Red Meat & Breast Cancer: Dietary Protein Sources in Early Adulthood and Breast Cancer Incidence | 22% Risk Increase for Red Meat Eaters, Substituting Poultry Normalizes Risk

Red (meat) breast cancer alert!

It's not the first study and I am pretty sure it's not going to be the last study to link red meat and cancer, but in view of the fact that I am pretty sure that the results Maryam S Farvid and colleagues present in their latest paper in the British Journal of Medicine are going to be all over the place this week, I feel that it's worth to give you an unbiased overview of the results before you are confronted with the sensational press release celebrating the newest "Harvard science" - a source people trust, one that "propagates the truth" and one that is (ab-)used by press release writers to generate the impression that each and every word they write is true.

Well the truth is that we are dealing with yet another prospective epidemiological study which does not have the power to reveal causal links between parameter (a), in this case the dietary protein sources in early adulthood and parameter (b), which is the incidence of breast cancer.
The average ignoramus will still read the headlines as "red meat" causes cancer and think of poultry, fish, eggs, legumes, and nuts as "the cure". In contrast to the red meat intake which was associated with a 22% risk increase in the 2830 documented cases of breast cancer the scientists had been collecting and following for 20 years, the a higher intake of poultry, fish, eggs, legumes, and nuts was
not just unrelated to breast cancer, in postmenopausal women, a high poultry intake was even associated with a -27% reduced breast cancer risk.

Energy intake and cancer risk expressed relative to lowest intake quintile for red meat (Farvid. 2014)

The latter observation gives rise to one of the (imho) hilarious substitute this for that equations, where the "estimating the effects of exchanging different protein sources, substituting one serving/day of legumes for one serving/day of red meat was associated with a 15% lower risk of breast cancer among all women (0.85, 0.73 to 0.98) and a 19% lower risk among premenopausal women (0.81, 0.66 to 0.99). Substituting poultry for red meat was even associated with 17% and 24% lower breast cancer risk in all and postmenopausal women.

Unlike the fooled readers of the press release, the researchers are obviously aware of the weaknesses of the study, in the discussion of the results, Farvid et al. point out that "potential limitations need to be considered":

• participants were predominantly white, educated US adults, they cannot determine whether our findings are generalizable to other race or ethnic groups
• dietary intake was assessed by food frequency questionnaires, some degree of measurement error is inevitably present, and thus to reduce measurement error they used the cumulative average of
  multiple measurements in a sensitivity analysis
• residual confounders are always of concern in any observational studies; although they adjusted for a wide range of potential confounders for breast cancer, they still could not rule out the possibility that other unmeasured or inadequately measured factors have confounded the true association
• they only estimated the effects of substitution of legumes, poultry, and other protein sources for red meat on risk of breast cancer, when trials on dietary modification would be ideal to support these substitutions

In addition, the scientists made multiple comparisons (different food groups and nutrients, premenopausal and postmenopausal subgroups, and subtype of tumors) in this analysis, and can thus not exclude the possibility of type I errors. But (sarcasm) this is not so much of concern, "the central finding of an association with red meat was [after all] a prior hypothesis." (Farvid. 2014) - In other words: What do you want people, we've just made sure we confirm our hypothesis.
Reference:

• Farvid, et al. "Dietary protein sources in early adulthood and breast cancer incidence: prospective cohort study." BMJ 2014;348:g3437 doi: 10.1136/bmj.g3437