PDB28 A Real-World Effectiveness of Udca Added to Oha on Glycemic Control in Type 2 Diabetes Patients with LIVER Disease

This study aimed to compare the effects of oral hypoglycemic agents (OHAs) alone and OHAs prescribed with ursodeoxycholic acid (UDCA) on lowering HbA1c in T2DM patients with liver disease. Among 70,655 patients whose first diagnosis of T2DM recorded in Severance Hospital's EMR from January 2010 to December 2017, selection criteria including liver disease (K75.8, K70, K76.0) were applied to specify qualifying target patients with at least 90-day prescription of OHAs during the observation period. Two different analytical methods were approached: Level 1 (Lv1) was a method analyzed in the entire observation period for 1 year and the change in HbA1c from the baseline at diagnosis to the last observation. Level 2 (Lv2) analyzed the specific duration of drug treatment of interest within 2-year long period by measuring HbA1c before and after treatment. The matching variables between test groups and control groups were all factors including patient demographics, laboratory test values at diagnosis, the period between the start of prescription after diagnosis and prescription duration. In Lv1, 1:1 matching using the PS method showed no statistical significance, but it was found that the test group further decreased 0.35% HbA1c (n=108, 1.25 ± 1.57 vs. 0.90 ± 2.19, p>0.05). In addition, the multiple linear regression analysis with adjusting the matching variables (n=108(test) vs. n=5,528(control)) showed significant reduction by 0.419% in the test group compared to the control (p=0.002). In the same analysis method, Lv2 (n=138(test) vs. n=258(control)) significantly decreased by 0.349% in the test group (p=0.012). In a retrospective DB study, an additional lowering effect on HbA1c when UDCA combined with OHA was consistently observed in subgroups (by OHA components). It is also expected that matching variables, by HbA1c change was statistically significant between two groups, can be used as stratification when designing large-scale prospective clinical observation study.