The susceptibilities of 265 strains of Corynebacterium species and other non-spore-forming gram-positive bacilli to 18 antimicrobial agents were tested. Most strains were susceptible to vancomycin, doxycycline, and fusidic acid. Corynebacterium jeikeium and Corynebacterium urealyticum were the most resistant organisms tested. Resistance to ␤-lactams, clindamycin, erythromycin, azythromycin, ciprofloxacin and gentamicin was common among strains of Corynebacterium xerosis and Corynebacterium minutissimum. Ampicillin resistance among Listeria monocytogenes was more prevalent than previously reported. Optochin, fosfomycin, and nitrofurantoin showed very little activity against most organisms tested, but the use of nitrofurantoin as a selective agent in culture medium may prevent the recovery of some isolates. Except for the unvarying activity of vancomycin against Corynebacterium species, the antimicrobial susceptibilities of the latter to other antibiotics are usually unpredictable, such that susceptibility tests are necessary for selecting the best antimicrobial treatment.During the last two decades a renewed interest in Corynebacterium species and other non-spore-forming gram-positive bacilli has emerged among clinicians and microbiologists alike (2,8,9,32). Infections caused by these organisms are emerging, new species are being recognized (2, 9), and infections by toxigenic and nontoxigenic Corynebacterium diphtheriae strains are also being described with increasing frequency, indeed, in countries where diphtheria had been totally or almost eradicated (7, 26). However, this renewed interest has not been followed by an in-depth study to determine the antimicrobial susceptibilities of such organisms. Most available data come from scattered case reports, studies on a particular organism, or very old reports, sometimes published before 1960 (9).The aim of the study described here was to determine the antimicrobial susceptibilities of a variety of organisms, most of which were isolated from clinical specimens, against 18 antimicrobial agents. MATERIALS AND METHODSBacterial strains. We tested 265 strains of Corynebacterium species and other non-spore-forming gram-positive bacilli obtained from the following different sources: clinical samples (n ϭ 141), skin (n ϭ 25), bacterial collections (n ϭ 19), and referrals to our laboratory from other institutions (n ϭ 80). Strains from clinical samples were isolated during the period from 1985 to 1993. All strains were identified by conventional methods (2) and also by using the API Coryne system (25), were stored frozen in 10% skim milk, and were maintained at Ϫ70ЊC until use.Antimicrobial agents. The following antibiotics were kindly provided by the manufacturers as powders for in vitro study: ampicillin and oxacillin (SmithKline Beecham Laboratories, Worthing, United Kingdom), cephalothin and vancomycin (Lilly, S.A., Madrid, Spain), cefuroxime (Glaxo Laboratories, Madrid, Spain), imipenem (Merck Sharp & Dohme, Madrid, Spain), tetracycline, doxycycline, and azithromycin (Pf...

The efficacy of amoxicillin/clavulanate and cefuroxime was determined in a gerbil model of otitis media with a mixed Streptococcus pneumoniae plus Haemophilus influenzae middle ear (ME) infection. Results were compared with those obtained in a previous single H. influenzae model. All untreated animals inoculated with the mixed inoculum developed acute otitis media (AOM), whereas 86.7% of those inoculated with H. influenzae developed otitis media with effusion (OME). Antibiotics eradicated H. influenzae from the ME more efficiently in AOM than in OME, and this difference was highly significant (P80% of animals developed culture-negative OME.

The laboratory records of patients with bacillus isolates identified as Corynebacterium xerosis were reviewed in an attempt to establish the clinical significance of the isolates, and the isolated strains were reidentified. Of the 22 strains available for reidentification, four were identified as Corynebacterium striatum and 18 as Corynebacterium amycolatum. Forty-one patients were considered to have Corynebacterium amycolatum isolates, and in 13 (31.7%) of these patients a genuine clinical infection occurred, comprising catheter-related infection in seven cases, surgical wound infection in five cases, and pilonidal cyst infection in one case. Most patients were treated with antimicrobial agents (vancomycin in seven cases and amoxicillin/clavulanic acid in four cases). All patients were cured. Corynebacterium amycolatum can cause genuine infection, usually hospital-acquired, and the clinical significance of isolates must be determined to ensure proper management of patients.

Thirty-two isolates of Corynebacterium urealyticum, isolated between 1991 and 1995, were studied by biochemical tests, phospholipid content, analysis of fatty and mycolic acids, ribotyping, whole-cell protein patterns and antimicrobial susceptibility to six antibiotics. Nineteen isolates were from human and human-related sources (HHRS); the remainder were from animal and animal-related sources (AARS). Most C. urealyticum isolates were similar in their biochemical and whole-cell protein pro®les, although most HHRS isolates were alkaline phosphatase-positive (84%) and produced almost identical protein patterns, whereas AARS isolates were quite diverse. The qualitative composition of cellular fatty acids was identical for all isolates examined. Twelve different ribotypes were obtained with HindIII producing four-to-seven bands. Ribotypes 8, 9 and 10 were predominant in isolates from HHRS, whereas in isolates from AARS, ribotypes 5 and 6 predominated. AARS isolates were signi®cantly less antibiotic-resistant, in comparison with HHRS isolates. Ribotyping appeared to be the most useful tool for strain characterisation.

In an attempt to determine the susceptibility breakpoints for amoxycillin, co-amoxiclav and cefotaxime in pneumococcal pneumonia, a neutropenic mouse model was established and tested with two strains having different susceptibility to penicillins and cefotaxime. With a penicillin-sensitive strain (MIC/MBC = 0.01/0.01 mg/L) the minimum dosage tested achieving significant cure was 2 mg/kg for amoxycillin, co-amoxiclav and cefotaxime. For the penicillin-insensitive strain (MIC/MBC = 1/2 mg/L), the minimum dosage tested giving significant cure was 50 mg/kg for amoxycillin and co-amoxiclav but 100 mg/kg for cefotaxime. Our results support the belief that MICs of amoxycillin, co-amoxiclav and cefotaxime for pneumococcal strains of < or = 0.5 or < or = 1 mg/L can be considered as clinically relevant susceptibility breakpoints.

Acute respiratory infection with penicillin-insensitive Streptococcus pneumoniae (MIC and MBC, 1 and 2 micrograms/ml, respectively) was established in guinea pigs. Intratracheal instillation of 0.5 ml of an overnight culture of S. pneumoniae concentrated 25 times (approximately 3 x 10(9) CFU) induced a bacteremic and fatal pneumonia in > 85% of untreated animals within 46 h, with a mean +/- standard deviation bacterial count of 8.83 +/- 1.11 log10 CFU in lung homogenates. This model was used to evaluate the efficacies of two doses each of amoxicillin, cefotaxime, and meropenem given 1 h after bacterial inoculation. The antibiotics were given at 8-h intervals for up to a total of four injections. The dose of 50 mg of any antibiotic per kg of body weight gave 66.6% survival, compared with 5.05% survival for untreated control animals (P < 0.001). A dose of 200 mg/kg gave a survival rate of 77.8% for meropenem and 83.3% for amoxicillin and cefotaxime, while survival for untreated controls was 11.1% (P < 0.001). Although antibiotic treatment decreased mortality compared with that in untreated controls, the antibiotics contributed to a high early (less than 9 h after bacterial inoculation) mortality, being 53.5% compared with only 6.06% for the untreated controls (P < 0.001). Quantitative cultures of the lungs of animals that died during the 46-h observation period or that were killed after this time showed a significant reduction in the numbers of organisms among treated animals compared with numbers among the control animals (P < 0.001). The described model is an appropriate system for evaluating antibiotic efficacy in invasive pulmonary infection caused by penicillin-insensitive S. pneumoniae.

A gerbil model of otitis media induced by a beta-lactamase producing and non-serotypeable isolate of Haemophilus influenzae was used to assess the in-vivo efficacy of co-amoxiclav and cefuroxime at low (5 mg/kg) and high (20 mg/kg) doses. The MIC of the antibiotics tested against the pathogen was 1 mg/L (1/0.5 mg/L for co-amoxiclav). The organism was inoculated (+/-10(6) cfu) by transbullar challenge directly in the middle ear and antibiotic treatment was commenced 2 h post-inoculation and continued at 8 h intervals for three doses. Only high dose co-amoxiclav significantly reduced the number of culture-positive specimens as compared with untreated animals or with other treatment groups (91.7% as compared with 36.7% for high dose cefuroxime). The results obtained in any treatment group were related to middle ear antibiotic level/MIC. Antibiotic concentrations in the middle ear 90 min after administration were about 10% of serum levels at 15 min, probably related to a slight inflammatory response. Only after high dose co-amoxiclav did the concentration in the middle ear exceed the MIC by a factor of four. In otitis media with effusion, if indicated, antibiotics active in vitro should be administered in high doses and, to avoid side effects, probably in short courses.

A new selective medium (CBU agar) was compared with blood agar (BA) medium for primary isolation of Corynebacteriurn urealyticurn from urine and skin samples of hospitalised patients. Overall, the CBU agar detected C. urealyticum in 14 (4.6%) of 302 urine samples and the BA medium detected the organism in four (1.3%), but most cultures which were positive only on CBU agar had < lo4 cfu/ml. Six strains of C. urealyticum were isolated from 60 skin samples with CBU agar, whereas none was detected with BA. Although most skin samples had heavy inocula, the selective agar facilitated the recognition of low colony counts ( S 10 cfu/plate) of C. urealyticurn by reducing the growth of competing flora. Challenge of the selective medium with reference and clinical strains showed that CBU agar was inhibitory for gram-negative bacteria and reduced the gram-positive flora, allowing the growth of C. urealyticum strains. The new selective medium appears to be a useful epidemiological tool to study urinary and skin colonisation by C. urealyticum.