Colchicine for primary biliary cirrhosis

Abstract

Background

Colchicine has been used for patients with primary biliary cirrhosis because of its immunomodulatory and antifibrotic potential. The therapeutical responses to colchicine in randomised clinical trials were inconsistent.

Objectives

To evaluate the beneficial and harmful effects of colchicine in patients with primary biliary cirrhosis.

Search methods

We identified trials through electronic searches of The Cochrane Hepato‐Biliary Group Controlled Trials Register, The Cochrane Central Register of Controlled Trials on The Cochrane Library, MEDLINE, EMBASE (September 2003), and manual searches of bibliographies. We contacted authors of trials and pharmaceutical companies.

Selection criteria

Randomised clinical trials comparing colchicine with any kind of control therapy were included irrespective of language, year of publication, and publication status.

Data collection and analysis

The primary outcomes were the number of deaths and the number of death and/or patients who underwent liver transplantation. Dichotomous outcomes were reported as relative risk (RR) with 95% confidence interval (CI). We examined intervention effects by using both a fixed effect model and a random effects model. Heterogeneity was investigated by subgroup analyses and sensitivity analyses.

Main results

Eleven randomised clinical trials involving 716 patients with primary biliary cirrhosis fulfilled the inclusion criteria. No significant differences were detected between colchicine and placebo/no intervention on the number of deaths (RR 1.21, 95% CI 0.71 to 2.06), the number of deaths and/or patients who underwent liver transplantation (RR 1.00, 95% CI 0.67 to 1.49), liver complications, liver biochemical variables, liver histological measurements, and adverse events. Trial methodology was generally low and some trials had high drop‐out rate. A best‐worst‐case‐scenario analysis showed no significant effect of colchicine on mortality (RR 0.59, 95%CI 0.30 to 1.15), while a worst‐best‐case‐scenario analysis showed a significant detrimental effect of colchicine on mortality (RR 2.28, 95% CI 1.17 to 4.44). Colchicine significantly decreased the number of patients without improvement of pruritus (RR 0.75, 95% CI 0.65 to 0.87). However, this estimate was based on only 156 patients from three trials. The effect of the combined treatment with ursodeoxycholic acid was not significantly different from that of colchicine alone.

Authors' conclusions

We did not find evidence either to support or refute the use of colchicine for patients with primary biliary cirrhosis. As we are not able to exclude a detrimental effect of colchicine, we suggest that it is only used in randomised clinical trials.

Plain language summary

No convincing evidence either to support or refute the use of colchicine for patients with primary biliary cirrhosis

Primary biliary cirrhosis is a rare, chronic liver disease of unknown etiology. Colchicine, a plant alkaloid, has been used to treat patients with primary biliary cirrhosis and was tested in randomised clinical trials. When all identified trials were combined, colchicine appeared to be not significantly different from placebo/no intervention in respect to mortality, mortality and/or patients who underwent liver transplantation, liver complications, liver biochemistry, liver histology, and the occurrences of adverse events. Colchicine may reduce pruritus, but this finding may be due to bias. The addition of ursodeoxycholic acid did not significantly influence the effect of colchicine.

Background

Primary biliary cirrhosis is an uncommon chronic progressive liver disease of unknown etiology. Ninety per cent of patients with primary biliary cirrhosis are females and the majority are diagnosed after the age of 40 years (James 1981). The earlier description was published in 1949 (MacMahon 1949). Later, Ahrens and co‐workers comprehensively described primary biliary cirrhosis in 1950 (Ahrens 1950). A progressive granulomatous hepatitis destroys small septal and interlobular bile ducts, eventually leading to cholestasis and biliary cirrhosis. Primary biliary cirrhosis is classically defined on the basis of the triad: antimitochondrial antibodies, which are found in over 95% of patients with primary biliary cirrhosis (Fregeau 1989; Lacerda 1995; Invernizzi 1997; Turchany 1997; Mattalia 1998); abnormal liver function tests that are typically cholestatic (with raised activity of alkaline phosphatases being the most frequently seen abnormality); and characteristic liver histological changes (Scheuer 1967) in the absence of extrahepatic biliary obstruction (Kaplan 1996). Patients may either be diagnosed during a symptomatic phase (the common symptoms being pruritus, fatigue, jaundice, liver enlargement, signs of portal hypertension, sicca complex, and scleroderma‐like lesions), in which case survival is significantly decreased, or during an asymptomatic phase of the disease, which has a relatively favourable prognosis (Beswick 1985; Balasubramaniam 1990). However, 40 to 100% of these patients will subsequently develop symptoms of primary biliary cirrhosis (Nyberg 1989; Metcalf 1996; Prince 2000).

Although the etiology remains unknown, primary biliary cirrhosis is in many respects analogous to the graft‐versus‐host syndrome in which the immune system is sensitised to foreign proteins. Most primary biliary cirrhosis patients have increased class II human leukocyte antigen (HLA) histocompatibility expression on bile duct cells (Ballarardini 1984; Van den Oord 1986), and the bile duct epithelium is infiltrated by cytotoxic T‐cells (Yamada 1986). Lacrimal and pancreatic glands, for example, with a high concentration of HLA class II antigens on their epithelium, may be involved in the disease process (Epstein 1982).

Patients with primary biliary cirrhosis have been subjected to many drugs. Ursodeoxycholic acid, a bile acid, is the most extensively used drug in these patients (Verma 1999). However, a recent systematic Cochrane Review was unable to demonstrate any significant effect of ursodeoxycholic acid on mortality or liver transplantation (Gluud 2002). Over the years, a number of other drugs have been evaluated for primary biliary cirrhosis. Earlier attempts to treat primary biliary cirrhosis using immunomodulatory and other agents such as azathioprine (Heathcote 1976; Christensen 1985), prednisolone (Mitchison 1992), chlorambucil (Hoofnagle 1986), cyclosporine (Wiesner 1990), D‐penicillamine (Epstein 1981; Matloff 1982; Dickson 1985; Neuberger 1985), methotrexate (Kaplan 1991; Lindor 1995), or colchicine have resulted in clinical effects that have not led to widespread acceptance of these drugs for primary biliary cirrhosis patients (Kaplan 1994).

Colchicine is a plant alkaloid. It is effective against gouty arthritis and other forms of rheumatic diseases (rheumatoid arthritis, familial Mediterranean fever, Bechet's disease, etc.) (Ben‐Chetrit 1998). The basis for effect of colchicine is inhibition of the migration of granulocytes into inflamed areas and decreased metabolic and phagocytic activity of granulocytes. Further, colchicine is an anti‐mitotic (Shi 1998) and anti‐fibrotic agent. Colchicine retards the microtubule mediated transport of procollagen (Ehrlich 1972) and enhances collagenase activity (Harris 1971).

Colchicine has been used for primary biliary cirrhosis patients because of its immunomodulatory and antifibrotic potential. Colchicine has been reported to slow the rate of progression of primary biliary cirrhosis (Kaplan 1997) and to produce improvements in liver function tests and immunoglobulin levels (Warnes 1987; Vuoristo 1995; Kaplan 1999). However, colchicine does not affect clinical symptoms or liver histology (Kaplan 1986). The effect of combination therapy with colchicine and ursodeoxycholic acid in patients with primary biliary cirrhosis has been reported, but the results have been conflicting (Shibata 1992; Ikeda 1996; Poupon 1996; Almasio 2000; Battezzati 2001). We have been unable to identify meta‐analyses or systematic reviews on the beneficial and harmful effects of colchicine for primary biliary cirrhosis patients.

Objectives

The objectives were to assess the beneficial and harmful effects of colchicine for patients with primary biliary cirrhosis.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised clinical trials irrespective of language, year of publication, and publication status. We excluded studies using quasi‐randomisation (e.g., allocation by date of birth).

Types of participants

Patients with primary biliary cirrhosis, i.e., patients having at least two of the following: elevated serum activity of alkaline phosphatases (or other markers of intrahepatic cholestasis), and/or a positive result for serum mitochondrial antibody, and/or liver biopsy findings diagnostic for or compatible with primary biliary cirrhosis.

Types of interventions

Administration of any dose of colchicine versus placebo or no intervention or other drugs. Co‐interventions were allowed as long as all intervention arms of the randomised clinical trial received similar co‐interventions. Therefore, we analysed the following comparisons: 
 1) Colchicine versus placebo/no intervention (monotherapy). 
 2) Colchicine plus ursodeoxycholic acid versus placebo/no intervention plus ursodeoxycholic acid (combination therapy). 
 3) Colchicine versus other drugs.

Types of outcome measures

Primary outcome measures were:

• Number of deaths.

• Number of deaths and/or patients who underwent liver transplantation.

Secondary outcome measures were:

• Number of patients who underwent liver transplantation.

• Pruritus: number of patients without improvement of pruritus and/or pruritus score.

• Fatigue: number of patients without improvement of fatigue and/or fatigue score.

• Incidence of complications: number of patients developing variceal bleeding, ascites, hepatic encephalopathy, jaundice, or hepato‐renal syndrome.

• Liver biochemistry: serum (s‐)bilirubin; s‐alkaline phosphatases; s‐gamma‐glutamyltransferase; s‐aspartate aminotransferase; s‐alanine aminotransferase; s‐albumin; s‐cholesterol (total); plasma immunoglobulin M.

• Liver biopsy findings: deterioration of liver histological stage or score.

• Quality of life: broad nature of a concept that includes physical functioning (ability to carry out activities of daily living such as self‐care), psychological functioning (emotional and mental well‐being), social functioning (relationships with others and participation in social activities), and perception of health, pain, and overall satisfaction with life.

• Adverse events. The adverse events are defined as any untoward medical occurrences in patients in either of the two arms of the included randomised clinical trials, which did not necessarily have a causal relationship with the treatment, but did, however, result in a dose reduction, discontinuation of treatment, or registration of the event as an adverse event/side effect. The adverse events are subdivided into non‐serious and serious, according to the ICH‐GCP guidelines (ICH‐GCP 1997). A serious adverse event is any event that leads to death, is life‐threatening, requires inpatient hospitalisation or prolongation of existing hospitalisation, results in persistent or significant disability or congenital anomaly/birth defect, or any important medical event, which may jeopardize the patient or requires intervention to prevent it.

• Health economics: the estimated costs connected with the interventions are weighed against any possible health gains.

Search methods for identification of studies

Relevant randomised clinical trials were identified by searching The Cochrane Hepato‐Biliary Group Controlled Trials Register (September 2003), The Cochrane Central Register of Controlled Trials on The Cochrane Library (Issue 3, 2003), MEDLINE (January 1966 to September 2003), and EMBASE (January 1980 to September 2003). See 'Appendix 1' for the search strategies that were applied to the individual electronic databases.

Further trials were identified by reading the reference lists of the identified studies. We wrote to the principal authors of the identified randomised clinical trials and to researchers active in the field to inquire about additional randomised clinical trials they might know of. We also wrote to the pharmaceutical companies that sponsored colchicine in identified trials to obtain unidentified or unpublished randomised clinical trials.

Data collection and analysis

The review was performed following the published protocol (Gong 2003) and the recommendations given by the Cochrane Reviewers' Handbook (Clarke 2003).

Trials selection 
 Identified trials were listed and two contributors (YG and CG) independently evaluated whether the trials fulfilled the inclusion criteria. Excluded trials were listed with the reasons for exclusion. Disagreements were resolved by discussion.

Data extraction 
 YG extracted the data and CG validated the data extraction. Disagreements were solved by discussion. YG wrote to the authors of all the included trials on colchicine for primary biliary cirrhosis and asked them to specify data, had they not been reported clearly in the articles.

Assessment of methodological quality of included trials 
 The methodological quality of the randomised clinical trials was assessed using four components (Schulz 1995; Moher 1998; Kjaergard 2001):

Generation of the allocation sequence

• Adequate, if the allocation sequence was generated by a computer or random number table. Drawing of lots, tossing of a coin, shuffling of cards, or throwing dice will be considered as adequate if a person who was not otherwise involved in the recruitment of participants performed the procedure;

• Unclear, if the trial was described as randomised, but the method used for the allocation sequence generation was not described;

• Inadequate, if a system involving dates, names, or admittance numbers were used for the allocation of patients. These studies are known as quasi‐randomised and were excluded from the present review.

Allocation concealment

• Adequate, if the allocation of patients involved a central independent unit, on‐site locked computer, identically appearing numbered drug bottles or containers prepared by an independent pharmacist or investigator, or sealed envelopes;

• Unclear, if the trial was described as randomised, but the method used to conceal the allocation was not described;

• Inadequate, if the allocation sequence was known to the investigators who assigned participants or if the study was quasi‐randomised. Such studies were excluded from the present review.

Blinding (or masking)

• Adequate, if the trial was described as double blind and the method of blinding involved identical placebo or active drug;

• Unclear, if the trial was described as double blind, but the method of blinding was not described;

• Not performed, if the trial was not double blind.

Follow‐up

• Adequate, if the numbers and reasons for dropouts and withdrawals in all intervention groups were described or if it was specified that there were no dropouts or withdrawals;

• Unclear, if the report gave the impression that there had been no dropouts or withdrawals, but this was not specifically stated;

• Inadequate, if the number or reasons for dropouts and withdrawals were not described.

Characteristics of patients 
 Number of patients randomised; patient inclusion and exclusion criteria; mean (or median) age; sex ratio; number of patients lost to follow‐up; drop‐outs; withdrawals.

Characteristics of interventions 
 Type, dose, and form of colchicine intervention; type of intervention in the control group and collateral interventions; duration of treatment, length of follow‐up.

Characteristics of outcomes 
 All outcomes were extracted from each included trial when possible.

We analysed mortality and/or liver transplantation at maximum follow‐up. We analysed other outcomes, which were repeatedly observed on patients (like liver biochemistry, clinical symptoms, etc.) at maximum follow‐up. However, where possible, we also extracted data on primary outcome measures from the maximal follow‐up in each randomised clinical trial, and if available, including data from after the patients were switched from blinded to open label therapy.

Statistical methods 
 We used RevMan Analyses 1.0.1 provided by The Cochrane Collaboration. Dichotomous data were presented as relative risk (RR) with 95% confidence interval (CI), and continuous outcomes were presented by weighted mean differences (WMD) with 95% CI. The analyses for the primary outcomes were performed according to the intention‐to‐treat analyses, which means that the participants in the trials were to be analysed in the groups to which they were randomised, regardless of whether they received or adhered to the allocated intervention. We computed a 'reported scenario' analysis. However, we placed most weight on the 'likely scenario' analysis (see Subgroup analyses and sensitivity analyses below).

We examined intervention effects by using both a fixed effect model (DeMets 1987) and a random effects model (DerSimonian 1986) with the level of significance set at P ≤ 0.05. If the results of the two analyses led to the same conclusion, only the result of the fixed effect model analysis was given in the text. In case of significant discrepancies of the two models, results from both models were reported and discussed. The presence of statistical heterogeneity was explored by the chi‐squared test with significance set at P ≤ 0.10 and measured the quantities of heterogeneity by I².

Subgroup analyses and sensitivity analyses 
 We performed subgroup analyses, in which trials were grouped according to the stage of disease; duration of treatment; adequacy of generation of the allocation sequence; allocation concealment; blinding; and whether the trial reported used intention‐to‐treat analysis. The cut‐off for duration of treatment was determined by comparing the intervention effect of the group of trials lasting for no more than the median treatment duration with that of the group of trials lasting for more than the median duration. The differences between subgroups were estimated according to Altman 2003.

Regarding the binary outcomes, patients with incomplete or missing data were included in sensitivity analyses by imputing them:

• Likely scenario: worst‐case scenario for both colchicine and control.

• Best‐case scenario: best‐case scenario for colchicine and worst‐case scenario for control.

• Reported scenario: best‐case scenario for both colchicine and control.

• Worst‐case scenario: worst‐case scenario for colchicine and best‐case scenario for control.

For secondary outcomes we adopted 'available case analysis', i.e., include data on only those patients, whose results are known, using for denominator the total number of patients who completed the trial for the particular outcome in question. Thus, in the review, the number of patients as the denominator might change according to the secondary outcomes investigated.

Bias detection 
 Funnel plot was used to provide a visual assessment of whether treatment estimates are associated with study size. The performance of the available methods of detecting publication bias and other biases (Begg 1994; Egger 1997; Macaskill 2001) vary with the magnitude of the treatment effect, the distribution of study size, and whether a one‐ or two‐tailed test is used (Macaskill 2001). Therefore, we decided to use the most appropriate method having good trade‐off in the sensitivity and specificity, based on characteristics of the trials to be included in this review.

Results

Description of studies

Search results 
 We identified a total of 559 references through electronic searches of The Cochrane Hepato‐Biliary Group Controlled Trials Register (n = 29), The Cochrane Controlled Trials Register on The Cochrane Library (n = 212), MEDLINE (n = 148), and EMBASE (n = 170). We excluded 465 duplicates and clearly irrelevant references through reading abstracts. Accordingly, 95 references were retrieved for further assessment. Of these, we excluded 57 because they were reviews, meta‐analyses, or observational studies. Among the 57 references, the three observational studies were listed under 'Characteristic of excluded studies' with reasons for exclusion. The remaining 38 references referred to 11 randomised trials, which fulfilled our inclusion criteria of this review.

Two of the 11 randomised clinical trials were published as abstracts only (Goddard 1995; Warnes 1996). One trial (Raedsch 1993) was published in symposia proceedings. The remaining eight randomised clinical trials were published in peer‐reviewed journals.

Included studies 
 All the included trials reported random allocation of 716 patients with primary biliary cirrhosis to:

• colchicine versus placebo (Kaplan 1986; Warnes 1987; Bodenheimer 1988; Goddard 1995; Vuoristo 1995; Warnes 1996)

• colchicine plus ursodeoxycholic acid versus placebo/no intervention plus ursodeoxycholic acid (Raedsch 1993; Goddard 1995; Ikeda 1996; Poupon 1996; Almasio 2000)

• colchicine versus ursodeoxycholic acid (Goddard 1995; Vuoristo 1995)

• colchicine versus methotrexate (Kaplan 1999).

Vuoristo 1995 had three intervention arms: colchicine versus ursodeoxycholic acid versus placebo, and Goddard 1995 had four intervention arms: colchicine, ursodeoxycholic acid, colchicine plus ursodeoxycholic acid, and placebo. We were not able to extract data from Goddard 1995 and Warnes 1996 for our meta‐analyses because they were published only as abstracts and correspondence with the authors did not lead to additional information. Accordingly, data from nine trials involving 599 patients with primary biliary cirrhosis were pooled in our meta‐analyses.

The entry criteria varied across trials, but were generally well‐defined, making it highly likely that all or almost all patients did in fact have primary biliary cirrhosis. The dosage of colchicine varied slightly, from 1 mg daily (n = 7) to 1.2 mg daily (n = 3). Only Warnes 1996 did not report the dosage. The duration of colchicine treatment varied from one to two years. Following the stipulated follow‐up, two trials (Kaplan 1986; Bodenheimer 1988) continued colchicine‐patients on open label colchicine (colchicine → colchicine) and offered open label colchicine to the patients originally receiving placebo (placebo → colchicine). One subsample of Poupon 1996 trial, which was published as an abstract, continued patients in both groups with open label colchicine plus ursodeoxycholic acid. However, we were not able to retrieve additional data after our correspondence with the principal author.

Risk of bias in included studies

The methods to generate the allocation sequence were considered adequate in four randomised clinical trials (Kaplan 1986; Warnes 1987; Vuoristo 1995; Almasio 2000) and unclear or inadequate in the remaining seven. The methods to conceal allocation were considered adequate in six (Kaplan 1986; Warnes 1987; Vuoristo 1995; Ikeda 1996; Kaplan 1999; Almasio 2000) and unclear or inadequate in the remaining five. Blinding was adequate in seven trials, unclear in three (Raedsch 1993; Goddard 1995; Warnes 1996) and not performed in Ikeda 1996 trial. The Kaplan 1999 trial, which compared colchicine to methotrexate, employed the double dummy technique to maintain the double‐blinding. The description of the placebo was, however, not sufficient ‐ i.e., some of the trials employing placebo only stated that the placebo tablets were identical in appearance or indistinguishable, but did not mention smell and taste. The other randomised clinical trials (Raedsch 1993; Vuoristo 1995; Warnes 1996) did not give any description of the placebo used. There was generally a fair description of follow‐up and withdrawal/drop‐out, in which eight trials were regarded as adequate and three inadequate. Two trials had high rates of loss of follow‐up and withdrawals/drop‐outs, 22.8% in Bodenheimer 1988 trial and 15.6% in Warnes 1987 trial.

To note, only two out of the 11 randomised clinical trials (Kaplan 1986; Kaplan 1999) provided pre‐trial sample size estimation based on the rates of success defined by the authors. None of the trials used mortality to calculate sample size estimation.

Effects of interventions

Colchicine versus placebo/no intervention (monotherapy or combination therapy) 
 Number of deaths 
 Seven randomised clinical trials involving 398 patients reported data on number of deaths. In the colchicine group 25/200 (12.5%) patients died versus 20/198 (10.1%) patients in the control group. Combining the results of individual trials demonstrated no significant difference in the number of deaths (RR 1.21, 95% CI 0.71 to 2.06) (Analysis 1.1).

1.1. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 1 Number of deaths.

We performed sensitivity analyses regarding the number of deaths (Analysis 6.1). Neither the reported‐scenario nor the likely‐scenario analyses showed any significant difference between colchicine and placebo/no intervention. The best‐worst‐case‐scenario analysis did not show any significant difference either. The worst‐best‐case‐scenario analysis detected a significant detrimental effect of colchicine on mortality.

6.1. Analysis.

Comparison 6 Sensitivity analyses ‐ colchicine versus placebo/no intervention, Outcome 1 Number of deaths.

Including data from the extended follow‐up during treatment with colchicine → colchicine versus placebo → colchicine into the analyses demonstrated a RR of 1.15 (95% CI 0.76 to 1.73) (Analysis 2.1).

2.1. Analysis.

Comparison 2 Colchcine ‐ colchicine versus placebo ‐ colchicine (including open label period), Outcome 1 Number of deaths.

Number of deaths and/or patients who underwent liver transplantation 
 Eight randomised clinical trials involving 455 patients reported data on 'number of deaths and/or patients who underwent liver transplantation'. We detected 36/228 (15.8%) deaths and patients who underwent liver transplantation in the colchicine group versus 36/227 (15.9%) in the control group. Combining the results of the eight trials demonstrated no significant difference in this outcome measure (RR 1.00, 95% CI 0.67 to 1.49) (Analysis 1.2).

1.2. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 2 Number of deaths and/or patients who underwent liver transplantation.

We performed sensitivity analyses. Neither the reported‐scenario nor the likely‐scenario analyses showed any significant difference between colchicine and placebo/no intervention. The best‐worst‐case‐scenario analysis showed a significant effect favouring colchicine, while the worst‐best‐case‐scenario analysis showed a significant effect favouring placebo/no intervention.

Including data from the extended follow‐up during treatment with colchicine → colchicine versus placebo → colchicine demonstrated a RR of 1.02 (95% CI 0.72 to 1.46) (Analysis 2.2).

2.2. Analysis.

Comparison 2 Colchcine ‐ colchicine versus placebo ‐ colchicine (including open label period), Outcome 2 Number of deaths and/or patients who underwent liver transplantation.

Subgroup analyses 
 The subgroup analyses, taking the dose and duration of colchicine into consideration, did not reveal differing results (Analysis 5.1; Analysis 5.2). The trials where colchicine was administered with 1 mg/day (RR 1.36, 95% CI 0.73 to 2.52) versus 1.2 mg/day (RR 0.83, 95% CI 0.28 to 2.44) did not reveal any significant influence on the relative risk of mortality. Test of interaction between the two estimates showed no significant difference (P = 0.44). The trials where colchicine was administered for no longer than two years (RR 1.04, 95% CI 0.56 to 1.93) did not differ significantly from the trials where colchicine was administered for longer than two years (RR 1.82, 95% CI 0.62 to 5.39). Test of interaction between the two groups detected no significant difference (P = 0.38).

5.1. Analysis.

Comparison 5 Subgroup analyses ‐ colchicine versus placebo/no intervention, Outcome 1 Number of deaths ‐ dose variation.

5.2. Analysis.

Comparison 5 Subgroup analyses ‐ colchicine versus placebo/no intervention, Outcome 2 Number of deaths ‐ treatment duration.

Subgroup analyses on mortality stratifying the seven trials according to their methodological quality were performed. The adequacy of generation of the allocation sequence, allocation concealment, and blinding did not change this estimate significantly (P = 0.15, 0.15, and 0.26, respectively) (Analysis 5.3 to Analysis 5.5). Follow‐up was adequate in all the trials, which provided mortality data.

5.3. Analysis.

Comparison 5 Subgroup analyses ‐ colchicine versus placebo/no intervention, Outcome 3 Number of deaths ‐ generation of the allocation sequence.

5.5. Analysis.

Comparison 5 Subgroup analyses ‐ colchicine versus placebo/no intervention, Outcome 5 Number of deaths ‐ blinding.

Subgroup analyses stratifying the trials according to monotherapy or combined treatment, i.e., colchicine plus ursodeoxycholic acid, did not change this estimate (Analysis 1.1):

• colchicine versus placebo (RR 0.92, 95% CI 0.48 to 1.75);

• colchicine plus ursodeoxycholic acid versus placebo/no intervention plus ursodeoxycholic acid ( RR 2.14, 95% CI 0.78 to 5.87).

Test of interaction between the two groups showed no significant difference (P = 0.17).

Similar findings applied to the risks of mortality or liver transplantation (Analysis 1.2; Analysis 5.6 to Analysis 5.10).

5.6. Analysis.

Comparison 5 Subgroup analyses ‐ colchicine versus placebo/no intervention, Outcome 6 Number of deaths and/or patients who underwent liver transplantation ‐ dose variation.

5.10. Analysis.

Comparison 5 Subgroup analyses ‐ colchicine versus placebo/no intervention, Outcome 10 Number of deaths and/or patients who underwent liver transplantation ‐ blinding.

Pruritus and fatigue 
 Pooling the data from three trials demonstrated that colchicine significantly decreased the number of patients without improvement of pruritus (RR 0.75, 95% CI 0.65 to 0.87). One trial reported the data of the 'number of patients without improvement of fatigue', and it was not significantly different in the colchicine group and the control group (RR 0.86, 95% CI 0.72 to 1.02).

Liver complications 
 Overall, no significant difference was detected on liver complications between colchicine and control group (RR 0.37, 95% CI 0.12 to 1.10). Neither the number of patients with development of varices (RR 0.31, 95% CI 0.08 to 1.19), gastrointestinal bleeding (RR 0.50, 95% CI 0.05 to 5.28), nor the number of patients developing hepatic encephalopathy (RR 1.07; 95% CI 0.07 to 16.31) were significantly affected by colchicine. We were not able to extract data on jaundice.

Biochemical variables 
 Colchicine did not lead to any significant effect on the following biochemical variables (Analysis 1.7 to Analysis 1.15):

1.7. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 7 Appearence of liver complications.

1.15. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 15 Plasma immunoglobulin M (g/L).

• s‐bilirubin: WMD (arithmetic mean) ‐1.35 µmol/L, 95% CI ‐4.52 to 1.82; WMD (geometric mean) ‐1.55 µmol/L, 95% CI ‐2.72 to 1.13;

• s‐alkaline phosphatases: WMD (arithmetic mean) ‐55.35 international units (IU)/L, 95 CI ‐158.56 to 47.85; WMD (geometric mean) ‐1.26 IU/L, 95 CI ‐1.80 to 1.14;

• s‐gamma‐glutamyltransferase: WMD ‐25.38 IU/L, 95% CI ‐73.26 to 22.50;

• s‐aspartate aminotransferase: WMD ‐10.10 IU/L, 95% CI ‐22.91 to 2.71;

• s‐alanine aminotransferase: WMD ‐2.05 IU/L, 95% CI ‐8.79 to 4.68;

• s‐albumin: WMD 0.09 g/dL, 95% CI ‐0.03 to 0.21;

• s‐total cholesterol: WMD (arithmetic mean) 0.10 mmol/L, 95% CI ‐0.88 to 1.08; WMD (geometric mean) ‐1.02 mmol/L, 95% CI ‐1.20 to 1.15;

• plasma immunoglobulin M: WMD ‐0.49 g/L, 95% CI ‐1.03 to 0.06;

• prothrombin time: WMD ‐0.03 seconds, 95% CI ‐0.75 to 0.69.

The Kaplan 1986 trial reported bilirubin, cholesterol, and alkaline phosphatases using geometric mean (Analysis 5.11; Analysis 5.12; Analysis 5.13), and we reported them as log transformed geometric mean for the sake of comparison.

5.11. Analysis.

Comparison 5 Subgroup analyses ‐ colchicine versus placebo/no intervention, Outcome 11 S‐bilirubin (µmol/L) ‐ reported as arithmetic mean or geometric mean.

5.12. Analysis.

Comparison 5 Subgroup analyses ‐ colchicine versus placebo/no intervention, Outcome 12 S‐cholesterol (total) (mmol/L) ‐ reported as arithmetic mean or geometric mean.

5.13. Analysis.

Comparison 5 Subgroup analyses ‐ colchicine versus placebo/no intervention, Outcome 13 S‐alkaline phosphatase (ALP) (IU/L) ‐ reported as arithmetic mean or geometric mean.

Liver histology 
 There was no significant influence of colchicine on the number of patients experiencing worsening of histological stage (RR 0.85, 95% CI 0.41 to 1.75), fibrosis (RR 0.60, 95% CI 0.24 to 1.49), piecemeal necrosis (RR 0.58, 95% CI 0.23 to 1.44), parenchymal inflammation (RR 0.69, 95% CI 0.28 to 1.72), or parenchymal necrosis (RR 1.00, 95% CI 0.31 to 3.18). In addition, the Warnes 1987 trial reported no significant effect of colchicine on the number of patients who underwent worsening of cholestasis and granulomas. Poupon 1996 demonstrated no significant effects of colchicine on the number of patients with worsening of ductular proliferation and cholangitis. However, Poupon 1996 observed a significant lower incidence of patients with worsening of lobular inflammation in the colchicine group (RR 0.16, 95% CI 0.03 to 0.80) (Analysis 1.17). No significant effects on histological score were observed (WMD 0.56, 95% CI ‐0.24 to 1.36) in colchicine patients when compared to the patients in the control group (Analysis 1.18).

1.17. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 17 Liver biopsy findings ‐ dichotomous variables.

1.18. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 18 Liver biopsy findings ‐ histological score.

Quality of life 
 None of the trials examined specific quality‐of‐life scales or health economics.

Adverse events 
 In the colchicine group, 39/228 (17.1%) patients had adverse events (mostly transient diarrhoea) versus 26/227 (11.5%) patients in the control group (Analysis 1.19). This was not significantly different (RR 1.45, 95% CI 0.94 to 2.25). Also, no significantly different occurrences of serious adverse events were observed (RR 1.17, 95% CI = 0.50 to 2.75) (Analysis 1.20).

1.19. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 19 Number of patients with adverse events.

1.20. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 20 Number of patients with serious adverse events.

Colchicine versus ursodeoxycholic acid 
 Vuoristo 1995 compared colchicine versus ursodeoxycholic acid. They observed that 5/29 patients died in the colchicine group versus 0/30 patients in the ursodeoxycholic acid group (RR 11.37, 95% CI 0.66 to 196.74) and no one underwent liver transplantation. The number of patients without improvement of fatigue was significantly less in the colchicine group than in the ursodeoxycholic acid group (RR 0.83, 95% CI 0.70 to 0.98). Regarding liver biochemical outcomes, only the levels of s‐alkaline phosphatases and gamma‐glutamyltransferase were significantly higher in the colchicine group than in the ursodeoxycholic acid group (WMD 378.00 IU/L, 95% CI 116.91 to 639.09; WMD 459.00 IU/L, 95% CI 157.57 to 760.43, respectively). For other outcomes (i.e., number of patients without improvement of pruritus, number of patients developing liver complications, number of patients with adverse events), no significant differences were detected.

Colchicine versus methotrexate 
 Kaplan 1999 compared colchicine versus methotrexate. This study observed that 9/43 patients died or underwent liver transplantation in the colchicine group versus 11/42 patients in the methotrexate group (RR 0.80, 95% CI 0.37 to 1.73). The pruritus score was significantly higher in patients receiving colchicine than methotrexate (WMD 0.68, 95% CI 0.25 to 1.11). Regarding liver biochemical outcomes, only the levels of s‐alkaline phosphatases and plasma immunoglobulin M were significantly higher in the colchicine group than in the methotrexate group (WMD 0.41, 95% CI 0.12 to 0.70; WMD 0.47, 95% CI 0.20 to 0.74), respectively. For other outcomes (i.e., fatigue score, liver histology, number of patients with adverse events) no significant differences were detected.

Bias detection 
 We did not perform funnel plot analysis and did not apply the three statistical methods to detect publication bias and other biases because the power of those would have been low and inconsistent because of the small number of included trials.

Discussion

We found no significant difference on mortality or mortality and liver transplantation between colchicine and placebo/no intervention for patients with primary biliary cirrhosis. These observations were robust to subgroup analyses taking methodological quality of trials, dose, and treatment duration into consideration. It has been reported that trials with inadequate methodological quality do significantly overestimate the effect of interventions (Schulz 1995; Moher 1998; Kjaergard 2001). However, we found that our results are not sensitive to the adequacy of generation of the allocation sequence, allocation concealment, blinding status, and the use of intention‐to‐treat analysis, probably due to the relatively small sample size and low number of trials included.

Our systematic review may have a number of limitations. Firstly, our systematic review regarding the comparison of colchicine versus placebo/no intervention on mortality analysed only seven trials involving 398 patients. This is a low number of patients (Ioannidis 2001). Additionally, compared to the natural history of primary biliary cirrhosis, most of the trials had relatively short period of medication and follow‐up. Thus, the risk of type 2 error (the risk of overlooking an effect if it really exists) is present and a potential beneficial effect of colchicine on survival cannot be reliably excluded. Secondly, the present meta‐analyses on mortality and mortality or liver transplantation were based on number of events per randomised patients from the individual trial, not on individual patient data analysis based on time‐to‐event data. An individual patient data analysis takes time and censored data into consideration and may offer potential advantages. However, the use of meta‐analysis based on aggregate data extracted from published and unpublished reports can be considered a useful approach and seems to reach similar conclusions (Liberati 1996). Thirdly, since we could not stratify summary data of included trials according to the patients' baseline stage of primary biliary cirrhosis, we do not know whether the effect of colchicine was associated with the severity of primary biliary cirrhosis. Fourthly, we performed a high number of statistical tests, which increases the risk of 'mass significance' (i.e., spurious significant findings due to repetitive testing). Therefore, significant findings ought to be conservatively interpreted. Fifthly, although we employed considerable search strategies and applied no publication status or language limitations, we are concerned about the existence of publication bias and other biases, which leads us to identify 'positive' studies more easily than 'negative' ones (Gluud 1998).

Our findings regarding primary outcome measures did not seem to be sensitive to missing data. Neither the reported scenario nor the likely scenario analyses, which are attempts to fill in missing data in a realistic manner, showed any significant difference between colchicine and placebo/no intervention. The best‐case scenario did not show any significant difference. The worst‐case scenario detected a significant detrimental effect of colchicine. Although the best‐worst‐case‐ and worst‐best‐case‐scenario analyses are extreme and unlikely, it is more probable that the treatment effect did not favour colchicine but placebo/no intervention. Additionally, we found that the effect of colchicine on mortality and liver transplantation (favouring colchicine in the best‐worst‐case‐scenario analyses and favouring placebo/no intervention in the worst‐best‐case‐scenario analysis) to be heavily depended on the Bodenheimer 1988 trial, in which the rate of loss of follow‐up in the colchicine group was 28.5%. After excluding this trial, we got a non‐significant difference (Analysis 6.3). In addition, when we included data from 114 patients from two trials switched from blinded to open label colchicine therapy, these differences were not significant either on mortality or on mortality or liver transplantation.

6.3. Analysis.

Comparison 6 Sensitivity analyses ‐ colchicine versus placebo/no intervention, Outcome 3 Number of deaths and/or patients who underwent liver transplantation (excluding Bodenheimer 1988).

In order to examine the effects of colchicine in a broader context, we expanded our analyses by including trials on colchicine versus placebo/no intervention for alcoholic and non‐alcoholic liver fibrosis and cirrhosis (Rambaldi 2003). The pooled results showed no significant difference on mortality. In the colchicine group 109/786 (13.9%) patients died versus 106/762 (13.9%) patients in the control group (see Figure 1).

1.

Relative risk of mortality in patients with primary biliary cirrhosis, alcoholic, and non‐alcoholic liver fibrosis and cirrhosis randomised to colchicine versus placebo/no intervention

The Goddard et al. trial was a 2 multiplied by 2 factorial designed trial, which could have investigated the possible interaction between colchicine and UDCA. However, the trial was only published as an abstract and the author did not reply to our request for further information. A synergistic effect was claimed based on a non‐randomised study (Shibata 1992). However, our subgroup analyses, stratifying the included trials into monotherapy (i.e., colchicine versus placebo/no intervention) and combination therapy (i.e., colchicine plus UDCA versus placebo/no intervention plus UDCA) did not suggest additional effect of colchicine introduced by the combination with UDCA in the identified trials.

We found that colchicine had a significant beneficial effect on pruritus. This finding was from three trials involving only 156 patients. A number of arguments may contradict this observation. First, lack of efficient blinding of trials (Kjaergard 2001) and the subjective nature of pruritus assessment could have biased the estimate. Second, pruritus usually reflects indices of cholestasis (e.g., serum alkaline phosphatases) and a correlation between the severity of pruritus and the presence of florid bile duct lesions in the liver has been reported (Poupon 1999). Our analyses did not show any significant effect of colchicine on any plasma indices of cholestasis or on liver histology. Furthermore, due to the large number of statistical comparisons having been performed some of the comparisons might have come out with a significant difference simply due to 'mass significance'. Therefore, we are not convinced that the improvement of pruritus was due to colchicine. The potential beneficial effect of colchicine on pruritus might be worth exploring in future high‐quality randomised trials.

We did not find any significant difference on liver biochemical parameters between colchicine treatment and placebo/no intervention. It appeared that the use of colchicine was associated with improvement in hepatic biochemistries in three early randomised clinical trials (Kaplan 1986; Warnes 1987; Bodenheimer 1988). In those three trials, however, the protocol violations regarding per cent of randomised patients who were: (i) lost to follow‐up; (ii) refused liver biopsy; (iii) were noncompliant, and (iv) were withdrawn due to adverse events or disease progression ‐ were: 33%, 13%, and 38%, respectively. Only one trial (Warnes 1987) stated having employed the intention‐to‐treat principle.

Primary biliary cirrhosis is a pathological process starting with portal inflammation, which progresses towards three irreversible stages: a stage of compensated cirrhosis, a stage of decompensated cirrhosis (defined by high bilirubin levels (greater than 100 µmol/L), ascites, and variceal bleeding), and a terminal stage, in which death occurs unless liver transplantation is performed (Gluud 2002). The purpose of the randomised clinical trials assessing colchicine for primary biliary cirrhosis has not been to evaluate whether colchicine could reverse the decompensated stage or the terminal stage of the disease, but rather, if colchicine could slow the progression towards the cirrhotic stage and the more advanced stages. It is, therefore, interesting to study the effect of colchicine on liver histology. In this review, we were not able to identify any significant effect of colchicine on a number of histological variables. The Almasio 2000 trial reported a significant reduction in histological grading score in patients administrating colchicine plus ursodeoxycholic acid; however, the proportion of patients having liver biopsy was very low (15 patients out of 90). Thus, its significance could be biased by impact of missing data on liver histology.

Vuoristo et al. performed comparison of colchicine versus ursodeoxycholic acid, the most widely used drug in the treatment of primary biliary cirrhosis. No significant difference was detected regarding mortality (Vuoristo 1995). Colchicine appeared to relieve fatigue, but the effect size was small. For the liver biochemical outcomes, the significant difference detected on s‐alkaline phosphatases and gamma‐glutamyltransferase was suggestive of a favourable effect of ursodeoxycholic acid. This is in accordance with trials comparing ursodeoxycholic acid with placebo/no intervention (Gluud 2002). Overall, we were not able to suggest beneficial effect of colchicine compared to ursodeoxycholic acid.

Compared to methotrexate, a folic acid antagonist that blocks nucleic acid synthesis, colchicine seemed to be less effective against methotrexate regarding severity of pruritus and level of s‐alkaline phosphases and plasma immunoglobulin M. The data we extracted were from a two‐year interim analysis of the ten‐year Kaplan 1999 trial. The trial is finished, but the published data are not available presently (October 2003).

Regarding the safety issue of colchicine treatment in primary biliary cirrhosis, this systematic review could not demonstrate that colchicine was associated with an increase or decrease of non‐serious adverse events (mainly transient diarrhoea, usually resolved by lowering the dose of colchicine) or serious adverse events. We were not able to identify data on the effects of colchicine concerning quality of life and health economics.

Authors' conclusions

Implications for practice.

We did not find convincing evidence showing that colchicine had significant beneficial effects on patients with primary biliary cirrhosis when compared to placebo or no intervention. The combination of colchicine and UDCA did not significantly change the effects of colchicine. We are not able to exclude the possibility that colchicine may reduce mortality by 70%. On the other hand, it may increase mortality by 344%. We therefore cannot recommend the use of colchicine outside randomised clinical trials.

Implications for research.

If researchers have an interest to investigate colchicine for primary biliary cirrhosis, they may consider the following:

• due to the chronic progression of primary biliary cirrhosis and thanks to the low toxicity of colchicine, long‐term follow‐up is needed and seems feasible;

• to have an independent data monitoring and safety committee, which can follow the data and stop the trial should it start to demonstrate harmful effects of colchicine

• to study in detail the potential effect of colchicine on pruritus;

• to ensure that enough patients with primary biliary cirrhosis are kept followed to undergo liver biopsy in order to obtain more data on liver histology;

• to include quality‐of‐life and health economics analyses;

• to adhere to the Consort Statement (www.consort‐statement.org).

Feedback

Colchicine for primary biliary cirrhosis

Summary

Date of Submission: 13‐Jun‐2006 
 Name: Roger Pepin 
 Email Address: r.pepin@elsevier.com 
 Personal Description: Occupation EBM Editor

Feedback: Could someone clarify the relationship of this review and the review published in American Journal of Gastroenterology, 2005 Aug;100(8):1876‐85. In particular, could they comment on the differing numbers of trials included in each review, and the observations made of trial quality in each abstract.

Submitter agrees with default conflict of interest statement: 
 I certify that I have no affiliations with or involvement in any organization or entity with a financial interest in the subject matter of my feedback.

Reply

We published the systematic review in The Cochrane Library 2004, Issue 2, based on which we published a modified version in The American Journal of Gastroenterology in 2005. The Cochrane Hepato‐Biliary Group has endorsed a co‐publication agreement with The American Journal of Gastroenterology.

The trials included in the review published in The American Journal of Gastroenterology compare colchicine versus placebo/no intervention. The review published in The Cochrane Library is broader since it also includes trials comparing colchicine versus ursodeoxycholic acid and colchicine versus methotrexate.

We stated in the The American Journal of Gastroenterology the following: 
 'This review is published as a Cochrane Review in The Cochrane Library 2004, Issue 2. Cochrane Reviews are regularly updated as new evidence emerges and in response to comments and criticisms. The Cochrane Library should be consulted for the most recent version of the Review.'

Contributors

Yan Gong, primary author 
 23 August 2006 
 Copenhagen, Denmark

Christian Gluud, author and Criticism Editor 
 23 August 2006 
 Copenhagen, Denmark

What's new

Date	Event	Description
24 September 2008	Amended	Converted to new review format.

Appendices

Appendix 1. Search Strategies

Database	Period	Search Strategy Used
The Controlled Trial Register of The Cochrane Hepato‐Biliary Group	September 2003	#1 = 'RCT' and ' PRIMARY BILIARY CIRRHOSIS' and ' COLCHICINE'
The Cochrane Library (CENTRAL)	2003 Issue 3	#1 = LIVER‐CIRRHOSIS‐BILIARY*: MESH #2 = (PRIMARY and BILIARY and CIRRHOSIS) or PBC #3 = COLCHICINE: MESH #4 = IMMUNOSUPPRES* : MESH #5 = URSODEOXYCHOLIC‐ACID: MESH #6 = COLCHICINE or IMMUNOSUPPRES* or (URSODEOXYCHOLIC and ACID) #7 = #3 or #4 or #5 or #6 #8 = (#1 and #7) #9 = (#2 and #7) #10 = (#8 or #9)
MEDLINE	January 1966 to September 2003	#1 = LIVER‐CIRRHOSIS‐BILIARY: MESH #2 =(PRIMARY and BILIARY and CIRRHOSIS) or PBC #3 = "PRIMARY BILIARY CIRRHOSIS" or PBC #4 = #2 or #3 #5 = COLCHICINE #6 = IMMUNOSUPPRES* #7 = URSODEOXYCHOLIC #8 = ACID #9 = #5 or #6 or (#7 and #8) #10 = COLCHICINE: MESH #11 = IMMUNOSUPPRESS*: MESH #12 = URSODEOXYCHOLIC‐ACID: MESH #13 = #9 or #10 or #11 or #12 #14 = #1 and #13 #15 = #4 and #13 #16 = #14 or #15 #17 = random* #18 = placebo* #19 = blind* #20 = meta‐analysis #21 = #17 or #18 or #19 or #20 #22 = #16 and # 21
EMBASE	January 1980 to September 2003	#1 = LIVER‐CIRRHOSIS‐BILIARY: MESH #2 = (PRIMARY and BILIARY and CIRRHOSIS) or PBC #3 = "PRIMARY BILIARY CIRRHOSIS" or PBC #4 = #2 or #3 #5 = COLCHICINE #6 = IMMUNOSUPPRES* #7 = URSODEOXYCHOLIC #8 = ACID #9 = #5 or #6 or (#7 and #8) #10 = COLCHICINE: MESH #11 = IMMUNOSUPPRESS*: MESH #12 = URSODEOXYCHOLIC‐ACID: MESH #13 = #9 or #10 or #11 or #12 #14 = #1 and #13 #15 = #4 and #13 #16 = #14 or #15 #17 = random* #18 = placebo* #19 = blind* #20 = meta‐analysis #21 = #17 or #18 or #19 or #20 #22 = #16 and # 21

Data and analyses

Comparison 1. Colchicine versus placebo/no intervention.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of deaths	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	1.21 [0.71, 2.06]
1.1 Colchicine versus placebo	3	184	Risk Ratio (M‐H, Fixed, 95% CI)	0.92 [0.48, 1.75]
1.2 Colchicine + UDCA versus placebo/no intervention + UDCA	4	214	Risk Ratio (M‐H, Fixed, 95% CI)	2.14 [0.78, 5.87]
2 Number of deaths and/or patients who underwent liver transplantation	8	455	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.67, 1.49]
2.1 Colchicine versus placebo	4	241	Risk Ratio (M‐H, Fixed, 95% CI)	0.85 [0.54, 1.32]
2.2 Colchicine + UDCA versus placebo/no intervention + UDCA	4	214	Risk Ratio (M‐H, Fixed, 95% CI)	1.78 [0.69, 4.57]
3 Number of patients who underwent liver transplantation	5	274	Risk Ratio (M‐H, Fixed, 95% CI)	0.34 [0.06, 2.10]
3.1 Colchicine versus placebo	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.36 [0.04, 3.23]
3.2 Colchicine + UDCA versus placebo/no intervention + UDCA	4	214	Risk Ratio (M‐H, Fixed, 95% CI)	0.32 [0.01, 7.63]
4 Number of patients without improvement of pruritus	3	156	Risk Ratio (M‐H, Fixed, 95% CI)	0.75 [0.65, 0.88]
4.1 Colchicine versus placebo	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.63 [0.47, 0.83]
4.2 Colchicine + UDCA versus placebo/no intervention + UDCA	2	96	Risk Ratio (M‐H, Fixed, 95% CI)	0.84 [0.71, 0.99]
5 Number of patients without improvement of fatigue	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.86 [0.72, 1.02]
5.1 Colchicine versus placebo	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.86 [0.72, 1.02]
5.2 Colchicine + UDCA versus placebo/no intervention + UDCA	0	0	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 Number of patients developing liver complications	3	156	Risk Ratio (M‐H, Fixed, 95% CI)	0.37 [0.12, 1.10]
6.1 Colchicine versus placebo	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.36 [0.04, 3.23]
6.2 Colchicine + UDCA versus placebo/no intervention + UDCA	2	96	Risk Ratio (M‐H, Fixed, 95% CI)	0.38 [0.11, 1.30]
7 Appearence of liver complications	3		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
7.1 Development of varices	3	153	Risk Ratio (M‐H, Fixed, 95% CI)	0.31 [0.08, 1.19]
7.2 Gastrointestinal bleeding	2	96	Risk Ratio (M‐H, Fixed, 95% CI)	0.5 [0.05, 5.28]
7.3 Development of ascites	2	82	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
7.4 Hepatic encephalopathy	2	82	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [0.07, 16.31]
8 S‐bilirubin (µmol/L)	4	202	Mean Difference (IV, Fixed, 95% CI)	‐1.35 [‐4.52, 1.82]
8.1 Colchicine versus placebo	1	60	Mean Difference (IV, Fixed, 95% CI)	8.5 [‐2.04, 19.04]
8.2 Colchicine + UDCA versus placebo/no intervention + UDCA	3	142	Mean Difference (IV, Fixed, 95% CI)	‐2.33 [‐5.66, 0.99]
9 S‐alkaline phosphatases (ALP)(IU/L)	4	200	Mean Difference (IV, Fixed, 95% CI)	‐55.35 [‐158.56, 47.85]
9.1 Colchicine versus placebo	1	60	Mean Difference (IV, Fixed, 95% CI)	66.0 [‐197.69, 329.69]
9.2 Colchicine + UDCA versus placebo/no intervention + UDCA	3	140	Mean Difference (IV, Fixed, 95% CI)	‐77.31 [‐189.46, 34.85]
10 S‐gamma‐glutamyltransferase (GGT)(IU/L)	4	200	Mean Difference (IV, Fixed, 95% CI)	‐25.38 [‐73.26, 22.50]
10.1 Colchicine versus placebo	1	60	Mean Difference (IV, Fixed, 95% CI)	221.0 [‐127.15, 569.15]
10.2 Colchicine + UDCA versus placebo/no intervention + UDCA	3	140	Mean Difference (IV, Fixed, 95% CI)	‐30.13 [‐78.47, 18.21]
11 S‐aspartate aminotransferase (AST)(IU/L)	2	82	Mean Difference (IV, Fixed, 95% CI)	‐10.10 [‐22.91, 2.71]
11.1 Colchicine versus placebo	1	60	Mean Difference (IV, Fixed, 95% CI)	‐2.0 [‐26.95, 22.95]
11.2 Colchicine + UDCA versus placebo/no intervention + UDCA	1	22	Mean Difference (IV, Fixed, 95% CI)	‐13.0 [‐27.93, 1.93]
12 S‐alanine aminotransferase (ALT)(IU/L)	4	201	Mean Difference (IV, Fixed, 95% CI)	‐2.05 [‐8.79, 4.68]
12.1 Colchicine versus placebo	1	60	Mean Difference (IV, Fixed, 95% CI)	‐5.0 [‐32.86, 22.86]
12.2 Colchicine + UDCA versus placebo/no intervention + UDCA	3	141	Mean Difference (IV, Fixed, 95% CI)	‐1.87 [‐8.82, 5.07]
13 S‐albumin (g/dL)	4	235	Mean Difference (IV, Fixed, 95% CI)	0.09 [‐0.03, 0.21]
13.1 Colchicine versus placebo	2	117	Mean Difference (IV, Fixed, 95% CI)	0.11 [‐0.04, 0.27]
13.2 Colchicine + UDCA versus placebo/no intervention + UDCA	2	118	Mean Difference (IV, Fixed, 95% CI)	0.06 [‐0.12, 0.24]
14 S‐cholesterol (total) (mmol/L)	1	60	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.88, 1.08]
14.1 Colchicine versus placebo	1	60	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.88, 1.08]
14.2 Colchicine + UDCA versus placebo/no intervention + UDCA	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
15 Plasma immunoglobulin M (g/L)	4	198	Mean Difference (IV, Fixed, 95% CI)	‐0.49 [‐1.03, 0.06]
15.1 Colchicine versus placebo	1	60	Mean Difference (IV, Fixed, 95% CI)	‐1.5 [‐3.58, 0.58]
15.2 Colchicine + UDCA versus placebo/no intervention + UDCA	3	138	Mean Difference (IV, Fixed, 95% CI)	‐0.41 [‐0.98, 0.15]
16 Prothrombin time (second)	1	57	Mean Difference (IV, Fixed, 95% CI)	‐0.03 [‐0.75, 0.69]
16.1 Colchicine versus placebo	1	57	Mean Difference (IV, Fixed, 95% CI)	‐0.03 [‐0.75, 0.69]
16.2 Colchicine + UDCA versus placebo/no intervention + UDCA	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
17 Liver biopsy findings ‐ dichotomous variables	3		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only
17.1 Worsening of histological stage	3	145	Odds Ratio (M‐H, Fixed, 95% CI)	0.85 [0.41, 1.75]
17.2 Worsening of fibrosis	2	95	Odds Ratio (M‐H, Fixed, 95% CI)	0.60 [0.24, 1.49]
17.3 Worsening of piecemeal necrosis	2	95	Odds Ratio (M‐H, Fixed, 95% CI)	0.58 [0.23, 1.44]
17.4 Worsening of parenchymal inflammation	2	95	Odds Ratio (M‐H, Fixed, 95% CI)	0.69 [0.28, 1.72]
17.5 Worsening of parenchymal necrosis	2	95	Odds Ratio (M‐H, Fixed, 95% CI)	1.00 [0.31, 3.18]
17.6 Worsening of cholestasis	1	37	Odds Ratio (M‐H, Fixed, 95% CI)	0.81 [0.18, 3.69]
17.7 Worsening of granulomas	1	37	Odds Ratio (M‐H, Fixed, 95% CI)	1.4 [0.12, 16.98]
17.8 Worsening of ductular proliferation	1	58	Odds Ratio (M‐H, Fixed, 95% CI)	0.30 [0.06, 1.58]
17.9 Worsening of lobular inflammation	1	58	Odds Ratio (M‐H, Fixed, 95% CI)	0.16 [0.03, 0.80]
17.10 Worsening of cholangitis	1	58	Odds Ratio (M‐H, Fixed, 95% CI)	0.57 [0.13, 2.52]
18 Liver biopsy findings ‐ histological score	1	50	Mean Difference (IV, Fixed, 95% CI)	0.56 [‐0.24, 1.36]
18.1 Colchicine versus placebo	1	50	Mean Difference (IV, Fixed, 95% CI)	0.56 [‐0.24, 1.36]
18.2 Colchicine + UDCA versus placebo/no intervention + UDCA	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
19 Number of patients with adverse events	8	455	Risk Ratio (M‐H, Fixed, 95% CI)	1.45 [0.94, 2.25]
19.1 Colchicine versus placebo	4	241	Risk Ratio (M‐H, Fixed, 95% CI)	1.33 [0.82, 2.15]
19.2 Colchicine + UDCA versus placebo/no intervention + UDCA	4	214	Risk Ratio (M‐H, Fixed, 95% CI)	2.02 [0.72, 5.69]
20 Number of patients with serious adverse events	8	455	Risk Ratio (M‐H, Fixed, 95% CI)	1.17 [0.50, 2.75]
20.1 Colchicine versus placebo	4	241	Risk Ratio (M‐H, Fixed, 95% CI)	1.06 [0.43, 2.59]
20.2 Colchicine + UDCA versus placebo/no intervention + UDCA	4	214	Risk Ratio (M‐H, Fixed, 95% CI)	3.0 [0.13, 71.34]

1.3. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 3 Number of patients who underwent liver transplantation.

1.4. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 4 Number of patients without improvement of pruritus.

1.5. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 5 Number of patients without improvement of fatigue.

1.6. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 6 Number of patients developing liver complications.

1.8. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 8 S‐bilirubin (µmol/L).

1.9. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 9 S‐alkaline phosphatases (ALP)(IU/L).

1.10. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 10 S‐gamma‐glutamyltransferase (GGT)(IU/L).

1.11. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 11 S‐aspartate aminotransferase (AST)(IU/L).

1.12. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 12 S‐alanine aminotransferase (ALT)(IU/L).

1.13. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 13 S‐albumin (g/dL).

1.14. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 14 S‐cholesterol (total) (mmol/L).

1.16. Analysis.

Comparison 1 Colchicine versus placebo/no intervention, Outcome 16 Prothrombin time (second).

Comparison 2. Colchcine ‐ colchicine versus placebo ‐ colchicine (including open label period).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of deaths	2	117	Risk Ratio (M‐H, Fixed, 95% CI)	1.15 [0.76, 1.73]
2 Number of deaths and/or patients who underwent liver transplantation	2	117	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.72, 1.46]

Comparison 3. Colchicine versus ursodeoxycholic acid.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of deaths	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	11.37 [0.66, 196.74]
2 Number of deaths and/or patients who underwent liver transplantation	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	11.37 [0.66, 196.74]
3 Number of patients who underwent liver transplantation	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	3.1 [0.13, 73.14]
4 Number of patients without improvement of pruritus	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	0.78 [0.55, 1.09]
5 Number of patients without improvement of fatigue	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	0.83 [0.70, 0.99]
6 Appearance of liver complications	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
6.1 Development of varices	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.2 Development of ascites	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.3 Hepatic encephalopathy	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	3.1 [0.13, 73.14]
7 S‐bilirubin (µmol/L)	1	59	Mean Difference (IV, Fixed, 95% CI)	3.40 [‐13.26, 20.06]
8 S‐alkaline phosphatases (ALP)(IU/L)	1	59	Mean Difference (IV, Fixed, 95% CI)	378.0 [116.92, 639.08]
9 S‐gamma‐glutamyltransferase (GGT)(IU/L)	1	59	Mean Difference (IV, Fixed, 95% CI)	459.00 [157.57, 760.43]
10 S‐aspartate aminotransferase (AST)(IU/L)	1	59	Mean Difference (IV, Fixed, 95% CI)	19.0 [‐8.86, 46.86]
11 S‐alanine aminotransferase (ALT)(IU/L)	1	59	Mean Difference (IV, Fixed, 95% CI)	24.00 [‐8.62, 56.62]
12 S‐albumin (g/dL)	1	59	Mean Difference (IV, Fixed, 95% CI)	0.04 [‐0.14, 0.22]
13 S‐cholesterol (total) (mmol/L)	1	59	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐1.14, 1.14]
14 Plasma immunoglobulin M (g/L)	1	59	Mean Difference (IV, Fixed, 95% CI)	0.70 [‐0.99, 2.39]
15 Number of patients with adverse events	1		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
15.1 Number of patients with adverse events	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	9.3 [0.52, 165.39]
15.2 Number of patients with serious adverse events	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	5.17 [0.26, 103.21]

3.1. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 1 Number of deaths.

3.2. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 2 Number of deaths and/or patients who underwent liver transplantation.

3.3. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 3 Number of patients who underwent liver transplantation.

3.4. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 4 Number of patients without improvement of pruritus.

3.5. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 5 Number of patients without improvement of fatigue.

3.6. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 6 Appearance of liver complications.

3.7. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 7 S‐bilirubin (µmol/L).

3.8. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 8 S‐alkaline phosphatases (ALP)(IU/L).

3.9. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 9 S‐gamma‐glutamyltransferase (GGT)(IU/L).

3.10. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 10 S‐aspartate aminotransferase (AST)(IU/L).

3.11. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 11 S‐alanine aminotransferase (ALT)(IU/L).

3.12. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 12 S‐albumin (g/dL).

3.13. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 13 S‐cholesterol (total) (mmol/L).

3.14. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 14 Plasma immunoglobulin M (g/L).

3.15. Analysis.

Comparison 3 Colchicine versus ursodeoxycholic acid, Outcome 15 Number of patients with adverse events.

Comparison 4. Colchicine versus methotrexate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of deaths and/or patients who underwent liver transplantation	1	85	Risk Ratio (M‐H, Fixed, 95% CI)	0.80 [0.37, 1.73]
2 Pruritus score	1	85	Mean Difference (IV, Fixed, 95% CI)	0.68 [0.25, 1.11]
3 Fatigue score	1	85	Mean Difference (IV, Fixed, 95% CI)	‐0.11 [‐0.54, 0.32]
4 S‐bilirubin (µmol/L) (presented as logtransformed geometric mean)	1	85	Mean Difference (IV, Fixed, 95% CI)	‐0.07 [‐0.37, 0.23]
5 S‐alkaline phosphatases (ALP)(IU/L) (presented as logtransformed geometric mean)	1	85	Mean Difference (IV, Fixed, 95% CI)	0.41 [0.12, 0.70]
6 S‐aspartate aminotransferase (AST)(IU/L) (presented as logtransformed geometric mean)	1	85	Mean Difference (IV, Fixed, 95% CI)	0.06 [‐0.18, 0.30]
7 S‐alanine aminotransferase (ALT)(IU/L) (presented as logtransformed geometric mean)	1	85	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐0.26, 0.26]
8 S‐albumin (g/dL)	1	85	Mean Difference (IV, Fixed, 95% CI)	0.17 [‐0.01, 0.35]
9 S‐cholesterol (total) (mmol/L) (presented as logtransformed geometric mean)	1	85	Mean Difference (IV, Fixed, 95% CI)	0.09 [‐0.03, 0.21]
10 Plasma immunoglobulin M (g/L) (presented as logtransformed geometric mean)	1	83	Mean Difference (IV, Fixed, 95% CI)	0.47 [0.20, 0.74]
11 Prothrombin time (second)	1	85	Mean Difference (IV, Fixed, 95% CI)	‐0.14 [‐0.54, 0.26]
12 Liver biopsy findings ‐ histological stage	1	55	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.60, 0.80]
13 Liver biopsy findings ‐ histological score	1	55	Mean Difference (IV, Fixed, 95% CI)	1.40 [‐1.53, 4.33]
14 Number of patients with adverse events	1	85	Risk Ratio (M‐H, Fixed, 95% CI)	1.47 [0.45, 4.82]

4.1. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 1 Number of deaths and/or patients who underwent liver transplantation.

4.2. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 2 Pruritus score.

4.3. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 3 Fatigue score.

4.4. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 4 S‐bilirubin (µmol/L) (presented as logtransformed geometric mean).

4.5. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 5 S‐alkaline phosphatases (ALP)(IU/L) (presented as logtransformed geometric mean).

4.6. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 6 S‐aspartate aminotransferase (AST)(IU/L) (presented as logtransformed geometric mean).

4.7. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 7 S‐alanine aminotransferase (ALT)(IU/L) (presented as logtransformed geometric mean).

4.8. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 8 S‐albumin (g/dL).

4.9. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 9 S‐cholesterol (total) (mmol/L) (presented as logtransformed geometric mean).

4.10. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 10 Plasma immunoglobulin M (g/L) (presented as logtransformed geometric mean).

4.11. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 11 Prothrombin time (second).

4.12. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 12 Liver biopsy findings ‐ histological stage.

4.13. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 13 Liver biopsy findings ‐ histological score.

4.14. Analysis.

Comparison 4 Colchicine versus methotrexate, Outcome 14 Number of patients with adverse events.

Comparison 5. Subgroup analyses ‐ colchicine versus placebo/no intervention.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of deaths ‐ dose variation	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	1.21 [0.71, 2.06]
1.1 Colchicine ‐ 1mg/day	6	338	Risk Ratio (M‐H, Fixed, 95% CI)	1.36 [0.73, 2.52]
1.2 Colchicine ‐ 1.2mg/day	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	0.83 [0.28, 2.44]
2 Number of deaths ‐ treatment duration	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	1.21 [0.71, 2.06]
2.1 Colchicine for <=2 years duration	5	280	Risk Ratio (M‐H, Fixed, 95% CI)	1.04 [0.56, 1.93]
2.2 Colchicine for > 2 years duration	2	118	Risk Ratio (M‐H, Fixed, 95% CI)	1.82 [0.62, 5.39]
3 Number of deaths ‐ generation of the allocation sequence	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	1.21 [0.71, 2.06]
3.1 Adequate generation of allocation schedule	4	274	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.58, 1.80]
3.2 Unclear or inadequate generation of allocation schedule	3	124	Risk Ratio (M‐H, Fixed, 95% CI)	5.0 [0.61, 41.04]
4 Number of deaths ‐ allocation concealment	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	1.21 [0.71, 2.06]
4.1 Adequate allocation concealment	5	296	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.58, 1.80]
4.2 Unclear or inadequate allocation concealment	2	102	Risk Ratio (M‐H, Fixed, 95% CI)	5.0 [0.61, 41.04]
5 Number of deaths ‐ blinding	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	1.21 [0.71, 2.06]
5.1 Adequate	5	348	Risk Ratio (M‐H, Fixed, 95% CI)	1.12 [0.65, 1.94]
5.2 Unclear or not perfomed	2	50	Risk Ratio (M‐H, Fixed, 95% CI)	5.0 [0.26, 95.61]
6 Number of deaths and/or patients who underwent liver transplantation ‐ dose variation	8	455	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.67, 1.49]
6.1 Colchicine ‐ 1mg/day	6	338	Risk Ratio (M‐H, Fixed, 95% CI)	1.14 [0.65, 2.00]
6.2 Colchicine ‐ 1.2mg/day	2	117	Risk Ratio (M‐H, Fixed, 95% CI)	0.85 [0.48, 1.50]
7 Number of deaths and/or patients who underwent liver transplantation ‐ treatment duration	8	455	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.67, 1.49]
7.1 Colchicine for <= 2 years duration	5	280	Risk Ratio (M‐H, Fixed, 95% CI)	0.94 [0.53, 1.67]
7.2 Colchicine for > 2 years duration	3	175	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [0.61, 1.86]
8 Number of deaths and/or patients who underwent liver transplantation ‐ allocation sequence generation	8	455	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.67, 1.49]
8.1 Adequate generation of allocation schedule	4	274	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.53, 1.52]
8.2 Unclear or inadequate generation of allocation schedule	4	181	Risk Ratio (M‐H, Fixed, 95% CI)	1.19 [0.64, 2.20]
9 Number of deaths and/or patients who underwent liver transplantation ‐ allocation concealment	8	455	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.67, 1.49]
9.1 Adequate allocation concealment	5	296	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.53, 1.52]
9.2 Unclear or inadequate allocation concealment	3	159	Risk Ratio (M‐H, Fixed, 95% CI)	1.19 [0.64, 2.20]
10 Number of deaths and/or patients who underwent liver transplantation ‐ blinding	8	455	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.67, 1.49]
10.1 Adequate	6	405	Risk Ratio (M‐H, Fixed, 95% CI)	0.94 [0.63, 1.42]
10.2 Unclear or not performed	2	50	Risk Ratio (M‐H, Fixed, 95% CI)	5.0 [0.26, 95.61]
11 S‐bilirubin (µmol/L) ‐ reported as arithmetic mean or geometric mean	5		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
11.1 Arithmetic mean	4	202	Mean Difference (IV, Fixed, 95% CI)	‐1.35 [‐4.52, 1.82]
11.2 Geometric mean	1	57	Mean Difference (IV, Fixed, 95% CI)	‐0.44 [‐1.00, 0.12]
12 S‐cholesterol (total) (mmol/L) ‐ reported as arithmetic mean or geometric mean	2		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
12.1 Arithmetic mean	1	60	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.88, 1.08]
12.2 Geometric mean	1	57	Mean Difference (IV, Fixed, 95% CI)	‐0.02 [‐0.18, 0.14]
13 S‐alkaline phosphatase (ALP) (IU/L) ‐ reported as arithmetic mean or geometric mean	5		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
13.1 Arithmetic mean	4	200	Mean Difference (IV, Fixed, 95% CI)	‐54.53 [‐156.45, 47.38]
13.2 Geometric mean	1	57	Mean Difference (IV, Fixed, 95% CI)	‐0.23 [‐0.59, 0.13]

5.4. Analysis.

Comparison 5 Subgroup analyses ‐ colchicine versus placebo/no intervention, Outcome 4 Number of deaths ‐ allocation concealment.

5.7. Analysis.

Comparison 5 Subgroup analyses ‐ colchicine versus placebo/no intervention, Outcome 7 Number of deaths and/or patients who underwent liver transplantation ‐ treatment duration.

5.8. Analysis.

Comparison 5 Subgroup analyses ‐ colchicine versus placebo/no intervention, Outcome 8 Number of deaths and/or patients who underwent liver transplantation ‐ allocation sequence generation.

5.9. Analysis.

Comparison 5 Subgroup analyses ‐ colchicine versus placebo/no intervention, Outcome 9 Number of deaths and/or patients who underwent liver transplantation ‐ allocation concealment.

Comparison 6. Sensitivity analyses ‐ colchicine versus placebo/no intervention.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Number of deaths	7		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
1.1 Likely scenario	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	1.21 [0.71, 2.06]
1.2 Best‐worst‐case scenario	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	0.59 [0.30, 1.15]
1.3 Reported scenario	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	1.13 [0.52, 2.47]
1.4 Worst‐best‐case scenario	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	2.28 [1.17, 4.44]
2 Number of deaths and/or patients who underwent liver transplantation	8		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
2.1 Likely scenario	8	455	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.67, 1.49]
2.2 Best‐worst‐case scenario	8	455	Risk Ratio (M‐H, Fixed, 95% CI)	0.42 [0.24, 0.74]
2.3 Reported scenario	8	455	Risk Ratio (M‐H, Fixed, 95% CI)	0.74 [0.39, 1.39]
2.4 Worst‐best‐case scenario	8	455	Risk Ratio (M‐H, Fixed, 95% CI)	1.75 [1.07, 2.86]
3 Number of deaths and/or patients who underwent liver transplantation (excluding Bodenheimer 1988)	7		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
3.1 Likely scenario	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	1.06 [0.65, 1.75]
3.2 Best‐case scenario	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	0.54 [0.28, 1.01]
3.3 Reported scenario	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.44, 1.82]
3.4 Worst‐case scenario	7	398	Risk Ratio (M‐H, Fixed, 95% CI)	1.76 [0.97, 3.18]

6.2. Analysis.

Comparison 6 Sensitivity analyses ‐ colchicine versus placebo/no intervention, Outcome 2 Number of deaths and/or patients who underwent liver transplantation.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Almasio 2000.

Methods	Generation of the allocation sequence: adequate, computer‐generated list. Allocation concealment: adequate, central unit. Blinding: adequate, double‐blinding, indistinguishable placebo. Follow‐up: adequate, 6/90 patients dropped out: 2 on UDCA plus placebo, 4 on UDCA plus colchicine.
Participants	Country: Italy. 90 patients (9 males and 81 females, being 55.5 ± 10.9 years in the UDCA/P group and 53.3 ± 10.2 years in UDCA/C group). Inclusion criteria: 1. An established diagnosis of primary biliary cirrhosis according to Taal et al. 2. Pruritus. 3. Serum bilirubin exceeding 2 mg/dL. 4. Histological diagnosis of cirrhosis. Exclusion criteria: 1. Ascites. 2. Gastrointestinal bleeding or encephalopathy. 3. Serum bilirubin levels exceeding 10 mg/dL. 4. Evidence of malignant conditions or of other major diseases unrelated to PBC. 5. Alcohol abuse. 6. Previous treatment with colchicine or immunosuppressant agents. 7. Low compliance.
Interventions	a) Colchicine plus UDCA: 1 mg/day colchicine plus 250 mg UDCA twice daily. b) Placebo plus UDCA: placebo plus 250 mg UDCA twice daily. Duration of medication: 3 years.
Outcomes	1. Biochemical variables. 2. Ig M. 3. Mayo score. 4. Major clinical events: death, liver transplantation, decompensation of liver disease, doubling of bilirubin. 5. Liver biopsy findings.
Notes	1. It was a multicenter‐study (six centres). 2. Sent letter (4 Nov. 2002). P. L. Almasio responded and provided additional data on liver biochemical variables. 3. This trial included the 44 patients described by Battezzati 2001, which followed patients for up to 10 years of treatment: 6/44 patients dropped out: 4 in UDCA+placebo, 2 in UDCA+colchicine.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Bodenheimer 1988.

Methods	Generation of the allocation sequence: unclear. Allocation concealment: unclear. Blinding: adequate, double‐blinding, identically appearing placebo. Follow‐up: adequate, 14/57 withdrew and lost to follow‐up during the blind period of the trial (6 in placebo, 8 in colchicine). In addition, one patient in the control group lost to follow‐up in the opened label period. Sample size estimation: no.
Participants	Country: USA. 57 patients (5 males and 52 females; mean age: 53 years in colchicine group and 51 years in placebo group). Inclusion criteria: 1. History of chronic cholestatic liver disease. 2. Liver biopsy results compatible with PBC.
Interventions	a) Colchicine: 0.6 mg, twice daily. b) Placebo: Identically appearing placebo. Duration of medication: 4 years.
Outcomes	1. Biochemical variables. 2. Immunological variables. 3. Histologic parameters proven by liver biopsy. 4. Number of death and number of patients undergoing liver transplantation. 5. Adverse events: diarrhoea, etc.
Notes	1. Patients assigned to placebo at entry were crossed to opened label colchicine for 4 additional years after the first 4‐year double blind interventions. The results of this trial were published by Zifroni 1991. 2. Sent letter (4 Nov. 2002), but no response received.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Goddard 1995.

Methods	Generation of the allocation sequence: unclear. Allocation concealment: unclear. Blinding: unclear. Follow‐up: inadequate.
Participants	Country: UK Sample size: 57. Inclusion criteria: unclear.
Interventions	a) Colchicine: 1mg/day colchicine. b) UDCA: 10mg/kg/day UDCA. c) Colchicine plus UDCA. d) Placebo. Duration of treatment: 30 months.
Outcomes	Biochemical variables.
Notes	1. Published as an abstract. 2. Sent letter (4 Nov. 2002), but no response received.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Ikeda 1996.

Methods	Generation of the allocation sequence: inadequate, a consecutive case number. Allocation concealment: adequate, sealed envelope. Blinding: not performed. Follow‐up: adequate, no patients withdrew/lost to follow‐up/drop‐out.
Participants	Country: Japan. 22 patients (3 males and 19 females; being 59.5 ± 3 years in UDCA/C group and 66.5 ± 3 years in UDCA group). Inclusion criteria: 1. Elevation of alkaline phosphatase over the upper limit of normal. 2. AMA. 3. Compatible histological appearance of liver biopsy specimens. 4. Radiological or ultrasonographic evidence that the bile ducts were patent.
Interventions	a) Colchicine plus UDCA: 1 mg/day colchicine plus 600 mg/day UDCA. b) UDCA alone: 600 mg/day UDCA. Duration of combined medication: 2 years.
Outcomes	1. Biochemical variables. 2. Adverse events: diarrhoea. 3. Clinical findings: pruritus, oesophageal varices. 4. Major clinical events: death, liver transplantation.
Notes	1. Before randomisation, all the patients were treated 600 mg/day UDCA for 30 months. 2. Sent letter (4 Nov. 2002). T. Ikeda responded and provided the information on trial design, clinical findings, adverse events, and liver biochemical variables.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Kaplan 1986.

Methods	Generation of the allocation sequence: adequate, randomisation scheme. Allocation concealment: adequate, a single study monitor. Blinding: adequate, double‐blinding identically appearing placebo. Follow‐up: adequate, 8/60 patients were classified as drop‐outs: five in placebo group, three in colchicine group.
Participants	Country: USA. 60 patients (3 males and 57 females; mean age was not given). Inclusion criteria: 1. A positive test for antimitochondrial antibody. 2. Liver‐biopsy proven PBC. 3. Radiologic or ultrasonographic evidence that bile ducts were patent.
Interventions	a) Colchicine: 0.6 mg colchicine twice daily. b) Placebo: Identically appearing placebo. Duration of blinded medication: two years. Duration of open label medication: the following two years.
Outcomes	1. Biochemical variables. 2. Clinical findings. 3. Liver histology score. 4. Cumulative mortality. 5. Adverse events: diarrhoea.
Notes	1. At the end of the two‐year double‐blind period, each patient was placed in an open‐lable trial of colchicine, 0.6 mg twice daily, for additional two years. 2. Sent letter (4 Nov. 2002), but no response received.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Kaplan 1999.

Methods	Generation of the allocation sequence: unclear. Allocation concealment: adequate, a single study monitor. Blinding: adequate, double‐blinding and double‐dummy. Follow‐up: inadequate.
Participants	Country: USA. 85 patients ( 3 males and 82 females; being 51 ± 1.4 years in colchicine group and 51 ± 1.5 years in methotrexate group). Inclusion criteria: 1. Serum ALP level of at least 2 times greater than the upper limit of normal. 2. Serum bilirubin level not greater than 10 mg/dL. 3. Liver biopsy performed consistent with PBC. 4. Radiological or ultrasonic evidence.
Interventions	a) Colchicine: 0.6 mg colchicine twice daily. b) Methotrexate: 15 mg/week, 5 mg every 12 hours 3 times. Duration of medication: 10 years.
Outcomes	1. Biochemical variables. 2. IgM. 3. Pruritus and fatigue. 4. Liver histological evidence.
Notes	1. It is an interim analysis of a ten‐year trial. 2. 2/87 withdrew from the trial immediately after randomisation before they received any drugs, did not return for follow‐up testings, and were not included in the analyses. Ten patients dropped out of the trial. The reasons were specified, but the number in each group was not given. 3. Sent letter (4 Nov. 2002). M. Kaplan responded, but did not provide additional information. The final results of this ten‐year trial are waiting publication.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Poupon 1996.

Methods	Generation of the allocation sequence: unclear. Allocation concealment: unclear. Blinding: adequate, double‐blinding, identically appearing placebo. Follow‐up: adequate, 2 patients dropped out: 2 in UDCA + colchicine.
Participants	Countries: France and Canada. 74 patients (11 males and 63 females; being 55 ± 2 years in UDCA/C group and 52 ± 2 years in UDCA/P group). Inclusion criteria: 1. Biopsy‐proven PBC. 2. No less than eight months previous treatment with UDCA(13‐15 mg/kg/day). 3. ALP activity more than 1.5 times the upper limit of normal. Exclusion criteria: 1. Drug therapy (except UDCA) for PBC during the 6 months (colchicine, azathioprine, chlorambucil, corticosteroids, D‐penicillamine, and cyclosporine). 2. Serum bilirubin concentration greater than 100umol/L. 3. A serum albumin concentration less than 25 g/L. 4. Past or active bleeding form oesophageal varices. 5. Ascites. 6. Other identified cause of liver of biliary diseases. 7. Excessive alcohol consumption (greater than 50 g/day). 8. Severe intercurrent disease. 9. Age older than 75 years.
Interventions	a) Colchicine plus UDCA: 1 mg/day colchicine, 5 days/week plus UDCA (13 to 15 mg/kg/day). b) Placebo plus UDCA: identically appearing placebo plus UDCA (13 to 15 mg/kg/day). Duration of intervention and follow‐up: 2 years.
Outcomes	1. Clinical findings. 2. Laboratory findings, including bilirubin level. 3. Serum markers of liver fibrosis. 4. Histologic parameters, including the degree of fibrosis. 5. Sulphobromophthalein pharmakinetics. 6. Clinical complications. 7. Adverse events: peripheral polyneuropathy.
Notes	1. This was a multicenter trial (10 study centres) and it included a subsample (22/74 patients) trial designed by Huet 1996 (only published as an abstract) in which all patients were given colchicine plus UDCA for additional 2 years at the end of the two‐year double‐blind period. 2. Sent letter (4 Nov. 2002), but no response received.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Raedsch 1993.

Methods	Generation of the allocation sequence: unclear. Allocation concealment: unclear. Blinding: unclear. Follow‐up: adequate, 2/28 patients dropped out: 2 in UDCA plus colchicine.
Participants	Country: Germany. All 28 patients were females with a mean age of 54 years. Inclusion criteria: 1. Blood biochemistry. 2. Specific AMA. 3. Compatible liver histology.
Interventions	a) Colchicine plus UDCA: 1 mg/day colchicine plus 10 to 12 mg/kg/day UDCA. b) Placebo plus UDCA: placebo plus 10 to 12mg/kg/day UDCA . Duration of medication: 3 years.
Outcomes	1. Biochemical variables. 2. Immunological variables. 3. Clinical symptoms. 4. Histological parameters.
Notes	1. All patients were pretreated with UDCA 10 to 12 mg/kg/day for 12 months. 2. Sent letter (4 Nov. 2002), but no response received.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Vuoristo 1995.

Methods	Generation of allocation sequence: adequate, computerized randomisation number. Allocation concealment: adequate, sealed envelopes. Blinding: adequate, double‐blinding, placebo with identical looking and film‐coated. Follow‐up: adequate, 6/90 drop‐outs: 3 in the placebo group, 3 in the colchicine group.
Participants	Country: Finland. 90 patients (16 males, 74 females; mean age: 57, 56 and 52 years in placebo, colchicine and UDCA group, respectively). PBC defined as: elevated alkaline phosphatases, liver biopsy compatible with PBC, and positive AMA. End‐stage PBC and patients treated with drugs that might affect prognosis were excluded.
Interventions	a) Colchicine: 1 mg/day colchicine. b) UDCA: 12 to 15 mg/kg/day UDCA. c) Placebo. Duration of medication: two years.
Outcomes	1. Major clinical events: death, liver transplantation, etc.. 2. Clinical findings. 3. Liver biochemistry. 4. Liver histology.
Notes	1. Sent letter (4 Nov. 2002). Vuoristo responded and provided additional information on trial design, clinical findings and liver biochemical variables.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Warnes 1987.

Methods	Generation of the allocation sequence: adequate, random tables. Allocation of concealment: adequate, staff pharmacist. Blinding: adequate, double‐blinding, identical placebo. Follow‐up: adequate, 10/64 patients withdrew: 8 on colchicine, 2 on placebo.
Participants	Country: UK. Sample size: 89. Inclusion criteria: 1. A raised serum ALP. 2. A positive AMA test. 3. Liver histology compatible with, or diagnostic of PBC.
Interventions	a) Colchicine: 500ug, twice daily. b) Placebo: Identical placebo. Duration of medication is 12 months. Median duration of follow‐up at the time of analysis was 23 months in the colchicine group and 15 months in the placebo group.
Outcomes	1. Biochemical findings. 2. Immunological findings. 3. Liver histological findings. 4. Survival data. 5. Adverse events: diarrhoea, upper gastrointestinal symptoms, peripheral neuropathy, proteinuria, etc.
Notes	1. Pair‐matched study, patients being matched on the basis of age and serum bilirubin. 2. Biochemical, immunological and histological findings at 12 months were compared, whilst survival data were compared up to 18 months. 3. Patients' age and sex ratio were not described. 4. Sent letter (4 Nov. 2002), but no response received.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Low risk	A ‐ Adequate

Warnes 1996.

Methods	Generation of the allocation sequence: unclear. Allocation concealment: unclear. Blinding: unclear. Follow‐up: inadequate.
Participants	Country: UK Sample size: 89. Inclusion criteria: unclear.
Interventions	a) Colchicine b) Placebo.
Outcomes	Biochemical findings: serum bilirubin, galactose elimination capacity and serum albumin.
Notes	1. Published as an abstract. 2. Sent letter (4 Nov. 2002), but no response received.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment?	Unclear risk	B ‐ Unclear

Ig: immunoglobulin 
 UDCA: ursodeoxycholic acid 
 PBC: primary biliary cirrhosis 
 AMA: antimitochondrial antibody 
 ALP: alkaline phosphatases

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Klion 1990	An observational study. It compared the risk score (R) using the Mayo model for a group of patients treated with colchicine using their pre‐treatment period as control.
Koldinger 1980	A case series of five patients with PBC for periods ranging from 12 to 40 months.
Shibata 1992	A non‐randomised trial. They divided twelve patients with PBC into two groups, one with UDCA and one with colchicine for three months. After three months both groups received combination therapy.

Contributions of authors

YG performed the searches, selected trials for inclusion, wrote to authors and pharmaceutical companies, performed data extraction and data analyses, and drafted the protocol and the systematic review. CG formulated the idea of this review and revised the protocol, selected trials for inclusion, validated, solved discrepancy of data extraction, and revised the review.

Sources of support

Internal sources

• Copenhagen Trial Unit, Centre for Clinical Intervention Research, H:S Rigshospitalet, Denmark.

External sources

• No sources of support supplied

Declarations of interest

None known. We have no affiliations or financial contracts with companies producing the drugs reviewed in this review.

Edited (no change to conclusions)