Abstract Background Detection of haemodynamically significant shunting from atrial septal defect (ASD) can be identified by using the pulmonary-systemic flow ratio (Qp/Qs). However, calculation of Qp/Qs by echocardiography relies on geometric assumptions that outflow tracts are circular, suffers from squared linear measures, and accurate measurement of right ventricular outflow tract (RVOT) diameter is challenging. Whilst adults with ASD should have overt right ventricular dilatation, RV sizing on echocardiography is often subjective and underappreciated. We evaluated whether a dimensionless index of flow (VTI) differences between the RVOT and left ventricular outflow tract (LVOT) could assist in identifying patients with ASD, and compared it with relative atrial index (RAI), a parameter previously assessed in identifying atrial shunting. Methods Data from 64 consecutive patients who underwent ASD closure and had no concomitant lesions, were compared with 63 normal controls. RVOT VTI, LVOT VTI, LVOT diameter, and atrial areas were measured. Results Between controls and ASD patients, there was no difference in LVOT VTI or forward stroke volume, but ASD patients had significantly higher RVOT VTI. The RVOT-LVOT VTI ratio was 1:1 in controls and 1.5:1 in ASD patients. Area under the ROC curve analysis of RVOT-LVOT VTI ratio was 0.83, and a ratio of 1.1:1 predicted patients with ASD with 86% sensitivity and 73% specificity. In comparison, the area under the ROC curve analysis of RAI was 0.70, and an RAI of 1.05:1 predicted patients with ASD with 77% sensitivity and 44% specificity. Conclusion Calculation of the dimensionless RVOT-LVOT VTI ratio is simple, and may be a useful additional semi-quantitative tool to assist cardiologists and sonographers in detecting atrial shunting, particularly in patients with borderline or overt right heart dilatation, and identify who patients should undergo further evaluation. Funding Acknowledgement Type of funding sources: None.

Abstract Background Pharmacological stress testing is commonly performed for cardiovascular risk stratification in potential renal transplant candidates due to a perceived inability for these patients to exercise. We have previously reported that exercise stress testing is feasible in renal transplant candidates but the prognostic utility of exercise capacity in this patient group is not known. Purpose This study prospectively evaluated the effect of exercise capacity on the risk of major adverse cardiovascular events (MACE), defined as a composite of cardiac death, non-fatal myocardial infarction, and stroke, in renal transplant candidates undergoing exercise stress echocardiography (ESE) for pre-transplant cardiovascular assessment. Methods We evaluated 898 consecutive patients with chronic kidney disease stage 4/5 who underwent symptom-limited treadmill ESE over 5-year mean follow-up. Exercise capacity was measured by age and sex predicted metabolic equivalents (METs). The primary outcome was achievement of predicted METs with first MACE. Cox proportional hazard multivariable modelling was used to determine MACE predictors with transplantation treated as a time-varying covariate. We also performed secondary analysis using a 7 MET threshold. Results There were 106 MACE with an annual cumulative risk of 2.4%. During follow-up, 525 (58%) received transplantation. Achievement of predicted METs (48%) (hazard ratio (HR) 0.49, 95% confidence interval (CI) 0.29–0.82, p=0.007) and transplantation (HR 0.52, 95% CI 0.30–0.91, p=0.02) were independently associated with reduced MACE. Similar results were observed using a 7 MET threshold achieved by 734 (82%) patients. Patients achieving predicted METs had no difference in MACE regardless of subsequent transplantation (HR 0.78, 95% CI 0.32–1.92, p=0.59). Patients who achieved predicted METs and did not receive transplantation had similar outcomes to those that did not achieve predicted METs and received transplantation (HR 0.97, 95% CI 0.42–2.25, p=0.95). Conclusions Exercise capacity is associated with reduced long-term MACE in renal transplant candidates undergoing ESE for pre-transplant cardiovascular assessment. Achievement of age and sex predicted METs confers excellent prognosis independent of subsequent transplantation. Funding Acknowledgement Type of funding sources: None.

Abstract Background Indigenous populations globally are known to have lower revascularisation rates following acute coronary events and higher mortality partly due to inequitable access to specialised care like cardiac catheterisation. Whether these disparities persist when access is readily available is unclear. Purpose We compared the rates of percutaneous coronary intervention (PCI), cardiac surgery, 30-day and long-term all-cause mortality in Indigenous (Aboriginal and Torres Strait Islanders) and non-Indigenous Australians in Far North Queensland (FNQ) – a region with a large Indigenous population and 24/7 cardiac catheterisation facilities. Method All public patients in FNQ having their first inpatient angiogram from November 2012 to October 2019 were identified. The primary study outcomes were rates of PCI or cardiac surgery and all-cause mortality at 30 days and long term. Secondary study outcomes were significant left ventricular dysfunction (ejection fraction <50%) and valvular disease (moderate to severe) in the echocardiogram subset. Other differences in baseline characteristics, including age, gender, body mass index, postcode and indication for angiography were accounted for using logistic and cox regression analysis. Results We identified 4489 patients (mean age, 61.7±13.0 years, 64.9% male, median follow-up 1045 days). 1042 (23.2%) self-identified as Indigenous. Indigenous patients were younger (53.7±11.6 vs 64.1±12.5 years, p<0.001), more likely female (45.5% vs 32.0%, p<0.001) and had small differences in angiography indications, ST elevation myocardial infarction (STEMI) 19.1% vs 18.1%, non-STEMI 45.7% vs 41.8%, angina 26.3% vs 28.0%, cardiac arrest 3.1% vs 3.7% and other 5.8% vs 8.4%, p=0.02. Rates of PCI or surgery 35.6% vs 38.5%, p=0.17, 30-day mortality 1.9% vs 2.7%, p=0.17 and long-term mortality 11.0% vs 11.5%, p=0.71 were similar in unadjusted data. 2959 patients (mean age, 62.1±13.0 years, 23.1% Indigenous, 64.9% male) were included in the echocardiogram subgroup. In unadjusted data Indigenous patients had similar rates of ventricular dysfunction 33.3% vs 31.3%, p=0.33 and valvular disease 19.4% vs 19.3%, p=0.93. After adjustment for other baseline characteristics, Indigenous patients had higher rates of PCI or cardiac surgery, OR 1.39 (95% CI, 1.18–1.64, p<0.001), ventricular dysfunction, OR 1.31 (95% CI, 1.07–1.60), p=0.01 and valvular disease, OR 1.93 (95% CI, 1.50–2.48), p<0.001. 30-day mortality was similar but Indigenous patients had higher adjusted long-term hazard of mortality, HR 1.80 (95% CI, 1.42–2.27), p<0.001. Conclusion When cardiac catheterisation was readily available Indigenous patients had higher rates of PCI and cardiac surgery and similar 30-day mortality to non-Indigenous patients. Equitable access to healthcare improves outcomes but the nearly double long-term mortality of Indigenous patients shows more is required to help close the gap for disadvantaged populations. Funding Acknowledgement Type of funding source: None

We present the first case of triple coronary artery ectasia in a previously well Indigenous Australian presenting with ST-elevation myocardial infarction. A 26 year old previously well Indigenous Australian male was admitted with anterior ST elevation myocardial infarction associated with an out of hospital ventricular fibrillation arrest. The patient was loaded with dual antiplatelet therapy and thrombolysed with 50mg tenecteplase prior transfer to our tertiary centre for rescue percutaneous coronary intervention due to <50% ST segment resolution and ongoing chest pain. Coronary angiography demonstrated thrombotic occlusion of the proximal left anterior descending artery with prominent vascular ectasia of all three arteries. Thrombus aspiration was not performed due to perceived high risk of systemic cardioembolic events. PCI was not attempted due to significant vascular ectasia preventing stent apposition within the grossly dilated vessel wall. He was managed with surgical thrombectomy and coronary artery bypass graft of the left anterior descending artery. There is no consensus approach in the management of isolated coronary artery ectasia presenting with acute coronary syndrome due to the paucity of prospective trials in its management. The management principles currently are to administer aspirin and to relieve the underlying obstructive lesion by percutaneous coronary intervention and failing that, cardiac surgery.

Indigenous populations globally have a higher burden of cardiovascular disease and increased mortality after acute coronary events, partly due to inequitable access to specialised care like cardiac catheterisation. Gender differences in revascularisation rates have been well described in non-Indigenous patients. Whether this applies to Indigenous patients when cardiac catheterisation facilities are readily available is unclear. We compared the rates of percutaneous coronary intervention (PCI), cardiac surgery, 30-day and long-term all-cause mortality in Indigenous (Aboriginal and Torres Strait Islanders) patients in Far North Queensland (FNQ) – a region with a large Indigenous population and 24/7 cardiac catheterisation facilities. All patients who presented to the tertiary referral center for FNQ, for their first inpatient angiogram between November 2012 and October 2019 were identified. The primary study outcomes were rates of PCI or cardiac surgery and all-cause mortality measured at 30 days and long term. Secondary study outcomes were significant left ventricular dysfunction (ejection fraction <50%) and valvular disease (moderate to severe) in patients who had an echocardiogram. Other differences in baseline characteristics, including age, gender, body mass index, postcode and indication for angiography were accounted for using logistic and cox regression analysis. 1042 patients (mean age 53.7±11.6 years, 45.5% female, median follow-up 1092 days) self-identified as Indigenous. Indigenous women were older 54.8±11.4 vs 52.8±11.7 years, p=0.005 and had different angiography indications. For Indigenous women and men respectively, rates of ST elevation myocardial infarction (STEMI) were 14.6% vs 22.9%, non-STEMI 44.3% vs 46.8%, angina 32.7% vs 21.0%, cardiac arrest 2.7% vs 3.3% and other 5.7% vs 6.0%, p<0.001. Indigenous women had significantly lower rates of PCI or cardiac surgery, 40.5% vs 60.7%, p<0.001, but similar 30-day mortality, 1.5% vs 2.3% p=0.34 and long-term all-cause mortality rates 11.2% vs 10.9%, p=0.89, in unadjusted data. 685 patients (mean age 53.8±11.5 years, 45.5% female) were included in the echocardiogram subgroup. Indigenous women had significantly more valvular disease, 23.3% vs 16.3%, p=0.022 but similar rates of left ventricular dysfunction, 30.2% vs 35.8%, p=0.12. Following adjustment for other baseline characteristics female gender independently predicted lower rates of PCI or cardiac surgery, OR 0.49 (95% CI 0.38–0.64) and higher rates of valvular disease, OR 1.60 (95% CI 1.07–2.39). Rates of ventricular dysfunction, 30-day and long-term all-cause mortality were similar. Indigenous women had significantly different indications for angiography, lower rates of PCI or cardiac surgery and higher rates of clinically significant valvular disease despite presenting in gender ratios similar to the general population in FNQ. Type of funding source: None

Superior vena cava (SVC) syndrome is a rare but documented complication of ascending aortic dilatation. In the past 30 years, there have only been 17 reports of this occurrence internationally. Here, we describe a case of chronic SVC syndrome due to ascending aortic pseudoaneurysm from previous aortic cannulation to highlight the importance of recognising this condition clinically. A 72 year old man was admitted with a five week history of exertional dyspnoea and facial swelling, on the background of previous coronary artery bypass graft surgery five years prior. On examination, there was plethoric facial swelling with markedly distended external jugular veins. There were no clinical signs of cardiac failure and his heart sounds were dual with no added sounds or murmurs. Chest X-Ray revealed a widened mediastinum without features of left ventricular failure or any mass lesions. Transthoracic echocardiogram showed normal left ventricular ejection fraction (65%) and a severely dilated ascending aorta (8.1cm). Computed tomography thoracic angiography confirmed the presence of focal aortic dilatation measuring 7.8cm by 8cm due to a localised Type A aortic dissection. The dilated ascending aorta was noted to be compressing the SVC just superior to the right atrium. The patient underwent urgent re-do cardiothoracic surgery which confirmed the diagnosis of localised aortic dissection with areas of pseudoaneurysm formation. The patient underwent ascending aortic repair without complication and was discharged from hospital one week later.

Abstract Background Although LV hypertrophy and dysfunction are associated with atrial fibrillation (AF), AF often occurs in the absence of LV hypertrophy or reduced ejection fraction. The effect of subclinical LV dysfunction on AF has not been fully studied. Purpose We sought the association between subclinical LV dysfunction (measured with global longitudinal strain, GLS) and new-onset AF. Methods This observational study evaluated 531 consecutive patients (median age, 67 years [interquartile range, 56 to 78]; 56% male), without a history of AF who underwent strain echocardiography after cryptogenic stroke. The CHARGE-AF score was used to calculate the 5-year risk of developing AF. Standard echocardiographic parameters were measured, and speckle-tracking was used to measure LA (reservoir strain, pump strain, and conduit strain) and LV strain (GLS). A strain analysis was conducted using a dedicated software package, using R-R gating. The baseline clinical and echocardiographic parameters of the patients who developed AF and those who did not were compared. Results Over 2.5 years of follow-up, 61 patients (11%) had new-onset AF. Patients who developed AF were older, had a higher CHARGE-AF score, larger LA volume, worse LA strain, and worse GLS than those who did not. Areas under the receiver-operating curve for GLS (0.84) was comparable to CHARGE-AF (0.79), LA pump strain (0.83), and LA reservoir strain (0.85). In the nested Cox models, GLS demonstrated an independent and incremental predictive value over the clinical and LA parameters (Figure). Moreover, adding GLS to the combined clinical and LA parameters model resulted in a significantly improved reclassification (net reclassification improvement, 0.32; p=0.016). Importantly, the predictive value of GLS was confirmed in patients with abnormal LA volumes (LA volume index≥34ml/m2) but not in patients with normal LA volumes. Figure 1 Conclusion GLS is associated with new-onset AF, especially in patients with abnormal LA volumes. This effect is independent of and incremental to the clinical and LA parameters.

Background: Atrial fibrillation (AF) is the most common cardiac arrhythmia and a significant cause of morbidity. It is most prevalent in individuals 65+ years and those with comorbidity. One third of individuals with AF are asymptomatic and unaware of their elevated risk of complications. Therefore, a targeted, community-based screening programme is likely to be clinically- and cost-effective. Aim: To investigate the effectiveness of a community-based AF screening program utilising a hand-held single-lead ECG device (Remon RM-100) for AF detection in older individuals (65+ years). Methods: We screened 106 individuals with no prior diagnosis of AF and comprehensively assessed their risk and clinical profile at a baseline clinic visit. Participants recorded their heart rate and rhythm three times per day for two weeks using the self-activated hand-held ECG device. Results: Mean age of participants was 70 ± 4 years at baseline and 67% were female. Mean CHA2DS2-VASc score was 3 ± 1 indicative of high thromboembolic risk and almost 40% were at greater than 5% risk of developing AF within the next 5 years due to their risk factor profile. Overall, 31% of the cohort were obese, 8% had type 2 diabetes, 12% had sleep apnoea and almost 50% were taking antihypertensive medication. Mild cognitive impairment was identified in 35% (MoCA score >26). Following two weeks of screening, AF was detected in three individuals (2.8%). Conclusion: A targeted AF screening program has the potential for widespread application and may detect subclinical AF prior to symptoms and/or the occurrence of AF-related complications.

Abstract Background Left atrial strain in the reservoir phase (LASr) measures passive LA stretch and is a sensitive marker of left ventricular diastolic dysfunction (DD). However, reduced LASr has not been prospectively validated against clinical heart failure (HF) endpoints and its place in diastology evaluation is unclear. Aim We sought whether DD grades defined by previously validated ranges of LASr predicted incident HF and whether reclassifying indeterminate diastolic function based on reduced LASr could facilitate assessment of HF risk. Methods Community dwelling elderly subjects were recruited and underwent baseline clinical and echocardiographic assessment. Where imaging was suitable, speckle-tracking echocardiography assessed LASr and subjects were assigned DD grades based on published ranges: normal >35%, grade 1 24–35%, grade 2 19–24%, grade 3 <19%. Current ASE/EACVI recommendations were used to identify those with indeterminate function; LASr-defined DD (LASr-DD) was defined as LASr ≤23%. Follow-up was ≤2 years and incident HF adjudicated by Framingham criteria. Results Of 610 subjects (age 71±5 yrs., 46% male) LASr analysis was feasible in 590 (97%); average LASr was 39% (IQR 34–43%). Incident HF was associated with LASr-DD grade, occurring in 8 (36%) with grade ≥2, 14 (10%) with grade 1 and 39 (9%) with normal function (p<0.001). Adjusted odds ratio for incident HF for LASr-DD grade ≥2 was 3.12 (95% CI 1.06–9.1, p=0.038) Diastolic function was indeterminate in 147 (24%) subjects; of 144 (98%) with LAS analysis, 6 (75%) of those with LASr-DD vs. 15 (11%) with normal LASr experienced incident HF (p<0.001). Univariable Multivariable* OR (95% CI) p-value OR (95% CI) p-value LASr-DD grade: 1 1.13 (0.59–2.15) 0.7 0.84 (0.42–1.69) 0.63 ≥2 5.7 (2.26–14.5) <0.001 3.12 (1.06–9.1) 0.038 *Adjusted for age, hypertension, diabetes, BMI, global longitudinal strain, E/e', LA volume index, LV mass index (all p<0.1 on univariable analysis). Incorporating LA strain in practice Conclusion DD defined by LASr is predictive of HF for grades ≥2 independent of other diastolic measures. Indeterminate diastolic function with LASr ≤23% is associated with incident HF. LASr may complement current diastolic function assessment recommendations. Acknowledgement/Funding Baker Heart and Diabetes Institute

Background: Undiagnosed atrial fibrillation (AF) is common, and screening for AF has been proposed to permit early diagnosis and aggressive risk factor modification. The presence of regional variation might provide a rationale for targeting these efforts. It was hypothesised that there would be differences in AF detection in the community in Tasmania and Victoria.

A healthy Taiwanese 28-year-old male presented with 1 month of dyspnoea, non-productive cough and night sweats that were unresponsive to multiple courses of antibiotics. On presentation he was febrile, tachycardic and tachypnoeic, with an elevated jugular venous pressure, systolic murmur and reduced air entry at both lung bases. His C-reactive protein (120 mg/L) and lactate dehydrogenase (LDH) (443 U/L) were elevated. Large bilateral pleural effusions were confirmed on chest x-ray. Urgent echocardiography demonstrated a large pericardial effusion with features of tamponade. The right ventricle (RV) was also noted to be dilated. Successful pericardiocentesis was performed without improvement in tachycardia or inferior vena cava (IVC) dimension.

Objective: Low socioeconomic status (SES) is associated with cardiovascular diseases, and an association with atrial fibrillation (AF) could guide screening. This study investigated if indices of disadvantage (IAD), education or occupation (IEO), and economic resources (IER) were associated with incident AF, independent of risk factors and cardiac function.

Background: Subclinical left ventricular (LV) function can be recognised by reduced global longitudinal strain (GLS). This can cause left atrial (LA) remodelling, which is associated with the pathogenesis of atrial fibrillation (AF). We investigated if reduced LV strain was associated with AF.

Background: Despite multi-parametric assessment, diastolic dysfunction (DD) may be designated as indeterminate. Left atrial strain-defined DD (LAS-DD) is a single measure, cut-offs for which have been validated for DD grade discrimination using 2009 American Society of Echocardiography (ASE) guidelines. However, LAS-DD grades have not been validated against outcomes nor compared with 2016 ASE recommendations.

Background: Atrial fibrillation (AF) is a leading cause of stroke in the elderly. Screening programs using portable ECG monitoring devices are cost effective. However, the practical aspects of implementing a program in elderly patients has not been studied. We implemented a pilot community AF screening program and explored some of the challenges that may be experienced with mass screening. Methods: 169 participants (mean age 70.1 ± 4.2yrs, males 53%) ≥65 years from the community (with 1 or more risk factors for AF) were followed up for incident AF. Patients were asked to use a portable ECG monitoring device (Remon) to perform 5 × 1 min recordings/day for 1 week. Patients were referred to their local doctor for anticoagulation if AF was diagnosed. We recorded compliance rates and technical issues. A questionnaire addressing the practical aspects of the device was sent to all participants. Results: AF was diagnosed in 21 participants (13%). 42% of participants completed less than 5 recordings/day. 17 (10%) contacted research staff due to technical issues with the device (4 reported that the device had stopped working). 77/169 questionnaires were completed (46%). 6 (8%) reported that the ECG device was difficult to use. 24 (31%) reported difficulty with holding the device to perform recordings and 24 (31%) had some degree of anxiety regarding the results. Conclusion: Although a mass screening program for AF is attractive and cost effective, there remains additional challenges in the elderly population which must be considered including manual dexterity, cognitive impairment, patient anxiety and use of technology.

Background: Left atrial(LA) and left ventricular(LV) function can be assessed using strain imaging. We investigated if reduced LA strain was associated with AF. Methods: 162 participants ≥65 years from the community (with 1 or more risk factors for AF) were followed up for incident AF(median 12 months). Echocardiography was done and strain analysis was conducted using external software (Image-Arena, TomTec). AF was diagnosed by 12 lead ECG during outpatient clinics or using a portable ECG monitoring device (5 × 1 min recordings over a 1 week period). Patients with incident AF (device and clinical diagnosis) were compared to those in sinus rhythm. Logistic regression analysis was used to assess the association between LA and LV strain with AF. Results: 42(26%) with mean age 71.7 ± 5.2yrs,male 69% were diagnosed with AF (19(12%) using portable ECG monitoring). Those with AF were older and had higher clinical AF risk(p < 0.05). Patients with AF had reduced ejection fraction (EF),reduced GLS and increased LA volume(p < 0.05). LA strain was similar in all groups(p > 0.05). There was no differences in demographics/echo parameters between those with device or clinically diagnosed AF. Increased age (OR 1.1,95%C.I 1.0-1.2), male gender(OR 2.2,95%C.I 1.1-4.7), reduced GLS (OR 1.24,95%C.I 1.1-1.4), increased LA volume(OR 1.04,95%C.I 1.0-1.1) and higher CHARGE-AF score (OR 1.1,95%C.I 1.0-1.2) were all associated with AF(p < 0.05). Reduced LA strain was not associated with increased risk of AF(p > 0.05). Conclusion: Abnormal LV function, not LA function is more predictive of incident AF.

Introduction: The CHARGE-AF score gives a 5-year atrial fibrillation(AF) risk assessment. Heart Failure(HF) risk can be assessed using the ARIC score. We aimed to compare the CHARGE-AF score to the ARIC score in prediction of HF and AF. Methods: A community based study of 503 participants(mean ± SD age 70.8 ± 4.7yrs,male 48% with median follow up 12 months) ≥ 65 years were recruited if they had presence of 1 or more risk factor. HF and AF risk was assessed using the CHARGE-AF and ARIC scores. ECG/echocardiography was performed and HF was diagnosed as per ESC guidelines. AF was diagnosed by local doctors, by 12 lead ECG during outpatient clinics or using a portable ECG monitoring device(5 × 1 min daily recordings over 1 week). ROC curves were compared between both scores using the Hanley and McNeil method. Results: During the follow-up period 55 patients developed HF. 173/503 participants completed follow-up for AF of which 43(25%) were newly diagnosed. Patients with HF were older with higher rates of diabetes and hypertension(p < 0.05). Patients with AF were older and had higher CHARGE-AF score(p < 0.05). In patients with HF and AF, echocardiography showed impaired global longitudinal strain and increased LA volume(p < 0.05). For HF and AF, there was only modest discriminative ability with no significant difference in the AUC between both scores(Table). Conclusion: The CHARGE-AF score is a useful clinical tool in predicting both HF and AF in patients with risk factors.Table 1AUC CHARGE-AF95% C.I p ValueAUC ARIC95% C.I p ValueComparison of AUC (p)HF0.650.58-0.73, p < 0.0010.650.58-0.73, p < 0.0010.89AF0.610.51-0.71, p = 0.040.590.48-0.69, p = 0.110.77 Open table in a new tab

Background: Atrial fibrillation (AF) is a major risk factor for ischemic stroke and anticoagulation therapy (vitamin K antagonists (VKA) and non-vitamin K oral anticoagulants (NOACs)) is recommended for primary and secondary stroke prophylaxis in high-risk AF patients.However, little is known regarding pre-and poststroke antithrombotic treatment pattern in prevalent AF patients presenting with ischemic stroke.Methods: Using Danish nationwide registries all prevalent atrial fibrillation patients in Denmark with a pre-stroke CHA2DS2-VASc >1 who presented with an ischemic stroke from January 2004 to September 2016 were identified.Antithrombotic therapy 180 days prior to stroke hospitalisation was assessed.In those who survived until 90 days after discharge post-stroke antithrombotic therapy was assessed.Results: 25,789 patients comprised the study population; 8779 (34.0%) received anticoagulation therapy with VKA or a NOAC (with or without antiplatelet agent), 10,338 (40.1%) received antiplatelet therapy alone, and 6672 (25.9%) did not receive any antithrombotic treatment prior to their stroke diagnosis.Those who did not receive anticoagulation therapy prior to their stroke diagnosis were in general older than those receiving anticoagulation therapy (median age 81 years, interquartile range (IQR) 72-87 and median age 79 years, IQR 72-84, respectively) and had similar risk scores for stroke as those treated with anticoagulation therapy (mean CHA2DS2-VASc 3.5, standard deviation (std) 1.1 and mean CHA2DS2-VASc 3.6, std 1.1, respectively).Post-stroke, 10,280 (51.1%) were not treated with anticoagulants. Conclusion:In AF patients with a CHA2DS2-VASc score >1 who presented with an ischemic stroke, 66.0% did not receive anticoagulant therapy prior to their stroke hospital admission as recommended by guidelines, while 51.1% did not receive anticoagulant therapy after discharge.Our findings suggest a substantial opportunity for improving primary and secondary stroke prophylaxis in AF patients.