Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Sep 9.
Published in final edited form as: Metabolism. 2013 Aug 27;62(12):1788–1794. doi: 10.1016/j.metabol.2013.07.008

Circulating 25-hydroxyvitamin D and insulin resistance in older adults: The Cardiovascular Health Study

John Danziger 1, Mary L Biggs 2, Matt Niemi 1, Joachim H Ix 3,4, Jorge R Kizer 5, Luc Djoussé 6, Ian H de Boer 7, David S Siscovick 2, Bryan Kestenbaum 2, Kenneth J Mukamal 1
PMCID: PMC4159161  NIHMSID: NIHMS592957  PMID: 23987236

Abstract

Background

Despite extensive study, the role of vitamin D in insulin resistance and secretion remains unclear.

Objective

To examine the cross-sectional and longitudinal relationships between 25-hydroxyvitamin D (25(OH)D) concentrations and indices of insulin resistance and secretion in older adults.

Methods and Results

Among 2134 participants of the Cardiovascular Health Study who were free from cardiovascular disease, we measured serum 25(OH)D concentrations in samples collected in 1992–1993. We examined insulin resistance and secretion using Homeostasis Model Assessment (HOMA) estimates cross-sectionally and among 1469 participants who had repeated HOMA measures four years later (1996–1997). In cross-sectional analysis, each 10 ng/mL increment in 25(OH)D concentration was associated with a 0.09 lower adjusted HOMA-IR [95%CI (−0.17, −0.02), p=0.01]. However, baseline 25(OH)D concentrations were not associated with change in HOMA-IR over 4 years of follow up (p=0.48). 25(OH)D concentrations were not associated with insulin secretion, as determined by HOMA-β, in either cross-sectional or longitudinal analysis.

Conclusions

Circulating 25(OH)D concentrations are associated with lower insulin resistance in cross-sectional but not longitudinal analyses. Whether this reflects residual confounding in cross-sectional analyses or the short-term nature of the relationship between vitamin D and insulin sensitivity will require trials with repeated measures of these factors.

Keywords: Vitamin D, 25(OH)D, insulin resistance

Introduction

A growing awareness of the pleiotropic effects of vitamin D and the prevalence of vitamin D deficiency has prompted intense research efforts to understand the biological importance of this hormone. The preponderance of cross-sectional1–4 and longitudinal5–7 data suggests a role for vitamin D deficiency in insulin resistance, and a recent large meta-analysis has demonstrated an inverse association between vitamin D concentrations and incident diabetes8. Although mechanistic studies suggest that vitamin D deficiency impairs pancreatic insulin secretion9, explanations for insulin resistance are lacking10. In randomized controlled studies, vitamin D supplementation has not uniformly improved insulin resistance10–12. The most robust randomized trial using a three-hour hyperglycemic clamp to characterize insulin resistance failed to find a protective role of vitamin D administration on insulin sensitivity in vitamin D-deficient individuals13.

The basis for the disparity in findings across trials and observational studies is uncertain. Two positive trials were conducted in South Asian and Indian individuals, while negative studies have included Norwegians and North Americans, suggesting that race could be important. Further, important confounders known to be associated with insulin resistance, including obesity14,15, seasonal variation, and physical inactivity16, have not necessarily been well accounted for in previous observational studies, and few have included longitudinal data on insulin resistance.

To understand the relationship between vitamin D and insulin resistance further, we studied participants enrolled in the Cardiovascular Health Study (CHS), a richly characterized sample of community-dwelling elderly individuals with extensive information on potential confounders including anthropomorphic measures of obesity, several biochemical markers of adiposity and inflammation, and detailed questionnaires on exercise amount and intensity. In addition, CHS includes repeated measures of fasting glucose and insulin with which to estimate insulin resistance and secretion.

Methods

Participants

CHS is a community-based study of older adults designed to evaluate risk factors for cardiovascular disease. The study design has been described previously17,18. In brief, eligibility included age ≥65 years, expectation to remain in one’s current area of residence for three years, no active malignancy, and the ability to provide consent. Between 1989 and 1990, 5201 participants were recruited from four U.S. communities through the use of Medicare eligibility lists. An additional 687 predominantly African-American individuals were recruited from 1992 to 1993. All participants provided written informed consent.

Serum 25(OH)D, fasting insulin, and fasting glucose concentrations were measured in 2316 individuals without prevalent cardiovascular disease (CVD) at the 1992–1993 (herein considered “baseline”) study visit after an eight-hour overnight fast19,20. We excluded from the analyses 143 participants taking diabetes medications, 28 who reported <8 hours fasting or were missing confirmation of fasting time, and 11 who had extreme, likely non-fasting values (HOMA-IR >20), leaving a total sample of 2134 participants.

Follow-up laboratory values were obtained from the 1996–1997 examination. Of the original 2134, 170 participants were deceased, 79 did not participate in the follow-up examination, and 323 were missing either a glucose or an insulin measurement. Another 38 individuals had unknown fasting times and 5 had HOMA-IR >20 and were excluded from analysis. In addition, 50 individuals began hypoglycemic medication during follow-up and were excluded in the primary analysis. In a sensitivity analysis, these individuals were included and their HOMA-IR set to the 99% percentile of HOMA-IR.

25(OH)D Concentration

Fasting serum was collected from CHS participants and stored at −70 degrees Celsius. 25(OH)D was measured by using high performance liquid chromatography tandem mass spectrometry on a Quattro micro mass spectrometer (Waters Corporation, Milford, Massachusetts) in 200820. The inter-assay coefficient of variation was less than 3.4%.

Determination of Insulin Resistance and Secretion

We used the Homeostasis Model Assessment to estimate insulin resistance (HOMA-IR) and secretion (HOMA-β). HOMA-IR was calculated as: fasting glucose [(millimoles per liter) * fasting insulin (milliunits per liter)]/22.5. HOMA-β was calculated as: [20 * fasting insulin (milliunits per liter)]/[fasting glucose (millimoles per liter)−3.5}21.

Other covariates

Covariates from the 1992–1993 examination, including age, sex, years of education, smoking status, and alcohol consumption were based on self report. Leisure time activity (kilocalories per week) was assessed using a modified Minnesota Leisure-Time Activities questionnaire, as well as the quantification of blocks walked per week, and amount of exercise intensity (none, low, moderate, or high). The Center for Epidemiological Studies Depression (CES-D) scale was used to assess for depressive symptoms. Study staff measured weight, height, and waist circumference using standard protocols. Missing values for height for the 1992–93 examination were carried forward from previous years, if available. Adiponectin was measured using an enzyme-linked immunosorbent assay (Millipore, Billerica, MA), plasma fatty acid binding protein-4 (FABP4) was measured using standard ELISA kits (Biovendor ELISA), fetuin-A was measured by ELISA (Epitope Diagnostics, San Diego, CA), non-esterified fatty acids (NEFA) were measured by the Wako enzymatic method, and cystatin C using a particle enhanced immunonephelometric assay (Dade Behring, Newark, DE). Self-reported measures of sleep included trouble falling asleep, awakening several times at night, and waking too early.

Statistical Analysis

Baseline characteristics of study participants were summarized according to categories of vitamin D status (≥30, 15–29, and <15 ng/ml), thresholds commonly used and previously shown to have biologic consequence5,20. We used natural cubic splines to evaluate the functional form of the association between 25(OH)D concentrations and HOMA measures. We used linear regression with robust variance estimates to examine the association of 25(OH)D concentrations with HOMA-IR and HOMA-B measured at baseline, as well as with change in HOMA-IR and HOMA-B over four years of follow-up.

In cross-sectional and longitudinal analyses, we modeled baseline 25(OH)D concentrations in units of 10 ng/mL increments, a difference in 25(OH)D concentration readily achieved with a moderate-dose vitamin D supplement. In addition, we modeled 25(OH)D concentrations using the 3 categories described above, with ≥30 ng/ml as the reference level. Given known seasonal variability in 25(OH)D concentrations22, season was defined harmonically1. Model I included variables for age, sex, race (African-American, non-African-American), clinic site, and season. Model II additionally adjusted for body mass index (BMI) and waist circumference. Model III added alcohol consumption (any vs. none), physical activity (kcal/per week), exercise intensity, C-reactive protein, cystatin-C, CES-D score, albumin, NEFAs, FABP4, fetuin-A, and total adiponectin.

To investigate if the association between 25(OH)D and insulin resistance differed by race, we created a multiplicative interaction term between race and 25(OH)D concentrations. To investigate the influence of health status, we repeated cross sectional and longitudinal analyses after restricting the sample to participants who characterized their health as “excellent, very good, or good”. We also tested additional adjustment for sleep variables with no change in our results.

We tested for a non-linear association of vitamin D and HOMA-IR by using the Stata fracpoly command to compare the deviance of a best-fit fractional polynomial model of degree 2 to a model with vitamin D fit as a linear term.

We conducted all analyses with Stata, version 12.1 (StataCorp LP, College Station, Tx.)

Results

Table I presents baseline characteristics of the subjects, stratified by their 25(OH)D concentrations. Concentrations of 25(OH)D varied seasonally, with highest levels reported in the fall. Vitamin D deficient participants tended to be male, African American, less active, and more likely to smoke. Although both BMI and waist circumferences were greater in those with vitamin D deficiency, levels of adiponectin, fetuin-A, FABP4, and C-reactive protein did not differ by vitamin D category. The mean(±SD) values of HOMA-IR according to vitamin D status were 2.5(1.6) for ≥30 ng/ml, 2.6(2.0) for 15–29 ng/ml, and 3.3(2.6) for <15 ng/mL.

Table 1.

Characteristics of CHS participants at 1992–93 examination by 25-hydroxyvitamin D concentrations.

25-hydroxyvitamin D, ng/ml
≥30 15–29 <15 p-value
No. 667 1,130 337
Age, years 73.33 ± 4.25 74.13 ± 4.88 74.22 ± 5.45 0.001
Male, n (%) 276 (41.4%) 296 (26.2%) 70 (20.8%) <0.001
African-American, n (%) 27 (4.0%) 138 (12.2%) 122 (36.2%) <0.001
Field center, n (%) <0.001
  North Carolina 187 (28.0%) 346 (30.6%) 87 (25.8%)
  California 201 (30.1%) 238 (21.1%) 85 (25.2%)
  Maryland 145 (21.7%) 301 (26.6%) 66 (19.6%)
  Pennsylvania 134 (20.1%) 245 (21.7%) 99 (29.4%)
Season, n (%) <0.001
  Winter 144 (21.6%) 265 (23.5%) 58 (17.2%)
  Spring 104 (15.6%) 285 (25.2%) 151 (44.8%)
  Summer 128 (19.2%) 264 (23.4%) 84 (24.9%)
  Fall 291 (43.6%) 316 (28.0%) 44 (13.1%)
Body mass index, kg/m2 25.41 ± 3.81 26.92 ± 4.75 27.55 ± 5.44 <0.001
Waist circumference, cm 90.25 ± 11.58 92.75 ± 13.30 94.65 ± 15.05 <0.001
Physical activity, kcal/wk 2024.72 ± 2034.49 1374.56 ± 1570.61 846.56 ± 1159.71 <0.001
Any alcohol use, n (%) 343 (51.4%) 501 (44.3%) 144 (42.7%) 0.003
Never-smoker, n (%) 310 (46.5%) 601 (53.2%) 154 (45.7%) 0.57
Total Adiponectin, ng/mL 15380.38 ± 8001.72 14726.74 ± 7774.01 14012.38 ± 7660.63 0.008
Fetuin A, g/L 0.49 ± 0.10 0.48 ± 0.09 0.47 ± 0.10 0.013
FABP4, ng/mL 30.66 ± 17.01 33.97 ± 15.74 37.01 ± 18.22 <0.001
C-reactive protein, mg/l 4.14 ± 8.27 4.58 ± 8.00 5.72 ± 8.69 0.006
Open in a new tab

Mean values ± standard deviation provided.

Cross sectional analysis

Table II presents sequential models of 25OH(D), defined continuously and categorically, and insulin resistance. In a model that included age, sex, race, study site, and season, 25(OH)D concentration was strongly inversely associated with insulin resistance. The addition of BMI and waist circumference substantially attenuated this association, but it remained statistically significant. The addition of adipocyte-derived hormones and physical activity had minimal further effect on the association. The fully adjusted association between 25(OH)D concentrations and HOMA-IR is illustrated in Figure 1. To provide context to the observed values, each 10 ng/mL increment in 25(OH)D concentration was associated with a 0.09 lower HOMA-IR [95%CI (−0.16, −0.01)], while each 1-kg/m2 increment in BMI was associated with a 0.09 higher HOMA-IR [95% CI (0.06,0.13)]. When modeled as a categorical variable, the lowest category of 25(OH)D (<15 ng/ml) was significantly associated with higher HOMA-IR in fully adjusted analysis (p=0.04)

Table 2.

Cross-sectional association between 25-hydroxyvitamin D and HOMA-IR and HOMA-β at 1992–93 examination.

Mean difference in HOMA measure associated with concentrations
of 25-hydroxyvitamin D
Categories of 25-hydroxyvitamin D
concentrations, ng/ml		 Per 10 ng/ml
greater
concentration of
25-hydroxyvitamin
D
≥30 15–29 <15
HOMA- IR
Model 11 0 (Ref.) 0.39 (0.21, 0.57)
p<0.001		 0.72 (0.41, 1.03)
p<0.001		 −0.25 (−0.34, −0.17)
p<0.001
Model 22 0 (Ref.) 0.08 (−0.09, 0.25)
p=0.35		 0.33 (0.05, 0.60)
p=0.02		 −0.10 (−0.18, −0.03)
p=.009
Model 33 0 (Ref.) 0.08 (−0.07, 0.24)
p=0.30		 0.27 (0.02, 0.53)
p=0.04		 −0.09 (−0.17, −0.02)
p=0.01
HOMA- β
Model 11 0 (Ref.) 6.50 (0.81, 12.19)
p=0.03		 7.10 (−1.73. 15.92)
p=0.12		 −1.54 (−4.45, 1.37)
p=0.24
Model 22 0 (Ref.) 0.45 (−5.10, 6.00)
p=0.88		 −0.50 (−8.87, 7.87)
p=0.91		 1.46 (−1.35, 4.26)
p=0.31
Model 33 0 (Ref.) 0.61 (−4.89, 6.11)
p=0.83		 −1.60 (−9.72, 6.51)
p=0.70		 1.58 (−1.24, 4.40)
p=0.27
Open in a new tab
1

Model 1 adjusts for age, sex, race (African-American, non-African-American), clinic site, season (as harmonic)

2

Model 2 adjusts for age, sex, race (African-American, non-African-American), clinic site, season (as harmonic), alcohol consumption (any vs. none), BMI, and waist circumference

3

Model 3 adjusts for age, sex, race (African-American, non-African-American), clinic site, season (as harmonic), alcohol consumption (any vs. none), BMI, waist circumference, physical activity (kcal), C-reactive protein, cystatin-C, exercise intensity, CES-D score, serum albumin, free fatty acids, FABP4, fetuin-A, total adiponectin.

Figure 1.

Figure 1

Open in a new tab

The fully adjusted cross-sectional association between HOMA-IR and 25(OH)D concentrations with sample density. 25(OH)D concentrations were modeled using a natural cubic spline with 3 knots. The solid line represents predicted HOMA-IR from the fully adjusted cross-sectional model, with dotted 95% confidence intervals.

The association of vitamin D and insulin resistance did not differ significantly by race (multiplicative interaction p=0.45) or by health status. 25(OH)D was not significantly associated with HOMA-β in cross section throughout the sequence of models.

Longitudinal analysis

A total of 1469 individuals had follow-up examinations four years after their baseline examination; attrition was not related to baseline 25(OH)D (p=0.34). The mean change in HOMA-IR between baseline and the follow-up examination was −0.15, [95%CI (−0.07, −0.23)]. The mean HOMA-IR [95% CI] change per baseline 25(OH)D category (≥30,15–29, <15 ng/mL) was −0.15[−0.26, −.05], −0.13[−0.25, −0.02], and −0.16[−0.44,0.11], respectively. In partially and fully adjusted models, baseline 25(OH)D concentrations were not associated with a change in either HOMA-IR or HOMA-β (Table III). Including the fifty individuals who developed diabetes between examinations and setting HOMA-IR values to the 99th percentile for these individuals did not significantly alter the findings.

Table 3.

Longitudinal association between 25-hydroxyvitamin D measured on specimens collected at 1992–93 examination and change in HOMA-IR and HOMA-β between 1992–93 and 1996–97.

Mean change in HOMA measures associated with concentrations of
25-hydroxyvitamin D
Categories of 25-hydroxyvitamin D
concentrations, ng/ml		 Per 10 ng/ml
greater
concentration of
25-hydroxyvitamin
D
≥30
(n=484)		 15–29
(n=780)		 <15
(n=205)
HOMA- IR
Model 11 0 (Ref.) 0.07 (−0.08, 0.23)
p=0.36		 0.10 (−0.20, 0.41)
p=0.50		 −0.03 (−0.11, 0.04)
p=0.38
Model 22 0 (Ref.) 0.07 (−0.09, 0.23)
p=0.40		 0.11 (−0.18, 0.40)
p=0.45		 −0.03 (−0.11, 0.05)
p=0.48
HOMA- β
Model 11 0 (Ref.) −1.06 (−7.39, 5.26)
p=0.74)		 1.01 (−8.16, 10.18)
p=0.83		 −0.73 (−3.43, 1.97)
p=0.60
Model 22 0 (Ref.) −1.94 (−8.37, 4.49)
p=0.55		 −0.14 (−9.14, 8.87)
p=0.98		 −0.14 (−2.84, 2.56)
p=0.92
Open in a new tab
1

Model 1 adjusts for age, sex, race (African-American, non-African-American), clinic site, season (as harmonic)

2

Model 2 adjusts for age, sex, race (African-American, non-African-American), clinic site, season (as harmonic), alcohol consumption (any vs. none), BMI, waist circumference, physical activity (kcal), C-reactive protein, cystatin-C, exercise intensity, CES-D score, serum albumin, free fatty acids, FABP4, fetuin-A, total adiponectin.

Discussion

In a large, community-based sample of elderly individuals without cardiovascular disease, higher circulating vitamin D concentrations were associated with less insulin resistance, but not with changes in insulin resistance over four years of follow up. We observed an insignificant association of 25(OH)D concentration with insulin secretion as estimated by the HOMA-β index. When modeled as a categorical variable, only severe 25(OH) deficiency (<15 ng/mL) was associated with insulin resistance, and no category was associated with a longitudinal change in insulin resistance.

Given the associations of 25(OH)D concentrations with multiple outcomes but lack of clear effect of vitamin D supplementation in randomized trials to date, understandable concern exists about the possibility that associations such as that between 25(OH)D and insulin resistance are confounded. Indeed, in CHS, the association between 25(OH)D and insulin resistance was attenuated by the addition of BMI and waist circumference, although the subsequent addition of measures of physical activity or adipocyte-derived hormones had little additional attenuating effect. Although the residual effect of 25(OH)D after adjustment remained strong and significant, with each 10 ng/mL difference in 25(OH)D concentration having a similar effect size to a one kg/m2 difference in BMI, over half of the original association appeared attributable to confounding. At the same time, the rich covariate set available in CHS necessarily limits the range of possibly omitted confounders that could explain this association.

Although a previous longitudinal analysis associated 25(OH)D deficiency with the development of insulin resistance7, our longitudinal analysis did not find a significant association. It is plausible that our shorter duration of follow-up (four years versus ten) might explain the disparate results. In addition, we studied older adults in whom individuals with the greatest drop in insulin sensitivity may have been the most likely to be lost to attrition; indeed, average insulin sensitivity appeared to increase over time in our analyses. Furthermore, given the likelihood that the effect of 25(OH)D on insulin metabolism has a short, if any, latency, and concentrations are affected by short-term dietary intake and seasonal exposure, differential changes in 25(OH)D concentrations in the follow-up period might account for the lack of longitudinal association in our data. Further studies that contain repeated measures of both 25(OH)D concentrations and HOMA-IR, as well as rigorous intervention studies, would be useful to address these issues.

We did not find strong effect modification by race in these analyses. Scragg et al suggested an absence of an association between 25(OH)D deficiency and insulin resistance in African Americans4 and others have suggested a beneficial effect of vitamin D intake on insulin sensitivity in African American, but not European American, women23. A putative mechanistic explanation for this observation is not clear. Furthermore, recent findings suggest that total 25(OH)D concentrations do not accurately reflect biologically available vitamin D in African Americans dialysis patients, further complicating the interpretation of a modifying effect of race24. However, the number of African Americans in our sample may be too small to detect interactions of modest effect.

Strengths of this study include its large sample size, biracial and geographic diversity, richly characterized phenotypic descriptions, and longitudinal data. However, as with any observational data, residual confounding likely remains, particularly given that the effects of obesity, dietary intake, and physical activity are difficult to fully characterize. Emerging data suggests that novel dietary factors, such as fiber25 and magnesium26,27, as well as sleep patterns28,29, can effect insulin sensitivity, and were not fully accounted for in our analyses. Furthermore, we did not have information on specific vitamin D supplementation at the time of 25(OH)D measurement. In addition, our primary endpoint was defined using fasting glucose and insulin measurements, rather than more precise methods that better characterize insulin resistance. 25(OH)D concentrations were only available at baseline, and the potential effects of changing 25(OH)D concentrations were not captured. The number of participants with very low 25(OH)D levels was small, including only one with a value <5, and hence we were unable to examine associations in this subset. Finally, although CHS participants derive from a community-based study with few baseline exclusions, the expected attrition during follow-up may introduce survival bias into any longitudinal study of older adults.

Conclusion

Higher circulating 25(OH)D is not significantly associated with insulin secretion, and is associated with lower insulin resistance in cross-sectional but not longitudinal analyses. Whether this reflects residual confounding in cross-sectional analyses or the short-term nature of the relationship between vitamin D and insulin sensitivity will require well designed interventional studies.

Acknowledgements

The study authors were designated in the following roles:

Study concept and design: Danziger, Niemi, Mukamal

Acquisition of data and statistical analysis: Ix, Kizer, Djoussé, de Boer, Siscovick, Kestenbaum, Mukamal

Analysis and interpretation of data: Biggs

Drafting of the manuscript and study supervision: Danziger

Critical revision of manuscript for important intellectual content: Niemi, Biggs, Ix, Kizer, Djoussé, de Boer, Siscovick, Kestenbaum, Mukamal

Funding: This research was supported by contracts HHSN268201200036C, HHSN268200800007C, N01 HC55222, N01HC85079, N01HC85080, N01HC85081, N01HC85082, N01HC85083, N01HC85086, and grants HL094555 and HL080295 from the National Heart, Lung, and Blood Institute (NHLBI), with additional contribution from the National Institute of Neurological Disorders and Stroke (NINDS). Additional support was provided by AG023629 from the National Institute on Aging (NIA). A full list of principal CHS investigators and institutions can be found at CHS-NHLBI.org.

Abbreviations

25(OH)D

25-hydroxyvitamin D

CHS

Cardiovascular Health Study

HOMA

Homeostasis Model Assessment

BMI

Body mass index

CES-D

Center for Epidemiological Studies Depression

Footnotes

Potential Conflicts of Interest: None disclosed.

References

  • 1.Shoben AB, Kestenbaum B, Levin G, et al. Seasonal variation in 25-hydroxyvitamin D concentrations in the cardiovascular health study. Am J Epidemiol. 2011;174:1363–1372. doi: 10.1093/aje/kwr258. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Liu E, Meigs JB, Pittas AG, et al. Plasma 25-hydroxyvitamin d is associated with markers of the insulin resistant phenotype in nondiabetic adults. J Nutr. 2009;139:329–334. doi: 10.3945/jn.108.093831. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Kositsawat J, Freeman VL, Gerber BS, Geraci S. Association of A1C levels with vitamin D status in U.S. adults: data from the National Health and Nutrition Examination Survey. Diabetes Care. 2010;33:1236–1238. doi: 10.2337/dc09-2150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Scragg R, Sowers M, Bell C. Serum 25-hydroxyvitamin D, diabetes, and ethnicity in the Third National Health and Nutrition Examination Survey. Diabetes Care. 2004;27:2813–2818. doi: 10.2337/diacare.27.12.2813. [DOI] [PubMed] [Google Scholar]
  • 5.Forouhi NG, Luan J, Cooper A, Boucher BJ, Wareham NJ. Baseline serum 25-hydroxy vitamin d is predictive of future glycemic status and insulin resistance: the Medical Research Council Ely Prospective Study 1990–2000. Diabetes. 2008;57:2619–2625. doi: 10.2337/db08-0593. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Grimnes G, Emaus N, Joakimsen RM, et al. Baseline serum 25-hydroxyvitamin D concentrations in the Tromso Study 1994–95 and risk of developing type 2 diabetes mellitus during 11 years of follow-up. Diabet Med. 2010;27:1107–1115. doi: 10.1111/j.1464-5491.2010.03092.x. [DOI] [PubMed] [Google Scholar]
  • 7.Gagnon C, Lu ZX, Magliano DJ, et al. Low serum 25-hydroxyvitamin D is associated with increased risk of the development of the metabolic syndrome at five years: results from a national, population-based prospective study (The Australian Diabetes, Obesity and Lifestyle Study: AusDiab) J Clin Endocrinol Metab. 2012;97:1953–1961. doi: 10.1210/jc.2011-3187. [DOI] [PubMed] [Google Scholar]
  • 8.Forouhi NG, Ye Z, Rickard AP, et al. Circulating 25-hydroxyvitamin D concentration and the risk of type 2 diabetes: results from the European Prospective Investigation into Cancer (EPIC)-Norfolk cohort and updated meta-analysis of prospective studies. Diabetologia. 2012;55:2173–2182. doi: 10.1007/s00125-012-2544-y. [DOI] [PubMed] [Google Scholar]
  • 9.Norman AW, Frankel JB, Heldt AM, Grodsky GM. Vitamin D deficiency inhibits pancreatic secretion of insulin. Science. 1980;209:823–825. doi: 10.1126/science.6250216. [DOI] [PubMed] [Google Scholar]
  • 10.Mitri J, Dawson-Hughes B, Hu FB, Pittas AG. Effects of vitamin D and calcium supplementation on pancreatic beta cell function, insulin sensitivity, and glycemia in adults at high risk of diabetes: the Calcium and Vitamin D for Diabetes Mellitus (CaDDM) randomized controlled trial. Am J Clin Nutr. 2011;94:486–494. doi: 10.3945/ajcn.111.011684. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Nagpal J, Pande JN, Bhartia A. A double-blind, randomized, placebo-controlled trial of the short-term effect of vitamin D3 supplementation on insulin sensitivity in apparently healthy, middle-aged, centrally obese men. Diabet Med. 2009;26:19–27. doi: 10.1111/j.1464-5491.2008.02636.x. [DOI] [PubMed] [Google Scholar]
  • 12.von Hurst PR, Stonehouse W, Coad J. Vitamin D supplementation reduces insulin resistance in South Asian women living in New Zealand who are insulin resistant and vitamin D deficient - a randomised, placebo-controlled trial. Br J Nutr. 2010;103:549–555. doi: 10.1017/S0007114509992017. [DOI] [PubMed] [Google Scholar]
  • 13.Grimnes G, Figenschau Y, Almas B, Jorde R. Vitamin D, Insulin Secretion, Sensitivity, and Lipids: Results From a Case-Control Study and a Randomized Controlled Trial Using Hyperglycemic Clamp Technique. Diabetes. 2011 doi: 10.2337/db11-0650. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.McGill AT, Stewart JM, Lithander FE, Strik CM, Poppitt SD. Relationships of low serum vitamin D3 with anthropometry and markers of the metabolic syndrome and diabetes in overweight and obesity. Nutr J. 2008;7:4. doi: 10.1186/1475-2891-7-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Biggs ML, Mukamal KJ, Luchsinger JA, et al. Association between adiposity in midlife and older age and risk of diabetes in older adults. JAMA. 2010;303:2504–2512. doi: 10.1001/jama.2010.843. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Balkau B, Mhamdi L, Oppert JM, et al. Physical activity and insulin sensitivity: the RISC study. Diabetes. 2008;57:2613–2618. doi: 10.2337/db07-1605. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Fried LP, Borhani NO, Enright P, et al. The Cardiovascular Health Study: design and rationale. Ann Epidemiol. 1991;1:263–276. doi: 10.1016/1047-2797(91)90005-w. [DOI] [PubMed] [Google Scholar]
  • 18.Tell GS, Fried LP, Hermanson B, Manolio TA, Newman AB, Borhani NO. Recruitment of adults 65 years and older as participants in the Cardiovascular Health Study. Ann Epidemiol. 1993;3:358–366. doi: 10.1016/1047-2797(93)90062-9. [DOI] [PubMed] [Google Scholar]
  • 19.Cushman M, Cornell ES, Howard PR, Bovill EG, Tracy RP. Laboratory methods and quality assurance in the Cardiovascular Health Study. Clin Chem. 1995;41:264–270. [PubMed] [Google Scholar]
  • 20.Kestenbaum B, Katz R, de Boer I, et al. Vitamin D, parathyroid hormone, and cardiovascular events among older adults. J Am Coll Cardiol. 2011;58:1433–1441. doi: 10.1016/j.jacc.2011.03.069. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28:412–419. doi: 10.1007/BF00280883. [DOI] [PubMed] [Google Scholar]
  • 22.Wang Y, Jacobs EJ, McCullough ML, et al. Comparing methods for accounting for seasonal variability in a biomarker when only a single sample is available: insights from simulations based on serum 25-hydroxyvitamin d. Am J Epidemiol. 2009;170:88–94. doi: 10.1093/aje/kwp086. [DOI] [PubMed] [Google Scholar]
  • 23.Alvarez JA, Bush NC, Choquette SS, et al. Vitamin D intake is associated with insulin sensitivity in African American, but not European American, women. Nutr Metab (Lond) 2010;7:28. doi: 10.1186/1743-7075-7-28. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Bhan I, Powe CE, Berg AH, et al. Bioavailable vitamin D is more tightly linked to mineral metabolism than total vitamin D in incident hemodialysis patients. Kidney Int. 2012;82:84–89. doi: 10.1038/ki.2012.19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Du H, van der AD, Boshuizen HC, et al. Dietary fiber and subsequent changes in body weight and waist circumference in European men and women. Am J Clin Nutr. 2010;91:329–336. doi: 10.3945/ajcn.2009.28191. [DOI] [PubMed] [Google Scholar]
  • 26.Cahill F, Shahidi M, Shea J, et al. High dietary magnesium intake is associated with low insulin resistance in the Newfoundland population. PLoS One. 2013;8:e58278. doi: 10.1371/journal.pone.0058278. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Hruby A, Ngwa JS, Renstrom F, et al. Higher magnesium intake is associated with lower fasting glucose and insulin, with no evidence of interaction with select genetic loci, in a meta-analysis of 15 CHARGE Consortium Studies. J Nutr. 2013;143:345–353. doi: 10.3945/jn.112.172049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Pamidi S, Wroblewski K, Broussard J, et al. Obstructive sleep apnea in young lean men: impact on insulin sensitivity and secretion. Diabetes Care. 2012;35:2384–2389. doi: 10.2337/dc12-0841. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Van Cauter E. Sleep disturbances and insulin resistance. Diabet Med. 2011;28:1455–1462. doi: 10.1111/j.1464-5491.2011.03459.x. [DOI] [PubMed] [Google Scholar]

RESOURCES