Hurtado-Guerra, Juan José

Abstract Background Infective endocarditis (IE) secondary to Staphylococcus aureus bacteremia (SAB) has high morbidity and mortality. The systematic use of echocardiography in SAB is controversial. We aimed to validate VIRSTA and Predicting Risk of Endocarditis Using a Clinical Tool (PREDICT) scores for predicting the risk of IE in Colombian patients with SAB and, consequently, to determine the need for echocardiography. Methods Cohort of patients hospitalized with SAB in 2 high complexity institutions in Medellin, Colombia, between 2012 and 2018. The diagnosis of IE was established based on the modified Duke criteria. The VIRSTA and PREDICT scores were calculated from the clinical records, and their operational performance was calculated. Results The final analysis included 922 patients, 62 (6.7%) of whom were diagnosed with IE. The frequency of IE in patients with a negative VIRSTA scale was 0.44% (2/454). The frequency of IE in patients with a negative PREDICT scale on day 5 was 4.8% (30/622). The sensitivity and negative predictive value (NPV) of the VIRSTA scale was 96.7% and 99.5%, respectively. For the PREDICT scale on day 5, the sensitivity and NPV were 51.6% and 95.1%, respectively. The discrimination, given by the area under the receiver operating characteristic curve, was 0.86 for VIRSTA and 0.64 for PREDICT. Conclusions In patients with negative VIRSTA, screening echocardiography may be unnecessary because of the low frequency of IE. In PREDICT-negative patients, despite the low frequency of IE, it is not safe to omit echocardiography.