Embolic Risk in Atrial Fibrillation that Arises from Hyperthyroidism

Abstract

Atrial fibrillation, the most common cardiac complication of hyperthyroidism, occurs in an estimated 10% to 25% of overtly hyperthyroid patients. The prevalence of atrial fibrillation increases with age in the general population and in thyrotoxic patients. Other risk factors for atrial fibrillation in thyrotoxic patients include male sex, ischemic or valvular heart disease, and congestive heart failure. The incidence of arterial embolism or stroke in thyrotoxic atrial fibrillation is less clear. There are many reports of arterial thromboembolism associated with hyperthyroidism, including cases of young adults without coexisting risk factors other than thyrotoxic atrial fibrillation.

The use of anticoagulative agents to prevent thromboembolic sequelae of thyrotoxic atrial fibrillation is controversial: national organizations provide conflicting recommendations in their practice guidelines. Herein, we review the medical literature and examine the evidence behind the recommendations in order to determine the best approach to thromboembolic prophylaxis in patients who have atrial fibrillation that is associated with hyperthyroidism.

Atrial fibrillation (AF) is a common cardiac sequela of hyperthyroidism. The use of anticoagulation to prevent thromboembolic complications of thyrotoxic AF is controversial: national organizations provide conflicting recommendations in their practice guidelines.^1,2 Herein, we examine the evidence behind those recommendations in order to determine the best approach to thromboembolic prophylaxis in the patient with AF that is caused by hyperthyroidism.

Epidemiology

The prevalence of hyperthyroidism is estimated to be 2% in women and 0.2% in men.³ Common causes of hyperthyroidism include Graves disease (in approximately 80% of cases), toxic nodular goiter, or the thyrotoxic phase of subacute thyroiditis.⁴ Less common causes include the early phase of Hashimoto thyroiditis, iodine or drug-induced hyperthyroidism, and exogenous intake of thyroid hormone. Rarely, neoplasms (for example, germ cell or pituitary tumors) or pituitary resistance to thyroid hormone can also lead to a hyperthyroid state.

Atrial fibrillation is the most common cardiac complication of hyperthyroidism, occurring in an estimated 10% to 25% of overtly hyperthyroid patients⁵; in comparison, 0.4% of the general population has AF. The prevalence of AF in both populations increases with age.⁶ High-normal thyroid levels or subclinical hyperthyroidism is also associated with an increased risk of developing AF.⁷ Other cardiac manifestations of hyperthyroidism are symptomatic heart failure (which occurs in approximately 6% of patients)⁸ and pulmonary arterial hypertension.⁹ Less than 1% of patients with hyperthyroidism develop dilated cardiomyopathy with left ventricular systolic dysfunction.⁸

In a Danish registry of more than 40,000 patients with hyperthyroidism, approximately 8.3% were diagnosed with AF or atrial flutter within 30 days of the diagnosis of hyperthyroidism.¹⁰ Risk factors for AF in thyrotoxic patients included male sex (odds ratio [OR], 1.8; 95% confidence interval [CI], 1.6–1.9), increasing age (OR, 1.7/10-year increment; 95% CI, 1.7–1.8), ischemic heart disease (OR, 1.8; 95% CI, 1.6–2.0), valvular heart disease (OR, 2.6; 95% CI, 1.9–3.4), and congestive heart failure (OR, 3.9; 95% CI, 3.5–4.4).

The incidence of arterial embolism or stroke in thyrotoxic AF is less clear. There are many case reports of arterial thromboembolism associated with hyperthyroidism, including cases of young adults without coexisting risk factors other than thyrotoxic AF.^11,12 The overall incidence of thromboembolism in the larger trials mentioned below varies widely (range, 8%–40%).^5,13–16 The recommendations for thromboembolic prophylaxis are inferred mainly from the results of these trials.

Pathophysiology

The mechanism of thyroid hormone-induced dysfunction is multifactorial. The heart rate increases due to increased sinoatrial activity, a lower threshold for atrial activity, and shortened atrial repolarization.^17,18 These last 2 factors also create a favorable substrate for the generation of AF, and a similar effect on ventricular myocardium has been linked with ventricular arrhythmias.¹⁹ In addition, volume preload increases due to activation of the renin-angiotensin system,²⁰ contractility increases due to increased metabolic demand and the direct effect of triiodothyronine on cardiac muscle,²¹ and systemic vascular resistance decreases because of triiodothyronine-induced peripheral vasodilation.²² The sum of these effects is a dramatic increase in cardiac output, frequently to more than 7 L/min.²³

In addition, thyroid hormone has numerous effects on coagulation. Studies indicate that hyperthyroidism is associated with increased thrombotic risk. Coagulation abnormalities such as shortened activated partial thromboplastin time, increased fibrinogen levels, and increased factor VIII and factor X activity,^24,25 and clinical sequelae such as stroke⁵ are seen frequently in patients in sinus rhythm with thyrotoxicosis. In a recent large study, more than 3,000 young adults with hyperthyroidism and more than 25,000 euthyroid young adults were monitored prospectively for 5 years.²⁶ After controlling for multiple stroke risk factors (including AF), the investigators noted a 1.44-times' greater chance of ischemic stroke among the thyrotoxic population. Although undocumented paroxysmal AF may contribute to embolic phenomena, these studies^24–26 suggest that hyperthyroidism is associated with a prothrombotic state and ischemic stroke independent of atrial arrhythmias.

Clinical Evidence in Regard to Atrial Fibrillation, Thyrotoxicosis, and Emboli

Staffurth and colleagues¹³ studied 262 patients who had a history of thyrotoxicosis and AF. Of those patients, 21 (8%) had 26 episodes of arterial embolism. Embolic events were noted in 11 patients who had concurrent AF and hyperthyroidism; in 3 of these patients, the diagnosis of thyrotoxicosis predated that of AF by as long as 9 years. An additional 3 patients experienced embolic events within 2 weeks of reverting to sinus rhythm—in 1 patient, this occurred a day after successful cardioversion following 3 years of AF. The remaining 7 patients experienced a thromboembolic event while in chronic AF, but from 3 months to 10 years after becoming euthyroid. In the study, the use of anticoagulation and the presence of congestive heart failure were not documented. Furthermore, no control group of thyrotoxic patients in sinus rhythm was included. Although this brings into question whether many of these thromboembolic events can be attributed to the thyrotoxic state, the authors recommended anticoagulation for most patients with thyrotoxic AF.

From among 210 thyrotoxic patients, Yuen and colleagues¹⁴ reported a case series of 21 patients with thyrotoxicosis and AF. Systemic embolism was noted in 5 patients (24%), all while the patients were still hyperthyroid. One event occurred in a patient who underwent electrical cardioversion and was not receiving anticoagulation. The outcomes of the 189 patients who did not have AF were not reported. An addendum to the paper noted that 9 more thyrotoxic patients developed AF during long-term follow-up, and 2 additional thromboembolic events occurred in this group. Accordingly, the overall incidence of systemic emboli was 23%. The authors recommended anticoagulation in all patients with thyrotoxic AF.

Bar-Sela and associates¹⁵ reported a case series of 30 patients with AF from among 142 thyrotoxic patients. Twelve (40%) had documented embolic events. No thyrotoxic patient who remained in sinus rhythm had an embolic event. There was no significant difference in baseline characteristics between those with and without embolic events; however, the presence or absence of heart failure was not specifically categorized. The authors recommended that until a controlled prospective trial could be carried out, thyrotoxic patients should receive anticoagulation.

Hurley and co-authors¹⁶ described their experience over 6 years with 381 patients who had thyrotoxicosis. Among these, 70 developed AF or atrial flutter and 39 reverted to sinus rhythm during antithyroid therapy. In 8 cases of arterial embolism, 6 patients showed evidence of congestive heart failure (which was not further characterized by the authors), including 2 who were found at autopsy to have mitral stenosis. The higher thromboembolic risk in these patients may have been due to these underlying, established risk factors and not necessarily to the hyperthyroid state. The authors therefore recommended anticoagulation only in patients with overt heart failure.

In the most statistically rigorous clinical trial, Petersen and co-investigators⁵ retrospectively analyzed 610 patients with thyrotoxicosis, 91 of whom had developed AF. A large age disparity was noted: more than 25% of those older than 60 years developed AF, compared with less than 1% who were younger than 50 years. Only the risk of cerebrovascular events was determined; other arterial-embolic events were not included in the study, nor was the presence of heart failure. The unadjusted risk of stroke in the AF and sinus groups was, respectively, 6.4% and 1.7% in the 1st year, and 13% versus 2.9% over a mean follow-up period of 39 months. These results implied that AF carried a higher risk of cerebrovascular events. However, the authors used logistic regression methods to analyze age, sex, and the presence of AF as independent variables. They determined that only age was a significant independent risk factor, and that a tendency toward more events in persons with AF was seen. According to the authors, either an unadjusted-for age discrepancy or the lack of a control group in sinus rhythm would have accounted for the disparate results in the earlier studies.^13–16 They suggested that the recommendation of prophylactic anticoagulative therapy should await further studies.

Current Recommendations for Anticoagulation in Thyrotoxicosis

In the medical literature, clinical evidence for the anticoagulation of thyrotoxic patients with AF comes predominantly from case series or retrospective cohort trials conducted before 1990. The trials discussed above^5,13–16 documented the incidence of thromboembolism in populations of patients with AF and thyrotoxicosis. However, the studies lacked a control group or were designed to compare patients with thyrotoxic AF to a population with thyrotoxicosis in sinus rhythm. A study that compares “lone” AF with thyrotoxic AF would be beneficial in exploring the independent thromboembolic potential of hyperthyroidism. In addition, the use of thromboprophylaxis or controlling for conventional thromboembolic risk factors (such as with use of the well-established and widely used CHADS₂ risk score²⁷) is sparsely and incompletely documented. Finally, patients with existing AF were not excluded from the earlier trials. To date, no trials involving the risk–benefit ratio or the efficacy of using anticoagulation in patients with thyrotoxic AF have been conducted.

In an analysis of the aforementioned clinical trials, Presti and Hart²⁸ observed that the data suggested an increased rate of embolism; that most emboli affect the central nervous system, which is particularly devastating; and that emboli seem to occur during the early course of thyrotoxic AF, particularly in patients with heart failure. The authors concluded that, in the absence of further evidence, the decision to administer short-term anticoagulation should be made on an individual basis after carefully analyzing age, associated cardiac disease, and the risks of therapy.

The controversial nature of this topic is apparent in the disparity between 2 national guidelines. The American College of Chest Physicians guidelines¹ acknowledge that some studies report a high incidence of thromboembolic complications in thyrotoxic AF, yet significant methodologic problems complicate the interpretation of that finding.^13–16 The guideline authors concluded that thyrotoxicosis does not appear to be a validated risk factor in stroke, and they recommended antithrombotic therapy regardless of whether hyperthyroidism is present.¹

The American College of Cardiology/American Heart Association (ACC/AHA) guidelines² state that, although the topic is controversial and increased risk has not been definitively proved, anticoagulative therapy is recommended “in the absence of a specific contraindication, at least until a euthyroid state has been restored and heart failure has been cured.”

It should be noted that the anticoagulative effect of vitamin K antagonists is affected by the patient's thyroid status—elevated thyroid levels increase the anticoagulative effect, and antithyroid agents (such as propylthiouracil or methimazole) may diminish the effect. In view of this dynamic situation, careful monitoring is required to keep the international normalized ratio within recommended levels during the treatment of patients with AF and hyperthyroidism.

Conclusion

Many of the aforementioned studies have shown an alarming incidence of arterial embolism with thyrotoxicosis and AF. However, the most evidence-based study⁵ did not find that the trend toward increased embolic risk was a statistically significant independent risk factor when other known risk factors were considered. This finding may be due in part to the unusually high incidence of stroke in the control groups, small numbers of patients, and lack of control for the presence of heart failure. Given the lack of clear evidence, the ACC/AHA classification of thyrotoxicosis as a moderate thromboembolic risk factor seems to be reasonable, and the recommendation to initiate anticoagulation when there are no contraindications appears to be warranted. More evidence-based trials are necessary to clarify this issue.