Atrial Fibrillation: Diabetes as a Risk Factor
Diabetes as a Risk Factor for Atrial Fibrillation
A clear relationship of type I DM (T1DM) with AF was not established until recently, with initial suggestions that long-term hyperglycaemia in T1DM does not itself promote AF.[3] Although studies on T1DM concerning AF risk are rare, recent large analyses eventually confirmed that T1DM was also independently associated with a higher incidence of AF.[3–5] When comparing patients with T1DM from the Swedish National Diabetes Register to controls, the risk of AF with T1DM was increased with an adjusted HR of 1.13 (95% CI 1.01–1.25, P = 0.029) in men and 1.50 (95% CI 1.30–1.72, P < 0.0001) in women (P = 0.002 for interaction).[3]
Considering T2DM, AF is not uncommon, and for example, 7.6% of patients had AF at baseline in the ADVANCE study (Action in Diabetes and Vascular Disease: Preterax and Diamicron-MR Controlled Evaluation).[6] The prevalence of AF is at least two-fold higher in patients with DM compared with people without DM,[7] and AF incidence may be even higher in patients with microvascular complications (retinopathy, renal disease).[8] In Huxley's meta-analysis, which included seven prospective studies and four case–control studies on 1.7 million subjects, 109 000 of whom had incident AF, it was found that DM was associated with about 40% higher risk of AF [relative risk (RR) 1.39; 95% confidence interval (CI) 1.10–1.75]. However, after adjusting for possible confounding factors, the effect appeared to be more limited with an increased risk of only 24%.[9] The authors assumed that the proportion of AF that was attributable to DM in the population was only 2.5%. In Ostgren's study, patients with both T2DM and high blood pressure had a three-fold higher risk of associated AF compared to non-diabetic non-hypertensive patients, but this association was no longer significant after adjusting for the presence of insulin resistance, which therefore appeared to be a major determinant for the development of AF.[10] Insulin resistance is also a mechanism by which hypertension and obesity might be associated with an increased risk of AF.[11] Diabetes mellitus and glucose intolerance are also associated with an increased risk of left ventricular hypertrophy (LVH) which could be one of the factors promoting AF.[7] Long-term inflammation is also a suspected link between DM and AF: inflammation may be found in the two conditions, it can partly predict the development of DM and of AF, and anti-inflammatory treatments have shown to be efficient on a few indirect parameters for some patients with DM and/or with AF (Figure 1).[12]
Figure 1.
Pathophysiology linking diabetes, AF, and risk of stroke. AF, atrial fibrillation; DOACs, direct oral anticoagulants; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; LA, left atrium; LVH, left ventricular hypertrophy; VKA, vitamin K antagonist.
In a subsequent analysis carried out in 35 000 healthcare professionals, and followed for 16 years, the 3% of women with T2DM had a doubled risk of developing AF during follow-up but again, the attributable risk was quite markedly reduced after adjusting for possible confounding factors in the patient characteristics: the risk was only 15% after multivariable analysis.[13]
This would overall suggest that patients with T2DM have a higher risk of prevalent AF (around 15%) (Figure 2), and then of incident AF (around 0.8%/year), which however seems partly related to other associated risk factors frequently encountered in these patients.[17] It also appears that DM would be an independent risk factor for AF when one specifically considers patients aged <75.[18] In contrast, when searching for AF after a stroke, since DM may be more closely associated with non-cardioembolic ischaemic stroke, the search may not find as much AF.[19,20]
Figure 2.
Interplay between atrial fibrillation (AF), diabetes (DM), and chronic kidney disease (CKD) showing the proportion of AF, CKD, and diabetes seen as co-morbidities with each other. For example, about 35% of AF patients have CKD and 30% have diabetes. Percentages vary depending on the definitions of CKD, diabetes, and atrial fibrillation and on specific populations but the figures are reasonable generalizations.14–16
Both T1DM and T2DM have been associated with atrial electrical and structural remodelling that may underly increased vulnerability for AF.[21] Patients with T1DM have more fractionated atrial electrograms and diastolic dysfunction. Animal models of T1DM demonstrate reduced Na+ current in the atrial cardiomyocyte with a subsequent reduced upstroke of the action potential and lower conduction velocity.[22] Also, increased structural remodelling through interstitial fibrosis formation has been demonstrated in T1DM models.[23] In patients with T2DM Ca2+ handling was impaired, which could add to AF induction.[24,25] Furthermore, atrial fibroblasts from T2DM patients had a more profibrotic phenotype compared to fibroblasts of patients without DM, evident from increased expression of collagen type 1, which adds to structural remodelling of the atrium.[26] The rate of new-onset AF is not obviously affected by intensive glycaemic control,[27] but may be affected by DM therapy (see Diabetes subtypes, diabetes severity, and the risk of atrial fibrillation section).
Diabetes as a Risk Factor for Mortality, Complications, and More Symptoms in Atrial Fibrillation
Diabetes mellitus and AF frequently coexist (Figure 2),[14–16] and when this occurs there is a substantially higher risk of all-cause death, CV death, stroke, CKD disease, and heart failure (HF).[28–30] Risk of death may be 25–60% higher in AF patients with DM (vs. AF and no DM).[29,30] This was confirmed with the observations from the ORBIT-AF (Outcomes Registry for Better Informed Treatment of AF) registry showing the incremental worsening of AF prognosis over all-cause death, CV death, and several other CV endpoints when DM was present.[31] Risk factors commonly associated with DM (and not fully dissociable, e.g. hypertension and obesity) are also likely to worsen prognosis. Diabetes mellitus and declined renal function are also closely interlinked with increased CV risk. Importantly, the combination of DM and renal impairment is associated with a higher risk of CV events and mortality than either co-morbidity alone[32] underlining the importance of CV prevention in these vulnerable patients when AF is also present (Figure 1). Atrial fibrillation is associated with substantially increased risks of death and CV events in patients with T2DM, meaning that AF further increases the already elevated risk of CV disease in these patients. Atrial fibrillation identifies individuals with DM who are likely to obtain greater absolute benefits from blood pressure-lowering treatment.[6] Atrial fibrillation in diabetic patients should be regarded as a marker of adverse outcomes suggesting prompt and aggressive management of all risk factors.
The detection of AF in patients with DM has major clinical consequences because the risk of stroke is markedly higher in these patients. In the absence of other co-morbidities, the annual risk of stroke can be estimated at 2.2% per year in isolated DM.[33] It is also higher in diabetic patients due to the common association with other risk factors for stroke, among which the existence of arterial hypertension, an age over 65 or 75, associated vascular disease, or HF. All of them are identified in the items of the Congestive HF, Hypertension, Age [≥75 years; 2 points], Diabetes mellitus, Stroke/transient ischaemic attack [2 points], Vascular disease, Age [65–74 years], Sex category [female] (CHA2DS2-VASc) score. Although there may be some variations in the literature, several studies found that the presence of DM was an independent predictor for stroke in patients with AF.[34] However, DM may not be a significant risk factor for stroke in the elderly.[35] More specifically, the Stroke in AF Working Group attributed a RR: 1.7 (95% CI 1.4–2.0) for stroke in diabetic patients with AF and an absolute stroke risk of 2–3.5%/year in the same population for non-anticoagulated patients.[36] Overall, DM is probably not the most potent independent risk factor for stroke in AF compared to the other items in the CHA2DS2-VASc score, the widely used risk stratifier adopted to guide anticoagulation therapy, but it is included in this risk stratification tool, giving a point as well as for most other items.[37–41]
Beyond the higher risk of mortality and CV events, DM in AF patients is likely to modify symptoms, particularly breathlessness, which is linked to the higher prevalence of obesity and higher risk of HF. However, the adverse effect of DM in AF patients might be related to the possible masking of AF-related palpitations, which can cause a delay in initiation of treatment and therefore poorer outcomes in diabetic patients.[1,42] This characterization of symptoms has been well described in the ORBIT-AF registry with the AF symptom checklist. Compared to AF patients with no DM, those with DM less often reported palpitations (29.2% vs. 33.6%) or syncope (3.3% vs. 4.9%), but they reported significantly more frequent dyspnoea on exertion (29.7% vs. 26.6%) or at rest (11.1% vs. 9.8%), exercise intolerance (11.9% vs. 9.2%), and fatigue (27.9% vs. 25.6%). Although these differences were statistically significant, they may not be obvious from a clinical point of view. Overall, AF patients with DM had worse functional status as reflected by higher European Heart Rhythm Association (EHRA) scores (although % of asymptomatic patients seemed similar) and also had a slightly lower overall quality-of-life scores than those without DM.[31] Diabetes mellitus has also been identified as an independent predictor of AF recurrence after AF ablation.[43] While AF ablation usually reduces palpitations and improves New York Heart Association (NYHA) class in patients without DM, patients with DM may report less improvement of NYHA class despite a reduction of palpitations.[44]
Screening for Atrial Fibrillation in Diabetes
Due to the high prevalence of AF in DM and an estimated development rate of nearly 1%/year, more active screening of diabetic patients for AF may influence therapeutic management. For symptomatic patients with palpitations (or possibly fatigue and dyspnoea) related to persistent or permanent AF, the diagnosis is likely to be made during an outpatient visit or in the hospital if the patient consults emergently. However, AF may also be asymptomatic in 30–50% of cases, as mentioned above. It can be found in different ways: on a conventional 12-lead electrocardiogram at a routine consultation or in hospital, on a 24-h (sometimes 72-h) ambulatory ECG recording, or in a patient with a pacemaker or a defibrillator by analysing the diagnostic functions (used as a simplified long-term ambulatory ECG) during a check in-office or during remote monitoring. Subcutaneously implantable loop recorders with long-term information over 24–36 months have the technical capacity to diagnose symptomatic or asymptomatic AF, but their usefulness is only recognized in most countries for the evaluation of patients with unexplained syncope or for those with ischaemic stroke and no known AF. However, AF detection with an implantable loop recorder in patients with DM may have similar clinical consequences with regard to the need for oral anticoagulation. Nowadays, AF may be detected by consumer devices and wearables that detect an irregular pulse (e.g. photoplethysmography), with the need in these cases to confirm the diagnosis of AF by means of an ECG recording, or by direct recording of an ECG tracing (e.g. using the Apple watch).[41]
Potential advantages and disadvantages of detecting previously undiagnosed AF through screening have been reported in the 2020 European Society of Cardiology (ESC) guidelines for the diagnosis and management of AF.[41] There are currently no specific recommendations on strategies and tools for routinely screening AF in diabetic patients (without known AF) who are asymptomatic, but data are emerging that is consistent on two points:
when DM is associated with other risk factors, in other words when the CHA2DS2-VASc score is high, the risk of developing AF increases during follow-up, is reaching nearly 40% within 2 years in some high-risk subgroups.[45]
simultaneously, in these patients with AF, the annual thrombo-embolic risk increases exponentially, which can reach 8–10% per year.[46,47]
It would seem quite logical on a population scale to systematically seek AF in patients with a CHA2DS2-VASc score of 2 or higher in order to reduce the risk of stroke which is the major complication to be feared in this context. However, this approach has not yet been validated in a randomized study: such a strategy would require extensive screening, and then the treatment of a large group of patients where AF would be detected and then treated with oral anticoagulant for obtaining a possible statistically lower risk of stroke. Such a design would be challenging. However, the ESC guidelines for the management of AF do recommend opportunistic screening for any patient over 65 years and systematic screening for patients at high risk of developing AF.[41] Thus, both opportunistic and systemic screening for AF in diabetic patients is clearly appropriate.
The association between AF and DM appears to be well established at the population level.[5] The causal link between DM and AF needs to be further clarified. Diabetes mellitus associated with AF results in a worse prognosis and several reports or statements from some health authorities indicate that efforts must be made for the prevention of AF, but there is not yet a strong level of evidence that this strategy will result in a significant benefit in terms of morbidity (the most important of which is stroke) or mortality.
Europace. 2021;23(12):1873-1891. © 2021 Oxford University Press
Copyright 2007 European Heart Rhythm Association of the European Society of Cardiology (ESC). Published by Oxford University Press. All rights reserved.
