Evaluation of network meta-analysis in clinical practice guidelines for percutaneous coronary intervention: A systematic review

Tags: Poster
Zhang L1, Chen W1, Kwong JS1, Xu Y1, Chen S1, Chen Q2, Huang D1, Zhang M1
1West China Hospital of Sichuan University, 2West China School of Public Health, Sichuan University

Background:Clinical practice guidelines (CPGs) have an important role in guiding choices among the numerous options in percutaneous coronary interventions (PCIs). Meta-analysis provides much valuable clinical evidence for Clinical practice guideline. but little is known about the influence of network meta-analysis (NMA).

Objectives:To assess whether NMA in clinical practice guidelines for PCI are consistent with current evidence and whether the consistency of the guidelines depends on the quality of guideline development.

Methods:Web of Science, MEDLINE, Embase, and Cochrane Library were searched from inception to August 2016, Two reviewers independently screened citations to identify English-language guidelines on PCI. Reviewers assessed whether the guidelines addressed and agreed with conclusions from these network meta-analysis. Two reviewers independently rated NMA quality by using AMSTAR.

Results:Of the 803 screened citations, 29 NMA met the inclusion criteria. Most of NMAs (86.2%) published from America and Europe. Twelve NMAs (41.4%) were cited by CPGs. Two NMAs included observations study and the rest all the randomised-controlled trials(RCTs). All-cause mortality, myocardial infarction, target vessel revascularisation and stent thrombosis are the main outcome with high attention. NMAs with higher quality are much more likely to be cited by CPGs and which were high consistent with the evidence-based conclusions.

Conclusions:Not all clinical practice guidelines on PCI were consistent with available evidence from NMA. Guidelines judged to be of higher quality contained more recommendations consistent with evidence-based conclusions. The quality of guideline development processes varied substantially.