In this post I discuss a recently published paper on the possible origins of cancer.  

book cancerBased on Cancer research UK’s 2010 statistics, one in three of us will develop a cancer over our lifetime.

Here’s what most of us know about cancer:

  • By and large, cancer is caused by ‘bad’ habits: smoking, alcohol, even unprotected sex, as Michael Douglas so tactfully brought to our attention recently.
  • We also know that being subjected to toxins, like asbestos or other poisons in our environment, in our food etc can cause a cancer to develop.
  • If family members have had cancer, we know that our risk of developing one increases.

But then, we also have examples in our day-to-day lives where people have developed cancer, and none of the above applies. Cancer has strong symbolic representations for many of us (the big “C”) and is riddled with some rather unpleasant connotations: pain, suffering and death. Basically it is perceived as a nasty piece of work – and with good reason.

In a recently published open-access paper, my colleagues and I conclude that stressful events occurring in childhood are associated with cancer in adulthood – even after taking into account lots of traditional cancer risk factors, and other variables, like social class (Kelly-Irving et al. 2013a).

This is by no means a claim of undeniable cause-and-effect, but rather a modest proposal: that new parameters in understanding how and why cancer occurs are worthy of investigation. Below, I discuss the work itself and some issues encountered. 

Can a different framework be applied to research on cancer?

We decided to study cancer within a lifecourse framework – the tenets of which I have described in a previous post. We tested whether psychosocial conditions early in life may be  linked to the development of an adult cancer, via 1) a biological pathway (alterations to physiology), 2) a socioeconomic pathway (ex: environmental and occupational exposures) and 3) a behavioural pathway (ex: health behaviours). This oversimplified model helps us to understand how these different conditions and events are potentially linked together over time, and associated with a cancer in adult life.

…but cancer is not one disease

Our work was met with ambivalence and scepticism by some researchers. The main reason for this is that cancer is considered to be not one, but many diseases referred to under one umbrella term. The mutations in a cell that cause it to become cancerous are varied: from virus induced cancers, to hormonal cancers, and cancers caused by exposures to toxins. Therefore it is impossible to study them all together – it makes no biological sense.

To this argument, we answer that cancers may indeed be aetiologically different, but the mechanism of cancerogenesis is shared regardless of the trigger (virus, toxin etc.). In other words, the failure of the immune system to kill a cancer cell is common to all cancer types. It is that particular biological common stem that we are initially interested in.

Our research question:

Are conditions occurring in early life linked to developing cancer in adulthood?

We know that conditions (socioeconomic, nutritional, psychosocial etc.) in early life have been associated with development of other chronic diseases in adulthood, such as diabetes, cardiovascular disease, poor lung function and even death.  Is it also relevant to think about cancer in this way?

Stress and cancer: plausible evidence?

In another paper published previously, we reviewed the evidence from the animal literature where experiments have been carried out to see whether physiological stress is related to the development of cancer (Kelly-Irving et al. 2013b). From the literature we identified plausible physiological pathways linking stress responses to the development and proliferation of cancer: neuroendocrine responses and epigenetic responses being the two main groups.

We also discussed and highlighted why events and conditions causing stress earlier in life during sensitive periods of development may be more likely to lead to permanent biological alterations.

It is true to state though, that epidemiological studies on associations between stress and cancer show a mixed bag of results, and few have managed to look at early life stressful experiences and cancer.

Adverse childhood experiences (ACEs)


Most often in the literature, ACEs are measured retrospectively. In other words, adults are asked about trauma, violence (physical & sexual), harassment, witnessing violence, and living with family members with mental health problems growing up. Much of the work on the topic of ACEs comes from the Adverse Childhood Experiences study based in San Diego (Felitti, 1998).

Using data collected prospectively, throughout childhood, we were able to create a different type of measure of ACEs based on the following definition:

intra-familial events or conditions in the child’s immediate environment causing chronic stress responses.” (Kelly-Irving et al. 2013a)

We distinguish between ACEs and conditions linked to the socioeconomic and material environment, to which they are closely linked. We used information collected from parents and teachers during childhood on events and conditions likely to have been stressful to the child like: being separated from their parents; being put in a care home; living with a mentally-ill parent etc (Kelly-Irving et al. 2013c).

What did we find?

We used a well-known cohort study, the 1958 birth cohort study, to look at the link between these adverse experiences and cancer later in adulthood. Our measure of cancer was less than perfect: we relied on people reporting whether they had been diagnosed with a cancer.

We found that among women, those who had experienced two or more adversities in childhood had a higher risk of developing cancer in adulthood compared to those who had not experienced any of the defined adverse events. In fact, the risk of having a cancer before the age of 50 years increased twofold for those who had accumulated two or more of these adverse events compared to those who had not experienced any. These findings were adjusted for the typical behavioural risk factors, like smoking and drinking in adulthood, but also for education level, anxiety and depression, and age at first birth.

Among men the associations observed were not statistically significant. This could be due to a lack of association, but in this study few men had reported having a cancer before 50, and the lack of significance could also be due to insufficient power.


In our paper we go into the limitations of the work at length. Here are a few of the more salient points:

  • Yes, the cancer variable used is self-reported – and so inadequate compared to a doctor’s diagnosis or a register notification. But, though people are often unreliable at reporting the type of cancer they have had, they are reliable at reporting having had cancer overall.
  • Yes, the time lapse between the childhood events and adult cancer is long, so many other elements may have intervened in between to affect the development of cancer. But these ‘other intervening elements’ are either: independently linked to the cancer, and therefore the association between ACEs and cancer remains – OR – they are operating along a pathway where adverse experiences lead individuals towards these factors that ultimately cause cancer.
  • Yes, there may be a biological, social or other factor upstream of adversity that explains the observed association. We know that mere statistical adjustment is insufficient in establishing whether or not an association is ‘real’. Carrying out analyses in different studies is key.
  • Are we saying that all children who have had bad experiences are going to develop cancer? No, of course not. We are saying that chronic stress in childhood may lead to a heightened risk of cancer on average at the population level.

So what?

Simply put, we should not restrict the ‘causes’ of cancer to the traditional risk factors I mentioned at the beginning. A susceptibility to cancer probably originates well before we have ever been exposed to such factors. Maybe that a lifecourse framework is worth using to study cancer further.

From this we might hypothesise that stressful events leading to certain biological responses, though they might be adaptive in the short-term, could be harmful if they last. They may result in biological changes that lead to a greater susceptibility to diseases like cancer. The findings also point towards childhood and adolescence as being times where such physiological changes might have a strong impact on our biological make-up.

These finding also lead us to think about cancer differently. To think about the causes of the causes – why are some people more likely to take-up behaviours that are bad for their health? Why are some people more likely to be exposed to dangerous toxins?

What do we do now?

This work contributes to a growing literature pointing to early life as a key period where positive environments have a hugely important impact in constructing the future, and negative ones can leave a lasting mark well into the future. It is never too early, and it is never too late to act to improve health and wellbeing. But understanding the origins of diseases could change how we attempt to prevent them.


What we can do is decide as a society to invest in early life and childhood. I mean make a real decision, not one determined by political office. Doing so is likely to improve the childhoods of many, but importantly, it will have a lever-arm effect on future wellbeing, health and physical functioning.

The ageing process starts from conception. Early experiences are strong determinants of how that ageing process is going to play-out.

By Michelle Kelly-Irving


Felitti, V. J., R. F. Anda, et al. (1998). “Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study.” Am J Prev Med 14(4): 245-258.

Kelly-Irving, M., B. Lepage, et al. (2013a). “Childhood adversity as a risk for cancer: findings from the 1958 British birth cohort study.” BMC Public Health 13(1): 767.

Kelly-Irving, M., L. Mabile, et al. (2013b). “The embodiment of adverse childhood experiences and cancer development: potential biological mechanisms and pathways across the life course.” International Journal of Public Health 58(1): 3-11.

Kelly-Irving, M., B. Lepage, et al. (2013c). “Adverse childhood experiences and premature all-cause mortality.” European Journal of Epidemiology: 1-14.