Atrial fibrillation (AF) is the most common sustained dysrhythmia encountered in clinical practice in North America and Europe, accounting for approximately one-third of all hospitalizations for a cardiac rhythm abnormality [1]. The presence of AF markedly increases the patient’s risk for developing arterial embolism and stroke, depending on the presence of other clinical conditions, such as hypertension and diabetes [2]. AF is associated with a fivefold increased risk for stroke, and is estimated to cause 15% of all strokes. The rate of ischemic stroke among patients with non-valvular AF averages 5% per year, 2–7 times that of people without AF. Additionally, when transient ischemic attacks (TIAs) and clinically “silent” strokes are considered, the rate of brain ischemia accompanying non-valvular AF exceeds 7% per year. Patients with AF frequently have several risk factors for atherosclerosis, including hypertension, diabetes, and dyslipidemia [3, 4]. Accordingly, systemic signs of atherosclerosis can be detected in AF patients, and these likely account for an enhanced risk of coronary heart disease (CHD).

In addition to cerebrovascular disease, patients with AF may suffer from coronary events including myocardial infarction (MI), but the rate of MI in AF patients seems to be variable, but often underestimated. Surprisingly few studies with antithrombotic drug therapy include MI as an end-point.

We reasoned that analysis of the MI rate in AF could be useful for future interventional trials with anti-thrombotic drugs. Thus, we evaluated the rate of MI in the clinical trials where this clinical end-point was taken into account, and tried to relate it with the common atherosclerotic risk factors.

Searches were conducted in computerized publication database (Medline) with hand searches of publications from January 1995 to September 2008. Bibliographies of all selected articles and review articles were reviewed for other relevant articles.

We used search terms including “atrial fibrillation” and “cardiovascular risk”, “atrial fibrillation” and “myocardial infarction”, “atrial fibrillation” and “coronary heart disease”, “atrial fibrillation” and “mortality”, “atrial fibrillation” and “cardiovascular events”. The search was confined to human studies and English language restricted.

We obtained 278 articles. A trial was included if the overall study population included AF patients, it was published in a peer-reviewed journal, and major adverse coronary events (MACE) were recorded through the follow-up. A trial was excluded from our analysis if the AF population was lower than 150 subjects. Thus, 13 studies were selected according to the above-mentioned inclusion and exclusion criteria [5–17] (Table 1). In only two of these, MACE were included as primary end-points [5, 6] (Table 1A).

Table 1 Studies characteristics
Full size table

In the first study, a very high incidence of new MACE was found in AF patients (74% during a mean follow-up of 30 months) compared with patients with a sinus rhythm (42%) or a supraventricular tachycardia (43%) [5]. These patients had a mean age of 81 years, and a clinical history complicated by CHD in >50% of cases. After controlling for other prognostic variables, patients with AF and CHD showed 2.2 times higher probability of developing new coronary events than other patients with CHD and no AF. The second study was performed on AF patients with no clinical history of CHD [6]. Of the 2,768 patients included, 17% had a first MACE during a mean follow-up of 6 years on average. The unadjusted incidence was 3.1% person-years. In these patients of advancing age, male gender, higher systolic blood pressure, history of systemic hypertension, diabetes mellitus and PAD were significant independent predictors of a first MACE.

Among the remaining 11 studies [7–17], in which MACE had been recorded during the follow-up, we observed a high variability (Table 1B). In fact, the rate of fatal and non-fatal MI ranged from 0.5 to 4%/year.

We analyzed if there were a clinical history of CHD that could account for such variability. A cut-off of 19% (median) of a positive history for previous CHD was used to analyze the data. This value is in agreement with the prevalence of CHD reported in the general population. In fact, the NHANES III study [18], conducted on American adults, reports an incidence of CHD of 18.7% in subjects aged 65 years or older. In the six studies in which more than 19% of patients had previous CHD, the average incidence of MI was 6.9% [7–12]. In contrast, in the remaining 5 studies in which a clinical history of CHD was detected in <19% of cases, MI occurred significantly less frequently(1.4%; Test for difference between two proportions: P < 0.0001) [13–17].

Atherosclerotic risk factors may be another important variable explaining the large variability of MI, but, as shown by the Table, data analysis is difficult because of incomplete report of risk factors prevalence. However, from the data of the two trials with adequate analysis of risk factors [9, 14], a relation seems to exist because MI occurs more frequently in groups of patients with a higher prevalence of risk factors such as diabetes and hypertension.

There are also anatomic and clinical studies suggesting a true link between AF and the risk of CHD. In a recent study, a correlation between coronary artery calcification, a marker of coronary atherosclerosis and MACE, and enlarged left atrium (LA), left atrial appendage (LAA) and pulmonary veins (PVs) was observed [19]. The coexistence of systemic atherosclerosis is documented by anatomical changes typical of atherosclerotic damage in the vascular tree. Thus, complex aortic plaque identified by transoesophageal echocardiography (TEE) is common and occurs in up to 57% of patients with AF, of whom about 25% have complex plaque (i.e., thicker than 4 mm and with ulceration, pedunculation, or mobile elements) [20]. The presence of complex plaque on the descending aorta is a risk factor for stroke [21]. Aortic plaque aids identification of patients who are at high stroke risk by virtue of the presence of associated vascular risk factors or atherothrombotic disease, in addition to AF. Coexistence of peripheral arterial disease (PAD) is a relevant clinical sign of systemic atherosclerosis. Thus, two large studies in patients with AF document the existence of PAD in about 3–5% of patients [22, 23]. It is possible, however, that such an incidence has been underestimated as only symptomatic patients were considered as affected by PAD. As PAD is an important marker of systemic atherosclerosis, its association with AF reinforces the concept that patients with AF have systemic atherosclerosis that potentially account for coronary complications.

AF is also associated with biomarkers that are typically observed in patients with atherosclerotic disease [24–27]. In particular, we recently reported that sCD40L, a marker of platelet activation, is elevated in AF patients, and predicts the recurrence of both stroke and MI [28]. These findings are likely to account for the small but significant reduction of vascular events with aspirin in this setting [29].

Based on these considerations, we believe that MI may have an important impact in the clinical progression of AF depending on the type of AF patients. The clinical history of CHD may be a simple approach to identify patients at high risk, but an analysis of other laboratory and clinical variables including the coexistence of PAD, could provide more information to stratify the atherosclerotic risk. This may help to explain the large variability of MI rate so far reported in an AF population, and to better evaluate the atherothrombotic risk in this clinical setting. In the era of new antithrombotic agents, exclusion or poor consideration of MI in the end-points of AF trials cannot fully reveal information from the clinical history of AF.