Numerous studies have shown that physically active individuals tend to be healthier. Aerobic fitness and habitual physical activity are associated with reduced all-cause mortality, metabolic risk and cardiovascular disease, even in very active individuals [1–4]. Longitudinal cohort studies have confirmed that changes in physical activity are reflected in altered metabolic risk and mortality over time . Trials of interventions to increase physical activity and fitness among older people have tended to be restricted to specific high-risk subgroups. Several trials have shown that lifestyle interventions reduce the risk of progression to diabetes in individuals with impaired glucose tolerance (IGT) and that the beneficial effects of such interventions persist for several years [6–8]. Exercise interventions also reduce cardiovascular and all-cause morbidity and mortality and lead to subjective improvements in quality of life, even in older individuals with established vascular disease [9, 10].
While at-risk populations such as those with IGT or vascular disease have been shown to benefit from exercise programmes, the efficacy of such interventions to improve health in the general population remains unproven. In particular, strong evidence that aerobic exercise interventions are beneficial in healthy older people is lacking. Notwithstanding this, public health initiatives encourage people to exercise at least five times per week, irrespective of age or disease risk . While such encouragement seems reasonable given our current knowledge of the benefits of physical activity from observational studies, the evidence base for aerobic exercise interventions in healthy older people ought to be more robust. In practice, old age is often an exclusion criterion for such studies.
Many previous studies that have assessed aerobic exercise interventions have yielded physiologically informative data but have inherent methodological limitations in terms of assessing the efficacy of these interventions per se. Some compare different exercise modalities, without including a true "control" group . Where studies have included controls, often only within-group comparisons are reported, rather than considering changes in outcomes in exercisers in the context of changes in the control group : Using the latter approach allows consideration of any measurement effect and would yield a more valid measure of the efficacy of such interventions. Some efficacy analyses have not included study dropouts, or report only on participants who adhered to the intervention in question, rather than conducting an intention-to-treat analysis . Often, attrition rates from exercise interventions are high, thus reducing power . Indeed, inadequate power is frequently cited as an explanation for negative findings in these studies, even where participant retention has been good [16, 17].
Some people respond better than others to aerobic exercise interventions in relation to reductions in body fat, improved fitness and changes in metabolic parameters. Several factors underlie this heterogeneity in the response to exercise. In the diabetes prevention trials, older participants gained more benefit from lifestyle interventions than younger ones did [8, 18]. Certain genotypes are known to affect the metabolic response to exercise . Individuals with a family history of diabetes derive more benefit from exercise than those without such a history . Metabolic improvements associated with exercise are more pronounced in those with higher metabolic risk . However the factors that influence the response to aerobic exercise in healthy older individuals have not been fully described.
Low Birth Weight, Fitness and Metabolic Risk
The thrifty phenotype hypothesis proposes that an adverse intrauterine environment (as manifest by a low birth weight) leads to alterations in the structure and function of various tissues that ultimately confer an increased risk of certain chronic diseases later in life . These alterations arise only when environmental influences occur during certain critical periods of development, when the organism is sensitive to such influences. Studies among individuals exposed to the Dutch "hunger winter" famine of 1944–45 showed that the babies of mothers who were exposed to famine in mid- or late-gestation were born smaller and subsequently had lower glucose tolerance as adults than those whose mothers were only exposed to famine in early gestation or not at all . In animal models, diet restriction in utero or prior to weaning reduces longevity, while dietary restriction after weaning has the opposite effect . Thus, the same environmental exposure can lead to different outcomes, depending on the phase of development in which it occurs. The emergence of programmed changes is modified by adult life factors such as obesity, ageing and physical activity. The highest cardiovascular risk is seen in people who are born small but become overweight as adults . In rural Gambia, even severe malnutrition in early childhood did not lead to adverse metabolic profiles in adults who were lean, fit and consuming a low fat diet . Hence the impact of pre- and post-natal growth on adult disease risk must be considered in the context of adult environmental exposures.
Body fat distribution is a key determinant of cardiovascular risk and foetal growth restriction is known to impact on fat deposition in utero, leading to central adiposity later in life [27, 28]. Whether such changes in body fat distribution persist into late adulthood is unknown. Children born small who showed catch-up growth in the first two years of life had a higher percentage body fat and more central adiposity compared to normal birth weight children at age five years . Birth weight predicts subsequent lean mass at different ages, independently of gestational age, and has been shown to correlate with grip strength in older individuals [30, 31]. While aerobic fitness is known to reduce cardiovascular risk, it also modulates the association between small size at birth and cardiovascular risk, such that the risk associated with low birth weight is more pronounced in less fit individuals . The association between birth weight and muscle strength persists independently of muscle mass, suggesting that cellular and molecular mechanisms are also modulated by birth weight . Whether aerobic exercise-induced changes in muscle function are differential with respect to birth weight is not known. A previous observational study suggested that exercise has a protective effect against progression to diabetes in low birth weight individuals , but higher level evidence from intervention studies is currently lacking.
Myocellular and hepatic lipid deposition
Skeletal muscle is the major site of insulin mediated glucose disposal. Lipid accumulation in skeletal muscle is strongly implicated in the pathogenesis of insulin resistance and type 2 diabetes . Impaired mitochondrial oxidative phosphorylation may also contribute to this problem . Magnetic resonance spectroscopy studies have suggested that intramyocellular lipid content (IMCL) is a major determinant of muscle insulin sensitivity . Exercise modulates several metabolic pathways in skeletal muscle, enhancing glucose uptake and glycogen synthesis and inducing mitochondrial biogenesis. However, exercise has been shown to increase IMCL, and elevated IMCL levels have been found in endurance trained athletes, despite normal insulin sensitivity [37, 38]. Potentially, IMCL is a marker for other lipid intermediates know to suppress insulin sensitivity rather than having a direct insulin desensitising effect in muscle per se . Exercise induced changes in muscle metabolic pathways have previously been shown to markedly improve glucose uptake and mitochondrial function, in all age groups . Whether exercise-induced changes in IMCL in older people are modulated by birth weight has not previously been determined.
Hepatic fat deposition is also a major determinant of insulin sensitivity. While one study in rats suggested that concurrent exercise may prevent liver steatosis induced by a high fat diet, other animal studies suggest there is no exercise effect [40, 41]. Similar negative findings have recently emerged in human studies . The effect of aerobic exercise on intrahepatic lipid (IHL) stores in older individuals has not previously been described. MR spectroscopy is considered the best non-invasive method for quantifying intrahepatic lipid content . However, the technique is relatively time consuming and expensive. The search for more practical alternative methods for detecting liver steatosis has begun in earnest, and ultrasound has shown promise in this regard . Whether it is a reliable method for quantifying milder degrees of liver fat deposition remains to be determined. A secondary objective of this study is to assess the reliability, validity and feasibility of ultrasound measures of liver steatosis that may subsequently be applied in larger epidemiological studies.
As individuals get older, body weight declines primarily because of loss of lean tissue , while IMCL and liver steatosis tend to increase . The decline in muscle mass results from loss of protein content and individual muscle fibres with a preferential atrophy of type II or fast twitch fibres . The inverse relationship between age and physical activity may account for reduced muscle function in older people . Whether interventions to increase physical activity have an effect on muscle structure and function in this age group has not previously been described. Furthermore, how birth weight affects the relationship between muscle function and physical activity remains unknown.