LAUSR

W e thank Ishak et al. [1] for sharing their concerns and offering us the opportunity to emphasize several results of our review. In the methods of our review, we state that ‘the accuracy of BP measurements over a sleeve breaks down to possible bias (systematic error) and possible imprecision (random error)’ [2]. Our results convincingly show that the bias is small, if any, with thin sleeves. Ishak et al. [1] are worried by the imprecision of BP measurements over a sleeve in individual patients that may lead to misclassifying them. This was analyzed in a single study (376 patients) and these results are reported in Table S6 of our review: the variability between measurements on a bare arm and over a thin sleeve at the individual level was similar to the variability between two successive measurements on a bare arm at the individual level [3]. The imprecision induced by thin sleeves is, therefore, also small, if any. Taken together, these two points support our conclusion that BP measurement over a thin sleeve is acceptable when the shirt can not be removed, and preferred to measurement below the rolled-up sleeve, which might be less reliable. The risk of misclassification because of BP measurement errors is an important issue but based on the best available evidence, measurements over a thin sleeve should not be our greatest fear with this regard. Anecdotal evidence is prone to confounding. It is unlikely that a thin sleeve has induced a measurement error above 30 mmHg for SBP and other factors should be considered, such as the activities performed by the patient in the afternoon, before he replaced the cuff once quiet at home in the evening.