Background Although several meta‐analyses have examined the effects of off‐label underdosing of nonvitamin K antagonist oral anticoagulants (NOACs) compared with their recommended doses in patients with atrial fibrillation (AF), they combined different kinds of NOACs in their primary analyses. Herein, we first conducted a meta‐analysis to separately assess the effects of off‐label underdosing versus on‐label dosing of four individual NOACs on adverse outcomes in the AF population. Methods The PubMed and Embase database were systemically searched until November 2021 to identify the relevant studies. Adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) were pooled by utilizing a random‐effects model. Results A total of nine studies with 144,797 patients taking NOACs were included in the meta‐analysis. In the pooled analysis, off‐label underdosing of rivaroxaban was related to an increased risk of stroke or systemic embolism (HR = 1.31, 95% CI 1.05–1.63; p = .02), whereas off‐label underdosing of apixaban was associated with a higher risk of all‐cause death (HR = 1.21, 95% CI 1.05–1.40; p = .01). When comparing off‐label underdosing versus on‐label dosing of dabigatran or edoxaban, no differences were found in the primary and secondary clinical outcomes. Conclusion Off‐label underdosing of rivaroxaban may increase the risk of stroke or systematic embolism, whereas off‐label underdosing of apixaban may heighten the incidence of all‐cause death.

BackgroundSeveral published studies have examined the association of coffee consumption with atrial fibrillation (AF) risk, but their findings are still controversial. Therefore, we performed a systematic review and dose–response meta-analysis of prospective studies to determine the relationship between coffee consumption and the risk of incident AF.MethodsWe systematically retrieved the PubMed and Embase databases until October 2021 for pertinent studies that reported the association of coffee consumption (caffeinated or decaffeinated coffee) with AF risk. A cubic spline random-effects model was used to fit the potential dose–response curve. The effect estimates were expressed as adjusted risk ratios (RRs) and 95% CIs.ResultsA total of 10 prospective studies (11 cohorts) involving 30,169 AF events and 723,825 participants were included. In the dose–response analysis, there was a linear inverse association between coffee intake and risk of AF although not statistically significant (Pnon–linearity = 0.25). Compared with participants with no coffee consumption, the RRs (95% CI) of AF risk estimated directly from the dose–response curve were 1.01 (0.98–1.03), 1.00 (0.97–1.04), 0.99 (0.92–1.02), 0.95 (0.89–1.01), 0.94 (0.87–1.01), 0.89 (0.79–1.02), and 0.87 (0.76–1.02) for 1–7 cups of coffee per day, respectively. One cup per day increased in coffee consumption was associated with a 2% reduced risk of AF (RR = 0.98, 95% CI: 0.97–1.00, P = 0.02).ConclusionsOur evidence from this meta-analysis suggested that coffee consumption had a trend toward reducing the risk of AF in a dose–response manner. Further studies could be conducted to reinforce our findings.