November 9, 2010
from this study, BNP correctly classified 74% of the patients with an intermediate probability of AHFS.33 When BNP was added to clinical judgment after routine evaluation, the area under the receiver operating characteristic curve (AUC) rose significantly from 0.86 to 0.93 (P0.0001). Similarly, a single-center investigation evaluated the diagnostic utility of NT-proBNP in the ED in 600 patients with dyspnea.31 The AUC rose from 0.90 to 0.96 when NT-proBNP was added to clinical judgment. The authors suggest a single cut point of 300 pg/mL to rule out AHFS, but 2 cut points to rule in AHFS depending on age: 50 years old (450 pg/mL)and 50 years old (900 pg/mL). Subsequent studies suggested even further delineation as follows: (1) either an age-independent cutoff of 900 pg/mL, or (2) the more accurate (but more complex) age-stratified approach of 450/900/1800 for pa- tients aged 50/50 to 75/75 years.34,35 Other smaller studies have also demonstrated the diagnostic utility of BNP and NT-proBNP for AHFS.29,30,36,37
The majority of studies suggest that BNP and NT-proBNP are of equal diagnostic utility. However, subtle differences in patient characteristics may favor one biomarker over the other. BNP and NT-proBNP both can be elevated in patients with renal insufficiency, which is more commonly found in older patients.38,39 Levels of NT-proBNP appear to be more affected by renal function.40 Four studies have directly compared the diagnostic utility of BNP and NT-proBNP.29,36,41,42 Both natriuretic peptides demonstrated similar accuracy in 3 stud- ies, but in 1 study BNP was superior to NT-proBNP.42 The AUC for the diagnosis of AHFS was 0.80 for NT-proBNP and 0.85 for BNP, P0.05. This was mostly a consequence of the lower specificity of NT-proBNP (76%) when com- pared with BNP (91%). In this study, only patients 65 years old were enrolled, suggesting that BNP may be superior in older patients. This finding will need to be confirmed in other studies. The natriuretic peptides are particularly good at ruling out AHFS; the negative likelihood ratio of BNP at 100 pg/mL is 0.13,28 and of NT-proBNP at 300 pg/mL is 0.015.31 However, the positive likelihood ratio of the natriuretic peptides is more limited (3.8 and 3.1, respectively, for BNP and NT-proBNP) because they can be elevated in numerous conditions including sepsis, pulmonary hypertension, older age, renal insufficiency, atrial fibrillation, and pulmonary embo- lism.43–47 Obesity is actually associated with disproportionately low BNP levels.48 Mechanisms that have been postulated for these low BNP levels include reduced peptide synthesis and/or secretion in subjects with obesity; increased expression of natriuretic peptide clearance receptors in adipose tissue; and increased circulating neutral endopeptidases, which are secreted by adipocytes, may contribute to a lesser extent.49 Patients with a history of HF can have chronically elevated BNP or NT-proBNP levels. An elevation above their baseline, or dry weight level, may help identify a patient with AHFS. What constitutes a significant change above the baseline level in any particular patient is uncertain at the present time. Biological variability further complicates this situation. Studies suggest that BNP may need to change by at least 70% and NT-proBNP may need to change by 50% to identify a patient with a diagnostically meaningful change.50–53
The clinical utility and resource utilization of BNP testing were evaluated in a single-center randomized trial of 453 patients with dyspnea in an ED in Switzerland.32 Two hundred twenty-five patients were randomly assigned to a standard diagnostic strategy, and 227 patients were randomly assigned to a standard diagnostic strategy plus BNP measure- ment. In comparison with the standard strategy, BNP testing led to reductions in the number of patients hospitalized (75% versus 85%, P0.008), time to discharge (8.0 days versus 11.0 days, P0.001), cost ($5410 versus $7264, P0.006), and time to treatment (63 minutes versus 90 minutes, P0.03) In a separate analysis from the same trial, the cost-effectiveness of BNP measurement in the ED was maintained at 180 days.54 However, the dramatically different lengths of stay compared with centers in the United States makes extrapolation of these results problematic. Another trial of 500 patients with dyspnea presenting to EDs in Canada randomly assigned 250 patients to a standard diag- nostic strategy and 250 patients to a standard diagnostic strategy plus NT-proBNP measurement.55 The AUC of the emergency physician’s diagnostic accuracy without knowl- edge of NT-proBNP results was 0.83 (95% CI 0.80 to 0.84), which increased to 0.90 (95% CI 0.90 to 0.93, P0.001) with knowledge of NT-pro BNP results. Although there were no clinically meaningful differences in ED or hospital length of stay or costs, there was a significant difference in 60-day rehospitalization and costs favoring the NT-proBNP group. However, randomized trials investigating the use of an initial BNP to aid in diagnostic accuracy or serial BNP levels to dictate therapy in the acute setting found no improvement in diagnostic accuracy or clinically important outcomes such as length of stay, mortality, and readmission.56,57 These random- ized trials do not clearly identify whether the potential improved diagnostic accuracy of natriuretic peptides can lead to more appropriate therapy in a cost-effective manner. Further research, preferably in the way of a multicenter trial, is indicated to address this issue.
In summary, the measurement of BNP or NT-proBNP in the ED patient being evaluated for possible AHFS improves diag- nostic accuracy when compared with standard diagnostic strat- egies. Either BNP or NT-proBNP should be measured in patients in whom there is clinical uncertainty concerning the diagnosis.
Current Therapy: Heterogeneous Presentations Met With Homogeneous Therapy Although dyspnea, the principal symptom in AHFS, is attributed to the common pathophysiologic denominator, increased left ventricular end-diastolic pressure, not all patients have the same etiology or precipitating factor.58,59 Regardless of the baseline cardiac pathophysiology, critical presenting features such as hemodynamic status, presence (or absence) of myocardial ische- mia, and renal dysfunction greatly influence management. Widespread appreciation of this phenotypic variability is lack- ing,60–62 perhaps because AHFS is viewed as a single disease entity rather than as a multifaceted disorder.58
Furthermore, symptoms related to congestion are what prompt patients with AHFS to seek care.63 The current goals of ED therapy are to relieve congestion, balance hemodynam-