Using Biomarkers to Guide Heart Failure Therapy

Being able to target specific therapies to specific disease entities is the goal of all precision medical initiatives. Such approaches have been implemented in malignancies with individual examples of great responses. Such successes have fueled enthusiasm to apply this approach to other forms of chronic disease. Ideal implementation would include a companion diagnostic (a blood test), which would specifically signal when initiation, reduction, or intensification of a therapy is indicated. Testing could then be used serially to titrate therapy. This ideal state has not yet fully developed for cardiovascular diseases but has been pursued in some areas of heart disease management such as using cardiac troponin to determine therapy with antiplatelet drugs, inflammatory biomarkers to define the need for lipid lowering agents, and cardiovascular stress biomarkers to guide heart failure (HF)3 management (1). Most data in chronic HF have focused on patients with reduced ejection fraction (HFrEF). Such patients are treated with a variety of defined treatment strategies predicated on guideline-supported therapies that reduce mortality (2). Although these therapies have been proven in randomized controlled trials, inadequacies in their application exist in nontrial settings. This may be because it is difficult for physicians to evaluate changes in functional status during patient visits and thus decide when to increase or decrease therapies. Furthermore, the polypharmacy in HFrEF is titrated to target doses, rather than physiological goals. Accordingly, understanding the timing and appropriateness of titration may be challenging for less experienced clinicians especially if patients insist they feel fine. For these reasons, multiple biomarkers have been suggested to assist objectively in the titration of therapy. The ones most utilized have been the natriuretic peptides (NPs). Increases in NP concentrations provide an objective signal concerning the compensation of HF patients. Application or intensification of these therapies has been shown to reduce NP concentrations in association with better outcomes (3, 4). The approach of biomarker guided HF care has been evaluated in several preliminary trials. The Pro-B-type Natriuretic Peptide Outpatient Tailored Chronic HF Therapy (PROTECT) trial provided useful insights for optimal application of such guided therapy (5). Despite being stopped prematurely, the results suggested care guided by N-terminal probrain natriuretic peptide (NT-proBNP) might reduce risk of cardiovascular events with the following caveats: Marked reductions in NP values must occur. In contrast to other trials, there were very marked reductions (nearly a halving) in NT-proBNP in the guided group compared to minor reductions in the standard therapy group. This was associated with marked improvement in outcomes. Therapy adjustment must be made. Key to the trial was the fact that there was a push to augment therapy whenever NT-proBNP was less than ideal even if patients appeared clinically stable. Therapies must be up-titrated beyond that of standard-of-care. If comparable care can be delivered without a guide, the approach would not be of value. These caveats indicate that guiding therapy with a biomarker is, in fact, a strategy whose success depends heavily on clinician's response to an abnormal biomarker result. Thus, an underresponse to an abnormal result, or overtreatment in the “usual care” arm could lead to similar outcomes between arms. Preliminary results about the use of biomarker to guide HF therapy summarized in several meta-analyses suggested that benefit was perhaps present (6). This finding led to a larger, pivotal trial design, the Guiding Evidence-Based Therapy Using Biomarker Intensified Treatment in Heart Failure (GUIDE IT) study (NCT01685840) (7). This trial has recently been stopped due to futility. The reasons for this are unclear but the lessons learned from PROTECT experience might help us to understand these results. It could be that sophisticated clinicians titrated therapies equally aggressively in both biomarker guided and usual care arms, or that clinicians did not respond aggressively to increased NT-proBNP concentrations. It also may well be that the biologic variation of NPs (25%–80% in patients with HF) might have made it difficult to titrate therapy, rendering changes in therapy less efficacious. Finally, because NP increases do not provide guidance in regard to specific therapies, although increased therapy was administered, less than ideally efficacious therapies were chosen. The results of GUIDE IT create opportunity to consider other biomarkers for guiding HF care. There are several ways in which new biomarkers might help guide care. We propose the following: The biomarker should be mechanistically linked to a prognostic process in HF. A clear prognostic threshold for the biomarker must be defined. Increases in the biomarker should lead to a clearly defined and specific therapeutic response. The biomarker should optimally “respond” to therapies so it can serve as a monitor for the response. The biomarker should be suitable for serial measurement, meaning it changes fairly rapidly to support rapid titration of therapy. Data concerning the optimal timing for remeasurement must be known, and how much change in the biomarker is needed before there is assurance that a definitive change has occurred must be established. Suitable assays for use must be available with defined imprecision and sufficient sensitivity for clinical use. To date, there are a handful of biomarkers worthy of study (Table 1) (1). Candidate “next” biomarkers for guiding chronic HF therapy.a Markers are graded on a scale from − to ++. References are available in the online Supplemental Data File. Candidate “next” biomarkers for guiding chronic HF therapy.a Markers are graded on a scale from − to ++. References are available in the online Supplemental Data File. Into this arena comes a recent report suggesting that carbohydrate antigen-125–guided therapy may be helpful in the care of HFrEF (8). Carbohydrate antigen-125 is a large membrane associated mucin with 3 domains, including a single transmembrane domain and 2 heavily glycosylated extracellular domains found in several tissues including the epithelia of the abdominal cavity. Carbohydrate antigen-125 concentrations are widely measured to monitor ovarian cancer but the biomarker may also be increased in patients with HF where the increases are thought to reflect increased blood volumes. Inflammation also appears to influence the biomarker concentrations. Concentrations of carbohydrate antigen-125 rise and fall in patients undergoing therapy for HF, and the biomarker values may be prognostic for adverse outcomes. Published data suggest a cutoff value of >35 U/mL optimally distinguishes patients with volume overload from those without volume expansion. With this background, a randomized trial testing the use of carbohydrate antigen-125 was implemented in 380 patients in Spain, using increases to influence titration of diuretic treatment; intravenous treatments were particularly enhanced. Patients in the carbohydrate antigen-125 arm were seen more frequently and had more adjustments in diuretics (including mineralocorticoid receptor antagonists and oral and intravenous loop diuretics). In recognition of a possible link to inflammation, those in the carbohydrate antigen-125-guided arm also had more titration of statin therapy. After 1 year, carbohydrate antigen-125-guided HF care reduced death/repeat HF hospitalization as a first event (66 events vs. 84 events; P = 0.02) and recurrent events (85 events vs. 165 events; incidence rate ratio: 0.49; 95% CI: 0.28 to 0.82; P = 0.008). The study was not powered to evaluate mortality as a solitary end point and therefore the results were predominately driven by an improvement in hospital readmissions. All in all, the success of carbohydrate antigen-125 to guide HF care followed lessons learned from the PROTECT study, which we consider critical. Besides uncertainty regarding its mechanism, several things are unclear about carbohydrate antigen-125 as a biomarker in HF. We do not know the biologic variation of carbohydrate antigen-125; the magnitude of changes needed to anticipate benefit is lacking. Biomarkers with lower biologic variation such as troponin, soluble ST2, and galectin-3 also might be good markers for guiding and titrating therapy for that reason (9, 10). This may be true with carbohydrate antigen-125 as well although given the assumption that its physiology is volume-dependent, one might think that its biologic variation would be high. Nonetheless, if increases of carbohydrate antigen-125 are mostly volume-dependent, then such changes may allow clinicians to move toward a specific therapy, i.e., diuretic therapy. Another concern is that carbohydrate antigen-125 is heavily glycosylated. It is unclear whether the glycosylation status of carbohydrate antigen-125 changes in HF, and if such glycosylation affects analytical performance. Guiding therapy using biomarkers, whether NPs, troponins, soluble ST2 or carbohydrate antigen-125, has substantial potential to markedly improve chronic HF care. If one could titrate therapies predicated on biomarkers, one might have an inexpensive and effective approach to therapy that would benefit patients. This would contrast substantially with some of the other approved approaches such as the use of invasive cardiac monitoring, which provides day to day information but at a very substantially increased expense. Several key issues need to be considered. These include: Who should monitor therapy? Experienced HF clinicians may not require a “guide.” For less experienced clinicians who manage most HF patients, there might be a greater value to guided therapy. Where should therapy be monitored? A remote testing strategy such as in the home, may provide unique opportunities to deploy a biomarker guided HF therapy approach. It is inevitable that assays using fingerstick blood samples on portable analyzers will be developed soon for these biomarkers. When/how should therapy be monitored? A recent trial of daily BNP values suggested trends in results were superior to single measurements to predict events (11). This needs to be defined for all of these approaches. Which therapies should be considered? Understanding exactly which therapies should be adjusted in the face of an abnormal biomarker value is critical in our opinion. Ideally, one would want to develop biomarkers that would lead to consideration of increases in therapies with proven mortality benefit, such as beta blockers or vasodilators (12). More targeted tissue approaches where one could potentially begin to probe organ-specific values for emerging biomarkers would be helpful. Though the road ahead is less clearly defined with the result of GUIDE IT, there is nonetheless reason for enthusiasm for guiding therapy using biomarkers. We would argue that the idea of guided therapy is still likely to be successful if deployed correctly, resulting in identification of unrecognized therapeutic opportunities, with consequent response by the clinician. In addition, while nearly all data have focused on therapy of HFrEF, we suggest efforts should also include therapeutic approaches to better target therapies for HF with preserved ejection fraction (13). This may require a different conceptual approach to the management of HF as will application of this sort of approach to acute HF (14, 15). To date, biomarkers have been developed and approved simply because they are a marker of disease or of an adverse response, without specifically triggering a therapeutic response. We would argue that this is no longer enough, and that modern development of biomarkers needs to focus on the therapeutic implications of an abnormal result and how one could leverage the biomarker to improve patient care. The commitment to doing randomized trials is important. This commitment should come from the in vitro diagnostics industry in partnership with the pharmaceutical industry, collaborating with trialists and regulators to provide targeted therapy strategies for the highest priority diagnoses in modern medicine. The high prevalence, high cost, and high morbidity of HF makes it a continued attractive target for this effort. heart failure HF with reduced ejection fraction natriuretic peptides Pro-B-type Natriuretic Peptide Outpatient Tailored Chronic HF Therapy N-terminal probrain natriuretic-peptide Guiding Evidence-Based Therapy Using Biomarker Intensified Treatment in Heart Failure.