Characterization of the inoculum effect with Haemophilus influenzae and beta-lactams.

An inoculum effect is defined as a four-fold or greater increase in MIC with an increase in bacterial inocula. Haemophilus influenzae was tested for an inoculum effect with ampicillin, cefuroxime, and amoxicillin/clavulanate using the standard initial inocula (5 x 10(5) CFU/mL) and a higher initial inocula (1 x 10(7) CFU/mL). An inoculum effect was observed with both beta-lactamase (TEM-1, ROB-1) positive and beta-lactamase negative strains of H. influenzae when MICs were determined based on turbidity. MICs based on viable cell counts however, demonstrated that only beta-lactamase positive strains of H. influenzae produced an inoculum effect. These observations suggest that MICs determined based on turbidity, using high initial inocula, are not reliable when examining the inoculum effect in H. influenzae. The magnitude of the inoculum effect with beta-lactamase positive strains was beta-lactam dependent (ampicillin > amoxicillin/clavulanate > cefuroxime). beta-lactam kill-curves confirmed the aforementioned results. Addition of the beta-lactamase inhibitor clavulanate completely reversed the inoculum effect in beta-lactamase (TEM-1 and ROB-1) positive strains of H. influenzae with all beta-lactams tested. Introduction of the beta-lactamase gene TEM-1 on plasmid vector pLS88 into a beta-lactamase negative strain of H. influenzae (Rd) produced an inoculum effect based on viable cell counts. In conclusion, our results suggest that the beta-lactam inoculum effect demonstrated by H. influenzae is the result of beta-lactamase production and is poorly assessed by turbidity.