Making best use of NT-proBNP to test for heart failure

Moira is a 84-year-old with frailty, chronic kidney disease (CKD) and a BMI of 19kg/m2. She was diagnosed with atrial fibrillation (AF) over a decade ago and had an echocardiogram in 2021 that did not show evidence of heart failure. However, she has become more easily tired in recent weeks and has some ankle oedema. You check her NTproBNP and it is 620ng/L. Her ECG shows that Moira is in AF with a heart rate of 105bpm. 

Steve is a 61-year-old man with a history of hypertension treated with ramipril and indapamide. He has a BMI of 37kg/m2. Steve has been getting more quickly short of breath when he takes his dog for a daily walk and has also new ankle oedema. Steve’s NTproBNP is also 620ng/L. 

How would you interpret this same result in Steve and Moira? 

In the UK, NICE last updated their heart failure guidance in 2025 and continue to recommend that we use a NT-proBNP threshold of 400ng/L, above which anyone with suspected heart failure should be referred for assessment. This is substantially higher than the cut-off of 125ng/L recommended in European Society of Cardiology (ESC) guidelines. So, how good are natriuretic peptide tests for identifying people with heart failure and what is the rationale behind the different thresholds? 

DIAGNOSE-NP was a project that looked at patterns of NT-proBNP test use in primary care to help answer this question. Within the cohort over a quarter of a million people had a natriuretic peptide test done in primary care, of whom around 10% were diagnosed with heart failure in the next six months. The results showed what an effective rule-out test NT-proBNP is; the negative predictive value of NTproBNP was excellent at both the ESC (98.9%) and NICE (97.7%) thresholds. Using the NICE threshold means far fewer patients are referred for specialist assessment, though will miss a small proportion of people with heart failure. Given the long waiting times in many parts of the UK for both echocardiography and cardiology clinics, using the NICE threshold helps prioritise those most likely to have heart failure for referral. 

However, we know lots of factors affect NT-proBNP levels, including factors linked to higher NT-proBNP (e.g. age, infection, CKD or AF) and others linked to suppressed NT-proBNP (e.g. increasing BMI and RAAS medications). How can we take account of these factors when we interpret NT-proBNP levels? Overall, NT-proBNP is a less accurate test for heart failure in people with with AF (area under the curve 0.877) compared to without AF (0.743) and levels are typically higher in people with AF, even if they don’t have heart failure. To take account of this, we could use higher cut-off levels for a positive test in people with AF, such as using a threshold of 660ng/L. NT-proBNP levels tend to rise as we get older so thresholds could also take account of age, such as having the lower ESC threshold of 125ng/L in people aged <50 years, increased to ≥250ng/L for people aged 50-74 and ≥500ng/L for the ≥75 years. In contrast, for people with a BMI >30kg/m2, a recent study found a lower threshold of 150ng/L was better than 300ng/L. Reflecting this, the ESC Heart Failure Association suggest reducing NT-proBNP cut-offs by 25% in people with a BMI between 30 and 35kg/m2, increasing to a 40% reduction in people with a BMI >40kg/m2. 

In this context, Steve’s NT-proBNP result looks more concerning; it is likely to be suppressed by his BMI and his blood pressure treatment, whilst it is almost three-times above the threshold for his age. He certainly needs referral for a heart failure assessment. In contrast, Moira’s result is likely to be explained by her AF, CKD and age and referral may not be necessary. Deciding who to refer will always require clinical judgement and we need to consider how likely we think heart failure is based on our clinic assessment as well as the NT-proBNP result alone. 

The DIAGNOSE-NP studies showed that use of NT-proBNP has increased over time in primary care in the UK, but despite this three-quarters of patients with a new diagnosis of heart failure had not had a prior NT-proBNP test done. This suggests we might be able to use NT-proBNP testing much more widely to improve the early diagnosis of heart failure and start people on treatments that can improve their quality of life and help them live longer. As we use NT-proBNP testing more, trying to take account of the wider patient context will make sure we refer those most likely to have heart failure so that we can make the best use of secondary care resources. 

If you are looking for a Cardiology Update, register for our on demand Primary Care Cardiology and ECG Course where we discuss the up-to-date management of common cardiovascular conditions, including ECG analysis and interpretation.