Distribution of thromboembolism in valvular versus non-valvular atrial fibrillation

Abstract

The most feared consequence of atrial fibrillation (AF) is thromboembolism, either to the brain causing stroke or to the non-cerebral circulation. Valvular atrial fibrillation (VAF) and non-valvular atrial fibrillation (NVAF) differ not only by morphological substrate of arrhythmia but also by the rate of thromboembolic complications, predisposing factors and destination of embolism. In the setting of VAF, there is a higher risk of thromboembolism and a higher prevalence of thrombus location within the body of the left atrium compared to NVAF. VAF is also associated with a proportionally higher propensity for non-cerebral thromboemboli than in NVAF. The distribution of non-cerebral thromboemboli appears to be similar in VAF and NVAF; however, more research needs to be done in this area, particularly with regard to VAF.

Keywords::

• distribution
• non-valvular atrial fibrillation
• risk factors
• thromboembolism
• valvular atrial fibrillation

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, with a lifetime risk of 22–26% in people aged 40–55 years. The point prevalence of AF in the US general population is 1–2% and in patients ≥80 years of age is 9% [1–3]. However, the actual prevalence may be higher as up to 35% of AF cases are subclinical [4]. In addition, the aging of the US population has increased the general incidence of this disease, which tends to be a dysrhythmia of the elderly [1–4].

AF carries an increased risk of first lifetime stroke, recurrent stroke and greater stroke-related disability and mortality compared to strokes of other etiologies [1–4]. While cardioembolic stroke is the most common complication of AF, peripheral emboli also occur. AF-associated cardioembolism accounts for 60–95% of acute limb ischemia, 31% of splenic infarction, 55% of renal infarction and 47% of acute mesenteric ischemia [5,6]. Notably, the mortality rates for patients with acute limb ischemia due to AF-associated embolization is considerably higher compared to those with either embolism in the setting of myocardial infarction or thrombosis in situ [6].

Definitions, incidence & prevalence of valvular atrial fibrillation & non-valvular atrial fibrillation

Although AF is one disease entity, it has long been recognized that different subtypes of AF are associated with different risks of complications. One important distinction in AF etiology is that of valvular atrial fibrillation (VAF) and non-valvular atrial fibrillation (NVAF). VAF is defined as ‘AF in the association with rheumatic mitral stenosis, a mechanical or bioprosthetic heart valve or mitral valve repair’ [7]. In contrast to NVAF, there are relatively few published studies on VAF, particularly in the recent literature. This disparity may in part be explained by the comparatively higher prevalence of NVAF in the Northern Hemisphere. In a cohort of 814 Italian patients with first-ever stroke and underlying AF, only 7% had VAF [4]. Within a group of 46,650 patients with AF seen at Mayo Clinic between 2004 and 2010, 2099 (4.5%) were diagnosed with VAF [8].

Rheumatic heart disease (RHD) as a component of VAF is seen mainly in developing countries [9]. Even though the incidence of RHD has fallen substantially in industrialized countries, valvular disease, mostly of degenerative etiology, is still present [10,11]. Within a group of 16,501 adult residents of Olmsted County, who had echocardiographic studies, valvular heart disease was present in 615 (5.2%, 95% CI: 4.8–5.6) participants [11].

Risk of thromboembolism in VAF & NVAF

Evidence from the Framingham study indicates that patients with RHD and AF have an 18-fold higher risk of stroke than age and blood pressure matched controls, while NVAF is associated with a five- to sixfold increase in stroke risk [1].

Observational studies from the 1950s and 1960s reported an annualized rate of thromboembolism in AF patients with RHD of 4–5% [12,13]. However, these results included a large number of patients followed for many years during a non-vulnerability stage of RHD (less than moderate mitral disease) and before development of AF, thus providing inaccurate stroke risk assessment. Once RHD patients developed AF, the observed rate of thromboembolism was astonishingly high. In a group of 37 VAF patients who suffered thromboembolism after new onset AF, 13 had an event within the first month of AF and 23 (62%) had an event within the first 12 months [13]. Not surprisingly, therefore, the authors postulated that the onset of AF in the presence of RHD should be regarded as a medical emergency.

In developed countries, mechanical or biological valve prosthesis contributes substantially to VAF populations. Although the risk of thromboembolism in patients with a bioprosthetic valve and sinus rhythm is low [14], the incidence of thromboembolism in AF patients with bioprosthetic valves was reported to be as high as 16% at 31–36 months [15]. Similarly, the incidence of thromboembolism is most likely higher when mechanical heart valve is accompanied by AF. The presence of mechanical valve(s) of all types or locations without AF is associated with an 8.6% overall risk of thromboembolism, including a 1.8% risk of valve prosthesis thrombosis and a 4.0% risk of major embolism [16]. Mitral location is associated with a 2.4-fold higher rate of thromboembolism compared to aortic location [16]. However, for VAF patients with mechanical heart prosthesis and arrhythmia, the natural history of thromboembolic complications is shrouded by the near universal use of anticoagulation and antiplatelet therapy.

The risk of thromboembolism in NVAF has been studied extensively and varies widely depending on the presence of several factors. This research has led to the development of multiple risk scoring systems to guide anticoagulation in NVAF patients. Among these, CHADS₂ (the acronym originating from the composition of risk factors: Congestive heart failure, Hypertension, Age >75 years, Diabetes mellitus and prior Stroke or TIA) revealed an adjusted stroke rate of 1.9% per year with a score of 0 and stroke rate of 18.2% with a maximum score of 6 [17]. By incorporating history of vascular disease, age 65–74 years and female gender in addition to the traditional CHADS₂ risk factors, the CHA₂DS₂-VASc system was able to better identify NVAF patients with an annual stroke risk of approximately zero. However, because of the very low number of patients with a high score (>4) in the original cohort of derivation, it is not able to reliably demonstrate the maximal risk of stroke [18]. These two scoring systems as well as all other stroke risk assessment tools have not been validated in VAF patients. Despite extensive data on several factors that are associated with the risk of stroke in NVAF, the mechanism of this process remains unclear. Echocardiographic features representing a ‘prothrombotic milieu’ in AF include enlarged left atrium, left atrial appendage with decreased emptying velocity and the presence of spontaneous echocardiographic contrast. These features point to blood stagnation as an important mechanism of thrombosis. Echocardiographic contrast, however, not only represents blood stasis but also reflects the ‘density’ of blood determined by ‘inflammatory components.’ These inflammatory components include fibrinogen concentration and von Willebrand factor level, which are both known blood markers of thromboembolism [19]. In the case of VAF, an additional substrate of either native atrial pathology (particularly rheumatic stenosis of mitral valve) or mechanical prosthesis is present and is most likely responsible for the observed threefold higher risk of stroke compared to the average risk associated with NVAF only [1]. There are currently four US FDA-approved oral anticoagulants for NVAF patients (warfarin, dabigatran, rivaroxaban and apixaban) with very well-documented efficacy and safety [7]. However, for VAF, only warfarin is recommended and there is only indirect evidence of its effectiveness (no specific studies have been conducted with VAF patients) [12,13,20]. The efficacy of anticoagulants in reducing the risk of thromboembolism in NVAF and VAF patients clearly indicates that the mechanism of thrombosis in both groups is at least partially dependent on the coagulation cascade. The fact that dabigatran, a direct thrombin inhibitor, was found to be ineffective in preventing thromboembolism in patients with mechanical valves [7] but effective in NVAF patients seems to indicate that the mechanisms of thrombosis in VAF may have some specific features not covered by the agent with this single target potential.

Distribution of thromboemboli in VAF & NVAF

Systemic review of literature revealed that in patients with VAF, more than half the thrombi (56%) were located in the left atrial cavity, while in the NVAF group only 11% were located outside of the left atrium appendage [21]. In our study of 414 consecutive NVAF patients who had transesophageal echocardiography, all 20 documented thrombi were found to be in the left atrial appendage [19].

There is an important difference in embolic destination in NVAF compared to VAF, specifically with regard to CNS and non-CNS circulation. Clinical series have shown that in NVAF patients, 94–100% of thromboemboli affect CNS circulation and 0–6% affect non-CNS arteries [22,23]. In VAF patients, however, there is a much higher proportion of emboli occurring in non-CNS circulation. Wood et al. [24] demonstrated that in RHD patients with AF, approximately 60% of thromboemboli occurred in cerebral territories, while 40% occurred in other locations. Szekely showed a very similar distribution of embolic events, with 63% being cerebral and 38% non-cerebral [13]. A relatively more recent study of VAF patients demonstrated 56% cerebral and 44% extracerebral location of embolic complications [20].

Because of a paucity of data, particularly in VAF patients, it is difficult to compare the distribution of non-cerebral emboli in NVAF compared to VAF patients. However, current evidence seems to suggest that it is likely similar. In a group of 72 NVAF patients who were treated at Mayo Clinic between 1995 and 2005 for non-cerebral embolism, 85% of embolic events occurred in the arteries of lower and upper extremities (2:1 ratio) and 15% occurred in the arteries of the abdominal organs, most often in the mesenteric arteries (45% of abdominal emboli) [25]. The distribution of non-CNS emboli in this series was consistent with prior reports for this group of patients [5,6]. In a study of RHD patients with VAF, 26 non-cerebral thromboemboli occurred, with 22 (85%) located in circulation in the extremities and three in the visceral arteries [13]. Another study of VAF patients revealed that out of 11 non-cerebral emboli, nine lodged in the extremities (82%), while one occurred in the renal circulation and another one in the mesenteric circulation [24]. The summary of thromboembolism distribution in VAF and NVAF is provided in Table 1.

Table 1. Distribution of thrombombolism in valvular atrial fibrillation vs non-valvular atrial fibrillation.

	Valvular atrial fibrillation	Non-valvular atrial fibrillation
Annual rate of thromboembolic complications	5–62%	0–18%
Distribution of intracardiac thrombi (left atrium vs left atrial appendage)	56 vs 44%	89 vs 11%
Distribution of thromboembolism (cerebral vs non-cerebral circulation)	60 vs 40%	94 vs 6%
Distribution of non-cerebral thromboemboli (extremities vs organs)	85 vs 15%	85 vs 15%

Several clinical features distinguish NVAF patients with non-cerebral emboli from NVAF patients with stroke. Patients with non-cerebral emboli are more likely to be >75 years of age, with a higher prevalence of hypertension, and have heart failure. Additionally, severe left ventricular dysfunction, spontaneous echo contrast and left atrial thrombus were two- to threefold more common in NVAF patients with non-cerebral emboli [25]. Unfortunately, we were unable to find any similar comparative studies on VAF.

The difference in the destination of emboli in VAF and NVAF patients is intriguing but unclear. Left atrial body thrombi, which are more often found in VAF patients, are larger than those that develop in the relatively small space of the multi-lobulated left atrial appendage [12–14,25]. It could therefore be speculated that thrombotic material originating in VAF patients might be too massive to pass through the orifices of the carotid and vertebral arteries and are thus more prone to lodge in the non-cerebral circulation.

Financial & competing interests disclosure

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.