N-terminal pro–B-type natriuretic peptide modestly improved measures of CVD risk prediction in women with numerous CV events, according to new data published in the Journal of the American College of Cardiology.
The addition of NT-proBNP levels to traditional risk factors and Reynolds Risk Score covariables also led to a small but significant improvement in CVD risk prediction.
Researchers evaluated data collected from 1,821 women who developed incident CVD during the Women’s Health Initiative observational study, along with a randomly selected group of 1,992 women without baseline CVD. All participants were postmenopausal women aged 50 to 79 years at enrollment. The researchers used a prospective case-cohort within the WHI to select 1,821 incident cases of CVD, including 746 MIs, 754 strokes, 160 hemorrhagic strokes and 161 other CV deaths, which occurred during a mean follow-up of 9.9 years.
The primary endpoint was a composite of nonfatal MI, nonfatal stroke and CV-related death.
At study entry, women with incident CVD had significantly higher median levels of NT-proBNP compared with controls (120.3 ng/L vs. 100.4 ng/L; P<.0001). After adjustment for traditional risk factors, women in the highest quartile of NT-proBNP levels (≥140.8 ng/L) had a 53% greater risk for CVD compared with women in the lowest quartile (HR=1.53; 95% CI, 1.21-1.94). Further adjustment for Reynolds Risk Score covariables yielded similar results (HR=1.53; 95% CI, 1.2-1.95).
The association between elevated NT-proBNP and CVD risk persisted in separate analyses of women who experienced CV-related death (HR=2.66; 95% CI, 1.48-4.81 for highest vs. lowest quartile), stroke (HR=1.6; 95% CI, 1.22-2.11) and MI (HR=1.39; 95% CI, 1.02-1.88), after adjustment for traditional risk factors and Reynolds Risk Score covariables.
The researchers noted that the addition of NT-proBNP to traditional risk covariables for multivariate analysis significantly improved c-statistic (P=.0003), categorical net reclassification (P<.0001) and integrated discrimination (P=.0105) compared with the use of traditional variables alone. The addition of NT-proBNP to Reynolds Risk Score for multivariate analysis was associated with similar results.
Pamela S. Douglas, MD, and G. Michael Felker, MD, MHS, from the Duke Clinical Research Institute at Duke University School of Medicine, noted that “the present study contributes important information to help refine risk prediction.” However, “many questions remain, including those related to pathophysiology and clinical translation … It is unclear why NT-proBNP should be predictive of the primary endpoint components, especially those not clearly etiologically related to this physiology, such as hemorrhagic stroke,” they wrote in a related editorial.
For more information:
Douglas PS. J Am Coll Cardiol. 2014;64:1798-1800.
Everett BM. J Am Coll Cardiol. 2014;64:1789-1797.
Disclosure: See the study and editorial for relevant financial disclosures.