Clinical outcomes of atrial fibrillation with hyperthyroidism.

BACKGROUND: Atrial fibrillation (Afib) is a common cardiac manifestation of hyperthyroidism. The data regarding outcomes of Afib with and without hyperthyroidism are lacking. HYPOTHESIS: We hypothesized that patients with Afib and hyperthyroidism have better clinical outcomes, compared with Afib patients without hyperthyroidism.
METHODS: We queried the National Inpatient Sample database for years 2015-2017 using Validated ICD-10-CM codes for Afib and hyperthyroidism. Patients were separated into two groups, Afib with hyperthyroidism and without hyperthyroidism.
RESULTS: The study was conducted with 68 095 278 patients. A total of 9 727 295 Afib patients were identified, 90 635 (0.9%) had hyperthyroidism. The prevalence of hyperthyroidism was higher in patients with Afib (0.9% vs 0.4%, P < .001), compared with patients without Afib. Using multivariate regression analysis adjusting for various confounding factors, the odds ratio of Afib with hyperthyroidism was 2.08 (CI 2.07-2.10; P < .0001). Afib patients with hyperthyroidism were younger (71 vs 75 years, P < .0001) and more likely to be female (64% vs 47%; P < .0001) as compared with Afib patients without hyperthyroidism. Afib patients with hyperthyroidism had lower prevalence of CAD (36% vs 44%, P < .0001), cardiomyopathy (24.1% vs 25.9%, P < .0001), valvular disease (6.9% vs 7.4%, P < .0001), hypertension (60.7% vs 64.4%, P < .0001), diabetes mellitus (29% vs 32%, P < .0001) and obstructive sleep apnea (10.5% vs 12.2%, P < .0001). Afib with hyperthyroidism had lower hospitalization cost ($14 968 ± 21 871 vs $15 955 ± 22 233, P < .0001), shorter mean length of stay (5.7 ± 6.6 vs 5.9 ± 6.6 days, P < .0001) and lower in-hospital mortality (3.3% vs 4.8%, P < .0001. The disposition to home was higher in Afib with hyperthyroidism patients (51% vs 42; P < .0001).
CONCLUSION: Hyperthyroidism is associated with Afib in both univariate and multivariate analysis. Afib patients with hyperthyroidism have better clinical outcomes, compared with Afib patients without hyperthyroidism.

INTRODUCTION

Atrial fibrillation (Afib) is the most common arrhythmia in the United States. Studies have reported that two thirds of Afib patients who visit the emergency department are admitted to the hospital; this results in higher healthcare costs compared to patients who experience other types of arrhythmias. , Hyperthyroidism, coronary arterial disease, congestive heart failure, valvular disease, chronic pulmonary disease, hypertension, diabetes mellitus, obstructive sleep apnea, obesity, RA/collagen vascular disease, and old age are risk factors for Afib. , , , , , Each risk factor has varying effects on the clinical outcomes, such as hospitalization cost, length of stay, in‐hospital mortality, and discharge disposition for Afib patients. Our study examines the clinical outcomes of Afib patients with hyperthyroidism. The association of Afib and hyperthyroidism is well known; however, the clinical outcomes as mentioned above have not been reported.

METHODS

This study was conducted using the National Inpatient Sample (NIS) database from October 2015 to December 2017. NIS is the largest inpatient care database in the United States that was developed by the Healthcare Cost and Utilization Project (HCUP) and is the sponsored Agency for Healthcare Research. It represents more than 95% of the US population from 47 states and consists of a 20% stratified sample of all inpatient discharges from US hospitals excluding rehabilitations and long‐term acute care hospitals. The self‐weighting design used has the advantage of producing stable and precise data. Its principle of inclusivity, such as incorporating all‐payers (Medicare, Medicaid, primary payers and uninsured), allows for estimates on trends for more than 36 million inpatient hospital stays nationally. Each admission contains information on patient demographics, primary and secondary diagnosis, comorbidities, length of stay, hospital costs, hospital characteristics, and discharge disposition. ,

The study cohort was derived from a de‐identified and publicly available database; therefore, it qualified for exemption from formal approval of the Institutional Review Board (IRB). The International Classification of Disease, 10th revision, Clinical Modification (ICD 10‐CM) codes were used to identify the relevant diagnoses for our study in the NIS database which are list in Table 1. The primary outcome was to determine the hospitalization cost, length of stay, in‐hospital mortality, and discharge disposition for Afib patients due to hyperthyroidism. Our study included patients 18 years and older. The baseline characteristics, namely, age, gender, and race, were obtained using ICD‐10‐CM codes. The individual comorbidities were obtained from the Elixhauser comorbidity software.

TABLE 1

ICD 10 codes

Atrial fibrillation	I48.0, I48.1, I48.2, I48.91
Coronary arterial disease	I20, I21, I22, I24.0, I24.8, I24.9, I25.1
OSA	G47.33
Hyperthyroidism
Thyrotoxicosis with diffuse goiter	E05.0
Thyrotoxicosis with toxic single thyroid nodule	E05.1
Thyrotoxicosis with toxic multinodular goiter	E05.2
Thyrotoxicosis from ectopic thyroid tissue	E05.3
Thyrotoxicosis factitia	E05.4
Other thyrotoxicosis	E05.8
Thyrotoxicosis, unspecified	E05.9
Cardiomyopathy	I42.0, I42.1, I42.2, I42.5, I42.6, I42.7, I42.8, I42.9, I43
Myocarditis	I40, I41

STATISTICAL ANALYSIS

The data analysis and extraction were done using SAS statistical software version 9.4. All continuous variables, such as length of stay, in‐hospital mortality, and hospitalization cost were compared using Student's t test. These variables were presented as a mean and standard deviation (SD) for normally distributed variables, while median and interquartile ranges were used for non‐Gaussian distributed variables. On the other hand, categorical variables were analyzed using the Pearson chi‐square test. These variables were presented as a weighted frequency in percentages. A P < .05 was considered statistically significant. The propensity score matching method, as demonstrated in Table 4, was used to adjust for cofounders and aid in assessment of the clinical outcomes of patients with AF with and without hyperthyroidism.

TABLE 4

Clinical characteristics of Afib with hyperthyroidism versus Afib without Hyperthyroidism in 2015q4‐2017 patients

Characteristics	Afib with hyperthyroidism	Afib without hyperthyroidism	P value
N = 9 727 295	N = 90 635 (0.9%)	N = 9 636 660 (99.1%)
Age
Mean years (SD)	70.9 ± 14	75.1 ± 11.8	<.0001
Gender
Male	32 730 (36.1%)	5 051 763 (52.4%)	<.0001
Female	57 885 (63.9%)	4 582 378 (47.6%)
*Missing—2540
Age groups
18‐34	1500 (1.6%)	44 390 (0.5%)	<.0001
35‐49	5750 (6.3%)	230 980 (2.4%)
50‐64	19 630 (21.7%)	1 465 279 (15.2%)
65‐79	34 900 (38.5%)	3 882 973 (40.3%)
≥80	28 855 (31.8%)	4 013 038 (41.6%)
Race
Caucasians	61 625 (68%)	7 569 156 (78.5%)	<.0001
African‐Americans	15 070 (16.6%)	825 195 (8.6%)
Others	13 935 (15.4%)	1 242 204 (12.9%)
*Missing—110
Elixhauser comorbidities
Coronary arterial disease	33 015 (36.4%)	4 218 583 (43.8%)	<.0001
Cardiomyopathy	21 805 (24.1%)	2 504 884 (25.9%)	<.0001
Valvular disease	6250 (6.9%)	711 315 (7.4%)	<.0001
Chronic pulmonary disease	23 360 (25.8%)	2 552 609 (26.5%)	<.0001
Hypertension	55 025 (60.7%)	6 203 407 (64.4%)	<.0001
Diabetes mellitus	26 240 (28.9%)	3 064 419 (31.8%)	<.0001
Obesity	12 800 (14.1%)	1 430 389 (14.8%)	<.0001
Collagen vascular disease	2605 (2.9%)	289 370 (3%)	<.02
Obstructive sleep apnea	9535 (10.5%)	1 177 355 (12.2%)	<.0001
Outcomes
In‐hospital mortality	3035 (3.3%)	466 990 (4.8%)	<.0001
*Missing—8085			<.0001
Length of stay, days, median (IQR)	5.7 ± 6.6	5.9 ± 6.6	<.0001
Total hospitalization costs, $, median (IQR)	14 968 ± 21 871	15 955 ± 22 233	<.0001
Disposition
Discharge to home	45 810 (50.6%)	4 023 383 (41.8%)	<.0001
Transfer other: includes skilled Nursing Facility (SNF), Intermediate Care Facility (ICF), and another type of facility	21 675 (23.9%)	2 835 168 (29.4%)
Home health care	16 755 (18.5%)	1 978 349 (20.5%)
Against medical advice (AMA)	885 (0.9%)	66 195 (0.7%)
*Missing—8085

Abbreviations: A. fib, atrial fibrillation; IQR, interquartile range; SD, standard deviation.

RESULTS

This study included data from the NIS database for patient hospital admissions for the years 2015 to 2017. After excluding patients <18 years of age, the total number of patients included in the study was 68 095 278. Patients were first divided into two groups, Afib and no Afib; 9 727 295 patients were in the Afib group and 58 367 983 patients were in the no Afib group. The overall incidence of Afib was 14.3% (shown in Figure 1). An extensive comparison of baseline patient‐level characteristics between patients with Afib and patients without AFib is shown in Table 2. Patients with Afib were older with a mean age of 75.1 and more likely to be male (52.3 vs 40.3, P < .001), when compared to patients without Afib. The prevalence of hyperthyroidism in patients with Afib was 0.9%, whereas, in patients without Afib, it was 0.4%, P < .001. In addition, Table 2 showed that other comorbidities such as, coronary arterial disease, congestive heart failure, valvular disease, chronic pulmonary disease, hypertension, obstructive sleep apnea, obesity and rheumatic/collagen vascular disease were more prevalent in patients with Afib compared to patients without Afib. The association of Afib and hyperthyroidism was further assessed using multivariate regression analysis (Table 3). After adjusting for all other comorbidities, hyperthyroidism association with Afib remained statistically significant (odds ratio 2.08, CI 2.07‐2.10, P < .001).

FIGURE 1

The flow chart of the sample included in the study

TABLE 2

Clinical characteristics of Afib versus without Afib in 2015q4‐2017 patients

Characteristics	Afib	No Afib	P value
N = 68 095 278	N = 9 727 295 (14.3%)	N = 58 367 983 (85.7%)
Age
Mean years (SD)	75.1 ± 11.8	54.8 ± 19.9	<.0001
Gender
Male	5 084 493 (52.3%)	23 503 849 (40.3%)	<.0001
Female	4 640 263 (47.7%)	34 842 419 (59.7%)
*Missing—24 255
Age groups
18‐34	45 890 (0.5%)	13 005 590 (22.3%)	<.0001
35‐49	236 730 (2.4%)	9 611 285 (16.5%)
50‐64	1 484 909 (15.3%)	14 771 668 (25.3%)
65‐79	3 917 873 (40.3%)	13 930 404 (23.9%)
≥80	4 041 893 (41.5%)	7 049 036 (12.1%)
Race
Caucasians	7 630 781 (78.4%)	36 527 564 (62.6%)	<.0001
African‐Americans	840 265 (8.6%)	9 090 433 (15.6%)
Others	1 256 139 (12.9%)	12 748 826 (21.8%)
*Missing—1270
Elixhauser comorbidities
Coronary arterial disease	4 251 598 (43.7%)	9 786 820 (16.8%)	<.0001
Cardiomyopathy	2 526 689 (25.9%)	3 948 298 (6.8%)	<.0001
Valvular disease	717 565 (7.4%)	1 261 254 (2.2%)	<.0001
Cardiomyopathy	951 130 (9.8%)	1 318 200 (2.3%)	<.0001
Chronic pulmonary disease	2 575 969 (26.5%)	9 842 246 (16.9%)	<.0001
Hypertension	6 258 432 (64.3%)	26 191 277 (44.9%)	<.0001
Diabetes mellitus	3 090 659 (31.8%)	12 682 714 (21.7%)	<.0001
Hyperthyroidism	90 635 (0.9%)	252 270 (0.4%)	<.0001
Obesity	1 443 189 (14.8%)	7 980 097 (13.7%)	<.0001
Alcohol abuse	260 575 (2.7%)	2 755 463 (4.7%)	<.0001
Collagen vascular disease	291 975 (3%)	1 497 259 (2.6%)	<.0001
Obstructive sleep apnea	1 186 890 (12.2%)	3 285 019 (5.6%)	<.0001

Abbreviations: Afib, atrial fibrillation; BMI, body mass index; SD, standard deviation.

TABLE 3

Multiple regression analysis of Atrial fibrillation with all of below risk factors in the table

	ODDS ratio	95% Wald confidence limits	P value
Hyperthyroidism	2.23	2.21–2.25	<.0001
Coronary artery disease (CAD)	1.83	1.82–1.84	<.0001
Cardiomyopathy	3.61	3.60–3.63	<.0001
Valvular disease	1.76	1.75–1.77	<.0001
Hypertension (HTN)	1.97	1.96–1.98	<.0001
Diabetes mellitus (DM)	1.04	1.03–1.06	<.0001
OSA(Obstructive sleep apnea)	1.92	1.90–1.92	<.0001
Collagen vascular disease	0.92	0.0.91–0.93	<.0001
Chronic pulmonary disease	1.26	1.25–1.27	<.0001
Alcohol abuse	1.12	1.11–1.13	<.0001
Age	1.07	1.05–1.08	<.0001

The Afib group was further divided into two groups; with hyperthyroidism and without hyperthyroidism, to assess the impact of the presence of hyperthyroidism. Mean hospitalization cost, length of stay, in‐hospital mortality, were lower in Afib patients with hyperthyroidism compared to patients with Afib related to other comorbidities (Table 4). Afib patients with hyperthyroidism had a higher discharge disposition to home compared to patients with Afib related to the comorbidities listed above (50.6 vs 41.8, P < .0001). The data in Table 4 also demonstrated that Afib patients without hyperthyroidism tend to have a higher prevalence of other comorbidities such as coronary arterial disease, congestive heart failure, valvular disease, chronic pulmonary disease, hypertension, diabetes mellitus, obstructive sleep apnea, obesity and rheumatic/collagen vascular disease compared to Afib patients with hyperthyroidism. Pateints with Afib and hyperthyroidism were younger as shown in Table 4.

The annual trends of mean hospitalization cost, length of stay and in‐hospital mortality were compared between Afib with hyperthyroidism and Afib without hyperthyroidism, as shown in Figures 2, 3, 4, respectively. Overall, Afib with hyperthyroidism had lower hospitalization cost, length of stay and in‐hospital mortality when compared to Afib without hyperthyroidism. The percent mortality for Afib with and without hyperthyroidism decreased from 2015 to 2017. However, for periods 2016 and 2017, the mortality in Afib patients with hyperthyroidism was significantly lower than Afib patients without hyperthyroidism as shown in Figure 4.

FIGURE 2

Comparison of the mean cost of hospitalization costs between patients with Afib and hyperthyroidism and Afib patients without hyperthyroidism

FIGURE 3

Comparison of length of stay (LOS) between patients with Afib and hyperthyroidism and Afib without hyperthyroidism

FIGURE 4

Depiction of in‐hospital mortality among patients with Afib and hyperthyroidism and Afib patients without hyperthyroidism

DISCUSSION

There was an association between Afib and hyperthyroidism in a large patient sample size after multivariate regression analysis. The prevalence of hyperthyroidism in Afib was higher when compared to patients without Afib. The study demonstrated that hospitalization costs, mean length of stay and in‐hospital mortality were lower in Afib patients with hyperthyroidism compared with Afib patients without hyperthyroidism. We also found that the disposition to home was higher in Afib with hyperthyroidism patients compared with Afib patients without hyperthyroidism.

The pathophysiology of Afib in hyperthyroid patients is related to the effect of triiodothyronine (T3), which is the biologically active form of thyroid hormone, directly on atrial myocytes, or indirectly by altering peripheral vascular resistance, or by influencing the sympathoadrenergic system. T3 results in increased sarcoplasmic calcium ATPase, myosin heavy chain α, voltage gated potassium channels, sodium channels and β1 adrenergic receptors. These cause increased heart rate, systolic blood pressure, ventricular contractility and cardiac hypertrophy. , There are a few proposed mechanisms by which thyroid hormones increase the risk of atrial fibrillation, including, shortening of the action potential duration, activation of arrhythmogenic foci, increasing the left atrial pressure due to increased left ventricular mass and impaired relaxation, and ischemia secondary to increased resting heart rate. , , , , ,

Risk factors for Afib include, but are not limited to, coronary arterial disease, congestive heart failure, valvular disease, chronic pulmonary disease, hypertension, diabetes mellitus, obstructive sleep apnea, obesity, rheumatic/collagen vascular diseases and old age. , , , , , From our analysis, we identified that Afib patients with hyperthyroidism had a low prevalence of these coexisting comorbidities (Table 4). Afib patients with hyperthyroidism were younger and more likely to be female compared with Afib patients without hyperthyroidism. The increased prevalence in females supports the finding that thyroid diseases are more common in women, which may be related to the increased prevalence of autoimmune diseases, as well as greater hormonal fluctuations (menstruation, pregnancy) in women. ,

Hospitalization cost, mean length of stay and in‐hospital mortality were consistently lower in Afib patients with hyperthyroidism compared with Afib patients without hyperthyroidism over the period of 2015 to 2017, as shown in Figures 2, 3, 4, respectively. These findings are likely due to the reversibility of the underlying disease. A higher percentage of these patients were able to be discharged home when compared to Afib patients without hyperthyroidism.

Anti‐thyroid medications such as carbimazole/methimazole, propylthiouracil or radio‐iodine are used to achieve a euthyroid state, subsequently, thyroidectomy is performed as definitive treatment. Until a euthyroid state is achieved, beta blockers are used to control heart rate and reduce cardiac failure. , Up to two‐third of patients who are treated for hyperthyroidism convert to sinus rhythm. The ease of conversion to sinus rhythm is also likely higher due to a lower prevalence of coexisting risk factors in Afib patients with hyperthyroidism (Table 4). Studies have shown that Afib patients with other risk factors, such as coronary artery disease, hypertension and cardiomyopathy have worse clinical outcomes, including increased mortality. , ,

LIMITATIONS

Our study sample size is large and representative of 95% of hospitals in the US. However, there are limitations to this study. We were not able to discern whether patients developed Afib before or after the development of the hyperthyroidism given the nature of the database. We relied primarily on diagnosis codes for hyperthyroidism and Afib, which could potentially lead to exposure and outcome misclassification. We did not investigate disease severity, which could affect the development of Afib.

CONCLUSION

From our study, it was determined that patients with Afib and hyperthyroidism have reduced hospitalization costs, shorter length of stay and lower in‐hospital mortality when compared to Afib without hyperthyroidism. We also found that patients with hyperthyroidism have a 3‐fold increased risk of Afib compared to those without hyperthyroidism. Given the potential for reversibility of this condition, this may account for the better clinical outcomes observed in these patients.

CONFLICT OF INTEREST

The authors declare no conflict of interests for this article.