Michiharu Seki

Efficacy of tolvaptan in patients with refractory ascites in a clinical setting

AIM To elucidate the efficacies of tolvaptan (TLV) as a treatment for refractory ascites compared with conventional treatment. METHODS We retrospectively enrolled 120 refractory ascites patients between January 1, 2009 and September 31, 2014. Sixty patients were treated with oral TLV at a starting dose of 3.75 mg/d in addition to sodium restriction (> 7 g/d), albumin infusion (10-20 g/wk), and standard diuretic therapy (20-60 mg/d furosemide and 25-50 mg/d spironolactone) and 60 patients with large volume paracentesis in addition to sodium restriction (less than 7 g/d), albumin infusion (10-20 g/wk), and standard diuretic therapy (20-120 mg/d furosemide and 25-150 mg/d spironolactone). Patient demographics and laboratory data, including liver function, were not matched due to the small number of patients. Continuous variables were analyzed by unpaired t-test or paired t-test. Fishers exact test was applied in cases comparing two nominal variables. We analyzed factors affecting clinical outcomes using receiver operating characteristic curves and multivariate regression analysis. We also used multivariate Coxs proportional hazard regression analysis to elucidate the risk factors that contributed to the increased incidence of ascites. RESULTS TLV was effective in 38 (63.3%) patients. The best cut-off values for urine output and reduced urine osmolality as measures of refractory ascites improvement were > 1800 mL within the first 24 h and > 30%, respectively. Multivariate regression analysis indicated that > 25% reduced urine osmolality [odds ratio (OR) = 20.7; P < 0.01] and positive hepatitis C viral antibodies (OR = 5.93; P = 0.05) were positively correlated with an improvement of refractory ascites, while the total bilirubin level per 1.0 mg/dL (OR = 0.57; P = 0.02) was negatively correlated with improvement. In comparing the TLV group and controls, only the serum sodium level was significantly lower in the TLV group (133 mEq/L vs 136 mEq/L; P = 0.02). However, there were no significant differences in the other parameters between the two groups. The cumulative incidence rate was significantly higher in the control group with a median incidence time of 30 d in the TLV group and 20 d in the control group (P = 0.01). Cox hazard proportional multivariate analysis indicated that the use of TLV (OR = 0.58; P < 0.01), uncontrolled liver neoplasms (OR = 1.92; P < 0.01), total bilirubin level per 1.0 mg/dL (OR = 1.10; P < 0.01), and higher sodium level per 1.0 mEq/L (OR = 0.94; P < 0.01) were independent factors that contributed to incidence. CONCLUSION Administration of TLV results in better control of refractory ascites and reduced the incidence of additional invasive procedures or hospitalization compared with conventional ascites treatments.

Sa1097 Analysis of Weekend Admission on Upper Gastrointestinal Bleeding; Is Weekend Effect Demonstrated in Japan?

Background and Aims: Whether weekend admission influences the mortality of upper gastrointestinal bleeding(UGIB) remains unclear in Japan. In Europe and North-America, some previous studies identified worse outcomes of weekend admission, however the opposite results were reported from Hong-Kong, Korea and Taiwan. The outcomes may vary considerably between each nation. This study was conducted to assess the weekend effect in patients with UGIB on the basis of a 10-year single-center experience in Japan. Methods and Patients:595 hospitalized patients, who were diagnosed with UGIB between January 2004 and March 2014, were enrolled. We divided those patients into two groups, variceal and nonvariceal UGIB. Bivariate analyses and multivariate logistic regression models were used to evaluate predicting variables of the in-hospital mortality, included patient characteristics, waiting time for endoscopy, and weekend admission. Results: 595 patients consisted of 484 nonvariceal and 111 variceal UGIB. 169 out of 484 Patients with nonvariceal UGIB were weekend admission and had similar mortality rates compared to the weekday group (2.96% vs 2.54%;odds ratio[OR],1.17;confidence interval[CI],0.38-3.63; P=0.79). The waiting time for endoscopy was also not different (9.97 hours versus 9.19 hours, P=0.55). There was no statistical difference in baseline characteristics. In the multivariate logistic regression model, mortality was correlated with Blatchford score ([OR],1.32 per point;[CI],1.05-1.72;P= 0.015) and waiting time to endoscopy ([OR],1.04 per hour;[CI],1.009-1.07; P=0.014). 34 out of 111 Patients with variceal UGIB were weekend admission and had relatively higher mortality rates compared to the weekday group (17.7% vs 10.4%;[OR],1.85;[CI],0.595.81;P=0.29), but not significant. The waiting time for endoscopy was not different (3.35 hours versus 5.3 hours, P =0.09). There was no statistical difference in baseline characteristics, except Child-Pugh scores. Child-Pugh scores of patients with variceal UGIB on weekend admission were significantly higher than those on weekday admission (9.29 vs 8.3; P=0.01). In the multivariate logistic regression model, mortality was correlated with Child-Pugh scores ([OR],1.83 per point;[CI],1.16-3.22; P=0.008). Conclusion:In Japan, quality of care did not appear to differ between week/weekend admission of patients with UGIB and weekend effect was not demonstrated in this study. Patient characteristics and waiting time for endoscopy were correlated with the risk of the in-hospital mortality.