Objectives

1)To determine the impact of adolescent disordered eating behaviours on bone accrual and bone development throughout adolescence and young adulthood.

2) To determine the role of factors such as caloric restriction, delayed puberty, excessive exercise, gender in explaining the effect of disordered eating on bone health in adolescence.

3) To determine the impact of adolescent disordered eating behaviours on stress fractures in young adulthood.

Background

Eating disorders (ED) are chronic disorders and affect about 5-10% of the population (Hoek and Van Hoeken, 2003).

Adolescent ED and bone development: ED have a peak incidence in adolescence, a crucial time for physical and skeletal development. Pubertal years are the period in the life course when bone mass reaches its highest level. The level of bone mass attained at this age is a key determinant of long-term bone health and risk of osteoporotic fractures in later life (Ott, 1991).

Past research focusing on bone density in ED has shown that restrictive ED (such as anorexia nervosa (AN) and atypical anorexia nervosa/ Eating Disorder not otherwise specified-EDNOS) are associated with reduced bone mass. This is probably secondary to low weight, reduced fat mass and hormonal abnormalities secondary to poor nutrition (Misra, 2008; Mehler et al, 2011). Whilst recovery from the ED has been shown to lead to improvements in bone density, several studies show that bone density remains lower than in individuals who have not experienced an ED, particularly in cases where the ED onset was in adolescent years (Biller et al., 1989; Hartman et al., 2000; Wentz et al., 2007; Bolton et al., 2005).

Binge- type ED (such as bulimia nervosa and binge eating disorder) have been less well studied, and appear not to affect bone density to the same extent as restrictive ED, particularly if normal weight is maintained (Misra, 2008). However, higher fracture risk was also shown in these women compared to controls (Vestergaard et al., 2003).

Studies investigating bone density in adolescents with restrictive ED as measured by dual energy X-ray absorptiometry (DXA) have shown low bone density at the spine, hip and femoral neck (Misra, 2008); as well as reduced bone turnover. These effects are apparent in both girls and boys (Castro et al., 2002). Follow-up studies of adolescents have reported some improvement in bone density with nutritional restoration, mainly in areas of trabecular bone, such as the spine. Several studies, however, have shown fewer improvements in bone density at cortical sites despite recovery from the ED ((Herzog et al., 1993; Mika et. al, 2007). Given that the adolescent years provide a narrow window of opportunity in which to optimize bone mass accrual, disruption during these years might lead to permanent deficits.

Stress fractures:Although stress fractures are relatively uncommon, they affect as many as 15% of young women athletes and military recruits. Among females, other suspected risk factors for developing a stress fracture include disordered eating (Rosen et al, 1986) and irregular menstrual cycles, both which may result in a deficient estrogen status that can counteract the beneficial effects of exercise on bone density (Wolman et al., 1990; Lloyd et al. 2000; Bass et al.,1998).The combination of disordered eating, amenorrhea, and low bone density is characterized as the Female Athlete Triad. The recognition of the triad has served to increase awareness of bone health concerns among young female athletes (Yeager et al., 1993; Otis et al., 1997). Thus, it is extremely important to identify modifiable predictors of stress fracture so that prevention programs could be developed for high-risk populations.

Gaps in the literature:Most studies in the field to date rely on small clinical samples, and are likely to represent more severely ill subjects attending in- or out-patient services and not be generalisable to all ED. Length of follow-up has often been limited to 1-2 years. Moreover, DXA provides an overall estimate of bone mass but does not directly measure aspects like cortical thickness and cross sectional area that determine overall bone strength. Techniques like peripheral quantitative computed tomography (pQCT) can provide detailed information about cortical bone geometry and strength, but only one study to date has used this method to study bone changes in adolescents with ED (Milos et al.,2007). In light of recent finding that fat mass is an important positive determinant of cortical bone size and thickness (Sayers & Tobias, 2009), we are particularly interested in examining whether adolescent ED predicts weaker cortical bones due to reduced fat mass. More than half of adult bone calcium is acquired during adolescence and a woman's peak bone mineral density, a major determinant of her long-term risk of osteoporosis, is thought to be achieved by early adulthood. Therefore assessing bone density in young adulthood, as well as throughout adolescence, allows a better assessment of peak bone density. This additional investigation will allow a better understanding of the longitudinal effects of adolescent ED behaviours on bone development.

To our knowledge there are: no studies on a general population sample of adolescents/young adults able to link temporal relationships between disordered eating behaviours and bone development. Both would allow a clear identification of causal biological mechanisms, and take into account the role of confounders. The current lack of evidence impacts on available prevention and early treatment for patients with ED.

This study is unique in that data have already been collected prospectively and independently on bone density and ED behaviours in about 6,000 adolescents from the Avon Longitudinal Study of Parents and Children (ALSPAC). This longitudinal prospective study will allow: (1) determining precise temporal relationships between predictor (ED behaviours) and outcomes (bone density, cortical bone size and thickness, peak bone mass, and stress fractures); (2) focusing on causal biological mechanism taking into account the role of confounders (thanks to the wealth of data available in ALSPAC).

Methodology

Theoretical/conceptual framework: This is a longitudinal study, which will rely on data prospectively collected as part of the ALSPAC study. The ALSPAC study is a longitudinal prospective cohort of 14,000 mothers and their children. Women were enrolled in the study in pregnancy. Children have been followed up at regular intervals from birth up to age 18.

Research questions:do adolescent disordered eating behaviours (not only clinical ED) negatively affect bone accrual and bone development during puberty and into young adulthood, including individuals with subclinical disorders not currently viewed as being at risk? Do reductions in fat mass contribute to these deleterious effects of ED behaviours on bone development particularly those on cortical bone? This being the case, are reductions in fat mass in the context of ED behaviours particularly harmful, reflecting the fact that they are achieved by dietary restriction as opposed to by increased physical activity? What is the impact of disordered eating behaviours on peak bone mass?

Methods:

Outcomes:

1) Total DXA scans have been performed on the children/adolescents at ages: 11.5 (n = 7159), age 13.5 (n = 6147), age 15.5 (n= 5509) and age 17.5 (approximately 4000). In addition, hip DXA scans have been performed at age 13.5 and 17.5, and pQCT scans of the mid tibia at 15.5 and 17.5. Standard DXA and pQCT parameters related to bone development have been derived and will serve as the main outcomes. Data on fractures has also been collected at regular intervals.

We are seeking funding for a new wave of data collection to collect:

1. a 1-page questionnaire on stress fractures;

2. a two page questionnaire (or computerised assessment) on disordered eating behaviours and participation in exercise

3. DXA (hip and total body) and pQCT (tibia plus radius)

4. objective measurement of weight and height

Predictors:

1) Data on adolescent ED behaviours have been collected at age 13, 14, 16, and 18 adolescents

Other explanatory variables (mediators and/or confounders):

1)- Data are also available on pubertal timing and anthropometric measures at all the above ages on ~6,000 adolescents.

-Total body fat and lean mass as measured by total body DXA

-Physical activity as measured in the children by accelerometry at multiple time points

-Detailed information on adolescent diet using a combination of diet diaries and food frequency questionnaires

-Extensive information on socio economic status

Data collection: we envisage the new data collection can be part of an ALSPAC clinic, alternatively we will select individuals on the basis of their exposure (any disordered eating in adolescence, N=600-800) and unexposed individuals (no disordered eating in adolescence, N=600-800) to assess.

Analyses: Univariate and multivariate logistic models will be used to determine the effect of relevant predictors on outcomes. Longitudinal modelling will be used for longitudinal repeated data and to model hypothesised relationships.

Aim: To carry out a genome-wide association study that takes into account not just lifetime asthma status but also the presence of hayfever.

Hypotheses: The power of asthma genome-wide association studies (GWAS) can be improved -- and so new susceptibility loci can be identified -- by defining cases and controls based on diagnostic criteria that are more refined than a simple doctor diagnosis of asthma. We hypothesise that the presence of hayfever in addition to asthma defines a subgroup of asthma patients that will help map new loci for allergic disease.

Independent variable: Individual genotyped or imputed SNPs located in the autosomes or X-chromosome, coded as allelic dosage.

Dependent variable: Binary phenotype (ie case-control status) with affected individuals defined as those individuals who reported in any of the available ALSPAC surveys (7.5 and 14 years) to have been diagnosed by a doctor with asthma *AND* reported in any of the available surveys (11 and 14 years) to have hayfever. So lifetime self-reported asthmatics with hayfever. Unaffected individuals are defined as those individuals who never reported in any survey to have asthma or hayfever.

Confounding variable: None.

Analysis plan: Perform association analysis of SNP dosage after applying standard QC filters (eg. MAFgreater than 1%, HW P-value greater than 10-6, call rate greater than 95%, imputation info/r2 greater than =0.3) and excluding samples of non-European ancestry. Analyse males and females separately and include SNPs on the autosomes and the X-chromosome.

This proposal will examine the prospective association between antisocial behavior in childhood with increased cardiometabolic risk in adolescence.

Specific aims:

1. To test whether childhood psychopathology, including antisocial behavior and conduct disorder predict higher cardiometabolic risk in adolescence as indicated by higher central obesity, hypertension, glucose intolerance, and high levels of inflammation markers.

2. To examine whether a prospective association between conduct disorder in childhood and cardiometabolic risk in adolescence varies by gender.

3.To examine whether a prospective association between conduct disorder in childhood and cardiometabolic risk in adolescence is modified by comorbidity with depression.

Exposure variables: Antisocial behavior, conduct disorder in childhood

Outcome variables: Blood presure, BMI, central adiposity, glucose intolerance, lipids profile, CRP, IL6 at ages 14-16

Covariates (confounders and effect modifiers): Depression, trauma, emotional problems, social class, IQ, parenting, psychosocial stress in pregnancy, parental history of psychiatric disorders and substance use, parental history of cardiometabolic disorders.

Aims

1) To explore associations between early parenting and parental knowledge/monitoring of teenage behaviour.

2) To investigate associations between parenting and adolescent risk behaviour, with a view to establishing whether parental knowledge/monitoring mediates any associations between early parenting and adolescent risk involvement.

Background

Parental monitoring and knowledge of teenage behaviour have been widely identified as important protective elements limiting teenage risk involvement [1-8]. However, there is no consensus on processes covered by the terms parental knowledge or monitoring. Stattin and Kerr's work gave primacy to parental knowledge, emphasising the key role of children's disclosure of information [9-11]. Other work has indicated that parent-led behaviours, such as parental solicitation of information and rule-setting, may also be involved [12-14].

Comparatively little is known about what leads some parents to have lower levels of knowledge about their teenagers, or to monitor their children less attentively. A recent review of the relationship between parental monitoring and knowledge and conduct problems [15] has called for further research in this area. This review identified three main groups of possible antecedents to parental monitoring and knowledge: contextual, child characteristics and aspects of early parenting. Contextual factors influencing the level of monitoring/knowledge include the child's gender, family status, SES, ethnicity and neighbourhood; while child characteristics include early conduct problems, early antisocial behaviour and a difficult or "resistant to control" temperament. Both groups of factors have been relatively well studied. The focus of the current study is on the third group of antecedents, early parenting.

Various aspects of the parent-child relationship (parental trust in their child, authoritative parenting style, parental sensitivity/responsiveness) and time spent in family activities have been found to be associated with greater parental knowledge or teenage disclosure, but these have been studied only in teenagers [12, 13, 16-19]. Little is known about how parenting of young children may relate to parental monitoring and knowledge of teenagers. Two (related) studies have suggested the importance of early "proactive" parenting: averting potential problems in early childhood before they become entrenched [20, 21]. While anticipating behavioural problems seems to be a useful parental strategy that may lay foundations for successful monitoring of teenagers, the teenage research suggests additional qualities related to early parent-child connectedness and behaviour management may also be important.

Hypotheses:

1. Early parent-child connectedness and behaviour management will be associated with (respectively) greater teenage disclosure of information and greater parental control of a teenager's behaviour.

2. Teenage disclosure and parental control will both be associated with teenage risk involvement.

3. Early parenting will be associated with teenage risk, directly and/or indirectly via parental knowledge and monitoring.

Research using ALSPAC data set

Outcome measures

Stattin and Kerr measures of parental knowledge, parental solicitation, parental control and child disclosure were collected from the 13-year old clinic sample and from their parents. Child-reported teenage risk behaviour (antisocial behaviour, substance use and sexual behaviour) was collected from the age 15 clinic sample and questionnaires at a comparable age.

Main exposures

Measures of early parenting were collected from mothers at ages 2 1/2, 3 1/2 and 6 years, covering important aspects of parenting such as connection, conflict, autonomy, rule-setting, and discipline.

Potential confounders

Likely important confounders of associations between early parenting and parental monitoring and knowledge of teenagers include contextual factors (gender, SES, ethnic group, and family status), and childhood behavioural problems, risky friendships and involvement in antisocial behaviour. Maternal mental health, parental substance use, quality of partner relationship and child pubertal development will also be explored as potentially important confounders of associations between early parenting and adolescent risk outcomes.

References

1. Hoeve, M., et al., The Relationship Between Parenting and Delinquency: A Meta-analysis. Journal of Abnormal Child Psychology, 2009. 37(6): p. 749-775.

2. Dishion, T.J. and R.J. McMahon, Parental Monitoring and the Prevention of Child and Adolescent Problem Behavior: A Conceptual and Empirical Formulation. Clinical Child and Family Psychology Review, 1998. 1(1): p. 61-75.

3. Li, X.M., S. Feigelman, and B. Stanton, Perceived parental monitoring and health risk behaviors among urban low-income African-American children and adolescents. Journal of Adolescent Health, 2000. 27(1): p. 43-48.

4. DiClemente, R.J., et al., Parental monitoring: Association with adolescents' risk behaviors. Pediatrics, 2001. 107(6): p. 1363-1368.

5. Boyer, T.W., The development of risk-taking: A multi-perspective review. Developmental Review, 2006. 26(3): p. 291-345.

6. Li, X.M., B. Stanton, and S. Feigelman, Impact of perceived parental monitoring on adolescent risk behavior over 4 years. Journal of Adolescent Health, 2000. 27(1): p. 49-56.

7. Huebner, A.J. and L.W. Howell, Examining the relationship between adolescent sexual risk- taking and perceptions of monitoring, communication, and parenting styles. Journal of Adolescent Health, 2003. 33(2): p. 71-78.

8. Rai, A.A., et al., Relative influences of perceived parental monitoring and perceived peer involvement on adolescent risk behaviors: An analysis of six cross-sectional data sets. Journal of Adolescent Health, 2003. 33(2): p. 108-118.

9. Stattin, H. and M. Kerr, Parental monitoring: A reinterpretation. Child Development, 2000. 71(4): p. 1072-1085.

10. Kerr, M. and H. Stattin, What parents know, how they know it, and several forms of adolescent adjustment: Further support for a reinterpretation of monitoring. Developmental Psychology, 2000. 36(3): p. 366-380.

11. Kerr, M., H. Stattin, and W.J. Burk, A Reinterpretation of Parental Monitoring in Longitudinal Perspective. Journal of Research on Adolescence, 2010. 20(1): p. 39-64.

12. Fletcher, A.C., L. Steinberg, and M. Williams-Wheeler, Parental influences on adolescent problem behavior: Revisiting Stattin and Kerr. Child Development, 2004. 75(3): p. 781-796.

13. Soenens, B., et al., Parenting and adolescent problem behavior: An integrated model with adolescent self-disclosure and perceived parental knowledge as intervening variables. Developmental Psychology, 2006. 42(2): p. 305-318.

14. Wight, D., L. Williamson, and M. Henderson, Parental influences on young people's sexual behaviour: A longitudinal analysis. Journal of Adolescence, 2006. 29(4): p. 473-494.

15. Racz, S.J. and R.J. McMahon, The relationship between parental knowledge and monitoring and child and adolescent conduct problems: a 10-year update. Clin Child Fam Psychol Rev, 2011. 14(4): p. 377-98.

16. Smetana, J.G., et al., Disclosure and secrecy in adolescent-parent relationships. Child Development, 2006. 77(1): p. 201-217.

17. Keijsers, L., et al., Reciprocal Effects Between Parental Solicitation, Parental Control, Adolescent Disclosure, and Adolescent Delinquency. Journal of Research on Adolescence, 2010. 20(1): p. 88-113.

18. Willoughby, T. and C. Hamza, A Longitudinal Examination of the Bidirectional Associations Among Perceived Parenting Behaviors, Adolescent Disclosure and Problem Behavior Across the High School Years. Journal of Youth and Adolescence, 2011. 40(4): p. 463-478.

19. Darling, N., et al., Predictors of adolescents' disclosure to parents and perceived parental knowledge: Between- and within-person differences. Journal of Youth and Adolescence, 2006. 35(4): p. 667-678.

20. Pettit, G.S., et al., Predicting the developmental course of mother-reported monitoring across childhood and adolescence from early proactive parenting, child temperament, and parents' worries. Journal of Family Psychology, 2007. 21(2): p. 206-217.

21. Pettit, G.S., et al., Antecedents and Behavior-Problem Outcomes of Parental Monitoring and Psychological Control in Early Adolescence. Child Development, 2001. 72(2): p. 583-598.

The ALSPAC dataset can answer the research questions we have efficiently and quickly, and with many major scientific strengths. It is the largest sample of its kind which is longitudinal, has good measures of the relevant exposures (total volume of habitual physical activity and intensity of physical activity measured by accelerometry at age 11 and 13y), of confounders, and of outcome measures (cognitive measures, academic attainment, mental health and well-being as measured by the SMFQ and the SDQ). Almost all previous studies to address our research questions have suffered from a combination of small sample size, cross-sectional design, crude or biased measures of exposures and/or outcomes, limited consideration of confounders.Please note that we already have the data required, as part of an existing collaboration with ALSPAC on a Bupa Foundation funded secondary analysis (ALSPAC coapplicants Prof Ness, Dr Leary, Dr Joinson; data buddy Dr Northstone) of associations between physical activity and cognition, academic attainment, and mental health in typically developing 11-13y olds in ALSPAC. The current proposal adddresses essentially the same research questions, but in study participants excluded from our current study on the grounds that they have developmental disorders.

Objectives:

1. To develop normal reference ranges (normograms) for blood pressure across pregnancy based on pre/early-pregnancy maternal risk factors, conditional on the first blood pressure measurement, and to assess the ability of deviations from these trajectories to predict PE, preterm birth and SGA.

2. To investigate associations of patterns of blood pressure change in pregnancy with offspring perinatal outcomes (preterm birth, birth weight and SGA) and cardiovascular risk factors, growth and changes in adiposity in the offspring during childhood and adolescence.

3. To identify maternal and fetal genetic variants related to blood pressure changes in pregnancy.

Aims: This systematic review aims to determine the prevalence of combined parental depression during the prenatal period and first year post-partum

Published and unpublished observational studies (cohort and cross-sectional studies) were included in the investigation if they reported the prevalence of combined parental depression during pregnancy and/or the first year postpartum. Studies were considered if they were published between January 1980 and December 2011.

Parents aged 15 to 50 years old, who had a primary diagnosis of depressive disorder according to DSM or ICD criteria assessed by a clinical interview, or who scored as depressed on a validated questionnaire, during pregnancy or within 1 year of the birth of their child were included.

Outcome measures

1. Prevalence of combined parental depressive disorder according to DSM or ICD criteria

2. Prevalence of heightened risk of depression according to a validated questionnaire (e.g. Edinburgh Postnatal Depression Scale or Beck Depression Inventory) simultaneously in both parents

Potential confounding variables:

Parental age/SES/eduction, study quality/calendar year/locality, measure of depression

As for ALSPAC data specifically, I would be interested in data on the number of couples assessed prenatally (at 18 weeks gestation), and postnatally (8 weeks and 8 months postpartum) in which both the mother and her partner scored more than 12 on the Edinburgh Postnatal Depression Scale (EPDS). Also, for reference, I would like to know the total number of couples assessed at these time points and numbers of mothers and partners individually scoring greater than 12 on the EPDS.

Aims

Human phenotypes exhibit significant levels of variation among individuals. Recently, genome-wide association studies (GWAS) have been performed to identify the genetic loci that control these variations. Despite the success of GWAS to identify a set of genetic loci, it has become widely recognized that these loci cannot fully explain the variance of a phenotype. One possible explaination is that there is phenotypic stochastic noise which has not been taken into account in the typical GWAS. Normally, only one measurement of phenotype from each individual has been used in the typical GWAS which could be considered as the average phenotype of each individual. However, if we measure one phenotype several times in one individual, we will find that in some people the phenotype is relatively stable while in other people the phenotype is relatively variable. In other words, people show different levels of phenotypic robustness (canalization). The genetic control on phenotypic robustness has been reported in plants, yeast and mice. However, there is still scarce data in human.

Hypothesis

We hypothesized that genetic components controlling phenotype robustness may exist in humans. Since typical GWAS's focus on the genetic variation impact upon average phenotypes, the findings from genetic control on phenotype stochastic noise would improve our understanding on the genotype-phenotype relationship. In TwinsUK dataset, we observed different levels of phenotypic robustness on glucose and lipids across multiple measures on each individual. In the initial GWA analysis on glucose level robustness, we found a trend of association on several genetic loci across the genome. In order to verify our finding, we will request to use the ALSPAC data to perfrom GWA on glucose level robustness. Then we will perform meta-analysis using GWAS results from TwinsUK and ALSPAC. We will carry out genome-wide meta-analysis on lipids levels as well.

Request Data

Exposure variables

The genome-wide genotypes are considered as exposure variables. Both genotyped and imputed SNPs will be used for analysis. We suggest to employ the following threshold for quality control of the SNPs:

a. Call rate greater than = 95%

b. Minor Allele Frequency greater than =1%

c. For the imputed SNPs, MACH r2greater than =3 or IMPUTE prop infogreater than =0.4

d.HWE pvalue greater than 10e-6

Outcome variables

Glucose levels and lipids levels (including HDL/LDL cholesterol, triglyceride) will be used to generate outcome variables. The individuals with greater than =2 measurements will be included in the analysis. The standard deviation of multiple measurements of phenotypic levels in each individual will be used as outcome variables. Age of phenotype collection will be included in the analysis model as covariate.

We are requesting access to the age data on these variables at age 17-19 in order to undertake initial descriptive analyses of the all of outcomes at this age, followed by more complex analyses exploring the extent to which and ways in which these outcomes are associated with clusters of risk behaviours at age 15/16 identified using latent class analysis. Adjustment for missing data will be undertaken using multiple imputation techniques. The work is being undertaken as part of the DECIPHer programme of work on multiple risk behaviours in adolescence. The work is directed by a multi disciplinary group of researchers, led by Prof Rona Campbell and Prof Matt Hickman, with involvement from Dr Ruth Kipping, Dr Jon Heron and Dr Michele Smith. Dr Michele Smith will undertake the descriptive analysis and Dr Jon Heron will undertake the latent class analysis. Michele Smith is supported by Jon Heron and hence there is no need to incur the time (and therefore costs) of a data buddy to facilitate the use of the data.