Ampicillin killing curve patterns for ampicillin-susceptible nontypeable Haemophilus influenzae strains by the agar dilution plate count method.

Ampicillin killing curve patterns for 20 strains of ampicillin-susceptible nontypeable Haemophilus influenzae were determined by the agar dilution plate count method. The paradoxical effect was detected in the 24-h killing curve patterns for each strain. For the biphasic effect, minimum survivor percentages (maximum killing) occurred over a narrow range of ampicillin concentrations immediately above the MIC, with survivor percentages then rising rapidly to peak at approximately 1-log10-unit increment higher. The 24-h minimum survivor percentages for the 20 strains ranged from approximately 0.01% (rapid killing) to greater than 10% (slow killing). In comparison with the previous results for typeable strains, the present findings suggest that nontypeable stains are, on average, killed much more slowly. Based on the initial 24-h killing curve patterns for the 20 strains, 4 strains were selected as putative representatives of the range of bactericidal responses encountered. These strains were then studied to examine the reproducibility of the 24-h patterns and to determine sequential killing curves. These patterns were found to be reproducible and served to characterize the relative killing responses of the strains. In the sequential studies of three of the four strains, tiny colonies having the gross and microscopic characteristics of L-forms were found to be present on the agar dilution plate count plates prior to the application of penicillinase at 48 and 72 h. Such colonies reverted to vegetative forms within 24 to 48 h after application of penicillinase to the panels. Of particular interest was the observation that the paradoxical effect was manifested both by the L-form colonies and by the reverted vegetative colonies. The late development of L-forms was observed for both rapidly and slowly killed strains.