Vancomycin remains the drug of choice for serious invasive infections caused by MRSA. Although there are newer agents active clinically against MRSA, none have proven consistently superior to vancomycin in head-to-head studies. Unfortunately the accumulating evidence suggests that vancomycin does not provide the same degree of efficacy against MRSA achieved with ß-lactams against methicillin-susceptible strains. Reasons for the apparently inferior efficacy have not yet been elucidated. Possibilities include factors that affect pharmacokinetics such as renal function and obesity, reduced penetration to certain infection sites such as the lungs (epithelial lining fluid), organism tolerance, the mergence of heterogeneous and homogeneously resistance vancomycin intermediate strains, so-called MIC-creep, the slower killing rate and the slower clearance of bacteraemia by vancomycin compared to ß-lactams, and insufficient drug exposure in many patients to reach the pharmacodynamic target. The evidence for each or any of these variables is patchy, leaving us uncertain about what interventions should be considered to increase efficacy. Combinations of vancomycin with other agents have not yet been shown to improve efficacy significantly, and agents such as rifampicin in particular have notable problems with resistance emergence and drug interactions. The substantial differential in acquisition costs between vancomycin and newer agents suggests that switching away from vancomycin altogether would be expensive and for uncertain clinical gain. Until the cost benefit of newer agents can be proven, strategies for vancomycin use are needed even though we believe it to be suboptimal. Avoiding its use wherever possible for serious methicillin-susceptible infection seems prudent. For serious MRSA infections, confident ruling-out of hetero-resistance and optimising drug exposure are likely to be the most important strategies. Monitoring levels to ensure pharmacodynamic targets are achieved, rather than reduce toxicity risk, should be considered in all patients with serious MRSA infection, ideally by a method that estimates AUC.