Embolic complications in a large contemporary cohort of infective endocarditis: do we need score model?

Abstract

Infective endocarditis (IE) is a heterogeneous disease undergoing epidemiological changes. Whether those changes have an impact on the correlates of embolic events (EE) remains unclear. We analyzed the correlates of EE and proposed a diagnostic score model in a large contemporary cohort. This is a retrospective observational study including patients with definite valve IE admitted between 2000 and 2023. EE were defined as acute complications causing overt clinical manifestations. The study primary aim was to identify independent correlates of EE. 715 valve-IE cases were included. EE occurred in 41.4% (n = 296) of patients. S. aureus etiology (OR 2.708[1.268-5.786]; p = 0.010), C-reactive protein > 6.7mg/dL (OR 2.415[1.371-4.252]; p = 0.002), and splenomegaly (OR 2.858[1.620-5.403]; p < 0.001) were independently associated with EE. VS ≥ 14mm (OR 1.575[0.925-2.682]; p = 0.061) and D-dimers > 747 ng/mL (OR 1.677[0.976-2.881]; p = 0.061) showed a trend for independent association. These variables were included in a diagnostic score model. A stepwise increase of EE occurrence was found stratifying patients into 3 categories (score 0-2-22%; score 3-5-53%; score 6-8-78%;p < 0.001). A cut-off of 2 (< 2 vs. ≥ 2) showed a sensitivity of 83% and a specificity of 50% (AUROC 0.732; p < 0.001). EE were independently associated to Staphylococcus aureus, C-reactive protein and splenomegaly and less strongly linked to vegetation size in our cohort. These results may be explained by a change in embolic complications correlates linked to the epidemiological shift. The discriminative ability of our score was only fair. At present, clinicians should rely upon clinical and imaging data to diagnose EE.