Cardiovascular (CV) diseases contribute to a significant proportion of comorbidity conditions and mortality in patients with end-stage kidney disease (ESKD) [1]. Risk factors for CV diseases in patients undergoing regular peritoneal dialysis (PD) are classified as traditional (hypertension (HTN), diabetes mellitus (DM), dyslipidemia, age, sex, and family history), nontraditional (inflammation, oxidative stress, endothelial dysfunction, and protein carbamylation), and uremia-specific (uremic toxins, volume overload, hyperparathyroidism, hyperphosphatemia, and vascular calcification) factors [2]. Increased arterial stiffness (AS) is a critical estimator of CV events, CV mortality, and death in patients with ESKD [3, 4]. Pathogenetic factors underlying arterial stiffening included a change in the elastin-to-collagen ratio, altered lipid metabolism, oxidative stress, genetic mutations, and epigenetic regulation [5]. Furthermore, a close association was reported between AS and atherosclerosis [6, 7].

The cardio-ankle vascular index (CAVI) is a noninvasive parameter of global AS from the aorta to the tibial arteries. Compared with the carotid–femoral pulse wave velocity (PWV), the CAVI was considered to be less operator dependent and not influenced by blood pressure during the measurement [8, 9]. The CAVI was reported as being positively associated with the left atrial chamber size, indices of left ventricular diastolic dysfunction, such as the E/A ratio, and the left ventricular mass index [10]. In a small cross-sectional study in Taiwan, patients receiving PD regularly demonstrated a high prevalence of AS, when measured by the CAVI [11]. Compared with the general population, patients on hemodialysis have a significantly higher CAVI [12], which can be used to predict all-cause deaths in individuals on regular hemodialysis therapy [13].

Lipoprotein(a) (Lp(a)), which is produced primarily in the hepatic parenchyma, comprises a low-density lipoprotein core complexed with glycoprotein apo(a). High plasma Lp(a) concentrations indicate an increased risk for CV diseases [14, 15]. The proatherogenic, prothrombotic, and proinflammatory effects of Lp(a) accelerate atherosclerotic plaque formation [15]. Furthermore, the serum Lp(a) concentrations increase significantly as the glomerular filtration rate declines. Patients on chronic dialysis therapy have higher Lp(a) levels than those without kidney failure [16]. Histologically, high plasma Lp(a) concentrations were found to be positively correlated with arterial Lp(a) content and atherosclerosis progression in patients with ESKD [17]. In a cohort study that focused on PD recipients, the upper tertile of Lp(a) levels had the highest CV mortality rate but not all-cause mortality [18].

Thus, this study aimed to assess the association between the serum Lp(a) levels and AS in ESKD individuals on regular PD using the CAVI.

The baseline parameters of the 86 patients undergoing PD are shown in Table 1. Major chronic medical illnesses included DM (n = 39; 45.3%) and HTN (n = 68; 79.1%). A total of 35 patients exhibited a high CAVI. Compared with the group with a normal CAVI, the patients with a higher CAVI had lower serum albumin concentrations (p = 0.026) and higher total cholesterol (p = 0.003), triglyceride (p = 0.044), CRP (p 0.001), and Lp(a) (p 0.001) levels. Blood hemoglobin and serum levels of fasting glucose, blood urea nitrogen, creatinine, total calcium, phosphorus, iPTH, age, sex, BMI, blood pressure, total and dialysate clearance of solutes, smoking status, anti-DM drugs used, anti-HTN drugs used, and proportions of patients with underlying DM, HTN, and CAD did not differ significantly between the groups.

Multivariate logistic regression analysis was performed to identify the aspects independently correlated with AS and adjusted for albumin, total cholesterol, triglyceride, CRP, Lp(a), age, and gender. Lp(a) [odds ratio (OR) 1.025; 95% confidence interval (CI) 1.010–1.040; p = 0.001], total cholesterol (OR 1.042; 95% CI 1.005–1.081; p = 0.027), and CRP (each increase 0.1 mg/dL, OR 1.217; 95% CI 1.008–1.469; p = 0.041) were the independent risk factors for AS development among the study patients (Table 2).

Table 3 represents the link between the baseline parameters and CAVI (left and right) as well as serum log-transformed Lp(a) (log-Lp(a)) levels by Spearman’s correlation analysis. Both the left and right CAVIs showed a significant and positive correlation with serum log-Lp(a) levels (r = 0.588, p 0.001; r = 0.639, p 0.001, respectively). Moreover, serum log-Lp(a) levels were positively associated with serum total cholesterol (r = 0.218, p = 0.043) and log-CRP (r = 0.270, p = 0.012) levels in addition to BMI (r = –0.243, p = 0.024). Both the left and right CAVIs had a significant positive correlation with serum log-CRP levels (r = 0.349, p = 0.001; r = 0.373, p 0.001, respectively) and a negative correlation with serum albumin levels (r = –0.255, p = 0.018; r = –0.223, p = 0.039, respectively).

The most important findings in the study were that the Lp(a), total cholesterol, and CRP serum concentrations were independently related to AS in patients undergoing chronic PD therapy. Both the left and right CAVIs are also positively associated with the serum log-Lp(a) levels. It is possible to use Lp(a) to predict adverse CV outcomes in these patients.

The pathophysiology of the Lp(a) contribution to the atherosclerotic process is complex. Several proposed mechanisms explain the contributing role of Lp(a) in atherogenesis, including direct Lp(a) deposition in arterial walls, increased foam cell transformation, and induction of endothelial dysfunction [15]. The apolipoprotein(a) (apo(a)) domain of Lp(a) shows homology to plasminogen; hence, Lp(a) interferes with the ability of plasminogen to facilitate fibrinolysis and promote thrombi formation [15, 21]. High serum Lp(a) levels were positively associated with platelet aggregation, which is an alternative explanation of the thrombogenic effect of Lp(a) [22, 23]. Lp(a) accelerates coagulation by promoting tissue factor expression and inhibiting the activity of the tissue factor pathway inhibitors [24, 25, 26]. Furthermore, Lp(a) has a proinflammatory role in atherosclerosis. Lp(a) and associated oxidized phospholipids upregulate proinflammatory cytokines, such as interleukin-6 (IL-6), IL-1$\beta$, and tumor necrosis factor-$\alpha$, increase the expression of adhesion molecules, and induce monocyte chemotaxis and migration [26, 27]. In summary, Lp(a) has several roles in atherosclerosis, including direct deposition in arterial walls, increased foam cell transformation, induction of endothelial dysfunction, and promotion of thrombi formation. It also promotes coagulation by promoting tissue factor expression and inhibiting the activity of the tissue factor pathway inhibitors. In this study, the serum log-Lp(a) levels in PD recipients were also positively associated with log-CRP levels.

Both Lp(a) and AS are important biomarkers of atherosclerosis and CV diseases. The association between Lp(a) and AS has been extensively studied. In a cross-sectional study that recruited hypertensive patients, AS was determined by measuring the carotid–femoral PWV and calculating the brachial augmentation index. The study concluded that logarithmically transformed Lp(a) and CRP were independently associated with an increased index but not with carotid–femoral PWV [28]. Furthermore, in older patients with DM, the positive correlation between serum Lp(a) and AS, determined by measuring the aortic PWV, was independent of age, sex, glycemic control, and the use of lipid-lowering medications [29]. In another small study, which included female participants, among the participants with hypertension, only oxidized Lp(a) concentrations, and not Lp(a) levels, were positively and independently correlated with the CAVI [30]. However, whether oxidized Lp(a) has a better predictive role in AS or adverse CV outcomes is still under investigation [31]. Moreover, the efficacy of Lp(a)-lowering therapy on CV endpoints is still not well established [32].

Cholesterol is an important constituent in atherosclerotic plaques; hyperlipidemia is a traditional CV risk factor. Establishing a positive association between hypercholesterolemia and AS was based on patients with familial hypercholesterolemia [33]. In a large prospective community-based study, total cholesterol concentrations were positively correlated with the brachial–ankle PWV and cholesterol could mediate hypertension through AS [34]. However, in a cohort study of older patients with hypertension, the total cholesterol levels did not correlate with large AS, when measured by aortic distensibility, increased index, and systemic arterial compliance [35]. On the contrary, serum triglyceride levels were significantly higher in the high CAVI group; however, triglycerides failed to independently predict AS in PD recipients. In a prospective community-based study, triglyceride levels were associated with AS independently of multiple CV risk factors and statin therapy [36].

The contributing role of malnutrition and inflammation in atherosclerosis is well-established [37]. The pathophysiology is characterized by the expression of adhesion molecules in endothelial cells, monocyte migration and penetration into vessel walls, macrophage transformation into foam cells, generation of proinflammatory cytokines and reactive oxygen species, activation of the coagulation cascade with a tendency of thrombus formation, and further activation of T lymphocytes [38, 39]. Moreover, CRP is not only a marker of inflammation but also participates in the progression of atherosclerosis and intravascular thrombosis [40]. Among PD recipients, higher CRP levels are an independent risk factor for the development of CV diseases, while high CRP levels were an independent predictor of the increased incidence of CV events [41, 42]. Several studies have found an association between CRP concentrations and AS, with the study participants including healthy and older individuals [43, 44]. PD recipients with protein-energy wasting have a reverse correlation between albumin levels and CRP, and IL-6 levels [45]. In addition, the aortic PWV is inversely correlated with albumin in PD recipients [46]. In this study, the left and right CAVIs were significantly and positively linked to the serum log-CRP levels and negatively linked to the serum albumin levels. Furthermore, serum CRP levels in patients undergoing chronic PD therapy, after adjusting for albumin, total cholesterol, triglycerides, and Lp(a), were independently associated with AS.

Previous investigations indicated that low Lp(a) levels could predict new-onset diabetes when the serum Lp(a) was inversely proportional to insulin concentrations and degree of insulin resistance [47, 48]. However, in the present study, we did not find any significant correlation between fasting plasma glucose or underlying DM and Lp(a) serum levels among patients undergoing PD.

As seen in Table 3, the serum calcium concentrations were slightly positively correlated with the right CAVI, whereas phosphorus concentrations were slightly negatively correlated, although neither reached statistical significance. There was no association between the iPTH and CAVI. Vascular calcifications are a critical component of chronic kidney disease mineral and bone disorder (CKD-MBD). Previous studies revealed that higher serum calcium levels, as a higher iPTH, were positively linked to markers of vascular stiffening, such as PWV [49]. In another study on PD patients, hyperphosphatemia was found to independently predict the deterioration of coronary artery calcifications [50]. The mixed results explained the complexity of CKD-MBD; various factors such as fibroblast growth factor-23, vitamin D, and other promoters or inhibitors of vascular calcifications could also contribute to the process.

There are a few limitations in the current investigation. Firstly, there was a limited number of samples and the study was conducted in a single center. Secondly, the cross-sectional design limited the ability of the study to determine causal relationships between clinical variables and AS. Thirdly, we did not measure the serum oxidized Lp(a) concentrations in the study participants; however, whether oxidized Lp(a) has a better predictive value for AS than conventional Lp(a) needs further investigation.

In this study on patients with ESKD undergoing maintenance PD, along with serum total cholesterol and CRP levels, the most significant clinical implication was that serum Lp(a) levels were positively correlated to the CAVI and can be used as a potential novel biomarker of AS. Currently, whether higher Lp(a) concentrations adversely affect CV morbidity and long-term survival among these patients is unknown. Thus, further prospective studies with a longitudinal design are warranted to determine the causal role of Lp(a) on clinical outcome parameters.

The data presented in this study are available on request from the corresponding author.

BGH, HJL, YLL, and CHW conceived and designed the experiments. HJL, YLL, and CHW performed the experiments. PYH, BGH, and JPT contributed reagents and analyzed the data. PYH, HJL, YLL, and CHW wrote the original draft preparation. BGH, and JPT review and editing the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

The research protocol was approved by the human research ethics committee of the Hualien Tzu Chi Hospital, Buddhist Tzu Chi Medical Foundation (approval ID: IRB108-219-A). Informed consent was obtained from the patient’s guardian(s). The study was performed in accordance with the Declaration of Helsinki.

We are grateful to all participating patients for their cooperation and willingness.

This research was funded by Hualien Tzu Chi Hospital, Buddhist Tzu Chi Medical Foundation, Hualien, Taiwan, grant number TCRD-112-021 and TCMF-CP 110-02.

The authors declare no conflict of interest.