoholic patient. N Engl J Med 1966; 275: 61–6. 2. Bird G, Lau JYN, Koskinas K, Wicks C, Williams R. Insulin and glucagon infusion in acute alcoholic hepatitis: a prospective randomised controlled trial. Hepatology 1991; 14: 1097– 101. 3. Sherlock S. Diseases of the Liver and Biliary System, 8th ed. Oxford: Blackwell; 1989. 4. Imperiale TF, McCullough AJ. Do corticosteroids reduce mortality from alcoholic hepatitis? A meta-analysis of the randomised trials. Ann Intern Med 1990; 113: 299–307. 5. Carithers RL Jnr, Herlong HF, Diehl AM, Shaw EW, Combes B, Fallon HJ, et al. Methylprednisolone therapy in patients with severe alcoholic hepatitis: a randomised multicentre trial. Ann Intern Med 1989; 110: 685–90. 6. Ramond M-J, Poynard T, Rueff B, Mathurin P, Theodore C, Chaput J-C, et al. A randomised trial of prednisolone in patients with severe alcoholic hepatitis. N Engl J Med 1992; 326: 507–12. 7. Mauger J-P, Claret M. Calcium channels in hepatocytes. J Hepatol 1988; 7: 278–82. 8. Thomas CE, Reed DJ. Current status of calcium in hepatocellular injury. Hepatology 1989; 10: 375– 84. 9. Landon EJ, Naukam RJ, Sastry BVR. Effects of calcium channel blocking agents on calcium and centrilobular necrosis in the liver of rats treated with hepatotoxic agents. Biochem Pharmacol 1986; 35: 697–705. 10. Liang D, Thurman RG. Protective effects of the calcium antagonists diltiazem and TA3090 against hepatic injury due to hypoxia. Biochem Pharmacol 1992; 44: 2207–11. 11. Iimuro Y, Ikejima K, Rose ML, Bradford BU, Thurman RG. Nimodipine, a dihydropyridine-type calcium channel blocker, prevents alcoholic hepatitis caused by chronic intragastric ethanol exposure in the rat. Hepatology 1996; 24: 391–7. 12. Buss WC. Savage DD, Stepanek J, Little SA, McGuffee LJ. Effect of calcium channel antagonists on calcium channel uptake and release by isolated rat cardiac mitochondria. Eur J Pharmacol 1988; 152: 247–53.
198
13. Stein HJ, Oosthuizen MM, Hinder RA. Effect of verapamil on hepatic ischaemic-reperfusion injury in normal and glutathione-depleted rats [abstract]. Gastroenterology 1989; 96: 663. 14. Bird GLA, MacSween RNM. Pathogenesis of alcoholic liver disease: immune mechanisms. In: Hall P, editor. Alcoholic Liver Disease. London: Edward Arnold; 1995. 15. Sayeed MM, Maitra SR. Effect of diltiazem on altered cellular calcium regulation during endotoxic shock [review]. Am J Physiol 1987; 253: 549–54. 16. Hijioka T, Goto M, Lemasters JJ, Thurman RG. Effect of short-term ethanol treatment on voltage-dependent calcium channels in Kupffer cells. Hepatology 1993; 18: 400–5. 17. Littleton J, Little H. Current concepts of ethanol dependence. Addiction 1994; 89: 1397–412. 18. Deakin CD, Fagan E, Williams R. Cytoprotective effects of calcium channel blockers. J Hepatol 1991; 12: 251–5. 19. Istin: Product Monograph. Sandwich, UK: Pfizer; 1990. 20. Royal College of Physicians Working Party on Alcohol. Medical Consequences of Alcohol Abuse: a Great and Growing Evil. London: RCP; 1987. p. 1–30. 21. Baptista A, Bianchi L, de Groote R, Desmet VJ, Gedigk P, Korb G, et al. Alcoholic liver disease: morphological manifestations. Lancet 1981; i: 707–11. 22. Young MJ, Bresnitz, EA, Strom BL. Sample size nomograms for interpreting negative clinical studies. Ann Intern Med 1983; 99: 248–51. 23. Trinchet J-C, Balkau B, Poupon RE, Heintzmann F, Callard P, Gotheil C, et al. Treatment of severe alcoholic hepatitis by infusion of insulin and glucagon: a multicentre sequential trial. Hepatology 1992; 15: 76–81. 24. Bennett WG, Bird GLA, Lau JYN. Prognostic models in acute alcoholic hepatitis (AAH) [abstract]. Hepatology 1994; 20: 320. 25. Vlavianos P, Bird G, Portman B, Westaby D, Williams R. Transjugular liver biopsy: use in a selected high risk population. Eur J Gastroenterol Hepatol 1991; 3: 469–72.