Reliability of Oscillometric Pulse Wave Analysis

Abstract

Measurement of central hemodynamic parameters, including central systolic blood pressure (cSBP) and arterial wave reflection (augmentation index, AIx), may provide clinicians with important information beyond that provided by peripheral blood pressure. These parameters can be obtained using oscillometric pulse wave analysis (PWA) devices, which have recently entered the market. Prior to clinical implementation, further study is required to determine measurement precision (reliability), including under conditions which may vary in the clinical setting. PURPOSE: This study sought to examine whether the between-day reliability of oscillometric cSBP and AIx readings: (i) exceed the criterion for acceptable reliability; (ii) are affected by posture (supine, seated) and fasting state. METHODS: Twenty healthy adults (50% F, 27.9 y, 24.2 kg/m2) were tested on six different mornings: three days fasted, three days non-fasted. On each occasion participants were tested in supine and seated postures, resulting in a total of 12 measurements per person, and a total of 240 data points. Oscillometric pressure waveforms were recorded on left upper arm using the SphygmoCor XCEL (AtCor Medical, Sydney, Australia). RESULTS: Repeated measures analysis of variance indicated a non-significant interaction effect (posture X fasting state) for cSBP (P = 0.479) and AIx (P = 0.207). For cSBP there was non-significant main effect for fasting state (P = 0.819) but there was a main effect for posture (P = 0.002). Conversely, for AIx there was non-significant main effect for posture (P = 0.537) but there was a large main effect for fasting state (P = <0.001). The criterion intra-class correlation coefficient value of 0.75 was exceeded for both variables when participants were assessed under the combined supine-fasted condition. For cSBP the reliability coefficient (RC) was lowest (best) when supine-fasted (6.8 mmHg) and greatest (worst) when seated-non-fasted (8.6 mmHg). For AIx, the RC was lower for the supine (11.4 - 11.7 %) compared to seated (14.0 - 15.2 %) posture. CONCLUSIONS: Findings from this study suggest that oscillometric assessments of central hemodynamic variables: (i) exceed the criterion for acceptable reliability; (ii) are most reliable when participants are evaluated while supine and fasted.