From alcohol and deodorants, to the contraceptive pill and HRT: what are the real risks for breast cancer?
By Guest Blogger Mr Simon Walsh, Consultant Breast Surgeon...
When I started my medical training, in Cambridge in 1982, we were told a woman’s lifetime risk of developing breast cancer was 1 in 14. In 2017 that risk is 1 in 9. And this is not just due to increased awareness of the disease or breast screening - it is a real increase. So, what has changed? Why is breast cancer nearly twice as common as it was 35 years ago?
What do we mean by ‘risk’?
To make sense of some of the factors that might make breast cancer more common, we need to first understand what “risk” means. Once we have done that we can consider things that we know affect the risk, things that might affect and those things that probably don’t. And despite all that, we need to remember that knowing what affects breast cancer risk tells us nothing at about the fundamental causes of breast cancer at all!
There are two basic types of risk to consider: absolute and relative.
Absolute risk tells you the overall likelihood of something happening. This is the equivalent of the “1 in 14” or the “1 in 9” figures I mentioned earlier. What it doesn’t do is to tell you whether something will actually happen, or not. The other type of risk, much beloved by the tabloid press, is the relative risk. This tells you how much more likely something is to happen in one group compared with another. It says nothing about the likelihood of something actually happening or not, so it tells you nothing about the absolute risk. This type of risk allows the newspapers to publish headlines such as: “HRT can double the risk of breast cancer”, and “One alcoholic drink a day raises breast cancer risk”. This type of headline can be really misleading, and the following example will help to illustrate this.
After the Fukushima nuclear accident, The Guardian reported that there had been a 70% increase in cancers. That sounds like a huge amount. Conversely, the Wall Street Journal reported a 0.5% rise. So, who is right? Both are! The relative risk rose from 0.77% before the nuclear accident to 1.29% after it (a 70% increase), but the absolute risk (a much more meaningful figure in reality) was up by only 0.5%, reflecting that these cancers are normally relatively rare. What was really interesting was that both papers changed their headlines on seeing the report in the other!
Now, let’s turn to breast cancer and start by considering the things we know increase the risk of breast cancer.
Risk factors for breast cancer
Somewhat obviously, the biggest risk factor to developing breast cancer is being a woman. The relative risk compared with being a man is 200. That means that around 0.5% of breast cancers occur in men.
We have already seen that the lifetime risk for a woman is 1 in 9, but age also plays an important part, with breast cancer being more common in older women. For example, there is only a 1 in 15000 chance of developing breast cancer before the age of 25 - this rises to 1 in 200 before the age of 40 and 1 in 50 before the age of 50. This analysis shows us that breast cancer under the age of 40 is not particularly common. Essentially, breast cancer could be considered to be a disease of older women.
Family history and genetics is the next area to consider. We know that 5% of breast cancers are “hereditary”, and in these cases the family history is very strong, often with many family members developing the disease under the age of 40. There may also be ovarian cancer in the family. We know that mutations in two genes (BRCA1 and BRCA2) account for most of these cases. Although the absolute risk of developing breast cancer with a faulty BRCA1 or BRCA2 gene may be as high as 80%, it’s not 100% - this shows us that something else needs to happen before a breast cancer can start. We don’t yet know what that something else is. We know that 10% of breast cancers are “familial”; several breast cancers have occurred in the same family over the generations, but there is no known genetic abnormality. Whilst, in some cases, this may reflect the fact that there are other genes yet to be discovered, it also suggests that environmental factors have a role to play.
People brought up in the same geographical area, who eat the same foods and are exposed to the same environment may all be exposed to, as yet unknown, factors that can affect their risk. This potential environmental effect is well seen in the difference between the incidence of breast cancer in different countries. In Japan, breast cancer is much less common, but stomach cancer is much more common. If Japanese women move to western countries the second generation of women have low stomach cancer rates and higher breast cancer ones. Since both Japan and the UK have things like nuclear power stations, vehicle pollution, mobile ‘phones and overhead power lines, it may be suggested that the major difference is in the diets in the two countries.
The next group of risk factors to consider is “ovulatory”, or “hormonal” factors. We know that if a woman’s periods start early (e.g. aged 10) and finish late (e.g. aged 55) then their relative risk is increased. If they have had no children, or were older when they had their first child, this relative risk increases a bit more. On the other hand, having your first child in your late teens and breast-feeding reduces the relative risk. All of these factors seem to relate to the number of normal ovulatory cycles a woman has in her lifetime. Periods starting early and finishing late, with no children, means more cycles and a higher relative risk. Obviously a woman does not have periods when she is pregnant and in most cases they do not occur when breast feeding, resulting in a lower relative risk. In fact, whilst breast feeding does seem to have a small effect on reducing breast cancer risk, this is only in young onset breast cancer, which is rare anyway.
Hormone-replacement therapy (HRT)
This leads us onto HRT as a possible risk a factor (much beloved by the tabloid press). The first important thing to know is that the average age of menopause in this country is 51, so taking HRT below this age (for example if a woman has had her ovaries removed or simply has an early menopause) does not affect the relative risk. The relative risk for those women over the age of 51 who are taking HRT, or within 5 years of stopping it, is 1.023 for each year of use, which doesn’t sound nearly as bad as “double”! In fact this extra risk is exactly the same as having a delayed menopause by the same number of years. This effect is only seen with “combined” HRT, which contains both oestrogen and progesterone. If a woman has had a hysterectomy then she can have oestrogen-only HRT and this may actually decrease the relative risk of breast cancer. Ultimately, whether to take HRT or not is a choice for each individual woman, many of whom will find that the improved quality of life they experience outweighs the potential risks.
Whilst for HRT there does seem to be a definite, although small, effect, for the oral contraceptive pill the risk is not so clear-cut. Some studies have suggested a small increase in breast cancer while a woman is on the pill, while others do not show an increased risk. And remember, breast cancer is very rare in younger women, so any effect will be very difficult to prove. Having said that, the pill may actually protect against some other cancers, such as that of the ovary of the uterus.
Whilst any type of x-ray could potentially cause a breast cancer, the risk for a standard chest x-ray and, particularly, a modern digital mammogram, is very low. Mammograms are deliberately very lose dose to make having them regularly as safe as possible. A CT scan covering the chest, abdomen and pelvis gives around 250 times the x-ray dose of a mammogram, for example. Women who have had breast cancer treated by breast conserving surgery (a “lumpectomy” or “wide local excision”) will also nearly all have radiotherapy to the breast as part of their treatment. In theory this could cause a new cancer in another part of the breast, but this risk is hugely outweighed by the reduction in the risk of recurrence of their original cancer that radiotherapy produces.
Lymphoma tends to occur in younger people. Part of the treatment for lymphoma involves radiotherapy and if the lymph glands in the chest are affected then radiotherapy must be given to the chest area. In women this will mean that the breasts are subjected to high doses of radiotherapy. Whilst lymphoma is often curable, following such treatment we know that the relative risk of a woman getting breast cancer after ‘mantle radiotherapy’ is around 3 times higher. This risk is enough that these women are now called for early breast screening, under the NHS Breast Screening Program.
Now we move onto the three things that might be responsible for much of the increase we see in breast cancer over the last three decades; smoking, drinking and obesity. Smoking, particularly if a woman started smoking before the age of 20, increases the relative risk by 1.5 and there is a similar increase for regular drinking (albeit on the heavy side). Young women did both of these things increasingly in the 1990s and it may be that this is feeding in to the rates of breast cancer we are now seeing. Similarly, we are all getting fatter, and having a BMI above 35 at the menopause can double the relative risk of developing breast cancer. The relationship between obesity and breast cancer probably relates to the fact that after the menopause (when the ovaries stop making oestrogen) oestrogen is still produced in fatty tissue, so if you are fatter you will have relatively higher levels of oestrogen. This is similar to the increase seen with combined HRT (but, interestingly not oestrogen-only HRT). As smoking, drinking and obesity are all lifestyle factors which we can control, it is encouraging to think that by recognising these risks we can do something about them.
Pre-cancerous changes in the breast
We also know that there are some changes that we see in the breast tissue that might, if not treated, lead to cancer in future. Ductal carcinoma in situ (DCIS) is the main example of this type of condition, but there are other changes that may also be on the rungs of the ladder that leads to breast cancer (albeit much lower down). DCIS is usually found on a routine mammogram as tiny areas of calcification (called micro calcification). Other sorts of calcification, that are more common, can also occur and most calcifications are completely harmless. DCIS comes in three main types. Almost everyone agrees that high grade DCIS will develop into cancer at some point (perhaps in a few years) and should be treated. Most people think that intermediate grade DCIS will also become cancerous, but over a much longer time period. There is also low grade DCIS, and we are not sure whether this will ever develop into cancer or not. There is currently a large trial being carried out to see whether this type of DCIS needs treating at all.
Other medical conditions
Other medical conditions that seem to be related to a slight increase in the relative risk of breast cancer include diabetes, high blood pressure and thyroid problems. Now, it could be argued that people with these conditions tend to be overweight which would increase their risk, but it probably isn’t that straightforward and there may be other hormonal interactions occurring that we do not, yet, understand.
There have also been some things published in the media as increasing the risk of breast cancer, which actually don’t. Lack of sleep and shift working had been suggested as increasing the relative risk, but recent data suggest that this is not the case. Similarly, there is no evidence that underwired coffee, underwired bras, hair dyes or overhead powerlines increase the risk. The risk from deodorants was much publicised a few years ago. In instances like this it is often quite informative to read the original scientific paper, rather than rely on a newspaper report. In this case it was observed that most breast cancers occur in, what is termed, the “upper outer quadrant”, towards the armpit. A tentative association was then postulated with the use of deodorants in the armpit, but in fact no evidence was presented and no mechanism of how this could cause cancer was given. In fact, the reason that most breast cancers occur in this part of the breast is, quite simply, because this is where most of the breast gland tissue is (the rest of the breast being mostly fatty tissue). When I was training, it was a requirement that a surgeon spent 2 years or so in research before becoming a consultant. I spent 2 years studying the molecular biology of breast cancer and the knowledge and techniques I acquired during that time have been invaluable in helping to unravel potentially misleading science.
Despite all the information we now have about the risks for breast cancer, the cause remains obscure. And perhaps that is the problem; there is no single cause. It seems that for a breast cancer to develop, many different things have to occur in a particular order and these can be genetic, hormonal, environmental or completely unknown. Looking for “the cause” may be fruitless. Similarly, breast cancer is a collection of many diseases that happen to occur in the breast. Whilst the media (and some scientific publications) would have us believe there are ten different kinds, the truth is that there are as many different types as people who get it. Each cancer is a unique combination not only of changes with a cell but also of how an individual’s immune system responds to the changes. There is theory that suggests we are all developing potentially cancerous cells all the time, but in the vast majority of cases our own immune system recognises the abnormal cells and stops them developing into a cancer.
We still have much to learn. But in the meantime we should concentrate on what we know can reduce the risk. So, simply put, to reduce the risk of breast cancer (and almost every other disease!): don’t smoke, don’t drink (too much) and don’t be overweight!
Anderson, K.N et al. Reproductive risk factors and breast cancer subtypes: a review of the literature. Breast Cancer Res Treat (2014) 144: 1. doi:10.1007/s10549-014-2852-7
Dumitrescu RG et al. Understanding breast cancer risk - where do we stand in 2005? Journal of Cellular and Molecular Medicine. January 2005. DOI: 10.1111/j.1582-4934.2005.tb0035
Ferlay J et al. Global burden of breast cancer. Breast Cancer Epidemiology: 1-19. 2009
Gierisch JM et al. Oral Contraceptive Use and Risk of Breast, Cervical, Colorectal, and Endometrial Cancers: A Systematic Review. Cancer Epidemiology Biomarkers and Prevention. DOI: 10.1158/1055-9965.EPI-13-0298
Hardefeldt PJ et al. Deodorant use and breast cancer risk.. Epidemiology. 2013 Jan;24(1):172
Hardefeldt PJ et al. Diabetes increases the risk of breast cancer: a meta-analysis. Endocr Relat Cancer. 2012 Nov 19;19(6):793-803. doi: 10.1530/ERC-12-0242. Print 2012 Dec.
Kamdar, B.B et al. Night-shift work and risk of breast cancer: a systematic review and meta-analysis. Breast Cancer Res Treat (2013) 138: 291. doi:10.1007/s10549-013-2433-1
Key TJ et al. Alcohol, diet, and risk of breast cancer. BMJ 2016; 353 doi: https://doi.org/10.1136/bmj.i2503
Lu Chen et al. Bra Wearing Not Associated with Breast Cancer Risk: A Population-Based Case–Control Study. Cancer Epidemiology Biomarkers and Prevention. DOI: 10.1158/1055-9965.EPI-14-0414 Published October 2014
King et al. Breast and Ovarian Cancer Risks Due to Inherited Mutations in BRCA1 and BRCA2. Science 24 Oct 2003: 302(5645): 643-646
McPherson K et al. Breast cancer – epidemiology, risk factors and genetics. In: ABC of Breast Diseases. Ed. Dixon MJ. BMJ Books
Naro SA. Hormone replacement therapy and the risk of breast cancer. Nature Reviews Clinical Oncology 8, 669-676 (November 011) doi:10.1038/nrclinonc.2011.110
Pinder SE et al. The diagnosis and management of pre-invasive breast disease: Ductal carcinoma in situ (DCIS) and atypical ductal hyperplasia (ADH) – current definitions and classification. Breast Cancer Research 20035:254
Swerdlow J et al. Breast Cancer Risk After Supradiaphragmatic Radiotherapy for Hodgkin’s Lymphoma in England and Wales: A National Cohort Study. Journal of Clinical Oncology 30(22). 2012
Network Locum (now Lantum) is a GP app and website for connecting directly with practices old and new, and managing all session related admin.
We work alongside the NHS who recognise our tech as a solution to efficiently connecting supply and demand in primary care. We are a GP-led health tech start-up whose mission is to support and empower GPs throughout their profession. Join us today.