Hypertrophic cardiomyopathy and atrial fibrillation: the Cardiomyopathy/Myocarditis Registry of the EURObservational Research Programme of the European Society of Cardiology

Statistical analysis

Analysis of baseline demographic and clinical variables and predefined clinical endpoints was stratified by the presence of any AF. Univariate analysis was applied to both continuous and categorical variables. Continuous variables were reported as mean±SD. Among-group comparisons were made using the non-parametric test (Kruskal-Wallis test). Categorical variables were reported as percentages. Among-group comparisons were made using the χ² test or a Fisher’s exact test if any expected cell count was less than five. Stepwise multivariable logistic regression analyses were performed to establish the relationship between the patient characteristics and the presence of AF, including into the model, all the candidate variables (variables with p<0.10 in univariate). The candidate variables for the prediction of AF included age at enrolment, age at first evaluation in the centre, body mass index (BMI), NYHA class, palpitations, history of sustained ventricular tachycardia, history of atrioventricular block, history of bundle branch block, history of stroke or TIA, arterial hypertension, diabetes mellitus, hyperlipidaemia, renal impairment, chronic obstructive pulmonary disease, anaemia and level of physical activity. A significant level of 0.05 was required for a variable to stay in the model. No interaction was tested. To verify that the models were optimal, Hosmer and Lemeshow Goodness-of-Fit test and per cent concordant were calculated. A two-sided p value of less than 0.05 was considered as statistically significant. All analyses were performed using SAS statistical software V.9.4 (SAS Institute, Cary, North Carolina).

Baseline characteristics

1739 adult patients with HCM (40.9% women; median age: 55.5) were recruited. Follow-up data at 1 year were obtained in 1420 patients (87.7%) (table 1).

Rate of AF in HCM population

AF was present at baseline in 478 (27.5%) individuals. Paroxysmal AF was the most common form of AF in patients with HCM (54.7%). Newly diagnosed AF was registered during follow-up in 48 (2.8%) individuals with HCM. The proportion of patients with AF at baseline and at follow-up in different cardiomyopathies is presented in table 1.

Demographic and clinical characteristics of AF and non-AF HCM patients at baseline

Baseline characteristics of HCM patients with and without AF was outlined in table 2. Age at enrolment (59.6±13.8 vs 50.8±16.1 years, p<0.001) and age at the first evaluation in the centre (53.9±15.5 vs 46.9±17.0 years, p<0.001) were greater in HCM subjects with AF. Patients with AF had larger BMI (27.7±5.1 vs 26.6±4.6 kg/m², p<0.001). New York Heart Association (NYHA) class was more advanced in AF (NYHA I/II/III/IV: 19.3/55.8/22.5/2.4%) than in non-AF subjects (NYHA I/II/III/IV: 39.2/47.0/12.9/0.8%, p<0.001). History of arrhythmias: sustained ventricular tachycardia (10.8 vs 6.4%, p=0.001), atrioventricular block (12.3 vs 8.4%, p=0.043) and bundle branch block (19.3 vs 12.3%, p=0.003) were more frequent in AF than in non-AF HCM population. History of stroke/TIA was positive more frequently in AF population (11.5% vs 3.3%, p<0.001).

The following comorbidities were more prevalent in patients with AF: arterial hypertension (43.4 vs 34.6%, p<0.001), diabetes mellitus type I or II (14.6 vs 8.2%, p<0.001), hyperlipidaemia (42.0 vs 34.1%, p=0.002), renal impairment (15.4 vs 6.4%, p<0.001), chronic obstructive pulmonary disease (6.7 vs 2.6%, p<0.001) and anaemia (7.5 vs 3.4%, p<0.001). Patients with AF reported less physical activity than those without AF (40.0 vs 53.3%, p<0.001).

Echocardiographic characteristics of AF and non-AF HCM patients at baseline

Patients with AF had lower LVEF (59.5±12.3 vs 63.5±10.7%, p<0.001), left atrium (LA) dilatation (48.9±9.1 vs 42.4±7.7 mm, p<0.001), increased pulmonary artery systolic pressure (37.8±13.7 vs 29.6±12.6 mm Hg, p<0.001), other pattern of LV hypertrophy (AF vs non-AF: septal 70.9 vs 74.6%, concentric 15.6 vs 11.4%, apical 8.4 vs 8.2%, other 5.1 vs 5.8%; p=0.031), more advanced LV diastolic dysfunction (AF vs non-AF: normal 14.8 vs 27.5%, grade I/impaired relaxation 38.3 vs 43.6%, grade II/pseudonormalisation 29.8 vs 24.0%, grade III/restriction 17.2 vs 5.0%; p<0.001) and more frequent LV outflow tract resting gradient (38.8 vs 30.7%, p<0.001).

Predictors of AF in HCM population

Multivariate logistic regression analysis revealed that the independent predictors of AF in the HCM population were age at enrolment (OR 1.068, p<0.001), symptom—palpitation (OR 2.151, p<0.001), left ventricular ejection fraction (LVEF) (OR 0.978, p<0.001) and LA diameter (OR 1.094, p<0.001) (table 3).

OAC and antiplatelet therapy in HCM population with AF

OAC was administered at baseline in 69.5% of HCM patients with AF: 48.5% patients were treated with vitamin K antagonists (VKA) and 21.0% with DOAC (figure 1). Antiplatelet therapy was administered in 20.3% of AF patients, of whom 90.7% received aspirin. Detailed data on anticoagulant and antiplatelet therapy in HCM patients with AF are presented in table 4.

Clinical endpoints in AF and non-AF patients with HCM

At 1-year follow-up, the incidence of stroke/TIA was higher in patients with AF compared with non-AF patients (2.6 vs 0.9%, p=0.009). There was a trend towards increased death from any cause in patients with AF (3.4 vs 1.7%, p=0.053). Comparison of other endpoints—death from ischaemic stroke, death from haemorrhagic stroke and death from systemic haemorrhage did not differ between the AF and non-AF HCM populations (table 5).

OAC and antiplatelet therapy in AF population

Almost a third of patients with HCM and AF did not receive anticoagulants during the observation period. This is consistent with data from the EORP-AF pilot registry,21 in which OAC was used in 80.1% of patients with AF. Other contemporary registries presenting data on OAC in the general AF population report rates of anticoagulation that vary from 46% to 97%.22 23 Anticoagulation is associated with a lower incidence of thromboembolic events24 and the CHA₂DS₂-VASc score is used as a method of stratifying patients with AF for therapy.8 However, retrospective evidence in HCM suggests that CHA₂DS₂-VASc performs suboptimally with respect to stroke prediction.7 25 Currently, HCM is incorporated into CHA₂DS₂-VASc score and constitutes a surrogate of heart failure in the scoring system.8 Given that AF increases the risk of thromboembolic events in patients with HCM to a greater extent than in the general population, the first occurrence of AF should be an indication for lifelong OAC and HCM.26 27 While there is more experience with VKA, DOAC can currently be used in these patients.26 Although antiplatelet agents are not indicated for prevention of thromboembolic events,8 17.5% of our AF patients received this therapy.

We may suspect only that in some patients stroke/TIA in case history, or coexistence of diabetes mellitus were the indications for the therapy. Probably in most cases, the antiplatelet therapy was used improperly instead of OAC. The registry highlights the need for further education on OAC in prevention of thromboembolic events in HCM-AF population.

Clinical endpoints and risk factors for stroke/TIA

The EORP Registry confirmed impaired prognosis for the population with HCM and concurrent AF. The present results suggest more than three times higher annual incidence of stroke/TIA in AF than in non-AF patients with HCM (2.65 vs 0.85%), and the annual incidence of stroke/TIA is comparable to the adjusted stroke rate (3.2% /year) in non-valvular AF population with a CHA₂DS₂-VASc score of 3 (population not receiving OAC).8 Our results are concordant with the meta-analysis by Guttmann et al,7 which demonstrated the annual incidence of thromboembolic events of 3.75% in patients with HCM and AF. According to a Korean database,25 the risk of stroke in HCM population with AF and without any CHA₂DS₂-VASc risk factors was similar to that of AF general population with CHA₂DS₂-VASc score of 3. In the study by Maron and coworkers on 900 patients with HCM, overall annual incidence of systemic thromboembolic events in HCM population was 0.8%/year, 1.9%/year in patients >60 years and 2.5%/year in patients with AF.5 This data confirm the requirement for prompt introduction of chronic OAC in patients with HCM and AF irrespective of the risk estimated using CHA₂DS₂-VASc score. The clinical vigilance in terms of screening for silent AF is vital and should be performed according to current ESC recommendations.28

Comparison of other clinical endpoints, that is, death from ischaemic stroke, death from haemorrhagic stroke, death from systemic haemorrhage that may be related to AF or anticoagulation did not reveal any differences, which may be linked to a limited number of events.

It is well-documented that AF coexists and interacts with other cardiovascular risk factors both in general9 and in cardiomyopathy population.22 Thus, AF itself and AF as an element of complex interactions may worsen the prognosis for AF population. We found that classic cardiovascular risk factors, that is, age, diabetes mellitus and renal impairment were associated with the incidence of stroke/TIA in the HCM population. The AF at baseline, previous incidence of stroke and anaemia were independent risk factors for the stroke/TIA on follow-up. It suggests that both monitoring for AF diagnosis and complex ‘up-stream therapy’ including modification of life-style risk factors and treatment of comorbidities are necessary to prevent cerebral events in patients with cardiomyopathies.

Limitations

The limitations for the analysis of the EORP Cardiomyopathy/Myocarditis Registry data have been described previously.9 Considering the selection bias related with tertiary reference centres involved in the registry, the use of anticoagulation in patients with HCM and AF might have been overestimated. The actual use of OAC in this population may thus be lower with further deleterious impact on the risk of thromboembolic events. The protocol did not impose on centres the strict requirement for screening for AF, thus, the registry represents real-world data on the prevalence of AF in this population.

The data in the EORP Registry have been collected between 2012 and 2016,9 and the anticoagulation was administered according to the current recommendations. However, we should be aware that specific recommendations have been set for patients with HCM and AF.2 26 Nowadays, we administer anticoagulation for all HCM subjects with AF since HCM is considered as surrogate of HF in CHA2DS2-VASc score.

Data on appropriate anticoagulation are important for the assessment of the risk of thromboembolic events. The registry did not include information about OAC doses or INR in VKA patients. We also have no data on genetic tests or familiar HCM appearance that would be of great importance regarding a potent separate analysis of predisposing factors for AF in a sarcomeric form of cardiomyopathy.

The relatively short follow-up constitutes another limitation of the study. Since there are not many subjects who experienced a stroke/TIA during follow-up, the analysis of risk factors for stroke/TIA was limited.