Is too much salmon increasing your risk for prostate cancer?

An article published on-line in the Journal of the National Cancer Institute appears to confirm earlier data suggesting that men with the highest levels of omega-3 fatty acids in their blood are at increased risk for a diagnosis of prostate cancer. It is getting lots of media coverage. However, …

As is carefully pointed out in comments about this article by Dr. Anthony D’Amico in a report on the HealthDay web site:

All of these studies on associations, which is what this is, are hypothesis-generating because they are looking back in time. It’s not a cause and effect.

Dr. D’Amico goes on to note that this study would have needed to account for other risk factors for prostate cancer before it could be considered definitive. Such other risk factors include things like family history, age, race, etc.

The new data reported by Blasky et al. are based on a retrospective analysis of data from the earlier SELECT trial that showed no benefit from taking either selenium of vitamin E supplements in the prevention of prostate cancer.

Based on the data generated by their analysis, the authors do clearly show that men in the group with the highest 25 percent of levels of omega-3 fatty acids (i.e., the highest quartile) were at relatively greater risk for prostate cancer than the men in the lowest quartile. Specifically, the men in the highest quartile had:

  • Heightened risk for low-grade prostate cancer (hazard ratio [HR] = 1.44)
  • Heightened risk for high-grade prostate cancer (HR = 1.71)
  • Heightened risk for any type of prostate cancer (HR = 1.43)

In contrast, men with relatively elevated levels of omega-6 fatty acids (linoleic acid) had:

  • Reduced risk for low-grade prostate cancer (HR = 0.75)
  • Reduced risk for any type of prostate cancer (HR = 0.77)

As Dr. D’Amico so accurately observes, it is very difficult to know what to make of these data from a clinical perspective. The “New Prostate Cancer InfoLink would suggest that what this study may be telling us is that excessive reliance on any one type of dietary component or supplement is not the best idea in the world — for all sorts of reasons.

Omega-3 fatty acids are taken in two predominant forms: as actual supplements (often by men with a history of heart disease) and through the consumption of oily fish like salmon, trout, and fresh tuna. Men who are at known risk for prostate cancer might want to be careful about excessive consumption of these types of oily fish several times a week. Equally, men who are taking high levels of omega-3 fatty acids may want to discuss the possible implications of this study with their doctors — particularly if they are at known risk for prostate cancer because of things like race or family history.

It would require a very long and expensive prospective trial to resolve whether there is any likelihood of a “real” effect of omega-3 fatty acid levels on risk for prostate cancer, with careful comparison of the risk factors of all the patients involved in the trial. Frankly, we don’t see anyone being able to invest in such a trial in the near future. So the sensible thing for men at known risk for prostate cancer — and the rest of us too — is probably just to not consume excessive amounts of omega-3 fatty acids, but to cut back our consumption into a reasonably “normal” range.

21 Responses

  1. As regards D’Amico’s comment re associations, frankly I read very little on this site or anywhere else regarding prostate cancer that is anything other than statistical association. Why is this pointed out only when the association is apparently disappointing?

  2. Dear Setrack:

    We can’t speak for other sites, but we have quite certainly pointed out regularly on this site occurrences of positive associations that need to be treated with equal caution.

  3. There are multiple studies showing benefits of fish/fish oil. (Click here to see “Dr. Geo’s take”.)

  4. Indeed you do and I obscured my main point to wit with all the research that is evidently conducted we seem very much in the dark regarding cause and effect. Much treatment seems to be based on speculative and often spurious statistical correlations. Best there is I suppose.

  5. This is true, which is exactly why it is hard to know what to make of the studies that show evidence of risk.

  6. The basic problem is that we still have no real clarity about: (a) why many specific cancers are initiated in particular individuals; (b) how cancers actually spread and metastasize; or (c) whether it is really possible to actually stop some of them from growing. This is true even in the case of a cancer like testicular cancer that we can actually cure in many men — even after metastasis.

    One of the fundamental problems with medicine is that when you don’t know why things happen, it can be pretty hard to do things to change the course of a disease with a high level of confidence … leaving us with statistical probabilities as opposed to “Yes/No” answers. Think of the similarities to the way we treated infections prior to our understanding that there were infectious organisms like bacteria and viruses!


    For years many of us prostate cancer patients have diligently consumed fish and fish oil or other marine-based supplements to increase our omega-3 fatty acids, while also trying to reduce omega-6 in our nutrients. We have been aware that there is a sharp difference between plant- and fish-based omega-3s, with the likelihood that plant-based omega-3 was harmful and fish-based beneficial.

    This latest paper is alarming. At first I thought the researchers might have studied plant-based omega-3s, such as those found in flaxseed oil, which many of us have long considered harmful to prostate cancer patients. However, the abstract, in its background section, specifically mentions several fish-based omega-3s as raising concern in recent studies.

    On the other hand, the abstract does not clarify whether those fish-based omega-3s were the omega-3s analyzed in the study, or whether all omega-3s were lumped together. This is a key point, as, during the period of the SELECT trial, flaxseed oil was a popular choice for men concerned about prostate cancer. If the researchers in this study were picking up all omega-3s, that would undermine their findings, in my view. I would also like to know whether the team looked at plant-based omega 3s in the SELECT cohort patients.

    Statistical concern: As is clear from the abstract, the statistical confidence intervals are rather wide. That suggests that the distribution of data points indicates chance could be playing a large role in these results. (No p values, which provide insight into the likely role of chance versus a true finding, are provided.) Also, the abstract does not give us the absolute risk levels. Therefore, we don’t know whether the hazard ratio of 1.71 for high-grade prostate cancer, for instance, applies to a very low baseline risk for the low omega-3 group or a higher risk. If the absolute risk is very low, then the relative risk reflected in that eye-catching hazard ratio of 1.71 would not mean so much.

    I am concerned that, as indicated in the cited news article, the research team did look at actual consumption of fish and fish oil, finding that: “The investigators found that men eating the most fatty fish and taking the most fish oil supplements had an overall 43 percent increase in risk for all prostate cancer, compared with men eating the least fish or taking the fewest supplements. The risk for aggressive prostate cancer was 71 percent higher; for non-aggressive prostate cancer, the risk was 44 percent greater.”

    There is one conclusion I will state with great confidence: we have not heard the last about this!

  8. Dear Jim:

    I think you are over-analyzing the abstract. Why not just contact the authors and ask for a copy of the paper?

  9. The discussion doesn’t mention two other related comments:

    (1) The study also doesn’t say anything about the effects of fish oil on men who already have cancer. “This study is not about men with prostate cancer,” Brasky said, noting that some studies have suggested fish oil might be beneficial in men who already have cancer.

    (2) But, according to a report on the NBC News web site, the researchers point out that recent studies have shown taking extra omega-3 has little effect on heart disease — including a study published in the New England Journal of Medicine in May.

    The first statement is important for most of the readers here, as they already have prostate cancer. The second statement seems important for everybody and suggests that there are some exceptions to “a heart healthy diet is also a good prostate cancer diet” — for men who already have the disease.

  10. Dear Wolfram:

    This is all debatable. I think this study is hypothesis-generating as opposed to definitive. Remember that this study is exclusively limited to men diagnosed with high-risk disease, and we know nothing at all about their other risk factors or what the rest of their diets were like.


    I just received an electronic copy of the complete paper from the outstanding library staff at my nearby large, community hospital. I’ll be going over it shortly.

    In addition to the concerns I raised previously, here’s another, perhaps the most important. I’m stating it before scrutinizing the paper as a way of presenting a hypothesis for mental testing. The SELECT cohort was selected some years ago, and likely were men more concerned about preventing than the average man based on the fact that they volunteered to participate in a clinical trial. They also were likely disposed to the idea that nutritional lifestyle tactics were probably worthwhile in preventing and coping with prostate cancer. Once diagnosed, especially if diagnosed with high-risk disease, I believe these men would have reacted as I did after the turn of the year in January 2000: pull out all the stops to try to stay otherwise healthy and survive. I’m thinking they would have consumed fish and fish oil, as well as other nutrients considered possibly beneficial.

    There may be ways to tease this out. If that hypothesis is correct, then we would expect to see signs of the presence or absence of other nutrients considered potentially helpful at the time to be elevated. For instance, the ACS published a suggested listed of recommended lifestyle tactics for those diagnosed with prostate cancer under the acronym CASTLES at that time (received my copy in early 2000), which stood for avoiding Cigarettes (and smoking), avoiding Animal fat, consuming Soy and Selenium supplements — specifically stated “supplements,” drinking green Tea, consuming Lycopene, consuming vitamin E and Exercising, and getting Screened for colorectal cancer. (Source: Dr. Gabriel Feldman, former Director of Prostate and Colorectal Cancer at NHO; all of the recommendations used the word “may” to indicate the likelihood of help, except for “should” regarding colorectal cancer.). As I mentioned earlier, flaxseed oil was popular among some men concerned with prostate cancer, and therefore ALA (alpha-linolenic acid) may also have been elevated.

    Okay, on to the actual paper.

  12. Dear Jim:

    (1) In my experience most men don’t come near to your level of obsession about trying to manage their health. Maybe they just don’t have the time, or maybe they just want to get on with life without trying to understand every detail about prostate cancer.

    (2) You are making a lot of other assumptions that need to be eliminated from your analysis of the available data. When one makes assumptions like this, one stops looking impartially at the actual data because one is trying to make the data fit one’s assumptions. Please try to let the assumptions go.

  13. I have not been diagnosed with prostate cancer but I am at an elevated risk because my father died of prostate cancer and because I have always had a high PSA for my age — though it has been pretty stable (around 1.5 ng/ml since age 40, with relatively minor deviations). I have been taking fish oil, in addition to eating a very healthy diet, as part of an effort to minimize my risk of prostate cancer, and just generally to stay healthy. I was taking about 1.1 g per day, though in the last year I have been taking only half that amount. I also eat fish several times a week.

    Dr. Strum recommended fish oil as part of a prevention/health regimen, on P2P, and I read some other articles that suggested it, including one by a Dr. Charles “Snuffy” Meyers, though that one may have been focused on people who already have prostate cancer. My EPA and DHA levels are pretty high (EPA is over 300 nmol/ml; AA/EPA ratio about 2:1 and EPA about 4% of total fatty acids). Now I am very concerned that I made a mistake and have actually increased my risk. I am losing sleep over this. (Part of the problem, to be perfectly honest, is that I have an anxiety disorder, so I tend to worry about things more than most people.) But, even so, sounds like I may have made a mistake.

  14. Dear Jonathan:

    So this is a case in which your anxiety disorder may be more of a problem than your diet or the fish oil.

    We do not know anything here is any real certainty. This type of retrospective analysis of data collected for a quite different study is always difficult to interpret, and (as others have been quick to point out) there have, over the years, been a lot of other data strongly suggestive of the path that you have been following.

    It’s easy for me to say because I don’t have your anxiety disorder and there is very little that I ever lose sleep over, but I think you need to just go on doing what you have been doing and give science time to provide some rather better data on this topic.

    If you are really anxious about all this and feel the need to “just do something”, then maybe you could stop taking the fish oil 2 days a week out of 7. Give yourself Saturday and Sunday “off” from the fish oil, for example. It probably won’t make a huge difference one way or another from a prostate cancer risk point of view, but it might help your anxiety.

  15. Hi Jonathan,

    Like you, I am a heavy consumer of omega 3s (and limiter of omega 6s) in my diet, with the same objective: to help reduce the risk of prostate cancer. Unlike you though, I was diagnosed 13.5 years ago with a highly challenging case. Shortly afterward I became aware of the information favorable to marine-based omega 3s, and I started eating a lot more fish as well as consuming supplements, which I do to this day. Like you, I was jolted when I saw Sitemaster’s initial review of this paper.

    I am almost done with a fairly thorough response that documents a likely critical flaw in the study that renders the findings an artifact, not something to worry about. Probably. Your post could not have fit my thesis better, so I hope you won’t mind reappearing in what I write, hopefully posted later today. I think I’ve not only seen the smoke but found the smoking gun. I would be interested in your reaction to what I post.

  16. Jim,

    If it’s in the study, would you list the actual amount of marine omega-3s in the blood, in the study, in various quintiles ?

    Also, do they mention how that relates to marine omega-3s in the prostate tissues?



  17. Jim:

    I look forward to seeing your analysis. I heard an interview with Dr. Anthony D’Amico at Brigham and Women’s Hospital who expressed considerable skepticism about the study, largely because of its apparent failure to control for other potentially relevant factors. (But the interviewer did not ask Dr. D’Amico what he thinks about the idea of taking fish oil.) I can deal with bad luck, such as getting a disease. I find it harder to deal with the fact that I may be doing something in an effort to improve my odds than is actually making them worse.

  18. We have received Jim’s analysis of the paper in question and will post it as a “guest blog” entry later today once we have dealt with issues related to the accurate formatting of the tables Jim has included. We will also post links to the guest blog from this page and from the blog back to this page and the related discussions.

  19. Jim’s analysis of the full text of this paper is now available as a new blog post elsewhere on this site.

  20. There should be a real conclusion on weather omega 3s cause prostate cancer and make existing prostate cancer worse. You can also read that omega-3s do not cause prostate cancer or make it worse. Dr. Weil stated this,

  21. Dear Mr. Collins. Much as you might like proof of this concept one way or the other, it is near to impossible to achieve such proof because there are far too many other concurrent variables that can affect the result of such a study.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: