Thyroid Function and Body Composition in Children: cause or effect? A study using Mendelian Randomization
Dr Peter Taylor BSc MBChB MRCP SCE (Diab/Endo), Centre for Endocrine and Diabetes and Sciences, University Hospital Wales
Final Report Summary
View the full report (PDF) here.
We were very grateful to receive the £10,000 research award from the British Thyroid Research Foundation in November 2011. This enabled us to perform 1,200 thyroid function tests in children in the Avon Longitudinal Study of Parents and Children (ALSPAC) from stored samples taken at ages 7 and 15.
ALSPAC is regarded as Europe's premier observational study in children, and to our knowledge, it is the only cohort in the world, which has stored blood samples from early childhood and serial detailed measurements of body composition such as height and fat mass as well as key measurements of bone development throughout childhood. ALSPAC also has detailed information regarding the children's genetic make-up and their environment (e.g. social class). Overall this project provided us with the unique opportunity to study the relationships between thyroid hormone levels, body composition and bone development at different stages of childhood development. A greater understanding of the impact of thyroid horomone on both body composition and bone architecture is essential in understanding future causes of adult ill health such as obesity, development of the metabolic syndrome and osteoporosis.
Our results showed that there were key changes in one of the thyroid hormones, free tri-iodothyronine (FT3) over childhood, which varied considerably over time in the same individual, unlike free thyroxine (FT4) and thyroid stimulating hormone (TSH) which were more stable. Younger children also appear to have substantially higher levels of FT3 than adults, which merits further study.
Only FT3 was substantially associated with body composition and bone development in childhood. Surprisingly FT3 was positively associated with fat mass, i.e. children with higher levels of fat mass had higher levels of FT3; which is perhaps counter intuitive given that higher levels of thyroid hormones are usually associated with weight loss. Statistical and simple genetic analyses raised the possibility that fat mass was involved in the conversion of thyroid hormone; further genetic analyses and laboratory studies are required to clarify this. This relationship between fat mass and FT3 is integral to the relationship between FT3 and bone as the association between FT3 and bone was lost after accounting for an individual's fat mass.
Furthermore we also found preliminary evidence that the genes which influence thyroid function are different between children and adults. This has two fold relevance i) Genetic association studies of thyroid function are becoming increasingly common and children and adults should potentially be analysed separately. ii) It also has important implications for our understanding of the role of thyroid hormone in childhood development, as thyroid hormone may influence developmental processes differently over childhood.
Overall, by following the same children from age 7 to age 15 through the stored blood samples from the ALSPAC cohort, we have shown that FT3 levels are higher in pre-pubertal children and fall as the children approach adulthood. The higher FT3 levels were surprisingly associated with greater fat mass and appeared to be controlled by different genes from those that play a role in older children. Taken together, this is the first large prospective cohort study to indicate that FT3 may play a different role in children and may be linked to changes in body composition that are relevant later on in life.
We were very grateful to receive the £10,000 research award from the British Thyroid Foundation in November 2011. This enabled us to perform additional full thyroid function in children in the Avon Longitudinal Study of Parents and Children (ALSPAC) from stored samples taken at ages 7 and 15.
ALSPAC is regarded as Europe’s premier birth cohort, and to our knowledge is the only cohort in the world which has stored blood samples from early childhood and serial detailed measurements of body composition such as height and fat mass as well as key measurements of bone architecture throughout childhood. This, combined with its extensive detailed information regarding their genetic make-up and their environment (e.g. social class) has provided us with the unique opportunity to study the relationships between thyroid hormone levels and body composition and bone development at different stages of childhood development.
A greater understanding of the determinants of both body composition and bone architecture is essential in understanding future causes of adult ill health such as obesity, development of the metabolic syndrome and osteoporosis. By having thyroid function at age 7 and 15 we were also able to clarify when thyroid hormones influence body and bone development, but also assess the degree of variation in thyroid hormone levels over childhood in the same individual. In addition we explored whether reference ranges widely used were appropriate at different stages in childhood.
Our results showed there was very strong evidence that only one of the thyroid hormones (T3) was associated with both body composition and bone development with no evidence for an association with (T4). Surprisingly T3 was positively associated with fat mass, i.e. children with higher levels of fat mass had higher levels of T3 which is perhaps counter intuitive especially as this persisted after adjusting for height. This raises the possibility that the higher T3 levels are a response to higher levels of adiposity which may be a compensatory mechanism, although the possibility remains that the association between free T3 is different between children and adults and we will be exploring this using genetic studies in the second half of this project. It also raises questions regarding the appropriateness of T4 only replacement which will lower the T3 : T4 ratio.
With regard to bone development, higher levels of free T3 at age 7 were associated with increased bone mass and bone area. Interestingly this effect appeared to be mediated through changes in height and fat mass. The greatest effect was in weight bearing bones particularly the legs. 15.3% of children age 7 had a higher free T3 than the reference range, given that this had the strongest association with both body composition and bone architecture, it is important to consider whether the current reference range is appropriate for younger children.
Finally we also assessed the longitudinal stability of free T3 and free T4. Whilst free T4 was tightly correlated between ages 7 and 15, free T3 was not. Furthermore we also found preliminary evidence that the genetic architecture of thyroid function is different between children and adults. This will be the focus of further work as part of my PhD and the latter stage of this project. This has relevance two-fold: i) separate genetic association studies may be required for children and ii) it may explain why thyroid hormone encourages bone development in children, but is associated with reduced bone mineral density and osteoporosis in adults.
Personal Statement from Dr Taylor
My interest in thyroid disease started in 2006 when I worked as an SHO in Diabetes and Endocrinology at the Edinburgh Royal Infirmary under Dr Toft. I had one of those wonderful “light bulb” moments in medicine, during a very late running winter clinic. Every patient I had seen had had very different symptoms, ranging from worsening mood and energy levels to difficulties with their weight and bowels. Many patients also commented that they still felt “not quite right” despite having hormone levels that were in the “normal range”. Given that the thyroid also influences hidden aspects of health including osteoporosis risk (bone thickness) and cholesterol levels I wanted to explore the health impact of modest variation in thyroid hormone levels and if it could be better treated.
As a result of this, I decided to undertake a career in academic medicine specializing in thyroid based research - a far cry from my childhood ambitions of being an astronaut. I relocated to the South West and every colleague I discussed my research career with recommended I approach Professor Dayan. Through him I forged links with the Avon Longitudinal Study of Parents and Children (ALSPAC) which has detailed information on almost 15,000 children and their parents. They even have a detailed collection of the children’s teeth and hair. These children have been repeatedly interviewed, prodded, scanned and even had detailed studies undertaken of their DNA.
These detailed studies mean I have the opportunity to explore the differences in the genes and DNA between people that determine a person’s thyroid hormone levels and how in turn this influences their health. Most excitingly we may be able to use genetic techniques to study the intracellular thyroid hormone pathway (inside cells) which is not possible with traditional blood tests. There is increasing evidence that this intracellular pathway may influence mood, osteoarthritis and responsiveness to thyroid hormone replacement therapy. It’s a project that will certainly keep me busy and more importantly keep my low boredom threshold at bay (I still read books while cleaning my teeth – and yes it’s definitely easier with a Kindle).
Ultimately from our research we propose to identify individuals who might benefit from thyroid hormone supplementation based on both their serum hormone levels and their genetic profile. When I’m not researching into these genetic polymorphisms, or writing research grants I can generally be found worrying about the genetic determinants of the two illnesses that run in my family (diabetes and supporting Man City) the last is sadly incurable I fear, but at least of late it has not been too bad, bar the actions of a certain Mr. C. Tevez.
Once again a very big thanks to everyone in the British Thyroid Foundation for this support and if anyone has a good recommendation for getting toothpaste off a Kindle screen please do let me know.