Childhood stature and adult cancer risk: the Boyd Orr cohort

Whitley, E. , Martin, R.M., Davey Smith, G., Holly, J.M.P. and Gunnell, D. (2009) Childhood stature and adult cancer risk: the Boyd Orr cohort. Cancer Causes and Control, 20(2), pp. 243-251. (doi: 10.1007/s10552-008-9239-1)

Full text not currently available from Enlighten.


Cancer risk in adulthood may be influenced by aspects of childhood diet. In the absence of direct dietary data, indirect measures of childhood diet and nutritional status, such as anthropometric measurements, may be useful in investigating diet–cancer associations. Previous studies suggest that taller adults may have increased cancer risk. Peak growth for different anthropometric measures occurs at different times and so differential associations with cancer risk may indicate periods of development that are particularly important in determining future risk. 2,642 traced members of the Boyd Orr cohort had measures of foot length, shoulder breadth, height, and leg length made when they were aged 2–14 years; trunk length was derived from the difference between overall height and leg length. Subjects were followed-up over 59 years to determine all-cause (n = 547) and site-specific (n = 97 for lung, 69 breast, 59 colorectal, 33 prostate, 320 not related to smoking) cancer deaths and registrations. There were no strong associations between childhood anthropometric measurements and adult cancer risk. Odds ratios (ORs) were broadly consistent with a slight increase in risk with increasing childhood stature but no single measure was of particular importance. The strongest associations were seen for breast cancer (OR per standard deviation increase in foot length: 1.16 (95% CI: 0.90, 1.51), shoulder breadth: 1.16 (0.91, 1.49) and trunk: 1.26 (1.00, 1.60), and prostate cancer (OR for foot length: 1.22 (0.86, 1.75)). There was little effect of adjustment for confounding factors and very limited evidence that associations differed with measures made prior to the onset of puberty (comparing the associations in children aged <8 vs. 8+ years). There was no evidence that any of the five indicators of childhood growth was more strongly related to cancer risk than the other measures.

Item Type:Articles
Glasgow Author(s) Enlighten ID:Whitley, Dr Elise
Authors: Whitley, E., Martin, R.M., Davey Smith, G., Holly, J.M.P., and Gunnell, D.
College/School:College of Medical Veterinary and Life Sciences > School of Health & Wellbeing > MRC/CSO SPHSU
Journal Name:Cancer Causes and Control
ISSN (Online):1573-7225

University Staff: Request a correction | Enlighten Editors: Update this record