Height: A Universal Cancer Risk Factor?

“…quantification of the risks of increasing height on various cancers may provide information useful for planning for future cancer screening and treatment needs.”

Historically, being taller has been associated with greater status, privilege and better health; however, results from the Million Women Study suggest that one disadvantage of greater stature is an increased risk for almost all common cancers affecting women [1]. In their recent publication, Green and colleagues reported that for every 10 cm increase in height, the risk of incident cancer increased by 16% (95% CI: 1.15–1.18), and their meta-analysis of prospective studies shows a similar effect for total cancer incidence and mortality in women (15% increase per 10 cm increase in height) and men (10% increase per 10 cm increase in height) [1].

An association between cancer occurrence and height has been documented in many previous studies [2], so the overall finding is not surprising. However, no previous studies have had sufficient power to present precise adjusted estimates for so many different cancer sites. The authors were able to consider the relationship between height and cancer incidence in women at 17 cancer sites. Of these, risk increased significantly with increasing height in ten sites; was elevated nonsignificantly in another five; and was essentially null in the remaining two. Risk was not significantly reduced with height for any cancer. Of note, the risk estimates for individual sites were of a very similar magnitude varying between 1.14 and 1.32 per 10 cm increase for those cancer sites with significant associations. The other finding of interest was the effect of smoking status on the height–cancer relationship. The overall association between height and cancers considered unrelated to smoking was the same in current and never smokers. However, the effect of a 10 cm increase in height on smoking-related cancers was approximately 10% greater in never smokers than in current smokers.

The reason for the association between greater height and cancer is not known. In Hill's seminal paper on causal associations, effect size and specificity of association were two criteria he proposed as part of an approach for assessing the likelihood of a causal relationship [3]. The small effect size and lack of specificity (in both cancer type and effect size) for the association between height and cancer might therefore immediately create suspicion. If the association is seen for most common cancers, then is it perhaps due to a common source of bias or fault in study design? This possibility has been discussed at length in previous reviews on this topic [2,4], and most sources of error have been dismissed as unlikely explanations for the relationship. For example, one might consider the possibility that taller people live longer and thus are more likely to eventually be diagnosed with cancer. However, cancers that affect younger people, such as premenopausal breast cancer, have also been consistently associated with greater height [5], and those investigating the association between height and cancer have mostly used analytic strategies to effectively account for the influence of age. If the association is real then the consistency of the effect across cancers suggests a very fundamental influence most likely acting in early life.

So what is the relevance of these findings? While being taller is an extremely prevalent cancer risk factor, modifying height for the purposes of cancer prevention is neither feasible nor desirable. An understanding of the association is, however, valuable in two other respects. First, average heights are continuing to increase in most countries, particularly in those undergoing rapid socioeconomic improvement, so this information may allow some anticipation of future health needs and planning of resource allocation accordingly. Second, an understanding of the reasons for this association may provide valuable insights into carcinogenic processes.

Tallness is, in itself, unlikely to cause cancer but rather be a marker of a causal process or exposure. Possible explanations for the association have been generated from an understanding of established determinants and correlates of attained height. Final adult height is determined by genes as well as a variety of environmental influences that, in essence, determine the quantity and quality of nutrition a child receives [6]. Childhood illness can influence the uptake of nutrients irrespective of other circumstances. Socioeconomic status in childhood is also inextricably linked to attained height, as demonstrated by the marked secular changes in height that have accompanied socioeconomic development worldwide [7–9]. On an individual level, factors such as higher parental levels of education, greater household wealth and lower sibling number have consistently been associated with greater height [10]. So, what in this mix of effects is likely to cause cancer?

“There are many questions about the association between height and cancer that remain unanswered.”

In their comprehensive reviews of this relationship, Gunnell and Okasha and colleagues have presented a number of plausible possibilities [2,4], including close linkage between height-related and cancer-causing genes, amount of exposure to childhood infections and subsequent immune system development, and associations between attained height and sex-hormone levels. Probably the most frequently cited explanation for the link between cancer and height involves IGF-1. IGF-1 is a peptide growth factor that acts as an inhibitor of apoptosis (programmed cell death) and also stimulates cell division (mitosis). Levels in childhood are strongly correlated with height [11] and levels of nutrition. While some prospective studies have found that high IGF-1 levels in adulthood increase the subsequent risk of some cancers [12,13], it is not as clear that adult levels of IGF-1 are strongly related to attained height [14,15]. A number of childhood cancers have also been related to attained height at the time of diagnosis [16,17]. Together these finding suggest that the relationship between height and cancer reflects events that occur early in life. One intriguing proposition is that extra nutrition in early life results in a general increase in cell production (secondary to IGF-1); that is, tall people simply have more cells and therefore have a greater chance of carcinogenic mutations arising in one or more of these extra cells. Taller people conceivably have, on average, greater skin surface area (and therefore more melanocytes), longer colons, larger kidneys and larger thyroids with more cells that could undergo neoplastic change. It is less clear that taller people have larger brains or that taller women have more breast tissue, although it may be that height is correlated with stem cell number rather than organ size per se.

There are several cancer sites where one might expect that height would be inversely associated with cancer. Stomach cancer is one example. Helicobacter pylori is a known causative factor for stomach cancer and Helicobacter infection is associated with many factors that are also strongly associated with lower attained adult height, such as a larger number of siblings and lower socioeconomic status [18]. A few studies have found an inverse relationship between height and stomach cancer [19,20]; however, a larger number of studies, including the Million Women Study [1], have not observed an association, either inverse or positive, between height and stomach cancer. It is possible that different correlates of height have competing effects in stomach cancer and it would of interest to see whether the influence of height varies with the presence or absence of Helicobacter infection (if, indeed, stomach cancer occurs in the absence of Helicobacter infection). It may be that subdividing the data in this way would reveal that increased height can positively influence risk of stomach cancer in much the same way that subdividing the data by smoking status showed an increased risk of smoking-related cancers in never smokers within the Million Women Study. The close relationship that socioeconomic status has with height and with causal factors for some cancers, such as smoking and Helicobacter infection, makes disentangling the important components of the height–cancer association particularly challenging.

There are many questions about the association between height and cancer that remain unanswered. For example, is there a period of growth that is particularly important to the relationship. Do taller people have more stem cells than shorter people? If IGF-1 levels are important, is it levels in childhood or adulthood that are the drivers of cancer? And, ultimately, is there any kind of intervention that can modify this risk in taller people? A series of well-designed, mostly longitudinal studies will be required to answer these questions. In the meantime, perhaps the most important public health implications of the relationship between height and many cancer types are for countries undergoing rapid health transitions. For them, quantification of the risks of increasing height on various cancers may provide information useful for planning for future cancer screening and treatment needs.