Mortality in dialysis patients is extremely high, with an annual death rate of ∼23%. Sudden cardiac death (SCD) is the single largest cause of death in dialysis patients accounting for ∼60% of all cardiac deaths and 25% of all-cause mortality. Interventions aiming at reducing cardiovascular mortality, especially SCD, in dialysis patients are therefore extremely important and clinically highly relevant. The purpose of this review is to give an outline of the epidemiology of SCD in dialysis patients and to provide a comprehensive overview of several interventional strategies (medical therapies, changing dialysis modality, and revascularization). Furthermore, it will discuss the current knowledge regarding the value of preventive implantable cardioverter defibrillator implantation and address future implications of the interventional strategies mentioned.
Currently, ∼350 000 people are treated with dialysis in the USA. It is estimated that this number will increase to >500 000 by the year 2020.1 In Europe, the number of dialysis patients is also increasing annually.2,3 The mortality in dialysis patients is very high with a mean death rate of ∼23% per year. This rate is mainly influenced by age, with a yearly mortality of 4% in patients younger than 20 years rising to 35% in patients older than 65 years.1 In dialysis patients, sudden cardiac death (SCD) is the major cause of death, accounting for ∼25% of all-cause mortality.4 The risk of SCD in dialysis patients is also strongly associated with patient age, with a 3-year probability of developing cardiac arrest of ∼12% in patients younger than 20 years rising to ∼34% in patients older than 75 years.4 Because the increase in number of patients treated with dialysis for end-stage renal disease (ESRD) can mainly be attributed to older patients, the total burden of SCD is rapidly increasing.1
Given the high event rate of SCD in dialysis patients, identification of risk factors and finding preventive interventions seem highly desirable. The objective of this review is (i) to provide an outline of the epidemiology of SCD in this patient group and furthermore (ii) to provide a comprehensive overview regarding several preventive interventions aiming at reducing cardiovascular mortality. We will especially focus on SCD, since this is the single largest cause of death in this highly vulnerable patient group.
Epidemiology of sudden cardiac death in dialysis patients
Approximately 60% of all cardiac deaths and 25% of all-cause mortality in patients on dialysis are due to SCD.4 In several large survival trials with dialysis patients, such as the Hemodialysis (HEMO) study and the German Diabetes and Dialysis (4D) study, similar incidences for SCD were found.5,6 These studies will be discussed in more detail later on. This high incidence may even be an underestimation as was demonstrated by Bleyer et al.7 They reviewed death notification forms from 1995 to 2003, obtained from five US dialysis centres. According to an accepted definition of SCD, they found that in total 88 of 228 deaths (39%) could be classified as sudden. The patient's primary nephrologist had only classified 59 of these 88 deaths as sudden, i.e. 26% of all deaths. The remaining 29 sudden deaths were initially not classified as sudden, but as acute myocardial infarction (6 patients), atherosclerotic heart disease (4 patients), cardiomyopathy (3 patients), pulmonary oedema (2 patients), other (2 patients), valvular heart disease (1 patient), and unknown (11 patients).
Factors relating to sudden cardiac death
The mechanisms that underlie SCD in dialysis patients are complex and many factors are involved. In addition to the traditional risk factors associated with SCD in the general population, such as ischaemic heart disease, there are several more specific factors and circumstances in dialysis patients which may contribute to the risk of SCD. These factors include: left ventricular hypertrophy (LVH), rapid electrolyte and fluid shifts in haemodialysis (HD) patients, and abnormalities in myocardial ultrastructure and function, including endothelial dysfunction, interstitial fibrosis and sympathetic overactivity.8–12
Peritoneal dialysis vs. haemodialysis
A recent study compared the survival of Dutch HD and peritoneal dialysis (PD) patients. A survival advantage for PD compared with HD was documented which decreased over time, with age and in the presence of diabetes mellitus as primary disease.13 The initial advantage of PD compared with HD may be explained by the fact that PD patients have lower co-morbidity at initiation of dialysis therapy14 and a better preservation of the remaining kidney function.15 With regard to SCD, there is no apparent difference between PD and HD.4
It has been documented that dialysis patients exhibit sustained activation of the sympathetic nervous system (SNS) and that the diseased kidneys themselves are the trigger of this overactivity.16,17 This sympathetic nerve overactivity is associated with mortality and worse cardiovascular outcomes,18 and it is likely that SCD in dialysis patients is associated with the overactivation of the SNS.19
β-Blocker therapy interferes with the deleterious actions of the SNS on cardiac endpoints20 and is a well-established and evidence-based intervention in hypertension21 and after myocardial infarcation.22 In the general population, much of the benefit conferred by β-blocker therapy can be attributed to the prevention of SCD.19,23
A large observational study by Foley et al.24 indicates that β-blocker therapy has a robust association with survival in dialysis patients. Despite this observation, only a few trials regarding β-blocker therapy in dialysis patients have been conducted so far. In a recent placebo-controlled trial in HD patients with dilated cardiomyopathy, Cice et al.25 showed that carvedilol gave a significant reduction (52 vs. 73%) in mortality. With regard to SCD in dialysis patients, it has been documented that in HD patients, β-blocker use at the time of a cardiac arrest is associated with higher survival.26
Nevertheless, although a potential benefit of β-blocker therapy has been indicated, it has been documented that β-blocker therapy is used in <30% of patients on HD.27 Furgeson and Chonchol19 suggest four major reasons for this low utilization: (i) therapeutic nihilism for these chronically ill patients, (ii) the unconventional epidemiology of cardiovascular disease in this population, (iii) the paucity of efficacy data in patients with serum creatinine >177 µmol/mL (2 mg/dL), and (iv) the potential for higher rates of adverse effects, including hypotension, hyperkalaemia and glycaemic abnormalities. In their review, they conclude that risks of dangerous side effects appear to be rare and manageable and that long-term clinical trials are desperately needed to evaluate the safety and efficacy of β-blockers in chronic dialysis patients.
When defining the role of β-blockers as a preventive intervention for SCD, it should also be considered that in several high-risk patient groups, prophylactic implantable cardioverter defibrillator (ICD) implantation has proven to be superior compared with β-blocker therapy.28–32
Statin therapy safely reduces the 5-year incidence of major coronary events, coronary revascularization, and stroke in a wide range of individuals by reducing LDL cholesterol.33 However, until recently, with regard to dialysis patients, only limited prospective data on the effectiveness of statin therapy were available. In 2005, the results of the German 4D study were published. The 4D study was the first prospective trial to evaluate the effectiveness of statins in 1255 patients receiving chronic HD with type II diabetes mellitus. This study showed that atorvastatin, despite its ability to lower LDL cholesterol, had no beneficial effect on cardiovascular death, non-fatal myocardial infarction, and stroke. It was, therefore, speculated by the investigators that the pathogenesis of vascular events in patients with diabetes mellitus who are receiving HD may, at least in part, be different from that in patients without ESRD.6
A recently published meta-analysis by Strippoli et al.34 demonstrated that statins significantly reduce lipid concentrations and cardiovascular endpoints in patients with chronic kidney disease. A significant reduction of ∼20% in the risk of cardiovascular mortality was documented, and in addition to this, compared with placebo, statin therapy also significantly decreased the risk of non-fatal cardiovascular events by ∼20%. However, this meta-analysis also showed no benefit on all-cause mortality.
The specific effect of statins on SCD remains doubtful as the mechanisms involved may be less amenable to cholesterol lowering. This is confirmed by the fact that statin therapy had no effect on SCD in the 4D study6 and in other statin trials in non-dialysis patient groups, such as in heart failure patients in the CORONA study.35
A relationship between anaemia and the development of LVH in ESRD patients has been demonstrated.36 Left ventricular hypertrophy is associated with a lower survival in ESRD patients37 and especially the worsening of LVH, independent of left ventricular (LV) mass, is a strong predictor of SCD in this patient group.38 Considering the relationship between LVH and anaemia and the strikingly high prevalence of LVH in dialysis patients,39 a strong beneficial effect of restoring anaemia with erythropoietin could be anticipated. However, the optimal level of haemoglobin (Hb) correction with erythropoietin remains uncertain. Although several studies showed that complete correction of anaemia improved several cardiovascular prognosis parameters in dialysis patients,40 at this moment there is insufficient published literature to generalize the risks or benefits of Hb levels >7.2 mmol/L (120 g/L).41,42
It should also be taken in account that correction of anaemia with erythropoietin to higher Hb-levels may increase blood pressure, the risk of vascular access thrombosis, and may lead to an increased number of adverse cardiovascular events,42 thus counterbalancing the potential positive effects. Considering the above, more information about the ideal Hb target level is needed and therefore only partial correction of anaemia by erythropoietin is nowadays recommended.43
Changing dialysis modality
Dialysis therapy itself is probably an important risk factor for SCD. For instance, it has been observed that SCD is temporally related to the HD procedure7 and several treatment-related factors, such as dialysis dose and the size of molecules that are removed, are implicated in mortality and morbidity among patients undergoing HD.44,45 The fact that dialysis should be considered pro-arrhythmic is supported to a certain extent by the significant decline of SCD rate after renal transplantation.4 Considering these observations, interventions aiming at altering the dialysis therapy may prove to be useful in reducing all-cause and cardiovascular mortality. In the last decade, many studies have been conducted in order to evaluate the effect of different alterations in dialysis therapy on mortality and cardiovascular morbidity. With regard to these alterations in dialysis therapy, changing dialysis dose, haemodiafiltration, and increasing dialysis frequency will be discussed below.
Changing the dialysis dose
Several observational studies reported significant mortality reductions in patients treated with a higher dialysis dose compared with a control group receiving standard dialysis dose.46,47 The HEMO study was the first randomized controlled trial designed to evaluate the effect of the dialysis dose on all-cause mortality and cardiovascular mortality. In this study, 1846 patients were randomized to receive either a standard dose (urea eKt/V 1.05) or high dose (urea eKt/V 1.45) and either a low-flux or high-flux dialyzer in a 2 × 2 factorial design with equal allocation. The results of the low- vs. high-flux dialyzer will be discussed in the next paragraph. No significant difference was observed between the standard dose vs. the high-dose group in all-cause or cardiovascular mortality.5 More recent observational data consisting of large study cohorts (n > 4000), however, indicate that dialysis dose and session length are associated with mortality risk.48,49 Therefore, the debate regarding the optimal dialysis dose and session length remains ongoing.
Haemodiafiltration and high-flux HD, which is considered a form of haemodiafiltration, are treatment modalities which not only remove small molecules (<5 kDa) but are also capable of obtaining considerable clearance of middle molecular weight substances (5–50 kDa).50 Middle molecular weight molecules include markers of inflammation, complement factor D, and other molecules that might be relevant in the pathogenesis of cardiovascular morbidity and mortality.51
Although large observational data indicate a significant reduction in mortality when comparing haemodiafiltration with low-flux HD,52 until now this has not been confirmed by randomized controlled trials. The earlier mentioned HEMO study, which also compared low-flux HD with high-flux HD, showed that in the high-flux group, there was an 8% reduction of all-cause mortality. This was, however, not statistically significant. Also a 20% reduction in cardiac death was documented, which was not statistically significant after adjustment for the multiple comparisons performed. It is therefore concluded that the overall pattern is consistent with a possible benefit for high-flux dialysis, which was too small to be detected given the power of the study.53 At this moment, several trials are ongoing which will further evaluate the value of haemodiafiltration with mortality and cardiovascular morbidity and mortality as primary or secondary outcome.54
Increasing dialysis frequency
Already in 1969, it was documented that daily dialysis improved several clinical outcome parameters in a selected patient group.55 However, at this moment, the majority of the HD patients depend on two to three dialysis sessions a week. Considering the potential benefit of a higher frequency, several possibilities have been proposed to increase the frequency of dialysis therapy. These include frequent nocturnal HD and short daily HD.
A recently published trial that compared the effects of frequent nocturnal HD vs. conventional HD on change in LV mass and health-related quality of life over 6 months showed promising results. Fifty-two patients were randomized in a 1:1 fashion to nocturnal HD or conventional HD. There was a regression in LV mass in the treatment group, whereas there was an increase in LV mass in the control group. This difference was statistically significant. Also a statistically significant improvement in some measures of mineral metabolism and selected measures of quality of life in the treatment group was documented.56
Promising results have also been reported regarding short daily HD. A recently published observational study showed that survival in 416 patients (52 ± 15 years) on short daily HD (frequency 5.8 ± 0.5 times weekly, duration 136 ± 35 min) was two to three times better than that of matched three times weekly HD patients reported by the United States Renal Data System (USRDS). The 5-year cumulative survival was 68 ± 4.1%.57
Although the results for both nocturnal HD and short daily HD are indeed promising, prospective data on significant changes in all-cause and/or cardiovascular mortality are lacking at this moment. Adequately powered trials with hard endpoints are therefore warranted to define the role of these treatment modalities.
In the general population, coronary artery disease (CAD) is present in ∼80% of the patients suffering from SCD.58 The HEMO investigators pointed out that CAD was the largest contributor to SCD also in dialysis patients.59,60 The prevalence of CAD in dialysis patients is ∼40%.61 However, despite the high prevalence of CAD in this highly vulnerable patient group, the utilization of invasive cardiac procedures in dialysis patients is clearly underused. For instance, the use of both diagnostic angiography and revascularization after MI is significantly lower in dialysis patients compared with patients with normal renal function.62 The underutilization of these therapies may be a reflection of ‘therapeutic nihilism’. Another explanation is the lack of evidence of benefit in this patient group due to the exclusion of dialysis patients in most clinical trials.
In a recent observational study, Hemmelgarn et al.63 compared patients receiving CABG, PCI, or no revascularization after coronary angiography and found that in dialysis patients, survival was significantly higher after CABG or PCI compared with no revascularization. However, in a population of dialysis patients with ischaemic heart disease, treated with optimal surgical coronary revascularization, Herzog et al.64 found that the probability of all-cause mortality and arrhythmically mediated death was not lower than that reported for the entire US dialysis population. They documented that, in 2002, the 2-year probability of all-cause death was 40% and the probability of SCD was 14% in prevalent US dialysis patients. In US dialysis patients receiving CABG, they found a 2-year all-cause mortality of 43% and a mortality attributed to arrhythmic mechanisms of 14%. These data do not suggest that coronary revascularization in dialysis patients is not efficacious, rather it is proposed that coronary revascularization may be a particularly incomplete therapy for cardiac disease in ESRD patients and that additional treatment strategies targeting the ‘non-ischaemic’ contributors to SCD may be necessary, as a large untreated hazard of arrhythmic death may remain despite revascularization.
Implantable cardioverter defibrillator therapy
Several trials in multiple patient groups have shown that ICD therapy is superior to medical therapy in primary and secondary prevention for all-cause mortality, almost exclusively by reducing SCD.28–32 Almost all ICD trials, however, excluded dialysis patients or did not publish subgroup analyses on this group of patients. Thus, only very limited literature is available with regard to ICD therapy in dialysis patients.
The few studies that evaluated the benefit of ICD therapy in dialysis patients, however, seem to indicate a possible benefit: Herzog et al.65 documented that after aborted cardiac arrest, ICD therapy was associated with a 42% reduction in death risk in dialysis patients, Hreybe et al.66 found that renal insufficiency is a strong predictor for appropriate ICD shocks and that the incidence of appropriate shocks in HD patients at 1 year is significantly higher when compared with non-dialysis patients and Robin et al.67 concluded that ESRD is the greatest predictor for ICD therapies.
Despite these observations, ICD therapy in dialysis patients is probably underutilized. Only 8% of the dialysis patients who survived a cardiac arrest episode receive prophylactic ICD implantation in the USA.65 This underutilization may be due to several concerns regarding prophylactic ICD implantation in this patient group. An important concern regards the effectiveness of prophylactic ICD implantation. In previous ICD studies, survival was reduced in the ESRD population compared with patients with a normal renal function.67,68 One explanation is the possibility that, in dialysis patients, ventricular arrhythmias are intermittently refractory to ICD therapies because of metabolic derangements.68 However, SCD rates in dialysis patients with an ICD have never been assessed. Another explanation is that co-morbidities in ESRD patients meeting current implantation indications may reduce the survival benefit of ICD placements in this population. Also an important concern regarding ICD therapy is that patients with ESRD may have higher rates of cardiac device-related complications, potentially offsetting some benefits of prophylactic ICD therapy. Dasgupta et al.69 found that dialysis patients indeed had higher complication rates from cardiac device implantation, such as infection and bleeding, but these complications did not result in death.
Considering these possible benefits and risks, further studies are required to assess the potential impact of ICD therapy in dialysis patients. We recently started such a study, the Implantable Cardioverter Defibrillator in Dialysis patients (ICD2) study, which will prospectively evaluate the impact of ICD therapy on SCD in dialysis patients. In addition to that, it will also focus on the feasibility of a larger trial which could evaluate the effect on all-cause mortality.70
The number of prevalent dialysis patients is annually increasing worldwide. Mortality rates among these patients are extremely high, with SCD being the largest contributor to death in this patient group. Interventions aiming at reducing SCD in dialysis patients are therefore very desirable. In the past decade, several interventional methods have been investigated. These methods included medical interventions (β-blockers, statins, and erythropoietin), altering dialysis therapy and revascularization.
Although some observational studies indicate potential survival benefit of the mentioned interventional methods, at this moment prospective data to confirm this benefit are lacking and more adequately powered prospective trials are warranted. With specific regard to their benefit in reducing SCD, the role of these interventional methods remains doubtful. For instance, important trials such as the 4D study did not find a significant reduction in SCD in the treatment arm of their study cohort and it has been observed that after optimal revascularization therapy in dialysis patients with ischaemic heart disease, SCD rates are similar when compared with the general dialysis population.
Prophylactic ICD implantation has become an important preventive intervention for SCD over the past decades. Because prophylactic ICD implantation is effective for prevention of SCD in several high-risk patient groups, it is hypothesized that this may also be valuable in preventing SCD in dialysis patients. This potential benefit of prophylactic ICD implantation was confirmed by several observational studies. However, this has not yet been evaluated prospectively in dialysis patients and therefore a prospective trial is warranted to define the value of prophylactic ICD implantation in this highly vulnerable patient group. The ICD2 study will be the first prospective randomized study that will evaluate the effects of prophylactic ICD implantation in dialysis patients.
Conflict of interest: none declared.