Search tips
Search criteria 


Logo of wtpaEurope PMCEurope PMC Funders GroupSubmit a Manuscript
Int J Epidemiol. Author manuscript; available in PMC 2010 August 1.
Published in final edited form as:
PMCID: PMC2818638

Cohort Profile

The Hertfordshire Ageing Study (HAS)
HE Syddall, MSc, Medical Statistician, SJ Simmonds, Bsc, Research Assistant, HJ Martin, PhD, Research Assistant, Clare Watson, MBBS, DRCOG, Research Doctor, EM Dennison, PhD FRCP, Reader in Rheumatology, C Cooper, DM FRCP FMedSci, Professor of Rheumatology and Director, A Aihie Sayer, PhD FRCP, Professor of Geriatric Medicine, and Hertfordshire Cohort Study Group

How did the study come about?

We have previously described the discovery of a unique set of infant records collected in Hertfordshire between 1911 and 19481. The historical data for births between 1911-30 were linked to the NHS central register and showed that small size at birth and during infancy was associated with increased risk of cardiovascular mortality in men and women.2;3 These were the first studies based on individual data to suggest that adverse environmental influences acting in utero and during infancy might increase the risk of cardiovascular disease mortality in later life. Detailed physiological investigations of surviving men and women born in Hertfordshire between 1920-30 and still living there in the early 1990’s were subsequently conducted and allowed investigation of morbidity. Studies based on men and women resident in East Hertfordshire demonstrated that small size at birth and during infancy was associated with increased risk of developing coronary heart disease and type II diabetes mellitus4;5, the metabolic syndrome and insulin resistance6 and osteoporosis7. During the same time period (1994-5), 717 men and women resident in North Hertfordshire attended clinics where a wide range of markers of ageing were characterised. These clinics comprised the first follow-up of what will herein be referred to as the Hertfordshire Ageing Study (HAS). This study was the first to demonstrate that size in early life was associated with markers of ageing in older people8.

In 2004, mortality studies were extended to include the whole 1911-39 birth cohort9 and a second follow-up of the HAS cohort was carried out approximately 10 years after the first one. Fieldwork for the second HAS follow-up was conducted at the same time as fieldwork for the larger Hertfordshire Cohort Study1 (HCS) which comprised 3,000 younger men and women who were born in Hertfordshire between 1931-39 and who still lived in the county. The HCS and HAS birth cohorts are completely distinct (with birth years ranging 1931-39 for HCS and 1920-30 for HAS) and have different, although related, research objectives. The principal objective of the Hertfordshire Cohort Study is to evaluate the interactions between the genome; the intra-uterine and early postnatal environment; and adult diet and lifestyle in the aetiology of chronic diseases in later life (cardiovascular disease, type II diabetes mellitus and obesity; osteoporosis, osteoarthritis and sarcopenia). The principal objective of the Hertfordshire Ageing Study (as described below) is to examine life course influences on healthy ageing, with sarcopenia, frailty, falls, physical performance and cognitive function as the principal outcomes of interest. The rest of this article describes the Hertfordshire Ageing Study in detail.

Principal investigators and funding

The Hertfordshire Ageing Study was established by Prof. Avan Aihie Sayer (principal investigator) and members of the Hertfordshire Cohort Study Group based at the Medical Research Council Epidemiology Resource Centre (MRC ERC), University of Southampton. The Medical Research Council and the University of Southampton were the principal sources of funding.

What does the Hertfordshire Ageing Study cover?

The principal objective of the Hertfordshire Ageing Study is to examine life course influences on healthy ageing. The outcomes of interest are sarcopenia, frailty, falls, physical performance and cognitive function. The main potential determinants of healthy ageing under investigation are intrauterine and postnatal size and growth, genetic characteristics, and adult physical activity, diet, lifestyle and socio-economic status. Details of the data collection are provided later in this paper.

Who is in the sample?

The recruitment of participants to the Hertfordshire Ageing Study is shown in Figure 1. The ledgers contained records for 6 803 live singleton births in North Hertfordshire between 1920 and 1930. The details of these boys and girls were sent for tracing by the National Health Service Central Register (NHSCR) in Southport and 3 477 (51.1%) were traced and flagged for ongoing follow-up, including notification of deaths. In 1994, 2 621 (75.4%) men and women were identified as still alive in the UK, of whom 1 428 (54.5%) were resident in North Hertfordshire; these men and women comprised the target population for the first HAS follow-up. 824 (57.7%) men and women agreed to a home interview with a trained research nurse and 717 (87.0%) subsequently attended a clinic for detailed investigations. Ethical approval for HAS was obtained from the Bedfordshire and Hertfordshire Local Research Ethics Committee and all subjects have given written informed consent.

Figure 1
Recruitment of Hertfordshire Ageing Study participants

Response bias between the HAS target population and participation in the first HAS follow-up

Birth weight and weight at one year of age were similar in men and women in the HAS target population but who did not participate in the first follow-up home interview [mean and standard deviation (s.d.) birth weight and weight at one year: 3.5kg(0.6), 10.2kg(1.1) in 318 men and 3.3kg(0.5) and 9.6kg(1.0) in 286 women], and among those who were in the target population and did participate in the home interview [mean (s.d.) birth weight and weight at one year: 3.5kg(0.5), 10.2kg(1.2) in 459 men and 3.4kg(0.5) and 9.7kg(1.0) in 365 women]. In addition, weights at birth and one year of age were comparable between men and women who only participated in the first follow-up home interview, or who participated in the home interview and clinic (data not shown).

To further characterise response bias in HAS a short postal questionnaire was mailed to the 604 men and women in the HAS target population who declined the invitation to participate in the study. Two hundred and eighty six (47%) of these “non-responders” returned the questionnaire and their characteristics are contrasted with those of the 717 HAS follow-up 1 clinic participants in Table 1. Male “non-responders” were: more likely to have memory problems; less likely to be of a high social class; and were more likely to state that they felt average rather than young for their age, than participants in the HAS follow-up 1 clinic. Female “non-responders” were: more likely to have had a heart attack; less likely to have ever smoked; and were more likely to state that they felt old or average rather than young for their age, than participants in the HAS follow-up 1 clinic.

Table 1
Characteristics of HAS follow-up 1 non-responders versus clinic participants

How often have they been followed up?

Figure 1 shows that the first HAS follow-up was conducted in 1994-5 when the participants ranged in age from 63 to 73 years (mean 67). The cohort are flagged with NHSCR for ongoing notification of deaths, but their second formal follow-up was conducted in 2003-5 after a median time of 9 years between follow-ups (range 8.2 to 10.0 years) when participating members of the cohort were aged 72 to 83 years (mean 76).

What is attrition like?

Figure 1 shows that 359 (50.0%) of the men and women who participated in the first follow-up also took part in the home interview for the second one. 122 (17.0%) men and women who participated in the first follow-up had died before the time of the second one. In addition, 236 (32.9%) men and women were lost to follow-up owing to failure to trace them across the intervening period (n=55), or because the individual was traced but declined the invitation to take part in the second follow-up (n=181).

Response bias between the first and second HAS follow-ups

The ageing, medical, lifestyle and socio-economic characteristics of the cohort at the first follow-up were compared according to subsequent mortality status between follow-ups. Unsurprisingly, among men, mortality in the intervening period was associated with: poorer cognition (AH4 and Mill Hill); poorer grip strength; greater likelihood of having no teeth; and a higher prevalence of smoking, memory problems, fracture since the age of 50 years and slow walking speed at the first follow-up (data not shown). Among women, mortality was associated with: somewhat poorer cognition on the AH4 and Mill Hill tests; a higher hearing threshold; and a poorer assessment of how the woman felt for her age at the first follow-up (data not shown).

We also examined response bias by contrasting the follow-up 1 characteristics of surviving men and women at the time of recruitment for follow-up 2 (n=595) according to whether (n=359) or not (n=236) they agreed to participate in follow-up 2. Among surviving men, higher social class, better cognition (AH4 and Mill Hill), and a greater number of teeth at follow-up 1 were associated with participation in follow-up 2 (data not shown). Among surviving women, the only characteristic from follow-up 1 which was associated with agreement to participate in follow-up 2 was better cognition (data not shown).

Comparison of participants in the second HAS follow-up and the 2005 Health Survey for England

We have described attrition and response bias at all stages of the HAS study and, as in any epidemiological study, the representativeness of the participants in the second HAS follow-up is open to question, not least because we have unsurprisingly demonstrated the presence of a healthy participant/survivor bias in HAS. We have addressed this by comparing the ageing characteristics of participants in the second HAS follow-up with those in the nationally representative Health Survey for England (HSE).

We accessed the 2005 HSE dataset10 ( We restricted the HSE dataset to match the age range of participants in the second HAS follow-up and focused attention on ageing markers that were assessed using comparable questionnaire items and measurement protocols in the two studies (namely grip strength, falls in the past year, prescribed medications and chair rises time). The ageing characteristics of the second HAS follow-up and HSE study populations were compared using t, Mann-Whitney and chi-squared tests. Analyses were carried out using Stata 1011.

Overall, the ageing characteristics of the participants in HAS and HSE were broadly comparable. Table 2 shows that the prevalence of falls in the past year, and number of prescribed medications, were not significantly different between participants in HAS and HSE. Grip strength was slightly higher among the HAS participants, but this could have been partly due to use of the Jamar dynamometer in HAS and the Smedley in HSE. Finally, chair rises times were slightly slower in HAS than HSE. Hence, overall there was no evidence to suggest that participants in HAS were consistently more or less aged than their counterparts in the HSE.

Table 2
Characteristics of participants in the Hertfordshire Ageing Study (HAS) second follow-up in comparison with those in the Health Survey for England (HSE)

What has been measured?

Table 3 outlines the extensive range of ageing, medical, lifestyle and socio-economic data available in the HAS cohort study. The characteristics of the participants in the follow-up 1 home interview (n=824) and clinic (n=717) have been described previously8.

Table 3
Data available from the Hertfordshire Ageing Study baseline and follow-ups

What has it found? Key findings and publications

The first HAS follow-up showed that lower birth weight was related to sarcopenia8 in later life, and lower weight at one year of age was related to sarcopenia, lens opacity, skin thickness and hearing threshold8. This phase of the study also demonstrated that different systems of the same individual do not age together, suggesting that further research was, and still is, needed to identify the different processes underlying ageing12. Further work explored whether chronological age or grip strength might function as useful single markers of frailty13 and concluded that, within the relatively narrow age range studied, grip strength was associated with more markers of frailty than chronological age; raising the possibility that grip strength might serve as a useful marker of frailty in clinical practice. The first HAS follow-up included a detailed eye examination (visual acuity using LogMAR chart; intra-ocular pressure using tonometry and nuclear lens opacity using LOCSIII grading and retinal photography). These data were used to demonstrate that reduced growth in the first year of life was associated with age-related nuclear lens opacities14 and retinal vasculature15 but no associations were identified between early size and growth and intra-ocular pressure16 or visual acuity17. Finally, analysis of HAS follow-up 1 data showed that polymorphism of the IGF2 gene and birth weight have independent effects on adult grip strength in men18.

Preparation of the HAS follow-up 2 dataset was completed in 2007 and analyses are currently underway. The principal purpose of ongoing research using these data is to examine life course influences on healthy ageing, with a particular focus on sarcopenia, frailty, physical performance, cognitive function and falls as the outcomes of interest, and intrauterine and postnatal size and growth, genetic characteristics, and adult physical activity, diet, lifestyle and socio-economic status as potential determinants of healthy ageing.

What are the main strengths and weaknesses?


Firstly, a detailed ageing, medical, lifestyle and socio-economic characterisation of HAS participants is available, including data for two follow-ups in later life. HAS therefore provides the opportunity to explore lifecourse determinants of both cross-sectional ageing markers and longitudinal rates of ageing. Secondly, sera are stored for future measurements and DNA has been extracted. Thirdly, prospectively collected data on birth weight, weight at one, and infant feeding and illnesses, in a cohort who are now aged, are a unique resource for examining lifecourse influences on healthy ageing. Fourthly, follow-up for mortality is ongoing and follow-up for incident clinical events is feasible. Fifth, survivor and responder biases mean that the cohort are no longer representative of the wider HAS target population, but the participants do nonetheless constitute a group of men and women who are ageing relatively successfully and in whom the determinants of healthy ageing can be studied. Lastly, all measurements were made by a trained research team working to strict study protocols. Data entry, record keeping, computer processing and statistical analyses have been carried out to an exceptionally high standard and an experienced multi-disciplinary research team has ensured preparation of a high quality research database. Overall, the Hertfordshire Ageing Study constitutes a high quality resource with which to explore lifecourse determinants of healthy ageing.


Firstly, HAS participants are local to Hertfordshire and losses have occurred at several stages of follow-up (Figure 1). However, we have previously shown that mortality patterns in the wider Hertfordshire cohort are broadly similar to England and Wales as a whole9 and, where comparison was possible, the ageing characteristics of HAS participants are broadly similar to those in the nationally representative Health Survey for England (Table 2). On this basis, we suggest that it is not unreasonable to generalise results from HAS to the wider population of older men and women in England. Secondly, “healthy” survivor and responder biases were apparent in HAS; this was unsurprising 19;20. Such response bias would only be a major concern if the relationships between potential lifecourse determinants and ageing outcomes differed systematically in men and women who participated fully in HAS and those who did not; this seems unlikely. Thirdly, the cohort is fairly small and the number of men and women available for future follow-ups will inevitably decrease over time. However, the HAS database is a rich research resource as it stands and will inform studies of ageing in younger birth cohorts such as the Hertfordshire Cohort Study1 and the National Survey of Health and Development (the 1946 birth cohort)21. Fourthly, birth weight is only a proxy for adaptations that a fetus may make to its body’s structure, physiology and metabolism in response to undernutrition in utero. Finally, with the benefit of hindsight it would have been desirable to include a broader panel of items that were measured at both follow-ups, in particular to have asked about falls, SF-36 self-assessed health, physical activity and performance at the first HAS follow-up. However, research themes develop and are in themselves informed by previous studies; hopefully our experience will help future researchers to decide which measurements to include in their studies.


We thank the men and women who have participated in the Hertfordshire Ageing Study, the Hertfordshire General Practitioners, and the nurses and doctors who conducted the home interviews and clinics at both follow-ups.

We acknowledge: the National Centre for Social Research and University College London (Department of Epidemiology and Public Health) as the principal investigators and original creators of the HSE 2005 dataset, and the depositors of this dataset with the UK Data Archive, University of Essex, Colchester. The Information Centre for Health and Social care was the original source of funding for the HSE 2005. These organisations bear no responsibility for our analysis and interpretation of the HSE 2005 data.


Data access and further information

The Hertfordshire Cohort Study Group has a long and successful history of conducting collaborative research5;14;22-24 and we welcome specific and detailed proposals for new collaborations. Initial enquiries should be made to Prof. Avan Aihie Sayer (igator, e-mail: Further information about the research work conducted by the MRC ERC, including the Hertfordshire Ageing Study and the Hertfordshire Cohort Study, may be found on the MRC ERC website ( and an information booklet is available on request.


1. Syddall HE, Sayer AA, Dennison EM, Martin HJ, Barker DJ, Cooper C. Cohort Profile: The Hertfordshire Cohort Study. Int J Epidemiol. 2005;34:1234–42. [PubMed]
2. Barker DJP, Osmond C, Winter PD, Margetts BM, Simmonds SJ. Weight in infancy and death from ischaemic heart disease. Lancet. 1989;2:577–80. [PubMed]
3. Osmond C, Barker DJP, Winter PD, Fall CHD, Simmonds SJ. Early growth and death from cardiovascular disease in women. BMJ. 1993;307:1519–24. [PMC free article] [PubMed]
4. Fall CHD, Osmond C, Barker DJP, Clark PMS, Hales CN, Stirling Y, et al. Fetal and infant growth and cardiovascular risk factors in women. BMJ. 1995;310:428–32. [PMC free article] [PubMed]
5. Hales CN, Barker DJP, Clark PMS, Cox C, Fall C, Osmond C, et al. Fetal and infant growth and impaired glucose tolerance at age 64 years. BMJ. 1991;303:1019–22. [PMC free article] [PubMed]
6. Barker DJP, Hales CN, Fall CHD, Osmond C, Phipps K, Clark PMS. Type 2 (non-insulin-dependent) diabetes mellitus, hypertension and hyperlipidaemia (Syndrome X): Relation to reduced fetal growth. Diabetologia. 1993;36:62–7. [PubMed]
7. Cooper C, Fall C, Egger P, Hobbs R, Eastell R, Barker D. Growth in infancy and bone mass in later life. Ann Rheum Dis. 1997;56:17–21. [PMC free article] [PubMed]
8. Aihie Sayer A, Cooper C, Evans JR, Rauf A, Wormald R, Osmond C, et al. Are rates of aging determined in utero? Age Ageing. 1998;27:579–83. [PubMed]
9. Syddall HE, Sayer AA, Simmonds SJ, Osmond C, Cox V, Dennison EM, et al. Birth Weight, Infant Weight Gain, and Cause-specific Mortality: The Hertfordshire Cohort Study. Am.J Epidemiol. 2005;161:1074–80. [PubMed]
10. National Centre for Social Research and University College London. Department of Epidemiology and Public Health . Health Survey for England, 2005. Colchester, Essex: UK Data Archive; Jul, 2007. computer file. distributor. SN: 5675.
11. Statacorp . Stata Statistical Software Release 10. College Station, TX: StataCorp LP; 2007.
12. Aihie Sayer A, Osmond C, Briggs RSA, Cooper C. Do all systems age together? Gerontology. 1999;45:83–6. [PubMed]
13. Syddall H, Cooper C, Martin F, Briggs R, Aihie Sayer A. Is grip strength a useful single marker of frailty? Age Ageing. 2003;32:650–6. [PubMed]
14. Evans JR, Rauf A, Aihie Sayer A, Wormald RP, Cooper C. Age-related nuclear lens opacities are associated with reduced growth before 1 year of age. Invest Ophthalmol Vis Sci. 1998;39:1531–4. [PubMed]
15. Chapman N, Mohamudally A, Cerutti A, Stanton A, Sayer AA, Cooper C, et al. Retinal vascular network architecture in low-birth-weight men. J Hypertens. 1997;15:1449–53. [PubMed]
16. Foss AJE, Rauf A, Sayer AA, Evans JR, Cooper C, Wormald RPL, et al. Does raised intraocular pressure begin in utero? Br J Ophthalmol. 1998;82:1125–30. [PMC free article] [PubMed]
17. Hall N, Aihie Sayer A, Evans J, Rauf A, Cooper C, Wormald R. Visual acuity and its relationship to early growth, eye disease and aging in North Hertfordshire. Ophthal Epidemiol. 1998;5:1–8. [PubMed]
18. Sayer AA, Syddall H, O’Dell SD, Chen XH, Briggs PJ, Day IN, et al. Polymorphism of the IGF2 gene, birth weight and grip strength in adult men. Age Ageing. 2002;31:468–70. [PubMed]
19. Breslow NE, Day NE. Statistical Methods in Cancer Research. Volume II - The design and analysis of cohort studies. Lyon: World Health Organisation, International Agency for Research on Cancer; 1987. [PubMed]
20. Coggon D, Rose G, Barker DJP. Epidemiology for the uninitiated. London: BMJ Publishing Group; 2003.
21. Wadsworth M, Kuh D, Richards M, Hardy R. Cohort Profile: The 1946 National Birth Cohort (MRC National Survey of Health and Development) Int.J.Epidemiol. 2006;35:49–54. [PubMed]
22. Dennison EM, Arden NK, Keen RW, Syddall H, Day IN, Spector TD, et al. Birthweight, vitamin D receptor genotype and the programming of osteoporosis. Paediatr Perinat Epidemiol. 2001;15:211–9. [PubMed]
23. Fawcett KA, Wareham NJ, Luan J, Syddall H, Cooper C, O’Rahilly S, et al. PARL Leu262Val is not associated with fasting insulin levels in UK populations. Diabetologia. 2006;49:2649–52. [PMC free article] [PubMed]
24. Day IN, Chen XH, Gaunt TR, King TH, Voropanov A, Ye S, et al. Late life metabolic syndrome, early growth, and common polymorphism in the growth hormone and placental lactogen gene cluster. J Clin Endocrinol Metab. 2004;89:5569–76. [PubMed]