Courses of urinary and serum NT-proBNP-levels as predictive marker for progression of heart failure and mortality in a real-life ICD-cohort

Background: Serum NT-proBNP is an established marker for heart failure and congestive heart failure.  Although not established in clinical routine, urinary NT-proBNP has also shown prognostic value in heart failure. However, there is few data regarding the prognostic value of courses of NT-proBNP, especially in patients with ICD.

Aim: The current study aims to evaluate the prognostic value of courses of serum and urinary NT-proBNP in regard of mortality and progression of heart failure, taking the course of ejection fraction (EF) into account.

Methods: 289 patients with ICDs who received regular follow-up in our out-patient clinic were included in this prospective study. Upon recruitment, blood and urine samples were collected and echocardiography was performed by trained staff members. Clinical follow-up was performed every 6 months and follow-up echocardiography was performed at least once per year (median 8 follow up echocardiographies). Additional blood and urine samples were acquired in a complementary study follow up (median follow up time 29 months). According to initial EF and changes in EF, patients were classified as preserved (EF initially and in every echocardiography afterwards >40%; n=75), recovered (initial EF ≤ 40%, in any follow up echocardiography and every subsequent echocardiography EF >40%; n=32), reduced (EF in every echocardiography ≤ 40%; n=40) or undulating (multiple changes in EF between > and ≤ 40%; n=34). Deceased patients were analysed separately.

Results: Median follow-up time was 29 months. 107 patients died during follow-up. Median age was 68 years (IQR 59-77 years) and 44 patients (15.2%) were women. 168 patients suffered from symptomatic heart failure (NYHA ≥ II, 58.3%). Median GFR was 65 ml/min/1.73m2 (IQR 47-86 ml/min/1.73m2). 149 patients (49.5%) received their ICD because of primary prevention and 150 patients (53.4%) showed an EF ≤ 35% upon study inclusion. Median levels of serum NT-proBNP were 819 pg/ml (IQR 284-2060 pg/ml) and of urinary NT-proBNP were 57.6 ng/g urinary creatinine (IQR 20.4-167.1 ng/g urinary creatinine).
Patients with preserved EF showed significantly lower levels of urinary and serum NT-proBNP compared to patients with reduced or undulating EF upon recruitment and follow-up (each p < 0.05).  Patients who died during follow up had significantly higher levels of serum NT-proBNP as well as urinary NT-proBNP than each other cohort (each p < 0.05). Furthermore, patients who died after submitting blood and urine samples for follow up showed significantly higher changes in NT-proBNP levels than every other cohort (median increase in serum NT-proBNP 767pg/ml and in urinary NT-proBNP 52.1 ng/g urinary creatinine, each p < 0.05). There was no significant difference in changes of serum or urinary NT-proBNP levels between the other cohorts (p=n.s.). In ROC-analyses serum and urinary NT-proBNP showed promising AUCs for predicting mortality (serum NT-proBNP AUC 0.81; urinary NT-proBNP AUC 0.78).

Conclusion: Higher levels of serum NT-proBNP and urinary NT-proBNP were associated with higher mortality in ICD-patients. Especially a significantly higher increase over 2 years in NT-proBNP was associated with higher morality. ICD-patients with preserved EF showed significantly lower levels of NT-proBNP.