The plasmid-mediated extended-spectrum beta-lactamases (ESBL) confer resistance to oxymino-cephalosporins, such as cefotaxime, ceftazidime, and ceftriaxone and to monobactams such as aztreonam. It is well known fact that ESBL producing bacteria exhibit a pronounced inoculum effect against broad spectrum cephalosporins like ceftazidime, cefotaxime, ceftriaxone and cefoperazone. The aim of this investigation was to determine the effect of inoculum size on the sensitivity and specificity of double-disk synergy test (DDST) which is the test most frequently used for detection of ESBLs, in comparison with other two methods (determination of ceftazidime MIC with and without clavulanate and inhibitor potentiated disk-diffusion test) which are seldom used in clinical laboratories. The experiments were performed on a set of K. pneumoniae strains with previously characterized beta-lactamases which comprise: 10 SHV-5 beta-lactamase producing K. pneumoniae, 20 SHV-2 + 1 SHV 2a beta-lactamase producing K. pneumoniae, 7 SHV-12 beta-lactamase producing K. pneumoniae, 39 putative SHV ESBL producing K. pneumoniae and 26 K. pneumoniae isolates highly susceptible to ceftazidime according to Kirby-Bauer disk-diffusion method and thus considered to be ESBL negative. According to the results of this investigation, increase in inoculum size affected more significantly the sensitivity of DDST than of other two methods. The sensitivity of the DDST was lower when a higher inoculum size of 10(8) CFU/ml was applied, in distinction from other two methods (MIC determination and inhibitor potentiated disk-diffusion test) which retained high sensitivity regardless of the density of bacterial suspension. On the other hand, DDST displayed higher specificity compared to other two methods regardless of the inoculum size. This investigation found that DDST is a reliable method but it is important to standardize the inoculum size.