Connecting the C's: Coronaries, Creatinine, Compliance, CRUSADE

Chronic kidney disease (CKD) is highly prevalent in the United States, affecting 1 in 6 individuals over the age of 20.1 Left untreated, CKD can result in end-stage renal disease (ESRD) necessitating either hemodialysis or kidney transplantation. The National Kidney Foundation (NKF), along with the American College of Cardiology (ACC) and the American Heart Association (AHA), recognize CKD as a coronary heart disease (CHD) risk equivalent.2, 3 Many of the risk factors for CKD, such as diabetes, hypertension, and obesity, also underlie coronary artery disease (CAD). Therefore, it should not be surprising that CAD is highly prevalent among CKD patients and accounts for significant morbidity and mortality.4, 5, 6, 7 Furthermore, the risk of cardiovascular events, including mortality, correlate with the severity of the CKD.4, 8 These findings underlie the importance of aggressively diagnosing and treating CAD in patients with CKD.

The ACC/AHA guidelines for management of patients with acute coronary syndromes (ACS)—both unstable angina/non–ST-elevation myocardial infarction (MI) and ST-elevation MI—are similar for patients with and without CKD, with the exception of the need to adjust the dosing of drugs that are excreted by the kidney.3, 9 These guidelines provide the scientific evidence underlying the use of specific pharmacologic and procedural therapies among patients with ACS. Standard pharmacologic therapies include the prescription of aspirin, a thienopyridine (clopidogrel, ticlopidine), anticoagulation (unfractionated or low-molecular-weight heparin), β-blockers, angiotensin-converting enzyme (ACE) inhibitors or angiotensin receptor blockers, and lipid-lowering agents.10 The administration and timing of reperfusion therapy (thrombolytic therapy or primary coronary intervention) is an additional measure among patients with an ST-elevation MI. There is increasing evidence that physicians are less likely to comply with these quality indicators in patients with CKD.11, 12, 13, 14, 15

In this issue of the American Journal of Kidney Disease, Patel et al16 examine the interaction between CKD and hospital quality indicator performance on the utilization of evidence-based acute MI (AMI) therapies. The source of the data is the CRUSADE (Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes With Early Implementation of the ACC/AHA Guidelines) registry, an observational cohort of 81,374 patients with non–ST-elevation acute coronary syndromes treated at 327 hospitals between 2003 and 2006. Hospitals were ranked into 4 groups based on the proportion of patients without contraindications that were provided standard pharmacologic therapies; the “leading” group had top performance and “lagging” group had worst performance. Both in-hospital prescriptions (aspirin, β-blockers, clopidogrel, heparin, and glycoprotein 2b3a inhibitors) and discharge prescriptions (aspirin, clopidogrel, ACE inhibitor, and lipid-lowering therapies) were examined. Patients from hospitals within the same rank were then divided into deciles based on the estimated glomerular filtration rate (eGFR) and the proportion of patients who received standard pharmacologic therapies was calculated. The authors assessed the role of eGFR, hospital rank, and the interaction between these variables on compliance with medication guidelines in leading versus lagging hospitals. Better-performing hospitals had lower prescribing rates for several therapies with lower kidney function. In contrast, lower-performing hospitals had similar prescribing rates across the range of kidney function.

The most noteworthy finding of this observational study was the underutilization of standard AMI therapies among patients with CKD. Both in-hospital prescriptions (aspirin, clopidogrel, heparin, and GP2b3a inhibitors) and discharge prescriptions (aspirin, clopidogrel, lipid-lowering agents, and ACE inhibitors) decreased with lower kidney function. These findings are consistent with prior observational studies of ACS patients.11, 12, 13, 14, 15, 17 In one of the earliest investigations, McCullough et al reported that the combined use of aspirin and β-blocker in ST-elevation MI patients decreased from 64% in patients with preserved kidney function to 35% in patients with an eGFR less than 46 mL/min/1.73 m2. Shiplak et al analyzed data from the Cooperative Cardiovascular Project (CCP), a retrospective study of AMI patients aged 65 and older, and observed a decrease in use of both pharmacotherapies and cardiovascular procedures with progressive decline in kidney function. In another analysis of AMI patients from the CCP, Berger et al observed a marked reduction in the use of aspirin, β-blockers, and ACE inhibitors in ESRD patients compared with non-ESRD patients.13 Wright et al, in a single-center analysis, found the use of standard AMI therapies steadily decreased with worsening category of kidney function, reaching a nadir in the small proportion of patients receiving hemodialysis.14 These findings were most recently confirmed by Santopinto et al in an analysis of the GRACE (Global Registry of Acute Coronary Events), a prospective observational study of ACS patients.15

Clearly, the underutilization of pharmacologic therapies in ACS patients with CKD is now well established; the unanswered question is which factors explain this phenomenon. First and foremost, AMI therapies have not been well studied in patients with CKD and are not accepted as “standard of care” by many physicians. Patients with CKD were either excluded or underrepresented from early randomized clinical trials of ACS.18 Few registries prospectively sought to assess cardiovascular complications among patients with CKD or renal complications among patients with cardiovascular disease. Hematologic, metabolic, and endocrinologic abnormalities in CKD patients may increase the incidence or severity of side effects of standard AMI therapies, thereby limiting their utilization.19 A classic example is the concern regarding antiplatelet therapy (aspirin and thienopyridines) in CKD patients, who tend to manifest anemia and platelet dysfunction. Second, heterogeneity regarding the definition of CKD has led to inconsistency in the literature. Remote studies utilized serum urea nitrogen or creatinine to assess the severity of disease. More recent studies have adopted creatinine clearance rate (estimated by the Cockcroft-Gault formula) or GFR (estimated by the Modification of Diet in Renal Disease Study equation) to assess the level of kidney function. Some studies included patients with ESRD, whereas other studies specifically excluded these patients. In fact, it was not until 2002 that the KDOQI (Kidney Disease Outcomes Quality Initiative) clinical practice guidelines standardized the categorization of CKD.20 Finally, a pragmatic issue results from the fact that patients with CAD and CKD are often comanaged by cardiologists and nephrologists with different practice styles and different guidelines. Consequently, the utilization of specific medical therapies might vary by physician specialty and inpatient therapies may not be transitioned to postdischarge therapies.

Patel et al stratify their analysis by hospital performance on cardiovascular quality indicators. This methodology of identifying discrepancies in quality of care has fundamental merit because it shifts the paradigm from the physician to the health care system. Identifying top-performing hospitals can be accomplished by either examining processes (performance on quality indicators) or health care outcomes (mortality, rehospitalization).10, 21, 22 Bradley et al assessed hospital performance in the CMS/JCAHO (Centers for Medicare & Medicaid Services/ Joint Commission on the Accreditation of Healthcare Organizations) AMI core process measures using data from 2002 through 2003 concerning 962 hospitals participating in the National Registry of Myocardial Infarction and correlated these measures with hospital-level, risk-standardized, 30-day mortality rates derived from Medicare claims data.21 Although several of the process measures were significantly correlated with risk-standardized, 30-day mortality rates, together they explained only 6% of hospital-level variation in the mortality. Thus, quality indicators alone cannot suffice as a surrogate of mortality. Wang et al examined the ranking of hospitals for “heart and heart surgery” in the US News & World Report and found the 30-day risk-standardized mortality rates for AMI were lower for ranked than nonranked hospitals.22 An observational analysis of the CRUSADE registry found the composite adherence rate was significantly associated with in-hospital mortality, with observed mortality decreasing from 6.3% for the lowest-adherence quartile to 4.2% for the highest quartile.23 After risk adjustment, every 10% increase in composite adherence at a hospital was associated with an analogous 10% decrease in its patients' likelihood of in-hospital mortality.

Applying the aforementioned methodology, top-performing (leading) hospitals would be expected to deliver key therapies independent of kidney function while hospitals with poor performance (lagging) would be less likely to provide individual therapies to patients with worse kidney function. Unfortunately, the patterns between hospital performance and level of kidney function identified by Patel et al were inconsistent. In fact, the hypothesis that higher-performing hospitals would have greater rates of prescribing at all levels of eGFR was only established for ACE inhibitors. For many of the medications, stochastic significance is demonstrated, but the small magnitude of difference in the odds ratio leads one to question the clinical significance of the findings. In examining the interaction between level of kidney function and hospital performance on the utilization of various therapies, one would expect to see a “dose effect” across levels of hospital performance. Unfortunately, this does not appear to be the case for the various medications, and in fact, the statistical analysis excludes the interquartile groups and compares only the leading and lagging hospitals.

A major limitation of the analysis by Patel et al was the exclusion of patients with ESRD. Although no specific mention is made regarding this high-risk population, the eGFR range had a lower limit of 20 mL/min/1.73 m2 and therefore presumably excluded dialysis patients. The incidence of AMI in patients with prevalent CKD and prevalent ESRD are 4.2% and 7.7%, respectively.24 Following an AMI, the 1-year mortality in dialysis patients is 59%.25 There is a paucity of data regarding therapy among these patients, in particular, because they were historically excluded from randomized clinical trials.13 It would have been interesting to have known how leading and lagging hospitals had performed on the various quality indicators.

In summary, patients with CKD are at high risk of developing CAD as well as other manifestations of atherosclerosis. The CRUSADE registry reinforces the perception of compliance failure, showing standard AMI therapies remain underutilized in patients with CKD. Unfortunately, the relationship between worsening kidney function and lack of adherence to guidelines cannot be adequately explained by individual hospital performance. Only when physicians “connect the C's”, that is, recognize the risk of premature CAD in patients in CKD and crusade for compliance with the ACS guidelines, will patient outcomes be improved.