Abstract
BACKGROUND: Rapid detection of bloodstream infections (BSIs) can be lifesaving. We investigated the sample processing and assay parameters necessary for highly-sensitive detection of bloodstream bacteria, using Staphylococcus aureus as a model pathogen and an automated fluidic sample processing-polymerase chain reaction (PCR) platform as a model diagnostic system. METHODOLOGY/PRINCIPAL FINDINGS: We compared a short 128 bp amplicon hemi-nested PCR and a relatively shorter 79 bp amplicon nested PCR targeting the S. aureus nuc and sodA genes, respectively. The sodA nested assay showed an enhanced limit of detection (LOD) of 5 genomic copies per reaction or 10 colony forming units (CFU) per ml blood over 50 copies per reaction or 50 CFU/ml for the nuc assay. To establish optimal extraction protocols, we investigated the relative abundance of the bacteria in different components of the blood (white blood cells (WBCs), plasma or whole blood), using the above assays. The blood samples were obtained from the patients who were culture positive for S. aureus. Whole blood resulted in maximum PCR positives with sodA assay (90% positive) as opposed to cell-associated bacteria (in WBCs) (71% samples positive) or free bacterial DNA in plasma (62.5% samples positive). Both the assays were further tested for direct detection of S. aureus in patient whole blood samples that were contemporaneous culture positive. S. aureus was detected in 40/45 of culture-positive patients (sensitivity 89%, 95% CI 0.75-0.96) and 0/59 negative controls with the sodA assay (specificity 100%, 95% CI 0.92-1). CONCLUSIONS: We have demonstrated a highly sensitive two-hour assay for detection of sepsis causing bacteria like S. aureus directly in 1 ml of whole blood, without the need for blood culture.