Aims and research questions:

1. What is the dietary intake of vitamin D at 6-7 years or 8-9 years old and how does this relate to serum concentration of 25OHD at 7, 9 or 11 years old?

Question 1. Method: These relationships will be investigated using multivariable regression analyses, with dietary intake of vitamin D as the exposure variable and serum concentration of 25OHD2 and 25OHD3 as outcome variables.

From initial analyses of serum 25(OH)D2 we know that this has a non-normal distribution with a large proportion of participants below the lower limit of detection of the assay and therefore for 25(OH)D2 we will generate a categorical variable consisting of those below the lower limit of detection and thirds of the remainder of the distribution and use ordinal linear regression to examine associations with this outcome. 25(OH)D3 has an approximate normal distribution and linear regression will be used here.

To examine possible effect modification by season we will define season as 'vitamin D winter' (October to March) and 'vitamin D summer' (April to September) as defined by Webb et al (1988) for Edmonton (52 degrees North; Bristol is 51.5 degrees North) and examine associations within each seasonal subgroup. We will formally test for differences between subgroups with a likelihood ratio test.

Potential confounding by age, sex, socioeconomic position, puberty, body composition, physical activity, time outdoors, and other dietary characteristics will be taken into account in the multivariable analyses.

2. What is the relationship between dietary intake of vitamin D at 8-9 and 12-13 years old and DXA, PQCT and fracture variables as outcomes?

Question 2. Method: Multivariate linear regression will be used to explore the relationship between dietary intake of vitamin D with bone quality (DXA and PQCT variables) and fractures as outcome variables.

Potential confounding by age, sex, socioeconomic position, puberty, anthropometry, body composition, physical activity, time outdoors and other dietary characteristics will be taken into account in the multivariable analyses.

1. SACN position statement, 2007

2. Gregory et al (2000) National Diet and Nutrition Survey: Young People Aged 4 to 18 years. London: The Stationary Office.

3. Webb et al (1988) Journal of Clinical Endocrinology & Metabolism, 67: 373-378

4. Ashwell et al (2010), British Journal of Nutrition, 104: 603-611

5. Davies et al (1999) European Journal of Clinical Nutrition, 53: 195-198

6. Bates et al (2003) Osteoporosis International, 14: 152-159

7. Hill et all (2008) British Journal of Nutrition, 99, 1061-1067

8. Cashman (2007), Post Graduate Medical Journal, 83: 230-235

9. Holick (2007) New England Journal of Medicine, 357:266-281

10. Ginty et al (2004) European Journal of Clinical Nutrition, 58, 1257-1265

11. Lehtonen-Veromaa et al (2002) American Journal of Clinical Nutrition, 76: 1446-1453

Cranney et al (2007) Effectiveness and Safety of Vitamin D in Relation to Bone Health. Evidence Report/Technology Assessment No.158. Rockville, MD: Agency for Healthcare Research and Quality (http://www.ahrq.gov/downloads/pub/evidence/pdf/vitamind/vitad.pdf).