Strains were isolated from antral puncture specimens obtained from adult patients with sinusitis between 1994 and 1998 and were identified by standard criteria (9, 13). The numbers and species of clinical isolates tested are given in Table 1.

The suppliers of the following standard laboratory powders were as indicated: amoxicillin-clavulanate and amoxicillin, SmithKline Beecham Laboratories, Philadelphia, Pa.; azithromycin, Pfizer Inc., New York, N.Y.; levofloxacin, R. W. Johnson Pharmaceutical Research Institute, Raritan, N.J.; clarithromycin, Abbott Laboratories, Abbott Park, Ill.; and cefuroxime, Glaxo-Wellcome, Research Triangle Park, N.C.

Frozen cultures were transferred at least twice to ensure purity and good growth. Susceptibility testing was performed according to National Committee for Clinical Laboratory Standards guidelines (10, 11). Streptococcus pneumoniae ATCC 49619, Haemophilus influenzae ATCC 49247, Staphylococcus aureus ATCC 29213, Escherichia coli ATCC 25922, and Bacteroides fragilis ATCC 25285 were tested simultaneously with the appropriate plates and environments. Brucella agar supplemented with hemin, vitamin K₁, and 5% laked sheep blood was the basal medium used for anaerobic species. Mueller-Hinton agar was used for staphylococci, Mueller-Hinton agar supplemented with 5% sheep blood was used for pneumococci and Moraxella isolates, and Mueller-Hinton agar supplemented with yeast extract, hematin, and NAD (Hemophilus test medium) was used for Haemophilus species. Antimicrobial agents were reconstituted according to the manufacturers’ instructions. Serial twofold dilutions of antimicrobial agents were prepared on the day of the test and added to the media at various concentrations (micrograms per milliliter).

The agar plates were inoculated with 10⁴ CFU per spot for aerobes and 10⁵ CFU for anaerobes by using a Steers replicator (Craft Machine Inc., Chester, Pa.). The MIC was defined as the lowest concentration of an agent that yielded no growth or a marked change in the appearance of growth as compared to the growth control plate. All peptostreptococci were tested and were negative for β-lactamase production by both cefinase disk and acidometric methods.

Because antimicrobial resistance is increasingly frequent, newer agents are being evaluated for the therapy of sinusitis (2, 14). Thornsberry et al. (14) tested 9,190 general clinical isolates of S. pneumoniae and found susceptibilities as follows: 67% to penicillin, 82% to amoxicillin clavulanate, 75% to cefuroxime, and 97% to levofloxacin. Doern et al. (3) studied 1,527 pneumococcal isolates collected in the 1994 to 1995 winter months from 30 U.S. reference centers and found 26% overall resistance to penicillin, but 39% of sinus aspirate isolates were penicillin resistant. They also noted that 12% of sinus aspirate isolates were resistant to cefuroxime. Our sinus isolates differed in that 94% were susceptible to amoxicillin, amoxicillin-clavulanate, azithromycin, and clarithromycin; 78% were susceptible to cefuroxime; and 100% were susceptible to levofloxacin. Our study did not include any pediatric strains, which may be more resistant than adult strains.

While the mechanism of resistance was not elucidated in their study, Thornsberry et al. (14) noted clarithromycin resistance in 29% of pneumococci. Our pneumococcal sinus isolates differed in that 94% were susceptible to azithromycin and clarithromycin, which was similar to the findings of Doern et al. (3), who found 10% resistance to azithromycin, clarithromycin, and erythromycin among pneumococci. The comparability of macrolide resistance of pneumococci may be related to the test system used by the investigators because of the potential effect of CO₂ in the incubation atmosphere for the E test and disk diffusion methods but not the agar or broth microdilution method (5), which we used.

Thornsberry et al. (14) noted 67 and 58% of H. influenzae were amoxicillin and clarithromycin susceptible, respectively. In our study, all H. influenzae isolates (100%) were β-lactamase producers and resistant to amoxicillin but susceptible to amoxicillin-clavulanate. Almost all were resistant to clarithromycin but susceptible to cefuroxime, azithromycin, and levofloxacin. All of our S. aureus isolates were β-lactamase producers, and 44% were also resistant to both macrolides.

The transition of acute into chronic sinusitis may be related to the emergence of resistant anaerobes (2, 8, 12) such as Prevotella species, fusobacteria, and peptostreptococci. Sixty-five percent (11 of 17) of our Prevotella isolates were resistant to amoxicillin (MIC ≥ 1 μg/ml), as were 57% of the Veillonella species (for which there is no amoxicillin breakpoint, but the ampicillin breakpoint is ≤0.5 μg/ml for susceptibility). All anaerobes were susceptible to amoxicillin-clavulanate. No Food and Drug Administration or National Committee for Clinical Laboratory Standards interpretive criteria have been established for the activity of levofloxacin, azithromycin, clarithromycin, or cefuroxime against anaerobic bacteria. Using concentrations similar to those established for most aerobes, overall 8% of anaerobes were resistant to levofloxacin (MIC > 2 μg/ml), especially some peptostreptococci. In contrast, 59% of Prevotella species, 66% of Peptostreptococcus magnus, and 86% of Veillonella species isolates were resistant to cefuroxime (MIC > 4 μg/ml). All P. magnus and fusobacteria isolates and 71% of Veillonella species isolates were resistant to clarithromycin (MIC ≥ 2 μg/ml). Azithromycin was often more active than clarithromycin against anaerobes but still had poor activity against 27% of fusobacteria, 29% of P. magnus (for 33 of 35, the MICs were ≥2 μg/ml), and 64% of Veillonella species isolates. The effect of CO₂ in the atmosphere of incubation has been shown to decrease the activity of macrolides against anaerobes but the clinical relevance has not been determined (6, 7).