Early life determinants of skeletal maturation, body composition and endocrine health in young adults (EPIPEAK): protocol for a nationwide birth cohort study

Abstract

Introduction

Early-life exposures, such as nutritional deficiencies, stress, smoking, toxins, medications, diseases, infections and inflammation may affect multiple physiological and metabolic systems in the offspring, including hormonal regulation, bone metabolism and mineralisation, and body composition. Moreover, the effect of these early-life exposures on later health may potentially be mediated through adverse neonatal epigenetic reprogramming of bone-related genes affecting health later in life, especially skeletal development and bone density. Thus, to advance this research further, the overall aim of the project is to investigate if (a) neonatal epigenetic and genetic signature; (b) maternal risk factors during preconception and pregnancy, such as medicine use, diseases, socioeconomic status, major life events, weight, growth and lifestyle; (c) risk factors at birth, such as instrumental delivery, mode of delivery, medicine use, injuries, diseases, weight, size for gestational age, ponderal index, gestational age; and (d) childhood risk factors, such as diseases, medicine use, major life events, weight, growth and lifestyle are associated with hormonal status, lipids, bone turnover markers, bone mineral density, fat mass and lean body mass at age 18–19 years.

Methods and analysis

Population-based, nationwide, cross-sectional clinical study with potential for longitudinal reassessment. Danish women and men aged 18–19 years old will be selected at random from the Danish National Population Registry and invited if they have available neonatal dried blood spot cards. A total of 2000 individuals will be enrolled. The study combines register data, and neonatal epigenetic and genetic analyses from stored blood with clinical and survey data. Body composition will be measured using dual-energy X-ray absorptiometry. Adult blood and hair samples will be obtained to assess hormonal status, lipids and bone turnover markers. Height, weight, waist and hip circumference, and blood pressure will be measured. Questionnaires on well-being, sleep patterns, dietary and exercise habits, onset of puberty, use of cannabis, nicotine, alcohol and pain medication will be included. Information on medicine use, diseases, socioeconomic status, major life events, weight, growth and lifestyle will be obtained from the national administrative and health registers at the time of conception and during pregnancy for the parents, as well as from the participants throughout their lifetime. Health registries include the Danish Medical Birth Register, the National Patient Register, the Danish National Prescription Register, the National Child Health Register and Statistics Denmark. Multivariate regression analyses will be performed.

Ethics and dissemination

This nationwide study has been approved by the Regional Committees on Health Research Ethics for Southern Denmark (S-20230105). The study participants will be enrolled in the study following their informed written consent. Results will be submitted for publication. The Strengthening the Reporting of Observational Studies in Epidemiology Statement guidelines will be used for reporting.

Trial registration number

NCT06509776.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• Data quality and completeness will be very good because the study combines data from a clinical examination at age 18–19 years with data from routinely collected neonatal biomaterial and national administrative health registers at an individual level in a large nationwide random sample of young adults and their parents.

• The Danish administrative and health registers are known to be of high quality and validity and comprise the entire Danish population, and the clinical examination will use proven and reproducible standardised methods.

• It is a limitation that questionnaires are subject to information bias, as the data are self-reported, but this is reduced by linkage to detailed register data.

• The demographics of the study participants will be compared with the background population to verify external validity; however, we expect that bias can arise as young adults with some demographics may be more likely to accept the invitation.

• For confounding variables that are not sufficiently granulated or for which we lack information, we will use proxy variables, for example. For sociodemographic factors, we will use, for example, educational level and parental age, and for behavioural factors, we will use, for example, parity and maternal smoking during pregnancy.

Introduction

From the second trimester of pregnancy and until early adulthood, bone is gradually developed and shaped, with longitudinal growth dominating the later stages of fetal life, infancy and childhood.¹ This is followed by a period of rapid bone mineral accrual occurring up to and during puberty, with bone mass accretion ultimately reaching a plateau (peak bone mass (PBM)) in young adult life, around age 18–19 years or possibly later. It is unknown if the amount of bone laid down during adolescence is pre-programmed as a chrono-switch or a mechanostat, where PBM reflects mechano-sensing, so that bone gain ends due to the skeleton reaching a ‘sufficient’ mechanical competence to meet the demands, rather than due to a certain age being reached. Thus, it is unclear, for example, if low bone mass due to anorexia nervosa, other illnesses or pharmacological treatments during adolescence can be recovered during the third decade of life or if a chronostat mechanism prevents further bone gain.

In recent years, it has become increasingly clear that intrauterine and early-life exposures have consequences for adult health. The textbook example of an epigenetic theory is the Barker ‘thrifty phenotype’ hypothesis, also known as ‘the study of developmental origins of health and disease’, where poor nutrition or other stressors during pregnancy and/or critical periods of growth may lead to changes in physiology and metabolism. Changes that may be beneficial short term, since they ensure the survival of the fetus, may lead to long-term adverse conditions, and the most well-known illustration of this is the linkage of low birth weight to metabolic syndrome.² A less widely known example involves skeletal health and links low birth weight to hip fracture risk in the elderly.³ Focusing on epigenetics rather than genetics or chronic suboptimal nutrition during pregnancy is crucial for several reasons. First, unlike fixed genetic changes, epigenetics is dynamic, which may be relevant in understanding how intrauterine environment shapes health outcomes later in life. Second, studies of epigenetics may provide mechanistic explanations for how stressors can lead to health issues, for example, changes in DNA methylation or histone modification in relation to bone development or skeletal fragility. Finally, understanding epigenetic influences may open opportunities for targeted interventions. Evidence from nature, animal experiments and increasingly from human observations has suggested that epigenetic mechanisms may be critical for the early life influences on not only the trajectory of bone,^{4 5} but also the trajectory of adverse weight development, and thus later risk of osteoporosis, overweight and obesity.⁶,¹³ It is strongly suggested that epigenetic variation at birth can result from differences in maternal health, lifestyle, nutrition, smoking and maternal medication usage and result in longer-term changes—reversible or irreversible—in gene expression and metabolism.^{6 14} While epigenetic modifications could be transient, differences in DNA methylation of genes known to be associated with bone mass have been demonstrated between bone tissue from healthy postmenopausal women and women with osteoporosis, suggesting persistence through adult life.¹⁵ Moreover, though the consequences of severe hormonal excess or deficits in infancy are well known, not least for the thyroid axis,¹⁶ the effects of more subtle and/or transient differences on subsequent adult health are mostly unknown.

Research has been challenging due to the scarcity of data on early life, other than birth weight records, for diseases that become manifest some decades into adult life. Our systematic review that summarised the literature on prenatal environmental exposures and bone health among young adults showed that diet (high concentrations of maternal vitamin D; low fat intake; and high intakes of calcium, phosphorus and magnesium) during pregnancy, as well as young maternal age, is associated with higher bone density among young people aged 16–30 years.¹⁷ We found no studies of other common prenatal factors such as hyperemesis gravidarum, medications used during pregnancy and development of gestational diabetes.

In summary, factors driving persistent hormonal changes, body composition and bone mass in young adults are very poorly understood. Thus, to advance this research, studies that make use of material collected at birth and life course measures of bone density, growth and endocrine health are needed.

The overall aim of the project is to investigate if (a) neonatal epigenetic and genetic signature; (b) maternal risk factors during preconception and pregnancy, such as medicine use, diseases, socioeconomic status, major life events, weight, growth and lifestyle; (c) risk factors at birth, such as instrumental delivery, mode of delivery, medicine use, for example, during delivery or to the new born, injuries, diseases, weight, size for gestational age, ponderal index, gestational age; and (d) childhood risk factors, such as diseases, medicine use, major life events, weight, growth and lifestyle are associated with young adult hormonal status, lipids, bone turnover markers, bone mineral density (BMD), fat mass and lean body mass at age 18–19 years (figure 1).

Figure 1. Study diagram.

The research questions are as follows:

1. Bone health in young adulthood: is whole body and regional BMD and body composition influenced by epigenetic profiles at birth?

2. Endocrine maturation in young adulthood: do maternal medication use and health issues, such as medicine use, diseases, socioeconomic status, major life events, weight, growth and lifestyle prior to pregnancy or during pregnancy affect endocrine health in young adults and can we identify epigenetic changes driving this?

3. Obesity and cardiometabolic health in young adulthood: do suboptimal health conditions (eg, diseases, socioeconomic status, major life events, weight, growth and lifestyle) prior to pregnancy, during pregnancy, at birth and during childhood adversely influence the establishment of a healthy body composition and lipid status in young adulthood?

We hypothesise that exposure to suboptimal conditions and risk factors prior to pregnancy, during pregnancy and birth, and subsequent lifestyle, poor health and adverse socioeconomics in childhood will adversely influence risk markers of bone health, sarcopenia, obesity and cardiovascular disease. Specifically, endocrine epigenetic variation at birth may result from differences in maternal health, lifestyle, nutrition and medication usage and account for persistent epigenetic changes in gene expression and metabolism in the offspring.

Methods and analysis

Study design

This is a population-based, nationwide cohort study that combines prospectively collected registry data and archived biomaterial with a cross-sectional clinical examination with the potential for longitudinal reassessment.

Inclusion and exclusion criteria

Inclusion criteria are individuals born in 2006 and 2007 with available neonatal dried blood spot (DBS) cards and over the age of 18 years. Individuals who are pregnant or breastfeeding will be excluded.

The Danish Civil Registration System will be used to randomly select a sample of 60 000 individuals (about half of all births during the period) born in Denmark from 1 January 2006 to 31 December 2007 and alive at the start of the study, excluding those born or currently living in Faroe Islands or Greenland, and those who have emigrated or no longer have a permanent address in Denmark. Using the unique personal identification numbers of the random sample, the Statens Serum Institut will identify the individuals with available neonatal DBS cards containing sufficient material in the Danish Biological Specimen Bank for Neonatal Screening. Expectedly, about 82% will have sufficient material.¹⁸ Only individuals who consent to the use of their neonatal DBS cards and data from the national Danish administrative and health registers will be included.

Procedures

The study participants will be enrolled in the study following their informed written consent from January 2025 to December 2026. After concluding that there is available blood from the DBS cards, invitations will be sent through the secure digital mailbox (global.e-boks.com) in blocks until 2000 participants have been enrolled. If a participant does not reply, one reminder will be sent through the secure digital mailbox. To ensure equal sex distribution in the participants, we will monitor and adjust the recruitment dynamically based on real-time participant data.

On-site clinical examination is conducted at eight collaborating clinical research centres based at hospitals and health research facilities across Denmark (Copenhagen, Aarhus, Aalborg, Odense, Frederiksberg, Holbæk, Køge and Hvidovre). Should dual-energy X-ray absorptiometry (DXA) capacity become a limiting factor, we have the option to invite additional Danish hospitals to participate.

To track changes in bone mass, the participants will be invited to an additional clinical assessment approximately 2 years after the initial assessment, pending research funding.

Databases

Information on prenatal and childhood exposures will be obtained through linkage to national Danish administrative and health registers using the unique personal identification. Prenatally, this will include conditions leading to hospital contact for parents, maternal and paternal redemption of prescribed medication in the 3 months preceding conception, maternal smoking during pregnancy, redemption of prescribed medication and hospital contacts during pregnancy. Immediate perinatal information in the form of mother and child height (length) and weight, placental weight, complications during delivery and neonatal health will also be included. Moreover, child health examinations in general practice at ages 5 weeks, 5 months, 1, 2, 3, 4 and 5 years will also be included. We will use information in the Danish Medical Birth Register,¹⁹ the National Patient Register,²⁰ the Danish National Prescription Registry,²¹ the Danish National Child Health Register²² and Statistics Denmark (www.dst.dk).

Epigenetic profiles

Epigenetic profiles will be analysed using the cohort’s already collected neonatal blood samples. Neonatal DBS cards by heel prick have been collected for all newborns in Denmark since the initiation of routine screening for congenital disorders in 1981.²³ For those born in the years 2006 and 2007 the DBS cards were collected up to 7 days after birth. Our group has previously demonstrated the feasibility of using archived DBS cards for examining vitamin D status at birth in relation to various childhood and adulthood health outcomes,¹⁸²⁴,²⁹ including childhood fractures.¹⁸ Epigenetic analyses from the blood of the DBS cards are also feasible.³⁰,³² Epigenetic mechanisms include DNA methylation, histone modifications and non-coding RNAs. The reliability of DNA extraction and DNA methylation measurement in array technologies has been demonstrated previously.³³ We will employ an updated EPIC array and/or Blood Methylation Array for maximum efficiency and we will prioritise epigenetic changes in genes involved directly in endocrine signalling (hormones, receptors, co-receptors and downstream—eg, CASR, AP2S1, CYP27B1, ESR1, HGR, GHRH, GHSR, GNAS, HSD11B1, HSD11B2, IGF1, IGF1R, PTH, TG, THRA, THRB, TRH, TRHR, VDR and others) and those previously linked to low BMD and with methylation changes in osteoporosis (eg, MEPE, RANKL, WNT16, SOST, WIF1, DKK1).

We will also collect new DBS cards to enable the possibility of conducting an epigenetic profile at age 18–19 years, pending research funding.

Outcome assessment

List of variables is provided in online supplemental material 1.

Following local standard procedures, BMD of the spine and hip, and whole body will be examined using DXA (primary outcome). Thus, DXA will provide information about total and regional BMD. Whole-body DXA will also provide total body lean mass (kg) and fat mass (kg) measurements. Hologic or Lunar devices will be used. In academic institutes with high-resolution peripheral quantitative computed tomography (HRpQCT) equipment available, HRpQCT of the distal radius and tibia will also be performed. This will provide additional structural information including cortical thickness, trabecular number, trabecular thickness and vBMD. For pragmatic reasons, local standard operating procedures at the eight collaborating centres will be followed for DXA and HRpQCT.

Weight will be measured to the nearest 0.1 kg on an electronic scale with participants wearing light clothing. Height will be measured to the nearest 0.1 cm using a stadiometer. Both measurements will be carried out barefoot. The waist and hip circumference will be measured using a measuring tape to the nearest 0.5 cm using standardised methods from WHO. At least two measurements will be performed and a third will be made if the two differ by more than 1 cm. After resting in a supine position with arms resting alongside the body, palms up and legs uncrossed for 5 min blood pressure of the participants will be measured. Blood pressure is measured automatically, with at least five readings taken at 2-minute intervals until three stable values are recorded.

Up to 24 mL blood is drawn for analyses of endocrine status and biomarkers of bone metabolism (eg, estradiol, sex hormone binding globulin, follicle-stimulating hormone, luteinising hormone, insulin-like growth factor 1, insulin-like growth factor binding protein 3, insulin, 25-hydroxyvitamin D3, procollagen type I N-terminal propeptide, bone-specific alkaline phosphatase, thyroid stimulating hormone, free thyroxine, free triiodothyronine, parathyroid hormone, testosterone, dehydroepiandrosterone sulphate, adiponectin, leptin, haemoglobin A1c, lipids and glucose). For pragmatic reasons, the participants are not fasting when the blood samples are obtained. The same certified laboratory will analyse the blood samples according to standard procedures.

Hair samples will be collected for assessment of cortisol concentrations over the past 2 months (equivalent to 2 cm hair). The hair samples will be cut from the posterior apex as close to the scalp as possible (minimum 20 mg/2 cm hair length per participant). The samples will be stored in aluminium foil until the time of analysis, and the proximal scalp end of the hair sample will be carefully marked. Cortisol extracted from hair will be analysed using radioimmunoassay.

All clinical examinations will be performed by trained health professionals, who are blinded to the exposures. All data will be anonymised by a coding procedure and transferred to the Danish National Archives (www.rigsarkivet.dk) on completion of the research project, having allowed a reasonable time period for publication of the findings and responding to subsequent scientific queries before archival.

Questionnaires

Current demographic and lifestyle-related information will be collected by electronic questionnaire (in Danish), covering habitual physical activity, habitual sleep patterns, dietary habits (modified food frequency questionnaire), alcohol consumption, cannabis use, tobacco/nicotine product use and non-prescription analgesic use. In addition, the participants will provide information on timing of puberty, quality of life (WHO 5) and experience of tinnitus. All questionnaires will be sent through the secure digital mailbox, and the responses are automatically registered and secured in the database (REDCap electronic data capture tools).^{34 35} A list of variables is provided in online supplemental material 2.

Sample size calculation

With a study population of 2000, the power (given α=0.05) available to detect a 0.2 SD effect on a continuous outcome such as BMD for a risk factor with a population prevalence of 20% is 90%. Examples of risk factors with a prevalence of more than 20%, including treatment with prescribed medicine during pregnancy³⁶ or childhood, either as polypharmacy or as redemption of a single type of medication, for example, melatonin³⁷ or antibiotics. With the inclusion of 1500 subjects, the corresponding study power is 90% for detection of an effect size of 0.24 SD or 80% for detection of an effect size of 0.2 SD. The study needs this resolving power to be able to address multiple contributing factors and to include factors in the model that may have a population prevalence below 20%. We will focus on more common risk factors, as demonstrating an effect of these would have a greater impact if they could be modified at the population level.

Data analysis plan

Multivariate regression analyses will be performed (statistical analysis plan is under development). Variables will be checked for normal distribution, and non-normally distributed variables will be standardised to achieve normally distributed variables with a mean of 0 and an SD of 1. Diagnostic plots will be used to check for heteroscedasticity. To account for potential non-linear time trends, we will use cubic splines. The analyses will be adjusted for confounders (identified using directed acyclic graphs) and stratified by sex using the sex allocated in the unique personal identification numbers (if feasible, given the power). Where appropriate, 95% CI will be reported and a p value of <0.05 will be considered statistically significant.

The statistical guidance on DNA methylation research proposed by Mansell et al³⁸ will be considered for the epigenetic data.³⁸

Patient and public involvement

None.

Ethics and dissemination

Written informed consent will be obtained from participants before participation in the study. The study has been approved by the Regional Committees on Health Research Ethics for Southern Denmark (S-20230105), including that the study participants will receive Kr300~€40 after completing the study visit as compensation for transport and lost earnings.

All data collection and management of the project follows the principles of the Declaration of Helsinki and the EU’s 2018 Personal Data Regulation, and all stakeholders in the project are responsible for compliance with Danish legislation on data protection. The parties have entered a contractual obligation of joint data responsibility under Danish law (details available on request).

The results will be published in relevant international peer-reviewed journals of high impact and presented at national and international conferences and to the national and international press. Both positive, negative and inconclusive results will be published. All information will be delivered in an anonymised form. Funding sources will appear in all disclosures. The Strengthening the Reporting of Observational Studies in Epidemiology statement guidelines will be used for reporting.³⁹ The results will also be presented to participants and relevant stakeholders, including the Danish health authorities and patient organisations; in addition, nationwide dissemination of the results will be achieved through (a) contact to relevant health, administrative and political bodies; (b) contact to relevant news outlets; and (c) reports to funding bodies.

A summary of publications and findings will also be available to study participants and to the public on the study website (www.epipeakstudy.dk), which is also used to provide background information to the participants, contact information for the research group and for booking participant information meetings.

Discussion

Our current knowledge about the importance of exposures before and during pregnancy and early childhood for adult health remains limited because such data, unlike, for example, birth weight and maternal age, are not easy to obtain. In this collaborative project across Denmark, we will establish a new national cohort of 2000 young adults randomly selected from the population register. Measurements of hormonal status, lipids, bone turnover markers, BMD, fat mass and lean body mass at age 18–19 years will be linked to already prospectively collected data from the Danish administrative and health registers regarding previous medical and environmental exposures early in life, as well as epigenetic information obtained from analyses of DBS cards taken and archived at birth. In this way, we can examine the role prenatal and early postnatal environment plays for the endocrine profile, bone and body composition in adolescence (around the time of peak growth) with the aim of preventing or delaying the later development of osteoporosis, sarcopenia, obesity and cardiovascular disease. Our group has shown that cohort effects rather than period effects explain the shifts in fracture epidemiology in the elderly population in both Denmark and Sweden,⁴⁰ and provided proof of concept for enriching epidemiology studies with bio-banked neonatal DBS cards.¹⁸²⁴,²⁹

The project is expected to provide knowledge of potential long-term health benefits to young adults at risk of having their health and life expectancy affected over the coming decades by osteoporotic fractures, sarcopenia, obesity and cardiovascular disease, debilitating non-communicable diseases that globally affect millions of people each year,⁴¹ and which are commonly assessed to be consequences of a poor lifestyle in adult life. However, where early-life exposures are found to explain the risk, primary prevention in the strictest sense is confined to the next generation and secondary prevention can be targeted to the younger segments of our population today. Additionally, understanding the persistent endocrine effects of exposures in utero and during early postnatal life will improve our ability to explain ‘idiopathic’ disturbances, and potentially whole genome sequencing will become a standard tool in diagnosing severe endocrine and metabolic diseases in genomic medicine. Obtaining a higher PBM can be shown mathematically to have a substantial impact on the long-term risk of osteoporosis and osteoporotic fractures. Hence, previous research has demonstrated that a delay of the onset of osteoporosis by up to 13 years can be expected with a mere increase of 10% in PBM.⁴² Importantly, the findings from the present study will inform not only determinants of health in young adults, which could be based on a range of exposures, including perinatal epigenetic profile, but may also identify novel therapeutic targets early in life to reduce the burden of lifestyle diseases.

Although the study fills a recognised and clinically relevant knowledge gap, by bringing together an in-depth examination of skeletal and endocrine health in young adults and adding already prospectively collected early life indicators and neonatal DNA material, there are some limitations to the study. It is a limitation that questionnaires are always subject to information bias, as the information is self-reported. However, in the present study, this is reduced by linkage to similar information from Danish detailed register data, such as educational level. To verify external validity, we will furthermore compare the demographics of the study participants with the background population; however, we expect that bias can arise as young adults with specific demographics may be more likely to accept the invitation to participate in the study.

Most noticeably, the study is limited by the information available in the Danish health registries. This is especially relevant in relation to residual confounding from variables that are not sufficiently granulated or from which we lack information, such as maternal food consumption or vitamin supplementation use during fetal life of our participants. We will use proxy variables, for example, for sociodemographic factors, for example, educational level and parental age, and for behavioural factors, we will use, for example, parity, maternal smoking during pregnancy, hospital contacts with alcohol-related diagnoses or redeemed prescriptions for alcohol dependence treatment. As proxy measures of stress, we will use registry information on major life events and bereavement, such as severe disease(s), accidents, trauma, convictions, unemployment, divorce or deaths of close relatives.