The isolate was from stool samples from a 6-month-old girl, seen in the outpatient clinic of Hospital SAMIC Oberá in northern Argentina. The child had vomiting and laboratory-confirmed bloody diarrhea approximately 20 days after a 1-week hospital stay for diarrhea and primary malnutrition. During the hospital stay, she received gentamicin for 1 week, but no bacterial enteropathogen was isolated.

The child lived in Oberá, a subtropical area in Misiones that has no running water. In this region, Shigella spp. is three times more prevalent than Salmonella spp. and other enteropathogens as the cause of pediatric diarrhea, and S. sonnei resistance to ampicillin and TMP-SMX is approximately 43% and 74%, respectively.

S. sonnei was isolated on eosin-methylene blue agar plates. The isolate did not produce gas and was hydrogen sulfide-negative and nonmotile. In triple sugar iron agar, it was negative for citrate, phenylalanine deaminase, and indol production. It was methyl-red positive, Voges-Proskauer negative, ornithine decarboxylase positive, arginine dihydrolase negative, lysine decarboxylase negative, and urease (Christensen) negative. (All media were from Britania, Argentina.) Confirmatory serotyping was carried out with antisera from the Instituto Nacional de Microbiología Dr. Carlos Malbran.

Confirmatory susceptibility tests followed conventional methods (12,13). Briefly, MICs were determined by the agar dilution method, using Mueller-Hinton agar (Britania) and inoculums of 10⁴ CFU per spot; plates were incubated 18 hours at 35°C. Escherichia coli ATCC 25922 and E. coli ATCC 35218 were included as quality controls. Antibiotics tested were ampicillin, ampicillin + clavulanate (CLA), cefoxitin, cefotaxime (CTX), CTX+ CLA (CTX/CLA), ceftazidime, and ceftazidime + CLA (ceftazidime/CLA); a fixed concentration of 4µg/mL lithium CLA was used when combined with beta-lactam drugs. Antimicrobial drugs were provided by Argentia, Argentina (Ampicillin, CTX), Sigma Chemical Co., St. Louis, USA (ceftazidime), Roemmers, Argentina (CLA) and Merck Sharp & Dohme, Argentina (cefoxitin). MICs were as follows: ampicillin, >1,024 µg/mL; ampicillin + CLA 16 µg/mL; cefoxitin, 2 µg/mL; CTX, 16 µg/mL; CTX/CLA, < 0.063 µg/mL; ceftazidime, 1 µg/mL; ceftazidime/CLA, < 0.063 µg/mL.

Resistance to gentamicin, amikacin, and TMP-SMX was detected by agar diffusion (13); quality controls also included E. faecalis ATCC 29212.

The presence of an ESBL was confirmed by microbiologic and biochemical tests using different third-generation cephalosporins as substrates for the enzymes present in bacterial sonicates and an iodometric detection system (9). A microbiologic confirmation test for ESBLs was performed according to recommendations of the National Committee for Clinical Laboratory Standards for E. coli and Klebsiella spp (13). After isoelectric focusing (9), crude bacterial extracts rendered two bands that hydrolyzed 500 µg/mL ampicillin (isoelectric points 5.4 and 8.2), the latter also active on 1,000 µg/mL ceftriaxone. As this enzyme was likely CTX-M-2 (the band comigrated with authentic CTX-M-2 samples) and the first probably corresponded to TEM-1 (or another related enzyme), plasmid DNA obtained by the method of Birnboin and Doly (14) was used as the template for polymerase chain reaction amplification, with specific primers for CTX-M-2 (bla_CTX-M-2 I: 5'-TTAATGATGACTCAGAGCATTC-3'; bla_CTX-M-2 II: 5'-GATACCTCGCTCCATTTATTG-3') and TEM-1 (bla_TEM-1 I: 5'-ATAAAATTCTTGAAGACGAAA-3'; bla _TEM-1 II 5'-GACAGTTACCAATGCTTAATCA-3'). Two fragments of 0.9 kbp and 1.2 kbp were obtained; they showed 100% agreement with theoretical and experimentally obtained fragments from bona fide CTX-M-2 and TEM-1 producing strains, respectively (8,10).

No resistant Shigella was isolated from the other 124 pediatric patients admitted that month for nonsurgical (85 patients) or surgical reasons (29) nor was a nosocomial outbreak caused by a third-generation cephalosporin-resistant enterobacteria detected. The patient recovered. The isolated microorganism was not the likely cause for the patient's first hospitalization as it could not be isolated at that time. Likely alternatives for its acquisition are 1) intestinal selection or acquisition of a resistant enterobacterium and in vivo transference to a freshly acquired Shigella or 2) direct acquisition of the resistant strain from contaminated water (not sustained, as no other resistant Shigella was obtained from the same community in the following year).

These findings, which may portend the spread of serious resistance in Shigella throughout Argentina and beyond, suggest the need for susceptibility testing of all Shigella spp. whenever financially feasible.