| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||
COMMENTARIES |
Section on Nephrology Wake Forest University School of Medicine Winston-Salem, NC, USA
e-mail: ableyer{at}wfubmc.edu
Cardiovascular mortality is the leading cause of death in the dialysis population (1). Yet most practicing nephrologists can recall only a few patients in their practice who have died of acute myocardial infarction. Instead, we are only too familiar with the call received during the night or early in the morning that one of our patients has passed away suddenly at home or, even worse, in the dialysis center. These instances of sudden death are frequently predictable based on risk factors such as age and underlying coronary disease (2), but they are unpredictable in terms of their timing and the 1 patient out of many who will be affected. One might suspect that a better recognition of temporal risk factors that trigger sudden death could help in its avoidance.
Sudden cardiac death is the result of a stressful stimulus being applied to a heart with underlying structural damage (3). The young man shoveling snow experiences a significant stress but is unlikely to experience a cardiac event. Similarly, his grandfather with underlying coronary disease is unlikely to experience an event while watching the snow being shoveled. It is only when the grandfather (with underlying cardiac disease) begins to shovel the snow (stressful stimulus) that a cardiac event is likely to occur.
Almost every dialysis patient has underlying cardiac disease, placing him at risk for sudden death. Hypertension, present in greater than 90% of individuals on dialysis, has been found to be a risk factor for sudden death (4). Myocardial hypertrophy, coronary artery disease, and congestive heart failure are more important risk factors for sudden death and are also extremely common in the dialysis population (5). While structural disease is important, baseline problems with cardiac electrical activity are also very common in dialysis patients. The QT interval — a measure of re-polarization and predictor of sudden death — is often prolonged in dialysis patients (6). The QT dispersion — a measure of the differences in time of the QT interval in different leads of the electrocardiogram — is also increased in dialysis patients and has been found to be a risk factor for sudden death (7). As dialysis vintage increases, cardiac abnormalities tend to worsen. Continued hypertension leads to increased hypertrophy, and the prevalence of cardiomyopathy increases with time on dialysis (8). While nephrologists are busily adjusting dialysis dose, cardiac disease is less well attended. In the end, worsening cardiac disease, and not underdialysis, usually leads to premature death.
It is easy to understand why dialysis patients are especially prone to sudden death based on just the underlying prevalence of heart disease, but that is only half the story, for end-stage kidney disease also results in stressors to which the general population with heart disease is not exposed. For the hemodialysis (HD) patient, incredible fluid fluxes occurring with dialysis result in repeated bouts of predialysis volume overload and intra- or postdialysis hypotension, each being a potent stressor. Similarly, potassium and other electrolytes undergo large fluxes with HD and hyperkalemia or hypokalemia precipitate cardiac arrest (9). In our study of cardiac arrest in HD patients, we found that Monday–Wednesday–Friday HD patients were more likely to suffer cardiac arrest on Monday, and Tuesday–Thursday–Saturday HD patients were more likely to have cardiac deaths on Tuesday (10). When individual death events were studied, cardiac arrest occurred more frequently during the 8 hours beginning at the start of dialysis and also during the extra weekend day (Sunday for Monday–Wednesday–Friday patients and Monday for Tuesday–Thursday–Saturday patients) (5).
Peritoneal dialysis (PD) patients would seem to have far fewer problems with the volume and electrolyte fluxes that occur in HD patients. In a retrospective study looking at the day of death for dialysis patients, death events were more regularly dispersed in the PD population compared to the HD population (10). One might expect that dialysis deaths would not be temporally related to the PD regimen, as PD is a much smoother and more gradual process. Indeed, QT dispersion and ventricular ectopy have been found to increase during the HD procedure (11,12), and QT dispersion has been found to be greater in HD than in PD patients (13).
The recent study by Chow et al. (14), in this issue of Peritoneal Dialysis International, is the first to systematically examine PD patients that suffered cardiac arrest. The authors used a standard definition of sudden death: unexpected, non-traumatic death occurring within 1 hour of symptoms and without any previous condition that would seem fatal. Specifically defining sudden death is important because, in most large end-stage renal disease databases, sudden death has been arbitrarily defined by nephrologists and can include deaths of almost any cause. This frequently inaccurate definition of sudden death unfortunately necessitates the abandonment of large databases and requires the study of individual centers, with smaller numbers of patients. The fact that in most centers there are few PD patients increases the task of collecting a sample of sudden deaths. Thus, while it may seem that only a small number of sudden deaths were identified in the study by Chow, this was a labor-intensive study spanning 4 years in which 175 deaths were reviewed to obtain 24 instances of certain sudden death. The fact that this study was carried out at one center allows for a more standard definition of sudden death. Unfortunately, it also decreases the applicability of the results, especially where the treatment of anemia is different than in this Chinese unit.
Diabetes, male gender, and tobacco usage were all associated with sudden death, as has been noted in previous studies. Coronary artery disease was noted to be more frequent in the sudden death population (38% vs 23%, p = 0.27) but may not have been significant due to the small number of patients. Unfortunately, data were not presented regarding congestive heart failure or ejection fraction. The QT interval was prolonged in patients suffering cardiac arrest compared to controls (456 ± 23 ms vs 436 ± 48 ms, p = 0.06).
Of significant interest, the authors noted blood transfusion to be a very important risk factor for sudden death, with 17% of individuals having a blood transfusion during the month prior to dialysis compared to only 2% in individuals without sudden death. Of note, transfusion was used instead of erythropoietin to maintain hematocrit in individuals that were not on the transplant list, and the mean hemoglobin of patients was only approximately 9 g/dL. Why was there such a significant difference? It is possible that low hemoglobin was a marker of general poor health and that these patients were predisposed to sudden death from acute illness, or the transfusions themselves may have led to inflammatory events proceeding to sudden death. Further studies in this population could be helpful in evaluating this finding.
Kt/V was also found to be lower in the sudden death population than in the non-sudden death population. Again, this could be related to selection bias, with sicker patients having lower Kt/V. However, future studies looking at sudden death in the PD population must further analyze this finding.
Prevention of sudden cardiac death is in many ways similar to the treatment of underlying cardiac disease. The use of beta-blockers and angiotensin-converting enzyme (ACE) inhibitors has been shown to increase survival in patients with cardiomyopathy, and they should be used in dialysis patients with these conditions. In a recent trial in HD patients with a left ventricular ejection fraction <35% and systolic blood pressure >90 mmHg, carvedilol was found to improve survival (15). Unfortunately, nephrologists have been shown to not place enough emphasis on the treatment of cardiac disease. In the study by Chow, only 63% of patients that experienced sudden death received beta-blockers, and only 25% received ACE inhibitors, although we do not know if these patients had a baseline medical condition that wound have prompted therapy with these agents. In a single-center study we performed (5), only 28% of HD patients with cardiomyopathy who suffered a sudden death were receiving a beta-blocker. In studies of patients post myocardial infarction, the usage of ACE inhibitors and beta-blockers has also been poor (16). Increasing the usage of these medications and developing systems to make sure the appropriate patients are receiving these therapies is crucial to improving survival. A structured approach to dealing with cardiac factors — similar to the structured approach we use to evaluate dialysis dosage, anemia, and calcium phosphate metabolism — could have a marked impact on survival. Periodic electrocardiograms to identify prolongation of the QT interval and echocardiograms to determine the presence of cardiomyopathy may help direct treatment.
For the patient with cardiomyopathy starting dialysis, it would seem that PD would be much preferred to HD as a means to prevent sudden death. The temporal association of cardiac arrest with HD is likely related to large volume shifts, frequent hypotension, and large electrolyte shifts, problems that are not present in PD. While the United States Renal Data System showed similar rates of cardiac arrest in PD and HD patients (10), these deaths were so frequently misclassified that it is difficult to tell what the true prevalence was.
Another reason that PD may be preferable to HD in the prevention of sudden death is that the placement of implantable cardiac defibrillators is a problem for HD patients. Placement of these devices is usually through a percutaneous intravascular route. The intravascular lead placement can lead to a decrease in the number of sites where venous access can be placed. In addition, if these leads become infected (from catheter-related bacteremia) (17), removal of the defibrillator and its leads can be a dangerous and morbid procedure. The use of implantable defibrillators in dialysis patients is a matter of some controversy. These defibrillators likely lead to an increase in survival in dialysis patients (18), although not as long as in patients that are not on dialysis (19); there have been no randomized trials specifically addressing this issue. In the relatively healthy PD patient, I believe they should be considered as prophylaxis against cardiac arrest if the patient suffers from cardiomyopathy.
Better characterization of the death event is a tremendous opportunity for research in the study of mortality in the dialysis population. Most studies concentrate on the baseline risk factors and description of the cohort involved, but there is generally very little information provided about the death event, or even the weeks preceding death. It is not uncommon to see studies published where the death event is linked to comorbid conditions obtained a number of years prior to the death event. I believe that studies that look at the characteristics at the time of death would provide unique insights that might be helpful in the prevention of these events. When the detective is confronted with a homicide, he starts at the scene of the crime; in health research, we tend to leave the scene of the crime unexamined. This problem is heightened in PD because, as individual physicians, we each see very few deaths in the PD population and it is difficult to form hypotheses as to what could be done to prevent death in that critical period before death occurs.
The study by Chow et al. provides us insight into cardiac arrests in the peritoneal dialysis population and should serve as a call to arms for further research in this area.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
