Statins for non-alcoholic fatty liver disease and non-alcoholic steatohepatitis

  • Review
  • Intervention

Authors

  • Layli Eslami,

    Corresponding author
    1. Golestan University of Medical Science, Taleghani Hospital, Gonbad e Kavous, Golestan province, Iran
    • Layli Eslami, Golestan University of Medical Science, Taleghani Hospital, North Khayyam crossroad, East Taleghani Avenue, Gonbad e Kavous, Golestan province, 49791-31983, Iran. laylieslami@gmail.com.

    Search for more papers by this author
  • Shahin Merat,

    1. Shariati Hospital, Tehran University of Medical Sciences, Digestive Diseases Research Centre, Tehran, Tehran, Iran
    Search for more papers by this author
  • Reza Malekzadeh,

    1. Shariati Hospital, Tehran University of Medical Sciences, Digestive Diseases Research Centre, Tehran, Tehran, Iran
    Search for more papers by this author
  • Siavosh Nasseri-Moghaddam,

    1. Shariati Hospital, Tehran University of Medical Sciences, Digestive Diseases Research Centre, Tehran, Tehran, Iran
    Search for more papers by this author
  • Hermineh Aramin

    1. Shariati Hospital, Tehran University of Medical Sciences, Digestive Diseases Research Centre, Tehran, Tehran, Iran
    Search for more papers by this author

Abstract

Background

Non-alcoholic fatty liver disease (NAFLD) and non-alcoholic steatohepatitis (NASH) are common causes of elevated liver enzymes in the general population. NASH and to some extent NAFLD have been associated with increased liver-related and all-cause mortality. No effective treatment is yet available. Recent reports have shown that the use of hydroxymethylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors (statins) in patients with elevated plasma aminotransferases may result in normalisation of these liver enzymes. Whether this is a consistent effect or whether it can lead to improved clinical outcomes beyond normalisation of abnormal liver enzymes is not clear.

Objectives

To assess the beneficial and harmful effects of statins (that is, lovastatin, atorvastatin, simvastatin, pravastatin, rosuvastatin, and fluvastatin) on all-cause and liver-related mortality, adverse events, and histological, biochemical, and imaging responses in patients with NAFLD or NASH.

Search methods

We performed a computerised literature search in the Cochrane Hepato-Biliary Group Controlled Trials Register, Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library, MEDLINE, EMBASE, and Science Citation Index Expanded up to March 2013. We did fully recursive searches from the reference lists of all retrieved relevant publications to ensure a complete and comprehensive search of the published literature. We did not apply any restrictions regarding language of publication or publication date.

Selection criteria

All randomised clinical trials using statins as the primary treatment for NAFLD or NASH versus no treatment, placebo, or other hypolipidaemic agents.

Data collection and analysis

Data were extracted, and risk of bias of each trial was assessed independently by two or more review authors. Meta-analyses were performed whenever possible. Review Manager 5.2 was used.

Main results

When the described search method was used and the eligibility criteria of the search results were applied, 653 records were found. Only two of these were randomised clinical trials that were considered eligible for inclusion. We assessed both trials as trials with high risk of bias. One of the trials was a pilot trial in which 16 participants with biopsy-proven NASH were randomised to receive simvastatin 40 mg (n = 10) or placebo (n = 6) once daily for 12 months. No statistically significant improvement in the aminotransferase level was seen in the simvastatin group compared with the placebo group. Liver histology was not significantly affected by simvastatin.

The other trial had three arms. The trial compared atorvastatin 20 mg daily (n = 63) versus fenofibrate 200 mg daily (n = 62) versus a group treated with a combination of the two interventions (n = 61). There were no statistically significant differences between any of the three intervention groups regarding the week 54 mean activity levels of aspartate aminotransferase, alanine aminotransferase, gamma-glutamyl transpeptidase, and alkaline phosphatase. The triglyceride levels seemed higher in the fenofibrate group compared with the atorvastatin group. Liver histology was not assessed in this trial. The presence of biochemical and ultrasonographic evidence of NAFLD seemed to be higher in the fenofibrate group compared with the atorvastatin group (58% versus 33%). Three patients discontinued treatment due to myalgia and elevated serum creatine kinase activity; one from the atorvastatin group and two from the combination group. Another patient from the atorvastatin group discontinued treatment due to alanine aminotransferase activity that was over three times the upper normal limit.

No data for all-cause mortality and hepatic-related mortality were reported in the included trials.

Authors' conclusions

Based on the findings of this review, which included two trials with high risk of bias and a small numbers of participants, it seems possible that statins may improve serum aminotransferase levels as well as ultrasound findings. Neither of the trials reported on possible histological changes, liver-related morbidity or mortality. Trials with larger sample sizes and low risk of bias are necessary before we may suggest statins as an effective treatment for patients with NASH. However, as statins can improve the adverse outcomes of other conditions commonly associated with NASH (for example, hyperlipidaemia, diabetes mellitus, metabolic syndrome), their use in patients with non-alcoholic steatohepatitis may be justified.

Résumé scientifique

Statines pour la stéatose hépatique non alcoolique et la stéatohépatite non alcoolique

Contexte

La stéatose hépatique non alcoolique (NAFLD) et la stéato-hépatite non alcoolique (NASH) sont des causes fréquentes des enzymes hépatiques dans la population en général. La NASH et, dans une certaine mesure, la NAFLD ont été associées à une augmentation de la stéatose hépatique liée au foie et à la mortalité toutes causes confondues. Aucun traitement efficace n’est encore disponible. Des rapports récents ont montré que l'utilisation d'inhibiteurs de l’hydroxymethylglutaryl-coenzyme (HMG-CoA) réductase (statines) chez les patients souffrant d’aminotransférase plasmatique pourrait entraîner une normalisation de ces enzymes hépatiques. Il n’est pas clairement déterminé s’il s’agit d’un effet constant ou si cela peut conduire à une amélioration des résultats cliniques au-delà de la normalisation des enzymes hépatiques anormaux.

Objectifs

Évaluer les effets bénéfiques et délétères des statines (lovastatine, atorvastatine, simvastatine, pravastatine, rosuvastatine et fluvastatine) sur la mortalité toutes causes confondues et sur la mortalité liée au foie, sur les effets indésirables et histologiques, sur les réponses biochimiques et d'imagerie chez les patients atteints NAFDL ou de NASH.

Stratégie de recherche documentaire

Nous avons effectué une recherche de littérature informatisée dans le registre des essais contrôlés du groupe Cochrane sur les affections hépato-biliaires, le registre des essais contrôlés de la centrale Cochrane (CENTRAL) dans la bibliothèque Cochrane, MEDLINE, EMBASE et Science Citation Index Expanded jusqu' à mars 2013. Nous avons effectué des recherches récursives des références bibliographiques de toutes les publications pertinentes pour garantir une recherche complète et exhaustive de la littérature publiée. Nous n'avons appliqué aucune restriction concernant la langue ou la date de publication.

Critères de sélection

Tous les essais cliniques randomisés utilisant des statines comme principal traitement de la NAFLD ou de la NASH par rapport à l'absence de traitement, un placebo, ou d'autres agents hypolipidémiants.

Recueil et analyse des données

Les données ont été extraites et le risque de biais de chaque essai a été évalué indépendamment par deux ou plusieurs auteurs de la revue. Les méta-analyses ont été réalisées lorsque cela était possible. Review Manager 5.2 a été utilisé.

Résultats principaux

Lorsque la méthode de recherche décrite a été utilisée et les critères d'éligibilité des résultats ont été appliquées, 653 dossiers ont été trouvés. Seuls deux étaient des essais cliniques randomisés et ont été considérés comme éligibles pour l'inclusion. Nous avons évalué les deux essais avec un risque de biais élevé. Un des essais était un essai pilote dans lequel 16 participants atteints de NASH confirmée par biopsie ont été randomisés pour recevoir 40 mg de simvastatine (n =10) ou un placebo (n =6) une fois par jour et pendant 12 mois. Aucune amélioration statistiquement significative dans le niveau d’aminotransférase n’était observée dans le groupe sous simvastatine par rapport au groupe sous placebo. L’histologie hépatique n'était pas significativement affectée par la simvastatine.

L'autre essai possédait trois groupes. L'essai comparait 20 mg d’atorvastatine par jour (n =63) versus 200 mg de fénofibrate (n =62) versus un groupe traité avec une combinaison des deux interventions (n =61). Il n'y avait aucune différence statistiquement significative entre l'un des trois groupes d'intervention concernant les activités moyennes de l’aspartate aminotransférase, de l'alanine aminotransférase du gamma-glutamyl transpeptidase et de la phosphatase alcaline pendant 54 semaines. Le niveau des triglycérides semblait plus élevé dans le groupe sous fénofibrate que dans le groupe sous atorvastatine. L'histologie hépatique n'était pas évaluée dans cet essai. La présence de preuves biochimiques et ultrasonographiques de NAFLD semblait être plus élevée dans le groupe sous fénofibrate par rapport aux deux groupes d'intervention. Trois patients ont arrêté le traitement en raison de myalgie et d’un taux élevé de l'activité de la créatine kinase sérique, un patient appartenait au groupe de l'atorvastatine et deux au groupe de combinaison. Un patient dans le groupe de l'atorvastatine avait arrêté le traitement en raison d'une augmentation de l'activité d'alanine aminotransférase, qui était plus de trois fois supérieure à la limite normale.

Aucune donnée sur la mortalité toutes causes confondues et sur la mortalité liée à l’hépatique n’a été rapportée dans les essais inclus.

Conclusions des auteurs

Basé sur les résultats de cette revue, qui comprenait deux essais présentant un risque de biais élevé et un faible nombre de participants, il semble possible que les statines puissent améliorer les niveaux d’aminotransférase sérique, ainsi que les résultats de l'échographie. Aucun des essais n’a rapporté d'éventuels changements histologiques, de morbidité liée au foie ou de mortalité. Des essais avec des tailles d'échantillon plus importantes et un faible risque de biais sont nécessaires avant de pouvoir proposer des statines comme traitement efficace chez les patients atteints de la NASH. Cependant, étant donné que les statines peuvent améliorer les critères de jugement indésirables couramment associés à d'autres troubles de la NASH (tels que l'hyperlipidémie, le diabète, le syndrome métabolique), leur utilisation chez les patients atteints de stéato-hépatite non alcoolique peut être justifiée.

Plain language summary

Statins for non-alcoholic steatohepatitis

Non-alcoholic fatty liver disease (NAFLD) and non-alcoholic steatohepatitis (NASH) in patients with no or very little alcohol consumption is characterised by hepatic histological changes similar to those associated with alcohol-induced liver injury. A range of histological changes can be seen. Some patients have fat accumulation in hepatocytes without significant inflammation or fibrosis (simple hepatic steatosis or NAFLD), but others have hepatic steatosis with prominent necro-inflammatory changes with or without associated fibrosis (this is NASH). Although NAFLD and NASH are common conditions, no effective medical treatment is available to correct the abnormal liver enzymes and adverse outcomes associated with them. This systematic review identified two randomised clinical trials with very small numbers of participants. One of the trials was a pilot trial and compared simvastatin with placebo, and the other trial assessed atorvastatin versus fenofibrate versus a combination of the two. The small pilot trial (n = 16 patients) assessing simvastatin versus placebo in NASH patients did not show significant effects on liver enzyme activities or liver histology. No adverse events were reported. The other trial compared atorvastatin versus fenofibrate versus a group receiving both interventions in 186 patients with NAFLD. There were no statistically significant differences between any of the three intervention groups regarding the 54 week mean activities of aspartate aminotransferase, alanine aminotransferase, gamma-glutamyl transpeptidase, or alkaline phosphatases (liver enzymes) in the blood. The triglyceride levels seemed higher in the fenofibrate group compared with the atorvastatin group. Liver histology was not assessed in this trial. The presence of biochemical and ultrasonographic evidence of NAFLD seemed higher in the fenofibrate group compared with the other two intervention groups. Three patients discontinued treatment due to myalgia and elevated serum creatine kinase activity, one from the atorvastatin group and two from the combination group. Another patient from the atorvastatin group discontinued treatment due to raised alanine aminotransferase activity, over three times the upper normal limit. Both trials were at high risk of bias (that is, overestimation of benefits and underestimation of harms). Furthermore, the groups were small raising the risks of random errors (that is, play of chance). Accordingly, we did not find evidence to support or refute the use of statins for patients with NAFLD or NASH. Further unbiased trials with larger numbers of patients looking explicitly at patient-related outcomes of interest (for example, quality of life, development of cirrhosis, and mortality) are needed to assess the effects of statins on NAFLD or NASH.

Résumé simplifié

Les statines en cas de stéatohépatite non alcoolique

La stéatose hépatique non alcoolique (NAFLD) et la stéatohépatite non alcoolique (NASH) chez les patients avec une absence ou une consommation minimum d'alcool est caractérisée par des changements histologiques hépatiques similaires à ceux associés à une lésion hépatique induite par l'alcool. Un éventail de changements histologiques peut être observé. Certains patients ont accumulé des triglycérides dans les hépatocytes sans inflammation ni fibrose significative (simple stéatose hépatique non alcoolique ou NAFLD), mais d'autres souffrent de stéatose hépatique avec d’importants changements nécro-inflammatoires avec ou sans fibrose associée (NASH). Bien que la NAFLD et la NASH soient des affections courantes, aucun traitement médical disponible n’est efficace pour corriger les enzymes hépatiques anormales et les effets indésirables qui sont associés. Cette revue systématique a identifié deux essais cliniques randomisés avec un très petit nombre de participants. Un des essais était un essai pilote et comparait la simvastatine avec un placebo. L'autre essai évaluait l'atorvastatine versus le fénofibrate versus une association des deux. Le petit essai pilote (n = 16 patients) évaluant la simvastatine par rapport à un placebo chez les patients atteints de NASH n'a démontré aucun effet significatif sur l'activité des enzymes hépatiques ou sur l'histologie hépatique. Aucun effet indésirable n'a été signalé. L'autre essai comparait l'atorvastatine versus le fénofibrate versus un groupe recevant les deux interventions chez 186 patients atteints de NASH. Il n'y avait aucune différence statistiquement significative parmi les trois groupes d'intervention concernant les activités moyennes de l’aspartate aminotransférase, de l'alanine aminotransférase et du gamma-glutamyl transpeptidase, ou des phosphatases (enzymes hépatiques) dans le sang pendant 54 semaines. Les niveaux de triglycérides semblaient plus élevés dans le groupe sous fénofibrate que celui sous atorvastatine. L'histologie hépatique n'était pas évaluée dans cet essai. La présence de preuves biochimiques et ultrasonographiques de NAFLD semblait être plus élevée dans le groupe sous fénofibrate par rapport aux deux groupes d'intervention. Trois patients ont arrêté le traitement en raison de myalgie et d’un taux élevé d'activité de créatine kinase sérique, un patient appartenait au groupe de l'atorvastatine et deux au groupe de combinaison. Un patient dans le groupe de l'atorvastatine avait arrêté le traitement en raison d'une augmentation de l'activité d'alanine aminotransférase, qui était plus de trois fois supérieure à la limite normale. Les deux essais étaient à risque de biais élevé (surestimation des effets bénéfiques et sous-estimation des effets néfastes). De plus, les groupes étaient de petite taille, augmentant le risque d'erreurs aléatoires (jeux de hasard). En conséquence, nous n'avons pas trouvé de preuves pour soutenir ou réfuter l'utilisation de statines chez les patients atteints de NAFLD ou de NASH. D'autres essais avec un plus grand nombre de patients portant explicitement sur des critères de jugement liés aux patients (par exemple, la qualité de vie, le développement d'une cirrhose et la mortalité) sont nécessaires pour évaluer les effets des statines sur la NAFLD ou la NASH.

Notes de traduction

Traduit par: French Cochrane Centre 14th January, 2014
Traduction financée par: Financeurs pour le Canada : Instituts de Recherche en Santé du Canada, Ministère de la Santé et des Services Sociaux du Québec, Fonds de recherche du Québec-Santé et Institut National d'Excellence en Santé et en Services Sociaux; pour la France : Ministère en charge de la Santé

Background

Non-alcoholic fatty liver disease (NAFLD) ranges in spectrum from steatosis only to steatohepatitis (non-alcoholic steatohepatitis (NASH)) and cirrhosis (Brunt 2001). According to the last guideline developed by the American Gastroenterological Association, the American Association for the Study of Liver Diseases, and the American College of Gastroenterology (AGA 2012), "the diagnosis of NAFLD requires that (a) there is hepatic steatosis by imaging or histology, (b) there is no significant alcohol consumption, (c) there are no competing etiologies for hepatic steatosis, and (d) there are no co-existing causes for chronic liver disease". NASH is a distinct clinical entity characterised by chronically elevated plasma liver enzymes (namely, aspartate aminotransferase (AST) and alanine aminotransferase (ALT)) that is unexplained by the presence of any of the other known causes of chronically elevated AST or ALT, or both (that is, viral, autoimmune, metabolic, and drug or toxin-induced causes) and liver biopsy findings that are similar to those seen in alcoholic hepatitis. Patients with NASH, however, do not consume alcohol in the quantities known to be able to cause liver injury (Ludwig 1980). Analysis of liver biopsy specimens is the gold standard for diagnosis of NASH, although this approach is not always used. Hepatic morphological findings may range from mild fatty changes to inflammation and hepatocyte injury, fibrosis, and cirrhosis with or without the presence of Mallory-Denk bodies (Brunt 2005; Yeh 2007).

The aetiology of NASH is multi-factorial, and its pathogenesis is not completely elucidated as yet. It is frequently associated with obesity, type 2 diabetes mellitus, hyperlipidaemia, and other features of the metabolic syndrome (Ludwig 1980; Lee 1988; Powell 1990; Wang 2006; Brunt 2007; Fujita 2009). Hyperlipidaemia (that is, hypertriglyceridaemia, hypercholesterolaemia, or both), which is frequently associated with both obesity and type 2 diabetes, has been reported in 20% to 80% of participants with NASH (Ludwig 1980; Lee 1988; Cello 1990; Powell 1990; Bacon 1994). Among these metabolic disturbances, obesity and insulin resistance are most commonly associated with NASH and are strongly linked to abnormal lipid metabolism resulting in increased hepatic accumulation of triacylglycerol (Hsiao 2007; Villanueva 2009). These abnormalities are associated with dysregulation of several related cellular processes and signalling pathways, including insulin signalling, de novo lipogenesis, mitochondrial dysfunction, oxidative stress, and endoplasmic reticulum stress (Sanyal 2001; Wang 2006; Wei 2006).

The liver maintains the whole body energy homeostasis by metabolising glucose and fatty acids. It can store a significant amount of lipid in the form of triacylglycerol in conditions associated with prolonged energy excess and impaired fatty acid metabolism. In hepatic steatosis, plasma levels are increased and hepatic accumulation of triacylglycerol is seen (Brunt 2001; Puri 2007). Our current knowledge of NASH pathogenesis indicates that statins may theoretically be beneficial in this condition (Ahmed 2006).

Description of the condition

NAFLD is usually an asymptomatic disease and the liver enzymes can be normal or elevated. Some patients may complain of right upper quadrant pain, which has not been shown to be related to NAFLD. Some may have an enlarged liver on physical examination. NASH patients may have symptoms similar to NAFLD and some may have symptoms suggestive of advanced liver damage (for example, jaundice, ascites, variceal bleeding).

NASH is a distinct clinical entity characterised by chronically elevated plasma liver enzymes and liver biopsy findings similar to those seen in alcoholic hepatitis. These patients do not consume alcohol in quantities known to cause liver injury (Sanyal 2001). It is widespread clinical experience that patients often under-report alcohol consumption due to cultural or religious reasons. Therefore, a number of patients suspected of NAFLD or NASH may indeed have an alcoholic origin to their disease.

Description of the intervention

Hydroxymethylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors are lipid-altering agents. These agents are called statins and are competitive inhibitors of HMG-CoA reductase, the rate-limiting step in cholesterol biosynthesis (Istvan 2001). Statins encompass drugs like lovastatin, atorvastatin, simvastatin, pravastatin, rosuvastatin, and fluvastatin.

How the intervention might work

Statins occupy a portion of the binding site of HMG-CoA, blocking access of the substrate (3-hydroxy-3-methylglutaryl CoA (HMG-CoA) to the active site on the enzyme. A reduction in intrahepatic cholesterol leads to an increase in low-density lipoprotein (LDL) receptor turnover that results from an enhanced rate of hepatic LDL receptor cycling (Ness 1996). Recent studies have described several important cellular processes and signalling pathways that are affected by abnormal lipid metabolism, resulting in specific biochemical, histological, and clinical changes associated with NAFLD (Cheung 2008). It could be that statins, as lipid-lowering agents, may reduce intrahepatic cholesterol and affect the abnormal lipid metabolism seen in patients with NASH.

Why it is important to do this review

Recent reports have indicated that HMG-CoA reductase inhibitors (statins) may normalise liver enzymes and elevated levels of liver aminotransferases in patients with NAFLD or NASH (Kiyici 2003; Hatzitolios 2004; Rallidis 2004; Athyros 2005). Whether this is a consistent effect or whether it can lead to improved clinical outcomes beyond normalisation of abnormal liver enzymes is not clear. This systematic review attempted to summarise the available data, assess the risk of bias of included trials, and strengthen available evidence to answer these questions. To the best of our knowledge no previous systematic review or meta-analysis has addressed these questions in patients with NAFLD or NASH.

Objectives

To assess the beneficial and harmful effects of statins (that is, lovastatin, atorvastatin, simvastatin, pravastatin, rosuvastatin, and fluvastatin) on all-cause and liver-related mortality, adverse events, and histological, biochemical, and imaging responses in patients with NAFLD or NASH.

Methods

Criteria for considering studies for this review

Types of studies

Randomised clinical trials comparing statins versus no intervention, placebo, or other interventions such as other lipid-lowering agents, no matter the numbers of participants randomly assigned or the publication status, year of publication, or language of publication. Non-randomised studies, if retrieved through searches for randomised clinical trials, were checked for reports of adverse events.

Types of participants

Participants of any age, sex, or ethnic origin diagnosed with NAFLD or NASH were included on the basis of the following criteria.

  • Imaging (ultrasound, computed tomography, or magnetic resonance imaging) or histology showing hepatic steatosis with or without elevated plasma liver enzymes and without other causes of hepatic steatosis.

  • Imaging (ultrasound, computed tomography, or magnetic resonance imaging) showing evidence of hepatic steatosis together with increased serum ALT above 1.5 times the upper normal limit or liver biopsy showing histological alterations suggestive of NASH, or both.

  • For both NAFLD and NASH, daily alcohol intake should be less than 10 g in women and less than 20 g in men.

We excluded trials that included participants:

  • diagnosed with hepatitis B, hepatitis C, autoimmune hepatitis, or genetic liver diseases, such as Wilson’s disease and haemochromatosis;

  • presenting with one or more causes commonly associated with secondary NAFLD (for example, drugs, surgical procedures, environmental toxins, total parenteral nutrition).

Types of interventions

The experimental intervention was a statin (lovastatin, atorvastatin, simvastatin, pravastatin, rosuvastatin, or fluvastatin) administered at any dose for a minimum of three months. Participants given the experimental intervention should be compared with a control group receiving no statin intervention, placebo, or other interventions such as another lipid-lowering agent. Only oral therapies were considered. Co-interventions were allowed if used equally in all intervention groups.

Types of outcome measures

Primary outcomes
  • All-cause mortality.

  • Hepatic-related mortality.

  • Adverse events: numbers and types of adverse events, defined as any untoward medical occurrence not necessarily having a causal relationship to the treatment. A serious adverse event was defined, according to the International Conference on Harmonisation (ICH) Guidelines (ICH-GCP 1997), as any event that led to death, was life threatening, required in-patient hospitalisation or prolongation of existing hospitalisation, or resulted in persistent or significant disability, as well as any important medical event that might have endangered the participant or required intervention to prevent its occurrence. All other adverse events were considered non-serious.

Secondary outcomes
  • Histological response: numbers of participants with histological deterioration in the degree of fatty liver infiltration, inflammation, and fibrosis based on one of the two scoring systems (that is, Brunt 1999 or Kliener 2005).

  • Biochemical response: numbers of participants with any reported improvement or deterioration in serum levels of AST or ALT, or both.

  • Imaging response: degree of fatty liver infiltration as assessed by ultrasound, computed tomography, or magnetic resonance imaging. The severity of steatosis is graded as mild, moderate, or severe (Rofsky 1995; Saadeh 2002; Poonam 2007; Schwenzer 2009).

Search methods for identification of studies

Electronic searches

We performed a computerised literature search of the Cochrane Hepato-Biliary Group Controlled Trials Register (Gluud 2013), Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library, MEDLINE, EMBASE, and Science Citation Index Expanded (Royle 2003). Search strategies used along with the time spans of these searches are given in Appendix 1. We limited our search to studies in humans. We applied no language restrictions for searching or trial inclusion. We read through the reference lists of retrieved publications to identify additional trials.

Searching other resources

We handsearched the abstracts of the American Digestive Disease Week (DDW), published in Gastroenterology, and the United European Gastroenterology Week (UEGW), published in Gut, from 1995 to March 2013. We scanned reference lists of retrieved publications to identify further relevant trials. We asked the authors of trials published only as abstracts to contribute full data sets or completed papers.

Data collection and analysis

Selection of studies

We screened the titles and abstracts of all potentially relevant studies for their relevance before retrieval of full articles. Where the title and abstract were ambiguous, we assessed full articles for relevance. Relevant trials were determined by consensus of the four review authors. Two review authors independently applied the selection criteria according to the pre-stated eligibility criteria; when disagreements occurred, they were resolved by consensus.

Data extraction and management

Two review authors independently extracted the following data.

  • Trial characteristics: date, location, and funding of the trial; trial design; length of follow-up; intention-to-treat analysis; and publication status.

  • Participant characteristics: number of participants randomly assigned; inclusion and exclusion criteria, mean (or median) participant age; and sex ratio.

  • Intervention characteristics: dose and duration of administration of various statins and additional intervention(s) and of other lipid-lowering agents or placebo in the control groups.

  • Outcomes: numbers of events in the intervention group and in the control group for each of the outcomes we have specified in our protocol.

Assessment of risk of bias in included studies

The authors followed the instructions given in The Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) and the Cochrane Hepato-Biliary Group Module (Gluud 2013) to assess the risk of bias of included trials.

Because of the risk of biased overestimation or underestimation of intervention effects in randomised trials with inadequate methodological quality (that is, risk of bias) (Schulz 1995; Moher 1998; Kjaergard 2001; Wood 2008; Lundh 2012; Savovic 2012; Savovic 2012a), we looked at the influence of the validity of included trials on study results by evaluating the bias risk domains (Higgins 2011). The risk of bias of each trial was assessed using the following domains.

Allocation sequence generation
  • Low risk of bias: sequence generation was achieved using computer random number generation or a random number table. Drawing lots, tossing a coin, shuffling cards, and throwing dice are adequate if performed by an independent research assistant not otherwise involved in the trial.

  • Uncertain risk of bias: the method of sequence generation was not specified.

  • High risk of bias: the sequence generation method was not random. 

Allocation concealment
  • Low risk of bias: participant allocations could not have been foreseen in advance of, or during, enrolment. Allocation was controlled by a central and independent randomisation unit. The allocation sequence was unknown to the investigators (for example, if the allocation sequence was hidden in sequentially numbered, opaque, sealed envelopes).

  • Uncertain risk of bias: the method used to conceal the allocation was not described, so that intervention allocations may have been foreseen in advance of, or during, enrolment.

  • High risk of bias: the allocation sequence was likely to be known to the investigators who assigned participants. 

Blinding of participants, personnel, and outcome assessors
  • Low risk of bias: blinding was performed adequately, or assessment of outcomes was not likely to be influenced by lack of blinding.

  • Uncertain risk of bias: information was insufficient to permit assessment of whether blinding was likely to introduce bias into study results.

  • High risk of bias: no blinding or incomplete blinding was provided, and assessment of outcomes was likely to be influenced by lack of blinding. 

Incomplete outcome data
  • Low risk of bias: missing data were unlikely to make treatment effects depart from plausible values. Sufficient methods, such as multiple imputation, were employed to handle missing data.

  • Uncertain risk of bias: information was insufficient to permit assessment of whether missing data in combination with the method used to handle missing data were likely to introduce bias into study results.

  • High risk of bias: study results were likely to be biased as the result of missing data. 

Selective outcome reporting
  • Low risk of bias: all outcomes were predefined (eg, in a published protocol) and reported, or all clinically relevant and reasonably expected outcomes were reported.

  • Uncertain risk of bias: it is unclear whether all predefined and clinically relevant and reasonably expected outcomes were reported.

  • High risk of bias: one or more clinically relevant and reasonably expected outcomes were not reported, and data on these outcomes were likely to have been recorded.

Other bias
  • Low risk of bias: the trial appears to be free of other components (industry or academic bias) that could put it at risk of bias.

  • Uncertain risk of bias: the trial may or may not be free of other components that could put it at risk of bias.

  • High risk of bias: other factors in the trial could put it at risk of bias (industry or academic bias).

Trials with low risk of bias were considered as those trials having 'low risk of bias' in all of the above domains. Trials judged as trials with 'uncertain risk of bias' or 'high risk of bias' were considered as 'trials with high risk of bias'.

We also reported on whether the investigators had performed a sample size calculation or had used intention-to-treat analysis (Gluud 2001).

Measures of treatment effect

We planned to calculate the pooled P values of all trials based on the Mantel-Haenszel method. Risk ratio (RR) with 95% confidence interval was the effect measure of choice for dichotomous data for all effect sizes. As the included trials compared different interventions, we could not perform meta-analyses. We could not obtain the raw data of two included trials to reconfirm the trials' analyses. This is why we provided the published results of the trials. In both included trials, a 95% confidence interval was used and a two-tailed P value less than 0.05 was considered significant. Parametric data within and between groups were compared using analysis of variance (ANOVA), and nonparametric data were compared using the Kruskal-Wallis test.

Unit of analysis issues

The intervention groups in a randomised clinical trial, or in a cluster-randomised trial, or during the first period of cross-over trials.

Dealing with missing data

All analyses of outcomes were planned to be based on the intention-to-treat principle.

Assessment of heterogeneity

If confidence intervals for the results of individual trials show poor overlap, this generally indicates the presence of statistical heterogeneity. We planned to use a Chi2 test to assess whether observed differences in results are compatible with chance alone. In the cases in which high levels of heterogeneity were detected (the P value is less than 0.1), we planned to put more weight on the random-effects model meta-analysis than on the fixed-effect model meta-analysis. Because the included trials compared different outcomes, we could not perform meta-analyses for the outcomes; tests for heterogeneity were not used.

Assessment of reporting biases

We did not use any analysis to investigate the possibility of publication bias.

Data synthesis

A funnel plot was not used to illustrate the trial data as only two randomised clinical trials fulfilled the inclusion criteria.

Subgroup analysis and investigation of heterogeneity

No data were available for subgroup analyses. In future updates we plan to assess subgroups according to risks of bias (trials without risk of bias compared to trials with risk of bias) and according to most likely diagnostic group (NAFLD compared to NASH).

Sensitivity analysis

No data were available for sensitivity analyses.

Results

Description of studies

See Characteristics of included studies; Characteristics of excluded studies.

Results of the search

After exclusion of duplicate publications, a total of 653 records were identified (Table 1). After exclusion of clearly irrelevant publications, 41 records were identified for further assessment. Six randomised clinical trials were identified (Athyros 2006; Lewis 2007; Nelson 2009; Athyros 2010; Carnelutti 2012; Han 2012). Only two trials were considered eligible for inclusion in the review (Athyros 2006; Nelson 2009). One of the trials was a small pilot trial that assessed simvastatin versus placebo (Nelson 2009). The other trial assessed atorvastatin versus fenofibrate versus a combination of the two interventions (Athyros 2006). Because the two eligible trials were different, we could not perform any meta-analyses. The reasons for exclusion of the remaining four randomised trials and the other clinical trials are explained in the Characteristics of excluded studies section of the review.

Table 1. Search results
Source of the searchNumber of records
CHBG Controlled Trial Register51
Cochrane Central Register of Controlled Trials78
MEDLINE (Ovid)116
EMBASE (Ovid)518
Science Citation Index EXPANDED91
Total number of references identified851
Number of duplicates excluded198
Number of references in final list653

Included studies

Only two trials were included (Athyros 2006; Nelson 2009) (Characteristics of included studies).

The Nelson 2009 pilot trial was designed to assess the effects of simvastatin versus placebo in participants with biopsy-proven NASH over a 12 month period. Serum aminotransferases and repeated liver biopsies were used to assess improvement.

In the Athyros 2006 trial, participants with both biochemical and ultrasonographic evidence of NAFLD at baseline were included and were randomly allocated to receive atorvastatin 20 mg/day (the intervention group in our analysis) or micronised fenofibrate 200 mg/day (the control group in our analysis), or both drugs, for a 54 week period.

The authors of both trials provided data on the total number of randomised participants. Follow-ups have been completed.

Excluded studies

Characteristics and reasons for exclusion of clinical trials are summarised in the Characteristics of excluded studies table.

Risk of bias in included studies

Both of the included randomised trials had high risk of bias as judged by the previously described criteria. Figure 1 and Figure 2 fully explain how the bias risk of trials was assessed.

Figure 1.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Figure 2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Allocation

Generation of allocation sequence and allocation concealment were unclear in Nelson 2009. The Athyros 2006 trial used computer-generated random numbers for generation of allocation sequence, but the method of allocation concealment was not explained; therefore, the trial was considered as having high risk of bias.

Blinding

The method of blinding was not reported in Nelson 2009 and was inadequate in Athyros 2006 because it was an open-label trial.

Incomplete outcome data

Dropouts and withdrawals were precisely reported in one trial (Athyros 2006), but the reasons for withdrawal were unclear in the other trial (Nelson 2009).

Selective reporting

Not all outcome data were available for the included trials.

Other potential sources of bias

The sample size calculation was not reported in any of the trials. None of the trials were stopped prematurely.

Effects of interventions

Simvastatin versus placebo

Primary outcomes
  • All-cause mortality

No all-cause mortality was reported.

  • Hepatic-related mortality

No hepatic-related mortality was reported.

  • Adverse events

No adverse events were reported in any of the intervention groups.

Secondary outcomes
  • Histological response

No histological improvement in hepatic steatosis, necro-inflammation, or fibrosis could be observed when comparing simvastatin versus placebo.

  • Biochemical response

Baseline and post-treatment laboratory variables were assessed in the trial (Table 2). There were no significant differences between the simvastatin and the placebo group regarding the 12 month activities of aspartate aminotransferase, alanine aminotransferase, and alkaline phosphatases.

Table 2. Baseline and after treatment laboratory variables in Nelson et al trial
  1. ALT = alanine aminotransferase
    AST = aspartate aminotransferase
    ALP = alkaline phosphatase
    TG = triglycerides

 Statin (n = 10)Placebo  (n = 6)
Before treatmentAfter treatmentMean decreaseBefore treatmentAfter treatmentMean decrease
ALT (U/L)70.449.520.966.875.3-8.5
AST (U/L)43.336.56.842.849.3-6.5
ALP (U/L)86.189.7-3.674.3731.3
TG (mg/dL)388.7490-101.3335.3361.7-26.4
  • Imaging response

The effect of statins on imaging manifestations of NASH was not assessed in this trial.

Atorvastatin versus fenofibrate versus the combination of the two interventions

We mainly focused on the comparison of atorvastatin versus fenofibrate and disregarded the combination group except for adverse events.

Primary outcomes
  • All-cause mortality

No all-cause mortality was reported.

  • Hepatic-related mortality

No hepatic-related mortality was reported.

  • Adverse events

Liver-related adverse events: transaminase activity greater than three times the upper limit of normal was reported in one patient treated with atorvastatin, who had the intervention subsequently discontinued. Three patients discontinued treatment due to myalgia and elevated serum creatine kinase activity, one from the atorvastatin group and two from the combination group.

Secondary outcomes
  • Histological response

The histological response was not studied in this trial.

  • Biochemical response

Baseline and post-treatment laboratory variables were assessed in this trial (Table 3). There were no statistically significant differences between any of the three intervention groups regarding the week 54 mean activities of aspartate aminotransferase, alanine aminotransferase, gamma-glutamyl transpeptidase, or alkaline phosphatase. The triglyceride levels seemed higher in the fenofibrate group compared with the atorvastatin group.

Table 3. Baseline and after treatment laboratory variables in Athyros et al trial
  1. ALT = alanine aminotransferase
    AST = aspartate aminotransferase
    ALP = alkaline phosphatases
    TG = triglycerides

 Statin (n = 63)Fenofibrate  (n = 62)
Before treatmentAfter treatmentMean decreaseBefore treatmentAfter treatmentMean decrease
ALT (U/L)543222523616
AST (U/L)382513392712
ALP (U/L)11075351087830
TG (mg/dL)203.8142.261.6194.9115.679.3
  • Imaging response

The effects of statins on improvement in ultrasonographic manifestations of NASH were assessed in combination with aminotransferase activity in this trial. At the end of the study the proportion of participants without evidence of NAFLD was higher in the atorvastatin group (42/63, 66%) versus the fenofibrate group (26/62, 42%) (P = 0.007 in Fisher's exact test). However, it was unclear if this was a post hoc analysis. No independent data were available for analysis of the ultrasonographic data without the biochemical findings.

Discussion

Although non-alcoholic fatty liver disease (NAFLD) is the most common cause of liver disease, no pharmacological agent has been established as definitive for its management. Statins, which are among the most widely prescribed drugs in the world, play a significant role in decreasing cardiovascular risk associated with non-alcoholic steatohepatitis (NASH), including the metabolic syndrome and type 2 diabetes mellitus. We have attempted to assess the effects of statins on all-cause mortality, liver–related mortality, liver histology, plasma liver enzyme activity, and ultrasonographic findings in patients with NAFLD or NASH.

Summary of main results

Two trials were eligible for inclusion in this systematic review. Both were judged as having high risk of bias. The primary outcomes (that is, all-cause mortality and liver-related morbidity) were not addressed in any of the included trials. The decrease in the mean level of serum aminotransferases was greater in participants who had received a statin compared with the control group, but this decrease did not reach statistical significance. There were no significant effects of simvastatin versus placebo on plasma activities of enzymes or on liver histology in NASH patients. However, this was only a pilot trial and hence the statistical power was modest. In the NAFLD trial, atorvastatin versus fenofibrate (a non-statin lipid-lowering drug) did not show significant differences regarding plasma liver enzyme activities or ultrasonographic findings, but when these two outcomes were combined into one atorvastatin seemed superior to fenofibrate.

Statins may cause myositis and rhabdomyolysis. Persistent (that is, longer than 12 weeks) elevations in aminotransferases (Kashani 2006), myalgias, myositis (that is, serum creatine kinase (CK) elevation to greater than 10 times the normal value in association with muscle symptoms) (Tobert 1988), and rhabdomyolysis with or without acute renal failure (Ballantyne 2003) were also considered adverse reactions. The proportions of adverse events were not statistically different between groups, but four patients out of a total of 124 patients on atorvastatin had their treatment discontinued due to adverse effects. No serious adverse events were reported in any of the groups.

In addition, we considered the Excluded studies when assessing the adverse effects of statins. Statin use in participants with the diagnosis of NASH seemed safe, and no adverse events were reported (Kiyici 2003; Gomez 2006; Lewis 2007; Tavakkoli 2009). Other excluded studies did not discuss adverse effects (Harlander 2001; Hatzitolios 2004; Rallidis 2004; Antonopoulos 2006).

Overall completeness and applicability of evidence

According to the findings of this systematic review, adequate evidence to clarify the role of statins in NAFLD and NASH is lacking. The trial that included NASH patients was very small. The trial that included NAFLD patients had no placebo control and assessed three intervention groups, which creates problems with the statistical analyses with multiplicity being a prominent problem. Furthermore, only non-validated surrogate outcomes were assessed (Gluud 2007). None of the trials assessed patient relevant outcomes like quality of life, cirrhosis development, liver-related morbidity, liver-related mortality, or all-cause mortality. So we are not able to recommend or refute statins for patients with NAFLD or NASH.

Patients with NAFLD or NASH often suffer from other conditions, for example raised serum cholesterol, or are at risk of cardiovascular diseases that may by themselves serve as an indication for statin use (Vale 2011; Taylor 2013).

We did not find that the existence of NAFLD or NASH should constitute a contraindication to statins that may seem indicated for other reasons. However, the evidence base is still small and further data may be needed.

Quality of the evidence

Based on The Oxford System Centre for Evidence Based Medicine (http://www.cebm.net/index.aspx) (Lomas 1991), we have summarised the level of evidence and the grade of recommendation for each outcome in Table 4.

Table 4. Level of evidence for each outcome
Outcome

Biochemical improvement

 

Radiological improvementNo improvement in histologySerious adverse effects
EvidenceLow-quality systematic reviewLow-quality randomised clinical trialSmall sample size in low-quality randomised clinical trial

Low-quality systematic review

 

Level of evidence2222
Grade of recommendationBBBB

Potential biases in the review process

Although during our work on this review we contacted the authors of the included trials to clarify some ambiguities in their methodology, we did not obtain clear answers. Meta-analyses for outcomes were not possible because the two included trials compared different entities, one trial had compared simvastatin versus placebo (Nelson 2009) and the other trial had compared atorvastatin versus fenofibrate (Athyros 2006). The second included trials compared three intervention groups, atorvastatin was considered the intervention group in our analysis, fenofibrate was considered the control group, and the third group received both atorvastatin and fenofibrate (data from this group were not used in our review apart from for adverse events) (Athyros 2006).

Agreements and disagreements with other studies or reviews

We found no significant effect of statins on serum activities of plasma liver enzymes. In a non-randomised clinical trial, Kiyici et al demonstrated significant decreases in AST, ALT, and ALP after six months of treatment with atorvastatin (Kiyici 2003). Similar results have been reported in other clinical studies, such as Hatzitolios 2004, Antonopoulos 2006, and Gomez 2006. Improvement in ultrasonographic evidence of NASH with statins, which was indicated in one of the included trials (Athyros 2006), has also been seen in two other clinical studies (Kiyici 2003; Antonopoulos 2006). In one of the included trials (Nelson 2009), no histological improvement was observed on liver biopsy after one year of treatment with statins. This finding contrasts with the findings of uncontrolled studies. Harlander et al, in their uncontrolled clinical study, reported significant improvement in the degree of liver steatosis as well as liver fibrosis in seven participants after one year of treatment with atorvastatin (Harlander 2001). The results of another uncontrolled clinical study support this finding (Rallidis 2004). Whether we are observing the usual overestimation of benefits in uncontrolled studies compared to more conservative or neutral findings in randomised clinical trials or if statins do not possess therapeutic benefits in NAFLD and NASH patients remains to be determined (Jakobsen 2013).

Use of simvastatin in participants with NASH did not seem to be accompanied by serious adverse effects, but the trial was very small. The use of atorvastatin in NAFLD patients did result in a few patients with likely adverse effects. These observations are in accordance with the findings of non-randomised controlled studies (Rallidis 2004; Ahmed 2006; Gomez 2006; Tavakkoli 2009).

Authors' conclusions

Implications for practice

Based on two trials with high risk of bias, it seems that statins may improve serum transferase levels as well as ultrasound findings. We cannot reach any conclusions about their safety as neither of the two trials reported on mortality or hepatic-related morbidity. Thus, it is not clear whether these effects of statins can lead to improved clinical outcomes. In the light of other beneficial effects of statins, especially in improving cardiovascular outcomes in conditions frequently associated with NASH (for example, hyperlipidaemia, diabetes mellitus, metabolic syndrome), their use in patients with non-alcoholic fatty liver disease or non-alcoholic steatohepatitis may be justified for other reasons.

Implications for research

Given the high risk of bias in the included trials for this systematic review and the fact that these studies have not addressed clinically important outcomes, further long-term trials with larger sample sizes, longer duration, and low risk of bias are needed to address these issues. Such trials ought to be designed according to the SPIRIT guidelines and reported according to the CONSORT guidelines.

Acknowledgements

Our special thanks goes to Dimitrinka Nikolova and Sarah Louise Klingenberg in The Cochrane Hepato-Biliary Group (CHBG) for their unlimited help in preparing this review.

Peer reviewers: Mona H Ismail, Saudi Arabia; Yusuf Yilmaz, Turkey; Raj Vuppalanchi, USA.
Contact editor: Christian Gluud, Denmark.

Data and analyses

Download statistical data

This review has no analyses.

Appendices

Appendix 1. Search strategies

Database Time Span Search Strategy
Cochrane Hepato-Biliary Group Controlled Trials RegisterMarch 2013(statin* OR Lovastatin OR Atorvastatin OR Simvastatin OR Parvovastatin OR Rosuvastatin OR Fluvastatin OR Pravastatin OR 'reductase inhibitor*') AND (liver* OR hepat* OR steatohepat* OR NAFL* OR NASH*)
Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library (Wiley)Issue 1, 2013#1 Statins /explode all trees (MeSH)
#2 HMG-CoA reductase inhibitors/ explode all trees (MeSH)
#3 Hydroxymethylglutaryl-CoA reductase inhibitors / explode all trees (MeSH)
#4 Lovastatin or Pravastatin or Rosuvastatin or Atorvastatin or Simvastatin
#5 (#1 or #2 or #3 or #4)
#6 Fatty liver / explode all trees (MeSH)
#7 Liver steatosis / explode all trees (MeSH)
#8 (#6 or #7)
#9 (#5 AND #8)
MEDLINE (OvidSP)1946 to March 2013

1. exp Hydroxymethylglutaryl-CoA Reductase Inhibitors/ or exp Lovastatin/ or exp Pravastatin/

2. (statin* or Lovastatin or Atorvastatin or Simvastatin or Parvovastatin or Rosuvastatin or Fluvastatin or Pravastatin or 'reductase inhibitor*').mp. [mp=title, original title, abstract, name of substance word, subject heading word, unique identifier]

3. 1 or 2

4. exp Fatty Liver/

5. ((fatty and (liver* or hepat*)) or steatohepat* or NAFL* or NASH*).mp. [mp=title, original title, abstract, name of substance word, subject heading word, unique identifier]

6. 4 or 5

7. 3 and 6

8. (random* or blind* or placebo* or meta-analysis).mp. [mp=title, original title, abstract, name of substance word, subject heading word, unique identifier]

9. 7 and 8

EMBASE (OvidSP)1974 to March 2013

1. exp hydroxymethylglutaryl coenzyme A reductase inhibitor/

2. (statin* or Lovastatin or Atorvastatin or Simvastatin or Parvovastatin or Rosuvastatin or Fluvastatin or Pravastatin or 'reductase inhibitor*').mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer name]

3. 1 or 2

4. exp fatty liver/

5. ((fatty and (liver* or hepat*)) or steatohepat* or NAFL* or NASH*).mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer name]

6. 4 or 5

7. 3 and 6

8. (random* or blind* or placebo* or meta-analysis).mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer name]

9. 7 and 8

Science Citation Index Expanded (http://apps.isiknowledge.com)1900 to March 2013

#1 TS=(statins OR HMG-CoA reductase inhibitors OR Hydroxymethylglutaryl-CoA reductase inhibitors OR Lovastatin OR Pravastatin OR Rosuvastatin OR Atorvastatin OR Simvastatin)
#2 TS=(Fatty liver OR Liver steatosis)
#3 #2 AND #1
#4 TS=(random* OR blind* OR placebo* OR meta-analysis)

#5 #4 AND #3

Contributions of authors

Formulating the research question: Reza Malekzadeh (RM), Layli Eslami (LE), and Siavosh Nasseri-Moghaddam (SNM).
Writing up the protocol: LE with comments from SNM and Shahin Merat (SM).
Searching: Cochrane Hepato-Biliary Group (CHBG).
Selecting and reviewing eligible studies: LE and SNM.
Performing methodological assessment and extraction of data: LE.
Providing data entry and management: LE.
Performing data analysis: LE.
Preparing the manuscript: LE, Hermineh Aramin (HA), and SNM.

Declarations of interest

None known.

Sources of support

Internal sources

  • Tehran University of Medical Sciences (TUMS), Iran.

External sources

  • No sources of support supplied

Differences between protocol and review

As the result of recent changes in Cochrane requirements, the order of outcomes in this review has been changed so that they reflect the needs of consumers.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Athyros 2006

MethodsRandomised clinical trial with three parallel groups
Participants

Participants were free of diabetes mellitus (DM) (ie, fasting glucose levels < 7 mmol/L; 126 mg/dL) and cardiovascular disease (diagnosed on the basis of personal history, clinical examination findings, and non-invasive methods).

The inclusion criteria were (a) the presence of the metabolic syndrome (MetS) (NCEP ATP III definition) (NCEP ATP III report), (b) low-density lipoprotein cholesterol (LDL-C) > 3.4 mmol/L (130 mg/dL), (c) ultrasonographic evidence of fatty liver, and (d) elevated serum aspartate aminotransferase (AST) and/or alanine aminotransferase (ALT) activity.

Other causes of liver disease were excluded.

Interventions

189 participants were randomly allocated to atorvastatin 20 mg/day (n = 63) versus micronised fenofibrate 200 mg/day (n = 62) versus both drugs (n = 61). All participants had both biochemical and ultrasonographic evidence of NAFLD at baseline.

All participants received:

  • lifestyle advice, including exercise (walking for at least 30 min, 5 days a week, or equivalent exercise) and a low-fat and low-calorie diet (NCEP ATP III report); and

  • treatment for hypertension (mainly inhibitors of the renin-angiotensin system), impaired fasting glucose (metformin), obesity (orlistat), and dyslipidaemia (randomly allocated to atorvastatin 20 mg/day (n = 63) or micronised fenofibrate 200 mg/day (n = 62) or both drugs (n = 61)).

Outcomes

At the end of treatment, 67% of participants taking atorvastatin, 42% taking fenofibrate, and 70% taking combination treatment no longer had biochemical plus ultrasonographic evidence of NAFLD (P < 0.05 versus baseline for all comparisons).

The percentage of participants who no longer had evidence of NAFLD was significantly higher (P < 0.009) in the atorvastatin and combination groups compared with the fenofibrate group.

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskComputer-generated random numbers were used for generation of allocation sequence.
Allocation concealment (selection bias)Unclear riskThe method of allocation concealment was unclear.
Blinding of participants and personnel (performance bias)
All outcomes
High riskIt was an open-labelled trial.
Blinding of outcome assessment (detection bias)
All outcomes
High riskIt was an open-labelled trial.
Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcomes, dropouts, and withdrawals were pointed out precisely in the trial.
Selective reporting (reporting bias)Low riskAll relevant outcomes were available in the trials. No reporting bias was detected.
Other biasUnclear riskIt is not clear whether sample size calculations were performed.

Nelson 2009

MethodsDouble-blind randomised placebo-controlled trial with two parallel groups.
ParticipantsSixteen adults, 18 years of age or older, with documented NASH based on liver biopsy using criteria established by Brunt et al (Brunt 2005), were included. All participants had compensated liver disease with haemoglobin values of 12 g/dL or greater in females and 13 g/dL or greater in males; a white blood cell count > 3000/mm3, neutrophil count > 1500/mm3, platelets > 70,000/mm3, albumin > 3.0 g/dL, in addition to normal total bilirubin, prothrombin time, and International normalised ratio. Other requirements included serum creatinine < 1.4 mg/dL and elevated serum lipid panel manifested by total cholesterol > 200 mg/dL, LDL > 130 mg/dL, or TGs > 200 mg/dL.
InterventionsParticipants were randomly assigned to receive simvastatin 40 mg (n = 10) versus placebo (n = 6) once daily for 12 months.
OutcomesFourteen participants completed the trial, and 10 underwent 1 year repeated liver biopsy. Although a 26% reduction in low-density lipoprotein was seen in the simvastatin group compared with the placebo group, no statistically significant improvement in serum aminotransferases, hepatic steatosis, necro-inflammatory activity, or stage of fibrosis was noted within or between groups.
Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskThe method was not reported.
Allocation concealment (selection bias)Unclear riskThe method was not reported.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskThe trial was described as double-blind, but it was not reported who was blinded.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskThe trial was described as double-blind, but it was not reported who was blinded.
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskReported data about the withdrawals were unclear.
Selective reporting (reporting bias)High riskNot all outcomes were reported.
Other biasUnclear riskIt is not clear whether sample size calculations were performed.

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
Antonopoulos 2006Inappropriate study design (uncontrolled clinical trial).
Athyros 2010A few participants in the control group received statin.
Carnelutti 2012Only an abstract of the study was available.
Gomez 2006Inappropriate study design (uncontrolled clinical trial).
Han 2012Both intervention and control groups received statin.
Harlander 2001Inappropriate study design (uncontrolled clinical trial).
Hatzitolios 2004Inappropriate study design (controlled before and after clinical trial).
Kiyici 2003Inappropriate study design (uncontrolled clinical trial).
Lewis 2007Inappropriate participants (patients with chronic liver disease).
Rallidis 2004Inappropriate study design (uncontrolled clinical trial).
Tavakkoli 2009Inappropriate study design (uncontrolled clinical trial).

Ancillary