Abstract
Atrial fibrillation is associated with hyperthyroidism. Patients with primary aldosteronism have an increased prevalence of atrial fibrillation. However, the prevalence of primary aldosteronism in the atrial fibrillation population is unknown.
This nationwide case–control study aimed to compare the prevalence of primary aldosteronism and thyroid disorders in patients with atrial fibrillation with that of age- and sex-matched controls.
We identified all atrial fibrillation cases in Sweden between 1987 and 2013 (n = 713,569) by using the Swedish National Patient Register. A control cohort without atrial fibrillation was randomly selected from the Swedish Total Population Register with a case to control ratio of 1:2. This control cohort was matched for age, sex and place of birth (n = 1,393,953).
The prevalence of primary aldosteronism in December 2013 was 0.056% in the atrial fibrillation cohort and 0.024% in controls. At the same time, the prevalence of hypothyroidism was 5.9% in the atrial fibrillation cohort and 3.7% in controls. The prevalence of hyperthyroidism was 2.3% in the atrial fibrillation cohort and 0.8% in controls.
This study shows, for the first time, a doubled prevalence of primary aldosteronism in a large cohort of patients with atrial fibrillation compared with the general population. There is also an increased prevalence of hypo- and hyper-thyroidism in patients with atrial fibrillation compared with the general population.
Introduction
Atrial fibrillation (AF) is a major public health issue with an increased risk of all-cause mortality, stroke and heart failure.1,2 AF affects almost 3% of the adult population and there are strong indications that the total AF burden in the population is increasing.3,4 Therefore, effective prevention and treatment of AF are important public health concerns. Current guidelines suggest that stratifying patients with AF by the underlying cause could improve management of AF.5
Structural heart disease and hypertension are common causes of AF. Other less frequent causes of AF are endocrine disorders, particularly hyperthyroidism, which is associated with an increased risk of AF. However, hypothyroidism is either protective or has no association with AF.6,7 Primary aldosteronism (PA), one of the most common causes of secondary hypertension, is generally not implicated as a cause of AF. However, previous studies have found a 5- to 12-fold higher risk of AF in patients with PA.8,9 The prevalence of PA among hypertensive patients is estimated to be between 4% and 10%, but the prevalence of PA in the AF population and the general population is unknown.10–12
Therefore, this study aimed to assess the prevalence of diagnosed PA in individuals with and without AF. A secondary aim of this study was to examine the prevalence of hypothyroidism and hyperthyroidism in patients with and without AF.
Methods
This case–control study was based on information from linking several national registers in Sweden. The case cohort included all patients with a registered diagnosis of AF in the Swedish National Patient Register (NPR) during 1987–2013.13 The NPR includes data on hospital discharge diagnoses since 1964. Since 1987, the NPR includes all inpatient care in Sweden, and from 2001, it also includes outpatient visits from both private and public caregivers. AF was defined as any first hospitalisation or outpatient visit with an International Classification of Disease (ICD) code of ICD-10 (I48), ICD-9 (427.31) and ICD-8 (427.92). The control cohort was based on randomly selected individuals from the Swedish Total Population Register without a diagnosis of AF from 1980 to 2014. We selected a case to control ratio of 1:2 and matched for age, sex and place of birth.
To assess the prevalence of PA, hypothyroidism and hyperthyroidism, we extracted diagnostic data for each individual case/control from the NPR starting 7 years prior to the time of diagnosis of AF until the end of follow-up on 31 December 2013. The prevalence was reported as the proportion of incident cases of PA, hypothyroidism and hyperthyroidism in patients who survived until 31 December 2013. PA was defined as the ICD-10 codes E26.0, E26.8 and E26.9, the ICD-9 code 255.B and the ICD-8 code 255.00. Hypothyroidism was defined as the ICD-10 codes E03.4, E03.5, E03.8 and E03.9, the ICD-9 codes 244.W and 244.X and the ICD-8 codes 244.00 and 244.09. Congenital hypothyroidism, iatrogenic hypothyroidism and hypothyroidism after surgery, radiation and iodine treatment, which were all defined by separate codes, were not included. Hyperthyroidism was defined as the ICD-10 code E05, the ICD-9 code 242 and the ICD-8 codes 242.00, 242.09, 242.10 and 242.20. Congenital hyperthyroidism, atoxic goitre and adenoma (separate codes) were not included. The prevalence of PA, hypothyroidism and hyperthyroidism was calculated for living individuals on 31 December each year between 1987 and 2013. Individuals who were not alive at the end of the follow-up were identified through the National Cause of Death Register.14 The total cardiovascular comorbidity burden of the two cohorts was characterised according to the CHA2DS2-VASc (Congestive heart failure [ point] Hypertension [ point] Age >75 years [points] Diabetes mellitus [1 point] previous Stroke/transient ischemic attack [2 points], Vascular disease [1 point], Age 65–74 years [1 point[, female Sex [1 point]) score.15 Components of the CHA2DS2-VASc score were defined by heart failure (ICD-10: I42, I43, I50; ICD-9: 425, 428; ICD-8: 425, 427.00, 427.10, 428.99, 782.40), hypertension (ICD-10: I10–I13, I15; ICD-9: 401–405; ICD-8: 400–404), age >65 or >75 years at inclusion, diabetes mellitus (ICD-10: E10–E11, E13–E14; ICD-9: 250; ICD-8: 250), previous ischaemic stroke, unspecified stroke, transient ischaemic attack or systemic emboli (ICD-10: I63–I64, G45, I74; ICD-9: 434–436, 444; ICD-8: 434–436, 444), female sex and vascular disease (prior myocardial infarction, peripheral arterial disease [ICD-10: I21, I252, I70–I73; ICD-9: 410, 412, 440–443; ICD-8: 410, 440–443]). The study protocol was approved by the Regional Ethical Review Board at the University of Gothenburg.
Statistical analysis
Descriptive statistics are shown as numbers with percentages for categorical variables and as means (±SD) for continuous variables. The prevalence rates are presented as percentages of patients who were alive on 31 December 2013. Logistic regression models were used to estimate odds ratios (ORs) for the prevalence of PA, hypothyroidism and hyperthyroidism. Logistic regression models were adjusted for age, sex, hypertension, ischaemic heart disease, heart failure, diabetes mellitus, cerebrovascular disease and cancer. ICD codes for the adjusted diagnoses are shown in Table 1. A two-tailed p-value of 0.05 was considered significant. Data management and analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).
International Classification of Disease (ICD) codes for adjusted covariates in logistic regression analyses.
| Diagnoses | ICD codes |
|---|---|
| Diabetes mellitus | ICD-10: E10–E11 and E13–E14 ICD-9: 250 ICD-8: 250 |
| Heart failure | ICD-10: I42, I43, I50 ICD-9: 425, 428 ICD-8: 425–428, 7824 |
| Ischaemic heart disease | ICD-10: I20–I25 ICD-9: 410–414 ICD-8: 410–414 |
| Cerebrovascular disease | ICD-10: I60–I69, G45 ICD-9: 430–437 ICD-8: 430–438 |
| Hypertension | ICD-10: I10–I13, I15 ICD-9: 401–405 ICD-8: 400–404 |
| Cancer | ICD-10: C00–C97 ICD-9: 140–209 ICD-8: 140–207 |
| Diagnoses | ICD codes |
|---|---|
| Diabetes mellitus | ICD-10: E10–E11 and E13–E14 ICD-9: 250 ICD-8: 250 |
| Heart failure | ICD-10: I42, I43, I50 ICD-9: 425, 428 ICD-8: 425–428, 7824 |
| Ischaemic heart disease | ICD-10: I20–I25 ICD-9: 410–414 ICD-8: 410–414 |
| Cerebrovascular disease | ICD-10: I60–I69, G45 ICD-9: 430–437 ICD-8: 430–438 |
| Hypertension | ICD-10: I10–I13, I15 ICD-9: 401–405 ICD-8: 400–404 |
| Cancer | ICD-10: C00–C97 ICD-9: 140–209 ICD-8: 140–207 |
ICD codes include primary and secondary diagnoses.
International Classification of Disease (ICD) codes for adjusted covariates in logistic regression analyses.
| Diagnoses | ICD codes |
|---|---|
| Diabetes mellitus | ICD-10: E10–E11 and E13–E14 ICD-9: 250 ICD-8: 250 |
| Heart failure | ICD-10: I42, I43, I50 ICD-9: 425, 428 ICD-8: 425–428, 7824 |
| Ischaemic heart disease | ICD-10: I20–I25 ICD-9: 410–414 ICD-8: 410–414 |
| Cerebrovascular disease | ICD-10: I60–I69, G45 ICD-9: 430–437 ICD-8: 430–438 |
| Hypertension | ICD-10: I10–I13, I15 ICD-9: 401–405 ICD-8: 400–404 |
| Cancer | ICD-10: C00–C97 ICD-9: 140–209 ICD-8: 140–207 |
| Diagnoses | ICD codes |
|---|---|
| Diabetes mellitus | ICD-10: E10–E11 and E13–E14 ICD-9: 250 ICD-8: 250 |
| Heart failure | ICD-10: I42, I43, I50 ICD-9: 425, 428 ICD-8: 425–428, 7824 |
| Ischaemic heart disease | ICD-10: I20–I25 ICD-9: 410–414 ICD-8: 410–414 |
| Cerebrovascular disease | ICD-10: I60–I69, G45 ICD-9: 430–437 ICD-8: 430–438 |
| Hypertension | ICD-10: I10–I13, I15 ICD-9: 401–405 ICD-8: 400–404 |
| Cancer | ICD-10: C00–C97 ICD-9: 140–209 ICD-8: 140–207 |
ICD codes include primary and secondary diagnoses.
Results
A total of 713,569 patients (53% men, mean age: 74 years) received a first hospital diagnostic code of AF between 1987 and 2013. The matched controls constituted a cohort of 1,393,953 individuals. The diagnosis of PA was found in 384 (0.054%) patients with AF and in 335 (0.024%) controls. Hypothyroidism was diagnosed in 42,417 (5.9%) cases and in 53,116 (3.8%) controls, while hyperthyroidism was diagnosed in 16,488 (2.3%) cases and in 11,588 (0.8%) controls.
The characteristics of individuals who were alive on 31 December 2013 are shown in Table 2. At the end of the follow-up, 662,165 patients with AF and 1,328,583 controls were alive. At the pre-specified date (31 December 2013), the prevalence of PA in the AF population was 0.056% and it was 0.024% in the control population (Figure 1). At the same time, the prevalence of hypothyroidism was 5.9% in the AF population and 3.7% in the control population (Figure 2). The prevalence of hyperthyroidism was 2.3% in the AF population and 0.8% in the control population (Figure 3).
Descriptive statistics of patients with atrial fibrillation and controls who were alive on 31 December 2013.
| Atrial fibrillation | Controls | |
|---|---|---|
| n | 662,165 | 1,328,583 |
| Male sex | 350,371 (52.9) | 695,581 (52.4) |
| Age at diagnosis of atrial fibrillation (years), mean (±SD) | 74 (±12.4) | 74 (±12.3) |
| Primary aldosteronism | 368 (0.056) | 319 (0.024) |
| Hypothyroidism | 39,053 (5.9) | 49,808 (3.8) |
| Hyperthyroidism | 15,249 (2.3) | 10,919 (0.8) |
| Hypertension | 200,408 (30.3) | 158,095 (11.9) |
| Ischaemic heart disease | 173,946 (26.3) | 135,519 (10.2) |
| Heart failure | 175,580 (26.8) | 61,162 (4.6) |
| Diabetes mellitus | 60,348 (9.1) | 61,912 (4.7) |
| Cerebrovascular disease | 118,342 (17.9) | 103,176 (7.8) |
| Cancer | 87,193 (13.2) | 136,792 (10.3) |
| CHA2DS2-VASc score, mean (±SD) | 3.2 (±1.8) | 2.3 (±1.5) |
| Atrial fibrillation | Controls | |
|---|---|---|
| n | 662,165 | 1,328,583 |
| Male sex | 350,371 (52.9) | 695,581 (52.4) |
| Age at diagnosis of atrial fibrillation (years), mean (±SD) | 74 (±12.4) | 74 (±12.3) |
| Primary aldosteronism | 368 (0.056) | 319 (0.024) |
| Hypothyroidism | 39,053 (5.9) | 49,808 (3.8) |
| Hyperthyroidism | 15,249 (2.3) | 10,919 (0.8) |
| Hypertension | 200,408 (30.3) | 158,095 (11.9) |
| Ischaemic heart disease | 173,946 (26.3) | 135,519 (10.2) |
| Heart failure | 175,580 (26.8) | 61,162 (4.6) |
| Diabetes mellitus | 60,348 (9.1) | 61,912 (4.7) |
| Cerebrovascular disease | 118,342 (17.9) | 103,176 (7.8) |
| Cancer | 87,193 (13.2) | 136,792 (10.3) |
| CHA2DS2-VASc score, mean (±SD) | 3.2 (±1.8) | 2.3 (±1.5) |
Data are shown as number (%) unless otherwise indicated; CHA2DS2-VASc: Congestive heart failure (1 point), Hypertension (1 point), Age >75 years (2 points), Diabetes mellitus (1 point), previous Stroke/transient ischemic attack (2 points), Vascular disease (1 point), Age 65–74 years (1 point), female Sex (1 point).
Descriptive statistics of patients with atrial fibrillation and controls who were alive on 31 December 2013.
| Atrial fibrillation | Controls | |
|---|---|---|
| n | 662,165 | 1,328,583 |
| Male sex | 350,371 (52.9) | 695,581 (52.4) |
| Age at diagnosis of atrial fibrillation (years), mean (±SD) | 74 (±12.4) | 74 (±12.3) |
| Primary aldosteronism | 368 (0.056) | 319 (0.024) |
| Hypothyroidism | 39,053 (5.9) | 49,808 (3.8) |
| Hyperthyroidism | 15,249 (2.3) | 10,919 (0.8) |
| Hypertension | 200,408 (30.3) | 158,095 (11.9) |
| Ischaemic heart disease | 173,946 (26.3) | 135,519 (10.2) |
| Heart failure | 175,580 (26.8) | 61,162 (4.6) |
| Diabetes mellitus | 60,348 (9.1) | 61,912 (4.7) |
| Cerebrovascular disease | 118,342 (17.9) | 103,176 (7.8) |
| Cancer | 87,193 (13.2) | 136,792 (10.3) |
| CHA2DS2-VASc score, mean (±SD) | 3.2 (±1.8) | 2.3 (±1.5) |
| Atrial fibrillation | Controls | |
|---|---|---|
| n | 662,165 | 1,328,583 |
| Male sex | 350,371 (52.9) | 695,581 (52.4) |
| Age at diagnosis of atrial fibrillation (years), mean (±SD) | 74 (±12.4) | 74 (±12.3) |
| Primary aldosteronism | 368 (0.056) | 319 (0.024) |
| Hypothyroidism | 39,053 (5.9) | 49,808 (3.8) |
| Hyperthyroidism | 15,249 (2.3) | 10,919 (0.8) |
| Hypertension | 200,408 (30.3) | 158,095 (11.9) |
| Ischaemic heart disease | 173,946 (26.3) | 135,519 (10.2) |
| Heart failure | 175,580 (26.8) | 61,162 (4.6) |
| Diabetes mellitus | 60,348 (9.1) | 61,912 (4.7) |
| Cerebrovascular disease | 118,342 (17.9) | 103,176 (7.8) |
| Cancer | 87,193 (13.2) | 136,792 (10.3) |
| CHA2DS2-VASc score, mean (±SD) | 3.2 (±1.8) | 2.3 (±1.5) |
Data are shown as number (%) unless otherwise indicated; CHA2DS2-VASc: Congestive heart failure (1 point), Hypertension (1 point), Age >75 years (2 points), Diabetes mellitus (1 point), previous Stroke/transient ischemic attack (2 points), Vascular disease (1 point), Age 65–74 years (1 point), female Sex (1 point).
Prevalence of primary aldosteronism in atrial fibrillation cases and in controls without atrial fibrillation from 1987 to 2013 by calendar year.
Prevalence of hypothyroidism in atrial fibrillation cases and in controls without atrial fibrillation from 1987 to 2013 by calendar year.
Prevalence of hyperthyroidism in atrial fibrillation cases and in controls without atrial fibrillation from 1987 to 2013 by calendar year.
After adjusting for age, sex, hypertension, ischaemic heart disease, heart failure, diabetes mellitus, cerebrovascular disease and cancer, the AF population had a significantly higher risk of being diagnosed with PA compared with controls (OR 1.65; 95% confidence interval [CI] 1.40–1.94). The same pattern was found for hypo- and hyper-thyroidism with an increased risk in the AF population (hypothyroidism: OR 1.42; 95% CI 1.39–1.44; hyperthyroidism: OR 2.79; 95% CI 2.71–2.86) compared with the control population. The results of the multiple-adjusted logistic regression analysis are shown in Table 3
Adjusted associations between atrial fibrillation and primary aldosteronism, hypothyroidism and hyperthyroidism compared with controls.
| Odds ratio (95% confidence interval) | p-value | |
|---|---|---|
| Primary aldosteronism | 1.65 (1.40–1.94) | <0.0001 |
| Hypothyroidism | 1.42 (1.39–1.44) | <0.0001 |
| Hyperthyroidism | 2.79 (2.71–2.86) | <0.0001 |
| Odds ratio (95% confidence interval) | p-value | |
|---|---|---|
| Primary aldosteronism | 1.65 (1.40–1.94) | <0.0001 |
| Hypothyroidism | 1.42 (1.39–1.44) | <0.0001 |
| Hyperthyroidism | 2.79 (2.71–2.86) | <0.0001 |
Adjusted for age, sex, hypertension, ischaemic heart disease, heart failure, diabetes mellitus and cerebrovascular disease.
Adjusted associations between atrial fibrillation and primary aldosteronism, hypothyroidism and hyperthyroidism compared with controls.
| Odds ratio (95% confidence interval) | p-value | |
|---|---|---|
| Primary aldosteronism | 1.65 (1.40–1.94) | <0.0001 |
| Hypothyroidism | 1.42 (1.39–1.44) | <0.0001 |
| Hyperthyroidism | 2.79 (2.71–2.86) | <0.0001 |
| Odds ratio (95% confidence interval) | p-value | |
|---|---|---|
| Primary aldosteronism | 1.65 (1.40–1.94) | <0.0001 |
| Hypothyroidism | 1.42 (1.39–1.44) | <0.0001 |
| Hyperthyroidism | 2.79 (2.71–2.86) | <0.0001 |
Adjusted for age, sex, hypertension, ischaemic heart disease, heart failure, diabetes mellitus and cerebrovascular disease.
We performed subgroup analysis of the AF population to compare cardiovascular comorbidity with and without PA. We found that patients with AF and PA had a mean CHA2DS2-VASc score of 2.5, while patients with AF without PA had a similar mean CHA2DS2-VASc score of 2.6 (Table 4).
Cardiovascular comorbidity in relation to the prevalence of primary aldosteronism in patients with atrial fibrillation who were alive on 31 December 2013.
| AF patients with PA (n = 368) | AF patients without PA (n = 661,797) | p-value | |
|---|---|---|---|
| Hypertension | 233 (63) | 200,175 (30) | <0.001 |
| Ischaemic heart disease | 73 (20) | 173,873 (26) | <0.01 |
| Heart failure | 66 (18) | 177,514 (27) | <0.001 |
| Diabetes mellitus | 58 (16) | 87,141 (12) | <0.01 |
| Cerebrovascular disease | 51 (14) | 118,291 (18) | 0.04 |
| CHA2DS2-VASc score, mean (±SD) | 2.5 (±1.6) | 2.6 (±1.7) | 0.1 |
| AF patients with PA (n = 368) | AF patients without PA (n = 661,797) | p-value | |
|---|---|---|---|
| Hypertension | 233 (63) | 200,175 (30) | <0.001 |
| Ischaemic heart disease | 73 (20) | 173,873 (26) | <0.01 |
| Heart failure | 66 (18) | 177,514 (27) | <0.001 |
| Diabetes mellitus | 58 (16) | 87,141 (12) | <0.01 |
| Cerebrovascular disease | 51 (14) | 118,291 (18) | 0.04 |
| CHA2DS2-VASc score, mean (±SD) | 2.5 (±1.6) | 2.6 (±1.7) | 0.1 |
Data are shown as number (%) unless otherwise indicated.
AF: atrial fibrillation; PA: primary aldosteronism; CHA2DS2-VASc: Congestive heart failure (1 point), Hypertension (1 point), Age >75 years (2 points), Diabetes mellitus (1 point), previous Stroke/transient ischemic attack (2 points), Vascular disease (1 point), Age 65–74 years (1 point), female Sex (1 point).
Cardiovascular comorbidity in relation to the prevalence of primary aldosteronism in patients with atrial fibrillation who were alive on 31 December 2013.
| AF patients with PA (n = 368) | AF patients without PA (n = 661,797) | p-value | |
|---|---|---|---|
| Hypertension | 233 (63) | 200,175 (30) | <0.001 |
| Ischaemic heart disease | 73 (20) | 173,873 (26) | <0.01 |
| Heart failure | 66 (18) | 177,514 (27) | <0.001 |
| Diabetes mellitus | 58 (16) | 87,141 (12) | <0.01 |
| Cerebrovascular disease | 51 (14) | 118,291 (18) | 0.04 |
| CHA2DS2-VASc score, mean (±SD) | 2.5 (±1.6) | 2.6 (±1.7) | 0.1 |
| AF patients with PA (n = 368) | AF patients without PA (n = 661,797) | p-value | |
|---|---|---|---|
| Hypertension | 233 (63) | 200,175 (30) | <0.001 |
| Ischaemic heart disease | 73 (20) | 173,873 (26) | <0.01 |
| Heart failure | 66 (18) | 177,514 (27) | <0.001 |
| Diabetes mellitus | 58 (16) | 87,141 (12) | <0.01 |
| Cerebrovascular disease | 51 (14) | 118,291 (18) | 0.04 |
| CHA2DS2-VASc score, mean (±SD) | 2.5 (±1.6) | 2.6 (±1.7) | 0.1 |
Data are shown as number (%) unless otherwise indicated.
AF: atrial fibrillation; PA: primary aldosteronism; CHA2DS2-VASc: Congestive heart failure (1 point), Hypertension (1 point), Age >75 years (2 points), Diabetes mellitus (1 point), previous Stroke/transient ischemic attack (2 points), Vascular disease (1 point), Age 65–74 years (1 point), female Sex (1 point).
Discussion
In this large case–control study, we describe for the first time the prevalence of PA among patients with AF in the general population, and found that it was 0.056%. We found a 65% higher risk of PA in the AF population compared with the control population. Similarly, we found that patients with AF had a 1.4-times higher risk of hypothyroidism and a 2.8-times higher risk of hyperthyroidism compared with controls. The prevalence of hypothyroidism (5.9%) and hyperthyroidism (2.3%) in the current AF population is in agreement with previously reported data of 6% and 2%, respectively.16 The prevalence of hypothyroidism (3.8%) and hyperthyroidism (0.8%) in the controls was found to be very similar to what has been previously reported (3.0–4.2% and 0.8%, respectively) in the general population.17,18 In our study, the prevalence of PA in the AF population appeared to be much lower than that in previously studied hypertensive populations. However, our prevalence of PA in the AF population was still twice as high as that in the control population without AF. Therefore, ruling out PA in the AF population, especially in hypertensive patients and other selected cases, may be of clinical value.
PA prevalence
In our study, the prevalence of PA in the AF population was low, particularly when taking into account that hypertension was present in approximately 30% of the individuals. A potential explanation of the higher prevalence of PA in previous studies that aimed to estimate PA in patients with hypertension may be because of active screening for PA in these studies in combination with selected referral of hypertensive patients.10,19 Potentially, active screening for PA in the AF population results in a higher prevalence of PA. That was obvious in a previous study, where we found that active screening for PA in a AF population <65 years of age revealed PA in 2.6% of the population.20 The ongoing PAPPHY study, which aims to examine the prevalence of PA in the AF population with hypertension, may clarify this issue further.21 Furthermore, we noticed an increasing PA prevalence during the 1990s and early 2000s. We believe that this observation was not a result of higher PA incidence, but rather was an effect of the rising awareness of hypertension in combination with the wide introduction of the aldosterone to renin screening possibility. Indeed, after 2007, the PA prevalence stabilised at the same plateau.
Comorbidities
In the subgroup analysis of the AF population, there was no significant difference between the CHA2DS2-VASc score in patients with AF with and without PA. However, as might have been expected, patients with AF and PA had a significantly higher prevalence of hypertension compared with patients with AF but without PA. In contrast, patients with AF but without PA had a significantly higher prevalence of heart failure and ischaemic heart disease. This indicate that the total burden of cardiovascular disease is equivalent between the two groups (i.e. same numerical CHA2DS2-VASc score), but the predominant diseases are different between the groups.
Relationship between AF and PA
The current study was not designed to show a causal relationship between AF and PA. However, existing data suggest theoretical mechanisms of how PA promotes the development of AF. It is well known that PA promotes AF indirectly through cardiac remodelling caused by hypertension.22 Moreover, aldosterone excess, in addition to causing hypertension, seems to induce a substrate for atrial arrhythmias directly through atrial fibrosis, myocyte hypertrophy and conduction disturbances.23 This is supported by animal models where mineralocorticoid receptor antagonists attenuate aldosterone-induced cardiac fibrosis.24 Moreover, in animal models, aldosterone induces cardiac electrical remodelling by increasing the Ca2+ current density and by causing alternations in the K+ currents, leading to shortening of the action potential.25
Evidence about a possible causal relationship between PA and AF could be provided by a Mendelian randomisation study that uses genetic knowledge (e.g. genome sequencing) to control for reverse causality and confounding factors. Genetic markers for PA have been identified, but their presence is only found in a minority of cases, which limits the possibility of performing such an analysis.26
Relationship between AF and thyroid disorders
The effects of thyroid hormones on the heart are mediated by triiodothyronine (T3). Through a number of intermediate steps, T3 activates renin, which, together with T3-stimulated erythropoietin secretion, results in increased blood volume and cardiac output. Another T3 effect is the upregulation of the β-adrenergic receptors that promotes inotropic and chronotropic cardiac effects.27,28 In hypothyroidism, the most common cardiac manifestations are hypertension and sinus bradycardia.29 Furthermore, hypothyroidism is associated with a higher risk of atherosclerosis due to hyperlipidaemia and hypertension. The role of hypothyroidism in AF development is controversial. In a large Danish cohort study, subclinical hypothyroidism was found to be associated with lower risk of AF,6 while hypothyroidism was not associated with the 10-year risk of new-onset AF in the Framingham Heart Study.7 Moreover, Bruere et al. demonstrated that a history of hypothyroidism was 300% more common than that of hyperthyroidism in 8962 AF patients over 10 years.16 The current study reproduced the results of higher hypothyroidism prevalence in the AF population. Whether there is a causal association between hypothyroidism and AF is still unclear. As described above, hypothyroidism is associated with atherosclerosis, which is a leading cause of ischaemic heart disease and heart failure. This is a possible pathway of how hypothyroidism can create an AF substrate. Another possible mechanism is through hypothyroidism-promoted hypertension, the leading cause of AF.30,31
In hyperthyroidism, common cardiac manifestations are palpitations, systolic hypertension, increased left ventricular mass, increased cardiac output, angina pectoris and AF.27,29,32,33 Hyperthyroidism is believed to promote AF due to a decreased atrial refractory period and increased sympathetic tone with decreased heart rate variability.34 In addition, hyperthyroid patients have been found to have an increased incidence of supraventricular depolarisations, a possible trigger for AF.35
Thus, both hypo- and hyper-thyroidism can lead to AF through different mechanisms. This is in line with animal experiments that found that both hypo- and hyper-thyroidism can induce sympathetic remodelling, which is important for the occurrence and maintenance of AF, leading to increased AF vulnerability.36,37
Strengths and limitations
The main strength of our study is the large study population, including all patients with a registered hospital diagnosis of AF in Sweden over 26 years. However, a critical issue is the identification of AF cases using the NPR. In Sweden, most patients with AF are referred to specialist clinics at some stage during their lifetime and should therefore be included in the NPR, even though some may be seen in primary care only and may therefore have been missed. The cases identified should be correctly diagnosed to a significant extent since validation studies have shown AF diagnoses in the NPR to be correct in 97% of cases.13 The issue of undiagnosed AF is, however, a global problem that is not specific to this particular study. The retrospective design is sensitive to undiagnosed PA, the proportion of which in the general population is unknown. Regardless, we believe that we have identified most of the known PA cases in Sweden during the studied period since all specialized endocrine clinics responsible for PA diagnoses are included in the NPR. Even so, not all cases of PA will have been identified because screening studies usually find a higher prevalence than those reported in unselected cohorts.
The prevalence of hypertension in both cases and controls was almost 50% lower compared to established prevalence numbers, suggesting that hypertension is systematically but equally underdiagnosed in our study. The higher prevalence of hypertension among AF patients might be a result of a higher rate of screening for secondary hypertension in these patients, which in turn may lead to more cases being diagnosed with PA. A low prevalence of comorbidity in the control population, coupled with a lower likelihood of being investigated for endocrine disorders, could introduce a bias. The size of this bias is difficult to estimate, even with adjustment for comorbidities, and some residual confounding cannot be excluded. We found that PA is rare in the population, with fewer than 700 cases identified in almost 2 million people over 25 years, suggesting a large proportion of undiagnosed cases. Given the comparatively low prevalence of PA, screening-based population surveys in order to resolve this issue are unlikely to be performed.
Conclusion
In this large, Swedish, nationwide, case–control study, the prevalence of diagnosed PA in the general AF population was 0.056% in December 2013. PA was twice as common in the AF population compared with matched controls without AF, suggesting an association between PA and AF. The results indicate that screening for PA in selected patients with newly diagnosed AF, particularly in the presence of hypertension, might be of interest.
Author contributions
GM, SAE, AR, LB, GJ and KM were involved in the concept and design of this study. SAE and MA performed the statistical analyses. GM, SAE, AR, LB, MA, GJ and KM interpreted the results. GM drafted the manuscript. SAE, AR, LB, MA, GJ and KM critically revised the manuscript. All authors read and approved the final manuscript.
Acknowledgements
We thank Ellen Knapp from Edanz Group (www.edanzediting.com/ac) for editing a draft of this manuscript.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: this work was supported by grants from the Swedish state under an agreement concerning research and education of doctors (grant number ALFGBG-427301), the Health & Medical Care Committee of the Regional Executive Board, Region Västra Götaland, Sweden, the Swedish Heart and Lung Foundation (grant number 2015-0438) and the Swedish Research Council (grant numbers 2013-5187, 2013-4236). GJ reports personal fees from AstraZeneca and Novo Nordisk and grants and personal fees from Novartis, Pfizer and Shire.
Comments