Proposal summaries
B918 - Is physical activity a protective factor for emotional wellbeing in young people - 03/12/2009
Is physical activity a protective factor for emotional wellbeing in young people?
Background
WHO(1) state that mental health problems are a worldwide public health concern and consequently, promotion of mental health is a key component of UK public health policy(2,3). Foresight(5) recently published a report which has drawn together evidence in this field and provides a framework for further developments.
Mental illness has a long-term impact on people and is a risk factor for mortality and morbidity(5), costing the UK economy £100 billion a year(6). In 2004, 10% of children and young people aged 5-16 years had a clinically diagnosed mental disorder, the majority of which were either emotional (anxiety or depression) or conduct disorders(7). However, many more young people are thought to suffer from poor emotional wellbeing. This was demonstrated in a report by UNICEF(8) in which the UK ranked bottom for children's wellbeing in comparison with North America and 18 European countries. A number of policies have been developed with a specific focus on the health and wellbeing of young people (9-11) and recently, NICE published guidance for promoting social and emotional wellbeing in primary(12) and secondary(13) schools. These key documents clearly demonstrate the need for a focus on improving the emotional wellbeing in young people in the UK.
Knowledge about factors that protect against mental health problems and improve emotional wellbeing in young people is required. There is some evidence of a positive although guarded association between physical activity and emotional health outcomes in adults(14,15), but young people are seldom studied. There are two relevant Cochrane reviews, the first by Larun et al(16) who examined the evidence for physical activity reducing or preventing common mental health disorders (anxiety and depression) amongst children and adolescents. They concluded that there was some evidence for a beneficial effect although the evidence was not robust enough, given the heterogeneity in the populations studied, to draw any conclusions. Ekeland et al(17), in a related Cochrane review, concluded that exercise had positive short-term effects on self-esteem in children and adolescents. However, neither review directly examined the association between physical activity and general emotional wellbeing. Similarly, some longitudinal studies have examined the relationship between depression and physical activity in adolescence(18-20) yet only two have explored the relationship between physical activity and emotional wellbeing(21,22). Predominately, the evidence is based on cross-sectional studies(23-26).
Clear conclusions about the association between physical activity and emotional wellbeing are difficult to make largely due to the use of different definitions and measures of physical activity and failure to investigate the dose-response relationship - the optimal amount and type of physical activity (incorporating information on the frequency, duration and intensity of such activity) to achieve emotional health benefits(27,28). Moreover, there have been a number of terms used to refer to emotional wellbeing (e.g. psychosocial health, mental health) and a range of different self-report questionnaires for measuring emotional wellbeing, which also makes it difficult to compare findings between studies. Furthermore, little is understood about the mechanisms that may underlie an association between physical activity and emotional wellbeing(22).
The proposed analysis will form part of an NIHR doctoral fellowship supervised by Professor Rona Campbell (co-supervised by Professor David Gunnell and Dr Ashley Cooper).
The research will analyse data already collected as part of ALSPAC to address the following aims:
* To determine whether physical activity is a protective factor for emotional wellbeing in young people
* To investigate the dose-response relationship between physical activity and emotional wellbeing in young people
* To determine whether emotional wellbeing is a barrier to physical activity in young people
Hypotheses will be tested through analysis of data from 2000 children in the ALSPAC cohort and 1000 children in the AHEAD cohort. Baseline physical activity was measured in the ALSPAC cohort at 12 years and then at follow-up age 14 then 16 years; AHEAD at 12 years then 15 years. Emotional wellbeing was measured at baseline in ALSPAC at 11 years and then at follow-up age 13 years; AHEAD at 12 years then 15 years.
Outcome data
The primary outcome is total difficulties score as measured by the SDQ (Strengths and Difficulties Questionnaire) http://www.sdqinfo.com/b1.html (29). This scale was selected as it is a valid and reliable instrument suitable for young people aged 11-16 and it can also be completed by parents or teachers (30). The 25 items in the SDQ comprise of 5 scales (emotional symptoms, conduct problems, hyperactivity/inattention, peer relationship problems, and pro-social behaviour) with 5 items in each. A 'total difficulties score' is calculated for the first four sub-scales (range 0-40). A high score (20+) on the total difficulties score can be used to identify those that may have mental health disorders. The 5 sub-scales will also be investigated independent of each other (each sub-scale ranges from 0-10). Child self-reported SDQ data collected from the AHEAD cohort and maternal and teacher reported data from the ALSPAC cohort will be analysed and also compared to explore the quality and content of the SDQ data from different informants.
Exposure data
Accelerometers are objective physical activity measurement devices which provide precision of measurement, as they overcome children's lack of ability to recall and quantify physical activity (31). The Actigraph accelerometer was selected for use with both the ALSPAC and AHEAD cohorts as it has been validated in both children and adolescents. Two main physical activity variables will be derived- total physical activity and time spent in moderate or vigorous physical activity (MVPA). Appropriate cut-points will be used to determine moderate and sedentary levels of activity.
Linear regression analysis (for continuous outcomes) and random effects logistic analysis (for dichotomous outcomes) will be used. Potential confounding factors will be controlled for e.g. age, gender, socio-economic markers and ethnicity in AHEAD and a number of additional confounders available from the ALSPAC data. Gender-specific effects will be formally tested.
In the AHEAD cohort (n=1000), an odds ratio of 1.36 can be detected with 80% power (at 5% significance). ALSPAC will have a minimum of 2000 participants in any analysis (due to a smaller number with physical activity data at 16 years compared to 12 and 14 years). If n=2000, an odds ratio of 1.24 can be detected with 90% power (at 5% significance).
The possibility of using the longitudinal nature of the data more fully can be explored by using repeated measures analyses. The role of e.g. social support as a potential mediator of this association (using number of close friends reported in the ALSPAC data, and peer nominations in the AHEAD data) could be investigated and the potential pathways using path analysis examined.
B917 - The associations between feeding difficulties and behaviours and dietary patterns at 2 years of age - 03/12/2009
We are proposing to extract dietary patterns using prinicipal components analysis from the FFQ data collected when the children were two years of age. The proposed work will build on the work already publised throughout childhood from 3 to 9 years of age (1-3) and the dietary pattern scores obtained can be used as a predictor for many other outcomes.
Two years of age is a crucial time for developing dietary habits. Weaning will have been achieved and it is a period when more adult type foods are being introduced and children begin to develop their own preferences. Several studies have shown that diet in the early years is important for later behavious. It can be a challenging time for parents and carers and we therefore propose to examine the associations between the dietary patterns obtained and current feeding difficulties and behaviours as reported by the main carer.
To our knowledge only one other study has examined dietary patterns in such a young age - Robinson et al examined infant dietary patterns at 6 and 12 months of age (4). It is therefore important to try to fill the gap in the knowledge for this age group.
B924 - A comparison of the caries risk of prematurely born children and full term children - 03/12/2009
The aim of this study is determine the caries risk in 5year old children born prematurely compared to full-term children. The type of study we will be a retrospective cohort study.
Our project will consist of:
1. A thorough literature search on publications related to our project. From the searches completed already we have found conflicting conclusions. An example of a paper which shows there is a relationship is 'National Pathfinder Survey on children's oral health in Italy: Pattern and Severity of caries disease in 4 year olds. Campus et al: Caries res 2009;43:155-162'. An example of a paper, which shows there isn't a relationship is 'Dental caries in pre-term and low birth-weight children and related factors. Ghasempour et al: J contemp Dent Pract.2009:10.'
2. From the papers collected in the literature search, we can compare how each of the studies were conducted and compare the anaylsis undertaken; and therefore how valid there conclusions are.
3. We will then be able to choose the appropriate confounders.
3. Obtaining the ALSPAC data would allow us to identify our variables, which will also include confounding factors. We would then anaylse the data using logistic regression, adjusting for the confounding variables, using the data package STATA.
4. We would then discuss our results, and draw a conclusion showing our outcome in the context of the existing literature.
Time-permitting we will further our research to investigate gestation as a continuous variable, following a similar outline to above.
B919 - Sedentary behaviour and academic performance in early adolescence - 01/12/2009
Background: Sedentary behavior and low levels of physical activity may predispose youth to increased risk of obesity and other chronic diseases in adulthood (3). Although research has shown lower educational achievement among adults with obesity, little is known whether the same association holds for children and adolescents (4). As recommended by the 2008 Physical Activity Guidelines Advisory Committee, one of the top research needs regarding the health benefits of physical activity among youth is to investigate the association between physical activity and academic achievement in children and adolescents (5). However, the guidelines do not specifically highlight the need to reduce sedentary behavior in terms of academic achievement, which is important given that children can be excessively sedentary while meeting the physical activity guidelines (6).
Purpose: To examine the association between sedentary behavior and academic performance in early adolescence.
Study Sample:
- Children ages 11-yr and 13-yr old
Dependent Variable:
- National tests of academic performance/achievement
- Key stage 2&3 assessments- standardized test scores in English, Math, and Science
- National curriculum level awarded in English, Math, and Science tests
- Total marks achieved in English, Math, and Science tests
- Tiers at which the exam was taken (for Math and Science)
Independent Variable:
- Sedentary behavior [accelerometer; mins/d less than 200cpm]
Covariates:
- Moderate-to-vigorous physical activity (MVPA) [accelerometer; minds/d >=3,600]
- Obesity [Percentiles: 90th (obese), 80th (overweight), 10-79th (normal)]
- BMI/DXA
- Child's birth weight
- Child's head circumference at birth
- Parental education
- Parental occupation
- Length of gestation
- Mother's age at birth of study child
- Mother breastfed
- Mother's smoking status during pregnancy
Data Analysis:
- Raw scores achieved in English, Math, and Science will be recalibrated using information from the key stage level achieved and the tier at which the exam was taken (for Math and Science) to obtain a transformed score in fraction of a level (method developed by Levacic et al. 2005. The same approach was also used by Leon Feinstein and Ricardo Sabates in their ALSPAC project entitled "Impact of adult learning on their children's school attainment").
- The transformations of the Math and Science scores are important because it takes into account the tier at which pupils enter the exam and their potential influence on the maximum score.
- Results will be interpreted as the amount of progress made by the student in months or years.
- Ordinary Least Squares (OLS) regression techniques, all regressions will be performed separately for English, Math, and Science test scores.
B930 - Replicate a GWAS study analysing 22 SNPS across 15 genes that may relate to alcohol phenotypes - 01/12/2009
Common genetic variation from candidate genes hypothesised to be related to alcohol metabolism, intake or predisposition to heavy and addictive consumption will be analysed in association with teenage drinking dimensions in univariate and multivariate analyses. Individual SNPs will be analysed one at a time, then we will attempt to create 'scores' grouping the SNPs showing the most marked association.
Phenotypes to be analysed will include age at initiation, age at first binge, frequency and severity of binge drinking episodes (based on the AUDIT q.aire), total amount drunk in average week. Both cross-sectional measures and derived life-time consumption measures will be used. Where possible and appropriate, trajectories of use will be defined as part of an ongoing collaboration with the SUGAR cluster interest in alcohol use.
See Appendix for details on 18 SNPs in 13 genes selected to be genotyped and supporting evidence (highlighted SNPs). We propose that these be genotyped in all available samples from young persons in ALSPAC. However, where information on these SNPs is available as part of GWAS and imputed data on 3000 individuals, no further genotyping would be needed.
B921 - A lifecourse study of epigenetic variation disease and ageing - 01/12/2009
Epigenetic markings play a major role in the expression of genetic traits but are not encoded in the DNA sequence of the genome, instead epigenetic information is carried in the form of chemical modifications to DNA, including methylation.
Epigenetic patterns change across the lifecourse and appear to be especially plastic in early life. Changes are also observed in later life although these appear to be limited mainly in non-coding regions of the genome harbouring repeat sequence motifs. The functional relevance of these various epigenetic alterations requires investigation. Evidence is mounting to illustrate that the epigenome 'captures' environmental exposures but little is known about how relevant this is to the pathogenesis of common complex disease or to the ageing trajectory.
Differentially methylated regions (DMRs) will be identified using MeDIP-sequencing on sets of 20 samples from multiple sample points (e.g. 0, 7y, 15y, 30y, 40y, 50y) to capture the full age distribution across the cohort and will include serial samples from the same ALSPAC mothers and children. A 384-plex panel of age-related DMRs will be taken forward for analysis on as many of the ALPSAC cohort as is plausible within the budget, currently estimated at 2,000-3,000 samples at multiple timepoints (n=7).
Data analysis will include modelling of epigenetic changes across the lifecourse as well as the development of MR-based analysis tools to incorporate DNA methylation data. The data set arising will be used to address the determinants of epigenetic patterns at specific time points, the factors driving their change over time and the subsequent consequences for health and disease in later life.
B911 - Breast tissue composition in young women and its pre-natal and early life determinant a study nested within ALSPAC - 26/11/2009
We propose to use magnetic resonance (MRI) to measure the water content of the breast, which, like mammographic radiodense tissue, reflects the amount of fibroglandular tissue in the breast. The main aim of the proposed study is to examine MRI breast tissue composition in relation to markers of pre-natal, childhood and adolescent growth and development. Specific aims are:
(i) To examine MRI breast tissue composition in young women (aged ~18 years) in relation to markers of foetal, childhood and adolescent growth and development;
(ii) to relate umbilical blood cord plasma levels of sex steroid hormones to MRI breast tissue composition in young women;
(iii) to examine MRI breast tissue composition in young women in relation to concurrent levels of endogenous sex hormones, prolactin, and growth factors
(iv) to compare MRI breast tissue composition in young women to mammographic breast density in their mothers
In the long-run, the proposed study has the potential to form the basis for an investigation into longitudinal changes in breast tissue composition from young adulthood, their determinants, and their implications for breast cancer risk and prevention.
B913 - Modelling the effects of residential mobility on children and young people - 23/11/2009
(No proposal received).
B914 - To determine if the omega-3 intake and blood status at various stages influence the adolescents drinking - 21/11/2009
This proposal builds on studies already completed or ongoing within ALSPAC regarding essential fatty acid status and behavioural outcomes during development. Several levels of data predict that greater omega-3 fatty acid intake and greater serum status will be beneficial in protecting individuals from developing substance misuse disorders, in particular alcohol misuse. We have reported that deficient omega-3 intake during pregnancy increases risk for low verbal IQ, suboptimal motor development and suboptimal social interactions. These parameters have been described as early markers of an adverse developmental trajectory.
In addition, a phenotype characterized by a low level of response to first exposures to alcohol has been well characterized in the ALSPAC cohort by Marc Schuckit and shown to be predictive of alcohol misuse at age 15. A plausible biological mechanism partially underlying this phenotype is hyper activity of the endocannabinoid system because it is associated with increased tolerance to alcohol and lower levels of response in both animal models and pilot clinical studies. Endocannabinoids are produced from phosholipid precursors in membranes and are comprised of arachidonic acid, an omega-6 essential fatty acid. We have demonstrated in animal studies that dietary depletion of omega-3 fatty acids causes reciprocal replacement with omega-6 fatty acids in phospholipids which, in turn, causes 2-3 fold greater endocannabinoid activity in liver and brain.
Aims: Determine if essential fatty status of RBC's during pregnancy, or in umbilical serum, or in 7 or 13 year old serum samples predict parameters of alcohol use or pheonotypic response to alcohol, throughout development.
Long chain essential fatty acid status is determined by a combination of direct dietary intakes and genetic variation in the FADS 1-2 gene complex, which enzymes that desaturate short chain precursors to long chain polyunsaturates. Thus, these determinants of serum essential fatty acid composition will be examined as predictive determinants of the parameters of alcohol use and response. These genetic variants are currently being examined as determinants of serum and maternal RBC fatty acid composition. Potential confounders including parameters of social, educational, smoking and nutritional status will be examined.
All data are or will be available, with the exception of fatty acid levels at age 13.We assume plasma will be available at this stage and a request has gone to the SR.If approved, the plasma will be assayed for fatty acid composition using the automated robotic high-throughput methodology established by the Section of Nutritional Neurosciences, NIAAA. NIH for the n = 4,090 umbilical cord and n = 6,500 7 year old serum samples.Validation of a method of high-throughput quantitative analysis of endocannabinoids in 250 ul of plasma is currently underdevelopment in this laboratory.If validated, endocannabinoid measures, in addition to the fatty acid compositional measures will be added to the rich ALSPAC database.
B915 - Data cleaning of diet diaries at 10 13 years - 18/11/2009
(No proposal received).
B907 - The relationship between insulin levels and skeletal development in ALSPAC - 16/11/2009
Background
Based on the positive association which we previously found between total body fat mass and bone mass as measured in ALSPAC children at age 10, we postulated that fat mass exerts a positive influence on bone mass accrual in childhood (1). A subsequent mendelian randomisation study in which genetic markers of obesity were used as instrumental variables supported a causal relationship between fat mass and bone mass accrual (2). In a recent study where we examined the relationship between fat mass and cortical bone geometry by pQCT, higher fat mass was associated with a greater overall cross section of bone, but reduced expansion of the inner surface, of which the latter effect was particularly marked in females (A Sayers et al, submitted for publication).
The biological pathways by which fat mass stimulates bone mass accrual in childhood are currently unclear, but may involve factors related to insulin resistance (3). For example, adiponectin is thought to be inversely related to insulin resistance, fat mass and BMD. However, in a recent study based on ALSPAC, whereas adiponectin was inversely related to a range of measures of skeletal development, this association only explained a relatively small proportion of the relationship between fat mass and bone development (A Sayers et al, submitted for publication). In terms of alternative pathways by which fat mass stimulates bone mass accrual in childhood, the insulin/IGF1 axis is another good candidate. IGF1 is known to increase overall bone cross section by stimulating periosteal bone growth in animal models (4). We also previously observed a positive relationship between IGF-1 and BMD measures at age 10 in the children in focus subgroup (J Tobias unpublished observations).
Aims and methods
We aim to establish whether insulin levels, as measured in children at age 16 (TF3) are related to bone mass accrual and cortical geometry as assessed at the same age. If a positive association is observed as expected, we will go on to determine whether this relationship helps to explain associations between the same skeletal parameters and fat mass and adiponectin as previously described. Insulin levels at age 16 will be analysed in relation to bone mass as measured by total body DXA scans at the same age, as will related factors such as fasting blood sugar and lipid levels measured at the same time. We will analyse relationships between total body less head bone mineral content, bone area, bone mineral density and bone mineral content adjusted for bone area. We will also investigate associations with indices of cortical geometry derived from pQCT scans of the mid tibia. Assuming positive associations are observed, we will then examine whether associations between fat mass and bone mass can be explained by those with insulin, by comparing fat-bone mass associations with or without adjustment for insulin. Equivalent analyses will be performed for adiponectin, with the exception that for these analyses, adiponectin levels will consist of those measured at age 10.
B906 - Prevalence course and age of onset of anxiety disorders in longitudinal samples - 16/11/2009
The project follows on from a systematic review of published longitudinal data appertaining to the prevalence, course and onset of anxiety disorders in childhood and adolescence submitted as part of a doctoral thesis to the University of Manchester (Gallagher, 2008). From the limited published data in the field this review concluded that it was likely that many adult anxiety disorders had their onset in childhood and adolescence, and that childhood anxiety was as prevalent as anxiety disorders in adulthood, yet receives only a small proportion of the research attention, despite a lack of specific developmentally appropriate treatment models. Due to the scarcity of published data on diagnosed anxiety disorders in these populations, these conclusions could not be sufficiently substantiated.
The current project has secured access to large longitudinal data sets from Germany and the US that have not yet been fully analysed in terms of prevalence, course and age at onset of anxiety disorders. Once these datasets are fully analysed then this data will be used in a metaanalysis to determine at what age onset of each anxiety disorder is most likely, whether early onset confers greater risk for later disorders, and in samples which have data into adulthood what proportion of adults had their first onset in childhood or adolescence.
The data requested from ALSPAC would therefore include diagnostic status according to the DAWBA at age 7, 10 and 13(14). These data would be analysed to generate prevalence rates, new onset and continuity of disorders across the three follow ups. The combination of mother and child response (if available) will be dictated in part by the methods used in existing data sets as this will then be compared to data from two published data sets and the two data sets that we have analysed. This information will also be statistically combined in order to generate mean prevalence and median age at onset for anxiety disorders across the samples. As the ALSPAC data is currently collected into adolescence it would not be suitable for inclusion in the final metaanalysis examining age at onset of adult disorders. However, this data will add to the current proposal in two major areas. 1) Longitudinal data from Britain was not previously included due to a lack of published data and ALSPAC data will give insight into patterns of the development of anxiety in a British population. 2) Only one of the other studies included has collected data before the age of 11, the inclusion of prevalenc data for seven year olds will add to the understanding of the course from childhood into adolescence.
B908 - Genetics of Osteoarthritis - ARCOGEN study - 15/11/2009
Osteoarthritis is the most common form of arthritis and has a huge clinical and economic impact. The disease has a complex aetiology but genetic factors play an important role in pathogenesis. Identification of the loci that predispose to OA susceptibility will lead to a much better understanding of the cause of the disease as well as identifying new molecular targets for therapeutic intervention. Alleles that predispose to OA could also be used, in individual cases, to determine prognosis and the response to surgical or drug treatments.
We have undertaken a genome-wide association study for OA susceptibility alleles in a large collaborative study of 8000 OA cases involving 5 UK centres using Illumina's 610K array. Analysis of the first 4000 cases was performed using publically available common control sets from circa 5000 UK individuals from the 1958 British Birth Cohort (58BC), the UK Blood Service collection of the Wellcome Trust Case Control Consortium and the TwinsUK. In the second stage of the study we need additional UK controls preferably analysed using the same platform. We have obtained permission to use non overlapping sets of 58BC (2500 samples) typed by the T1D Geneetic Consortium and TwinsUK (2000 samples). To further increase the control sample we seak permission to include the genotypic data of the 3000 ALSPAC samples profiled with the 317K and 660K arrays at the Sanger Institute. The genotypic data are available at the Sanger Institute and therefore there is no need to resend. Analyses may require to correct for height and weight and therefore we are requesting this data for the oldest available age. Genotypic data will be used as common controls in the main discovery GWA analysis as well as part of further meta-analyses of other OA scans.
We would like to request
Sex Child
Body weight Child Clinic latest age
Height Child '' latest age
Genotypes (317K or 660K array) Child ''
B905 - CORNET a consortium to identify genes influencing variation in plasma cortisol - 15/11/2009
Variations in morning plasma cortisol are associated with quantitative traits including blood pressure and glucose, serum lipid profile, mood and memory. Morning plasma cortisol is a heritable trait but little is known of the genes that contribute. To identify such genes we have undertaken cortisol measurements and genome wide association studies (GWAS) in ~3000 participants from 3 cohorts. Initial results are promising, identifying two loci in which the top hit SNPs are associated with cortisol to pless than 10-8. We have therefore established the CORtisol NETwork (CORNET) in order to: (i) extend GWAS with morning cortisol to n=10,000; (ii) genotype top hit SNPs in at least two replication cohorts, in which additional measurements of dynamic control of cortisol are available; and (iii) resequence loci around top hits to identify rarer variants contributing to cortisol variation. The results will provide novel insights into pathways leading to cardiometabolic and neuropsychiatric disease.
B904 - The effects of common genetic variation in the thyroid hormone pathway on growth and development - 07/11/2009
We intend to analyse the genotypes TSH-R, PDE8B and DIO1 and their phenotypic effects on bone and neurological and emotional development in the entire ALSPAC cohort, as genotype data are already available from previous work on childhood growth. We will also use GWA data to assess the phenotypic effects of other SNPs known to be reasonable thyroid related candidates, but which do not satisfy strict thresholds for "genomewide significance".
B902 - Mechanisms linking early nutrition growth and ageingthe role of DNA methylation LINKED TO B0903 and 0921 - 04/11/2009
The influence of early nutrition and growth on ageing processes and the role of DNA methylation in mediating this will be explored by integrating transcriptomic and epigenomic profiling of a rodent model and two human cohorts. Information arising from these analyses will be replicated in several human cohorts of different age strata inclduing ALPSAC.
As outlined above, quantitative DNA methylation analysis will be undertaken on one or more platforms. DNA quantity and volme will therefore need to be stipulated when the platform is confirmed but will be in line with requirements already familiar to the ALSPAC lab team as the platforms are those used for MRC CAiTE pilot epigenetics work. Assays will be multiplxed where possible and DNA used for multiple assays from the same bisulphite conversion where ever possible.
The possible range of phenotypic data required is outlined above and will be specified once the discovery phase of the proposed project has been completed and any associations between DNA methylation and phenotypic traits have been confirmed in the discovery cohorts.
Exposure data requested will include birth weight, growth during infancy (conditional weight; birth to 3 months, 6 months and 12 months) .
Data requested may include existing DNA methylation information held by ALSPAC at the time the project comes online, including that generated by CAiTE epigenetics pilot work or proposed human methylation 27K array analysis.
B899 - Developmental origins of bone phenotypes - 02/11/2009
There is increasing evidence for a developmental origin of bone mass and skeletal development. Peak bone mass, together with later life bone mass loss, is a major determinant of osteoporosis risk in later life.(1) Lower birth weight and earlier gestational age are associated with reduced bone mineral density (BMD) and bone mineral content (BMC) in infancy, childhood and adulthood and with adult fracture risk in a number of studies.(2-5) These findings have led to the exploration for modifiable pre-natal risk factors that affect later bone health.
Maternal pregnancy diet/nutrition and offspring bone health
Variation in maternal vitamin D during pregnancy has been positively associated with offspring total and spinal BMC and with total BMD in a small study (N=198) from the Southampton Women's Survey.(6) In ALSPAC (N=6995) maternal UVB exposure in pregnancy (used as proxy for vitamin D level (7)) is related to bone size at mean age 9.9 years independently of height and lean mass.(8) We will further explore maternal vitamin D in pregnancy with offspring bone phenotypes as agreed with ALSPAC exec. as part of the MRC Vitamin D grant (DAL PI) and so this is not discussed further in this application.
Whilst other constituents of maternal diet/nutrition (other than vitamin D) have been implicated in offspring bone health, a detailed analysis in ALSPAC suggested that variation in folate intake during pregnancy was the only important predictor of future offspring BMC after adjustment for child size and other potential convariables.(9) In a separare ALSPAC publication the child's own C667T MTHFR genotype was found to be associated with spinal BMD at mean age 9.9 years, with each additional T allele of that variant being associated with a 0.10SD decrease in spinal BMD on average (but with no association to total body bone phenotypes).(10) There was some evidence in that study that this association was considerably weakened in children whose mother's diet during pregnancy, and whose own diet in childhood was high in vitamin B, suggesting that high levels of homocysteine adversely affect accrual of trabecuar bone.
Maternal smoking in pregnancy
In the Southamton Women's Survey (N=145 and in later study = 841) maternal smoking in pregnancy was associated with decreased whole body BMC at birth,(11,12) and in the Tasmanian Infant Health Survey (N=330) maternal smoking in pregnancy was associated with reduced size adjusted bone mass at the lumbar spine and femoral neck (but not whole body) at mean age 8 years in children who were born at term.(13) The association of maternal smoking in pregnancy with offspring bone phenotypes has not been examined in detail in the ALSPAC cohort, to our knowledge, though in one publication concerned with diet and bone phenotypes it was noted (but results not presented) that maternal smoking was not associated with DXA determined bone phenotypes at age 9.9 years.(9) Any association of maternal pregnancy smoking with offspring bone phenotypes might be due to intrauterine mechanisms or might be due to the association of maternal smoking in pregnancy with other lifestyle characteristics, that are shared by the offspring (including earlier initiation of smoking) that affect bone development (though for associations with bone phenotype at birth this is unlikely). One way of exploring this that has been used in ALSPAC with other offspring outcomes is to compare the association of maternal smoking in pregnancy with offspring bone phenotype to that of paternal smoking in pregnancy to bone phenotype; a stronger maternal association would be expected if intrauterine mechanisms are important.
Maternal adiposity and weight gain in pregnancy
In the Southampton Women's Survey (N=841) independent predictors of whole body bone area and BMD included maternal own birthweight, height and triceps skinfold thicknes in pregancy.(12) One conclusion of this study was that offspring of women with greater fat stores in pregnancy had better bone development and greater BMC at birth. They speculated that maternal fat stores might influence intrauterine bone mineral acrual through several mechanisms including effects on nutrient availability and endocrine factors such as leptin and oestrogen. The latter would be influenced in pregnancy by both maternal adiposity pre-pregnancy and greater acquisition of fat during pregnancy. This potential advantage of greater maternal fat stores in pregnancy requires further exploration since other studies suggest that greater maternal adiposity in pregnancy might be detremental to offspring future cardiovascular and metabolic risk.(13) As with other phenotypes comparing maternal pre-pregnancy adiposity associations to those of paternal prepregnancy adiposity associations can help to distinguish intrauterine mechanisms from shared familial genetic and lifestyle characteristics.
Given that much of the work to date on potential modifiable pre-natal influences on offspring bone phenotypes has come from the Southampton Women's Survey with a relatively small sample size and with general non-specificity (variation in associations with different bone phenotypes) there is a need for further exploration of this area. This is particulalry important with respect to maternal adiposity and fat stores in pregnancy where associations with potential advantages for bone health may be in the opposite direction to those for cardiovascular health.
Objectives
1. To examine the association of maternal smoking in pregnancy with offspring bone phenotypes and compare these associations with equivalent associations for paternal smoking in pregancy
2. To examine the association of maternal pre-pregnancy BMI with offspring bone phenotypes and compare these associations with equivalent associations for paternal prepregancy BMI
3. To examine the associations of maternal gestational weight gain with offspring bone phenotype
Methods
All applicants will contribute to the analysis protocol. DAL will compile a dataset. CM will complete analyses with supervision from DAL. All applicants will contribute to interpretation of results and writing of papers.
References
1. Hernande CJ, et al. Osteoporosis Int 2003;14:843-847
2. Jones G, Dwyer T. Calcif Tissue Int 2000;67:304-308
3. Yarbrough DE, et al. Osteoporosi Int 2000;11:626-630
4. Dennison EM, et al. Pediatric Research 2005;57:582-86
5. Cooper C, et al. Osteoporosis Int 2001;12:623-629
6. Javaid MK, et al. Lancet 2006;367:36-43
7. Sayers A, et al. IJE; 2009 doi:10.1093/ije/dyp237
8. Sayers A & Tobias JH. J Clin Endocrinol Metab 2009;94:765-771
9. Tobias JH, et al. Osteoporos Int 2005;16:1731-1741
10. Steer CD, et al. J Bone & Mineral Res 2009;24:117-124
11. Godfrey K, et al. J Bone & Mineral Res 2001;9:1694-1703
12. Harvey K, et al. J Developmental Origins of Health & Disease 2009, in press
13. Jones G, et al. J Bone & Mineral Res 1999;14:146-151
14. Viswanthan M et al. Evidence Report/Technology Assessment No. 168. AHRQ Publication No. 08-E009 ed. Rockville, MD: Agency for Healthcare Research and Quality, 2008.
B916 - 10000 UK genome sequences accessing the role of rare genetic variants in health disease - 01/11/2009
The 10,000 UK Genome Sequence project is an expansion on the GWAS (common gene variant) studies that we have already performed on our twin cohort. Our primary objective is to investigate in a systematic genome-wide fashion the contribution of low frequency and rare genetic variants to medical traits, based on genome-scale sequencing of phenotyped samples. We will study multiple groups of related phenotypes so as to explore rare variants in different types of disease process, and add exomes from extreme sample sets to increase power and compare clinical extreme to population cohort designs. The sequence data set we generate will provide a public genotype resource an order of magnitude deeper than available currently that will empower future human genetic research in the UK and beyond.
We do not plan to do provide the twins with the results of their genetic data, which is in keeping with the procedures we have maintained in our previous genetic projects.
B900 - Increasing the breadth of understanding about the factors that affect pupil attainment - 30/10/2009
DCSF has exploited the National Pupil Database to investigate the kinds of things that are related to pupils' attainment. The PLASC variables have allowed the department to disentangle the effects of, say, ethnicity, gender and SEN status. However, the range of data available for use in this way is rather narrow, and the same kind of information is generally available every year. With only NPD available, it was not possible to examine the effects of other factors about which information is not collected, or to assess the extent to which the effects of the NPD variables appear inflated due to the lack of this other information. Further, there is still a large unexplained variation at pupil level which it is presumed must be partly due to those variables we don't have information for.
Schools Analysis and Research Division has formed a new team which has among its responsibilities the task of investigating other data sources which can broaden the Department's understanding of what affects attainment. This team has done some work using other longitudinal data sources where factors such as parental education, material deprivation and family composition have been unpicked in terms of their influence on raw GCSE scores and also on progression from Key Stage 2 to Key Stage 4. It has also been possible to look at how important each of these factors is in terms of explaining gaps in attainment and progression between certain groups of pupils. For example, by combining the effects of different levels of aspirations with the prevalence of each level amongst both Free School Meal pupils and their counterparts, it is possible to attribute some of the FSM gap to aspirations. Doing the same for other factors allows comparison of the relative sizes of the contributions to the gaps.
Given that ALSPAC has also been matched into NPD, it offers two key things that the other data sources have not been able to provide, and which would enable similar work to be carried out and provide answers to similar questions. First, there is a wide range of other data collected, such as birth weights, handedness and child development scores. These would allow us to look at the relative importance of these factors in explaining gaps in attainment and progression alongside the other family and school factors available on NPD. Second, the young people involved in ALSPAC have had information collected about them from before birth, and should all have completed Key Stage Four. This gives a huge depth of information about the individuals as well as breadth. It would enable analysis of the factors that contribute to earlier attainment than has been looked at so far, Key Stage Two results, and allow greater likelihood of attributing causality, as it will allow factors from an earlier age to be taken into account.
Indeed, analysis could look separately at attainment at Key Stage Two, attainment at Key Stage Four and progression between the two, using data collected at ages appropriate for each part. Results from this work would be used to inform policy development inEngland, particularly on narrowing the attainment gap between disadvantaged children and others. Analysis would be carried out internally by government statisticians using the GSS code of practice. Results may be published as departmental research reports, possibly alongside similar analysis from other sample data sources.
B898 - Replication study A common variant on chromosome 11q13 is associated with atopic dermatitis - 26/10/2009
Common variant (rs7927894) has a reasonably strong effect on atopic dermatitis [1]. We propose two
analyses: 1) to look at the association of this genetic variant with atopic dermatitis in ALSPAC and 2) to look at the association of this genetic variant with asthma, given the effect of filaggrin on asthma.
References
[1] Gordillo et al. A common variant on chromosome 11q13 is associated with atopic dermatitis Nat Genet. 2009 May;41(5):596-601.