Abstract & Commentary
By Michael H. Crawford, MD
Sources: deFilippi CR, et al. Dynamic cardiovascular risk assessment in elderly people. J Am Coll Cardiol. 2010;55:441-450; Berger R, et al. N-terminal pro-B type natriuretic peptide-guided, intensive patient management in addition to multidisciplinary care in chronic heart failure. J Am Coll Cardiol. 2010;55:645-653.
Natriuretic peptide levels (BNP, NT-proBNP) are of prognostic value in general populations, but whether they add information to other known risk factors for cardiovascular outcomes is less clear. Thus, deFilippi et al studied almost 3,000 older adults free of heart failure by measuring NT-proBNP at baseline and then 2-3 years later. The endpoints were new-onset heart failure and cardiovascular death. After adjustment for confounders, they sought an association between initial NT-proBNP and changes in NT-proBNP and the primary endpoints.
Results: The highest quintile of NT-proBNP (> 268 pg/mL) was independently associated with new-onset heart failure (HR = 3.1, 95% CI: 2.5-3.8) and CV death (HR = 3.0, CI: 2.4-3.9), as compared to the lowest quintile (< 48 pg/mL). The cut-point for increased risk was a NT-proBNP level of 190 pg/mL. A change in NT-proBNP > 25% was associated with higher or lower risk, depending on the direction of change, compared to those with unchanged levels. The authors concluded that NT-proBNP levels independently predict new-onset heart failure and CV death in older adults, and that changes over time reflect a change in these risks.
Berger et al investigated whether heart failure patient management using NT-proBNP levels (BM) was superior to multi-disciplinary (MD) or standard management (SM). They randomized 278 hospitalized patients to BM vs. MD vs. SM. MD included consultations from a heart-failure specialist and homecare by a heart failure-trained nurse. BM used NT-proBNP levels to aggressively up titrate medications. The end-points were heart-failure rehospitalization and death. At one year, more of the BM group was on triple therapy (beta blocker, spironolactone, ACEI/ARB) at target doses. BM reduced heart-failure hospitalization days (BM 488, MD1254, SM 1588, p < .001). The combined endpoint of heart-failure rehospitalization and death was lower in the BM group (37%, MD 50%, SM 65%, p < .05). Death was similar in BM and MD, 22%, but higher in SM, 39% (p < 0.02). The authors concluded that heart-failure management, using NT-proBNP levels to guide therapy, was superior to multi-disciplinary care and standard management for reducing heart-failure recurrence and death.
These two studies support a role for natriuretic peptide (NP) level measurements in the primary and secondary prevention of heart failure. In the elderly group studied by deFilippi et al, NP levels were independent of age, traditional risk factors, ECG, and echocardiographic measures in predicting heart failure and CV death. This is important information since traditional risk factors and biomarkers are known to be less predictive of outcomes in the elderly (i.e., lipids, CRP). Thus, in patients at high risk for heart failure (post MI, hypertension), NP may signal an opportunity to intensify therapy to prevent heart failure from developing. It would not make sense to test everyone but to focus on those most likely to develop heart failure.
NP levels reflect measure ventricular wall stress, which, when elevated, can lead to heart failure. Although all patients with heart failure should be put on all medications known to improve symptoms and survival, and they should be titrated up to at least the doses used in their respective trials, this does not happen often, as was demonstrated in the standard management group of Berger's trial. Thus, NP levels allow one to focus resources on the highest-risk patients who need this up titration and maximization of medical therapy. Also, maximum therapy can be dangerous in some patients with lower blood pressure and other comorbidities. By focusing on high-risk patients, medicine maximization can be accomplished more safely. Whether BNP would work as well as NT-proBNP is not known, but the principle should be the same.