Aortic pressure and forward and backward wave components in children, adolescents and young-adults: Agreement between brachial oscillometry, radial and carotid tonometry data and analysis of factors associated with their differences.

Non-invasive devices used to estimate central (aortic) systolic pressure (cSBP), pulse pressure (cPP) and forward (Pf) and backward (Pb) wave components from blood pressure (BP) or surrogate signals differ in arteries studied, techniques, data-analysis algorithms and/or calibration schemes (e.g. calibrating to calculated [MBPc] or measured [MBPosc] mean pressure). The aims were to analyze, in children, adolescents and young-adults (1) the agreement between cSBP, cPP, Pf and Pb obtained using carotid (CT) and radial tonometry (RT) and brachial-oscillometry (BOSC); and (2) explanatory factors for the differences between approaches-data and between MBPosc and MBPc.1685 subjects (mean/range age: 14/3-35 y.o.) assigned to three age-related groups (3-12; 12-18; 18-35 y.o.) were included. cSBP, cPP, Pf and Pb were assessed with BOSC (Mobil-O-Graph), CT and RT (SphygmoCor) records. Two calibration schemes were considered: MBPc and MBPosc for calibrations to similar BP levels. Correlation, Bland-Altman tests and multiple regression models were applied. Systematic and proportional errors were observed; errors´ statistical significance and values varied depending on the parameter analyzed, methods compared and group considered. The explanatory factors for the differences between data obtained from the different approaches varied depending on the methods compared. The highest cSBP and cPP were obtained from CT; the lowest from RT. Independently of the technique, parameter or age-group, higher values were obtained calibrating to MBPosc. Age, sex, heart rate, diastolic BP, body weight or height were explanatory factors for the differences in cSBP, cPP, Pf or Pb. Brachial BP levels were explanatory factors for the differences between MBPosc and MBPc.

Introduction

The independent prognostic value of central aortic blood pressure (cBP) and aortic wave-derived parameters has been demonstrated[1-6]. That, together with the growing interest towards improving risk estimates contributes to explain the explosive development in the last decades, of methods and devices aiming at providing cBP, aortic wave components (e.g. forward and backward aortic pressure wave amplitude, Pf and Pb respectively) and/or derived parameters levels[3-7]. Available non-invasive devices estimate central systolic BP (cSBP) from pressure or surrogate signals obtained from peripheral arteries (e.g. carotid, brachial or radial), recorded using a variety of techniques (e.g. applanation tonometry, brachial oscillometry). From the obtained signals, and after calibrating them, the devices quantify cSBP directly (e.g. direct calibration of carotid waves) and/or indirectly, applying generalized transfer functions (GTF), low-pass filters or wave analysis[8,9]. Due to the differences in sites of measurement, signals recorded and in the methodological approaches used to assess central hemodynamics, data obtained from different approaches could differ and inter-device agreement would vary, depending on the parameter considered[8,9]. Although these approaches are already used in research involving children and adolescents [10-12], the extent to which they provide similar data on cBP and/or wave-derived parameters is unknown.

It is recognized that the pressure waveform is modified and the aortic or central pulse pressure (cPP) is amplified towards the periphery[13]. The centre-periphery changes in pressure wave are associated with age. The differences between cSBP and peripheral systolic blood pressure (pSBP)are greater in young subjects than in old adults, and may be particularly important in children and adolescents. Then, the relationship between central and peripheral hemodynamic parameters would vary depending on subjects´ age. This could affect the accuracy of estimating central parameters from peripheral data obtained with a methodological approach and/or the agreement between devices or methods. Hence, it would be valuable to analyze the agreement of methods used to estimate cBP and wave derived parameters considering data from subjects of different ages. It would be interesting to identify other subjects´ characteristics (e.g. demographic, anthropometric) that could contribute to explain the degree of agreement between data from different approaches.

An additional relevant issue to evaluate is the extent to which wave-derived parameters and/or cBP values depend on the calibration scheme considered: (1) calibrating to mean blood pressure (MBP) measured by oscillometry (MBPosc) or (2) calibrating to MBP calculated (MBPc), obtained from pSBP and peripheral diastolic blood pressure (pDBP). To know this, as well as to clarify which variables may influence the difference between MBPc and MBPosc would be valuable at the time of assessing and analyzing accurately cBP levels and/or wave-derived parameters.

This work aims were: 1) to determine the agreement of cSBP, cPP, Pf and Pb data obtained in children, adolescents and young adults using different methodological approaches; 2) to analyse subjects´ characteristics that could contribute to explain the differences (a) between methods used to assess central hemodynamic parameters and (b) between MBPosc and MBPc.

Materials and methods

Study population

In this work we considered data from a total of 1685 subjects (mean/range age: 14.4/3-35 y.o.; 854 females). The subjects´ records are part of CUiiDARTE Database, which includes data from longitudinal (cohort)and cross-sectional studies developed in Uruguay from February 2012until July 2019[10,14-16]. The subjects or their families were selected (random sampling), mainly from their reference health institutions, educational and/or work centres, and were invited to participate through personal interviews. Included subjects were part of cohorts representative of their corresponding populations (e.g. children cohort, adolescents cohort) [16], but the whole group included in this work could not be considered (in rigorous terms) as representative of the entire Uruguayan population. Interviews, anthropometrical measurements and cardiovascular evaluations were performed in the ambulatory and/or office non-invasive vascular laboratories of CUiiDARTE. Subjects included in this work met the following criteria at the time of the evaluation: (a) all were asymptomatic and in stable clinical conditions, (b) none had congenital, chronic or infectious diseases, and (c) none was taking vasoactive drugs. Exclusion criteria included rhythm other than sinus rhythm and valvular heart disease.

All procedures agreed with the Declaration of Helsinki (1975; reviewed in 1983). The study protocol was approved by Institutional Ethic Committee (Comité de Ética en Investigación del Centro Hositalario Pereira-Rossell). Written informed consent was obtained from participants or from parents in case of subjects aged <18y.o., who gave informed assent before data collection.

Clinical interview and anthropometric measurements

Before vascular evaluation a brief clinical interview together with the anthropometric evaluation enabled to evaluate the exposure to cardiovascular risk factors (CRFs). Subjects’ body weight (BW) and height (BH) were measured and body mass index (BMI) obtained as BW-to-squared BH ratio. For subjects <18 y.o. BMI was converted into z-scores[17]. Obesity was defined as a BMI z-score ≥2 for subjects aged <18 y.o. and as BMI≥30kg/m2 for subjects ≥18 y.o. Regular smokers (defined as usually smoking at least one cigarette per week) were identified. Sedentary life style was considered present if the subject´s physical activity (PA) was less than the recommended in terms of frequency and/or intensity[18]. To assess this, we applied questionnaires asking about characteristics of the developed PA (e.g. duration, frequency, pattern, type, intensity). Following the World Health organization (WHO) guidelines, 3 or 4 y.o. children were considered active when spending ≥180 min/day in a variety of PAs that involved different intensities; of which ≥60 min/day corresponded to moderate-to-vigorous PA. Subjects aged 5–17 y.o. that did not accumulate ≥60 min/day of moderate-to-vigorous intensity PA were considered sedentary. The concept of accumulation refers to meeting the goal of 60 min/day by performing activities in multiple shorter bouts spread throughout the day (e.g. 2 bouts of 30 minutes). Adults (≥18 y.o.) who performed ≥150 min/week of moderate-intensity aerobic PA or ≥75 min/week of vigorous-intensity aerobic PA were considered active. There could be multiple ways to reach the total of 150 min/week. The concept of accumulation refers to meeting the goal of 150 min/week by performing activities in multiple shorter bouts, of ≥10 minutes each, spread throughout the week. Dyslipidemia, diabetes and history of hypertension or high BP levels (HBP) was considered present if it had been previously diagnosed by referring physicians[19-21]. Subjects <16 y.o. who had pSBP and/orpDBP > 95th percentile for sex, age and BH during the study, were considered with hypertensive BP levels (disregard of previous diagnosis of hypertension). For subjects aged ≥16 y.o., hypertensive BP levels were defined using cutoff values similar to those for adults (pSBP≥140 mmHg and/or pDBP≥90 mmHg)[19-21].

Central blood pressure and wave components levels

Participants were asked to avoid exercise, tobacco, alcohol, caffeine and food-intake four hours before evaluation. All haemodynamic measurements were performed in a temperature-controlled environment (21–23°C), with the subject in supine position and after resting for at least 10–15 minutes. Heart rate (HR) and brachial pSBP and pDBP were recorded in supine position using the validated oscillometric device (HEM-433INT; Omron Healthcare Inc., Illinois, USA) simultaneously and/or immediately before or after each non-invasive tonometric(radial and carotid applanation tonometry [RT and CT], respectively] and brachial oscillometry [BOSC]) recording. Peripheral pulse pressure (pPP; pPP = pSBP–pDBP) and MBPc (MBPc = pDBP+pPP/3) were obtained.

Central BP and wave components (Pf and Pb) were assessed (random order) using two commercially available devices: SphygmoCor-CvMS (SCOR; v.9, AtCor-Medical, Australia) and Mobil-O-Graph PWA-monitor system(MOG; I.E.M.-GmbH, Stolberg, Germany) [Fig 1][9,10,11]. Both devices and systems enable doing pulse wave analysis (PWA) and wave separation analysis (WSA)[6,11,12,14].

Fig 1

Instrumental approach employed to obtain aortic blood pressure and wave components.

CCA: common carotid artery. BA: brachial artery. RA: radial artery.

Radial and carotid pressure waves were obtained by applanation tonometry with SCOR. The acquired waves were calibrated toMBPc and pDBP(HEM-433INT; Omron Healthcare Inc., Illinois, USA). Central BP waves were derived from radial recordings (using a GTF) and cSBP and cPP were quantified. Carotid artery pulse waves were assumed to be identical to the aortic ones (due to the proximity of the arterial sites). Thus, a GTF was not applied to obtain central waves from carotid records. Considering a triangular flow model (using WSA), Pf and Pb components of the obtained aortic waves were separated [2]. Only accurate waveforms on visual inspection and high-quality recordings (in-device quality index>75%) were considered.

Brachial BP levels and waveforms were obtained using the MOG (brachial cuff-based oscillometric device, BOSC)[21]. The device determined cBP levels and waveforms from peripheral recordings using a validated GTF. Then, by means of PWA and WSA, Pf and Pb were obtained[10,14]. Only high quality records (index equal to 1 or 2) and satisfactory waves (visual inspection) were considered. A step-by-step explanation of the method used to carry out WSA based on recorded (carotid wave, SCOR) and mathematically-derived aortic waveform (SCOR and MOG) was included as Supplementary Material (S1 Appendix). Absolute and relative intra (repeatability) and inter-observer (reproducibility) variability of cSBP, cPP, Pf and Pb was evaluated [Supplementary Material, S1 Appendix]. No significant differences were observed in cSBP, cPP, Pf or Pb absolute levels either within each visit, between two records or between records obtained by investigators; indicating excellent repeatability, as well as reproducibility. In all cases, the relative inter- and intraobserver variability was <6%.

Data and statistical analysis

A stepwise data analysis was done. First, ANOVA plus Bonferroni post-hoc tests were done to compare (for each age-related group) mean values obtained with the different methods (RT vs. CT vs. BOSC). Second, the relationships between cSBP, cPP, Pf and Pb data obtained with the different approaches (RT, CT and BOSC) were assessed (correlation analyses). Third, Bland-Altman analyses were performed to evaluate the agreement (equivalence) between methods. Bland-Altman plots correspond to the mean of the methods considered (x-axis; e.g. RT and CT mean) against their difference (y-axis; e.g. RT minus CT). The corresponding linear regression equations were obtained. Systematic error (bias) was considered present if mean error was significantly different from zero, whereas proportional error was considered present if the slope of the linear regression was statistically significant. Fourth, multiple linear regression models (MLR; stepwise method) were considered to analyze the association between the differences in cSBP, cPP, Pf and Pb (ΔcSBP, ΔcPP, ΔPf and ΔPb, respectively; absolute values) between methods [dependent variable] and age, sex, BH, BW, pDBP and HR[independent variables]. Fifth, the degree of equivalence between MBPoscandMBPc was analyzed (correlation and Bland-Altman test), and explanatory variables for the differences were identified (MBPosc minus MBPc; correlation analysis and MLR model[enter and stepwise method]). After identifying independent variables (those with p<0.01 in bivariate analysis) two different MLR models were tested. Model 1: independent variables were all significant variables in bivariate analysis (age, sex, BH, BW, BMI, HBP, pPP and pSBP; pDBP was excluded due to multicollinearity). Model 2: independent variables were pDBP and pSBP (pPP was excluded due to multicollinearity). In all MLR analyses, a variance inflation factor (VIF) <5 was selected to evaluate (discard) significant collinearity among variables.

The described analysis was done considering all the studied subjects (entire group), as well as three age-related groups: children: 3–12 y.o. (n = 728), adolescents: 12–18 y.o. (n = 361) and young adults: 18–35 y.o. (n = 596). The analysis was done calibrating peripheral signals (carotid, radial and brachial)to pDBP and MBPc data obtained at the time of signals recording ("instantaneous blood pressure"). Thereafter, the analysis was done considering carotid, radial and brachial signals calibrated to the same pressure levels, taking into account data obtained from the Mobil-O-Graph and different calibration methods: pDBP/MBPc and pDBP/MBPosc (subsample). This analysis enabled to evaluate the differences in cSBP, cPP, Pf and Pb data obtained with the different methods (RT, CT, BOSC) taking into account (i.e. with independence of)potential differences in pressure values considered for calibration and/orin MBP data used to calibrate signals (i.e. MBPc vs. MBPosc).

According to the central limit theorem, a normal distribution was considered (considering Kurtosis and Skewness coefficients distribution and number of subjects, with sample size >30)[22]. Data analyses were done using MedCalc (v.14.8.1, MedCalc Inc., Ostend, Belgium) and IBM-SPSS Statistical Software (v.20, SPSS Inc., Illinois, USA). A p value<0.05 was considered statistically significant.

Results

Agreement between cBP data obtained from RT, CT and BOSC

Table 1 (and S1 Table) shows characteristics of the studied subjects.

Table 1

Clinical features and cardiovascular risk factors for the entire and age-related subgroups.

	Entire group[3–35 y.o; n = 1685]		Children[3–12 y.o; n = 728]		Adolescents[12–18 y.o; n = 361]		Young adults[18–35 y.o; n = 596]
	MV	SD	MV	SD	MV	SD	MV	SD
Age (years)	14.45	7.5	7.2	1.9	15.4	2.1	22.6	4.7
Sex female, n (%)	854	[50.7]	332	[45.6]	175	[48.5]	347	[58.1]
Bodyheight (m)	1.47	0.24	1.23	0.13	1.63	0.10	1.68	0.09
Bodyweight (kg)	49.0	22.6	29.2	12.9	61.8	16.5	66.2	14.1
BMI (Kg./m2)	21.3	5.0	18.7	4.2	23.1	5.2	23.4	4.0
z-BMI* (kg/m2)	1.30	1.88	1.36	1.88	1.07	1.88	-	-
Hypertension and/or HBP, n [%]	206	[12.2]	85	[11.7]	54	[15]	67	[11.3]
Diabetes, n [%]	7	[0.4]	4	[0.6]	3	[0.8]	0	[0]
Dyslipidemia, n [%]	113	[6.7]	35	[4.8]	23	[6.4]	55	[9.2]
Obesity, n [%]	304	[18.7]	204	[28.5]	60	[17.6]	40	[7]
Smoking, n [%]	145	[8.8]	0	[0]	14	[3.9]	131	[23.3]
Family history of CV disease [%]	0	[0]	0	[0]	0	[0]	0	[0]
Sedentarylifestyle, n [%]	512	[36.7]	159	[22.4]	146	[47.9]	207	[54.3]

MV: mean value. SD: standard deviation. BMI: body mass index. z-BMI:

*z-score of BMI calculated only for under 18 years old (y.o.). HBP: high blood pressure state. CV: cardiovascular.

Table 2 (and S2 Table) shows peripheral and central pressure, Pf, Pb and HR levels, calibrating data to pDBP/MBPcvalues registered at the time of RT and CT assessment(device: HEM-433INT) or during cBP measurement using BOSC (device: Mobil-O-Graph, self-calibration). Within each group there was a wide-range of BP and HR values, which enabled analyzing different hemodynamic states. There was an age-related increase in pSBP, pDBP and pMBPcvalues used to calibrate RT, CT and BOSC[Table 2, Fig 2].

Table 2

Haemodynamic and aortic wave-derived parameters measured with three different methods in the entire and age-related groups.

Entire group [3–35 y.o.; n = 1685]
	RT (Scor)		CT (Scor)		BOSC (MOG)		P value
	MV	SD	MV	SD	MV	SD	RT vs CT	RT vs OSC	CT vs OSC
pSBP (mmHg)	114.7	12.9	114.4	14.3	113.3	12.0	1.00	0.04	0.21
pDBP (mmHg)	64.1	8.7	62.8	8.1	63.7	8.0	0.001	0.88	0.06
MBPc (mmHg)	80.8	9.2	80.0	8.8	80.0	8.5	0.08	0.08	1.00
HR (beats/minute)	75.8	14.2	75.5	14.4	78.9	15.2	1.00	<0.001	<0.001
cSBP (mmHg)	98.6	12.3	105.9	15.2	99.7	14.4	<0.001	0.19	<0.001
cPP (mmHg)	33.3	9.8	43.0	13.7	34.6	11.6	<0.001	0.04	<0.001
Pf (mmHg)	31.1	10.0	43.2	14.2	23.5	7.8	<0.001	<0.001	<0.001
Pb (mmHg)	13.3	4.0	15.4	4.7	13.2	5.1	<0.001	1.00	<0.001
Children [3–12 y.o.; n = 728]
pSBP (mmHg)	104.8	10.0	104.7	11.7	106.7	1.0	1.00	0.01	0.01
pDBP (mmHg)	60.0	7.4	59.6	6.8	60.3	6.4	1.00	1.00	0.32
MBPc (mmHg)	74.7	7.0	74.6	7.1	75.5	1.0	1.00	0.11	0.12
HR (beats/minute)	85.0	14.0	84.6	0.8	86.9	14.0	1.00	0.25	0.11
cSBP (mmHg)	87.0	8.9	95.4	11.3	89.9	8.8	<0.001	<0.001	<0.001
cPP (mmHg)	24.4	7.3	35.8	10.5	28.3	7.1	<0.001	0.003	<0.001
Pf (mmHg)	24.5	7.6	35.8	10.3	19.5	4.9	<0.001	<0.001	<0.001
Pb (mmHg)	11.4	4.5	12.6	3.6	10.5	3.0	<0.001	0.004	<0.001
Adolescents [12–18 y.o.; n = 361]
pSBP (mmHg)	116.9	10.9	117.3	12.5	118.5	11.1	1.00	0.27	0.72
pDBP (mmHg)	63.1	7.6	62.3	7.9	65.6	7.4	0.60	<0.001	<0.001
MBPc (mmHg)	80.8	8.3	80.6	7.9	83.0	7.7	1.00	0.005	0.003
HR (beats/minute)	73.8	13.6	72.5	12.9	73.2	12.9	0.70	1.00	1.00
cSBP (mmHg)	100.5	9.9	109.6	13.8	107.2	12.1	<0.001	<0.001	0.09
cPP (mmHg)	36.1	9.0	47.3	13.4	40.1	11.7	<0.001	<0.001	<0.001
Pf (mmHg)	34.2	9.0	47.2	15.0	27.3	8.1	<0.001	<0.001	<0.001
Pb (mmHg)	13.3	3.3	15.9	4.7	15.3	5.4	<0.001	<0.001	0.50
Young adults [18–35 y.o.; n = 596]
pSBP (mmHg)	120.3	11.7	120.7	13.1	120.5	10.7	1.00	1.00	1.00
pDBP (mmHg)	67.5	8.7	66.0	8.2	68.4	8.3	0.01	0.62	0.003
MBPc (mmHg)	85.1	8.5	84.2	8.2	85.4	8.3	0.29	1.00	0.27
HR (beats/minute)	71.5	12.0	69.9	11.8	69.4	11.0	0.13	0.08	1.00
cSBP (mmHg)	104.5	10.2	112.1	14.3	110.4	12.7	<0.001	<0.001	0.3215
cPP (mmHg)	36.1	9.5	46.1	13.9	40.8	12.4	<0.001	<0.001	<0.001
Pf (mmHg)	33.5	10.0	45.6	14.0	27.1	8.3	<0.001	<0.001	<0.001
Pb (mmHg)	14.4	3.6	16.9	4.6	16.0	5.6	<0.001	<0.001	0.060

MV: mean value. SD: standard deviation. y.o.: years old. RT and CT: radial and carotid applanation tonometry (SphygmoCor device). BOSC: brachial oscillometry/plethysmography (Mobil-O-Graph device). pSBP, pDBP, MBPc: peripheral systolic, diastolic and mean (calculated) blood pressure. HR: heart rate. cSBP, cPP: central systolic and pulse blood pressure. Pf and Pb: forward and backward wave height. Significance: p<0.05 (ANOVA+Bonferroni post-hoc test).

Fig 2

Haemodynamic parameters obtained for the entire group and age-related groups.

Scor and MOG: SphygmoCor and Mobil-O-Graph. pSBP, pDBP: peripheral systolic and diastolic blood pressure. MBPc: calculated mean blood pressure. HR: heart rate. cSBP, cPP: central systolic and pulse pressure. Pf, Pb: forward and backward wave height. *p<0.05 with respect to Children; +p<0.05 with respect to Adolescents.

Tables 3 and 4 shows data from the analyses of association (correlation) and agreement (Bland-Altman) between methods used to obtain cSBP, cPP, Pfand Pb (calibration: pDBP/MBPc).

Table 3

cSBP, cPP, Pf and Pb: Correlation and agreement among values obtained with three different recording methods (entire group and children).

	Entire group [3–35 y.o.; n = 1685]			Children [3–12 y.o.; n = 728]
	RT- CT	RT -BOSC	CT—BOSC	RT- CT	RT -BOSC	CT—BOSC
cSBP
R	0.82	0.79	0.72	0.73	0.64	0.50
P	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
ME (mmHg)	-8.0	-5.2	3.1	-7.5	-3.9	4.2
ME, p value	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
ME, SD (mmHg)	8.8	9.2	11.2	7.1	7.4	9.8
Regressionequation	y = 12.9–0.2x	y = 9.1–0.1x	y = -5.4+0.08x	y = 10.0–0.2x	y = 7.3–0.1x	y = -11.5+0.2x
p (Slope)	<0.001	<0.001	0.03	<0.001	0.038	0.037
cPP
R	0.66	0.64	0.50	0.57	0.63	0.46
P	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
ME (mmHg)	-9.2	-2.4	7.3	-8.9	-2.9	6.9
ME, p value	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
ME, SD (mmHg)	9.2	8.1	10.8	7.5	6.2	8.7
Regressionequation	y = 2.6–0.3x	y = 0.9–0.09x	y = -0.5+0.2x	y = 0.7–0.3x	y = -2.1–0.03x	y = -2.0+0.3x
p (Slope)	<0.001	0.003	<0.001	<0.001	0.644	0.002
Pf
R	0.57	0.57	0.35	0.54	0.62	0.44
P	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
ME (mmHg)	-10.9	7.1	18.3	-11.0	4.4	15.6
ME, p value	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
ME, SD (mmHg)	10.2	7.5	10.9	8.4	6.0	8.9
Regressionequation	y = 1.9–0.4x	y = -3.5+0.4x	y = -6.2+0.7x	y = -0.3–0.4x	y = -7.9+0.6x	y = -9.2+0.9x
p (Slope)	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
Pb
R	0.69	0.58	0.44	0.39	0.57	0.40
P	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
ME (mmHg)	-1.9	-0.7	1.0	-0.9	0.3	1.3
ME, p value	<0.001	<0.001	<0.001	<0.001	0.093	<0.001
ME, SD (mmHg)	3.8	3.8	4.5	3.5	2.8	3.6
Regressionequation	y = 3.8–0.4x	y = 4.9–0.4x	y = 3.2–0.1x	y = 2.8–0.3x	y = 1.7–0.1x	y = 1.5–0.02x
p (Slope)	<0.001	<0.001	0.006	0.001	0.078	0.880

y.o.: years old. RT, CT: radial and carotid tonometry (SphygmoCor), respectively. BOSC: brachial oscillometry/plethysmography (Mobil-O-Graph). cSBP, cPP: central systolic and pulse pressure, respectively. Pf, Pb: forward and backward wave amplitude, respectively. R: correlation (Pearson) coefficient. ME: mean or systematic error. β: slope of regression equation. Significance: p<0.05. Bland-Altman: "x" was considered the mean of both methods compared (e.g. (RT+CT)/2); "y" the difference among first and second method (e.g. RT minus CT).

Table 4

cSBP, cPP, Pf and Pb: Correlation and agreement among values obtained with three different recording methods (adolescents and young adults).

	Adolescents [12–18 y.o.; n = 361]			Young adults [18–35 y.o.; n = 596]
	RT- CT	RT—BOSC	CT—BOSC	RT- CT	RT—BOSC	CT—BOSC
cSBP
R	0.70	0.60	0.53	0.68	0.47	0.49
P	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
ME (mmHg)	-9.3	-6.5	3.6	-7.4	-5.5	1.5
ME, p value	<0.001	<0.001	<0.001	<0.001	<0.001	0.07
ME, SD (mmHg)	9.80	9.89	12.6	9.9	10.3	11.0
Regressionequation	y = 30.6–0.4x	y = 15.4–0.2x	y = -21.6+0.2x	y = 26.8–0.3x	y = 9.2–0.1x	y = -17.7+0.2x
p (Slope)	<0.001	0.003	0.004	<0.001	0.10	0.045
cPP
R	0.67	0.47	0.49	0.65	0.62	0.52
P	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
ME (mmHg)	-10.7	-2.6	8.6	-8.3	-1.6	6.5
ME, p value	<0.001	<0.001	<0.001	<0.001	0.015	<0.001
ME, SD (mmHg)	10.0	9.0	12.5	9.8	9.1	11.0
Regressionequation	y = 9.6–0.5x	y = 6.5–0.2x	y = -7.7+0.4x	y = 6.4–0.4x	y = 5.0–0.2x	y = 5.2+0.03x
p (Slope)	<0.001	0.001	<0.001	<0.001	0.012	0.724
Pf
R	0.57	0.59	0.44	0.60	0.58	0.47
P	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
ME (mmHg)	-11.8	8.3	20.2	-10.3	9.1	19.0
ME, p value	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
ME, SD (mmHg)	11.0	7.7	12.2	10.7	7.9	10.8
Regressionequation	y = 10.4–0.6x	y = 2.2+0.2x	y = -9.8+0.8x	y = 4.9–0.4x	y = 1.4+0.3x	y = -5.1+0.7x
p (Slope)	<0.001	0.007	<0.001	<0.001	<0.001	<0.001
Pb
R	0.56	0.56	0.40	0.57	0.54	0.41
P	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
ME (mmHg)	-2.2	-1.6	0.6	-2.4	-0.9	1.2
ME, p value	<0.001	<0.001	0.15	<0.001	0.002	0.002
ME, SD (mmHg)	3.8	4.1	4.9	3.8	4.3	4.9
Regressionequation	y = 4.3–0.4x	y = 5.1–0.5x	y = 2.4–0.1x	y = 3.5–0.4x	y = 7.3–0.6x	y = 5.5–0.3x
p (Slope)	<0.001	<0.001	0.283	<0.001	<0.001	0.008

y.o.: years old. RT and CT: radial and carotid tonometry (SphygmoCor), respectively. BOSC: brachial oscillometry/plethysmography (Mobil-O-Graph). cSBP, cPP: central systolic and pulse pressure, respectively. Pf, Pb: forward and backward wave amplitude, respectively. R: correlation (Pearson) coefficient. β: slope of regression equation. ME: mean or systematic error. Significance: p<0.05. Bland-Altman: "x" was considered the mean of both methods compared (e.g. (RT+CT)/2); "y" the difference among first and second method (e.g. RT minus CT).

Mean error data are shown in Fig 3, ordered by means of the obtained values. Additional information about the described analyses can be found in Supplemental Material [S3–S7 Tables; S1–S4 Figs]. cSBP, cPP, Pf and Pb data from the different methodological approaches were positively associated (p<0.001) [Tables 3 and 4]. Significant mean error levels were obtained when analyzing methods´ agreement for cSBP, cPP, Pf and Pb data. The only exceptions were cSBPCT-BOSC in the 18–35 y.o. group (p = 0.07), PbRT-BOSC in the 3–12 y.o. group (p = 0.09) and PbCT-BOSC in the 12–18 y.o. group (p = 0.15) [Tables 3 and 4; Fig 3].

Fig 3

Systematic (mean) error obtained in Bland-Altman test, reported as mean value and its confidence interval (95%).

RT and CT: radial and carotid applanation tonometry. BOSC: brachial oscillometry. cSBP, cPP: central (aortic) systolic and pulse pressure. Pf, Pb: forward and backward wave height.

In turn, with few exceptions, cSBP, cPP, Pf and Pb data obtained from the different methodological approaches showed proportional errors [Tables 3 and 4; S3–S7 Tables; S1–S4 Figs]. Disregard of the age-group, for cSBP, cPP and Pf values analyzed, the greater the measurements mean, greater the differences in data from the different approaches (RT, CT, BOSC). In other words, the higher the cSBP, cPP or Pf, greater the observed differences between methods [S1 Fig].

Mean errors obtained for cSBP, cPP, Pf and Pb when comparing the methodological approaches (RT, CT and BOSC) are shown in Fig 3. Absolute values observed for mean errors ranged from1.5to9.3 mmHg for cSBP (BOSC-CT 18–35 y.o. and RT-CT 12–18 y.o.);1.6and 10.7 mmHg forcPP (RT-BOSC 18–35 y.o. and RT-CT 12–18 y.o.); 4.4and20.2 mmHg for Pf (BOSC-RT 3–12 y.o. and BOSC-CT 12–18 y.o.) and from 0.3and 2.4 mmHg for Pb (RT-BOSC 3–12 y.o. and RT-CT 18–35 y.o.) [Fig 3]. The lowest absolute values observed for mean error levels for cSBP data were obtained when analyzing CT and BOSC: 3.1 mmHg (entire group) and 1.5, 3.6 and 4.2 mmHg (young adults, adolescents and children, respectively)[Fig 3]. For all age-groups, the highest mean error values between cSBP data were obtained when comparing CT and RT [Fig 3].

When considering cPP, data from RT and BOSC were the ones with the greatest similarity (least absolute mean error values), 2.4 mmHg (entire group), 1.6, 2.6 and 2.9 mmHg (young adults, adolescents and children, respectively) [Fig 3]. Again, the greatest absolute difference was observed when comparing RT and CT data (e.g. 9.2 mmHg, entire group) [Fig 3]. For all age-groups, Pf values obtained with RT and BOSC showed the most similitude (e.g. mean error 7.1 mmHg, entire group), whereas the major absolute errors were obtained when comparing BOSC and CT (e.g. 18.3 mmHg, entire group) [Fig 3]. When analyzing the entire group, the least differences in Pb were observed between BOSC and tonometry-derived data (e.g. 1.0 mmHg), whereas RT and CT data showed the greatest differences (e.g. 1.9 mmHg). When considering the different age-groups, findings were heterogeneous (e.g. in children the absolute difference in Pb between RT and CT was just 0.9 mmHg) [Fig 3].

When data were calibrated to identical pDBP and MBP, and regardless of the calibration method used[S8 and S9 Tables], Bland-Altman analyses [S10–S19 Tables; S5–S12 Figs] showed that beyond some changes in the “ranking” of the comparisons in the different age-groups, there were no substantial changes when analyzing the entire population[S13–S16 Figs]. In this regard, even when the statistical significance of the comparisons was lost, minor mean errors were observed between: (1) BOSCT-CT for cSBP; (2) RT-BOSC for cPP; (3) BOSC-RT for Pf and (4) BOSC-CT for Pb. Additionally, even when calibrating to identical values and considering calibration methods, the greatest absolute errors were still observed between: (1) RT-CT for cSBP; (2) RT-CT for cPP; (3) BOSC-CT for Pfand (4) RT-CT for Pb (without statistical significance compared to RT-BOSC) [S13–S16 Figs].

Fig 4 (S9 Table) shows that generally when calibrating to identical pressure levels and regardless of the calibration method, the highest cSBP, cPP and Pf values were obtained from CT whilst the lowest cSBP and cPP levels were observed when using RT. Independently of the methodological approach (RT, CT or BOSC), the parameter(cSBP, cPP, Pf and Pb) and the age, higher levels were obtained when calibrating top DBP/MBPosc [Fig 4, S9 Table].

Fig 4

Haemodynamic parameters obtained for the entire and age-related groups, when calibrating to peripheral diastolic (pDBP) and calculated mean blood pressure (MBPc) or measured mean blood pressure (MBPosc).

Scor and MOG: SphygmoCorandMobil-O-Graph device. RT and CT: radial and carotid applanation tonometry. BOSC: brachial oscillometry. cSBP, cPP: central (aortic) systolic and pulse pressure. Pf, Pb: forward and backward wave height.

Explanatory factors for the differences in cSBP, cPP, Pfand Pb data obtained with RT, CT and BOSC

For a given variable, the explanatory factors for the absolute differences between methods used in its measurement varied, depending on the methodological approaches that were compared (e.g. RT vs. CT and RTvs. BOSC) (Tables 5 and 6). Factors explaining the differences in cSBP and cPP data varied depending on the methods considered. In general terms, the differences in cSBP or cPP were associated with sex (major differences in males), age (negatively), pDBP (negatively), BH and/or BW (positively). Depending on the methods compared, HR was positively or negatively associated with the differences in cSBP or cPP [Tables 5 and 6]. Absolute differences in Pf were associated with age (negatively), sex (higher differences in males), BW (positively) and/or HR (positively) [Tables 5 and 6]. The differences in Pb were mainly explained by age (positively; just for RT-CT pDBP/MBPc comparison), sex (higher difference for males), HR and/or pDBP (negatively) and/or BH (positively). Differences in Pf were explained by BW, rather than by BH; the opposite was observed when analyzing Pb.

Table 5

Multiple regression models for absolute differences in cSBP and cPP levels measured with three different methods in the entire group, calibrated with identical pressure levels: pDBP/MBPc and pDBP/MBPosc.

	Calibration method	Variable	β	p-value (β)	VIF	R2	R2a	p-value (model)
|ΔcSBP|
RT-CT	pDBP/MBPc	Sex	-2.14	0.003	1.06	0.04	0.04	0.003
		HR	0.06	0.016	1.06
	pDBP/MBPosc	Age	-0.45	<0.001	1.47	0.05	0.04	0.014
		BodyWeight	0.08	0.012	1.47
RT-BOSC	pDBP/MBPc	BodyHeight	10.21	<0.001	1.25	0.25	0.24	<0.001
		pDBP	-0.18	0.003	1.13
		HR	-0.25	<0.001	1.14
	pDBP/MBPosc	BodyHeight	12.86	0.001	1.25	0.27	0.26	<0.001
		pDBP	-0.24	0.002	1.13
		HR	-0.35	<0.001	1.14
CT-BOSC	pDBP/MBPc	Sex	-2.64	0.008	1.00	0.03	0.02	0.008
	pDBP/MBPosc	Sex	-3.35	0.014	1.00	0.02	0.02	0.014
|ΔcPP|
RT-CT	pDBP/MBPc	Age	-7.91	0.012	2.29	0.13	0.11	<0.001
		BodyWeight	0.09	<0.001	2.19
		Sex	-2.63	<0.001	1.06
		HR	0.12	<0.001	1.16
	pDBP/MBPosc	Sex	-3.76	<0.001	1.06	0.09	0.08	<0.001
		HR	0.18	<0.001	1.06
RT-BOSC	pDBP/MBPc	BodyHeight	10.26	<0.001	1.25	0.25	0.24	<0.001
		pDBP	-0.20	<0.001	1.13
		HR	-0.21	<0.001	1.14
	pDBP/MBPosc	BodyHeight	14.43	<0.001	1.01	0.11	0.10	<0.001
		Sex	-3.94	0.001	1.01
CT-BOSC	pDBP/MBPc	Sex	-2.44	0.009	1.00	0.03	0.02	0.009
	pDBP/MBPosc	Sex	-3.58	0.005	1.06	0.04	0.03	0.008
		HR	0.10	0.042	1.06

cSBP, cPP: central systolic and pulse blood pressure, respectively. HR: heart rate. RT and CT: radial and carotid, tonometry, respectively, obtained with SphygmoCor device (SCOR). BOSC: brachial oscillometry obtained with Mobil-O-Graph device (MOG). |ΔcSBP|, |ΔcPP|: absolute values of difference of cSBP or cPP obtained by resting the two methods of measurement as are shown in columns. For linear regression models the dependent variable were ΔcSBP and ΔcPP, while independent variables were age (y.o.), body height (m), body weight (Kg.), sex (female: 1, male:0), peripheral (brachial) diastolic blood pressure (pDBP, mmHg) and heart rate (HR, beats/minute) entered with stepwise method. MBPc: mean blood pressure calculated as pDBP+(pSBP-pDBP*1/3). MBPosc: mean blood pressure measured by oscillometry. β: slope of regression equation. VIF: variance inflation factor. R2a: adjusted R2. Significance level: p value <0.05.

Table 6

Multiple regression models for absolute differences in Pf or Pb levels measured with three different methods in the entire group, calibrated with identical pressure levels: pDBP/MBPc and pDBP/MBPosc.

	Calibration method	Variable	B	p-value (β)	VIF	R2	R2a	p-value (model)
|ΔPf|
RT-BOSC	pDBP/MBPc	Age	-0.20	0.011	1.43	0.11	0.10	<0.001
		BodyWeight	0.11	<0.001	1.43
	pDBP/MBPosc	Age	-0.29	0.007	1.43	0.11	0.10	<0.001
		BodyWeight	0.14	<0.001	1.43
CT-BOSC	pDBP/MBPc	BodyWeight	0.09	0.001	1.03	0.14	0.13	<0.001
		Sex	-2.87	0.010	1.08
		HR	0.21	<0.001	1.11
	pDBP/MBPosc	Age	-0.49	0.006	1.45	0.10	0.06	<0.001
		BodyWeight	0.15	0.002	1.37
		HR	0.15	0.022	1.09
|ΔPb|
RT-CT	pDBP/MBPc	Age	0.05	0.025	1.10	0.09	0.08	<0.001
		HR	-0.03	0.007	1.10
	pDBP/MBPosc	BodyHeight	2.09	0.047	1.10	0.08	0.07	0.001
		HR	-0.04	0.016	1.10
RT-BOSC	pDBP/MBPc	BodyHeight	6.72	<0.001	1.24	0.26	0.25	<0.001
		pDBP	-0.10	<0.001	1.13
		HR	-0.09	<0.001	1.13
	pDBP/MBPosc	BodyHeight	10.57	<0.001	1.25	0.31	0.30	<0.001
		Sex	-1.19	0.049	1.10
		pDBP	-0.12	<0.001	1.15
		HR	-0.11	<0.001	1.19
CT-BOSC	pDBP/MBPc	BodyHeight	3.67	0.007	1.13	0.14	0.13	<0.001
		HR	-0.07	<0.001	1.13
	pDBP/MBPosc	BodyHeight	5.45	<0.001	1.12	0.13	0.12	<0.001
		HR	-0.06	0.013	1.12

Pf, Pb: forward and backward wave height (amplitude) at aortic level, respectively. HR: heart rate. RT and CT: radial and carotid, tonometry, respectively, obtained with SphygmoCor device (SCOR). BOSC: brachial oscillometry obtained with Mobil-O-Graph device (MOG). |ΔPf|, |ΔPb|: absolute values of difference of Pf and Pb obtained by resting the two methods of measurement as are shown in columns. For linear regression models the dependent variable were ΔPf or ΔPb, while independent variables were age (y.o.), body height (m), body weight (Kg.), sex (female: 1, male:0), peripheral (brachial) diastolic blood pressure (pDBP, mmHg) and heart rate (HR, beats/minute) entered with stepwise method. MBPc: mean blood pressure calculated as pDBP+(pSBP-pDBP*1/3). MBPosc: mean blood pressure measured by oscillometry. β: slope of regression equation. VIF: variance inflation factor. R2a: adjusted R2. Significance level: p value <0.05.

Explanatory factors for the differences between MBPosc and MBPc

MBPosc and MBPc showed a strong positive association (R = 0.9887; p<0.0001). MBPosc values were higher than MBPc levels. Systematic (6.9±1.3 mmHg, p = 0.0003) and proportional errors were observed. The differences increased in association with higher MBP levels (p = 0.0003) [S20 Table, S17 Fig]. Table 7 (bivariate models) and Table 8 (MLR) show explanatory variables for the differences between MBPosc and MBPc.

Table 7

Differences between MBPosc and MBPc: association with demographic data, anthropometric data, risk factors and haemodynamic properties.

	MBPosc—MBPc (mmHg)
Demographic and anthropometric variables	R	P
Age (years)	0.165	0.007
Sex (1: female, 0: male)	-0.247	<0.0001
Bodyheight (m)	0.283	<0.0001
Bodyweight (kg)	0.365	<0.0001
BMI (m/kg2)	0.297	<0.0001
z-BMI* (standard deviation)	0.276	0.001
Cardiovascular RiskFactors	R	P
Hypertension and/or HBP [1: yes, 0: no]	0.179	0.005
Diabetes [1: yes, 0: no]	-0.045	0.48
Dyslipidemia[1: yes, 0: no]	-0.008	0.90
Smoking[1: yes, 0: no]	0.073	0.25
Sedentarylifestyle[1: yes, 0: no]	0.046	0.48
Haemodynamic parameters	R	P
HR (beats/minute)	-0.099	0.11
pDBP (mmHg)	-0.205	<0.001
pSBP (mmHg)	0.653	<0.0001
pPP (mmHg)	0.934	<0.0001

MBP: mean blood pressure. BMI: body mass index. HBP: high blood pressure. HR: heart rate. pDBP, pSBP, pPP: peripheral diastolic, systolic and pulse pressure. R: pearson coefficient. Significance level: p-value <0.05.

*z-BMI: z score of BMI calculated only for under 18 y.o.

Table 8

Multiple regression models for differences between MBPosc and MBPc: Association with respect to demographic, anthropometric, risk factors and haemodynamic properties.

Dependent variable: MBPosc-MBPc	Variable	β	p-value (β)	VIF	R2	R2a	p-value (model)
Model 1 (Enter method)	Age	0.019	0.049	2.69	0.883	0.880	<0.001
	Sex	-0.078	0.223	1.16
	BodyHeight	-0.610	0.030	2.82
	BMI	0.008	0.220	1.27
	pSBP	0.003	0.444	2.41
	pPP	0.124	<0.001	2.03
Model 1 (Stepwise method)	pPP	0.127	<0.001	1.00	0.880	0.879	<0.001
Model 2 (Enter method)	pSBP	0.127	<0.001	1.42	0.874	0.873	<0.001
	pDBP	-0.121	<0.001	1.42
Model 2 (Stepwise method)	pSBP	0.127	<0.001	1.42	0.874	0.873	<0.001
	pDBP	-0.121	<0.001	1.42

BMI: body mass index. pDBP, pSBP, pPP: peripheral (brachial) diastolic, systolic and pulse pressure. HTA and HBP: Hypertension and/or high blood pressure. MBPc: mean blood pressure calculated as pDBP+(pSBP-pDBP*1/3). MBPosc: mean blood pressure measured by oscillometry [mmHg]. For all multiple linear regression models the difference between MBPosc and MBPc was the dependent variable. Model 1: independent variables were age [y.o], sex(female: 1, male: 0), body height(m), body weight(Kg.), body mass index (BMI; Kg./m2), HTA/HBP(yes: 1, no: 0), pDBP, pPP and pSBP(mmHg). Body weight, HTA/HBPand pDBP; pDBPwas excluded due to multicollinearity defined as variance inflation factor (VIF)>5. Model 2: independent variables were pDBP, pSBP, and pPP;pPP was excluded due to multicollinearity. β: slope of regression equation. Significance level: p value <0.05.

The differences between MBPosc and MBPc were associated with sex (major differences in males), pDBP (negatively) and age, BH, BW, BMI, z-BMI, pSBP, pPP, hypertension and/or HBP (positively) [Table 7]. However, the multivariate analysis showed that when demographic, anthropometric, haemodynamic and CRFs variables were jointly considered (Model 1; stepwise and enter), the variable with major explanatory capacity was pPP (positive association), but when considering the enter method, the associations for age (positive, p = 0.049) and BH (negative, p = 0.030) were statistically significant (p<0.05). The R2 showed little variation when considering the three variables (pPP, age, BH) in the equation(0.880 for pPP vs. 0.883 for pPP, age and BH). When analyzing pDBP, pSBP and pPP (Model 2), it was observed that the differences between MBPosc and MBPc were associated with pSBP (positive association)and with pDBP (negative association) [Table 8].

Discussion

Our main results were: First, systematic and proportional errors were observed when analyzing methods agreement. Statistical significance and errors values varied according to the parameter analyzed, the age group considered and the methods compared. When analysing cSBP and cPP data the methods with the greatest similarity varied, but for both variables the greatest differences were obtained when comparing RT and CT data [Tables 3 and 4, Fig 3].

Second, with few exceptions, for cSBP, cPP, Pf or Pb data the methodological approaches with major and least similarities did not vary in association with variations in the calibration scheme considered (pDBP/MBPc vs. pDBP/MBPosc) [Tables 3 and 4, Fig 3]. Regardless of the calibration scheme when data were calibrated to similar pBP, the highest cSBP, cPP and Pf levels were obtained from CT, whereas the lowest cSBP and cPP values were obtained from RT [Fig 4, S9 Table]. Disregard of the technique (RT, CT or BOSC), parameter (cSBP, cPP, Pf or Pb) or age group (children, adolescents or young adults) considered, higher absolute values were obtained when data were calibrated to pDBP/MBPosc [Fig 4, S9 Table].

Third, age, sex, HR, pDBP, BW and BH were explanatory factors for the differences in cSBP, cPP, Pf or Pb [Tables 5 and 6]. Considering a given central parameter (e.g. cPP), explanatory factors for the differences between data obtained from different approaches would vary depending on the methods compared (e.g. RT vs. CT and RT vs. BOSC) [Tables 5 and 6]. Regardless of demographic (age, sex) and anthropometric (BW, BH and BMI) variables and CRFs exposure (e.g. HBP), pPP, pSBP and pDBP, were explanatory factors for the difference between MBPosc and MBPc [Table 8].

Our results showed that the greatest differences were observed between data obtained using a similar technique (i.e. applanation tonometry) and calibrating signals to similar BP levels. In this regard, as described, the greatest differences for cSBP and cPP data were obtained when comparing RT and CT. On the other hand, the agreement between methods varied according to the parameter studied (cSBP, cPP, Pf or Pb). Then, an adequate analysis of the equivalence between data obtained with different techniques and devices requires considering individually the different parameters, being aware that results obtained for a given variable cannot be extrapolated to others. Although there were variations in the absolute differences between data from different methods, in general terms, the "ordering" of approaches defined by the degree of agreement between data did not vary depending on the calibration method considered.

Higher cSBP, cPP and Pf levels were obtained with CT, whereas minor cSBP and cPP values were obtained with RT. Then, for the methodological approaches analyzed it could be said that, the closer to the aortic root the register is achieved, the higher cSBP and cPP levels obtained. This issue could be explained by the differences in pBP measured invasive and non-invasively[23]. About this, regardless of the technique used (e.g. oscillometric or auscultatory) cuff BP under-estimated intra-arterial pSBP (and pPP), at the same time it over-estimated intra-arterial pDBP[24]. Consequently, when using pDBP and MBP (calculated or measured) to calibrate arterial signals (e.g. carotid, brachial or radial), if pSBP is obtained first (and thereafter cSBP), greater errors or differences (underestimation) could be expected when considering peripheral arteries records[8,23].

For the different methodological approaches, parameters and age-groups, the highest absolute levels were obtained when calibrating to pDBP/MBPosc [Fig 4, S9 Table]. This finding is in agreement with the expected. Compared to other calibration methods (i.e. pSBP/pDBP, pDBP/MBPc), calibrating to MBPosc (pDBP/MBPosc) enables obtaining higher cSBP levels which in turn would be more alike those measured invasively[3]. On the other hand, in works that mostly used BOSC (i.e. Mobil-O-Graph) in adults, it was shown that MBPosc would be closer to the true invasive MBP and therefore compared to MBPc, the MBPosc would be more accurate to calibrate signals[25]. cSBP obtained calibrating to DBP/MBPosc has shown: (1) better correlation with cardiac hypertrophy when using 24-hour ambulatory cSBP data[26], (2) superior discriminatory capability, associated with significant improvement in reclassification to identify cardiac structural abnormalities in community-based patients with stage A heart failure[27], (3) association with clinical outcomes in patients with chronic kidney disease[28], (4) enhanced association with cardiac structural features in children, adolescents and young adults [29]. Additionally, comparatively, cSBP obtained calibrating to DBP/MBPosc showed a weaker association with pSBP, which would increase the independent predictive capacity[29,30,31]. The value of considering cSBP apart from pSBP has been discussed and considered over the last two decades. It has been stated that their association a-priori limits the incremental clinical value of cSBP[32]. However, it has been demonstrated that the association could be modified by the measurement procedure and particularly by the calibration method[8,27,28,33]. By using pSBP (i.e. within the MBPc equation) as a direct input variable for cSBP estimation, an intrinsic mathematical connection is established systematically and an association between variables is predetermined. Therefore, cSBP obtained using oscillometric data for calibration, avoids the use of pSBP and considers measured MBP (MPBosc) instead. Then, the impact of pSBP is attenuated[8,27,28,33,34]. In this work, we showed for the first time in children and adolescents, that cSBP levels obtained calibrating to MBPosc would be higher than those obtained when calibrating to MBPc (Form factor: 0.33), regardless of the technique considered: RT, CT, BOSC.

For a given hemodynamic variable (e.g. cPP) the explanatory factors for the differences between data obtained with the different approaches varied depending on the methods compared and/or calibration schemes. Looking for explanation to these findings it could be proposed that the age and/or HR-association described for pulse amplification and arterial stiffness (mainly in central arteries)[13], contributes to understand the age and/or HR dependence observed for the differences in cSBP, cPP and Pf data obtained from central and peripheral tonometric records (RT-CT). Additionally, sex and BH explained the differences between some methods, which could be related: (a) with the algorithms (not disclosed) used by the devices (e.g. that could incorporate BH for the assessment of cSBP) and/or (b) to the fact that an average GTF is used for all subjects. Regarding the latter, some devices (e.g. SphygmoCor) frequently use an average GTF (a frequency-dependent transformation) to correlate measured radial BP waves to measured cBP waves, obtained in a group of subjects. Then, the GTF is applied to radial records from new subjects to estimate cBP waves and parameters. Despite it has been demonstrated that GTF could yield good agreement with invasive cBP measurements, since the GTF is a population average, it could assume that PP amplification is just a fixed value. Hence, the GTF may not adapt to the aforesaid inter-subject variability in PP amplification and therefore it could yield non-trivial cBP errors when PP amplification is non-uniform [35]. In this sense, since in general males show greater center-periphery amplification of the pulse and/or present greater BH than females, the use of a single GTF could explain the differences found between methods. Consequently, in addition to the "direct" effects that biological factors (e.g. subjects´ characteristics) may have on the methods and records (e.g. CT limitations in obese subjects, children or women) that would determine or contribute to the differences between data from different methods, they could also be integrated into the equations (usually not given to the users) used by the devices to obtain hemodynamic parameters. Therefore, biological factors could also contribute "indirectly" (mathematically) to the differences between data from different methodological approaches.

Our data suggest that the differences between MBPosc and MBPc [S20 Table, S17 Fig], that contribute to explain differences in central parameters obtained calibrating to pDBP/MBPosc or to pDBP/MBPc were not associated with demographic or anthropometric variables, nor with HR. On the other hand, they were associated with pSBP, pPP or pDBP levels [Table 8]. Higher pSBP (or pPP) and lower pDBP levels were associated with higher differences between MBPosc and MBPc. This is in agreement with Kiers et al. [36], who after bivariate and multivariate analysis (data from subjects without cardiovascular disease) reported that the differences between MBPosc and MBPc were not associated with sex, age, BMI or HR, but with pSBP and pDBP. In turn, Bos et al. showed that MBPc (MBPc = pDBP+1/3pPP) underestimates "real" MBP (invasively measured), with larger underestimations at higher pressure levels [37]. That indirectly suggests that if MBPosc really approximates more to "real" MBP than MBPc, then the greater the pressure levels, the greater the MBPosc-MBPc difference. The result of the univariate analysis is in agreement with that reported by Smulyan et al. who evaluated the difference between MBPc and MBPosc (patients who underwent a coronary angiography), and observed an association ("a weak correlation") between MBP differences and age (r = 0.32; p<0.001); unfortunately, the authors did not perform multivariate analysis including blood pressure, which precludes a full comparison [25].

Strengths and limitations

Due to the characteristics of the studied population invasive data were not obtained. However, characteristics of the studied sample comply with what was stated by Sharman et al. (2017): participants should have a sex distribution of at least 30% male and female, in sinus rhythm; devices should be tested over a range of BP and across a range of HRs (i.e. 60–100 b.p.m.)[9]. In this study, females were within 46–58% of the entire population and age-related groups, all the subjects were in sinus rhythm, pSBP and pDBP range: 70–217 mmHg and 42–106 mmHg, respectively, and HR range was 42–151 beats/minute. We did not measure MBP invasively, so we cannot conclude on the best way to quantify non-invasively MBP to be used for calibration. On the other hand, because the measurement algorithm used by the oscillometric device is unknown, we cannot explain the differences between oscillometric and calculated MBP levels as derived from the algorithm. In spite of this, the comparative analysis of MBP has important practical implications. Researchers using an oscillometric device for obtaining MBP should be aware of the differences between calculated and measured MBP. Additionally, researchers should describe the method used to obtain MBP with an oscillometric device precisely and whether measured or calculated MBP was used should be specially indicated. We do want to underline the importance of describing the method used to determine MBP when using an oscillometric device. A problem arises when oscillometric devices (e.g. the Mobil-O-Graph) do not report (in the display) MBPosc. Consequently, users usually calculate MBP from pSBP and pDBP values given by the device. It is clear that despite they are calculated using the same device and record, MBPosc and MBPc (form factor equal to 0.33) obtained are not similar. Furthermore, it is to note that there are other methods to quantify MBPc that we did not analyze since they are not used by methodological approaches considered in this work (Mobil-O-Graph).

Conclusions

Systematic and proportional errors were observed. When analysing cSBP and cPP, there were differences in the techniques with the greatest similarity, but for both variables the greatest differences were obtained when comparing RT and CT data. In general terms, for cSBP, cPP, Pf or Pb data the methods (RT, CT and BOSC) with major and least similarities did not vary in association with variations in the calibration scheme considered. Regardless of the calibration scheme, when data were calibrated to similar pBP, the highest cSBP, cPP and Pf levels were obtained from CT, whereas the lowest cSBP and cPP values were obtained using RT. Higher cSBP, cPP, Pf or Pb absolute values were obtained when data were calibrated to pDBP/MBPosc. Age, sex, HR, pDBP, BW and/or BH were explanatory factors for the differences in cSBP, cPP, Pf or Pb. For a given central parameter (cSBP, cPP, Pf, Pb) the explanatory factors for the differences between data obtained from different approaches would vary depending on the methods compared (e.g. RT vs. CT and RT vs. BOSC). pSBP and pPP (positively) and pDBP (negatively) were explanatory factors for the differences between MBPosc and MBPc.

A. Wave separation analysis (WSA) [SphygmoCor and Mobil-O-Graph]. B. cSBP, cPP, Pf and Pb absolute and relative intra (repeatability) and inter-observer (reproducibility) variability.

(DOCX)

Click here for additional data file.

Bland-Altman graphs for cSBP (Calibration: pDBP/MBPc; instantaneous levels): Entire and age-related groups.

(DOCX)

Click here for additional data file.

Bland-Altman graphs for cPP(Calibration: pDBP/MBPc; instantaneous levels): Entire and age-related groups.

(DOCX)

Click here for additional data file.

Bland-Altman graphs for Pf(Calibration: pDBP/MBPc; instantaneous levels): Entire and age-related groups.

(DOCX)

Click here for additional data file.

Bland-Altman graphs for Pb (Calibration: pDBP/MBPc; instantaneous levels): Entire and age-related groups.

(DOCX)

Click here for additional data file.

Bland-Altman graphs for cSBP(Calibration: pDBP/MBPc; equal levels): Entire and age-related groups.

(DOCX)

Click here for additional data file.

Bland-Altman graphs for cSBP(Calibration: pDBP/MBPosc; equal levels): Entire and age-related groups.

(DOCX)

Click here for additional data file.

Bland-Altman graphs for cPP(Calibration: pDBP/MBPc; equal levels): Entire and age-related groups.

(DOCX)

Click here for additional data file.

Bland-Altman graphs for cPP(Calibration: pDBP/MBPosc; equal levels): Entire and age-related groups.

(DOCX)

Click here for additional data file.

Bland-Altman graphs for Pf(Calibration: pDBP/MBPc; equal levels): Entire and age-related groups.

(DOCX)

Click here for additional data file.

Bland-Altman graphs for Pf(Calibration: pDBP/MBPosc; equal levels): Entire and age-related groups.

(DOCX)

Click here for additional data file.

Bland-Altman graphs for Pb (Calibration: pDBP/MBPc; equal levels): Entire and age-related groups.

(DOCX)

Click here for additional data file.

Bland-Altman graphs for Pb (Calibration: pDBP/MBPosc; equal levels): Entire and age-related groups.

(DOCX)

Click here for additional data file.

Bland-Altman-derived cSBP Mean Error levels obtained with three different recording methods and calibration schemes: pDBP/MBPc (instantaneous levels), pDBP/ MBPc (equal levels) and pDBP/MBPosc (equal levels).

(DOCX)

Click here for additional data file.

Bland-Altman-derived cPP Mean Error levels obtained with three different recording methods and calibration schemes: pDBP/MBPc (instantaneous levels), pDBP/ MBPc (equal levels) and pDBP/MBPosc (equal levels).

(DOCX)

Click here for additional data file.

Bland-Altman-derived Pf Mean Error levels obtained with three different recording methods and calibration schemes: pDBP/MBPc (instantaneous levels), pDBP/ MBPc (equal levels) and pDBP/MBPosc (equal levels).

(DOCX)

Click here for additional data file.

Bland-Altman-derived Pb Mean Error levels obtained with three different recording methods and calibration schemes: pDBP/MBPc (instantaneous levels), pDBP/ MBPc (equal levels) and pDBP/MBPosc (equal levels).

(DOCX)

Click here for additional data file.

Bland-Altman-derived mean and systematic error levels for MBPosc and MBPc differences.

(DOCX)

Click here for additional data file.

Clinical features and cardiovascular risk factors for the entire and age-related subgroups (Extended table).

(DOCX)

Click here for additional data file.

Haemodynamic and aortic wave-derived parameters measured with three different methods in the entire and age-related subgroups.

(DOCX)

Click here for additional data file.

cSBP, cPP, Pf and Pb: Correlation and agreement among values obtained with three different recording methods (Summary Table).

(DOCX)

Click here for additional data file.

cSBP: Correlation and agreement among values obtained with three different recording methods.

(DOCX)

Click here for additional data file.

cPP: Correlation and agreement among values obtained with three different recording methods.

(DOCX)

Click here for additional data file.

Pf: Correlation and agreement among values obtained with three different recording methods.

(DOCX)

Click here for additional data file.

Pb: Correlation and agreement among values obtained with three different recording methods.

(DOCX)

Click here for additional data file.

Clinical features and cardiovascular risk factors for the entire and age-related subgroups: Subsample.

(DOCX)

Click here for additional data file.

Haemodynamic and aortic wave-derived parameters measured with three different methods in the entire and age-related subgroups, calibrated with identical peripheral blood pressure levels obtained by oscillometry, using two different calibration schemes: pDBP/MBPc and pDBP/MBPosc.

(DOCX)

Click here for additional data file.

cSBP and cPP: Agreement among parameters measured with three different methods in the entire and age-related subgroups, calibrated with identical peripheral blood pressure levels obtained by oscillometry, using two different calibration schemes: pDBP/MBPc and pDBP/MBPosc [Summary table].

(DOCX)

Click here for additional data file.

Pf and Pb: Agreement among parameters measured with three different methods in the entire and age-related subgroups, calibrated with identical peripheral blood pressure levels obtained by oscillometry, using two different calibration schemes: pDBP/MBPc and pDBP/MBPosc [Summary table].

(DOCX)

Click here for additional data file.

cSBP: Agreement among parameters measured with three different methods in the entire and age-related subgroups, calibrated with identical peripheral blood pressure levels obtained by oscillometry (Calibration scheme: pDBP/MBPc) [Extended table].

(DOCX)

Click here for additional data file.

cPP: Agreement among parameters measured with three different methods in the entire and age-related subgroups, calibrated with identical peripheral blood pressure levels obtained by oscillometry (Calibration scheme: pDBP/MBPc) [Extended table].

(DOCX)

Click here for additional data file.

Pf: Agreement among parameters measured with three different methods in the entire and age-related subgroups, calibrated with identical peripheral blood pressure levels obtained by oscillometry (Calibration scheme: pDBP/MBPc) [Extended table].

(DOCX)

Click here for additional data file.

Pb: Agreement among parameters measured with three different methods in the entire and age-related subgroups, calibrated with identical peripheral blood pressure levels obtained by oscillometry (Calibration scheme: pDBP/MBPc) [Extended table].

(DOCX)

Click here for additional data file.

cSBP: Agreement among parameters measured with three different methods in the entire and age-related subgroups, calibrated with identical peripheral blood pressure levels obtained by oscillometry (Calibration scheme: pDBP/MBPosc) [Extended table].

(DOCX)

Click here for additional data file.

(DOCX)

Click here for additional data file.

Pf: Agreement among parameters measured with three different methods in the entire and age-related subgroups, calibrated with identical peripheral blood pressure levels obtained by oscillometry (Calibration scheme: pDBP/MBPosc) [Extended table].

(DOCX)

Click here for additional data file.

Pb: Agreement among parameters measured with three different methods in the entire and age-related subgroups, calibrated with identical peripheral blood pressure levels obtained by oscillometry (Calibration scheme: pDBP/MBPosc) [Extended table].

(DOCX)

Click here for additional data file.

Agreement among oscillometry-derived and calculated mean blood pressure (MBPosc and MBPc, respectively) obtained with Mobil-O-Graph device (MOG).

(DOCX)

Click here for additional data file.

18 Sep 2019

PONE-D-19-22199

Aortic pressure and forward and backward wave components in children, adolescents and young-adults: agreement among data from brachial oscillometry, and radial and carotid tonometry, and factors associated with their differences

PLOS ONE

Dear Dr. Bia,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Nov 02 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Giacomo Pucci

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. In your Methods section, please provide additional information about the participant recruitment method and the demographic details of your participants. Please ensure you have provided sufficient details to replicate the analyses such as: a) the recruitment date range (month and year), b) a description of any inclusion/exclusion criteria that were applied to participant recruitment, c)  a statement as to whether your sample can be considered representative of a larger population, d) a description of how participants were recruited, and e) descriptions of where participants were recruited and where the research took place. Moreover, please specify how the levels of physical activity were assessed.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This is an interesting study on the possible differences between different methods of assessing central blood pressure levels in children, adolescents and young adults. Τhe authors aimed to analyze, in children, adolescents and young-adults (1) the agreement between cSBP, cPP, Pf and Pb obtained using carotid (CT) and radial tonometry (RT) and brachial-oscillometry (BOSC); and identify (2) explanatory factors for the differences between the approaches-data. 1685 subjects (mean/range age: 14/3-35 y.o.) assigned to three age-related groups (3-12; 12-18; 18-35 y.o.) were included. cSBP, cPP, Pf and Pb were assessed with BOSC (Mobil-O-Graph), CT and RT (SphygmoCor) records. Two calibration schemes were considered: MBPc and MBPosc for calibrations to similar BP levels. Correlation, Bland-Altman tests and multiple regression models were applied. Systematic and proportional errors were observed; errors´ statistical significance and values varied depending on the parameter analyzed, methods compared and group considered. The explanatory factors for the differences between data obtained from the different approaches varied depending on the methods compared. The highest cSBP and cPP were obtained from CT; the lowest from RT. The study is well-performed. There are several issues that if addressed could improve the manuscript.

Major comments

1. The length of the manuscript must be significantly reduced and especially the Discussion.

2. Please avoid having the figure legends and tables between the text because it gets really confusing trying to read through the manuscript. Please have a separate part with Tables and Figure legends.

3. In Table 1, for clarity and space reasons please remove the min and max columns as well as the interquartile range unless it is necessary due to a non-normal distribution. Moreover, include in the same column the mean and SD values. If this is done please present the overall and the other 3 smaller populations in a single table.

4. In Table 2, please include the mean and SD values in the same column. Please report the decimal values in the values in Table 2, because the p-values are different for similar differences and thus this means there is information missing. Moreover, when the value of p is ≥0.01, please report only 2 decimals.

5. Please place all the supplemental material in a different to the main manuscript file.

6. Table 3 needs to be reduced in size as well. I would keep only the r and p-values of the correlations and have the rest of the data in a supplemental table.

7. The authors reports errors as follows: “~2 to ~9”. The exact values should be included throughout the manuscript instead of using the ~ sign.

8. “Male sex showed the greatest differences between methods.” How can the authors explain this result?

9. “Differences in cSBP between RT and BOSC were explained by the BH of the subject.” How do the authors explain this result? Could be the fact that BOSC uses an algorithm that incorporates height for the assessment of cBPs?

10. The authors should report also their intra- and inter-observer reproducibility, because there is a chance that these errors in assessment between techniques are due to large variability in the assessment.

Minor comments

1. In the study population criteria please change “cardioactive” to “vasoactive”.

2. Please correct “have shown to met” to “have shown to meet”.

3. Please correct “("intantaneous blood pressure")” to “("instantaneous blood pressure")”.

4. Please correct “Pf y Pb” to “Pf and Pb”.

5. Please improve the resolution of all images because the cannot be properly assessed in their current format.

6. “Fig 4 (data included in S6 Table)”. Please report these data either on text or figure but not both.

7. Please change “were observed when using TR.” to “were observed when using RT.”.

8. Please report in the supplement the exact step by step method you used to conduct the wave separation analysis based on the observed aortic waveforms in each device and which given parameters you used from each device.

9. “When analyzing the age impact, the magnitude of the associations and differences varied depending on the approaches and calibration methods considered. For example, an increase in age equal to 10 years could be associated with a reduction in the differences in Pf equal to ~2 mmHg, ~4 mmHg, or ~6 mmHg, depending on whether RT-BOSC (cal. pDBP/MBPc), CT-BOSC (cal. pDBP/MBPc) or CT-BOSC (cal. pDBP/MBPosc) are compared.”

“A BH increase equal to 1 meter during growth would associate an increase in Pb differences equal to: ~2 mmHg for RT-CT (calibration: pDBP/MBPosc); ~7 mmHg and ~11 mmHg for RT-BOSC (calibration: pDBP/MBPc and pDBP/MBPosc, respectively), and ~5 and ~7 mmHg for CT-BOSC (calibration: pDBP/MBPc and pDBP/MBPosc, respectively).”

These parts seem more like a discussion rather than a result. If the authors want to keep it, they could move it in the Discussion section.

10. The authors in the beginning of their Discussion report more than 5 different results. They must choose which ones of them are the most essential and report only them, because it get really confusing with so many results in such a short space.

11. Please change “by Sharma et al.” to “by Sharman et al.”.

Reviewer #2: This is an interesting study exploring agreement between different methods (carotid and radial

tonometry and brachial oscillometry) in measuring central blood pressure and pulse wave analysis parameters in a large cohort of children, adolescents and young-adults.

The study is well conducted. The results highlighting the highest central BP levels obtained from carotid tonometry and the lowest by using radial tonometry are helpful when interpreting and comparing results of different studies.

I have two requests for the authors, which require further analysis which may improve the quality of the paper.

1) It seems that the way of calculating mean BP, necessary for calibration of all methods, may significantly change the estimation of central BP parameters. Is it possible to clarify which variables may influence the difference between MBPc and MBPosc, using a multiple regression model?

2) When interpreting the variables explaining absolute differences between methods (Table 4), a possibility is that actual BP levels (and particularly DBP) and heart rate may have a significant influence. My suggestion is to repeat the analysis in Table 2 by introducing these two parameters in the multiple regression models.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

7 Nov 2019

Academic Editor: Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming.

Authors: We believe that the manuscript meets all the requirements. Thanks for the comments and suggestions that have contributed to improve the manuscript.

Academic Editor: In your Methods section, please provide additional information about the participant recruitment method and the demographic details of your participants. Please ensure you have provided sufficient details to replicate the analyses such as: a) the recruitment date range (month and year), b) a description of any inclusion/exclusion criteria that were applied to participant recruitment, c) a statement as to whether your sample can be considered representative of a larger population, d) a description of how participants were recruited, and e) descriptions of where participants were recruited and where the research took place. Moreover, please specify how the levels of physical activity were assessed.

Authors: As requested, in the revised version we provided additional methodological information.

-------------------------------------------------------------------------------------------------------------------------------

Reviewer #1: This is an interesting study on the possible differences between different methods of assessing central blood pressure levels in children, adolescents and young adults. Τhe authors aimed to analyze, in children, adolescents and young-adults (1) the agreement between cSBP, cPP, Pf and Pb obtained using carotid (CT) and radial tonometry (RT) and brachial-oscillometry (BOSC); and identify (2) explanatory factors for the differences between the approaches-data. 1685 subjects (mean/range age: 14/3-35 y.o.) assigned to three age-related groups (3-12; 12-18; 18-35 y.o.) were included. cSBP, cPP, Pf and Pb were assessed with BOSC (Mobil-O-Graph), CT and RT (SphygmoCor) records. Two calibration schemes were considered: MBPc and MBPosc for calibrations to similar BP levels. Correlation, Bland-Altman tests and multiple regression models were applied. Systematic and proportional errors were observed; errors´ statistical significance and values varied depending on the parameter analyzed, methods compared and group considered. The explanatory factors for the differences between data obtained from the different approaches varied depending on the methods compared. The highest cSBP and cPP were obtained from CT; the lowest from RT. The study is well-performed. There are several issues that if addressed could improve the manuscript.

Authors: Thanks for your revision and comments.

Reviewer #1: The length of the manuscript must be significantly reduced and especially the Discussion.

Authors: Since we were asked (1) by the Editor to include more methodological information and (2) by Reviewer 2 to perform new analyses (which we did) it was difficult to reduce the length of the article.

However, as was requested, and without reducing the quality of the information provided, a significant part of the new analyses and the requested information (e.g. description of the wave separation analysis - WSA -) was included as Supplementary Material.

Reviewer #1: Please avoid having the figure legends and tables between the text because it gets really confusing trying to read through the manuscript. Please have a separate part with Tables and Figure legends. Please place all the supplemental material in a different to the main manuscript file.

Authors: We do understand the reviewer concern. However, that is what is indicated in the "Instructions for authors" (see below). We could have changed it, but the Editor has asked us to: "Please ensure that your manuscript meets PLOS ONE's style requirements, including .... ".

Supplementary Tables and Figures are in separate files.

The only thing we have included at the end of the manuscript file (taking into account the "Instructions for authors") was the description of the supplementary files ("Supporting information captions").

Reviewer #1: In Table 1, for clarity and space reasons please remove the min and max columns as well as the interquartile range unless it is necessary due to a non-normal distribution. Moreover, include in the same column the mean and SD values. If this is done please present the overall and the other 3 smaller populations in a single table. In Table 2, please include the mean and SD values in the same column. Please report the decimal values in the values in Table 2, because the p-values are different for similar differences and thus this means there is information missing. Moreover, when the value of p is ≥0.01, please report only 2 decimals. Table 3 needs to be reduced in size as well. I would keep only the r and p-values of the correlations and have the rest of the data in a supplemental table.

Authors: As requested, in Table 1 we eliminated the minimum, maximum as well as the 25th and 75th percentile values. In addition, data was unified in a single Table.

As suggested, Table 2 was modified (decimal values were added).

Table 3 was significantly shortened (its full version was included as a Supplementary Table). In the new (shortened) Table 3 we chose to keep the value of R and p of the correlations, and fundamental information of Bland-Altman tests, which allowed the existence of systematic (mean error, ME) and/or proportional errors to be assessed. In other words, "ME, C.I. 95% U.L. (mmHg)", "ME, C.I. 95% L.L. (mmHg)", "C.I. 95%, Upper limit (mmHg)" and "C.I. 95%, Lower limit (mmHg)" data (a total or 32 rows) were eliminated.

We kept mean and SD values in different columns due to we considered this enhances the visual clarity of the Table (mainly in the more extensive Tables - e.g. Table 2 -).

Reviewer #1: The authors reports errors as follows: “~2 to ~9”. The exact values should be included throughout the manuscript instead of using the ~ sign.

Authors: The requested change was done.

Reviewer #1: “Male sex showed the greatest differences between methods.” How can the authors explain this result? “Differences in cSBP between RT and BOSC were explained by the BH of the subject.” How do the authors explain this result? Could be the fact that BOSC uses an algorithm that incorporates height for the assessment of cBPs?

Authors: The issue was analyzed in greater detail in the revised version. Furthermore, the (additional) data analysis done in response to Reviewer #2 comments also contributed to explain the finding.

Reviewer #1: The authors should report also their intra- and inter-observer reproducibility, because there is a chance that these errors in assessment between techniques are due to large variability in the assessment.

Authors: The requested information was included summarily in the revised version. Detailed data was included in the Supplementary Materials´ section.

Reviewer #1: In the study population criteria please change “cardioactive” to “vasoactive”. Please correct “have shown to met” to “have shown to meet”. Please correct “("intantaneous blood pressure")” to “("instantaneous blood pressure")”. Please correct “Pf y Pb” to “Pf and Pb”. Please change “were observed when using TR.” to “were observed when using RT.”. Please change “by Sharma et al.” to “by Sharman et al.”.

Authors: The text was revised. Mistakes were corrected. Thanks.

Reviewer #1: Please improve the resolution of all images because the cannot be properly assessed in their current format.

Authors: As suggested, all figures were evaluated and corrected using the recommended system (Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/.)

Reviewer #1: 6. “Fig 4 (data included in S6 Table)”. Please report these data either on text or figure but not both.

Authors: We believe that both are useful, since Figure 4 is in the article (main text), highlighting most relevant haemodynamic information, while "Table S6" is a Supplementary Table included in Annexes (there, interested authors can find detailed numerical information). That is, Figure 4 does not show all the information included in Table S6. On the other hand, the inclusion of Table S6 in the manuscript (main text) would not make sense as the information contained is extensive and could be considered excessive in that context.

Reviewer #1: Please report in the supplement the exact step by step method you used to conduct the wave separation analysis based on the observed aortic waveforms in each device and which given parameters you used from each device.

Authors: As requested, a special document was included as Supplementary Material explaining the exact, step by step, method used to carry out wave separation analysis based on recorded carotid (SphygmoCor) and mathematically-derived aortic (SphygmoCor and Mobil-O-Graph) waveforms. We included examples of the analysis of records obtained in a real subject, using both softwares.

Reviewer #1: “When analyzing the age impact, the magnitude of the associations and differences varied depending on the approaches and calibration methods considered. For example, an increase in age equal to 10 years could be associated with a reduction in the differences in Pf equal to ~2 mmHg, ~4 mmHg, or ~6 mmHg, depending on whether RT-BOSC (cal. pDBP/MBPc), CT-BOSC (cal. pDBP/MBPc) or CT-BOSC (cal. pDBP/MBPosc) are compared.” “A BH increase equal to 1 meter during growth would associate an increase in Pb differences equal to: ~2 mmHg for RT-CT (calibration: pDBP/MBPosc); ~7 mmHg and ~11 mmHg for RT-BOSC (calibration: pDBP/MBPc and pDBP/MBPosc, respectively), and ~5 and ~7 mmHg for CT-BOSC (calibration: pDBP/MBPc and pDBP/MBPosc, respectively).” These parts seem more like a discussion rather than a result. If the authors want to keep it, they could move it in the Discussion section.

Authors: As suggested, the text was modified and included in the Discussion section.

Reviewer #1: The authors in the beginning of their Discussion report more than 5 different results. They must choose which ones of them are the most essential and report only them, because it get really confusing with so many results in such a short space.

Authors: As was requested, without reducing the quality of the information provided, the initial paragraph was modified (simplified) in order to clarify the message given.

-----------------------------------------------------------------------------------------------------------------------------------

Reviewer #2: This is an interesting study exploring agreement between different methods (carotid and radial tonometry and brachial oscillometry) in measuring central blood pressure and pulse wave analysis parameters in a large cohort of children, adolescents and young-adults. The study is well conducted. The results highlighting the highest central BP levels obtained from carotid tonometry and the lowest by using radial tonometry are helpful when interpreting and comparing results of different studies.

Authors: Thanks for the comments and suggestions that have contributed to improve the article.

Reviewer #2: I have two requests for the authors, which require further analysis which may improve the quality of the paper. 1) It seems that the way of calculating mean BP, necessary for calibration of all methods, may significantly change the estimation of central BP parameters. Is it possible to clarify which variables may influence the difference between MBPc and MBPosc, using a multiple regression model?

Authors: The requested analysis was performed in three steps. First, Bland-Altman analyses were performed to evaluate the agreement between MBPosc and MBPc (Supplementary Materials). Second, we identified demographic, anthropometric, cardiovascular risk factors and/or haemodynamic characteristics associated with the differences between MBPosc and MBPc (simple and point bi-serial correlation analysis) (new Table 5). Third, we included the selected independent variables (those with a p˂0.1) in multiple linear regression models (MLR) (new Table 6).

Reviewer #2: When interpreting the variables explaining absolute differences between methods (Table 4), a possibility is that actual BP levels (and particularly DBP) and heart rate may have a significant influence. My suggestion is to repeat the analysis in Table 4 by introducing these two parameters in the multiple regression models.

Authors: The suggestion was accepted. Thank you for your contribution. The requested analysis was performed and results were included in the new (modified) Table 4.

Submitted filename: Response to reviewers.docx

Click here for additional data file.

26 Nov 2019

PONE-D-19-22199R1

Aortic pressure and forward and backward wave components in children, adolescents and young-adults: agreement between brachial oscillometry, radial and carotid tonometry data and analysis of factors associated with their differences

PLOS ONE

Dear Dr. Bia,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Specifically, please shorten and edit the manuscript according to recommendations of Reviewer #1.

We would appreciate receiving your revised manuscript by Jan 10 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Giacomo Pucci

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have adequately answered most of the Editor’s and Reviewers’ comments. This paper has been improved. However, there are some small issues that still need their attention.

Comments

1. It is weird that the resolution in the Supplem. Figures is far better than the one in the main Figures. The labels are too blur and not visible for most of the Figures of the main manuscript.

2. The manuscript as commented in my first review is extremely long (reaching close to 10,000 words) and thus not easily readable. I was hoping that the authors would be able to shorten it after the revision, but despite their great efforts the length is not fit for publication. In fact, the number of supplemental tables (20) and figures (17) is huge. Despite, the good scientific background of the study the authors have not managed to report their results in a presentable manner. I think the authors should try and shorten the manuscript or if they think this is not feasible try to resubmit their findings in more than 1 articles.

3. One example of ways to reduce the size is the analysis presented in Table 6. There is no reason to use both enter and stepwise analysis in the models. Only the stepwise model would be sufficient.

Reviewer #2: The authors adequately answered to my previous comments.

The paper has improved from the previous version.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Andrea Grillo

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

29 Nov 2019

29 -Nov-2019

Dear, Giacomo Pucci

Academic Editor

PLOS ONE

I am sending the second revised version of the original manuscript (PONE-D-19-22199): "Aortic pressure and forward and backward wave components in children, adolescents and young-adults: agreement between brachial oscillometry, radial and carotid tonometry data and analysis of factors associated with their differences", to be considered for publication as full length article in Plos One.

All reviewers’ suggestions were considered in the new version ('Response to Reviewers'; below). We thank the interest in our manuscript, and the reviewers’ comments that allowed us to increase the clarity of the manuscript.

Sincerely,

Prof. Dr. Daniel Bia

Physiology Department, School of Medicine.

Universidad de la República, Montevideo, Uruguay

Phone: 5982-9243414 (3313); e-mail: dbia@fmed.edu.uy

---------------------------------------------------------------------------------------------------------------------------------

Reviewer #1: The authors have adequately answered most of the Editor’s and Reviewers’ comments. This paper has been improved. However, there are some small issues that still need their attention.

Authors: Thanks for the comments and suggestions that contributed to improve the manuscript.

Reviewer #1: 1. It is weird that the resolution in the Supplem. Figures is far better than the one in the main Figures. The labels are too blur and not visible for most of the Figures of the main manuscript.

Authors: As requested, in the revised version (1) Figures were improved (e.g. font size and graphics symbols were increased and the legends became more visible).

Reviewer #1: 2. The manuscript as commented in my first review is extremely long (reaching close to 10,000 words) and thus not easily readable. I was hoping that the authors would be able to shorten it after the revision, but despite their great efforts the length is not fit for publication. In fact, the number of supplemental tables (20) and figures (17) is huge. Despite, the good scientific background of the study the authors have not managed to report their results in a presentable manner. I think the authors should try and shorten the manuscript or if they think this is not feasible try to resubmit their findings in more than 1 articles. One example of ways to reduce the size is the analysis presented in Table 6. There is no reason to use both enter and stepwise analysis in the models. Only the stepwise model would be sufficient.

Authors: We understand the position of the Reviewer. In addition, we understand that while Plos One does not limit words, authors must be concise (*). In this context, we reduced the length of the article (approximately in 3,000 words). Then we consider it currently has a number of words that we believe is not excessive (Main text: Introduction + Methods + Results + Discussion + Conclusion: 5,136 words). See below the word count details.

It should be noted that we were asked for additional information and new (complementary) data analysis in the first revision of the original text. Consequently, the effort to shorten the work has been important.

We consider ed the example given by the reviewer, but eliminating ENTER method from Table 6 would only account for a reduction equal to 25-30 words. Taking into account this and that eliminating the analysis would reduce the quality of the manuscript and remove data already approved by Reviewer 2, we did not modify Table 6.

* Ploos One Instructions for authors: "Length: Manuscripts can be any length. There are no restrictions on word count, number of figures, or amount of supporting information.We encourage you to present and discuss your findings concisely."

Word count

Section Previous revised version New (Actual) revised version

Abstract 250 250

Introduction 491 487

Material and Methods (*) 1561 1493

Results (*) 1266 1181

Discussion 1907 1783

Conclusion 195 192

SUBTOTAL (Introduction - Conclusion) 5420 5136

Acknowledgments 17 17

References 1214 (37 references) 1188 (37 references)

Supporting information 3512 805

TOTAL WORD FILE 10413 7396

* without considering tables and legend of figures

Submitted filename: Response to reviewers.docx

Click here for additional data file.

5 Dec 2019

Aortic pressure and forward and backward wave components in children, adolescents and young-adults: agreement between brachial oscillometry, radial and carotid tonometry data and analysis of factors associated with their differences

PONE-D-19-22199R2

Dear Dr. Bia,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Giacomo Pucci

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

11 Dec 2019

PONE-D-19-22199R2

Aortic pressure and forward and backward wave components in children, adolescents and young-adults: agreement between brachial oscillometry, radial and carotid tonometry data and analysis of factors associated with their differences

Dear Dr. Bia:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Giacomo Pucci

Academic Editor

PLOS ONE