The last two decades of life course epidemiology, and its relevance for research on ageing
The term ‘life course epidemiology’ was coined in 1997 with the publication of the first edition of A Life Course Approach to Chronic Disease Epidemiology.1 This book reviewed the pre-adult risk factors for cardiometabolic and respiratory disease, the catalyst being the imaginative research on the fetal origins of adult disease being driven forward at that time by Professor David Barker. We defined life course epidemiology as ‘the study of long-term biological, behavioural and psychosocial processes that link adult health and disease risk to physical or social exposures acting during gestation, childhood, adolescence, earlier in adult life or across generations’.1 Although our definition of life course epidemiology has stood the test of time, the field has evolved and there have been conceptual developments, methodological innovations which facilitate efforts to test these concepts, and an increasing corpus of empirical research demonstrating how factors from earlier life are associated with later life health and disease, as well as the pathways and biological mechanisms that may be involved. These developments have generated further ideas and challenges to life course models in an iterative process. As the theme of this special issue suggests, one important development has been the gradual shift of research focus from clinical disease endpoints to multi-faceted traits and longitudinal trajectories of functional phenotypes that can be assessed well before any clinical threshold is reached. This has naturally led on to the application of a life course epidemiological approach to ageing. The purpose of this overview is therefore to assess the development and current state of the field of life course epidemiology, including its recent application to the study of ageing as the focus of this special issue. Life course epidemiology is now recognized as a field of research in its own right and was added to the fifth edition of the Dictionary of Epidemiology.2 The original life course book was to become the first in a series on a life course approach to adult health,3–8 as well as a journal glossary9 and chapters in a range of public health, epidemiology and other academic textbooks.10–14 The original International Journal of Epidemiology editorial on life course epidemiology, published in 2002,15 is the fifth most highly cited article in the journal’s history (search run on Web of Science core collection 19 February 2016). In order to chart key developments in the field, we have drawn on our own experiences and the results of three literature searches: (i) a cited reference search of four key conceptual life course publications,1,4,9,15 and Medline searches for publications including the terms (ii) ‘life course’ and (iii) ‘developmental origin’ or ‘fetal origin’ (see Box 1). Box 1. Reviewing the published literature To assess developments in life course epidemiology since the inception of this term in 1997, we aimed to capture relevant papers published in this area up to the end of 2014 using two searches (undertaken in February and March 2015): 1. A cited reference search of four publications on life course epidemiology (from the date of their publication to 31 December 2014): I. Kuh D, Ben Shlomo Y (eds). A Life Course Approach to Chronic Disease Epidemiology. 1st edn. Oxford, UK: Oxford University Press,. 1997. II. Kuh D, Ben Shlomo Y (eds). A Life Course Approach to Chronic Disease Epidemiology. 2nd edn. Oxford, UK: Oxford University Press, 2004. III. Ben-Shlomo Y, Kuh D. A life course approach to chronic disease epidemiology: conceptual models, empirical challenges and interdisciplinary perspectives. Int J Epidemiol 2002;31:285–93 IV. Kuh D, Ben-Shlomo Y, Lynch J, Hallqvist J, Power C. Life course epidemiology. J Epidemiol Community Health 2003;57:778–83 2. A free-text search to identify all literature indexed in Medline which: included the term: life course (i.e. ‘life ADJ course’ or ‘life-course’ or ‘lifecourse’), in the title or abstract; was published between 1 January 1990 and 31 December 2014; was published in the English language and indexed as study of humans. To enable us to draw comparisons with developments in the Developmental Origins of Health and Disease (DOHaD) we undertook a third search (in March 2015): 3. A free-text search to identify all literature indexed in Medline which: included the term: ‘developmental origin*’ OR ‘fetal origin*’ OR ‘foetal origin*’ in the title or abstract; was published between 1 January 1990 and 31 December 2014; was published in the English language and indexed as study of humans. The results of each of these searches were downloaded to reference management software and duplicates were removed. See Supplementary Data (available as Supplementary Data at IJE online) for number of search results returned. Each result from search 2 was screened and classified by Y.B.S., R.C. or D.K. using standard criteria based on its relevance, type of paper, outcome, exposure and analyses, timing of exposures and setting. We have, from the outset, highlighted the long-term history of life course concepts and the fact that disciplines outside epidemiology, such as demography, sociology, anthropology and psychology, have considered these broad ideas for longer.3,9,16,17 (see commentary by Alwin18). It was thus no surprise that our literature searches, even when restricted to Medline, returned empirical as well as narrative and review papers from journals in these disciplines as well as epidemiology and public health, highlighting the broad appeal of a life course perspective. Alwin18,19,20 discusses the conceptual and methodological developments and the growth of empirical evidence in these other fields and draws parallels with the developments in epidemiology. Life course perspectives in epidemiology were eclipsed in the mid 20th century by the almost exclusive focus on the adult lifestyle model of disease causation.21 The catalyst for the revival of life course thinking in epidemiology was the seminal work of David Barker and colleagues on the fetal origins of adult disease (FOAD)22 from the late 1980s, which was subsequently extended to the developmental origins of health and disease (DOHaD) at the turn of the century.23,24 The first of three phases in the development of life course epidemiology reflects this: a focus on the associations of early life factors with clinical disease endpoints. Life course epidemiology is often incorrectly interpreted as being synonymous with DOHaD, but there are both similarities and differences (see below) and interestingly there was only limited overlap found between publications in these two fields (see Figure 1). Venn diagram showing overlap between search results. We consider the second main phase in the development of life course epidemiology to be the expansion of the field to consider a wider range, and more integrated study, of outcomes, including different aspects of physiological function and their natural history across life, and of biological, psychological and social risk factors acting independently, cumulatively or synergistically across different life stages. These developments were possible because of the growing number of birth cohort and other longitudinal studies with increasingly rich and the development of life course that such The study of functional is further and this of cohort studies with relevant has led to the third and most recent phase in the development of life course the application of a life course epidemiological approach to the study of ageing. a for life course epidemiology almost and of the ideas and subsequently and This is by evidence of a in the number of papers that one of the four key or ‘life course’ in their title or (see Figure is as there has been a in the number of publications in and we that the term ‘life course’ is and This the by in review of life course concepts in social that life course become a all to to all of different aspects of experiences in In was that a number of papers the term ‘life course’ in their this have that this may be in In of our own that we as of life course epidemiology, were by the searches, as and of epidemiological papers focus on study exposures and and the conceptual approach a with current and of publications by literature searches 1 and 2 by publication these we in the published literature which our on the key phases of development in the field, and evidence of the between papers empirical studies and other of In the of empirical we to capture the of and exposures and the in which the studies are and relevant and publications by these searches are cited the of the papers by our Medline search for publications that included the term ‘life course’ in their title (i.e. search empirical of of Supplementary as Supplementary Data at IJE these empirical papers there was evidence that studies of and and were the most across all as were studies of there has been a in the life course or life perspective. was evidence of a in the of being time, with more recent empirical papers life course of a wider range of including health, and and were a number of papers on as and physical were but other included and with these behavioural trajectories different life In terms of there has been an the in the of empirical studies which model exposures as or trajectories life stages. The of an in and on later an because such exposures more to being a life course have been across life and can be by a of The of early psychosocial and and behavioural risk factors to be In terms of the of studies are based in the and the which has been by the of from birth and cohort in the there has been an in the number of studies in other and a in relevant studies being in and (see commentary by and The of study is to which life course associations are and which are A but increasing number of papers (see below) which of these We have the of the or model as a for life course epidemiology. We this as one of and models that to be In as and social differences in disease risk is as well as more mechanisms that may epidemiological We that the term ‘life course’ has in the at the on DOHaD, and the term ‘life course has been by is important to are the similarities and differences in The model that to the of any can be to different time (i) is the (ii) in earlier life and the of and and (iii) natural Chronic disease may a is in the a to a in be the adult is and to type This is to the model of with early be by even these at later as as the on these to The is when there are on that models that later life has no on the of or This may be for can an important in the health and of and in and the of have of on for We with a recent review which that the early of the term was in an of DOHaD, and was by the empirical with life course epidemiology or the approach research in how developmental factors are associated with adult function and disease, but there are important Life course epidemiology is in the of function or its of and the factors across the of life that The focus of is on developmental mechanisms an but this is a and is relevant to outcomes, research has on the developmental origins of chronic cardiometabolic In life course epidemiology with its public health in both social and biological and has been to a wider of outcomes, as by our literature DOHaD, its growing the has on in or in with a and has been of the for later life to disease risk or disease A life course approach is and to empirical evidence to the most timing of any be at a or to The approach has been more in the but the life course approach has from interdisciplinary with for research to the growth and how has life course studies in in The commentary by and the to life course epidemiology with from We that both have to each the limited overlap in the literature to date (see Figure and further in how ideas from both can be to research (see commentary by and for further In our original work we conceptual models such as models, as by DOHaD, and or of risk models which are more in the social epidemiology We these ideas using at this time we were of the growing literature on (see commentary by and for a more of in to life course concepts and approach to exposure in the of and these and models were as different of models, we have since that is more to or models as special of an model when the of an exposure the of exposures different time up in a but in the of in a or in a that can only be in one time by In our we an of risk as the of different exposures with and highlighted the issue of of our literature searches that most papers that to this model the exposure of a such as of social in a This is a standard epidemiological approach and of to or epidemiology. as to how to model the of and A life course approach the of timing and We have this with a (see Figure the between of exposure and A model be with an to a model consider of during or the risk on and course there may be other pathways that early and may exposure to at the time of disease and the may be but by early for evidence of or in models of the between and risk of papers have trajectories of using a of models, models, models, reference for a review of this area and commentary by and to different exposure are associated with As an one can the as to the chronic disease risk associated with is different for have been at the end of the and have were and in or early or mid has been to a approach which as for behavioural or the of and of and The of a was early life, by It is increasingly being that biological and social during other life are to be exposures may have more at other as birth cohort studies have and with rich on adolescence, and mid have the timing of the growth to be important for of growth and in across early and mid are associated with a and of in early can be but be by a of trajectories (see as or in a may the of an this is at and driven by Life course models have to between These have and life course model which and is by the of and a of a which is by the exposure and as an This is to life course epidemiology but is more when exposures acting a time for when function in is associated with the number of social and behavioural that in different The original life course models a outcome, such as disease at a in as trajectories of functional phenotypes are increasingly being as This shift in focus has been driven in by our work which forward from review of function in the original life course These how the disease be development or or a of that time these were but recent have that of all of chronic disease in with adult function but the other development of function with an This highlighted two important (i) that consider such as or function and phenotypes such as as well as and (ii) that disease endpoints at one time in later life, the increasing This risk associations are only of for but The has been for in to functional has the of the study of disease and risk risk functional may have more the of disease with of as well as and can be for clinical to early of development an of functional before is is to a life course because an to the of the of functional these trajectories with social and time may all important and for further the of functional are time, of both trajectories of exposure and and their is from the birth cohort has been to model how in during may to trajectories between to using In this there is no with as trajectories when functional trajectories that may be for and physical is which be as the and which the both may each The to test may be for and such as the when are of to the challenges of the in these The research evidence on functional a life course is (see this for evidence of in cardiometabolic It has been and more to to the for and this may be to this has in important for to earlier life factors that have been to be associated with the of adult functional or is the to the developmental is associated with but this to with that the is of In early social are associated with in in mid an on of the of functional trajectories and their is the for longitudinal across the life which are the of from a study with a range, or the of a birth cohort by longitudinal from different with have been to on across To one longitudinal across different as was by using from cohort in the only cohort a more mid life in social this was to from different have been to with in a for and but have a from from different or using a we are that there are no or cohort which may be This can be when different birth with across the are with each of from birth from the in and and that the of being or in the three the of in the two at were in the three and the in or trajectories across these was to being in the more recent The earlier of and in more recent can be to disease and on health with of are but as well as the important when be life course ideas to ageing research has us and of our life course one and ageing with in of biological as to These can be biological of a or but are more often that of In each the is to that the has added in other outcomes, such as or and The of these may at different of the life course and may and as with any other of traits that and be from a life course perspective. We have that there is a of is by the term because are to two any of the health and of of the health study with including biological ageing from how or function to study their ageing is as a to or functional to and biological ageing three key longitudinal to in the of and of for the of We three of that all a with physical and at the and is defined as the to the physical and of the physiological of each and the of and the trajectories of these different aspects of their time, and their and has become a focus in life course epidemiology. and have a model for or life that on four key ageing (i) or (ii) or in pathways that both the and and in (iii) and and and We have these an integrated life course model of ageing (see Figure We have added ageing considered by the of development and from an and health which may be a of ageing but is important for social and of A physiological to the of and and as and early that may as for later disease and integrated life course model of ageing. Kuh D, Ben-Shlomo A Life Course Approach To Oxford, Oxford University Press, We phases of life and we the of or social The other are we to as the and the to the social and physical which in turn and with the This may physiological to or or of the growth growth in to to the with long-term or these are with that can in both and to and or to and a in ageing own epidemiological of physical and and of as these are to in epidemiological disease such as chronic or disease on these as well as and is important for function but in physical and and The that ageing is associated with across be or This may be as the of may across that one may to be more even physical is evidence for the a recent found only evidence and between The of ageing that this integrated life course model is to with but may differences in This model three (i) (ii) and (iii) to of across the life of growth in childhood, and in and or management of in (see commentary by Alwin18). This may for an with with a possible by the It is possible to this by the to be This with in The in in later life, this is a is to but more reflects a different such as on this may be for all and and longitudinal of the more to in the first of life, to and growth before the are to of and of The of both processes result in an early the and a later to different the a and the a from We there is in biological ageing and our models, to be of any to be to this in to challenges with The to and was to by as considered this to only in the phase between that of the adult and the We that this is limited and in fact this can life the and to may well and is to with as A life course approach is in the between the and or or different risk factors there an between these phases and clinical disease or functional is the of we have as with development of a have further to before any threshold and with disease at a later or may of In our original this was by the of function at the We to the to from exposures and how this with we now term as (see Figure Although such be there may or may be in to ageing. and this as type 2 chronic The of and and on a life course the of to a only on the of the and how was before this but the in and may or may be to an empirical that to developmental or is of to to the of in early In order to capture is to the This may be a natural as in the in on or a exposure such as the test to and to a standard studies have of other such as to in to or from an such as or A recent was a study of the of on longitudinal (i) but no in with and (ii) with and both and with is associated with often health evidence the of the physical in an in an of how functional trajectories clinical of have most to from an only health but and often for and of the often of mid life risk factors in and in mid life but at both are associated with with this reflects a by and or a of which is with in to the of to be In how an to result in which may be or by the a as the and model to In and by the biological of the life the of and This of to and be by a of psychological and social and factors acting on the and social and may facilitate with which in turn and or function of in the of The other of our of different is that this the and to which may added for for a risk of because is a of which of or is a of mechanisms that a limited have been for in in which a standard of and reflects epidemiological studies have been to capture on a because of the and in now this a our of the life course of such is these are is we from the that physical in can the of this was associated with a in This review has that the field of life course epidemiology has a on how different aspects of function the and across life, and on their risk and of this is with our colleagues in DOHaD, in ageing research (see commentary by and in (see commentary by and we our and be further This from and the of an ageing on health and social a on the for these research A developments which in turn for life course epidemiology and for epidemiology more We a or increasingly us to capture of and which on the of long-term are growing for and have been which and of This has added in that an in issue in any longitudinal study, and of as with from the A is to to capture the pathways and the between and their of health and social that may functional for how such as and the at that time, trajectories of in later Life course and longitudinal are being to the to this in the that biological pathways earlier exposures to later function be and functional and disease development IJE special on and review of life course epidemiology to the of these in these current and developments in life course epidemiology and in life course are since the there has been a in terms of the growing number of with increasingly rich and for life course this has to and the (see editorial by We a number of challenges for cohort studies in a time of Each cohort study its as well as its from with the of In the both of study are as are but in the for limited to cohort challenges in are for and to a of to in cohort and assess and life course of that the of longitudinal on more exposure and can be is to the growing of evidence on the on adult health, disease and ageing or and studies to the (see by and and and and increasingly a life course and evidence from cohort this is an early life on and both important in their own research and consider that the is Life course epidemiology has on studies the long-term of It may be that the from life course for the long-term of and psychosocial may be to that are an is important to that life course may be highly in both time and from to and comparisons are important and before of epidemiology one of our that the concepts life course epidemiology, were that the and in these concepts may our We have this but have been by the of in to this We that in this of and research with such as and we are in a to on these The from the to and with current and cohort us to the of and factors on development and of health and function as well as disease R.C. and D.K. are by the and was by the for Health The are of the and of the the or the of is a the of The of and the between the of and and The for in a between and of