Omega-3s and prostate cancer risk

Omega-3s and Prostate Cancer Risk

Review of:
“Plasma Phospholipid Fatty Acids and Prostate Cancer Risk in the SELECT Trial” Theodore M. Brasky, Amy K. Darke, Xiaoling Song, Catherine M. Tangen, Phyllis J. Goodman, Ian M. Thompson, Frank L. Meyskens Jr, Gary E. Goodman, Lori M. Minasian, Howard L. Parnes, Eric A. Klein, Alan R. Kristal

Published ahead of print in Journal of the National Cancer Institute, July 10, 2013.

As many of you know, a study was published earlier this week linking omega-3s to an increased risk of prostate cancer. GOED has reviewed the study an prepared the analysis of the scientific inaccuracies and study design issues below. In addition, Adam Ismail wrote a more consumer-friendly blog column that was published this morning on Virgo's Omega-3 Insights page. We thought this would be useful for GOED members to share with concerned customers and consumers.

GOED Comment on Experimental Design and Results

The test cohort was the same as used for the abandoned SELECT trial (Selenium and Vitamin E Cancer Prevention Trial) enlisting men over the age of 55 from the USA, Canada and Puerto Rico from 2001 to 2004. The cohort was relatively healthy and intelligent in relation to the general population. Fatty acids were analyzed from a single blood serum sample taken on enlistment. The absolute individual fatty acids were not reported; rather they were expressed as % of total fatty acids.
The study used a case-cohort design nested with the SELECT cohort. Cases of prostate cancer were reported as low or high grade according to defined clinical parameters.
The incidence of prostate cancers in relation to individual blood serum fatty acids was assessed by estimating hazard ratios using a Cox proportional hazard (CPH) statistical model. In such models, the covariate is multiplicative with respect to the hazard rate. This model assumes a linear response with time elapsed.
Only total omega-3 PUFA (EPA +DHA +DPAn-3) gave a statistically significant association (p< 0.05). The hazard ratios (95% CI) are given in the table below:

Omega 3% of FA Low grade prostate cases(n) Hazard ratio for Low grade High grade prostate cases(n) Hazard ratio for High grade
<3.68 146 1 26 1
3.68-4.41 159 1.1 35 1.39
4.42-5.31 176 1.26 52 1.87
>5.31 203 1.44 43 1.71
Using the CPH model and assuming a 50% increase in risk with time, the authors suggest that omega-3 PUFA’s are associated with a 22-25% prostate cancer risk.

GOED Takeaways

While any recommendation for supplementation should consider any associated risks, we believe the authors' conclusion is irresponsible and blatantly ignores the totality of the scientific evidence that has been collected over multiple decades regarding the health benefits of marine omega-3 fatty acids.

  • This study was not specifically designed to look at the exact relationship between omega-3 fatty acid intake and prostate cancer.
  • The difference in mean blood plasma phospholipid fatty acids blood level for omega-3s was 4.66% in the combined cancer group versus 4.48% in the control. They are basing their results on just ca. 0.2% difference in omega-3 levels.
  • If the findings were true, then prostate cancer would be rampant in any country with high seafood consumption (Scandinavia, Japan etc) and conversely, low level consumption should be protective. Clearly this is not the case.
  • Plasma phospholipid fatty acids as measured in this study are not a good index of long term intake and are influenced dramatically by a single meal, or even timing of a fish oil dose. A single fish oil dose massively increases LC omega 3 (typically increasing levels by 100% or more) in about 4-12 hours and then washes out around 48 hours.
  • Absolute serum levels of EPA, DHA, DPAn-3 were not reported.
  • No documentation was provided in the paper regarding intake of fish or fish oil in the study group.
  • The study was not designed to look at omega-3 and confounded with selenium and Vitamin E used in the treatment arms.
  • It is very possible the statistical model used (Cox proportional hazards) was not appropriate for this study. This model is suitable for a drug taken at the same time every day or exposure to a toxin throughout the day at the same level, but not for something like fish or fish oil intake where the levels in blood serum will vary considerably depending on food and supplement consumption patterns.
  • The association between individual omega-3 fatty acids, EPA, DPA and DHA and prostate cancer was NOT statistically significant within the CPH model.
  • The DPA:high grade data violated the CPH model assumptions (as did 3 other fatty acids associations).
  • The so called “meta-analysis” of earlier studies carried out at the end of the paper actually only included three previous studies, one of which was his own (Brasky, 2011). Another by Park et al. (2009) used the same nested case control design, while the remaining study by Chavarro et al (2007) showed a strong benefit for marine omega 3 fatty acids reducing the risk of prostate cancer.
  • In a similar study two years ago, the same lead author (Brasky, 2011) used the same methodology, though a different cohort, to attempt to find a link between omega-3’s and prostate cancer. In this case, no association with EPA was found, and though a real association with DHA was found, it was not linear with time/dose dependent and applied to high grade tumors only. The actual mean DHA levels, reported as a % of total serum fatty acids, for the control group, low grade prostate cancer group and high grade prostate cancer group respectively were 2.84, 2.89, 2.99 – an absolutely minimal real difference.
  • The test cohort included sick and healthy people. It is possible that sick people were taking fish oil supplements at a higher rate than the healthy individuals.

What Else Should You Know?

  • A recent meta-analysis of fish consumption and prostate cancer by Szymanski et al. (2010) reported a large reduction in late stage or fatal prostate cancer among cohort studies.
  • Several population based studies have shown a benefit of increased omega-3 fatty acid intakes to reducing prostate cancer risk (Lietzman et al., 2004, Terry et al., 2001).
  • A recent meta-analysis by Zheng (2013) of 16 independent cohort studies reported the association between marine n-3 PUFA (reported as intake or biomarker data) and risk of breast cancer, involving 16,178 breast cancer events and 527,392 participants. Marine n-3 PUFA was significantly inversely associated with risk (relative risk 0.86, 95% confidence interval 0.78 to 0.94). Dose-response analysis indicated that a 0.1g/day increment of dietary marine n-3 PUFA was associated with 5% lower risk of breast cancer.

Suggested Citation

Global Organization for EPA and DHA Omega-3s (2013). Omega-3s and Prostate Cancer Risk [Peer commentary on the paper “Plasma Phospholipid Fatty Acids and Prostate Cancer Risk in the SELECT Trial” by Brasky T.M. et al, 2013, JNCI, doi:10.1093/jnci/djt174.].


Brasky TM, Till C, White E, et al., Serum phospholipid fatty acids and prostate cancer risk: results from the prostate cancer prevention trial, Am J Epidemiol. 2011;173(12):1429–1439.
Chavarro JE, Stampfer MJ, Li H, et al. A prospective study of polyunsaturated
fatty acid levels in blood and prostate cancer risk. Cancer Epidemiol Biomarkers Prev. 2007;16(7):1364–1370.
Lietzman MF, Stampfer MJ, Michaud DS, et al. Dietary intake of n-3 and n-6 fatty acids and the risk of prostate cancer. Am J Clin Nutr 2004;80:204-216.
Park SY, Wilkens LR, Henning SM, et al. Circulating fatty acids and prostate cancer risk in a nested case–control study: the Multiethnic Cohort. Cancer Causes Control. 2009;20(2):211–223.
Szymanski KM, Wheeler DC, Mucci LA. Fish consumption and prostate cancer risk: a review and meta-analysis. Am J Clin Nutr. 2010;92(5):1223–1233
Terry P, Lichtenstein P, Feychting M, et al. Fatty fish consumption and risk or prostate cancer. Lancet, 2001,357:1764-6.
Ju-Sheng Zheng, Xiao-Jie Hu, Yi-Min Zhao, Jing Yang, Duo Li . Intake of fish and marine n-3 polyunsaturated fatty acids and risk of breast cancer: meta-analysis of data from 21 independent prospective cohort studies. BMJ 2013;346:f3706 doi: 10.1136/bmj.f3706