Introduction Atrial arrhythmias are strongly associated with thromboembolism with 1 in 5 strokes being attributed to atrial fibrillation (AF). Patients with adult congenital heart disease (ACHD) are at an increased risk of developing atrial arrhythmias compared to those with structurally normal hearts either due to the inherent cardiac anomaly or due to scar formation from surgery. The CHA₂DS₂VASC score is validated for stroke risk stratification for those with acquired heart disease. Its role in adult congenital heart disease however, is unclear.

Methods A retrospective review of the electronic ACHD database of all patients at a large quaternary specialist ACHD Centre with approximately 4500 patients under regular follow up. Data spanned from year 2000–2015 and data collected from referral letters, clinic letters, ECG and echocardiograms.

Results 376 patients with atrial arrhythmias were identified. Mean age was 57.7 ±  15.8 years with 51% female. 88 (23%) of patients had complex disease (including those with a Fontan circulation) with the remainder having simple or moderate forms of congenital heart disease. In this cohort, 52 (13.8%) patients had at least one thrombo-embolic event with 77% being either transient ischaemic attacks or strokes. The remainder of events were embolic peripheral artery occlusions, deep vein thromboses and pulmonary emboli. The mean age of first occurrence of thromboembolic event was 44.2 ±  19 years and mean CHA₂DS₂VASC score was 1.04 ± 0.7 prior to the event. In 46% of cases the thromboembolic event occurred after the atrial arrhythmia had been diagnosed and 71% of patients were not anticoagulated at the time of their first event. Of those patients anticoagulated at the time of their event (n = 9), 2 patients had a documented subtherapeutic INR, 2 were non-compliant with their medication and 1 was on a Novel Oral Anticoagulant (NOAC). It was unclear in 11% of cases whether they were anticoagulated before the event. 88% of patients with a CHA₂DS₂VASC score of 2 or more were anticoagulated, with 93% on warfarin and 5% on NOACs and 2% on low molecular weight heparin. Of those with a CHA₂DS₂VASC score of 0–1, 74% were on either warfarin or a NOAC. There were 4 known deaths from complications related to anticoagulation (2 gastrointestinal bleeds and 2 from massive haemoptysis).

Discussion Atrial arrhythmias are common in ACHD and there is a high incidence of stroke in those with AF. In those groups with higher incidence of AF, loop recorders may be beneficial for detecting asymptomatic disease. The CHA₂DS₂VASC scoring system does not appear to be applicable for stroke risk stratification in ACHD. The mean pre- event CHA₂DS₂VASC was 1. Another marker for patients with  ACHD is required to differentiate low risk from high risk patients. The role of NOACs in this group of patients is unclear and needs further evidence. 1 patient had a stroke on a NOAC despite being compliant.