Cardiac myosin-binding protein C as a candidate biomarker in heart failure: rational but not revolutionary.

Abstract

This article refers to ‘Cardiac myosin-binding protein C in the diagnosis and risk stratification of acute heart failure’ by N. Kozhuharov et al., published in this issue on pages 716–725. Biomarkers have become a well-established part of the diagnosis and management of heart failure (HF). The natriuretic peptides, particularly B-type natriuretic peptide (BNP) and N-terminal pro-B-type natriuretic peptide (NT-proBNP), are endorsed by all authoritative guidelines as a highly recommended aid in the diagnosis of acute HF (AHF).1, 2 They are reflective of intra-cardiac haemodynamic load. Their widespread application has made a major impact on clinical decision-making in both acute and chronic HF worldwide. Their best proven use is in the timely distinction of AHF among patients presenting with new onset of shortness of breath. In addition, plasma concentrations of cardiac troponins, T and I, now both measurable by highly sensitive assays, provide a semi-quantitative estimate of the extent of ongoing cardiomyocyte injury in HF and give independent prognostic information in both acute and chronic HF.3 However, these existing tools, though providing considerable power to rule out AHF, have imperfect positive predictive diagnostic value, which for NT-proBNP sits at about 70%, in distinguishing AHF from other causes of recent onset dyspnoea.3 Furthermore, despite consistent reports suggesting that plasma natriuretic peptides and cardiac troponins are independently predictive of important clinical outcomes in acute and chronic HF, the best comprehensive multivariable models incorporating these markers still offer no more than modest prognostic accuracy or precision with receiver operator-derived c statistics seldom exceeding 0.75. Hence discovery, assessment and validation of alternative or additional markers remains necessary. In this issue of the Journal, Kozhuharov et al.4 provide a powerful assessment of the diagnostic and prognostic performance of an attractive candidate marker in AHF, cardiac myosin-binding protein C (cMyC). This cardiac-specific protein is an essential structural component of the cardiac sarcomere. Normal median plasma concentrations have been reported at 12 ng/L with a 99th centile upper limit of 87 ng/L and plasma cMyC is measurable in well over 99% of normal subjects using a recently validated sensitive assay.5 Hence cMyC is some fivefold more abundant in plasma than cardiac troponins. In vitro studies suggest release of cMyC from as little as 40 mg of myocardium would lead to clear elevations in circulating cMyC.6 In the context of human acute coronary syndromes, cMyC appears to peak in the plasma earlier than troponin and its diagnostic performance for acute myocardial infarction is comparable to that of troponin I.7, 8 On this background, Kozhuharov et al. have extended investigations of cMyC into the realms of diagnosis and prognosis in AHF. Plasma marker concentrations were assayed in 1083 participants in the Basics in Acute Shortness of Breath EvaLuation (BASEL V) study recruited after presentation to the emergency department with acute dyspnoea. Of these, 548 (51%) received a diagnosis of AHF as adjudicated by two independent cardiologists. Patients were followed up systematically for 1 year and key events captured including deaths and hospital admissions for HF. Median plasma cMyC concentrations were higher in AHF compared with other causes of breathlessness (72 vs. 22 ng/L). The strength of plasma cMyC in distinguishing AHF from other causes of acute dyspnoea as assessed by the area under the receiver-operating characteristic curve (AUC) was 0.81 compared to 0.79 for high-sensitivity troponin I (hsTnI) (P = NS) and 0.91 for NT-proBNP (P < 0.001). At 16 ng/L the sensitivity of cMyC was 95% with a negative predictive value of 88%, with rule-out of AHF in 21% of presentations. NT-proBNP applied at the widely used threshold of 300 pg/mL yielded sensitivity and negative predictive values of 98% and 97%, respectively, with rule-out of 28% of presentations. The prognostic performance of all three of NT-proBNP, hsTnI and cMyC was similarly modest with above median levels of the three markers conferring hazard ratios of 1.5 to 2.5 for incurring any of the nominated endpoints. C statistics for prediction of events out to 1 year fell below 0.7 for all markers. Addition of cMyC to either NT-proBNP and hsTnI did not augment prognostic performance. Furthermore, cMyC was not independently associated with outcomes in multivariable models incorporating well-recognized predictors such as age, left ventricular ejection fraction and renal function. Therefore cMYC is strongly diagnostic of AHF amongst dyspnoeic patients and modestly prognostic for events over 1 year after presentation with AHF. However, it does not match NT-proBNP diagnostically nor does it offers any additional prognostic power beyond the modest predictive performance already provided by existing markers and clinical predictors. Notably, median levels of cMyC in AHF (72 ng/L) fell below the normal reference 99th centile value (87 ng/L). Median cMyC levels in AHF were about threefold median non-AHF values. Both findings stand in contrast to the behaviour of NT-proBNP in this setting. Plasma NT-proBNP is typically 40-fold higher in AHF than other conditions and median AHF values (∼5000 ng/L) markedly exceed the 99th centile observed in healthy people.3 Further plausible reasons for this marker to under-perform in comparison to BNP or NT-proBNP in the diagnosis of AHF include the distinct mechanisms underlying the deposition of these markers into the circulation. Synthesis and secretion of cardiac natriuretic peptides is driven primarily by intra-cardiac distending pressures and the integrated effect of cardiomyocyte stretch. Natriuretic peptides are therefore a direct marker of the cardiac haemodynamic overload which characterizes AHF. They are secreted as part of a compensatory response and exert vasodilating, natriuretic and neurohumoral suppressant actions beneficial in HF. In contrast, like troponin, cMyC reflects cardiomyocyte injury with inadvertent leak into the circulation, providing only an indirect association with the acuity and severity of AHF. The bioactivity, if any, of circulating cMyC in HF is unknown. The candidate marker was further disadvantaged by the study design in that adjudication of the diagnosis of AHF included consideration of plasma BNP or NT-proBNP concentrations. Thus, natriuretic peptide values were intrinsic to the diagnosis. A more appropriate standard would have been the assessment of all three markers by adjudicators blind to values of all three markers. This report is a cogent indicator of the high bar to be crossed in order to supersede or substantially augment the diagnostic and prognostic utility already offered by plasma concentrations of NT-proBNP and cardiac troponins in HF. cMyC is cardiac-specific, abundant, has fast release kinetics and can be measured by a sensitive assay and yet does not immediately add obvious diagnostic or prognostic value in AHF. It is interesting to note that history has robbed cMyC of potential early claims to utility. If it had been discovered and applied before either the natriuretic peptide or troponin performance would have been considered remarkable, it may have become a benchmark laboratory test for AHF. It still remains possible the marker can offer advantage in highly specific and challenging circumstances. The authors report, ‘Notably, in the subgroup of patients with renal dysfunction, cMyC showed better prognostic performance for the short-term prognosis as compared to NT-proBNP (AUCs at 15 days 0.826 vs. 0.686, P = 0.007).’ This echoes the prognostic advantage of mid-regional pro-adrenomedullin over NT-proBNP for outcomes in the first 30 days after onset of acute decompensated HF.9 It is a post-hoc sub-analysis worthy of further investigation and corroboration. Furthermore, it is possible the early release kinetics of cMyC may add information to NT-proBNP with respect to estimating the timing of onset of HF. The investigators also suggest ‘…the change in [cMyC] concentrations may be more relevant for therapy guidance. The dynamic nature of cMyC, with its fast release kinetics … may provide a particular advantage in this setting.’ Circulating biomarkers of increased diagnostic specificity and predictive power in HF constitute an unmet need reflecting the shortcomings inherent in the clinical utility of the powerful yet imperfect markers currently in use. cMyc held promise and, as discussed, met several cogent criteria for selection as a rational candidate choice for diagnostic and prognostic application in HF. Although the current study indicates it does not obviously match or augment existing markers, its possible niche applications remain to be fully explored. Meanwhile the search for other more perfectly sensitive and specific markers for detection and risk stratification of HF will continue. Conflict of interest: A.M.R. reports grants, advisory board fees and support in kind from the in vitro Diagnostics Industry, including Roche Diagnostics, Abbott Laboratories, Thermo Fisher, Critical Diagnostics and Sphingotec.