Does serum B type natriuretic peptide (BNP or NT-pro BNP) testing and monitoring in patients with heart failure benfit patients and the NHS

Heart failure affects about 900,000 people in the UK. It happens when the heart becomes damaged, for example after a heart attack, and cannot pump blood around the body properly. B-type natriuretic peptide, and another chemical derived from it (collectively referred to as BNP), are hormones secreted by the heart in response to injury and their levels in the blood are raised in people with heart failure. High levels of BNP predict people who will experience a faster deterioration in their health from heart failure and who have a higher risk of serious disease events, such as death or deterioration in health requiring emergency admission to hospital. There is some evidence that people with heart failure do better when BNP is measured regularly over time, so that doctors can adjust patients’ drugs (used to treat heart failure, such as beta-blockers or ACE inhibitors) to try to lower BNP levels. Most of this evidence comes from randomised controlled trials, in which patients are allocated at random (by chance) to receive either BNP-guided treatment or standard treatment without having BNP measured. As part of our proposed project, we will bring together the results from all trials (in effect, averaging and improving the precision of the overall result) to determine whether those patients with heart failure who have BNP-guided treatment experience better outcomes than those who receive standard treatment (STUDY 1).

However, we expect it to be difficult to conclude that the overall finding from these trials should be applied to the general population with heart failure, whether managed by their GPs or hospital doctors, since the patients participating in clinical trials are not representative of the general population with heart failure – they are younger, with more men than women, having only a certain type of heart failure (reduced ventricular ejection fraction) and no other conditions. Also, most trials have been conducted outside the UK, making it difficult to determine whether measuring BNP is cost effective in the UK setting.

Therefore, we propose to supplement the IPD meta-analysis with analyses of a representative group of patients with heart failure in the UK (STUDY 2). We will create this patient group by linking data from the Clinical Practice Research Database (CPRD, which contains patient data from GP practices) and the UK National Heart Failure Audit (NHFA, which contains data on patients admitted to hospital with heart failure), as well as hospital record data and death registry data. These data will allow us to profile how heart failure patients are cared for in the NHS, from diagnosis, through treatment, to outcome. Both databases have a subset of patients who have had BNP values recorded. We propose to match these patients with a similar group of patients who have not had a BNP value recorded (taken from the same databases). We will compare groups with and without BNP measurements for differences in risk of death, hospital admission/readmission and length of hospital stay (for those admitted to hospital), prescribed medications, number of outpatient appointments and patient management. We will also undertake a health economic analysis to determine whether measuring BNP is cost effective in the NHS (STUDY 3).