Yuan-Ta Shih

Quantification of the Calibration Error in the Transfer Function-Derived Central Aortic Blood Pressures

BACKGROUND The accuracy of the central aortic systolic (SBP-C) and pulse (PP-C) blood pressures estimated noninvasively by a generalized transfer function technique has been questioned. The purpose of the present study was to quantify precisely the impact of the input errors (differences between the oscillometric (SBP-O, DBP-O, PP-O) and invasive (SBP-B, DBP-B, PP-B) brachial systolic, diastolic, and pulse blood pressures) on the output errors (differences between the estimated and invasively measured SBP-C and PP-C). METHODS Invasive high-fidelity right brachial and central aortic pressure waveforms, and SBP-O, DBP-O, and PP-O (=SBP-O - DBP-O) were obtained simultaneously in 40 patients during cardiac catheterization. A generalized transfer function was applied on the individual brachial pressure waveform to derive predicted SBP-C and PP-C. RESULTS Observed input errors were -2.3 ± 5.8 mm Hg from SBP-O, 8.1 ± 5.3 mm Hg from DBP-O, and -10.4 ± 7.1 mm Hg from PP-O, respectively. The output errors were -2.2 ± 6.4 mm Hg and -10.3 ± 8.0 mm Hg for SBP-C and PP-C, respectively, when the brachial pressure waveforms were recalibrated using SBP-O and DBP-O. The outputs were determined by the inputs according to the Equation (1): SBP-C error = 0.97 × SBP-O error + 0.03 (r = 0.88, P < 0.01); and the Equation (2): PP-C error = 0.96 × PP-O error - 0.30 (r = 0.86, P < 0.01). CONCLUSIONS Noninvasive application of the generalized transfer function techniques produces estimates of SBP-C and PP-C with errors equivalent to those of the oscillometric blood pressure monitor in the estimation of SBP-B and PP-B. The output errors can be predicted from input errors of SBP-O and DBP-O.

Measurement Accuracy of a Stand-Alone Oscillometric Central Blood Pressure Monitor: A Validation Report for Microlife WatchBP Office Central

BACKGROUND The superiority of prognostic value of blood pressure (BP) measured at central aorta (CBP) over conventional brachial BP measured by cuff-based BP monitors has reignited the development of new noninvasive techniques for estimating CBP. The present study validated the accuracy of CBP measured by a newly developed stand-alone CBP monitor. METHODS The CBP monitor provided readings of brachial systolic BP (SBP), brachial diastolic BP (DBP), central SBP, and central pulse pressure (PP). Brachial PP and central DBP were calculated from the relevant readings. The accuracy of the brachial and central SBP, PP, and DBP was validated against the simultaneously recorded invasively measured central aortic SBP, PP, and DBP, according to the invasive standard requirements for the noninvasive brachial BP monitors from the Association for the Advancement of Medical Instrumentation (AAMI) in 85 subjects (255 measurements; age range, 30-93 years). RESULTS The mean differences of cuff BP with reference to the invasively measured central SBP, PP, and DBP were -2.6 ± 9.0, -8.6 ± 11.2, and 6.1 ± 7.0 mm Hg, respectively, with the former two being obviously underestimated at high CBP and overestimated at low CBP. In contrast, the corresponding differences for the central SBP, PP, and DBP measured by the CBP monitor were -0.6 ± 5.5, -0.4 ± 7.0, and -0.2 ± 6.5 mm Hg, respectively, without obvious systematic bias. The distribution of measurement errors for central SBP, PP, and DBP surpassed the AAMI criteria. CONCLUSION Central SBP, PP, and DBP can be measured accurately by a stand-alone automatic BP monitor.

Measurement of central aortic pulse pressure: noninvasive brachial cuff-based estimation by a transfer function vs. a novel pulse wave analysis method.

BACKGROUND The prognostic value of central aortic pulse pressure (PP-C) may have been underestimated due to its measurement inaccuracy. We aimed to investigate the accuracy of noninvasive brachial cuff-based estimation of PP-C by a generalized transfer function (GTF) or a novel pulse wave analysis (PWA) approach to directly estimate PP-C. METHODS Invasive high-fidelity right brachial and central aortic pressure tracings, and left brachial pulse volume plethysmography (PVP) waveforms from an oscillometric blood pressure (BP) monitor were all digitized simultaneously in 40 patients during cardiac catheterization. An aortic-to-brachial GTF and a PWA multivariate prediction model using the PVP waveforms calibrated to brachial cuff systolic BP (SBP) and diastolic BP(DBP) were constructed. Accuracy of the two methods was examined in another 100 patients against invasively measured PP-C. RESULTS The error of cuff PP in estimating PP-C was 1.8 ± 12.4 mm Hg. Application of the GTF on noninvasively calibrated PVP waveforms produced reconstructed aortic pressure waves and PP-C estimates with errors of -3.4 ± 11.6 mm Hg (PP-C = reconstructed aortic SBP - aortic DBP) and -2.3 ± 11.4 mm Hg (PP-C = reconstructed aortic SBP - cuff DBP), respectively. The observed systematic errors were proportional to the magnitudes of PP-C. In contrast, the error of the PWA prediction model was 3.0 ± 7.1 mm Hg without obvious proportional systematic error. CONCLUSIONS Large random and systematic errors are introduced into the PP-C estimates when PP-C is calculated as the difference between the estimated central SBP and central or cuff DBP. The accuracy can be improved substantially with the novel PWA approach.

Measurement of Central Systolic Blood Pressure by Pulse Volume Plethysmography With a Noninvasive Blood Pressure Monitor

BACKGROUND Central systolic blood pressure (SBP) can be estimated by an oscillometric method developed from a pulse volume plethysmography (PVP) device. The present study applied this novel method to a noninvasive blood pressure monitor (NBPM). METHODS We enrolled 50 patients (37 men, age range 30-84 years) referred for cardiac catheterization. Invasive right brachial and central aortic pressures (using a dual-sensor pressure catheter), and noninvasive left brachial SBP and diastolic blood pressure (DBP), and PVP waveform (using a customized NBPM) were simultaneously recorded. Central SBP was estimated by analysis of the PVP waveform calibrated to the noninvasive SBP and DBP, using both the original (CSBP-O) and the newly generated (CSBP-N) regression equations. The reproducibility of the invasive central SBP by CSBP-O and CSBP-N was examined using the concordance correlation coefficient. RESULTS Overall, the invasive central aortic SBP ranged 86-176 with a mean of 124 ± 21 mm Hg. The mean differences between the estimated and the invasive central SBP were -1.3 ± 6.7 mm Hg for CSBP-O and 0.0 ± 6.2 mm Hg for CSBP-N, respectively. The concordance correlation coefficients for CSBP-O and CSBP-N were 0.94 (95% confidence interval (CI): 0.93-0.94) and 0.95 (95% CI: 0.95-0.96), and both were significantly better than that for the noninvasive brachial SBP (0.87, 0.84-0.91) indicated by non-overlapping CIs. CONCLUSIONS The PVP method for noninvasive estimation of central SBP can be applied to a commonly used NBPM. Whether the NBPM-derived central SBP is superior to the noninvasive brachial SBP in the prediction of cardiovascular risks remains to be investigated.

P-004 PULSE PRESSURE AMPLIFICATION OVER UPPER ARM: THE ROLE OF LOCAL WAVE REFLECTION

Objectives We decomposed the lower body and upper limb reflections from the brachial PW to quantify their relative contributions to the aorta-to-brachial Pulse pressure amplification (PPA). Background PPA between central aorta and brachial artery is affected by wave reflections from both lower body and local upper limb. Methods High fidelity central aortic and right brachial PWs were simultaneously recorded at baseline and 3 minutes after sublingual nitroglycerin (NTG) administration, using a custom-made dual pressure sensor catheter in 54 patients (66.2 ± 13.0 year; 45 men). Lower body reflected PW (amplitude = Pb) was decomposed from the aortic PW. Local upper limb reflected PW (amplitude = P0) was obtained by subtraction of the aortic PW from the foot-aligned brachial PW. Aorta-to-brachial pulse wave velocity (abPWV) was calculated from the simultaneously recorded aortic and brachial PWs. Results PPA was 5.4 ± 5.9 mmHg and increased to 10.9 ± 6.1 mmHg after NTG. Pb decreased by 39.6 ± 18.7 % and P0 by 6.7 ± 29.0 % after NTG. By forward multivariate linear regression analysis, the total variance (r2 = 0.76) of PPA could be explained by P0 (54.9%), Pb (33.0%), total arterial compliance (5.7%), abPWV (3.4%), and reflected wave transit time (3.0%), whereas the total variance (r2 = 0.53) of P0 could be explained by abPWV (57.3%), Pb (37.2%), and heart rate (5.5%). Conclusions Local wave reflection contributes substantially to the aorta-to-brachial PPA, an emerging mechanical biomarker of cardiovascular risk.

E-005 ESTIMATION OF CENTRAL PULSE PRESSURE FROM PERIPHERAL PRESSURE WAVEFORMS: CALIBRATION MATTERS

Objectives The aim of this study was to identify the major source of errors and to improve the accuracy for the noninvasive estimation of central aortic pulse pressure (PP). Background The clinical value of central PP may have been underestimated due to the inaccuracy of its measurement. Methods Invasive high-fidelity right brachial and central aortic pressure tracings and left-brachial pulse volume plethysmography (PVP) waveforms from an oscillometric blood pressure monitor were all digitized simultaneously in 40 patients during cardiac catheterization. Invasive brachial pressure waveforms were used to estimate the central PP by both the generalized transfer function (GTF) and the pressure waveform analysis (PWA) approaches. A noninvasive PWA prediction model was also constructed using the PVP waveforms calibrated by the oscillometric brachial pressures. Results Errors between the central PP and invasive estimates by GTF and PWA were −0.8 ± 3.8 mmHg and 0.0 ± 3.5 mmHg, respectively. However, the oscillometric brachial PP underestimated the invasive brachial PP by −7.4 ± 11.7 mmHg. When the brachial pressure waveforms were purposefully calibrated using oscillometric brachial pressures, the errors of the estimation of central PP were markedly inflated (−7.3 ± 8.4 mmHg by GTF, and −11.7 ± 8.6 mmHg by PWA, respectively). In contrast, the error of the noninvasive PVP model was 0.0 ± 6.7 mmHg. Conclusions Large errors are introduced into the central PP estimates from the oscillometric brachial pressures. The accuracy may be improved when the oscillometric pressures calibrated peripheral pressure waveforms are used in the construction of the prediction models.

E-004 ESTIMATION OF CENTRAL AORTIC BLOOD PRESSURE USING A BRACHIAL OSCILLOMETRIC PHLETHYSMOGRAPHIC WAVEFORM AND A GENERALIZED TRANSFER FUNCTION

Background Central aortic systolic (SBP-C) and pulse (PP-C) blood pressures can be accurately estimated from an invasively derived brachial pressure waveform and an aortic-to-brachial generalized transfer function (A2BGTF). The purpose of the present study was to investigate if the A2BGTF or a generalized transfer function (A2PGTF) between the central aortic pressure waveform and a brachial pulse volume plethysmography (PVP) waveform calibrated to the noninvasive brachial systolic and diastolic blood pressures can be used to estimate SBP-C and PP-C from a PVP waveform. Methods A2BGTF and A2PGTF were generated from 40 patients in whom aortic and brachial pressure waveforms recorded by a high-fidelity dual pressure sensor catheter, and a PVP waveform recorded by an automatic oscillometric blood pressure monitor, were obtained simultaneously during cardiac catheterization. The A2BGTF and A2PGTF were then validated in another 100 patients in whom aortic and brachial pressure waveforms were recorded by a high-fidelity single pressure sensor catheter, and PVP waveform was recorded by another automatic oscillometric blood pressure monitor. Results Mean differences between the estimated and recorded SBP-C and PP-C were −0.6 ± 7.7 mmHg and −8.7 ± 8.7 mmHg, respectively, when using the A2BGTF; and −0.7 ± 7.8 mmHg and −1.3 ± 8.6 mmHg, respectively, when using the A2PGTF. Bland-Altman analysis showed a systematic bias in the estimation of PP-C when using either A2BGTF or A2PGTF. Conclusion Both A2BGTF and A2PGTF can be used to estimate SBP-C but not PP-C from a noninvasive blood pressure monitor-derived PVP waveform.

Estimation of Central Aortic Blood Pressure using a Noninvasive Automatic Blood Pressure Monitor

Automatic blood pressure monitors are the most important healthcare products. The abnormal blood pressure values and waveform could indicate many cardiovascular diseases. The pressures measured using oscillometric method, which compared with invasive blood pressure devices, are often slight overestimation in SBP and underestimation in DBP.Recent studies indicated that the continuous pulse volume recording (PVR) could predict the central aortic blood pressure using a transfer function. Therefore, this study was implementing a novel device using the oscillometric method to predict the central aortic blood pressure to measure central blood pressure as well as the SBP, MBP and DBP automatically. The ensemble averaging 10 heart-beats of continuous PVR was used to extract the value of (cSBP), central SBP2 (cSBP2), central pulse pressure (cPP) and incisura point (inciP) for central blood pressure calculation. The method estimating the central blood pressure was correlated with the invasive measurement of pressure at aortic root at catheter Lab with high accuracy.