Abstract 12220: Association of CT-derived Left Atrial Size with Major Cardiovascular Events in the General Population: The Heinz Nixdorf Recall Study

Abstract

Background: Computed tomography (CT) imaging of the heart is emergently used for risk stratification purposes. With the advent of cardiac CT, information on the left atrium (LA) is readily available and may improve prognostic implication of this imaging modality. Objective: We aimed to determine, whether LA size, as quantified from non-contrast enhanced cardiac CT, is associated with the incidence of major cardiovascular (CV) events in the general population.

Methods: Subjects aged 45-75 years were drawn from the population based Heinz Nixdorf Recall Study with enrollment between 2000 and 2003. For this analysis, we excluded participants with prevalent CV disease at baseline. LA size was quantified as the area at the level of the mitral valve and left ventricular outflow tract from non-contrast enhanced cardiac CT. Major CV events, defined as incident myocardial infarction, stroke, or CV death, were assessed during follow up. Cox-regression analysis was performed to determine association of LA size with events in unadjusted models, adjusting for traditional CV risk factors and ancillary for coronary artery calcification (CAC)-score.

Results: Overall, 3958 subjects (59.2±7.7 years, 53% female) were included in this analysis. Mean LA area was 17.64 ± 4.22 cm² (range: 7.16 - 44.13 cm²). During a follow-up period of 8.0 ± 1.5 years, 221 subjects developed a major CV event. Subjects with CV events had significantly larger LA size at baseline (19.49 vs. 17.53 cm², p<0.0001). In univariate analysis, increase of LA size by 1 standard deviation was associated with nearly 50% excess events (HR (95%CI): 1.48 (1.32-1.65), p<0.0001). Associations remained statistically significant after adjustment for CV risk factors (HR (95%CI): 1.25 (1.09-1.43), p=0.001) and when further adjusting for CAC (HR (95%CI): 1.22 (1.07-1.40), p=0.004). Associations for LA size with CV events were stronger for low CAC-scores (HR (95%CI): CAC-score=0: 1.60 (1.08-2.39), CAC-score >400: 1.24 (1.01-1.51)).

Conclusion: CT-derived LA size is associated with incident major CV events independent of traditional risk factors and CAC-score in a European general population cohort. Our results encourage quantification of LA size from thoracic CT as it may help to improve risk prediction of CV events.