Aims: Current guidelines recommend echocardiography in all episodes of Staphylococcus aureus bacteraemia (SAB). This study aimed to determine whether a very low-risk group of patients with SAB could be found in whom echocardiography was of no incremental diagnostic value for the diagnosis of infective endocarditis.
Methods: Using the ANZCOSS dataset, we identified 574 eligible episodes of adult SAB at Auckland District Health Board (ADHB) between 2007 and 2012, and retrospectively obtained additional microbiological and clinical data. Prevalence of IE was determined using the modified Duke's criteria for diagnosis of IE. Multivariate logistic regression analysis was used to determine whether risk factors were independently associated with IE, and we also assessed their negative predictive value (NPV).
Results: Transthoracic and/or transoesophageal echocardiography was performed in 370 (65%) episodes of SAB. The prevalence of clinically definite and clinically possible IE was 5.6% and 8.5%, respectively. Thirty day all-cause mortality was 11.7%. The factors with the highest NPV when absent in hospital-acquired SAB were non central venous line-associated bacteraemia (100%), persistent bacteraemia (96%), and presence of a prosthetic valve or cardiac rhythm management (CRM) device (95%). When none of these three criteria were present the NPV was 100% (99-100%).
Conclusions: A group of very low risk patients was found in our study: central line-associated SAB without prosthetic valves / CRM devices and without persistent bacteraemia. These patients had no episodes of IE and echocardiography is of no incremental diagnostic benefit.