There’s been a lot of hullabaloo recently about a study concerning fish oil and prostate cancer. If the breathless reporting can be believed, consuming fish oil increased your chances of prostate cancer.
This could be true, but the study in question did not show that at all.
The Study In Question
This will get science-heavy. If it’s too heavy for you, just skip to the next section.
In short, this study initially looked at participants of the SELECT trial and got a sample of persons who were diagnosed with prostate cancer (n=834) and made note of how many had advanced cancer (n=156), then 1393 persons from the SELECT study who did not have prostate cancer were selected for comparison. The researchers then measured serum omega-3 fatty acids (EPA, DHA, and their intermediate DPA) and stratified the groups into quartiles to see if there was an association.
The results showed that persons who had prostate cancer were more likely to have higher circulating omega-3 fatty acid levels (excluding ALA, which was not associated) and that omega-6 was unrelated to prostate cancer. Trans fatty acids were mostly unrelated aside from a possible positive relationship with palmitelaidic acid (16:1). When comparing the quartiles against one another (the lowest quartile being set at 1.00 as a reference), the highest levels of fish based omega-3 fatty acids (collectively) was associated with increased risk as assessed by Hazard Ratio for low (1.07-1.40), medium (1.07-1.43), and high (1.00-1.54) grade prostate cancer. While DHA had an HR showing an association with low (1.07-1.37), medium (1.06-1.38), and high (1.03-1.54) grade prostate cancer, DPA was only associated with low (1.03-1.46) and medium (1.08-1.57) while EPA was not significantly associated with an increased risk. Alpha-linolenic acid, omega-6 fatty acids, and trans fatty acids were not related.
Note: The above ranges are known as the 95% confidence interval, and show the range of values that the researchers are 95% confident that the true value lies in. So for a range of 1.03-1.54 this means a possible 3-54% increase (with a 5% margin for error), and if the interval crosses 1.00 (the zero point where lower means reduced risk) then the observation is not considered significant
When adjusting for the omega-3 to omega-6 ratio, it seems that the HRs for low (0.98-1.36) and medium (0.97-1.36) grade prostate cancer became nonsignificant while there was still a significant relation with the high (1.40; 1.03-1.92) grade. The average value here, 1.40, may be where media sources are claiming a 40% increase in prostate cancer risk. The ’71%’ referenced in most publications was a direct comparison of the risk in the highest quartile against the lowest (with a confidence interval of 0-192%) with a 43% (9-88%) overall increase in risk.
Important Note: The actual association for people with high-grade prostate was in the range of 3% to 92% – this gets averaged to 40%. When comparing the highest quartile (25% of sample) to the lowest, the average risk was increased to 71% but became more variable at 0-192%.
The variables that were made note of in the analysis were education, history of diabetes, family history of prostate cancer, and SELECT intervention assignment (so, placebo or vitamin E groups). So despite the increased risk seen with vitamin E previously in SELECT it likely does not influence the results.
This study found that, when comparing the lowest 25% of subjects (assessed by how much fish oil was in their blood) against the highest 25% that the higher group had a higher frequency of prostate cancer at the time of measurement. They conducted a one-time measurement of blood lipids, and there was no supplemental intervention
What This Means
First of all, the study was a retrospective study. What that means is that it heavily surveyed a lot of people – about what they ate, their lifestyle, their diet, and so forth. The problem with such a study is that trusting people’s memories tends to be a bad idea – people’s recall of their dietary and fitness habits can be quite awful.
Thus, as this was a retrospective study, it was not an intervention with a control/placebo group and a drug group. What does this mean? Normally you want a double-blind trial – half your test group takes a placebo, and the other half takes your drug (in this case, fish oil). No one, not even the people running the test, know who is taking what. All things being the same, you then see if there were differences between the two groups. Without this kind of explicit control, it’s hard to know what caused what (cause and effect is very important). This is extremely critical as one of the recommended things to do when you have prostate cancer is to supplement with fish oil! To be fair, there are ethical considerations, so you cannot directly induce cancer either!
Most importantly, this study simply measured the amount of omega 3s in the bloodstream, not fish oil consumption itself!
The real question that this study poses is why is there higher omega 3s in people with more aggressive prostate cancer? Is it because people are supplementing with fish oil? Is it your body’s response to the cancer? Or is something else going on?
We don’t know. But to state that fish oil causes prostate cancer is about as lazy reporting as you can get (though it is equally wrong to say that fish oil does not cause prostate cancer).
Author: Sol Orwell
Sol is an author at examine.com