Can We STOP Heart Failure? Using a Screening BNP to Guide Care


Written By

Regina Arellano, PharmD
Christie Schumacher, PharmD, BCPS, BC-ADM, CDE

Reviewed By

Dawn Fuke, PharmD, BCPS
Oralia Bazaldua, PharmD, BCPS


Ledwidge M, Gallagher M, Conlon C, et al. Natriuretic peptide-based screening and collaborative care for heart failure. The STOP-HF randomized trial. JAMA. 2013;310: 66-74.

A 65 year old patient presents to your clinic with an A1C of 7.2%, a blood pressure of 146/80 mmHg and a LDL cholesterol of 103 mg/dL. Should you order a BNP level to determine the patient’s risk of developing heart failure?

Brain-type natriuretic peptide (BNP) is a biomarker which has an established role in the diagnosis of heart failure (HF).1 It is released by the cardiomyocytes in response to volume overload and myocardial stretch. The ACCF/AHA guidelines state that BNP is useful for establishing the prognosis of HF and is also effective to assess disease severity (Class I, Level of Evidence A).1  The usefulness of serial measurements of BNP to reduce hospitalizations or mortality in patients with HF is not well established.  Is BNP a useful biomarker to identify and treat patients at an increased risk of developing HF? The STOP-HF trial is the first prospective randomized trial to study the effectiveness of a BNP-guided collaborative care model in reducing the risk of newly diagnosed HF, asymptomatic left ventricular dysfunction, and emergency cardiovascular hospitalizations.2

The St Vincent’s Screening to Prevent Heart Failure (STOP-HF) study was a prospective, randomized trial conducted in Ireland to determine the efficacy of a screening program using BNP and a collaborative care model in an at-risk population to reduce  the rate of new onset HF.  The study enrolled patients older than 40 years (mean age = 64.8 [SD, 10.2]) with one or more cardiovascular risk factor.  Most patients had at least 2 risk factors with hypertension being the most prevalent. Patients were excluded if they had evidence of left ventricular systolic dysfunction, symptomatic HF, or a diagnosis that would limit survival over the study period.

A total of 1374 patients were randomized 1:1 to a control group, which received routine care through their primary care provider (PCP), or an intervention group, which received a BNP-driven collaborative care model of care.  Group assignment was not blinded to study participants or to clinical investigators. All patients in the study underwent BNP blood testing at the start of the study and annually thereafter. Patients in the intervention group with a BNP value of 50 ng/L or higher were referred to the cardiology service for an echo and ongoing management.  This included optimal risk factor management with the most appropriate therapy and coaching by a specialist nurse.  The nurse emphasized the importance of medication adherence and lifestyle behaviors. Patients in the control group were referred to a cardiologist only if requested by the PCP.

The primary endpoint was prevalence of left ventricular dysfunction (asymptomatic left ventricular systolic dysfunction or asymptomatic left ventricular diastolic dysfunction) with or without newly diagnosed HF. Left ventricular diastolic dysfunction (LVD) was added as an endpoint as enrollment in the study was slower than anticipated. The secondary endpoints included emergency hospitalization for any major adverse cardiovascular event including arrhythmia, transient ischemic attack (TIA), stroke, myocardial infarction, pulmonary embolism, or HF.  Patients were followed for a mean of 4.2 [SD, 1.2] years.

Table 1. Primary Outcome Analysis


N = 677

N = 697

Odd Ratio
(95% CI)

P - value

All Patients

HF or LVD, n (%)

59 (8.7)

37 (5.3)

0.55 (0.37 – 0.82)


Asymptomatic LVD, n (%)

45 (6.6)

30 (4.3)

0.57 (0.37 – 0.88)


Heart failure, n (%)

14 (2.1)

7 (1)

0.48 (0.20 – 1.20)


Patients with BNP > 50 pg/mL

HF or LVD, n (%)

44 (18.7)

25 (9.5)

0.44 (0.26 – 0.73)


Asymptomatic LVD, n (%)

32 (13.6)

20 (7.6)

0.47 (0.27 – 0.83)


Heart failure, n (%)

12 (5.1)

5 (1.9)

0.43 (0.15 – 1.19)



The unadjusted incidence rates of emergency hospitalization for MACE were 40.4 per 1000 patient-years in the control group and 22.3 per 1000 patient-years in the intervention group (IRR, 0.60; 95% CI, 0.45-0.81). This finding was driven by a reduction in stroke or TIA (0.53 [0.29-0.96]) likely due to a 10 point reduction in systolic blood pressure (p < 0.001) in the intervention group. These data suggest that BNP might be an effective screening tool and that collaborative care may be an effective means of reducing hospitalizations in high-risk patients.

Since the study was conducted in Ireland, its generalizability could certainly be questioned.  We need a second study in a more diverse patient population to confirm the findings to know with certainty that BNP-guided care improves outcomes.  Moreover, the collaborative care model will be difficult to implement in the United States.  In integrated health settings, such as the Veteran’s Administration or staff-model HMOs, the intervention could be adopted. But in most settings where patients receive their care (in the US at least), payment for coaching services (by nurses or pharmacists) is not the norm.  However, the study findings clearly suggests there is an expanded role for pharmacists in HF PREVENTION. Patients are unlikely to remain adherent to preventative therapies without ongoing follow-up and coaching. Since the study did not discern which component of the intervention (collaborative care or BNP testing) was responsible for the improvement, it is difficult to tell whether BNP testing alone would improve outcomes.

Another potential limitation — the primary endpoint was changed midway through the study. Further, more than half of the patients assessed for eligibility were excluded due to inability to meet enrollment criteria.   Thus, it is unclear how many patients may actually benefit from BNP-guided collaborative care.

An estimated 5.7 million people in the United States have a diagnosis of HF.3,4  It is the primary cause of death in more than 55,000 people each year.  The medical costs, including the cost of health care services, medication, and loss of productivity associated with HF are estimated at $34.4 billion.5  Adopting a BNP-guide approach to care would be an additional cost.  But if the study results are valid and this intervention reduces hospitalizations by 20-50% over a 4-year period, it could be a wise investment of resources.

Serial BNP testing does not influence treatment selection in patients with heart failure.  But is it a viable screening tool to determine how we manage patients at risk for heart failure?  Should we spend money on a lab test to determine who is at highest risk for developing HF?  Or should we spend time with EVERY patient with cardiovascular risk factors educating them on the importance of adherence to appropriate treatments to prevent the development of HF?  Would early identification of high-risk patients lead to more pharmacist-physician collaboration?  What are your thoughts?  Should we adopt this approach to care?



Nice job summarizing key aspects of STOP-HF. This study was also flawed in that the study investigators were not blinded. Furthermore, this integrated approach would be difficult to duplicate in a wide variety of settings and I don't think BNP monitoring, alone, would have the same results. An interesting idea that warrants further study.