Intended for healthcare professionals


Height, early energy intake, and cancer

BMJ 1998; 317 doi: (Published 14 November 1998) Cite this as: BMJ 1998;317:1331

Evidence mounts for the relation of energy intake to adult malignancies

  1. Demetrius Albanes, Senior investigator
  1. Cancer Prevention Studies Branch, Division of Clinical Sciences, National Cancer Institute, 321, Bethesda, Maryland, 20892-7058, USA

    Papers pp 1350, 1351

    Study of the relation between human body size and cancer risk has its origins in both human observational studies and animal experiments. In his late 19th century treatise Air, Food, and Exercises based on observations in England and continental Europe Rabagliatti concluded, “Overfeeding is the predisposing cause of cancer.”1Since then the restriction of total energy intake has been shown over decades of research to be one of the most powerful nutritional interventions capable of lowering cancer rates in laboratory animals—and the only factor known actually to increase lifespan in these same systems. 2 3Such rodent experiments resulted in lighter, leaner animals—that is, with decreased body weight and adiposity—and in many instances, depending on the timing of the energy restriction, stunted growth and shortened overall animal length. Although lifelong dietary restriction has had the greatest impact, reduced food consumption early during development also effectively inhibits tumorigenesis.3Human data are available from epidemiological studies of the relation between energy intake, body size, and cancer rates or risk. In these, overweight and obesity are associated with higher overall incidences of cancer4and, in particular, with cancer of the breast (after the menopause), prostate, large bowel, endometrium, ovary—that is, the major non-smoking related malignancies—and kidney.5

    A growing number of investigations have also established greater adult height as a risk factor for several of these same cancers. Data provided by two reports in this week's issue corroborate the previously observed direct relation between height, in this instance assessed as adult stature or childhood leg length, and the development of cancer. 6 7Using nearly 30 years of follow up of the Whitehall study of London civil servants, and based on more than 2000 cancers, Davey Smith et al observed 36% more non-smoking related malignancies among men taller than 6 ft (1.8 m) than in men 5 ft 6 in (1.65 m) or shorter, with men of intermediate height having intermediate risk (p 1351).6Similarly, Gunnell et al evaluated prepubertal height and leg length in the Boyd Orr cohort and found that during a 50 year follow up nearly 80% more deaths from cancers unrelated to smoking occurred for every 3-4 mm increment (age and trunk length corrected) in leg length (p 1350).7Underlying both these prospective investigations is the view that height is a proxy for early nutrition (particularly of energy intake)and childhood growth and development, with taller people more likely to have been exposed to a greater surfeit of dietary energy during maturation than shorter ones, notwithstanding the influence of heredity. This interpretation is bolstered bythe fact that leg length may be the more nutritionally sensitive component of height.

    Two other aspects of the studies deserve mention. Firstly, the relation between height and cancer was not evident for organ sites associated with tobacco smoking, a finding consistent with most previous studies, suggesting that the carcinogenic effects of smoking can overshadow those from other behavioural exposures such as early nutrition. Secondly, statistical adjustment was made for socioeconomic status, which greatly improves the likelihood that the observed associations reflect an independent effect of height and its other determinants, and are not confounded by other socioeconomic correlates—or consequences—of height.

    Although controlled trials of early energy intake are not possible, natural experiments have occurred. For example, women subjected during puberty to wartime food deprivation experienced reduced lifetime breast cancer rates compared with younger and older cohorts.8Furthermore, direct evidence for the role of childhood energy intake itself in human cancer risk was reported earlier this year from one of these cohorts, with a 20% higher rate of cancers not related to smoking observed for every megajoule increase in total energy intake.9Taken together, the human data are consistent with the laboratory experiments and suggest that excess energy intake relative to requirements, manifest as increased height or early maturation (from early excess) or overweight (from cumulative excess), has negative consequences with regard to several major cancers.

    Elucidation of the biological basis for these observations will both provide the plausibility needed to impute causality and improve our understanding of how nutrition affects the malignant transformation of cells. Candidate mechanisms include greater exposure to mitogenic factors such as growth hormone, insulin, insulin-like growth factors, and sex steroids that could result from accelerated maturation mediated by the hypothalamic-pituitary axis. Positive energy balance during development may led to a higher subsequent risk of malignancy not only through increased cumulative cell proliferation but also by expanding organ specific stem cell populations.10Effects of excess energy intake on oxidative stress, detoxification enzyme systems, or immune function are also possible.

    Affluent industrialised populations have advanced beyond concerns about adequate energy intake that originated from uncertain sustenance and periods of starvation. Ironically, the new, highly prevalent threat is posed by overnutrition—or more specifically, energy consumption beyond that expended.11This hazard applies not only to the risk of cancer but also to cardiovascular disease as well as other chronic illnesses (such as type 2 diabetes). Although the risks appear to be understood by most clinical, public health, and research communities, they have yet to be fully acted on. This may be a function of the unpopularity and challenging nature of the inherent proscriptive message—eat less, be less inactive—when compared with the facility and palatability of their alternatives.


    View Abstract