Table 1.
Current recommendation for the use of aspirin in patients with non-valvular atrial fibrillation.
Fig 1.
None of the studies reported a standardized risk assessment score (such as the CHADS-2) to estimate the risk of thromboembolic for their population. The AFASAK II [13] was the only study enrolling patients with a prior history of stroke or transient ischemic attack, and it also included a large number of patients with heart failure (>70%). See Table 1 for studies characteristics and Table 2 for quality assessment. Both the PATAF [11] and Vemmos trial [12] excluded patients with a prior history of stroke or transient ischemic attack. The most common risk factor for thromboembolic event in both trials was hypertension. See Table 2 for studies characteristics and Table 3 for quality assessment.
Table 2.
Characteristics of the studies included in the main analysis.
Table 3.
Quality assessment.
Fig 2.
Meta-analysis of ischemic stroke or systemic embolism.
There was no difference in the rate of major bleeding [OR 1.06 (95% CI 0.42–2.62); I2 0%] or vascular death [OR 1.04 (95% CI 0.61–1.75); I2 1%] but patients treated with aspirin had an increased risk in all-cause mortality [OR 1.66 (95% CI 1.12–2.48); I2 0%] (Fig 3). The difference in all-cause mortality was driven by an increased risk in non-vascular death in patients treated with aspirin [OR 3.20(95% CI 1.31–7.82); I2 0%], whereas the risk for death from unknown causes not significantly different [OR 1.525 (95% CI 0.65–3.55; I2 0%]. Table 4 provides the number of events in each study.
Table 4.
Event rates in individual studies.
Fig 3.
All cause mortality meta-analysis.
The addition of a study arm from the AFASAK study [13] comparing aspirin vs. low-intensity anticoagulation plus aspirin did not modify any of the estimates including the reduction in all-cause mortality [OR 1.66(95% CI 1.15–2.38); I2 0%]. Table 3 presents a summary of the number of individual events from each study.