Performance of a semi-automated antibiotic susceptibility testing system (ABAC)

The ABAC system for antibiotics susceptibility testing was compared with an agar diffusion method in 14 960 tests, including 23 antibacterial agents. Identical breakpoints were used. Only 3% major discrepancies (M.d.; sensitive vs resistant) and 19% minor discrepancies (m.d.; intermediate vs sensitive or resistant) were noted. Major discrepancies were mainly found for methicillin (Staphylococcus aureus), netilmicin (Pseudomonas aeruginosa), chloramphenicol, sulphamethoxazole and trimethoprim (Proteus sp.) and were checked by quantitative susceptibility tests. These showed ABAC to be at fault in 41--47% of discrepancies, the diffusion test in 21--32% and 21--37% were intermediate. Half of the m.d. involved beta-lactams, which is explained by two low breakpoints. Except for methicillin and netilmicin the overall results showed ABAC to be equal to the agar diffusion method. Technical faults, like leakage and incorrect filling of cups in the plastic rotors of ABAC, occurred in 14% of the rotors.