Oral selenium supplementation has no effect on prostate-specific antigen velocity in men undergoing active surveillance for localized prostate cancer

Abstract

The Nutritional Prevention of Cancer trial showed a 52% lower incidence of prostate cancer in men supplemented with selenium. As a result, our study was designed to assess whether selenium supplementation attenuates the progression of prostate cancer. A phase 2 randomized, double-blind, placebo-controlled clinical trial was conducted in men with localized nonmetastatic prostate cancer who had elected to forgo active treatment and be followed by active surveillance. A total of 140 men were randomized to placebo (n = 46), 200 microg/d (n = 47), or 800 microg/d (n = 47) selenium p.o. (as selenized yeast) and followed every 3 months for up to 5 years. Prostate-specific antigen (PSA) velocity was used as a marker of prostate cancer progression and was estimated using mixed-effects regression. Adjusting for age, body mass index, baseline selenium, smoking, baseline PSA, race, PSA method, and Gleason score, PSA velocities for the 200 microg/d and 800 microg/d treatment groups were not statistically significantly different from placebo (P = 0.32 and P = 0.61, respectively). In the highest quartile of baseline selenium, men supplemented with 800 microg selenium showed statistically significantly higher PSA velocity as compared with placebo (P = 0.018). Selenium supplementation did not show a protective effect on PSA velocity in subjects with localized prostate cancer. On the contrary, supplementation with high-dose selenium was observed to be a risk factor for increased PSA velocity in men with high baseline plasma selenium concentrations.

Figure 1

Participant cohort distribution

Figure 2. Point estimate and 95% confidence intervals for difference in PSA trajectories of selenium 200 μg/day and selenium 800 μg/day treatment groups as compared to placebo (reference)

Above models were adjusted for age, body mass index, race, Gleason score, pack-years of smoking, baseline PSA, baseline selenium, and type of assay used to estimate PSA. Confidence intervals crossing zero indicate no statistically significant difference at p = 0.05 level. (* Represents the regression coefficient for the interaction term between years on study and treatment group.)

Figure 3. Point estimate and 95% confidence intervals for difference in PSA trajectories of selenium 200 μg/day (green) and selenium 800 μg/day (orange) treatment groups as compared to placebo (reference) stratified by quartiles of baseline selenium

Above models were adjusted for age, body mass index, race, Gleason score, pack-years of smoking at enrollment, baseline PSA, and type of assay used to estimate PSA. Confidence intervals crossing zero indicate no statistically significant difference at p = 0.05 level. (* Represents the regression coefficient for the interaction term between years on study and treatment group.)

Figure 4