Alcohol intake and risk of stroke in atrial fibrillation: the lesser the better, but this is not enough

This editorial refers to ‘Lower risk of stroke after alcohol abstinence in patients with incident atrial fibrillation: a nationwide population-based cohort study’, by S.R. Lee et al., doi:10.1093/eurheartj/ehab315. Risk factors and impact of lifestyle modifications to reduce atrial fibrillation (A) Multiple risk factors for atrial fibrillation are shown with implications for lifestyle modifications to prevent arrhythmia recurrence, including control of hypertension, sleep apnoea, diabetes mellitus, hyperlipidaemia, weight loss, exercise, and abstinence from alcohol. (B) Potential mechanisms by which alcohol may cause atrial fibrillation and contribute to stroke risk are shown. There has been increased awareness about the importance of lifestyle management and risk factor control in the treatment and prevention of atrial fibrillation (AF).1 This includes treatment of obesity, hypertension, diabetes mellitus, and obstructive sleep apnoea, as well as exercise to improve cardiovascular health. Sustained weight loss is associated with a significant reduction of AF burden and improves arrhythmia-free survival in patients with obesity and AF,2 and a tailored exercise programme to improve cardiorespiratory fitness adds to the beneficial effects of weight loss.3 In patients undergoing AF ablation, aggressive risk factor management improves the long-term success of the procedure.4 The importance of addressing modifiable risk factors has been highlighted in recent AF practice guidelines.5 , 6 Alcohol intake has been demonstrated to increase the risk of incident AF, and this should also be addressed as a modifiable risk factor.7 , 8 A randomized controlled trial in Australia demonstrated that abstinence or a substantial reduction in alcohol intake can reduce AF recurrence and also reduce the proportion of time spent in AF in individuals who are regular drinkers.9 While alcohol intake has been demonstrated to increase the risk of incident AF, its association with stroke risk has been controversial and varied between studies.10 Consistent conclusions about the relationship between alcohol intake and stroke may be limited by follow-up duration, type of stroke (ischaemic vs. haemorrhagic), amount of alcohol consumed, pattern of drinking, and adjustment for other risk factors in various studies. While the CHA2DS2-VASc score is a standard and convenient risk assessment tool recommended in practice guidelines, there are many other factors that can influence stroke risk that are not assessed by this score. Some of these additional risk factors include renal dysfunction, left atrial (LA) volume, LA flow velocity, spontaneous echo contrast or smoke in the left atrium, LA fibrosis, LA appendage morphology, and AF burden, to name a few. Should alcohol intake be considered another factor in thrombo-embolic risk assessment? In the current issue of the European Heart Journal, Lee et al. demonstrated that continued alcohol intake in patients with newly diagnosed AF was not only associated with a higher risk of ischaemic stroke compared with non-drinkers, but there was a linear dose–response relationship between the amount of current alcohol intake and the risk of stroke.11 In addition, abstinence from alcohol was associated with a reduced risk of stroke in patients with newly diagnosed AF.11 The authors used data from the Korean National Health Insurance Service (NHIS) that provides insurance coverage by the government for the entire Korean population. This database includes nationwide claims and health examination information, including information about biennial health screening that incorporates blood pressure measurement, laboratory studies, and completion of a self-reported questionnaire on health behaviours such as smoking, alcohol consumption, and physical activity. Data were extracted on a total of 97 869 subjects with newly diagnosed AF from 2010 to 2016. Patients were categorized into three groups according to alcohol consumption before and after AF diagnosis, namely ‘non-drinkers’, ‘abstainers from alcohol after AF diagnosis’, and ‘current drinkers’. The primary outcome was incident ischaemic stroke during follow-up. After balancing for differences in baseline characteristics using propensity scores, inverse probability of treatment weighting analysis was used to compare the three groups. During the period of follow-up, 3120 patients were diagnosed with incident ischaemic stroke, and current drinkers had a significantly higher risk of ischaemic stroke compared with non-drinkers. In addition, abstinence from alcohol after the diagnosis of AF was associated with a 14% reduction in risk of ischaemic stroke [hazard ratio (HR) 0.86, 95% confidence interval (CI) 0.77–0.96] compared with current drinkers. Why might abstinence from alcohol be associated with reduced stroke risk? There are several potential explanations. (i) Alcohol intake may directly increase the risk for ischaemic stroke in some manner, and alcohol cessation may reduce this risk. For example, alcohol consumption may be associated with electrical and structural changes in the left atrium, with direct cellular effects on atrial myocytes and influences on autonomic function, creating a milieu for onset and maintenance of AF.12 Moderate alcohol consumption has been associated with lower atrial voltage and conduction slowing identified during high-density LA mapping in patients presenting for AF ablation.13 Adverse effects may be dose related, and, while there is a perception that low doses of alcohol may actually offer some cardioprotective benefit, these benefits do not appear to extend to AF. Alcohol cessation may impact on stroke risk by a reduction in AF burden. It has been previously identified that a greater AF burden is associated with a higher risk of ischaemic stroke in the absence of anticoagulation, and this is independent of other known risk factors for stroke, including CHA2DS2-VASc score.14 It is currently unclear what burden of AF is required to increase the risk for stroke; however, one study suggested that this duration may be >24 h.15 Patients who had subclinical AF detected on implantable device interrogation lasting 6 min to 24 h had a similar risk of stroke or systemic embolism as patients without subclinical AF.15 (ii) Alcohol cessation may be associated with other healthy lifestyle changes (such as concomitant smoking cessation, weight loss, and increased exercise) or better control of other risk factors for stroke (such as control of hypertension or diabetes mellitus) that are associated with the reduced risk for ischaemic stroke. (iii) Alcohol cessation may result in better adherence to medication, including oral anticoagulation when indicated for thrombo-embolic prophylaxis or use of antihypertensive therapy, that is associated with reduced risk for ischaemic stroke. The authors acknowledge some important study limitations. A self-reported questionnaire on health behaviours, including smoking, alcohol consumption, and physical activity, was utilized. It is possible that alcohol use could have been under-reported or may be subject to recall bias. While there was adjustment for differences in baseline characteristics between current drinkers and other groups, including a higher percentage of smoking in current drinkers and a lower use of oral anticoagulation in current drinkers, other residual confounding cannot be excluded. In addition, alcohol intake may secondarily impact on other comorbidities (such as hypertension and obesity), and the impact of alcohol cessation may have been at least partially related to control of other risk factors with residual confounding despite adjustment. It is also well established that AF burden can impact on stroke risk, and there was no measure of AF burden in this study. As this was an Asian population—with a reported overall lower prevalence of AF compared with Caucasians—these results are not necessarily applicable to other cohorts. Targeting modifiable risk factors should be an essential component of AF care (Graphical Abstract). In addition to rate control, rhythm control, and anticoagulation, risk factor management is now considered the ‘fourth pillar’ of AF care.1 Alcohol consumption is a modifiable risk factor for AF. While more research is needed to identify mechanisms by which alcohol cessation may result in improved outcomes and stroke reduction, an integrated and comprehensive approach to care is essential for patients with AF. Practice guidelines and professional society recommendations have highlighted the importance of a patient-centred and an integrated team-based approach to care.5 , 6 , 16 The authors should be commended on further highlighting the importance of modifiable risk factors, including alcohol intake, in an integrated approach to healthcare delivery for the treatment of AF. In addition to provider recognition of the importance of lifestyle and risk factor management in the treatment of AF, patient education and engagement will provide a foundation for successful risk factor modification. Conflict of interest: none declared. The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.