Inadequate study designs for the evaluation of blood pressure monitoring devices and their potential misleading conclusions.

We read with great interest the study by Tan et al. [1]. Their timely investigation aims to provide further insights on the ability of a specific cuffless blood pressure (BP) device (Aktiia Bracelet G1; Aktiia SA, Switzerland) to track BP changes induced by circadian rhythm and medication. We were pleased to read the authors reached similar results to recently published data when comparing daytime BP measurements performed by Aktiia and by a 24-h ambulatory BP monitoring (ABPM, WatchBP O3; MicroLife, Widnau, Switzerland) [2]. These results consolidate Aktiia as a valid alternative to ABPM for daytime BP measurements and for hypertension diagnosis. Although we agree with the results showing that the nighttime dip measured by Aktiia was significantly smaller compared to ABPM, we contend the authors’ conclusion that Aktiia does not accurately track nighttime BP decline. We have recently shown that the magnitude of BP changes captured by Aktiia is highly correlated with those of an ABPM, albeit smaller in amplitude (approximately threefold) [3]. Aktiia and ABPM measure BP based on different principles and, consequently, differences in BP readings between the two modalities are to be expected at different conditions. Firstly, from a physiological perspective, BP readings performed by ABPM oscillometric cuffs are affected by changes in hydrostatic and gravitational forces at different body positions [4]. Conversely, BP readings performed by the proprietary algorithm embedded in Aktiia devices have been demonstrated be robust to those changes [5]. Secondly, from a metrological perspective, the existing protocols for the evaluation of oscillometric cuffs do not test accuracy/precision at the lying position and, consequently, for sleeping [6]. Therefore, data from oscillometric cuffs collected in body positions other than sitting must be treated with prudence. Aktiia, on the contrary, has demonstrated no bias compared with auscultatory BP readings in different body positions [5]. We also consider precipitated the authors’ conclusions regarding the ability of Aktiia to track medication-induced BP changes. The authors have used both Aktiia and home BP monitoring (HBPM) to follow up on medication titration for BP control in three patients. The authors concluded that Aktiia was unable to track medication-induced BP decline compared with HBPM. Such a small sample population might generate inconclusive – if not misleading – results. More data are needed to clarify whether the changes perceived by the HBPM are indeed a real response to the titration – that is, an actual decrease of the systemic BP – or an artefactually augmented measured decrease of BP due to the oscillometric response to local brachial vasoconstriction phenomena. Again, caution is needed when using oscillometric cuffs as a reference for evaluation studies, in particular during drug titration. To the best of our knowledge, no data exist that demonstrate the reliability of oscillometric-measured BP changes as compared to a ground truth during medication titration. Tan et al. [1] should be praised for their work. Cuffless BP devices represent a dramatic shift in the paradigm of BP monitoring and hypertension management and should, therefore, be thoroughly investigated for their translation into clinical practice and research. However, the proposed study design is inadequate to support any conclusion regarding their accuracy or precision. Case in point, the authors compared the readings of a commercial device (Aktiia device) against the readings of another commercial device for which no performance data is available under the tested conditions (oscillometric device). We believe that the definition of an adequate ground truth as a reference for BP measurements is key for a grounded scientific evaluation of novel BP modalities, and we therefore criticize the inadequate study design of the published work, which is a potential source of misleading conclusions. ACKNOWLEDGEMENTS The authors are employed by Aktiia SA. Conflicts of interest There are no conflicts of interest.