Background: Antibiotic combinations are frequently used in order to obtain wide-spectrum effects in the treatment of serious infections such as septicemia and endocarditis, and also to produce an in vivo effect against strains which are defined as resistant to the known inhibiting or fatal dose of one antibiotic. The synergistic effects of combinations such as aminoglycoside + beta-lactam, aminoglycoside + quinolone and quinolone + beta-lactam on Pseudomonas aeruginosa have been revealed in different studies. The multiple resistance rate of nosocomial P. aeruginosa strains isolated from intensive care units (ICUs) has been reported as high in many studies. Methods: In this study, the effects of various combinations of antibiotics (aminoglycoside + beta-lactam and aminoglycoside + quinolone) against 101 multiresistant P. aeruginosa strains which were isolated from the ICUs of three different hospitals in Istanbul were investigated using the E-test method. The combinations for which the highest synergistic effects were determined by the E-test method were also tested with the checkerboard method, i.e. in addition to the E-test method, in 19 of a total of 23 strains. Results: When the synergistic results which were obtained with the combinations of aminoglycoside + beta-lactam were compared with those of the aminoglycoside + quinolone combinations, they were determined to be higher for the two aminoglycosides gentamicin (GM) and tobramycin (TM). We determined the synergistic rates to be 23, 21, 19, 18, 16, 14, 10 and 10% for GM + ceftriaxone (TX), GM + piperacillin (PP), GM + ceftazidime (TZ), TM + PP, TM + TX, TM + TZ, GM + ciprofloxacin (CI) and TM + Cl, respectively. The GM + TX combination - for which the highest synergistic effects were determined with the E-test stripes - was also determined as synergistic with the checkerboard method in 19 of a total of 23 strains (23%), and the agreement rate between the two methods was 100% (kappa > 0.7). The highest synergistic effects against strains which were sensitive to both of the antibiotics which constitute the combinations were found for the GM + TX combination, as 50%, whereas for strains which were resistant to both of the antibiotics, this was found for the TM + PP combination, also as 50%. Conclusions: We consider that the minimal inhibitory concentration values of antibiotics are not sufficient alone in order to constitute a combination for multiresistant strains and it would be advisable to begin a treatment by applying a combination study. The E-test method has been evaluated as a good alternative for combination investigations because of its ease both of application and evaluation and also for its good agreement with the standard checkerboard method. Copyright (C) 2002 S. Karger AG, Basel.