Bloodstream infections (BSIs) are life-threatening conditions with a high mortality rate [1, 2]. In the present prospective study, we evaluated a real-time polymerase chain reaction (PCR) approach for the rapid diagnosis of bacteremia caused by Staphylococcus aureus and the detection of methicillin resistance in 384 patients from the University Hospital of Ulm with clinical suspicion of BSI, including children and adults with hematological malignancies, various forms of severe immunodeficiency, and intensive care unit (ICU) patients. PCR results were compared to standard blood culture results.

Criteria for the inclusion of patients in the study were as follows: ICU patients with systemic inflammatory response syndrome (SIRS) or sepsis [3], hematology/oncology patients with fever >38.5°C in neutropenia, or patients with other forms of hereditary or acquired immunodeficiency and fever >38.5°C. Within one infectious episode (IE), subsequent samples were included only when obtained within 7 days of sampling of the initial sample. A patient was eligible for the study a second time if the new IE occurred more than 4 weeks after the initial inclusion of the patient into the study and if a blood culture (BC)-confirmed infection with an identical species had been ruled out. The study was approved by the ethics committee of the University of Ulm (175/07). Blood samples for PCR were collected in 5-, 2.5-, or 1.4-ml citrate tubes. The blood collection team was advised to obtain the citrate blood sample and a 20-ml blood sample for the inoculation of a pair of aerobic and anaerobic BACTEC™ Plus/F blood cultures (in children BACTEC™ PED bottle only) from the same venipuncture. DNA isolation from blood samples was performed Monday to Friday twice daily with the MolYsis Complete5 kit (Molzym, Bremen, Germany), allowing the digestion of human DNA by using a DNAse and enrichment of bacterial DNA [4, 5]. The kit was used as indicated by the manufacturer with following adaptations: (i) volumes of buffers CM and DB1 were adapted to the sample volume: 250 μl of each buffer was added to 1-ml blood samples, 800 μl each to 2-ml samples, and 2 ml each to 5-ml samples; (ii) before incubation with BugLysis® solution at 37°C for 30 min (contained in the kit), 1.4 μl β-mercaptoethanol was added. The isolated DNA was stored at −20°C until use in the PCR. With each kit, a water ‘negative’ control of DNA isolation was investigated. DNA isolation was done under a HEPA-filtered PCR working bench, decontaminated daily by UV radiation, and strictly separated from PCR working places. For the detection of Staphylococcus aureus, a nuc gene-specific real-time PCR was run on a LightCycler® [6]. For the detection of the mecA gene, a previously published real-time PCR assay was used [7].

A total of 902 blood samples from 384 patients (329 adults and 55 children <18 years of age) were included in the study. The samples were obtained in 468 IEs (379 from adults and 89 from children). In 197 IEs, a single blood sample, in 126 IEs two, in 134 IEs three, and in 11 IEs four or more blood samples were obtained.

In 12 of the 902 blood cultures, S. aureus was cultured. The S. aureus-specific PCR was positive in eight of the 12 corresponding blood samples. In all BC-positive but PCR-negative samples, PCR inhibition was excluded by spiking of the sample with 1 pg/µl S. aureus DNA. The 12 positive BCs were obtained from six patients (Table 1). Defining a PCR-positive patient as a patient with a positive PCR result in at least one blood sample of all samples obtained within the same IE, four of the six BC-positive patients were positive in the S. aureus-specific PCR (Table 1), corresponding to a sample and IE-related sensitivity of 66.6%. In all PCR-negative but BC-positive isolates, presence of the nuc gene was confirmed by PCR from the cultured isolate. While the growth of S. aureus in BCs was detected one to two days after the arrival of the BCs in the laboratory, i.e., after a mean incubation time of 13.5 h in the BACTEC™ system, the PCR result can already be obtained within 4 h, including DNA preparation and PCR analysis. Thereby, PCR analysis enables significantly earlier administration of pathogen-specific antimicrobial therapy and may, thus, have a positive impact on the outcome of the patient. Compared to the gold standard BC diagnostics, the sensitivity of the PCR assay was quite low. It was, however, comparable to the sensitivity of a recently published real-time TaqMan PCR assay for the detection of S. aureus in whole blood samples of ICU patients [8]. While Peters et al. [8] took a sample volume of only 200 μl of whole blood for DNA isolation, we used a volume of 5 ml in adults and 2.5 ml or 1.4 ml in children in order to increase sensitivity. In addition, for DNA isolation, the new kit MolYsis Complete5 was used, which allows the enrichment of bacterial DNA [4, 5]. Nevertheless, the sensitivity of the PCR approach was still lower than the sensitivity of BCs. This lower sensitivity may be caused by the larger blood volume contained in the BC bottles compared to the volume investigated in one PCR analysis, which involves a 1 to 20 dilution of the blood sample for PCR. Considering a staphylococcal concentration in patients with BSI of 1–30 CFU per ml of blood [9], this low concentration will be sufficient for obtaining growth in BC bottles, but may be below the sensitivity of a PCR assay, despite a sensitivity of the S. aureus-specific PCR assay used in this study of 1 CFU/PCR reaction (data not shown).

Table 1 Characteristics and polymerase chain reaction (PCR) results of patients with Staphylococcus aureus-positive blood cultures
Full size table

In addition to an earlier detection of the causative agent of BSI, the PCR approach used in this study also allowed the determination of methicillin resistance in all cases of MRSA BSI which were positive in the S. aureus PCR. It was correctly negative in patients with BSI caused by methicillin-susceptible isolates (Table 1). PCR was falsely negative in one patient with BC-confirmed but S. aureus PCR-negative MRSA infection, suggesting a low concentration of bacterial DNA in the sample. Since the mecA gene targeted by the PCR can also be found in methicillin-resistant coagulase-negative Staphylococcus (CNS), a false-positive identification of MRSA is possible in samples containing both methicillin-susceptible S. aureus (MSSA) and methicillin-resistant CNS, such as patient no. 812. To avoid false-positive and false-negative results, the mecA gene PCR should only be done in samples which are positive in the S. aureus PCR.

A positive S. aureus PCR result was also obtained in 13 of 890 BC-negative samples and in ten of 462 BC-negative IE of ten patients, respectively (Table 2). In the dependency of the clinical condition of the patient and the results of other microbiological specimens, we judged the PCR result as ‘likely’ or ‘possible,’ or ‘indefinite.’ In patients judged as ‘likely’ (n = 3), a focus of S. aureus infection was documented (Table 2). The negative mecA gene PCR in patient no. 241 may be explained by the lower sensitivity of the mecA gene PCR (50 CFU/PCR [10]) compared to the S. aureus PCR (1 CFU/PCR). Out of the remaining seven culture-negative but PCR-positive patients, all but one (no. 263) had clinical infection without the detection of other pathogens. Six patients were receiving systemic antimicrobial therapy with efficacy against most strains of MSSA at the day of blood sampling and the last 48 h before, and one patient was treated with broad-spectrum antimicrobials until two days before the day of blood sampling (Table 2). Therefore, at least in some of these patients, the detection of S. aureus DNA may be attributed to the presence of bacteria in the bloodstream that did not grow in the simultaneously obtained BCs due to the administered antimicrobial therapy. In addition, the detection of cases of S. aureus bacteremia by PCR only has already been described earlier [8]. In patient no. 263, bacteremia by Serratia marcescens was detected, but cross-reaction of the S. aureus-specific PCR was excluded since a eubacterial PCR followed by sequencing (SepsiTest™, Molzym, Bremen, Germany) confirmed the presence of S. aureus DNA. The mecA gene PCR was negative in all samples of the ten BC-negative patients.

Table 2 Characteristics of S. aureus blood culture-negative, PCR-positive patients
Full size table

In conclusion, the PCR approach evaluated in this study is a valuable additional tool for the rapid detection of S. aureus BSI and methicillin resistance in patients at risk for S. aureus bacteremia. Nevertheless, due to the lower sensitivity, it cannot replace BCs in the microbiological workup of suspected BSIs. The main limitation of our study is the small number of BC-documented S. aureus BSIs. Therefore, the diagnostic utility and the clinical benefit of a rapid PCR diagnostic should be investigated in further large studies.