Introduction There is growing evidence that paroxysmal atrial fibrillation (PAF) carries the same thromboembolic stroke risk as permanent AF. With up to half of episodes asymptomatic PAF frequently eludes detection on routine 12-lead electrocardiogram (ECG) as well as Holter monitor testing. The loop recorder (LR) is emerging as a powerful tool in the diagnosis of asymptomatic PAF with impressive sensitivity and specificity, detecting paroxysmal AF in those with a normal ECG and Holter recording. LR use is not however entirely without associated morbidity as well as significant cost implications. The aim of this study is to evaluate features of the baseline ECG at LR implantation which might delineate increased propensity to paroxysmal atrial fibrillation.

Methods We retrospectively studied a population of 192 adult patients who underwent LR implantation from January 2010 to March 2013 for the investigation of syncope or palpitations. LRs were checked every 3 months or whenever patients activated the LR. Information regarding demographics and co-morbidities were collected. Baseline ECG characteristics were collected at implant date. All the ECGs and EGMs were scrutinised by two blinded investigators. PAF incidence was defined as episodes lasting >30 sec recorded in the LR memory. Multivariate logistic regression was used.

Results Our LR population consists of 192 patients, 99 (52%) male, with a median age of 64 years (range 18–88), in whom LR was implanted for the investigation of syncope (65%), palpitations (17%) or both (18%). PAF was detected by LR in 39 patients. Median age of this group was 76 (range 38–88), 25 (65%) were male. 24 (62%) recorded a symptomatic event on the LR.

Multivariate logistic regression analysis identified age (OR 1.05, 95% CI 1.01–1.08, p = 0.004), cigarette smoking (OR 3.82, 95% CI 1.45–10.02, p = 0.007) and RBBB (OR 6.57, 95% CI 2.23–19.36, p = 0.0007) as significant predictors of AF. The major finding was the increased prevalence of RBBB at baseline, detected in 15 (38%) of patients who went on to demonstrate PAF (5 complete and 10 incomplete RBBB).

Conclusion In our study we looked at incidence and burden of paroxysmal episodes of AF in a loop recorder population. Predictive factors included advancing age, cigarette smoking and right bundle branch block. Incomplete right bundle branch block has already been postulated as a predictive factor for atrial fibrillation, and our study supports that hypothesis, using a larger sample population. Knowledge that these three factors may predispose to atrial fibrillation may be a useful guide when deciding whether to further investigate individuals to minimise thromboembolic risk and to develop an optimal anti-arrhythmic strategy.Future research should be directed toward identifying the relationship between RBBB and PAF.