Pharmacological interventions for primary biliary cholangitis

Summary of findings for the main comparison. Ursodeoxycholic acid (UDCA) versus no intervention for primary biliary cholangitis

UDCA versus no intervention for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: UDCA Comparison: no intervention
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	No intervention	UDCA
Mortality at maximal follow‐up Follow‐up: 12 to 89 months	208 per 1000	206 per 1000 (136 to 301)	OR 0.99 (0.60 to 1.64)	734 (6 trials)	⊕⊝⊝⊝ Very low^1,2
Serious adverse events (proportion) Follow‐up: 12 to 41 months	There were no events in either group			380 (3 trials)	⊕⊝⊝⊝ Very low^1,2,3
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio; UDCA: ursodeoxycholic acid.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels). ³ There was moderate heterogeneity (downgraded by one level).

Summary of findings 2. Azathioprine versus no intervention for primary biliary cholangitis

Azathioprine versus no intervention for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: azathioprine Comparison: no intervention
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	No intervention	Azathioprine
Mortality at maximal follow‐up Follow‐up: 63 months in 1 trial and not stated in 1 trial	208 per 1000	128 per 1000 (78 to 205)	OR 0.56 (0.32 to 0.98)	224 (2 trials)	⊕⊝⊝⊝ Very low^1,2
Serious adverse events (proportion)	None of the trials reported this outcome.
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels).

Summary of findings 3. Colchicine versus no intervention for primary biliary cholangitis

Colchicine versus no intervention for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: colchicine Comparison: no intervention
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	No intervention	Colchicine
Mortality at maximal follow‐up Follow‐up: 12 to 24 months	208 per 1000	168 per 1000 (78 to 327)	OR 0.77 (0.32 to 1.85)	122 (2 trials)	⊕⊝⊝⊝ Very low^1,2
Serious adverse events (proportion) Follow‐up: 12 months	There were no events in either group			64 (1 trial)	⊕⊝⊝⊝ Very low^1,2,3
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels). ³ There was moderate heterogeneity (downgraded by one level).

Summary of findings 4. Ciclosporin versus no intervention for primary biliary cholangitis

Ciclosporin versus no intervention for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: ciclosporin Comparison: no intervention
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	No intervention	Ciclosporin
Mortality at maximal follow‐up Follow‐up: 31 to 35 months	208 per 1000	188 per 1000 (118 to 283)	OR 0.88 (0.51 to 1.50)	390 (3 trials)	⊕⊝⊝⊝ Very low^1,2
Serious adverse events (proportion)	None of the trials reported this outcome.
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels).

Summary of findings 5. D‐Penicillamine versus no intervention for primary biliary cholangitis

D‐Penicillamine versus no intervention for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: D‐penicillamine Comparison: no intervention
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	No intervention	D‐Penicillamine
Mortality at maximal follow‐up (Follow‐up 24 to 66 months)	208 per 1000	191 per 1000 (130 to 274)	OR 0.90 (0.57 to 1.44)	423 (5 trials)	⊕⊝⊝⊝ Very low^1,2,3
Serious adverse events (proportion) (Follow‐up 24 months)	4 per 1000	104 per 1000 (6 to 679)	OR 28.77 (1.57 to 526.67)	52 (1 trial)	⊕⊝⊝⊝ Very low^1,2,3
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels). ³ There was moderate heterogeneity (downgraded by one level).

Summary of findings 6. Colchicine plus ursodeoxycholic acid (UDCA) versus UDCA for primary biliary cholangitis

Colchicine plus UDCA versus UDCA for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: colchicine + UDCA Comparison: UDCA
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	UDCA	Colchicine + UDCA
Mortality at maximal follow‐up Follow‐up: 24 months in 1 trial; not reported in 1 trial	110 per 1000	185 per 1000 (45 to 524)	OR 1.84 (0.38 to 8.91)	158 (2 trials)	⊕⊝⊝⊝ Very low^1,2
Serious adverse events (proportion) Follow‐up: not stated	14 per 1000	42 per 1000 (2 to 526)	OR 3.08 (0.12 to 78.14)	74 (1 trial)	⊕⊝⊝⊝ Very low^1,2,3
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio; UDCA: ursodeoxycholic acid.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels). ³ There was moderate heterogeneity (downgraded by one level).

Summary of findings 7. Methotrexate plus ursodeoxycholic acid (UDCA) versus UDCA for primary biliary cholangitis

Methotrexate plus UDCA versus UDCA for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: methotrexate + UDCA Comparison: UDCA
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	UDCA	Methotrexate + UDCA
Mortality at maximal follow‐up Follow‐up: 11 to 91 months	110 per 1000	126 per 1000 (64 to 237)	OR 1.17 (0.55 to 2.51)	290 (2 trials)	⊕⊝⊝⊝ Very low^1,2
Serious adverse events (proportion)	None of the trials reported this outcome.
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio; UDCA: ursodeoxycholic acid.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels). ³ There was moderate heterogeneity (downgraded by one level).

Background

Description of the condition

Primary biliary cholangitis (previously named primary biliary cirrhosis) is a chronic liver disease caused by the destruction of small intrahepatic bile ducts resulting in stasis of bile (cholestasis), liver fibrosis, and liver cirrhosis (NCBI 2014). There is global variation in the incidence and prevalence of primary biliary cholangitis with annual incidence varying from 1.6 to 3.2 per 100,000 people and prevalence varying from 5 to 38 per 100,000 people, with a trend of increasing incidence and prevalence in many countries (Metcalf 1997; Boberg 1998; Kim 2000; Sood 2004; Lazaridis 2007; Pla 2007; Rautiainen 2007; Myers 2009; Baldursdottir 2012; Boonstra 2014). It is more common in women, particularly aged 25 to 40 years (Metcalf 1997; Kim 2000; Gershwin 2005; Pla 2007; Myers 2009; Baldursdottir 2012). The mean age at diagnosis is 40 to 60 years (Kim 2000; Parikh‐Patel 2001; Gershwin 2005; Myers 2009; Baldursdottir 2012).

The aetiology of primary biliary cholangitis is unclear. The associations with primary biliary cholangitis include family history of primary biliary cholangitis, Sjögren's syndrome (autoimmune disease characterised by dry mouth and dry eyes), systemic lupus erythematosus (autoimmune connective tissue disorder), autoimmune thyroid disease, multiple sclerosis (autoimmune disorder of the central nervous system), scleroderma (autoimmune disease affecting the skin and internal organs), polymyositis (chronic inflammation of the muscles, possibly an autoimmune disease), history of cigarette smoking, history of hair dye use, and urinary tract infections (Parikh‐Patel 2001; Gershwin 2005; Lazaridis 2007; Prince 2010; Lammert 2013). People with primary biliary cholangitis have other coexisting autoimmune disorders such as rheumatoid arthritis, systemic lupus erythematosus, autoimmune thyroid disease, multiple sclerosis, scleroderma, and polymyositis (Parikh‐Patel 2001; Gershwin 2005; Prince 2010; Lammert 2013). Although the strong association between personal and family history of autoimmune diseases suggests that primary biliary cholangitis may have an autoimmune aetiology, the clustering of primary biliary cholangitis in certain areas and associations between primary biliary cholangitis and hair dye use, past smoking, and history of urinary tract infections have prompted people to consider environmental factors such as toxins and infections as possible aetiologies or triggering factors for primary biliary cholangitis (Leung 2005; Dronamraju 2010; Prince 2010; Selmi 2010).

A significant proportion of people with primary biliary cholangitis are asymptomatic at the time of diagnosis (up to about 60% in some studies (Pla 2007)). Itching and fatigue are the most common symptoms (Pla 2007; Myers 2009). Other ways of clinical presentation include Raynaud's syndrome (bluish discolouration of the fingers and toes due to vasospasm in response to cold or emotional stress); features of portal hypertension; osteoporosis; high cholesterol (particularly high ratio of high‐density lipoprotein cholesterol (which is considered protective for the heart) to low‐density lipoprotein cholesterol); and rarely deficiencies of vitamin A, vitamin D, vitamin E, and vitamin K (Kim 2000; Gershwin 2005; Pla 2007; Myers 2009; Baldursdottir 2012). Approximately 3% to 8% of people require liver transplantation in about five to six years from diagnosis (Kim 2000; Lindor 2009; Myers 2009; Baldursdottir 2012). Approximately 3% to 4% of people with primary biliary cholangitis die every year, usually because of liver‐related causes such as decompensated liver disease or hepatocellular carcinoma (Rautiainen 2007; Myers 2009). Overall, approximately 21% to 50% of people are dead in about 10 to 11 years from diagnosis (Kim 2000; Rautiainen 2007; Myers 2009; Floreani 2011; Baldursdottir 2012).

The diagnosis of primary biliary cholangitis is made in the presence of any two of the following three criteria (Lindor 2009).

Elevation of alkaline phosphatases.
Presence of antimitochondrial antibody (AMA).
Liver biopsy demonstrating non‐suppurative destructive cholangitis and destruction of interlobular bile ducts.

Some variations of primary biliary cholangitis are AMA‐negative primary biliary cholangitis that requires liver biopsy for establishing the diagnosis and the primary biliary cholangitis ‐ autoimmune hepatitis overlap syndrome (Lindor 2009). However, there is currently no strong evidence that the course of the disease is different between the classic primary biliary cholangitis and these variants (Lindor 2009).

Description of the intervention

Various pharmacological interventions have been tried to treat people with primary biliary cholangitis. These include bile acids such as ursodeoxycholic acid (UDCA) (Kaplan 2004; Combes 2005; Rautiainen 2005; Rudic 2012a); fibrates such as bezafibrate (Kurihara 2000; Rudic 2012b); immunosuppressants or immunomodulators such as glucocorticosteroids (Prince 2005; Rautiainen 2005), colchicine (Almasio 2000; Gong 2004a; Kaplan 2004), methotrexate (Kaplan 2004; Combes 2005; Giljaca 2010), azathioprine (Gong 2007a), ciclosporin (Gong 2007b), chlorambucil (Li Wei 2012), mycophenolate mofetil (Jones 1999; Talwalkar 2005), and thalidomide (McCormick 1994); and copper‐chelating agents such as D‐penicillamine (Gong 2004b) and tetrathiomolybdate (Askari 2010). Several other interventions such as bisphosphonates and hormonal replacement to prevent or treat osteoporosis (Ormarsdottir 2004; Rudic 2011a; Rudic 2011b; Guanabens 2013); antidepressants such as fluoxetine and fluvoxamine to overcome fatigue (Ter Borg 2004; Talwalkar 2006); cholesterol‐lowering agents such as simvastatin to decrease the high cholesterol (Cash 2013); and cholestyramine, rifampicin, and S‐adenosyl methionine for pruritus (Bergasa 2000) have been evaluated for control of various symptoms. Liver transplantation is performed in some people with decompensated liver disease due to primary biliary cholangitis (Kim 2000; Lindor 2009; Myers 2009; Baldursdottir 2012).

How the intervention might work

Certain bile acids are protective while other bile acids are harmful to hepatocytes (liver cells), cholangiocytes (cells that line the bile duct), and gastrointestinal cells lining the oesophagus and stomach (Perez 2009). Bile acids such as UDCA may protect the cholangiocytes from the damage caused by hydrophobic bile acids by decreasing the oxidative stress (by direct antioxidant effect or an increase in antioxidant defences) (Paumgartner 2002; Perez 2009). Bile acids also stimulate the secretion of bile acids from hepatocytes, thereby decreasing their stasis and the resulting damage to the cells and inhibit apoptosis (programmed cell death) (Paumgartner 2002; Perez 2009). Fibrates inactivate hydrophobic bile acids and, therefore, decrease the damage to the cells (Kurihara 2000). Since primary biliary cholangitis is considered an autoimmune disorder, altering the immunity and inflammatory response using glucocorticoids and other immunosuppressants may decrease the damage resulting from the inflammatory response. D‐Penicillamine and tetrathiomolybdate might remove the excess copper, thereby protecting the cells from the damage caused by copper accumulation. They also have antifibrotic properties (Song 2008). In this Cochrane Review, we included only pharmacological interventions aimed at controlling the liver disease (i.e. we excluded symptomatic treatments, lifestyle modifications, and liver transplantation).

Why it is important to do this review

The optimal pharmacological treatment of primary biliary cholangitis is unknown. Currently, both the European Association for the Study of the Liver (EASL) and American Association for the Study of Liver Diseases (AASLD) recommend UDCA for the management of primary biliary cholangitis (EASL 2009; Lindor 2009). However, one Cochrane Review that compared UDCA versus placebo or no intervention reported that there was no survival or symptomatic benefit for UDCA (Rudic 2012a). Therefore, there is clearly a discordance between the evidence and guideline recommendation. Network meta‐analysis allows combination of the direct evidence and indirect evidence, and allows ranking of different interventions in terms of the different outcomes (Salanti 2011; Salanti 2012). There has been no Cochrane Review on the different pharmacological interventions for primary biliary cholangitis. This systematic review and attempted network meta‐analysis provides the best level of evidence for the role of different interventions used in the treatment of people with primary biliary cholangitis.

Objectives

To assess the comparative benefits and harms of different pharmacological interventions in the treatment of primary biliary cholangitis through a network meta‐analysis and to generate rankings of the available pharmacological interventions according to their safety and efficacy. However, it was not possible to assess whether the potential effect modifiers were similar across different comparisons. Therefore, we did not perform the network meta‐analysis, and, instead, assessed the comparative benefits and harms of different interventions using standard Cochrane methodology.

When more trials become available with adequate description of potential effect modifiers, we will attempt to conduct network meta‐analysis to generate rankings of the available interventions according to their safety and efficacy. This is why we retained the planned methodology for network meta‐analysis in our Appendix 1. Once data appear allowing for the conduct of network meta‐analysis, this Appendix 1 will be moved back into the Methods section.

Methods

Criteria for considering studies for this review

Types of studies

We considered randomised clinical trials only for this network meta‐analysis, irrespective of the language, publication status, or date of publication. We excluded studies of other design because of the risk of bias in such studies. We are all aware that such exclusions make us focus much more on potential benefits and not fully assess the risks of serious adverse events as well as risks of adverse events.

Types of participants

We included randomised clinical trials with participants with primary biliary cholangitis irrespective of the method of diagnosis of the disease or the presence of symptoms. We excluded randomised clinical trials in which participants had undergone liver transplantation previously.

Types of interventions

Any of the following pharmacological interventions that are possible treatments used either alone or in combination for primary biliary cholangitis and can be compared with each other or with placebo or no intervention.

The interventions that we considered were:

UDCA;
obeticholic acid;
bezafibrate;
glucocorticosteroids;
colchicine;
methotrexate;
azathioprine;
ciclosporin;
chlorambucil;
mycophenolate mofetil;
thalidomide;
D‐penicillamine;
tetrathiomolybdate.

The above list was not exhaustive. If we identified pharmacological interventions that we were not aware of, we considered them as eligible and included them in the review if they were used primarily for the treatment of primary biliary cholangitis.

Types of outcome measures

We assessed the comparative benefits and harms of available pharmacological interventions aimed at treating people with primary biliary cholangitis for the following outcomes.

Primary outcomes

Mortality at maximal follow‐up.
Mortality:
- short‐term mortality (up to one year);
- medium‐term mortality (one to five years).
Adverse events (within three months after cessation of treatment). Depending on the availability of data, we attempted to classify adverse events as serious or non‐serious. We defined a non‐serious adverse event as any untoward medical occurrence not necessarily having a causal relationship with the treatment but resulting in a dose reduction or discontinuation of treatment (any time after commencement of treatment) (ICH‐GCP 1997). We defined a serious adverse event as any event that would increase mortality; was life threatening; required hospitalisation; resulted in persistent or significant disability; was a congenital anomaly/birth defect; or any important medical event that might jeopardise the person or require intervention to prevent it. We used the definition used by study authors for non‐serious and serious adverse events:
- proportion of participants with serious adverse events;
- number of serious adverse events;
- proportion of participants with any type of adverse event;
- number of any type of adverse event.
Health‐related quality of life as defined in the included trials using a validated scale such as EQ‐5D or 36‐item Short Form (SF‐36) (EuroQol 2014; Ware 2014):
- short‐term (up to one year);
- medium‐term (one to five years);
- long‐term (beyond five years).

We considered long‐term quality of life more important than short‐term or medium‐term quality of life, although short‐term and medium‐term quality of life are also important primary outcomes.

Secondary outcomes

Liver transplantation (maximal follow‐up):
- proportion of participants with liver transplantation;
- time to liver transplantation.
Decompensated liver disease (maximal follow‐up):
- proportion of participants with decompensated liver disease;
- time to liver decompensation.
Cirrhosis (maximal follow‐up):
- proportion of participants with cirrhosis;
- time to cirrhosis.
Hepatocellular carcinoma (maximal follow‐up).

Search methods for identification of studies

Electronic searches

We searched the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, Embase, and Science Citation Index Expanded (Royle 2003) from inception to 27 February 2017 for randomised clinical trials comparing two or more of the above interventions without applying any language restrictions. We searched for all possible comparisons formed by the interventions of interest. To identify further ongoing or completed trials, we also searched the World Health Organization International Clinical Trials Registry Platform Search Portal (apps.who.int/trialsearch/), which searches various trial registers, including ISRCTN and ClinicalTrials.gov. Appendix 2 shows the search strategies we used.

Searching other resources

We searched the references of the identified trials and existing Cochrane Reviews on primary biliary cholangitis to identify additional trials for inclusion.

Data collection and analysis

Selection of studies

Two review authors (KG and FS) independently identified the trials for inclusion by screening the titles and abstracts. We sought full‐text articles for any references that at least one of the review authors identified for potential inclusion. We selected trials for inclusion based on the full‐text articles. We listed the excluded full‐text references with reasons for their exclusion in the Characteristics of excluded studies table. We have also listed any ongoing trials identified primarily through the search of the clinical trial registers for further follow‐up. We resolved discrepancies through discussion.

Data extraction and management

Two review authors (KG and FS or LHE) independently extracted the following data.

Outcome data (for each outcome and for each treatment arm whenever applicable):
- number of participants randomised;
- number of participants included for the analysis;
- number of participants with events for binary outcomes, mean and standard deviation for continuous outcomes, number of events for count outcomes, and the number of participants with events and the mean follow‐up period for time‐to‐event outcomes;
- definition of outcomes or scale used if appropriate.
Data on potential effect modifiers:
- participant characteristics such as age, sex, comorbidities, proportion of symptomatic participants, proportion with AMA‐positive status, proportion of participants with overlap syndrome, and responders;
- details of the intervention and control (including dose, frequency, and duration);
- risk of bias (assessment of risk of bias in included studies).
Other data:
- year and language of publication;
- country in which the participants were recruited;
- year(s) in which the trial was conducted;
- inclusion and exclusion criteria;
- follow‐up time points of the outcome.

If available, we planned to obtain the data separately for symptomatic participants and asymptomatic participants from the report. If available, we also planned to obtain the data separately for people with AMA‐positive status and people with AMA‐negative status and for responders and non‐responders separately. We sought unclear or missing information by contacting the trial authors. If there was any doubt whether trials shared the same participants, completely or partially (by identifying common authors and centres), we attempted to contact the trial authors to clarify whether the trial report was duplicated. We resolved any differences in opinion through discussion.

Assessment of risk of bias in included studies

We followed the guidance given in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) and described in the Cochrane Hepato‐Biliary Module (Gluud 2017) to assess the risk of bias in included trials. Specifically, we assessed the risk of bias in included trials for the following domains using the methods below (Schulz 1995; Moher 1998; Kjaergard 2001; Wood 2008; Savović 2012a; Savović 2012b; Lundh 2017).

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 were adequate if performed by an independent person not otherwise involved in the trial.
Unclear risk of bias: the method of sequence generation was not specified.
High risk of bias: the sequence generation method was not random.

Allocation concealment

Low risk of bias: 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 (e.g. if the allocation sequence was hidden in sequentially numbered, opaque, and sealed envelopes).
Unclear 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 and personnel

Low risk of bias: any of the following: no blinding or incomplete blinding, but the review authors judged that the outcome was not likely to be influenced by lack of blinding; or blinding of participants and key study personnel ensured, and it was unlikely that the blinding could have been broken.
Unclear risk of bias: any of the following: insufficient information to permit judgement of 'low risk' or 'high risk'; or the trial did not address this outcome.
High risk of bias: any of the following: no blinding or incomplete blinding, and the outcome was likely to be influenced by lack of blinding; or blinding of key study participants and personnel attempted, but likely that the blinding could have been broken, and the outcome was likely to be influenced by lack of blinding.

Blinding of outcome assessors

Low risk of bias: any of the following: no blinding of outcome assessment, but the review authors judged that the outcome measurement was not likely to be influenced by lack of blinding; or blinding of outcome assessment ensured, and unlikely that the blinding could have been broken.
Unclear risk of bias: any of the following: insufficient information to permit judgement of 'low risk' or 'high risk'; or the trial did not address this outcome.
High risk of bias: any of the following: no blinding of outcome assessment, and the outcome measurement was likely to be influenced by lack of blinding; or blinding of outcome assessment, but likely that the blinding could have been broken, and the outcome measurement was likely to be influenced by lack of blinding.

Incomplete outcome data

Low risk of bias: missing data were unlikely to make treatment effects depart from plausible values. Sufficient methods, such as multiple imputation, were employed to handle missing data.
Unclear 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: the trial reported at least the following predefined outcomes: mortality, decompensated liver disease, requirement for transplantation, or treatment‐related adverse events. If the original trial protocol was available, the outcomes should have been those called for in that protocol. If the trial protocol was obtained from a trial registry (e.g. www.clinicaltrials.gov), the outcomes sought should have been those enumerated in the original protocol if the trial protocol was registered before or at the time that the trial was begun. If the trial protocol was registered after the trial was begun, those outcomes were not considered to be reliable.
Unclear risk: not all predefined, or clinically relevant and reasonably expected, outcomes were reported fully, or it was unclear whether data on these outcomes were recorded or not.
High risk: one or more predefined or clinically relevant and reasonably expected outcomes were not reported, even though data on these outcomes were likely to have been available and even recorded.

For‐profit bias

Low risk of bias: the trial appeared to be free of industry sponsorship or other type of for‐profit support that may manipulate the trial design, conductance, or results of the trial.
Unclear risk of bias: the trial may or may not have been free of for‐profit bias as no information on clinical trial support or sponsorship was provided.
High risk of bias: the trial was sponsored by industry or received other type of for‐profit support.

Other bias

Low risk of bias: the trial appeared to be free of other components (e.g. inappropriate control or dose or administration of control) that could put it at risk of bias.
Unclear risk of bias: the trial may or may not have been free of other components that could put it at risk of bias.
High risk of bias: there are other factors in the trial that could put it at risk of bias (e.g. inappropriate control or dose or administration of control).

We considered a trial at low risk of bias if we assessed the trial to be at low risk of bias across all domains. Otherwise, we considered trials to be at unclear risk of bias or at high risk of bias regarding one or more domains as at high risk of bias.

Measures of treatment effect

For dichotomous variables (e.g. short‐term and medium‐term mortality, liver transplantation, proportion of participants with adverse events, decompensated liver disease, cirrhosis, or hepatocellular carcinoma), we calculated the odds ratio (OR) with 95% confidence intervals (CI). For continuous variables (e.g. quality of life reported on the same scale), we planned to calculate the mean difference with 95% CI. We planned to use standardised mean difference values with 95% CI for quality of life if included trials used different scales. For count outcomes (e.g. number of adverse events), we calculated the rate ratio with 95% CI. For time‐to‐event data (e.g. mortality at maximal follow‐up or requirement for liver transplantation, time to liver decompensation, and time to cirrhosis), we planned to use the hazard ratio (HR) with 95% CIs. We also calculated Trial Sequential Analysis‐adjusted CI to control random errors (Thorlund 2011).

Unit of analysis issues

The unit of analysis was people with primary biliary cholangitis according to the intervention group to which they were randomly assigned.

Cluster randomised clinical trials

We found no cluster randomised clinical trials. However, if we had found them, we would have included them provided that the effect estimate adjusted for cluster correlation was available.

Cross‐over randomised clinical trials

If we found cross‐over randomised clinical trials, we included the outcomes after the period of first intervention only since primary biliary cholangitis is a chronic disease and the interventions could potentially have a residual effect.

Trials with multiple treatment groups

We collected data for all trial intervention groups that met our inclusion criteria.

Dealing with missing data

We performed an intention‐to‐treat analysis whenever possible (Newell 1992). Otherwise, we used the data that were available to us (e.g. a trial may have reported only per‐protocol analysis results). As such per‐protocol analyses may be biased, we planned to conduct best‐worst case scenario analysis (good outcome in intervention group and bad outcome in control group) and worst‐best case scenario analysis (bad outcome in intervention group and good outcome in control group) as sensitivity analyses whenever possible.

For continuous outcomes, we planned to impute the standard deviation from P values according to guidance given in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). If the data were likely to be normally distributed, we planned to use the median for meta‐analysis when the mean was not available. If it was not possible to calculate the standard deviation from the P value or the CIs, we planned to impute the standard deviation using the largest standard deviation in other trials for that outcome. This form of imputation may decrease the weight of the study for calculation of mean differences and may bias the effect estimate to no effect for calculation of standardised mean differences (Higgins 2011).

Assessment of heterogeneity

We assessed clinical and methodological heterogeneity by carefully examining the characteristics and design of included trials. We assessed the presence of clinical heterogeneity by comparing effect estimates in the presence or absence of symptoms, the presence or absence of AMA, responders versus non‐responders, and the doses of the pharmacological interventions. Different study designs and risk of bias may contribute to methodological heterogeneity. We used the I² test and Chi² test for heterogeneity, and overlapping of CIs to assess heterogeneity.

Assessment of reporting biases

We planned to use visual asymmetry on a funnel plot to explore reporting bias in the presence of at least 10 trials that could be included for a direct comparison (Egger 1997; Macaskill 2001). In the presence of heterogeneity that could be explained by subgroup analysis, we planned to produce a funnel plot for each subgroup in the presence of an adequate number of trials (at least 10 trials). We planned to use the linear regression approach described by Egger 1997 to determine funnel plot asymmetry.

We also considered selective reporting as evidence of reporting bias.

Data synthesis

We performed the meta‐analyses according to the recommendations of Cochrane (Higgins 2011), using the software package Review Manager 5 (RevMan 2014). We used a random‐effects model (DerSimonian 1986) and a fixed‐effect model (DeMets 1987). In the case of a discrepancy between the two models, we reported both results; otherwise, we reported only the results from the fixed‐effect model.

Calculation of required information size and Trial Sequential Analysis

For calculation of the required information size, see Appendix 3. We performed Trial Sequential Analysis to control the risk of random errors when there were at least two trials included for mortality at maximal follow‐up, serious adverse events (proportion) and health‐related quality of life, the three outcomes that determine whether an intervention should be used (Wetterslev 2008; Thorlund 2011; TSA 2011; Wetterslev 2017). We used an alpha error as per guidance of Jakobsen 2014, power of 90% (beta error of 10%), a relative risk reduction of 20%, a control group proportion observed in the trials, and the diversity observed in the meta‐analysis.

Subgroup analysis and investigation of heterogeneity

We planned to assess the differences in the effect estimates between the following subgroups.

Trials at low risk of bias compared to trials at high risk of bias.
Participants with symptomatic compared to participants with asymptomatic primary biliary cholangitis.
AMA‐positive participants compared to AMA‐negative participants.
Responders compared to non‐responders to bile acids.
Different doses of pharmacological interventions. For example, various doses of UDCA used in randomised clinical trials include 5 mg/kg to 7 mg/kg, 13 mg/kg to 15 mg/kg (moderate dose), and 23 mg/kg to 25 mg/kg (high dose) (Angulo 1999a; Lindor 1997).

We planned to use the Chi² test for subgroup differences to identify subgroup differences.

Sensitivity analysis

If a trial reported only per‐protocol analysis results, we planned to re‐analyse the results using the best‐worst case scenario and worst‐best case scenario analyses as sensitivity analyses whenever possible.

Presentation of results and GRADE assessments

We reported mortality, serious adverse events, and health‐related quality of life, the three most important outcomes that determine the use of an intervention in a 'Summary of findings' table format, downgrading the quality of evidence for risk of bias, inconsistency, indirectness, imprecision, and publication bias using GRADE (Guyatt 2011). We have presented the 'Summary of findings' tables for all comparisons in which two trials were included for one of mortality at maximal follow‐up, serious adverse events, or health‐related quality of life.

Results

Description of studies

Results of the search

We identified 5592 references through electronic searches of CENTRAL (n = 1104), MEDLINE (n = 2383), Embase (n = 604), Science Citation Index Expanded (n = 1362), World Health Organization International Clinical Trials Registry Platform (n = 88), and ClinicalTrials.gov (n = 51). After the removal of 1249 duplicates we obtained 4343 references. We then excluded 3973 clearly irrelevant references through screening titles and reading abstracts. We retrieved 370 references for further assessment. No references were identified through scanning reference lists of the identified randomised trials. We excluded 117 references for the reasons stated in the Characteristics of excluded studies table. Nine references are an ongoing trial without any interim data (ChiCTR‐IPR‐16008935; EUCTR2015‐002698‐39‐GB; NCT02308111; NCT02701166; NCT02823353; NCT02823366; NCT02937012; NCT02943447; NCT02965911). We were unable to obtain the full texts for two references (O'Brian 1990; Zaman 2006). In total, 242 references (74 trials) met the inclusion criteria. The reference flow is summarised in the study flow diagram (Figure 1).

Figure 1

Study flow diagram.

Included studies

The 74 trials that met the inclusion criteria for this review included 5902 participants. Some 28 trials did not contribute any information for this review leaving 4274 participants included in one or more outcomes in the review (Bodenheimer 1988; Arora 1990; Oka 1990; Smart 1990; Poupon 1991a; Senior 1991; Battezzati 1993; Manzillo 1993a; Manzillo 1993b; Bobadilla 1994; Goddard 1994; Lim 1994; McCormick 1994; Steenbergen 1994; Lindor 1997; Kaplan 1999; Leuschner 1999; Nakai 2000; Mazzarella 2002; Ueno 2005; Iwasaki 2008a; Iwasaki 2008b; Askari 2010; Liberopoulos 2010; Cash 2013; Bowlus 2014; Kowdley 2014a; Mayo 2015). In the main review unstratified by the dose of UDCA or obeticholic acid, 4060 participants were included in one or more outcomes in the review. The mean or median age of the participants ranged from 46 to 64 years in the trials that reported this information. The proportion of females ranged from 77.8% to 100% in the trials that reported this information. The proportion of participants with symptoms varied from 19.9% to 100% in the trials that reported this information. The proportion of participants who were AMA positive ranged from 80.8% to 100% in the trials that reported this information. Ten trials included non‐responders to bile acids only (Van Hoogstraten 1998; Wolfhagen 1998; Kanda 2003; Ueno 2005; Iwasaki 2008b; Mason 2008; Liberopoulos 2010; Hirschfield 2015; Hosonuma 2015; Nevens 2016). The remaining trials did not state whether they included responders or non‐responders, or both. However, it appeared that most trials included participants who had not received previous treatments or regardless of the previous treatments received. The interventions, controls, number of participants included in each trial, and the follow‐up period reported in the different trials are listed in Table 1.

Table 1. Characteristics of included studies arranged by comparison

Study name	No participants randomised	Post‐randomisation dropouts	No participants for whom outcome was reported	Intervention(s)	Control	Mean follow‐up period (months)
Smart 1990	20	Not stated	20	Antioxidants	No intervention	Not stated
Christensen 1985	248	63	185	Azathioprine	No intervention	63
Heathcote 1976	45	6	39	Azathioprine	No intervention	Not stated
Hoofnagle 1986	24	0	24	Chlorambucil	No intervention	52
Bodenheimer 1988	57	10	47	Colchicine	No intervention	33
Kaplan 1986	60	3	57	Colchicine	No intervention	24
Warnes 1987	64	Not stated	64*	Colchicine	No intervention	19 (median)
Bobadilla 1994	40	Not stated	40	Colchicine + UDCA	No intervention	12
Lombard 1993	349	0	349	Ciclosporin	No intervention	31 (median)
Minuk 1988	12	0	12	Ciclosporin	No intervention	Not stated
Wiesner 1990	40	11	29	Ciclosporin	No intervention	35 (median)
Dickson 1985	309	82	227	D‐Penicillamine	No intervention	60 (median)
Epstein 1979	98	Not stated	98	D‐Penicillamine	No intervention	66
Macklon 1982	60	0	60	D‐Penicillamine	No intervention	37
Matloff 1982	52	0	52	D‐Penicillamine	No intervention	24
Neuberger 1985	189	Not stated	189	D‐Penicillamine	No intervention	Not stated
Taal 1983	24	Not stated	24	D‐Penicillamine	No intervention	18
Triger 1980	35	Not stated	35	D‐Penicillamine	No intervention	Not stated
Mitchison 1989	36	0	36	Glucocorticosteroids	No intervention	36
Ueno 2005	20	Not stated	20	Lamivudine	No intervention	Not stated
Mitchison 1993	104	3	101	Malotilate	No intervention	25 (median)
Hendrickse 1999	60	Not stated	60	Methotrexate	No intervention	68
Steenbergen 1994	14	Not stated	14	Methotrexate + UDCA	No intervention	24
Mayo 2015	45	3	42	NGM282	No intervention	Not stated
Bowlus 2014	216	Not stated	216	Obeticholic acid	No intervention	12
Hirschfield 2015	165	0	165	Obeticholic acid	No intervention	3
Kowdley 2014a	59	Not stated	59	Obeticholic acid	No intervention	Not stated
Manzillo 1993a	32	Not stated	32	S‐Adenosyl methionine	No intervention	1
Manzillo 1993b	6	Not stated	6	S‐Adenosyl methionine	No intervention	2
Cash 2013	21	8	13	Simvastatin	No intervention	12
Askari 2010	28	0	28	Tetrathiomolybdate	No intervention	Not stated
McCormick 1994	18	0	18	Thalidomide	No intervention	Not stated
Arora 1990	9	Not stated	9	UDCA	No intervention	5
Battezzati 1993	88	2	86	UDCA	No intervention	6
Combes 1995a	151	0	151	UDCA	No intervention	24
Eriksson 1997	116	15	101	UDCA	No intervention	24
Heathcote 1994	222	Not stated	222	UDCA	No intervention	24
Leuschner 1989	20	0	18	UDCA	No intervention	12
Lim 1994	32	Not stated	32	UDCA	No intervention	Not stated
Lindor 1994	180	10	170	UDCA	No intervention	24
Oka 1990	52	7	45	UDCA	No intervention	Not stated
Papatheodoridis 2002	92	6	86	UDCA	No intervention	89
Pares 2000	192	0	192	UDCA	No intervention	41 (median)
Poupon 1991a	149	3	146	UDCA	No intervention	Not stated
Senior 1991	20	1	19	UDCA	No intervention	18
Turner 1994	46	0	46	UDCA	No intervention	24
Goddard 1994	57	Not stated	57	Intervention 1: UDCA Intervention 2: colchicine Intervention 3: colchicine + UDCA	No intervention	15
Wolfhagen 1998	50	Not stated	50	Azathioprine + glucocorticosteroids + UDCA	UDCA	12
Iwasaki 2008a	45	Not stated	45	Bezafibrate	UDCA	12
Kurihara 2000	24	Not stated	24	Bezafibrate	UDCA	Not stated
Hosonuma 2015	27	0	27	Bezafibrate + UDCA	UDCA	96
Iwasaki 2008b	22	Not stated	22	Bezafibrate + UDCA	UDCA	12
Kanda 2003	22	0	22	Bezafibrate + UDCA	UDCA	7
Nakai 2000	23	Not stated	23	Bezafibrate + UDCA	UDCA	12
Almasio 2000	90	6	84	Colchicine + UDCA	UDCA	Not stated
Ikeda 1996	22	0	22	Colchicine + UDCA	UDCA	24
Poupon 1996	74	Not stated	74	Colchicine + UDCA	UDCA	24
Raedsch 1993	28	8	20	Colchicine + UDCA	UDCA	24
Yokomori 2001	11	Not stated	11	Colestilan + UDCA	UDCA	Not stated
Liberopoulos 2010	10	Not stated	10	Fenofibrate + UDCA	UDCA	Not stated
Leuschner 1999	40	0	39	Glucocorticosteroids + UDCA	UDCA	24
Rautiainen 2005	77	8	69	Glucocorticosteroids + UDCA	UDCA	36
Gao 2012	79	Not stated	79	Intervention 1: glucocorticosteroids + UDCA Intervention 2: azathioprine + UDCA	UDCA	Not stated
Mason 2008	59	0	59	Lamivudine + zidovudine + UDCA	UDCA	6
Combes 2005	265	0	265	Methotrexate + UDCA	UDCA	91 (median)
Gonzalezkoch 1997	25	Not stated	25	Methotrexate + UDCA	UDCA	11
Nevens 2016	217	Not stated	216	Obeticholic acid + UDCA	UDCA	12
Ferri 1993	30	0	30	TUDCA	UDCA	6
Ma 2016	199	8	191	TUDCA	UDCA	6
Kaplan 1999	87	2	85	Colchicine	Methotrexate	24
Comparison of doses
Lindor 1997	150	Not stated	150	Intervention 1: UDCA (high) Intervention 2: UDCA (moderate)	UDCA (low)	12
Angulo 1999a	155	Not stated	155	Intervention 1: UDCA (high) Intervention 2: UDCA (moderate)	UDCA (low)	12
Van Hoogstraten 1998	61	2	59	UDCA (moderate)	UDCA (low)	Not stated
Mazzarella 2002	42	Not stated	42	UDCA (high)	UDCA (moderate)	72

TUDCA: taurodeoxycholic acid; UDCA: ursodeoxycholic acid.

Source of funding: nine trials receive no additional funding or were funded by parties with no vested interest in the results (Heathcote 1976; Hoofnagle 1986; Almasio 2000; Nakai 2000; Iwasaki 2008a; Iwasaki 2008b; Askari 2010; Cash 2013; Hosonuma 2015). Thirty‐one trials were partially or fully funded by the pharmaceutical companies that would benefit based on the results of the trial (Triger 1980; Matloff 1982; Christensen 1985; Dickson 1985; Bodenheimer 1988; Minuk 1988; Oka 1990; Wiesner 1990; Poupon 1991a; Senior 1991; Lombard 1993; Mitchison 1993; Heathcote 1994; Lindor 1994; McCormick 1994; Combes 1995a; Poupon 1996; Eriksson 1997; Van Hoogstraten 1998; Wolfhagen 1998; Leuschner 1999; Pares 2000; Papatheodoridis 2002; Combes 2005; Rautiainen 2005; Mason 2008; Bowlus 2014; Kowdley 2014a; Mayo 2015; Ma 2016; Nevens 2016). The source of funding was not available from the 34 remaining trials.

Excluded studies

The reasons for exclusion are summarised in the Characteristics of excluded studies table. While the reasons for exclusion for most references were self‐explanatory, the reasons for exclusion of 15 references required some explanation (Poupon 1994; Lindor 1995a; Emond 1996; Lindor 1996; Angulo 1999b; Angulo 1999c; Degott 1999; Corpechot 2000; Jorgensen 2002; Kaplan 2004; Combes 2005b; Leung 2010; Leung 2011; Kowdley 2015; Carbone 2016). These 15 references were long‐term follow‐up reports of included trials, but the randomisation was not maintained and the 'no intervention' group received the intervention. While this is acceptable if some participants crossed over for specific reasons in an intention‐to‐treat analysis, it is not acceptable if the cross‐over from one group to another was done in a systematic manner. Therefore, we excluded these references.

Risk of bias in included studies

The risk of bias is summarised in Figure 2, Figure 3, and Table 2.

Figure 2

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

Figure 3

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

See: Summary of findings for the main comparison Ursodeoxycholic acid (UDCA) versus no intervention for primary biliary cholangitis; Summary of findings 2 Azathioprine versus no intervention for primary biliary cholangitis; Summary of findings 3 Colchicine versus no intervention for primary biliary cholangitis; Summary of findings 4 Ciclosporin versus no intervention for primary biliary cholangitis; Summary of findings 5 D‐Penicillamine versus no intervention for primary biliary cholangitis; Summary of findings 6 Colchicine plus ursodeoxycholic acid (UDCA) versus UDCA for primary biliary cholangitis; Summary of findings 7 Methotrexate plus ursodeoxycholic acid (UDCA) versus UDCA for primary biliary cholangitis

Table 2. Risk of bias arranged according to comparisons

Name of studies	Intervention(s)	Control	Random sequence generation	Allocation concealment	Blinding of participants and health professionals	Blinding of outcome assessors	Missing outcome bias	Selective outcome reporting	For‐profit bias	Other bias
Smart 1990	Antioxidants	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Christensen 1985	Azathioprine	No intervention	Unclear	Unclear	Low	Low	High	High	High	Low
Heathcote 1976	Azathioprine	No intervention	Unclear	Unclear	High	High	High	High	Low	Low
Hoofnagle 1986	Chlorambucil	No intervention	Low	Low	High	High	Low	Low	Low	Low
Bodenheimer 1988	Colchicine	No intervention	Unclear	Unclear	Low	Low	High	High	High	Low
Kaplan 1986	Colchicine	No intervention	Unclear	Unclear	Unclear	Unclear	High	High	Unclear	Low
Warnes 1987	Colchicine	No intervention	Low	Low	Low	Low	Unclear	Low	Unclear	Low
Bobadilla 1994	Colchicine + UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Lombard 1993	Ciclosporin	No intervention	Unclear	Unclear	Low	Low	Low	Low	High	Low
Minuk 1988	Ciclosporin	No intervention	Unclear	Unclear	Low	Unclear	Unclear	Low	High	Low
Wiesner 1990	Ciclosporin	No intervention	Unclear	Unclear	Low	Low	Unclear	Low	High	Low
Dickson 1985	D‐Penicillamine	No intervention	Low	Low	Low	Low	High	High	High	High
Epstein 1979	D‐Penicillamine	No intervention	Unclear	Unclear	High	High	Unclear	High	Unclear	Low
Macklon 1982	D‐Penicillamine	No intervention	Unclear	Unclear	Unclear	Unclear	Low	Low	Unclear	Low
Matloff 1982	D‐Penicillamine	No intervention	Unclear	Unclear	Unclear	Unclear	Low	Low	High	Low
Neuberger 1985	D‐Penicillamine	No intervention	Unclear	Low	Low	Low	Unclear	High	Unclear	Low
Taal 1983	D‐Penicillamine	No intervention	Unclear	Unclear	Low	Low	Unclear	Low	Unclear	Low
Triger 1980	D‐Penicillamine	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	High	Low
Mitchison 1989	Glucocorticosteroids	No intervention	Low	Low	High	High	Low	High	Unclear	Low
Ueno 2005	Lamivudine	No intervention	Unclear	Unclear	Low	Low	Unclear	High	Unclear	Low
Mitchison 1993	Malotilate	No intervention	Low	Low	Low	Low	High	Low	High	Low
Hendrickse 1999	Methotrexate	No intervention	Low	Low	Unclear	Unclear	Unclear	High	Unclear	Low
Steenbergen 1994	Methotrexate + UDCA	No intervention	Low	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Mayo 2015	NGM282	No intervention	Unclear	Unclear	Unclear	Unclear	High	High	High	Low
Bowlus 2014	Obeticholic acid	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	High	Low
Hirschfield 2015	Obeticholic acid	No intervention	Low	Unclear	Low	Low	Low	High	Unclear	High
Kowdley 2011	Obeticholic acid	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	High	Low
Manzillo 1993a	S‐Adenosyl methionine	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Manzillo 1993b	S‐Adenosyl methionine	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Cash 2013	Simvastatin	No intervention	Unclear	Low	High	High	High	High	Low	High
Askari 2010	Tetrathiomolybdate	No intervention	Low	Low	Low	Low	Low	High	Low	High
McCormick 1994	Thalidomide	No intervention	Unclear	Unclear	Low	Low	Low	High	High	Low
Arora 1990	UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Battezzati 1993	UDCA	No intervention	Low	Low	Low	Low	High	High	Unclear	Low
Combes 1995a	UDCA	No intervention	Unclear	Unclear	Low	Low	Low	High	High	Low
Eriksson 1997	UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	High	High	High	Low
Heathcote 1994	UDCA	No intervention	Unclear	Low	Low	Low	Unclear	High	High	Low
Leuschner 1989	UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	High	Low	Unclear	Low
Lim 1994	UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Lindor 1994	UDCA	No intervention	Unclear	Unclear	Low	Low	High	Low	High	Low
Oka 1990	UDCA	No intervention	Unclear	Low	Low	Low	High	High	High	Low
Papatheodoridis 2002	UDCA	No intervention	Low	Low	High	High	High	High	High	High
Pares 2000	UDCA	No intervention	Unclear	Unclear	Low	Low	Unclear	Low	High	Low
Poupon 1991a	UDCA	No intervention	Unclear	Unclear	Low	Low	High	High	High	Low
Senior 1991	UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	High	High	High	Low
Turner 1994	UDCA	No intervention	Unclear	Unclear	Low	Low	Low	High	Unclear	Low
Goddard 1994	Intervention 1: UDCA Intervention 2: colchicine Intervention 3: colchicine + UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Wolfhagen 1998	Azathioprine + glucocorticosteroids + UDCA	UDCA	Low	Low	Low	Low	Unclear	High	High	Low
Iwasaki 2008a	Bezafibrate	UDCA	Unclear	Low	High	High	Unclear	High	Low	Low
Kurihara 2000	Bezafibrate	UDCA	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Hosonuma 2015	Bezafibrate + UDCA	UDCA	Low	Low	High	High	Low	Low	Low	Low
Iwasaki 2008b	Bezafibrate + UDCA	UDCA	Unclear	Low	High	High	Unclear	High	Low	Low
Kanda 2003	Bezafibrate + UDCA	UDCA	Unclear	Unclear	Unclear	Unclear	Low	High	Unclear	Low
Nakai 2000	Bezafibrate + UDCA	UDCA	Unclear	Unclear	Unclear	Unclear	Unclear	High	Low	Low
Almasio 2000	Colchicine + UDCA	UDCA	Low	Low	Low	Low	High	High	Low	Low
Ikeda 1996	Colchicine + UDCA	UDCA	Unclear	Unclear	Unclear	Unclear	Low	High	Unclear	High
Poupon 1996	Colchicine + UDCA	UDCA	Unclear	Unclear	Low	Low	Unclear	Low	High	Low
Raedsch 1993	Colchicine + UDCA	UDCA	Unclear	Unclear	Unclear	Unclear	High	High	Unclear	Low
Yokomori 2001	Colestilan + UDCA	UDCA	Unclear	Unclear	High	High	Unclear	High	Unclear	Low
Liberopoulos 2010	Fenofibrate + UDCA	UDCA	Unclear	Unclear	High	High	Unclear	High	Unclear	Low
Leuschner 1999	Glucocorticosteroids + UDCA	UDCA	Low	Unclear	Unclear	Unclear	High	High	High	Low
Rautiainen 2005	Glucocorticosteroids + UDCA	UDCA	Unclear	Unclear	High	High	High	High	High	Low
Gao 2012	Glucocorticosteroids + UDCA Azathioprine + UDCA	UDCA	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Mason 2008	Lamivudine + zidovudine + UDCA	UDCA	Low	Low	Low	Low	Unclear	High	High	Low
Combes 2005	Methotrexate + UDCA	UDCA	Unclear	Unclear	Unclear	Unclear	Low	High	High	Low
Gonzalezkoch 1997	Methotrexate + UDCA	UDCA	Unclear	Low	Unclear	Unclear	Unclear	Low	Unclear	Low
Nevens 2016	Obeticholic acid + UDCA	UDCA	Low	Low	Low	Low	High	Low	High	Low
Ferri 1993	TUDCA	UDCA	Unclear	Unclear	Unclear	Unclear	Low	High	Unclear	Low
Ma 2016	TUDCA	UDCA	Low	Low	Low	Low	Unclear	High	High	Low
Kaplan 1999	Colchicine	Methotrexate	Unclear	Unclear	Low	Low	High	High	Unclear	Low
Comparison of doses
Lindor 1997	Intervention 1: UDCA (high) Intervention 2: UDCA (moderate)	UDCA (low)	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Angulo 1999a	Intervention 1: UDCA (high) Intervention 2: UDCA (moderate)	UDCA (low)	Low	Low	Low	Low	Unclear	Low	Unclear	Low
Van Hoogstraten 1998	UDCA (moderate)	UDCA (low)	Low	Low	High	High	Unclear	High	High	Low
Mazzarella 2002	UDCA (high)	UDCA (moderate)	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low

TUDCA: taurodeoxycholic acid; UDCA: ursodeoxycholic acid.

Allocation

Twenty trials were at low risk of bias due to random sequence generation (Dickson 1985; Hoofnagle 1986; Warnes 1987; Mitchison 1989; Battezzati 1993; Mitchison 1993; Steenbergen 1994; Van Hoogstraten 1998; Wolfhagen 1998; Angulo 1999a; Hendrickse 1999; Leuschner 1999; Almasio 2000; Papatheodoridis 2002; Mason 2008; Askari 2010; Hirschfield 2015; Hosonuma 2015; Ma 2016; Nevens 2016). The remaining trials were at unclear risk of bias.

Twenty‐four trials were at low risk of bias due allocation concealment (Dickson 1985; Neuberger 1985; Hoofnagle 1986; Warnes 1987; Mitchison 1989; Oka 1990; Battezzati 1993; Mitchison 1993; Heathcote 1994; Gonzalezkoch 1997; Van Hoogstraten 1998; Wolfhagen 1998; Angulo 1999a; Hendrickse 1999; Almasio 2000; Papatheodoridis 2002; Iwasaki 2008a; Iwasaki 2008b; Mason 2008; Askari 2010; Cash 2013; Hosonuma 2015; Ma 2016; Nevens 2016). The remaining trials were at unclear risk of bias.

Sixteen trials were at low risk of both random sequence generation bias and allocation concealment bias (Dickson 1985; Warnes 1987; Mitchison 1989; Battezzati 1993; Mitchison 1993; Van Hoogstraten 1998; Wolfhagen 1998; Angulo 1999a; Hendrickse 1999; Almasio 2000; Papatheodoridis 2002; Mason 2008; Askari 2010; Hosonuma 2015; Ma 2016; Nevens 2016); these trials were considered to be at low risk of selection bias. The remaining trials were at unclear risk of selection bias.

Blinding

Thirty trials were at low risk of performance bias (Taal 1983; Christensen 1985; Dickson 1985; Neuberger 1985; Warnes 1987; Bodenheimer 1988; Minuk 1988; Oka 1990; Wiesner 1990; Poupon 1991a; Battezzati 1993; Lombard 1993; Mitchison 1993; Heathcote 1994; Lindor 1994; McCormick 1994; Turner 1994; Combes 1995a; Poupon 1996; Wolfhagen 1998; Angulo 1999a; Kaplan 1999; Almasio 2000; Pares 2000; Ueno 2005; Mason 2008; Askari 2010; Hirschfield 2015; Ma 2016; Nevens 2016). Thirteen trials were at high risk of performance bias (Heathcote 1976; Epstein 1979; Hoofnagle 1986; Mitchison 1989; Van Hoogstraten 1998; Yokomori 2001; Papatheodoridis 2002; Rautiainen 2005; Iwasaki 2008a; Iwasaki 2008b; Liberopoulos 2010; Cash 2013; Hosonuma 2015). The remaining trials were at unclear risk of performance bias.

Twenty‐nine trials were at low risk of detection bias (Taal 1983; Christensen 1985; Dickson 1985; Neuberger 1985; Warnes 1987; Bodenheimer 1988; Oka 1990; Wiesner 1990; Poupon 1991a; Battezzati 1993; Lombard 1993; Mitchison 1993; Heathcote 1994; Lindor 1994; McCormick 1994; Turner 1994; Combes 1995a; Poupon 1996; Wolfhagen 1998; Angulo 1999a; Kaplan 1999; Almasio 2000; Pares 2000; Ueno 2005; Mason 2008; Askari 2010; Hirschfield 2015; Ma 2016; Nevens 2016). Thirteen trials were at high risk of detection bias (Heathcote 1976; Epstein 1979; Hoofnagle 1986; Mitchison 1989; Van Hoogstraten 1998; Yokomori 2001; Papatheodoridis 2002; Rautiainen 2005; Iwasaki 2008a; Iwasaki 2008b; Liberopoulos 2010; Cash 2013; Hosonuma 2015). The remaining trials were at unclear risk of detection bias.

Twenty‐nine trials were at low risk of performance bias and detection bias (Taal 1983; Christensen 1985; Dickson 1985; Neuberger 1985; Warnes 1987; Bodenheimer 1988; Oka 1990; Wiesner 1990; Poupon 1991a; Battezzati 1993; Lombard 1993; Mitchison 1993; Heathcote 1994; Lindor 1994; McCormick 1994; Turner 1994; Combes 1995a; Poupon 1996; Wolfhagen 1998; Angulo 1999a; Kaplan 1999; Almasio 2000; Pares 2000; Ueno 2005; Mason 2008; Askari 2010; Hirschfield 2015; Ma 2016; Nevens 2016). Thirteen trials were at high risk of performance bias and detection bias (Heathcote 1976; Epstein 1979; Hoofnagle 1986; Mitchison 1989; Van Hoogstraten 1998; Yokomori 2001; Papatheodoridis 2002; Rautiainen 2005; Iwasaki 2008a; Iwasaki 2008b; Liberopoulos 2010; Cash 2013; Hosonuma 2015). The remaining trials were at unclear risk of performance and detection bias.

Incomplete outcome data

Fifteen trials were at low risk of attrition bias (Macklon 1982; Matloff 1982; Hoofnagle 1986; Mitchison 1989; Ferri 1993; Lombard 1993; McCormick 1994; Turner 1994; Combes 1995a; Ikeda 1996; Kanda 2003; Combes 2005; Askari 2010; Hirschfield 2015; Hosonuma 2015). Twenty‐two trials were at high risk of attrition bias due to dropouts which may have been related to the intervention that the participant received (Heathcote 1976; Christensen 1985; Dickson 1985; Kaplan 1986; Bodenheimer 1988; Leuschner 1989; Oka 1990; Poupon 1991a; Senior 1991; Battezzati 1993; Mitchison 1993; Raedsch 1993; Lindor 1994; Eriksson 1997; Kaplan 1999; Leuschner 1999; Almasio 2000; Papatheodoridis 2002; Rautiainen 2005; Cash 2013; Mayo 2015; Nevens 2016). The remaining trials were at unclear risk of attrition bias.

Selective reporting

We were unable to find any protocols published prior to the full study reports. Seventeen trials were at low risk of due to selecting outcome reporting (Macklon 1982; Matloff 1982; Taal 1983; Hoofnagle 1986; Warnes 1987; Minuk 1988; Leuschner 1989; Wiesner 1990; Lombard 1993; Mitchison 1993; Lindor 1994; Poupon 1996; Gonzalezkoch 1997; Angulo 1999a; Pares 2000; Hosonuma 2015; Nevens 2016). The remaining trials were at high risk of bias due to selective reporting (reporting bias).

Other potential sources of bias

For profit bias: nine trials receive no additional funding or were funded by parties with no vested interest in the results and were at low risk of for‐profit bias (Heathcote 1976; Hoofnagle 1986; Almasio 2000; Nakai 2000; Iwasaki 2008a; Iwasaki 2008b; Askari 2010; Cash 2013; Hosonuma 2015). Thirty‐one trials partially or fully funded by the pharmaceutical companies that would benefit based on the results of the trial were at high risk of for‐profit bias (Triger 1980; Matloff 1982; Christensen 1985; Dickson 1985; Bodenheimer 1988; Minuk 1988; Oka 1990; Wiesner 1990; Poupon 1991a; Senior 1991; Lombard 1993; Mitchison 1993; Heathcote 1994; Lindor 1994; McCormick 1994; Combes 1995a; Poupon 1996; Eriksson 1997; Van Hoogstraten 1998; Wolfhagen 1998; Leuschner 1999; Pares 2000; Papatheodoridis 2002; Combes 2005; Rautiainen 2005; Mason 2008; Bowlus 2014; Kowdley 2014a; Mayo 2015; Ma 2016; Nevens 2016). The remaining trials were at unclear risk of for‐profit bias.

Six trials were at high risk of other bias: authors presented the results of only a subgroup of participants without explaining the reason for this approach (Dickson 1985; Ikeda 1996); a significant proportion of participants crossed over from placebo to UDCA (Papatheodoridis 2002); it was unclear whether the participants continued to take UDCA in both groups (Askari 2010); participants continued to take varying doses of UDCA (Hirschfield 2015); and participants were allowed to continue previous prescriptions for primary biliary cholangitis (it was unclear whether this was balanced across groups) (Cash 2013). The remaining trials were at low risk of other bias.

Overall risk of bias

All trials were at high risk of bias in one or more domains.

Effects of interventions

Mortality at maximal follow‐up

Twenty‐eight trials including 2823 participants reported mortality at maximal follow‐up (Heathcote 1976; Epstein 1979; Macklon 1982; Matloff 1982; Taal 1983; Christensen 1985; Neuberger 1985; Hoofnagle 1986; Kaplan 1986; Warnes 1987; Minuk 1988; Leuschner 1989; Mitchison 1989; Wiesner 1990; Lombard 1993; Mitchison 1993; Heathcote 1994; Lindor 1994; Turner 1994; Poupon 1996; Gonzalezkoch 1997; Hendrickse 1999; Almasio 2000; Pares 2000; Papatheodoridis 2002; Combes 2005; Hosonuma 2015; Nevens 2016). The period of follow‐up in these trials varied between 11 and 96 months. The proportion of people with mortality (maximal follow‐up) was higher in the methotrexate group (adjusted proportion: 23.3%) than in the no intervention group (1/30 (3.3%)) (OR 8.83, 95% CI 1.01 to 76.96; 60 participants; 1 trial). The proportion of people with mortality (maximal follow‐up) was lower in the azathioprine group (adjusted proportion: 53.5%) than in the no intervention group (72/107 (67.3%)) (OR 0.56, 95% CI 0.32 to 0.98; 224 participants; 2 trials; I² = 0%). There was no evidence of a difference in any of the remaining comparisons (Analysis 1.1).

Mortality (up to one year)

Eight trials including 655 participants reported mortality (up to year) (Heathcote 1976; Neuberger 1985; Warnes 1987; Minuk 1988; Leuschner 1989; Gonzalezkoch 1997; Almasio 2000; Nevens 2016). There was no evidence of a difference in any of the comparisons (Analysis 1.2).

Mortality (one to five years)

Twenty trials including 2168 participants reported mortality (one to five years) (Epstein 1979; Macklon 1982; Matloff 1982; Taal 1983; Christensen 1985; Hoofnagle 1986; Kaplan 1986; Mitchison 1989; Wiesner 1990; Lombard 1993; Mitchison 1993; Heathcote 1994; Lindor 1994; Turner 1994; Poupon 1996; Hendrickse 1999; Pares 2000; Papatheodoridis 2002; Combes 2005; Hosonuma 2015). The proportion of people with mortality (one to five years) was higher in the methotrexate group (adjusted proportion: 23.3%) than in the no intervention group (1/30 (3.3%)) (OR 8.83, 95% CI 1.01 to 76.96; 60 participants; 1 trial). There was no evidence of a difference in any of the remaining comparisons (Analysis 1.3).

Serious adverse events (proportion)

Eleven trials including 1076 participants reported serious adverse events (proportion) (Matloff 1982; Warnes 1987; Leuschner 1989; Lindor 1994; Poupon 1996; Kurihara 2000; Pares 2000; Kanda 2003; Mason 2008; Hirschfield 2015; Nevens 2016). The period of follow‐up varied from three to 41 months. The proportion of people with serious adverse events (proportion) was higher in the D‐penicillamine group (adjusted proportion: 28.8%; based on a control group proportion of 1%) versus the no intervention group (0/26 (0.0%)) (OR 28.77, 95% CI 1.57 to 526.67; 52 participants; 1 trial). The proportion of people with serious adverse events (proportion) was higher in the obeticholic acid plus UDCA group (adjusted proportion: 4.1%) versus the UDCA group (19/143 (13.3%)) (OR 3.58, 95% CI 1.02 to 12.51; 216 participants; 1 trial). There was no evidence of a difference in any of the remaining comparisons (Analysis 1.4).

Serious adverse events (number of events)

One trial including 216 participants reported serious adverse events (number of events) (Nevens 2016). The period of follow‐up was 12 months. There was no evidence of a difference between the UDCA plus obeticholic acid versus the UDCA groups (Analysis 1.5).

Adverse events (proportion)

Nineteen trials including 1652 participants reported adverse events (proportion) (Macklon 1982; Dickson 1985; Minuk 1988; Leuschner 1989; Wiesner 1990; Ferri 1993; Lombard 1993; Mitchison 1993; Raedsch 1993; Lindor 1994; Ikeda 1996; Gonzalezkoch 1997; Kurihara 2000; Pares 2000; Yokomori 2001; Kanda 2003; Rautiainen 2005; Gao 2012; Hirschfield 2015). The proportion of people with adverse events (proportion) was higher in the ciclosporin group (adjusted proportion: 76.2%) versus the no intervention group (97/189 (51.3%) (OR 3.04, 95% CI 1.98 to 4.68; 390; 3 trials; I² = 27%), D‐penicillamine group (adjusted proportion: 50.6%) versus the no intervention group (25/135 (18.5%)) (OR 4.51, 95% CI 2.56 to 7.93; 287 participants; 2 trials; I² = 0%); malotilate group (adjusted proportion: 19.2%) versus the no intervention group (1/49 (2.0%)) (OR 11.43, 95% CI 1.40 to 93.04; 101 participants; 1 trial); and obeticholic acid group (adjusted proportion: 96.1%) versus the no intervention group (32/38 (84.2%)) (OR 4.58, 95% CI 1.31 to 15.95; 165 participants; 1 trial). The proportion of people with adverse events (proportion) was higher in the glucocorticosteroids plus UDCA (adjusted proportion: 15.8%) versus the UDCA group (2/61 (3.3%) (OR 5.54, 95% CI 1.35 to 22.84; 135 participants; 2 trials; I² = 0%) and methotrexate plus UDCA (adjusted proportion: 100.0%) versus the UDCA group (0/12 (0.0%)) (OR 115.00, 95% CI 4.98 to 2657.48; 25 participants; 1 trial). The proportion of people with adverse events (proportion) was higher in the taurodeoxycholic acid (TUDCA) group (adjusted proportion: 60.0%) versus the UDCA group (1/15 (6.7%)) (OR 21.00, 95% CI 2.16 to 204.61; 30 participants; 1 trial). There was no evidence of a difference in any of the remaining comparisons (Analysis 1.6).

Adverse events (number)

Fourteen trials including 1304 participants reported adverse events (number) (Matloff 1982; Taal 1983; Dickson 1985; Hoofnagle 1986; Minuk 1988; Wiesner 1990; Lombard 1993; Mitchison 1993; Ikeda 1996; Gonzalezkoch 1997; Wolfhagen 1998; Hirschfield 2015; Hosonuma 2015; Ma 2016). The number of adverse events was higher in the chlorambucil group (adjusted rate: 57.9 events per 100 participants) versus the no intervention group (3/11 (27.3 events per 100 participants)) (rate ratio 3.67, 95% CI 1.04 to 12.87; 24 participants; 1 trial); ciclosporin group (adjusted rate: 84.4 events per 100 participants) versus the no intervention group (128/189 (67.7 events per 100 participants)) (rate ratio 2.58, 95% CI 1.26 to 5.31; 390 participants; 3 trials; I²= 69%); D‐penicillamine group (adjusted rate: 48.4 events per 100 participants) versus the no intervention group (37/155 (23.9 events per 100 participants)) (rate ratio 2.99, 95% CI 1.04 to 8.63; 303 participants; 3 trials; I² = 75%), malotilate group (adjusted rate: 20.7 events per 100 participants) versus the no intervention group (2/49 (4.1 events per 100 participants)) (rate ratio 6.13, 95% CI 1.38 to 27.14; 101 participants; 1 trial); and obeticholic acid group (adjusted rate: 175.0 events per 100 participants) versus the no intervention group (96/38 (252.6 events per 100 participants)) (rate ratio 1.41, 95% CI 1.13 to 1.75; 76 participants; 1 trial); ; ; ; . The number of adverse events was higher in the methotrexate plus UDCA group (adjusted rate: 30.6 events per 100 participants) versus the UDCA group (0/12 (0.0 events per 100 participants)) (rate ratio 30.64, 95% CI 1.84 to 510.76; 27 participants; 1 trial). There was no evidence of a difference in any of the remaining comparisons (Analysis 1.7).

Health‐related quality of life

None of the trials reported health‐related quality of life at any time point.

Liver transplantation

Eleven trials including 1561 participants reported liver transplantation (Neuberger 1985; Wiesner 1990; Lombard 1993; Heathcote 1994; Lindor 1994; Turner 1994; Eriksson 1997; Hendrickse 1999; Papatheodoridis 2002; Combes 2005; Hosonuma 2015). There was no evidence of a difference in any of the comparisons (Analysis 1.8).

Decompensated liver disease

Seven trials including 909 participants reported decompensated liver disease (Taal 1983; Combes 1995a; Almasio 2000; Papatheodoridis 2002; Combes 2005; Gao 2012; Nevens 2016). There was no evidence of a difference in any of the comparisons (Analysis 1.9).

Cirrhosis

Three trials including 103 participants reported cirrhosis (Heathcote 1976; Turner 1994; Wolfhagen 1998). There was no evidence of a difference in any of the comparisons (Analysis 1.10).

Hepatocellular carcinoma

None of the trials reported hepatocellular carcinoma.

Subgroup analysis

All the trials were at high risk of bias for one or more domains. None of the trials reported separate data for symptomatic and asymptomatic participants, AMA‐positive and AMA‐negative participants, or for responders and non‐responders to bile acids. A secondary analysis performed by stratifying for the doses of UDCA and obeticholic acid revealed no differences between the main analysis except for the following.

There was no evidence of differences in the proportion of people with adverse events when stratified by the dose of obeticholic acid (obeticholic acid (high) versus no intervention: OR 16.60, 95% CI 0.90 to 305.59; 79 participants; 1 trial; obeticholic acid (moderate) versus no intervention: OR 8.81, 95% CI 1.01 to 76.73; 86 participants; 1 trial; and obeticholic acid (low) versus no intervention: OR 1.59, 95% CI 0.41 to 6.17; 76 participants; 1 trial). In addition, when stratified by dose, only obeticholic acid (high) had higher number of adverse events than no intervention (rate ratio 1.91, 95% CI 1.50 to 2.44; 79 participants; 1 trial). It also had higher number of adverse events than obeticholic acid (moderate) and obeticholic acid (low) (obeticholic acid (moderate) versus obeticholic acid (high): rate ratio 0.66, 95% CI 0.53 to 0.81; 89 participants; 1 trial; obeticholic acid (low) versus obeticholic acid (high): rate ratio 0.55, 95% CI 0.43 to 0.70; 79 participants; 1 trial).

Sensitivity analysis

We did not perform a sensitivity analysis of imputing information based on different scenarios because of paucity of data to carry out these analyses. We did not impute standard deviation; therefore, we did not perform a sensitivity analysis to assess the impact of imputing the standard deviation.

Reporting bias

We did not assess reporting bias by creating a funnel plot because of the few trials included under each comparison.

Using fixed‐effect model versus random‐effects model

The interpretation of results was not altered based on the model used for analysis.

Quality of evidence

The overall quality of evidence was very low for all the outcomes (summary of findings Table for the main comparison; summary of findings Table 2; summary of findings Table 3; summary of findings Table 4; summary of findings Table 5; summary of findings Table 6; summary of findings Table 7). This was because of the high risk of bias in all the trials (downgraded by two levels); small sample sizes for all outcomes and wide CIs (downgraded by two levels for imprecision) and heterogeneity (downgraded by two levels) for some of the outcomes.

Sample size calculations and Trial Sequential Analysis

The required sample size for identifying a 20% relative risk reduction in the different outcomes based on an alpha error of 5%, a beta error of 20%, and the control group proportion observed across trials were as follows.

Mortality (up to one year) (control group proportion: 25.2%): 2166 participants.
Mortality (one to five years) (control group proportion: 20.0%): 2894 participants.
Mortality at maximal follow‐up (control group proportion: 20.8%): 2758 participants.
Serious adverse events (proportion) (control group proportion: 0.4%): 175,996 participants.
Adverse events (proportion) (control group proportion: 27%): 1978 participants.
Liver transplantation (control group proportion: 7.4%): 8910 participants.
Decompensated liver disease (control group proportion: 20.8%): 2758 participants.
Cirrhosis (control group proportion: 55.6%): 632 participants.

The above mentioned are sample sizes uncorrected for heterogeneity. In the presence of heterogeneity, for example, in the presence of a heterogeneity of 27%, the required information size for adverse events (proportion) is 1978/(1 ‐ 0.27) = 2710 participants.

As shown in Figure 4, Figure 5, and Figure 6, the accrued sample sizes were only small fractions of the diversity‐adjusted required information size (DARIS) and therefore, the boundaries could not be drawn. There was a high risk of random errors. The TSA‐adjusted CI could not be calculated as there was too little information for the calculation (i.e. the CIs were wide).

Figure 4

$Trial Sequential Analysis of mortality at maximal follow‐up: azathioprine versus no intervention and colchicine versus no intervention.Based on an alpha error of 2.5%, power of 90% (beta error of 10%), a relative risk reduction (RRR) of 20%, a control group proportion observed in the trials (Pc = 20%), and diversity observed in the analyses (0%), the accrued sample size (224 for azathioprine versus intervention and 122 for colchicine versus no intervention) was only a small fraction of the diversity adjusted required information size (DARIS) (4580 for both comparisons); therefore, the trial sequential monitoring boundaries were not drawn. The Z‐curve (blue line) crossed the conventional boundaries (dotted green line) favouring azathioprine for azathioprine versus no intervention, but did not cross the conventional boundaries for colchicine versus no intervention. This indicates that there is a high risk of random errors in both these comparisons.$

Trial Sequential Analysis of mortality at maximal follow‐up: azathioprine versus no intervention and colchicine versus no intervention.

Based on an alpha error of 2.5%, power of 90% (beta error of 10%), a relative risk reduction (RRR) of 20%, a control group proportion observed in the trials (Pc = 20%), and diversity observed in the analyses (0%), the accrued sample size (224 for azathioprine versus intervention and 122 for colchicine versus no intervention) was only a small fraction of the diversity adjusted required information size (DARIS) (4580 for both comparisons); therefore, the trial sequential monitoring boundaries were not drawn. The Z‐curve (blue line) crossed the conventional boundaries (dotted green line) favouring azathioprine for azathioprine versus no intervention, but did not cross the conventional boundaries for colchicine versus no intervention. This indicates that there is a high risk of random errors in both these comparisons.

Figure 5

$Trial Sequential Analysis of mortality at maximal follow‐up: ciclosporin versus no intervention and D‐penicillamine versus no intervention.Based on an alpha error of 2.5%, power of 90% (beta error of 10%), a relative risk reduction (RRR) of 20%, a control group proportion observed in the trials (Pc = 20%), and diversity observed in the analyses (82% for ciclosporin versus no intervention and 61% for D‐penicillamine versus no intervention), the accrued sample size (394 for ciclosporin versus no intervention and 423 for D‐penicillamine versus no intervention) was only a small fraction of the diversity adjusted required information size (DARIS) (25,098 for ciclosporin versus no intervention and 11,623 for D‐penicillamine versus no intervention); therefore, the trial sequential monitoring boundaries were not drawn. The Z‐curve (blue line) crossed the conventional boundaries (dotted green line) favouring ciclosporin for ciclosporin versus no intervention, but did not cross the conventional boundaries for D‐penicillamine versus no intervention. This indicates that there is a high risk of random errors in both these comparisons.$

Trial Sequential Analysis of mortality at maximal follow‐up: ciclosporin versus no intervention and D‐penicillamine versus no intervention.

Based on an alpha error of 2.5%, power of 90% (beta error of 10%), a relative risk reduction (RRR) of 20%, a control group proportion observed in the trials (Pc = 20%), and diversity observed in the analyses (82% for ciclosporin versus no intervention and 61% for D‐penicillamine versus no intervention), the accrued sample size (394 for ciclosporin versus no intervention and 423 for D‐penicillamine versus no intervention) was only a small fraction of the diversity adjusted required information size (DARIS) (25,098 for ciclosporin versus no intervention and 11,623 for D‐penicillamine versus no intervention); therefore, the trial sequential monitoring boundaries were not drawn. The Z‐curve (blue line) crossed the conventional boundaries (dotted green line) favouring ciclosporin for ciclosporin versus no intervention, but did not cross the conventional boundaries for D‐penicillamine versus no intervention. This indicates that there is a high risk of random errors in both these comparisons.

Figure 6

$Trial Sequential Analysis of mortality at maximal follow‐up: colchicine plus UDCA versus UDCA.Based on an alpha error of 2.5%, power of 90% (beta error of 10%), a relative risk reduction (RRR) of 20%, a control group proportion observed in the trials (Pc = 7.8%), and diversity observed in the analyses (0%), the accrued sample size (160 participants) was only a small fraction of the diversity adjusted required information size (DARIS) (13,316); therefore, the trial sequential monitoring boundaries were not drawn. The Z‐curve (blue line) did not cross the conventional boundaries (green dotted line). This indicates that there is a high risk of random errors in both this comparison.$

Trial Sequential Analysis of mortality at maximal follow‐up: colchicine plus UDCA versus UDCA.

Based on an alpha error of 2.5%, power of 90% (beta error of 10%), a relative risk reduction (RRR) of 20%, a control group proportion observed in the trials (Pc = 7.8%), and diversity observed in the analyses (0%), the accrued sample size (160 participants) was only a small fraction of the diversity adjusted required information size (DARIS) (13,316); therefore, the trial sequential monitoring boundaries were not drawn. The Z‐curve (blue line) did not cross the conventional boundaries (green dotted line). This indicates that there is a high risk of random errors in both this comparison.

Discussion

Summary of main results

We included 74 trials (5902 participants) in this review, and included 4274 participants from 46 trials in one or more outcomes in this review. We did not perform the planned network meta‐analysis because there was no closed loop (i.e. outcomes for which direct and indirect estimates were available to allow us estimation of inconsistency). Therefore, we reported only the direct comparisons. We reported the results from frequentist meta‐analysis only as it is more familiar to people.

Although mortality at maximal follow‐up was lower in people who received azathioprine versus no intervention, there was no evidence of any reduction in mortality by any intervention, either at less than one year or between one and five years. However, this evidence is unreliable because the Christensen 1985 trial excluded a large proportion of participants (25%) (i.e. only 185/224 participants were included in the meta‐analysis). The Trial Sequential Analysis showed that only a small proportion of the required information size was reached and the risk of random errors was high. In addition to the risk of systematic errors and random errors, the proportion of people who died was high (71.3%) in the no intervention group of the Christensen 1985 trial compared to the other trials (the overall mortality at maximal follow‐up was 20.8%). Although this difference could be due to the shorter follow‐up periods in some of the other trials, the mortality observed in this trial was much higher than that observed in the other trials with similar or longer follow‐up such as Epstein 1979; Hendrickse 1999; and Papatheodoridis 2002. The general care of people with cirrhosis is likely to have improved since the 1980s and it is unlikely to be as high as that mortality observed in Christensen 1985. This is another reason why there is no need to recommend azathioprine routinely in people with primary biliary cholangitis.

There was no evidence of a decrease in liver transplantation, decompensated liver disease, or cirrhosis in any of the interventions compared with no intervention. However, several interventions increased the number of people with, and total number of, adverse events. Although the Trial Sequential Analysis revealed that only a small proportion of the required information size was reached, the risk of random errors was high. Thus, concluding that there were more adverse events in some of these comparisons is only of academic interest because none of the interventions appeared to result in clinical benefit.

However, it has to be pointed out that the periods of follow‐up in the trials were sufficiently long to identify any differences in clinical outcomes because primary biliary cholangitis is a slowly progressive disease. Trials with sufficient follow‐up (e.g. five or 10 years) are required to detect any differences in clinically important outcomes.

Overall completeness and applicability of evidence

The trials included symptomatic and asymptomatic primary biliary cholangitis, AMA‐positive and AMA‐negative primary biliary cholangitis, treatment‐naive people, and people regardless of the treatments that they had received previously. However, majority of the trials excluded people with advanced liver cirrhosis and primary biliary cholangitis in people with other liver diseases. Therefore, this review is applicable to people with primary biliary cholangitis without advanced liver cirrhosis or with coexisting other liver diseases.

Quality of the evidence

The overall quality of evidence was very low for all the outcomes. The major reasons for this were the high risk of bias in the trials, in particular, exclusion of participants from the analysis after randomisation, small sample size, and imprecision. Overall, there were serious concerns about whether the effect estimates observed were the true effect estimates.

Potential biases in the review process

We followed the guidance of the Cochrane Handbook for Systematic Reviews of Interventions with two review authors independently selecting trials and extracting data (Higgins 2011). We performed a thorough search of the literature. However, the search period includes the premandatory trial registration era and it is possible that some trials on treatments that were not effective or were harmful were not reported at all.

We excluded studies which compared variations in the different treatments. Hence, this review does not provide information on whether one variation is better than another.

We only included randomised clinical trials which are known to focus mostly on benefits and do not collect and report harms in a detailed manner. Therefore, we might have missed a large number of studies that addressed the reporting of harms. Accordingly, this review is biased towards benefits ignoring harms. We did not search for interventions and trials registered at regulatory authorities (e.g. US Food and Drug Administration; European Medicines Agency, etc.). This may have overlooked trials and as such trials usually are unpublished, the lack of inclusion of such trials may make our comparisons look more advantageous than they really are. However, this is of academic interest only because there is no evidence of benefit of any treatment in people with primary biliary cholangitis (i.e. there is no reason to suggest that any of the treatments should be used in routine clinical practice regardless of the adverse event profile of the intervention).

We planned to perform a network meta‐analysis. However, it was not possible to assess whether the potential effect modifiers were similar across different comparisons. Performing a network meta‐analysis in this scenario can be misleading. Therefore, we did not perform the network meta‐analysis, and assessed the comparative benefits and harms of different interventions using standard Cochrane methodology.

Agreements and disagreements with other studies or reviews

We identified three network meta‐analyses on this topic (Zhu 2015a; Zhu 2015b; Zhu 2015c). We disagreed with the authors of these reviews that UDCA in combination with corticosteroids or methotrexate are effective interventions in the treatment of primary biliary cholangitis. The disagreements were probably due to considering mortality and liver transplantation separately in this review compared to Zhu 2015a and Zhu 2015b and only including evidence prior to cross‐over in our review. In particular, the cross‐over was not true cross‐over where the interventions were swapped but all the participants belonging to the 'no intervention' were switched over to the intervention. Therefore, it was not possible to obtain the effect estimate adjusted for intra‐participant correlation either. It should be also noted that the decision to switch the no intervention to intervention was based on improvement of some laboratory parameters which are invalidated surrogate outcomes. This can only be considered as observational evidence. We disagree with current EASL and AASLD guideline recommendations that UDCA should be used for the management of primary biliary cholangitis (EASL 2009; Lindor 2009). Again, these recommendations were based on observational evidence and invalidated surrogate outcomes (Gluud 2007).

We agreed with several systematic reviews that showed that none of the interventions are effective in improving survival or other major clinical outcomes such as cirrhosis or liver transplantation (Giljaca 2010; Rudic 2012a; Rudic 2012b; Yin 2015; Zhang 2015).

Figure 1

Study flow diagram.

Figure 2

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

Figure 3

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

Figure 4

Trial Sequential Analysis of mortality at maximal follow‐up: azathioprine versus no intervention and colchicine versus no intervention.

Figure 5

Trial Sequential Analysis of mortality at maximal follow‐up: ciclosporin versus no intervention and D‐penicillamine versus no intervention.

Based on an alpha error of 2.5%, power of 90% (beta error of 10%), a relative risk reduction (RRR) of 20%, a control group proportion observed in the trials (Pc = 20%), and diversity observed in the analyses (82% for ciclosporin versus no intervention and 61% for D‐penicillamine versus no intervention), the accrued sample size (394 for ciclosporin versus no intervention and 423 for D‐penicillamine versus no intervention) was only a small fraction of the diversity adjusted required information size (DARIS) (25,098 for ciclosporin versus no intervention and 11,623 for D‐penicillamine versus no intervention); therefore, the trial sequential monitoring boundaries were not drawn. The Z‐curve (blue line) crossed the conventional boundaries (dotted green line) favouring ciclosporin for ciclosporin versus no intervention, but did not cross the conventional boundaries for D‐penicillamine versus no intervention. This indicates that there is a high risk of random errors in both these comparisons.

Figure 6

Trial Sequential Analysis of mortality at maximal follow‐up: colchicine plus UDCA versus UDCA.

Analysis 1.1

Comparison 1 Main analysis, Outcome 1 Mortality at maximal follow‐up.

Analysis 1.2

Comparison 1 Main analysis, Outcome 2 Mortality (< 1 year).

Analysis 1.3

Comparison 1 Main analysis, Outcome 3 Mortality (1 to 5 years).

Analysis 1.4

Comparison 1 Main analysis, Outcome 4 Serious adverse events (proportion).

Analysis 1.5

Comparison 1 Main analysis, Outcome 5 Serious adverse events (number of events).

Analysis 1.6

Comparison 1 Main analysis, Outcome 6 Adverse events (proportion).

Analysis 1.7

Comparison 1 Main analysis, Outcome 7 Adverse events (number).

Analysis 1.8

Comparison 1 Main analysis, Outcome 8 Liver transplantation.

Analysis 1.9

Comparison 1 Main analysis, Outcome 9 Decompensated liver disease.

Analysis 1.10

Comparison 1 Main analysis, Outcome 10 Cirrhosis.

Analysis 2.1

Comparison 2 Stratified by dose, Outcome 1 Mortality at maximal follow‐up.

Analysis 2.2

Comparison 2 Stratified by dose, Outcome 2 Mortality (< 1 year).

Analysis 2.3

Comparison 2 Stratified by dose, Outcome 3 Mortality (1 to 5 years).

Analysis 2.4

Comparison 2 Stratified by dose, Outcome 4 Serious adverse events (proportion).

Analysis 2.5

Comparison 2 Stratified by dose, Outcome 5 Serious adverse events (number of events).

Analysis 2.6

Comparison 2 Stratified by dose, Outcome 6 Adverse events (proportion).

Analysis 2.7

Comparison 2 Stratified by dose, Outcome 7 Adverse events (number).

Analysis 2.8

Comparison 2 Stratified by dose, Outcome 8 Liver transplantation.

Analysis 2.9

Comparison 2 Stratified by dose, Outcome 9 Decompensated liver disease.

Analysis 2.10

Comparison 2 Stratified by dose, Outcome 10 Cirrhosis.

Summary of findings for the main comparison. Ursodeoxycholic acid (UDCA) versus no intervention for primary biliary cholangitis

UDCA versus no intervention for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: UDCA Comparison: no intervention
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	No intervention	UDCA
Mortality at maximal follow‐up Follow‐up: 12 to 89 months	208 per 1000	206 per 1000 (136 to 301)	OR 0.99 (0.60 to 1.64)	734 (6 trials)	⊕⊝⊝⊝ Very low^1,2
Serious adverse events (proportion) Follow‐up: 12 to 41 months	There were no events in either group			380 (3 trials)	⊕⊝⊝⊝ Very low^1,2,3
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio; UDCA: ursodeoxycholic acid.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels). ³ There was moderate heterogeneity (downgraded by one level).

Summary of findings for the main comparison. Ursodeoxycholic acid (UDCA) versus no intervention for primary biliary cholangitis

Summary of findings 2. Azathioprine versus no intervention for primary biliary cholangitis

Azathioprine versus no intervention for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: azathioprine Comparison: no intervention
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	No intervention	Azathioprine
Mortality at maximal follow‐up Follow‐up: 63 months in 1 trial and not stated in 1 trial	208 per 1000	128 per 1000 (78 to 205)	OR 0.56 (0.32 to 0.98)	224 (2 trials)	⊕⊝⊝⊝ Very low^1,2
Serious adverse events (proportion)	None of the trials reported this outcome.
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels).

Summary of findings 2. Azathioprine versus no intervention for primary biliary cholangitis

Summary of findings 3. Colchicine versus no intervention for primary biliary cholangitis

Colchicine versus no intervention for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: colchicine Comparison: no intervention
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	No intervention	Colchicine
Mortality at maximal follow‐up Follow‐up: 12 to 24 months	208 per 1000	168 per 1000 (78 to 327)	OR 0.77 (0.32 to 1.85)	122 (2 trials)	⊕⊝⊝⊝ Very low^1,2
Serious adverse events (proportion) Follow‐up: 12 months	There were no events in either group			64 (1 trial)	⊕⊝⊝⊝ Very low^1,2,3
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels). ³ There was moderate heterogeneity (downgraded by one level).

Summary of findings 3. Colchicine versus no intervention for primary biliary cholangitis

Summary of findings 4. Ciclosporin versus no intervention for primary biliary cholangitis

Ciclosporin versus no intervention for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: ciclosporin Comparison: no intervention
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	No intervention	Ciclosporin
Mortality at maximal follow‐up Follow‐up: 31 to 35 months	208 per 1000	188 per 1000 (118 to 283)	OR 0.88 (0.51 to 1.50)	390 (3 trials)	⊕⊝⊝⊝ Very low^1,2
Serious adverse events (proportion)	None of the trials reported this outcome.
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels).

Summary of findings 4. Ciclosporin versus no intervention for primary biliary cholangitis

Summary of findings 5. D‐Penicillamine versus no intervention for primary biliary cholangitis

D‐Penicillamine versus no intervention for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: D‐penicillamine Comparison: no intervention
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	No intervention	D‐Penicillamine
Mortality at maximal follow‐up (Follow‐up 24 to 66 months)	208 per 1000	191 per 1000 (130 to 274)	OR 0.90 (0.57 to 1.44)	423 (5 trials)	⊕⊝⊝⊝ Very low^1,2,3
Serious adverse events (proportion) (Follow‐up 24 months)	4 per 1000	104 per 1000 (6 to 679)	OR 28.77 (1.57 to 526.67)	52 (1 trial)	⊕⊝⊝⊝ Very low^1,2,3
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels). ³ There was moderate heterogeneity (downgraded by one level).

Summary of findings 5. D‐Penicillamine versus no intervention for primary biliary cholangitis

Summary of findings 6. Colchicine plus ursodeoxycholic acid (UDCA) versus UDCA for primary biliary cholangitis

Colchicine plus UDCA versus UDCA for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: colchicine + UDCA Comparison: UDCA
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	UDCA	Colchicine + UDCA
Mortality at maximal follow‐up Follow‐up: 24 months in 1 trial; not reported in 1 trial	110 per 1000	185 per 1000 (45 to 524)	OR 1.84 (0.38 to 8.91)	158 (2 trials)	⊕⊝⊝⊝ Very low^1,2
Serious adverse events (proportion) Follow‐up: not stated	14 per 1000	42 per 1000 (2 to 526)	OR 3.08 (0.12 to 78.14)	74 (1 trial)	⊕⊝⊝⊝ Very low^1,2,3
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio; UDCA: ursodeoxycholic acid.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels). ³ There was moderate heterogeneity (downgraded by one level).

Summary of findings 6. Colchicine plus ursodeoxycholic acid (UDCA) versus UDCA for primary biliary cholangitis

Summary of findings 7. Methotrexate plus ursodeoxycholic acid (UDCA) versus UDCA for primary biliary cholangitis

Methotrexate plus UDCA versus UDCA for primary biliary cholangitis
Patient or population: people with primary biliary cholangitis Settings: secondary or tertiary care Intervention: methotrexate + UDCA Comparison: UDCA
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (trials)	Quality of the evidence (GRADE)
	Assumed risk	Corresponding risk
	UDCA	Methotrexate + UDCA
Mortality at maximal follow‐up Follow‐up: 11 to 91 months	110 per 1000	126 per 1000 (64 to 237)	OR 1.17 (0.55 to 2.51)	290 (2 trials)	⊕⊝⊝⊝ Very low^1,2
Serious adverse events (proportion)	None of the trials reported this outcome.
Serious adverse events (number of events)	None of the trials reported this outcome.
Health‐related quality of life	None of the trials reported this outcome.
The basis for the assumed risk* is the mean control group proportion across all the trials. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; OR: odds ratio; UDCA: ursodeoxycholic acid.
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
¹ Risk of bias in the trial(s) was high (downgraded by two levels). ² Sample sizes were small and 95% confidence intervals overlapped clinically significant and clinically insignificant or no effect (downgraded by two levels). ³ There was moderate heterogeneity (downgraded by one level).

Summary of findings 7. Methotrexate plus ursodeoxycholic acid (UDCA) versus UDCA for primary biliary cholangitis

Table 1. Characteristics of included studies arranged by comparison

Study name	No participants randomised	Post‐randomisation dropouts	No participants for whom outcome was reported	Intervention(s)	Control	Mean follow‐up period (months)
Smart 1990	20	Not stated	20	Antioxidants	No intervention	Not stated
Christensen 1985	248	63	185	Azathioprine	No intervention	63
Heathcote 1976	45	6	39	Azathioprine	No intervention	Not stated
Hoofnagle 1986	24	0	24	Chlorambucil	No intervention	52
Bodenheimer 1988	57	10	47	Colchicine	No intervention	33
Kaplan 1986	60	3	57	Colchicine	No intervention	24
Warnes 1987	64	Not stated	64*	Colchicine	No intervention	19 (median)
Bobadilla 1994	40	Not stated	40	Colchicine + UDCA	No intervention	12
Lombard 1993	349	0	349	Ciclosporin	No intervention	31 (median)
Minuk 1988	12	0	12	Ciclosporin	No intervention	Not stated
Wiesner 1990	40	11	29	Ciclosporin	No intervention	35 (median)
Dickson 1985	309	82	227	D‐Penicillamine	No intervention	60 (median)
Epstein 1979	98	Not stated	98	D‐Penicillamine	No intervention	66
Macklon 1982	60	0	60	D‐Penicillamine	No intervention	37
Matloff 1982	52	0	52	D‐Penicillamine	No intervention	24
Neuberger 1985	189	Not stated	189	D‐Penicillamine	No intervention	Not stated
Taal 1983	24	Not stated	24	D‐Penicillamine	No intervention	18
Triger 1980	35	Not stated	35	D‐Penicillamine	No intervention	Not stated
Mitchison 1989	36	0	36	Glucocorticosteroids	No intervention	36
Ueno 2005	20	Not stated	20	Lamivudine	No intervention	Not stated
Mitchison 1993	104	3	101	Malotilate	No intervention	25 (median)
Hendrickse 1999	60	Not stated	60	Methotrexate	No intervention	68
Steenbergen 1994	14	Not stated	14	Methotrexate + UDCA	No intervention	24
Mayo 2015	45	3	42	NGM282	No intervention	Not stated
Bowlus 2014	216	Not stated	216	Obeticholic acid	No intervention	12
Hirschfield 2015	165	0	165	Obeticholic acid	No intervention	3
Kowdley 2014a	59	Not stated	59	Obeticholic acid	No intervention	Not stated
Manzillo 1993a	32	Not stated	32	S‐Adenosyl methionine	No intervention	1
Manzillo 1993b	6	Not stated	6	S‐Adenosyl methionine	No intervention	2
Cash 2013	21	8	13	Simvastatin	No intervention	12
Askari 2010	28	0	28	Tetrathiomolybdate	No intervention	Not stated
McCormick 1994	18	0	18	Thalidomide	No intervention	Not stated
Arora 1990	9	Not stated	9	UDCA	No intervention	5
Battezzati 1993	88	2	86	UDCA	No intervention	6
Combes 1995a	151	0	151	UDCA	No intervention	24
Eriksson 1997	116	15	101	UDCA	No intervention	24
Heathcote 1994	222	Not stated	222	UDCA	No intervention	24
Leuschner 1989	20	0	18	UDCA	No intervention	12
Lim 1994	32	Not stated	32	UDCA	No intervention	Not stated
Lindor 1994	180	10	170	UDCA	No intervention	24
Oka 1990	52	7	45	UDCA	No intervention	Not stated
Papatheodoridis 2002	92	6	86	UDCA	No intervention	89
Pares 2000	192	0	192	UDCA	No intervention	41 (median)
Poupon 1991a	149	3	146	UDCA	No intervention	Not stated
Senior 1991	20	1	19	UDCA	No intervention	18
Turner 1994	46	0	46	UDCA	No intervention	24
Goddard 1994	57	Not stated	57	Intervention 1: UDCA Intervention 2: colchicine Intervention 3: colchicine + UDCA	No intervention	15
Wolfhagen 1998	50	Not stated	50	Azathioprine + glucocorticosteroids + UDCA	UDCA	12
Iwasaki 2008a	45	Not stated	45	Bezafibrate	UDCA	12
Kurihara 2000	24	Not stated	24	Bezafibrate	UDCA	Not stated
Hosonuma 2015	27	0	27	Bezafibrate + UDCA	UDCA	96
Iwasaki 2008b	22	Not stated	22	Bezafibrate + UDCA	UDCA	12
Kanda 2003	22	0	22	Bezafibrate + UDCA	UDCA	7
Nakai 2000	23	Not stated	23	Bezafibrate + UDCA	UDCA	12
Almasio 2000	90	6	84	Colchicine + UDCA	UDCA	Not stated
Ikeda 1996	22	0	22	Colchicine + UDCA	UDCA	24
Poupon 1996	74	Not stated	74	Colchicine + UDCA	UDCA	24
Raedsch 1993	28	8	20	Colchicine + UDCA	UDCA	24
Yokomori 2001	11	Not stated	11	Colestilan + UDCA	UDCA	Not stated
Liberopoulos 2010	10	Not stated	10	Fenofibrate + UDCA	UDCA	Not stated
Leuschner 1999	40	0	39	Glucocorticosteroids + UDCA	UDCA	24
Rautiainen 2005	77	8	69	Glucocorticosteroids + UDCA	UDCA	36
Gao 2012	79	Not stated	79	Intervention 1: glucocorticosteroids + UDCA Intervention 2: azathioprine + UDCA	UDCA	Not stated
Mason 2008	59	0	59	Lamivudine + zidovudine + UDCA	UDCA	6
Combes 2005	265	0	265	Methotrexate + UDCA	UDCA	91 (median)
Gonzalezkoch 1997	25	Not stated	25	Methotrexate + UDCA	UDCA	11
Nevens 2016	217	Not stated	216	Obeticholic acid + UDCA	UDCA	12
Ferri 1993	30	0	30	TUDCA	UDCA	6
Ma 2016	199	8	191	TUDCA	UDCA	6
Kaplan 1999	87	2	85	Colchicine	Methotrexate	24
Comparison of doses
Lindor 1997	150	Not stated	150	Intervention 1: UDCA (high) Intervention 2: UDCA (moderate)	UDCA (low)	12
Angulo 1999a	155	Not stated	155	Intervention 1: UDCA (high) Intervention 2: UDCA (moderate)	UDCA (low)	12
Van Hoogstraten 1998	61	2	59	UDCA (moderate)	UDCA (low)	Not stated
Mazzarella 2002	42	Not stated	42	UDCA (high)	UDCA (moderate)	72
TUDCA: taurodeoxycholic acid; UDCA: ursodeoxycholic acid.

Table 1. Characteristics of included studies arranged by comparison

Table 2. Risk of bias arranged according to comparisons

Name of studies	Intervention(s)	Control	Random sequence generation	Allocation concealment	Blinding of participants and health professionals	Blinding of outcome assessors	Missing outcome bias	Selective outcome reporting	For‐profit bias	Other bias
Smart 1990	Antioxidants	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Christensen 1985	Azathioprine	No intervention	Unclear	Unclear	Low	Low	High	High	High	Low
Heathcote 1976	Azathioprine	No intervention	Unclear	Unclear	High	High	High	High	Low	Low
Hoofnagle 1986	Chlorambucil	No intervention	Low	Low	High	High	Low	Low	Low	Low
Bodenheimer 1988	Colchicine	No intervention	Unclear	Unclear	Low	Low	High	High	High	Low
Kaplan 1986	Colchicine	No intervention	Unclear	Unclear	Unclear	Unclear	High	High	Unclear	Low
Warnes 1987	Colchicine	No intervention	Low	Low	Low	Low	Unclear	Low	Unclear	Low
Bobadilla 1994	Colchicine + UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Lombard 1993	Ciclosporin	No intervention	Unclear	Unclear	Low	Low	Low	Low	High	Low
Minuk 1988	Ciclosporin	No intervention	Unclear	Unclear	Low	Unclear	Unclear	Low	High	Low
Wiesner 1990	Ciclosporin	No intervention	Unclear	Unclear	Low	Low	Unclear	Low	High	Low
Dickson 1985	D‐Penicillamine	No intervention	Low	Low	Low	Low	High	High	High	High
Epstein 1979	D‐Penicillamine	No intervention	Unclear	Unclear	High	High	Unclear	High	Unclear	Low
Macklon 1982	D‐Penicillamine	No intervention	Unclear	Unclear	Unclear	Unclear	Low	Low	Unclear	Low
Matloff 1982	D‐Penicillamine	No intervention	Unclear	Unclear	Unclear	Unclear	Low	Low	High	Low
Neuberger 1985	D‐Penicillamine	No intervention	Unclear	Low	Low	Low	Unclear	High	Unclear	Low
Taal 1983	D‐Penicillamine	No intervention	Unclear	Unclear	Low	Low	Unclear	Low	Unclear	Low
Triger 1980	D‐Penicillamine	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	High	Low
Mitchison 1989	Glucocorticosteroids	No intervention	Low	Low	High	High	Low	High	Unclear	Low
Ueno 2005	Lamivudine	No intervention	Unclear	Unclear	Low	Low	Unclear	High	Unclear	Low
Mitchison 1993	Malotilate	No intervention	Low	Low	Low	Low	High	Low	High	Low
Hendrickse 1999	Methotrexate	No intervention	Low	Low	Unclear	Unclear	Unclear	High	Unclear	Low
Steenbergen 1994	Methotrexate + UDCA	No intervention	Low	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Mayo 2015	NGM282	No intervention	Unclear	Unclear	Unclear	Unclear	High	High	High	Low
Bowlus 2014	Obeticholic acid	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	High	Low
Hirschfield 2015	Obeticholic acid	No intervention	Low	Unclear	Low	Low	Low	High	Unclear	High
Kowdley 2011	Obeticholic acid	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	High	Low
Manzillo 1993a	S‐Adenosyl methionine	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Manzillo 1993b	S‐Adenosyl methionine	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Cash 2013	Simvastatin	No intervention	Unclear	Low	High	High	High	High	Low	High
Askari 2010	Tetrathiomolybdate	No intervention	Low	Low	Low	Low	Low	High	Low	High
McCormick 1994	Thalidomide	No intervention	Unclear	Unclear	Low	Low	Low	High	High	Low
Arora 1990	UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Battezzati 1993	UDCA	No intervention	Low	Low	Low	Low	High	High	Unclear	Low
Combes 1995a	UDCA	No intervention	Unclear	Unclear	Low	Low	Low	High	High	Low
Eriksson 1997	UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	High	High	High	Low
Heathcote 1994	UDCA	No intervention	Unclear	Low	Low	Low	Unclear	High	High	Low
Leuschner 1989	UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	High	Low	Unclear	Low
Lim 1994	UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Lindor 1994	UDCA	No intervention	Unclear	Unclear	Low	Low	High	Low	High	Low
Oka 1990	UDCA	No intervention	Unclear	Low	Low	Low	High	High	High	Low
Papatheodoridis 2002	UDCA	No intervention	Low	Low	High	High	High	High	High	High
Pares 2000	UDCA	No intervention	Unclear	Unclear	Low	Low	Unclear	Low	High	Low
Poupon 1991a	UDCA	No intervention	Unclear	Unclear	Low	Low	High	High	High	Low
Senior 1991	UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	High	High	High	Low
Turner 1994	UDCA	No intervention	Unclear	Unclear	Low	Low	Low	High	Unclear	Low
Goddard 1994	Intervention 1: UDCA Intervention 2: colchicine Intervention 3: colchicine + UDCA	No intervention	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Wolfhagen 1998	Azathioprine + glucocorticosteroids + UDCA	UDCA	Low	Low	Low	Low	Unclear	High	High	Low
Iwasaki 2008a	Bezafibrate	UDCA	Unclear	Low	High	High	Unclear	High	Low	Low
Kurihara 2000	Bezafibrate	UDCA	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Hosonuma 2015	Bezafibrate + UDCA	UDCA	Low	Low	High	High	Low	Low	Low	Low
Iwasaki 2008b	Bezafibrate + UDCA	UDCA	Unclear	Low	High	High	Unclear	High	Low	Low
Kanda 2003	Bezafibrate + UDCA	UDCA	Unclear	Unclear	Unclear	Unclear	Low	High	Unclear	Low
Nakai 2000	Bezafibrate + UDCA	UDCA	Unclear	Unclear	Unclear	Unclear	Unclear	High	Low	Low
Almasio 2000	Colchicine + UDCA	UDCA	Low	Low	Low	Low	High	High	Low	Low
Ikeda 1996	Colchicine + UDCA	UDCA	Unclear	Unclear	Unclear	Unclear	Low	High	Unclear	High
Poupon 1996	Colchicine + UDCA	UDCA	Unclear	Unclear	Low	Low	Unclear	Low	High	Low
Raedsch 1993	Colchicine + UDCA	UDCA	Unclear	Unclear	Unclear	Unclear	High	High	Unclear	Low
Yokomori 2001	Colestilan + UDCA	UDCA	Unclear	Unclear	High	High	Unclear	High	Unclear	Low
Liberopoulos 2010	Fenofibrate + UDCA	UDCA	Unclear	Unclear	High	High	Unclear	High	Unclear	Low
Leuschner 1999	Glucocorticosteroids + UDCA	UDCA	Low	Unclear	Unclear	Unclear	High	High	High	Low
Rautiainen 2005	Glucocorticosteroids + UDCA	UDCA	Unclear	Unclear	High	High	High	High	High	Low
Gao 2012	Glucocorticosteroids + UDCA Azathioprine + UDCA	UDCA	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Mason 2008	Lamivudine + zidovudine + UDCA	UDCA	Low	Low	Low	Low	Unclear	High	High	Low
Combes 2005	Methotrexate + UDCA	UDCA	Unclear	Unclear	Unclear	Unclear	Low	High	High	Low
Gonzalezkoch 1997	Methotrexate + UDCA	UDCA	Unclear	Low	Unclear	Unclear	Unclear	Low	Unclear	Low
Nevens 2016	Obeticholic acid + UDCA	UDCA	Low	Low	Low	Low	High	Low	High	Low
Ferri 1993	TUDCA	UDCA	Unclear	Unclear	Unclear	Unclear	Low	High	Unclear	Low
Ma 2016	TUDCA	UDCA	Low	Low	Low	Low	Unclear	High	High	Low
Kaplan 1999	Colchicine	Methotrexate	Unclear	Unclear	Low	Low	High	High	Unclear	Low
Comparison of doses
Lindor 1997	Intervention 1: UDCA (high) Intervention 2: UDCA (moderate)	UDCA (low)	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
Angulo 1999a	Intervention 1: UDCA (high) Intervention 2: UDCA (moderate)	UDCA (low)	Low	Low	Low	Low	Unclear	Low	Unclear	Low
Van Hoogstraten 1998	UDCA (moderate)	UDCA (low)	Low	Low	High	High	Unclear	High	High	Low
Mazzarella 2002	UDCA (high)	UDCA (moderate)	Unclear	Unclear	Unclear	Unclear	Unclear	High	Unclear	Low
TUDCA: taurodeoxycholic acid; UDCA: ursodeoxycholic acid.

Table 2. Risk of bias arranged according to comparisons

Comparison 1. Main analysis

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Mortality at maximal follow‐up Show forest plot	28		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

1.1 Azathioprine versus no intervention	2	224	Odds Ratio (M‐H, Fixed, 95% CI)	0.56 [0.32, 0.98]
1.2 Chlorambucil versus no intervention	1	24	Odds Ratio (M‐H, Fixed, 95% CI)	0.14 [0.01, 3.28]
1.3 Colchicine versus no intervention	2	122	Odds Ratio (M‐H, Fixed, 95% CI)	0.77 [0.32, 1.85]
1.4 Cyclosporin versus no intervention	3	390	Odds Ratio (M‐H, Fixed, 95% CI)	0.88 [0.51, 1.50]
1.5 D‐Penicillamine versus no intervention	5	423	Odds Ratio (M‐H, Fixed, 95% CI)	0.90 [0.57, 1.44]
1.6 Glucocorticosteroids versus no intervention	1	36	Odds Ratio (M‐H, Fixed, 95% CI)	0.64 [0.14, 2.92]
1.7 Malotilate versus no intervention	1	101	Odds Ratio (M‐H, Fixed, 95% CI)	2.0 [0.47, 8.48]
1.8 Methotrexate versus no intervention	1	60	Odds Ratio (M‐H, Fixed, 95% CI)	8.83 [1.01, 76.96]
1.9 UDCA versus no intervention	6	734	Odds Ratio (M‐H, Fixed, 95% CI)	0.99 [0.60, 1.64]
1.10 Bezafibrate plus UDCA versus UDCA	1	27	Odds Ratio (M‐H, Fixed, 95% CI)	9.67 [0.45, 207.78]
1.11 Colchicine plus UDCA versus UDCA	2	158	Odds Ratio (M‐H, Fixed, 95% CI)	1.84 [0.38, 8.91]
1.12 Methotrexate plus UDCA versus UDCA	2	290	Odds Ratio (M‐H, Fixed, 95% CI)	1.17 [0.55, 2.51]
1.13 Obeticholic acid plus UDCA versus UDCA	1	216	Odds Ratio (M‐H, Fixed, 95% CI)	1.55 [0.06, 38.46]
2 Mortality (< 1 year) Show forest plot	8		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.1 Azathioprine versus no intervention	1	39	Odds Ratio (M‐H, Fixed, 95% CI)	0.58 [0.16, 2.10]
2.2 Colchicine versus no intervention	1	64	Odds Ratio (M‐H, Fixed, 95% CI)	0.86 [0.22, 3.33]
2.3 Cyclosporin versus no intervention	1	12	Odds Ratio (M‐H, Fixed, 95% CI)	0.14 [0.01, 3.63]
2.4 D‐Penicillamine versus no intervention	1	189	Odds Ratio (M‐H, Fixed, 95% CI)	0.71 [0.35, 1.42]
2.5 Ursodeoxycholic acid (UDCA) versus no intervention	1	18	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.6 Colchicine plus UDCA versus UDCA	1	84	Odds Ratio (M‐H, Fixed, 95% CI)	1.0 [0.13, 7.45]
2.7 Methotrexate plus UDCA versus UDCA	1	25	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.8 Obeticholic acid plus UDCA versus UDCA	1	216	Odds Ratio (M‐H, Fixed, 95% CI)	1.55 [0.06, 38.46]
3 Mortality (1 to 5 years) Show forest plot	20		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

3.1 Azathioprine versus no intervention	1	185	Odds Ratio (M‐H, Fixed, 95% CI)	0.56 [0.30, 1.04]
3.2 Chlorambucil versus no intervention	1	24	Odds Ratio (M‐H, Fixed, 95% CI)	0.14 [0.01, 3.28]
3.3 Colchicine versus no intervention	1	58	Odds Ratio (M‐H, Fixed, 95% CI)	0.71 [0.22, 2.25]
3.4 Cyclosporin versus no intervention	2	378	Odds Ratio (M‐H, Fixed, 95% CI)	0.94 [0.54, 1.64]
3.5 D‐Penicillamine versus no intervention	4	234	Odds Ratio (M‐H, Fixed, 95% CI)	1.10 [0.59, 2.08]
3.6 Glucocorticosteroids versus no intervention	1	36	Odds Ratio (M‐H, Fixed, 95% CI)	0.64 [0.14, 2.92]
3.7 Malotilate versus no intervention	1	101	Odds Ratio (M‐H, Fixed, 95% CI)	2.0 [0.47, 8.48]
3.8 Methotrexate versus no intervention	1	60	Odds Ratio (M‐H, Fixed, 95% CI)	8.83 [1.01, 76.96]
3.9 UDCA versus no intervention	5	716	Odds Ratio (M‐H, Fixed, 95% CI)	0.99 [0.60, 1.64]
3.10 Bezafibrate plus UDCA versus UDCA	1	27	Odds Ratio (M‐H, Fixed, 95% CI)	9.67 [0.45, 207.78]
3.11 Colchicine plus UDCA versus UDCA	1	74	Odds Ratio (M‐H, Fixed, 95% CI)	5.28 [0.24, 113.87]
3.12 Methotrexate plus UDCA versus UDCA	1	265	Odds Ratio (M‐H, Fixed, 95% CI)	1.17 [0.55, 2.51]
4 Serious adverse events (proportion) Show forest plot	11		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

4.1 Colchicine versus no intervention	1	64	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.2 D‐Penicillamine versus no intervention	1	52	Odds Ratio (M‐H, Fixed, 95% CI)	28.77 [1.57, 526.67]
4.3 Obeticholic acid versus no intervention	1	165	Odds Ratio (M‐H, Fixed, 95% CI)	1.83 [0.21, 15.73]
4.4 UDCA versus no intervention	3	380	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.5 UDCA versus bezafibrate	1	24	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.6 Bezafibrate plus UDCA versus UDCA	1	22	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.7 Colchicine plus UDCA versus UDCA	1	74	Odds Ratio (M‐H, Fixed, 95% CI)	3.08 [0.12, 78.14]
4.8 Lamivudine plus zidovudine plus UDCA versus UDCA	1	59	Odds Ratio (M‐H, Fixed, 95% CI)	0.47 [0.04, 5.43]
4.9 Obeticholic acid plus UDCA versus UDCA	1	216	Odds Ratio (M‐H, Fixed, 95% CI)	3.58 [1.02, 12.51]
5 Serious adverse events (number of events) Show forest plot	1		Rate Ratio (Fixed, 95% CI)	Subtotals only

5.1 Obeticholic acid plus UDCA versus UDCA	1	216	Rate Ratio (Fixed, 95% CI)	1.66 [0.75, 3.66]
6 Adverse events (proportion) Show forest plot	19		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

6.1 Cyclosporin versus no intervention	3	390	Odds Ratio (M‐H, Fixed, 95% CI)	3.04 [1.98, 4.68]
6.2 D‐Penicillamine versus no intervention	2	287	Odds Ratio (M‐H, Fixed, 95% CI)	4.51 [2.56, 7.93]
6.3 Malotilate versus no intervention	1	101	Odds Ratio (M‐H, Fixed, 95% CI)	11.43 [1.40, 93.04]
6.4 Obeticholic acid versus no intervention	1	165	Odds Ratio (M‐H, Fixed, 95% CI)	4.58 [1.31, 15.95]
6.5 UDCA versus no intervention	3	380	Odds Ratio (M‐H, Fixed, 95% CI)	1.45 [0.50, 4.25]
6.6 Azathioprine plus UDCA versus UDCA	1	42	Odds Ratio (M‐H, Fixed, 95% CI)	19.67 [0.94, 413.50]
6.7 Bezafibrate versus UDCA	1	24	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.8 Bezafibrate plus UDCA versus UDCA	1	22	Odds Ratio (M‐H, Fixed, 95% CI)	3.29 [0.12, 89.81]
6.9 Colchicine plus UDCA versus UDCA	2	42	Odds Ratio (M‐H, Fixed, 95% CI)	6.20 [0.63, 60.80]
6.10 Colestilan plus UDCA versus UDCA	1	11	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.11 Glucocorticosteroids plus UDCA versus UDCA	2	135	Odds Ratio (M‐H, Fixed, 95% CI)	5.54 [1.35, 22.84]
6.12 Methotrexate plus UDCA versus UDCA	1	25	Odds Ratio (M‐H, Fixed, 95% CI)	115.0 [4.98, 2657.48]
6.13 TauroUDCA versus UDCA	1	30	Odds Ratio (M‐H, Fixed, 95% CI)	21.0 [2.16, 204.61]
6.14 Glucocorticosteroids plus UDCA versus azathioprine plus UDCA	1	50	Odds Ratio (M‐H, Fixed, 95% CI)	0.40 [0.08, 2.12]
7 Adverse events (number) Show forest plot	14		Rate Ratio (Random, 95% CI)	Subtotals only

7.1 Chlorambucil versus no intervention	1	24	Rate Ratio (Random, 95% CI)	3.67 [1.04, 12.87]
7.2 Cyclosporin versus no intervention	3	390	Rate Ratio (Random, 95% CI)	2.58 [1.26, 5.31]
7.3 D‐Penicillamine versus no intervention	3	303	Rate Ratio (Random, 95% CI)	2.99 [1.04, 8.63]
7.4 Malotilate versus no intervention	1	101	Rate Ratio (Random, 95% CI)	6.13 [1.38, 27.14]
7.5 Obeticholic acid versus no intervention	1	76	Rate Ratio (Random, 95% CI)	1.41 [1.13, 1.75]
7.6 Azathioprine plus glucocorticosteroids plus UDCA versus UDCA	1	50	Rate Ratio (Random, 95% CI)	1.32 [0.88, 1.97]
7.7 Bezafibrate plus UDCA versus UDCA	1	29	Rate Ratio (Random, 95% CI)	11.79 [0.65, 213.14]
7.8 Colchicine plus UDCA versus UDCA	1	24	Rate Ratio (Random, 95% CI)	5.91 [0.28, 123.08]
7.9 Methotrexate plus UDCA versus UDCA	1	27	Rate Ratio (Random, 95% CI)	30.64 [1.84, 510.76]
7.10 TauroUDCA versus UDCA	1	191	Rate Ratio (Random, 95% CI)	1.17 [0.81, 1.71]
8 Liver transplantation Show forest plot	11		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

8.1 Cyclosporin versus no intervention	2	378	Odds Ratio (M‐H, Fixed, 95% CI)	0.86 [0.43, 1.72]
8.2 D‐Penicillamine versus no intervention	1	189	Odds Ratio (M‐H, Fixed, 95% CI)	0.93 [0.06, 15.05]
8.3 Methotrexate versus no intervention	1	60	Odds Ratio (M‐H, Fixed, 95% CI)	0.17 [0.02, 1.58]
8.4 UDCA versus no intervention	5	640	Odds Ratio (M‐H, Fixed, 95% CI)	0.90 [0.48, 1.68]
8.5 Bezafibrate plus UDCA versus UDCA	1	27	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
8.6 Methotrexate plus UDCA versus UDCA	1	265	Odds Ratio (M‐H, Fixed, 95% CI)	0.70 [0.35, 1.39]
9 Decompensated liver disease Show forest plot	7		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

9.1 D‐Penicillamine versus no active treatment	1	24	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
9.2 UDCA versus no intervention	2	237	Odds Ratio (M‐H, Fixed, 95% CI)	1.60 [0.86, 2.98]
9.3 Azathioprine plus UDCA versus UDCA	1	42	Odds Ratio (M‐H, Fixed, 95% CI)	0.52 [0.05, 5.18]
9.4 Colchicine plus UDCA versus UDCA	1	84	Odds Ratio (M‐H, Fixed, 95% CI)	0.21 [0.04, 1.07]
9.5 Glucocorticosteroids plus UDCA versus UDCA	1	66	Odds Ratio (M‐H, Fixed, 95% CI)	0.55 [0.11, 2.69]
9.6 Methotrexate plus UDCA versus UDCA	1	265	Odds Ratio (M‐H, Fixed, 95% CI)	1.34 [0.77, 2.33]
9.7 Obeticholic acid plus UDCA versus UDCA	1	216	Odds Ratio (M‐H, Fixed, 95% CI)	1.55 [0.06, 38.46]
9.8 Glucocorticosteroids plus UDCA versus azathioprine plus UDCA	1	50	Odds Ratio (M‐H, Fixed, 95% CI)	1.06 [0.10, 11.18]
10 Cirrhosis Show forest plot	3		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

10.1 Azathioprine versus no intervention	1	31	Odds Ratio (M‐H, Fixed, 95% CI)	0.79 [0.18, 3.41]
10.2 UDCA versus no intervention	1	22	Odds Ratio (M‐H, Fixed, 95% CI)	0.15 [0.01, 1.53]
10.3 Azathioprine plus glucocorticosteroids plus UDCA versus UDCA	1	50	Odds Ratio (M‐H, Fixed, 95% CI)	0.28 [0.03, 2.90]

Comparison 1. Main analysis

Comparison 2. Stratified by dose

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Mortality at maximal follow‐up Show forest plot	29		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

1.1 Azathioprine versus no intervention	2	224	Odds Ratio (M‐H, Fixed, 95% CI)	0.56 [0.32, 0.98]
1.2 Chlorambucil versus no intervention	1	24	Odds Ratio (M‐H, Fixed, 95% CI)	0.14 [0.01, 3.28]
1.3 Colchicine versus no intervention	2	122	Odds Ratio (M‐H, Fixed, 95% CI)	0.77 [0.32, 1.85]
1.4 Cyclosporin versus no intervention	3	390	Odds Ratio (M‐H, Fixed, 95% CI)	0.88 [0.51, 1.50]
1.5 D‐Penicillamine versus no intervention	5	423	Odds Ratio (M‐H, Fixed, 95% CI)	0.90 [0.57, 1.44]
1.6 Glucocorticosteroids versus no intervention	1	36	Odds Ratio (M‐H, Fixed, 95% CI)	0.64 [0.14, 2.92]
1.7 Malotilate versus no intervention	1	101	Odds Ratio (M‐H, Fixed, 95% CI)	2.0 [0.47, 8.48]
1.8 Methotrexate versus no intervention	1	60	Odds Ratio (M‐H, Fixed, 95% CI)	8.83 [1.01, 76.96]
1.9 UDCA (low) versus no intervention	2	64	Odds Ratio (M‐H, Fixed, 95% CI)	0.33 [0.03, 3.47]
1.10 UDCA (moderate) versus no intervention	4	670	Odds Ratio (M‐H, Fixed, 95% CI)	1.05 [0.62, 1.77]
1.11 UDCA (low) versus UDCA (high)	1	106	Odds Ratio (M‐H, Fixed, 95% CI)	1.04 [0.06, 17.06]
1.12 UDCA (moderate) versus UDCA (high)	1	103	Odds Ratio (M‐H, Fixed, 95% CI)	0.36 [0.01, 9.05]
1.13 UDCA (low) plus colchicine versus UDCA (low)	1	84	Odds Ratio (M‐H, Fixed, 95% CI)	1.0 [0.13, 7.45]
1.14 UDCA (low) plus methotrexate versus UDCA (low)	1	25	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.15 UDCA (moderate) versus UDCA (low)	1	101	Odds Ratio (M‐H, Fixed, 95% CI)	0.35 [0.01, 8.72]
1.16 Bezafibrate plus UDCA (moderate) versus UDCA (moderate)	1	27	Odds Ratio (M‐H, Fixed, 95% CI)	9.67 [0.45, 207.78]
1.17 Colchicine plus UDCA (moderate) versus UDCA (moderate)	1	74	Odds Ratio (M‐H, Fixed, 95% CI)	5.28 [0.24, 113.87]
1.18 Methotrexate plus UDCA (moderate) versus UDCA (moderate)	1	265	Odds Ratio (M‐H, Fixed, 95% CI)	1.17 [0.55, 2.51]
1.19 Obeticholic acid (low) plus UDCA (moderate) versus UDCA (moderate)	1	216	Odds Ratio (M‐H, Fixed, 95% CI)	1.55 [0.06, 38.46]
2 Mortality (< 1 year) Show forest plot	9		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

2.1 Azathioprine versus no intervention	1	39	Odds Ratio (M‐H, Fixed, 95% CI)	0.58 [0.16, 2.10]
2.2 Colchicine versus no intervention	1	64	Odds Ratio (M‐H, Fixed, 95% CI)	0.86 [0.22, 3.33]
2.3 Cyclosporin versus no intervention	1	12	Odds Ratio (M‐H, Fixed, 95% CI)	0.14 [0.01, 3.63]
2.4 D‐Penicillamine versus no intervention	1	189	Odds Ratio (M‐H, Fixed, 95% CI)	0.71 [0.35, 1.42]
2.5 UDCA (low) versus no intervention	1	18	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.6 UDCA (low) versus UDCA (high)	1	106	Odds Ratio (M‐H, Fixed, 95% CI)	1.04 [0.06, 17.06]
2.7 UDCA (moderate) versus UDCA (high)	1	103	Odds Ratio (M‐H, Fixed, 95% CI)	0.36 [0.01, 9.05]
2.8 Obeticholic acid (low) plus UDCA (moderate) versus UDCA (moderate)	1	216	Odds Ratio (M‐H, Fixed, 95% CI)	1.55 [0.06, 38.46]
2.9 UDCA (low) plus colchicine versus UDCA (low)	1	84	Odds Ratio (M‐H, Fixed, 95% CI)	1.0 [0.13, 7.45]
2.10 UDCA (low) plus methotrexate versus UDCA (low)	1	25	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.11 UDCA (moderate) versus UDCA (low)	1	101	Odds Ratio (M‐H, Fixed, 95% CI)	0.35 [0.01, 8.72]
3 Mortality (1 to 5 years) Show forest plot	20		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

3.1 Azathioprine versus no intervention	1	185	Odds Ratio (M‐H, Fixed, 95% CI)	0.56 [0.30, 1.04]
3.2 Chlorambucil versus no intervention	1	24	Odds Ratio (M‐H, Fixed, 95% CI)	0.14 [0.01, 3.28]
3.3 Colchicine versus no intervention	1	58	Odds Ratio (M‐H, Fixed, 95% CI)	0.71 [0.22, 2.25]
3.4 Cyclosporin versus no intervention	2	378	Odds Ratio (M‐H, Fixed, 95% CI)	0.94 [0.54, 1.64]
3.5 D‐Penicillamine versus no intervention	4	234	Odds Ratio (M‐H, Fixed, 95% CI)	1.10 [0.59, 2.08]
3.6 Glucocorticosteroids versus no intervention	1	36	Odds Ratio (M‐H, Fixed, 95% CI)	0.64 [0.14, 2.92]
3.7 Malotilate versus no intervention	1	101	Odds Ratio (M‐H, Fixed, 95% CI)	2.0 [0.47, 8.48]
3.8 Methotrexate versus no intervention	1	60	Odds Ratio (M‐H, Fixed, 95% CI)	8.83 [1.01, 76.96]
3.9 UDCA (low) versus no intervention	1	46	Odds Ratio (M‐H, Fixed, 95% CI)	0.33 [0.03, 3.47]
3.10 UDCA (moderate) versus no intervention	4	670	Odds Ratio (M‐H, Fixed, 95% CI)	1.05 [0.62, 1.77]
3.11 Bezafibrate plus UDCA (moderate) versus UDCA (moderate)	1	27	Odds Ratio (M‐H, Fixed, 95% CI)	9.67 [0.45, 207.78]
3.12 Colchicine plus UDCA (moderate) versus UDCA (moderate)	1	74	Odds Ratio (M‐H, Fixed, 95% CI)	5.28 [0.24, 113.87]
3.13 Methotrexate plus UDCA (moderate) versus UDCA (moderate)	1	265	Odds Ratio (M‐H, Fixed, 95% CI)	1.17 [0.55, 2.51]
4 Serious adverse events (proportion) Show forest plot	12		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

4.1 Colchicine versus no intervention	1	64	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.2 D‐Penicillamine versus no intervention	1	52	Odds Ratio (M‐H, Fixed, 95% CI)	28.77 [1.57, 526.67]
4.3 Obeticholic acid (high) versus no intervention	1	79	Odds Ratio (M‐H, Fixed, 95% CI)	5.14 [0.57, 46.17]
4.4 Obeticholic acid (low) versus no intervention	1	76	Odds Ratio (M‐H, Fixed, 95% CI)	0.32 [0.01, 8.22]
4.5 Obeticholic acid (moderate) versus no intervention	1	86	Odds Ratio (M‐H, Fixed, 95% CI)	0.79 [0.05, 13.01]
4.6 UDCA (low) versus no intervention	1	18	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.7 UDCA (moderate) versus no intervention	2	362	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.8 UDCA (low) versus bezafibrate	1	24	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.9 Obeticholic acid (low) versus obeticholic acid (high)	1	79	Odds Ratio (M‐H, Fixed, 95% CI)	0.09 [0.00, 1.61]
4.10 Obeticholic acid (moderate) versus obeticholic acid (high)	1	89	Odds Ratio (M‐H, Fixed, 95% CI)	0.15 [0.02, 1.37]
4.11 Obeticholic acid (moderate) versus obeticholic acid (low)	1	86	Odds Ratio (M‐H, Fixed, 95% CI)	2.43 [0.10, 61.39]
4.12 Lamivudine plus zidovudine plus UDCA (moderate) versus UDCA (moderate)	1	59	Odds Ratio (M‐H, Fixed, 95% CI)	0.47 [0.04, 5.43]
4.13 UDCA (moderate) versus obeticholic acid (low) plus UDCA (moderate)	1	216	Odds Ratio (M‐H, Fixed, 95% CI)	0.28 [0.08, 0.98]
4.14 Bezafibrate plus UDCA (low) versus UDCA (low)	1	22	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.15 UDCA (moderate) versus UDCA (low)	1	59	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4.16 Colchicine plus UDCA (moderate) versus UDCA (moderate)	1	74	Odds Ratio (M‐H, Fixed, 95% CI)	3.08 [0.12, 78.14]
5 Serious adverse events (number of events) Show forest plot	1		Rate Ratio (Fixed, 95% CI)	Subtotals only

5.1 Obeticholic acid (low) plus UDCA (moderate) versus UDCA (moderate)	1		Rate Ratio (Fixed, 95% CI)	1.66 [0.75, 3.66]
6 Adverse events (proportion) Show forest plot	20		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

6.1 Cyclosporin versus no intervention	3	390	Odds Ratio (M‐H, Fixed, 95% CI)	3.04 [1.98, 4.68]
6.2 D‐Penicillamine versus no intervention	2	287	Odds Ratio (M‐H, Fixed, 95% CI)	4.51 [2.56, 7.93]
6.3 Malotilate versus no intervention	1	101	Odds Ratio (M‐H, Fixed, 95% CI)	11.43 [1.40, 93.04]
6.4 Obeticholic acid (high) versus no intervention	1	79	Odds Ratio (M‐H, Fixed, 95% CI)	16.6 [0.90, 305.59]
6.5 Obeticholic acid (low) versus no intervention	1	76	Odds Ratio (M‐H, Fixed, 95% CI)	1.59 [0.41, 6.17]
6.6 Obeticholic acid (moderate) versus no intervention	1	86	Odds Ratio (M‐H, Fixed, 95% CI)	8.81 [1.01, 76.73]
6.7 UDCA (low) versus no intervention	1	18	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.8 UDCA (moderate) versus no intervention	2	362	Odds Ratio (M‐H, Fixed, 95% CI)	1.45 [0.50, 4.25]
6.9 Glucocorticosteroids plus UDCA (moderate) versus azathioprine plus UDCA (moderate)	1	50	Odds Ratio (M‐H, Fixed, 95% CI)	0.40 [0.08, 2.12]
6.10 UDCA (low) versus bezafibrate	1	24	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.11 Obeticholic acid (low) versus obeticholic acid (high)	1	79	Odds Ratio (M‐H, Fixed, 95% CI)	0.09 [0.00, 1.78]
6.12 Obeticholic acid (moderate) versus obeticholic acid (high)	1	89	Odds Ratio (M‐H, Fixed, 95% CI)	0.38 [0.02, 9.62]
6.13 Obeticholic acid (moderate) versus obeticholic acid (low)	1	86	Odds Ratio (M‐H, Fixed, 95% CI)	5.53 [0.59, 51.70]
6.14 Bezafibrate plus UDCA (low) versus UDCA (low)	1	22	Odds Ratio (M‐H, Fixed, 95% CI)	3.29 [0.12, 89.81]
6.15 Colestilan plus UDCA (low) versus UDCA (low)	1	11	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.16 Methotrexate plus UDCA (low) versus UDCA (low)	1	25	Odds Ratio (M‐H, Fixed, 95% CI)	115.0 [4.98, 2657.48]
6.17 UDCA (moderate) versus UDCA (low)	1	59	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6.18 Azathioprine plus UDCA (moderate) versus UDCA (moderate)	1	42	Odds Ratio (M‐H, Fixed, 95% CI)	19.67 [0.94, 413.50]
6.19 Colchicine plus UDCA (moderate) versus UDCA (moderate)	2	42	Odds Ratio (M‐H, Fixed, 95% CI)	6.20 [0.63, 60.80]
6.20 Glucocorticosteroids plus UDCA (moderate) versus UDCA (moderate)	2	135	Odds Ratio (M‐H, Fixed, 95% CI)	5.54 [1.35, 22.84]
6.21 TauroUDCA (moderate) versus UDCA (moderate)	1	30	Odds Ratio (M‐H, Fixed, 95% CI)	21.0 [2.16, 204.61]
7 Adverse events (number) Show forest plot	15		Rate Ratio (Fixed, 95% CI)	Subtotals only

7.1 Chlorambucil versus no intervention	1		Rate Ratio (Fixed, 95% CI)	3.67 [1.04, 12.87]
7.2 Cyclosporin versus no intervention	3		Rate Ratio (Fixed, 95% CI)	1.87 [1.51, 2.32]
7.3 D‐Penicillamine versus no intervention	3		Rate Ratio (Fixed, 95% CI)	2.64 [1.78, 3.91]
7.4 Malotilate versus no intervention	1		Rate Ratio (Fixed, 95% CI)	6.13 [1.38, 27.14]
7.5 Obeticholic acid (high) versus no intervention	1		Rate Ratio (Fixed, 95% CI)	1.91 [1.50, 2.44]
7.6 Obeticholic acid (low) versus no intervention	1		Rate Ratio (Fixed, 95% CI)	1.05 [0.80, 1.39]
7.7 Obeticholic acid (moderate) versus no intervention	1		Rate Ratio (Fixed, 95% CI)	1.25 [0.97, 1.62]
7.8 Obeticholic acid (low) versus obeticholic acid (high)	1		Rate Ratio (Fixed, 95% CI)	0.55 [0.43, 0.70]
7.9 Obeticholic acid (moderate) versus obeticholic acid (high)	1		Rate Ratio (Fixed, 95% CI)	0.66 [0.53, 0.81]
7.10 Obeticholic acid (moderate) versus obeticholic acid (low)	1		Rate Ratio (Fixed, 95% CI)	1.19 [0.93, 1.53]
7.11 UDCA (low) versus UDCA (high)	1		Rate Ratio (Fixed, 95% CI)	2.08 [0.78, 5.53]
7.12 UDCA (moderate) versus UDCA (high)	1		Rate Ratio (Fixed, 95% CI)	0.73 [0.21, 2.60]
7.13 UDCA (low) plus methotrexate versus UDCA (low)	1		Rate Ratio (Fixed, 95% CI)	30.64 [1.84, 510.76]
7.14 UDCA (moderate) versus UDCA (low)	1		Rate Ratio (Fixed, 95% CI)	0.35 [0.11, 1.10]
7.15 Azathioprine plus glucocorticosteroids plus UDCA (moderate) versus UDCA (moderate)	1		Rate Ratio (Fixed, 95% CI)	1.32 [0.88, 1.97]
7.16 Bezafibrate plus UDCA (moderate) versus UDCA (moderate)	1		Rate Ratio (Fixed, 95% CI)	11.79 [0.65, 213.14]
7.17 Colchicine plus UDCA (moderate) versus UDCA (moderate)	1		Rate Ratio (Fixed, 95% CI)	5.91 [0.28, 123.08]
7.18 TauroUDCA (moderate) versus UDCA (moderate)	1		Rate Ratio (Fixed, 95% CI)	1.17 [0.81, 1.71]
8 Liver transplantation Show forest plot	12		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

8.1 Cyclosporin versus no intervention	2	378	Odds Ratio (M‐H, Fixed, 95% CI)	0.86 [0.43, 1.72]
8.2 D‐Penicillamine versus no intervention	1	189	Odds Ratio (M‐H, Fixed, 95% CI)	0.93 [0.06, 15.05]
8.3 Methotrexate versus no intervention	1	60	Odds Ratio (M‐H, Fixed, 95% CI)	0.17 [0.02, 1.58]
8.4 UDCA (low) versus no intervention	2	162	Odds Ratio (M‐H, Fixed, 95% CI)	0.99 [0.24, 4.06]
8.5 UDCA (moderate) versus no intervention	3	478	Odds Ratio (M‐H, Fixed, 95% CI)	0.88 [0.44, 1.76]
8.6 UDCA (low) versus UDCA (high)	1	106	Odds Ratio (M‐H, Fixed, 95% CI)	3.17 [0.13, 79.71]
8.7 UDCA (moderate) versus UDCA (high)	1	103	Odds Ratio (M‐H, Fixed, 95% CI)	3.37 [0.13, 84.70]
8.8 UDCA (moderate) versus UDCA (low)	1	101	Odds Ratio (M‐H, Fixed, 95% CI)	1.06 [0.06, 17.47]
8.9 Bezafibrate plus UDCA (moderate) versus UDCA (moderate)	1	27	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
8.10 Methotrexate plus UDCA (moderate) versus UDCA (moderate)	1	265	Odds Ratio (M‐H, Fixed, 95% CI)	0.70 [0.35, 1.39]
9 Decompensated liver disease Show forest plot	7		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

9.1 D‐Penicillamine versus no intervention	1	24	Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
9.2 UDCA (moderate) versus no intervention	2	351	Odds Ratio (M‐H, Fixed, 95% CI)	1.33 [0.84, 2.12]
9.3 Obeticholic acid (low) plus UDCA (moderate) versus UDCA (moderate)	1	216	Odds Ratio (M‐H, Fixed, 95% CI)	1.55 [0.06, 38.46]
9.4 UDCA (low) plus colchicine versus UDCA (low)	1	84	Odds Ratio (M‐H, Fixed, 95% CI)	0.21 [0.04, 1.07]
9.5 Azathioprine plus UDCA (moderate) versus UDCA (moderate)	1	42	Odds Ratio (M‐H, Fixed, 95% CI)	0.52 [0.05, 5.18]
9.6 Glucocorticosteroids plus UDCA (moderate) versus UDCA (moderate)	1	66	Odds Ratio (M‐H, Fixed, 95% CI)	0.55 [0.11, 2.69]
9.7 Methotrexate plus UDCA (moderate) versus UDCA (moderate)	1	151	Odds Ratio (M‐H, Fixed, 95% CI)	2.00 [0.79, 5.04]
9.8 Glucocorticosteroids plus UDCA (moderate) versus azathioprine plus UDCA (moderate)	1	50	Odds Ratio (M‐H, Fixed, 95% CI)	1.06 [0.10, 11.18]
10 Cirrhosis Show forest plot	3		Odds Ratio (M‐H, Fixed, 95% CI)	Subtotals only

10.1 Azathioprine versus no intervention	1	31	Odds Ratio (M‐H, Fixed, 95% CI)	0.79 [0.18, 3.41]
10.2 UDCA (low) versus no intervention	1	22	Odds Ratio (M‐H, Fixed, 95% CI)	0.15 [0.01, 1.53]
10.3 Azathioprine plus glucocorticosteroids plus UDCA (moderate) versus UDCA (moderate)	1	50	Odds Ratio (M‐H, Fixed, 95% CI)	0.28 [0.03, 2.90]

Comparison 2. Stratified by dose