Teicoplanin resistance in methicillin-resistant Staphylococcus aureus in an intensive care unit

Sir During a randomized trial of linezolid versus teicoplanin in the treatment of Gram-positive infection in the critically ill, isolates of Staphylococcus aureus were tested by Etest (AB Biodisk, Solna, Sweden) for susceptibility to teicoplanin and linezolid. The breakpoints for teicoplanin recommended by the British Society for Antimicrobial Chemotherapy are susceptible ≤4 mg/L and resistant ≥8 mg/L; those of the NCCLS are susceptible ≤8 mg/L, intermediate 16 mg/L and resistant ≥32 mg/L. S. aureus with reduced susceptibility to teicoplanin was isolated from four of 204 patients enrolled in the study, two of whom were given teicoplanin and two linezolid, all within a period of 8 weeks. Another patient not in the study was found to harbour a strain with reduced teicoplanin susceptibility. Only one patient acquired the organism during his ICU stay. The organisms were detected on nasal/perineal screens in all five patients and in the sputum from two; bacteraemia occurred in one patient. Twenty-five isolates from these five patients demonstrated teicoplanin MICs of at least 8 mg/L, as determined with Etests; the median MIC was 16 mg/L (range 8–32 mg/L). Disc testing of teicoplanin did not detect resistance. Subsequent testing of the isolates at the Antibiotic Resistance Monitoring and Reference Laboratory by agar dilution showed only seven isolates from two patients with teicoplanin MICs of 16 mg/L and one isolate from another patient with an MIC of 8 mg/L. Of these eight isolates, five were from screens, one from the patient’s environment, one from sternal tissue and one from blood. All 25 isolates were susceptible to linezolid (MIC 1–1.5 mg/L), vancomycin (1–2 mg/L), gentamicin (≤1 mg/L), rifampicin (≤0.25 mg/L), tetracycline (1–2 mg/L), clindamycin (≤0.25 mg/L) and fusidic acid (1–2 mg/L). The methicillin MICs were all >32 mg/L. All eight isolates with confirmed teicoplanin resistance and seven other of the 25 isolates (MIC 2–4 mg/L by reference laboratory testing) were EMRSA-16 and were indistinguishable on sub-typing (lysed by phage 52,75,77). Four isolates were urease positive. Pulsed-field gel electrophoresis (PFGE) of SmaI (Boehringer-Mannheim, Mannheim, Germany) macrodigested genomic DNA was consistent with a single strain. Of the two patients with isolates confirmed to be teicoplaninresistant (or intermediate on NCCLS criteria), the index case (case 1) was a 62-year-old man with diabetes, pulmonary oedema and peripheral vascular disease. MRSA had been isolated from a perianal abscess 10 days earlier and he had then been treated with a low dose of teicoplanin for 6 days (200 mg every 2 days) followed by flucloxacillin. He was treated with linezolid for 1 week after MRSA was re-isolated, and subsequently recovered. Patient 2 was present in the ICU at the same time as patient 1, and had an infected sternal wound for which he had received a 10-day course of teicoplanin. The teicoplaninresistant MRSA strain was detected in his wound, sputum and pleural fluid 3 weeks after that in the first patient. Two other strains of MRSA, susceptible to teicoplanin, were also present during this time (Table 1). He was treated with linezolid for 1 week and then vanco-