In the medical community there is not general consensus in either recommending or advising against OAT in HF. Several evidence have recommended OAT in HF patients with AF, as prevention for the high cardioembolic risk observed in this population. Differently, the majority of HF patients due to ischemic etiology takes aspirin for secondary prevention of coronary heart disease. Therefore, the aim of this present meta-analysis has been to assess potential differences in efficacy and safety between these two therapeutic modalities in HF patients in sinus rhythm. Our results indicate that warfarin does not show a better efficacy-safety profile compared to aspirin in preventing cardioembolism in HF patients without AF. However, when compared to aspirin, warfarin was associated with a significant reduction (OR 0.49) of ischemic stroke incidence with a relative risk reduction that is comparable to that reported in HF patients with AF treated with OAT 
. Nevertheless, due to the low annual stroke rate observed in HF patients in sinus rhythm (between 0.8% and 3.2% per year), the advantages deriving from routine anticoagulation cannot overcome the increased risk of bleedings related to warfarin use 
. Consistently, our data show that OAT was associated with a more than doubled bleeding risk, with a trend to increase also the risk of intracerebral bleedings. It is important to underline, that only in the HELAS study, HF patients were dived for ischemic or non ischemic cardiomyopathy, thus no conclusions can be drown about efficacy or safety in these different patient subpopulations.
In HF patients with non ischemic etiology and without AF, no studies are available about efficacy and safety of OAT or antiplatelet therapy compared to placebo, whereas several line of evidence support the use of aspirin in HF patients due to ischemic etiology as secondary prevention of coronary artery disease. Thus, RCTs comparing aspirin vs. placebo in HF patients in sinus rhythm and non ischemic etiology might be helpful to guide therapy in this specific HF subpopulation. In the present analysis the hospital admission rate for worsening HF was lower in warfarin group only after exclusion of WARCEF study. Its exclusion reduces the heterogeneity among studies but unfairly reduces the available information since this is the largest trial with the longest follow-up. The important issue regarding discordant results on the incidence of hospital admissions for worsening HF in patients treated with warfarin or aspirin is not addressable. Unfortunately, despite it would be of great interest whether warfarin treatment impacts quality of life (especially in terms of anxiety burden related to OAT monitoring), the identified trials did not investigate this relevant aspect which certainly would be an important argument for future trial investigations.
Strengths and Limitations of this Meta-analysis
The main strengths of our review include the systematic strategy and the high score at Cochrane quality assessment for all trials included. Our meta-analysis has one major shortcoming. It was not carried out on individual patients data exploring subgroups with higher thromboembolic risk. Moreover, there are some limitations in the outcome evaluation due to the too small sample size and the too short follow up (i.e. 2 years) period, resulting in a low number of events in the trials included. The exclusion of gray literature could be a limitation of our search strategy. It is important to underline that in WATCH trial, aspirin dosage is lower than that used in the other trials, thus explaining the lower efficacy of treatment reported. Although differences in trial definitions of outcomes among different trials should be considered for the interpretation of the overall result, it is important to mention that mortality and ischemic stroke, the principal efficacy outcomes evaluated in our study, are hard endpoint not affected by study definitions. The absence of heterogeneity in the most part of analysis supports the strength of our results. Further studies are needed to better identify high risk HF subgroups. The definition of an HF risk stratification score, similar to that available for ischemic risk assessment (CHADS2 or CHA2DS2-VASc scores), or bleeding risk assessment (HAS-BLED) 
, will be useful to identify HF patients at higher risk. Finally, no evidences are provided regarding the use of new oral anticoagulants (oral direct thrombin inhibitors, oral Factor Xa inhibitors) which seem to offer a different risk–benefit profile compared to warfarin and might induce a reduction in ischemic stroke rates with less risk of major bleeding. Thus, an head to head comparison between warfarin and new anticoagulants (rivaroxaban, apixaban and dabigatran), with antiplatelet therapy might be of great interest in HF patients in sinus rhythm.
In patients with HF in sinus rhythm, warfarin and aspirin seem to be similar in reducing mortality. Warfarin reduces the incidence of ischemic stroke, but increases major bleedings. Thus, it is possible to speculate that aspirin could be indicated in patients with high risk of bleeding, whereas warfarin could be preferred in patients with high thromboembolic risk. However, further studies are needed to clarify the role of antitrombotic therapy in HF patients in sinus rhythm, particularly in the subpopulation with non ischemic etiology.