The research, published in PLOS Medicine, looked at over 155,000 people who had an NT-proBNP blood test ordered by their GP between 2004 and 2018. NT-proBNP helps GPs decide whether someone with symptoms such as breathlessness or tiredness might have heart failure and should be referred for specialist assessment. Currently, there is a single threshold used to decide whether or not an NT-proBNP is elevated and usually only those with a positive result are referred on for further assessment.
The study used anonymised electronic health records from the Clinical Practice Research Datalink (CPRD), which includes data from GP practices across England, linked with hospital and mortality records. It forms part of the DIAGNOSE-NP project, supported by the National Institute for Health and Care Research (NIHR) through the Oxford Applied Research Collaboration (ARC) and the Community Healthcare Medtech and In Vitro Diagnostics Cooperative.
What the researchers found
AF and heart failure frequently co-exist and within those assessed in the study, around one in ten people undergoing an NT-proBNP test (17,400) had AF. Additionally, people with AF were more than twice as likely to be diagnosed with heart failure within six months of testing as those without AF (24% compared with 8%).
However, people with AF often had higher NT-proBNP levels even when they did not have heart failure. This is because the irregular, fast rhythm in AF can also raise NT-proBNP levels in the blood above the current threshold to refer for heart failure, which makes it harder to interpret the results.
As a result, the test was less accurate in people with AF. It could still help rule out heart failure when levels were low but this was unusual. The current referral threshold would mean most patients need referring for further testing of heart failure– putting pressure on secondary care services and leading to more false positive results.
The researchers found that using a higher cut-off point for people with AF (for example, raising the referral threshold from 125 to 660 pg/mL) could reduce false positives by around a quarter, meaning fewer unnecessary referrals. However, they caution it might also mean a small number of true cases of heart failure are missed or diagnosed later.
Why this matters
Lead author Dr Nicholas Jones, from the Nuffield Department of Primary Care Health Sciences, explained:
“These findings help provide more granular detail regarding the extent to which NT-proBNP levels are affected by atrial fibrillation. The NT-proBNP test still has an important role, but GPs need to interpret results in the context of AF, where significantly higher readings are common even without heart failure. Other factors are also relevant when interpreting the result, including age, kidney function and body mass index. Our results suggest that adjusting the referral threshold could make testing more efficient and reduce unnecessary referrals, though clinicians should be aware this would also mean some patients who do have heart failure would be missed.”
Looking ahead
The authors highlight that the data did not always show whether patients were in AF at the exact time of their test, and that future studies should consider how factors such as age, weight, kidney function, and medication might affect results. However, the findings could help shape future guidance on how NT-proBNP test results are used to diagnose heart failure in people with AF, particularly where secondary care resources such as access to echocardiogram, are limited. They suggest further research could further explore whether using different NT-proBNP thresholds for people with AF as well as other co-morbidities could improve diagnosis and make better use of NHS testing and specialist resources.
By analysing data from over 155,000 people tested in primary care, this study provides one of the clearest pictures yet of how NT-proBNP performs in people with AF, offering valuable evidence to support more accurate and efficient diagnosis of heart failure in everyday practice.
Read the full paper, 'NT-proBNP testing for heart failure diagnosis in people with atrial fibrillation: A diagnostic accuracy study', in PLOS Medicine.