Background: Former studies showed that urinary N-terminal pro-brain natriuretic peptide (NT-proBNP) was comparable to plasma NT-proBNP regarding diagnosis of acute and chronic heart failure. Most commonly ejection fraction (EF) is used for assessment of left ventricular (LV) function. However, impairment of longitudinal LV function measured by global longitudinal strain (GLS) often precedes deterioration of EF.  There is only limited evidence whether urinary NT-proBNP correlates with impairment of GLS. Therefore, we investigated GLS, urinary NT-proBNP and plasma NT-proBNP in a prospective cohort of ICD patients.

Methods: We included 148 patients with ICD and an echocardiographic exam performed within six months before or after blood and urine sample collection. Levels of urinary NT-proBNP were related to urinary creatinine to avoid dilution effects.

Results: Out of 148 patients, 82% were male. 51% had an implanted ICD device for primary, 49% for secondary prevention of sudden cardiac death. 50% had coronary artery disease (CAD), 22% dilatative cardiomyopathy. Mean LVEF was 43.6%, mean GLS was 12.5%.

GLS showed a significant and negative correlation to plasma NT-proBNP (r=-0.555, P<0.001) as well as urinary NT-proBNP (r=-0.475, P<0.001).

Patients with GLS values ≥-16% corresponding to impairment of longitudinal LV function showed significantly higher values for plasma NT-proBNP (1888 pg/mL for GLS ≥-16% vs. 775 pg/mL for GLS <-16%, P<0.001) as well as for urinary NT-proBNP (165 pg/mg crea for GLS ≥-16% vs. 120 pg/mg crea for GLS <-16%, P<0.001).

Regarding receiver operating characteristic (ROC) analysis, plasma and urinary NT-proBNP could predict impairment of GLS (>-16%) with an area under the curve (AUC) of 0.83 for plasma NT-proBNP (P<0.001) and an AUC of 0.81 for urinary NT-proBNP (P<0.001). This resulted in a sensitivity of 83% and a specificity of 77% for plasma NT-proBNP (cut-off value 421 pg/mL). For urinary NT-proBNP, sensitivity was 80% and specificity was 81% at a cut-off value of 21 pg/mg crea.

Conclusion: In our prospective cohort of ICD patients, plasma as well as urinary NT-proBNP were capable of predicting impairment of longitudinal LV function. Detecting subclinical LV restriction via urine testing could be a clinical application that merits further investigation.