Any study focused on fats vs. carbs, heart disease and death, encompassing several countries, and that challenges existing dietary guidelines around the world is bound to generate news coverage — as well as a slew of eye-catching headlines:
Low-fat diet could kill you, major study shows (The Telegraph)
The Low-Fat vs. Low-Carb Diet Debate Has a New Answer (TIME)
Recommended fat intake should increase, Canadian researchers say (CBC)
The study, published in this week’s issue of The Lancet medical journal, uses questionnaires to document the dietary intake for roughly 135,000 people in 18 countries (3 high-income, 11 medium-income, and 4 low-income).
The findings can be summarized as:
- HIGH CARB intake (>60% energy intake) was associated with a higher risk of total mortality compared with low carb diets
- HIGH FAT intake (~ 35% energy intake) was associated with a lower risk of total mortality
- NEITHER HIGH CARB nor HIGH FAT were associated with a risk of cardiovascular disease or cardiovascular deaths.
In threatening death, promising a ‘new answer,’ and recommending a sweeping change in dietary guidelines these headlines are jumping the gun. Badly. Here are two of the major reasons why.
First, this is an observational study (in which cause and effect can’t be established). So headlines like CTV’s — which says that “moderate fat consumption may lower risk of death” — are clearly out of bounds. This study can’t tell us whether changes in fat consumption will cause any change in someone’s risk of dying early. By contrast, STAT did a nice job of calling attention to this limitation and at least mentioning the biases that are common to observational studies.
Second, it’s based on self-reporting of eating habits using food frequency questionnaires (FFQ’s are notoriously unreliable).
But there are some other reasons why journalists should challenge any suggestion that this study warrants an overhaul of our current nutritional guidelines, as suggested by lead investigator Mahshid Dehghan PhD of McMaster University.
Some important considerations missed by most of the coverage
The study authors appropriately mentioned — but very few reporters discussed — the confounding issue of socioeconomic status at play in this international study. It’s well established that those with limited financial resources are less able to afford the higher quality fats, carbohydrates and proteins as those study subjects with more money. Not to mention substantial differences in access to health care, exercise, and health information between vastly different socioeconomic groups. These very real confounding influences make it even more difficult for the study authors to conclusively state that high carbs or low fats were solely responsible for the published results.
Not a single reporter (and I’ll include myself here, since I missed this as well) brought up the fact that, although these results seem impressive at the population level, they are minuscule at the individual level. Larry Husten of the blog CardioBrief caught this and quoted one of the study authors — Andrew Mente PhD of McMaster University — as pointing out:
“The effects are modest effects, in the neighborhood of a 20% reduction in relative risk. So if the annual [absolute] risk of mortality is 1%, it would be reduced to 0.8%. At the individual level it is tiny. And nowhere near what you find for smoking and lung cancer — about 200 times smaller in fact.”
Ethan Weiss MD
A final limitation of the study is that it didn’t document what foods the macronutrients came from. For example, we don’t know if the high-carbohydrate consumers were disproportionately eating refined sugars and grains rather than whole grains — which is quite likely. Nor did the study look at trans fats, which are ubiquitous in the processed foods that are becoming increasingly popular in low and middle income countries. Trans fats are strongly associated with increased risk of heart disease and cardiovascular death.
“I think this study should be seen as hypothesis-generating, provocative, but not result in any sort of large-scale policy changes regarding the absolute daily recommended values for these major macronutrients,” says Ethan Weiss MD, a cardiologist at UC-San Francisco. His other impressions:
“As the authors point out, a lot of the carbohydrate association is driven by Asian countries where diets are rich in cheap, processed carbohydrates. This study does add to existing evidence that excessive carbohydrate consumption (> 60%) is unhealthy; but how that occurs and whether that remains true outside an observational study remains to be determined. So, for my patients I’ll keep saying: ‘eat lots of olive oil and nuts (monounsaturated fats), and limit consumption of processed carbohydrates’.”
Time and time again we see news stories about nutrition being hyped. It’s an area of health care where it’s extremely difficult to design definitive studies and, therefore, make clear-cut conclusions. Even with the significant limitations listed above this is an important study. At the population level it helps identify trends that can guide future controlled studies that may provide more definitive evidence.
But it also serves as a reminder how easy it is for journalists and readers alike to label preliminary findings as definitive. Or, even go so far as to suggest widespread policy changes be made based on very limited data. All the more reason we recommend these 5 Questions Journalists Should Ask when covering nutritional or diet research.
Especially since the fat vs. carbs merry-go-round isn’t going to stop spinning anytime soon.
