Two fully automated oscillometric devices have become available for the noninvasive assessment of central aortic blood pressure (BP). They tend, however, to underestimate SBP. It has been proposed that calibration by mean/diastolic instead of systolic/diastolic brachial BP may reduce this bias. The present work compares the accuracy of these two calibrations in the Mobil-O-Graph.
Post-hoc analysis of the largest validation study on noninvasive assessment of central BP so far. Data on both calibration approaches were available in 159 patients without atrial fibrillation, who underwent simultaneous invasive and noninvasive assessment of central BP. Noninvasive BP measurements were conducted using the SphygmoCor XCEL (calibration by systolic/diastolic brachial BP only) and the Mobil-O-Graph (calibration by both systolic/diastolic and mean/diastolic brachial BP).
Measurements of both devices and both calibration methods revealed highly significant correlations for systolic and diastolic central BP with invasively assessed BP (P < 0.001 each). Calibration by mean/diastolic and systolic/diastolic BP yielded similar correlations for central DBP (R2 0.56 vs. R2 0.55, P = 0.919). Correlation of central SBP, however, was significantly lower using calibration by mean/diastolic brachial BP (R2 0.86 vs. R2 0.74, P = 0.002). Numerically, the SphygmoCor device revealed the highest correlation (R2 0.92 for central SBP and 0.72 for central DBP; P < 0.001 each). Calibration by systolic/diastolic brachial BP was associated with an underestimation of central SBP using both the SphygmoCor and the Mobil-O-Graph. Calibration by mean/diastolic brachial BP, instead, was associated with an overestimation, which was numerically comparable (4.8 ± 11.3 vs. −4.2 ± 8.0). The calibration method had little effects on the biases of diastolic measurements.
Calibration by mean/diastolic instead of systolic/diastolic brachial BP led to an overestimation instead of underestimation of central SBP without improving accuracy. Hence, mean/diastolic calibration is not necessarily superior to systolic/diastolic calibration and the optimal approach has to be determined in a device-specific manner.