Epidemiology of healthy ageing and the idea of more refined outcome measures

Most of us reading this article will reach deep old age and would prefer to do so disease-free and fully functional (Fig. 1). As virtually all industrialized countries and many less developed nations are now facing rapid population ageing, healthy ageing is also an important topic for society. Unsurprisingly, the need for reliable evidence to inform policy on healthy ageing has spawned a large number of longitudinal studies of ageing over the past 15 years. This issue of the IJE profiles three of them—the Longitudinal Aging Study Amsterdam (LASA)1, The Irish Longitudinal Study on Ageing (TILDA)2 and the Bambuí Cohort Study of Ageing3—and in our editorial we explore ideas for potential next steps in this field of research. Many of the new longitudinal studies of ageing are tailor-made for the purpose, and mostly1–6 but not exclusively,7 focus on older adults. LASA, TILDA and the Bambuí Study are excellent examples of such studies. In common with many studies of ageing in high-income countries LASA and TILDA were set up to examine determinants of change in the physical, cognitive, social and emotional components of functioning.1,2,8,9,10 The Bambuí Study is from Brazil where, in common with other Latin American and Caribbean countries, demographic change is happening quickly. What this means is that many of those now at risk of chronic conditions in old age have survived epidemics of infectious diseases in childhood. The study's aim is to examine the combined effects of Trypanosoma cruzi infection (T. cruzi causes Chagas disease which can result in cardiac damage) and non-communicable diseases on health outcomes in old age.3 Another group of ageing studies is comprised of ‘new’ longitudinal investigations on birth cohorts and other studies originally set up to examine a different set of questions. The Editor's Choice in February's issue of the IJE commented on a web-based, updated cohort profile of the National Survey of Health and Development (NSHD). Also known as the British 1946 Birth Cohort, it is a perfect example of an existing study which has, almost by default become a study of ageing.11 As a birth cohort, it is particularly well placed to look at early life influences on healthy ageing as well as how experience and exposures across the whole life course affect the chance of healthy ageing. The Doetinchem Cohort Study, in common with other longitudinal studies like Framingham that started off with a focus on cardiovascular disease, has also refocused on healthy ageing.12 A wide range of health-related outcomes are covered by Doetinchem, including the major chronic diseases, bone density, osteoarthritis, osteoporosis and fractures, other chronic diseases (such as diseases of the bowel, liver and thyroid gland), incontinence, muscle strength and tests of physical functioning. Expectations for studies in this research area are obviously high. However, major studies on longevity have already found that people with few risk factors for fatal diseases, such as cardiovascular disease, cancer and diabetes, tend to live longer. As healthy ageing means the absence of disease, critically the absence of fatal disease, one may wonder what the findings from these new ageing studies will add to what is already known about risk factors for those diseases that are the commonest causes of death. Or, put another way, how can the new studies on ageing avoid merely duplicating conclusions from the epidemiology of major chronic conditions? Previous success stories in disease epidemiology may provide some guidance. Discovering the association between smoking and lung cancer13 or cholesterol and coronary heart disease are great victories of this discipline.14 Nevertheless, stroke may be a more useful example for us, as the findings on potential risk factors, such as major lipids, tended to be inconsistent until researchers started to take into account that stroke is a composite outcome. There is a clear link between dyslipidaemia and stroke when haemorrhagic stroke is excluded from the outcome, reflecting differences in the aetiology of the ischaemic and haemorrhagic stroke subtypes.15 To date, findings from ageing studies are far from consistent. Definitions of healthy ageing include a mixture of different outcomes, encompassing traditional ideas related to long lifespan and freedom from disease, as well as broader considerations, including functioning and independence, quality of life and participation in society and the workforce. If the lessons from stroke research apply here, refinement of these rather broad definitions of healthy ageing could facilitate future research in this field. First, it is important to recognize that those who live long are not a uniform group: many of them are ‘survivors’ (those living with extended morbidity due to age-related disease diagnosed before old age), but some are ‘delayers’ (age-related disease diagnosed after the average life expectancy of their birth cohort) or ‘escapers’ (those who attain very old age without any major disease). Kaplan et al.16 have illustrated this nicely using data from repeat measures of self-rated health in the Alameda County study over a period of 30 years; Group A in Figure 2 represents the escapers and Group C the survivors (for present purposes ignore Group B, as we are only interested in long lifespans). A smiling old man (Reproduced with permission www.istockphoto.com) Three possible health tranectories over decades. Reproduced with the kind permission of Oxford University Press Already in the 1980s Fries’ classic ‘compression of morbidity’ hypothesis had suggested that the focus of ageing studies be narrowed to the delayers and escapers, that is, an examination of factors compressing the burden of lifetime illness into a shorter period before the time of death.17 More recently, the National Institute on Aging Advisory Board (2004) promoted a more stringent definition, ‘exceptional survival’, i.e. disability free, in addition to morbidity free survival.18 Further to refine this definition, one has to drill down into the concept of healthy ageing and identify a subset of people who experience exceptionally healthy ageing: living until old age not only without major morbidity and disability but also retaining a high level of functioning. Identification of such people needs to take into account that age-related decline in functioning begins early in life. In six US cohorts, for example, age-related decline in walking speed was already evident after the age of 30 years.19 Correspondingly, upward systolic blood pressure trajectories after the age of 40 years have characterized adult populations participating in major UK studies, such as the Caerphilly study, NSHD and the Whitehall II study.20 Analyses have additionally shown surprisingly early age-related decline in various domains of cognitive function, such as memory, reasoning and fluency, starting by the age of 45 years.19–21 Exceptionally healthy ageing is therefore characterized by a slower and smaller decline in functional trajectories across the adult life course. The determinants of this ideal ageing phenotype, both modifiable and non-modifiable, are certainly of general interest. In animal models, for example, mechanisms such as preserving high telomere length (complex DNA protein structures that cap and stabilize the physical ends of chromosomes) or function of the insulin/IGF-1 pathway have been implicated in extended lifespans, but so far success in replicating these findings in humans has been limited.22 It might be worth looking at whether studies using a more stringent definition of healthy ageing will be more successful. A recent investigation from the US Physician's Health Study showed smoking abstinence, weight management, blood pressure control and regular exercise to be associated with an enhanced lifespan,23 and in a study of Japanese American men, avoidance of hyperglycemia, hypertriglyceridemia and heavy drinking were additionally associated with extended longevity.24 National guidelines recommend target levels for all such conventional risk factors for the prevention of specific diseases. It would be useful to know whether these target levels need to be lower, or within a narrower range (cf. body mass index), if the goal is exceptionally healthy ageing. Some readers may wonder whether the additional criteria we have used to define exceptionally healthy ageing have led us to sketch an unfeasible concept. A quick look at empirical data suggests the answer is no. At least in the Whitehall II study of British civil servants an exceptionally healthy ageing phenotype, with good functional status across the adult life course, exists, although it is rarer than the conventionally defined healthy ageing phenotype. A total of 30% of this relatively privileged cohort was free of long-term illness and disability at the age of 60 years or older. However, only 7% additionally preserved good musculoskeletal, cardiovascular and metabolic functioning (indicated by above median age-standardized test scores) plus good cognitive function and mental health. We urge new and existing ageing studies to consider conducting analyses of more refined outcome measures, such as exceptionally healthy ageing, and anticipate that such approaches might result in identification of new intervention targets to facilitate healthy ageing beyond those already identified by chronic disease epidemiology and traditional ageing research.