Comparison of Characteristics Between Chinese Patients Taking Glucagon-like Peptide 1 Receptor Agonists and Insulin: A Cross-sectional Database Analysis Ke Wang, PhD 1 ; Yun Chen, PhD 1 ; Alena Strizek, MS 2 ; Kristina Boye, PhD 3 ; Liqun Gu, MD 1 ; Yanjun Liu, MS 4 ; and Shuli Qu, MPH 4 1 Lilly Suzhou Pharmaceutical Company, Ltd., Shanghai, China; 2 Eli Lilly and Company, Sydney, Australia; 3 Eli Lilly and Company, Indianapolis, IN, USA; and 4 Real World Insights, IQVIA, Shanghai, China ABSTRACT Purpose: In China, although insulin has been prescribed for decades, glucagon-like peptide 1 receptor agonists (GLP-1-RAs) have been available as an injectable treatment for patients with type 2 diabetes mellitus (T2DM) since 2009. GLP-1 RAs are listed as second-line treatment in the 2017 Chinese Guideline for patients with T2DM in whom prior oral antidiabetic therapy has failed. This study compares the baseline characteristics of Chinese patients with T2DM taking different prescriptions of first injectable therapy (GLP-1-RA or insulin). Methods: The IQVIA Patient Diary Study database, which captures data from a patient medical recordebased physician online survey, was the data source used in this study. Cross-sectional patient data were collected from hospitals in 15 major Chinese cities from June 1, 2016, to June 30, 2018. Adults with T2DM commencing either GLP-1-RA or insulin use as their first injectable antidiabetic therapy were included. Baseline demographic and clinical characteristics were compared between the GLP-1-RA and insulin treatment groups, using t tests and c 2 or Fisher exact tests. Findings: Overall, 563 patients using GLP-1-RAs and 2387 using insulin were identified. In general, patients using GLP-1-RA were younger (mean [SD], 49.6 [10.8] years vs 59.3 [10.9] years), had lower mean (SD) glycosylated hemoglobin levels (8.5% [1.2%] vs 9.6 [1.7%]), had lower mean (SD) fasting plasma glucose levels (9.0 [1.9] mmol/L vs 10.8 [2.6] mmol/L), higher mean (SD) body mass indexes (29.4 [3.9] kg/m 2 vs 24.6 [3.1] kg/m 2 ), had higher comorbidity of obesity (75% vs 15%), had a higher occurrence of hyperlipidemia (63% vs 44%), and had lower occurrence of neuropathy (13% vs 34%) when compared with those using insulin (P < 0.0001 for all). The results of multivariate logistic regression model indicate that when controlling other variables in the multivariate logistic regression model, a higher fasting plasma glucose level and a longer diagnosis duration are associated with higher odds of insulin therapy commencement, but higher body mass index and some comorbidities, such as obesity and hyperlipidemia, are associated with higher odds of being a GLP-1-RA user. Implications: Significant differences were identified between selected baseline characteristics of patients initiating GLP-1-RA and insulin therapy, suggesting that these medicines are more likely to be prescribed to different types of patients with T2DM in China. These findings may help to inform Chinese physicians regarding the characteristics of those patients with T2DM who are initiating treatment with a GLP-1- RA or insulin. Because the Patient Diary Study data were collected from hospitals in 15 major cities in China, one noteworthy limitation is that the results may not represent the overall treatment pattern in rural areas of China. (Clin Ther. xxxx;xxx:xxx) © 2019 Elsevier Inc. All rights reserved. Keywords: cross-sectional study, GLP-1-RA, insu- lin, patient characteristics, type 2 diabetes mellitus. Accepted for publication August 1, 2019 https://doi.org/10.1016/j.clinthera.2019.08.003 0149-2918/$ - see front matter © 2019 Elsevier Inc. All rights reserved. ▪▪▪ xxxx 1 Clinical Therapeutics/Volume xxx, Number xxx, xxxx