Progression in Central Blood Pressure and Hemodynamic Parameters and Relationship With Cardiovascular Risk Factors in a Spanish Population: EVA Follow-Up Study

Abstract

BACKGROUND

The progression of central blood pressure (CBP) values and central hemodynamic parameters and its relationship with cardiovascular risk factors is quite unknown. We sought to investigate this association in a Spanish adult population without cardiovascular diseases.

METHODS

Prospective observational research with a 5-year follow-up. Randomly sampled 501 individuals (mean age 56 ± 14 years, 50.3% women). After 5 years, 480 individuals had a follow-up. Measurements taken using the SphygmoCor (AtCor Medical Pty Ltd., Head Office, West Ryde, Australia), following all the recommendations established in the “International task force” (Sharman JE, Avolio AP, Baulmann J, Benetos A, Blacher J, Blizzard CL, Boutouyrie P, Chen CH, Chowienczyk P, Cockcroft JR, Cruickshank JK, Ferreira I, Ghiadoni L, Hughes A, Jankowski P, Laurent S, McDonnell BJ, McEniery C, Millasseau SC, Papaioannou TG, Vlachopoulos C. Validation of non-invasive central blood pressure devices: ARTERY Society task force consensus statement on protocol standardization. Eur Heart J 2017; 38:2805–2812), giving an estimate of CBP relative to measured brachial blood pressure (type 1 device).

RESULTS

Progressions during follow-up: central systolic blood pressure (cSBP): 4.16 ± 13.71 mm Hg; central diastolic blood pressure: 2.45 ± 11.37 mm Hg; central pulse pressure: 1.72 ± 12.43 mm Hg; pulse pressure amplification (PPA): 2.85 ± 12.20 mm Hg; ejection duration: 7.00 ± 47.87 ms; subendocardial viability ratio (SEVR): −8.04 ± 36.24%. In multiple regression analysis: cSBP positively associated with: body mass index (BMI) (β = 0.476); waist size (β = 0.159); number of cigarettes per day (β = 0.192). Inversely associated with peripheral systolic blood pressure (β = −0.282). Central diastolic blood pressure increase positively associated with number of cigarettes per day (β = 0.174). Inversely associated with peripheral diastolic blood pressure (β = −0.292). Central pulse pressure increase positively associated with BMI (β = 0.330). Inversely associated with peripheral pulse pressure (β = −0.262). Pulse pressure amplification increase positively associated with: BMI (β = 0.276); number of cigarettes per day (β = 0.281). Ejection duration progress inversely associated with basal plasma glucose (β = −0.286).

CONCLUSIONS

All measures increased except for SEVR. Progressions in CBP and PPA were positively associated with anthropometric parameters and number of cigarettes and CBP inversely associated with peripheral blood pressure, although this association was different according to sex.

Graphical Abstract

Open in new tabDownload slide

The measurement of central blood pressure (CBP) by non-invasive techniques is important for assessing the pressure on vital organs and vessels prone to atherosclerosis. Central blood pressure levels, like central sytolic blood pressure (cSBP) and pulse pressure amplification (PPA), obtained with noninvasive approaches generally underestimate the true values of cSBP and PPA, and the higher the invasive cSBP or PPA, the greater the underestimation.^1,2 These differences have made it necessary to create an “international task force” with the aim of providing recommendations on the accuracy (validation) of CBP devices.³ Currently the non-invasive measurement of central hemodynamic parameters is increasingly used in research studies, given the impossibility of performing invasive measurement on a general basis due to patient discomfort and cost. Furthermore, values and response to treatment differ from peripheral blood pressure (PBP).^4,5 Central blood pressure is influenced by arterial stiffness and pulse wave reflection, which are associated with age, gender and vascular aging.⁵ Recent evidence suggests that CBP are independent predictor of cardiovascular disease and all-cause mortality.⁶ However, there is debate on whether CBP is more predictive than PBP.^4,5 Most recent papers suggest that CBP and parameters derived from pulse wave analysis are better at predicting cardiovascular diseases, target organ damage, brain injury or predicting new onset hypertension than PBP.^4–10 Hence, the incremental prognostic value of CBP over PBP is unclear.

Papers describing baseline values of CBP and central hemodynamic parameters are sparse. Herbert et al.¹¹ analyzed the cSBP values in a total of 45,436 healthy subjects and showed that obesity had a direct association with CBP. Paiva et al.¹² analyzed age- and sex-stratified baseline values in subjects aged 18–90 years. Nevertheless, few studies have analyzed the central hemodynamic parameters in one same sample. Weber et al. 2019¹³ examined data from 2,721 Germans without cardiovascular illness. Reference values for CBP and pulse wave analysis were reported in the Gomez-Sanchez et al.¹⁴ Lastly, Weber et al.¹⁵ published references of CBP measured over 24 hours with the Mobilograph device. However, to our knowledge, no research has examined the relationship between the rise in CBP and central pulse wave parameters during a 5-year period in the same sample.

Thus, this study has the following objectives: To analyze the progression in CBP and central hemodynamic parameters both overall and by sex, and to study their relationship with cardiovascular risk factors (CRF) in a Spanish adult population free of cardiovascular disease.

METHODS

The “association between different risk factors and vascular accelerated aging study (EVA study)”¹⁶ is an observational prospective study with a 5-year follow-up. It was carried out at the Primary Care Research Unit of Salamanca (APISAL). It includes a baseline assessment and a 5-year follow-up.

Between June 2016 and November 2017 five urban health centers were used to pick 501 participants, 100 in each group (50 men and 50 women), aged between 35 and 75 years (reference population: 43,946), using stratified random sampling with replacement, by age group (35, 45, 55, 65, and 75 years) and sex. Four hundred eighty participants were followed up between May 2021 and October 2022.

Inclusion criteria: To sign the informed consent form and age from 35 to 75. Exclusion criteria: Subjects with terminal illness or cardiovascular disease, unable to travel to health centers, a glomerular filtration rate of <30%, chronic inflammatory disease or an acute inflammatory process within the last three months, or undergoing treatment with estrogen, testosterone or growth hormone.

The 501 participants in this study enabled the population’s central systolic blood pressure (cSBP) average to be determined with an accuracy of 1.35 mm Hg and a standard deviation of 15.5 mm Hg, assuming an 95% confidence interval.

The two measurements performed in this study were made using the SphygmoCor device (AtCor Medical Pty Ltd., West Ryde, Australia)^17,18(classified by the “International task force”³ as a type 1 device) which measures CBP and the aortic pulse wave. Measurements were performed by tonometry, the most common noninvasive method to estimate CBP, by placing a sensor on the radial artery.¹⁷ We followed the following steps: 1.- Preparation of the patient (previous rest for at least 5 minutes, supine position with the arm where the measurement was taken relaxed and at heart level, without ingesting caffeine or smoking the previous 5 minutes, in a quiet and noise-free room to avoid interference; 2.- Placement of the tonometry sensor on the radial artery, making sure it is well aligned and applying light but sufficient pressure to flatten the artery without occluding it; 3.- Taking the measurement: the sensor is held until a clear pulse wave is obtained repeatedly, then the device records the wave for 10-20 seconds taking the peripheral blood pressure so that the SphygmoCor estimates the CBP; 4.- Analysis of the pulse wave: the device software uses specific algorithms to convert the measured pulse wave into an estimate of the arterial pressure. 5.- Results: SphygmoCor also provides results for other parameters such as the central augmentation index; 6.- Repeat measurements: In all cases where the signal quality is not adequate or there is any doubt about the reliability of the measurement; 7.- Device maintenance: The calibration, cleaning and care of the device have been carried out following the manufacturer’s instructions.

At baseline and at 5 years were measured: cSBP; cDBP; cPP; PPA as an absolute value (peripheral pulse pressure-central pulse pressure); ED in ms and SEVR% calculated as the ratio between the wave area in diastole/wave area during systole. The patient was sitting in a quiet, temperature-controlled room with between 22 and 28 °C, after a rest of 5 minutes.

The peripheral blood pressure was measured using an OMRON model M10-IT (Kyoto, Japan) sphygmomanometer, following the recommendations of the European Society of Hypertension.¹⁸ Body weight and height were measured using a calibrated scale (Seca 770, Birmingham, UK) and a stadiometer (Seca 222, Birmingham, UK). Body mass index (BMI) was estimated as weight in kg/height in m2. After inspiration, the subject’s waist circumference was measured with a tape parallel to the ground, at the midpoint between the last rib and the iliac crest.

Venous blood was drawn at 8am, fasting and without smoking or having consumed alcohol or caffeine for 12 hours. Plasma glucose, total cholesterol, high-density lipoprotein cholesterol, and triglycerides were determined using enzymatic methods. Low-density lipoprotein cholesterol was determined using the Friedewald formula.

A patient was considered hypertensive if they were taking antihypertensive drugs, or had PBP readings ≥140/90 mm Hg; with type 2 diabetes mellitus if they were taking hypoglycemic drugs, or had blood glucose levels ≥126 mg/dl, or HbA1c ≥6.5%; with dyslipidemia if they were taking lipid-lowering drugs or had total cholesterol levels ≥240 mg/dl, or low-density lipoprotein cholesterol ≥160 mg/dl, or high-density lipoprotein cholesterol ≤40 mg/dl in men and ≤50 mg/dl in women, or triglycerides ≥200 mg/dl; with obesity if the BMI was ≥30 kg/m², with abdominal obesity if the waist circumference was ≥102 cm in men or ≥88 cm in women, and smoker if they were smoking at the time of evaluation or had quit smoking during the last year.

Statistical analysis

We used Matlab R2023b through a Campus-Wide License. We observed missing data at 5 years due to the death of some individuals and difficulties in accessing the health center because of the COVID. We found 8.3% of missing men and 10.4% of missing women in the cSBP, cDBP, cPP, ED, and SEVR variables. We addressed this problem using a logistic regression imputation method,¹⁹ which estimates the missing values. The root mean squared error was below 10 and the R² above 60%, except in the estimation of ED, which were slightly worse in both men and women.

Continuous variables were expressed as mean ± standard deviation. Categorical variables as number and percentage. The comparison of measures between two independent groups was performed with Student’s t-tests, and the comparison of proportions with the χ² test. The differences in means by age groups were analyzed with the ANOVA test. To analyze the association between CRF and the progression of vascular function parameters, several models of multiple regression and logistic regression were performed. Six multiple regression models were conducted using the increments of the analyzed measures as dependent variables. Baseline values (systolic blood pressure, diastolic blood pressure, pulse pressure, BMI, waist circumference, total cholesterol, and number of cigarettes per day) and age in years, were used as independent variables, while sex (1 = male; 0 = female) and treatment with antihypertensive drugs (1 = yes; 0 = no) were used as adjustment variables. In logistic regression models, the same dependent variables were used (increments of measures ≥P50 = 1, increments of measures <P50 = 0). Independent variables included CRF such as hypertension, diabetes mellitus, general obesity, abdominal obesity, dyslipidemia, and tobacco consumption (1 = presence of risk factor, 0 = absence of risk factor), with adjustments similar to those in multiple regression. All analyses were conducted overall and by sex. In the hypothesis tests, a risk α of 0.05 was established as statistical significance.

Ethical considerations

The “Committee on Ethics of Research with Medicines in Salamanca” approved the study for baseline evaluation on 05 April 2015, and for follow-up on 13 September 2020 (CEIm reference code. PI 2020 10 569). All participants signed informed consent before inclusion. The Helsinki Declaration recommendations were followed.

RESULTS

Study population

The initial flow diagram is shown in Supplementary Figure S1, reflecting the subjects included, excluded and causes of exclusion by age groups and sex. The flow diagram of the 5 years follow-up is shown in Supplementary Figure S2, reflecting the number of deaths and losses by age group and sex. In the initial evaluation, 501 subjects were included, and 480 subjects were evaluated at 5 years of follow-up. The mean follow-up time was 4.94 ± 0.53 years, in men 4.90 ± 0.51 years and in women 4.97 ± 0.55 years (P = 0.174). The characteristics of the subjects dead and lost at follow-up, as well as the causes, are shown in Supplementary Table S1. During the follow-up period, 10 subjects died (6 men and 4 women). We could not contact 11 subjects (6 men and 5 women).

The clinical variables in the baseline evaluation and at 5 years of follow-up are shown overall, in men and in women in Supplementary Tables S2–S4.

Differences between the two measurements of CBP and parameters derived from the pulse wave

The mean differences, both overall and by sex, between follow-up and baseline evaluations are presented in Table 1. In the overall analysis, differences in the analyzed measures were significant (P < 0.05), all showing values higher after 5 years except for SEVR, which was lower. The progressions in the analyzed measures per year of follow-up were as follows: cSBP = 0.88 ± 2.86 mm Hg, cDBP = 0.53 ± 2.40 mm Hg, cPP = 0.35 ± 2.61 mm Hg, PPA = 0.60 ± 2.55 mm Hg, ED = 1.42 ± 9.37ms, and SEVR = -1.67 ± 7.52 %. In the sex-specific analysis, the results were similar, although in men, the increases in cSBP (P = 0.094) and cPP (0.893) were not significant. When comparing by sex, the increase in cSBP (P = 0.006) and cPP (P < 0.001) was greater in women, as reflected in Supplementary Table S5.

The values of the 10th, 25th, 75th, and 90th percentiles of the progressions in the analyzed parameters, by age groups, both overall and by sex, are shown in Figure 1 and Supplementary Table S6. The mean values of the progressions in the analyzed parameters, by age groups and by sex, are shown in Figure 2 and Supplementary Table S7, revealing differences between age groups in women for cSBP (P = 0.023), cPP (P = 0.010), and SEVR% (P = 0.043). In women, the mean in cSBP and cPP increased, while SEVR% decreased with age. Figure 3 shows the differences between sexes by age groups of the analyzed parameters. Increases were greater in women in cSBP in the 70 and 80-year-old groups, in cPP in the 70 and 80-year-old groups, and PPA in the 60-year-old group. The decrease in SEVR was lower in men.

Relationship of analyzed parameters with CRF and antihypertensive

Table 2 summarizes the differences in progressions in central pressures and different pulse wave parameters between subjects with and without CRF overall and by sex. Hypertensive individuals showed smaller increases than non-hypertensive individuals in cSBP in men and cDBP in the overall analysis and in men. Women with type 2 diabetes mellitus showed smaller progressions than non-diabetic individuals in SEVR%. People with dyslipidemia showed greater increases than those without dyslipidemia in cSBP in women, cPP overall and in women, and smaller progressions in SEVR% in women. Female smokers showed smaller increases than non-smokers in cSBP. Obese individuals showed greater increases than non-obese individuals in cSBP, cPP in men, and PPA overall, in men and in women.

Table 3 shows variations in CBP between participants receiving and not receiving antihypertensive treatment overall and by sex. Antihypertensive treatment subjects had higher increases in cPP and lower increases in cDBP in men than did untreated subjects.

Association of CRF with CBP and hemodynamic parameters

The results of multiple regression analysis overall and by sex are shown in Table 4. In the overall analysis, the increase in cSBP showed a positive association with the number of cigarettes per day (β = 0.192), BMI (β = 0.476), and waist circumference (β = 0.159), and an inverse association with peripheral systolic blood pressure (PSBP) (β = −0.282). The increase in cDBP showed a positive association with the number of cigarettes per day (β = 0.174) and an inverse association with peripheral diastolic blood pressure (PDBP) (β = −0.292). The increase in cPP showed a positive association with BMI (β = 0.330) and an inverse association with peripheral pulse pressure (PPP) (β = −0.262). The increase in PPA showed a positive association with the number of cigarettes per day (β = 0.281) and with BMI (β = 0.276). The increase in ED showed an inverse association with baseline glucose levels (β = −0.282). In men, the increase in cSBP showed a positive association with BMI (β = 0.845) and waist circumference (β = 0.278) and an inverse association with PSBP (β = −0.300). The increase in cDBP showed an inverse association with PDBP (β = −0.278). The increase in cPP showed a positive association with BMI (β = 0.608) and waist circumference (β = 0.160) and an inverse association with clinical PPP (β = −0.217). The increase in PPA showed a positive association with BMI (β = 0.421). In women, the increase in cSBP showed an inverse association with PSBP (β = −0.287). The increase in cDBP showed a positive association with the number of cigarettes per day (β = 0.388) and an inverse association with PDBP (β = −0.319). The increase in PPA showed a positive association with the number of cigarettes per day (β = 0.569). The increase in ED showed a positive association with the number of cigarettes per day (β = 0.673) and an inverse association with waist circumference (β = −0.661). The progression in SEVR showed a positive association with PSBP (β = 0.539).

Overall findings from the logistic regression analysis are displayed in Figure 4, Supplementary Figure S3 for males, and Supplementary Figure S4 for women.

DISCUSSION

At baseline assessment, cSBP, cDBP, and cPP values were lower than those reported by other authors,^11,13 with a behavior, by age and sex, similar to data published by Weber et al.¹³ Progressions in CBP were lower in hypertensive subjects, with more pronounced differences in men, possibly due to different changes caused by antihypertensive drugs, known to have variable effects on CBP and PBP among individuals.⁵ Moreover, CBP is a more direct measurement of the pressure to which major target organs are exposed, more accurately determining myocardial²⁰ oxygen demand and supply than PBP, potentially improving diagnostic and therapeutic value compared to PBP.⁷ Sex differences could be explained because women have higher CBP than men with similar PBP, mainly influenced by shorter stature.²¹ This difference in CBP and PBP, considering that decisions on pharmacological treatment are based on PBP figures, may suggest unrecognized insufficient treatment of hypertensive women, explaining why CBP reduction was lower in them.²¹ These findings may justify the need for different PBP targets by sex. Furthermore, greater progressions in CBP in women over 60 corroborate earlier research demonstrating that, for comparable ages, men had higher blood pressure than women; however, the opposite happens after menopause, due to an estrogen decrease. However, some studies²² have found that among women and men with similar PBP, women have a higher cardiovascular risk, suggesting that this increased risk could be due to greater unrecognized CBP in women compared to men with similar PBP.²¹ Therefore, CBP measurement, especially in women, may improve identification and management of patients with elevated cardiovascular risk.^13,21

Increases in cSBP were higher in obese men and women with dyslipidemia. Differences are likely related in the former case to height differences and in the latter to lipid-lowering drugs.¹² However, only obesity in men was retained in the analysis of associations. Previous studies have shown sex differences between normotensive and hypertensive individuals in CBP. One study showed that male sex was associated with lower SBP in hypertensive individuals, while in normotensive individuals it was the opposite, suggesting a strong interaction between age, sex, and PBP value with CBP.¹¹

This study shows a greater increase in PPA in men under 50 years of age and a greater increase in women over 60 years of age. The increase in PPA was higher in obese individuals and showed a positive association in overall analysis and in men but not in women. One possible explanation is that young men with hyperkinetic hearts may have greater reactivity to stress than young women. Additionally, PPA degree varies within and between individuals, influenced by factors such as age, sex, height, heart rate, and drugs.⁵ The age–sex difference in PPA is consistent with previous studies indicating that PPA increases in menopausal women older than 55 years compared to younger women. Furthermore, at ages over 55 years, the impact of the evolution of PPA on cardiovascular (CV) mortality is three times greater in women than in men,²³ probably due to the loss of estrogenic action on the arterial wall. Contrary to other studies, we have not found an association with hypertension or diabetes mellitus, which shows that hypertensive diabetic subjects can increase cardiovascular risk, regardless of age and sex.²⁴ These differences may be justified by different classes of antihypertensive drugs that have different effects on PPA; angiotensin-converting enzyme (ACE) inhibitors, angiotensin receptor blockers, dihydropyridine calcium blockers, and nitrates may have a more beneficial effect on PPA than diuretics and beta-blockers.⁴ Pulse pressure amplification amplification is a complex hemodynamic mechanism and is mainly related to heart rate and wave reflections, and we have not taken into account the heterogeneity between treatments with different beta-blockers. In fact, new beta-blocker treatments with peripheral vasodilatory effects could act differently on PPA and could reduce the reflection of pressure waves and partially counteract the effect of heart rate deceleration.²⁴ However, we must not forget that PPA amplification is probably one of the main measures of central hemodynamic evaluation and cardiovascular risk and can avoid the variations in calibration shown by other measures.¹ Furthermore, PPA is considered an independent cardiovascular risk and mortality factor and is a stronger predictor of cardiovascular risk in subjects with arterial hypertension.^24–26 On the other hand, PPA and aortic stiffness are partially independent and may appear as complementary CRF and are related to changes in aortic stiffness and reflection waves along the arterial tree.^8,24

In this study, the increase in ED was lower in men than in women treated with antihypertensive drugs, probably due to greater diastolic ventricular dysfunction.²⁰ We did not find differences in this parameter by age group or sex. However, there are some studies²⁰ that have found an increase in ED with age and higher values in women.

The balance between SEVR and myocardial oxygen demand can be measured with precision using SEVR. In this study, it decreases with age and is the only measure showing values lower than at baseline. Unlike studies that found an association with age and waist circumference, in this study, the relationship with waist circumference was not significant.²⁰

In summary, there is still no consensus on which measurement we should use in clinical practice to estimate central arterial pressures and pulse wave parameters. Non-invasive measurements generally underestimate the true values of cSBP and PPA. Furthermore, the higher the invasive cSBP or PPA, the greater the underestimation.^1,2 On the other hand, PPA is a dimensionless parameter and overcomes the effect related to calibration.² The findings of this work show that over time the levels of CBP and peripheral blood pressure progress differently depending on sex and age. The data it provides suggest that central arterial pressures and pulse wave parameters could provide new information on the estimation of the risk of cardiovascular diseases.

Strengths and limitations

As far as we know, this is the first prospective study to examine the rise in PBC and other pulse wave-derived hemodynamic parameters over a 5-year period in a sample of people without cardiovascular disease. The SphygmoCor device, type 1 device, was used for all measurements, by a consistent methodology. The results can be extrapolated to the reference population because the sample was drawn at random from the population and was stratified by sex and age group. This study is not without limits, however. The first related to the limitations of non-invasive measurement of central arterial pressures and central hemodynamic parameters. Current evidence suggests that calibration with mean arterial pressure and DBP may provide a more accurate assessment of CBP than calibration with systolic blood pressure (SBP) and diastolic blood pressure (DBP)²⁷, calibration form that is used by our device. On the other hand, the sample studied is small, as is the number of subjects with a diagnosis of hypertension, diabetes and subjects with pharmacological treatment, which precludes studies by different subgroups. Finally, the individuals are from an urban population and could not be entirely typical of all the Spanish population.

CONCLUSIONS

At 5 years of follow-up, all CBP and measures generated from pulse waves increased, with the exception of SEVR, which declined. Progressions in CBP were inversely associated with peripheral blood pressure and positively associated with anthropometric measures and the number of cigarettes per day, with variations depending on sex.

SUPPLEMENTARY DATA

Supplementary materials are available at American Journal of Hypertension (http://ajh.oxfordjournals.org).

Trial Registration Number: NCT02623894

(https://classic.clinicaltrials.gov/ct2/show/NCT02623894)

ACKNOWLEDGMENTS

We thank the patient advisers who collaborated with the investigators, Raquel Pizarro-Juárez for reviewing the use of English and all the EVA investigator group (named in the Supplementary Material).

FUNDING

This study was funded by different public entities: the Gerencia Regional de Salud de Castilla y León through research projects (GRS 1193/B/15; GRS2303/B/21), the Instituto de Salud Carlos III (ISCIII) of the Ministry of Science, through RD21/0016/0010 and INT22/00007, Red de Investigación en Cronicidad, Atención Primaria y Promoción de la Salud (RICAPPS), Facility for Recovery and Resilience (MRR) and PI21/00454, Junta de Castilla y León (INT/M/02/17 and INT/M/04/15), the Instituto de Investigación Biomedico de Salamanca (IBSAL) (IBI21/00001). None of them played any role in the study design, data analysis, reporting results, or the decision to submit the manuscript for publication.

CONFLICT OF INTEREST

The authors declared no conflict of interest.

Data Availability Statement

The data underlying this article will be shared on reasonable request to the corresponding author.

REFERENCES

Author notes