The recent systematic review of epidemiological studies of the associations between long-chain (LC) n-3 PUFA and the risk of several cancers highlights the need for additional prospective studies that use valid and unbiased measures of dietary intake⁽Reference Gerber¹⁾. Specifically, it was noted that additional prospective studies using blood biomarker-based assessments of LC n-3 PUFA exposure would minimise the measurement error inherent in all measures based upon dietary recall. We agree with this. The conclusions of the review of the associations between LC n-3 PUFA and prostate cancer risk are based on two case–control and four prospective studies; five out of six of these studies used retrospective dietary assessment using FFQ. Missing from the review, however, are findings from several prospective biomarker studies⁽Reference Crowe, Allen and Appleby^2–Reference Park, Wilkens and Henning⁴⁾ including the two largest blood biomarker studies published to date on the topic⁽Reference Crowe, Allen and Appleby^2, Reference Brasky, Till and White³⁾. In a small (n _cases 376) case–control study nested within the Multiethnic Cohort, no association was reported between LC n-3 PUFA and prostate cancer risk⁽Reference Park, Wilkens and Henning⁴⁾. However, in a much larger case–control study (n _cases 962) nested within the European Prospective Investigation into Cancer and Nutrition, a 31 % (relative risk (RR) 1·31, 95 % CI 0·96, 1·81) and 39 % (RR 1·39, 95 % CI 1·02, 1·90) increase in prostate cancer risk in the highest compared with the lowest quintile of plasma phospholipid EPA and DHA was seen, respectively⁽Reference Crowe, Allen and Appleby²⁾. There was a 100 % increased risk (RR 2·00, 95 % CI 1·07, 3·76) of high-grade cancer contrasting the highest to lowest quintiles of EPA⁽Reference Crowe, Allen and Appleby²⁾. In a case–control study nested within the Prostate Cancer Prevention Trial (n _cases 1658), which was a unique study because the presence or absence of prostate cancer was determined by biopsy for all participants, a 150 and 99 % increased risk of high-grade prostate cancer in the highest v. the lowest quartile of serum phospholipid DHA (OR 2·50, 95 % CI 1·34, 4·65) or EPA+DHA (OR 1·99, 95 % CI 1·08, 3·68), respectively, was found, although there was no association for EPA alone⁽Reference Brasky, Till and White³⁾. Although the findings from the Prostate Cancer Prevention Trial were based on screen-detected cancers, which identifies cases that might never have become clinically relevant, the strength of the associations for high-grade cancer suggests that the findings are indeed clinically relevant.

The addition of these prospective and biomarker-based studies to the studies reviewed recently⁽Reference Gerber¹⁾ would certainly modify the conclusion on whether or not LC n-3 PUFA could reduce the risk of prostate cancer. Indeed, the findings from these biomarker-based studies challenge the generally accepted notion that increasing consumption of foods high in LC n-3 PUFA uniformly reduces chronic disease risk⁽Reference Moyad⁵⁾.