Biomarker predicts outcome but no effect on ARNI efficacy

When the findings from the PARAGON-HF were reported in September 2019, many in the heart failure community were disappointed by the findings and almost immediately began the process of drilling down into the data seeking explanations.

The preserved ejection fraction heart failure (HFpEF) trial narrowly missed its efficacy endpoint — a difference of just 7 patients would have demonstrated significant benefit, while at the same time sacubitril/valsartan did demonstrate a benefit in women and among subjects whose ejection fractions were more mid-range than frankly “preserved.”

Sacubitril/valsartan is known to reduce NT-proBNP, which lead Jonathan W. Cunningham, MD, of Brigham and Women’s Hospital, Division of Cardiovascular Medicine, Boston, and colleagues to hypothesize that the key to the puzzling results of PARAGON-HF might be the potential prognostic significance of baseline NT-proBNP.

Their analysis confirmed the prognostic value of NT-proBNP for outcomes — lower is better — however, “screening NT-proBNP did not modify the treatment effect of sacubitril/valsartan compared with valsartan (P interaction = 0.96),” they wrote in JACC Heart Failure.

In PARAGON-HF, NT-proBNP was measured at baseline screening in all 4,796 patients and before, between, and after a sequential run-in with valsartan and sacubitril/valsartan, as well as at 16 and 48 weeks post-randomization in more than 2,700 patients.

At baseline, the median NT-proBNP was 911 pg/ml, and this screening NT-proBNP “was strongly associated with the primary endpoint, total HF hospitalizations and cardiovascular death (rate ratio [RR]: 1.68 per log increase in NT-proBNP, 95% confidence interval [CI]: 1.53 to 1.85; P < 0.001). “

The “relationship was stronger in patients with atrial fibrillation (adjusted RR: 2.33 [95% CI: 1.89-2.87] versus 1.58 [95% CI: 1.42 to 1.75] in patients without atrial fibrillation; P interaction <0.001) and weaker in obese patients (adjusted RR: 1.50 [95%CI: 1.31-1.71] versus 1.92 [95% CI: 1.70 to 2.17] in nonobese patients; P interaction <0.001),” they wrote.

Of note, at 16 weeks, sacubitril/valsartan treatment significantly reduced NT-proBNP by 19% (P <0.001) compared with valsartan, and the reductions were similar in men (20%) and women (18%), as well as in patients with LVEF of 57% or lower (20%) and those with LVEF of more than 57% (18%).

And, those reductions were beneficial, as the “primary endpoint rate was 11.2 (95% CI: 8.7-14.5) per 100 patient-years in the quartile of patients with greatest NT-proBNP decline (>38%), and 15.8 (95% CI: 13.1-19.1) per 100 patient-years in the quartile of patients who had NT-proBNP increased >25%. Adjusted reductions were similar for the components of the primary endpoint, CV death alone (RR: 0.62 per log decrease, 95% CI: 0.50-0.76; P < 0.001) and recurrent HF hospitalizations (RR: 0.62 per log decrease, 95% CI: 0.52-0.72; P< 0.001),” they wrote.

Cunningham and colleagues concluded, “These data validate the prognostic importance of NT-proBNP in this contemporary HFpEF population and support consistent treatment effects of sacubitril/valsartan in reducing NT-proBNP levels overall and in key subgroups of patients with HFpEF.”

But in an editorial accompanying the study, James L. Januzzi Jr., MD, of Massachusetts General Hospital/Harvard Medical School, and Peder L. Myhre, MD, PhD, of Akershus University Hospital, Lørenskog, Norway and the Institute of Clinical Medicine, University of Oslo, Oslo, Norway, said it is a bit more complicated than that when the patients have HFpEF.

“NT-proBNP is a less perfect biomarker in HFpEF than it is in those with reduced LVEF,” they wrote. “Not only is NT-proBNP considerably influenced by key clinical features of HFpEF such as AF, obesity, and renal impairment, but the range of NT-proBNP in HFpEF (lower than in HFrEF) extends down into the normal range in some patients, leaving one to speculate how to judge their risk. Efforts to adjust for such differences in the trial enrollment criteria will probably help overcome some, but not all, of these barriers. Indeed, the 3-fold higher NT-proBNP threshold for AF in the PARAGON-HF trial seems to be an accurate approach for capturing comparable risk as patients in sinus rhythm. It is, however, critically important to be aware of comorbidities and how they may shift (e.g., weight loss, worsening renal function, treatment of AF) may confound interpretation of serial NT-proBNP measurements.”

Januzzi and Myhre suggested that it may not be possible to adjust for the changes they list and cited that impossibility as a likely explanation for the “inconsistency and differences between trials.”

The path forward, they suggested, may be a strategy that enlists a second biomarker to be used with NT-proBNP. “A combined biomarker strategy may also be applicable as an enhanced surrogate endpoint to improve the translation to clinical outcome measures,” they wrote. “There are several biomarker candidates for such strategy, including cardiac troponins and insulin-like growth factor-binding protein-7. Whether this is a path forward or not, it is worth considering next steps to better manage this complex and challenging population of patients.”

  1. In the PARAGON-HF trial of sacubitril/valsartan compared with valsartan in HFpEF, baseline NT-proBNP and change in NT-proBNP over time strongly predicted HF events.

  2. Further studies are needed to identify patients with HFpEF who benefit most from sacubitril/valsartan and to investigate the mechanisms of NT-proBNP reduction with neprilysin inhibition.

Peggy Peck, Editor-in-Chief, BreakingMED™

PARAGON-HF trial was sponsored by Novartis.

Cunningham is supported by Heart, Lung, and Blood Institute (NHLBI) T32 postdoctoral training grant T32HL094301.

Januzzi is supported in part by the Hutter Family Professorship; is a trustee of the American College of Cardiology; has received grant support from Novartis Pharmaceuticals; has received consulting income from Abbott, Janssen, Novartis, Pfizer, Merck, and Roche Diagnostics; and participates in clinical endpoint committees/data safety monitoring boards for Abbott, AbbVie, Amgen, Bayer, Siemens, and Takeda.

Myhre is supported by a research grant from South-Eastern Norway Regional Health Authority and the University of Oslo; has served on advisory boards for Novartis and Novo Nordisk; and has consulted for Novartis, AmGen, and Novo Nordisk.

Cat ID: 3

Topic ID: 74,3,791,730,3,6,192,916,925