Frailty can be broadly defined as the vulnerable health status that occurs in older adults. With the ageing population understanding frailty is becoming an increasingly important issue. While no consensus exists on the exact definition of frailty, two models have become prominent in geriatric research. These are the frailty phenotype, a model that measures frailty according to the syndromic aggregation of 5 physical criteria and the frailty index (FI), a broad index of age related health deficits. While recent years have seen a substantial increase in research into frailty, there remains a relative paucity of data from European studies and studies in men. Of the many mechanisms suggested to contribute to frailty there has been particular interest in the role of declining levels of anabolic hormones, partly because replacement of these hormones represents a potential strategy for managing this condition. The broad aim of this thesis was to explore the condition of frailty and its relationship to anabolic hormones, particularly testosterone (T) in ageing European Men. This project involved analysis of data from 2 studies: The European Male Ageing Study (EMAS), a longitudinal cohort study of 3369 men from 8 European centres and a trial of T treatment in 262 men with low testosterone and symptoms of frailty. A set of phenotypic frailty criteria were developed for use in the EMAS, using this model the prevalence of frailty was 2.6%. This increased with age from 0.1% in men aged 40-49 up to 6.7% in men aged 70-79. This model was compared against an FI, the correlation between the two models was moderate, r=0.41, and both models were related to incident falling at 2 year follow up; Ordinal OR (95% CI), 3.15 (1.75 to 5.66) for the frailty phenotype and 5.28 (3.35 to 8.32) for the FI in adjusted analyses. In the hormone analyses frailty was related to lower free T according to both models, Ordinal OR (95% CI); 1.19 (1.02 to 1.39) for the phenotype and β-coefficient (95% CI); 0.006 (0.003 to 0.009) for the FI (FI values range from 0-0.7). Free T was particularly related to the sarcopenia criteria OR (95% CI); 1.40 (1.09 to 1.80). Frailty was also related to LH, FSH and SHBG. Deficiency in multiple anabolic hormones was related to frailty, in adjusted analyses each additional deficiency was associated with an RRR (95% CI); 1.71 (1.38 to 2.13) increased risk of phenotypic frailty and a β-coefficient (95% CI); 0.016 (0.012 to 0.02) increase in FI score. The trial analyses focussed on a 6 month post treatment follow up phase. It was found that gains in lean mass and muscle strength were not maintained at 6 months post treatment. The adjusted difference between groups at 6 months post treatment for knee extensor strength was 4.0 (-3.9 to 11.9) Nm compared to 8.1 (-0.2 to 16.5) Nm at the end of treatment, similarly the difference in lean mass declined from 1.2 (0.8 to 1.7) kg at end of treatment to 0.3 (-0.1 to 0.8) kg at 6 months post treatment. In summary, the frailty phenotype was adapted and validated for use in the EMAS study. Analyses using this model and the trial follow up analyses are supportive of an influence of T on lean body mass in ageing men. The other hormone relationships seen suggest frailty may be broadly related to changes in the endocrine system in ageing men. The lack of sustained benefit from T treatment combined with the relationships with multiple endocrine markers suggests more complex management strategies may be required for this condition.