2. METHODS
In this retrospective cohort study, we reviewed aggregate de-identified data via the TriNetX Research Network from https://live.trinetx.com. Inclusion criteria was: (1) patients having been diagnosed with AF using ICD-10 codes (see Supplemental Table 1), (2) AF diagnosis present for at least one month, and (3) patients being 18 years or older. For the liraglutide cohort, we included (4) patients on liraglutide and (5) that these patients were on liraglutide for at least month as of November 13, 2021. A 1:1 propensity-scored matching was then conducted and controlled for age, gender, race, ethnicity, hypertensive heart disease, diabetes, metabolic diseases, overweight and obesity, genitourinary disease, respiratory disease, atrial fibrillation, ischemic heart disease, neoplasms, heart failure, vascular disease, cerebrovascular disease, rheumatic valvular disease, nonrheumatic valvular disease, cardiovascular procedures, cardiovascular medications, and blood glucose lowering medications.
Statistical analysis was conducted via TriNetX. Categorical variables were compared using chi-square tests while continuous variables were compared utilizing independent sample t-tests. Cox proportioned hazards ratio with 95% confidence intervals (CI) for incidence of all-cause mortality, ischemic stroke, hemorrhagic stroke, acute heart failure episode, and acute coronary syndrome episode were generated after propensity score matching. Kaplan-Meier analysis was conducted to estimate all-cause mortality. Statistical significance was set at p < 0.05.