SUBJECTS AND METHODS: Between Aug 2002 and May 2004, 433 patients found to have systolic HF or moderate to severe RI were included. The patients were divided into 3 groups (group I; only HF, group II; only RI, group III; HF and RI). The severity of RI was graded according to the calculated creatinine clearance (Ccr); moderate 30≤Ccr<60, severe 15≤Ccr<30 or end stage renal disease (ESRD) Ccr<15 mL/min.

RESULTS: The mean age of the patients was 67.6±12, and 49% were male. There were significant differences in the mean BNP levels between group III and the other two groups (p<0.001); group I (n=65, 837.3±884), group II (n=137, 1049.4±1332) and group III (n=231, 1738.3±1501 pg/mL). A weak negative correlation was note between BNP and Ccr (r=-0.335, p<0.001) in patients with RI. As the renal function deteriorated, the mean BNP of groups II and III was found to be elevated (moderate 625.5±574, 1183.0±1056; severe 760.5±1211, 2205.4±1470; ESRD 2157.6±1831, 3209.9±1900 pg/mL, p<0.05), with the mean BNP of group III being higher than that of group II for each grade (p<0.05). From the ROC curve, the optimal cut-off point of BNP for the diagnosis of systolic HF in patients with RI was 829 pg/mL (accuracy 68%, sensitivity 66% and specificity 70%, p<0.001).

CONCLUSION: In the case of patients with moderate to severe RI, a higher BNP cut-off point for the diagnosis of systolic HF and a relatively lower diagnostic accuracy of BNP should be considered.