BNP has diagnostic value in patients with suspected HF
MedWire News: B-type natriuretic peptide has "considerable" value for diagnosing heart failure (HF) in primary care, a meta-analysis reveals.
Despite this promising finding, however, the study authors stress that elevated BNP at currently recommended cutoff levels cannot be used to make a reliable diagnosis of HF.
Johannes Kelder (University Medical Center, Utrecht, The Netherlands) performed a meta-analysis of individual patient data with the aim of quantifying the added diagnostic value of BNP for the diagnosis of HF in primary care.
For the derivation phase of the study, the team analyzed information on 276 patients with suspected HF referred to a rapid-access diagnostic outpatient clinic. Although all patients had signs or symptoms compatible with HF, just 85 (30.8%) were ultimately diagnosed with the condition.
Using univariate and multivariate analysis, Kelder's team identified factors that were significantly associated with a HF diagnosis. This set of nine clinical variables (age, gender, known coronary artery disease, diabetes, orthopnea, elevated jugular venous pressure, crackles, pitting edema and S3 gallop) had a c-statistic of 0.79.
Adding log-transformed BNP to the model increased the c-statistic significantly, to 0.92.
Neither chest x-ray nor electrocardiogram data improved the discriminative ability of the model beyond that seen with the clinical variables plus BNP, the authors remark.
The team then sought to validate the "clinical plus BNP" model in a comparable primary-care population. Again, the model performed well, with a c-statistic of 0.91.
In both cohorts, the model had excellent calibration.
Finally, the team applied the "rule in" and "rule out" cutoff levels, and the lower alternative rule, as described in current European guidelines. They found that this substantially reduced the proportion of patients with a correct diagnosis.
"BNP can be considered a robust additional marker," Kelder et al write in Heart. "It can be concluded that the addition of BNP results in a significant improvement in the discrimination between those with and without HF and yields more adequate probability estimations of the presence of HF."
However, they stress that the purpose of their study was not to provide clinicians with a validated diagnostic rule, and say that the next step is "to develop a diagnostic rule, obtained from the logistic regression model, offering the clinician an estimation of the probability of the presence of HF."
MedWire (www.medwire-news.md) is an independent clinical news service provided by Springer Healthcare Limited. © Springer Healthcare Ltd; 2011
By Joanna Lyford