Holter monitoring is useful.

welcomes Letters to the Editor and will publish them, if suitable, as space permits. They should not exceed 750 words (excluding references) and may be subject to editing or abridgment. Please submit letters in duplicate, typed double-spaced. Include a fax number for the corresponding author and a completed copyright transfer agreement form (published in every issue). It is important to examine accepted standards of care critically and ignore past assumptions as Schaer et al have done. 1 We also routinely perform Holter monitoring for 24 hours on patients with stroke and transient ischemic attack (TIA), looking for evidence of paroxysmal atrial fibrillation (PAF), and like Schaer et al, we find it a test with a relatively low yield. Nonetheless, we disagree with their conclusion that routine Holter monitoring is not useful. First, in their article, Schaer et al suggest that the prognosis of PAF is not known and that there is little evidence to guide its treatment. However, the Stroke Prevention in Atrial Fibrillation studies and, more recently, a patient-level meta-analysis of 6 randomized controlled trials (comparing aspirin with anticoagu-lation) have clearly shown that patients with PAF and previous ischemic stroke or TIA carry a risk of recurrent stroke and benefit from anticoagulation in an identical manner to patients with chronic AF. 2,3 Furthermore, recent guidelines jointly issued by the American College of Cardiology, the American Heart Association, and the European Society of Cardiology state that the criteria used to select antithrombotic medication for patients with persistent and paroxysmal AF should be the same. 4 As such, we believe that the suggestion by Shaer et al that, after ischemic stroke or TIA, the diagnosis of PAF has an uncertain impact on choice of antithrombotic medication to be very misleading. Second, the analysis by Shaer et al appears to have focused on the cost of Holter monitoring rather than on its cost-effectiveness. Consider: the risks of sustaining a further and disabling stroke in patients with AF and a previous stroke or TIA are large; 2 the costs associated with their care, both in-hospital and over the long-term in the community, are very high; yet anticoagulation is capable of preventing a considerable proportion of these recurrent events. 3 Hence, even at a yield of only 2%, the estimated cost of about $10 000 per case of PAF found would appear to be good value for the money. Indeed, Shaer et al excluded …