Modern life is bad for us. If you ignore natural limits, fail to rein in your arrogance, then you create your own doom. People in tribal societies millennia ago had a natural and intuitive understanding of their world, living in healthy harmony. If you want evidence of how modern pressures cause harm, evidence of how they twist and scar us physically, look no further than atherosclerosis. Like high blood pressure, it was absent in our ancestors. We bring it on ourselves with our overindulgence and our diets, even though we’ve long known we should reduce our consumption of fat and alcohol. Our heart attacks, strokes, and vascular disease—they are nature’s response to hubris.
As opening paragraphs go, that one hinges on statements which are fashionable, widely accepted . . . and entirely false. A glimpse of reality—
Atherosclerosis was found to be as common in ancient mummies who had lived on fish as in those who feasted on steak.
and a lesson in how difficult it can be to clearly see it—comes from two recent studies. Both are beautiful, though one is flawed, and the flaws as well as the beauty are worthy of attention. One is called HORUS, after an Egyptian god, and the other PREDIMED, a tortured acronym derived from its full name, Prevención con Dieta Mediterránea.
Atherosclerosis ties the studies together. It has grown in importance to become the biggest killer in the developed world. It could easily be called a disease of affluence, but our tendency to associate it with overindulgence and other aspects of modernity misses the point.
The HORUS study cannot tell us anything about the superiority of one diet over another, but it does reveal that when it comes to tackling atherosclerosis by altering diet and lifestyle, there may not be a magic preventive or cure. Far from being the peculiar side effect of modernity, problems with blood vessels narrowing and hardening occur routinely with age in all human societies. The foggy idea that modern life fosters atherogenesis—a notion that for too long was accepted without having been properly examined—evaporates under the sunlight.
While atherosclerosis may not be a uniquely modern condition, never has it been more prevalent. What accounts for its massive rise? Could that be unnatural, some consequence of leading lives too far removed from those we were designed for? The authors of HORUS even raised the possibility, pointing out that in studies looking at “20th century hunter-foragers-horticulturalists, about 75% of mortality was attributed to infections, and only 10% from senescence.” For senescence, read mainly atherosclerosis and cancer. Those two conditions have exploded in importance. But that’s not because of the stresses and unnatural pressures of modern life, nor because our world is packed with novel industrial carcinogens. Heart disease, strokes, and other manifestations of atherosclerosis simply get more common with age, as does cancer. Those diseases we fear so much diminish our lives in ways that are terrible, but their dominance should also be cause for celebration. They show that we have survived to achieve a degree of ripeness our ancestors rarely attained.
The general medical wards in the hospital where I work have an atmosphere of elderly decay—not just from the decrepitude of the building, but because the beds hold patients with a median age of eighty-four. When I worked in Africa the wards were packed with the young, and the air was thick with the infections and traumas that were ending lives not yet properly begun. Diseases of aging are such a problem now because so many of us live to meet them. They are diseases of affluence arising from the wealth of longevity. The past two or three decades have seen a decline in cardiovascular disease in America and the rest of the developed world. The best estimates attribute about half of that to drugs and operations, half to the fact that over that period of time we have, through lifestyle and diet and a reduction in smoking, simply become healthier.
The second study, PREDIMED, is Spanish, and it looked at how to reduce the impact of atherosclerosis. It too reported a positive finding. “Among persons at high cardiovascular risk,” it found, “a Mediterranean diet supplemented with extra-virgin olive oil or nuts reduced the incidence of major cardiovascular events.” Over the five years for which the trial ran, those on a Mediterranean diet dropped their relative risk of death or of a heart attack or stroke by almost a third. That’s a lot—a huge amount relative to what medicine normally has to offer. It’s a significantly bigger impact than drugs like statins and aspirin can deliver, or even interventions such as coronary angiograms and openheart surgery.
PREDIMED took almost 7,500 people 55 and older, who were free of cardiovascular disease but had risk factors for it, and randomly divided them into three groups. One group, the control, was simply encouraged to eat a low-fat diet, reducing their intake of oil and fat, as well as of red meat, sodas, and processed foods. They did the worst. The other two groups, like the first, were urged to consume fruits and vegetables and to keep away from sodas and red meats, but in addition they were pressed to eat fatty fish, to take wine with their meals, and to gorge on nuts and oil—components of the traditional Mediterranean diet. The two experimental groups differed from each other in that one was given a weekly allowance of extra-virgin olive oil (a liter of it), and the other received 210 grams per week of walnuts, almonds, and hazelnuts. We’re used to dietary recommendations being puritanical and reinforcing our conviction that the pursuit of health involves battling our appetites. A liter of olive oil per person per week is no small amount. The idea that indulging some of our appetites might make us healthier takes getting used to. Should we accustom ourselves to doing so?
Perhaps. But the fact that we so easily believed atherosclerosis to be a uniquely modern condition was not an aberration. Seeing the natural world clearly is strikingly difficult. Unless our observations and our experiments are careful—with our chief protection often being the scientific method—they become, not windows onto reality, but mirrors of our own presumptions. When that most consummate observer of physical reality, Leonardo da Vinci, drew dissections of the human body, he was simply unable to see clearly what was in front of him. He recorded not what lay before him, but a blend of what Classical anatomists had written should be visible and what he had seen in other animals. Intuition and prejudice can both be helpful mental traits, but they have a treacherous influence on our ability to bring clarity to our understanding of the natural world.
PREDIMED claimed to be testing the impact of a Mediterranean diet, but as you might expect, a Mediterranean diet was fairly common in Spain. An editorial accompanying PREDIMED’s publication remarked that despite the encouragement to avoid fat and alcohol, “the control group appeared to consume a variant of the Mediterranean diet.” The study’s efforts at dietary counseling, in fact, led only to very small changes that may not have mattered at all. What really altered people’s diets was what you gave them over and above their daily choices: PREDIMED seems chiefly to have ended up testing not the Mediterranean diet, but the effects of handing out free oil or nuts.
That’s not necessarily a problem—not if those effects are more compelling than the drugs and surgeries of modern medicine. But there was another problem with PREDIMED, a modern version of da Vinci’s inaccurate drawings: PREDIMED suffered from its authors being overly enthusiastic about what they wanted to see. Understanding how the error happened involves considering the error of methodology, and that relates to the fact that the trial ended prematurely. After PREDIMED was underway, a monitoring committee got together regularly and reviewed the accumulated results, as is good practice. When the committee met in the summer of 2011, they decided enough was enough. The trial was showing such benefit in two of the groups—exactly the sort of benefit it had been set up to find—that it would be unfair to the third group and the wider world for the study to continue. People deserved to know. PREDIMED was stopped and two years later it was published.
With any long-term study you have a duty to check in on it. Your new treatment might turn out to be so good (or bad) that carrying on with the trial under the guise of needing to gather more information becomes dishonest. But you need to weigh against that the concern that the earlier you stop the trial, the less confident you can end up being about the results. You wouldn’t think it fair to stop a horse race, goes one analogy, the moment your favorite is a nose in front. The longer the race runs, the more sure you become. Not that this means uncertainty can ever be avoided—if we waited to see what drugs do over a period of an entire human life, which is what we really need to know, medical development would stagnate. How much doubt can we cope with? Tolerating some degree of uncertainty is an essential part of rational thought.
Unless we are careful our observations and our experiments become mirrors of our own presumptions.
The authors of PREDIMED had decided in advance to stop their trial if differences appeared that had a less than one in fifty likelihood of being due merely to the play of chance. When that happened, they called a halt. That left a real but small chance that a misleading snapshot has been mistaken for a true result. Even so, that isn’t the key problem. The odds are great that the olive oil and nut diets are better (although one might wish for a better than 1 in 50 chance of error if the world was going to permanently change its diet as a result), but the more fatal issue is that stopping early left it desperately unclear how much better the superior diets might be. Those who got the free olive oil had a hazard ratio of 0.70 (95 percent confidence interval of 0.53 to 0.91) for heart attacks or strokes compared to those on the “lowfat” diet. Those getting the nuts had similarly improved odds—a hazard ratio with an identical midpoint, 0.70, but a confidence interval running from 0.54 to 0.94. What that means is that there’s a 95 percent chance the oil and nut diets provide benefits that lie somewhere between reducing bad outcomes by almost half and reducing them by only 6 percent.
We’re left not knowing whether these diets really matter. PREDIMED will have no impact. It won’t persuade people to change their habits, nor persuade governments or HMOs to give out free nuts and oil. Continuing the trial would have led to those confidence intervals narrowing. Stopping it early was not a necessity, but an ethical and a scientific disaster. It did not protect the people in the third, control group from the harm of being excluded from mankind’s massive adoption of free nuts and oil—it didn’t protect them because it left the trial’s conclusions so vague that no one was ever going to adopt anything. Stopping early betrayed those wishing to know whether these diets achieve a worthwhile alteration in health. It betrayed those who participated in (and funded) the trial out of a belief that they were usefully adding to the world’s knowledge. The only people who are likely to be persuaded by PREDIMED are those working in the field, meaning that its only impact will be to prevent better trials taking place to look more closely at the same diets.
Science has made modern life healthier, happier, and longer. It remains our best tool for fighting illness and misery and for strengthening our grasp on truth. When the scientific or mental method we deploy is shaky and timid, asking a good question doesn’t protect us from ending up with an unhelpful mess of an answer. The more worthwhile your target, and the closer you are to hitting it, the more blame you deserve if your last-minute mistake means you miss completely. We wouldn’t wince at a 500-year-old sketch by a terrible artist that shows a mistaken version of the human body, but to find that da Vinci got something wrong—when he got so much right—is painful.
The PREDIMED trial, which was almost excellent, thus failed badly. The beauty was in its promise, and that beauty remains in its lingering and unfulfilled suggestion that randomized experiments are capable of differentiating fantasy and guesswork from reality when it comes to understanding the effect of dietary choices on human health. PREDIMED’s present utility, like that of the prejudices painting prehistoric and historical populations in a false glow of natural benevolence, lies in showing us what to avoid.--DB