Preoperative NT-proBNP Improves Risk Prediction Beyond EuroSCORE II in Cardiac Surgery
In 640 cardiac surgical patients followed for nearly three years, preoperative NT-proBNP independently predicted long-term mortality after adjustment for EuroSCORE II and significantly reclassified patient risk. A simplified model combining age, creatinine clearance, chronic pulmonary disease, and NT-proBNP matched the prognostic accuracy of the full EuroSCORE II. Notably, hs-TnT did not add incremental value beyond NT-proBNP, positioning natriuretic peptide measurement as the preferred biomarker for preoperative cardiac surgical risk stratification.
The original study
The predictive value of NT-proBNP and hs-TnT for risk of death in cardiac surgical patients.
- Authors
- Brynildsen J, Petäjä L, Pettilä V, Nygård S, Vaara ST, Linko R, et al.
- Journal
- Clinical biochemistry
- Type
- Journal Article, Multicenter Study, Observational Study
- PMID
- 29371002
Original abstract
BACKGROUND: European System for Cardiac Operative Risk Evaluation II (EuroSCORE II) is used for risk stratification before cardiac surgery, but whether N-terminal pro-B-type natriuretic peptide (NT-proBNP) and high-sensitivity troponin T (hs-TnT) may add prognostic information to EuroSCORE II is not known. METHODS: Preoperative (n=640) and postoperative (n=629) blood samples were available from cardiac surgical patients with 961-day follow-up (FINNAKI Heart study; cohort #1). The accuracy of a parsimonious risk model with NT-proBNP measurements was also tested in 90 patients with respiratory failure after cardiac surgery (FINNALI study; cohort #2). RESULTS: Sixty-one patients (9.5%) died during follow-up in cohort #1. Preoperative NT-proBNP and hs-TnT concentrations correlated (rho=0.58; p<0.001) and were higher in non-survivors compared to survivors: median 2027 (Q1-3 478-5387) vs. 373 (134-1354) ng/L [NT-proBNP] and 39 (16-191) vs. 13 (8-32) ng/L [hs-TnT]; p<0.001 for both. Preoperative NT-proBNP concentrations were associated with time to death after adjustment for EuroSCORE II (HR [lnNT-proBNP] 1.33 [95% CI 1.08-1.64]), p=0.008 and reclassified patients on top of EuroSCORE II (net reclassification index 0.39 [95% CI 0.14-0.64], p=0.003). Pre- and postoperative NT-proBNP concentrations were closely correlated (rho=0.80, p<0.001) and postoperative NT-proBNP concentrations were also associated with long-term mortality after adjustment for EuroSCORE II. A parsimonious risk model that included age, creatinine clearance, chronic pulmonary disease, and NT-proBNP measurements provided comparable prognostic accuracy as EuroSCORE II in cohort #1 and #2 for risk of long-term mortality. hs-TnT measurements did not add to NT-proBNP measurements CONCLUSION: NT-proBNP measurements could improve and simplify risk prediction in cardiac surgical patients.