Clinical metabolic studies have demonstrated that insulin action declines progressively with age in humans. In addition to its close association with Type 2 diabetes, which reduces life expectancy in older people, age-related insulin resistance is implicated in pathogenesis of several highly prevalent disorders for which ageing is a major risk factor. These include atherosclerotic cardiovascular disease, dementia, frailty and cancer. Accordingly, insulin resistance may be viewed as biomarker of age-related ill health and reduced lifespan. The rapidly rising number of older people, coupled with a high prevalence of insulin resistance resulting from obesity and sedentary lifestyles, presents unprecedented public health and societal challenges. Studies of centenarians have shown that preserved whole-body sensitivity to insulin is associated with longevity. The mechanisms through which insulin action is associated with age-related diseases remain unclear. Changes in body composition, i.e. sarcopenia and excess adiposity, may be more potent than age per se. Moreover, the impact of insulin resistance has been difficult to disentangle from the clustering of vascular risk factors that co-segregate with the insulin resistance-hyperinsulinaemia complex. Potentially modifiable mediators of age-related changes in insulin sensitivity include alterations in adipocytokines, impaired skeletal myocyte mitochondrial function and brown fat activity. The hypothesis that improving or maintaining insulin sensitivity preserves health and extends lifespan merits further evaluation. Practical non-pharmacological interventions directed against age-related insulin resistance remain underdeveloped. Novel metabolically active pharmacological agents with theoretical implications for some age-related disorders are entering clinical trials. However, recent adverse experiences with the thiazolidinediones suggest the need for a cautious approach to the use of insulin sensitizing drugs in older people. This could be particularly important in the absence of diabetes where the risk to benefit analysis may be less favourable.