Fluke or Trend? Omega 3 Fatty Acids vs Prostate Cancer

When it comes to new, anomalous findings, it's best not to change medical strategies abruptly but rather to wait for confirmation.  In other words, we look for trends, rather than change on a whim based upon a fluke.

Case in point is the relationship between omega 3 fatty acids (most commonly found in fish) and prostate cancer.  Just last week, a case control study was published linking high levels of docosahexaenoic acid (DHA) to prostate cancer, quite contrary to what numerous previous studies have concluded.  The authors found 1,658 participants from the Prostate Cancer Prevention Trial who had actually developed prostate cancer over 7 years of observation and compared them to 1,803 controls matched for age, treatment & family history.  They then analyzed serum omega 3 fatty acids, omega 6 fatty acids, and trans-saturated fatty acids.  Because of all the well known benefits associated w/omega 3 fatty acids, the authors were not expecting to find DHA associated w/high grade prostate cancer.

But as one of my patients asked me, does this study mean that he should stop taking his (high dose) fish oil supplements?  In other words, is this finding a fluke or a trend? It turns out that back in November, three other authors concluded that fish consumption was associated w/63% reduction in prostate-specific mortality after performing a meta-analysis of 12 case-control studies involving 5,777 cases & 9,805 controls.  Similarly, an April 2009 case-control study demonstrated that consumption of dietary omega 3 fatty acids was associated w/decreased risk of aggressive prostate cancer after comparing 466 cases of aggressive prostate cancer to 478 age- & ethnicity-matched controls. An even earlier case-control study published in July 2007 also came to a similar conclusion that higher serum omega 3 fatty acids are associated w/lower risk of prostate cancer after comparing 476 cases to 476 controls.

So what are we to make of this latest study?  First, case-control studies are relatively weak forms of evidence upon which to base medical therapy.  Just like observational & epidemiologic studies, they're good for demonstrating association but not causation.  Therefore, they're most useful in generating hypotheses rather than directing therapy.  And given the preponderance of the data, I would posit that this latest finding is more fluke than trend.  Personally, I'm sticking w/my fish oil supplements.