Amiodarone-Induced Cirrhosis of Liver: What Predicts Mortality?

Introduction. Amiodarone has been used for more than 5 decades for the treatment of various tachyarrhythmias and previously for the treatment of refractory angina. There are multiple well-established side effects of amiodarone. However, amiodarone-induced cirrhosis (AIC) of liver is an underrecognized complication. Methods. A systematic search of Medline from January 1970 to November 2012 by using the following terms, amiodarone and cirrhosis, identified 37 reported cases of which 30 were used in this analysis. Patients were divided into 2 subsets, survivors versus nonsurvivors, at 5 months. Results. Aspartate aminotransferase was significantly lower (P = 0.03) in patients who survived at 5-months (mean 103.33 IU/L) compared to nonsurvivors (mean 216.88 IU/L). There was no statistical difference in the levels of prothrombin time, total bilirubin, alanine aminotransferase, alkaline phosphatase, gamma-glutamyl transpeptidase, cumulative dose, and latency period between the two groups. The prevalence of DM, HTN, HLD, CAD, and CHF was similar in the two groups. None of the above-mentioned variables could be identified as a predictor of survival at 5 months. Conclusion. AIC carries a mortality risk of 60% at 5 months once the diagnosis is established. Further prospective studies are needed to identify predictors of AIC and of mortality or survival in cases of AIC.


Introduction
Amiodarone has been used since the 1960s for the management of various tachyarrhythmias and in the past for refractory angina. There are multiple reported and well-established side effects of amiodarone therapy such as effects on the thyroid, skin, lungs, nerves, and cornea.
The effect of amiodarone on the liver resulting in hepatotoxicity is a recognized complication of amiodarone, but this hepatotoxicity leading to cirrhosis of the liver is unfortunately an underrecognized side effect. Little has been written on amiodarone-induced cirrhosis (AIC) of the liver due to its rarity [1][2][3][4]. The purpose of this paper is to review what we know so far about AIC of the liver.

Selection of Studies.
A systematic search of Medline from January 1970 to November 2012 by using the following terms, amiodarone and cirrhosis, was performed; 37 reported cases were identified (Table 1) of which 30 were used in this analysis. We also searched the reference lists of all reported cases to identify citations that were not identified during the initial search. Data that were extracted for each patient included age, gender, latency period in years, whether ≥200 mg/day amiodarone dosage was used (it is thought that low-dose amiodarone has lesser side effects), cumulative dose, presence or absence of hypertension (HTN), diabetes (DM), hyperlipidemia (HLD), coronary artery disease (CAD), congestive heart failure (CHF), values of bilirubin, aspartate transaminase (AST), alkaline phosphatase (ALK P), albumin, alanine transaminase (ALT), and prothrombin time. The outcomes in terms of survival or mortality at 5 months were used. This specific timeline was selected because more than 50% of patients in these case reports died within 5 months after the diagnosis was established. Not all of the studies we found had all these data; however, they were included if the outcome was provided. In some 2 ISRN Cardiology  cases, the outcome was not available, authors were contacted, and the outcome was determined. Those cases, for which outcome could not be determined, were excluded from the final analysis. Of seven cases that were excluded from the final analysis, one case had indeterminate cause of cirrhosis and for the rest of cases, outcome could not be established. Some authors provided cumulative doses for amiodarone. If total cumulative dose was not given, a formula (cumulative dose in grams = 365 * no. of years * total daily dose/1000) was used to estimate the final cumulative doses, with an assumption that the patient had been 100% compliant with the medicine. Prothrombin time for some cases was given as percentage activity rather than in seconds. We could not find any convertor for percentage activity to seconds (which is a normally reported unit). For cases in which prothrombin time was reported as a percentage prothrombin activity, values were not used in this analysis.

Statistical Analysis.
Of 37 cases analyzed, outcome was available only for 31 cases. One case had indeterminate cause of cirrhosis and was excluded, and 30 cases were used for our analysis.
Survival at 5 months determined the patient subset and common pathophysiologic factors like DM, HTN, HLD, CAD, and CHF; lab values comprising of prothrombin time, total bilirubin, AST, ALT, ALK P, and GGT were compared between these two subsets. Logistic regression was used for comparing continuous independent variables, and chi-square test was used for comparing categorical variables. When the expected frequency was less than five, Fisher's exact test was used for categorical variables. Cox proportional hazards ratios were employed in determining prognosis. All statistical tests were two tailed with significance set at 95% level ( < 0.05). STATA 11 IC software was used for statistical analysis.

Results
AST was significantly lower ( = 0.03) in patients who survived at 5 months (mean 103.33 IU/L) compared to nonsurvivors (mean 216.88 IU/L). AST levels overall were raised in both groups and ranged between 64 IU/L and 734 IU/L (Tables 2 and 3). There was no significant difference in the levels of prothrombin time, total bilirubin, ALT, ALK P, and GGT between the two groups. The prevalence of DM, HTN, HLD, CAD, and CHF was similar in these two groups. Mean cumulative dose in cases of AIC was 280 g with a median latency period of 2.92 years and was statistically nonsignificant.
The risk of dying at 5 months was marginally higher in patients with high aspartate aminotransferase. Hazard ratio for death is 1.003 (95% CI ranging between 1.001 and 1.006). This finding was statistically significant ( = 0.03) ( Table 3). Prothrombin time, total bilirubin, ALT, ALK P, GGT, and coexistence of DM, HTN, HLD, CAD, and CHF did not predict survival. Results are summarized in Tables 2 and 3.
Median age at the time of diagnosis of AIC was 68 years for all cases, 69.5 among survivors, and 67.5 among nonsurvivors. Most cases of AIC were observed among females though most who died were males.

Pathogenesis of AIC.
Amiodarone is a lipophilic agent [5] and tends to accumulate in lipid-laden organelles such as the liver. Amiodarone causes liver damage by different pathogenic mechanisms, one of which may include phospholipidosis. There are two different mechanisms by which amiodarone causes phospholipidosis.
(I) Amiodarone and its metabolites (N-desmethylamiodarone) accumulate in lysosomes of hepatocytes, bile duct epithelium, and kupffer cells and leads to inhibition of phospholipase A1 and A2 [35,36], which thereby inhibits removal of lysosomal lipids and leads to phospholipidosis. (II) Amiodarone binds to phospholipids in lysosomes and forms a nondigestible complex [37,38], which leads to phospholipidosis.
The exact mechanism of phospholipidosis-induced liver damage is unclear. Phospholipidosis has been reported to occur within two months of starting amiodarone therapy [6,7]. Phospholipidosis occurs in a much larger percentage of patients receiving amiodarone [5] than actual hepatocellular damage (1% to 3%), which suggests that phospholipidosis may or may not have a role in the pathogenesis of AIC. In fact, development of phospholipidosis is considered a marker for accumulation of amiodarone [4] rather than a marker of hepatotoxicity.
Leakage [3,4,39,40] of proteolytic enzymes from abnormal lysosomes represents another pathogenic mechanism of amiodarone-induced liver damage. Leakage of proteolytic enzymes may contribute to the elevation of aminotransferases and may over time lead to hepatic necrosis, fibrosis, and eventually cirrhosis.
Immunologic mechanisms may be involved in pathogenesis in instances of amiodarone-induced acute hepatitis in patients with positive Coombs' test [41].
Amiodarone-induced inhibition of cellular respiration is another possible pathogenic mechanism for amiodaroneinduced liver damage. Impairment of mitochondrial -oxidation and uncoupling of oxidative phosphorylation leads to the formation of reactive oxygen species, which in turn has a role in the development of AIC [42][43][44].

Pathology.
On histologic examination of biopsy samples obtained from amiodarone-induced cirrhotic patients; leukocytic infiltrate and strikingly high Mallory's hyaline along with other usual pathologic findings of cirrhosis are noted. High Mallory's hyaline or Mallory's bodies are suggestive of AIC.
Mallory's hyaline is an eosinophilic inclusion made up of intermediate keratin filaments. Mallory's hyaline is not specific for AIC and may be seen in primary biliary cirrhosis, alcoholic cirrhosis or hepatitis, nonalcoholic cirrhosis, hepatocellular cancer, morbid obesity, and some other conditions. Mallory hyaline in AIC is present in zone 1 of acinus, whereas in alcoholic liver disease they are located usually in zone 3 [4].
Histologic findings in patients with amiodarone-induced hepatic damage are similar to those caused by alcohol [4,45]. The complete pathologic spectrum of alcoholic like liver injury due to amiodarone includes micro-and macrovesicular steatosis, steatonecrosis, mega mitochondria, portal inflammation, fibrosis, and cirrhosis. Amiodarone-associated epithelloid granulomas have also been reported [8].
Presence of phospholipids laden lamellar lysosomal inclusion bodies on electron microscopy [1,4,37] is another characteristic pathologic finding of AIC of liver. Yap et al. observed that amiodarone-induced lysosomal inclusions developed in nearly 100% of patients at a period of 2 weeks [46].

Diagnostic Workup. AIC is a diagnosis of exclusion.
Extensive workups are normally done to exclude other diagnoses including viral etiology, Wilson's disease, hemochromatosis, alpha 1 antitrypsin deficiency, alcoholic hepatitis, congestive liver damage, autoimmune liver pathologies, and hepatitis due to other drugs and toxins. There is no specific diagnostic lab test for AIC of the liver.
Furthermore, there is no specific imaging characteristic for AIC of the liver although sometimes increased liver density may be noted on a noncontract CT scan of liver. Increased liver density is thought to be secondary to increased iodine content in the liver. Amiodarone has two atoms of iodine that constitute 37% of molecular weight of the drug. Enhanced density due to amiodarone is reversible upon discontinuation of amiodarone [47].
Diagnosis of AIC is usually based on liver biopsy. Mallory hyaline and lamellar lysosomal inclusions are typical of amiodarone-induced liver damage.

Presentation.
Amiodarone-induced liver damage may present as Reye's syndrome in kids [9] and may present as asymptomatic elevation of liver enzymes in adults. Asymptomatic liver enzyme elevation occurs in 25% of the population treated with amiodarone [1] and is usually reversible upon discontinuation of therapy [45]. Normalization of liver enzymes may take place anywhere from three weeks to nine months [48]. Symptomatic hepatic dysfunction occurs in less than 1% of the population treated with amiodarone and includes acute and chronic liver injuries. Acute liver injury includes acute hepatitis (idiosyncratic reaction may be involved in pathogenesis), whereas chronic liver injury includes steatosis (macro and microvesicular steatosis) and cirrhosis.

General Discussion
Amiodarone is an iodinated benzofuran derivative, lipophilic drug with a half-life of 35-110 days and a very large volume of distribution (VD). Amiodarone comes as number eight in drugs that cause drug-related hepatic fatalities [9]. Major metabolite of amiodarone, N-desmethylamiodarone, is not only pharmacologically active but has a longer elimination half-life and a larger VD than the parent drug [1,10,[49][50][51]. Amiodarone and N-desmethylamiodarone may be detected even months after stopping the drug [51] as amiodarone accumulates in lipid reservoirs and is released slowly from these reservoirs. Due to this storage mechanism, amiodarone concentration in liver may be as high as 500-fold of serum level [52]. Based on these facts, damaging effects of amiodarone may persist up to one year after complete discontinuation of therapy [1]. Since amiodarone is mainly metabolized in the liver, any damage to the liver from any cause would hamper amiodarone metabolism and lead to a vicious cycle of accumulation of amiodarone and further amiodaroneinduced hepatic damage [53]. Why some patients develop cirrhosis or hepatic damage as a side effect of amiodarone is not entirely clear. It has been suggested that differing sensitivity to amiodarone toxicity in population may exist [54]. Most patients who developed AIC usually used amiodarone PO 200 mg or more per day for more than 1-2 years. In light of the above-mentioned fact, researchers propose that the total cumulative dose of amiodarone may be important in estimating the risk of irreversible liver injury [55]. A cumulative dose of 380 g is suggested to associate with hepatotoxicity leading to cirrhosis [50]. Other prospective studies showed that amiodarone hepatic toxicity correlates to steady state serum levels of amiodarone rather than daily or cumulative doses [54,[56][57][58]. For example, if daily dosage of amiodarone during long-term therapy is reduced, despite increasing lifetime cumulative dose, the steady state serum concentration will still be reduced and thus decreasing risk of hepatotoxicity from amiodarone. Researchers have suggested that amiodarone level less than 1.5 mg/L has a minimal risk of hepatotoxicity, whereas a level above 2.5 mg/L may have a risk up to 6% for hepatotoxicity [5]. Patients with lower eject fraction may be more prone to hepatotoxic effects of amiodarone as suggested by Tisdale et al. [59]. Although it seems logical that patients with preexisting liver damage may be more prone to amiodarone-induced AIC of the liver, results by Kum et al. suggest otherwise [60].
Besides chronic liver injury due to prolonged amiodarone use, acute hepatic side effects from amiodarone intravenous loading dose have been reported and are thought to be caused by polysorbate 80; a solvent used in drug preparation. This form of acute hepatotoxicity usually improves with discontinuation of medication, although fatalities have been reported [61].
To prevent amiodarone-induced cirrhosis, the amiodarone should be titrated to lowest effective dose. Patients should have baseline LFTs and then periodic monitoring of LFTs while on amiodarone (at 1, 3, and 6 months and then semiannually) [62]. Studies have suggested that baseline LFTs monitoring is performed only in 44% of patients, and a follow-up testing at 6 months and 1 year is done in only 41% and 35% of patients, respectively [63]. Patients found to have asymptomatic elevation of transaminases while on amiodarone should have a thorough investigation of the cause [56]; repeated testing may be necessary before labeling diagnosis of amiodarone-induced hepatotoxicity [45]. According to some authors, discontinuation of amiodarone for liver-related toxicity may not be necessary [57]. Kum et al. noted in their study that 50% of patients with increased transaminases while on amiodarone did not improve even after 1.5 years of drug withdrawal [60]. However, most authors suggest that if aminotransferases are two times above baseline value or above three times the upper normal limit, then amiodarone either should be reduced or discontinued [58,61,64,65]. Withdrawal of drug, when irreversible liver damage has already occurred, has a very little effect on restoration of liver function. Despite what is said, discontinuation of amiodarone for liver-related hepatotoxicity may not be necessary except in cases to prevent irreversible loss of liver function [11]. After discontinuation of amiodarone, monitoring period should continue for at least one year as the damaging effect of amiodarone might persist. Patients on amiodarone should be advised to avoid any potentially hepatotoxic agent to prevent additive hepatic damage. Discontinuation of amiodarone is reported to occur in nearly 20-40% of patients if changes in aminotransferases are detected during amiodarone therapy [66].
To date, the role of routine monitoring of amiodarone or its metabolite's serum levels for predicting hepatic damage is not well established. Nonetheless, there are numerous patients who may develop AIC without any abnormality in liver enzymes. Such patient population will not be detected until very late. Routine imaging may have a role in this subset of patients, but there is no study to support this.
Based on our review (Table 1), we noted that AIC is extremely rare. However, once the diagnosis is established, the mortality risk may be as high as 60% at 5 months. The most common cause of death among reported cases was due to liver-and GI-related complications. Ethnic predisposition to AIC could not be determined due to lack of the published literature. The most common symptoms reported were generalized weakness, abdominal pain, and abdominal distension.
A cumulative dose of 380 g has been suggested to associate with cirrhosis. By reviewing published cases, we realize that cirrhosis has been reported to occur with a cumulative dose as low as 55 g [9]. Why some patients with a cumulative dose of 200 g develop cirrhosis and others do not until cumulative dose crosses above 1000 g is unclear, but it does suggest that it may be a steady state concentration rather than cumulative dose that may be important in predicting risk of cirrhosis. The diagnosis in most patients was made by liver biopsy. Although Mallory hyaline and lysosomal inclusions are characteristic features of amiodarone-induced cirrhosis, the dilemma is that the above-mentioned pathologic characteristics may be identified in patients on amiodarone who do not have cirrhosis. Liver biopsy alone for diagnosis of AIC may not be enough. In fact, a thorough evaluation for all possible causes of cirrhosis should be done. Once all possible causes of cirrhosis have been ruled out, then only the presence of Mallory hyaline and lysosomal inclusions in cirrhotic liver may be suggestive of AIC of the liver. In all published cases, a thorough investigation for the cause of cirrhosis was carried out. In most published cases, obesity, DM, and ethanol consumption were given special attention to rule out the possibility of alcoholic and nonalcoholic steatohepatitis (NASH). Risk factors for development of amiodarone-induced hepatotoxicity and cirrhosis have not been clearly defined. Whether it is cumulative dose or steady state concentration of amiodarone that predicts risk of cirrhosis has not been determined. In most patients even after discontinuation of amiodarone, toxicity effects did not subside. Majority of patients have had amiodaronerelated hepatotoxicity before onset of cirrhosis leading to a decrease in dose or discontinuation of the drug. However, the drug was restarted for various reasons, which unfortunately lead to cirrhosis. Although AST and ALTs are said to be mainly elevated in amiodarone-related hepatotoxicity and not the ALK P or GGT [56], our analysis suggests otherwise. According to our analysis, the only statistically significant variable different among survivors versus nonsurvivors at 5 months was the level of AST at time of diagnosis of AIC. We suggest that if patients on amiodarone have a persistent increase in aminotransferases, a liver imaging and a liver biopsy should be done.
There is no treatment available for amiodarone-induced hepatotoxicity or cirrhosis besides discontinuation of the offending agent and switching to some other antiarrhythmic agent. Previously, antioxidants vitamin E and selenium have been tried without any success. We recommend against using antioxidants for counteracting amiodarone hepatotoxicity, as there is no strong scientific evidence for such practices.
We are aware that there are limitations in our study because of the smaller sample size, which may have affected our analysis. Future prospective studies using larger patient population may be able to identify predictors of survival in case of AIC.

Conclusion
Although amiodarone does have very serious and fatal effects on the liver, such effects are rare. With a closer monitoring and taking appropriate actions when prompted, amiodarone can be safely used on long-term basis. Further prospective studies are needed to identify predictors of AIC and of mortality or survival in cases of AIC. Role of routine imaging and biopsy of the liver in patients taking amiodarone is unclear; future studies are needed to address this issue.