Methods
This was a retrospective data analysis at SUNY Downstate Medical Center, an academic, tertiary care center that serves a largely African American and Afro-Caribbean population in Brooklyn, NY. Data was analyzed in patients that were admitted with an acute ischemic stroke that after standard evaluation by a stroke neurologist where not found to have a clear etiology for stroke and were denominated as cryptogenic stroke.
All patients received at least 24-hours of continuous telemetry monitoring that did not reveal the presence of AF and subsequently underwent implantable loop recorder (ILR) placement in standard left chest location. 6 Both Abbott and Medtronic ILR were implanted and programmed to AF duration detection of >30 sec and tachycardia detection > 130 bpm and manufactured specific setting allowing for “balanced” detection of AF. Patients were followed-up in the outpatient device clinic by in-person interrogations of ILR or remotely. Each event identified as AF was initially reviewed by experienced mid-level practitioner and verified by a senior electrophysiologist.
Descriptive statistics were collected for all variables. All continuous variables are provided with mean, standard deviations, and proportions as percentages. Means were compared using independent sample t-tests; and proportions were compared using Chi square analysis. Binary univariate and multivariate regression were performed to determine predictors for A-fib detection. All analysis was done using SPSS (IBM SPSS Statistics for Macintosh, Version 24.0. Armonk, NY) using a P-value of <0.05 for statistical significance. Kaplan Meier curve was created using Prism (Prism 8 for macOS, Version 8.4.2, San Diego, CA). Study design was approved by SUNY Downstate IRB.