Relationships between Cardio-Ankle Vascular Index and Peptides Related to Hypertensive Disorders of Pregnancy in Patients with Lower Extremity Arterial Disease

Lower extremity arterial disease (LEAD), previously included in peripheral arterial disease (PAD), is an obstructive/stenotic disease of large arteries in the lower extremities that is usually caused by atheromatous plaques and occurs at higher rates in smokers and patients with diabetes [1]. In our recent study, six peptides related to hypertensive disorders of pregnancy (HDP) were shown to be associated with leg ischemia in patients with LEAD and were proposed as biomarkers for evaluating leg arterial flow [2]. These peptides are fragments originating from fibrinogen-α (P-2091), kininogen 1 (P-2081, P-2127, P-2209), complement C4 (P-2378), and α-2-HS-glycoprotein (P-2858), and they were named on the basis of their mass/charge ratios as shown above in parentheses. Moreover, in healthy men, P-3156 (originating from inter-α-trypsin inhibitor heavy chain H4) as well as P-2091 and P-2378 were reported to be associated with cardiometabolic risk factors such as blood pressure, blood lipids, and body mass index (BMI) [3]. LEAD is usually caused by atherosclerosis, and patients suffering from LEAD are at a great risk for cardiovascular diseases including ischemic heart disease and stroke [4, 5]. There is a close link between atherosclerosis and arterial stiffness, in which a number of different hemodynamic, mechanical, metabolic, and enzymatic mechanisms are involved, and aging and hypertension are common major factors influencing processes of atherosclerosis and arterial stiffness [6]. However, it remains unknown whether and how the HDP-associated peptides are related to arterial stiffness in patients with LEAD.

Arterial stiffness is often evaluated by using pulse wave velocity (PWV) [7], and cardio-ankle vascular index (CAVI) is a more recent marker for arterial stiffness [8, 9]. CAVI is a blood pressure-independent measure of heart-ankle PWV. Ankle-brachial index (ABI), the ratio of arterial pressure at the knee joint to that in the cubital joint, is a standard measure for evaluating leg arterial flow, and low ABI (0.9 or lower) is used as a criterion for diagnosing LEAD [10]. Since arterial stiffness is closely related to atherosclerosis, which causes LEAD, PWV, and CAVI were expected to be inversely associated with ABI. In fact, PWV was reported to be inversely associated with ABI in the general population [11, 12]. However, paradoxical associations of PWV and CAVI with ABI have been reported in patients with LEAD: PWV and CAVI were not inversely but positively associated with ABI [13‒16]. This paradox may be due to lower mean arterial pressure distal to stenosis, leading to slow transit time of pulse wave in the legs, and causes the paradoxical relationship between CAVI and ABI, which can be impacted by lower pulse pressure amplification. Instead of PWV and CAVI themselves, the absolute value of the difference between right- and left-side measures of PWV or CAVI has been proposed as an indicator of leg ischemia in patients with LEAD [13, 17]. However, there has been little information on these new parameters, and thus, their clinical significance remains to be determined.

Both the new peptides (derived from kininogen, fibrinogen, complement C4 and α2-HS-glycoprotein) in serum and diff-CAVI (absolute value of the difference between right and left CAVI) have recently been proposed to be new markers of leg ischemia in patients with LEAD [2, 17]. The aim of this study was, therefore, to investigate relationships between the new peptides and diff-CAVI and to confirm the significance of these biomarkers. Moreover, the HDP-related peptides have recently been shown to be associated with ABI but not with carotid atherosclerosis, and thus, we also investigated the relationships of the peptides with CAVI, a marker of arterial stiffness, although a paradoxical association was reported between CAVI and ABI in patients with LEAD [16].

The participants of this study were 165 outpatients (145 men and 20 women) diagnosed as having LEAD by using the criterion of a low ABI (≤0.9) [1]. Most of the participants had already received medication therapy for LEAD including anticoagulation therapy, antiplatelet therapy, surgical therapy and/or endovascular therapy (Table 1).

The protocol of this study was approved by the Ethics Committee of Yamagata Saisei Hospital (Approval No.: 199 at the Ethics Committee since 2013). Written informed consent to participate in this study was obtained from all the participants. Information on lifestyles and medical histories, including illness and medication therapy, were obtained through a survey using questionnaires. Smokers were classified into three categories by daily average cigarette consumption of nonsmokers, light smokers (≤20 cigarettes), and heavy smokers (≥21 cigarettes). Alcohol drinkers were classified into three categories by frequency of weekly drinking of nondrinkers, occasional drinkers (≤4 days per week), and regular drinkers (≥5 days per week).

BMI was calculated as weight (kilograms) divided by the square of height (meters). Participants with obesity were defined as those with BMI of 25 kg/m² or higher [18].

Venous blood was collected from each participant in the morning after overnight fasting. Concentrations of hemoglobin A_1c were determined by using an automatic glyco-hemoglobin analyzer based on high-performance liquid chromatography (ADAMSTM A1c HA-8170, Sekisui Medical Co., Ltd, Tokyo, Japan). Calibration of hemoglobin A_1c values was performed using the formula by the Japan Diabetes Society [19]. Participants were diagnosed as having diabetes when they were receiving medication therapy for diabetes and/or showed high hemoglobin A_1c levels of 6.5% or higher, according to the criteria of the American Diabetes Association [20].

CAVI and blood pressure of the brachial arteries were measured by an oscillometric method using an automatic device (VaSera VS-1500, Fukuda Denshi, Tokyo, Japan) [8, 9]. An absolute value of the difference in CAVI on right- and left-side measurements was calculated and defined as diff-CAVI. A large diff-CAVI was defined as 1.05 or higher, which was determined in our previous study using the same database as that used in the present study [17]. ABI was also recorded by using the above instrument (VaSera VS-1500). An absolute value of the difference in ABI at right- and left-side measurements was calculated and defined as diff-ABI. Participants with hypertension were defined as those who showed high blood pressure (systolic blood pressure of ≥140 mm Hg and/or diastolic blood pressure of ≥90 mm Hg) [21] and/or those who were receiving medication therapy for hypertension. The mean arterial pressure level was defined as diastolic blood pressure level plus four-tenths of the difference between systolic and diastolic blood pressure levels [22].

The concentration of each peptide in serum was measured essentially according to the method described previously [3, 23]. Each serum sample of 50 μL was spiked with stable isotope-labeled (SI) synthetic standard peptides and heated at 100°C for 10 min. After centrifugation, the supernatant was applied to a graphite carbon GL-tip GC device (GL Sciences Inc., Tokyo, Japan) equilibrated with a solution containing 3% acetonitrile and 0.1% trifluoroacetic acid (TFA), and bound materials were eluted with 60 µL of a solution containing 30% acetonitrile and 0.1% TFA. The eluate was evaporated and reconstituted with a solution containing 2% acetonitrile and 0.1% TFA. The prepared sample corresponding to 7.5 µL of each serum sample was applied to the nano-flow reverse-phase C18 LC-MS/MS system (QTRAP 6500, SCIEX, Framingham, MA, USA) and quantified by the multiple reaction monitoring mode. The multiple reaction monitoring transitions were set into two periods according to the retention time of each peptide on LC, four peptides (P-2081, P-2091, P-2209, and P-2127) having an early retention time and 3 peptides (P-2378, P-2858, and P-3156) having a late retention time [24]. The mass chromatographic peak area of each peptide was calculated with the MQ4 algorithm of MultiQuant 3.0 software (SCIEX), and the peptide concentration was quantified on the basis of the ratio of the peak areas of the natural and SI peptides. In measurements of the seven peptides, each ratio of the area of a stable isotope-labeled peptide to the area of an isotope-unlabeled peptide was relatively well reproducible (coefficient of variation: 0.008–0.147).

A computer software program (IBM SPSS Statistics for Windows, version 25.0., IBM Corp, Armonk, NY, USA) was used for statistical analyses. Categorical variables are displayed as frequencies and percentages and were compared by using Pearson’s chi-square test. Continuous variables that showed normal distributions are displayed as means with standard deviations or means with 95% or 99.3% confidence intervals. In univariate correlation analysis, Spearman’s rank correlation coefficients were calculated for variables not showing normal distributions. In a multivariate analysis of linear regression, standardized partial regression coefficients (β) were calculated when variables showed normal distributions. Comparison of variables between two groups with and without large diff-CAVI was performed by using Student’s t test in univariate analysis and by using analysis of covariance followed by Student’s t test after Bonferroni correction in multivariate analysis. Tertile groups for each of the HDP-related peptides were prepared as follows: values of each peptide in all participants were arranged in ascending order, and then they were divided into three tertile groups of equal sizes. Although ABI and CAVI showed normal distributions, diff-CAVI and each of the peptide levels did not show a normal distribution, and they were, therefore, used after base-10 logarithmic transformation in some analyses as needed. Dichotomous variables such as large diff-CAVI were compared in the tertile groups of each peptide by using logistic regression analysis. In multivariate analyses, adjustment was performed using the following explanatory variables and covariates: age, gender, habit of smoking, habit of alcohol drinking, history of diabetes, adiposity (BMI), blood pressure (mean arterial pressure), and histories of medication therapy using insulin and medication therapy using anticoagulation and/or antiplatelet drugs. Bonferroni’s multiple comparison test was used in analyses of the seven HDP-related peptides. Statistical significance was considered when probability (p) values were less than 0.05.

The characteristics of the participants are shown in Table 1. The median age of the participants was 75 years, and 24.2% of the participants were smokers and 41.2% and 86.7% of the participants were suffering from diabetes and hypertension, respectively. About 80% of the participants were receiving anticoagulation therapy and/or antiplatelet therapy, and about 60% of the participants had received surgical and/or endovascular therapy for LEAD. The mean CAVI level of right-side measurements was slightly but not significantly higher than that of left-side measurements (mean with 95% confidence interval: right, 9.23 [8.88–9.58]; left, 9.05 [8.74–9.37], p = 0.461). The proportion of participants showing high diff-CAVI was 40.6%. The ranges of the seven peptide concentrations in serum varied: P-2858 levels were relatively high (135.6–4,571.7 ng/mL), while mean P-2081 and P-2127 levels were relatively low (<1 ng/mL).

CAVI and diff-CAVI were compared between the groups of participants with and without low ABI (≤0.9) (Table 2). CAVI was significantly lower in the group of participants with ABI of ≤0.9 than in the group of participants with ABI of >0.9. On the other hand, log-transformed diff-CAVI was significantly higher in the group of participants with ABI of ≤0.9 than in the group of participants with ABI of >0.9. Thus, in patients with LEAD, diff-CAVI, but not CAVI itself, is thought to reflect leg ischemia.

Table 3 shows correlation coefficients between peptide levels and mean levels of right and left CAVI in univariate and multivariate analyses. In multivariate analysis, age, gender, habit of smoking, history of diabetes, and mean arterial pressure were adjusted (Multivariate-1 in the table). In addition, BMI, habit of alcohol drinking, and histories of medication therapy using insulin and medication therapy using anticoagulation and/or antiplatelet drugs were adjusted in other analyses (Multivariate-2 in the table). Both in univariate and multivariate analyses, P-2127 levels showed significant positive correlations with CAVI, while P-2091 and P-2378 levels showed significant inverse correlations with CAVI. P-2081, P-2209, or P-2858 did not show significant correlations with CAVI in multivariate analysis with adjustment for the above nine explanatory variables (Multivariate-2). P-3156 did not show a significant correlation with CAVI in univariate or multivariate analyses.

Table 4 shows results for the correlation coefficients between peptide levels and diff-CAVI. In both univariate and multivariate analyses, P-2091 and P-2378 were significantly correlated with diff-CAVI, while the other peptides (P-2081, P-2127, P-2209, P-2858, and P-3156) did not show significant correlations with diff-CAVI except for a weak but significant inverse correlation between P-2127 and diff-CAVI in univariate analysis.

The mean levels of each peptide were compared in the groups with and without large diff-CAVI (≥1.05) in univariate and multivariate analyses. Since the peptide levels did not show normal distributions, they were used after base-10 logarithmical transformation. As shown in Table 5, P-2091 and P-2378 were significantly higher in the group with large diff-CAVI than in the group without large diff-CAVI in univariate and multivariate analyses. There were no significant differences in mean levels of the other peptides, including P-2081, P-2127, P-2209, P-2858, and P-3156, between the participant groups with and without large diff-CAVI in univariate and multivariate analyses, except for a significantly lower mean P-2127 level in the group with large diff-CAVI in univariate analysis and multivariate analysis (Multivariate-1).

The prevalence of large diff-CAVI was significantly different among the tertile groups of P-2091 and P-2378 but not among those of the other peptides including P-2081, P-2127, P-2209, P-2858, and P-3156 (Table 6). The prevalence was significantly higher in the 3rd tertile groups of P-2091 and P-2378 than in the corresponding 1st tertile groups and was significantly higher in the 2nd tertile group of P-2378 than in the 1st tertile group.

Table 7 and Figure 1 show the results of logistic regression analysis for the relationship between each peptide level and large diff-CAVI. Odds ratios (ORs) for large diff-CAVI of the 3rd vs. 1st tertiles of P-2091 and P-2378 were significantly higher than the reference level of 1.00. ORs for large diff-CAVI of the 2nd vs. 1st tertiles of P-2378 were also significantly higher than the reference level. The above results were obtained in both univariate and multivariate analyses with adjustment for the above nine explanatory variables. On the other hand, ORs for large diff-CAVI of the 2nd and 3rd vs. 1st tertiles of the other peptides (P-2081, P-2127, P-2209, P-2858, and P-3156) were not significantly different from the reference level in univariate and multivariate analyses.

Table 8 shows the correlation coefficients of each peptide levels with ABI and diff-ABI levels in univariate and multivariate analyses. P-2081, P-2127, and P-2209 levels showed significant positive correlations with ABI, while P-2091 and P-2378 levels were significantly inversely correlated with ABI. P-2091 and P-2378 showed significant positive correlations with diff-ABI, while the other five peptides did not show significant correlations with diff-ABI in multivariate analysis.

In patients with LEAD, P-2091 (derived from fibrinogen-α) and P-2378 (derived from complement C4) but not the other five HDP-related peptides were demonstrated to be associated with diff-CAVI, which has been proposed as a new indicator of leg ischemia [17]. This study is the first study showing the usefulness of diff-CAVI. Because both diff-CAVI and HDP-associated peptides have been reported to show associations with ABI [2, 17], it is reasonable that there were significant relationships between diff-CAVI and the two HDP-associated peptides. P-2091 and P-2378 were significantly positively correlated with diff-ABI, while the other five peptides did not show significant correlations with diff-ABI (Table 8). Thus, these relationships between the seven peptides and diff-ABI are similar to the relationships between these peptides and diff-CAVI (Table 4). These findings are reasonable because both diff-ABI and diff-CAVI are thought to reflect leg ischemia in patients with LEAD. The associations between diff-CAVI and the peptides were also found in multivariate analysis with adjustment for age, gender, habits of smoking and alcohol drinking, history of diabetes, blood pressure, BMI, and histories of therapy using insulin and anticoagulation and/or antiplatelet drugs (Table 4, Table 5, Table 7). Thus, P-2091 and P-2378 are associated with diff-CAVI, and the associations are independent of the above possible confounders including risk factors of LEAD.

This study is also the first study in which the relationships between HDP-related peptides and arterial stiffness were investigated. Progression of atherosclerosis, the major pathogenesis of LEAD, is closely linked with increased arterial stiffness [6], which results in an increase of CAVI. In patients with LEAD, CAVI is known to show a paradoxical positive association with ABI [15, 16], which is decreased by leg ischemia. ABI shows a U-shaped relationship with arterial stiffness when a considerable number of cases with high ABI (>1.4), resulting from highly calcified arterial lesions, are included in the participants. However, in this study, there were no participants showing abnormally high ABI (1.4 or higher), and consequently, ABI showed a significant positive correlation with CAVI (Pearson’s correlation coefficient: right, 0.364 [p < 0.01]; left, 0.328 [p < 0.01]). Therefore, no inclusion of participants with abnormally high ABI is the reason for a linear positive correlation but not a U-shaped relationship between ABI and CAVI in this study.

Theoretically, diff-CAVI can be perfectly normal in the presence of bilateral stenosis of the same severity in each side. However, there is a unilateral lower extremity predisposition to LEAD [25, 26], resulting in a larger difference between right and left CAVI levels. Thus, from the hydrodynamic viewpoint, diff-CAVI is thought to be a reasonable index for evaluation of leg ischemia in patients with LEAD. Moreover, diff-CAVI is especially helpful for evaluation of leg ischemia in LEAD patients with progressed calcification in the arterial wall, which causes an increase in ABI. diff-CAVI, instead of ABI, is, therefore, a useful marker for evaluation of leg ischemia in LEAD patients with highly calcified arterial lesions.

diff-CAVI was significantly smaller in participants with a history of surgical and/or endovascular therapy than in participants without such a history (median with interquartile range of diff-CAVI: 0.60 [0.20, 1.60] vs. 1.35 [0.43, 3.35], p < 0.001). In addition, levels of P-2091 and P-2378, which were associated with diff-CAVI, were compared between the participant groups with and without a history of surgical and/or endovascular therapy. P-2091 and P-2378 were lower in the group with the therapy history than in the group without it (P-2091 [ng/mL]: 1.46 [0.56, 5.44] vs. 5.99 [2.53, 17.98] [p < 0.01]; P-2378 [ng/mL]: 2.17 [0.78, 18.09] vs. 35.07 [4.70, 51.35] [p < 0.01]). These results suggest the usefulness of diff-CAVI, P-2091 and P-2378 as indicators for evaluating the effectiveness of interventional therapy for patients with LEAD.

As mentioned above, in a previous study in which leg arterial flow was evaluated by using ABI, leg ischemia was inversely associated with P-2081 (derived from kininogen 1) and P-2127 (derived from kininogen 1) and was positively associated with P-2091 and P-2378 [2]. On the other hand, in the present study, CAVI showed a positive association with P-2127 and inverse associations with P-2091 and P-2378. Thus, the relationships between these three HDP-related peptides and CAVI agree with the paradoxical relationship between CAVI and leg ischemia, which may be explained by lower mean arterial pressure distal to stenosis, leading to slow transit time in the legs. Thus, the paradoxical relationships of CAVI with the four peptides, P-2081, P-2091, P-2127, and P-2378, indirectly suggest the associations between these peptides and leg ischemia.

Although the HDP-related peptides were associated with ABI and diff-CAVI, none of the HDP-related peptides showed an association with intima-media thickness of the carotid artery in patients with LEAD in our recent study [27]. Thus, the peptides are thought to reflect not the degree of progression of atherosclerosis but the degree of decrease in leg arterial flow. This agrees with the results of the present study for the relationships of the peptides with CAVI and diff-CAVI: The peptide levels were paradoxically and logically associated with CAVI and diff-CAVI, respectively (Tables 3, 4). The paradox of the relationships between the peptides and CAVI support the hypothesis that the HDP-related peptides are not related to atherosclerosis.

Only two HDP-related peptides, P-2091 and P-2378, showed positive associations with diff-CAVI in the present study, while six HDP-related peptides, including P-2081, P-2091, P-2127, P-2209 (derived from kininogen 1), P-2378 and P-2858 (derived from α2-HS-glycoprotein), were reported to be associated with leg ischemia evaluated by ABI [2]. The dissociation of these findings is possibly explained by a difference in sensitivities of diff-CAVI and ABI to leg ischemia, and thus, ABI might be a more sensitive marker than diff-CAVI for LEAD. Another possible reason for the above dissociation of the findings regarding associations of the peptides with ABI and diff-CAVI is a difference in sensitivities of the peptides to leg ischemia. P-2091 and P-2378 might be more prone to be affected by a decrease of arterial flow than P-2081, P-2127, P-2209, and P-2858. In fact, P-2091 and P-2378 were reported to be more strongly correlated with ABI than were the other four peptides in the previous study (Spearman’s rank correlation coefficients: 0.227 [P-2081] vs. −0.349 [P-2091] vs. 0.279 [P-2127] vs. 0.229 [P-2209] vs. −0.342 [P-2378] vs. −0.135 [P-2858]) [2]. P-2081, P-2127, and P-2209 originate from the same parent protein of kininogen 1, while P-2091, P-2378, and P-2858 originate from fibrinogen-α, complement C4, and α2-HS-glycoprotein, respectively. Thus, P-2091 and P-2378, which showed associations with diff-CAVI, originate from different parent proteins. Therefore, further studies are needed to clarify the causality and mechanisms of the relationships between these peptides and leg ischemia due to LEAD. Interestingly, in patients with LEAD, P-3156 (derived from inter-α-TIHC H4) was not associated with diff-CAVI in the present study or with ABI in the previous study [2]. On the other hand, P-3156 was reported to be associated with risk factors of atherosclerotic disease including adiposity index and blood lipids in a general population [3]. Therefore, P-3156 is speculated to be not a biomarker for leg ischemia but a biomarker for progression of atherosclerosis.

There are limitations of this study as follows. The ages of 73% of the participants (n = 121) were 70 years or older, and male and female participants were not analyzed separately because the number of female participants was not large enough for analysis. Therefore, further studies using a database of younger or female patients are needed to confirm the findings of this study. CAVI reflects heart-ankle PWV but does not reflect stiffness of more peripheral foot arteries, which needs to be investigated in future studies. Most of the participants of this study had already received therapeutic intervention, and thus, P-2091 and P-2378 are thought to be biomarkers for evaluating post-therapeutic status of leg arterial flow. Therefore, further studies are needed to test the usefulness of the HDP-related peptides for initial diagnosis of LEAD. Major risk factors of LEAD including age, gender, smoking, adiposity, diabetes, and blood pressure [1] were adjusted in multivariate analyses; however, there are other possible confounding factors for the relationships between peptide levels and diff-CAVI, e.g., physical activity, nutrition, diet, and socio/economic factors including education and occupation, for which information was not included in the database used in this study. Since this study is cross-sectional in its design, further prospective studies are needed to discuss the causality of the relationships between the peptides and decreased leg arterial flow in LEAD.

Among the seven HDP-related peptides, P-2091 and P-2378 were positively associated with diff-CAVI and were inversely associated with CAVI. Since diff-CAVI has been shown to be associated with leg ischemia and a paradoxical relationship is known between CAVI and leg ischemia, P-2091 and P-2378 are thought to reflect leg ischemia and to be useful biomarkers for progression of LEAD. Mechanisms of the associations between these peptides and leg ischemia due to LEAD remain to be clarified.

The authors thank Dr. Mitsuaki Yanagida for measurement of the HDP-related peptides.

The protocol of this study was approved by the Ethics Committee of Yamagata Saisei Hospital (Approval No. 199 at the Ethics Committee since 2013). Written informed consent to participate this study was obtained from all of the participants.

The authors declare that they have no conflict of interest to disclose.

This study was supported by a Grant-in-Aid for Scientific Research (No. 21H03386) from the Japan Society for the Promotion of Science.

I.W. contributed to study conception and design, data analysis, and drafting and reviewing the manuscript, and Y.S. contributed to study conception and design, data collection, and reviewing the manuscript.

Data sets generated and analyzed for the current study are not publicly available due to their containing information that could compromise the privacy of research participants but are available from the corresponding author (I.W.) upon reasonable request.