Cardiovascular Disease and Dialysis: A Review of the Underlying Mechanisms, Methods of Risk Stratification, and Impact of Dialysis Modality Selection on Cardiovascular Outcomes

Abstract

Cardiovascular disease is the leading cause of morbidity and mortality in patients with end-stage kidney disease (ESKD). To date, no trial has been adequately powered to evaluate the effects of different dialysis modalities on cardiovascular events or mortality. To properly assess the risks and benefits of each modality for individual patients, it is crucial to understand the unique cardiovascular risk factors in patients undergoing dialysis. This review explores the existing literature on cardiovascular risk assessment in this population. It examines the mechanisms contributing to increased risk, including volume overload, blood pressure abnormalities, mineral bone disorder, vascular calcification, uremia, anemia, and chronic inflammation. Additionally, we discuss data from trials assessing cardiovascular outcomes and compare various dialysis modalities, including in-center hemodialysis, frequent dialysis, nocturnal dialysis, and peritoneal dialysis.

1. Introduction

Cardiovascular disease (CVD) is the leading cause of morbidity and mortality among patients with end-stage kidney disease (ESKD) [1]. When stratified by age, gender, race, and diabetes, death from CVD is approximately 20 times higher in patients undergoing dialysis than in the general population [2]. In addition to traditional risk factors for cardiovascular disease such as age, hyperlipidemia, diabetes, and smoking, patients suffering from ESKD have additional risk factors that will be outlined below. This increase in risk for CVD appears to be similar in both the glomerulonephritis (GN) population and in the non-GN population, suggesting that, regardless of cause, loss of renal function leads to cardiovascular complications [3].

Addressing this critical issue requires collaboration among patients, nephrologists, cardiologists, and primary care physicians to develop and implement patient-centered preventive strategies tailored to the ESKD population.

The impact of dialysis modality on cardiovascular disease warrants greater attention. Currently, no definitive evidence suggests that one dialysis modality provides superior cardiovascular outcomes compared to another [4,5]. Designing a randomized controlled trial sufficiently powered to address this question poses significant challenges [6]. However, understanding the risk factors contributing to cardiovascular disease and their interaction with dialysis modalities can help inform individualized treatment recommendations.

In this review, we explore the physiological mechanisms underlying the accelerated progression of cardiovascular disease in patients with ESKD, examine the rationale behind the potential benefits of frequent dialysis sessions on cardiovascular outcomes, and review the existing literature comparing cardiovascular outcomes between hemodialysis and peritoneal dialysis.

2. Methods

We searched PubMed, Medline, Google Scholar, Web of Science, and the reference lists of relevant citations for studies on cardiovascular disease in chronic kidney disease and end-stage renal disease. Search terms included cardiovascular disease, ESRD, CKD, peritoneal dialysis, home hemodialysis, left ventricular hypertrophy, coronary artery disease, stroke, and heart failure. We included randomized controlled trials (RCTs), observational studies, meta-analyses, evidence-based reviews, and clinical practice guidelines written in the English language. We excluded articles if no full-text document was available for review.

3. Risk Assessment

Quantifying the risk for CVD in patients undergoing maintenance dialysis is challenging. Atherosclerotic cardiovascular disease risk calculators, such as the American Heart Association Pooled Cohort Equation and the European Society of Cardiology Systematic Coronary Risk Evaluation, recommend factoring high-risk conditions such as chronic kidney disease (CKD) into each patient’s risk assessment [7,8]. However, how these conditions should be incorporated into calculations remains uncertain. As described below, chronic hypertension, volume overload, mineral bone disorders, anemia, and the uremic milieu are unique risk factors that contribute to the development of accelerated atherosclerosis in ESKD and may change the sensitivity and specificity of commonly used tests for identifying coronary artery disease (CAD), making estimating the risk of CVD in the ESKD population complex [9]. The following section will provide an overview of CVD risk prediction models and the accuracy of commonly used imaging modalities for diagnosing CAD.

Coronary artery calcium (CAC) scoring utilizes computed tomography to measure the density of calcification within the coronary vasculature and has been identified as a suitable surrogate for the total atherosclerotic burden [10]. CAC scores of zero are associated with low risk for CVD [11]. Whether the same diagnostic thresholds can be used in the CKD and ESKD populations remains to be seen. One study evaluating the degree of coronary artery calcification in patients with CKD found a median CAC score of over 200 [12]. The reliability of CAC scoring in this population is also questionable since there is evidence to suggest no significant difference in composite cardiovascular outcomes with the use of calcium-based and non-calcium-based binders despite increased CAC scores when using calcium-based binders [13,14]. The high baseline level of coronary artery calcification in the CKD population, compared to the non-CKD population, and the apparent lack of correlation between CAC score and CV outcomes in the CKD population may reduce the discriminatory power of the CAC test and the applicability of the information gathered [12].

A multifaceted imaging approach may perform better than any one test. A study by Winther et al. compared CAC scores, coronary computed tomography angiography (CTA), single-photon emission computed tomography (SPECT), and coronary angiography in 138 patients with chronic kidney disease being evaluated for kidney transplantation [15]. A total of 43% of the patients were on maintenance dialysis. The investigators found that 22% of the patients had obstructive coronary artery disease, defined by a greater than 50% reduction in luminal diameter on coronary catheterization. Using a threshold of 400 units for the CAC score, they found the sensitivity and specificity to be 67% and 77% for obstructive CAD.

Regarding coronary CTA, they observed a sensitivity and specificity of 93% and 63%. For SPECT, the sensitivity and specificity were 53% and 82%. In addition to the individual tests, they developed hybrid models for combinations of tests. For the CACS/SPECT hybrid, the sensitivity and specificity were 33% and 97%, and for the CTA/SPECT hybrid, the sensitivity and specificity were 67% and 86% [6].

Transthoracic echocardiography is another useful tool in the nephrologist’s imaging armamentarium. Compared with a prevalence of 15–21% in the general population, LVH affects 50–70% of patients during intermediate stages of CKD and up to 90% of patients by the time they reach dialysis [16]. Foley and colleagues found high left ventricular mass index and mass-to-volume ratio were associated with mortality (relative risks of 3.29 and 2.24, respectively) [17].

Cardiac MRI can assess myocardial fibrosis, left ventricular mass, intra-cardiac filling pressures, coronary artery flow, and myocardial perfusion [18,19]. Given the risk of gadolinium and nephrogenic systemic fibrosis (NSF), Rutherford and colleagues investigated whether T1 imaging could identify myocardial fibrosis in patients with ESKD [20]. They found that global, septal, and mid-septal T1 times were significantly higher in the hemodialysis group than in controls. These findings correlated with other non-invasive markers of myocardial fibrosis.

The risks and benefits of iodinated and gadolinium-based contrast agents must be weighed when counseling patients on which image or combination of imaging modalities to choose, though newer data may suggest the risk of NSF is lower with newer gadolinium-based contrast agents [21,22]. Other imaging modalities, such as ultrasound measurements of carotid intima-media thickness, may help further quantify the risk for CVD in the ESKD population, but more studies are needed [23].

4. Mechanisms of Increased Cardiovascular Disease

Figure 1 shows several of the proposed mechanisms for the development of cardiovascular disease in patients undergoing dialysis. As we look at these risk factors, we will discuss the common differences seen between patients on in-center, three-times-weekly hemodialysis (HD), which we will refer to as standard HD; more frequent HD, home hemodialysis (HHD; nocturnal hemodialysis; and peritoneal dialysis, which includes continuous ambulatory peritoneal dialysis (CAPD) and automated peritoneal dialysis (APD).

4.1. Volume Overload and Blood Pressure

Hypertension and volume overload are strongly associated with cardiovascular outcomes [24]. Chronic volume overload, increased sympathetic nervous system activity, renin–angiotensin–aldosterone system, and impaired response of the vascular endothelium to vasodilatory stimuli are some of the proposed mechanisms for hypertension in the ESKD population [25]. Volume overload is arguably the most important of these mechanisms [26].

The association between interdialytic weight gain and hypertension is complex and is a driver of left ventricular hypertrophy (LVH) [27]. As discussed above, cardiac MRI and transthoracic echocardiography are non-invasive means to measure left ventricular volume that can be re-evaluated over time. Therefore, assessing left ventricular size may help determine the risk for subsequent cardiovascular disease and the cardiovascular safety of various modes of dialysis. Zoccali et al. found that increasing left ventricular mass is associated with an increased risk for all-cause mortality in the dialysis population [28]. This is true for both the hemodialysis and the peritoneal dialysis populations [29]. Subsequent studies found that addressing hypertension and improving anemia reduced left ventricular mass and was associated with improved survival [30].

Agarwal et al. randomized 150 patients with hypertensive ESKD on hemodialysis to a conventional ultrafiltration strategy or to a dry weight challenge to investigate whether challenging dry weight improved blood pressure control. The primary outcome was reduced ambulatory systolic blood pressure at 8 weeks. On average, there was a 1 kg reduction in weight at 8 weeks in the challenge group compared to no change in weight in the conventional group. Reducing the dry weight resulted in a decrease in systolic blood pressure of 6.9 mm Hg (95% CI −1.3, −12.4 mm Hg, p = 0.016) at 4 weeks and a 6.6 mm Hg (95% CI −1.0, −12.2 mm Hg, p = 0.021) at eight weeks [31].

Work performed by Flythe and colleagues showed that reducing the ultrafiltration rate led to improved mortality [32]. Compared with standard HD, more frequent dialysis treatments like home hemodialysis or peritoneal dialysis allow for lower ultrafiltration rates and less ultrafiltration volume per treatment. In addition, the shortened interdialytic interval should theoretically decrease the time available to gain weight between sessions, which should further reduce the impact of volume-mediated hypertension and associated left ventricular hypertrophy.

Another potential intervention to improve blood pressure is to increase the number of dialysis sessions. Fagugli et al. found that short daily hemodialysis resulted in lower systolic blood pressure than standard HD (128 mmHg +/− 11.6 versus SHD, 148 +/− 19.2 mmHg; p = 0.01) [33]. Chan et al. studied a cohort of patients converting from conventional HD to nocturnal home dialysis. They found that after 2 months of home hemodialysis, systolic blood pressure fell by an average of 15 mmHg as measured by 24 h blood pressure monitoring, and the number of agents needed to maintain blood pressure was decreased [34]. This was associated with a decrease in left ventricular hypertrophy. Further investigations are necessary to prove that interventions that reduce LVH reduce cardiovascular death.

Patients on peritoneal dialysis typically have mild hypervolemia, and 30–80% of patients with PD are hypertensive, with 30% displaying treatment-resistant hypertension (defined as hypertension despite three anti-hypertensive medications or requiring four anti-hypertensive medications to meet blood pressure targets) [35,36,37,38]. More research is needed to uncover the optimal blood pressure control strategy in the peritoneal dialysis population.

Patients with ESKD on standard HD suffer from both chronic volume overload and from recurrent transient myocardial injury induced by excessive ultrafiltration [39]. Intradialytic hypotension (IDH) is more common on HD than PD and is related to fluid and osmolar shifts during dialysis. Over time, this can lead to organ ischemia and an increased risk of death. Intensive hemodialysis may lessen the risk of IDH per session compared to standard HD [40].

4.2. Mineral Bone Disorder and Vascular Calcification

Despite the great success of statins in the mitigation of cardiovascular disease in the general and CKD populations, effectiveness in patients on dialysis has yet to be demonstrated [41,42]. It is postulated that this is due to a difference in the mechanism of atherosclerosis between patients with CKD and non-CKD [43,44,45].

Patients with ESKD almost universally have elevated serum phosphorus, parathyroid hormone (PTH) levels, and fibroblast growth factor 23 (FGF-23) levels with concomitant derangement in Klotho metabolism [46]. Elevated phosphorous levels have been implicated in transforming vascular smooth muscle cells into osteoblast-like cells, leading to vascular calcification [47]. This alteration in the FGF-23/Klotho axis has been linked to the development of endothelial dysfunction, hypertension, left ventricular hypertrophy, and mortality [48].

Longer dialysis treatments have been shown to improve phosphate clearance, as demonstrated by subsequent analyses of the FHN trial, wherein the implementation of nocturnal hemodialysis led to a reduction in the daily dose of phosphate binders [49].

While longer hemodialysis sessions appear to improve phosphate clearance, the effectiveness of different PD strategies was uncertain. Courivaud et al. investigated phosphate transport kinetics and classified participants into fast, fast-average, slow, and slow-average peritoneal transporters. They then compared the weekly phosphate clearance between these groups and between continuous ambulatory peritoneal dialysis (CAPD) and automated peritoneal dialysis (APD). They found that CAPD had higher weekly phosphate clearance compared to APD [50].Increased phosphate removal helps reduce the stimulation of PTH and FGF-23, which are associated with increased mortality [51,52]. Opting for a dialysis modality that offers longer treatments could theoretically lead to improved CVD outcomes in part due to improved phosphate clearance; however, more research is needed to investigate the correlation between enhanced phosphate clearance and cardiovascular outcomes.

4.3. Anemia

Anemia is an independent risk factor for the development of cardiovascular disease, including LVH and recurrent or new-onset heart failure in the ESKD population [53,54]. This is likely due to chronically increased cardiac output and compensatory arterial intimal thickening [55]. In the ESKD population, the severity of anemia correlates with left ventricular hypertrophy [56,57]. London and colleagues found that decreasing blood pressure and targeting a hemoglobin level of 10–11 mg/dL led to a reduction in left ventricular mass and improved survival [30]. The administration of erythropoietin-stimulating agents (ESAs) has risks. Exposure to erythropoiesis-stimulating agents can also cause downstream effects on cardiovascular health [58].

4.4. Uremia

There is an ever-growing list of uremic toxins [59,60]. Some toxins, such as indoxyl sulfate and p-cresyl sulfate play a role in accelerating vascular calcification and endothelial damage in ESKD [61,62]. Other uremic toxins may contribute to the downregulation of Klotho levels [46]. Low serum Klotho may increase susceptibility to vascular calcification and cardiac fibrosis [63]. The uremic milieu also creates an imbalance between pro-inflammatory and anti-inflammatory cytokines, increases oxidative stress, and leads to malnutrition, all of which have been linked to cardiovascular disease [64,65,66]. In addition to endogenous toxin generation, the innate immune system is activated in response to exposure to various synthetic surfaces during the dialysis treatment [67].

It is unclear whether increasing dialysis frequency leads to improved middle molecule clearance [68]. Measures to preserve residual renal function may lead to improved middle molecule clearance. It is postulated that the increased residual uremic toxins may stimulate kidneys, suggesting a benefit of incremental dialysis in those with RKF, given that RKF has a strong correlation with survival benefit in dialysis studies [63]. Clinicians must, therefore, keep in mind that frequent hemodialysis has been associated with a more rapid loss of residual kidney function (RKF) [69].

5. Cardiovascular Trials

As described above, the benefits of one modality for a particular cardiovascular risk factor may be counterbalanced by a worsening of another risk factor. These complexities may explain the lack of difference between modalities. Comparing cardiovascular outcomes between dialysis modalities is challenging due to the inability to conduct blinded, randomized, controlled trials. Here, we will summarize retrospective studies that have evaluated the cardiovascular risk between dialysis modalities and point out trends that may warrant further investigation. Table 1 provides a summary of these retrospective studies.

The in-center hemodialysis six times per week versus three times per week trial conducted by the Frequent Hemodialysis Network (FHN) trial group found that more minutes of dialysis lowered the composite of death and 12-month change in left ventricular mass assessed by cardiac MRI. In the frequent hemodialysis group, the hazard ratio for death or increased left ventricular mass was 0.61; 95% confidence interval 0.46 to 0.82. In the frequent hemodialysis group, the hazard ratio for death or a decrease in the RAND-36 physical health composite score was 0.70; 95% CI, 0.53 to 0.92. The investigators also found that LV mass reduction was significant in the frequent HD group [70].

Weinhandl et al. found similar rates of hospitalizations in patients with HHD and standard HD though patients with HHD had more infection-related hospitalizations and patients with standard HD required more hospitalizations for cardiovascular complications [71].

Locatelli et al. compared survival and risk of de novo cardiovascular disease between patients with incident HD and peritoneal dialysis (PD). They found no significant difference in 4-year survival or the development of de novo CVD between the groups [72]. Ganesh et al. analyzed CMS and USRDS data from 107,000 patients who initiated dialysis between 1995 and 1997. They compared survival between patients with and without CAD on PD vs. those on HD. They found that patients with diabetes on peritoneal dialysis with pre-existing CAD had a 23% higher relative risk for death compared to those on HD. Surprisingly, even those without diabetes on peritoneal dialysis with pre-existing CAD had a 20% higher relative risk for mortality compared to those on HD. Survival was similar in patients without pre-existing CAD between HD and PD [73]. Stack et al. evaluated survival among patients with heart failure stratified based on diabetes status who either started on PD or HD [74]. They found that patients with heart failure, regardless of diabetes status who started PD were at increased risk for death compared to those who started HD. In contrast, Refaat et al. compared rates of CVD and survival for 157 patients with prevalent ESKD who either initiated HD or PD and found a non-statistically significant trend that PD had a lower risk for CVD [9].

Studies comparing rates of cardiovascular disease between home hemodialysis and peritoneal dialysis are limited. Weinhandl et al. assessed the differences in all-cause mortality, technique failure, and hospitalization between HHD and PD. They found that HHD had lower rates of all-cause mortality, hospitalizations, and technique failure compared to PD [75]. Suri et al. found similar results in their study comparing rates of hospitalizations for CVD, infection, access issues, bleeding, and modality failure between HHD and PD. They also found that HHD was associated with fewer CV hospitalizations than standard HD [76].

Shah et al. published a retrospective cohort study of 68,000 participants evaluating the composite of hospitalizations for acute coronary syndrome, heart failure, and stroke between peritoneal dialysis and home hemodialysis. When unadjusted for age, HHD had more cardiovascular events, heart failure, stroke, CV death, and all-cause death. However, the participants on home hemodialysis were older than those on peritoneal dialysis. Adjusting for age, there were lower rates of stroke, acute coronary syndrome, CV death, and all-cause death in the HHD arm [77].

A review in the American Journal of Kidney Diseases reports similar rates of arrhythmias in HHD, standard HD, and PD; more cerebrovascular and peripheral vascular disease in PD than in HD; more heart failure and fluid overload in standard HD than in PD and more in PD than in HHD; a trend toward more ischemic heart disease in standard HD than HHD; and more in HHD than in PD than in HHD, but without reaching statistical significance. Cardiac-related hospitalizations were similar in the standard HD and PD groups, with lower rates of hospitalization and cardiac death compared to both groups with more intensive HD [78].

Table 1. Studies comparing rates of cardiovascular disease and hospitalization between home modalities and iHD.

Study	Comparison	Key Results
Locatelli et al. (2001) [72]	Comparison of survival and risk of de novo CVD between HD and PD in 3120 patients	No significant difference in 4-year survival or in development of de novo CVD
Ganesh et al. (2003) [73]	CMS and USRDS data for 107,922 patients starting dialysis 1995–1997 comparing survival between patients with and without CAD on PD vs. HD	Patients with DM on PD w/CAD 23% higher RR for death (95% CI; 1.12 to 1.34) compared to HD Patients presenting as non-DM on PD w/CAD 20% higher RR of death (95% CI; 1.10 to 1.32) compared to patients with HD without CAD had similar survival rates on PD (RR 0.99 CI; 0.9 to 1.05)
Stack et al. (2003) [74]	Survival risk for 107,922 patients with ESRD with CHF on PD vs. HD	DM patients relative risk for death (1.30, 95% confidence interval 1.20 to 1.41) Patients presenting as non-DM relative risk for death (1.24, 95% CI; 1.14 to 1.35)
Refaat et al. (2016) [9]	Compared 157 patients with prevalent ESRD on HD vs. PD regarding de novo CVD and survival	Non-statistically significant trend that PD had lower risk of CVD
Weinhandl et al. (2015) [75]	All-cause mortality, hospitalizations, technique failure in 4201 patients with new-start PD vs 4201 patients with new-start HD	HHD had lower risk than PD for each of these outcomes
Suri et al. (2015) [76]	Rates of hospitalizations for CVD, infection, access issues, bleeding, modality failure between 1116 HHD and 2784 PD; compared 1187 HHD with 3173 iHD	HHD associated with fewer hospitalizations than PD; fewer CV hospitalizations for HHD than iHD

Table 1. Studies comparing rates of cardiovascular disease and hospitalization between home modalities and iHD.

Study	Comparison	Key Results
Locatelli et al. (2001) [72]	Comparison of survival and risk of de novo CVD between HD and PD in 3120 patients	No significant difference in 4-year survival or in development of de novo CVD
Ganesh et al. (2003) [73]	CMS and USRDS data for 107,922 patients starting dialysis 1995–1997 comparing survival between patients with and without CAD on PD vs. HD	Patients with DM on PD w/CAD 23% higher RR for death (95% CI; 1.12 to 1.34) compared to HD Patients presenting as non-DM on PD w/CAD 20% higher RR of death (95% CI; 1.10 to 1.32) compared to patients with HD without CAD had similar survival rates on PD (RR 0.99 CI; 0.9 to 1.05)
Stack et al. (2003) [74]	Survival risk for 107,922 patients with ESRD with CHF on PD vs. HD	DM patients relative risk for death (1.30, 95% confidence interval 1.20 to 1.41) Patients presenting as non-DM relative risk for death (1.24, 95% CI; 1.14 to 1.35)
Refaat et al. (2016) [9]	Compared 157 patients with prevalent ESRD on HD vs. PD regarding de novo CVD and survival	Non-statistically significant trend that PD had lower risk of CVD
Weinhandl et al. (2015) [75]	All-cause mortality, hospitalizations, technique failure in 4201 patients with new-start PD vs 4201 patients with new-start HD	HHD had lower risk than PD for each of these outcomes
Suri et al. (2015) [76]	Rates of hospitalizations for CVD, infection, access issues, bleeding, modality failure between 1116 HHD and 2784 PD; compared 1187 HHD with 3173 iHD	HHD associated with fewer hospitalizations than PD; fewer CV hospitalizations for HHD than iHD

6. Conclusions

Comparative studies between dialysis modalities have provided mixed results. While some studies indicate a potential cardiovascular benefit of intensive hemodialysis, more data are needed to determine differences in mortality outcomes and hospitalizations. The promise of potential benefits and the absence of any clear disadvantage do warrant efforts to increase the use of home modalities.

Ultimately, the choice of dialysis modality should be tailored to the individual patient’s preference. This review has looked at cardiovascular risk, which is the leading cause of morbidity and mortality of patients with ESKD. However, other comorbid conditions also affect the risks and benefits of these various dialysis modalities. Patients may benefit from using different modalities over the course of kidney failure management based on the circumstances at the time.

Ongoing research is needed to delineate further the optimal approach for managing cardiovascular risk in patients undergoing dialysis. Clinicians must stay informed about the latest evidence to provide the best possible care for this vulnerable population.