In Belgium, the vast majority of salmonellosis is caused by the serovars Enteritidis and Typhimurium, which represent together more than 80% of the isolates. Salmonella enterica serovar Kentucky is a very uncommon serovar that represented 0.02% to 0.5% of the total isolates during the last decade, but 0.8% in 2006. Indeed, this serovar showed a usual increase in Europe during the third quarter of 2006 with several of them acquired during travels to Northeast Africa and Turkey.1 We report here the first Belgian case of a travel-acquired multidrug-resistant Salmonella Kentucky resulting in a treatment failure because of a high resistance level to ciprofloxacin and secondarily acquired resistances to extended-spectrum cephalosporins (ESCs) and trimethoprim + sulfamethoxazole (co-trimoxazole).
In September 2005, a 77-year-old healthy man on a cruise along the North Africa coast developed febrile diarrhoea a few hours after a meal (chicken couscous and dates) in a small restaurant on the Libyan coast. Consequently, he was admitted for 48 h in a hospital in Cairo (Egypt) and treated with intravenous (iv) mezlocillin followed by co-trimoxazole given orally for a few more days whereby he slowly recovered.
Two weeks after his return to Belgium, he presented again with febrile diarrhoea and arthritis of the wrist. After examination of his stool, the Clostridium difficile stool toxin test was positive and Salmonella Kentucky was isolated. The Salmonella isolate was susceptible to co-trimoxazole, but resistant to ampicillin, tetracycline, streptomycin, sulphonamide, nalidixic acid and ciprofloxacin (full resistance) by the standard disc diffusion susceptibility method according to CLSI (formerly NCCLS) guidelines.2 MICs of nalidixic acid and ciprofloxacin determined by standard agar doubling dilution were >1024 and 4 mg/L, respectively. In order to identify the mechanisms responsible for ciprofloxacin resistance, the quinolone resistance-determining regions of gyrA, gyrB, parC and parE were amplified and sequenced as described previously.3 Two mutations leading to substitutions Ser-83 → Phe (TCC → TTC) and Asp-87 → Asn (GAC → AAC) were identified in gyrA. An additional substitution was observed in parC resulting in Ser-80 → Ile (AGC → ATC). In the presence of the efflux pump inhibitor Phe-Arg-β-naphthylamide, the MIC of ciprofloxacin was reduced 4-fold, suggesting that an active efflux mechanism was also present. The involvement of the AcrAB-TolC efflux system was determined by measuring AcrA expression with a method described previously.4 However, the expression of the AcrA protein was equivalent to the baseline production in a susceptible reference strain.
As ambulatory treatment, the patient received a prescription for co-trimoxazole (160 mg of trimethoprim/800 mg of sulfamethoxazole, twice a day) and metronidazole (500 mg, three times a day) given orally for 10 days. One week after drug discontinuation, diarrhoea and arthritis reappeared. A new stool culture was again positive for Salmonella Kentucky. In addition to the profile observed in the first isolate, this strain presented additional resistances to co-trimoxazole (MIC > 256 mg/L) and ESCs [cefotaxime MIC > 256 mg/L, ceftriaxone MIC > 256 mg/L, cefepime MIC = 16 mg/L and showed increased MIC of ceftazidime (4 mg/L)]. The strain was susceptible to meropenem (MIC = 0.5 mg/L). The C. difficile toxin test was again positive for the second stool culture. As diarrhoea persisted, the patient was this time admitted for iv treatment with meropenem (1 g three times a day for 5 days) combined with oral vancomycin (250 mg three times a day for 10 days). Although the stool cultures performed during the hospitalization remained positive for the C. difficile toxin, the diarrhoea disappeared but abdominal pain persisted. Therefore, a colonoscopy was performed and showed non-specific rectal ulcerations. Abdominal pain was resolved with mesalazine enema treatment.
To determine the nature of produced β-lactamases provoking the cephalosporin resistance, PCR amplifications specific for β-lactamase genes of the TEM, SHV and CTX-M families were performed as described previously.5 Only the CTX-M consensus PCR assay gave the expected PCR fragment (545 bp). Sequence analysis of the PCR product revealed 100% identity with the blaCTX-M-1 sequence (GenBank accession number X92506).
To determine the genetic relatedness between the two consecutive Salmonella Kentucky isolates, an XbaI PFGE analysis was performed (Figure 1). Two very similar profiles were observed differing only by one additional band of ∼100 kb in the cephalosporin-resistant strain. To the best of our knowledge, this is the first isolation of a Salmonella resistant to the three antibiotic classes recommended to treat invasive salmonellosis (fluoroquinolones, ESCs and co-trimoxazole) and associated with a treatment failure. The existence of ciprofloxacin-resistant but ESC-susceptible Salmonella Kentucky was also recently reported after their isolation from French patients having travelled in Northeast and Eastern Africa.3 Previously, Salmonella Kentucky was considered as an unsuccessful pathogen because it was rarely associated with human illness, although this serovar was widespread in the food supply (http://msc.tigr.org/salmonella/salmonella_enterica_subsp_enterica_serovar_kentucky_strain_cdc191/index.shtml). However, the spread of an Egyptian-imported Salmonella Kentucky case in two hospitals in the Slovac Republic highlights the pathogenicity and the dissemination and establishment features of this serovar.6 An increased surveillance of such strains should be recommended as it was proven that they could disseminate on a large scale and finally could jeopardize classical antibiotic therapy in patients at risk.
We are very grateful to L. Willems, D. Baeyens, H. Steenhout, F. De Cooman and D. Delbecq for their technical help. Parts of this work were financed by the Federal Public Service Health, Food Chain Safety and Environment and the Flemish and French Communities.
None to declare.