The study by Krahn et al1 confuses rather than clarifies the question of whether to perform thyroid function tests in patients with recent-onset atrial fibrillation. They contend that routine serum thyrotropin level analysis has a low yield, but methodological flaws and perplexing data presentation prohibit this conclusion.
They excluded from analysis patients with recent cardiovascular surgery, cardiac catheterization, or pacemaker implantation. These patients deserve assessment too, even if they appear to have a readily identifiable cause of their arrhythmia, since treatment of an underlying thyroidal cause simplifies treatment and reduces the risks for unnecessary or potentially toxic therapy.2 An underlying, undetected thyroid disease may have predisposed them to the atrial fibrillation.
They included patients with known thyroid disease. One performs screening tests in the absence of known disease, so including these patients in the analysis confounds the sensitivity and positive predictive value calculations. It is unclear from the