Breast Cancer is a serious diagnosis and devastating disease.
This 2019 https://acsjournals.onlinelibrary.wiley.com/doi/full/10.3322/caac.21583 publication documents over 300,000 new cases and an estimated 42,000 deaths in 2019. If there is a nutritional approach that could decrease either frequency of developing breast cancer or deaths from breast cancer, I would strongly support it.
A study that reported a 22% decrease in mortality from breast cancer with a low-fat diet was recently brought to my attention. I will explain why this study should not be taken at face value and present an alternative viewpoint for consideration.
The study in question was published in JAMA Oncology in May 2018. The authors reviewed data that was originally gathered as part of a large trial known as the Women’s Health Initiative (WHI). The study enrolled subjects from 1993 to 1998 and followed them until 2005 for an average of 8.5 years. 48,835 post-menopausal women participated in the study.
The WHI was a massive study and the data collected has been used to publish numerous papers. I have used this data from a different paper to demonstrate the limited effect of a low-fat dietary approach for weight loss.
In this publication they focused exclusively on the incidence (likelihood of developing) and mortality (death) from breast cancer. The women were randomized into two groups. One group was instructed to reduce their dietary fat to 20% of daily calories and eat more vegetable, fruits and grains. The second group was instructed to eat their “usual diet.” The low-fat group, known as the dietary intervention (DI) group was also provided counseling.
They went to 18 group sessions within their first year and had a solo counseling session to verify that they were eating as recommended. After that they went to quarterly group sessions for the remainder of the intervention. To determine what they ate they used a 4-day food record at the beginning of the study and used this 4-day data to make recommendations on their diet. They then used food frequency questionnaires after the first year and every 3 years after that. This likely meant that each participant was screened 3 times during an average 8.5 year time span.
The results are compelling at first glance! The authors report a significant reduction in mortality from breast cancer favoring the low-fat group. They use a hazard ration (HR) to statistically prove their point. The HR for this association is 0.78 or 0.67 depending on how they explain it. A hazard ratio under 1 means that the intervention group has less chance of dying from breast cancer over the time observed. At first glance this looks like a 33% reduction in death from breast cancer. However, this must be taken into context. There are several reasons why we should question this finding!
First, a hazard ration must be viewed in context with its confidence interval (CI). This calculated term means that the researchers can only be certain that the study findings could be extrapolated to the larger population within a certain range of hazard ratios. For example, the 0.67 HR has a CI of 0.46 to 0.97. That means that they can only be certain that the HR is true between that range. An HR of 1 would mean there is no difference and 0.97 has little clinical significance. So statistically there is very little certainty that this low-fat dietary approach made a difference. Let’s look at some of the bigger issues in the study.
The intervention was a low-fat diet compared to essentially a standard American diet. We know clearly that any deviation from highly process, high sugar, high carbohydrate and high inflammatory fat diet will result in improved health. There is no arm of this study looking at other eating approaches. Additionally, the participants did not achieve their 20% dietary fat goal and averaged 24%, still low-fat by our standards but only an 11% reduction from what they were eating before the study began.
The intervention of nutrition counseling and observation may have played a larger role in their improved health. The collection of data also presents a big problem. We don’t actually know what they ate. Food frequency questionnaires are notoriously inaccurate especially at 3-year intervals. Think back over the last 10 years. Could you give me an accurate description of what you ate?
Lastly there was no significant difference in incidence of breast cancers. Each group was similar in that number. Interestingly the cancer types and pre-existing risk factors were not the same between groups. The DI group had less aggressive cancers with a better prognosis. They also had less previously documented heart disease. Is it surprising that they had a better outcome? The deaths recorded included not only those from breast cancer but also from other causes including heart disease.
This study was published by an author that strongly believes in improving the standard American diet. We all probably agree that the standard American diet needs improvement, but with what approach is the question. This study, unfortunately, is not convincing in the use of a low-fat diet to improve death rates from cancer. There is much more compelling clinical data that I will present in a future post supporting different dietary approaches for those fighting cancer.
Douglas E Lucas, DO
Chief Science Officer
PHD Weight Loss and Nutrition