Biomarkers in Acute Heart Failure: From Risk Markers to Therapeutic Targets

Scientific and technological progress has led to the discovery of an unprecedented large number of circulating molecules (“biomarkers”) in the past decade. However, the role of these biomarkers in clinical practice is variable and, in many cases, uncertain. Specifically in heart failure (HF), biomarkers have been used to (1) identify pathophysiologic perturbations that either precede HF or result as a downstream consequence of HF and altered physiology of target organs; (2) aid in the diagnosis, differential diagnosis, and classification of HF; (3) guide therapy and management; and (4) refine risk stratification.1 However, although newer markers are now approved by the Food and Drug Administration (eg, galactin-3), there are currently no other uses for biomarkers in HF endorsed by national or international guidelines, with the exception of certain applications for B-type natriuretic peptide (BNP) or its precursor fragment N-terminal pro-BNP (NT-proBNP).2 A large number of molecules have been shown to correlate with or refine prognosis in acute and chronic HF. However, in the absence of a common framework for development of biomarkers in HF and the disconnect between projected risks, underlying biology, and therapeutic decisions, these associations between the various biomarkers and prognosis have had minimal impact on clinical practice.1 This discrepancy between discovery and clinical adoption of biomarkers in HF is also true for biomarkers of renal dysfunction, both in acute and chronic HF settings. Renal dysfunction is common in patients with HF, with a prevalence ranging from 20% to 57% in chronic HF and 30% to 67% in patients with acute HF.3 Although the prognostic impact of worsening renal function during an acute HF episode is still a matter of debate, and likely represents a manifestation with variable underlying etiology with differing impact on prognosis, there is little controversy that chronic renal dysfunction is associated with worse long-term outcomes and increased health care resources utilization among patients hospitalized with acute HF.4 The association of renal dysfunction with worse outcomes in acute HF has been consistently demonstrated both with routine clinical (eg, serum urea nitrogen [BUN], creatinine, estimated glomerular filtration rate [eGFR]) and novel (eg, neutrophil gelatinase-associated lipocalin, cystatin C) biomarkers.3 However, it is currently unclear which renal biomarker (or biomarkers) adds the most useful information from a clinical perspective for patients with acute HF and how to use these for clinical decisions. In this issue of the Journal, Kim and colleagues report that serum levels of both cystatin C and uric acid conferred superior prognostic information compared with other renal function biomarkers (BUN, creatinine, and eGFR) in a cohort of 587 patients admitted with acute HF.5 The authors appropriately acknowledge that cystatin C and uric acid may have excelled by a small margin among renal biomarkers as a result of chance, and that further studies would be needed to confirm the optimal biomarker of renal function for risk prediction in acute HF. Nevertheless, both cystatin C and uric acid led to significant improvements in risk reclassification for mortality and HF readmission in this acute HF cohort. Importantly, the risk for adverse events was more pronounced among patients presenting with elevated levels of both biomarkers. Several findings in this study merit further discussion. Cystatin C has been previously identified as a more accurate marker of renal dysfunction in various cohorts, including patients with acute HF.6 Similarly,6 Kim and colleagues have demonstrated that addition of cystatin C to a clinical model for prediction of clinical events after an acute HF episode improves not only model fit, but also risk classification, which is a more clinically relevant statistic. However, although risk reclassification has been adopted as the “holy grail” of clinical relevance for prognostic markers in cardiovascular medicine, it is often overlooked that meaningful reclassification requires a meaningful initial classification. From this perspective, the value of cystatin C as a prognosticator is limited by the lack of (1) a widely acceptable model for risk prediction in acute HF and (2) decision-making thresholds based on absolute risks in acute HF. For example, decisions for implantation of a circulatory assist device and/or listing for heart transplantation in stable, eligible patients with stage D HF may be facilitated by one-year mortality projections with multifactorial risk assessment tools such as the Seattle Heart Failure Model.7 Thus, prognosis refinement with biomarkers could impact decision making in this group.8 However, no model has found its way into practice to date or has been widely accepted as the “yardstick” against which new developments should be compared in acute HF. As a result, Kim and colleagues5 have resorted to an ad hoc, data-driven baseline prediction model upon which the incremental value of cystatin C was derived. This is a common approach in acute HF studies but comes with certain drawbacks. We have previously outlined the limitations of this approach.1 Importantly, there are no clinical decisions that are currently based on absolute projected risks of death or readmission in acute HF. Although there is currently a significant enhancement of activities to reduce the 30-day readmission rates among patients with acute HF, these are mostly driven by administrative concerns with little scientific basis, and definitely are not dependent on individual patient risk projections.9 The finding that elevated uric acid levels are independently associated with worse prognosis in acute HF,5 although not entirely novel as acknowledged by the authors, is more intriguing. In contrast to the other markers of renal function or damage, at least at this point of development, uric acid is potentially a clinically useful therapeutic target in HF.1 Not only elevated uric acid signifies increased risk, which qualifies this biomarker as a risk marker in acute HF, but also partially shares the properties of a risk factor. In patients with HF, hyperuricemia is associated with reduced vasodilator capacity and impaired peripheral blood flow, which is reversed by xanthine oxidase inhibition.10 The causal link is believed to be the increased activity of xanthine oxidase, which leads to production of reactive oxygen species and increased oxidative stress in HF.11 Xanthine oxidase inhibition leads to improvement in endothelium-dependent vasodilation and reduced levels of oxidative stress12 and improved myocardial efficiency13 in HF. In patients with systolic HF receiving optimal medical therapy, oxypurinol 600 mg once daily reduced serum uric acid by approximately 2 mg/dL compared with placebo and in a post hoc analysis was shown to improve the composite endpoint of HF morbidity, mortality, and quality of life among those patients with uric acid levels ≥9.5 mg/dL.14 Currently, a phase II, National Institutes of Health–funded study, the Xanthine Oxidase Inhibition for Hyperuricemic Heart Failure Patients (EXACT-HF) trial, is prospectively evaluating the efficacy of allopurinol treatment for high-risk chronic HF patients with low ejection fraction and uric acid ≥9.5 mg/dL.15 Importantly, uric acid can be serially tracked and potentially serve as surrogate of effective treatment in these patients. In the oxypurinol trial,14 uric acid response to oxypurinol correlated with clinical response. Two additional findings in the study by Kim and colleagues5 merit special attention. The uric acid level associated with worse outcome in the current study (6.6 mg/dL) is commonly regarded as “normal” and far below the threshold for enrollment in EXACT. Finally, the mean ejection fraction in this acute HF cohort was ≥50%, whereas prospective studies with xanthine oxidase inhibition have only included systolic HF patients thus far. Therefore, these features might be important for consideration in future research. In conclusion, it is currently unclear how to translate the increased risk associated with elevated renal biomarkers clinically in the acute HF setting. However, elevated uric acid may be a useful surrogate for increased oxidative stress and a potential therapeutic target in HF. This hypothesis is currently under active investigation in chronic HF. The study by Kim and colleagues generates several new hypotheses for xanthine oxidase inhibition in the setting of acute HF. First, should we test the efficacy of xanthine oxidase inhibition in patients with acute HF specifically? It is intriguing that intravenous administration of allopurinol acutely improves the concentrations of myocardial high-energy phosphates and adenosine triphosphate flux through creatine kinase in patients with nonischemic cardiomyopathy.16 Second, in such a hypothetical acute HF trial, should the uric acid threshold be the same as in chronic HF? And finally, should we include patients with HF and preserved ejection fraction? All these important questions need prospective investigation in rigorous clinical studies.