Who wants to live forever? Study in Nature questions benefits of calorie restriction in primates

Can scientific knowledge about the ideal diet help us increase our life-span? One of the options tested by members of the longevity movement is calorie restriction. A new study questions previous evidence for its health-effects in primates. Extreme fasting might not be a good idea for humans, after all.

The first time I had contact with the new science on dietary restriction and longevity was on a conference in Tucson, Arizona in 2006. The conference I attended was on consciousness research where I presented a paper on neuroethics but it happened to co-occur with a calorie restriction and longevity meeting. The other night I was invited to a millionaire’s party in the suite on the top floor of a nearby hotel and there attendants of both events mixed.

I remember talking to a young and exceptionally athletic and thin American woman who told me about her German ascendants – she had a German surname, though she pronounced it very differently – and then also explained some basic ideas of the longevity movement, something I had never heard about before. Indeed, there were people at the party who were very excited about exchanging recipes for zero (sic!) calorie chocolate and, correspondingly, low-calorie chocolate cake. Those people seemed so fascinated and charming, they almost convinced me – but then, they could not tell me much about tasty zero calorie beer, a serious bummer for my own longevity perspective.

Nevertheless, I followed the respective research for a couple of years and also used the review “Extending Healthy Life Span—From Yeast to Humans” published in Science in 2010 in my own teaching to demonstrate how some people generalize findings from yeast or rodents to human beings. This paper concluded with a balanced outlook, actually warning of negative effects of extreme dietary restriction:

Extreme dietary restriction can lead to several detrimental health effects such as amenorrhea, infertility, sarcopenia, osteoporosis, and immune deficiencies. Thus, it will be important to examine these negative side-effects in dietary-restricted subjects that are not malnourished. Indeed, experimental studies are required to evaluate the optimal calorie intake and macro- and micro- nutrient composition needed for healthy aging in humans, on the basis of age, sex, genotype, and energy expenditure. (Fontana, Partridge & Longo, 2010, Science, p. 326)

In the meantime, I have met people referring to respective findings in roundworms or rodents to argue that calorie restriction might also be beneficial for humans. There were even some who tried to compensate the possible psychological effects of this diet – for example, tiredness or a loss of motivation – with stimulant consumption. Because of the habituation and possible side-effects of stimulants, particularly when taken over a long period, I would personally strongly advise against such an “enhanced” calorie restriction diet.

Choose your control group wisely

One of the basic ideas of experimental research, particularly when working with living beings, is to choose one’s control condition or control group wisely. Randomization is one important step: Because no two living beings are perfectly alike, probably not even artificially standardized and genetically modified laboratory mice, individuals are assigned to two (or more) different treatment groups by chance. The reasoning is that individual variance will be random in both groups and not interfere systematically with the treatment effect to be measured (e.g. drug vs. placebo). A violation would be to assign, say, elderly individuals to one sample and young ones to the other. Then the factor age could obviously interact with the effect of treatment such that the drug might have different effects in the elderly than in the young. Of course, if age is of direct interest, this would be a recommended design, but if you are looking for a general, age-independent effect, then randomization of age between your groups is essential.

But even the results of a randomized study might be skewed, related to the baseline problem. If you modify the diet of one of your groups you will need to compare your findings to another sample. If we already knew the perfect diet, there would be no need to investigate possible benefits of a particular nutrition in the first place. That is, the effects of diet, for example, calorie restriction, will have to be interpreted relatively against the findings in another group with a “normal” diet; but of course this normality is something defined by the researcher. According to new findings it seems that one of the major scientific breakthroughs on the effects of calorie restriction in primates might have been biased by a bad baseline definition.

It was perceived as a major finding for the field when Ricki Colman from the Wisconsin National Primate Research Center (WNPRC) in Madison and colleagues reported 2009 in Science that “Caloric Restriction Delays Disease Onset and Mortality in Rhesus Monkeys”. Most importantly, at the time of gathering the final data, 80% of the monkeys in the target group, but only 50% of those in the control group were still alive. They had investigated them for almost twenty years. However, their control group was fed a diet that might not have been very healthy in the first place. It was made of purified compounds, offered in unlimited supplies, and contained 28.5% sucrose. Although I do not know much about monkey nutrition, this neither appears very natural nor healthy. Indeed, about 40% of them developed diabetes.

Fat rats skew research results

Only a few months later, Bronwen Martin from the Laboratory of Clinical Investigation in Baltimore and colleagues published a cautionary note in the Proceedings of the National Academy of Sciences in the USA based on their observations of the diet and other health-related factors of laboratory rodents. They warned that many control rats and mice are actually not good controls, in other words, that they are not a useful baseline to compare treatment effects against. This critique was referred to in Nature under the telling heading “Fat rats skew research results”.

Failure to recognize that many standard control rats and mice used in biomedical research are sedentary, obese, glucose intolerant, and on a trajectory to premature death may confound data interpretation and outcomes of human studies. […] Although overfed sedentary rodents may be reasonable models for the study of obesity in humans, treatments shown to be efficacious in these animal models may prove ineffective or exhibit novel side effects in active, normal-weight subjects. (Martin et al., 2010, p. 6127)

Thus, those animals are not representative of “normal” rodents but rather unhealthy ones. A treatment effect measured in another group and compared against this bad control might then actually not indicate an improvement relative to “normal” samples but rather a reduced impairment compared to the unhealthy population.

Did fat monkeys skew the original results?

The new study by July Mattison at the NIA and colleagues published online yesterday in Nature questions these findings empirically. They fed one monkey group a diet with 30% fewer calories but otherwise the same nutrients of a standard diet. This was made of natural ingredients and contained valuable components such as phytochemicals, minerals, and fat from soy and fish oils. Notably, while the diets at the NIA and WNPRC were comparable in terms of carbohydrates, the former contained only 3.9% sucrose compared to more than 25% in the latter. Accordingly, only 13% of their monkeys suffered from diabetes, much fewer than in the WNPRC sample.

When analyzing the data of their 23-year study, the researchers at the NIA did not find a significant effect of calorie restriction on years of life or age-related deaths. This poses a stark contrast to the previously reported findings in the Science paper. However, this does not mean that this dietary measure does not have any health benefit at all. Combining a number of variables such as tumor incidence, diabetes, or cardiovascular disease for a general analysis, they could at least find a statistical trend in favor of calorie restriction. The monkeys in the control group tended to have more tumors and at an earlier age, more frequently diabetes, but also less frequently cardiovascular disease.

Overall, Mattison and colleagues suggest that calorie reduction alone is too simple a measure to investigate the relation between nutrition and life span. The effects on health, morbidity, and mortality are complex and probably mediated by genetics, husbandry, diet composition, but also study design. Both balancing the amount of food intake and its composition might be a better health advice, after all, than the mere reduction of calories. Indeed, an epidemiological study published in the New England Journal of Medicine and based on the data of 1.46 Million white people reported that the mortality of those with a Body-Mass-Index between 20 and 25, thus people of normal weight, had the lowest overall mortality.

This whole story illustrates a couple of aspects on science and science communication: First, how some people are willing to adapt their lifestyle on the basis of preliminary findings within model organisms that can even be as remote from humans as yeast or rodents. Second, how difficult it can be to carefully select a baseline or control condition. Third, how important it is for science as well as for society that independent research groups try to replicate others’ findings. Fourth and finally, a more balanced diet might be more beneficial than a diet close to starvation.