SSRI

      Child-Pugh A Child-Pugh B Child-Pugh C
Citalopram  Safety no additional risks known no additional risks known no additional risks known
Dose no dose adjustment needed use half of the normal dose use half of the normal dose
Escitalopram  Safety no additional risks known no additional risks known no additional risks known
Dose no dose adjustment needed use half of the normal dose use half of the normal dose
Fluvoxamine Safety no additional risks known no additional risks known no additional risks known
Dose no dose adjustment needed start with half of the normal dose, no adjustment in maximum dosage use half of the normal dose (initial and maximum)
Paroxetine Safety no additional risks known no additional risks known unsafe
Dose no dose adjustment needed use half of the normal dose no dosing advice (unsafe)
Venlafaxine Safety no additional risks known no additional risks known unsafe
Dose use half of the normal dose use half of the normal dose no dosing advice (unsafe)
Fluoxetine  Safety no additional risks known unsafe unsafe
Dose use half of the normal dose no dosing advice (unsafe) no dosing advice (unsafe)
Sertraline Safety no additional risks known unsafe unsafe
Dose use half of the normal dose no dosing advice (unsafe) no dosing advice (unsafe)
Dapoxetine Safety no additional risks known unsafe unsafe
Dose no dose adjustment needed no dosing advice (unsafe) no dosing advice (unsafe)
Explanation
Selective serotonin reuptake inhibitors (SSRIs) are all metabolized by the liver and exposure increases in patients with cirrhosis. Especially sertraline, fluoxetine and dapoxetine show largely altered pharmacokinetics. Adverse events are mainly dose-dependent and can be prevented by using the suggested dose adjustment and slowly increase the dose. Dapoxetine is used for another indication but is also discussed in this assessment report.

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

 

Summary of literature

Considerations

All selective serotonin reuptake inhibitors (SSRIs) are metabolized by the liver and this is affected by liver cirrhosis. Adverse events are mainly dose-dependent and can be prevented by using the suggested dose adjustment and slowly increase the dose. Some case-reports mentioned serotonin syndrome, QT-prolongation and gastro-intestinal bleeding in patients using SSRIs. In almost all cases there was also another drug in use which could influence the risk (drug-drug interaction). In the following text all considerations of the expert panel are discussed per SSRI.

The pharmacokinetics of citalopram have been studied in 9 cirrhotic patients (Child-Pugh (CTP) A/B/C 2/4/3) in one study (evidence level 3) and in a modelling study (level 4). Clearance was reduced by one third and half-life doubled. In a study described in the product information, steady-state plasma levels were doubled in patients with cirrhosis. The modelling study predicts doubling of exposure in CTP B and C cirrhosis, but not yet in CTP A. The product information advises to start with 10 mg per day and use a maximum of 20 mg in these patients because of the risk of QT-prolongation. In the pharmacokinetic study and in another study (evidence level 3; n=2) citalopram was well tolerated. We classify citalopram as ‘no additional risks known’ and advise to use half of the normal dose in patients with CTP B and C cirrhosis.

Escitalopram is studied in one pharmacokinetic study (level 3) among 12 patients with cirrhosis (CTP A/B 6/6) and one modelling study (level 4). Exposure increased by approximately 50% in CTP A and 70% in CTP B compared to healthy controls. Escitalopram was well tolerated with no serious adverse events. In the modelling study the AUC was doubled in patients with CTP B and C cirrhosis. The product information advises to start with 5 mg in patients with cirrhosis and to use a maximum of 10 mg. Escitalopram is classified as ‘no additional risks known’ in all CTP-classes, in accordance with citalopram. For dosing, it is advised to use half of the normal dose in patients with CTP B and C cirrhosis.

In one pharmacokinetic study (level 3) among 12 patients with compensated cirrhosis and in a pharmacokinetic modelling study (level 4) fluoxetine was used. Fluoxetine has nonlinear pharmacokinetics and in the single-dose study clearance was approximately halved in patients compared to controls. Half-life was prolonged threefold to more than 6 days. In the modelling study, high increases in exposure were seen in patients with CTP B and C cirrhosis. The product information advises to use a lower dose or a larger dosing interval. In one retrospective study (evidence level 4) and in the pharmacokinetic study, the safety of fluoxetine was assessed. There were 2 cases of QT-prolongation and a case of serotonin syndrome noted with fluoxetine in combination with other drugs. Because of the predicted increases in exposure in CTP B and C and the long elimination half-life it is better not to use fluoxetine in these patients (unsafe). In CTP A cirrhosis, it can be used (no additional risks known) if the dose is adjusted to half of the normal dose.

Three pharmacokinetic studies (level 4) with fluvoxamine in 53 patients with cirrhosis (CTP A/B/C 20/0/20, rest unknown) were found and one modelling study (level 4). Fluvoxamine has nonlinear pharmacokinetics. During multiple dosing, trough plasma levels were a bit increased in CTP A and increased by 73% in CTP C. Exposure increased with 50% in CTP A and with 170% in CTP C. In the single-dose study, exposure was 60% higher (severity of cirrhosis unknown). In the modelling study exposure was comparable to healthy controls in CTP A, about 50% higher in CTP B and doubled in CTP C. We advise to start with half of the normal initial dose in patients with CTP B or C cirrhosis and in CTP C also use half of the normal maximum dose. In the three pharmacokinetic studies, mild adverse events were noted in the cirrhotics (drowsiness, nausea) which were a bit more frequent in CTP C patients. Fluvoxamine is classified as ‘no additional risks known’, if dose adjustments are followed.

Paroxetine was used by 24 cirrhotic patients (severity unknown) in two pharmacokinetic studies (level 3-4) and there was also a modelling study found. Paroxetine has nonlinear pharmacokinetics. In the multiple dose study, exposure was doubled in patients with cirrhosis which was also predicted in the modelling study in CTP B. In CTP A the predicted exposure was a bit increased (20-50%) and in CTP C it was predicted to increase more (+200-400%). Three studies (level 3-4) were found were the safety of paroxetine in cirrhosis was examined in 8 patients with among them two case-reports of rare adverse events. In one of the pharmacokinetic studies paroxetine was tolerated well, while in the multiple dose study three patients left the study because of severe nausea. Based on the pharmacokinetic studies, we advise a normal dose for CTP A, half of the normal dose for CTP B patients and classify it as ‘no additional risks known’. For CTP C patients, a higher exposure is predicted and there are safer alternatives so it is classified as ‘unsafe’.

Sertraline has a high hepatic extraction ratio. In two pharmacokinetic studies (level 3 and 4) in 19 patients with cirrhosis exposure was threefold increased in one (n=10) and fourfold in the other (n=9 CTP A+B patients) compared to controls. In the modelling study, exposure was predicted to be almost doubled in CTP A and more than doubled in CTP B and C. Two studies looked at the safety of sertraline (level 3-4) in 8 cirrhotic patients and so did two pharmacokinetic studies. Gastro-intestinal adverse events occurred in a few patients (nausea, vomiting and loose stools). In a case-report serotonin syndrome occurred in a patient with CTP B when the dose of sertraline was increased from 50 mg/d to 100 mg per day. The patient used no other serotonergic medication and there was a positive rechallenge with citalopram. For patients with CTP A, it is advised to use half of the normal dose and it is classified as ‘no additional risks known’. For patients with CTP B and C, exposure is fourfold or higher increased and it is classified as ‘unsafe’.

There was no pharmacokinetic study found with venlafaxine. There was a modelling study (level 4) that predicted a 50% increased exposure in CTP A, a 200% increase in CTP B and a 500% increase in CTP C. Two studies are described in the American product label among patients with CTP A+B showing a 50% decreases in clearance, but also decreases of 90% in patients with more severe cirrhosis. In one retrospective study (level 4) one patient had a QT-prolongation when using venlafaxine in combination with trimipramine. In this study also serotonin syndrome was noted in two patients using venlafaxine (one in combination with mirtazapine, and one in combination with tramadol). A serotonin syndrome was also described in a case-report (level 4) where the patient used 150 mg venlafaxine and started metaxalone. Two other case-reports were found where patients with cirrhosis on venlafaxine developed acute respiratory failure in one (in combination with propranolol) and persistent hypoglykemia in the other. In patients with CTP A and B, the advice from the product information is taken to use half of the normal dose and it is classified as ‘no additional risks known’. For the indication panic disorder it is not possible to half the dose so an alternative is preferred. Because the predicted 500% increase in exposure and the 90% decrease in clearance seen in a few patients, it is classified as ‘unsafe’ in patients with CTP C.

No study was found in literature where dapoxetine was used by patients with cirrhosis. The only data is from the product information, which describes a doubling in exposure to dapoxetine and the active metabolite in CTP B and an increase of 223% in exposure in CTP C. For patients with CTP A it is classified as ‘no additional risks known’ based on the limited data. In patients with CTP B and C dapoxetine is classified as ‘unsafe’ based on the increases in exposure, the relative impossibility to adjust the dose and the contra-indication in the product information.

Pharmacokinetic data

  • Absorption: Citalopram, escitalopram and fluoxetine are almost completely absorbed after oral administration (F≥80%) and have a low hepatic extraction ratio. Dapoxetine, fluvoxamine, paroxetine and venlafaxine are subject to first-pass metabolism and have a lower bioavailability (F=40-55%) with an intermediate hepatic extraction ratio. Sertraline has the lowest bioavailability and a high hepatic extraction ratio. The peak plasma levels of both paroxetine and sertraline were almost doubled in patients with cirrhosis compared to healthy controls. Fluoxetine, fluvoxamine and paroxetine exhibit nonlinear pharmacokinetics.
  • Distribution: Dapoxetine, fluoxetine, paroxetine and sertraline are highly bound to plasma proteins (fb>90%), while the other SSRIs have a lower protein binding. All SSRIs have a large volume of distribution. In the studies with citalopram and fluoxetine, the volume of distribution was larger in cirrhotic patients than in healthy controls.
  • Metabolism: All SSRIs are extensively metabolized by liver enzymes, especially CYP450-enzymes. Fluvoxamine and paroxetine have no active metabolites, while the metabolites of the other SSRIs are active, but mostly to a lesser extent than the parent drug. In studies where metabolites were measured, these metabolites frequently had a lower exposure and clearance in patients compared to controls.
  • Elimination: Citalopram has a predominant hepatic excretion, while sertraline is excreted both in urine and in bile. The other SSRIs are mostly excreted in urine.
  • Exposure:       
  • Citalopram: in one single-dose study clearance was reduced by one third in patients with cirrhosis compared to healthy controls and the elimination half-life more than doubled (83 h vs. 37 h). In a pharmacokinetic modelling study exposure to citalopram increased with severity of cirrhosis to an almost doubled AUC in CTP B and a more than doubled AUC in CTP C. The product information states that steady-state plasma levels are doubled in cirrhosis.
  • Escitalopram: a nonlinear mixed effect model was made in one study for escitalopram in patients with CTP A and B cirrhosis. AUC was increased by 52% in CTP A and 69% in CTP B compared to healthy controls. In another pharmacokinetic modelling study exposure to escitalopram increased with severity of cirrhosis to an almost doubled AUC in CTP B and a more than doubled AUC in CTP C.
  • Fluoxetine: in a single-dose study in 12 patients with compensated cirrhosis, clearance was more than halved compared to controls and elimination half-life was three times prolonged (6.6 days vs. 2.2 days). Clearance of the metabolite was also decreased and its elimination half-life doubled. In a pharmacokinetic modelling study exposure to fluoxetine increased with severity of cirrhosis to a doubled AUC in CTP A, a fourfold increase in CTP B and an increase of about tenfold in CTP C cirrhosis.
  • Fluvoxamine: three studies examined pharmacokinetics of fluvoxamine in cirrhotics. In a single-dose study exposure was increased by 50% and elimination half-life prolonged to 25 hours (vs. 16 h in controls). The two other multiple-dose studies included patients with CTP A and C cirrhosis. Exposure was increased by 50% in CTP A and trough levels were 30% higher after 5 days compared to controls. In CTP C patients, exposure was increased by 167% and trough levels were 73% higher. In a pharmacokinetic modelling study exposure to fluvoxamine increased with severity of cirrhosis to a doubled AUC in CTP C cirrhosis.
  • Paroxetine: one single-dose study and one multiple-dose study examined pharmacokinetics of cirrhotic patients. After multiple dosing, exposure was doubled in cirrhotic patients and both peak and trough levels were approximately doubled compared to controls. Half-life was prolonged from 36 h in controls to 83 h in patients. In a pharmacokinetic modelling study exposure to paroxetine increased with severity of cirrhosis to a doubled AUC in CTP B cirrhosis and a threefold increase in CTP C.
  • Sertraline: in a single dose study among patients with CTP A and B cirrhosis, half-life was prolonged in cirrhotics (82 h) versus 25 h in healthy controls. Exposure was 4-fold higher in cirrhotic patients. In a multiple-dose study peak plasma levels and exposure was approximately 3-fold higher in patients with cirrhosis and half-life prolonged to 44 hours. In a pharmacokinetic modelling study exposure to sertraline increased with severity of cirrhosis to a doubled AUC in CTP A and B cirrhosis and a tripled exposure in CTP C.
  • Venlafaxine: in two studies described in the product information, clearance decreased by 50% in patients with CTP A and B cirrhosis. In three patients with more severe cirrhosis, decreases in clearance of up to 90% were reported. In a pharmacokinetic modelling study exposure to venlafaxine increased with severity of cirrhosis to a tripled AUC in CTP B and a six-fold increase in exposure in CTP C cirrhosis.
  • Dapoxetine: in the product information a study was described in which exposure to dapoxetine and the active metabolite was similar to healthy controls in CTP A, increased by 100% in CTP B and increased by 223% in CTP C.

Safety data

In ten studies the safety of SSRIs in patients with cirrhosis was studied, among these were seven case-reports describing individual adverse events. Nine of the pharmacokinetic studies also reported adverse events during the study. In most (pharmacokinetic) studies, mild adverse events were noted. These were somewhat more frequent in cirrhotic patients, where exposure was also often increased. Gastro-intestinal adverse events (nausea and vomiting), dizziness and drowsiness were most often reported.

In the case-reports and case-series, a few adverse events occurred several times. In three case-reports and one retrospective study, serotonin syndrome developed in six cirrhotic patients using a SSRI. In almost all cases the SSRI was used in combination with another serotonergic drug. In one case-report only sertraline (100 mg/d) was used and rechallenge with citalopram (40 mg/d) also led to serotonergic toxicity. Both occurred just after a dose increase. A retrospective study observed 3 cases of QT-prolongation in patients on SSRIs (n=2 fluoxetine, n=1 venlafaxine), in combination with another QT-prolonging drug. Furthermore, in a case-report a patient with CTP A cirrhosis was admitted for severe gastro-intestinal bleeding, probably caused by a combination of paroxetine and aspirin. In the retrospective study, an increase bleeding risk was found in a patient on venlafaxine and low dose aspirin.

Two case-reports found more rare adverse events occurring in patients on venlafaxine: acute respiratory failure and persistent hypoglycaemia. In a last case-report a patient with CTP A cirrhosis started on paroxetine. After a month, worsening of liver function occurred presenting as ascites, icterus and increased liver function tests. After discontinuation of paroxetine, the patient recovered.