The Face of Cancer – A Documentary
Cancer affects genders, all races, rich and poor alike. The diagnosis is feared, as it is assumed (often correctly) to be a death sentence by those afflicted with it. Both the disease itself and its treatment are major causes of pain and distress. Treating cancer is a major burden on healthcare systems worldwide, and the disease is a significant cause of loss of productive capacity within the workforce due to premature death. We will take an overview in our papers about the cancer problem, focusing on some of the more common cancers to illustrate how numbers vary across the world. Any illness affecting so many people will also have major economic impacts, so will also highlight some of the ways in which the economy and health services interact, themes that will be developed further in later texts. Studying patterns in rates of cancer sheds very interesting light on the causes of cancer. Some of the most striking links will be highlighted as well.
Cancer care and cancer research are also important components of industrial activity. Half of all drugs in clinical trials are for cancer; the global market for all cancer drugs was estimated at $48 billion in 2008, up from $34.6 billion in 2006. Analysts expect growth from 2010 to 2015 to be above 10% annually. Every year, the pharmaceutical industry spends between $6.5 and $8 billion on research and development of cancer drugs. This spend dwarfs that from government and research charities on drug development, potentially meaning that new drugs are concentrated in areas with maximum commercial rather than public health impact. Pharmaceutical companies with successful cancer drugs are among the biggest corporations worldwide. Biotech companies without marketable products but with a promising ‘pipeline’ cancer drug can be worth billions of dollars simply because of the possibility that the drug may be licensed at some future date for treating cancer. At least 19 anticancer drugs exceeded $1 billion in sales in 2009, a major strain for health systems in even the richest economies charged with purchasing these drugs for their patients.
At the other end of the spectrum, around one-third of cancer patients have very limited access to effective treatments, rising to over half in the poorest countries. Moving forwards, with an ageing population and rising drug price trends, we may get to the situation when ‘state-of-the-art’ drug therapy will be available only to the richest strata in the richest economies. Alternatively, better prediction of response to therapy may allow individually targeted treatment choices, reducing costs from unnecessary or ineffective therapy. Unlike, say, cars or computers, which we expect to, work every time we use them, most cancer drugs work on only a proportion of patients. For those with advanced disease, for whom the aim is palliation of symptoms or improvement in quality of life, this proportion may be much less than 50%, hence the majority of treatments may be pointless, or indeed worse than useless, as they may cause side effects with no benefit. Being able to identify patients who may benefit ahead of therapy would be very cost and clinically effective and this is therefore a major focus of current cancer research.
Cancer has also fascinated the world’s academics and universities. In 1961, John F. Kennedy pledged to put a man on the moon by the end of the decade. Nine years later, Neil Armstrong and Buzz Aldrin walked on the moon. Ten years later, in 1971, Richard Nixon echoed this pledge by declaring a ‘war’ on cancer. Rather like the more recent ‘war on terror’, picking a fight with a multifaceted worldwide problem has been at best only partly successful. Nixon’s initial pledge was around $100 million, which seemed like a bonanza at the time, but has turned out to barely scratch the surface. Since 1971, billions more research dollars have followed, but more than 30 years later cancer remains one of the largest causes of death worldwide, with around 1 in 3 developing the disease in developed economies and 1 in 5 in the West dying from it. Curing cancer is clearly harder than ‘rocket science’.
Worldwide, huge amounts are spent on research into the causes and treatment of cancer. In 2009/10, the US National Cancer Institute spent $4.7 billion on cancer research; equivalent spend in Europe was around 1.4 billion. In the UK, the biggest spender is Cancer Research UK, one of the largest British charities, which in 2010 had an annual income from donations of more than £500 million, reflecting the importance attached to finding causes and cures for cancer among the wider population (the foremost recipients of public donations are, however, animals not people!). Despite this vast expenditure on research, we still do not really understand what causes a substantial proportion of cancers. Furthermore, despite the money spent on drugs and drug research, for the majority of patients cured of cancer, this is as a result of either surgery or radiotherapy..
Chemotherapy and other newer treatments such as monoclonal antibodies or targeted ‘small molecule’therapies, while growing in importance, still account for only a minority of cures but have a major role in palliation of advanced disease symptoms.
There are various ways of looking at the problem cancer poses. These range from the raw numbers –how many people diagnosed, how many people die – to the personal – what is your individual risk of getting a particular cancer? Population-based statistics can be presented in various ways, from rates for the whole population to rates adjusted by age to calculations on numbers of years of life lost.
These latter statistics are often expressed as years lost before the age of 70 – the biblical ‘three score years and ten’ – thereby assuming that deaths after 70 (or sometimes 75) essentially represent death from old age. A further complication is that deaths from cancer vary enormously by income, race, and country of residence. For example, breast and prostate cancers are much more common in Europe and North America than in Japan and China. Migrants from these countries to the United States progressively alter their risk of these cancers towards that of white Americans but retain a lower overall risk. This tells us that the lower rates of breast and prostate cancer in the Far East are partly down to environment and partly down to racial differences or some linked aspect of the environment that is portable – diet, for example.
To try and explore these concepts further, I will present samples of the raw statistics using a range of methods. The question of which statistic is most useful depends on your point of view. For example, doctors working in public health, responsible for planning healthcare provision for a local population, will not be very interested in the rates of a given cancer in another country. Conversely, researchers looking at the effect of diet on risk of cancer may well want to focus on differences in disease rates between societies, as they may shed light on which lifestyle factors are important in the development of a given cancer. Fundraisers for cancer research will tend to focus on diseases affecting large numbers in the target donating population – breast cancer is the best example of this in Europe and North America, but more recently fundraising for prostate cancer has tapped into the same vein of public opinion.
The raw figures
As already mentioned, around 13%, or 1 in 7, of all deaths worldwide are due to cancer. This rises to around 1 in 3-4 in the developed world, where risk of premature death from infections, malnutrition, or violence is comparatively much lower. It is clear that there are large variations by region, with cancers common in one part of the world not featuring in the list of common cancers in another. There are too many differences to cover every one in detail.
Lung cancer
Worldwide, lung cancer is the largest cause of cancer death, with 17% of all cancer deaths, amounting to 1.2 million people, due to this type. It is a highly lethal disease, with fewer than 1 in 10 diagnosed surviving 5 years in most countries. Even in the United States, which has the best treatment results, fewer than 1 in 5 survive long term. Furthermore, the worldwide death rate is rising rapidly, having doubled between 1975 and 2002. Lung cancer is among the major killers in all parts of the globe. There is a well-known, strong link between smoking and lung cancer. Differences in the rates of lung cancer not surprisingly therefore vary with rates of smoking. Mostly, lung cancer is diagnosed relatively late in life, reflecting consumption of large numbers of cigarettes over half a century in most cases (younger people who have had less exposure obviously can suffer from the disease, but these cases are relatively less common). It therefore follows that the rates of lung cancer and the trend in the rates (rising or falling) reflect smoking habits over the previous half century. If we know the trends in rates of smoking, we can predict the future trends in lung cancer rates for a population.
Proportions of people dying of cancer by continent
In Western Europe and North America, rates of smoking in men are declining and with them rates of lung cancer (and other smoking-related diseases). In contrast, in large tracts of the developing world, rates of smoking are increasing rapidly as countries industrialize. The effect this is likely to have on cancer rates is illustrated by trends in Japan, where the rate of lung cancer between 1960 and 1980 more than doubled as the effects of Japan’s industrialization took their toll. Similar changes are now being observed in countries like China. There are various reasons for this: the habit still has an aura of ‘coolness’ in these countries very different from the increasing pariah status of smokers in the West.
There are generally lower levels of awareness of the health issues attached to smoking, and the restrictions on tobacco promotion increasingly seen in Europe and North America are not present. Indeed, officials in one recession-hit Chinese province recently decreed that all adults had to smoke the local cigarettes in order to boost both the local growers and tax revenues. Looking forwards, therefore, we can see that just as lung cancer declines as a problem in the ‘developed’ world, the newly industrializing economies will face an increasing burden of smoking-related cancers (and other problems such as heart disease) unless there is rapid adoption of the sorts of smoking-prevention strategies now the norm in Western Europe and North America. At present, this seems unlikely, and thus the industrializing world is likely to acquire one of the less desirable trappings of the developed world.
Breast cancer
In terms of new cases, breast cancer is the commonest cancer in women, accounting for 21% of female cancer cases and 14% of female cancer deaths worldwide. The overall survival rate is, however, much better than for lung cancer, with three-quarters of sufferers in Europe and North America surviving 5 years. Even in less developed countries, over half of breast cancer patients will reach this milestone.
Variations in breast cancer diagnosis worldwide
A study of the patterns of occurrence of breast cancer also helps to illustrate some of the ways cancer statistics can shed light on the behaviour of the disease. The risk of getting breast cancer (as for most cancers) increases steadily with age. Very similar distributions will be found in all developed countries. If we look the actual numbers for each age the peak numbers occur in the 50–70 age range – although their risk is higher, there are fewer women in the 70+ age groups due to deaths from other causes. As can also be seen, few women aged under 40 are diagnosed with the disease, although fundraisers often use women from this age group in their promotional materials. Looks at the distribution of cases from another angle, that of social class demonstrates that wealthier, better-educated women are at significantly higher risk than the less well off. Middle-aged educated women are often formidable campaigners, having both the time and education to lobby effectively. As we shall see later in the book, neither cancer research nor treatment access are arranged purely on the basis of need but are often substantially influenced by lobby-group pressure on behalf of particular groups.
The figures on worldwide risk of breast cancer again show some striking trends. Looking at there is a clear suggestion that breast cancer is in some way associated with affluence – richer countries have higher rates than poorer ones. For the smoking/cancer link, there is a pretty clear relationship between consumption and risk. It is harder to see why higher average income should increase the risk of an illness – this is the reverse of most public health trends. So why should this be?
One factor is the age structure of the population. The risk of cancer increases with age. Hence a woman in a poor country with a low life expectancy may simply not live long enough to get breast cancer, having already died of another disease earlier in life. This does not account for the large range of risk seen, however. There are various theories about the observed underlying difference, and the most likely explanation relates to the effect of hormones on the breast tissue. For example, there are clear effects on cancer risk relating to age of first pregnancy and numbers of pregnancies. Late onset of puberty, early first pregnancy, and more frequent pregnancies are factors that appear to protect against breast cancer. In the West, puberty occurs earlier than in the past due to better nutrition and higher-protein diets, whereas pregnancy occurs later due to effective contraception, the increasing independence of women, and better education. In poorer countries, puberty occurs later and women have less control over their fertility. Whilst this situation of course brings all sorts of potential problems, it does appear to protect against breast cancer. Breast-feeding, which affects hormone levels post-delivery, also appears to protect against breast cancer, and being more prevalent among the better educated in the West may be predicted to skew the trend the other way. Fertility rates tend to drop and age at first pregnancy tends to rise as both national and personal income increases, so it may be expected that, as with lung cancer, increasing development will result in an increase in cases of breast cancer worldwide.
Clearly, the breast is an organ that changes throughout life in response to changes in hormone levels (arising from puberty, pregnancy, breast-feeding, menopause, or drug therapies such as oral contraception and hormone replacement therapy). It follows from the above observations that medical treatments that affect hormone levels may alter the risk of developing breast cancer. Hormone replacement therapy (HRT) is widely used for menopausal symptoms. It was hoped that, in addition to helping ameliorate symptoms such as hot flushes and loss of libido, HRT would prevent diseases that tend to occur with increasing frequency after the menopause such as heart disease and bone loss (osteoporosis) with consequent risk of fracture. While HRT is indeed effective in some of these aims, it also appears to increase risk of breast cancer with prolonged use. A similar effect is seen with the oral contraceptive pill, which again works by altering the normal hormone environment. These, then, are confusing effects: some hormone changes (those associated with pregnancy and breast-feeding) protect against breast cancer, while other changes (oral contraception and HRT) increase risk. Against this background, much laboratory research is focused on the role that hormones play in the causation of breast cancer and on the development of drugs that interfere with hormone pathways and thereby treat breast cancer. One of these drugs, tamoxifen, which acts mainly by blocking the effects of the hormone oestrogen, can be regarded as one of the most effective drugs of all time, having saved the lives of probably millions of women and helped prolong life for many more in the 25 or so years since it came into clinical use.
Finally, there is a perception, promoted to a degree by groups campaigning for better treatment and research, that breast cancer is a disease of young women. In general, as we have already seen, this is inaccurate. However, studies of patterns of risk of breast cancer revealed that some families appear to be at very high risk of breast cancer, with mothers, sisters, aunts all affected at an early age, often with disease in both breasts or associated with cancer of the ovaries, or of the prostate in the male relatives.
These families were obvious candidates for in-depth study and, given the very obvious risks to the families involved, sufferers were often very receptive to participation in research. Studies of the patterns of inheritance in such cases suggested that the risk of breast cancer was passed on from mother to child with a 50:50 risk, and suggested at least two common inherited forms of the disease plus a number of less common versions.
Liver cancer
Liver cancer is one of the commonest cancers worldwide but with a very different pattern of distribution to lung and breast cancer. It is of particular interest as a freely available vaccination (against hepatitis B) can effectively prevent development of the cancer. Overall, it is the sixth most common cancer in terms of new cases, but the third most common cause of cancer death, reflecting the highly aggressive nature of the disease. There are a number of key features to the pattern of cases of liver cancer that merit more detailed examination. It is between 5 and 7 times more common in parts of China and Africa than in Europe and North America. The disease is almost always lethal, partly because it occurs in parts of the world with less developed healthcare, but mostly because it arises as a result of serious damage to the liver by the hepatitis B virus.
Liver cancer is linked to chronic liver damage, and in Europe and North America this is generally caused by alcohol abuse. In the parts of the world where the cancer is more common, the more important factor is infection with the hepatitis B virus (HBV), first described in 1965 by Dr Baruch Blumberg, who received the Nobel Prize for his work. Epidemiological studies established the link between hepatitis and liver cancer some years ago. Subsequent work showed that the molecular biology of the virus was consistent with it having a direct causative role rather than this being a chance association. With the linkage between virus and cancer established, the possibility of a vaccine against a common cancer became a reality. Pleasingly for all concerned, HBV vaccination has been a great success, with benefits appearing in the highest-risk populations very rapidly.
Cancers of the gut
Gut cancers commonly occur either in the top end (stomach and oesophagus) or the bottom end (colon and rectum), with cancers of the middle bit (the small bowel) being comparatively rare. There are some interesting trends in the patterns of gut cancers which I will run through, starting at the top with stomach cancer.
Overall, almost a million people are diagnosed with stomach cancer each year with around two-thirds of those afflicted dying from the disease – at least 650,000. Stomach cancer has been steadily falling in incidence in the West over the last 50 years, moving from being a relatively common cancer to now being quite rare. In other parts of the world, incidence has also begun to fall but more recently. Various reasons have been proposed for this, ranging from the rise of cheap refrigerators to medical treatments for stomach ulceration, but at present the reasons for the changes are not fully understood.
Cancers of the large bowel also show large variations between populations. Broadly speaking, bowel cancer is common in Europe and North America, less common in the Far East and uncommon in Africa. It is thus predominantly a disease of the developed world. Altogether, around a million people are diagnosed with the disease each year, and around half of these patients will die from the disease.
Death rates are now declining in North America and Europe due to improved awareness, early diagnosis, and better treatment. Studies of migrants suggest that the differences are environmental rather than racial – migrants from low-risk to high-risk countries rapidly take on the risk pattern of their new homeland. In addition, countries with an increasingly westernized diet such as Japan are seeing a rise in the incidence of the disease. The prime candidate for this effect is therefore diet –differences in the environment of the lining of the lower bowel clearly arise from differences in what goes in at the top end! There thus appears to be some sort of reciprocal effect – changes in diet over the last 50 years have made stomach cancer increasingly rare but have led to an increased risk of cancer at the other end of the bowel. Studying these sorts of changes provides important clues to the origins of cancers and also can point the way to prevention strategies.
Prostate cancer
Prostate cancer is an interesting disease. In Europe and North America, it is the most frequently diagnosed cancer in men and one of the leading causes of cancer death in men. In 2007, worldwide there were 670,000 men diagnosed with the disease. Deaths are more difficult to ascertain as many men diagnosed with early prostate cancer die with rather than of the disease. Like breast cancer, there are major differences in rates between different countries. Some of these differences appear to be driven by differences in rates of use of a blood test for prostate-specific antigen (PSA) which will detect early cancers and can be used as a screening test.
Prostate-specific antigen is made by the prostate and is a protein whose normal function is to liquefy the fluid produced during ejaculation (an aside – rodents do not make PSA and produce a solid semen plug during intercourse, yet mice are widely used in prostate cancer research). PSA is found in small quantities in the blood in men without cancer. In the presence of a prostate cancer (but also in other diseases affecting the prostate), larger amounts are liberated into the bloodstream, enabling the measurement of PSA to be used as both an early diagnostic and monitoring test for prostate cancer. Since the early 1990s, the test has been increasingly widely available and used both for screening for undiagnosed cancer and as a tool for monitoring the response of the cancer to treatment. In the USA, the test has been widely available from a range of sources and is actively promoted to the public by the makers of test kits – knowledge of your PSA level has become something men need to be aware of in the same way that cholesterol used to be. In the UK, until recently, government policy discouraged ‘opportunistic’ PSA testing, and there was no systematic screening programme on the grounds that there was no evidence that early diagnosis of prostate cancer reduced death rates from the disease.
Recent data from screening trials suggest that PSA testing may reduce deaths from prostate cancer, but that around 40 men need to be treated for PSA-test-detected cancer in order to save life. Whether this level of benefit will result in screening programmes being set up remains to be seen. It should be noted that this is similar to the level of benefit from breast cancer screening. Although widely applied, the benefits of screening are therefore not nearly as clear cut as may be imagined from the very widespread application of breast cancer screening across the Western world. For the time being, PSA testing is variable across the world and largely consumer driven.
If we start by looking at diagnosis and death rates from prostate cancer, some very obvious differences are seen (Figure 8). Men living in Europe and North America have a strikingly higher death rate than men living in Indo-China, where the disease is relatively rare, as it is also in most of Africa. Within Europe and North America, there are further interesting variations, with increasing risk of death with increasing distance from the Equator, an effect best seen in the white populations of the United States and Australia where the ethnicity of the white population is fairly uniform. If we look at ethnic effects, there are also striking variations, with men of African origin having roughly doubled the risk of prostate cancer death than white men. In contrast, men of Indo-Chinese descent retain the lower risk of their regions of origin, similar to the effect seen with women and breast cancer.
How can we explain this? The best evidence suggests that the differences between the white and Asian populations are driven by differences in diet plus a difference in racial sensitivity to whatever causes prostate cancer (which is largely unknown). The variation with latitude is much harder to explain by diet and clearly is not explained by race, as it can be observed in Europe, North America, and Australia. The best explanation seems to be exposure to sunlight, with sun exposure being protective.
This is a very surprising conclusion, given the widespread public health campaigns aimed at reducing people’s exposure to the sun. How may sun exposure affect the risk of cancer in an internal organ about as protected from the sun as it is possible to be? The answer appears to be vitamin D. Lack of vitamin D leads to rickets and conjures up images of Victorian workhouses and deformed children, but the 21st-century version of the disease may be an increased risk of cancer.
Vitamin D, sunlight, and cancer
Vitamin D is closely involved in the growth and development of a whole range of tissues including glandular structures like the prostate and breast. Vitamin D metabolism is complex, but a key step occurs in the skin and requires sunlight. Lack of exposure to sunlight over prolonged periods may thus lead to a shortage of ‘active’ vitamin D – not enough to cause rickets, but enough to shade the odds of getting prostate cancer. This may also explain why men of African origin, who often have the darkest skins, may be at the highest risk of prostate cancer if living in temperate latitudes. If this hypothesis were true, it could be predicted that white people with the highest sun exposure in a population would have a lower risk of prostate cancer. A good index of sun exposure is skin cancer, and studies have been carried out of prostate cancer risk in those with skin cancer. As predicted, the risk of prostate cancer is reduced in those with the highest levels of sun exposure as evidenced by solar skin damage and skin cancer. What is more, with sun exposure, reduction in an individual’s risk of getting prostate cancer is substantial – one study estimates this may be as much as 40%. Even in those developing prostate cancer, sun exposure appears to delay diagnosis significantly – around 5 years, from an average age of 67 years for the least sun-exposed to 72 for the most sun-exposed. The central message thus appears to be very consistent – one of the biggest killers of the male population could be prevented by more sunbathing – yet public health policy advises against it!
If this effect is present with prostate cancer, clearly mediated by circulating factors generated in the skin, could it be seen with other cancers as well? The answer appears to be ‘yes’, and the effect size seems to be similar for pretty much all cancers of internal organs. The only cancers that are increased by sun exposure are those of the skin (specifically melanoma), which actually kill relatively few people. The study of prostate cancer death rates thus sheds all sorts of interesting light on the causation of common cancers, and has thrown up a very surprising connection that fundamentally challenges current standard public health advice. In the opinion of the author, the accepted wisdom on sun exposure is overdue for radical revision.
Prostate cancer incidence and mortality
There is a second striking set of differences in the diagnosis and death rates. If we compare, say, the UK and the USA, we see very similar death rates but very different diagnosis rates per 100,000 population, with more than twice as many cases diagnosed per death from prostate cancer in the USA as in the UK. Looked at another way, a far lower percentage of men with prostate cancer die from the disease in the USA than in the UK.
There are a number of possible explanations – prostate cancer may truly be more common in the USA, and the US healthcare system twice as good at treating it as the UK system. Whilst it is true that the UK healthcare system delivers slightly inferior outcomes compared to the US system, these differences for most cancers are of the order of a few percentage points and are unlikely to explain the apparent difference in cure rates. Furthermore, if we look at rates of detection for other common cancers, the UK and USA have generally similar numbers per 100,000 populations, suggesting that other factors are operating. The explanation lies in the PSA blood test.
Differences in public policy and availability of PSA tests have resulted in far fewer men being tested in the UK than the USA, with a consequently lower rate of diagnosis of the disease. However, most men diagnosed in the USA, where there are high rates of PSA blood testing, have clinically trivial disease. This may never have troubled them had they not been diagnosed with it, suggesting the large difference in incidence is largely driven by higher rates of diagnosis of low-grade, relatively nonlethal disease in the USA compared to the UK. Both sides of the Atlantic, a smaller number of men are diagnosed and eventually die from more aggressive forms of the disease. Since the late 1990s, death rates have been falling, but whether this is down to screening directly or to other factors is hotly debated.
