Serum magnesium and burden of atrial and ventricular arrhythmias: The Atherosclerosis Risk in Communities (ARIC) Study

Abstract

Introduction: Low serum magnesium (Mg) is associated with an increased incidence of atrial and ventricular arrhythmias. A richer phenotyping of arrhythmia indices, such as burden or frequency, may provide etiologic insights.

Objectives: To evaluate cross-sectional associations of serum Mg with burden of atrial arrhythmias [atrial fibrillation (AF), premature atrial contractions (PAC), supraventricular tachycardia (SVT)], and ventricular arrhythmias [premature ventricular contractions (PVC), non-sustained ventricular tachycardia (NSVT)] over 2-weeks of ECG monitoring.

Methods: We included 2513 ARIC Study visit 6 (2016-2017) participants who wore the Zio XT Patch-a leadless, ambulatory ECG-monitor-for up to 2-weeks. Serum Mg was modeled categorically and continuously. AF burden was categorized as intermittent or continuous based on the percent of analyzable time spent in AF. Other arrhythmia burdens were defined by the average number of abnormal beats per day. Linear regression was used for continuous outcomes; logistic and multinomial regression were used for categorical outcomes.

Results: Participants were mean ± SD age 79 ± 5 years, 58% were women and 25% black. Mean serum Mg was 0.82 ± 0.08 mmol/L and 19% had hypomagnesemia (<0.75 mmol/L). Serum Mg was inversely associated with PVC burden and continuous AF. The AF association was no longer statistically significant with further adjustment for traditional lifestyle risk factors, only the association with PVC burden remained significant. There were no associations between serum Mg and other arrhythmias examined.

Conclusions: In this community-based cohort of older adults, we found little evidence of independent cross-sectional associations between serum Mg and arrhythmia burden.

Keywords: Atrial arrhythmias; Continuous electrographic monitoring; Older adults; Serum magnesium; Ventricular arrhythmias.

Figure 1.

Predicted probability of atrial fibrillation (intermittent or continuous) detected over 2-weeks ambulatory ECG monitoring^a across serum magnesium with distribution of serum magnesium superimposed. ^a Predicted probabilities calculated from logistic regression modeling serum magnesium using restricted cubic splines with knots at the 5th, 27.5th, 50th, 72.5th, and 95th percentiles and with adjustment for age (mean 79 years), race-center (white-Minnesota), sex (male). Dashed line reflects average probability of atrial fibrillation equal to 0.07. Confidence intervals are pointwise prediction bands based on model-based standard errors.