Sarcopenia of age is prevalent and costly and proven pharmacological interventions are currently lacking. The pathophysiology of sarcopenia is incompletely understood but appears to involve multiple pathways, including inflammation, hormonal dysregulation, impaired regeneration, mitochondrial dysfunction and denervation. There are several ways in which we might select potential pharmacological interventions for testing in clinical trials. These include a 'bottom-up' approach using basic science to elucidate the molecular processes involved and identify potential targets from this knowledge-a strategy that has led to the development of myostatin inhibitors. A 'top-down' approach might use observational data to examine the association between physical function and use of certain medications, such as the association between angiotensin-converting enzyme inhibitors with slower decline in physical function. Once a pharmacological intervention has been proposed, efficacy must be demonstrated in this complex multi-morbid population. Both muscle mass and muscle function need to be measured as outcomes, but these outcomes require large sample sizes and sufficient follow-up to detect change. Biomarkers that can predict the response of sarcopenia to intervention after a short time would greatly assist our ability to select candidate interventions in short proof-of-concept trials. Further development of trial methods is required to accelerate progress in this important area of medicine for older people.