BNP vs NT-proBNP: The Heart-Failure Test Explained
Peptide biology
The natriuretic peptides are the body's own answer to the renin–angiotensin–aldosterone system — and medicine has extracted far more value from measuring them than from administering them. Three peptides built around the same conserved seventeen-amino-acid ring split into two entirely different careers: two of them defend the circulation against volume overload, and the third quietly runs long-bone growth. Along the way they produced one of the most-ordered blood tests in medicine, and a string of failed heart failure drugs.
BNP vs NT-proBNP in one line: both come from the same prohormone, but only one of them is the hormone — the other is the inert leftover, and that difference is why the two tests do not move together. This article is about the test. For the underlying biology of the whole family — the shared ring, the cyclase receptors and the counterbalance to RAAS — see Natriuretic peptides (ANP, BNP, CNP) explained.
Three peptides, one shared ring
Every natriuretic peptide is made as a preprohormone and processed down to a mature form carrying an obligatory seventeen-amino-acid disulfide ring at the C-terminus. Break that ring and the activity disappears — it is the family's defining structure, not decoration.
- ANP — 28 residues, released mainly from atrial granules in response to stretch, with a half-life of roughly two minutes in humans.
- BNP — 32 residues. It is present in atrial granules too, but reaches its highest levels in ventricles under stress, where it is controlled at the level of transcription rather than stored ready-made. Its clearance is biphasic; the slow component is around twenty minutes.
- CNP — 22 residues, first isolated from pig brain in 1990, now known to come mainly from endothelium and, in much higher concentrations, from chondrocytes.
The processing step for BNP is more contested than most reviews let on. Both furin and corin cleave proBNP in vitro, at different positions, and O-glycosylation near the cleavage site blocks processing entirely — so a portion of circulating proBNP never becomes anything. Corin's role in cutting proANP, by contrast, is firmly settled. This is the same precursor-processing logic that governs insulin and glucagon.
The receptors are not GPCRs
This is the detail most summaries skip. Natriuretic peptide receptors are not G-protein-coupled receptors at all. NPR-A and NPR-B are particulate guanylyl cyclases: a single protein chain that binds the peptide outside the cell and manufactures cyclic GMP inside it. There is no G protein, no second messenger cascade to get to the second messenger.
The selectivity was worked out in 1992 and has held up. For cyclic GMP production through NPR-A the order is ANP at least as potent as BNP, and both far ahead of CNP; through NPR-B, CNP dominates. NPR-C is the odd one out — it has only 37 intracellular amino acids, no cyclase domain, and its main job is to internalise and destroy ligand. Mice without it clear ANP roughly two-thirds more slowly, run mildly hypotensive, and grow elongated long bones.
The second clearance route is neprilysin, and its substrate preference is the key to everything that follows: CNP and ANP are cleaved readily, BNP poorly. The reason is chemical. Neprilysin cuts ANP and CNP at a cysteine–phenylalanine bond that opens the ring and destroys the peptide; in human BNP the equivalent site is a methionine–valine bond it handles about ten times less efficiently. Compare the resulting picture with the enzyme's own article, and note that even so, neprilysin inhibitors fail to block BNP degradation by human kidney membranes — the dominant human BNP-degrading protease has not been identified.
What they actually do
Activating NPR-A produces renal sodium and water excretion, vasorelaxation, direct antagonism of the renin–angiotensin–aldosterone system including suppression of adrenal aldosterone production, and one mechanism that is genuinely surprising: increased endothelial permeability.
CNP's job is different enough to look like a different family. It drives endochondral bone growth through NPR-B, and human genetics makes the dose–response unusually clean: loss-of-function mutations in the receptor gene cause acromesomelic dysplasia and, in carriers, short stature, while gain-of-function mutations that leave the cyclase constitutively producing cyclic GMP cause extreme tall stature without skeletal deformity. It is not cardiovascularly silent — endothelial CNP acting on pericyte NPR-B contributes about 10 mmHg of blood pressure control — but it acts locally rather than as a circulating hormone, more like endothelin than like ANP.
BNP vs NT-proBNP: the blood test built on a leftover
proBNP is cut into a 32-residue hormone and a 76-residue fragment, NT-proBNP, which does nothing whatsoever. That inertness is the point. NT-proBNP is not an NPR-C ligand and not a neprilysin substrate, so it persists roughly six times longer than BNP and accumulates to higher concentrations. The useless half is the more reliable measurement.
The evidence base is unusually solid. In the emergency-department study that established the test, BNP at 100 pg/mL identified heart failure with 90% sensitivity and 76% specificity and an area under the curve of 0.91, outperforming both clinical judgement and the Framingham criteria. The pooled NT-proBNP analysis that followed produced age-stratified rule-in thresholds — 450 pg/mL under 50, 900 pg/mL from 50 to 75, 1,800 pg/mL above 75 — and a single rule-out threshold of 300 pg/mL with a 98% negative predictive value.
Worth noting precisely: the 2022 US heart failure guideline recommends the test at its highest class and evidence level but publishes no numeric cut points. The numbers people quote come from the European guideline and from those pooled studies, not from the American one.
Everything else that moves the number
The confounders are not footnotes. In a UK primary-care study of 155,347 people, those with atrial fibrillation but no subsequent heart failure diagnosis had a median NT-proBNP of 724 pg/mL against 111 pg/mL in sinus rhythm — roughly six and a half times higher — and at the standard 125 pg/mL rule-out threshold, specificity in atrial fibrillation collapsed to 13%. Age and renal impairment push in the same direction.
Obesity pushes the other way. In Framingham participants without heart failure, adjusted mean BNP in men fell from 21.4 pg/mL when lean to 12.7 pg/mL when obese. The tempting explanation is faster clearance through NPR-C receptors on adipose tissue, and NPR3 expression genuinely is higher in obese human myocardium. But NT-proBNP, which NPR-C does not bind, falls in lockstep — and a clearance-only mechanism cannot do that. Human heart tissue from obese organ donors with normal ventricular function shows reduced transcription of both peptide genes. The heart is making less, and clearing more; the relative weighting is unresolved.
The drug that split the two markers apart
Sacubitril inhibits neprilysin. BNP is a neprilysin substrate; NT-proBNP is not. So in the trial that established the combination in 8,442 patients with reduced ejection fraction — a 20% relative reduction in cardiovascular death or heart failure hospitalisation, and a 16% reduction in all-cause mortality — NT-proBNP fell while BNP rose. BNP doubled in 18% of patients and tripled in 6% within ten weeks.
The European product information states the consequence outright: BNP is not a suitable heart failure biomarker in patients on sacubitril/valsartan, and NT-proBNP is the more suitable one. The US label carries only the neutral observation. Be careful how far you take this, though: a later analysis found BNP retained its prognostic accuracy on treatment. It is the diagnostic thresholds that stop meaning what they used to, not the signal itself.
Why giving these peptides has mostly failed
Nesiritide, recombinant human BNP, was approved in 2001 for acute decompensated heart failure. Two 2005 meta-analyses raised worsening renal function and a 74% relative increase in 30-day death that fell just short of significance. The definitive trial randomised 7,141 patients, of whom 7,007 received drug: death or rehospitalisation at 30 days was 9.4% against 10.1%, and the dyspnoea improvements, though nominally significant, missed the prespecified threshold. Symptomatic hypotension roughly doubled. FDA formally withdrew the approval in June 2020.
Ularitide — synthetic urodilatin, the four-residue-longer renal form of ANP from the same prohormone — was tested in 2,157 patients and produced a hazard ratio for cardiovascular death of 1.03. Carperitide, recombinant ANP, is marketed only in Japan, where observational analyses have suggested harm and the effect vanishes when the two largest database studies are removed. Trial design in this space has not been the limiting factor; the biology has.
The exception proves the point. Vosoritide, approved in 2021 and expanded in 2023 to children of any age with open growth plates, is a 39-residue CNP analogue for achondroplasia. It binds NPR-B and antagonises the overactive FGFR3 signalling that limits growth, adding about 1.57 cm per year over placebo. It works on the branch of the family that was never about the heart.
Frequently asked questions
Is BNP a "brain" peptide?
Only historically. It was first isolated from pig brain, which is where the B came from; it is now more usually written as B-type natriuretic peptide. CNP was also first found in brain and is likewise not a brain hormone.
Which is better, BNP or NT-proBNP?
Neither is universally superior for diagnosis, but they are not interchangeable and their reference ranges are different by an order of magnitude. NT-proBNP is the safer choice on a neprilysin inhibitor.
Why is my NT-proBNP high if my heart is fine?
Age, atrial fibrillation and reduced kidney function all raise it substantially, as do pulmonary embolism, sepsis and critical illness. A raised value is a reason to look further, not a diagnosis; interpretation belongs with a clinician who has the rest of the picture.
Are natriuretic peptides related to vasopressin?
Only by opposition. Vasopressin retains water; the natriuretic peptides get rid of it. They are structurally unrelated and use entirely different receptor machinery.
References
- Potter LR. Natriuretic peptide metabolism, clearance and degradation. FEBS J 2011;278(11):1808–1817. PMC4495883
- Maisel AS, Krishnaswamy P, Nowak RM, et al. Rapid Measurement of B-Type Natriuretic Peptide in the Emergency Diagnosis of Heart Failure. N Engl J Med 2002;347(3):161–167. nejm.org
- Januzzi JL, van Kimmenade R, Lainchbury J, et al. NT-proBNP testing for diagnosis and short-term prognosis in acute destabilized heart failure (ICON). Eur Heart J 2006;27(3):330–337. academic.oup.com
- McDonagh TA, Metra M, Adamo M, et al. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J 2021;42(36):3599–3726. academic.oup.com
- Entresto (sacubitril/valsartan) EPAR product information, sections 4.4 and 5.1. European Medicines Agency. ema.europa.eu
- O'Connor CM, Starling RC, Hernandez AF, et al. Effect of Nesiritide in Patients with Acute Decompensated Heart Failure (ASCEND-HF). N Engl J Med 2011;365(1):32–43. nejm.org
Informational only — not medical advice. This article describes published research, guidelines and regulatory documents; it is not guidance for interpreting your own test results or managing any condition. Consult a qualified healthcare professional. Intended for readers 21+.