Association of Obesity and Cardiovascular Outcomes in Atrial Fibrillation

This retrospective nationwide cohort study used administrative claims data from the Korean National Health Insurance Service (NHIS) and the combined health check-up database of the National Health Insurance Corporation between 2013 and 2020.

The investigators included oral anticoagulant (OAC) new users diagnosed with atrial fibrillation (AF) between January 2015 and December 2020. Patients aged <20 years, those who were regarded as valvular AF patients (patients with mitral stenosis or prosthetic heart valve), those with possible alternative indications for OAC (pulmonary embolism, deep vein thrombosis, or recent joint replacement surgery), and those with cancer were excluded from the analysis.

The follow-up period was defined as the time from the index date (date of OAC prescription) to each outcome event, date of death, or end of the study period (December 31, 2020), whichever came first.

Patients' demographic data, comorbidities, concomitant medications, and income level were collected from the Korean NHIS database. The recent health check-up data from the index date was also ascertained, including height, weight, waist circumference, blood pressure, health surveys, and laboratory exam. Health survey included family history, smoking history, alcohol history, and the level of individual physical activity. The investigators defined the proportion of medical use by calculating formula with the recuperation cost and the number of the visit to hospitals.

According to BMI following the World Health Organization recommendation for Asian population, study patients were categorized into 5 groups: underweight, <18.5 kg/m2; normal range, 18.5 to <23 kg/m2; overweight, 23 to <25 kg/m2; obese I, 25 to <30 kg/m2; and obese II, ≥30 kg/m2[4].

During the follow-up period, the investigators assessed 3 clinical outcomes, including thromboembolism, major bleeding, all-cause death, and a composite clinical outcome of thromboembolism+major bleeding+all-cause death. Major bleeding included the brain hemorrhage and gastrointestinal bleeding. Clinical outcomes were mainly defined by the the International Classification of Diseases, 10th revision (ICD-10). Patients were censored at the clinical outcomes, the discontinuation of index OAC treatment, or the end of the study period (December 31, 2020), whichever came first.

Continuous variables are presented as mean±SD or median (interquartile ranges). Categorical variables are presented as numbers and percentages. Baseline characteristics were compared across 5 BMI categories with a linear trend test using a generalized linear model for continuous variables and the Cochran-Armitage trend test for categorical variables. The incidence rates of clinical outcomes were calculated based on the number of events during the follow-up period divided by 100 person-years at risk.

For the primary analyses to evaluate the association between BMI and clinical outcomes in AF patients treated with OAC, the investigators conducted 2 types of analyses: (1) BMI as a continuous variable and (2) BMI as a categorical variable by predefined BMI ranges. First, the association between BMI as a continuous variable (per 5 kg/m2 increase) and clinical outcomes was evaluated using a Cox proportional hazard model to derive unadjusted and adjusted hazard ratios (HRs). Adjusted cubic spline curves were used to visualize the relationship between BMI and the risk of clinical outcomes. Second, the association between different BMI categories and clinical outcomes was explored using adjusted Cox proportional hazard models. Patients with a normal range of BMI defined as 23 to <25 kg/m2 were used as the reference group. Covariates in the multivariable analysis included age, sex, CHA2DS2-VASc score, comorbidities, renal function, use of antiplatelet agents, and OAC treatment (warfarin or non-vitamin K antagonist OAC [NOAC]).

All analyses were 2 tailed, and P<0.05 was considered significant. Statistical analyses were conducted with R-statistics.