Dismal long‐term survival of dialysis patients after acute myocardial infarction: can we alter the outcome?

The burden of ischemic heart disease in end‐stage renal disease

Cardiac disease is the major cause of death in dialysis patients, accounting for about 45% of all‐cause mortalities [1]. Approximately 20% of cardiac deaths are attributed to acute myocardial infarction (AMI) [1]. In the United States, the greatest increase in treated end‐stage renal disease (ESRD) has occurred in patients with the highest risk for cardiovascular disease, older patients and those with diabetic nephropathy. There were an estimated 281 000 dialysis patients in 2000, with a projected number of 520 000 U.S. dialysis patients by 2010 [2]. The burden of ischaemic heart disease in ESRD patients has been confirmed by the European Dialysis and Transplant Association (EDTA) registry. In the United Kingdom, the all‐age death rate from myocardial ischaemia and AMI for patients commencing renal replacement therapy in 1985 to 1990 was 16.6‐fold greater in men and 17.7‐fold greater in women, compared to the general population [3]. With this increasing burden of cardiovascular disease in dialysis patients, the number of cardiac events in ESRD patients is certain to increase.

The pathogenesis of vascular lessions

Chronic renal failure, a condition plagued by extraordinary cardiovascular risk, has been justly deemed a ‘vasculopathic state’ [4]. A host of potentially modifiable risk factors contributing to accelerated cardiovascular morbidity and mortality in renal patients (both pre‐ESRD and ESRD) include hypertension, dyslipidaemia, hyperglycaemia, smoking, physical inactivity, enhanced thrombogenicity, and hyperhomocystinaemia. The development of left ventricular hypertrophy may be promoted by anaemia and vascular non‐compliance. Aortic stiffness is an independent predictor of cardiovascular and all‐cause death in dialysis patients [5]. Premature coronary artery calcification has been detected in young dialysis patients, and the metabolic milieu of ESRD, including elevated calcium–phosphorus product, may be implicated [6]. Vascular endothelial dysfunction is likely to contribute to the expression of atherosclerotic disease, and even a single haemodialysis run may adversely affect endothelial function [7]. In diabetic patients, hyperglycaemia may cause endothelial dysfunction by promoting the formation of advanced glycation end products, which may oppose nitric oxide‐mediated endothelium‐dependent relaxation [8]. The composition of coronary plaques in patients with ESRD may be qualitatively different, with increased media thickness and marked calcification of the affected coronary arteries compared with non‐renal patients [9,10].

Mortality of dialysis and cardiac death

The largest cause of cardiac death identified in the United States Renal Data System (USRDS) database is ‘cardiac arrest, cause unknown’, which accounted for 47% of all cardiac deaths in U.S. dialysis patients from 1996 to 1998 [1]. The presence of left ventricular hypertrophy and the concomitant occurrence of abnormalities in myocardial ultrastructure and function, including interstitial fibrosis, decreased perfusion reserve, and diminished ischaemia tolerance [11–14], may make ESRD patients particularly vulnerable to sudden cardiac death. The unnatural prescription of routine haemodialysis schedules (usually thrice weekly in the United States) may also contribute to increased cardiac death, based on Bleyer et al.'s finding that significantly higher cardiac mortality occurs on a Monday or Tuesday [15]. It is unclear when these excess deaths temporally occur in relation to the scheduled haemodialysis run. It is tempting to implicate the non‐physiological effects of rapid volume and electrolyte shifts, but these can only be partly to blame as the cardiac death rate (158 deaths/1000 patient years) of diabetic patients receiving peritoneal dialysis (without the attendant rapid volume changes) is actually higher than for diabetic haemodialysis patients (126 deaths/1000 patient years) [1]. Preliminary data indicate worse long‐term survival after AMI for patients receiving peritoneal vs haemodialysis. In a comorbidity‐adjusted Cox model, the risk of all cause death after AMI was 19% lower for haemodialysis vs peritoneal dialysis [16]. The favourable Tassin (France) experience with long‐duration dialysis offers one scenario for better cardiovascular outcomes in ESRD patients. Hopefully, this entire issue relating to the control of volume status on cardiovascular survival will be clarified by a randomized, prospective trial of conventional vs long‐duration dialysis in a contemporary population of high cardiovascular risk dialysis patients.

Mortality of acute myocardial infarction in the patient on dialysis or after renal transplantation

Acute myocardial infarction in dialysis patients is a catastrophic event associated with dismal long‐term survival. Using USRDS data, we reported a one‐year mortality of 59% and a 73% two‐year mortality for 34189 dialysis patients in the U.S. sustaining AMI in 1977 to 1995. Even more striking was the poor survival of patients treated in the ‘era of reperfusion’: the one‐ and two‐year mortalities of patients with AMI in 1990–95 were, respectively, 62% and 74% [17]. Chertow et al. [18] reported a one‐year 53% mortality in 640 U.S. dialysis patients with AMI in 1994–95. These patients would be regarded as high risk by conventional clinical criteria, as 62% had congestive heart failure at admission. Although the definition of congestive heart failure in a dialysis patient is obfuscated by the confounding effect of volume status, there was a 40% increased death risk by univariate analysis for congestive heart failure. In Okinawa, Japan, the one‐year mortality of 61 dialysis patients with AMI was 63% [19]. Beattie et al. [20], using the Henry Ford Hospital (Detroit, Michigan, USA) clinical database, found a two‐year 65% mortality for 47 dialysis patients hospitalized for AMI in 1990–98. Interestingly, the early survival of patients with severe non‐dialysis‐dependent renal failure (Cr=2.7±2.6 mg/dl) was even worse than the dialysis group. In a survival analysis of 113 dialysis patients hospitalized for AMI at Hennepin County Medical Center (Minneapolis, Minnesota, USA) from 1977 to 1994, the in‐hospital mortality was 16% for patients with non‐transmural MI and 52% for transmural MI (with 50% for acute anterior MI and a surprisingly high 56% in‐hospital mortality for acute inferior MI) [21]. To put these survival data in perspective, in a cohort of 29249 U.S. Medicare‐eligible male patients (mean age 75.5±7.0 years) hospitalized for AMI in 1994–95, the one‐year mortality was 31.8% [22]. In a logistic regression model, patients with a serum creatinine of 1.5–7.0 mg/dl had a two‐fold increased death risk.

The survival rate of renal transplant recipients after AMI is considerably better than for dialysis patients. Using USRDS data, we reported a two‐year mortality of 34% in 4250 renal transplant patients hospitalized for AMI 1977–96 in the United States [23]. In a recent unpublished analysis, we compared the survival of 71 472 dialysis patients, 5573 transplant waiting list patients, and 6705 renal transplant recipients hospitalized for AMI in 1977–99. The two‐year survival was 27% for dialysis patients, 44% for transplant waiting list patients, and 67% for transplant recipients. These data suggest that the superior outcome of renal transplant recipients after AMI may reflect more than patient selection bias, particularly since the survival advantage of waiting list patients vs dialysis persists even after adjustment for demographics and comorbidity, with a 34% decreased long‐term death risk after AMI.

Are dialysis patients given optimal treatment after acute myocardial infarcation?

It is unclear to what degree the dismal long‐term survival of dialysis patients after AMI reflects the intrinsic vulnerability of ESRD patients to cardiac death or the consequences of therapeutic nihilism and deficiencies in the delivery of modern cardiac care to dialysis patients. In the United States there is a significant geographic regional variation in the survival of dialysis patients after AMI; after adjustment for demographic differences and comorbid conditions, there is a 30% variation in post‐AMI survival related to renal network [24]. There appears to be a striking underuse of thrombolytic therapy in dialysis patients with AMI. We have reported a two‐year 78% mortality in 33277 U.S. dialysis patients sustaining AMI in 1991–97, who received no coronary reperfusion therapy. Using claims data, we identified 176 patients receiving intravenous thrombolytic therapy with a two‐year mortality of 60%. In a comorbidity‐adjusted Cox model, thrombolytic therapy was associated with a significant 28% reduction in all‐cause death risk [25]. Beattie et al. [20] reported a relative underutilization of beta‐blocker therapy in dialysis patients admitted with ST‐segment elevation MI at their own institution (48% in dialysis patient vs 81% in non‐renal patients). The unfortunate exclusion of ESRD patients from past clinical trials on the treatment of acute coronary syndromes may have abetted a nihilistic approach to the treatment of dialysis patients, as data on the safety and efficacy of modern pharmacological agents (which have been shown to benefit non‐renal patients) for the treatment of acute coronary syndromes is non‐existent in dialysis patients. In past surveys of medical treatment of dialysis patients, the prescription of therapeutic agents known to improve cardiovascular survival has been found lacking in the United States. For example, in the USRDS Dialysis Morbidity and Mortality Wave 2 Study, a disproportionately high percentage (52%) of incident dialysis patients in 1996–97 were prescribed calcium channel blockers for hypertension, as compared to beta‐blockers (17%) and ACE inhibitors (24%) [26]. In addition, we have found an apparent underutilization of diagnostic testing related to the management of ischaemic heart disease in dialysis patients with AMI in the United States, although this has improved over time. We identified 33331 Medicare‐eligible incident dialysis patients sustaining AMI from 1991 to 1998. In 1991, coronary angiography was performed in 14.5% of these patients, 18.6% in 1994, and it had risen to 25.9% in 1998 [27]. In a Medicare‐eligible non‐dialysis cohort hospitalized in 1994–95 for AMI, cardiac catheterization was performed in 45.4% of patients [22]. In dialysis patients sustaining AMI in 1998, only one‐sixth had lipid testing within 6 months of the AMI hospitalization [27]. One lesson learned from past clinical trials is that patients at the highest risk will reap the largest absolute benefit from effective therapies. Unless there is strong evidence to the contrary, treatment strategies for AMI which have been validated in the non‐renal population should be implemented in dialysis patients, as it would be difficult to imagine worse outcomes than we have reported in dialysis patients receiving no treatment.

Can we alter the outcome?

The outlook for dialysis patients with AMI can be improved. I suspect that under‐recognition of acute ischaemic syndromes, coupled with under‐treatment, plays a role in magnifying the unfavourable impact of the ‘vasculopathic’ state present in patients with chronic renal failure (both ESRD and non‐ESRD). When a dialysis patient shows up for a Monday haemodialysis run 5 kg above ‘dry weight’ and complaining of dyspnoea (and sometimes angina), would a nephrologist order an electrocardiogram or a dialysis run first? Most of the time the symptoms will reflect circulatory congestion, and will remit with dialysis. The early recognition of an AMI, however, might be problematic in this clinical setting because the symptoms of volume overload (in these patients with stiff, hypertrophied (i.e. pre‐load sensitive) left ventricles) and AMI may be indistinguishable. Acute myocardial ischaemia from coronary thrombosis and acute ingestion of a pepperoni pizza (or Westphalian ham?) could cause the same symptoms on a Monday morning, but it is doubtful that dialysis is a good intervention for an AMI. It is conceivable that the high prevalence of left ventricular hypertrophy (and attendant repolarization abnormalities) may contribute to ambiguity in the electrocardiographic diagnosis of AMI. The initiation of dialysis may play a role in the triggering of AMI. We have reported an early hazard of AMI related to dialysis initiation, as 52% of infarcts occurred within two years of dialysis initiation, compared to 29% in the first two years after renal transplantation [17]. One potential clinical strategy would be to target dialysis patients for therapeutic coronary intervention shortly after dialysis initiation. It may be possible to identify a subset of dialysis patients at the highest risk for mortality, perhaps with outpatient testing for the presence of increased serum troponin I and troponin T, as both appear to prospectively identify dialysis patients at increased risk for mortality [28,29].

Ultimately, the most promising method to improve cardiovascular outcome in ESRD patients will be to utilize therapeutic strategies (and most importantly, pre‐ESRD patients) which are successfully used in the non‐renal population. Cardiovascular clinical trials targeting patients with renal failure will, hopefully, encourage a more active approach in the treatment of ESRD patients. Two noteworthy clinical trials in dialysis patients currently in progress are the ‘4‐D Study’ (a study of atorvastatin in German diabetic dialysis patients) and CHORUS (a study of cerivastatin in North American dialysis patients). The creation of clinical registries collecting outcome data of interest in patients with chronic renal failure would also facilitate the advancement of cardiovascular care in these patients. It is evident that patients with chronic renal failure are at high risk for cardiovascular disease, and, thus, have the most to gain from aggressive attempts at prevention, diagnosis, and treatment of cardiovascular disease.

References