Monkeying Around with the Data on Calorie Restriction?

Ronald Bailey | 7.13.2009 9:33 AM

Last week, two research teams reported to great fanfare that restricting the calories consumed by rhesus monkeys had extended their lifespans. Calorie restriction is thought to increase longevity by boosting DNA repair. The idea is that the mechanism evolved so that creatures on the verge of starvarion could live long enough to reproduce when food becomes plentiful again. But did the experiments really show the CR works?

In my earlier blogpost on the research results, I noted that some experts quoted in the New York Times were not convinced. Why? Because the difference in actual death rates between the dieting monkeys and the free feeding monkeys was not statistically significant.

Food policy blogger, Sandy Szwarc, looks at the reported results at Junkfood Science and finds them severely wanting. Among other things, Szwarc argues:

The lower mortality claimed among the monkeys on the calorie restricted diet were achieved only after eliminating 37% of the monkey deaths. They defined mortality as "age-associated deaths" and eliminated any cause of death they didn't believe was associated with aging. As the supplemental data explains, 16 deaths from "non-age-associated causes were censored and their age of death used as the time variable in the regression."

Science doesn't really work that way. Researchers can't simply ignore the evidence that doesn't support their hypothesis. That would be the difference between research done to build evidence to support a hypothesis, from science that is objectively studying a hypothesis…

The non-aging-related causes of death included monkeys who died while taking blood samples under anesthesia, from injuries or from infections, such as gastritis and endometriosis. These causes may not be aging-related as defined by the researchers, but they could realistically be adverse effects of prolonged calorie restrictions on the animals' health, their immune system, ability to handle stress, physical agility, cognition or behavior.

As we know, the most important endpoint in medical interventions is all-cause mortality. Selectively looking at only one cause of death, while ignoring that more patients died from something else, is not evidence to support the efficacy of a treatment. "The treatment worked, but the patient died" is not good medicine that considers the whole patient.

Assume that the dead dieting monkeys were properly excluded on the grounds that they died of non-age-related causes, there is other evidence that suggests that being too skinny is not healthy. A 2003 study by British researchers found that underweight people (body mass index below 18) are more likely to die (at a younger age*) than are average or slightly overweight people. That finding was bolstered by a new Canadian study reported last month. Not surprisingly, people who practice calorie restriction tend to have below average BMIs.

So while there is suggestive evidence that calorie restriction may increase longevity, the case is not yet proved.

Hat tip to Andrew Mayne.

*clarified