Plasmid-mediated quinolone resistance in nalidixic-acid-susceptible strains of Salmonella enterica isolated in Scotland

Sir, Chromosomal mutations in the topoisomerase II gene gyrA and altered drug accumulation were considered to be the only mechanisms of quinolone resistance until 1998, when the plasmidmediated quinolone resistance gene, qnrA, was described in an isolate of Klebsiella pneumoniae. Since then, other plasmidborne resistance genes, qnrB, qnrS (and their variants), aac(60)-Ib-cr and qepA, have been identified. The clinical importance of these qnr genes is that, although not conferring detectable resistance according to the criteria set by the CLSI or BSAC (currently 4 mg/L for ciprofloxacin), their presence facilitates the recovery of mutants with elevated levels of resistance to quinolones. The presence of qnr has frequently been detected in strains producing extended-spectrum b-lactamases (ESBLs). We investigated a total of 70 isolates of Salmonella enterica that were selected based on reduced susceptibility to ciprofloxacin ( 0.125 mg/L) for the presence of qnr genes. Duplicate samples were excluded, and only a single representative isolate from each outbreak was included. Of these 70 isolates, 53 were susceptible to nalidixic acid (,40 mg/L). The remaining 17 isolates exhibited additional resistance to cefotaxime (1 mg/L). Antimicrobial resistances were determined using an in-house breakpoint method. Isolates were screened for qnr genes by a touchdown multiplex PCR using primers described previously. Transconjugants of Escherichia coli J53 harbouring plasmids pMG252, pMG298 and pMG306 were used as positive controls for qnrA1, qnrB1 and qnrS1, respectively. The results of the study are summarized in Table 1. Of the 34 isolates positive for one of the qnr genes, 30 were of human origin; the remaining 4 were isolated from ovine (2), bovine (1) and environmental (1) sources. qnr genes were identified in 12 different serotypes. Salmonella Corvallis was the most common with 13 strains accounting for 38%, followed by 9 Salmonella Typhimurium (26%) and 2 each of Salmonella Stanley and Salmonella Enteritidis (6%). The remaining eight serotypes, Salmonella Virchow, Salmonella Haifa, Salmonella Agona, Salmonella Gaminara, Salmonella Blockley, Salmonella Rissen, Salmonella Colindale and a monophasic strain of Salmonella Java, were each represented by a single strain (3%). qnrA1 was identified in strains of Salmonella Virchow and Salmonella Stanley. qnrB1 was identified in Salmonella Colindale, with qnrB2 being identified in Salmonella Agona and Salmonella Haifa. qnrB5 was identified in Salmonella Gaminara. With the exception of these six strains, all other qnr genes were identified as qnrS1. These results are similar to those of other laboratories. Scientists in Denmark identified qnrS1 in every one of the 23 strains of Salmonella Corvallis they examined, while scientists in England identified the qnrS1 gene in all but 2 of the 39 isolates confirmed as carrying the qnr genes. Of the 18 strains isolated from patients with a history of foreign travel, 16 were identified as harbouring qnrS1, one, qnrA1 and another, qnrB2. With the exception of two strains from South America and Africa, all qnrS1-positive strains originated from the Far East. Reports on the identification of qnr genes have been increasingly frequent around the world; however, this is the first report of their presence in strains of Salmonella isolated in Scotland. The 34 qnr-positive strains belonged to 12 serotypes, including Salmonella Gaminara, Salmonella Rissen, Salmonella Agona, Salmonella Haifa, Salmonella Blockley, Salmonella Colindale and a monophasic strain of Salmonella Java. To the best of our knowledge, there have been no previous reports of the identification of qnr genes in these serotypes. Plasmid-mediated quinolone resistance has serious implications for the use of this group of antimicrobials in the future. More importantly, the presence of these genes on integrons together with ESBLs further supports the necessity for ongoing surveillance of resistance. Surveillance will be problematic, in that routine diagnostic laboratories rarely test for reduced susceptibility to fluoroquinolones. This problem was addressed by the recommendation that nalidixic acid disc screening was indicative of fluoroquinolone susceptibility. However, as this study has shown, isolates that are nalidixic-acid-susceptible can possess qnr genes and have an MIC of ciprofloxacin as high as 0.75 mg/L. However, two cefotaxime-resistant strains with an MIC of ciprofloxacin of 1 mg/L were cross-resistant to nalidixic acid. The elevated MIC may indicate the presence of gyrase mutations in these strains and warrants further investigation. Although these strains are not considered resistant by CLSI or BSAC criteria, elevated MICs of ciprofloxacin have been shown to contribute to treatment failures in the past. This new mechanism of reduced susceptibility in Enterobacteriaceae again raises the question of the appropriateness of the current breakpoints for fluoroquinolones.