The association between pulse wave velocity and heart failure: a systematic review and meta-analysis

Background The arterial stiffness measured by pulsed wave velocity (PWV) is associated with heart failure (HF). However, the effectiveness of arterial stiffness and PWV as prognostic indicators in patients with HFpEF and HFrEF is still unclear. In this systematic review and meta-analysis, we synthesized the prognostic value of PWV and arterial stiffness in HF patients. Methods Four databases, including Embase, PubMed, Scopus, and Web of Science, were systematically searched for published studies assessing the relationship between PWV and HF from inception up to August 31, 2023. The Newcastle-Ottawa Scale (NOS) was used to assess the quality of the included studies. The standardized mean difference (SMD) and their corresponding 95% confidence intervals (CI) were used to compare PWV in HF (HFrEF and HFpEF) and controls. Meta-regressions based on age, year of publication, sample size, and gender (male percentage) were also conducted. Results The systematic search yielded 5,977 results, of which 58 met our inclusion criteria and 24 were analyzed quantitatively. Studies included 64,687 patients with a mean age of 53.7 years, and 41,803 (67.3%) were male. Meta-analysis of 19 studies showed that PWV was significantly higher in HF patients compared to the controls (SMD 1.04, 95% CI 0.43–1.66, P < 0.001, I2 = 93%). Moreover, nine studies have measured PWV among HFrEF and HFpEF patients and found no significant difference (SMD −0.51, 95% CI −1.03 to 0.02, P = 0.057, I2 = 95%). Moreover, increased PWV was linked to an increased chance of developing new-onset HF in individuals with cardiovascular risk factors. Conclusions Patients with HF exhibit significantly higher arterial stiffness, as indicated by PWV, compared to the normal population. However, this association was not significant between HFrEF and HFpEF patients. Future research is warranted to establish the potential prognostic role of PWV in HF. Systematic Review Registration https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42023479683, PROSPERO (CRD42023479683).


Introduction
Heart failure (HF) is a diverse and potentially fatal syndrome impacting over 60 million people worldwide (1).HF is marked by a high rate of mortality and morbidity, a low quality of life, and a heavy financial and resource strain on healthcare systems (1).The already alarming HF epidemic is anticipated to worsen as the population ages (2).Diabetes, hypertension, and ischemic heart disease are the most prevalent etiologies of HF (3,4), followed by cardiomyopathies and infections such as viral myocarditis and Chagas' disease (5).Two predominant phenotypes of HF based on left ventricular ejection fraction (EF) have been recognized: HF with preserved EF (HFpEF) and HF with reduced EF (HFrEF).These subtypes differ in their underlying pathophysiology, clinical characteristics, and treatment response (6,7).
Arterial stiffness, an artery's decreased capacity to expand and contract in response to pressure changes, can predict left ventricular (LV) diastolic dysfunction and is associated with cardiovascular risk (8)(9)(10)(11).Greater arterial stiffness is connected to LV diastolic dysfunction and HF with preserved EF (12,13).The pulse wave velocity (PWV) has been acknowledged as the validated test for examining stiffness in large arteries in the consensus document on ventricular-arterial coupling in cardiac disease (14).PWV is determined by dividing the distance between two points by the time it takes for the pulse to travel between them (2).Increased arterial stiffness, as indicated by elevated PWV, is correlated with poor prognosis in several cardiovascular diseases, particularly in the case of HF (15,16).
HFpEF has not shown positive responses to conventional pharmacological interventions, except nitrate; hence, there is a growing interest in identifying novel prognostic markers and therapeutic targets for HF (2).This systematic review and metaanalysis aimed to assess the effectiveness of arterial stiffness and PWV as prognostic indicators in patients with HFpEF and HFrEF.By synthesizing data from relevant studies, we seek to elucidate the relationship between arterial stiffness, PWV, and different HF phenotypes, providing insights into their potential utility in risk stratification, management, and therapeutic decision-making for HF patients.

Materials and methods
This study was written based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 checklist (17).This systematic review and meta-analysis protocol was registered in PROSPERO (CRD42023479683).

Literature search strategy
Four databases, including Embase (Elsevier), PubMed (US National Library of Medicine), Scopus (Elsevier), and Web of Science, were systematically searched for published studies assessing the relationship between PWV and HF from inception up to August 31, 2023.The search terms we used in this study included: "Pulse Wave Analysis" OR "Ankle-Brachial Pulse Wave Velocit*" OR "Pulse Wave Velocity" AND "Heart Failure, Systolic" OR "Heart Failure" OR "Heart Failure, Diastolic" OR "Heart Failure" OR "Cardiac Failure" OR "Congestive Heart Failure" OR "Heart Failure, Reduced Ejection Fraction" OR "Heart Failure, Preserved Ejection Fraction".Additional details about the medical subject headings, keywords, and entered terms are presented in Supplementary Table S1.

Study selection and eligibility criteria
Two reviewers (Z.E.and Z.V.) assessed each record separately using EndNote 21 software (Tomson Reuters, New York, USA).First, duplicates were removed, and records were screened based on their titles and abstracts.Afterward, the full texts of the studies were assessed; selection of studies adhered to the inclusion and exclusion criteria.The third author (A.A.) acted as the facilitator of agreement meetings to address any potential disputes among reviewers.
Studies were included in this review using the following inclusion criteria: (1) subjects were adults (aged ≥18 years), ( 2) clinical studies that assessed and compared PWV in HF subtypes, (3) studies that assessed PWV in HF patients and compared them with the normal population, (4) papers that reported baseline PWV in the normal population and assessed for probable incident HF, and (5) studies that reported PWV in different stages of HF.Finally, duplicate publications, studies not reporting PWV, animal studies, case reports, abstracts, and reviews were excluded.

Data extraction and quality assessment
Two reviewers (P.B. and Z.V.) independently extracted the following data from the included studies into a pre-made Excel sheet: first author name, year of publication, study location, study design, sample size, study population (normal or HF), age, gender, and EF.The extracted data were admitted by a third reviewer for probable disparities (A.H.B.).
The Newcastle-Ottawa Scale (NOS) was used to assess the quality of the included studies.The Cochrane Handbook recommends and has created this tool for evaluating the quality of observational studies (18).In cohort studies, there are three key areas to evaluate: selection, comparability, and outcome, with ratings of up to four, two, and three stars; in cross-sectional studies, three aspects were evaluated: selection, comparability, and outcome, with maximus ratings of three, two, and two stars, respectively.A rating of ≥7 is viewed as top quality on this scale.Two separate writers (Z.V. and Z.E.) evaluated the characteristics, and if there was any conflict, a third author (A.K.) settled the matter.

Statistical analysis
Statistical analyses of this study were conducted using the R program [version 4.3.0].We used Hedges' g standardized mean difference (SMD) and their corresponding 95% confidence intervals (CI) to compare PWV in HF patients and controls and to compare PWV in different HF subtypes, including HFrEF and HFpEF patients (19).We conducted univariable meta-regression based on age, year of publication, sample size, gender (male percentage), and subgroup analysis among HF subtypes based on the locations and the devices used for PWV assessments.The heterogeneity of studies was assessed using Cochrane's Q and Higgins' I 2 tests.There was high inter-study heterogenicity if I 2 > 50% and P < 0.1 for the result of the Q test (20).The random-effects model was applied to accommodate the heterogeneity of the enrolled studies (20).P < 0.05 reflected statistical significance for all data analyses.Finally, Egger's statistical test and funnel plot were performed for publication bias (21).

Study selection
The systematic search of electronic databases yielded 5,977 results, including 624 from PubMed, 3,274 from Embase, 1,086 from Web of Science, and 993 from Scopus.After duplicate removal, 2,974 studies remained.Among those, 2,488 records were excluded during the initial screening based on their title and abstract, and 386 records underwent further full-text screening.The full texts of 378 studies were retrieved and went through an eligibility assessment, from which 319 were excluded.Finally, 58 studies met our inclusion criteria and remained for our qualitative evaluation (22-78).Twenty-four studies were analyzed quantitatively.Figure 1 demonstrates the study selection process in detail.

Baseline characteristics and quality assessment
A summary of the characteristics of the included studies is demonstrated in Table 1.A total of 64,687 patients were included in our study, with sample sizes ranging from 16 to 40,064 participants.Twenty-four studies were conducted in Europe, twenty-four in Asia, nine in America, and one in Africa.The mean age of the population was 53.67 years, and 41,803 (67.27%) were male.The range of mean EF was 21.8 to 68.4%.The carotid-femoral and ankle-brachial arteries were the most frequently assessed anatomical sites for measuring PWV.Carotid-femoral PWV was measured in 25 studies (14, 24, 25, 28, 29, 32-34, 37, 40, 44, 46-49, 51, 53, 54, 57, 61, 65-67, 70, 72, 77, 78), while 13 studies reported ankle-brachial PWV measurements (38-40, 43, 55, 56, 58, 64, 71-75).Of the included studies, 47 had a high-quality score with an overall score of seven or higher based on the NOS tool, while six and five studies scored 6 and 5, respectively (Supplementary Table S2).
Meta-regression revealed a significant association between gender and the PWV in two groups (β −0.0166, 95% CI −0.0315 to −0.0017], P = 0.02).However, no significant associations were found between the effect size and other investigated variables, including age, year, and sample size (Supplementary Table S3 and Figures S4-S7).
Publication bias was detected when comparing PWV between HF and the normal population, according to the funnel plot asymmetry and Egger's test (P = 0.04).The funnel plot can be observed in Figure 4.
Sensitivity analysis yielded significantly reduced pooled SMD by removing three studies, Alem et al. ( 45 S10) Meta-regression showed no statistically significant correlation between any investigated moderators and PWV in the two groups (Supplementary Table S4 and Figures S11-S14).

PWV and adverse outcomes in HF
Thirteen studies have investigated the role of PWV measurements on adverse outcomes in HF patients.

Mortality
Four studies (50,51,55,59) assessed the relationship between PWV and mortality in HF patients.PWV was associated with higher rates of mortality in the study by Demir et (55) found no significant association between mortality and PWV in the HF population.

Mortality and hospitalization
Four studies have explored the correlation of PWV with the composite outcome of death or hospitalization.PWV was associated with higher rates of death or hospitalization reported

Dyspneic patients
Two studies have explored PWV in patients with dyspnea without a diagnosis of HF compared to dyspneic HF patients.Cong et al. ( 73) assessed baPWV in a cohort of patients with acute dyspnea and found that HF patients presented a higher level of baPWV compared to dyspneic patients without HF diagnosis (OR 2.26, 95% CI 1.15-4.44).The same results were obtained by Weber et al. (78) (OR 1.57, 95% CI 1.28-1.93).These studies suggest using PWV as a predictor of HF in symptomatic patients.

Increased risk of incident HF with a higher PWV
There was a link between higher PWV and the risk of newonset HF in three studies by Zheng et

Main findings
The systematic review and meta-analysis aimed to assess arterial stiffness and PWV in patients with HF and uncovered various vital discoveries.First, the PWV was notably elevated in individuals with HF compared to the control population, mainly when measured in the carotid-femoral and brachial-ankle regions.However, the disparity in PWV between HFrEF and HFpEF was not statistically significant, especially in the carotidfemoral area.Significantly, elevated PWV levels were associated with higher chances of experiencing various cardiovascular events and adverse outcomes, such as death, hospitalization, and readmission.PWV demonstrated the potential to predict new cases of HF in patients experiencing dyspnea, indicating its potential value in detecting HF in symptomatic patients.Moreover, increased PWV was linked to a heightened likelihood of developing new-onset HF in individuals with cardiovascular risk factors, underscoring its importance as a prognostic indicator for HF onset.These results highlight how PWV is crucial as a predictive marker and could be valuable in evaluating cardiovascular risk and treating patients with HF.
Arterial stiffness is a predictor of cardiovascular events and allcause mortality (15).Since LV diastolic function is also a predictor of increased mortality from HF (79, 80), increased arterial stiffness may contribute to cardiovascular events by causing diastolic dysfunction.The pathophysiology of HFpEF is characterized by diastolic dysfunction, resulting in insufficient ventricular filling during diastole due to impaired ventricular relaxation and Frontiers in Cardiovascular Medicine 09 frontiersin.orgincreased stiffness.Even though the LV cavity size is usually normal, the LV wall is thick and stiff, leading to a higher LV mass to end-diastolic volume ratio (81).Increased arterial stiffness plays a crucial role in HFpEF, contributing to higher LV pressure afterload.This compromises ventricular-arterial coordination and exacerbates diastolic dysfunction.The rigidity of large elastic arteries decreases their capacity to absorb the rhythmic flow produced by the heart, resulting in higher LV filling pressures and lower aortic pressures in the relaxation phase, intensifying the pulse pressure (82).The rise in pulse pressure heightens the demand for oxygen in the heart muscle during contraction, leading to LV thickening, further hindering blood flow to the coronary arteries during relaxation (83).In contrast, HFrEF is defined by systolic dysfunction, where the heart muscle's weakened ability to contract leads to a lower ejection fraction and diminished cardiac output.This can result from a heart attack, inflammation of the heart muscle, or other heart conditions causing changes and expansion of the ventricles (84).In HFrEF, arterial stiffness plays a crucial role in the progression of the disease by exacerbating LV systolic dysfunction (85).Increased vascular resistance and reduced compliance, demonstrated by elevated PWV, lead to a compromised oxygen Funnel plot showing the presence of publication bias in the analysis of HF vs. normal population.supply-demand balance in the heart, contributing to adverse changes in cardiac structure.The connection between arterial stiffness and diastolic dysfunction is not solely due to the development of LV hypertrophy, as arterial stiffness is also linked to diastolic dysfunction even when ventricular hypertrophy is present (86, 87).
Arterial tonometry is a simple, non-invasive method to assess arterial rigidity that can be conducted at the patient's bedside (88).Various techniques to measure arterial stiffness include carotid-femoral pulse wave velocity (cfPWV), brachial-ankle pulse wave velocity (baPWV), cardio-ankle vascular index (CAVI), and augmentation index (AIx) (89)(90)(91).The baPWV is Funnel plot showing no publication bias in the analysis of hFrEF vs. HFpEF patients.obtained by dividing the distance between the arms and ankles, as determined by anthropometric data depending on a person's height, by the propagation time of the pulse wave between these two points, which is measured using occlusion cuffs (92).The cfPWV is calculated by manually measuring the distance between the carotid and femoral arteries, then dividing by the time it takes for the pulse wave to travel between the two locations (93).
The varying influence of PWV measurement sites on outcomes highlights the significance of considering the specific vascular region under evaluation.The carotid-femoral location is commonly seen as the best method for evaluating central arterial stiffness, offering information on the stiffness of major elastic arteries such as the aorta, which are important for dampening pulsatile flow (94).However, baPWV consistently demonstrates a 17%-20% increase compared to cfPWV (95), showing that baPWV assesses further elements of arterial stiffness.cfPWV focuses solely on the central arterial tree, providing insight into central arterial stiffness, which predominantly affects afterload.In contrast, baPWV considers both the central and peripheral arterial trees, providing a more comprehensive understanding of the afterload effects on diastolic dysfunction by reflecting the overall resistance and compliance of the arterial system.This comprehensive measurement may explain the different values obtained by cfPWV and baPWV.While central arteries primarily determine afterload, peripheral arteries become significant, especially when peripheral arterial disease is present (9).Additional factors that could account for the significant variability in this research may stem from the increased reliance on the person administering the test for cfPWV compared to baPWV, as well as the required use of a handheld pressure transducer to measure pulse waves at the neck and groin (96).Also, the lack of significant differences in PWV between HFrEF and HFpEF, particularly in the carotid-femoral location, may suggest that systemic arterial stiffness is a shared characteristic across both HF phenotypes.A similar systematic meta-analysis was conducted to study the association between arterial stiffness assessed by arterial tonometry and echocardiographic markers of diastolic dysfunction, which is essential for diagnosing heart failure with preserved ejection fraction (HFpEF).Twenty-seven studies included 6,626 patients.baPWV showed significant correlations with the E/A ratio, e0, and E/e0 ratio.Similarly, cfPWV was significantly correlated with the E/A and the E/e0 ratios, but not e0.AIx showed a strong relationship with E/A ratio (r = −0.356,95% CI −0.255 to −0.450), e0 (r = −0.313,95% CI −0.195 to −0.423), and E/e0 ratio (r = 0.321, 95% CI 0.250-0.388).CAVI had a strong correlation with the E/A ratio, e0, and baPWV showed a significant correlation with diastolic dysfunction compared to other tonometry techniques (97).

Strengths and limitations
Adding a significant number of research studies involving 64,687 patients increased the statistical strength and reliability of the meta-analysis.A significant percentage of the studies analyzed were rated as high quality based on their Newcastle-Ottawa Scale (NOS) scores.This enhances the credibility and accuracy of the results.Subgroup analyses were conducted to investigate possible reasons for diversity, like different locations for PWV measurement and HF subtype (HFpEF vs. HFrEF), as well as the devices used for the PWV measurement, leading to a more detailed understanding of the findings.Meta-regression analysis was used to examine how gender, age, and sample size affect observed associations, improving comprehension of potential moderators.
This study carries some limitations.Although attempts were made to deal with differences through subgroup analyses and sensitivity analyses, numerous meta-analyses still showed significant heterogeneity, which may restrict the applicability of the results.Differences in research structure, patient traits, and methods, like variations in PWV measurement methods and anatomical locations, HF subtype (HFpEF vs. HFrEF), and disease severity, probably played a part in the diversity.Also, we could not analyze PWV based on the blood pressure levels, as this information was not sufficiently reported in the included studies.Future research should aim to collect and report BP data to allow for a more comprehensive analysis.Even though the majority of the studies analyzed were well-conducted, a few had methodological flaws that may have influenced the results of the meta-analysis.

Conclusion
In conclusion, our systematic review and meta-analysis revealed that patients with HF exhibit significantly higher arterial stiffness, as indicated by PWV, compared to the normal population.This association was consistent across various anatomical sites of PWV measurement, including carotid-femoral and ankle-brachial arteries.Furthermore, PWV was found to be a predictor of adverse outcomes in HF patients, including mortality, hospitalization, and readmission.Although there was no statistically significant difference in PWV between patients with HFrEF and HFpEF, subgroup analyses indicated potential differences based on the site of PWV measurement.Additionally, PWV was associated with adverse outcomes in high-risk cardiovascular populations and individuals with dyspnea, underscoring its potential utility as a predictive tool for HF development.Future research should focus on elucidating the underlying mechanisms linking PWV to adverse outcomes in HF patients and exploring its role in risk stratification and therapeutic interventions to improve patient outcomes.

FIGURE 2 Forest
FIGURE 2Forest plot showing the PWV difference in HF vs. normal population.

FIGURE 3
FIGURE 3Subgroup analysis of PWV in HF vs. normal population based on the location of PWV measurement.

FIGURE 5 Forest
FIGURE 5Forest plot showing the PWV difference in hFrEF vs. HFpEF patients.

FIGURE 7
FIGURE 7Subgroup analysis of PWV in hFrEF vs. HFpEF patients based on the location of PWV measurement.

TABLE 1
Baseline characteristics of included studies assessing PWV in HF.