Atrial cardiopathy (AC), which refers to left atrial (LA) structural and functional disorders (independent of atrial fibrillation [AF]) [1,2], is increasingly considered a potential mechanism for embolic stroke of undetermined source (ESUS) [3]. AC has been independently associated with silent AF detection and stroke recurrence [4,5], suggesting an etiopathogenic role. Most studies on AC in ESUS were conducted before the recently proposed ESUS construct update [6]. Given the heterogeneous definition of ESUS, etiological misclassification may have limited our understanding of the link between AC and ESUS. After applying the proposed ESUS construct update, we first assessed whether any differences existed in the prevalence of AC in patients with ESUS classified according to the traditional versus revised criteria. After focusing on the revised classification, we investigated the clinical and radiological differences between ESUS with AC (AC(+)/ESUS) versus without AC (AC(-)/ESUS). Additionally, we investigated the association between AC and stroke severity and outcome and the role of AC in stroke recurrence and AF detected after stroke (AFDAS) [7].

This retrospective single-center study included all consecutive patients with acute ischemic stroke (AIS) diagnosed as ESUS (according to standard criteria) [8]. These patients were admitted to our stroke unit between January 2018 and December 2022. All diagnostic evaluations were reviewed for each patient, and recently proposed changes to the ESUS construct [6] were applied to redefine the ESUS classification. We excluded patients with (1) high-risk patent foramen ovale (PFO), (2) high-risk non-stenosing (<50%) ipsilateral supracardiac atherosclerosis, and (3) probable cancer-related hypercoagulability (see Supplementary Methods for details).

AC was defined as LA enlargement (LAE), measured using the LA volume index (LAVI) based on the standard criteria: LAVI >34 mL/m². The entire cohort was divided based on the presence of AC (AC(+)/ESUS if LAVI >34 mL/m²) versus its absence (AC(−)/ESUS if LAVI ≤34 mL/m²). AC was also categorized according to severity as mild (LAVI 35-41 mL/m²) or moderate/severe (LAVI ≥42 mL/m²) [9].

Radiological data, including the analysis of stroke lesions by location and site, were also collected. Stroke severity (measured by the baseline National Institutes of Health Stroke Scale [NIHSS] score) and functional status at 90 days (modified Rankin Scale [mRS] score of 0-2 and 0-3) were considered as clinical outcomes. Stroke recurrence and AFDAS were considered long-term follow-up outcomes.

Statistical analysis was performed using Stata statistical software (Version 17; StataCorp., College Station, TX, USA), with the significance level set at P<0.05. Univariate and multivariate logistic (or ordered logistic) regression analyses were performed to evaluate the association between AC (considered both continuous [LAVI] and dichotomous variables [AC(+) vs. AC(-); moderate/severe AC vs. mild AC/AC(-)]) and stroke severity, 90-day mRS score, stroke recurrence, and AFDAS. This study was approved by the local ethics committee (Comitato Etico Milano Area 3, n. 346-18052022), and informed consent was obtained from patients upon admission. Detailed information regarding the study population, diagnostic evaluations, and statistical analyses can be found in Supplementary Methods.

Among the 414 eligible ESUS patients (with available LAVI measurements), 116 (28%) were reclassified and excluded per the ESUS construct update, resulting in a final sample of 298 ESUS patients (Supplementary Figure 1). The prevalence of AC was higher in ESUS cases classified according to the revised criteria than in those classified according to the traditional criteria (42.0% vs. 36.2%) and significantly different from excluded ESUS cases (42.0% vs. 21.5%; P<0.001) (Table 1). The excluded patients with ESUS had a lower LAVI, were younger with fewer vascular risk factors, experienced milder strokes, and had better 3-month outcomes (Supplementary Table 1). The general characteristics of the final ESUS cohort (revised criteria) are presented in Table 2. Patients with AC(+)/ESUS were older and had more hypertension, coronary artery disease, and non-stenosing ipsilateral supra-cardiac atherosclerosis. Additionally, they suffered more frequently from cortico-subcortical strokes and had fewer small-isolated cortical lesions compared to patients with AC(-)/ESUS.

Over a median follow-up of 20 months (interquartile range [IQR] 8-32; available for 290 patients), recurrent stroke occurred in 17 patients (5.9%) and AFDAS in 28 patients (9.7%). The median time between the index and recurrent stroke was 5 months (IQR 3-20). In both univariate and multivariate logistic regression analyses, no significant associations were observed between the AC parameters (presence, severity, and LAVI), stroke severity, 90-day mRS, and stroke recurrence. The AFDAS was independently associated with all AC parameters (Table 3). Further analyses and results are reported in the Supplementary Results, including Kaplan-Meier survival analysis for the risk of stroke recurrence (Supplementary Figure 2).

In our study, we found that, following the recent ESUS update, patients classified as non-ESUS exhibited a significantly different echocardiographic profile than ESUS patients, with a higher AC incidence in the latter group (42.0%). This finding highlights the substantial heterogeneity in the echocardiographic profiles within these two groups and underscores the importance of precise ESUS patient classification.

The prevalence of AC varies in published studies, depending on the criteria used for its definition. Various AC biomarkers, categorized as electrophysiological, structural, hemodynamic, and serological, have been associated with stroke risk [2]. Our study defined AC as LAE by measuring the LAVI, which is now considered a superior indicator of LA dimensions compared to LA diameter. The LAVI has been demonstrated to be better associated with new-onset AF [4] and stroke recurrence [10] in patients with ESUS.

Consistent with previous studies, our findings indicate that patients with AC tended to be older and have a higher atherosclerotic burden. This finding aligns with the pathogenetic evidence indicating that LAE results from progressive cardiac wall remodeling due to aging, inflammation, oxidative stress, and stretching from pressure and volume overloads [1,2].

Additionally, the infarction pattern differs; AC(+)/ESUS exhibits more cortical-subcortical infarcts and fewer small isolated cortical lesions, suggesting a potential connection to the formation of larger thrombi in the larger left atria. Our study revealed no significant differences in stroke recurrence rates. However, given the limited sample size and recurrence rates, our results may be underpowered to draw meaningful conclusions. Notably, we found that AC was independently associated with AFDAS. This finding aligns with those of previous studies [4,7] and supports the adoption of the LAVI in future trials assessing the role of anticoagulant therapy in selected patients with ESUS at high risk of AFDAS.

Our study had several strengths. First, we focused on a carefully “selected” ESUS population, following the recently proposed update, and utilized LAVI measurement as a superior marker of LAE. However, acknowledging certain limitations is important. This was a retrospective, observational, single-center study. We restricted our analysis to structural cardiopathy, defining AC as LAE without measuring other markers of LA dysfunction. Further studies should include serological and electrophysiological biomarkers to ensure a more comprehensive evaluation of AC in ESUS. Finally, the initiation of anticoagulation therapy might have occurred after AFDAS, potentially influencing the observed stroke recurrence rate.

Despite these limitations, our study provides valuable insights for a deeper understanding of the role of AC in ESUS. Although we found an independent association between AC and AFDAS, no significant associations were observed with stroke severity, 90-day outcome, and stroke recurrence. Considering the recent failure of the ARCADIA trial (NCT03192215), our results may prove instrumental in the design of future trials aimed at demonstrating the benefits of anticoagulation therapy in these patients.