Sensitivity tests for the polymyxins.

Although the value of polymyxins B and E in the treatment of infections with Gram-negative bacilli is well recognized, they are not usually used unless the causal organisms are insensitive to antibiotics with wider activity. In consequence, sensitivity tests for them are not common laboratory procedures and it is not generally appreciated that the usual methods are not entirely satisfactory. Judged from experience with the commoner antibiotics, organisms often appear insensitive to the polymyxins when tested by the agar diffusion method, yet the dilution (inclusion) method may show them to be fully sensitive. This latter method too can give very variable results, especially when a fluid medium is used, and on a solid medium the strain may often appear to consist of resistant and sensitive variants when in fact it is homogeneous. The discrepancies in both of these methods are probably due to the fact that the mode of action of the polymyxins differs from those of the more commonly used antibiotics, and studies of the action have revealed two important facts. First, the polymyxins are always bactericidal (Brownlee and Bushby, 1948; Brownlee, Bushby, and Short, 1952) and bacteriolytic (White, Alverson, Baker, and Jackson, 1949; Bushby, 1952), apparently through being absorbed by the cell membrane and altering its permeability. They thus act in a manner similar to the cationic detergent; cetyltrimethylammonium bromide (Bushby, 1952; Few and Schulman, 1953). Secondly, the minimum bactericidal concentration of the polymyxins depends upon the number of organisms present (Brownlee et al., 1952) due to the fact that a certain quantity of polymyxin is expended in killing each organism. The amount absorbed is therefore roughly proportionalO to the sensitivity of the strain and to the number of organisms present (Bushby, 1952), but the relationship is complicated by the fact that complete saturation of the organisms is not necessary for the bactericidal effect. Incidentally the organisms do not have to be living for this absorption to occur. From these facts it follows that it is not possible to state an exact sensitivity figure for any given organism without defining the size of the inoculum used in the test. Swift and Bushby (1951) showed that the apparent sensitivity of a strain of Haemophilus influenzae ranged from 0.8 units to 25 units per ml., depending upon the size of the inoculum. The following experiments illustrate the poor correlation between the apparent sensitivities when measured by the diffusion and dilution methods. Three different-sized inocula of eight organisms of different genera were used.