Low prevalence of blaOXA-143 in private hospitals in Brazil.

We read with great interest C. S. Antonio et al.’s letter describing the high prevalence of Acinetobacter baumannii carrying blaOXA-143 in Brazilian hospitals (1). Recently, we carried out a similar study, and although the blaOXA-143 gene was identified, its frequency was lower than that reported by Antonio et al. (1). During 2008, a total of 803 Gram-negative bacillus isolates, 1 isolate per patient, were collected from 17 private hospitals located in eight cities from four distinct geographic Brazilian regions. Among them, 91 (11.3%) were A. baumannii isolates that were recovered mainly from the respiratory tract (70.3%) and bloodstream (24.2%). Susceptibility testing was performed by CLSI broth microdilution (3). The detection of metallo-lactamase (M L)and carbapenem-hydrolyzing class D -lactamase (CHDL)-encoding genes was performed by multiplex PCR (5, 7, 9) and confirmed by sequencing. The presence of the insertion sequence ISAba1 upstream of the CHDL-encoding genes was also investigated. Genetic relatedness among CHDL-producing A. baumannii isolates, including the first OXA-23producing A. baumannii clone isolated in Brazil (4), was evaluated by pulsed-field gel electrophoresis (PFGE). A total of 83/91 (91.2%) isolates were resistant to carbapenems. We also observed low rates of susceptibility to amikacin (18.7%), ceftazidime (12.1%), cefepime (8.8%), piperacillin-tazobactam (3.3%), and ciprofloxacin (3.3%). In contrast, most A. baumannii isolates were susceptible to polymyxin B (MIC90, 1 g/ml; 97.8% of the isolates were susceptible). M L-encoding genes were not identified in our study, as was also reported by Antonio et al. (1). However, we identified the blaOXA-23 gene in carbapenem-resistant isolates more frequently than in the former study (83.5% versus 41.7%). The blaOXA-23 gene was found in all carbapenem-resistant isolates from the cities of Belo Horizonte, Blumenau, Curitiba, and São Luís, followed by Rio de Janeiro (93.7%), Porto Alegre (80.0%), and São Paulo (69.0%). These results are in accordance with previous local reports (2, 4, 6) that emphasize that this gene is widespread in our country. The ISAba1 element was positioned upstream of blaOXA-23 in all isolates, whereas no insertion sequence was observed upstream of blaOXA-51. Although A. baumannii carrying blaOXA-58 and blaOXA-72 had recently been described in Brazil (1, 8), no isolates carrying these variants were found in our study. Nine distinct PFGE clones were identified among the 76 OXA-23-producing A. baumannii isolates. The predominance of a single clone (clone A [36.8% of the isolates]) was observed in isolates collected from six distinct Brazilian cities. This clone exhibited a PFGE profile similar to that of the first Brazilian clone producer of OXA-23 (4). A. baumannii belonging to clones B (17.1%) and D (9.2%) were also identified in isolates collected from distinct cities, while other genotypes were identified in specific locations. While Antonio et al. (1). reported a high prevalence of the blaOXA-143 gene (58.3%), we found that only 7 of 83 (8.4%) A. baumannii isolates carried this gene. These isolates were collected from a few hospitals located in the cities of São Paulo (n 6) and Rio de Janeiro (n 1). In both studies, the majority of OXA-143-producing A. baumannii isolates were recovered from cities located in São Paulo State. However, while Antonio et al. observed that 70% (21/30) of the isolates from this region carried the blaOXA-143 gene, in the present study, we identified this resistance determinant in only 20.7% (6/29) of isolates collected from São Paulo. Moreover, we have observed the predominance of a single PFGE clone among the seven OXA-143-producing A. baumannii isolates, which contrasts with results obtained by Antonio et al., in which 7 distinct enterobacterial repetitive intergenic consensus sequence (ERIC) PCR clones harbored the blaOXA-143 gene. Nevertheless, in their study, the high prevalence of OXA-143-producing isolates could also be partially justified by the intrahospital spread of a single clone, which corresponded to 57.1% of all OXA-143-producing isolates (1). The high prevalence of blaOXA-23 found in our study may also be justified by intraand interhospital spread of endemic clones. The results of these two studies show that the prevalence of CHDLs may vary according to the disseminated clone in a specific hospital or region and emphasize the importance of appropriate adherence to infection control measures. Thus, wide national surveillance studies are necessary to analyze the real prevalence of CHDLs in Brazilian hospitals.