Comparison of carotid intima-media thickness between hemodialysis and peritoneal dialysis patients

Abstract

BACKGROUND

Patients with chronic kidney disease (CKD) are at high risk of developing cardiovascular disease (CVD). Carotid intima-media thickness (cIMT) is a strong predictor of CVD. Klotho has been associated with increased atherosclerotic plaque stability in CKD patients.

AIM

To investigate the association between dialysis modality and cIMT in dialysis patients.

METHODS

A cross-sectional study of 83 dialysis patients, 40 hemodialysis (HD) patients and 43 peritoneal dialysis (PD) patients, who met the inclusion criteria. Anamnesis, carotid artery ultrasonography (USG), and laboratory tests were performed on all patients. cIMT measurements were performed in 6 carotid territories using arterial USG.

RESULTS

PD patients had lower cIMT [0.53 (0.47; 0.67) vs 0.63 (0.54; 0.70); P = 0.005]. Younger age in PD patients was independently associated with lower cIMT (β = 0.006; P = 0.003). PD patients had a lower FGF23/Klotho ratio compared to HD patients [26.26 (13.32; 69.45) vs 67.34 (16.56; 131.35); P = 0.039]. Similarly, in PD patients, a moderate positive correlation was found between calcium, phosphorus, and CaxP product levels with fibroblast growth factor 23 (FGF-23) levels. However, the strength of the correlation was reduced when correlated with the FGF-23/Klotho ratio.

CONCLUSION

This study showed that PD patients have lower cIMT than HD patients. Younger age was independently associated with lower cIMT. On the other hand, FGF-23 and the FGF-23/Klotho ratio were significantly positively correlated with CaxP product in PD patients, but not in HD patients. Klotho reduces the strength of this correlation, thus appearing play an important role in the lower cIMT in these PD patients.

Key Words: Calcium; Phosphorus; Carotid artery intima-media thickness; Fibroblast growth factor 23/Klotho ratio; Peritoneal dialysis

Core Tip: Chronic kidney disease patients are at high risk of developing cardiovascular disease (CVD). However, traditional risk factors cannot fully predict the incidence of CVD. Carotid artery intima-media thickness (cIMT) is a strong predictor of CVD. There is evidence of a relationship between dialysis modality and cIMT. In this study, peritoneal dialysis (PD) patients were shown to have lower cIMT than HD, making it a preferred option for patients with end-stage renal disease in the context of CVD risk. This may be due to better preservation of residual kidney function, no exposure to dialysis membranes, and the absence of hemodynamic fluctuations in PD patients.

INTRODUCTION

Patients with chronic kidney disease (CKD) have a substantially higher risk of developing cardiovascular disease (CVD). Among those with end-stage renal disease (ESRD), the likelihood of developing CVD is 3.5 to 50 times greater than in the general population, with vascular calcification as one of underlying causes[1]. The main contributor to this elevated mortality rate is dialysis patients, with CVD is responsible for 53% of fatalities[2]. In patients with CKD, the hallmark of atherosclerosis is more advanced than in the general population[3,4]. Atherosclerosis is an independent factor associated with cardiovascular events (CVE) in ESRD, and plaque involvement in each additional arterial territory increases the risk of CVE by 86%[4,5].

Traditional risk factors, including diabetes, hypertension, and lipid metabolism disorders are commonly found in patients with CKD and cannot fully predict CVE in these dialysis patients, which implies that other factors in CKD patients may accelerate the occurrence and progression of atherosclerosis. It is suspected that increased cardiovascular risk is related to factors related to CKD itself, such as dialysis modality, bone metabolism disorders, secondary hyperparathyroidism, increased calcium-phosphate production, oxidative stress and inflammation, resistance to fibroblast growth factor 23 (FGF-23), Klotho deficiency, and deficiency of vascular calcification (VC) inhibitors[6]. CKD causes bone metabolism disorders, leading to calcium and phosphate deposits in the blood vessel walls and causing VC in the intima and media layers. VC emerges early in CKD and increases markedly with declining kidney function, putting patients with CKD and ESRD at high risk for mortality and cardiovascular morbidity[7-9]. Multiple factors contribute to VC in CKD, including hypercalcemia, inflammatory cytokines, and especially, phosphate[10,11], and elevated phosphate recognized as a major trigger of VC.

Determination of carotid intima-media thickness (cIMT) has shown promising prognostic value in CKD populations[4,12]. cIMT is a well-established surrogate marker of atherosclerosis and an early marker of subclinical atherosclerosis[13], while biochemical markers such as calcium, phosphorus, CaxP products, FGF-23, and the FGF-23 to Klotho ratio reflect the body's mineral balance related to VC. There are significant differences between the two dialysis modalities, which may have differential effects on the determinants of cIMT. Several previous studies have demonstrated that dialysis modality affects cIMT[4,14-16]. It is well known that PD patients exhibit a potentially higher atherogenic profile than HD patients, resulting from glucose-based dialysis solutions and persistent peritoneal protein leakage. Conversely, HD patients exhibit poorer preservation of residual kidney function, as periodic fluctuations may result in inflammation, endothelial dysfunction, and VC. Furthermore, PD is associated with stable fluid status and blood pressure patterns compared with the periodic hemodynamic fluctuations observed in HD[4]. Therefore, this study aimed to determine the association between dialysis modality and cIMT in dialysis patients. With this understanding, clinicians can design more personalized management strategies for hemodialysis and peritoneal dialysis patients.

MATERIALS AND METHODS

Patients and study design

This is a cross-sectional study of patients undergoing continuous outpatient peritoneal dialysis (PD) and regular hemodialysis (HD). This study was conducted in the dialysis unit of the Department of Internal Medicine, Faculty of Medicine, Universitas Brawijaya, Dr. Saiful Anwar General Hospital, Malang, from July to December 2024. The inclusion criteria included patients aged > 18 who had undergone dialysis for > 3 months. All subjects were in good cardiovascular condition (i.e., independent in daily activities). The exclusion criteria were patients with acute diseases, such as severe infections, acute coronary syndrome, acute heart failure, cerebral infarction, or brain hemorrhage, as well as chronic diseases, such as coronary artery disease, congestive heart failure, chronic obstructive pulmonary disease, and other chronic inflammatory diseases. Patients with malignancies, immunosuppressant therapy, and previous CVD were also excluded from the study.

Data collected at recruitment included demographic information, anthropometric data, clinical history (diabetes, hypertension, and dyslipidemia), medication history, causes of ESRD, carotid ultrasound results, and laboratory results. All patients provided written and signed informed consent before participating in the study. This study has received approval from the Hospital Ethics Committee under registration No. 400/267/K.3/302/2023.

Dialysis methods

CAPD patients received four bags of PD solution containing 1.5% glucose per day. Those with insufficient ultrafiltration were given a PD solution containing 2.5% glucose. All PD solutions were manufactured by Baxter Healthcare Limited Company. HD patients underwent dialysis twice weekly (10 hours/week), with each session lasting 4-5 hours, with both conventional and polysulfone dialyzers.

Clinical and biochemical data

Blood samples were drawn from each patient in the morning after an overnight fast. Samples were analyzed for complete blood cell count, differential leukocyte count, serum creatinine, serum urea, hemoglobin, serum albumin, phosphorus, corrected serum calcium, iron, and serum transferrin, intact parathyroid hormone (iPTH), FGF-23, and Klotho. Laboratory sample analysis was performed in a centralized laboratory at the hospital. For HD patients, blood was drawn immediately before the dialysis session, while for PD patients, blood was drawn during an outpatient examination.

Carotid ultrasonography

Carotid ultrasonography was performed by a consultant radiologist blinded to clinical data using a Mindray DC-7 System equipped with a high-frequency linear-array probe (7-12 MHz). To ensure intra-observer reliability, cIMT was measured at end-diastole using the 'double-line’ sign on the far wall of the artery.

Measurements were captured at three standardized segments bilaterally: The common carotid artery (1 cm proximal to the bulb), the carotid bifurcation, and the internal carotid artery (1 cm distal to the bulb). At each segment, three separate readings were taken and averaged to minimize manual measurement error. The final composite cIMT was defined as the mean of these six anatomical sites. Intra-class correlation coefficients (ICC) were calculated from a subset of 20 images re-analyzed 14 days apart to validate intra-observer consistency and the ICC value obtained was 0.947.

Statistical analysis

Data are reported as proportions or means with standard deviation or medians with interquartile ranges. We performed the Pearson χ² test to compare categorical variables expressed as n (%). Normally distributed data were subjected to an unpaired t-test, while those not normally distributed were subjected to the Mann-Whitney U test. The Pearson or Spearman correlation test was used according to the data distribution to determine the relationship between parameters.

Variables with a P value < 0.25 or that were statistically significant in the univariate analysis were included in the multivariate regression analysis, with a 95% confidence interval. Statistically significant differences are indicated by a two-sided P value of less than 0.05. All calculations were performed using IBM SPSS Statistics for Windows software (version 25; IBM Corp., Armonk, NY, United States).

RESULTS

A total of 83 consecutive subjects with ESRD undergoing continuous outpatient PD (43 patients) and HD (40 patients) with complete documentation were included in this study (Table 1).

Table 1 Characteristics of study populations based on dialysis modality.

Characteristic	PD (n = 43)	HD (n = 40)	P value
Age (year), mean (SD)	41.14 (12.67)	51.85 (11.06)	< 0.0012,a
Male	22 (51.2)	21 (52.5)	0.9031
Female	21 (48.8)	19 (47.5)
Cause of ESRD
Diabetes	5 (11.6)	5 (12.5)	0.9321
Hypertension	30 (69.8)	24 (0.6)
Unknown	8 (18.6)	10 (0.25)
Duration of dialysis, month	40 (15; 52)	35.5 (15.5; 64.5)	0.4593
Hemoglobin, gr/dL	8.32 (2.13)	9.55 (2.26)	0.0122,a
NLR	4.02 (2.08)	2.98 (1.77)	0.0012,a
TSAT, %	29 (23; 43)	27 (23; 37.5)	0.9432
Albumin, g/dL	3.52 (0.69)	2.98 (1.77)	0.0682
Calcium, mg/dL	7.72 (1.54)	7.91 (5.81)	0.0672
Phosphor, mg/dL	4.55 (1.41)	3.80 (2.20)	0.8402
Calcium/phosphor product	35.97 (12.34)	35.02 (24.56)	0.8212
Klotho, ng/mL	6.12 (2.99; 9.11)	5.30 (1.88; 8.36)	0.2863
FGF-23, pg/mL	132.26 (77.82; 312.93)	148.66 (106.05; 457.42)	0.1553
FGF-23/Klotho ratio	26.06 (13.32; 69.45)	67.34 (16.56; 131.35)	0.0393,a
cIMT, mm	0.53 (0.47; 0.63)	0.63 (0.54; 0.70)	0.0053,a
iPTH, pg/mL	603.10 (347.00; 841.10)	209.05 (110.55; 423.87)	< 0.0013,a

¹χ² test.

²Independent t-test.

³Mann–Whitney U test.

^aSignificant values.

Data were presented in n (%), mean (SD), median (interquartile range). PD: Peritoneal dialysis; HD: Hemodialysis; NLR: Neutrophil-lymphocyte ratio; TSAT: Transferrin saturation; FGF-23: Fibroblast growth factor 23; cIMT: Carotid intima-media thickness; iPTH: Intact parathyroid hormone.

There were no significant differences between the PD and HD groups in terms of gender, cause of ESRD, dialysis duration, transferrin saturation, serum albumin levels, serum calcium and phosphorus levels, calcium-phosphorus product, serum Klotho levels, and FGF-23 levels (P > 0.05) (Table 1). PD patients were significantly younger and had lower hemoglobin and albumin levels than HD patients. Similarly, PD patients had lower cIMT compared to HD (P = 0.005), higher neutrophil-lymphocyte ratio (NLR) values (P = 0.001), and higher iPTH levels (P ≤ 0.001) (Table 1). However, based on multiple regression analysis, only age was independently associated with cIMT (β = 0.006; adjusted R² = 0.161, P = 0.003). The age-related increase in cIMT was found to be 0.006 mm/year (Table 2).

Table 2 Factors associated with carotid intima-media thickness in dialysis patients (n = 83).

Variables	Coefficient (β)	Standard error	t-value	95%CI	P value
Age, year	0.006	0.002	3.078	0.002-0.009	0.0031,a
Hemoglobin, g/dL	0.002	0.010	0.211	-0.018 to 0.022	0.833
NLR	-0.006	0.011	-0.547	-0.029 to 0.017	0.586
Albumin, g/dL	0.004	0.018	0.189	-0.034 to 0.041	0.851
Calcium, mg/dL	0.000	0.006	0.076	-0.011 to 0.012	0.940
FGF-23/Klotho ratio	-0.050	0.080	-0.563	-0.229 to 0.128	0.575
FGF-23, pg/mL	0.053	0.249	1.641	-0.011 to 0.118	0.105
iPTH, pg/mL	-0.025	0.024	-1.039	-0.074 to 0.023	0.302

¹Multivariate regression analysis.

^aSignificant values. The determinant value (adjusted R²): 0.161 and the β value: 0.006.

NLR: Neutrophil-lymphocyte ratio; FGF-23: Fibroblast growth factor 23; CIMT: Carotid intima-media thickness; iPTH: Intact parathyroid hormone.

cIMT in PD patients, but not HD, correlated with Klotho levels (Table 1), and had a lower FGF-23/Klotho ratio compared to HD patients [26.26 (13.32; 69.45) vs 67.34 (16.56; 131.35); P = 0.039] (Table 3). Calcium and Phosphorus levels, and CaxP product did not differ between PD and HD patients. However, in PD patients, a moderate positive correlation was found between CaxP product and FGF-23 and the FGF-23/Klotho ratio, with correlations of r = 0.592 (P ≤ 0.001), and r = 0.486 (P = 0.001), respectively (Table 3). On the other hand, HD patients had significantly lower iPTH levels than PD patients and a moderate positive correlation with FGF-23 levels.

Table 3 Correlation between several chronic kidney disease mineral bone disorder parameters between peritoneal dialysis and hemodialysis patients.

Variable	PD, r (P value)			HD, r (P value)
	Klotho, ng/mL	FGF-23, pg/mL	FGF-23/Klotho ratio	Klotho, ng/mL	FGF-23, pg/mL	FGF-23-Klotho ratio
Calcium, mg/dL	-0.056 (0.723)	0.465 (0.002)a	0.408 (0.007)a	0.056 (0.732)	0.078 (0.634)	0.016 (0.921)
Phosphor, mg/dL	-0.088 (0.575)	0.440 (0.003)a	0.341 (0.025)a	0.119 (0.465)	0.094 (0.563)	-0.013 (0.938)
Calcium/phosphor product	-0.098 (0.530)	0.592 (< 0.001)a	0.486 (0.001)a	0.065 (0.692)	0.144 (0.376)	0.033 (0.838)
cIMT, mm	0.345 (0.023)a	0.171 (0.273)	-0.100 (0.522)	0.148 (0.361)	0.096 (0.555)	-0.051 (0.756)
PTH, pg/mL	-0.178 (0.254)	0.069 (0.659)	0.165 (0.290)	0.309 (0.052)	0.459 (0.003)a	0.132 (0.415)
NLR	0.129 (0.409)	0.274 (0.075)	0.154 (0.324)	0.015 (0.927)	-0.014 (0.933)	0.009 (0.958)

^aSignificant values.

Spearman correlation. Neutrophil-lymphocyte ratio (NLR) correlated with calcium levels (r = 0.313; P = 0.049) and phosphorus levels (r = 0.384, P = 0.015) only in peritoneal dialysis (PD) patients. Overall [both PD and hemodialysis (HD) patients], NLR correlated only with phosphorus levels (r = 0.349, P = 0.001). There was no significant correlation between carotid intima-media thickness and intact parathyroid hormone in PD patients (r = -0.219; P = 0.158) and HD patients (r = -0.180; P = 0.267). PD: Peritoneal dialysis; HD: Hemodialysis; cIMT: Carotid intima-media thickness; PTH: Parathyroid hormone; NLR: Neutrophil-lymphocyte ratio; FGF-23: Fibroblast growth factor 23.

DISCUSSION

The leading contributor to cardiovascular morbidity in patients with ESRD is atherosclerosis. The earliest site affected in the process of atherosclerosis occurrence and development is the tunica intima-media[17]. Considering the fact that the carotid and coronary vessels have the same risk of developing atherosclerosis[18], increased cIMT in uremic patients serves as a marker of coronary atherosclerosis and prognostic factor for cardiovascular and cerebrovascular events.

The mechanisms underlying VC in dialysis patients are multifactorial. No single factor can explain the occurrence of VC in dialysis patients. Factors associated with VC include CKD-mineral bone disorder (CKD-MBD), characterized by elevated levels of phosphate, calcium, calcium-phosphate products, and the hormone PTH; deficiencies of VC inhibitors including Klotho, fetuin-A, and matrix protein Gla; resistance to FGF-23; inflammatory reactions and oxidative stress; and phenotypic changes in vascular smooth muscle cells (VSMCs)[6].

According to this study, PD patients had lower cIMT compared to those undergoing HD (Table 1). Multivariate regression analysis showed that only age was independently associated with cIMT (β = 0.006; adjusted R² = 0.161; P = 0.003). Previous data indicate that cIMT is strongly linearly associated with age, and is not influenced by CVD or risk factors. The age-related increase in cIMT is comparable between populations with and without CV risk factors (0.006 mm/year vs 0.008 mm/year)[19,20]. In this study, we found an age-related increase in cIMT of 0.006 mm/year (Table 2). In addition to age (adjusted R² = 0.161), several factors may contribute to cIMT, including preservation of residual renal function, exposure to dialysis membrane and apparatus, and hemodynamic fluctuations, which differ between the two dialysis modalities. Residual renal function facilitates fluid and electrolyte regulation and is associated with better nutritional status and survival. Residual renal function is typically better preserved in PD patients and takes part in slowing the progression of atherosclerosis in this population[21]. Contact between blood cells and the complement system with the dialysis membrane causes an inflammatory reaction and activation of the complement cascade. This is followed by the release of inflammatory mediators such as reactive oxygen species (ROS), several cytokines [interleukin (IL)-1, IL-6, IL-8, IL-12, IL-13, tumor necrosis factor alpha, monocyte chemoattractive peptide-1, and interferon][21]. Similarly, complement activation will induce trans endothelial migration of inflammatory cells to the site of inflammation, releasing ROS and degradative enzymes. The subsequent impact is tissue damage and increases the development of atherosclerosis[21]. PD patients exhibit more stable fluid volume status and lower blood pressure variability compared to HD patients. Each dialysis session causes rapid changes in intravascular volume and blood pressure, resulting in shear stress on the carotid artery wall. These hemodynamic fluctuations lead to dysfunction and accelerated atherogenesis through inducing the release of inflammatory mediators and activating signaling pathways that drive smooth muscle cell proliferation and extracellular matrix deposition, which result in increased intima-media thickness[21].

Another factor associated with lower cIMT is more stable phosphate levels in PD patients, due to better preservation of residual renal function compared to HD patients. Phosphate clearance in PD patients is continuous, sustained, and slower, resulting in lower fluctuations compared to HD patients. Low phosphate levels in HD patients do not always reflect true phosphate levels, as HD patients experience greater fluctuations in phosphate levels, with high levels before HD and low levels after HD[21].

Klotho and FGF-23 are CKD-MBD parameters that exhibit alterations earlier than other CKD-MBD parameters, such as phosphate and PTH. Klotho has multiple functions, including acting as a co-receptor for FGF-23 and protecting against VC and protection against VC and mineral bone disorder[7]. In our study, the FGF-23/Klotho ratio was lower in PD patients than in HD patients (Table 1). In PD patients, a moderate positive correlation was also found between calcium and phosphorus levels, and the CaxP product, with FGF-23 levels. Interestingly, the correlation strength decreased when calcium and phosphorus levels, and the CaxP product, were correlated with the FGF-23/Klotho ratio (Table 3). This suggests that Klotho exerts a protective effect against vascular remodeling and calcification. The FGF-23/Klotho ratio reflects both a risk factor (FGF-23) and a protective factor (Klotho) simultaneously. A low FGF-23/Klotho ratio is protective against vascular remodeling and calcification. On the other hand, a high FGF-23/Klotho ratio indicates the presence of FGF-23 resistance, decreased vascular protection and an increased risk of endothelial dysfunction and thickening of the tunica intima[22].

Hyperphosphatemia and a positive net calcium and phosphate balance play a crucial role in VSMC calcification and the development of VC[23,24]. However, the exact mechanism of phosphorus in the development and progression of VC remains unclear. Some mechanisms that promote and trigger the development of VC include: (1) The transition of VSMCs from a contractile phenotype to an osteochondrogenic phenotype and mineralization of the VSMC extracellular matrix; (2) Induction of VSMC apoptosis; (3) Inhibition of monocyte/macrophage differentiation into osteoclast-like cells; (4) Increased FGF-23 levels; and (5) Decreased Klotho expression[24]. On the other hand, Klotho plays a crucial role in protecting against VC. Reported functions of Klotho include inhibition of local phosphate transport in vascular cells, a shift in the phenotype of vascular cellular elements to bone-forming cells, reduced matrix mineralization and calcification, and preservation of endothelial functional properties and viability[8].

We also found a positive correlation between carotid intima-media thickness (CIMT) and Klotho, but only in PD patients (Table 3). This seemingly contradictory result is likely related to the role of Klotho itself in responding to hyperphosphatemia to remove phosphate in the presence of preserved residual renal function in PD, resulting in higher Klotho levels compared to HD (although not significantly). Increased Klotho levels, in turn, improve VC, as indicated by lower cIMT in PD patients. Another possibility is that Klotho acts as a counter-regulatory response to oxidative stress and vascular injury. The use of glucose-based dialysates is known to induce low-grade, active chronic inflammation[25], as demonstrated by significantly higher NLR in PD patients (Table 1). Conversely, Klotho is known to have protective effects against oxidative stress and has anti-inflammatory effects[26], which may counter the pathophysiological mechanisms that lead to increased cIMT and atherosclerosis. Increased Klotho levels help control oxidative stress and inflammation, in part by negatively regulating nuclear factor kappaB[27]. Thus, the presence of vascular stress may increase Klotho production as a protective response, even while intima-media thickening is already occurring. This creates a pseudo-positive correlation; the positive association reflects clinical complexity rather than a direct causal effect.

In our study, PTH levels were higher in PD patients than in those HD patients. However, there was no significant correlation between PTH levels and cIMT in either PD or HD patients (Table 3). cIMT is a longitudinal and accumulative process, while PTH is dynamic. Therefore, cIMT should be correlated with serial PTH levels (time-averaged PTH) for more accurate results[28].

PTH levels are influenced by many factors, including phosphate, calcium, and FGF-23 levels. PTH levels can remain elevated despite normal levels of calcium, phosphate, and calcium-phosphate products. This can occur due to autonomous PTH production, good phosphate and calcium control through a low-fat diet and/or the use of phosphate binders, and dialysis modalities. PD modalities allow for better phosphate control[28]. Increased PTH levels may indicate resistance to bone or parathyroid autonomy, not always due to increased CaxP products[29]. High PTH is not always accompanied by abnormal phosphate levels, because this condition can be caused by increased FGF-23 as an effort to increase phosphate excretion, but in fact the total phosphate levels in the body increase[28]. Intracellular and tissue phosphate deposits can accumulate in vascular walls, myocardium, and bone, leading to calcification and CKD-MBD despite normal serum phosphate levels. Furthermore, dialysis can also normalize phosphate levels, but it cannot eliminate total phosphate levels[28].

Study limitations

Some limitations of this study include the small sample size, single-time data collection, and cross-sectional design. Phosphate, calcium, iPTH, FGF-23, or Klotho levels tend to fluctuate depending on the time of collection, potentially underestimating the true levels and chronic exposure of these biomarkers to the vasculature. CIMT is a cumulative effect of VC and is longitudinal; therefore, correlating it with dynamic (time-averaged) phosphate, calcium, iPTH, FGF-23, or Klotho levels would provide more accurate results. Therefore, large-scale, prospective, multicenter randomized studies are needed to confirm our results.

CONCLUSION

This study shows that PD patients have lower cIMT than HD patients. Younger age independently associated with lower cIMT. On the other hand, FGF-23 and the FGF-23/Klotho ratio were significantly positively correlated with CaxP product in PD patients, but not in HD patients. Klotho reduces the strength of this correlation, thus appearing play an important role in the lower cIMT in these PD patients.

ACKNOWLEDGEMENTS

The authors would like to thank all the patients involved in this study and the staff at the dialysis unit at Dr. Saiful Anwar Hospital in Malang for their support. They also thank Winda and Luluk for their assistance in preparing this article.