Examples of systematic reviews Goran Poropat

Cochrane systematic reviews To make unmanageable amounts of information – manageable Identify, appraise and synthesize research-based evidence Present in an accessible format

Cochrane systematic reviews Clearly stated set of objectives Explicit, reproducible methodology Systematic search Assessment of validity Systematic presentation

Protocol for a Cochrane review Idea Register title Write protocol To minimize the potential for bias in the review process Changes possible

Bezafibrate for primary biliary cirrhosis Title* Protocol information: Authors* Contact person* Dates What’s new History The protocol: Background* Objectives* Methods: Criteria for selecting studies for this review: Types of studies* Types of participants* Types of interventions* Types of outcome measures* Search methods for identification of studies* Data collection and analysis* Acknowledgements References: Other references: Additional references Other published versions of this review Tables and figures: Additional tables Figures Supplementary information: Appendices Feedback: Title Summary Reply Contributors About the article: Contributions of authors Declarations of interest* Sources of support: Internal sources External sources Published notes

Bezafibrate for primary biliary cirrhosis

Chronic, progressive, inflammatory and autoimmune- mediated liver disease Survival in symptomatic patients = yr. Bilirubin – indipendent predictor of survival and prognosis Bezafibrate – inhibition of acetil-CoA carboxylase activity PPAR  - ATP-binding cassette (ABC) B4 transporter Increased secretion of phosphatidyl-choline

Bezafibrate for primary biliary cirrhosis Methods Types of studies RCTs assessing bezafibrate in patients with PBC Irrespective of blinding, language, publication status Quasi-randomised and observational studies – Exluded for report of benefit – Included for report of harm

Bezafibrate for primary biliary cirrhosis Types of participants Pts with PBC – Elevated alkaline phophatases – Positive anti-mitochondrial antibody – Compatible liver biopsy Types of interventions – bezafibrate at any dose or regimen vs. placebo or any other drug

Bezafibrate for primary biliary cirrhosis Types of outcome measures

Bezafibrate for primary biliary cirrhosis Search methods of identification of studies Cochrane Hepato-Biliary Group Controlled Trials Register The Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library MEDLINE EMBASE Science Citation Index Expanded LILACS Chinese Bio-medical Literature Database Reference lists of identified studies Pharmaceutical companies Clinicaltrials.gov WHO International Clinical Trials Registry Platform

Bezafibrate for primary biliary cirrhosis Data collection and analysis Two authors independently Disagreements resolved by discussion Risk of bias assessment Allocation sequence generation Allocation concealment Blinding Incomplete outcome data Selective outcome reporting Other bias

Bezafibrate for primary biliary cirrhosis Statistical analyses Fixed-effect and random-effect models meta-analyses Dichotomous outcomes – RR/RD (95% CI) Continuous outcomes – MD (95% CI) Dichotomous outcomes – intention-to-treat Continuous outcomes – case analysis and inclusion of known data Assesment of heterogeneity (chi-square test) Trial sequential analysis

Search results 91 PUBLICATIONS IDENTIFIED 65 PUBLICATIONS ASSESSED 7 PUBLICATIONS INCLUDED 6 TRIALS INCLUDED 26 DUPLICATES 58 PUBLICATIONS EXCLUDED 4 TRIALS (N=82) Bezafibrate vs. no intervention 2 TRIALS (N=69) Bezafibrate vs. UDCA

Characteristics of included studies All trials from Japan Parallel groups design (5 trials) Cross-over design (1 trial) >85% participants were female Non-advanced PBC (4 trials) PBC stage not known (2 trials)

Characteristics of included studies Bezafibrate 400 mg daily orally Duration of administration of bezafibrate – 6 moths (2 trials) – months (4 trials) UDCA 600 mg daily orally

Risk of bias judgement All six trials are at high risk of bias

Results: Trials assessing bezafibrate vs. no intervention

LIVER-RELATED MORBIDITY ALL-CAUSE MORTALITY

PRURITUS ADVERSE EVENTS

ALKALINE PHOSPHATASES MD -186, 95% CI -249 to -123; I 2 = 34%

Trial sequential analysis (ALP) A minimal relevant difference100 U/L (based on clinical judgement, more modest than the estimated effect seen in the meta-analysis) Standard deviation 200 U/L Risk of type I error5% Risk of type II error20% Required information size216

TSA – ALKALINE PHOSPHATASES

Other results IgM Bilirubin  GT ALT Total cholesterol Tryglicerids No effect on

Conclusion Bezafibrate vs. no intervention No significant effect on mortality, liver-related morbidity, adverse events, and pruritus Quality of life and fatigue could not be assessed Trial sequential analysis implies evidence for a beneficial effect on serum alkaline phosphatases activity

Results: Trials assessing bezafibrate vs. ursodeoxycholic acid (UDCA)

LIVER-RELATED MORBIDITY ALL-CAUSE MORTALITY

ALKALINE PHOSPHATASES ADVERSE EVENTS

Other results Fixed-effect analysis Random-effect analysis Significant decrease of ALP,  GT, ALT, IgM No effect on  GT, ALT, IgM

Conclusion Bezafibrate vs. UDCA No significant effect on mortality, liver-related morbidity, adverse events, and pruritus Quality of life, pruritus, and fatigue could not be assessed There is no firm evidence of effect on ALP,  GT, ALT, and IgM

CONCLUSIONS Treatment of primary biliary cirrhosis with bezafibrate can neither be supported nor rejected