News

The big causes of death from noncommunicable disease Richard Peto tells Andréia Azevedo Soares why efforts to reduce premature death from noncommunicable diseases should focus mainly on the big causes.

Q: You wanted to be an astronomer when you were at school. How did you Richard Peto, Professor of Medical Statistics and become interested in medical statistics? Epidemiology at Oxford University since 1992, is a A: By accident. At university I stumbled into statistics without really widely influential health statistician, jointly responsible knowing what it was, then in 1967 I for many large treatment trials, epidemiological studies had a job interview with Richard Doll and meta-analyses. He obtained a first degree in Natural (1912–2005), who had been one of the Science from Cambridge University in 1965 and an MSc first to show that smoking caused lung in Statistics from London University in 1967, but never cancer. Towards the end of the interview studied for a doctorate. he asked me why I wanted to work with Courtesy Peto of Richard Richard Peto him as a statistician, and I said: “I don’t know if I do want to – in fact, I don’t even know whether I want a job at all.” Q: What was the other study? Q: Why the focus on smoking rather than I remember his wife saying to me at the A: I was also working with Richard other possible causes of cancer? first Christmas party: “So you’re the Doll on the 20-year follow-up of his A: In the 1970s some people, believ- young man who isn’t sure whether he study of smoking and death in British ing that most cancers in non-smokers wants to work or not. Have you made doctors. In 1951 Doll had asked all the were due to occupational and environ- up your mind yet?” and I said I hadn’t. doctors in Britain whether they them- mental pollutants, hoped to use labora- At that time I still wasn’t sure that going selves smoked, and was following them tory tests on animals to find out which to work every day was the best thing to up to compare the death rates in smok- industrial chemicals were carcinogenic, do on this planet. ers, ex-smokers and never-smokers. ban those chemicals, and thereby greatly When British doctors read Doll’s find- reduce human cancer death rates. Doll Q: Why did you change your mind? ings they realized smoking was really and I thought this was unrealistic, A: A few months later I started to serious: it wasn’t just killing patients, it and that the big causes of cancer were get my first scientific results. It doesn’t was killing doctors too! So, most of the probably not industry-derived envi- matter how important they actually British medical profession accepted the ronmental pollutants. If so, a strategy were, but it is really exciting and inter- evidence of hazard, and British doctors of over-reliance on animal tests risked esting to get new results. From then on became the first major group of serious neglecting the few really important I worked more and more. In those first smokers with widespread permanent human hazards, like cigarette smoking. few months I would never take work smoking cessation, providing a nice Our focus on human evidence and on home with me. Now, I never go home natural demonstration that stopping the few causes of cancer that were known without a briefcase, but I was probably smoking saves lives. Later, cessation to be big was regarded by some as old- a better human being when I was young. spread to other professionals, then to the fashioned, even though in the United country as a whole. Doll and I published Kingdom in the 1970s cigarettes were Q: What kind of research were you doing? the 20-year results in the 1970s, the 40- a cause of well over half of all cancer A: I was working on many differ- year results in the 1990s and the 50-year deaths in men and an increasing pro- ent studies, but two were of particular results in the 2000s. portion of all cancer deaths in women. relevance to tobacco. Charles Fletcher (1911–1995) had written the 1962 report Q: How can we know what the big causes of the Royal College of Physicians on the are unless we study everything, including hazards of smoking, which was the first A moderate environmental contaminants? report on smoking by such a major body, A: Many causes of cancer have been reduction in a big and led directly to the highly influential “ discovered in the past few decades, and 1964 report of the US Surgeon General cause can prevent far others await discovery. For example, on smoking and health. I worked with more deaths than a big the risk of specific types of cancer is Fletcher on chronic obstructive pulmo- reduction in a small increased by certain chronic infections, nary disease, showing that some people cause. Smoking kills by particular types of mould on poorly who smoked cigarettes didn’t get much and stopping works, stored food, by traditional medicinal lung damage, but some suffered progres- but a billion people still herbs containing aristolochic acid, by sive loss of lung function over many smoke. some hormonal medicines, and by years, leading eventually to disability obesity and diabetes. It’s somewhat reas- and death. If, however, those getting into ” suring, however, that in recent decades trouble stopped smoking before their non-smoker cancer (and other) death health was severely affected, deteriora- rates in middle age have been decreas- tion slowed substantially. ing in many countries, so I don’t expect

Bull World Health Organ 2016;94:413–414 | doi: http://dx.doi.org/10.2471/BLT.16.030616 413 News

any new causes of noncommunicable Q: Did you see any effect of your research European Union, for example, there disease to be as big as smoking, which in China on cigarette sales? are 1.3 million deaths every year before still causes about a quarter of all can- A: That’s difficult to say. We did a age 70. More than a million of these cer deaths in developed countries. Of succession of nationwide studies in the are from noncommunicable diseases, course, it’s better if researchers don’t all 1980s, the 1990s and the 2000s involv- including 0.3 million caused by tobacco. follow the same ideas, as we might then ing a total of two million people, docu- The big causes of premature death from all be wrong. For me, however, what I menting how the epidemic of tobacco noncommunicable disease are smoking, wanted in the 1970s – and still want now deaths was evolving. In 1990, smoking blood pressure, blood lipids, diabetes – is for my work to be on the big causes was causing about 10% of all adult male and chronic infections, and these few big of premature death in many different deaths in China, in the 2010s it’s causing causes should not be obscured. populations. A moderate reduction in about 20%, and by the 2030s it will be a big cause can prevent far more deaths causing about 30%. Our October 2015 Q: What might obscure these big risks? than a big reduction in a small cause. Lancet paper contains the first reliable A: Lots of small or uncertain risks. Smoking kills and stopping works, but description of what is happening to For example, last year the International a billion people still smoke. male and female mortality from smok- Agency for Research on Cancer reported ing in China (increasing and decreasing, that red meat is “probably carcinogenic”. Q: What was it like to do research in respectively). I don’t know what the They didn’t say it’s definitely carcino- those days? effect of our work was or will be, but genic, only probably, so the lower limit A: At first we could do only me- at least the Chinese epidemic in recent of the range of reasonable uncertainty dium-sized studies, but now we’ve got decades has now been documented, as to how many cancers it causes is big studies in many different countries and systems are in place to ensure that zero. Nevertheless, this report gener- of major causes of chronic disease, in- it will continue to be documented in ated widespread media coverage, and cluding not only smoking but alcohol, future decades. probably increased general scepticism adiposity (overweight and obesity), about news reports about cancer haz- diabetes, high blood pressure and high ards, including those from tobacco. I’m blood lipids, which mainly affect vas- not saying we should concentrate only cular disease rather than cancer. Some Generating on tobacco, but we should not divert thought that studies of such old risk reliable“ evidence too much attention away from the big factors would not find anything new, and acting on it are causes of cancer that are proven and but there were always new things to both needed, but well established. learn – smoking predominated in Brit- should often be ain, alcohol in the Russian Federation, done by different Q: Should epidemiological researchers adiposity and diabetes in Mexico, and develop policies based on their findings individuals. in the present century the benefits of to change people’s behaviour? stopping smoking are even bigger than A: Not necessarily. For scientific early studies had suggested. ” results to be trusted, it may be best for those producing them not to be too Q: You conducted big epidemiological closely connected with the political studies in China. How did you become in- Q: How does your work relate to treat- process of how those results are used. terested in the Chinese tobacco epidemic? ment of noncommunicable diseases? Generating reliable evidence and acting A: As a statistician, I knew that 99% A: Vaccination, sanitation and on it are both needed, but should often of the world is not British, so I wanted treatment have transformed worldwide be done by different individuals. This to know more about the causes of pre- mortality rates from communicable is true for both the consequences and mature death in populous countries. I disease, and treatment has substantially the causes of smoking. For example, was lucky to get the opportunity to work reduced mortality from injuries and doubling real prices of a packet of ciga- with Chinese scientists in the 1980s, and noncommunicable disease, at least in rettes reduces consumption by about a I’ve done so ever since. I was expecting countries with good medical services. third. The World Health Organization communicable disease to predominate Indeed, half my work has involved col- and the United Nations have targeted as a cause of death, but our collabora- laboration with colleagues in Oxford a one-third reduction in smoking by tive surveys confirmed that it no longer on large international trials and meta- 2030, but the world’s governments earn did so. The surveys also showed a vast analyses of trials of various treatments about US$ 300 billion a year in tobacco increase in cigarette use by men, an in- for vascular disease and breast cancer. taxation and sales. If real cigarette prices crease that would take several decades stay constant and smoking decreases to have its main effects on mortality. Q: How well do scientists communicate by a third, then the governments of The Chinese tobacco epidemic was still health risks to lay audiences? the world would lose US$ 100 billion a at an early stage in the 1980s, and few A: Some try to emphasize how year. But if real prices double because were taking it seriously. I suggested set- much the big risks predominate, but of increased excise taxes, this itself will ting up systems to monitor the extent to some don’t. If we give the general public reduce consumption by a third, and the which cigarettes were killing the Chinese a list of a hundred possible causes of governments would gain US$ 100 billion people and to monitor the changes in cancer, it could divert attention from a year. This is the scientific evidence, this over decades, and the health min- things like smoking that cause vast but it’s up to governments and society ister at the time facilitated this. numbers of premature deaths. In the to decide what to do with it. ■

414 Bull World Health Organ 2016;94:413–414| doi: http://dx.doi.org/10.2471/BLT.16.030616