OVERVIEW — Dialysis patients are at extraordinarily high risk for death. In 2016, the annual mortality rate for all United States dialysis patients was 179 deaths/1000 patient-years [1,2].
Cardiac disease is the major cause of death, accounting for approximately 37 percent of all-cause mortality in patients receiving either hemodialysis or peritoneal dialysis [1]. In the United States Renal Data System (USRDS) database, the single, largest, specific cause of death is attributed to arrhythmic mechanisms or sudden cardiac arrest (SCA) [1]. (See "Patient survival and maintenance dialysis".)
The epidemiology, clinical manifestations, and evaluation of SCA and sudden cardiac death (SCD) in the dialysis population are provided in this topic review. Detailed discussions of treatment and prevention of SCA and SCD are presented separately. (See "Overview of sudden cardiac arrest and sudden cardiac death".)
DEFINITION AND EPIDEMIOLOGY — In the general population, the term "sudden cardiac death" (SCD) is commonly used to describe SCA in the setting of heart disease (although some have structurally normal hearts) with cessation of cardiac function whether or not resuscitation or spontaneous reversion occurs.
Previously, the term "sudden cardiac death" has been used even if a patient was successfully resuscitated. Such cases have been referred to as "aborted SCD" or "resuscitated SCD," and patients who experienced such events were said to be "sudden death survivors." Clearer and more rational definitions of SCA and SCD were proposed in 2006 by the American College of Cardiology/American Heart Association/Heart Rhythm Society (ACC/AHA/HRS) [3] (see "Overview of sudden cardiac arrest and sudden cardiac death"):
"[Sudden] cardiac arrest is the sudden cessation of cardiac activity so that the victim becomes unresponsive, with no normal breathing and no signs of circulation. If corrective measures are not taken rapidly, this condition progresses to sudden death. Cardiac arrest should be used to signify an event as described above, that is reversed, usually by CPR and/or defibrillation or cardioversion, or cardiac pacing. Sudden cardiac death should not be used to describe events that are not fatal."
Except where noted, we will use the terms "SCA" and "SCD" as defined in the 2006 ACC/AHA/HRS document. However, most of the available epidemiologic data were published prior to this standardization and are therefore based upon different standards.
In the United States Renal Data System (USRDS) database, the cause of death attributed to arrhythmic mechanisms is noted in the Centers for Medicare and Medicaid Services (CMS) death notification form 2746 by either "cardiac arrest/cause unknown" or arrhythmia. Based upon this definition, arrhythmias may therefore be responsible for [1]:
●Seventy-four percent of all cardiac deaths or 28 percent of all-cause mortality in peritoneal dialysis patients
●Seventy-eight percent of all cardiac deaths or 29 percent of all-cause mortality in hemodialysis patients
However, death attributed to arrhythmias that is based entirely upon the CMS death notification form has obvious limitations and may be inaccurate. Thus, in addition to this definition, the USRDS Cardiovascular Special Studies Center (CVSSC) has used a more complex method in the 2006 USRDS annual data report, which incorporates cause of death in the context of death location (eg, a patient succumbing to myocardial infarction on an ambulance run would be identified as sustaining SCD) and excludes patients with deaths occurring in the setting of sepsis, malignancy, hyperkalemia, and, importantly, withdrawal from dialysis. Using this method, it has been estimated that 29.7 percent of deaths in prevalent dialysis patients are related to SCD [4,5]. The USRDS CVSSC method, however, has not been applied to contemporary data.
We think that the overall best estimate is that SCD is responsible for 29 percent of all-cause mortality in dialysis patients. Similar findings on the relative contribution (22 to 26 percent) of sudden death to all-cause mortality in dialysis patients have been reported in multiple studies including HEMO, 4D trial, the Choices for Healthy Outcomes in Caring for End-Stage Renal Disease (CHOICE) cohort, and the Dialysis Outcomes and Practice Patterns Study (DOPPS) [6-9] and the Evaluation of Cinacalcet HCL Therapy to Lower Cardiovascular Events (EVOLVE) trial [10]. EVOLVE, which enrolled 3883 hemodialysis patients and provided formal adjudication of cause-specific mortality, was the largest randomized, clinical trial ever performed in hemodialysis patients. In EVOLVE, 25 percent of all deaths were adjudicated as "sudden deaths."
The rate due to SCD is approximately 54 deaths/1000 patient-years in combined 2013 to 2015 dialysis cause-specific mortality data, which is essentially identical to that reported in 2011 [1,11]. The CVSSC has also previously estimated that the rate of SCD for period prevalent dialysis patients in 2002 was 6.9 percent per year [4]. It is noteworthy that there has been a slow, steady decline in the overall cardiac mortality rate over time in the US dialysis population, and this finding is reflected in the most current estimate of the SCD rate in dialysis patients. One possible explanation is the increase in the use of "evidence-based therapies" (including beta blockers) in dialysis patients [11]. In the CHOICE cohort, it was 37 SCD events/1000 years (which excluded in-hospital deaths, potentially underestimating the "true" frequency of SCD) [8].
One noteworthy trend is the "disconnect" between the rates of SCD and cardiovascular death (figure 1 and figure 2). Between 2000 and 2013, there was a decline in the rates of all-cause and cardiovascular mortality in the US over approximately 13 years. Historically, approximately 25 to 27 percent of all-cause mortality was attributed to SCD. Since 2009 (figure 2), cardiovascular death rates continued to decline with little change in SCD rates. For this reason, the attributable percentage of SCD to cardiovascular and all-cause mortality has increased.
For whatever reason, the decade-long improvement in all-cause and cardiovascular mortality is no longer reflected in SCD rates in the current decade.
This rate is significantly higher than that observed in the general population. In one population-based study, for example, the overall incidence of out-of-hospital cardiac arrest was 1.89 per 1000 patient-years [12]. The risk for prevalent dialysis patients is roughly comparable with that observed in patients in the general population with a history of an adverse cardiovascular event. (See "Overview of sudden cardiac arrest and sudden cardiac death".)
There is also an enhanced risk of SCD in the first hemodialysis session of the week. Compared with the average risk of SCD, there is a 50 percent increased frequency of SCD on Monday (for patients dialyzing Monday, Wednesday, and Friday) and on Tuesday (for patients having hemodialysis Tuesday, Thursday, and Saturday) [13]. In addition, one study reported a threefold increased risk of sudden death in the 12 hours before the end of the long weekend interval and a 1.7-fold increased risk of SCD in the 12 hours starting with the dialysis procedure following this long interval [14].
The initiation of hemodialysis is a special period of heightened risk for both all-cause mortality and SCD. In 2015, the annualized death rate for US dialysis patients in the first 90 days after dialysis initiation was 312 deaths/1000 patient-years [1]. As shown in the figures (figure 3 and figure 4), the highest rate of sudden death is in the first month after dialysis initiation, with at least 30 percent of all deaths being attributable to SCD.
MECHANISMS — In the general as well as the dialysis population, most SCA events are believed to be due to ventricular arrhythmias; that is, ventricular tachycardia (VT) or ventricular fibrillation (VF). A minority may have been attributed to bradyarrhythmias. (See "Pathophysiology and etiology of sudden cardiac arrest".)
However, two independent groups of investigators using insertable cardiac monitors (sometimes referred to as implantable cardiac monitors or implantable loop recorders) may force a reconsideration of the importance of bradyarrhythmia as a contributing mechanism to SCD [15-17]. However, the study population of one of these studies had a mean time on dialysis of six years, potentially biasing the population away from one with more ventricular tachyarrhythmias. Additionally, we do not know if the patients died from bradyarrhythmias or with bradyarrhythmias [18]. It is our opinion that this older vintage cohort may be very different from newly incident hemodialysis patients, both in terms of structural heart disease and underlying arrhythmic mechanisms potentially leading to SCD.
Although SCA can occur in patients with structurally normal hearts, most patients with SCA have some form of underlying heart disease. A triggering event or condition interacts with the underlying substrate to produce the fatal arrhythmia [19].
Although many triggers have been identified, acute myocardial ischemia is felt to be the most common initiating event in the general population [20]. As will be discussed, the end-stage kidney disease (ESKD) patient has some unique factors that can both alter the underlying substrate as well as trigger ventricular arrhythmic events.
RISK FACTORS AND CAUSES — In the general population, abnormalities of the coronary arteries, myocardium, and cardiac conduction system are the most common underlying causes of the life-threatening arrhythmias that result in SCA. (See "Pathophysiology and etiology of sudden cardiac arrest" and "Overview of screening and diagnosis of heart disease in patients on dialysis" and "Valvular heart disease in patients with end-stage kidney disease".)
Predictors of SCD among hemodialysis patients were evaluated using data from the HEMO study [21]. Among 1745 enrolled hemodialysis patients, 808 died over a median follow-up of 2.5 years, 22 percent of which were due to SCD. Age, diabetes, peripheral vascular disease, ischemic heart disease, a low serum creatinine (reflecting decreased muscle mass and poor nutrition), and an elevated alkaline phosphatase predicted a higher risk for SCD. Traditional cardiovascular risk factors such as smoking and cholesterol did not, but risks conferred by these factors may have been incorporated into the overall increased risk associated with ischemic heart disease. This study did not adjust for dialysis-related risk factors (such as potassium dialysate) nor for other known risk factors such as left ventricular (LV) hypertrophy. The major results of the HEMO study are discussed elsewhere. (See "Prescribing and assessing adequate hemodialysis", section on 'Target Kt/V'.)
The following is a brief overview of the more important abnormalities found in the dialysis population that may underlie their increased incidence of SCA:
●Obstructive coronary artery disease is likely an important contributor to SCA. However, data from the United States Renal Data System Cardiovascular Special Studies Center (USRDS CVSSC) have found an unexplained high mortality due to arrhythmic mechanisms after successful coronary revascularization, suggesting that other factors must be significant [22,23]. In the dialysis population, for example, there was a two-year mortality of 48 and 43 percent after nondrug eluting coronary artery stents and coronary artery bypass surgery (CABG) incorporating internal mammary graft use, respectively. The annual mortality attributed to arrhythmic mechanisms was 8.5 and 7 percent after stenting and CABG, respectively, which is higher than that observed in the general population. This implies that reliance solely upon ameliorating myocardial ischemia by coronary revascularization may be an inadequate clinical strategy for the prevention of SCD in dialysis patients.
●There is a markedly increased incidence of myocardial abnormalities, such as LV hypertrophy (approximately 75 percent of dialysis patients) and alterations in myocardial ultrastructure and function (including endothelial dysfunction, interstitial fibrosis, decreased perfusion reserve, and diminished ischemia tolerance). (See "Overview of screening and diagnosis of heart disease in patients on dialysis" and "Hypertension in dialysis patients".)
●Rapid electrolyte shifts during hemodialysis sessions and the presence of hyperkalemia due to renal failure increase the risk of arrhythmias. This enhanced risk is not surprising given the profound fluid and electrolyte derangements before hemodialysis and the physiologic demands of the hemodialysis session. The nonphysiologic nature of conventional, thrice-weekly hemodialysis sessions may further increase the risk of SCA. This is supported by the observational data previously mentioned showing an increased risk surrounding the first hemodialysis session of the week. (See 'Definition and epidemiology' above.)
●The interplay between the type of renal replacement therapy and dialysis vintage appears to have an impact upon the risk of SCA, with the relative hazard of cardiac arrest in hemodialysis compared with peritoneal dialysis varying with time after initiation of renal replacement therapy. The rate of cardiac arrest is approximately 50 percent higher in hemodialysis patients three months after dialysis initiation, but they are similar at two years. Three years after dialysis initiation, the rate of cardiac arrest is higher in peritoneal dialysis patients (figure 5).
There is a strikingly higher adjusted mortality rate in hemodialysis versus peritoneal dialysis early after initiation of renal replacement therapy (figure 6).
Risk factors related to dialysis prescription — Risk factors related to the dialysis prescription were identified in two large cohorts [9,24]. In one case-control study of 43,200 patients, low-potassium dialysate (<2 mEq/L) was an independent risk factor for SCA [24]. The increased risk associated with low-potassium dialysate was greatest at lower levels of predialysis serum potassium. Increased ultrafiltration volume and low-calcium dialysate were also linked to SCA in this study. Compared with dialysate potassium ≥3 mEq/L, dialysate potassium concentrations ≤1.5 and ≤2 to 2.5 mEq/L were associated with increased risk of SCD (hazard ratios [HRs] 1.39, 95% CI 1.12-1.74 and 1.17, 95% CI 1.01-1.37, respectively). The magnitude of the association of SCD with dialysis potassium ≤1.5 was greater among patients with serum potassium <5 mEq/L. The Dialysis Outcomes and Practice Patterns Study (DOPPS) including 55,183 patients did not demonstrate a significant difference in clinical outcome related to the use of dialysate potassium concentrations of 2 versus 3 mEq/L [25].
In DOPPS, among 37,765 participants, SCD was associated with a treatment time <3.5 hours (HR 1.13, 95% CI 1.00-1.27), ultrafiltration volume >5.7 percent of postdialysis weight (HR 1.15, 95% CI 1.00-1.32), and Kt/V <1.2 (HR 1.06, 95% CI 1.00-1.12) [9].
Dialysate cooling has been advocated to reduce intradialytic hypotension and myocardial stunning; however, the putative benefit on reduction of subsequent cardiac events (including sudden death) remains to be established by adequately powered randomized clinical trials. Similarly, frequent hemodialysis (which reduces LV mass) has not been shown to reduce cardiac events, including sudden death [26,27].
These observational studies suggest that avoiding these prescriptions when possible may reduce the risk of SCA [24,28].
CLINICAL MANIFESTATIONS — In patients with and without end-stage kidney disease (ESKD), most individuals suffering from SCA become unconscious within seconds to minutes as a result of insufficient cerebral blood flow. There are usually no premonitory symptoms. Symptoms, if present, are nonspecific and include chest discomfort, palpitations, shortness of breath, and weakness. (See "Overview of sudden cardiac arrest and sudden cardiac death".)
EVALUATION
Survivor of sudden cardiac arrest — In the general population, the evaluation of the survivor of SCA includes the following (see "Cardiac evaluation of the survivor of sudden cardiac arrest"):
●Identification and treatment of acute reversible causes
●Evaluation for structural heart disease
●In patients without obvious arrhythmic triggers or cardiac structural abnormalities, an evaluation for primary electrical diseases
●Neurologic and psychologic assessment
●In selected patients with a suspected or confirmed heritable syndrome, evaluation of family members
The evaluation in the dialysis patient who survives SCA is generally the same as that in the patient without renal failure. However, close attention should be paid to the presence of myocardial dysfunction and/or ischemia (since they are so common), the possibility of improper medication dosing in the patient with renal failure, and the circumstances associated with the event, particularly if it occurred during and/or surrounding a hemodialysis session.
As examples:
●Since both myocardial ischemia and/or dysfunction are relatively common in the dialysis patient, their presence, either alone or in combination, may markedly enhance the risk of SCA. However, retrospective analysis has shown that 71 percent of dialysis patients who experienced SCD had either normal or only mild to moderate left ventricular (LV) dysfunction, suggesting that other factors may underlie SCA in many dialysis patients [29]. (See 'Identification of the high-risk dialysis patient' below.)
●Increased electrical instability resulting in SCA may have been due to fluid shifts, autonomic imbalance/increased sympathetic activity (including sleep apnea), acid/base disturbances, and electrolyte abnormalities [24,30-36]. An increased risk of cardiac arrest may be particularly associated with a low-potassium concentration in the dialysate. (See "Acute complications during hemodialysis".)
●Improper dosing of certain medications may predispose the patient with renal failure to brady/tachyarrhythmias and/or proarrhythmic effects, thereby causing SCA. (See "Overview of sudden cardiac arrest and sudden cardiac death".)
One study reported that low predialysis serum potassium (<4.3 mEq/L) was associated with an increased mortality hazard in hemodialysis patients receiving digoxin, suggesting that even low "normal" potassium levels may enhance the proarrhythmic risk of digoxin [37]. In this study, the mortality risk associated with digoxin was attenuated in dialysis patients with serum potassium >4.6 mEq/L. Digoxin should be used with extreme caution in such patients.
Identification of the high-risk dialysis patient — In the general population, it is known that reduced LV function is the strongest predictor of SCA. Clinically, the presence of heart failure also identifies patients who are at high risk of SCA, perhaps by additional arrhythmogenic factors, such as activation of the neurohormonal cascade and electrolyte shifts.
Identification of the high-risk patient is most useful if therapy can provide significant benefits. In the nondialysis population, most primary prevention implantable cardioverter-defibrillator (ICD) trials have shown significant improvement in survival in the following high-risk groups who received ICDs:
●The earlier primary prevention trials enrolled nondialysis patients with decreased systolic function, prior myocardial infarction, nonsustained ventricular tachycardia (VT), and positive electrophysiology (EP) study for induction of VT.
●In later trials, enrollment did not require nonsustained VT or positive EP studies. Instead, enrollment was based upon decreased systolic function and/or heart failure. In addition, nonischemic cardiomyopathy patients were included.
Thus, in the nondialysis population, primary prevention ICD trials have principally demonstrated survival improvement in groups with decreased systolic function who receive ICDs. (See "Primary prevention of sudden cardiac death in patients with cardiomyopathy and heart failure with reduced LVEF".)
However, it is unclear if dialysis patients with decreased systolic function also receive survival benefits with ICD. There are no prospective studies that have examined this issue, although retrospective studies suggest some benefit.
In addition, despite the very high annual mortality from SCD, the dialysis patient with reduced LV function is not typical of the general dialysis population. Fifteen percent or fewer dialysis patients have significantly decreased LV function [38-40]. In the largest study to date, among 1254 consecutive patients starting hemodialysis in Japan, 5 percent had LV ejection fractions of less than 40 percent [40].
Thus, other unique factors/circumstances may contribute to the general increased risk of SCA in end-stage kidney disease (ESKD) patients in dialysis. Possibilities include the following:
●There is additional substrate (myocardial) modification, such as interstitial fibrosis due to chronic uremia, microvascular disease, or endothelial dysfunction; increased calcium/phosphate deposition; and significant LV hypertrophy due to hypertension and/or anemia [41-46].
●Increased electrical instability may be present due to fluid shifts, autonomic imbalance/increased sympathetic activity (including sleep apnea), inflammatory state, acid/base disturbances, and/or electrolyte abnormalities [24,30-36,47].
It is likely that the increased risk of ventricular arrhythmias in ESKD patients is due to a combination of these many interacting factors (table 1).
In addition, while an ejection fraction of <35 to 40 percent is considered the major risk factor for SCA in nondialysis patients, it is likely that more mild LV dysfunction imparts a greater risk of cardiovascular events in the dialysis population, regardless of the etiology and despite optimal management. As an example, one study found that the best predictor of SCD risk in peritoneal dialysis patients was an LV ejection fraction of ≤48 percent [48].
Risk-stratification studies within the dialysis population are required to better identify those patients at highest risk of SCA and in whom prophylactic interventions may be beneficial.
We agree with recommendations concerning evaluation as noted in the 2005 National Kidney Foundation Dialysis Outcome Quality Initiative (KDOQI) Clinical Practice Guidelines for Cardiovascular Disease in Dialysis Patients [49]. We recommend that, at initiation of dialysis, all patients should undergo baseline echocardiography and electrocardiography. Echocardiography should be performed after dry weight is attained (which usually occurs after one to three months) and should be repeated routinely at three-year intervals. Additional evaluation of LV systolic function should be performed following a change in cardiac status or after an intercurrent cardiac event.
Findings of a decreased ejection fraction (<40 percent) or significant regional wall motion abnormalities with or without ischemic symptoms require an evaluation for the presence of coronary artery disease, which may underlie myocardial dysfunction in many dialysis patients.
If these surveillance guidelines are met, all dialysis patients with reduced LV function should be identified by echocardiography. Once identified, the main question is whether patients with significantly reduced LV function should be treated with aggressive primary prevention measures, including placement of an ICD. (See "Implantable cardioverter-defibrillators: Overview of indications, components, and functions".)
The use of biomarkers for identification of dialysis patients at high risk for SCD deserves further study. Cardiac troponin T is a strong independent predictor of all-cause mortality [50], and high-sensitivity C-reactive protein (CRP) is associated with the risk of SCD [8].
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Dialysis".)
SUMMARY AND RECOMMENDATIONS
●Cardiac disease is the major cause of death among dialysis patients. In the United States Renal Data System (USRDS) database, the single, largest, specific cause of death is attributed to arrhythmic mechanisms or sudden cardiac death (SCD). The overall best estimate is that SCD is responsible for approximately 29 percent of all-cause mortality in dialysis patients. (See 'Definition and epidemiology' above.)
●Many causes of and risk factors for sudden cardiac arrest (SCA) are shared among patients with and without end-stage kidney disease (ESKD), although dialysis patients frequently have a relatively increased incidence of abnormalities of the coronary arteries, myocardium, and cardiac conduction system. There are also issues unique to dialysis patients. (See 'Evaluation' above.)
●The evaluation in the dialysis patient who survives SCA is generally the same as that in the patient without renal failure. However, close attention should be paid to the presence of myocardial dysfunction and/or ischemia, the possibility of improper medication dosing in the patient with renal failure, and the circumstances associated with the event, particularly if it occurred during and/or surrounding a hemodialysis session. (See "Cardiac evaluation of the survivor of sudden cardiac arrest" and 'Evaluation' above.)
●Rapid electrolyte shifts during hemodialysis sessions increase the risk of arrhythmias. Variables in the hemodialysis prescription such as the use of low-potassium dialysate, low-calcium dialysate, or large ultrafiltration volumes may be modifiable risk factors for SCA. (See 'Risk factors and causes' above.)
●To help identify the dialysis patient at increased risk of SCA, we recommend that, at initiation of dialysis, all patients should undergo baseline echocardiography and electrocardiography, with echocardiography repeated routinely at three-year intervals. Additional evaluation of left ventricular (LV) systolic function should be performed following a change in cardiac status or after an intercurrent cardiac event. Findings of a decreased ejection fraction (<40 percent) or significant regional wall motion abnormalities (with or without ischemic symptoms) require an evaluation for the presence of coronary artery disease.
