Answers are grounded in this paper and cite their sources.
SIR—In a recent article by Thornsberry et al. [1], the authors compared the antimicrobial susceptibility profiles of strains of Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis isolated during 1999–2000 with the profiles of strains isolated during 1998–1999. In the Materials and Methods section of their article, the authors stated that the MICs of agents against S. pneumoniae and H. influenzae were interpreted on the basis of the recommended breakpoints of the National Committee for Clinical Laboratory Standards (NCCLS), and, furthermore, that the authors used the breakpoints that had been in place at the start of the respiratory season for each year (i.e., NCCLS documents M100-S8 for 1998–1999 and M100-S9 for 1999–2000) [2, 3]. However, the breakpoints for several of the agents tested were revised in 2000 (NCCLS document M100-S10) [4], and the revised breakpoints have been used to interpret data from other studies [5–8] conducted during the same time as their study. In fact, one of these other studies was performed at the same institution and by many of the same authors; in that study, on the basis of the breakpoint approved in 2000, the rate of resistance to amoxicillin/clavulanic acid was reported to be 1.0% of isolates [5]. In contrast, the rate of resistance to amoxicillin/clavulanic acid determined on the basis of the old breakpoint was 14.2% of isolates (table 2 in [1]), with regional variations of 6.7%–22.6% (table 4 in [1]). This means that readers who look only at the tables and not the text will be left with an incorrect, outdated assessment of the current susceptibility of S. pneumoniae to an agent commonly used to treat respiratory tract infection. Thornsberry et al. [1] do point out the changes made in January 2000 for amoxicillin/clavulanic acid breakpoints as well as the changes made in the cefuroxime axetil breakpoints, but they only report the effects of these changes in the text and not in the tables or abstract. The changes were very dramatic in the case of amoxicillin/clavulanic acid: with use of the old breakpoints (susceptibility, ⩽0.5 μg/mL; intermediate susceptibility, 1 μg/mL; and resistance, ⩾2 μg/mL), 77.7% of the isolates were susceptible, 8% had intermediate susceptibility, and 14.3% were resistant; with use the new breakpoints (susceptibility, ⩽2 μg/mL; intermediate susceptibility, 4 μg/mL; and resistance, ⩾8 μg/mL), 94% of the isolates were susceptible, 3.4% had intermediate susceptibility, and 2.6% were resistant. In the case of cefuroxime axetil, the distribution changed as follows: with use of the old breakpoints (susceptibility, ⩽0.5 μg/mL; intermediate susceptibility, 1 μg/mL; and resistance, ⩾2 μg/mL), 70.6% of the isolates were susceptible, 2% had intermediate susceptibility, and 27.4% were resistant; with use of the new breakpoints (susceptibility, ⩽1 μg/mL; intermediate susceptibility, 2 μg/mL; and resistance, ⩾4 μg/mL), 72.6% of the isolates were susceptible, 4.6% had intermediate susceptibility, and 22.8% were resistant. These changes in the breakpoints highlight the importance of continuing application of information from clinical studies and pharmacokinetic and pharmacodynamic (PK/PD) studies and of the introduction of new formulations of agents, to provide improved interpretation of susceptibility data. Therefore, patient care is best served by use of the most-current breakpoints rather than the breakpoints that had been in place at the time isolates were recovered, because the objective of collecting surveillance data is to guide future therapy. Furthermore, the use of even current NCCLS MIC interpretations for some oral agents used against S. pneumoniae and H. influenzae has been challenged [9–14]. Although, for most oral agents, the breakpoints for S. pneumoniae published from the year 2000 onwards [4, 15, 16] are in general agreement with the findings of bacteriological outcome studies and PK/PD parameters, current NCCLS breakpoints for H. influenzae, which have not been revised for >10 years, are considerably higher than those for S. pneumoniae (table 1) [16]. Examples include the breakpoints for amoxicillin/clavulanic acid (for S. pneumoniae: ⩽2 μg/mL, susceptibility; 4 μg/mL, intermediate susceptibility; and ⩾8 μg/mL, resistance; for H. influenzae: ⩽4 μg/mL, susceptibility; and ⩾8 μg/mL, resistance), cefuroxime axetil (for S. pneumoniae: ⩽1 μg/mL, susceptibility; 2 μg/mL, intermediate susceptibility; and ⩾4 μg/mL, resistance; for H. influenzae: ⩽4 μg/mL, susceptibility; 8 μg/mL, intermediate susceptibility; and ⩾16 μg/mL, resistance), azithromycin (for S. pneumoniae: ⩽0.5 μg/mL, susceptibility; 1 μg/mL, intermediate susceptibility; and ⩾2 μg/mL, resistance; for H. influenzae: ⩽4 μg/mL, susceptibility), and clarithromycin (for S. pneumoniae: ⩽0.25 μg/mL, susceptibility; 0.5 μg/mL, intermediate susceptibility; and ⩾1 μg/mL, resistance; for H. influenzae: ⩽8 μg/mL, susceptibility; 16 μg/mL, intermediate susceptibility; and ⩾32 μg/mL, resistance). Application of PK/PD breakpoints for H. influenzae to the data of Thornsberry et al. [1] would result in the rate of susceptibility to amoxicillin/clavulanic acid decreasing from 99.8% to 90%–98.8% of isolates, the rate of susceptibility to cefuroxime axetil decreasing from 99.9% to 50%–90%, the rate of susceptibility to azithromycin decreasing from 99.7% to <50%, and the rate of susceptibility to clarithromycin decreasing from 94.7% to <50% (data not shown). In a similar study, H. influenzae data were interpreted according to both NCCLS and PK/PD breakpoints, and susceptibility rates (NCCLS vs. PK/PD) for these same agents were as follows: amoxicillin/clavulanic acid, 100% versus 97.5% of isolates; cefuroxime, 98.8% versus 78.1%; azithromycin, 99.7% versus 0%; and clarithromycin, 76.6% versus 0.1% [9]. In addition, as noted by Thornsberry et al. [1], NCCLS currently does not have MIC interpretations for Moraxella catarrhalis, and application of S. pneumoniae breakpoints would yield very different susceptibility rates than would application of H. influenzae breakpoints. Changes in the National Committee for Clinical Laboratory Standards (NCCLS) breakpoints for Streptococcus pneumoniae and Haemophilus influenzae, 1998–2002. Therefore, for the reasons stated above, we believe that surveillance data for respiratory tract infections should be interpreted according to both current NCCLS criteria and PK/PD parameters when there are significant differences between the breakpoints for agents. Because old breakpoints were used in the tables, Thornsberry et al. [1] reported falsely high rates of S. pneumoniae resistance to amoxicillin/clavulanic acid and cefuroxime. In an era of increased prevalence of drug resistance and limited choices of antimicrobial agents, clinicians are ill served by this approach. Reports of falsely high resistance rates can result in inappropriate administration of such agents as the fluoroquinolones when certain β-lactam agents can be used.
No takes yet. Share an insight, caveat, or question.
Jacobs et al. (2002) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: