Visual assessment of korotkoff sounds further improves the accuracy of a validated professional automated auscultatory blood pressure monitor - inbody bpbio480kv

Objective: A novel automated auscultatory upper-arm cuff blood pressure (BP) monitor for office use (InBody BPBIO480KV, Seoul, S Korea) has been recently validated, which allows healthcare professionals to visually review the Korotkoff sounds curves generated for each BP reading. This study investigated whether this feature could further improve the accuracy of the device by identifying unreliable BP readings. Design and method: Three observers assessed independently the morphology of Korotkoff sound curves generated during a validation study according to the ISO 81060-2:2018 standard, and classified them as having good, fair, or poor signal quality (low signal amplitude or sound intensity, aberrant signal morphology, interference of background noise or other signal artifact, presence of auscultatory gap, or irregular rhythm), separately for the systolic and diastolic BP area. The observers were blinded to the automated and the reference BP measurement results. Disagreement among observers was resolved by a supervisor. Factors associated with poor quality signal were also investigated. Results: A total of 255 Korotkoff sound curves (test BP readings) by 85 participants (mean age 57.3±15.0 years, 53 men) were analyzed. For systolic BP, 82.0/12.2/7.5% were classified as good/fair/poor quality and for diastolic 76.9/12.2/11.0%. Inter-observer agreement in detecting poor quality curves was 84.7/83.1% (systolic/diastolic). Of poor quality curves 10.5/60.7% (systolic/diastolic) clustered in the same individuals. Young age and female sex increased the odds for poor quality signal. ISO validation Criterion 1 (mean BP difference ±SD) was satisfied for readings with good (0.1±4.9/0.3±3.8 mmHg, systolic/diastolic) and fair (-0.4±6.4/0.2±5.0) quality curves, good/fair quality curves (0.1±5.1/0.2±4.0 mmHg, systolic/diastolic), but not for those with poor quality signal (2.7±8.8/3.6±8.1 mmHg). In the ISO validation study, excluding readings with poor quality signal (systolic and/or diastolic area; 40 of 255) Criterion 1 was improved (0.2±4.9/0.2±3.9 mmHg versus 0.3±5.5/0.6±4.7 mmHg). Conclusions: The visual assessment of Korotkoff sounds generated by automated auscultatory BP measurement using the InBody BPBIO480KV professional office BP monitor is useful in identifying unreliable readings and further improves the device accuracy.