Description of the condition
Liver failure is characterised by hepatic encephalopathy, jaundice, and coagulopathy. The liver failure can lead to progressive multi-organ failure and death (Rahman 1999; Riordan 1999). Liver failure can be divided into two types: acute liver failure and acute-on-chronic liver failure (Khuroo 2004). Acute liver failure is defined as sudden or severe hepatic injury with coagulation abnormality and any degree of encephalopathy in an individual without previous liver disease. The duration of acute liver failure is usually less than 26 weeks (Polson 2005; Bernal 2010). Acute liver failure can be divided into severe acute liver failure and fulminant hepatic failure according to manifestations of liver injury (Bernuau 1986). Severe acute liver failure is defined as acute liver failure with 50% or more decrease in coagulation factors manufactured by the liver, but without hepatic encephalopathy. Fulminant hepatic failure is acute liver failure that is complicated by hepatic encephalopathy within eight weeks after the onset of illness. Subfulminant liver failure is applied to patients in whom hepatic encephalopathy develops into a longer illness lasting up to 26 weeks (Lee 1993; Patton 2012).
Acute-on-chronic liver failure refers to an acute deterioration of liver function in patients with previously well-compensated chronic liver disease over a period of two to four weeks, usually due to the effects of precipitating events, such as surgery, sepsis, or upper gastrointestinal bleeding (Jalan 2002; Sen 2002).
The causes of liver failure vary due to geographical environment, and the causes depend on the types of drug use and the prevalence of hepatotropic or non-hepatotropic virus infections (Acharya 2002; Ichai 2008). In the western countries, major causes of liver failure are alcohol and hepatotoxic drug-induced liver injuries. On the other hand, in eastern countries, such as China, infectious aetiologies predominate. Most patients are infected with hepatitis A, B, and E viruses.
Studies from developed countries report that acute liver failure is rare. The incidence of acute liver failure in developed countries is one to six people per million per year (Brandsaeter 2002; Bower 2007; Bernal 2010). However, the prognosis is poor, with an overall mortality of 33% and a transplant rate of 25% in the United States (Stravitz 2008). The prognosis is mainly affected by aetiology and symptoms. Liver failure mortality caused by acetaminophen overdose, hepatitis A, and ischaemia is 40%, whereas the mortality reaches 75% if the aetiology is hepatitis B or indeterminate cases. If complicated by hepatic encephalopathy, the mortality can be as high as 90% to 95%, despite advanced intensive care medical treatment (Lee 2008; Wang 2008). Liver failure is a heavy burden to the society because of high mortality, use of resources, and cost.
Currently, orthotopic liver transplantation (OLT) remains the gold standard treatment in the management of liver failure and has significantly increased survival (Polson 2005; Di Campli 2007). Unfortunately, due to the serious shortfall of donors and the expense, most patients die while waiting for organ transplantation. Effective and convenient management has been proposed in order to overcome these difficulties.
Description of the intervention
Granulocyte-colony stimulating factor (G-CSF) is a blood growth factor that stimulates the mobilisation and differentiation of bone marrow stem cell (BMSC) (Welte 1985). It has a short half-life and its elimination is primarily mediated by the G-CSF-receptor and renal clearance. G-CSF is widely used in cancer and haematological malignancy patients who have received chemotherapy as it can promote the production of neutrophils (Frampton 1994; Calhoun 2000). Current evidence suggest that it has a significant beneficial effect on hepatic failure in humans and animals (Theocharis 2003; Di Campli 2007; Garg 2012). G-CSF inhibits progressive hepatocyte necrosis and promotes liver regeneration.
Theocharis et al reported that G-CSF could ameliorate the histologically evident liver injury and reinforce the proliferate capacity of the hepatocytes in an animal model (Theocharis 2003). Furthermore, survival was statistically significantly improved in the G-CSF-treated group. Only 25% of rats with fulminant hepatic failure were alive in the control group at the end of the trial, while in G-CSF-treated group, the survival was increased to 60% (P < 0.01) (Theocharis 2003). Other experimental data have also demonstrated a beneficial effect of G-CSF administration on survival, liver synthetic function, and hepatic regenerative capacity in animal models of end-stage liver disease (Yannaki 2005; Liu 2006; Xu 2006). Two controlled clinical trials showed that treatment with G-CSF could induce CD34 mobilisation, increased the number of peripheral circulation stem cells, and changed some cytokine values in patients with severe liver dysfunction patients (Di Campli 2007; Spahr 2008).
The side effects of G-CSF therapy are minor: back pain, headache, and nausea are the most common. To our knowledge, there have been no reports of significant adverse events.
The most appropriate regimen of G-CSF administration is not clear; however, a dose of 10 to 15 µg/kg/day is the dose most often used in trials. Di Campli reported that high dosages of G-CSF (15 µg/kg/day) were associated with higher CD34+ cells mobilisation, and they did not increase adverse events (Di Campli 2007).
How the intervention might work
The effect of G-CSF is mainly on mobilising bone marrow cells (BMC) into the injured liver to differentiate into hepatocytes (Korbling 2002; Yannaki 2005). G-CSF has also immunomodulatory function (Saito 2002).
The mechanisms of G-CSF on liver failure are under investigation and have probably two aspects. On one hand, G-CSF regulates the cytokine gradient and adhesion molecule, decreases the release of cell-derived factor (SDF-1), and highly promotes the expression of CXCR4 on haematopoietic progenitor cells (HPC). The down-regulation of adhesion molecules and the interaction of SDF-1/CXCR4 contribute to the BMC mobilisation (Petit 2002; Gazitt 2004; Di Campli 2007). On the other hand, G-CSF can inhibit the release of pro-inflammatory cytokines such as interleukin (IL)-12, granulocyte macrophage colony-stimulating factor (GM-CSF), tumour necrosis factor (TNF)-α, and interferon (IFN)-γ. G-CSF can increase the antiinflammatory cytokine release capacity such as soluble tumour necrosis factor receptor (sTNF-Rs), interleukin-1 receptor antagonist (IL-Ira), and IL-10 which can ease the secondary injury of severe hepatitis (Hartung 1995; Saito 2002).
Why it is important to do this review
G-CSF may have a beneficial effect on patients with liver failure, but there are only a few small clinical trials, and its benefits and harms have not been comprehensively reviewed. This is why we have planned to carry out this systematic review. We have been unable to identify previous meta-analyses or systematic reviews on the topic.
To assess the benefits and harms of G-CSF and to determine the optimum regimen of G-CSF for acute and acute-on-chronic liver failure.
Criteria for considering studies for this review
Types of studies
We will include all randomised clinical trials irrespective of language of publication, publication status, or blinding. Quasi-randomised studies and observational studies will be excluded for the report of benefit but will be included for assessment of harm.
Types of participants
Patients with acute liver failure or acute-on-chronic liver failure, adults, regardless of sex, nationality, aetiology, or the severity of the liver disease.
Types of interventions
1. G-CSF of any type, dose, and duration versus placebo or no intervention.
2. G-CSF of any type, dose, and duration versus standard medical therapy.
2. G-CSF of any type, dose, and duration versus extracorporeal liver support systems.
Co-interventions will be permitted if used equally in all intervention groups of the trial.
Types of outcome measures
1. All-cause mortality: number of deaths irrespective of the cause.
2. Liver function (evaluated by the Child-Turcotte-Pugh (CTP) or the model for end-stage liver disease (MELD) scores).
3. Adverse events: number of patients with different types of adverse events. Serious adverse events are defined according to the International Conference on Harmonisation-Good Clinical Practice guidelines (ICH-GCP 1997) as any untoward medical occurrence that resulted in death or significant disability, is life-threatening, or any important medical event which has jeopardised the patient or requires intervention to prevent it. All other adverse events will be considered non-serious.
1. Quality of life.
2. Number of days in intensive care unit (ICU).
3. Number of days in hospital.
4. Development of new-onset complications such as hepatic encephalopathy, hepatorenal syndrome, gastrointestinal bleeding, and ascites.
Search methods for identification of studies
We will search The Cochrane Hepato-Biliary Group Controlled Trials Register (Gluud 2012), the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library, MEDLINE, EMBASE, and Science Citation Index Expanded (Royle 2003). We will also search the Chinese Biomedical Literature Database (CBM). The preliminary search strategies with the expected time span of the searches are described in detail in Appendix 1.
Searching other resources
We will also search the references of the identified trials to identify further relevant trials and contact pharmaceutical companies that produce G-CSF to obtain data from unpublished randomised clinical trials.
ClinicalTrials.gov register (ClinicalTrials.gov), Chinese Clinical Trial Register (http://www.chictr.org) and the metaRegister of Controlled Trials (www.controlled-trials.com/mrct/) will also be searched for unpublished and ongoing trials.
Data collection and analysis
Selection of studies
One review author (XM Chang) will develop the search strategies. Two authors (JC Peng and LJ Zhang) will independently select randomised clinical trials, quasi-randomised studies, and observational studies that come up with the searches by screening the titles and abstracts to be included in the review according to the prespecified selection criteria. We will also list the excluded trials with the reasons for exclusion. We will solve disagreements by discussion or arbitration by the third author (XM Chang). We will contact primary authors of the included trials for clarification if information in the publications is unclear or missing.
Data extraction and management
Two review authors (JC Peng and LJ Zhang) will independently extract the prespecified characteristics of all included trials into a data extraction sheet, and check the extracted data. Disagreements will be resolved by a third author (QM Yang). For cross-over randomised clinical trials, the review authors will only include data from the first period. Review authors will extract the following information: general information: publication title, primary author, publication status (published: journal, year and language of publication; unpublished: year in which the trial was conducted), country in which the trial was conducted, study sponsor; baseline characteristics; characteristics of participants: sample size, median age, sex, aetiology of the liver failure, liver function according Child-Pugh score, dropouts and losses; interventions: dose, frequency, routes, and duration of administration; methods: inclusion criteria, study sign, method of randomisation, allocation procedure, blinding( patients, caregivers and outcome assessors), follow-up, 'intention-to treat' analysis; outcomes and result: primary and secondary outcomes; changes in outcome reporting.
We will contact the primary authors of the trials if the relevant data are missing.
Assessment of risk of bias in included studies
Two authors (JC Peng and QM Yang) will assess the methodological quality, that is, risk of bias of each included trial independently, based on the instructions given in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) and the Cochrane Hepato-Biliary Group Module (Gluud 2012). A consensus method will be used if there are differences in the assessment of risk of bias in the trials. If information is not available about the study methods, we will contact the primary authors. We will evaluate bias risk of the trials using the domains given below (Schulz 1995; Moher 1998; Kjaergard 2001; Wood 2008; Savovic 2012; Savovic 2012a).
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.
- Low risk of bias: the 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, and 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 the participants.
Blinding of participants, personnel, and outcome assessors
- Low risk of bias: blinding was performed adequately, or the assessment of outcomes was not likely to be influenced by lack of blinding.
- Uncertain risk of bias: there was insufficient information to assess whether blinding was likely to induce bias on the results.
- High risk of bias: no blinding or incomplete blinding, and the assessment of outcomes were 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, have been employed to handle missing data.
- Uncertain risk of bias: there was insufficient information to assess whether missing data in combination with the method used to handle missing data were likely to induce bias on the results.
- High risk of bias: the results were likely to be biased due to missing data.
Selective outcome reporting
- Low risk of bias: all outcomes were pre-defined and reported, or all clinically relevant and reasonably expected outcomes were reported.
- Uncertain risk of bias: it is unclear whether all pre-defined 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.
For a trial to be assessed with low risk of bias in the selective outcome reporting domain, the trial should have been registered either on the www.clinicaltrials.gov web site or a similar register, or there should be a protocol, eg, published in a paper journal. In the case when the trial was run and published in the years when trial registration was not required, we will try to find and scrutinize all publications that are reporting on the trial to identify the trial objectives and outcomes. If usable data on all outcomes specified in the trial objectives are provided in the publications results section, then the trial can be considered low risk of bias trial in the Selective outcome reporting domain.
- Low risk of bias: the trial appears to be free of industry sponsorship or other kind of for-profit support that may manipulate the trial design, conductance, or results of the trial.
- Uncertain risk of bias: the trial may or may not be free of for-profit bias as no information on clinical trial support or sponsorship is provided.
- High risk of bias: the trial is sponsored by the industry or has received other kind of for-profit support.
- Low risk of bias, the study appears to be free of other sources of bias, Such as baseline imbalance.
- Uncertain risk of bias, insufficient information to assess whether the risk of bias exits.
- High risk of bias, there is at least one important risk of bias.
Trials having one or more domains with high or uncertain risk of bias will be grouped as trials with high risk of bias trials. All other trials will fall into the group of trials with low risk of bias.
Measures of treatment effect
We will define measures of treatment effects according to the recommendations of The Cochrane Collaboration (Higgins 2011).
For dichotomous variables, we will present the results as relative risk (RR) with 95% confidence interval (CI). For continuous variables, we will present the results as mean difference (MD) or standardised mean difference (SMD) with 95% CI.
Unit of analysis issues
The individual intervention group of the randomised trials.
Dealing with missing data
The following methods will be used to deal with missing data:
We will attempt to contact the original investigators for the missing data that are needed for the meta-analyses.
We will perform intention-to-treat analysis (ITT) whenever possible. We will use sensitivity analysis with missing imputation based on the worst-case/best-case scenarios (Higgins 2011).
Assessment of heterogeneity
Heterogeneity among trials will be explored using a Chi
Assessment of reporting biases
We will assess reporting biases by using funnel plots if at least ten trials are included.
We will use the software Review Manager 5 (RevMan 2012) provided by the Cochrane Collaboration for statistical analysis. All analyses will be performed according to the intention-to-treat principle. Both random-effects model and fixed-effect model will be used. If the results of the two analyses lead to the same conclusion, only the results of the fixed-effect model analysis will be reported. If I
Trial sequential analysis
We plan to perform trial sequential analysis (TSA) for the first three primary outcomes (CTU 2011) (http://ctu.dk/tsa). Trial sequential analysis aims to reduce the risk of random error in the setting of repetitive testing of accumulating data, thereby improving the reliability of conclusions (Brok 2008; Wetterslev 2008; Brok 2009; Thorlund 2009; Wetterslev 2009). Where possible, we will examine apparently significant beneficial and harmful intervention effects with trial sequential analyses (CTU 2011; Thorlund 2011) in order to evaluate if these apparent effects could be caused by random error (play of chance) (Brok 2008; Wetterslev 2008; Brok 2009; Thorlund 2009; Wetterslev 2009; Thorlund 2010). We plan to estimate the diversity-adjusted required information size based upon the occurrence of the outcome in the control intervention group, a relative risk reduction of 20%, and as observed in trials with low risk of bias, an alpha of 5%, a beta of 20%, and the diversity observed in the meta-analysis (Wetterslev 2008; Wetterslev 2009; Thorlund 2011).
Subgroup analysis and investigation of heterogeneity
We will perform the following subgroup analyses based on:
- risk of bias (trials with low bias risk compared to trials with high bias risk).
- type of liver failure( acute liver failure compared to acute-on-chronic liver failure).
- G-CSF dose.
- duration of treatment.
- trials with co-interventions compared to the trials without co-interventions.
In addition to the sensitivity analyses described in 'Dealing with missing data', we plan to perform sensitivity analyses to explore whether conclusions are robust by excluding trials with high risk of bias.
Summary of Findings Tables
We will also assess the quality of evidence at the outcome level across trials using GRADEpro. If possible, we will prepare summary of findings tables presenting the primary outcomes of our review (GRADE table) (http://ims.cochrane.org/revman/other-resources/gradepro).
We thank The Cochrane Hepato-Biliary Group for their guidance and support, especially Dimitrinka Nikolova. Thanks to Jin Hui Tian and Lun Li for giving useful suggestions. Thanks to Ronald Koretz, Shiv Kumar Sarin, and Vishal Garg for helpful comments.
Peer Reviewers: Shiv Kumar Sarin, India; Vishal Garg, India.
Contact Editor: Ronald Koretz, USA.
Appendix 1. Appendix 1
Contributions of authors
Formulated the idea for the review: Jucong Peng, Qimei Yang.
Drafted the protocol: Jucong Peng, Xinming Chang.
Developed the search strategy: Xinming Chang.
Revised the protocol: Jucong Peng, LingJuan Zhang.
Provided clinical advice: Xinming Chang.
Provided technical support: LingJuan Zhang, Qimei Yang.
Declarations of interest
Sources of support
- Medical College of Xi'an Jiaotong University, China.
- No sources of support supplied