a life course approach to reproductive health theory and methods

Mishra∗, Rachel Cooper, Diana Kuh MRC Unit for Lifelong Health and Ageing, Department of Epidemiology and Public Health, University College and Royal Free Medical School, 33 Bedford Plac

Contents lists available atScienceDirect

Maturitas

j o u r n a l h o m e p a g e :w w w e l s e v i e r c o m / l o c a t e / m a t u r i t a s

Review

A life course approach to reproductive health: Theory and methods

Gita D Mishra∗, Rachel Cooper, Diana Kuh

MRC Unit for Lifelong Health and Ageing, Department of Epidemiology and Public Health, University College and Royal Free Medical School, 33 Bedford Place,

London WC1B 5JU, United Kingdom

a r t i c l e i n f o

Article history:

Received 3 December 2009

Accepted 10 December 2009

Keywords:

Life course models

Women’s health

Critical periods

Sensitive periods

a b s t r a c t

Taking a life course approach to the study of reproductive health involves the investigation of factors across life and, also across generations, that influence the timing of menarche, fertility, pregnancy out-comes, gynaecological disorders, and age at menopause It also recognises the important influence of reproductive health on chronic disease risk in later life Published literature supports the use of an inte-grated life course approach to study reproductive health, which examines the whole life course, considers the continuity of reproductive health and the interrelationship between the different markers of this This

is in contrast to more traditional approaches that tend to focus only on contemporary risk factors and which consider each marker of reproductive health separately For instance, we found evidence linking early life factors such as growth, socioeconomic conditions, and parental divorce with ages at menarche and menopause, although the nature of the relationship differs We discuss the different theoretical mod-els that are used within life course epidemiology and which postulate pathways linking exposures across the life course to health outcomes, using examples of relevance to the study of reproductive health These highlight the importance of examining timing of exposures, such as during critical periods in early life, and the temporal order of exposures How life course frameworks of reproductive health can be devel-oped to help identify hypotheses to be tested is also demonstrated This approach has implications for the development of effective health policy that moves beyond identifying not only the type of intervention but also the most appropriate time across life to intervene

© 2009 Elsevier Ireland Ltd All rights reserved

Contents

1 Introduction 92

2 A life course approach to reproductive health 93

3 Life course epidemiology: theoretical models (with relevant examples) 93

3.1 Critical period versus sensitive periods 94

4 Life course framework to reproductive health 95

5 Methodological challenges encountered in studying the life course 95

6 Conclusions and future research 95

Contributors 96

Conflict of interest 96

Funding 96

Provenance 96

References 96

1 Introduction

Menarche heralds the beginning of a female’s reproductive life

with menopause signalling its end The timing of these two events

is often used to provide an indication of a woman’s reproductive

∗ Corresponding author Tel.: +44 (0)20 76705712.

E-mail address: g.mishra@nshd.mrc.ac.uk (G.D Mishra).

health Other events, such as childbirth and gynaecological surgery, and characteristics, such as menstrual regularity, gynaecological problems, pregnancy complications and offspring birthweight, are used to characterise reproductive health further [1] Each of the different components of a woman’s reproductive health can be, and often are, considered separately Many, however, are strongly associated with each other, as they act as indicators of the same underlying traits, for example endogenous hormone levels or sub-fertility, or, are on the same causal pathways

0378-5122/$ – see front matter © 2009 Elsevier Ireland Ltd All rights reserved.

In studying reproductive health it is first necessary to identify

which indicators to use and consider how they may be related to

each other If the aim is then to investigate factors which may

influ-ence these, it is important to consider that factors across life, from

conception onwards, have been shown to be associated with

repro-ductive health outcomes[1,2] This is exemplified in the findings

from a recent review of the literature on the early life predictors

of age at menarche and menopause[2] This found a range of

fac-tors in early life to be associated with the timing of menarche and

menopause For example, the main factors in early life found to be

associated with early menarche were: faster growth rates during

childhood[3,4]; higher childhood socioeconomic position[5–9];

family conflict and parental divorce[10–12]; presence of a

step-father[13]and; exposure to stressors such as war shortly before

menarche[14,15] This review also showed that while the

rela-tionships between many adult environmental factors and timing

of menopause have been investigated, only cigarette smoking and

nulliparity have consistently been related to an earlier menopause

early life factors and early menopause; these included not having

been breastfed, poor early growth, poor socioeconomic conditions,

lower childhood cognitive abilities, and parental divorce in

child-hood[22,23]

In addition to evidence of associations between factors across

life, from early life onwards, and reproductive health outcomes,

there is also growing evidence that a woman’s reproductive health

is linked to the reproductive characteristics of previous generations

[24] Taking timing of menarche and menopause as examples again,

there is evidence for a positive correlation between mother’s and

daughter’s age at menarche[24]and family and twin studies have

indicated that the genetic effect on timing of menopause is

con-siderable, with estimates of heritability ranging from 30% to 85%

cohort studies that a woman’s age at menopause is strongly

associ-ated with her mother’s reported age at menopause[19,23,25–29]

By taking all of this evidence into consideration we highlight the

importance of a life course perspective Of note however is that

while the field of life course epidemiology is considered to be

rel-atively new, having been coined as a term in the 1990s, there has

been a longstanding interest in the long-term effects of early life

exposures on adult disease risk[30]

Further justification for the use of a life course perspective

comes from the increasing recognition that, as well as being integral

to her overall health and wellbeing, a woman’s reproductive health

is a sentinel of chronic disease in later life[1,31–33] It has long been

acknowledged that earlier menarche, later menopause and other

reproductive characteristics, for example nulliparity, are

associ-ated with increased risk of some cancers, including breast[34,35]

and endometrial,[36–38]and lower risk of osteoporosis

Further-more, it has been shown that markers of reproductive health, such

as pregnancy complications including pre-eclampsia,

pregnancy-induced hypertension, gestational diabetes, preterm delivery, and

having a low birthweight baby, indicate future risk of

cardiovascu-lar disease[1,31,39]

In order to apply a life course perspective appropriately there is

a need to understand the theories and models that underpin this

approach

2 A life course approach to reproductive health

A life course approach examines how biological (including

genetics), behavioural and social factors throughout life, and across

generations[40], act independently, cumulatively and interactively

to influence health A life course approach to reproductive health

asks a range of questions that are relevant to the development of

health policy For example, does birthweight influence the age of menarche and of menopause? Does maternal stress during preg-nancy influence the development of polycystic ovary syndrome in female offspring? What is the influence of childhood growth on age at menopause and is this modified by adult body size? Could the link between reproductive health and other chronic diseases

be due to a common set of factors that affects them both and, if so, when and what is the best way to intervene? What is the impact of grandmother’s fertility rate on that of the granddaughter’s? Would preventing maternal gestational diabetes provide the most cost-effective means of reducing the risk of gestational diabetes in the offspring?

At the heart of this life course perspective lies a theoretical framework that assumes and tests for a temporal ordering of exposure variables and their inter-relationships with the outcome measure, both directly and through intermediary (mediating or modifying) variables[40,41]

3 Life course epidemiology: theoretical models (with relevant examples)

The underlying purpose of life course epidemiology is to build and test theoretical models that postulate pathways linking exposures across the life course to health outcomes[41] Given the wide range of exposures and the potential importance of their timing and duration, exposures may be acting to influence disease risk on a variety of different pathways Four broad hypothetical life course models that can operate for exposures acting at different points across the life course have been proposed (Fig 1)[40], these are the: critical period model; the critical period model with later life effect modifiers; accumulation of risk model; and chains of risk model each of which is described in more detail below

The critical period model pays attention to the timing of an expo-sure and assumes that the irreversible changes in body systems that occur during a particularly vulnerable phase of life, usually dur-ing early development, have implications for later health[30,41] The basic critical period model, also known as biological program-ming or as a latency model, underlies the fetal origins of adult disease hypothesis[41] An example of this, of relevance to the study of reproductive health, comes from the well-known random-ized double-blind prevention trial that showed intake of folic acid supplementation around the time of conception, but not later in pregnanacy, can prevent neural tube defects in the offspring[42] (Fig 1, model 1)

An expanded version of the critical period model includes the possibility of exposures in early life interacting with later life expo-sures, thereby either enhancing or decreasing the risk of chronic disease in later life; this model can be described as the critical period with later effect modifiers[40,41] For instance, the risk of cardiovascular disease in midlife of those with low birth weight (<800 g) could be due to the interactive effects of low birth weight

on the motor system together with a more inactive lifestyle (Fig 1, model 2)[43,44]

In contrast to the critical period models, the accumulation of risk model assumes that cumulative insults or exposures during the life course increase the risk of health later in life irrespective

of their timing This idea is similar to the notion of allostatic load [41] so that as the number, duration, and severity of exposures increase, there is cumulative damage to biological systems Risk exposures may cause long-term, gradual damage to health in sepa-rate and independent ways (accumulation model with independent and uncorrelated insults) For example, an individual may expe-rience a variety of unrelated exposures, such as having a higher growth rate, the presence of a stepfather, and being exposed to war,

Fig 1 Life course models with illustrative examples.

all of which cumulatively impact on age at menarche independently

of each other (Fig 1, model 3)[2]

While it is plausible that a range of different exposures will act

independently of each other and accumulate to influence disease

risk, it is more common for them to cluster together in socially

pat-terned ways The accumulation model with risk clustering takes

this into account For example, low childhood socioeconomic

posi-tion is associated with poorer growth, more family stress and

inadequate diet all of which may increase risk of earlier menopause

(Fig 1, model 4a)[30] Here, understanding the effects of childhood

socioeconomic position by identifying the specific aspects of the

early physical and psychosocial environments or possible

mech-anisms (such as nutrition, infection or stress) that are associated

with age at menopause may provide further etiological insights

The chain of risk model is a special version of the accumulation

model and refers to a sequence of linked events where one adverse

(or beneficial) exposure or experience tends to lead to another, and

so on For example, smoking may lead to early menopause, which in turn may increase the likelihood of developing heart disease Here, each exposure in a chain of risk may not only increase the risk of subsequent exposure in a probabilistic way, but may also have an independent additive effect on later health (Fig 1, model 4b) Alter-natively, it may be that only the final link in the chain leads to the adverse outcome, for example risky sexual behaviour is associated with increased risk of sexually transmitted infections which are associated with infertility[45](trigger effect) (Fig 1, model 4c) 3.1 Critical period versus sensitive periods

As discussed above, in life course epidemiology, different stages

of life and development are referred to as critical or sensitive peri-ods Making the distinction between these is important but often difficult to do To elaborate, a critical period is defined as a limited time window in which an exposure can have adverse or

protec-Fig 2 An example of a schematic representation of biological and social factors acting across the life course that may influence the timing of menarche and menopause.

tive effects on development and subsequent disease outcomes[41]

Outside this window, there is no excess disease risk associated with

the exposure

A sensitive period is a time period when an exposure has a

stronger effect on development and hence disease risk than it

would at other times For instance, in a recent review it was found

that when compared with nonsmokers, current smokers had a

greater reduction in age at menopause than former smokers[46]

This suggests that perimenopause is a sensitive period when the

effect of smoking may be more important than smoking history in

explaining an earlier onset of menopause

Critical periods may be more evident for chronic disease risk

associated with developmental mechanisms in biological

subsys-tems, whereas sensitive periods are likely to be more common in

behavioural development[41]

4 Life course framework to reproductive health

In most cases, once the specific life course associations to be

tested using empirical data and the theoretical models most likely

to underpin these have been identified it is important to place

these in the context of a wider life course framework In this

framework we are able to acknowledge, even if we cannot test all

the associations which are depicted, the potential role of factors

across life in explaining the specific associations under

investi-gation For example,Fig 2shows the pathways linking biological

and social factors across life to reproductive health Taking age at

menopause as our outcome of interest, paths (a), (b) and (d) depict

the effect of factors operating earlier in life, while paths (k) and

(l) represent factors operating closer to the time of menopause

(seeFig 2) If the associations of early life factors such as

breast-feeding, parental divorce and growth with timing of menopause

(path (d)) are to be tested, it is useful to consider not only the

dif-ferent pathways on which these factors may operate, but to take

into consideration the continuity of reproductive health across life

(paths (c), (g), and (j)) and other lifetime factors which may act as

confounders or mediators of these associations (paths (a) and (b))

(seeFig 2) Likewise, if we are to consider the association between

timing of menopause and cardiovascular disease risk in later life

it is important to make similar considerations, including taking

into account the effects of pre-existing cardiovascular risk prior to

menopause[47]

5 Methodological challenges encountered in studying the

life course

In recent years, there have been developments of new

statisti-cal approaches and epidemiologistatisti-cal thinking in relation to causal

models that can be usefully applied to etiological questions framed

within a life course paradigm[48–50] We demonstrated that the

critical period models, accumulation models, the effect modifica-tion models (such the critical period with later modifier) were each

a special case of the saturated model, which contains the effects of all possible combinations of the exposure measures across the life course[50] By comparing the model fit of a set of nested models – each corresponding to the accumulation, critical period, and effect modification hypotheses – to an all-inclusive (saturated) model, the model that best describes the data can be selected As the life course models described above are not mutually exclusive and may operate simultaneously, this standard approach to model building can provide further clues to the processes operating across the life course However, the development of standardised and acceptable methods of combining the different types of cumulative exposures are some of the many methodological challenges that still need addressing

It remains essential to have an understanding of the biolog-ical mechanisms underlying the effect of exposures on specific reproductive health outcomes upon which to base the statisti-cal modeling Regardless of which statististatisti-cal methods have been selected – including structural equations models, path analysis, G-estimation, and multi-level models – issues of measurement error, missing data, survival bias, and confounding factors are inevitable

in life course studies and hence results may still be biased and caution is required in their interpretation[41]

6 Conclusions and future research

In this article, we have highlighted how a life course frame-work encourages one to consider and test for a temporal ordering

of exposure variables across life and their inter-relationships with the outcome measure, both directly and through interme-diary variables We also suggest that the use of a life course approach may provide a better understanding of women’s repro-ductive health There is evidence that the study of reprorepro-ductive health would benefit from an integrated approach covering the whole life course, rather than an approach that restricts itself

to the study of contemporary risk factors, or which considers each reproductive outcome separately For instance, we found consistent evidence, though the nature of relationships differ, linking early life factors, such as growth, socioeconomic con-ditions, and parental divorce with both ages at menarche and menopause

As the value of the life course approach to health is increasingly recognised, a number of areas have emerged as important direc-tions for future research Relatively little work has been done, partly due to the lack of prospective data, to study the extent to which childhood nutrition underlies the relationship between growth and age at menarche or menopause Further work is also required to disentangle the associations between the different types of stres-sors in early life and pubertal timing The effects of factors across

the life course on markers of other reproductive health outcomes

such as polycystic ovary syndrome, fertility and gynaecological

disorders also need to be studied The full impact on

reproduc-tive health, of the rise in childhood obesity, and the delay in

the age at first birth in the industrialised world is also yet to be

realised

Family-based studies (intergenerational, sibling and twin

stud-ies) across the life course can be used to test specific causal

mechanisms and life course models as they can help

under-stand whether the timing of risk factors (critical and sensitive

periods) are important and establish the role of

heritabil-ity [51] By moving beyond associations, to understanding the

underlying mechanisms that determine reproductive health and

the relationships with chronic disease, we will strengthen our

ability to predict population outcomes and make timely

inter-ventions that can benefit current and future generations of

women

Contributors

GDM, RC and DK contributed to the design of the research, read,

edited, and approved the final manuscript

Conflict of interest

None declared

Funding

GM and DK are supported by the Medical Research Council RC

is funded by the New Dynamics of Ageing (RES-353-25-0001)

Provenance

Commissioned and externally peer reviewed

References

[1] Rich-Edwards J A life course approach to women’s reproductive health In: Kuh

D, Hardy R, editors A life course approach to women’s health Oxford: Oxford

University Press; 2002 p 23–43.

[2] Mishra GD, Tom SE, Cooper R, Kuh D Early life circumstances and their

impact on subsequent reproductive health: a review Women’s Health

2009;5(2):175–90.

[3] dos Santos Silva I, De Stavola BL, Mann V, Kuh D, Hardy R, Wadsworth ME.

Prenatal factors, childhood growth trajectories and age at menarche Int J

Epi-demiol 2002;31(2):405–12.

[4] Blell M, Pollard TM, Pearce MS Predictors of age at menarche in the newcastle

thousand families study J Biosoc Sci 2008;40(4):563–75.

[5] Attallah NL, Matta WM, El Mankoushi M Age at menarche of schoolgirls in

Khartoum Ann Hum Biol 1983;10(2):185–8.

[6] Billewicz WZ, Fellowes HM, Thomson AM Menarche in Newcastle upon Tyne

girls Ann Hum Biol 1981;8(4):313–20.

[7] Ulijaszek SJ, Evans E, Miller DS Age at menarche of European

Afro-Caribbean and Indo-Pakistani schoolgirls living in London Ann Hum Biol

1991;18(2):167–75.

[8] Rao S, Joshi S, Kanade A Height velocity, body fat and menarcheal age of Indian

girls Indian Pediatr 1998;35(7):619–28.

[9] Oduntan SO, Ayeni O, Kale OO The age of menarche in Nigerian girls Ann Hum

Biol 1976;3(3):269–74.

[10] Bogaert AF Age at puberty and father absence in a national probability sample.

J Adolesc 2005;28(4):541–6.

[11] Bogaert AF Menarche and father absence in a national probability sample J

Biosoc Sci 2008;40(4):623–36.

[12] Romans SE, Martin JM, Gendall K, Herbison GP Age of menarche: the role of

some psychosocial factors Psychol Med 2003;33(5):933–9.

[13] Mendle J, Turkheimer E, D’Onofrio BM, et al Family structure and age at

menar-che: a children-of-twins approach Dev Psychol 2006;42(3):533–42.

[14] Pesonen AK, Raikkonen K, Heinonen K, Kajantie E, Forsen T, Eriksson JG

Repro-ductive traits following a parent–child separation trauma during childhood:

a natural experiment during World War II Am J Hum Biol 2008;20(3):345–

51.

[15] Mul D, Oostdijk W, Drop SL Early puberty in adopted children Horm Res

[16] Bromberger JT, Matthews KA, Kuller LH, Wing RR, Meilahn EN, Plantinga P Prospective study of the determinants of age at menopause Am J Epidemiol 1997;145(2):124–33.

[17] Hardy R, Kuh D, Wadsworth M Smoking, body mass index, socioeconomic sta-tus and the menopausal transition in a British national cohort Int J Epidemiol 2000;29(5):845–51.

[18] van Asselt KM, Kok HS, Der Schouw YT, et al Current smoking at menopause rather than duration determines the onset of natural menopause Epidemiology 2004;15(5):634–9.

[19] Cramer DW, Xu H, Harlow BL Family history as a predictor of early menopause Fertil Steril 1995;64(4):740–5.

[20] Discigil G, Gemalmaz A, Tekin N, Basak O Profile of menopausal women in west Anatolian rural region sample Maturitas 2006.

[21] Hardy R, Kuh D Reproductive characteristics and the age at inception of the perimenopause in a British National Cohort Am J Epidemiol 1999;149(7): 612–20.

[22] Hardy R, Mishra G, Kuh D Life course risk factors for menopause and diseases in later life In: Keith L, editor Menopause, postmenopause and ageing London: Royal Society of Medicine Press Ltd.; 2005 p 11–9.

[23] Mishra G, Hardy R, Kuh D Are the effects of risk factors for timing of menopause modified by age? Results from a British birth cohort study Menopause 2007;14(4):717–24.

[24] Morton SMB, Rich Edwards J How family-based studies have added to under-standing the life course epidemiology of reproductive health In: Lawlor DA, Mishra GD, editors Family matters: designing, analysing and understanding family-based studies in life course epidemiology Oxford: Oxford University Press; 2009 p 295–315.

[25] Kok HS, van Asselt KM, van der Schouw YT, Peeters PH, Wijmenga C Genetic studies to identify genes underlying menopausal age Hum Reprod Update 2005;11(5):483–93.

[26] van Asselt KM, Kok HS, Pearson PL, et al Heritability of menopausal age in mothers and daughters Fertil Steril 2004;82(5):1348–51.

[27] de Bruin JP, Bovenhuis H, van Noord PA, et al The role of genetic factors in age

at natural menopause Hum Reprod 2001;16(9):2014–8.

[28] Murabito JM, Yang Q, Fox C, Wilson PW, Cupples LA Heritability of age at natural menopause in the Framingham Heart Study J Clin Endocrinol Metab 2005;90(6):3427–30.

[29] Torgerson DJ, Thomas RE, Reid DM Mothers and daughters menopausal ages:

is there a link? Eur J Obstet Gynecol Reprod Biol 1997;74(1):63–6.

[30] Kuh D, Ben-Shlomo Y A life course approach to chronic disease epidemiology Oxford University Press; 2004.

[31] Rich-Edwards JW Reproductive health as a sentinel of chronic disease in women Womens Health (Lond Engl) 2009;5(2):101–5.

[32] Hardy R, Mishra G, Kuh D In: Keith L, editor Life course risk factors for menopause and diseases in later life London: Royal Society of Medicine Press Ltd.; 2005 p 11–9.

[33] Barsom SH, Dillaway HE, Koch PB, Ostrowski ML, Mansfield PK The menstrual cycle and adolescent health Ann NY Acad Sci 2008;1135:52–7.

[34] Kelsey JL, Gammon MD, John EM Reproductive factors and breast cancer Epi-demiol Rev 1993;15(1):36–47.

[35] Hsieh CC, Trichopoulos D, Katsouyanni K, Yuasa S Age at menarche, age

at menopause, height and obesity as risk factors for breast cancer: associ-ations and interactions in an international case-control study Int J Cancer 1990;46(5):796–800.

[36] Wernli KJ, Ray RM, Gao DL, De Roos AJ, Checkoway H, Thomas DB Menstrual and reproductive factors in relation to risk of endometrial cancer in Chinese women Cancer Causes Control 2006;17(7):949–55.

[37] Xu WH, Xiang YB, Ruan ZX, et al Menstrual and reproductive factors and endometrial cancer risk: results from a population-based case-control study

in urban Shanghai Int J Cancer 2004;108(4):613–9.

[38] Pettersson B, Adami HO, Bergstrom R, Johansson ED Menstruation span—a time-limited risk factor for endometrial carcinoma Acta Obstet Gynecol Scand 1986;65(3):247–55.

[39] Sattar N, Greer IA Pregnancy complications and maternal cardiovascular risk: opportunities for intervention and screening? BMJ 2002;325(7356):157– 60.

[40] Kuh D, Ben Shlomo Y, Lynch J, Hallqvist J, Power C Life course epidemiology J Epidemiol Community Health 2003;57(10):778–83.

[41] 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.

[42] Prevention of neural tube defects: results of the Medical Research Council Vitamin Study MRC Vitamin Study Research Group Lancet 1991;338(8760): 131–7.

[43] Whitfield MF, Grunau RE Teenagers born at extremely low birth weight Pae-diatr Child Health 2006;11(5):275–7.

[44] Rogers M, Fay TB, Whitfield MF, Tomlinson J, Grunau RE Aerobic capac-ity, strength, flexibilcapac-ity, and activity level in unimpaired extremely low birth weight (<or=800 g) survivors at 17 years of age compared with term-born con-trol subjects Pediatrics 2005;116(1):e58–65.

[45] Gray RH, Wawer MJ, Serwadda D Sexually transmitted infections and health through the life course In: Kuh D, Hardy R, editors A life course approach to women’s health Oxford: Oxford University Press; 2002.

[46] Parente RC, Faerstein E, Celeste RK, Werneck GL The relationship between smoking and age at the menopause: a systematic review Maturitas 2008;61(4):287–98.

[47] Kok HS, van Asselt KM, van der Schouw YT, et al Heart disease risk

determines menopausal age rather than the reverse J Am Coll Cardiol

2006;47(10):1976–83.

[48] Pickles A, Maughan B, Wadsworth M Epidemiological methods in life course

research Oxford: Oxford University Press; 2007.

[49] De Stavola BL, Nitsch D, dos I SS, et al Statistical issues in life course

epidemi-ology Am J Epidemiol 2006;163(1):84–96.

[50] Mishra G, Nitsch D, Black S, De Stavola B, Kuh D, Hardy R A structured approach

to modelling the effects of binary exposure variables over the life course Int J Epidemiol 2009;38(2):528–37.

[51] Lawlor DA, Mishra GD Family matters: designing, analysing and understanding family based studies in life course epidemiology Oxford: Oxford University Press; 2009.