Abstract
Tolvaptan (TLV) carries the risk of serious side effects, and its introduction requires hospitalization. Therefore, it is important from the viewpoints of safety and medical economics to predict in advance, the patients for whom it will be effective and introduce it. The purpose of this study was to investigate the noninvasive and simple predictors for identifying TLV responders. We conducted a retrospective observational study of patients with heart failure who had TLV introduced at our hospital from January 1, 2017, to December 31, 2018. By using the body weight and BNP reduction as the effect indices, predictors of body weight and BNP reduction were extracted by logistic analysis. The sensitivity and specificity at the cutoff value obtained by ROC analysis were also examined. Among 85 subjects, urine sodium concentration >63 mEq/L [odds ratio (OR): 6.11, 95% confidence interval (CI): 1.36-27.4] was detected as a predictor of body weight reduction. The sensitivity at this cutoff value was 81%, and the specificity was 70%. Serum osmolarity>291 mOsm/L (OR: 3.76, 95% CI: 1.00-14.2), urine potassium concentration<21 mEq/L (OR: 4.45, 95% CI: 1.09-18.2), and urine sodium concentration>71 mEq/L (OR: 7.38, 95% CI: 2.05-26.6) were detected as predictors of BNP reduction. The sensitivities were 62%, 53%, and 73%, and the specificities were 58%, 68%, and 68%, respectively. Therefore, it was suggested that urine sodium concentration may be useful as a predictor of body weight and BNP decrease after TLV induction.
