By Dr. Geoffrey Modest
The PREDIMED study was a large dietary intervention trial which assessed different diets and looked at the subsequent development of invasive breast cancer (see JAMA Intern Med. 2015;175(11):1752-1760). This was a single-blind randomized controlled field trial done in Spain from 2003-2009 looking at several different outcomes (see below). Breast cancer was a prespecified secondary outcome in women without a prior history of breast cancer.
Details:
- 4282 women aged 60-80 at high cardiovascular risk (mean age 67.7, BMI 30.4, 3% on hormone replacement therapy) were randomized to a Mediterranean diet supplemented with EVOO (extra-virgin olive oil), Mediterranean diet supplemented with mixed nuts, or a control diet with advice to decrease fat intake (the Mediterranean diet groups got free supplementary foods; after the first 3 years the researchers became more aggressive in outreach to the control group in an attempt to make the contacts from researchers more equal) and were followed a median of 4.8 years.
- Achieved diet composition was pretty similar for total calories (approx 2000/d), % protein, % carbs, % fiber, % polyunsaturated fats, though there were strong trends for increased monounsaturated fats in both Mediterranean diets, a-linolenic acid in the nut diet, % EVOO in the EVOO diet (22%, vs 12.5% in the nut diet and 8.8% in low fat, but there was concomitant different % refined olive oil in the EVOO diet of 0.8% vs 5.9% in the nut diet and 6.7% in low fat: i.e., the total % of energy from olive oils was actually not so different between the groups). There was a pretty similar consumption of saturated fats (9.8% of energy in each of the Mediterranean diets and 8.8% in the low fat one).
Results:
- 35 confirmed cases of invasive breast cancer
- 1/1000 person-years in those on Mediterranean diet with EVOO
- 8/1000 person-years in those on Mediterranean diet with mixed nuts
- 9/1000 person-years in those on the low fat diet
- Multivariable-adjusted hazard ratios:
- Mediterranean diet with EVOO vs control: HR 0.32 (0.13-0.79)
- Mediterranean diet with nuts vs control: HR 0.59 (0.26-1.35) [ie, not significant]
- Analyzing by annual cumulative updated dietary exposure: each additional 5% of calories from EVOO: HR 0.72 (0.57-0.90)
I have sent out a few previous blogs on the PREDIMED diet: one showing that the Mediterranean diet supplemented with EVOO was associated with lower likelihood of developing atrial fibrillation (see https://stg-blogs.bmj.com/bmjebmspotlight/2014/06/24/primary-care-corner-with-geoffrey-modest-md-olive-oil-and-atrial-fibrillation/ ), one showing that this diet was associated with primary prevention of coronary artery disease (see https://stg-blogs.bmj.com/bmjebmspotlight/2014/01/21/primary-care-corner-with-geoffrey-modest-md-mediterranean-diet-and-cad-primary-prevention/ ), and one finding lower likelihood of developing diabetes (see https://stg-blogs.bmj.com/bmjebmspotlight/2014/01/21/primary-care-corner-with-geoffrey-modest-md-mediterranean-diet-for-diabetes-prevention/ ).There was another PREDIMED study finding improved cognition in those on either of the Mediterranean diets (see J Neurol Neurosurg Psychiatry 2013;84:1318–1325)
So, this present study does raise several issues:
- There were a small numbers of events, so the results need to be confirmed by larger, longer-term RCTs.
- The amount of EVOO consumed was pretty significant: there was a statistical benefit only in those consuming >15% of their cumulative energy intake from EVOO and increasing benefit with higher intakes.
- It is unclear from this study design whether the benefit was from the high mono-unsaturates associated with olive oils overall, or to the higher concentration of polyphenols in EVOO (though the actual achieved olive oil consumption, as noted above, supports more the role of EVOO itself, since the actual total % of energy from all olive oils was pretty similar between the groups). There are some other data suggesting that some of the olive oil monounsaturates are associated with antiproliferative effects on human oncogenes and with decreased DNA oxidative damage. But olive oil polyphenols from EVOO have some effect in “inhibition of tumor growth and proliferation, migration, and invasiveness of breast cancer cells in vitro and in vovo breast cancer models” as well as increased apoptosis of cultured breast cancer cells.
- One important point here is that this was a randomized intervention trial. Prior epidemiologic studies have found mixed results on the role of fat in increasing breast cancer risk, but these have been observational studies (the data are more consistent for alcohol as a “dietary” risk factor). The choice of the Mediterranean diet in PREDIMED was because of observational data that heart disease and breast cancer, for examples, are lower in Mediterranean countries than in the US or Northern/Central Europe. A secondary outcome of the old Lyon Diet Heart Study, a 4 year study of 605 patients with coronary heart disease, had found a 56% risk reduction for total deaths and a 61% risk reduction for cancers, though not specifically breast cancer, in those on Mediterranean diet vs low fat diet (see Arch Intern Med 1998; 158:1181).
- But this study does raise, I think, the important point that a healthy diet has an array of benefits throughout the body. We in medicine are often so focused on one part of the body (should we lower the cholesterol to prevent heart disease?) that we may miss the big picture (in this case, the Mediterranean diet seems to be good for the heart, the brain, the pancreas, etc.). And this myopia in part leads to the finding in patients who take statins that they actually eat worse than before the statins (well, after all, the patient is taking a statin, their lipids are pristine, so all is good) — instead of our continuing to reinforce the multitude of known (and unknown but likely) benefits of eating healthfully.