LAUSR

W e read with great interest the study by Grillo et al. [1] which critically evaluates currently used empirical formulas to estimate cuff-derived brachial mean arterial pressure (MAP) from SBP and DBP. The authors must be congratulated for this impressive study. The reference MAP was derived by integrating the pressure waveform recorded by applanation tonometry on the controlateral brachial artery in 1526 patients belonging to three distinct cohorts. Various empirical formulas [2,3] add to DBP an established percentage of pulse pressure (PP) (form factor). The authors conclude that currently used empirical formulas are unreliable to estimate brachial MAP from SBP and DBP [1]. On the contrary, in our opinion, one must acknowledge the extremely high accuracy, good precision and simplicity of at least two of these formulas [2,3]. The average difference between the formula we have previously proposed (MAP1⁄4DBPþPP 0.33þ 5mmHg) [3] and the reference MAP was 0.35 mmHg. This represents only 0.35% MAP and this is well within acceptable measurement error of validation protocols for noninvasive BP devices, namely the Association for the Advancement of Medical Instrumentation guidelines [4] or the European Society of Hypertension International Protocol-2010 [5]. Given such an extremely high accuracy and because data were obtained over a wide range of BP and age, we have difficulty to share the authors’ conclusion that ‘A more accurate evaluation of MAP based on the analysis of brachial PP curve is advisable’ [1]. The tested empirical formulas resulted in an essentially similar SD of the error of about 3 mmHg [1]. This reflects good precision of MAP estimates according to current standard and of the same range than SBP or DBP precision from cuffs [4,5]. It is true there was a problematic bias across the range resulting in a tendency toward slight overestimation/underestimation of MAP for lower/higher BP values [1], and we agree that a fixed form factor is therefore imperfect. To solve this issue, the authors propose complex, patient-specific formulas in which the form factor is corrected for sex and for DBP, and they remain to be tested on an independent sample. In an attempt to remain close from the valuable original formulas, a close look at Figure 3 [1] suggests that a slightly different formalism of these formulas at low or high BP level should be sufficient to correct for their pressure dependence, and this point deserves further studies. Finally, it is interesting to note that the form factor depends upon PP more than upon DBP, SBP or MBP at the central aortic level (3). Had the authors tested PP in their model?