Bloodstream infections S245 (78%, 144/185 strains versus 22%, 41/185 strains) while CoNS were the most common contaminants. The mean detection time for all microbes (435 isolates) was 24.13 h (range 2.88 h to 122 h) and 72% (313/435) of them were detected within the ﬁrst 24 h of incubation. The mean detection times for the Gram-negative and Gram-positive bacteria were 31.08 h and 21.9 h respectively (E. coli 9.3 h, Proteus spp. 9.2 h, Klebsiella spp. 9.2 h, P. aeruginosa 14.2 h, Brucella melitensis 65.39 h, S. aureus 15.5 h, CoNS 23.6 h, S. pneumoniae 10 h, Enterococcus spp. 19.8 h, Streptococcus spp. 36.6 h, Candida spp. 15.6 h). Conclusions: The recovery time of Enterobacteriaceae (9.2 h) was signiﬁcantly shorter of Gram-positive bacteria and 85% of Enterobac- teriaceae were isolated within the ﬁrst 24 h. All pathogenic microbes, including Brucella spp. were recovered within 5 days of incubation and the detection times of clinically signiﬁcant isolates were shorter than contaminants. A protocol of 5-days incubation period is possible to apply with the BACTEC 9120. P936 Is there any relation between time to positivity in blood cultures and source of infection in patients with Enterococcus spp. bacteraemia? J. Alava, C. Ezpeleta, G. Ezpeleta, V. Cabezas, I. Atutxa, C. Busto, E. Gomez, J. Unzaga, R. Cisterna (Bilbao, ES) Objective: The aim of this study is to know if, like in other bacteraemias, the time between blood culture incubation and growth detection measured by the time to positivity in a continuously monitored system correlates with the source of infection and the outcome of the patient who suffers and enterococcal bacteraemia. Methods: We performed a retrospective, observational study involving adult inpatients that had Enterococcus spp. bacteraemia between 1 October 2003 and 30 September 2006 at a University hospital. Measurements included time to positivity in initial blood culture series, duration of bacteraemia episode, gender, age, rate of metastatic infection, and outcome. Results: A total of 38 Enterococcus spp. bacteraemias (>1 positive blood culture result) were reported for patients with ages between 1 day–94 years (median age, 69 years); 5 (13.15%) bacteraemias were associated with endocarditis. The microbiological documentation of the source of infection was achieved only in half of the cases. The mortality rate was 21.8%. The duration of bacteraemia was 1–47 days (median duration, 8 days; average duration 11.65 days). The time to positivity ranged from 40 minutes to 1 day (median time to positivity, 8.45h). There was signiﬁcantly shorter for patients with an endocarditis or catheter related infection, compared with the other sources of bacteraemia (p = 0.05) but no statistical difference was observed when both endovascular sources of infections were compared. Analysis using logistic regression found that a short time to positivity was an independent predictor of an endovascular source of infection but not the outcome of the patient. In fact, all the deaths recorded in this study were non-infection related. Conclusions: Time to positivity in Enterococcus spp. bacteraemia may provide useful diagnostic information of the source of infection but not prognostic information. Meanwhile due to the reduced number of cases further studies are needed. P937 Time to positivity in blood cultures of patients with Staphylococcus aureus bacteraemia: possible correlation with the source of infection C. Ezpeleta, J.A. Alava, G. Ezpeleta, V. Cabezas, I. Atutxa, C. Busto, E. Gomez, J. Unzaga, R. Cisterna (Bilbao, ES) Objective: The aim of this study is to asses if the time between blood culture incubation and growth detection in Staphylococcus aureus bacteraemia measured by the time to positivity in a continuously monitored system correlates with the source of infection and the outcome of the patient. Methods: We performed a retrospective, observational study involving adult inpatients who had S. aureus bacteraemia between 1 October 2002 and 30 September 2006 at a University hospital. Measurements included time to positivity in initial blood culture series, duration of bacteraemia episode, gender, age, rate of metastatic infection, and outcome. Results: A total of 211 S. aureus bacteraemias (>1 positive blood culture result) were reported for patients with ages between 7–94 years (median age, 69 years); 21 (9.95%) bacteraemias were associated with endocarditis. The microbiological documentation of the source of bacteraemia was achieved only in half of the cases. The mortality rate was 21.8%. The duration of bacteraemia was 1–60 days (median duration, 8 days; average duration 9.8 days). The time to positivity ranged from 40 minutes to 3 days (median time to positivity, 11.19h) and was signiﬁcantly shorter for patients with an endovascular source of infection (endocarditis or catheter related infection), compared with the other sources of bacteraemia (p = 0.001). Analysis of the data using logistic regression revealed that a time to positivity shorter than 10 hours, was an independent predictor of an endovascular source of infection and outcome of the patient. Conclusions: Time to positivity in S. aureus bacteraemia may provide useful diagnostic and prognostic information. Growth of S. aureus within 10 h after the initiation of incubation may identify patients with a high risk of fatal infection. Meanwhile due to the reduced number of cases further studies are needed. P938 Time-to-report blood culture Gram stain and tentative susceptibility result G. Just-N¨ ubling, S. Franck, M. Hahn, V. Rickerts, P.M. Shah (Frankfurt am Main, DE) Objective: Rapid detection of pathogens in blood culture bottles shortens hospitalisation (Beekmann et al.) and appropriate empirical antimicrobial treatment improves outcome. (Behrendt et al. Hautala et al.) How long does it take to report Gram stain result, tentative categorisation (TC) “susceptible, intermediate or resistant” from directly inoculated agar plates and how often were very major (VME), major (ME) or minor errors (mE) when TC results were compared to the ﬁnal susceptibility results? Method: Bactec bottles (Plus aerobic/anaerobic) were inoculated on the wards as recommended by the manufacturer. All bottles were incubated in the BACTEC 9120/9140. When growth was reported to the laboratory personnel a Gram stain was performed and an aliquot was streaked on blood and chocolate agar plates. Depending on the Gram stain result appropriate antibiotic disks were placed on additional M¨ uller-Hinton or Columbia blood agar plates and TC was judged by using agar diffusion method.. The Gram stain result was reported to the clinician. Tentative susceptibility was interpreted after 6-8 hours of incubation and TC was compared with the ﬁnal categorisation obtained from MIC determined using the E-Test method. Results: 141 blood culture bottles were prospectively studied. Gram stain results were available in 0.6 (standard deviation [SD] 0.6, range 3.6–0.1) hours and it took 0.9 (SD 0.7, range 3.3–0.1) hours to report the result to the attending physician. 2,538 antibiotic interpretations (TC) were available within 22.5 (SD 16.4, range 97.8–5.1) hours. VME were seen in 9 (0.35%), ME in 4 (0.16%) and mE in 27 (1.06%). Conclusions: Gram stain result and direct susceptibility result are available within 23.1 hours and can be safely used to modify empirical treatment, as very major or major errors were recorded in less than 1%. Reference(s) Beekmann, S.E., et al. Effects of rapid detection of bloodstream infections on length of hospitalisation and hospital charges. J Clin Microbiol 41 (2003): 3119-25. Behrendt, G., et al. Inﬂuence of antimicrobial treatment on mortality in septicaemia. J Chemother 11 (1999): 179-86. Hautala, T., et al. Blood culture Gram stain and clinical categorization based empirical antimicrobial therapy of bloodstream infection. Int J Antimicrob Agents 25(4) (2005): 329-33.