Background: Natriuretic peptides (BNP/NT-proBNP) are routinely used for the diagnosis of heart failure and to predict outcome in patients with heart failure. Likewise, for patients without heart failure, a potential value of natriuretic peptides for the prediction of prognosis is suggested. For non-heart failure cohorts, however, no thresholds are established.

Methods: The present analysis is based on a registry of patients undergoing conventional coronary angiography at our department between 2004 and 2019. Patients with existing diagnosis of heart failure or elevated natriuretic peptides (BNP >100pg/nl, NT-proBNP >400pg/nl) were excluded. Moreover, patients with missing follow-up information or without BNP/NT-proBNP levels at admission were excluded. Random sampling was used to divide the cohort into a derivation cohort (2/3 of the cohort) and a validation cohort (1/3 of the cohort). Incidence of death of any cause during follow-up was recorded. As either BNP or NT-proBNP was available for singular patients and to adjust for the skewed distribution, we standardized BNP and NT-proBNP levels based on gender specific percentile rank in levels from 0 to 99. In the derivation cohort, cox regression analysis was used to determine the association of BNP/NT-proBNP rank as continuous variable with incident mortality in univariable and multivariable models (age, sex, systolic blood pressure, LDL-cholesterol, diabetes, and smoking status). Hazard ratios and 95% confidence intervals were calculated per 1 standard deviation increase in BNP/NT-proBNP rank. Receiver operating characteristics curve analysis was performed, with corresponding area under the curve, along with Youden’s J index assessment, to establish a threshold for prediction of survival. The association of this threshold with incident mortality was then tested in the derivation cohort and validated in the validation cohort.

Results: Overall, 3,687 patients (mean age: 62.9±12.5 years, 71% male) were included in our analysis. During a mean follow-up of 2.6±3.4 years, 169 deaths of any cause occurred. In the derivation cohort, BNP/NT-proBNP was significantly associated with mortality (univariable: 1.43 [1.17-1.74], p=0.0003, multivariable: 1.25 [1.01-1.54], p=0.04). Based on Youden’s J index, BNP-thresholds of 9.6 and 29 and NT-proBNP thresholds of 65 and 77 for men and women, respectively, were determined. In the derivation cohort, BNP/NT-proBNP levels above these thresholds were significantly associated with increased mortality (univariable: 3.19 [1.75-5-82], p<0.0001, multivariable: 2.44 [1.32-4.53], p=0.005). The predictive value of the determined thresholds was confirmed in the validation cohort (3.02 [1.43-6.35], p=0.004), independent of traditional cardiovascular risk factors (2.78 [1.26-6.14], p=0.01).

Conclusion: We here describe gender-specific BNP/NT-proBNP thresholds that allow prediction of impaired survival in patients without heart failure. Utilization of these thresholds in clinical routine may qualify for risk prediction in non-heart failure cohorts, independent of traditional cardiovascular risk factors.