This article is a rebuttal to the recently publicized negative study offish oil and prostate cancer compiled by a panel of Nutri- West experts: Lynn Toohey, PhD; Don Bellgrau PhD (NIH grant reviewer); Dan Murphy D, DABCO (years of research on omega 3s; part time faculty, Life Chiropractic College West), Brandon Lundell, DC, DABCI instructor, Assistant Professor, So. Cal. U of Health Sciences.

In a word: fiction. Recently, headlines touted a negative association of fish oil and prostate cancer (Brasky et al. 2013)', based on an observational study (not a clinical trial proving cause and effect) that analyzed (after the fact) previous data taken from a publication that did not adjust for what they call confounding factors - other factors that can cause prostate cancer The conclusions that they drew from their observations have been rebutted by experts in the field, and called everything from "silly" to "irresponsible". An association was not only mistakenly drawn, it was given wide media coverage, mostly because it goes against everything we have been told is true (based on good research) that fish oil is good for us. This misinterpretation of facts certainly does not negate all of the thousands of positive articles on fish oil, which include well-designed, randomized placebo controlled CLINICAL TRIALS, in which participants are actually given fish oil and then results are measured.

The Science of Interpreting Studies: It is taught in every first year doctorate program that when you read a study, in order to avoid drawing false conclusions, one must remember the axiom: Correlation Does Not Imply Causation. In this study, the flawed correlation of prostate cancer with Omega 3's in the blood does not in any way mean that Omega 3's cause prostate cancer. Even if a correlation did exist (which is unlikely as you will see below) the Omega 3's could be elevated because of the cancer and not the other way around. Cancer is known to alter metabolism in various ways, including fatty acid metabolism. It is more conceivable, given the preponderance of evidence of beneficial effects of Omega 3's, that cancer induces inflammatory chemicals which in turn release fatty acids into the blood stream. This could be a host —defense mechanism, or a tactic on part of the cancer cells to supply ample energy for tumorigenesis. In any case, one must remember the axiom "correlation does not imply causation".

This study in question was riddled with many flaws, including the fact that they did not measure fish oil intake, or even fish oil supplements for that matter (and we all know that there is a difference between farm raised fish and wild caught salmon!) What they measured was the level of fatty acids in the blood for one particular moment in time, a measurement that is known to have wide day-to-day variability. The omega-3 index, which measures both EPA and DHA within red blood cells, is a much more accurate indicator of long-term omega-3 intake and tissue status than is the plasma omega-3 level, which is subject to significant day-to-day variability. They also failed to account for body weight, a factor highly associated with prostate cancer (and more of the prostate cancer group was overweight). It is suggested that fat cells release inflammatory cytokines, and increased inflammation is a suspected cause of prostate cancer. Many factors contribute to omega 3 blood levels, including other dietary factors, metabolism, genetics etc., and some of the things not taken into consideration included that more than 50% of the prostate cancer group were smokers, consumed alcohol, and were overweight.

Anthony Victor D'Amico, MD, PhD, who specializes in prostate cancer research, notes that while the study adjusted for diabetes and family history, they left out huge factors associated with prostate cancer, such as race, age, PSA level, rectal exam, body mass index, and whether or not they were overweight (already mentioned to be highly correlated). According to Dr. D'Amico, unless you take these things into consideration, "You could associate almost anything with prostate cancer; something like "driving a Cadillac" could be associated with prostate cancer if you don't account for the major factors that cause prostate cancer." Dr. D'Amico also finds it interesting that Table one in the study lists the PSA level and race but those factors were not adjusted for in the model, "maybe because the association was lost once they did."

Even if the blood measurement of fatty acids was a gold standard, which it is not, the clinical relevance of such small differences (.04% EPA and 0.1% DHA) in each of the groups remains unknown. Dr. Don Bellgrau, an NIH grant reviewer, emphasizes that the assay is quite cumbersome and prone to error, for which they provide no data regarding validation. It involves extracting total lipids from plasma followed by separation of other lipids by chromatography. Most biomarker people would say the more you do to the sample prior to analysis the more likely you are to add error to the process. They also use pooled serum from "healthy ^volunteers as an internal control for each assay but do not show any data regarding day to day variability when assaying the same samples. They also express fatty acids as a weight percentage of total phospholipid fatty acids, rather than an absolute value, and using percentages is likely to be misleading."

Dr. Bellgrau further explains how the authors of the article even stated within the article that analyzing in this manner could lead to spurious (false) conclusions. He points out that the authors admit that "because some fatty acids represent greater proportions of the total weight than others"(page 4), they did some statistical gymnastics to make the claim that this was a valid approach even though they state later that "spurious associations arise because of measurement of fatty acids as a proportion of total weight rather than absolute concentration" (page 6). This approach is likely what gave them their conclusion (potentially spuriously) because as they state on page 8 'expressing fatty acids as weight proportions could create spurious associations because an increase in the percentage of one type of fatty acid requires a decrease in others". Dr. Bellgrau emphasized that they should have compared by absolute weight, not by percentage.

Another expert who questions the validity of inferring causality from observations that don't take confounding factors into consideration is Michael Savage, PhD. Dr. Savage's PhD is in epidemiology from U of C Berkeley, and he interviews Anthony Victor