Kouichi Shimakawa

Metallo-β-Lactamase-Producing Gram-Negative Bacilli: Laboratory-Based Surveillance in Cooperation with 13 Clinical Laboratories in the Kinki Region of Japan

ABSTRACT A total of 19,753 strains of gram-negative rods collected during two 6-month periods (October 2000 to March 2001 and November 2001 to April 2002) from 13 clinical laboratories in the Kinki region of Japan were investigated for the production of metallo-β-lactamases (MBLs). MBLs were detected in 96 (0.5%) of the 19,753 isolates by the broth microdilution method, the 2-mercaptopropionic acid inhibition test, and PCR and DNA sequencing analyses. MBL-positive isolates were detected in 9 of 13 laboratories, with the rate of detection ranging between 0 and 2.6% for each laboratory. Forty-four of 1,429 (3.1%) Serratia marcescens, 22 of 6,198 (0.4%) Pseudomonas aeruginosa, 21 of 1,108 (1.9%) Acinetobacter spp., 4 of 544 (0.7%) Citrobacter freundii, 3 of 127 (2.4%) Providencia rettgeri, 1 of 434 (0.2%) Morganella morganii, and 1 of 1,483 (0.1%) Enterobacter cloacae isolates were positive for MBLs. Of these 96 MBL-positive strains, 87 (90.6%), 7 (7.3%), and 2 (2.1%) isolates carried the genes for IMP-1-group MBLs, IMP-2-group MBLs, and VIM-2-group MBLs, respectively. The class 1 integrase gene, intI1, was detected in all MBL-positive strains, and the aac (6′)-Ib gene was detected in 37 (38.5%) isolates. Strains with identical PCR fingerprint profiles in a random amplified polymorphic DNA pattern analysis were isolated successively from five separate hospitals, suggesting the nosocomial spread of the organism in each hospital. In conclusion, many species of MBL-positive gram-negative rods are distributed widely in different hospitals in the Kinki region of Japan. The present findings should be considered during the development of policies and strategies to prevent the emergence and further spread of MBL-producing bacteria.

High Frequency of Antibiotic-Associated Diarrhea due to Toxin A-Negative, Toxin B-Positive Clostridium difficile in a Hospital in Japan and Risk Factors for Infection

Patients hospitalized in a hospital with a high incidence of antibiotic-associated diarrhea due to toxin A-negative, toxin B-positive (A−/B+) Clostridium difficile were retrospectively investigated to determine the clinical manifestations and risk factors for infection. Of 77 Clostridium difficile isolates obtained from 77 patients during the 1-year investigation period, 30 were A−/B+ and 47 were toxin A-positive, toxin B-positive (A+/B+). By pulsed-field gel electrophoresis analysis, 23 of the 30 A−/B+ strains were outbreak-related, suggesting nosocomial spread of a single type of bacterium, which mainly affected patients in the wards of respiratory medicine, hematology and neurology. Using regression analysis, three factors were found to be associated with infection by A−/B+ isolates: (i) exposure to antineoplastic agents (P=0.01, odds ratio [OR]=5.1), (ii) the use of nasal feeding tubes (P=0.008, OR=5.2), and (iii) assignment to a certain internal medicine ward (P=0.05, OR=3.0). Between patients with Clostridium difficile-associated diarrhea caused by A−/B+ strains and those with A+/B+ strains, no statistically significant difference was found in body temperature, serum concentration of C-reactive protein, leukocyte count in whole blood, frequency of diarrhea, or type of underlying disease. These results indicate that A−/B+ strains of Clostridium difficile can cause intestinal infection in humans and they spread nosocomially in the same manner as A+/B+ strains.

Evaluation of MicroScan ESBL confirmation panel for Enterobacteriaceae-producing, extended-spectrum β-lactamases isolated in Japan

We assessed use of the MicroScan ESBL confirmation panel (Dade Behring, Tokyo, Japan) for the detection of eight Enterobacteriaceae-producing extended-spectrum beta-lactamases (ESBL) species. Of 137 bacterial strains isolated from patients in 32 hospitals in the Kinki area of Japan, 91 produced ESBL and comprised 60 bacteria (of E. coli, K. oxytoca, and K. pneumoniae) targeted by the NCCLS ESBL test and 31 non-target bacteria such as chromosomal AmpC-producing bacteria (e.g., Serratia marcescens, Enterobacter spp.). Sensitivity and specificity of the MicroScan panel for the target bacteria were 92% and 93%, respectively; sensitivity and specificity for non-target bacteria were 52% and 100%, respectively. There were 20 ESBL-positive strains that were not inhibited by clavulanic acid in the MicroScan panel (3 of 32 ESBL-producing E. coli strains, 1 of 24 K. pneumoniae, 1 of 4 K. oxytoca, 8 of 13 E. cloacae, and 7 of 14 S. marcescens), and most of them were bacteria not targeted by the NCCLS test. In 19 of the 20 strains, the synergy effect of clavulanic acid was observed in the modified-double-disk synergy test using only the cefepime-disk. Because these strains had high MICs of > or = 16 microg/ml for cephamycins such as cefoxitin and cefmetazole, these strains might produce high levels of AmpC in addition to ESBL. The MicroScan ESBL confirmation panel showed excellent performance in detecting target, but not other bacteria. Addition of cefepime and clavulanic acid to the MicroScan panel may significantly improve detection of non-target bacteria.

Monte Carlo simulation for evaluation of the efficacy of carbapenems and new quinolones against ESBL-producing Escherichia coli

Extended-spectrum β-lactamase (ESBL)-producing bacteria are known to be resistant to penicillins, cephalosporins, and monobactams because of their substrate specificity, and these bacteria are sensitive only to a narrow range of antimicrobial agents. The present study was undertaken to evaluate the efficacy of carbapenems and the new quinolones against ESBL-producing Escherichia coli, using a Monte Carlo simulation based on the pharmacokinetic/pharmacodynamic (PK/PD) theory. The time above MIC (TAM, %) served as the PK/PD parameter for carbapenems, with the target level set at 40%. The AUC/MIC served as the PK/PD parameter for the new quinolones, with the target level set at more than 125. In the analysis of drug sensitivity, the MIC50 of all carbapenems other than imipenem was low (0.03 μg/ml), while the MIC50 of the new quinolones was higher (1–2 μg/ml). The probability of achieving the PK/PD target with carba penems after two doses at the usual dose level, as determined by the Monte Carlo simulation, was high for each of the carbapenems tested (99.0% for biapenem, 99.60% for meropenem, and 95.03% for doripenem), except for imipenem. Among the new quinolones, the highest probability of achieving the PK/PD target was obtained with pazufloxacin (42.90%). Thus, the results of the present study have revealed that carbapenems are effective at the regular dose and can be used as the first-choice antibiotics for ESBL-producing E. coli because the resistance ratios for carbapenems are low compared to those of the new quinolones.