OBJECTIVES: Patients with atrial fibrillation (AF) and a previous stroke or transient ischaemic attack (TIA) have a high risk of stroke and may have a different baseline risk than patients without previous stroke or TIA, which may act as a treatment effect modifier. Therefore, the comparative efficacy of new oral anticoagulants (NOACs) in terms of stroke or systemic embolism (SE) was assessed for the subgroup of patients with a previous stroke or TIA. METHODS: A Bayesian network meta-analysis (NMA) was performed for patients with previous stroke or TIA from three pivotal randomized controlled trials: ARISTOTLE, RE-LY, and ROCKET-AF, which compared apixaban, dabigatran, and rivaroxaban to warfarin, respectively. Parametric survival functions were used to model the hazard ratios (HR) over time for the compared interventions, and the difference in the shape and scale parameters of these functions was synthesized and indirectly compared. Results were compared to an analysis of constant HRs as well as to previous NMAs for this subgroup.  RESULTS: The time-varying HRs for the treatments versus warfarin suggest that each NOAC is at least as efficacious as warfarin with respect to stroke and SE. The HR for dabigatran 150mg was fairly constant over time (range: 0.80-0.68). The HR for dabigatran 110mg decreased slightly over time (range: 1.82-0.40), whereas the HRs increased slightly over time for rivaroxaban (range: 0.61-1.34) and apixaban (range: 0.62-0.94).  CONCLUSIONS: Based on the NMA of stroke or SE among patients with AF and a prior stroke or TIA, dabigatran 150mg and apixaban are expected to be comparable to warfarin; dabigatran 110mg is expected to be comparable to warfarin for the first 16 months and more efficacious up until 30 months; rivaroxaban is expected to more efficacious than warfarin for the first 4 months, and comparable to warfarin thereafter.