Incremental Value of a Cuff-Less Blood Measurement Device Compared to Clinic Blood Pressure Measurement

Objective: Cuffless, pulse-transit-time based 24-hour blood pressure measurement (BPM) lacks clinical validation based on recent protocols. It tends to higher values than cuff-based BPM, making standard hypertension cut-off values inapplicable. Additionally, the cuffless device exhibits a lower ability to track 24-hour blood pressure changes than standard devices. Due to these limitations, this study aims to evaluate if there is an advantage of cuffless BPM over clinic BPM (CBP). Design and method: 166 participants were prospectively enrolled for simultaneously performed cuffless (TestBP, Somnotouch NIBP) and cuff-based (RefBP, Spacelabs or Mobil-o-graph) 24h-BPM. CBP was defined as the mean of 3 cuff-based BPM: after 5 min of rest, BP was measured sequentially at both upper arms. Then, first, RefBPM was used as calibration for cuffless TestBP. We compared the TestBP and CBP to mean awake RefBP including calculation of test characteristics for the diagnosis of elevated BP using ROC curve analysis to identify ideal cut-off values for the TestBP. Gold standard for the detection of elevated BP was mean awake RefBP (>134/84 mmHg). Results: Median age was 51 years (IQR 40-63). Mean awake systolic and diastolic RefBP was 134±15 mmHg and 83±10 mmHg, respectively. Mean differences between TestBP and RefBP were 6±13/5±8 mmHg and mean absolute differences 11±9/7±6 mmHg, for systolic/diastolic values, respectively, whereas for CBP 7±13/5±7 mmHg and 12±10/7±5 mmHg. There was a linear relationship between TestBP/CBP and RefBP (Figure, panels A-D). In ROC analysis CBP and TestBP showed an AUC for the diagnosis of hypertension on awake RefBP of 0.845/0.873 and 0.862/0.877 for systolic/diastolic values, respectively. Best cut-off for hypertension was at 133/88 mmHg for mean 24h, 137/88 mmHg for awake and 136/80 mmHg for asleep cuffless TestBP values. Misclassifications compared to RefBP with these (ASV) and previously published (TS) cut-offs for TestBP and usual cut-offs for CBP are shown in the figure (panels E-F). Conclusions: We found a low incremental value of the TestBP device in relation to CBP, with comparable AUCs and correlation coefficients. With the application of the adapted awake BP cut-off of 137/88 mmHg, masked hypertension misclassification may be reduced