We evaluated, by an improved susceptibility testing method, the prevalence and significance of low-level glycopeptide resistance in methicillin-resistant Staphylococcus aureus (MRSA) isolates, which belonged to a previously described, retrospective cohort of patients treated for orthopedic device-related infections (ODRI) at the Geneva University Hospital between 2000 and 2008. Fifty-seven individual or multiple isolates were retrieved from 41 ODRI patients for glycopeptide susceptibility and clonality studies, including 20 patients with prosthetic joint (PJ) and 21 with osteosynthesis (OS) MRSA infections. Low-level glycopeptide resistance was detected by elevated teicoplanin or/and vancomycin minimum inhibitory concentrations (MICs = 4 mg/L), as determined by a previously validated combination of macrodilution and agar dilution assays of improved sensitivity. MRSA isolates with elevated teicoplanin MICs were detected in 20/41 (49 %) ODRI patients at the onset or during the course of glycopeptide therapy, namely, in 10 of 20 patients with PJ and 10 of 21 patients with OS infections. Only one isolate developed a concomitant increase in vancomycin MIC during therapy. 13/20 (65 %) glycopeptide-intermediate S. aureus (GISA)-infected patients, including 7/10 (70 %) with PJ and 6/10 (60 %) with OS, experienced treatment failure. In contrast, therapy failed in only 5/21 (24 %) ODRI patients with non-GISA isolates (p = 0.012), including 2/10 (20 %) with PJ and 3/11 (27 %) with OS infections. The emergence of low-level teicoplanin resistance could not be explained by teicoplanin administration, since only four patients received teicoplanin. The evaluation of low-level teicoplanin resistance may improve the detection of GISA isolates. Further studies are warranted to evaluate the impact of low-level teicoplanin resistance on the outcome of glycopeptide therapy.