Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Selenium and CHD

Systematic review
Comment

The apparently conflicting evidence about whether selenium is protective against heart disease has long intrigued Bandolier. There is something of a rational choice, in that it is involved in protection against oxidative damage. Foods that tend to be selenium rich include plants, meats and seafood, though the content varies with geography and use of fertiliser with selenium. Many people supplement their diets with selenium, often as yeast grown on selenium rich nutrients, often with 100 μg per day or so.

The question is whether such supplementation is helpful, and, indeed, whether selenium levels in the body have any relationship to actual outcomes, like coronary heart disease. A new systematic review and meta-analysis [1] pulls the evidence together, and poses some interesting if difficult questions.

Systematic review


The search examined MEDLINE and Cochrane up to March 2006 for studies that might be relevant. Importantly, it also did some extensive searching of references, which is often as good a source for observational studies as electronic searching. The aim was to identify observational studies assessing the association of selenium content of blood or other tissues with clinical coronary heart disease, as well as any randomised studies examining selenium supplementation.

The main results are shown in Table 1. Cohort studies found a bare statistical association between selenium levels and heart disease, though only one individual cohort study out of 14 individually reached statistical significance. Almost all of these (13/14) had adjusted for confounders like age, sex, smoking, BMI, and other factors.

By contrast, case control studies showed a large reduction in coronary heart disease associated with higher selenium levels (Table 1). Thirteen of 18 case control studies individually reached statistical significance. However, the three case control studies with at least 100 cases either failed to reach statistical significance, or only bare significance was achieved. Moreover, most case control studies did not make any adjustment for confounding factors, and only six of 18 made any adjustments.


Table 1: Relative risk of clinical coronary heart disease and selenium levels in tissue in observational studies, or supplementation in RCTs



Type of study
Number of studies
(cases)
Relative risk
(95% CI)
Cohort
14 (1,835)
0.85 (0.74 to 0.99)
Case control
16 (1,854)
0.43 (0.29 to 0.66)
RCTs
6 (452)
0.89 (0.68 to 1.2)



Randomised trials generally used 100 μg selenium supplementation daily from yeast. The four larger studies were of high quality and had a long duration, of three to eight years. Some had selenium combined with other vitamins or minerals. The results (Figure 1, Table 1) were not statistically significant.


Figure 1: Results for clinical coronary heart disease in randomised trials of selenium supplementation






Comment


There are some interesting lessons here, particularly the difference between case control studies that generally made no adjustments for confounding, and cohort studies that did. Of course, adjustment may not be the issue, but it is notable that both larger case control studies, and those that made adjustments for confounding factors had similar magnitude of effect as did cohort studies and randomised trials. Another, recent, randomised trial [2] found no effect of selenium supplementation on heart disease over 7.6 years of therapy.

This emphasises the fact that we have to look carefully at small observational studies, and not be overwhelmed by the size of an effect that could be just an artefact of how a study has been performed.

So what of selenium supplementation? We can probably be sure that this is no magic bullet to protect us against heart disease, especially if we have a varied and full diet. If there is an effect, it is small, probably no more than about 10%. That, of course, depends on one's point of view, and one persons 10% trivia could be someone else's 10% Grand Canyon. Selenium may be a bigger deal in other contexts.

Reference:

  1. G Flores-Mateo et al. Selenium and coronary heart disease: a meta-analysis. American Journal of Clinical Nutrition 2006 84: 762-773.
  2. S Stranges et al. Effect of selenium supplementation on cardiovascular disease incidence and mortality: secondary analyses in a randomized clinical trial. American Journal of Epidemiology 2006 163: 694-699.

previous or next story