Article Text

Download PDFPDF

D Diabetes mellitus generates the substrate for atrial fibrillation by causing a localised conduction block
  1. Klemen Ziberna,
  2. Alexandra S Mighiu,
  3. Alice Recalde,
  4. Barbara Casadei,
  5. Ricardo Carnicer
  1. Radcliffe Department of Medicine, Division of Cardiovascular Medicine, University of Oxford, Oxford, UK


Introduction Diabetes mellitus (DM) is an important risk factor for atrial fibrillation (AF) – the most common heart rhythm disorder. However, the mechanisms underlying this association are poorly understood. In addition, patients with DM often have comorbidities, such as obesity, hypertension, dyslipidaemia, that are also independently associated with increased risk for AF. The aim of this project was to explore whether prolonged hyperglycaemia alone is sufficient to increase the risk of AF.

Methods Multiple low-dose (50 mg/kg over 5 days) intraperitoneal injections of streptozotocin were used to induce DM in C57BL/6 mice. After 12 weeks of DM, left ventricular (LV) systolic and diastolic function was characterised using echocardiography. In vivo atrial electrophysiological properties and arrhythmia inducibility were assessed using transoesophageal atrial pacing. Atrial conduction time and action potential duration (APD) and conduction velocity were measured by optical mapping of isolated atria.

Results Diabetic mice had significantly higher probability of in vivo AF induction (15±3% vs. 6±1%, in controls, p=0.005, n=24–26/group), atrial ECG conduction abnormalities (longer PQ interval in diabetic mice: 44±1 ms vs. 39±1 ms in controls, p=0.001, n=24–26/group), but without any changes in the P wave duration, RR, QRS or QT intervals. Diabetic mice developed modest LV diastolic dysfunction (tissue Doppler E’/A’ 1.02±0.05 vs. 1.20±0.06 in controls, p=0.04, n=11–12/group), but no LV systolic dysfunction. In addition, diabetic mice also had larger relative LA size (LA area/body weight 0.28±0.01 mm2/g vs. 0.23±0.01 mm2/g in controls, p=0.001, n=11–12/group). Optical mapping revealed a two-fold greater conduction time in diabetic left atria (LA) (39±3 ms vs. 22±2 ms in controls, p=0.0006, n=6/group), without differences in the right atrial (RA) conduction time. The APD was not significantly different in RA or LA. The regional analysis of optical mapping recordings demonstrated a significant decrease in the conduction velocity in the medial part of the RA resulting in increased conduction wavefront roughness in the LA. Finally, the medial part of the diabetic RA also had a significantly higher total collagen content (5.5±0.5 µg/mg vs. 9.1±0.6 µg/mg in controls, p=0.0002, n=9–11/group), and a decrease in the connexin 43 levels (0.92±0.05 vs. 0.66±0.4 in controls, p=0.001, n=7–8/group), with no significant changes in other parts of the atria.

Conclusions Prolonged hyperglycaemia alone is sufficient to cause pathological remodelling making atria more susceptible for AF. Diabetic hearts also have modest LV diastolic dysfunction, relative LA enlargement, but no LV systolic dysfunction. In this study, we provide evidence of a novel mechanism where localized (rather than global) atrial fibrosis in a critical area causes a focal conduction defect and predisposes diabetic mice to AF.

  • Atrial fibrillation
  • diabetes mellitus
  • cardiac electrophysiology

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.