STOCKHOLM – Levels of four biomarkers appear to predict risk for cardiovascular events in patients with the triad of type 2 diabetes, moderate chronic kidney disease, and anemia, a group for whom most cardiovascular risk factors are not predictive.
In a new secondary analysis of a major randomized clinical trial, urinary protein/creatinine ratio and levels of C-reactive protein, troponin T, and N-terminal prohormone brain natriuretic peptide (NT-proBNP) each provided additional prognostic information and independently predicted cardiovascular outcome in this population of patients, "who, by the way, we’re all seeing more and more in ordinary cardiology practice," Dr. John J.V. McMurray said at the congress.
"The finding that surprised me the most was that 45% of these ambulatory patients with type 2 diabetes and relatively mild anemia and renal dysfunction not requiring dialysis had an elevated troponin T, and 38% had a clearly elevated NT-proBNP," added Dr. McMurray, professor of medical cardiology at the University of Glasgow (Scotland).
In this analysis, NT-proBNP was the single most powerful predictor of outcome. It was an even stronger predictor than advanced age or baseline heart failure. Moreover, cardiac troponin T (TnT), NT-proBNP, C-reactive protein (CRP), and urinary protein/creatinine ratio (UPCR) accounted for four of the top eight predictors of cardiovascular events in patients with the triad of type 2 diabetes, chronic kidney disease, and anemia.
Dr. McMurray presented a secondary, post hoc analysis of data from TREAT (Trial to Reduce Cardiovascular Events With Aranesp Therapy), an international, double-blind, randomized, placebo-controlled, phase III study involving 4,038 patients with the triad at 623 centers. They were assigned to darbepoetin alfa (Aranesp) or placebo in order to test the hypothesis that raising their hemoglobin level would lessen the rates of death, cardiovascular morbidity, and/or end-stage renal disease during a median 2.4 years of follow-up. It did not.
Indeed, TREAT was a negative study. Darbepoetin alfa therapy not only failed to reduce the primary cardiovascular end point, it doubled the stroke risk, which Dr. McMurray and coinvestigators concluded was not balanced by the modest quality of life benefit and reduced need for RBC transfusion in the active treatment arm (N. Engl. J. Med. 2009;361:2019-32).
In a post hoc analysis of the data, the TREAT investigators found that most of the standard cardiovascular risk factors were nonpredictive in the TREAT triad population. Among these nonpredictive variables were baseline blood pressure, lipid levels, smoking history, body mass index, and a history of peripheral artery disease, lung disease, diabetic complications, and gout. Yet the triad population was clearly at high cardiovascular risk, with roughly a 25% incidence of the composite end point of cardiovascular death, MI, stroke, or hospitalization for heart failure or myocardial ischemia during 2.4 years of follow-up.
The investigators then looked for novel predictors, including biomarkers. The UPCR and CRP markers immediately stood out. In an analysis that was adjusted for other predictive variables, TREAT participants in the bottom tertile for both CRP and UPCR had a 6% annual rate of the composite cardiovascular end point, rising to 20% in those in the highest tertile for both.
Next the investigators turned to NT-proBNP and TnT, which were measured in the first 1,000 subjects. After controlling for other outcome predictors, including UPCR and CRP, the researchers found that the adjusted annual risk of the composite end point in patients who were in the lowest tertile for NT-proBNP and had an undetectable TnT was 5%, climbing stepwise to 30% per year in those in the top tertile for NT-proBNP and a TnT greater than the median 0.028 ng/mL. That’s nearly a sevenfold difference in annualized risk, the cardiologist noted.
"We believe that in addition to aiding risk stratification, these plasma and urinary biomarkers may help in understanding the mechanisms of cardiac risk in this important patient population and suggest new therapeutic strategies in this population," Dr. McMurray concluded.
Discussant Dr. Karl Swedberg recalculated the TREAT data using a different statistical method and came up with a slightly different rank ordering of risk predictors. By his measure, the most powerful predictor of cardiovascular events was baseline CRP, with a chi-squared value of 19.1, followed closely by NT-proBNP, age, TnT, and heart failure. Thus, biomarkers accounted for three of the five strongest predictors, with UPCR lying further down the list, albeit still a significant predictor. But this rearrangement is a small matter, he added.
"I fully concur with the conclusions of Dr. McMurray. I think they’re important, valuable, and will probably be useful," said Dr. Swedberg, professor of medicine at the University of Gothenburg (Sweden).