S32 17th ECCMID / 25th ICC, Oral presentations of antibiotics in humans and their effects have been outlined in order to determine the optimal dosing interval. Antimicrobial drugs with a concentration-dependent activity, i.e. the newer quinolones and aminoglycosides, should be administered as a single daily dose to maximise the peak serum concentration/MIC ratio. AUC/MIC ratio is also an important parameter for PK/PD correlations. Existing data on fluoroquinolones suggest that a ratio of 100–125 correlates with high bacterial eradication and optimal clinical outcome in infections due to Gram-negative pathogens, while a ratio of 50 is associated with a high probability of eradication of S. pneumoniae strains. For this group of antibiotics the new concept of the so called mutant prevention concentration (MPC) and mutant selection window (MSW) may be helpful in restricting the enrichment of mutant subpopulations and consequently, at least partially, in controlling the spread of resistance. Beta-lactams, having time-dependent efficacy, usually do not have great post antibiotic effects (PAEs), and the parameter which seems to better correlate PD with PK is the time duration with concentrations higher than the MIC (T > MIC). These antibiotics need to be given with short dosing intervals or even by continuous infusion to maintain plasma levels exceeding the MIC for a sufficiently long period. In summary, we can theoretically optimise the dosage regimen of antibiotics (dose, administration route and interval between doses) in clinical practice by correlating the PK and PD pertaining to each antibiotic class, based on experimental animal studies and these PK/PD parameters may contribute to the containment of resistance for all drug classes and especially the most important ones used in serious infections in intensive care patients. Antibiotic resistance in developing countries: a Pandora’s Box S159 Surveillance of antimicrobial resistance in developing countries: methods and strategies A. Bartoloni (Florence, IT) Antibiotics are the most commonly purchased class of drugs in low- resource countries, where the infectious diseases are extremely frequent and the bacterial infections are the major cause of death, especially in childhood. The phenomenon of microbial drug resistance, which represents a global public health problem, is particularly serious in these countries, as resistance rates are even higher than in industrialised countries and the therapeutic options are often unavailable or too expensive. Surveillance of antibiotic susceptibility is a key element to provide updated information on the magnitude and trends in resistance, and to plan and monitor intervention strategies aimed at preserving the therapeutic efficacy of antibiotics. In low-resource countries, effective surveillance programmes are difficult to implement for a number of reasons, including scarce financial resources, lack of laboratory facilities and, where laboratories do exist, lack of quality control, reliable reagents and adequate supervision. In these settings, the development of reliable and low-cost alternative methods could facilitate the implementation of large-scale surveillance. There is an increasing agreement about the importance of extending the surveillance of antibiotic resistance to the commensal microbiota of humans and animals. This bacterial population, although not being a specific target, is continuously exposed to the selective pressure generated by antimicrobial chemotherapy and may become a potential reservoir of resistant strains that can cause infections, and of resistance determinants that can be transferred to pathogenic bacteria. Therefore, surveillance of antibiotic-resistant bacteria carried by healthy individuals is considered an indicator of the spread of antibiotic resistance that could also be useful to predict the emergence of resistance in pathogenic bacteria. In this perspective, resistance patterns of some members of the commensal microbiota, such as the faecal Escherichia coli, have been evaluated in various epidemiological settings. For this purpose, different microbiological approaches have been implemented and evaluated as useful tools to conduct large scale resistance surveillance studies and to monitor resistance control programmes in a cost-effective manner. Does detection of extended-spectrum b-lactamases matter? S160 Does detection of extended-spectrum b-lactamases matter? The No case J. Turnidge (North Adelaide, AU) The emergence of extended-spectrum b-lactamases (ESBLs) in Enter- obacteriaceae in the 1980s marked an important turning point in routine laboratory diagnostics. Soon after their discovery, phenotypic methods were developed for their detection and confirmation. Confirmation depended largely on the ability of clavulanate to inhibit the main culprit enzymes at the time, the TEM and SHV variants. Later this technique proved reliable for CTX-M type enzymes. Many susceptibility testing methods recommended the use of ESBL screening and confirmation tests on a routine basis, including CLSI, BSAC, CA-SFM, and SRGA. However, a number of issues have emerged with routine use over the years: 1. The problem of defining an adequate number of substrates to ensure sufficiently sensitive screening. Ideally, one should include a minimum of 4, namely cefpodoxime, ceftazidime, ceftriaxone or cefotaxime, and aztreonam, and at concentrations that often differ from those used for susceptibility breakpoints. 2. The lack of reliable phenotypic methods to detect ESBLs in species with inducible AmpC b-lactamases. Some of these species have been shown to be important reservoirs for ESBLs, and resistance to extended-spectrum cephalosporins cannot solely attributed to stable de-repression of AmpC. 3. The failure of current methods to provide advice on the interpretation of a positive screening test but a negative confirmation test, especially if the isolates are “susceptible” to extended-spectrum cephalosporins using method-recommended breakpoints. Such strains have been shown to harbour OXA enzymes, inhibitor-resistant TEM enzymes, or particularly plasmid-borne AmpC enzymes with significant frequency. Thus there is no current phenotypic or genotypic test that can be practically and effectively applied in the routine laboratory with sufficient sensitivity to detect the emerging range of transmissible enzymes. However, we can rely to a great extent on the selection of or change to appropriate susceptibility breakpoints. A range of recent studies has suggested that failures of treatment with extended- spectrum cephalosporins are likely when strains of Enterobacteriaceae have MICs elevated above the wild-type. Further, application of pharmacokinetic/pharmacodynamic (PK/PD) principles to the most widely recommended dosing schedules of cephalosporins suggest that the susceptibility breakpoints recommended by many methods are too high, and should to lowered to values that fortunately coincide wild-type cut- off values. Hence, lowering of susceptibility breakpoints for extended- spectrum cephalosporins to those defined by PK/PD will indicate the presence of an ESBL or plasmid-borne AmpC enzyme with sufficiently high likelihood to allow laboratories to report both resistance to these agents and provide advice about appropriate infection control procedures. Antibiotic usage O164 A planned dramatic drop in trimethoprim consumption in a 180,000 population did not result in a related decrease in trimethoprim resistance in Escherichia coli M. Sundqvist, M. Sj¨ olund, A. Runehagen, H. Cars, K. Abelson-Storby, D.I. Andersson, O. Cars, G. Kahlmeter (Uppsala, V¨ axj¨ o, SE) Objectives: Since antibiotic resistance often is associated with a biological fitness cost for bacteria it is assumed that a reduction of antibiotic use is followed by a reduction in resistance rates. Over the last 10 years trimethoprim (TRI) resistance in Escherichia coli has increased in Sweden. In Kronoberg county the TRI resistance has increased from