Microdilution methodology was used to study the interaction of imipenem with erythromycin and tetracycline, a combination therapy that might be used for the treatment of serious pelvic inflammatory disease. The combination of imipenem and erythromycin showed no antagonism for Escherichia coli and Haemophilus influenzae but was antagonistic for Staphylococcus aureus, Enterococcusfaecalis, and group B streptococci; the combination of imipenem and tetracycline was antagonistic for all strains except H. influenzae. Correlation between the results of kill curves and the measurement of fractional bactericidal concentration (FBC) indices was good, although FBC indices showed less antagonism than kill curves. Fractional inhibitory concentration indices showed poor correlation, rarely showing antagonism, and indeed showed synergy in three cases. If erythromycin or tetracycline is considered necessary in addition to imipenem in the treatment of pelvic inflammatory disease, it is probably more effective when given after the course of imipenem has been completed.Pelvic inflammatory disease (PID) is often a mixed infection of aerobic and anaerobic bacteria (1) necessitating broad-spectrum antibiotic therapy, and imipenem may be appropriate therapy in serious PID. However, the role of Chlamydia trachomatis in PID is being increasingly recognized (5, 6), although this organism is difficult to detect in the routine diagnostic laboratory. It may therefore be relevant, when treating patients for serious PID with imipenem, to use an antichlamydial agent such as erythromycin or tetracycline in addition because imipenem does not have significant antichlamydial activity. The combination of bacteriostatic agents such as tetracycline and erythromycin with bactericidal agents such as beta-lactams is often antagonistic (4, 9). If this is the case with imipenem, then it would argue against the simultaneous administration of erythromycin or tetracycline, which might be more effective when given subsequently. This study was undertaken to assess the interaction of these three antibiotics.( (190 ,u) was dispensed into each of the wells of 16 microdilution trays. These were covered with cling film until required.After 2, 4, 6, and 24 h of incubation, the contents of each well of the checkerboard were sampled after thorough mixing by suction into the pipette tip and expulsion four times. In addition, at 24 h the microdilution plates were placed on a shaker for 2 min to disperse any pellets that had formed. Colony counts were performed by transferring 10-,ul samples from each well to an Iso-Sensitest agar plate (Oxoid), the surface of which was dried prior to use to prevent coalescence of the drops. Second 10-pI samples were then transferred to the microdilution trays, and six serial 20-fold dilutions were prepared prior to transfer to Iso-Sensitest agar for incubation and counting. In the case of H. influenzae, the Iso-Sensitest agar was supplemented with 5% Fildes enrichment; for group B streptococci, it was supplemented with 5% lysed horse ...