**3. Results**

#### *3.1. Arterial System at 6 and 18 y: Comparative Analysis of the Association between Birth or Current Body Size and Early, Intermediate, Late or Global Growth-Related Body Size Changes*

The association between CV z-scores at 6 y and (a) body size at birth and 6 y or (b) body size changes during growing-up (0–2 or 0–6y) are shown in Table 2 (children cohort). Both, ΔBWH z-score 0–2y and Δz-BMI 0–6y, were associated with structural and hemodynamic parameters. Positive associations were found between z-pSBP, z-pMBP, z-cSBP, z-cDBP and the ΔBWH z-score 0–2y and Δ z-BMI 0-6y (*p* < 0.05). Negative associations were observed when z-AP, z-AIx and z-AIx@75 were considered. z-Pf was positively associated with Δz-BMI 0-6y (statistical threshold, *p* = 0.053), but the association with ΔBWH z-score 0–2 y did not reach significance (*p* = 0.09). The characteristics and statistical significance of the associations between anthropometric changes and structural properties varied, depending (among on other factors) on the structural parameter considered. No significant associations were found between stiffness z-scores and ΔBWH z-score 0-2 y and Δz-BMI 0–6 y (Table 2)


**Table 2.** Comparative analysis CV variables' association with anthropometric characteristics at birth, at 6 y and with anthropometric changes within that period

z-cfPWV 0.00 0.991 –0.03 0.775 –0.02 0.790 0.00 1.00 0.649 0.876 0.989 0.959 0.610 z: z-score. BW, BH: body weight and height. Δ: change in the analyzed period. BMI: body mass index. SBP, DBP, PP, MBP: systolic, diastolic, pulse and mean pressure (p: peripheral, c: central). AIx, AIx@75: aortic augmentation index without and with heart rate adjustment. AP: augmented pressure. Pf, Pb: forward and backward pressure components. R: Right. L: Left. CCA, CFA: common carotid and femoral artery. EM: elastic modulus. IMT: intima-media thickness. DD, SD: diastolic and systolic diameter. cfPWV: carotid-femoral pulse wave velocity. *R*: Pearson coefficient. *p* < 0.05 (red text) was considered significant. In Zero-order correlations columns, the numbers in square brackets define de groups. Then in "William Test (Comparison of Correlations)" (also in columns in the same table), groups are compared in pairs cnsidering groups numbers. Identical comment for next Tables.

Unlike BWH at birth, current z-BMI (6y) was associated with hemodynamic and structural parameters. Furthermore, whereas BWH at birth did not show significant associations with hemodynamic variables, current z-BMI was associated (*p* < 0.05) with almost all of them (pBP, cBPandwave-derived parameters) (Table 2).

Compared to ΔBWH z-score 0–2y and Δz-BMI 0–6y, current z-BMI levels (at 6 y) were more strongly associated with CV z-scores. For example, R coe fficients for the association between z-pSBP and ΔBWH z-score 0–2y, Δz-BMI 0–6y and current z-BMI (6 y) were 0.16, 0.20 and 0.32, respectively (*p* < 0.05). When compared, the association was stronger in case of z-BMI at6y(*p* = 0.019 and *p* = 0.009) (Table 2). The strength of association between arterial parameters at 6 y and ΔBWH z-score 0–2y and Δz-BMI 0–6y did not show statistical di fferences, with the only exception of z-Right CCA SD, DD and IMT that showed stronger association with 0–6y changes (*p* = 0.026, *p* = 0.002 and *p* < 0.01, respectively).

In adolescents, CVz-scores showed associations with current z-BMI, while almost no association was observed between arterial parameters at 18 y and prior body size characteristics (i.e., BWH at birth and Δz-BMI 0-6y) (Table 3). On the other hand, in case of variables associated with Δz-BMI 0–18y (z-pSBP, z-pPP, z-CCA SD and DD) or BWH z-score at birth (z-Right CCA EM and z-Left CCA EM), the comparative analysis (William's test) showed that the strongest associations were obtained when considering current z-BMI (18 y) (Table 3).

Jointly analyzing data from both cohorts it was observed that the strength of the associations between CV z-scores and current z-BMI (6 or 18 y), were always greater than those obtained for any change in body size between birth and the time of the study (0–2, 0–6 or 6–18 y) (Tables 2 and 3).

As mentioned, arterial properties at 6 y were associated with body size changes in that life period (i.e., 0–2 and 0–6 y) whereas the CV properties in subjects 18 y showed almost no association with prior (i.e., Δz-BMI 0–6, 6–18 and 0–18 y) anthropometric conditions (Table 4). When the cohorts were statistically compared, it was observed that for the same "body change" ( Δz-BMI 0–6 y), associations were significant for almost all the studied variables when subjects were 6 y, but not when they were 18 y (Table 4).

Thus, as the subject's age increases, the association between CV z-scores and prior anthropometric changes (i.e., during childhood) decreases (Table 4).

#### *3.2. Arterial Structure and Function at 6 y: Independent Association with* Δ*BWH z-Score 0–2 y*

Associations between hemodynamic and structural parameters at 6 y and ΔBWH z-score 0–2 y kept significant after controlling for BWH z-score at birth (Table 5). After adjusting for BWH z-score at birth and current z-BMI only associations with z-cSBP (but not with z-pSBP, *p* = 0.103) remained significant ( *R* = 0.14, *p* = 0.041) (Table 5). Thus, the association between ΔBWH z-score 0-2 y and cSBP, while weak, is independent of size at birth and at the time of CV study. AIx@75 (*p* = 0.009) and Pf (*p* = 0.052) showed significant associations after controlling for body size at birth (Table 5).

As mentioned, significant positive associations were found between ΔBWH z-score 0–2 y and structural parameters. Disregarding BWH z-scores at birth, ΔBWH z-score 0-2y values were positively associated with z-IMT (both CCA) and z-diameters (both CFA, left CCA). Then, the greater the body size change within the first 2 y, the higher the diameters and wall thickness (Table 5). The associations between ΔBWH z-score 0–2 y and z-IMT remained significant after adjusting for BWH at birth and current z-BMI (6 y). Then, regardless of nutritional status at birth and at the time of arterial evaluation, z-IMT levels at 6 y are influenced by ΔBWH z-score 0–2y (Table 5).

There were no significant associations between ΔBWH z-score 0–2 y and sti ffness z-scores, before (zero-order correlations) and after (partial correlations) adjusting for BWH z-scores at birth and/or current z-BMI (Table 5).


**Table 3.** Comparative analysis CV variables' association with anthropometric characteristics at birth, at 18 y and with anthropometric changes within that period

c: central). AIx, AIx@75: augmentation index without and with heart rate adjustment. AP: augmented pressure. Pf, Pb: forward and backwardpressure components. R: Right. L: Left. CCA, CFA: common carotid and femoral artery. EM: elastic modulus. IMT: intima-media thickness. DD, SD: diastolic and systolic diameter. cfPWV: carotid-femoral pulse wave velocity. R: Pearson coefficient. *p* < 0.05 (red) was considered significant.

#### *J. Cardiovasc. Dev. Dis.* **2019**, *6*, 33


L: Left. CCA, CFA: common carotid and femoral artery. EM: elastic modulus. IMT: intima-media thickness. DD, SD: diastolic and systolic diameter. cfPWV: carotid-femoral pulse wave

velocity. *R*: Pearson coefficient. *p* < 0.05 (red) was considered significant.

*J. Cardiovasc. Dev. Dis.* **2019**,

*6*, 33

#### *3.3. Arterial Structure and Function at 6 and 18 y: Independent Association with* Δ*z-BMI(Body Mass Index) 0–6 y*

Table 6 shows correlations between CV z-scores (at 6 and 18 y) and Δz-BMI 0-6y. In children, z-pSBP, z-pDBP, z-pMBP, z-cSBP, z-cDPB and z-cMBP correlated (positively) with Δz-BMI 0–6 y. Associations remained significant after controlling for BWH at birth, but showed dependence on current z-BMI. AIx and AIx@75 showed significant negative associations when controlling for BWH at birth but when controlling for z-BMI at 6 y, positive correlations were observed ( *R*=0.14, *p* = 0.034; *R*=0.19, *p* = 0.005). Both z-CCA IMT were associated with Δz-BMI 0–6 y, with independence of BWH at birth and current z-BMI. Similar results were observed for right and left CFA IMT and DD.

CV parameters (z-scores) at 18 ywere not associated (zero-order correlation) with Δz-BMI 0–6 y. However, several correlations became significant after adjusting for BWH z-scores at birth. Then, body size changes within 0–6 y would contribute to explain CV characteristics at 18 y. Unlike the observed at 6 y, CVparameters at 18 ywere not associated with Δz-BMI 0–6 y (except z-Pb and z-left CCA IMT) after adjusting for body size at birth and current z-BMI (Table 6)

#### *3.4. Arterial Structure and Function at 18 y: Independent Association with* Δ*z-BMI 0–18 y and 6–18 y*

Tables 7 and 8 show the association between CV z-scores and Δz-BMI 0–18 and 6–18 y. There were no independent associations between CV properties at 18 y and overall (0–18 y) or late (6–18 y) anthropometric (z-BMI) changes, being the only exceptions z-pPP (for Δz-BMI 0–18 y; *p* = 0.017) and z-Right CCA SD (for Δz-BMI 6–18 y; *p* = 0.035). Then, global changes in body size from birth (0–18 y) or childhood (6–18 y) until late in adolescence would not contribute to explain CV characteristics at the beginning of adulthood, with independence of birth size or current z-BMI.

#### *3.5. Hemodynamic and Arterial Properties at 6 and 18 y: Hierarchical Impact of Anthropometric Variables*

MLRmodels allowed analyzing whether growth-related body size changes contribute to explain CV properties, considering and comparing BWH z-score at birth, current z-BMI and the interaction of variables (Supplementary Tables S3–S12). In children, cBP and pBP variables (z-SBP, z-DBP, z-PP, z-MBP) were only explained by z-BMI, while wave-derived parameters were explained by z-BMI (z-AIx@75, z-Pf, z-Pb), ΔBWH z-score 0–2 y (z-AIx, z-AP), Δz-BMI 0–6 y (z-AIx) and by Δz-BMI 0–6 y and z-BMI interaction (z-AP). Structural variables were mainly explained by z-BMI; arterial sti ffness showed no association with body size parameters (Supplementary Tables S3–S6).

In the adolescents: (1) Δz-BMI 0–6 y showed explanatory capacity when interacting with current z-BMI (both z-CFA IMT) (Supplementary Tables S7 and S8); (2) Δz-BMI 0–18 y showed explanatory capacity when interacting with current z-BMI (z-cDBP, z-Right CFA diameters) or BWH z-score at birth (z-cDBP, z-cPP) (Supplementary Tables S9 and S10) and (3) Δz-BMI 6–18 ydid not showsignificant explanatory capacity for the studied variables (Supplementary Tables S11 and S12). Current z-BMI was always the variable with the greatest explanatory power. The explanatory capacity of the interactions between z-BMI (or z-BWH at birth) and Δz-BMI 0-6, 0–18 or 6–18 y was limited. Compared to current z-BMI, intermediate (0–6 y), late (6–18 y) or global (0–18 y) body-size changes showed almost no explanatory capacityfor interindividual variations in CV properties at 18 y.

The joint analysis of both cohorts showed that CV variables were mainly associated with current (i.e., 6 or 18 y) z-BMI. Body-size changes showed little individual explanatory power and their contribution was mainly relatedto the interaction with z-BMI at the time of CVevaluation.

#### *3.6. Arterial Function at 6 and 18 y: Impact of Body Size Changes vs. Anthropometric and Cardiovascular Risk Factors (CRFs)*

Table 9 and Supplementary Table S13 show explanatory models for CV parameters in children. pBP parameters were explained by current z-BMI, while cBP parameters and wave-derivedindexes were mainly associated with z-pSBP. Then, early (0–2 y) or intermediate (0–6 y) body size changes

contributed little to explain cBP or pBPvariations found at 6 y compared to the conditions associated with CV risk (i.e., BMI and pSBP) at the time of the CV study.



z: z-score. BW, BH: body weight and height. BMI: body mass index. SBP, DBP, PP, MBP: systolic, diastolic, pulse and mean pressure (p: peripheral, c: central). AIx, AIx@75: aortic augmentation index without and with heart rate adjustment. AP: augmented pressure. Pf, Pb: forward and backward pressure components. CCA, CFA: common carotid and femoral artery. EM: elastic modulus. IMT: intima-media thickness. DD, SD: diastolic and systolic diameter. cfPWV: carotid-femoral pulse wave velocity. *R*: Pearson coefficient. *p* < 0.05 (red) was considered significant. Partial correlations controlling for: 1 BW-for-BH (length) z-score at birth; 2 BW-for-BH (length) z-score at birth and z-BMI at the time of measurement (6 y).


**Table 6.** Associations between CV z-scores and z-BMIvariations between 0–6 y(both cohorts).

z: z-score. BMI: body mass index. SBP, DBP, PP, MBP: systolic, diastolic, pulse and mean pressure (p: peripheral, c: central). AIx, AIx@75: augmentation index without and with heart rate adjustment. AP: augmented pressure. Pf,Pb: forward and backward pressure components. R: Right. L: Left. CCA, CFA: common carotid and femoral artery. EM: elastic modulus. IMT: intima-media thickness. DD, SD: diastolic and systolic diameter. cfPWV: carotid-femoral pulse wave velocity. *R*: Pearson coefficient. *p* < 0.05 (red text) was considered significant. Partial correlations controlling for: 1 BW-for-BHz-score at birth; 2 BWH z-score at birth and z-BMI at the time of CV study; 3 BWH z-score at birth and z-BMI variation between 6 and 18 y.

**Table 7.** Associations between CVz-scores and z-BMI variations between 0–18y(adolescent cohort, *n* = 340).




z: z-score. BW, BH: body weight and height. BMI: body mass index. SBP, DBP, PP, MBP: systolic, diastolic, pulse and mean pressure (p: peripheral, c: central). AIx, AIx@75: augmentation index without and with heart rate adjustment. AP: augmented pressure. Pf, Pb: forward and backward pressure components. CCA, CFA: common carotid and femoral artery. EM: elastic modulus. IMT: intima-media thickness. DD, SD: diastolic and systolic diameter. cfPWV: carotid-femoral pulse wave velocity. *R*: Pearson coe fficient. *p* < 0.05 (red) was considered significant. Partial correlations controlling for: 1 BW-for-BH (length) z-score at birth; 2 BW-for-BH z-score at birth and z-BMI at the time of CV study.

**Table 8.** Associations between CV z-scores and z-BMI variations between 6–18 y (adolescent cohort, *n* = 340).


z: z-score. BW, BH: body weight and height. BMI: body mass index. SBP, DBP, PP, MBP: systolic, diastolic, pulse and mean pressure (p: peripheral, c: central). AIx, AIx@75: augmentation index without and with heart rate adjustment. AP: augmented pressure. Pf, Pb: forward and backward pressure components. CCA, CFA: common carotid and femoral artery. EM: elastic modulus. IMT: intima-media thickness. cfPWV: carotid-femoral pulse wave velocity. DD, SD: diastolic and systolic diameter. *R*: Pearson coe fficient. *p* < 0.05 (red) was considered significant. Partial correlations controlling for: 1 BW-for-BH (lenght) z-score at birth; 2 BW-for-BH z-score at birth and z-BMI at the time of CV study (18 y).


**Table 9.** Multiple regression analysis between CVvariables (dependent) and anthropometric and CRFsvariables (independent) in children cohort (*n* = 632).


**Table 9.** *Cont*.

βu and βs: un- and standardized coe fficients. R: Pearson coe fficient. Adj R2: adjusted squared R. SE: Standard Error. VIF: variance inflation factor. z-: z-score. BMI: body mass index. SBP, DBP, PP, MBP: systolic, diastolic, pulse and mean pressure (p: peripheral, c: central). AIx, AIx@75: augmentation index without and with heart rate adjustment. AP: augmented pressure. Pf,Pb: forward and backward pressure components. CCA, CFA: common carotid and femoral artery. EM: elastic modulus. IMT: intima-media thickness. DD, SD: diastolic and systolic diameter. cfPWV: carotid-femoral pulse wave velocity. *p* < 0.05 was considered statistically significant. Variables entered in the model (forward method): z-BMI, z-BWH at birth, ΔBWH z-score 0–2y, Δ-zBMI0–6y, Sex (1: female, 0: male), z-pSBP, Hypertension (yes:1, no: 0). Interaction between growth parameters and z-BMI and z-BWH at birth were entered in the model if they showed significant association (*p* < 0.005) in the multiple linear regression. Only significant (*p* < 0.05) variables entered in the model are shown.

Explanatory variables for variations in structural parameters showed heterogeneity. At 6 y inter-individual variations in CCA SD and DD were associated with current z-BMI and z-pSBP. When considering CFA z-diameters or z-IMT levels the explanatory variables were: (a) current z-BMI (for z-left CCA IMT, CFA diameter), (b) Δz-BMI 0-6y (for z-Right CFA IMT and diameters), (c) ΔBWH z-score 0–2 y (for z-Right CCA IMT) and (d) BWH z-score at birth (for z-Right CCA IMT). Sti ffness variations were not associated with anthropometric variations (Table 9). For the adolescents cohort (Table 10, Supplementary Table S14), the independent variables that remained significant in all the models (*p* < 0.05) were z-pSBP and current z-BMI (Table 10). Then, at 18 y nor BWH z-score at birth, nor the bodily changes during growth (early, intermediate, late or global), explained the CV interindividual variations (with the only exception being the interaction between current z-BMI and Δz-BMI 0-6 y for z-Right CCA IMT) (Table 10).

From the joint analysis of data from both cohorts and considering the exposure to CRFs, it can be asserted that: (1) current z-BMI was the variable mostly associated with CV characteristics; (2) the older the subject, CV properties (e.g., arterial structure) are less explained by changes in body size during the early growth phases.


**Table 10.** Multiple regression analysis between CVvariables (dependent) and anthropometric and CRFsvariables (independent) in adolescent cohort (*n* = 340).


**Table 10.** *Cont*.

βu and βs: un- and standardized coefficients. R: Pearson coefficient. Adj R2: adjusted squaredR. SE: Standard Error. VIF: variance inflation factor. z-: z-score. BMI: body mass index. SBP, DBP, PP, MBP: systolic, diastolic, pulse and mean pressure (p: peripheral, c: central). AIx, AIx@75: augmentation index without and with heart rate adjustment. AP: augmented pressure. Pf, Pb: forward and backward pressure components. CCA, CFA: common carotid and femoral artery. EM: elastic modulus. IMT: intima-media thickness. DD, SD: diastolic and systolic diameter. cfPWV: carotid-femoral pulse wave velocity. *p* < 0.05 was considered statistically significant. Variables entered in the model (forward method): z-BMI, BWH at birth, Δ-zBMI0–6y, Sex (1: female, 0: male), z-pSBP, Hypertension (1: yes, 0: no), Dislypemia (1: yes, 0: no), Smoking (1: yes, 0: no), Sedentarism (1: yes, 0: no). Interactions between growth parameters and z-BMI or z-BWH were entered in the model if they showed significant association in multiple linear regressions. Only significant (*p* < 0.05) independent variables entered in the models are shown.
