Conducting and interpreting results of network meta-analyses in type 2 diabetes mellitus: A review of network meta-analyses that include sodium glucose co-transporter 2 inhibitors. (February 2019)

• Record Type: Journal Article
• Authors: Willis, Michael; Asseburg, Christian; Neslusan, Cheryl
• Is Part Of: Diabetes research and clinical practice. Volume 148(2019)
• Page Start: 222
• Page End: 233
• Publication Date: 2019-02
• Subjects: Network meta-analysis -- Sodium glucose co-transporter 2 inhibitor -- Systematic review -- Type 2 diabetes mellitus; Diabetes -- Periodicals; Diabetes Mellitus -- Periodicals; 616.462
• DOI: 10.1016/j.diabres.2019.01.005
• Languages: English
• ISSNs: 0168-8227

Abstract

Highlights: Network meta-analyses provide comparative effectiveness data for diabetes therapies. We evaluated 35 network meta-analyses of sodium glucose co-transporter 2 inhibitors. Most network meta-analyses followed methodological guidelines. Use of these guidelines has improved the quality of data reporting. Diabetes-related guidance may improve network meta-analysis data interpretation/use. Abstract: Aims: Network meta-analyses (NMAs) are valuable ways to generate comparative effectiveness data for therapies available to treat type 2 diabetes mellitus (T2DM). This review assesses NMAs that evaluate sodium glucose co-transporter 2 (SGLT2) inhibitors for treatment of T2DM and discusses potential issues in conducting and interpreting NMAs. Methods: A systematic literature search was conducted on September 13, 2018 using the search terms "network meta-analysis, " "SGLT2, " variations of these terms, and individual SGLT2 inhibitor names. Extracted data included NMA objectives, methods, target populations, treatments, study endpoints, length of follow-up, and funding. Differences between NMAs were investigated. Results: Thirty-five full-length publications met criteria for inclusion. In most NMAs, the target population was defined by therapeutic regimen (e.g., combination with metformin). Follow-up intervals permitted in NMAs varied considerably (range, 4–208 weeks). Twenty-nine NMAs included dapagliflozin, 28 evaluated canagliflozin, and 27 evaluated empagliflozin. Nine NMAs used frequentist methods; 16 used Bayesian methods. Six NMAs were funded by pharmaceutical manufacturers. Heterogeneity across NMAs was seen in scope, time frame, and other aspects of analytic design. Conclusions: Although this review indicates that methodological guidelines for reporting NMAs were generally followed, it also emphasizes the need for T2DM-specific guidance requiring clear reporting of NMA scope and objectives to aid appropriate interpretation and use of NMA results.