Portal de Periódicos da CAPES

Understanding and making good use of changes in natriuretic peptide levels in chronic heart failure

2015; Elsevier BV; Volume: 17; Issue: 8 Linguagem: Inglês

10.1002/ejhf.322

1879-0844

Roberto Latini, Serge Masson,

Cardiac pacing and defibrillation studies

This article refers to ‘Changes in N-terminal pro brain natriuretic peptide levels and outcomes in heart failure with preserved ejection fraction: an analysis of the I-Preserve study’, by P.S. Jhund et al., published in this issue on pages XXX–XXX. The background of evidence supporting the use of levels of natriuretic peptides to help the cardiologist in the management of patients with heart failure (HF) goes back several decades. However, it was only ∼10 years ago that the concept of changes over time of natriuretic peptide levels was developed and tested in the first large population of >4000 patients with heart failure with reduced ejection fraction (HFrEF) of the Valsartan Heart Failure Trial (Val-HeFT).1 Several smaller studies have appeared in the literature since then, so that serial changes of natriuretic peptides in the management of HF have become a key issue of debate. It is worth directing the interested reader to the pros and cons debated in the series ‘Controversies in Cardiovascular Medicine’.2, 3 The study by Jhund et al.4 is worthy of a comment mainly due to two features, the patients have HF with preserved EF (HFpEF) and the sample is adequately sized to draw conclusions with confidence. Indeed, it is always a ‘scientific pleasure’ to comment on results of solid large studies even if the 2612 patients are a subgroup of the larger population enrolled in I-PRESERVE.5 Before reading the paper it is useful to remember that post-hoc analyses will never provide real evidence but rather hypotheses to be verified in prospective studies, collectively identified under the label of ‘natriuretic peptide-guided therapy’, which are not the subject of the present Editorial. The authors present results of three different types of analyses of the changes in NT-proBNP over time: continuous, by quartiles, and by categories (i.e. low–low, low–high, high–low, and high–high). These different analyses offer different ways of looking at the data, leading to complementary but not identical information which is summarized below. The cubic splines and their 95% confidence intervals clearly show that (i) the hazard ratio for cardiovascular death or heart failure hospitalization increases steeply from no change up to +1000 ng/L and then levels off; and (ii) decreases of NT-proBNP are associated with a lower but not statistically significant hazard ratio. Based on this last finding, one might wonder whether decreasing plasma levels of NT-proBNP is a clinically worthwhile target. The categorical analysis reassures the conscientious reader since patients in the high–low category do better than those in either the low–high or high–high category (see figure 3 of Jhud et al.4). The patients with the smallest 6-month changes in NT-proBNP, that are in quartiles Q2 and Q3 were younger, had fewer co-morbidities, less AF, and lower prescription rates of beta-blockers, digoxin, diuretics, and spironolactone, but more frequently were in NYHA class III or IV. Outcomes of Q2 and Q3 were also significantly milder than those of Q1 and Q4. This suggests that larger changes in NT-proBNP, no matter whether they are increases or decreases, carry a worse prognosis. These counterintuitive findings are probably driven by higher baseline NT-proBNP levels associated with the largest changes over time either way. Although AF was present in more than a quarter of the patients and in one-fifth at ECG exam, it did not significantly influence the relationship between NT-proBNP changes and outcomes. Given the well known prognostic role of AF, this finding cannot be easily explained. On the other hand, AF is significantly more frequent in patients with greater changes (decreases or increases) in NT-proBNP, suggesting that it may be a marker of haemodynamic/clinical instability. An increase over time in NT-proBNP carries a worse outcome in HF, no matter whether EF is decreased or preserved. This suggests that changes in NT-proBNP, possibly resulting from different pathophysiological mechanisms, lead to similar outcomes. Generally speaking, it appears that the changes in NT-proBNP are not specific for either HFrEF or HFpEF, although in general its concentrations are lower in HFpEF. The differential effect of ARBs on natriuretic peptides is also worth commenting upon. While in patients with HFrEF valsartan consistently reduced both BNP and NT-proBNP,6, 7 in the patients with HFpEF of I-PRESERVE irbesartan did not affect NT-proBNP. This could be due to either of two reasons: (i) the decrease in natriuretic peptides by ARBs is not a class effect; or (ii) ARBs do not affect natriuretic peptides in HFpEF, which is consistent with their lack of effect on outcomes in this disease. The scarcity of data on the effects of different ARBs on changes of natriuretic peptides in placebo-controlled studies in HF does not allow a class effect to be either proved or disproved. Indeed, no robust data on irbesartan are available in the literature, while the large Candesartan in Heart failure-Assessment of Mortality and Morbidity (CHARM) programme did not generate data on effects of candesartan on natriuretic peptides levels over time. When dealing with HFpEF, the lack of data is even more impressive. All these findings support the novelty of the data obtained from I-PRESERVE.4 The vast majority of patients have stable NT-proBNP levels over 6 months (83%), as can be appreciated looking at figure 3 of the study by Jhud et al.4 In Val-HeFT, the fraction of patients stable at 4 months was almost identical: 86%. This observation corroborates the recommendation in these two large-scale controlled trials of enrolling patients who were clinically very stable, already on optimized cardiac therapy. In conclusion, although the analyses presented are novel and allow for different insights into the complexity of HFpEF, NT-proBNP still does not emerge as a straightforward surrogate endpoint in HFpEF. The easy take-home message is that NT-proBNP-guided treatment of HFpEF is still an open area of research, while searching for an effective drug. In fact, the recent data on the new drug LCZ696 suggest that this dual antagonist can decrease NT-proBNP plasma concentrations over 12 weeks when compared with valsartan in 266 patients with HFpEF and high NT-proBNP circulating levels.8 Conflict of interest: none declared.