Patient population
The patient population consisted of patients admitted to the General Hospital of Tripolis, Tripolis, Greece, during the period of 16/05/2000 – 15/05/2005. The General Hospital of Tripolis is a 204-bed, tertiary health center that offers services of most medical specialties, including medicine, surgery, pediatrics, and obstetrics and gynecology. The hospital serves the population of the city of Tripolis (40,000 people) and the population of the surrounding rural area of about 250,000 people (local, mainly rural population of Central and South Peloponnesus, Greece). Our study did not involve any experimentation on patients. It is a retrospective analysis of microbiological data. The study was approved by the Department of Microbiology of General Hospital of Tripolis, Greece and by the Ethics Committee of the Alfa Institute of Biomedical Sciences, Athens, Greece.
Microbiological studies
In our study, isolates of the first positive blood cultures obtained for each admission for each patient were analyzed. The isolation of the microorganisms from blood culture specimens was performed using the Bact/Alert 3D 60 Select (Biomerieux) automated system, according to the manufacturer's instructions. We focused on Gram-negative and Gram-positive bacterial isolates. Identification of the microorganisms to the species level was performed with the API system (Biomerieux Vitek, Hazelwood, MO). Gram-negative microorganisms were identified by the API 20 E strip system, a self-contained system of 20 microtubes of dehydrated substrates designed for overnight incubation. Identification is performed by adding necessary reagents and then visually interpreting the results. The tests included in the system, are: o-nitrophenyl galactopyranoside (ONPG), arginine dehydrolase (ADC), lysine and ornithine decarboxylase (LDC and ODC respectively), citrate (CIT), H2S, urea (URE), tryptophan deaminase (TDA), indole (IND), Veges-Proscone (VP), gelatin (GEL), glucose (GLU), mannitol (MAN), inositol (INO), sorbitol (SOR), rhamnose (RHA), sucrose (SAC), melibiose (MEL), amygdalin (AMY), and arabinose (ARA). Numerical coding of results allows computerized interpretation of patterns, lists of which are available in a codebook. Streptococci were identified with the API 20 strep system. Staphylococcus aureus was differentiated from other catalase-positive, Gram-positive cocci by mannitol salt agar. Pseudomonas species were identified with the use of the oxidase test. The fungi isolates were Candida but they were not further identified to the species level.
Bacterial isolation was followed by antimicrobial susceptibility testing that was performed using the disc diffusion in Mueller-Hinton agar and measurement of the diameter of the inhibition zones technique. The antibiotic discs that we used were provided by MAST Diagnostics, Mast Group Ltd, Bootle, Merseyside, United Kingdom. The diameters of microbial inhibition were interpreted based on the relevant guidelines regarding the cutoff diameters of the Clinical and Laboratory Standards Institute (CLSI, formerly called NCCLS) for each antibiotic tested, depending on the microbial species (NCCLS volume 21-1, M2-A7 document). In our study we included data regarding the in vitro susceptibility of blood isolates to antimicrobial agents only if they were available for the whole 5-year study period, except for results of in vitro susceptibility testing of S. aureus and coagulase negative staphylococci to oxacillin and penicillin.
Statistical analysis
Differences in proportions were compared by x2 test or Fischer exact test if the numbers of the studied was small (less than 6 in any of the compared cells of observations). The statistical significance was set at the level p < 0.05. All statistical analyses were performed using SPSS 11.0 (SPSS Inc., Chicago, Illinois, USA).