Search

Management of complicated bloodstream infections requires more aggressive treatment than uncomplicated bloodstream infections. We assessed the value of follow-up blood culture in bloodstream infections caused by Staphylococcus aureus , Enterococcus spp., Streptococcus spp., and Candida spp. and studied the value of persistence of DNA in blood (using SeptiFast) for predicting complicated bloodstream infections. Patients with bloodstream infections caused by these microorganisms were enrolled prospectively. After the first positive blood culture, samples were obtained every third day to perform blood culture and SeptiFast analyses simultaneously. Patients were followed to detect complicated bloodstream infection. The study sample comprised 119 patients. One-third of the patients developed complicated bloodstream infections. The values of persistently positive tests to predict complicated bloodstream infections were as follows: SeptiFast positive samples (sensitivity, 56%; specificity, 79.5%; positive predictive value, 54%; negative predictive value, 80.5%; accuracy, 72.3%) and positive blood cultures (sensitivity, 30.5%; specificity, 92.8%; positive predictive value, 64%; negative predictive value, 75.5%; accuracy, 73.9%). Multivariate analysis showed that patients with a positive SeptiFast result between days 3 and 7 had an almost 8-fold-higher risk of developing a complicated bloodstream infection. In S. aureus , the combination of both techniques to exclude endovascular complications was significantly better than the use of blood culture alone. We obtained a score with variables selected by the multivariate model. With a cutoff of 7, the negative predictive value for complicated bloodstream infection was 96.6%. Patients with a positive SeptiFast result between days 3 and 7 after a positive blood culture have an almost 8-fold-higher risk of developing complicated bloodstream infections. A score combining clinical data with the SeptiFast result may improve the exclusion of complicated bloodstream infections.