Patients with high-normal thyroid function at increased risk for atrial fibrillation

Issue: December 2008

Eight-year follow-up results demonstrated a link between atrial fibrillation and high-normal thyroid function in patients enrolled in the prospective, population-based Rotterdam Study.

Researchers from the Netherlands investigated the association in 1,426 patients aged 55 years and older with normal thyroid-stimulating hormone levels (0.4 mU/L to 4.0 mU/L) who were free from AF at baseline. The researchers also examined the association between free thyroxine levels within the normal range (0.86 ng/dL to 1.94 ng/dL) and AF in 1,177 of the patients.

During a median of eight years follow-up, the researchers reported 105 (7.4%) new cases of AF in patients with normal thyroid function. Increased risk for AF was associated with TSH (HR=1.94; 95% CI, 1.13-3.34) and a graded association of free T₄ levels (HR=1.62; 95% CI, 0.84-3.14) after multivariate adjustments (see table).

“Within the normal range of serum thyroid function parameters, patients with high-normal thyroid function are at increased risk for AF. This finding requires confirmation in other studies,” the researchers wrote. – by Katie Kalvaitis

The Rotterdam Study: Association Among Atrial Fibrillation, Thyroid-Stimulating Hormone and Free Thyroxine Levels

	First quartile	Second quartile	Third quartile	Fourth quartile
TSH levels n=1,426	0.4 mU/L to 1.04 mU/L	1.05 mU/L to 1.51 mU/L	1.52 mU/L to 2.16 mU/L	2.17 mU/L to 3.98 mU/L
AF (patients per cohort	39/358	20/356	26/355	20/357
Free T₄ levels	11.0 pmol/L to 14.4 pmol/L	14.5 pmol/L to 15.9 pmol/L	16.0 pmol/L to 17.9 pmol/L	18.0 pmol/L to 25.0 pmol/L
AF (patients per cohort)	15/297	20/289	23/296	25/295

Arch Intern Med. 2008;168:2219-2224.

The authors argue that since hypothyroidism has never been shown to prevent atrial fibrillation, high-normal thyroid hormone levels may explain this observation. Why this result differs from U.S. observational studies (Sawin et al., 1994; Cappola et al., 2006) in which a higher incidence of atrial fibrillation was limited to those with markedly suppressed TSH levels or at least subnormal levels is not clear. Is an older person with a low-normal range of TSH in Rotterdam at greater risk for nontoxic multinodular goiter with incident autonomous function? If so, this could predict a higher risk for eventually developing subclinical hyperthyroidism or overt hyperthyroidism than seen in a similar age-matched cohort in the United States. In any event, this study — at least preliminary — supports avoiding low TSH levels in patients being treated for hypothyroidism or stage I differentiated thyroid cancer in addition to regularly monitoring patients with low-normal TSH values for thyrotoxicosis and AF.

– Jeffrey Garber, MD

Endocrine Today Editorial Board member