The authors have declared that no competing interests exist.
As an essential nutrient, Selenium (Se) is involved in many metabolic activities including mimicking insulin function. Data on Se in various biological samples and insulin resistance are contradictory, moreover there is no large study available regarding the relationship of dietary Se intake with insulin resistance in the general population. To investigate the association between dietary Se intake and variation of insulin resistance in a large population based study, a total of 2420 subjects without diabetes from the CODING (Complex Diseases in the Newfoundland Population: Environment and Genetics) study were assessed. Dietary Se intake was evaluated from the Willett Food Frequency questionnaire. Fasting blood samples were used for the measurement of glucose and insulin. Insulin resistance was determined with the homeostasis model assessment (HOMA-IR). Body composition was measured using dual energy X-ray absorptiometry. Analysis of covariance showed that high HOMA-IR groups in both males and females had the lowest dietary Se intake (μg/kg/day) (p < 0.01), being 18% and 11% lower than low HOMA-IR groups respectively. Insulin resistance decreased with the increase of dietary Se intake in females but not in males after controlling for age, total calorie intake, physical activity level, serum calcium, serum magnesium, and body fat percentage (p < 0.01). Partial correlation analysis showed that dietary Se intake was negatively correlated with HOMA-IR after adjusting for the Se confounding factors in subjects whose dietary Se intake was below 1.6 μg/kg/day (r = -0.121 for males and -0.153 for females, p < 0.05). However, the negative correlation was no longer significant when dietary Se intake was above 1.6 μg/kg/day. Our findings suggest that higher dietary Se intake is beneficially correlated with lower insulin resistance when total dietary Se intake was below 1.6 μg/kg/day. Above this cutoff, this beneficial effect disappears.
Selenium (Se) is an essential micronutrient element, and a key component of several selenoproteins with essential enzymatic functions that include redox homeostasis [
As early as 1990s, data from isolated rat adipocytes suggest that Se (as selenate) can mimic the effects of insulin, including stimulating glucose transport activity and enhancing insulin receptor kinase activity [
The association between Se nutritional status and diabetes is very complicated and intriguing. Insulin resistance is not only a hallmark but also a pathogenic factor of T2DM. However, the quantitative relationship between dietary Se intake and insulin resistance has been only reported in studies with very small sample size [
Therefore, we designed the present study to investigate the association between dietary Se intake and insulin resistance in a large general population with systematic control of major confounding factors.
All participants were from the ongoing CODING (Complex Diseases in the Newfoundland Population: Environment and Genetics) study. Eligibility of participants for the CODING study was based upon the following inclusion criteria: 1) ≥19 years of age; 2) at least a third generation Newfoundlander; 3) without serious metabolic, cardiovascular or endocrine diseases; 4) women were not pregnant at the time of the study [
Anthropometrics were performed following a 12-hour overnight fast. Trained personnel obtained these measurements for each subject using standard procedures. Standing height was measured using a fixed stadiometer (to the nearest 0.1 cm). After fully emptying their bladders, subjects wore standard hospital gowns for all weight measurements using a platform manual scale balance (Health O Meter, Bridgeview, IL; nearest 0.1 kg). Body mass index (BMI) (kg/m2) was calculated as weight in kilograms divided by height squared in meters. Waist circumference (WC) was defined as the midway point between the iliac crest and the lowest rib, and hip by the maximum circumference over the buttocks below the iliac crest. Waist-hip ratio (WHR) was the division of WC by hip circumference.
Body compositions including total body fat percentage (BF%), trunk fat percentage (TF%), android fat percentage (AF%) and gynoid fat percentage (GF%) were measured, in a supine position, utilizing dual energy X-ray absorptiometry (DXA: Lunar Prodigy; GE Medical Systems, Madison, WI) with the Lunar Prodigy software system. The Lunar Prodigy software system has the capacity to distinguish each of these regions. Trunk fat region is from the top of the shoulders to the top of the iliac crest, while the android fat region is the top of the second lumbar vertebra to the top of the iliac crest and the gynoid fat region extends down the iliac crest twice the height of the android area. The enCORE (Ver 12.2, 2008, GE Medical Systems, Madison, WI) software package was used for DXA data acquisition. Daily quality assurance was performed on the DXA scanner and the typical coefficient of variation was 1.3% during the study period [
Dietary intake of each participant was assessed using a 124 item semi- quantitative Willett food frequency questionnaire (FFQ) [
All participants completed a self-administered screening questionnaire, which was used to collect information of personal health history. Physical activity patterns were measured using the ARIC Baecke Questionnaire, which consists of a Work Index, Sports Index, and Leisure Time Activity Index [
Venous blood samples were collected in the morning after an overnight fast (12 hours). Serum samples were isolated from whole blood and stored at −80°C for subsequent analysis. FBG were measured on an Lx20 analyzer (Beckman Coulter Inc., Fullerton, CA) using Synchron reagents. Fasting insulin (FINS) was measured on an Immulite Immunoassay analyzer. Insulin resistance and β cell function were determined with the homeostasis model assessment (HOMA-IR and HOMA-β), as described by Matthews et al [
HOMA-IR = (Fasting Insulin [mU/L]×Fasting Glucose [mmol/L])/22.5
HOMA-β = (20×Fasting Insulin [mU/L])/(Fasting Glucose [mmol/L]—3.5)
All data are presented as mean ± standard error of the mean(SEM). FINS, HOMA-IR, HOMA-β, calorie intake and dietary Se intake were log-transformed in order to normalize data distributions to perform effective statistical analysis. Anthropometrics, body composition, dietary intake and biochemical measurements were compared between females and males with independent Student's t-test.
In order to analyze the variation of dietary Se intake in different status of insulin resistance, participants were divided into tertiles (low, medium, and high) of insulin resistance based upon HOMA-IR. Dietary Se intake was compared among the three groups with analyses of variance and covariates (ANCOVA) controlling for age, calorie intake and physical activity. The variations of insulin resistance in different dietary Se intakes were analyzed after participants were divided into tertiles (low, medium, and high dietary Se intake). FBG, FINS, HOMA-IR, and HOMA-β were compared among groups using ANCOVA controlling for age, calorie intake, physical activity, serum calcium, serum magnesium, and BF%. Serum calcium and magnesium were taken into consideration as well because our previous studies have shown that they were associated with insulin resistance [
Dietary Se intake below 0.4 μg/kg/day is considered as Se deficiency [
Partial correlation analysis, controlling for age, calorie intake, physical activity, serum calcium, serum magnesium, and BF%, was subsequently applied to further confirm the findings from ANCOVA. To control for possible influence of smoking, alcohol drinking, disease status medication use and menopausal, analyses were performed in participants with and without these confounding factors.
All statistical analyses were performed using SPSS 20.0 (SPSS Inc., Chicago, IL). All tests were two sided and p < 0.05 was considered to be statistically significant.
Clinical and dietary characteristics of the study subjects are presented in
Entire cohort (n = 2420) | Females (n = 1777) | Males (n = 643) | p value | |
---|---|---|---|---|
Age (yr) | 42.40 ± 0.27 | 43.42 ± 0.27 | 39.80 ± 0.47 | < 0.001 |
Weight (kg) | 74.04 ± 0.34 | 69.16 ± 0.29 | 86.64 ± 0.53 | < 0.001 |
BMI (kg/m2) | 26.56 ± 0.10 | 26.14 ± 0.11 | 27.67 ± 0.16 | < 0.001 |
WC (cm) | 91.57 ± 0.29 | 89.35 ± 0.29 | 97.32 ± 0.45 | < 0.001 |
WHR | 0.91 ± 0.001 | 0.89 ± 0.001 | 0.97 ± 0.002 | < 0.001 |
TF% | 36.11 ± 0.19 | 38.43 ± 0.19 | 30.11 ± 0.33 | < 0.001 |
AF% | 41.15 ± 0.23 | 43.20 ± 0.23 | 35.87 ± 0.40 | < 0.001 |
GF% | 40.08 ± 0.20 | 44.51 ± 0.14 | 28.59 ± 0.27 | < 0.001 |
BF% | 33.82 ± 0.19 | 37.18 ± 0.17 | 25.14 ± 0.28 | < 0.001 |
FBG (mmol/L) | 5.02 ± 0.01 | 4.96 ± 0.01 | 5.17 ± 0.02 | < 0.001 |
FINS (pmol/L) | 67.61 ± 0.89 | 65.77 ± 0.99 | 72.6 1± 1.88 | 0.008 |
HOMA-IR | 2.24 ± 0.03 | 2.17 ± 0.04 | 2.43 ± 0.07 | < 0.001 |
HOMA-β | 133.84 ± 1.98 | 136.54 ± 2.56 | 126.39 ± 3.72 | 0.002 |
Serum calcium (mmol/L) | 2.36 ± 0.002 | 2.35 ± 0.002 | 2.38 ± 0.003 | < 0.001 |
Serum magnesium (mmol/L) | 0.88 ± 0.002 | 0.88 ± 0.001 | 0.89 ± 0.002 | < 0.001 |
Physical activity | 8.28 ± 0.03 | 8.18 ± 0.03 | 8.53 ± 0.06 | < 0.001 |
calorie intake (kcal/day) | 1991.55 ± 18.49 | 1873.63 ± 17.62 | 2317.87 ± 38.32 | < 0.001 |
Se (μg/day) | 109.22 ± 1.18 | 102.34 ± 1.11 | 128.23 ± 2.78 | < 0.001 |
Se (μg/kg/day) | 1.53 ± 0.02 | 1.53 ± 0.02 | 1.53 ± 0.03 | 0.19 |
All data presented as mean ± SEM. BMI, Body mass index; WC, Waist circumference; WHR, Waist hip rate; TF%, trunk fat percentage; AF%, android fat percentage; GF%, gynoid fat percentage; BF%, total body fat percentage; FBG, fasting blood glucose; FINS, fasting insulin; HOMA-IR, homeostasis model assessment of insulin resistance; HOMA-β, homeostasis model assessment of β cell function.
Significant differences of dietary Se intake were revealed between the three groups with different insulin resistance after controlling for age, total calorie intake and physical activity (
Low | Medium | High | p trend | ||
---|---|---|---|---|---|
Females | 592 | 592 | 592 | - | |
HOMA-IR | 0.40~1.42 | 1.42~2.32 | 2.32~24.07 | - | |
Se (μg/day) | 102.19 ± 1.34 | 99.84 ± 1.32 | 101.23 ± 1.35 | 0.21 | |
Se (μg/kg/day) | 1.62 ± 0.02 | 1.49 ± 0.02 | 1.34 ±0.02 | < 0.001 | |
Males | 214 | 214 | 214 | ||
HOMA-IR | 0.40~1.60 | 1.60~2.74 | 2.74~16.37 | - | |
Se (μg/day) | 117.70 ± 3.15 | 117.83 ± 3.02 | 121.83 ± 3.10 | 0.62 | |
Se (μg/kg/day) | 1.47 ± 0.04 | 1.38 ± 0.04 | 1.30 ±0.04 | < 0.001 |
Data were assessed with Covariance controlling for age, total calorie intake, and physical activity. All values are presented as means ± SEMs. HOMA-IR, homeostasis model assessment of insulin resistance.
When subjects were grouped into tertiles according to dietary Se intake (low, medium and high), levels of FINS, HOMA-IR, HOMA-β presented a dose-dependent decline (high, medium and low) with the increase of dietary Se intake after controlling for age, total calorie intake, physical activity, serum calcium, serum magnesium and BF% in females (p < 0.05) not in males (
Low | Medium | High | p | ||
---|---|---|---|---|---|
Female | Number | 592 | 592 | 592 | - |
Se (μg/kg/day) | 0.16 ~1.12 | 1.22 ~ 1.66 | 1.66 ~ 8.89 | - | |
FBG (mmol/L) | 5.05 ± 0.02 | 4.99 ± 0.02 | 4.99 ± 0.02 | 0.13 | |
FINS (pmol/L) | 73.02 ± 1.86 | 62.47 ± 1.59 | 61.23 ± 1.96 | <0.001 | |
HOMA-IR | 2.41 ± 0.07 | 2.02 ± 0.06 | 2.04 ± 0.08 | <0.001 | |
HOMA-β | 142.22 ± 4.36 | 131.40 ± 3.72 | 126.77 ± 4.59 | 0.01 | |
Male | Number | 214 | 214 | 214 | - |
Se (μg/kg/day) | 0.22 ~ 1.05 | 1.05 ~ 1.61 | 1.61 ~ 7.19 | - | |
FBG (mmol/L) | 5.25 ± 0.04 | 5.23 ± 0.03 | 5.27 ± 0.04 | 0.69 | |
FINS (pmol/L) | 78.24 ± 3.88 | 69.76 ± 3.34 | 73.61 ± 4.62 | 0.30 | |
HOMA-IR | 2.63 ± 0.14 | 2.36 ± 0.12 | 2.51 ± 0.17 | 0.28 | |
HOMA-β | 129.75 ± 8.57 | 122.21 ± 7.37 | 135.42 ± 10.19 | 0.41 |
Data were assessed with covariates controlling for age, calorie intake, physical activity, serum calcium, serum magnesium, and BF%. Data presented as mean ± SEM. FBG, fasting blood glucose; FINS, fasting insulin; HOMA-IR, homeostasis model assessment of insulin resistance; HOMA-β, homeostasis model assessment of β cell function.
For each 1 μg/kg/day increase in dietary Se intake, average weight, BMI, WC, and WHR decreased by 8.39 kg, 2.98 kg/m2, 8.03 cm, and 0.02 in women, and by 8.85 kg, 2.34 kg/m2, 7.39 cm, and 0.02 in men, respectively. Likewise, TF%, AF%, GF% and BF% were reduced by 4.58%, 5.56%, 3.05% and 4.16% in women, and by 5.43%, 5.94%, 4.19% and 4.45% in men, respectively (
Subjects were divided into 10 groups based on dietary Se intake with an interval of 0.4 μg/kg/day, and the variations of FINS, HOMA-IR and HOMA-β with dietary Se intake were analyzed with covariates (
HOMA-IR, homeostasis model assessment of insulin resistance; HOMA-β, homeostasis model assessment of β cell function.
The correlations between dietary Se intake and FINS, HOMA-IR and HOMA-β are presented in
Dietary Se intake |
≤ 1.6 | > 1.6 | |||
---|---|---|---|---|---|
r0 (p) | r1(p) | r0 (p) | r1(p) | ||
Females | FBG | -0.157 (0.000) | -0.103 (0.001) | -0.042 (0.303) | -0.048 (0.240) |
FINS | -0.183 (0.000) | -0.148 (0.000) | -0.028 (0.492) | -0.049 (0.238) | |
HOMA-IR | -0.191 (0.000) | -0.153 (0.000) | -0.036 (0.375) | -0.064 (0.121) | |
HOMA-β | -0.107 (0.000) | -0.102 (0.001) | -0.001 (0.981) | -0.003 (0.939) | |
Males | FBG | -0.165 (0.001) | 0.022 (0.664) | -0.112 (0.150) | -0.043 (0.586) |
FINS | -0.216 (0.000) | -0.124 (0.014) | -0.177 (0.023) | -0.004 (0.958) | |
HOMA-IR | -0.227 (0.000) | -0.121 (0.017) | -0.186 (0.016) | -0.015 (0.853) | |
HOMA-β | -0.156 (0.002) | -0.121 (0.017) | -0.111 (0.154) | 0.041 (0.606) |
Partial correlations between dietary Se intake (μg/kg/day) and insulin resistance were controlling for age, total caloric intake, physical activity, serum calcium, serum magnesium, and body fat percentage. FBG, fasting blood glucose; FINS, fasting insulin; HOMA-IR, homeostasis model assessment of insulin resistance; HOMA-β, homeostasis model assessment of β cell function. r0: correlation coefficient; r1: partial correlation coefficient.
To further exclude the influence of additional covariates, data of participants who were non-smokers, non-drinkers, on no medication and otherwise healthy were analyzed. In
No Smoking | No Alcohol | No Medication | No Disease | |||||
---|---|---|---|---|---|---|---|---|
r1 | r2 | r1 | r2 | r1 | r2 | r1 | r2 | |
Female | ||||||||
FBG | -0.086 |
-0.045 | -0.078 | 0.055 | -0.081 |
0.012 | -0.128 |
-0.065 |
FINS | -0.137 |
-0.038 | -0.179 |
-0.062 | -0.106 |
-0.012 | -0.133 |
-0.042 |
HOMA-IR | -0.138 |
-0.055 | -0.180 |
-0.078 | -0.110 |
-0.015 | -0.140 |
-0.061 |
HOMA-β | -0.101 |
0.010 | -0.139 |
-0.027 | -0.075 |
-0.014 | -0.082 |
-0.007 |
Male | ||||||||
FBG | -0.063 | -0.065 | -0.154 | 0.268 | -0.027 | -0.035 | -0.010 | -0.046 |
FINS | -0.137 |
-0.024 | -0.279 |
0.220 | 0.105 |
-0.061 | -0.104 |
-0.014 |
HOMA-IR | -0.136 |
-0.037 | -0.280 |
0.211 | -0.096 |
-0.048 | -0.098 |
-0.024 |
HOMA-β | -0.122 |
0.031 | -0.238 | 0.233 | 0.124 | -0.095 | -0.112 | 0.030 |
Partial correlations between dietary Se intake (μg/kg/day) and insulin resistance controlling for age, total caloric intake, physical activity, serum calcium, serum magnesium, and body fat percentage. FBG, fasting blood glucose; FINS, fasting insulin; HOMA-IR, homeostasis model assessment of insulin resistance; HOMA-β, homeostasis model assessment of β cell function. r1: partial correlation coefficient, when dietary Se intake was below or equal to 1.6 μg/kg/day; r2: partial correlation coefficient, when dietary Se intake was greater than 1.6 μg/kg/day.
*, P<0.05
** P<0.01.
In the present study, we analyzed the associations between dietary Se intake and insulin resistance in the large CODING study with a wide range of dietary Se intake among 2,420 adult Newfoundlanders. To the best of our knowledge, this is the first large cross-sectional study specifically designed to analyze the association between dietary Se intake and insulin resistance in the general population. The most important finding is that dietary Se intake was significantly negatively associated with insulin resistance in females and males after controlling all major confounding factors, when dietary Se intake was ≤ 1.6 μg/kg/day most subjects fell in this range. However, when dietary Se intake was > 1.6 μg/kg/day, this reverse relationship was no longer significant. The findings suggest that the beneficial negative relationship between dietary Se intake and insulin resistance may have a dose ceiling.
Se may affect insulin resistance via multiple routes including insulin-like action, inflammatory cytokines and oxidative stress. Early studies showed that sodium selenate may mimic insulin to stimulate glucose uptake [
Insulin resistance is a complex pathophysiological condition. There are numerous factors that can potentially be involved in the development of insulin resistance. It is critical to identify and properly control or adjust major factors in a large population study because if these factors are not properly adjusted, they would potentially cause either false positive or negative results. Similarly a variety of factors may potentially affect dietary Se intake. Food choice and consumption and insulin resistance vary among different age and gender groups [
Previous studies had reported conflicting results for the relationships between Se in various biological samples and insulin resistance. It was reported that hair Se were negatively correlated with the HOMA-IR controlled for age and sex in a small Korean study [
Another important finding from the present study is the ‘threshold’ of the beneficial effect of dietary Se intake on insulin resistance. The beneficial association generally started from very low level to 1.6 μg/kg/day, approximately equivalent to 118 μg/day for person with average body weight (139 μg/day for males and 110 μg/day for females). This beneficial association was nearly linear within this range but became weaker and disappeared if dietary Se intake was above this level. This cut-off point is similar to the suggested dietary Se dose (100–150 μg/day) for tumor protection [
There are several potential limitations in the present study. Although the study has identified and adjusted many major confounding factors, it is inevitably other potential factors were not included, such as zinc and copper which have been reported to have effects on insulin sensitivity [
In conclusion, our findings revealed a significantly negative association of dietary Se intake with insulin resistance in the large CODING study with many major confounding factors adjusted, when dietary Se intake was below 1.6 μg/kg/day. Given the narrow margin between Se deficiency, adequacy, over-nutrition and toxicity, accurately determining this cutoff point is clinically important. People with an adequate or high Se status have already received the maximum benefit from Se and need not take additional Se supplementation, because of the potential risks, such as the increased risk of type 2 diabetes. Ultimately, this study aims to add to our present knowledge about the optimal constitution of an ideal diet to reduce insulin resistance and risk of type 2 diabetes in terms of specific macronutrient and micronutrient composition.
We would like to thank all of the volunteers who participated in this present study. We declare that we have no conflicts of interest.