Author + information
- John G.F. Cleland, MD⁎ (, )
- Iain Squire, MD and
- Leong Ng, MD
- ↵⁎University of Hull, Castle Hill Hospital, Castle Road, Kingston upon Hull HU16 5JQ, United Kingdom
The Heart ProtectionStudy (HPS) reported that the use of simvastatin reduced mortality in a broad range of patients with vascular disease or diabetes and did not exclude patients with heart failure, although this diagnosis was not recorded (1,2). The authors of the CORONA (Controlled Rosuvastatin Multinational Study in Heart Failure) study failed to identify a reduction in mortality with rosuvastatin in predominantly elderly patients with heart failure due to left ventricular systolic dysfunction secondary to ischemic heart disease (3). Overall, the annual mortality was approximately 3% in the HPS trial and 12% in the CORONA study.
Both studies reported plasma concentrations of amino-terminal pro-brain natriuretic peptide (NT-proBNP), and both identified that patients with higher levels had a substantially greater rate of mortality. The plasma concentration of NT-proBNP was measured in 20,536 patients in the HPS trial, with a reported median value of 1,091 pg/ml (interquartile range [IQR] 330 to 3,028 pg/ml) and in 3,664 patients in the CORONA study with a median value of 170 pmol/l (IQR 71 to 380 pmol/l), which corresponds to 1,438 pg/ml (IQR 600 to 3,214 pg/ml). In the HPS trial, analysis of the prognostic importance of NT-proBNP was conducted by dividing patients into 5 groups with an approximately equal number of cardiovascular events. The group with the lowest risk, comprising 28% of all patients, had NT-proBNP levels of <386 pg/ml, and their rate of major vascular events was 16.2% over the course of 5 years (about 4% per year), whereas the 12% of patients who were in the highest-risk group had NT-proBNP values >5,579 pg/ml, and their rate of events was 37.7% over the course of 5 years (about 7% per year). In the CORONA study, the annual rate of major vascular events was approximately 7.7% per year in the lower 2 tertiles (<2,348 pg/ml) combined compared with 20.5% per year in the upper tertile.
These data suggest that plasma concentrations of NT-proBNP in studies designed to have a widely different prevalence of heart failure were rather similar, which is surprising. This anomaly is explained by differences in the assays used. In the HPS trial, the authors used a locally developed assay and in the CORONA study, the authors used a commercially available one (Roche, Nutley, New Jersey). The conversion factor between assays (reporting in picograms per milliliter) is not linear. For instance, cut-offs used in the HPS trial of 386 and 5,758 pg/ml correspond to 68 and 728 pg/ml in the CORONA study, whose authors used the Roche assay. The median value and IQR in the HPS trial would have corresponded to approximately 168 pg/ml (IQR 58 to 422 pg/ml) when using the Roche assay. Therefore, the overlap between plasma NT-proBNP concentrations in the HPS trial and the CORONA study is modest (Fig. 1).
The authors of the HPS trial noted a substantial decrease in the relative but not absolute benefit of simvastatin as NT-proBNP increased. The authors of the CORONA study extend this observation into a higher range of NT-proBNP and suggest a further decrease in benefit from statins with greater levels.
Please note: Dr. Cleland was part of the CORONA Steering Committee for which he received an honorarium. Dr. Ng supervised the laboratory in which the HPS samples were assayed.
- American College of Cardiology Foundation