Epidemiology of and risk factors for atrial fibrillation progression

Background: Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, and the number of AF patients is estimated to double until 2060. Current thinking indicates that AF frequently progresses from short episodes to longer and more frequent attacks. Over years, a number of patients develop sustained forms that are less amenable to treatment and are thought to be associated with a worse outcome. However, the concept of AF progression is poorly understood and data on this important issue are mainly based on small studies in selected patients or studies with short follow-up. Based on those gaps in knowledge, the general aim of this PhD thesis was to systematically assess the incidence and predictors of AF progression. The specific aims were 1) to summarize the current evidence on the incidence of AF progression in a systematic review and meta-analysis of the previous literature, 2) to assess the incidence and predictors of AF progression in our own cohorts, taking into account the potential effect of rhythm control interventions (RCIs), and 3) to compare the incidence of adverse outcome events in patients with paroxysmal, persistent and permanent AF.
Methods: For the meta-analysis, we searched PubMed, EMBASE and the Cochrane Library. Random effect models were used to calculate cumulative incidence rates. Predictors related to between study variability were assessed using meta-regression analyses. The second and third manuscripts of the PhD thesis were based on two ongoing prospective cohort studies (BEAT-AF and Swiss-AF) in whom 3,968 patients with documented AF have been enrolled. At baseline and during yearly follow-up we assessed AF type, RCIs and adverse outcome events. We defined clinical AF progression as progression from paroxysmal to non-paroxysmal AF or as progression from persistent to permanent AF at the latest follow-up as compared to the baseline AF type. RCIs included electrical cardioversions (ECVs), pulmonary vein isolation (PVI) or initiation of amiodarone treatment. Multivariable adjusted Cox regression analyses were performed to assess potential predictors for clinical AF progression and RCIs. Prespecified adverse outcome events for the third manuscript were stroke/systemic embolism, incident hospitalization for heart failure (HF), bleeding, all-cause mortality and a combined outcome of ischemic stroke, myocardial infarction or cardiovascular death (MACE). Multivariable adjusted time-dependent Cox regression analyses were performed to compare hazard ratios (HRs) across patients with paroxysmal, persistent and permanent AF.
Results: For the meta-analysis, we identified 47 studies with 27,266 patients who were followed for 105,912 patient-years. The pooled incidence of AF progression was 8.1 per 100 patient-years (95% confidence interval [CI], 7.1; 9.1). Higher age and prevalence of hypertension were associated with a higher AF progression rate, follow-up duration and prevalence of paroxysmal AF with a lower AF progression rate. In the second analysis we found that 458 (15.9%) of 2,882 patients had clinical AF progression in our own cohorts (incidence 5.2 per 100 patient-years), and 613 (27.1%) had at least one RCI (incidence 10.9 per 100 patient-years) during a median (IQR) follow-up of 3 (2; 5) years. Increasing BMI (HR 1.03, [95% CI 1.00; 1.05], p=0.02) and higher blood pressure (HR 1.03 [95% CI 1.00; 1.05], p=0.05) were associated with a higher AF progression rate, whereas an inverse association was found for regular physical activity (HR 0.80 [95% CI 0.66; 0.98], p=0.03) and previous PVI (HR 0.68 [95% CI 0.52; 0.89], p=0.005). Important factors associated with a RCI were regular physical activity (HR 1.43 [95%CI 1.21; 1.69], p<0.001), presence of AF related symptoms (HR 1.83 [95%CI 1.46; 2.29], p<0.001) and younger age (HR per 5 years 0.89 [95%CI 0.85; 0.92], p<0.001). In the third project we found an incidence of stroke/systemic embolism of 1.0 per 100 patient-years. The incidence was 0.8, 1.0 and 1.5 per 100 patient-years for patients with paroxysmal, persistent and permanent AF, respectively. Compared with patients with paroxysmal AF, the HRs (95% CI) for stroke/SE were 1.26 [95%CI 0.79; 2.01], p=0.34) and 1.35 ([95% CI 0.88; 2.07], p=0.17) for persistent and permanent AF in age and sex adjusted time-updated models. After multivariable adjustment, the HRs were 1.16 ([95% CI 0.70; 1.91], p=0.57) and 1.29 ([95% CI 0.84; 2.00], p=0.25), respectively. Patients with permanent AF had a higher risk of heart failure (HR 1.68 [95% CI 1.31; 2.15[, p<0.001), MACE (HR 1.70 [95% CI 1.33; 2.17], p<0.001), all-cause mortality (HR 1.72 [95% CI 1.34; 2.21], p<0.001), clinically relevant non-major bleeding (HR 1.31 [95% CI 1.04; 1.66], p=0.02) and any bleeding (HR 1.23 [95% CI 1.00; 1.51], p=0.046) in age/sex adjusted models. In multivariable time-updated models, permanent AF remained significantly associated with MACE (HR 1.37 [95% CI 1.06; 1.78], p=0.02) and all-cause mortality (HR 1.38 [95% CI 1.07; 1.80], p=0.01), but not with heart failure (HR 1.24 [95% CI 0.96; 1.60], p=0.10), clinically relevant non-major bleeding (HR 1.00 [95% CI 0.78; 1.28], p=1.00) and any bleeding (HR 0.97 [95% CI 0.78; 1.20], p=0.76).
Discussion: Current evidence suggests that the incidence of AF progression is relatively low, and these numbers were confirmed in our own data. Our findings suggest that a healthy lifestyle with a lower BMI and regular physical activity may reduce the rate of AF progression. RCIs were commonly used, mainly predicted by symptoms, physical activity and lower age. There was no significant association between AF type and risk of stroke/systemic embolism, but patients with permanent AF had an increased risk of all-cause mortality and MACE. Co-morbidities and risk-factors seem to have an important influence on these relationships. The residual effect of AF type seems to be smaller than estimated previously.