In this rural elderly Chinese cohort with an extensive range of selenium distribution, we found that lower selenium levels were significantly associated with higher depressive symptoms adjusting for demographic and medical conditions. However, when cognitive function was included as an independent variable in the model, the relationship between selenium and depressive symptoms was no longer significant, suggesting that selenium’s association with depressive symptoms is explained in part by its association with cognitive function.
Several previous publications including observational studies as well as clinical trials had examined the association between selenium levels and depression in various populations. Two studies observed that low selenium diet was associated with worsening mood [2
]. In addition, there were studies showing that selenium supplementation or high dietary selenium was associated with improved mood with the most significant effect seen in participants with low dietary selenium [1
]. However, a recent randomized trial of selenium supplementation found that a 6-month selenium supplementation had no significant effect on mood or quality of life in 501 participants age 60 to 74 from the United Kingdom [5
]. The differences in study findings between the recent trial and previous results may be explained by differences in dietary selenium levels of study participants. The UK trial consisted of participants whose baseline selenium status was higher than the UK national average [5
]. It is known that the brain has a unique feature in storing selenium and it also receives priority supply of selenium in times of deprivation [27
]. Animal studies have shown that after generations of selenium deficiency when selenium concentrations in liver, skeletal muscle and blood dropped to below 1
% of normal level, the brain still retained 60
% of selenium as in normal controls.[7
] In addition, studies have shown that plasma GP-x activity plateaus after sufficient selenium supply indicating that selenium supplementation in high selenium individuals may not provide additional benefit [28
]. It is perhaps not surprising that the UK trial found no selenium effect as most of the participants would already have adequate selenium supply at baseline and additional supplementation was not able to increase mood measures. However, results of the UK trial still left open the question of whether low selenium is associated with depression in those with low dietary selenium levels at baseline.
The review of current literature suggests that long-term exposure to low selenium may be needed to impact brain function. The brain’s unique selenium metabolism may also make it more difficult to show the effect of short-term selenium exposure on brain function than on other organs.
The rural Chinese population offers an opportunity to investigate the association of long-term selenium exposure and depression by identifying populations where substantial numbers are exposed to low selenium diet for an extended length of time and comparing their depressive symptoms to those with normal selenium levels. The overwhelming majority of our study participants were life-long residents of the same towns where they were interviewed and very few take dietary supplements. Hence the ascertained selenium levels can be inferred as life-long exposure to selenium without the influence of supplements. In this stable cohort, we found a significant association between lower selenium levels and increased depressive symptoms which is consistent with previous reports in participants with low dietary selenium intake.
The mechanism underlying an association between selenium and depression is not yet clear, although published results point to the fact that selenium is important to brain functions. In animal models, selenium deficiency was shown to alter neurotransmitter turnover rate [29
] and sodium selenite protected against dopamine loss [30
]. In addition, the enzyme closely related to selenium activity (GP-x) protected against neuron loss [31
], cell death [32
], and decreased free radical generation or amyloid beta peptide [33
]. In human studies, selenium supplementation was shown to reduce intractable epileptic seizures [34
]. A recent cross-sectional study reported that whole blood GPX activity was significantly lower in depressed patients than in normal controls [36
]. More research is needed to elucidate the biological mechanism for selenium’s function in the brain.
To our knowledge, no previous studies on selenium and depression have examined the role of cognitive function. Numerous studies have reported the co-existence of cognitive impairment and depression [37
], although the underlying mechanism for such an association is not yet established. Some suggest that depressive symptoms are early manifestation of the dementia disorder [38
], while others believe that depression represents an independent risk factor for cognitive decline [25
]. We have previously reported a significant relationship between selenium level and cognitive function in the same cohort using data collected from baseline [8
]. The non-significant association between selenium and depressive symptoms after adjusting for cognitive function suggests that cognitive function plays an important role in the relationship between selenium and depression.
Our study has a number of strengths. Selenium levels were measured in nail samples reflecting stable long-term exposure. The study included participants from sites with extensive range of selenium exposure. In addition, the majority of our study participants were life-long residents of the same towns where they were interviewed and the participants were known to have not taken dietary supplements. Hence the ascertained selenium levels can be inferred as life-long exposure to selenium without the influence of supplements. A number of cognitive instruments were used so that a reliable composite measure of cognitive function can be derived.
There are also important limitations in this study. Our results were based on cross-sectional analyses in the sense that cognitive function and depressive symptoms were measured at the same time. Longitudinal follow-up of this cohort would be necessary to examine the casual direction of the association between cognitive function and depressive symptoms. A second limitation is that we relied on the GDS to assess depressive symptoms rather than clinical diagnosis of specific depressive disorders, an approach adopted by many large epidemiological studies [25
]. Further limitations include the exclusion of potentially important risk factors for depression such as disability or social support.