Lipid lowering medication and hepatotoxicity


Rolf Hultcrantz Department of Gastroenterology and hepatology, Karolinska hospital, S-176 76 Stockholm, Sweden (tel: +46 8 517 750 33; fax: +46 8 517 726 37; e-mail:

Drug-induced liver injuries represent an important challenge for physicians, the pharmaceutical industry and health authorities. Practically the whole spectrum of acute and chronic hepatobiliary disease can be reproduced by toxicity attributed to drugs. Drugs are a leading cause of fulminant hepatitis in Western countries and may also cause chronic liver disease with development of liver cirrhosis. Hepatotoxicity is the principal cause for withdrawal of drugs in clinical trials and after marketing for safety reasons [1]. One recent example illustrating the seriousness of this problem is the recent withdrawal of troglitazone because of hepatotoxicity.

Drug induced hepatotoxicity may mimic almost any type of hepatobiliary disease. Most drug toxicity to the liver are acute, presenting as cytotoxic hepatocellular hepatitis, cholestatic disease or mixed pictures. Less common reactions include chronic hepatitis, chronic cholestasis, granulomatous hepatitis, steatosis, steatohepatitis, veno-occlusive disease, neoplasia, fibrosis and cirrhosis [2]. Life threatening liver failure may develop not only in acute fulminant hepatitis, but also in drug reactions with an insidious course leading to cirrhosis.

Liver cirrhosis is the late stage of a longstanding process in which fibrosis accumulates in an attempt to encapsulate injury. Once cirrhosis occurs, patients may develop the end-stage complications of liver disease; portal hypertension, haemorrhage, hepatic insufficiency and primary liver cancer. An important step in the understanding of liver fibrosis has been the identification of hepatic stellate cells (HSC) as the main fibrogenic cell type in the liver [3]. HSC are normally located in the perisinusoidal space of Disse as quiescent cells rich in vitamin A. Following liver injury, HSC undergo a process known as ‘activation’, characterized by a phenotypic conversion from quiescent vitamin A-rich cells into highly proliferative myofibroblast-like cells producing connective tissue components generating fibrosis (Fig. 1). HSC activation is initiated by mediators released from injured hepatocytes and neighbouring endothelial and Kupffer cells (resident macrophages in the liver). Amongst numerous mediators involved in HSC activation, transforming growth factor-β has been identified as the principal fibrogenic mediator, whilst platelet-derived growth factor-BB is the most potent mitogenic factor. Hepatocytes are also a potent source of lipid peroxides, which may be important in many forms of liver fibrosis.

Figure 1.

 Hepatic stellate cell (HSC) activation. Following liver injury, HSC undergo a process known as ‘activation’, characterized by a transition from quiescent vitamin A-rich cells into highly proliferative and fibrogenic myofibroblast-like cells. HSC activation is initiated by mediators released from injured hepatocytes and neighbouring endothelial and Kupffer cells.

The toxic effects of drugs on the liver are most probably underestimated. For most drugs the prevalence of liver toxicity is relatively low, which explains why toxicity often is not detected during clinical development of the drug. Instead, the evaluation of adverse liver reactions relies to a large extent on spontaneous reporting of individual cases to drug safety authorities after marketing authorization has been obtained. Moreover, the diagnosis of drug hepatotoxicity is admitted to be difficult and requires a high level of suspicion and may therefore easily be overlooked in clinical practice.

Once suspicion of drug induced liver disease has been evoked in a case, the major challenge will be the attribution of causality. Except for those cases in which a positive rechallenge (unacceptably dangerous in most cases but sometimes accidental) confirms the hepatotoxic role of a certain drug, the diagnosis of drug induced liver disease is subjective and based on circumstantial evidence. The development of clinical scales for causality assessment, which generate a numerical score intended to identify the probability of causality, has helped to standardize and to improve the level of reliability and objectivity of the evaluation of drug hepatotoxicity. The system most widely used was developed by an international consensus group in 1990 [4]. The principles of this system relies on the attribution of scores to different parameters such as time to onset, course, risk factors, concomitant drugs, search for nondrug causes, previous information on hepatotoxicity of the drug and response to re-administration. The total score generated distinguishes the probability of causality of a given drug as highly probable, probable, possible, unlikely, or excluded.

The mechanisms of hepatotoxicity remain unknown for the majority of drugs [5]. Drug induced damage to the liver may be predictable, that is, they occur in all individuals exposed to a sufficient dose, such as toxic hepatitis after intake of overdoses of paracetamol. However, most adverse liver reactions are unpredictable and involve presumed metabolic idiosyncratic reactions, or immuno-allergic mechanisms, in which the compound triggers an adverse immune response directed against the liver tissue. In the latter case, main clinical features include association with hypersensitivity reactions such as fever, chills, skin rash, hypereosinophilia, a shortened delay upon rechallenge and occasionally presence of auto-antibodies.

In the current issue of this journal, Punthakee et al. describe two cases of liver injury attributed to treatment with lipid lowering drugs of the statin and fibrate classes. In both cases the causality was probable.

Fibric acid derivates or fibrates, cause enhanced oxidation of free fatty acids in muscle cells and reduce the rate of hepatic lipid generation [6]. Fibrates bind to peroxisome proliferator-activated receptors type α (PPARα), which are nuclear hormone receptors. These bind to deoxyribonucleic acid (DNA) at specific sites called peroxisome proliferator response elements and either activate or repress their expression. Fibrates are associated with several adverse effects. Liver enzyme elevations are not uncommon with fenofibrate. Although rare, cases with chronic active hepatitis of auto-immune origin and severe fibrosis have been reported after treatment with fenofibrate and ciprofibrate [7–9].

The 3-hydroxy-3-methyglutaryl coenzyme A (HMG CoA) reductase inhibitors, or statins, are a class of agents that interfere with the rate-limiting step in the synthesis of cholesterol [6]. The most common adverse reactions to statins include nausea and diarrhoea, muscle pain or weakness and increased transaminases. Only rare cases of symptomatic liver disease and even fewer cases of mild liver fibrosis have been reported. Combinations of statin and fibrate seem safe, although discontinuation of therapy because of elevated transaminases is necessary in a minority of patients [10].

The first case described in this report was treated with statin alone and presented with chronic active hepatitis with mild periportal fibrosis, whilst the second patient was treated with a combination of statin and fibrate and presented with chronic active hepatitis and severe bridging fibrosis. Only the patient treated with fibrates had severe fibrosis, which is in accordance with earlier reported cases from the statin and fibrate literature. In both cases the clinical picture indicated an auto-immune origin.

None of the patients had evidence of chronic liver disease prior to the acute onset of symptoms and increase in transaminases. This raises the question if fibrates and/or statins may lead to liver fibrosis in the absence of pathological liver function tests. If this is true, this could mean that surveillance of aminotransferases is not sufficient to detect hepatotoxicity in patients on cholesterol lowering agents. This would then have major implications for the surveillance of patients under lipid lowering therapy. Should these patients be followed by sequential liver biopsies and/or serum markers for fibrosis, in addition to monitoring of liver function tests, as is currently recommended for patients treated with methotrexate [11, 12]? The answer to this can only come from epidemiological and prospective studies that include examination of liver tissue, as suggested by Punthakee et al.

It is important to keep in mind that the risk of drug hepatotoxicity is influenced by various acquired and genetic factors and it should be important to identify such susceptibility factors for the development of fibrosis in the case of statins and fibrates. It must also be stressed that appropriateness of a therapy is fundamental to avoid unnecessary side-effects of medication in a population. In a recent study, overuse of statin therapy was found amongst 69% of patients undergoing primary prevention and amongst 47% undergoing secondary prevention [13]. This study also showed that monitoring of liver function varied widely in this group of patients.

In summary, the article by Punthakee et al. points at the difficult and ongoing problem of drug induced liver injury. Clearly a high level of attention must be attributed to suspected cases of drug hepatotoxicity. In the case of statins and fibrates, further studies are needed to answer the question of whether these patients or a subgroup of these patients should be monitored for liver fibrosis with more invasive methods.