Measurement of circulating natriuretic peptides to identify community-based patients at risk of incident heart failure - should we?

Abstract

This article refers to ‘Prediction of incident heart failure by serum amino-terminal pro-B-type natriuretic peptide level in a community-based cohort’ by D.J. Campbell et al., published in this issue on pages 449–459. Readers of this Journal do not need to be reminded that health care systems across the world are facing an ever-increasing burden associated with the burgeoning prevalence of heart failure.1 For the individual, the development of heart failure brings with it very clear detrimental impacts on life-expectancy and quality of life. However, and again as the readers of this Journal will recognize, the identification of patients with the heart failure syndrome is fraught with difficulty, often leading to delays in diagnosis and, for individuals with heart failure with reduced ejection fraction (HFrEF), delays in the application of evidence-based therapies. Importantly, evidence-based treatment for heart failure applies very largely to those with HFrEF. For the estimated 50% of patients with heart failure in the context of preserved left ventricular (LV) ejection fraction (HFpEF), no specific evidence-based treatments have been identified, with international guideline documents recommending appropriate management of the underlying aetiological factors.2 The identification of patients most likely to develop heart failure provides the potential opportunity to delay or prevent the development of the heart failure syndrome. Current European Society of Cardiology guidelines recommend the application of angiotensin-converting enzyme (ACE) inhibitors and beta-blockade in individuals with known LV systolic dysfunction, irrespective of symptoms, and the appropriate treatment of hypertension, dyslipidaemia, obesity, and dysglycaemia to reduce the risk of the development of heart failure.2 The very widespread application of primary percutaneous coronary intervention in ST-elevation myocardial infarction, and of ACE inhibition, beta-blockade, and statins in patients with any acute coronary syndrome, to a large extent reflects this desire to prevent future events, including heart failure. Numerous studies have addressed the potential for a huge variety of biomarkers to identify those individuals at the highest risk of developing heart failure. In the setting of acute coronary events, the natriuretic peptides have well-established value in this regard, and the potential additive predictive value of many other biomarkers has been examined.3, 4 However, it might be argued that the ‘blanket’ application of secondary prevention therapies in this setting makes further risk stratification less (cost) effective. With regard to ‘community-based’ rather than hospitalised patients, identification of those at highest risk of developing heart failure has explored a number of strategies largely including the use of circulating natriuretic peptide levels to guide intervention, with some evidence of success in reducing incident heart failure.5 In the current issue of the Journal, Campbell and colleagues describe the utility of measurement of amino-terminal pro-B-type natriuretic peptide (NT-proBNP) levels in identifying patients at risk of future development of heart failure.6 The authors report age- and sex-specific cut-off levels for circulating NT-proBNP which allow identification of such individuals. The results suggest the potential for the utility of a single, and of serial, measures of NT-proBNP for prediction of 5-year risk of incident heart failure, with sensitivity exceeding 70% and specificity in the range of 45–70%. While interesting, and potentially encouraging, the results merit close inspection. First of all, the concept of the use of age- and sex-specific cut-off values for circulating natriuretic peptides as biomarkers of risk is not new; indeed, my own group described such differences at over 10 years ago.7 Secondly, the authors have reported similar predictive value for incident heart failure for circulating NT-proBNP as compared to a 7-element clinic risk score (baseline age, log body mass index, diabetes, myocardial infarction, obstructive sleep apnoea, smoking status, and serum NT-proBNP quintile). The authors go on to report the sensitivity and specificity of NT-proBNP for incident heart failure, but not of the risk score. To my way of thinking, the relevance of this lies in the fact that these patients were members of a health care insurance scheme, and as such, it is likely that details of six of the seven parameters used in the clinical model (all except NT-proBNP) would be on record. With similar predictive value, the added value of NT-proBNP is unclear. In particular, the cost-effectiveness of the authors' approach has not been presented. With NT-proBNP measured in 3842 participants at baseline, and in 3053 and 2284 individuals at a second and third visit, the identification of 162 incident heart failure events over a median of 4.5 year follow-up clearly comes at considerable cost. Further to this, and importantly, of the 162 new cases of heart failure, only 53 (33%) were of HFrEF, for which evidence-based treatments are available. What the clinicians could have done about those individuals whose incident heart failure was valvular (n = 36, 22%) or HFpEF (n = 73, 45%) is not clear, beyond what the current guidelines recommend namely management of risk factors.2 Would it be appropriate that for patients without LV systolic dysfunction or heart failure, we amend their cardiovascular management based upon a single measurement of natriuretic peptide, with cut-off levels which will change as the patient ages? Would we treat patients with peptide levels just below the cut-off any differently to those just above? A number of years ago we identified that natriuretic peptide measurement, when added to consideration of simple clinical parameters (abnormal electrocardiogram and history of ischaemic heart disease) reduced by a factor of six the number of patients needed to undergo echocardiography in order to identify hitherto unknown LV systolic dysfunction.8 The current analysis from Campbell and colleagues adds to the body of knowledge illustrating the potential use of natriuretic peptides in risk stratification in patients with, or at risk of, cardiovascular disease. The authors have presented a well conducted study; comparison of the NT-proBNP levels with a model without NT-proBNP quintiles would be of value; further, their cost-effectiveness analysis is awaited with great interest. Conflict of interest: none declared.