For Peto's sake, give up

Fred Pearce talks to Richard Peto abouthis crusade to save 150 million lives by halving the number of adult smokers and why his research is under threat. Worldly Wise: 23

Few people can claim to have saved as many lives as Richard Peto has. Certainly not many mathematicians. But as one of the world's leading designers and analysts of large-scale studies of disease risks and treatments, he has a unique insight into why millions of people die in the prime of life each year - and how they might be saved.

Peto co-directs the University of Oxford's clinical trial service unit in an anonymous building at the back of the Radcliffe Infirmary. He went there 30 years ago with his mentor Sir Richard Doll, the grandfather of "big" epidemiology, who established the unit for the Imperial Cancer Research Fund.

Doll is still around, but today Peto's co-director is Rory Collins, medical doctor, statistician and the man who makes the grand stratagems happen. Together, with a staff of 120, they form the nerve centre for a worldwide network of mass clinical trials that has transformed medical thinking. And the unit, with its huge archive of data and blood samples, literally waiting for people to die before springing into action, is likely to carry on saving lives for decades to come.

Peto, whose shock of grey hair cannot disguise the zeal of a 1960s student, is proud of the successes: probably 20,000 lives saved each year from the discovery of the value of prescribing tamoxifen to women after breast cancer surgery; hundreds of thousands of deaths prevented from finding that aspirin helps prevent heart attacks; and who knows how many saved by his tireless documenting of the spread of the tobacco pandemic?

This last is the work most publicly associated with him, not just for his research but for his brutal way with statistics. A third of all adults worldwide smoke. Three million smokers die annually from their habit. Peto estimates that the death toll is likely to rise to ten million within 30 years. With the World Health Organisation, he is conducting a series of national studies in China, India, Cuba, Mexico, Egypt, South Africa, Russia and elsewhere to track the disease.

Although the studies are being carried out by local researchers, his hands-on approach ensures a heavy itinerary. We fitted in our interview between his trips to Beijing, Havana and Budapest.

The Chinese study published its first findings in November and attracted headlines worldwide. A unique look at an epidemic across the world's largest nation, it concluded that a third of all young men in China today are likely to die from smoking. That is 100 million men.

The study, designed by Peto and carried out by the Chinese Academy of Medical Science, was a prodigious effort. More than 500 field workers traced and interviewed the families of a million people who died between 1986 and 1988, establishing how they died and whether they smoked. "Where else could you interview a million dead people - well their relatives anyway?" says Peto, admiringly, of Chinese organisation. Another 250,000 living men aged over 40 were interviewed, tested, and will be monitored to the grave.

Such studies only confirm what we already know about tobacco but Peto's dogged pursuit brings home to governments the scale of the epidemic. It also throws up interesting results. The pattern of death attributable to tobacco in China is not the same as that in the West, for instance. The links are just as strong, but different.

In Britain, approaching half of tobacco victims die of lung cancer. In China it is only 15 per cent. Heart disease also seems to trouble Chinese smokers much less than in the West. Instead emphysema is the big killer. "This came as a complete surprise," says Peto.

There also seems to be a link to tuberculosis, both in China and India. Some early findings suggest smoking may quadruple the TB death rate in parts of India, a country where a million people die of the disease each year. Peto concludes that smoking appears to increase the incidence of death from already common diseases. He does not yet know why.

Nor are his forceful arguments against smoking confined to statistics. Two years ago Cambridge dons voted to accept a Pounds 12.6 million endowment from British-American Tobacco, the world's number two tobacco seller. Brushing aside university claims that it was not endorsing the company's products, Peto noted simply: "Half BAT's regular customers die so they have to spend a lot of money buying goodwill. I am sorry that Cambridge did not have the sense to turn down the money."

"People still don't really understand the hazards of smoking," he adds. "Packets say Smoking Kills. But they don't say that of 1,000 young adult Britons who smoke, one will be murdered, six will die in road accidents and 500 will die from tobacco."

It is probably no accident that the home of the world's leading investigators of the hazards of smoking is also the place where death rates from tobacco have fallen most steeply - by half in 25 years. But Peto worries about the fashion in public health education in Britain for concentrating on preventing children from taking up smoking. "If we concentrated on persuading adults to stop smoking it might work better," he says.

One reason is that his studies reveal that within a few years of giving up smoking, the risks of lung cancer and heart disease revert almost to normal. Another reason is that he is not sure that kids respond to fears about being killed. "But adults do. By the time you are in your 30s you know that death is real."

Moreover, if we are in the business of saving lives over the next 50 years, an assault on adult smoking would bring greater dividends. This reveals itself in some startling calculations, which are vintage Peto. He estimates that if we halve smoking among children in the next 20 years, we will avoid 15 million deaths worldwide between now and 2050. But if we halve adult smokers, that could save ten times as many lives in the same period.

Peto is destined to be best known for his work on tobacco. But Peto and Collins have been equally influential in breaking the mould of clinical research by undermining the obsession with small intense trials. Motivated by a desire to learn as much as they can about individual patients, doctors can end up learning nothing of statistical significance.

What is required, says Peto, are extremely large trials in which simplicity rules. An impressive example of this is therapy for dissolving blood clots in heart patients using the drug streptokinase. Trials conducted in the 1960s and 1970s failed to show significant improvements from the treatment. The only identifiable effect was heavier bleeding, leaving analysts to conclude that the treatment was "dangerous and ineffective".

But Peto and his colleagues re-examined data from all the trials. Though none of the studies individually found statistically significant benefits, when they combined the data they found a highly significant 20 per cent reduction in deaths among those given the drug as soon as possible after the onset of a heart attack.

They repeated the trick with tamoxifen, a drug widely tested on women who had undergone surgery for breast cancer. It seeks out and immobilises the cancer cells left behind. More than 50 studies among some 30,000 patients worldwide failed to identify a benefit for pre-menopausal women from the drug. But re-analysis by Peto and colleagues revealed that the drug halved their risk of getting a new cancer.

The future for the unit appears increasingly to be in blood. It has hundreds of thousands of blood samples from round the world in its freezers. Much of it has never been analysed - deliberately. "We store the blood from various trials and wait until people start dying," says Peto. Then his team analyses the blood from those who die of whatever disease is being investigated and compare it with a random sample of the remainder. The aim is to look for chemical or genetic markers in blood that predisposes for the disease in question. "It can be difficult to persuade people to fund us to collect blood when we don't even know what we are going to test it for," says Peto.

On the therapy side, the unit is helping organise a trial among a million Indian children into the potential benefits of a pill to kill intestinal worms. It seems that the pill, costing about 3p and administered once every six months, increases an infested child's year-on-year weight gain by 25 per cent. The critical test for Peto will be whether that weight gain is translated into a reduced death rate. "The treatment could be adopted as standard policy," he says.

But his main preoccupation remains reducing death in middle age. Today in the rich countries 70 per cent of people live to their three score years and ten. In poor countries it is about 50 per cent. Addressing key scourges could take that figure up to 90 per cent, he says. His work is charting the course.

There is one cloud on the horizon. New international guidelines on clinical practice in drugs trials are in danger of outlawing the Peto approach. Rules are being proposed that require very strict auditing of data. The aim is to remove the risk of fraud in trials for new pharmaceuticals. But one result would be to vastly increase the cost of his mass studies. The data audits alone could cost millions of pounds. "It would be a catastrophe," he says.

Medical research, page 26