Medicines used for alcohol use disorder

    Child-Pugh A Child-Pugh B Child-Pugh C
Acamprosate Safety no additional risks known no additional risks known unknown
Dose dose adjustment is not necessary dose adjustment is not necessary no dosing advice possible
Disulfiram Safety additional risks known unsafe unsafe
Dose start with half of the normal dose no dosing advice (unsafe) no dosing advice (unsafe)
Nalmefene Safety no additional risks known additional risks known unsafe
Dose dose adjustment is not necessary dose adjustment is not necessary no dosing advice (unsafe)
Naltrexone (PO) Safety unsafe unsafe unsafe
Dose no dosing advice (unsafe) no dosing advice (unsafe) no dosing advice (unsafe)

Most medicines used in patients with an alcohol use disorder are largely cleared by the liver. Pharmacokinetics also frequently change resulting in increased exposure to the medicines (i.e. disulfiram, nalmefene and naltrexone). Therefore a risk/benefit assessment is very important since continued alcohol use is also harmful. Acamprosate is not eliminated by the liver and pharmacokinetics are not altered in Child-Pugh A and B cirrhosis. It has not been studied in Child-Pugh C. A recent review describes that theoretically acamprosate could favour the development of hepatic encephalopathy because it antagonizes the glutamate receptor. As there are no data available on pharmacokinetics or safety, we classify it as ‘unknown’ in Child-Pugh C patients. If it is used, the patient should be carefully monitored. Disulfiram is extensively metabolised in the liver and pharmacokinetics are altered in cirrhosis possible resulting in a different effect, yet findings are inconsistent. Disulfiram could also cause (dose-dependent) hepatotoxicity; it can best be avoided in Child-Pugh A and should not be used in Child-Pugh B and C. Nalmefene and naltrexone are both highly cleared by the liver, resulting in large pharmacokinetic changes in cirrhosis for naltrexone when administered orally (≥ fivefold increased exposure). Naltrexone should therefore not be used in these patients as treatment for alcohol dependence. Pharmacokinetic changes are less profound with nalmefene, but exposure is threefold increased in Child-Pugh B. It is not known what the clinical effect of increased exposure is in these patients. In Child-Pugh C, even larger changes in exposure are expected and it is contraindicated in the product information, therefore it is classified as unsafe. Baclofen is also off-label used for this indication. Studies suggest that it could be an alternative if no other drug can be used. However, it is not an innocent drug (it could cause severe side effects) and we have not yet assessed the safety of baclofen, so we cannot provide advice on baclofen.  

Information about the safety classification and the recommended actions can be found here.

Summary of literature


The pharmacokinetics of acamprosate have been studied in 18 patients with cirrhosis (CTP A and B) in two studies (evidence level 4) after single and multiple dosing. These studies showed no alterations in pharmacokinetic parameters compared to healthy controls. This was also expected because of the marginal role of the liver in the pharmacokinetics of acamprosate. The safety was studied in a randomized controlled trial (level 2; n=12 CTP A/B) and it was well tolerated after a single dose. In CTP A and B acamprosate is classified as ‘no additional risks known’ and no dose adjustments are needed. In CTP C, the product information does not recommend its use. Given the pharmacokinetic characteristics of acamprosate, we expect a marginal role of the liver in the pharmacokinetics in CTP C patients. However, as described in a recent review, theoretically acamprosate could favour the development of hepatic encephalopathy because it antagonizes the glutamate receptor. As there are no data available on pharmacokinetics or safety, we classify it as ‘unknown’ in these patients.

Disulfiram is largely metabolized in the liver and has an active metabolite, (possibly) more potent than disulfiram itself. The pharmacokinetics were studied in two (old) articles (evidence level 4; n=12 severity unknown) after single and three day dosing. In the single dose study no alterations in absorption and excretion were seen in the patients with liver disease compared to the controls. In the multiple dose study, exposure to the active DDC metabolites was threefold increased after two and three days dosing. Three other studies (evidence level 3; n=41, severity unknown) examined the pharmacodynamic effect of disulfiram on acetaldehyde plasma levels and two of these also studied the disulfiram-alcohol-reaction. Inconsistent results were found. In one study there was an apparent lack of disulfiram-effect in the patients with cirrhosis compared to alcoholics without cirrhosis, while the other study found a disulfiram-effect more often in patients with cirrhosis (particularly in patients with more severe liver disease). This study also found higher acetaldehyde exposure, while the former found lower exposure compared to alcoholic controls without cirrhosis. The third study demonstrated that disulfiram increased acetaldehyde plasma levels in all subjects (endogenous; without ethanol ingestion), but the disulfiram-effect in the patients with cirrhosis was less striking. This study measured transaminase levels and noted that in 3 of 13 cirrhotic patients, a significant elevation (>1.5 times upper limit of normal) occurred, but no liver deterioration. In the product information hepatotoxicity is mentioned as possible adverse effect with the note that it (mostly) occurred as a consequence of a high dose or overdose. There were two case reports where disulfiram treatment led to acute liver failure in two cirrhotic patients (one used the normal dose of 500 mg/d and the other a supratherapeutic dose of 1500 mg/d). In two other case reports disulfiram induced neurotoxicity.

To conclude, alterations in pharmacokinetics of disulfiram are likely to occur in patients with cirrhosis, especially in advanced cirrhosis (CTP B and C). It is contraindicated in cirrhosis with ascites according to the product information. It is unsure if the pharmacokinetic changes will result in a more pronounced effect or a diminished effect of disulfiram because of its complicated metabolism and active metabolite(s). Studies show contradicting results. Because of the (possible) dose-dependent hepatotoxicity, the pharmacokinetic alterations, the unpredictable effects and the vulnerability of the patients, it is classified as ‘unsafe’ in CTP B and C cirrhosis. In CTP A cirrhosis, the pharmacokinetic alterations are expected to be less profound with less vulnerable patients. It is therefore classified as ‘additional risks known’ and it is advised to start low with half of the normal dose.

Nalmefene has a high hepatic extraction ratio. It has been studied in patients with cirrhosis in the context of market authorisation. These data are only available in the product information (number of subjects unknown). Exposure increased by 1.5-fold in CTP A patients compared to healthy controls and 2.9-fold in CTP B. The maximum plasma levels were 1.7 times higher in CTP B patients, while elimination half-life was comparable to controls. In one safety study (level 3; n=11 severity unknown) patients experienced a considerable withdrawal reaction after the first doses. In CTP A it is classified as ‘no additional risks known’. It is not known if subjects tolerated the threefold increase in exposure in CTP B well, therefore it is classified as ‘additional risks known’. We comply with the advice of the product information that no dose adjustment in CTP A is needed since it is often only used in an acute course as single dose (“dose as needed”). In CTP B, a high exposure have been found, but the tablets may not be divided and it is not known if a dose adjustment to a tablet every other day as needed, is effective. Therefore, again we do not recommend dose adjustment. Exposure is expected to increase even more in CTP C, therefore the advice from the product information is adopted to avoid the use (‘unsafe’).

Two studies assessed the pharmacokinetics of naltrexone in cirrhotic patients (evidence level 3, n=23; 9 CTP A, 8 CTP B and 6 CTP C). Naltrexone has a high hepatic extraction ratio. A study with intramuscular naltrexone noted no change in pharmacokinetics in patients with CTP A and B cirrhosis compared to healthy controls. Intramuscular naltrexone has not been classified, because this drug is not marketed in the Netherlands. After oral administration however, another article noted large differences in exposure (5-fold increase in CTP A and B, 10-fold increase in CTP C). This is probably caused by an increase in bioavailability through altered hepatic blood flow, seen by the even larger increases in peak plasma levels. In reviews, it is described that naltrexone hepatotoxicity is a concern when higher doses are used. Because of the large increases in exposure, it is classified as ‘unsafe’ in all stages of cirrhosis. Naltrexone is also sometimes used as treatment for pruritus, we did not specifically assess the safety of naltrexone for this indication, but advice to avoid the use in patients with cirrhosis.    

Pharmacokinetic data

  • Absorption: Only a small part of acamprosate is slowly absorbed from the gastrointestinal tract (F=11%). Peak plasma levels were not different in Child A and B cirrhosis compared to healthy controls. Disulfiram is well absorbed (F=80-90%). Nalmefene and naltrexone both have a low bioavailability (F=5-40%), caused by a high first-pass effect. Both have a high hepatic extraction ratio. The peak plasma levels of naltrexone after oral administration were 7- to 9-fold higher in patients with cirrhosis compared to healthy controls. Plasma levels after intramuscular injection of naltrexone were not different from controls. Plasma levels of nalmefene were 1.7 times higher in CTP B patients.
  • Distribution: Acamprosate does not bind to plasma proteins. Disulfiram, nalmefene and naltrexone all have a low protein binding (≤50%). Acamprosate has a small volume of distribution (70 L), this volume is unknown for disulfiram. Nalmefene and naltrexone both have a large volume of distribution
  • Metabolism: Acamprosate is not metabolized. Disulfiram is extensively metabolized in the liver into diethyldithiocarbamate (DDC). This is partly glucuronidated into the active metabolite DDC methylesther. In a study among cirrhotic patients, this metabolism was profoundly altered. Nalmefene is mainly metabolised by UGT2B7 and a small proportion is converted to nalmefene 3-O-sulphate by sulphate by sulphation. This active metabolite has a potency comparable to that of nalmefene, but is only present in a concentration less than 10% of nalmefene. Naltrexone is mainly metabolized by hydroxylation in the liver to 6-ß-naltrexole, which attributes to the pharmacological effect just like naltrexone itself. Plasma levels of this metabolite are decreased in cirrhosis when administered per os.
  • Elimination: Acamprosate is excreted by the kidneys and the elimination half-life is 33 hours because of the slow absorption. The metabolites of disulfiram are also renally excreted. Disulfiram has an elimination half-life of 7.5 hours and its active metabolite 22 hours. Nalmefene has a terminal half-life of 12.5 hours and is excreted in urine. This elimination half-life was not clinical relevantly altered in CTP B cirrhosis according to the product information. The half-life of naltrexone is 4 hours and of its main metabolite 13 hours and these are mainly renally excreted.
  • Exposure:        
  • Acamprosate: the product information and the study of Saivon et al. describe studies in CTP A and B patients in which the pharmacokinetics are not altered compared to healthy controls. Exposure was comparable between groups.
  • Disulfiram: plasma levels of disulfiram metabolites (carbon disulphide, DDC and disulphides) increased in a study within 3 days of treatment to about +50% for carbon disulphides and +200-250% for DDC and disulphides. The authors suggest to use low doses in cirrhotics. Another study found no difference in the absorption, urine excretion, breath excretion and total recovery of radioactive disulfiram between patients with and without liver disease (cirrhosis in 6/10). Three other studies looked at the pharmacodynamic effect of disulfiram in patients with cirrhosis. Two studied the disulfiram-ethanol reaction in cirrhotic alcoholics and non-cirrhotic alcoholics and found contrary results. The first found an apparent lack of disulfiram effect, in particular in the clinical findings and also lower acetaldehyde exposure compared to the controls. In the second study, the cirrhotic group had more frequently a (minimal) reaction to ethanol lasting longer and significantly higher acetaldehyde levels. The group with cirrhosis was further divided in less severe and severe cirrhosis. The last group also suffered more frequently from a reaction and longer and had higher acetaldehyde levels. The difference between these two contrary studies also was the dose of disulfiram (400 mg/d vs. 250 mg/d, resp.) and the ethanol dose (0.2 mg/kg and 64 mg/kg resp.) In a third study, disulfiram increased acetaldehyde levels in red blood cells and in plasma in both alcoholics with and without cirrhosis. But the effect on plasma acetaldehyde was less striking in cirrhotics.
  • Nalmefene: the EPAR describes a study in patients with CTP A and B in which exposure was significant higher in compared to healthy controls (1.5- and 2.9-fold, respectively).
  • Naltrexone: after intramuscular injection, exposure of naltrexone and 6-ß-naltrexole was not changed in CTP A and B cirrhosis compared to healthy controls. After oral administration exposure to naltrexone increased 5-fold in CTP A and B cirrhosis and 8-fold in CTP C cirrhosis compared to controls. The FDA label describes increases of 5-fold in compensated and 10-fold in decompensated cirrhosis.


Safety data

  • Acamprosate: in one randomized controlled trial a single dose of 666 mg was well tolerated by 12 cirrhotic patients (CTP A and B). No effect on the development of subclinical hepatic encephalopathy was noted. One case report described a rash in a patient with alcoholic cirrhosis after acamprosate treatment for 10 days. A review warned for the theoretical risk of hepatic encephalopathy with acamprosate use because it antagonizes the glutamate receptor.
  • Disulfiram: in an open-label trial disulfiram 250 mg was given daily to 13 cirrhotic alcoholic patients and 11 non cirrhotic alcoholics. In 3 patients with cirrhosis clinically significant elevations in transaminases (>1.5 x ULN) occurred but none developed any clinical sign of liver deterioration. Several case-reports were found where disulfiram was probably the causative agent of liver toxicity. In two patients hepatotoxicity presenting as acute liver failure occurred; in one of these patients with compensated cirrhosis the outcome was unknown, in the other patient a liver transplantation was needed. Both used 500 mg disulfiram daily. Two other case-reports describe neurotoxicity with psychiatric symptoms occurring in a patient on 400 mg disulfiram daily and a patient on 1500 mg daily. In the EASL guideline it is stated that disulfiram should be avoided in patients with severe alcoholic liver disease because of possible hepatotoxicity. Two other recent reviews also advise to avoid disulfiram in chronic liver disease (advanced alcoholic liver disease).
  • Nalmefene: in one open-label trial 11 patients with cirrhosis (severity unknown) were treated with (very) high doses of nalmefene (5 mg twice daily to start and increased to 20-40 mg thrice daily). All experienced a considerable (withdrawal) reaction starting within an hour of the first doses consisting of cerebral effects (sometimes with hallucinations), anorexia, nausea, colicky abdominal pain and constipation. 
  • Naltrexone: no safety data available, in the reviews the potential for hepatotoxicity is mentioned. Goh et al. describe that “hepatotoxicity has not emerged as a clinical problem with naltrexone in standard doses, although hepatotoxicity is a concern, in certain circumstances, when higher doses are used. Patients with advanced liver disease are particularly vulnerable to naltrexone-induced hepatotoxicity, so its use in this situation would need to be monitored carefully.”