We love demonizing and idolizing specific nutrients, particularly when it comes to their effects on our waistlines and longevity.
While carbohydrates have been catching their fair share of heat lately, fat has historically received the brunt of criticism in nutrition science.
You may have seen the following headlines, or similar variations, over the past week:
“Lower-fat diet reduces women’s risk of dying from breast cancer, study says” (The Washington Post)
“A Low-Fat Diet May Lower the Risk of Dying from Breast Cancer” (Time)
But what merit is there to these headlines?
What does the study upon which they are based actually say?
Read on to find out.
Disclaimer: this post is intended to discuss the referenced headlines and the study upon which they are based, not to determine the role of diet in cancer outcomes.
The referenced headlines were prompted by a recent analysis, published in JAMA Oncology in 2018, of data from study titled “Association of Low-Fat Dietary Pattern With Breast Cancer Overall Survival (1).
The study’s authors analyzed data from a trial called the Women’s Health Initiative Dietary Modification Trial, which started in 1993.
In the trial, 48,835 women were randomly assigned to one of two groups.
One group received no dietary instruction or support beyond “written diet-related education materials”.
The other group was instructed to implement what was called a “low fat dietary program” intervention.
The intervention was intended to reduce dietary fat intake below 20% of total energy and increase consumption of vegetables, fruit, and grains.
Neither calorie reduction nor weight loss were targets for the intervention.
During the first year of the trial, the intervention group attended 18 group sessions for instruction on this dietary pattern.
Over the second year of the study, each participant attended an individual session between weeks 12 and 16 and then group sessions were held each quarter.
Participants provided a 4-day food record at the beginning of the study, at the end of year one (all participants), and then every 3 years thereafter (33% rotating sample of participants).
The 2018 analysis doesn’t offer much information so far as what the participants did or didn’t eat following the intervention, but a 2006 analysis of the data up to that point offers some insight (2):
“Vegetable and fruit consumption was higher in the intervention group by at least 1 serving per day and a smaller, more transient difference was found for grain consumption.”
The authors found that, of participants that were diagnosed with breast cancer, 174 of 671 participants in the dietary intervention group had died and 342 of 1093 in the control group had died.
These values are for total deaths of any cause, not only breast cancer.
While the study authors presented this accurately, that the effect was seen not in breast cancer deaths but of all cause mortality, the news headlines do not.
The news headlines suggest that the dietary intervention was effective at reducing death from breast cancer.
The news headlines aren’t the only part of the picture that warrant a closer look.
There are additional points to consider regarding the study itself and how its results were presented.
Reduced fat or increased fruits and vegetables?
Any time there is a reduction of one thing in a diet, we would be prudent to look at what, if anything, replaced that which was reduced.
In this case, we might ask, “what replaced the fat?”
In this intervention, the participants didn’t simply reduce their fat consumption.
They also increased their fruit and vegetable consumption.
Due to their fiber, vitamin, mineral, and phytonutrient content, it’s not unreasonable to suggest that the increased fruit and vegetable consumption played a role in the observed effect on mortality (3, 4).
So, do these results suggest that fat is harmful, or that fruits and vegetables are beneficial?
What if instead of reducing fat, we reduced carbohydrate, but also increased fruits and vegetables?
At least one (really well freaking done) study suggests that both fat restriction and carbohydrate restriction can yield to positive health outcomes when an emphasis on food quality is prioritized (5).
Thus, we must ask ourselves if the benefit observed in the 2018 WHI DM analysis was due to dietary fat consumption or due to increased fruit and vegetable consumption.
To this point, one statement in the study is noteworthy:
“As the contribution of the individual components of dietary intake to the breast cancer outcome cannot be fully separated, the WHI-DM trial evaluated all consequences of the low-fat eating pattern.”
It is not clear why the authors were unable to parse out the effects of each dietary component.
Without such an analysis, it’s not reasonable to suggest that the reduction of fat, and not the increased fruit and vegetable consumption, was more responsible for the observed benefits.
To the authors’ credit, they do not explicitly state that fat reduction was responsible for the observed benefit.
Rather, they refer instead to the “eating pattern” and “dietary pattern”, and not a single characteristic of the pattern, being associated with positive health outcomes.
Prefacing these terms with the descriptor “low fat”, however, emphasizes that characteristic of the intervention diet.
Reduced fat or reduced weight?
There’s another confounding variable that may be playing a role in the observed benefit of the intervention diet.
Participants in the intervention group also lost weight.
The authors note that weight loss was “not an intervention target”, which would technically mean it was unintentional.
However, the difference between “intentional” and “unintentional” weight loss may reflect a difference between “intentional” weight loss driven by dietary and lifestyle change and “unintentional” weight loss driven by disease.
Thus, the reduced mortality effects of intentional weight loss may be more relevant to this study than the increased mortality observed of unintentional weight loss.
I’m curious if we would see similar benefit to survival in a low carbohydrate intervention that resulted in similar weight loss.
Finally, these results reflect only a small subset of a larger body of data.
This study reports observations of only 1764 – fewer than 4% – of 48,835 participants
It would be interesting to see what the difference in mortality was for all participants, including those not diagnosed with breast cancer.
If the 2018 paper was intended to be a follow up to those analyses, why not explore some of the endpoints explored in those analyses, like breast cancer incidence, before exploring more focused points such as all-cause mortality following breast cancer diagnosis?
It’s possible the authors reported only the results observed of this subset because it showed an effect not observed of the overall body of data.
Reporting only data that shows an effect while withholding data that doesn’t show an effect is a form of publication bias, which is a recognized problem in the scientific community (11).
It’s worth considering if that played a role in the way this analysis was conducted and presented.
This point is intended as a question for consideration and not an allegation of mal-intent.
This wasn’t the worst case of findings being presented inaccurately by any means, but it serves as an example that news headlines about nutrition nearly always deserve a second look.
The points above aren’t offered to suggest that the study was useless or that the dietary intervention was ineffective.
Rather, they’re to suggest that this study doesn’t tell us what, specifically, about the dietary intervention may have been responsible for the reduced mortality among those diagnosed with breast cancer.
Was the observed effect a result of increased intake of fruits and vegetables?
Were the benefits due to weight loss, independent of dietary characteristics?
Is there a relationship between level of adherence and health outcomes?
Additionally, it doesn’t tell us what effects, if any, the intervention had in the other 96% of the trial population.
These are all questions we might want to answer before attempting to apply this study’s findings in the form of dietary recommendations.
Most of us will probably see plenty of health improvements, regardless of whether we choose a low-carbohydrate or a low-fat approach, so long as we’re consistently eating quality food (12).
Focus on the basics.
Prioritize minimally-processed, nutrient-dense foods.
Sleep like a lion.
Get some fresh air and sunlight.
Do things you love with people you love.
Aim for consistency, rather than perfection.
Rinse and repeat.
You’ve got this.