Improved agreement and diagnostic accuracy of a cuffless 24-h blood pressure measurement device in clinical practice

Abstract

Abstract A cuffless blood pressure (BP) device (TestBP) using pulse transit time is in clinical use, but leads to higher BP values compared to a cuff-based 24 h-BP reference device (RefBP). We evaluated the impact of a recent software update on BP results and TestBP’s ability to differentiate between normo- and hypertension. 71 individuals had TestBP (Somnotouch-NIBP) and RefBP measurements simultaneously performed on either arm. TestBP results with software version V1.5 were compared to V1.4 and RefBP. Mean 24 h (± SD) BP for the RefBP, TestBP-V1.4 and TestBP-V1.5 were systolic 134.0 (± 17.3), 140.8 (± 20) and 139.1 (± 20) mmHg, and diastolic 79.3 (± 11.7), 85.8 (± 14.1) and 83.5 (± 13.0) mmHg, respectively (p-values < 0.001). TestBP-V1.5 area under the curve (95% confidence interval) versus RefBP for hypertension detection was 0.92 (0.86; 0.99), 0.94 (0.88; 0.99) and 0.77 (0.66; 0.88) for systolic and 0.92 (0.86; 0.99), 0.92 (0.85; 0.99) and 0.84 (0.74; 0.94) for diastolic 24 h, awake and asleep BP respectively. TestBP-V1.5 detected elevated systolic/diastolic mean 24 h-BP with a 95%/90% sensitivity and 65%/70% specificity. Highest Youden’s Index was systolic 133 (sensitivity 95%/specificity 80%) and diastolic 87 mmHg (sensitivity 81%/specificity 98%). The update improved the agreement to RefBP. TestBP was excellent for detecting 24 h and awake hypertensive BP values but not for asleep BP values.