Presentation (CTI)

OBJECTIVES: Retractions of scientific research have increased across different clinical fields, disease indications, and study designs. It is currently unknown to what extent retractions have appeared in the literature of indirect treatment comparisons (ITCs), particularly network meta-analyses (NMAs). This research aimed to describe the characteristics of retracted NMAs, identify reasons for retraction, and outline implications for health decision-making.
METHODS: We conducted a preregistered systematic review of retracted NMAs, using MEDLINE, Embase, PubMed, Retraction Watch, and RetractBASE, from inception up to 2026. Two hundred eighty-five abstracts and 23 full-text records were screened by two reviewers. Publication details, including reasons for retraction, were extracted.
RESULTS: We identified 19 retracted NMAs published between 2016 and 2024. NMAs incorporated a median 27 randomised trials (range: 5-101) in their networks, comparing a broad range of interventions (median: 8; range: 4-55) including pharmacotherapies, behavioural programmes, and alternative medicine, across a wide array of indications (including cancer, psychiatric disorders, neurological conditions). NMAs appeared in 17 unique journals, of which 9 were top-tier outlets based on journal ranking. NMAs remained in the evidence record for median 12.7 months (range: 0-48) before retraction. Errors were the reason for retraction in 7/19 NMAs, including errors in searches, data extraction, and/or computation. Evidence of misconduct, specifically the systematic manipulation of the peer review process, was found in 7/19 NMAs. Five NMAs did not report a reason for retraction. Median citation count was 4 (IQR: 0-8); one NMA notably had 321 citations.
CONCLUSIONS: Evidence for NMA retractions was found. Retracted NMAs spanned a variety of disease indications, interventions, and journal sources. Errors and scientific misconduct were reported reasons for extraction. NMAs were present in the evidence record for a year, leading to subsequent citations. Stakeholders should be made aware of retractions, not just of clinical trials but also of ITCs, as these can impact health decision-making.