Carotid‐Femoral Pulse Wave Velocity in Healthy Spanish Children: Reference Percentile Curves

Abstract

Carotid‐femoral pulse wave velocity (cfPWV) measurement is an appropriate method for determining arterial stiffness and is a useful tool for early detection of cardiovascular disease. However, the lack of reference values due to the difficulty in accessing healthy child populations, among other causes, has limited its use in clinical practice. The aim of this study was to create reference cfPWV percentile curves for healthy children. The initial sample consisted of 350 girls and boys aged 8 to 11 years. The cfPWV per age and sex were generated using the lambda‐mu‐sigma (LMS) technique. The effects of sex, age, arterial pressure, and body mass index were taken into account. The main result of this study is the age‐ and height‐specific cfPWV reference percentile curves for girls and boys. Curves were obtained for both sexes, since sexual differences were observed in growth and development rates that may affect cfPWV.

As people age, arterial walls lose elasticity. This is due to changes in morphology and in the composition of their main structural proteins, elastin and collagen, which favor hypertension development,1 one of the major risk factors in the onset of CV (CV) disease. Although these variations are mainly attributed to old age, there is increasing scientific evidence of the influence exerted by factors that are already active in the early stages of the life cycle, raising CV risk (CVR) in adult stages. Higher arterial stiffness values were observed in obese children,2, 3 those with type 1 diabetes mellitus4 and those with kidney disease.5 Given that most human organs and physiological systems begin to develop at early stages of fetal life and reach functional maturity after birth, interactions can occur between these developmental processes, along with a variety of environmental factors such as nutrition, exposure to toxic substances, and epigenetic alterations, which, in the case of the circulatory system, affect the final artery wall structure.6, 7, 8, 9, 10 Therefore, the conditions governing the processes of growth and development in the early prenatal and postnatal stages have an impact on CVR in adults.11, 12

The elastic fibers in arteries begin to form at the beginning of the fetal stage, and the rate of elastin synthesis peaks at the end of the third 3‐month period of gestation. From this moment, the rate of elastin synthesis decreases up to the conclusion of the process of growth and development at adolescence.13, 14 In adverse environmental conditions during this stage (eg, malnutrition, disease), the synthesis and organization of elastin in the blood vessels is affected and causes permanent alterations in their mechanical properties, predisposing patients to higher CVR in adulthood.1, 13, 14 The information available to date appears to support the hypothesis that preventive action against CVR should begin in the initial stages of the life cycle, and that it is highly useful in clinical practice to use noninvasive methods to detect preclinical vascular lesions.15

Until recently, the ease with which peripheral arterial pressure is measured and its correlation with the development of CV episodes has restricted the use of methods for the direct assessment of vessels. Lately, however, noninvasive methods have been developed for studying arterial damage, among which measuring arterial stiffness is a useful marker for early arterial lesion, and its predictive and prognostic value has been recognized in the literature.16, 17, 18

Measuring carotid‐femoral pulse wave velocity (cfPWV) by tonometry is an appropriate method for determining arterial stiffness of large arteries and its use is recommended by the European Society of Hypertension (ESH) and the European Society of Cardiology.16, 19, 20 It provides an important prognostic risk marker and is a useful tool for early detection of CV disease, allowing better characterization of actual CV damage and more appropriate and rapid intervention in affected individuals.

In adults, cfPWV is closely associated with the presence and extension of atherosclerosis and increases in the presence of other risk factors such as obesity and arterial hypertension (AHT).21, 22

The new ESH guideline published in 2013 provides reference cfPWV values for adults.23 In adult hypertensive patients, a cfPWV value exceeding 10 m/s may be interpreted as a marker of subclinical damage to the artery.20, 23, 24, 25 However, for healthy children and adolescents, cfPWV reference values are still scarce,26, 27 and are nonexistent for the Spanish child population. This lack of references obtained from healthy child populations has limited their usage in clinical practice.

The aim of this study was to create cfPWV tables and reference percentile curves for a healthy Spanish child population, controlling for the effects of sex, age, height, excess weight, and arterial pressure, factors known to affect cfPWV values in each individual.

Methods

Population studied and Sample Characteristics

Data were gathered at public and subsidized schools in the Autonomous Community of Madrid between 2010 and 2013. The educational centers were approached beforehand through the school board, which granted approval for the study to be conducted. Previously, and observing the ethical standards for research on human beings set forth in the Helsinki Protocol28 and Organic Law 15/1999 of 13 December on personal data protection, the research project report was submitted to the ethical committee of the Universidad Autónoma de Madrid for approval. Participation in the study was voluntary and the families, before data gathering commenced, signed informed consent, in which they allowed their sons and daughters to be included in this research. Only girls and boys with no history of artery diseases or hypertension and without current antihypertensive medication were included in the study. Girls with menarche onset were also excluded. Information on the history of CV disease, hypertension, and pharmacologic treatment due to these pathologies for each child was obtained from personal interviews with each family. Likewise, girls were questioned about the presence or absence of menarche and their answers were confirmed by families.

The initial sample consisted of 350 children (182 boys and 168 girls) with age ranging from 8 to 11 years, with an average age of 10.18 years (standard=1.07) for boys and 10.07 years (standard=1.14) for girls (t=0.911; P=.363).

The following variables were considered in the study:

Anthropometric Variables

Weight (kg) and height (cm) were measured according to the International Biological Program guidelines.29 All measurements were taken by trained personnel in the Department of Biology, Universidad Autónoma de Madrid, using certified and approved instruments. Height was measured with a GPM anthropometer (Zürich, Switzerland), with a range of 0 to 2100 mm and precision of 0.1 mm, and weight was measured on digital scales with a range of 0 to 130 kg and precision of 100 g.

Subcutaneous biceps (mm), triceps (mm), and subscapular (mm) and suprailiac (mm) fat skinfolds were measured using a Holtain skinfold caliper (Crosswell, UK) with a range of 0 to 48 mm and precision of 0.2 mm.

Anthropometric Indices

From the values for weight and height, the body mass index (BMI) was calculated.30 In addition, Z‐scores for weight, height, and BMI were calculated using the 2007 World Health Organization values as reference.31, 32 These Z‐values were used to classify the children according to underweight (<−2 standard deviations [SDs]), normal weight (≥−2 SD and <1 SD), excess weight (≥1 SD and <2 SD), and obese (≥2 SD).

The body fat mass percentage (BFM%) was calculated from biceps, triceps, and subscapular and suprailiac subcutaneous skinfolds, after calculating body density (D) with Brook's equation33:

$$\text{Boys: D}{\text{(kg/cm}}^3)=1.1690-0.0788\times \text{log}\left(\mathrm{\Sigma }\text{skinfolds}\right)$$

$$\text{Girls: D}{\text{(kg/cm}}^3)=1.2063-0.0999\times \text{log}\left(\mathrm{\Sigma }\text{skinfolds}\right)$$

Having obtained D, calculations were made to find the BFM% using the expression proposed by Siri34:

$$\text{BFM}\%=[(4.95/\mathrm{D})-4.50]\times 100$$

Arterial Pressure

Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured with a Panasonic Diagnostec EW‐BU30 automatic oscillometric tensometer (Osaka, Japan) (0 to 280 mm Hg). Measurements were taken in all children in the morning while in a sitting position, back supported, with the left arm unclothed and supported at heart level, legs uncrossed, feet resting on the floor, and after at least 5 minutes of rest beforehand. The arm circumference was measured at the midpoint between the acromion and olecranon for selecting the appropriate cuff. A Diagnostec child cuff arm with a circumference from 17 to 30 cm was used. Two consecutive measurements were registered and the mean of the two records was used as the clinical SBP and DBP. In cases yielding high blood pressure values, the family was informed and a visit to the pediatrician was recommended. The values obtained for each boy and girl were compared against the reference values for sex, age, and height given in The Fourth Report on the Diagnosis, Evaluation, and Treatment of High Blood Pressure in Children and Adolescents.35 Patients with blood pressure values below the standard percentile 95 (p95) were placed in the normotensive category, and children with SBP or DBP above the standard percentile 95 (≥p95) were considered hypertensive. Finally, mean arterial pressure (MAP) was calculated by the formula [SBP+(2DBP)]/3.

Excess Weight and Hypertension

Children with obesity and arterial hypertension (AHT) were identified, and a new variable was created and labelled CV risk factor (CVRF), comprising four categories: (1) presence of obesity, (2) presence of AHT, (3) presence of both, and (4) absence of both.

Arterial Stiffness

cfPWV was determined by means of the SphygmoCor system (AtCor Medical, West Ryde, Australia).36 cfPWV measurements were taken at a transcutaneous level by means of a tonometer, which, through applanation, gathers arterial pulse readings at two arterial points, between the carotid artery, a direct branch of the aorta, and the right femoral artery, combining the two at a fixed point in the cardiac cycle by means of an electrocardiographic (ECG) reading. Monitoring the ECG results during measurement ensured detection by the system of the pulse wave initiation point. Measurements were taken at the right carotid artery, 1 cm below the carotid bulb, with the individual in a supine decubitus position for 5 minutes before measurement, with the head turned to the left 45 degrees.16, 36 Previously, the distance in millimeters between the two arterial (carotid and femoral) points was measured on the body surface in order to relate this to the time lapse obtained and thus establish a distance/time ratio. The best distance estimate in adults (and probably also in children) is 80% of the full carotid‐femoral distance.23 Therefore, all cfPWV values should be recalculated to the standard distance to calculate the standard cfPWV.

To prevent any potential methodological bias, a single trained researcher performed all of the cfPWV measurements.

Creation of the Curves: cfPWV Standardization Using the LMS Method

From the 350 children, those presenting with obesity and AHT were excluded with the aim of controlling any possible effects on cfPWV from these variables. Therefore, the final sample used for the cfPWV percentile curves consisted of 277 children. cfPWV by sex and age was generated using the LMS method,37 which categorizes the distribution of a variable according to its median (M), the coefficient of variation (S, ie, the relation between SD and the mean), and the asymmetry (L), required to transform data to normal distribution by means of a maximum‐likelihood curve adjustment algorithm to the original data plotted over the independent variable. Decimal age was used, calculated from the date on which the survey was conducted and from the date of birth.

Statistical Analyses

With the variables obtained, a database was built and statistically analyzed using the SPSS 20.0 program (SPSS Inc, Armonk, NY). The Kolmogorov‐Smirnov test was used to verify, prior to statistical analyses, the normal distribution of quantitative variables. For the comparison of average weight (kg), height (cm), BMI (kg/m²), BFM%, DBP (mm Hg), SBP (mm Hg), and PWV (m/s) between sexes and for the comparison of average cfPWV values in patients with CVRF and those without CVRF, the Student t test was used for independent samples. The differences between average cfPWV values by age groups were analyzed by means of analysis of variance. The relationship between qualitative variables was analyzed with the chi‐square test.

The relationships between cfPWV and the remainder of anthropometric and hemodynamic variables were measured. Those variables displaying a significant correlation with cfPWV were treated as independent variables in multiple linear regression models in order to establish predictive models for cfPWV.

A P value <.05 was considered statistically significant.

Results

The anthropometric and hemodynamic variables for both sexes are described in Table 1. No significant differences were observed between boys and girls in age, cfPWV, SBP, DBP, weight, height, or BMI, and the percentile distribution was found to be similar in both sexes. However, mean BFM% values were found to be significantly higher for girls than for boys.

Table 1.

Anthropometric and Hemodynamic Variables and Their Percentile Distribution in Both Sexes (N=350 [182 Boys and 168 Girls])

Parameters	Mean	SD	P Value	3rd	10th	25th	50th	75th	90th	97th
Age, y
Boys	10.18	1.08	.363	8.33	8.76	9.31	10.08	11.09	11.63	11.93
Girls	10.07	1.14		8.47	8.69	9.13	9.81	11.22	11.75	11.88
cfPWV, m/s
Boys	5.00	0.83	.870	3.57	4.00	4.40	5.04	5.52	5.92	6.42
Girls	5.02	0.83		3.68	4.00	4.44	4.88	5.52	6.13	6.70
SBP, mm Hg
Boys	103.34	15.15	.777	71.04	81.60	94.00	104.00	114.00	121.00	133.84
Girls	102.87	14.23		76.68	85.00	93.00	105.00	112.00	120.00	130.64
DBP, mm Hg
Boys	62.44	10.88	.696	36.24	50.00	56.00	62.00	70.00	75.20	84.88
Girls	62.99	10.74		40.72	49.60	57.00	63.00	69.00	75.40	80.32
MAP, mm Hg
Boys	76.19	11.11	.861	53.00	62.53	70.67	76.67	83.00	89.93	94.99
Girls	76.11	10.57		56.13	63.00	69.00	77.33	83.67	88.33	92.88
Weight, kg
Boys	37.78	8.20	.638	24.92	27.68	31.30	37.00	42.80	49.12	55.88
Girls	37.33	9.17		22.97	26.56	29.90	35.90	43.30	51.02	55.76
Height, cm
Boys	142.47	8.04	.362	125.92	132.14	137.00	142.50	148.50	152.44	157.82
Girls	141.55	9.75		125.91	129.88	134.60	140.40	147.60	155.04	163.05
BMI, kg/m2
Boys	18.43	2.67	.939	14.44	15.42	16.48	17.95	19.96	22.15	24.35
Girls	18.41	3.00		14.21	15.00	16.11	17.61	20.50	23.06	24.99
BFM, %
Boys	23.05	6.93	.002a	11.95	14.22	18.06	22.47	27.69	33.69	37.17
Girls	25.57	7.74		11.50	14.66	20.38	24.81	30.96	35.96	39.82

Abbreviations: BFM, body fat mass; BMI, body mass index; cfPWV, carotid‐femoral pulse wave velocity; DBP, diastolic blood pressure; MAP, mean arterial pressure; SBP, systolic blood pressure; SD, standard deviation. ^a P<.05 difference between boys vs girls in mean value.

Table 2 shows the average cfPWV values according to sex and age. Significant average cfPWV value differences were observed between children aged 8, 9, and 10 years and those aged 11 years both in boys and girls (Table 2).

Table 2.

Mean Values of cfPWV by Sex and Age for the Total Sample (N=350 [182 Boys and 168 Girls])

	cfPWV, m/s
	8 y			9 y			10 y			11 y			P Value
	No.	Mean	SD	No.	Mean	SD	No.	Mean	SD	No.	Mean	SD
Boys	32	4.79	0.71	51	4.88	0.72	48	4.98	0.87	51	5.27a	0.93	.034a
Girls	37	4.82	0.83	54	4.92	0.78	33	4.91	0.75	44	5.39a	0.85	.006a
P value	.848			.814			.734			.528

Abbreviations: cfPWV, carotid‐femoral pulse wave velocity; SD, standard deviation. ^a P<.05 difference between 11‐year age group and the other ages.

The prevalence of total excess weight (excess weight and obesity) per age group is shown in Table 3, where significant differences among age groups were also not observed. Similarly, the prevalence of AHT in the sample was not significantly different in the two sexes, nor among age groups within each sex (Table 3).

Table 3.

Prevalence of Overweight, Obesity, and Hypertension in the Total Sample (N=350 [182 Boys and 168 Girls])

Age, y	No.	Body Mass Index			Arterial Pressure
		Normal Weight, % (No.)	Overweight, % (No.)	Obesity, % (No.)	Normal Blood Pressure, % (No.)	Hypertension, % (No.)
Boys
8	32	65.6 (21)	25.0 (8)	9.4 (4)	84.4 (27)	15.6 (5)
9	51	68.1 (35)	14.9 (7)	17.0 (9)	86.3 (44)	13.7 (7)
10	48	56.3 (27)	35.4 (17)	8.3 (4)	81.3 (39)	18.8 (9)
11	51	54.9 (28)	33.3 (17)	11.8 (6)	90.2 (46)	9.8 (5)
Total	182	61.0 (111)	26.9 (49)	12.1 (22)	85.7 (156)	14,3 (26)
P value		.455			.187
Girls
8	37	73.0 (27)	16.2 (6)	10.8 (4)	91.9 (34)	8.1 (3)
9	54	61.1 (33)	31.5 (17)	7.4 (4)	83.3 (45)	16.7 (9)
10	33	60.6 (20)	24.2 (8)	15.2 (5)	90.9 (30)	9.1 (3)
11	44	65.9 (29)	25.0 (11)	9.1 (4)	84.1 (37)	15.9 (7)
Total	168	64.9 (109)	25.0 (42)	10.1 (17)	86.9 (146)	13.1 (22)
P value		.686			.571

With regard to the variable CVRF, 79.5% of the sample presented with neither obesity nor hypertension, 6.8% presented with obesity, 9.3% with hypertension, and 4.3% with obesity and hypertension together. No significant differences were found for cfPWV among these four categories (F=1.938; P=.123). Nevertheless, when comparing the obesity‐ and hypertension‐free category with a category including any of these risk factors, higher cfPWV values were found in the latter, but among boys this was a tendency without statistical significance (Table 4).

Table 4.

Mean Values of cfPWV by Sex and Presence or Absence of CVRFs for the Total Sample (N=350 [182 Boys and 168 Girls])

		No CVRF (n=277)			CVRF (n=73) (Obesity and AHT)			P Value
		No.	Mean	SD	No.	Mean	SD
cfPWV, m/s	Boys	144	4.98	0.80	38	5.08	0.96	.539
	Girls	133	4.95	0.78	36	5.30	0.94	.022a

Abbreviations: AHT, arterial hypertension; cfPWV, carotid‐femoral pulse wave velocity; CVRFs, cardiovascular risk factors; SD, standard deviation. ^a P<.05.

For this reason, the cfPWV percentile curves were designed taking into consideration only children free from hypertension and obesity in both sexes.

Correlations of cfPWV with anthropometric and hemodynamic measurements were made by sex. In boys, cfPWV correlates positively and significantly with age (r=0.214; P<.001), height (r=0.211; P<.001), MAP (r=0.161; P<.05), and SBP (r=0.160; P<.05). In girls, cfPWV correlates positively and significantly with age (r=0.265; P<.001), height (r=0.228; P<.001), weight (r=0.183; P<.05), SBP (r=0.218; P<.001), DBP (r=0.217; P<.001), and MAP (r=0.245; P<.001).

Regression analyses were performed for boys and girls taking cfPWV as a dependent variable. Age, height, weight, SBP, DBP, and MAP were taken as explanatory variables. For boys, the only independent explanatory variable for cfPWV was height (β=0.022; standard coefficient=0.211; P=.004; R ²=0.054) and for girls, the model retained as explanatory variables MAP (β=0.019; standard coefficient=0.213; P=.005), and age (β=0.153; standard coefficient=0.205; P=.007), which jointly explained 9% of the variability observed in cfPWV (R ²=0.090).

Therefore, according to our data, the best estimate for cfPWV in Spanish children can be obtained with the following formula:

$$\text{Boys: cfPWV (m/s)}=1.854+\left(\text{height (cm)}\times 0.022\right)$$

$$\text{Girls: cfPWV (m/s)}=2.278+\left(\text{MAP}\times 0.017\right)+\left(\text{Age (years)}\times 0.153\right)$$

Reference cfPWV Values

Given that differences were observed both in the correlations and in the multiple regression for boys and girls, the LMS values and percentiles were determined according to sex, decimal age, and height, excluding individuals presenting with obesity or hypertension (Table 5). On the basis of these values, the percentile curves shown in the Figure were drawn up for cfPWV values in healthy boys and girls.

Table 5.

LMS Values and Specific Percentile Values for cfPWV According to Age and Height for Healthy Children, With No Presence of Obesity and Hypertension or Cardiovascular Risk Factors (N=277 [144 Boys and 133 Girls])

	L	M	S	3rd	10th	25th	50th	75th	90th	97th
Age, y
Boys
8.0	2.37	4.85	0.12	3.53	4.02	4.44	4.85	5.22	5.53	5.81
8.5	1.91	4.88	0.13	3.53	4.00	4.44	4.88	5.28	5.62	5.94
9.0	1.23	4.88	0.14	3.55	3.98	4.41	4.88	5.33	5.73	6.12
9.5	0.58	4.88	0.15	3.60	3.99	4.40	4.88	5.38	5.85	6.33
10.0	0.10	4.93	0.15	3.68	4.04	4.45	4.93	5.47	6.00	6.56
10.5	−0.15	4.97	0.16	3.74	4.09	4.48	4.97	5.53	6.09	6.71
11.0	−0.11	5.06	0.15	3.80	4.16	4.56	5.06	5.62	6.19	6.81
11.5	0.24	5.15	0.16	3.79	4.19	4.63	5.15	5.72	6.28	6.86
12.0	0.72	5.22	0.16	3.69	4.16	4.66	5.22	5.81	6.35	6.90
Girls
8.0	−0.18	4.48	0.16	3.34	3.66	4.03	4.48	4.99	5.50	6.06
8.5	−0.16	4.64	0.16	3.47	3.80	4.18	4.64	5.16	5.68	6.26
9.0	−0.11	4.86	0.15	3.66	4.00	4.39	4.86	5.39	5.93	6.52
9.5	−0.07	4.75	0.15	3.60	3.93	4.30	4.75	5.26	5.77	6.33
10.0	−0.03	4.67	0.15	3.54	3.87	4.23	4.67	5.16	5.65	6.18
10.5	0.02	4.83	0.15	3.65	3.99	4.37	4.83	5.33	5.83	6.36
11.0	0.06	5.06	0.15	3.81	4.18	4.58	5.06	5.59	6.10	6.64
11.5	0.10	5.18	0.15	3.90	4.27	4.69	5.18	5.71	6.23	6.78
12.0	0.14	5.15	0.14	3.91	4.27	4.67	5.15	5.67	6.18	6.73
Height, cm
Boys
120	−3.65E‐02	4.72	0.16	3.53	3.87	4.25	4.72	5.24	5.76	6.33
125	−3.65E‐02	4.73	0.16	3.53	3.88	4.26	4.73	5.25	5.77	6.34
130	−3.65E‐02	4.75	0.16	3.55	3.89	4.28	4.75	5.27	5.80	6.37
135	−3.65E‐02	4.78	0.16	3.58	3.92	4.31	4.78	5.31	5.84	6.41
140	−3.65E‐02	4.83	0.16	3.61	3.96	4.35	4.83	5.36	5.89	6.47
145	−3.65E‐02	4.93	0.16	3.69	4.04	4.44	4.93	5.47	6.02	6.61
150	−3.65E‐02	5.08	0.16	3.80	4.17	4.58	5.08	5.64	6.21	6.82
155	−3.65E‐02	5.26	0.16	3.93	4.31	4.74	5.26	5.84	6.42	7.05
160	−3.65E‐02	5.44	0.16	4.07	4.46	4.90	5.44	6.04	6.64	7.30
Girls
120	4.14E‐03	4.73	0.15	3.54	3.88	4.26	4.73	5.25	5.77	6.33
125	−3.87E‐03	4.76	0.15	3.56	3.90	4.29	4.76	5.28	5.80	6.36
130	−1.19E‐02	4.79	0.15	3.59	3.93	4.32	4.79	5.31	5.84	6.40
135	−1.99E‐02	4.83	0.15	3.62	3.97	4.36	4.83	5.36	5.88	6.46
140	−2.79E‐02	4.88	0.15	3.66	4.01	4.40	4.88	5.41	5.94	6.51
145	−3.59E‐02	4.91	0.15	3.69	4.04	4.43	4.91	5.45	5.98	6.56
150	−4.40E‐02	4.94	0.15	3.72	4.07	4.46	4.94	5.48	6.02	6.60
155	−5.20E‐02	5.00	0.15	3.76	4.12	4.51	5.00	5.54	6.08	6.67
160	−6.00E‐02	5.08	0.15	3.83	4.19	4.59	5.08	5.63	6.18	6.78
165	−6.80E‐02	5.19	0.15	3.91	4.28	4.68	5.19	5.74	6.30	6.91

Abbreviations: cfPWV, carotid‐femoral pulse wave velocity; L, indicates skewness; M, mean; S, coefficient of variation. The coefficient of variation (CV) is calculated as: CV=(standard deviation/mean).

Figure 1.

Carotid‐femoral pulse wave velocity (cfPWV) percentile curves according to sex and age (A) and height (B) in healthy children (N=277).

Discussion

In this study, reference cfPWV curves are proposed for a child population free from CRF. Scarce research of this kind exists for child populations,26, 27, 38 and therefore these reference values are potentially useful in clinical practice for diagnosing early CV risk in childhood and population studies. Several devices have been used to assess arterial stiffness in a clinical setting, the most traditional being the Complior (Alam Medical, Vincennes, France). In this study we used the SphygmoCor device because it allows us to obtain other hemodynamic variables, such as central pressures. These systems used different algorithms to identify the foot pulse waves recorded in the same heart beat at the two sites (carotid and femoral). Such characteristics may result in different values (5% to 15%) when data obtained with the Complior and SphygmoCor systems are compared39. Therefore, our reference values are related to the device used in our study and adjustments have to be made when using other devices as described.40

The percentile curves are presented disaggregated by sex. In other studies, no differences were found in cfPWV among boys and girls, for which reason joint curves were designed for both sexes.38, 41 In this study, likewise, no differences were found between boys and girls for average cfPWV values, but differences were detected in percentiles 90 and 97, where higher values were found for girls than for boys. Regarding height and BFM%, higher values for girls than for boys were also observed in percentiles 90 and 97. Therefore, and in accordance with recommendations by other authors,26, 42 functional measurements of major arteries must be taken, allowing for changes inherent to the process of growth and development in childhood, the juvenile stage, and adolescence and different maturation rhythms in boys and girls. In addition, the results obtained in the multiple linear regression models also indicate that the explanatory variables for cfPWV differ between the two sexes: while for boys, height is the variable that best explains the observed variability in cfPWV, for girls this variable is age and MAP. These results also support the design of separate curves for boys and girls.

Although the total sample size for each analyzed age subgroup may seem small, we considered it important to establish the curves for boys and girls separately. In addition, there is a sample with a small variability of age (8 to 11 years), which excluded obese children, children with high blood pressure, and girls who had reached menarche.

The study was conducted in schools located in neighborhoods with different socioeconomic levels of the community of Madrid. The sampling was not random, but can be considered so, since it included children representing a broad social spectrum of the population of Madrid.

Conclusions

Standard percentile cfPWV curves obtained with the SphygmoCor tonomer are presented for healthy boys and girls aged 8 to 11 years, by age and height. These curves have been designed to control the effects of hypertension and obesity on cfPWV. These curves represent an important contribution to the diagnosis of early vascular damage and may have significant applications in clinical practice as reference values and for the development of further studies. Previous studies have shown that PWV may more accurately predict CV outcomes in adults.27 It is conceivable that stiffer arteries in childhood may predict development of CV outcomes such as high blood pressure in adulthood. However, absence of reference values limits the use of this parameter in clinical settings.

J Clin Hypertens (Greenwich). 2017;19:227–234. 10.1111/jch.12899 © 2016 Wiley Periodicals, Inc.