OBJECTIVES: To critically appraise the published network meta-analyses (NMAs) evaluating the efficacy or safety of the new oral anticogulants (NOACs) dabigatran, rivaroxaban and apixaban for the prevention of stroke in patients with non-valvular atrial fibrillation (AF). METHODS: A systematic literature review was performed to identify the relevant NMAs using MEDLINE®, EMBASE®, Cochrane Library, Database of Abstracts of Reviews of Effects, and Health Technology Assessment. The synthesis studies were evaluated using the ‘Questionnaire to assess the relevance and credibility of the NMA’.  RESULTS: Eleven NMAs evaluating NOACs among adults with non-valvular AF were identified. Most NMAs included three large phase III RCTs, comparing NOACs to adjusted-dose warfarin (RE-LY, ROCKET-AF, ARISTOTLE). The main differences identified related to potential treatment effect modifiers regarding the mean time spent in therapeutic range (TTR) in the warfarin arm, the risk of stroke or systemic embolism across the trials (mean CHADS score: Cardiac failure, Hypertension, Age ≥ 75 years, Diabetes mellitus, Stroke, 2 two points for stroke) or primary versus secondary prevention, and type of populations used in the analysis. Kansal et al. appropriately adjusted the ROCKET-AF TTR to match the RE-LY population based on individual patient data. Meta-regressions are not expected to minimize confounding bias given limited data, whereas subgroup analyses had some impact on the point estimates for the treatment comparisons.   CONCLUSIONS: Results of the synthesis studies were generally comparable and suggested the NOACs had similar efficacy, although some differences were identified depending on the outcome. The extent to which the differences in the distribution of TTR, CHADS or primary versus secondary prevention biased the results remains unclear.