The four studies included in the analysis were RE-LY, ROCKET-AF, ARISTOTLE and ENGAGE AF-TIMI48; all used warfarin as the comparator. The baseline characteristics were different in the four trials, as were inclusion and exclusion criteria for concomitant use of antiplatelets. Altogether just over 30% of patients enrolled in the studies received concomitant aspirin, mainly low-dose (≤100mg).
The characteristics of the patients on aspirin therapy were different from the total study populations as they were older and had more cardiovascular diseases in previous history. Reporting on aspirin persistence and prevalence during the study varied considerably; the definitions of the various outcomes also differed slightly.
The main outcomes presented for use of NOAC as compared with VKA were:
• Lower risk for stroke or systemic embolism (HR 0.78; 95% CI 0.67-0.91) and vascular death (0.85; 0.76-0.93)• No difference for ischemic stroke (0.93; 0.73-1.18) • Numerically lower incidence of major bleeding but not statistically significant (0.83; 0.69-1.01)• Lower risk of intracranial haemorrhage (0.38; 0.26-0.56) • Numerically higher incidence of myocardial infarction but not statistically significant (1.16; 0.97-1.39)
The authors say their results suggest it may be safer and more effective to use NOACs as compared to vitamin K antagonists in patients with non-valvular AF who are on concomitant aspirin therapy. They do however acknowledge a number of limitations to their study, including differences between the studies in terms of baseline characteristics, design and conduct, the exclusion of important patient groups, the fact that aspirin use was not randomised (differing indications), lack of information on changes in aspirin therapy throughout the trial, lack of information on gastrointestinal bleeding, and the assumption that all NOACS are comparable in terms of safety and efficacy (as data were pooled).