Urinary NT-proBNP compared head-to-head to plasmatic NT-proBNP in a real life collective with an ICD.

Aims: Plasma NT-proBNP is an established marker of heart failure. Previous studies suggested urinary NT-proBNP has potential as marker of chronic heart failure as well. The objective of this study was to compare urinary NT-proBNP to plasma NT-proBNP in a real-life collective of patients with an ICD, especially regarding ICD-therapies.

Methods & results: NT-proBNP was assessed in plasma and fresh spot urine (the latter related to urinary creatinine) from 322 patients of our ICD outpatient clinic. 54 healthy individuals served as a control group. Follow-up regarding mortality and ICD therapies was performed after 32 months (IQR 5-35 months). Plasma and urinary NT-proBNP was positively correlated (r = 0.89, p < 0,001). According to ROC analysis urinary NT-proBNP detected LV dysfunction (EF<35 % vs. healthy CTRL) with very satisfying predictive values (AUC 0.95), but plasma NT-proBNP showed slightly better values (AUC 0.99). Patients who received appropriate ICD-shock-therapies showed significantly higher plasma (p < 0.001) as well as urinary NT-proBNP levels (p = 0.011) compared to patients without shock-therapy. In Kaplan-Meier analysis, plasma as well as urinary NT-proBNP levels > Youden-Index showed significantly higher event rates for appropriate ICD-shock therapies (p < 0.001 and p = 0.016) and the combined endpoint of all-cause-mortality and shock therapies (each p < 0.001). Urinary and plasma NT-proBNP were independent predictors for appropriate ICD-shock-therapies and for the combined endpoint of all-cause mortality and appropriate ICD-shock-therapies (each p < 0.001).

Conclusion: Urinary NT-proBNP as a marker for LV dysfunction and symptomatic heart failure showed promising predictive values. Associations between plasma as well as urinary NT-proBNP and ICD shock-therapies could be shown.