Assessment of different antibacterial effect measures used in in vitro models of infection and subsequent use in pharmacodynamic correlations for moxifloxacin.

A dilutional culture in vitro pharmacodynamic model of infection was used to assess the best measure of antibacterial effect for moxifloxacin at simulated human doses of 400 mg 24 hourly for 48 h. This was then related to two pharmacodynamic parameters, the drug area under curve: MIC ratio (AUC/MIC) and the length of time that the drug concentration remained above the MIC of the bacterium (T > MIC). Twenty-one bacterial strains (Streptococcus pneumoniae n = 6; Haemophilus influenzae n = 6; Moraxella catarrhalis n = 3; beta-haemolytic streptococci n = 3; Staphylococcus aureus n = 3; MIC range 0.06-3.6 mg/L) were tested in 69 individual simulations. The measures or parameters of antibacterial effect considered were log change in viable count in the initial inoculum at 12 h (triangle up12), 24 h (triangle up24), 36 h (triangle up36), 48 h (triangle up48), maximum reduction in count (triangle up(max)); time for bacterial counts to reduce by 100-fold from the initial density (T99) or 1,000-fold (T99.9); and area under the bacterial kill curve from 0 to 24 h (AUBKC(24)) or from 0 to 48 h (AUBKC(48)). triangle up12, triangle up24, triangle up36, triangle up48, triangle up(max), T99, T99.9 did not vary over the complete range of MICs; at high MICs, especially with Gram-positive bacteria the T99 and T99.9 values were >48 h while at low MICs, especially with Gram-negative bacteria, bacterial counts were reduced below the limit of detection with triangle up12, triangle up24, triangle up36, triangle up48 and triangle up(max) exceeding >6.5 log reduction. AUBKC(24) and AUBKC(48) varied more completely over the range of MICs and more importantly had the best within-strain reproducibility (median percentage coefficient of variation <15%). The relationship between the transformed AUBKC(24) and AUC/MIC could be described by a sigmoid Emax model but the relationship with T > MIC could not. Use of weighted least squares regression to examine the combined effect of AUC/MIC and T > MIC on AUBKC(24) indicated that AUC/MIC provided a good fit to the data (r(2) = 0.94) and adding T > MIC did not improve the model fit. Cox proportional hazards regression indicated that AUC/MIC was predictive of T99 and in a multivariate model although AUC/MIC predicted outcome after fitting AUC/MIC, T > MIC was not significant. AUBKC was thus shown to be the optimum measure of antibacterial effect to use in pharmacodynamic studies of moxifloxacin and AUC/MIC the best predictor of antibacterial effect as measured by AUBKC(24) or T99. These results are in good agreement with animal data on moxifloxacin pharmacodynamics and human data for some other fluoroquinolones.

External quality assessment of the serum bactericidal test: results of a methodology/interpretation questionnaire.

Two hundred microbiology laboratories in the UK took part in two separate experimental external quality assessment distributions related to the serum bactericidal test (SBT). In the first, Staphylococcus aureus NCTC 6571 (vancomycin MIC 1 mg/L), was tested against a human serum containing vancomycin 38 mg/L plus gentamicin 0.5 mg/L. In the second, Streptococcus oralis PAJ 112/4183 (penicillin MBC < or = 0.03 mg/L) and Streptococcus sanguis PAJ 107/4184 (penicillin MBC = 128 mg/L) were tested against human serum containing penicillin 15 mg/L. Respondents returned their laboratory results and a questionnaire on clinical interpretation and technical aspects. Most laboratories (194/199, 97.5%) recommend the use of the SBT in the management of infective endocarditis but only 48 (25.2%) often or always change therapy on the basis of the result. A wide range of interpretative criteria, definitions of bactericidal endpoints and methodologies are used. Performance in the first distribution was acceptable for 75% of laboratories but in the second only 34% could identify penicillin tolerance; 34 respondents reported an SBT result of < or = 2 for the tolerant strain, 81 laboratories reported one of > or = 16. Technical factors related to acceptable performance were: sonication of broth before counting the inoculum; knowing the inoculum size in cfu/mL; use of a 4-8 h broth culture to make the inoculum; incubation of recovery plates for > 36 h; use of a calibrated pipette to sample for surviving bacteria; use of measured volumes to add the inoculum. Use of uncalibrated pipettes or standard loops to recover survivors was related to poor performance. Microbiology departments in the UK should review the clinical need to perform the SBT in the light of their local circumstances and if they elect to continue to offer this test, revise their methodologies which could be producing misleading results when testing alpha-haemolytic streptococci.

Typing of Listeria monocytogenes by random amplified polymorphic DNA (RAPD) analysis.

The aim of the study was to determine the effectiveness of random amplified polymorphic DNA analysis in typing Listeria monocytogenes from human infections. Twenty-five L. monocytogenes serogroup 1/2 and 70 serogroup 4 including 14 serovar 4b(x) were typed by RAPD-PCR analysis. Six primers were used to type each L. monocytogenes isolate and the DNA amplification performed with supertaq DNA polymerase in a Hybaid Thermal Reactor. Each bacterial strain was analysed separately with all primers and the profiles were judged by eye and designated to a group by comparison to other strains. Bands were classified as major or minor. Based on analysis of major band patterns, the 25 serogroup 1/2 isolates gave rise to 12 different groups. The groups only contained serovar 1/2a or 1/2b with a single exception. Using minor bands all isolates could be distinguished. All the serogroup 4 isolates gave the same major band patterns. The 14 serovar 4b(x) isolates which were epidemiologically related gave identical profiles with the exception of one isolate. Of the remaining strains, 41 produced individual patterns on minor band analysis. RAPD analysis with multiple primers is low cost, discriminatory and is most ideally suited to testing small (< 50) numbers of strains. We have shown that serogroup 1/2 L. monocytogenes strains are a more diverse group than serovar 4b strains and RAPD-PCR will provide a technique of considerable value in typing L. monocytogenes in the future.

A critical assessment of the agar dilution chequerboard technique for studying in-vitro antimicrobial interactions using a representative beta-lactam, aminoglycoside and fluoroquinolone.

The agar dilution chequerboard technique of studying antimicrobial interactions was assessed by testing a representative beta-lactam (piperacillin/tazobactam), aminoglycoside (gentamicin) and fluoroquinolone (ciprofloxacin) against themselves, that is piperacillin/tazobactam plus piperacillin/tazobactam, gentamicin plus gentamicin and ciprofloxacin plus ciprofloxacin. In addition, combinations of piperacillin/tazobactam plus gentamicin or ciprofloxacin were also tested against Enterobacteriaceae and Acinetobacter spp. in triplicate. The agar dilution chequerboard technique did not reliably show addition when agents were combined with themselves, and there was also considerable variation when beta-lactam plus aminoglycoside or fluoroquinolone combinations when tested in triplicate. These observations, and problems with the design of the method, indicate that the chequerboard technique should be used only with adequate controls and replication, and then interpreted with extreme caution; ideally, it should not be used as a method of assessing antimicrobial interactions.