The Journal of Nutrition. First published ahead of print July 15, 2015 as doi: 10.3945/jn.115.211763. The Journal of Nutrition Nutritional Epidemiology
Serum Vitamin D Concentrations Are Not Associated with Insulin Resistance in Swiss Adults1–3 Pedro Marques-Vidal,4 Peter Vollenweider,4 Idris Guessous,5,8,9 Hugues Henry,6 Olivier Boulat,6 G´erard Waeber,4 and Franc¸ois R Jornayvaz7* 4
Department of Internal Medicine, 5Institute of Social and Preventive Medicine (IUMSP), 6Department of Laboratory Medicine, and Service of Endocrinology, Diabetes, and Metabolism, Lausanne University Hospital, Lausanne, Switzerland; 8Unit of Population Epidemiology, Division of Primary Care Medicine, Department of Community Medicine, Primary Care and Emergency Medicine, Geneva University Hospitals, Geneva, Switzerland; and 9Department of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, GA 7
Downloaded from jn.nutrition.org at UNIVERSITY ALABAMA BIRMINGHAM on July 20, 2015
Abstract Background: Low vitamin D status has been associated with an increased risk of developing future type 2 diabetes and insulin resistance (IR), although this has been recently questioned. Objective: We examined the association between serum vitamin D metabolites and incident IR. Methods: This was a prospective, population-based study derived from the CoLaus (Cohorte Lausannoise) study including 3856 participants (aged 51.2 6 10.4 y; 2217 women) free from diabetes or IR at baseline. IR was defined as a homeostasis model assessment (HOMA) index >2.6. Fasting plasma insulin and glucose were measured at baseline and at follow-up to calculate the HOMA index. The association of vitamin D metabolites with incident IR was analyzed by logistic regression, and the results were expressed for each independent variable as ORs and 95% CIs. Results: During the 5.5-y follow-up, 649 (16.9%) incident cases of IR were identified. Participants who developed IR had lower baseline serum concentrations of 25-hydroxyvitamin D3 [25(OH)D3 (25-hydroxycholecalciferol); 45.9 6 22.8 vs. 49.9 6 22.6 nmol/L; P < 0.001], total 25(OH)D3 (25(OH)D3 + epi-25-hydroxyvitamin D3 [3-epi-25(OH)D3]; 49.1 6 24.3 vs. 53.3 6 24.1 nmol/L; P < 0.001), and 3-epi-25(OH)D3 (4.2 6 2.9 vs. 4.3 6 2.5 nmol/L; P = 0.01) but a higher 3-epi- to total 25(OH)D3 ratio (0.09 6 0.05 vs. 0.08 6 0.04; P = 0.007). Multivariable analysis adjusting for month of sampling, age, and sex showed an inverse association between 25(OH)D3 and the likelihood of developing IR [ORs (95% CIs): 0.86 (0.68, 1.09), 0.60 (0.46, 0.78), and 0.57 (0.43, 0.75) for the second, third, and fourth quartiles compared with the first 25(OH)D3 quartile; P-trend < 0.001]. Similar associations were found between total 25(OH)D3 and incident IR. There was no significant association between 3-epi-25(OH)D3 and IR, yet a positive association was observed between the 3-epi- to total 25(OH)D3 ratio and incident IR. Further adjustment for body mass index, sedentary status, and smoking attenuated the association between 25(OH)D3, total 25(OH)D3, and the 3-epi- to total 25(OH)D3 ratio and the likelihood of developing IR. Conclusion: In the CoLaus study in healthy adults, the risk of incident IR is not associated with serum concentrations of 25(OH)D3 and total 25(OH)D3. J Nutr doi: 10.3945/jn.115.211763.
Keywords:
vitamin D, insulin resistance, prospective study, HOMA, type 2 diabetes, body mass index
Introduction The worldwide prevalence of type 2 diabetes (T2D)10 is increasing at an alarming rate and is estimated to affect 360 1 The CoLaus (Cohorte Lausannoise) study is supported by research grants from GlaxoSmithKline, the Faculty of Biology and Medicine of Lausanne, and the Swiss National Science Foundation (grants 33CSCO-122661, 33CS30-139468, and 33CS30-148401). Vitamin D measurements were supported by a research grant from the Loterie Romande. 2 Author disclosures: P Marques-Vidal, P Vollenweider, I Guessous, H Henry, O Boulat, G Waeber, and FR Jornayvaz, no conflicts of interest. 3 Supplemental Tables 1–3 are available from the ‘‘Online Supporting Material’’ link in the online posting of the article and from the same link in the online table of contents at http://jn.nutrition.org. * To whom correspondence should be addressed. E-mail: [email protected].
million individuals by 2030 (1). Apart from the classical risk factors for T2D such as obesity or physical inactivity, vitamin D status has recently been associated with T2D and insulin resistance (IR) (2, 3). Growing evidence suggests that low concentrations of vitamin D are associated with an increased risk of T2D and IR (3–8). These data are from a few prospective studies examining the association between serum 25-hydroxyvitamin D3 [25(OH)D3; 25-hydroxycholecalciferol] and future 10 Abbreviations used: CoLaus, Cohorte Lausannoise; CVD, cardiovascular disease; HOMA, homeostasis model assessment; IR, insulin resistance; T2D, type 2 diabetes; total 25(OH)D3, 25-hydroxyvitamin D3 + epi-25-hydroxyvitamin D3; VDR, vitamin D receptor; 3-epi-25(OH)D3 , 3-epi-25-hydroxyvitamin D 3; 25(OH)D 3, 25-hydroxyvitamin D 3 (25-hydroxycholecalciferol).
ã 2015 American Society for Nutrition. Manuscript received February 5, 2015. Initial review completed March 25, 2015. Revision accepted June 25, 2015. doi: 10.3945/jn.115.211763.
Copyright (C) 2015 by the American Society for Nutrition
1 of 6
Methods Recruitment. The CoLaus study was designed to assess the prevalence of cardiovascular risk factors and to identify new molecular determinants of these risk factors in the general population of the city of Lausanne, Switzerland. The study was approved by the Institutional Ethics Committee of the University of Lausanne, and all participants provided written informed consent. The sampling procedure of the CoLaus study was described previously (17). In summary, a simple, nonstratified random sample of the overall population of Lausanne was drawn. The following inclusion criteria were applied: 1) written informed consent, 2) willingness to take part in the examination and to provide blood samples, and 3) Caucasian origin. Recruitment began in June 2003 and ended in May 2006 and included 6733 participants. The evaluation included an interview, a physical examination, blood sampling, and a set of questionnaires. The follow-up visit was similar to the baseline evaluation and was performed between April 2009 and September 2012, 5.5 y, on average, after the collection of baseline data. Personal history and clinical data collection. All participants were evaluated in the morning after an overnight fast (minimum fasting time: 8 h). Participants were asked about their personal and family history of cardiovascular disease (CVD) and cardiovascular risk factors as well as their treatment. Physical activity was categorized into ‘‘none,’’ ‘‘moderate’’ (once per week), and ‘‘high’’ (twice or more per week). Personal history of CVD was considered if the participant reported having presented with angina, 2 of 6
Marques-Vidal et al.
myocardial infarction, or stroke or to have benefited from angioplasty or coronary artery bypass grafting. Body weight and height were measured with participants standing without shoes in light indoor clothes. Body weight was measured in kilograms to the nearest 100 g by using a Seca scale (Seca GmbH), which was calibrated regularly. Height was measured to the nearest 5 mm by using a Seca height gauge. Blood pressure was measured 3 times on the left arm, with an appropriately sized cuff, after a rest of at least 10 min in the seated position with the use of an Omron HEM-907 automated oscillometric sphygmomanometer (Omron Healthcare). The average of the last 2 measurements was used for analyses. Hypertension was defined as a systolic blood pressure $140 mm Hg and/or a diastolic blood pressure $90 mm Hg and/or presence of antihypertensive drug treatment. Biological data collection. Venous blood samples (50 mL) were drawn in the fasting state and kept at –80°C for further analysis. All biological assays were performed at the clinical laboratory of the Lausanne University Hospital. All measurements except for 25(OH)D3 were performed on plasma samples and conducted in a Modular P apparatus (Roche Diagnostics). Glucose was assessed by glucose dehydrogenase with a maximum interassay CV of 2.1% and a maximum intra-assay CV of 1.0%. Insulin was assessed by a solid-phase, 2-site chemiluminescent immunometric assay (Diagnostic Products Corporation) with a maximum intra-assay CV of 13.7%. HDL cholesterol was assessed by cholesterol oxidase-phenol-aminophenazone + polyethylene glycol + cyclodextrin with a maximum interassay CVof 3.6% and a maximum intra-assay CVof 0.9%. TGs were assessed by GPO-PAP with a maximum inter-assay CV of 2.9% and a maximum intra-assay CV of 1.5%. Metabolic dyslipidemia was defined as HDL cholesterol 2.82 mmol/L (250 mg/dL), which are indications for testing for diabetes on the basis of the American Diabetes Association recommendations (18). Glycated hemoglobin was not assessed at baseline. IR was assessed by the HOMA index according to Matthews et al. (19): insulin (mU/mL) 3 glucose (mmol/L)/22.5. IR was defined as a value >2.6 (20). Diabetes was defined as fasting plasma glucose $7.0 mmol/L and/or the presence of oral hypoglycemic or insulin treatment. Participants with incident diabetes but with normal IR values were not considered as presenting incident IR. An ultra-HPLC tandem-MS system was developed and validated for the quantification of vitamin D metabolites 25(OH)D3 and 3-epi25(OH)D3 in human serum samples (21). The calibrators, 3Plus1 Multilevel Serum Calibrator Set 25-OH-Vitamin D3/D2 (ChromoSystems), were standardized against the National Institute of Standards and Technology 972 reference material. Serum 25(OH)D3 and 3-epi-25(OH)D3 were expressed in nanomoles per liter (conversion factor: 1 nmol/ L = 0.4006 mg/L). The interday CV% was 4.6% at 40 nmol/L. No participant had vitamin D concentrations >375 nmol/L (hypervitaminosis). Statistical analysis. Statistical analysis was conducted by using Stata version 13.1 (StataCorp). Comparisons between subjects who developed IR and those who did not were conducted by using Kruskal-Wallis nonparametric test for skewed, non–Gaussian-distributed continuous variables; StudentÕs t test for Gaussian-distributed continuous variables; or chi-square test for categorical variables. The association of vitamin D metabolites (either quartiles or log-transformed values) with incident IR (defined as a binary variable) was analyzed by logistic regression, and the results were expressed for each independent variable as an OR and 95% CI. Analyses using quartiles were also performed for 25(OH)D3, total 25(OH)D3, 3-epi-25(OH)D3, and the 3-epi- to total 25(OH)D3 ratio. To account for vitamin D synthesis from sun exposure, regression models were adjusted for month of blood draw. Two multivariable models were used: model 1 adjusted for month, age, and sex and model 2 adjusted for month, age, sex, BMI, sedentary status, and smoking. The potential interaction between BMI and vitamin D was assessed by including interaction terms between categories of BMI and categories of vitamin D (‘‘##’’ interaction term in Stata syntax). Sensitivity analyses were conducted by using quartiles of vitamin D including participants with incident diabetes and normal IR in the incident IR group. Patients
Downloaded from jn.nutrition.org at UNIVERSITY ALABAMA BIRMINGHAM on July 20, 2015
glycemic status (4–12). Importantly, some studies relied on questionnaires for vitamin D assessment (11, 12) or measured only 25(OH)D3 (4–8, 10). 3-Epi-25-hydroxyvitamin D3 [3-epi25(OH)D3] is a stereoisomer that differs from 25(OH)D3 only at the C3 position; 25(OH)D3 is metabolized, in part, through the C3-epimerization and the resulting end product, 3-epi-25(OH)D3 binds to the vitamin D receptor (VDR) and exerts biological activity (13). Because mass and fragmentation patterns of 3-epi-25(OH)D3 equal those for 25(OH)D3, its presence results in an overestimation of 25(OH)D3 when both metabolites are not chromatographically resolved (14). None of the previous studies (3–10) took into account the 3-epi-25(OH)D3 form and thus likely overestimated 25(OH)D3 concentrations. To our knowledge, only 2 prospective studies assessed the association between vitamin D status and incidence of IR (4, 5). A small (n = 524) study that used the Ely population–based prospective data (15) showed that baseline 25(OH)D3 status was inversely associated with glycemic status, IR, and metabolic syndrome after 10 y of follow-up, independent of potential confounders such as age, sex, and BMI (4). The Australian Diabetes, Obesity, and Lifestyle Study included 6537 participants (5) and showed that at the 5-y follow-up, serum 25(OH)D3 concentrations were positively and independently associated with insulin sensitivity as evaluated by the homeostasis model assessment (HOMA) index (16). However, neither of these studies defined IR by using an HOMA index cutoff (4, 5). Finally, although most studies showed an inverse linear association between vitamin D and the risk of developing T2D, a recent prospective case-cohort study indicated that this association was no longer linear after adjusting for BMI (9). Thus, the aim of this study was to examine the association between serum concentrations of total 25(OH)D3 [i.e., 25(OH)D3 + 3-epi-25(OH)D3], 25(OH)D3, and 3-epi-25(OH)D3 and the 3-epi- to total 25(OH)D3 ratio with incident IR (defined by an HOMA index >2.6) at the 5.5-y follow-up by using data from a population-based prospective study [CoLaus (Cohorte Lausannoise) study].
with diabetes (with or without IR at baseline) were excluded from the analysis. Significance was set at P < 0.05.
Results
Associations between vitamin D and incidence of IR. After a follow-up of 5.5 y, 649 participants (16.9% of the included participants) developed IR. Characteristics of participants are shown in Table 1. Participants who developed IR were more frequently men, had a higher BMI, were more frequently sedentary, and presented more frequently with hypertension, dyslipidemia, or a personal history of CVD. Groups did not differ in terms of smoking status. Participants who developed IR had lower baseline concentrations of 25(OH)D3, total 25(OH)D3, and 3-epi-25(OH)D3 and a higher 3-epi- to total 25(OH)D3 ratio (Table 1). Similarly, participants in the lowest quartiles of vitamin D metabolites had higher incident IR (Table 2) and HOMA values at the 5.5-y follow-up, although this association was not significant after adjusting for potential confounders (Supplemental Table 2). Results of multivariable analysis of associations between serum vitamin D metabolites and incident IR are summarized in Table 3. After adjusting for month, age, and sex, a negative association was found between serum concentrations of 25(OH)D3, total 25(OH)D3, and 3-epi-25(OH)D3 and the 3-epi- to total 25(OH)D3 ratio with incident IR (Table 3). The association between total 25(OH)D3 (P-trend = 0.07), 25(OH)D3, and 3-epi-25(OH)D3 and IR was no longer significant after adjusting for BMI, sedentary status, and smoking, whereas the association with the 3-epi- to total 25(OH)D3 ratio remained significant (Table 3). There was no significant interaction between BMI and vitamin D. The use of vitamin D metabolites as log-transformed, continuous variables revealed similar findings [i.e., negative association with 25(OH)D3 and positive association with the 3-epi- to total 25(OH)D3 ratio], whereas the associations with the other metabolites were no longer significant after adjusting for BMI (Supplemental Table 3).
Discussion To our knowledge, this is one of the largest prospective cohorts analyzing the predictive effect of not only 25(OH)D3 and total 25(OH)D3 but also of 3-epi-25(OH)D3 and the 3-epi- to total 25(OH)D3 ratio on the development of IR (based on HOMA-IR). Our results suggest that in a homogenous healthy population from a single city, low concentrations of 25(OH)D3, total 25(OH)D3, and 3-epi-25(OH)D3 and a high 3-epi- to total 25(OH)D3 ratio are
n Women, n (%) Age, y BMI, kg/m2 BMI status, n (%) Normal Overweight Obese Smoking status, n (%) Never Former Current Physical activity, n (%) None Moderate High Hypertension, n (%) Dyslipidemia, n (%) History of cardiovascular disease, n (%) Serum vitamin D metabolites, nmol/L Total 25(OH)D3 25(OH)D3 3-epi-25(OH)D3 3-epi- to total 25(OH)D3 ratio
P
No IR
Incident IR
3192 1909 (59.8) 50.7 6 10.3 24.0 6 3.3
649 305 (47.0) 53.7 6 10.6 27.6 6 3.9
2078 (65.1) 983 (30.8) 131 (4.1)
148 (22.8) 350 (53.9) 151 (23.3)
1336 (41.9) 1011 (31.7) 845 (26.5)
288 (44.4) 214 (33.0) 147 (22.6)
953 (29.9) 335 (10.5) 1904 (59.7) 743 (23.3) 104 (3.3) 126 (4.0)
255 (39.3) 73 (11.3) 321 (49.5) 275 (42.4) 61 (9.4) 44 (6.8)
,0.001 ,0.001 0.001
6 6 6 6
,0.0012 ,0.0012 0.012 0.007
,0.001 ,0.001 ,0.001 ,0.001
0.13
,0.001
53.3 6 49.9 6 4.3 6 0.08 6
24.1 22.6 2.5 0.04
49.1 45.9 4.2 0.09
24.3 22.8 2.9 0.05
Values are means 6 SDs unless otherwise indicated. Statistical analysis was conducted by chi-square test for categorical variables or by StudentÕs t test for Gaussian-distributed continuous variables. IR, insulin resistance (defined as a homeostasis model assessment index .2.6); total 25(OH)D3, 25-hydroxyvitamin D3 + epi-25-hydroxyvitamin D3; 3-epi-25(OH)D3, 3-epi-25-hydroxyvitamin D3; 25(OH)D3, 25-hydroxyvitamin D3 (25-hydroxycholecalciferol). 2 Kruskal-Wallis test was used for skewed, non–Gaussian-distributed continuous variables. 1
associated with an increased risk of IR at the 5.5-y follow-up. However, these associations were no longer significant after adjusting for BMI. The molecular mechanisms by which vitamin D influences glucose metabolism are still unclear. VDRs are located in >30 tissues, including pancreatic b cells and insulin-sensitizing tissues (22). Furthermore, polymorphisms in the VDR gene have been associated with IR (23), suggesting that vitamin D likely contributes to glucose metabolism. Interestingly, although some studies on vitamin D supplementation suggest improved insulin sensitivity, the results are conflicting (24–27). Furthermore, a randomized clinical trial examining vitamin D supplementation and progression to T2D among subjects at risk is still ongoing (28); thus, the relation is yet unclear. Most studies assessing the association between vitamin D status and glucose metabolism used T2D as the primary endpoint, whereas we used IR (as defined by the HOMA index). We selected IR because it is a strong predictor of developing T2D (29, 30), although not all subjects with IR develop T2D. The use of the HOMA index to assess IR is valuable because it closely mirrors the gold-standard technique for assessing insulin sensitivity (hyperinsulinemic-euglycemic clamp) (19, 31). Contrary to most previous prospective studies (3–5), the inverse association between vitamin D metabolites and IR was no longer significant after adjusting for BMI. It should be noted that in the study by Forouhi et al. (4), only 54 of 524 (10%) subjects developed T2D and no HOMA index cutoff was used to define Vitamin D and insulin resistance risk
3 of 6
Downloaded from jn.nutrition.org at UNIVERSITY ALABAMA BIRMINGHAM on July 20, 2015
Characteristics of participants. Of the initial 6733 participants at baseline, 1626 (24.2%) had no follow-up data, 1132 (16.8%) were excluded because of diabetes or IR at baseline, 63 (0.9%) were excluded because of missing vitamin D at baseline, 36 (0.5%) were excluded because of other missing data at baseline, and 20 (0.3%) were excluded because of missing data for HOMA status at follow-up, leaving 3841 participants (57%) for analysis. Excluded participants were more often older, sedentary, overweight men who were current smokers (Supplemental Table 1). Participants excluded from the analysis were more frequently sampled in spring and summer than in autumn or winter. Baseline concentrations of 25(OH)D3, total 25(OH)D3, and 3-epi-25(OH)D3 were significantly lower in excluded subjects (Supplemental Table 1).
TABLE 1 Characteristics of healthy participants who developed and who remained free from IR at the 5.5-y follow-up1
TABLE 2 Distribution between quartiles of baseline serum concentrations of vitamin D metabolites in healthy participants who developed (n = 649) and who remained free (n = 3192) from IR at the 5.5-y follow-up1 Quartile First 25.0 725 192 23.0 767 195 1.8 745 172 0.05 766 150
(3.9–34.0) (23.9) (30.8) (4.3–31.6) (24.0) (30.1) (0.0–2.6) (24.5) (27.6) (0.0–0.06) (25.2) (24.0)
42.3 750 166 39.7 785 176 3.2 745 171 0.07 782 134
Third
(34.0–49.7) (24.7) (26.6) (31.6–46.9) (24.6) (27.1) (2.6–3.8) (24.5) (27.4) (0.06–0.08) (25.8) (21.5)
58.5 776 134 50.2 817 137 4.6 773 140 0.09 766 148
P
Fourth
(49.7–68.2) (25.6) (21.5) (46.9–64.2) (25.6) (21.1) (3.8–5.5) (25.5) (22.4) (0.08–0.10) (25.2) (23.7)
80.9 786 132 76.0 823 141 6.8 774 141 0.12 723 192
,0.001
(68.2–164) (25.9) (21.2) (64.2–149) (25.8) (21.7) (5.5–31.3) (25.5) (22.6) (0.10–0.69) (23.8) (30.8)
0.001
0.07
0.002
1 Values are medians (ranges) unless otherwise indicated. Statistical analysis was conducted by chi-square test; the P value corresponds to the chi-square comparing the distribution of participants with and without incident IR between quartiles of the different variables. IR, insulin resistance (defined as a homeostasis model assessment index .2.6); total 25(OH)D3, 25-hydroxyvitamin D3 + epi-25-hydroxyvitamin D3; 3-epi-25(OH)D3, 3-epi-25-hydroxyvitamin D3; 25(OH)D3, 25-hydroxyvitamin D3 (25-hydroxycholecalciferol). 2 n = 3037 with no IR and n = 624 with incident IR.
IR. Therefore, no clinical outcomes such as T2D or IR could be properly analyzed due to inadequate statistical power (4). A prospective study including 5200 Australian adult men and women reported a significant positive association between serum 25(OH)D3 and HOMA-insulin sensitivity (5). In this study, adjustment was performed for waist circumference and physical activity but not for BMI (although waist circumference and BMI are well correlated). Still, adjusting for waist circumference instead of BMI in our study led to similar conclusions (i.e., lack of linear association between quartiles of vitamin D metabolites and the incidence of IR; data not shown).
In our study, restricting the analysis to incident T2D led to conclusions similar to those for IR, although the inverse association between log-transformed vitamin D metabolite concentrations and incident T2D was still significant after adjusting for BMI. Adjusting for waist circumference instead of BMI also showed a significant protective effect of logtransformed 25(OH)D3 concentrations (OR: 0.79; 95% CI: 0.64, 0.97; P < 0.05) and a borderline significant effect of total 25(OH)D3 (OR: 0.82; 95% CI: 0.66, 1.01; P = 0.062). Although most studies argued for a potential link between vitamin D status and T2D risk, several points need to be considered. For
TABLE 3 Multivariable analysis of the association between quartiles of vitamin D metabolites and incident IR at the 5.5-y follow-up1 P
Quartile
Total 25(OH)D32 Incidence, cases/total n (%) Multivariable model 1 Multivariable model 2 25(OH)D3 Incidence, cases/total n (%) Multivariable model 1 Multivariable model 2 3-epi-25(OH)D3 Incidence, cases/total n (%) Multivariable model 1 Multivariable model 2 3-epi- to total 25(OH)D3 ratio Incidence, cases/total n (%) Multivariable model 1 Multivariable model 2
First
Second
Third
Fourth
Linear
Quadratic
Cubic
192/917 (20.9) 1 (ref) 1 (ref)
166/916 (18.1) 0.81 (0.63, 1.03) 0.88 (0.67, 1.14)
134/910 (14.7) 0.59 (0.45, 0.76) 0.74 (0.56, 0.99)
132/918 (14.4) 0.53 (0.40, 0.69) 0.78 (0.58, 1.06)
,0.001 0.07
0.56 0.56
0.45 0.21
195/962 (20.3) 1 (ref) 1 (ref)
176/961 (18.3) 0.86 (0.68, 1.09) 0.94 (0.72, 1.21)
137/954 (14.4) 0.60 (0.46, 0.78) 0.72 (0.54, 0.96)
141/964 (14.6) 0.57 (0.43, 0.75) 0.85 (0.63, 1.15)
,0.001 0.12
0.60 0.24
0.18 0.14
172/917 (18.8) 1 (ref) 1 (ref)
171/916 (18.7) 0.95 (0.75, 1.21) 1.14 (0.88, 1.49)
140/913 (15.3) 0.72 (0.56, 0.92) 0.95 (0.72, 1.25)
141/915 (15.4) 0.68 (0.53, 0.89) 1.02 (0.76, 1.37)
,0.001 0.80
0.99 0.75
0.25 0.18
165/916 (18.0) 1 (ref) 1 (ref)
142/916 (15.5) 0.78 (0.60, 1.00) 0.79 (0.60, 1.04)
160/914 (17.5) 0.86 (0.67, 1.10) 0.95 (0.73, 1.25)
211/915 (23.1) 1.24 (0.98, 1.57) 1.29 (0.99, 1.67)
0.06 0.03
0.03 0.03
0.96 0.60
1 Values are ORs (95% CIs) unless otherwise indicated. Statistical analysis was conducted by logistic regression. Multivariable model 1 adjusted for month, age, and sex; multivariable model 2 adjusted for month, age, sex, BMI, physical activity, and smoking status. IR, insulin resistance (defined as a homeostasis model assessment index .2.6); ref, reference; total 25(OH)D3, 25-hydroxyvitamin D3 + epi-25hydroxyvitamin D3; 3-epi-25(OH)D3, 3-epi-25-hydroxyvitamin D3; 25(OH)D3, 25-hydroxyvitamin D3 (25-hydroxycholecalciferol). 2 n = 3037 with no IR and n = 624 with incident IR.
4 of 6
Marques-Vidal et al.
Downloaded from jn.nutrition.org at UNIVERSITY ALABAMA BIRMINGHAM on July 20, 2015
Total 25(OH)D3,2 nmol/L No IR, n (%) Incident IR, n (%) 25(OH)D3, nmol/L No IR, n (%) Incident IR, n (%) 3-epi-25(OH)D3, nmol/L No IR, n (%) Incident IR, n (%) 3-epi to total 25(OH)D3 ratio No IR, n (%) Incident IR, n (%)
Second
results warrant further research in the assessment of a potential causal association between vitamin D and IR. Acknowledgments We thank Andrew A Dwyer for his careful reading of the manuscript. PM-V and FRJ analyzed the data and wrote part of the manuscript; PV performed the experiments, reviewed and edited the manuscript, and contributed to the discussion; IG reviewed and edited the manuscript; HH and OB performed vitamin D measurements and reviewed and edited the manuscript; GW performed the experiments, reviewed and edited the manuscript, and contributed to the discussion; and PM-V had full access to the data and is the guarantor of the study. All authors read and approved the final version of the manuscript.
References 1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care 2004;27:1047–53. Pittas AG, Chung M, Trikalinos T, Mitri J, Brendel M, Patel K, Lichtenstein AH, Lau J, Balk EM. Systematic review: vitamin D and cardiometabolic outcomes. Ann Intern Med 2010;152:307–14. Forouhi NG, Ye Z, Rickard AP, Khaw KT, Luben R, Langenberg C, Wareham NJ. 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 metaanalysis of prospective studies. Diabetologia 2012;55:2173–82. 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;57:2619–25. Gagnon C, Lu ZX, Magliano DJ, Dunstan DW, Shaw JE, Zimmet PZ, Sikaris K, Grantham N, Ebeling PR, Daly RM. Serum 25-hydroxyvitamin D, calcium intake, and risk of type 2 diabetes after 5 years: results from a national, population-based prospective study (the Australian Diabetes, Obesity and Lifestyle study). Diabetes Care 2011;34:1133–8. Knekt P, Laaksonen M, Mattila C, Harkanen T, Marniemi J, Heliovaara M, Rissanen H, Montonen J, Reunanen A. Serum vitamin D and subsequent occurrence of type 2 diabetes. Epidemiology 2008;19:666–71. Mattila C, Knekt P, Mannisto S, Rissanen H, Laaksonen MA, Montonen J, Reunanen A. Serum 25-hydroxyvitamin D concentration and subsequent risk of type 2 diabetes. Diabetes Care 2007;30:2569–70. Pittas AG, Sun Q, Manson JE, Dawson-Hughes B, Hu FB. Plasma 25hydroxyvitamin D concentration and risk of incident type 2 diabetes in women. Diabetes Care 2010;33:2021–3. Buijsse B, Boeing H, Hirche F, Weikert C, Schulze MB, Gottschald M, Kuhn T, Katzke VA, Teucher B, Dierkes J, et al. Plasma 25hydroxyvitamin D and its genetic determinants in relation to incident type 2 diabetes: a prospective case-cohort study. Eur J Epidemiol 2013;28:743–52. Scho¨ttker B, Herder C, Rothenbacher D, Perna L, Muller H, Brenner H. Serum 25-hydroxyvitamin D levels and incident diabetes mellitus type 2: a competing risk analysis in a large population-based cohort of older adults. Eur J Epidemiol 2013;28:267–75. Liu S, Song Y, Ford ES, Manson JE, Buring JE, Ridker PM. Dietary calcium, vitamin D, and the prevalence of metabolic syndrome in middle-aged and older U.S. women. Diabetes Care 2005;28:2926–32. Pittas AG, Dawson-Hughes B, Li T, Van Dam RM, Willett WC, Manson JE, Hu FB. Vitamin D and calcium intake in relation to type 2 diabetes in women. Diabetes Care 2006;29:650–6. Molnar ´ F, Sigueiro R, Sato Y, Araujo C, Schuster I, Antony P, Peluso J, Muller C, Mourino A, Moras D, et al. 1alpha,25(OH)2–3-epi-vitamin D3, a natural physiological metabolite of vitamin D3: its synthesis, biological activity and crystal structure with its receptor. PLoS One 2011;6:e18124. Singh RJ, Taylor RL, Reddy GS, Grebe SK. C-3 epimers can account for a significant proportion of total circulating 25-hydroxyvitamin D in infants, complicating accurate measurement and interpretation of vitamin D status. J Clin Endocrinol Metab 2006;91:3055–61.
Vitamin D and insulin resistance risk
5 of 6
Downloaded from jn.nutrition.org at UNIVERSITY ALABAMA BIRMINGHAM on July 20, 2015
instance, in a prospective study, an inverse association between 25(OH)D3 and T2D was reported in women but not in men (10). Similarly, an inverse association between baseline 25(OH)D3 and the risk of incident T2D was found in the NursesÕ Health Study (8), but T2D diagnosis was self-reported and not confirmed by blood sampling. A case-cohort study derived from the German arm of the EPIC (European Prospective Investigation into Cancer and Nutrition) study reported that, after adjusting for BMI and waist circumference, the inverse association between 25(OH)D3 and T2D was restricted to participants with 25(OH)D3 concentrations
Serum Vitamin D Concentrations Are Not Associated with Insulin Resistance in Swiss Adults1–3 Pedro Marques-Vidal,4 Peter Vollenweider,4 Idris Guessous,5,8,9 Hugues Henry,6 Olivier Boulat,6 G´erard Waeber,4 and Franc¸ois R Jornayvaz7* 4
Department of Internal Medicine, 5Institute of Social and Preventive Medicine (IUMSP), 6Department of Laboratory Medicine, and Service of Endocrinology, Diabetes, and Metabolism, Lausanne University Hospital, Lausanne, Switzerland; 8Unit of Population Epidemiology, Division of Primary Care Medicine, Department of Community Medicine, Primary Care and Emergency Medicine, Geneva University Hospitals, Geneva, Switzerland; and 9Department of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, GA 7
Downloaded from jn.nutrition.org at UNIVERSITY ALABAMA BIRMINGHAM on July 20, 2015
Abstract Background: Low vitamin D status has been associated with an increased risk of developing future type 2 diabetes and insulin resistance (IR), although this has been recently questioned. Objective: We examined the association between serum vitamin D metabolites and incident IR. Methods: This was a prospective, population-based study derived from the CoLaus (Cohorte Lausannoise) study including 3856 participants (aged 51.2 6 10.4 y; 2217 women) free from diabetes or IR at baseline. IR was defined as a homeostasis model assessment (HOMA) index >2.6. Fasting plasma insulin and glucose were measured at baseline and at follow-up to calculate the HOMA index. The association of vitamin D metabolites with incident IR was analyzed by logistic regression, and the results were expressed for each independent variable as ORs and 95% CIs. Results: During the 5.5-y follow-up, 649 (16.9%) incident cases of IR were identified. Participants who developed IR had lower baseline serum concentrations of 25-hydroxyvitamin D3 [25(OH)D3 (25-hydroxycholecalciferol); 45.9 6 22.8 vs. 49.9 6 22.6 nmol/L; P < 0.001], total 25(OH)D3 (25(OH)D3 + epi-25-hydroxyvitamin D3 [3-epi-25(OH)D3]; 49.1 6 24.3 vs. 53.3 6 24.1 nmol/L; P < 0.001), and 3-epi-25(OH)D3 (4.2 6 2.9 vs. 4.3 6 2.5 nmol/L; P = 0.01) but a higher 3-epi- to total 25(OH)D3 ratio (0.09 6 0.05 vs. 0.08 6 0.04; P = 0.007). Multivariable analysis adjusting for month of sampling, age, and sex showed an inverse association between 25(OH)D3 and the likelihood of developing IR [ORs (95% CIs): 0.86 (0.68, 1.09), 0.60 (0.46, 0.78), and 0.57 (0.43, 0.75) for the second, third, and fourth quartiles compared with the first 25(OH)D3 quartile; P-trend < 0.001]. Similar associations were found between total 25(OH)D3 and incident IR. There was no significant association between 3-epi-25(OH)D3 and IR, yet a positive association was observed between the 3-epi- to total 25(OH)D3 ratio and incident IR. Further adjustment for body mass index, sedentary status, and smoking attenuated the association between 25(OH)D3, total 25(OH)D3, and the 3-epi- to total 25(OH)D3 ratio and the likelihood of developing IR. Conclusion: In the CoLaus study in healthy adults, the risk of incident IR is not associated with serum concentrations of 25(OH)D3 and total 25(OH)D3. J Nutr doi: 10.3945/jn.115.211763.
Keywords:
vitamin D, insulin resistance, prospective study, HOMA, type 2 diabetes, body mass index
Introduction The worldwide prevalence of type 2 diabetes (T2D)10 is increasing at an alarming rate and is estimated to affect 360 1 The CoLaus (Cohorte Lausannoise) study is supported by research grants from GlaxoSmithKline, the Faculty of Biology and Medicine of Lausanne, and the Swiss National Science Foundation (grants 33CSCO-122661, 33CS30-139468, and 33CS30-148401). Vitamin D measurements were supported by a research grant from the Loterie Romande. 2 Author disclosures: P Marques-Vidal, P Vollenweider, I Guessous, H Henry, O Boulat, G Waeber, and FR Jornayvaz, no conflicts of interest. 3 Supplemental Tables 1–3 are available from the ‘‘Online Supporting Material’’ link in the online posting of the article and from the same link in the online table of contents at http://jn.nutrition.org. * To whom correspondence should be addressed. E-mail: [email protected].
million individuals by 2030 (1). Apart from the classical risk factors for T2D such as obesity or physical inactivity, vitamin D status has recently been associated with T2D and insulin resistance (IR) (2, 3). Growing evidence suggests that low concentrations of vitamin D are associated with an increased risk of T2D and IR (3–8). These data are from a few prospective studies examining the association between serum 25-hydroxyvitamin D3 [25(OH)D3; 25-hydroxycholecalciferol] and future 10 Abbreviations used: CoLaus, Cohorte Lausannoise; CVD, cardiovascular disease; HOMA, homeostasis model assessment; IR, insulin resistance; T2D, type 2 diabetes; total 25(OH)D3, 25-hydroxyvitamin D3 + epi-25-hydroxyvitamin D3; VDR, vitamin D receptor; 3-epi-25(OH)D3 , 3-epi-25-hydroxyvitamin D 3; 25(OH)D 3, 25-hydroxyvitamin D 3 (25-hydroxycholecalciferol).
ã 2015 American Society for Nutrition. Manuscript received February 5, 2015. Initial review completed March 25, 2015. Revision accepted June 25, 2015. doi: 10.3945/jn.115.211763.
Copyright (C) 2015 by the American Society for Nutrition
1 of 6
Methods Recruitment. The CoLaus study was designed to assess the prevalence of cardiovascular risk factors and to identify new molecular determinants of these risk factors in the general population of the city of Lausanne, Switzerland. The study was approved by the Institutional Ethics Committee of the University of Lausanne, and all participants provided written informed consent. The sampling procedure of the CoLaus study was described previously (17). In summary, a simple, nonstratified random sample of the overall population of Lausanne was drawn. The following inclusion criteria were applied: 1) written informed consent, 2) willingness to take part in the examination and to provide blood samples, and 3) Caucasian origin. Recruitment began in June 2003 and ended in May 2006 and included 6733 participants. The evaluation included an interview, a physical examination, blood sampling, and a set of questionnaires. The follow-up visit was similar to the baseline evaluation and was performed between April 2009 and September 2012, 5.5 y, on average, after the collection of baseline data. Personal history and clinical data collection. All participants were evaluated in the morning after an overnight fast (minimum fasting time: 8 h). Participants were asked about their personal and family history of cardiovascular disease (CVD) and cardiovascular risk factors as well as their treatment. Physical activity was categorized into ‘‘none,’’ ‘‘moderate’’ (once per week), and ‘‘high’’ (twice or more per week). Personal history of CVD was considered if the participant reported having presented with angina, 2 of 6
Marques-Vidal et al.
myocardial infarction, or stroke or to have benefited from angioplasty or coronary artery bypass grafting. Body weight and height were measured with participants standing without shoes in light indoor clothes. Body weight was measured in kilograms to the nearest 100 g by using a Seca scale (Seca GmbH), which was calibrated regularly. Height was measured to the nearest 5 mm by using a Seca height gauge. Blood pressure was measured 3 times on the left arm, with an appropriately sized cuff, after a rest of at least 10 min in the seated position with the use of an Omron HEM-907 automated oscillometric sphygmomanometer (Omron Healthcare). The average of the last 2 measurements was used for analyses. Hypertension was defined as a systolic blood pressure $140 mm Hg and/or a diastolic blood pressure $90 mm Hg and/or presence of antihypertensive drug treatment. Biological data collection. Venous blood samples (50 mL) were drawn in the fasting state and kept at –80°C for further analysis. All biological assays were performed at the clinical laboratory of the Lausanne University Hospital. All measurements except for 25(OH)D3 were performed on plasma samples and conducted in a Modular P apparatus (Roche Diagnostics). Glucose was assessed by glucose dehydrogenase with a maximum interassay CV of 2.1% and a maximum intra-assay CV of 1.0%. Insulin was assessed by a solid-phase, 2-site chemiluminescent immunometric assay (Diagnostic Products Corporation) with a maximum intra-assay CV of 13.7%. HDL cholesterol was assessed by cholesterol oxidase-phenol-aminophenazone + polyethylene glycol + cyclodextrin with a maximum interassay CVof 3.6% and a maximum intra-assay CVof 0.9%. TGs were assessed by GPO-PAP with a maximum inter-assay CV of 2.9% and a maximum intra-assay CV of 1.5%. Metabolic dyslipidemia was defined as HDL cholesterol 2.82 mmol/L (250 mg/dL), which are indications for testing for diabetes on the basis of the American Diabetes Association recommendations (18). Glycated hemoglobin was not assessed at baseline. IR was assessed by the HOMA index according to Matthews et al. (19): insulin (mU/mL) 3 glucose (mmol/L)/22.5. IR was defined as a value >2.6 (20). Diabetes was defined as fasting plasma glucose $7.0 mmol/L and/or the presence of oral hypoglycemic or insulin treatment. Participants with incident diabetes but with normal IR values were not considered as presenting incident IR. An ultra-HPLC tandem-MS system was developed and validated for the quantification of vitamin D metabolites 25(OH)D3 and 3-epi25(OH)D3 in human serum samples (21). The calibrators, 3Plus1 Multilevel Serum Calibrator Set 25-OH-Vitamin D3/D2 (ChromoSystems), were standardized against the National Institute of Standards and Technology 972 reference material. Serum 25(OH)D3 and 3-epi-25(OH)D3 were expressed in nanomoles per liter (conversion factor: 1 nmol/ L = 0.4006 mg/L). The interday CV% was 4.6% at 40 nmol/L. No participant had vitamin D concentrations >375 nmol/L (hypervitaminosis). Statistical analysis. Statistical analysis was conducted by using Stata version 13.1 (StataCorp). Comparisons between subjects who developed IR and those who did not were conducted by using Kruskal-Wallis nonparametric test for skewed, non–Gaussian-distributed continuous variables; StudentÕs t test for Gaussian-distributed continuous variables; or chi-square test for categorical variables. The association of vitamin D metabolites (either quartiles or log-transformed values) with incident IR (defined as a binary variable) was analyzed by logistic regression, and the results were expressed for each independent variable as an OR and 95% CI. Analyses using quartiles were also performed for 25(OH)D3, total 25(OH)D3, 3-epi-25(OH)D3, and the 3-epi- to total 25(OH)D3 ratio. To account for vitamin D synthesis from sun exposure, regression models were adjusted for month of blood draw. Two multivariable models were used: model 1 adjusted for month, age, and sex and model 2 adjusted for month, age, sex, BMI, sedentary status, and smoking. The potential interaction between BMI and vitamin D was assessed by including interaction terms between categories of BMI and categories of vitamin D (‘‘##’’ interaction term in Stata syntax). Sensitivity analyses were conducted by using quartiles of vitamin D including participants with incident diabetes and normal IR in the incident IR group. Patients
Downloaded from jn.nutrition.org at UNIVERSITY ALABAMA BIRMINGHAM on July 20, 2015
glycemic status (4–12). Importantly, some studies relied on questionnaires for vitamin D assessment (11, 12) or measured only 25(OH)D3 (4–8, 10). 3-Epi-25-hydroxyvitamin D3 [3-epi25(OH)D3] is a stereoisomer that differs from 25(OH)D3 only at the C3 position; 25(OH)D3 is metabolized, in part, through the C3-epimerization and the resulting end product, 3-epi-25(OH)D3 binds to the vitamin D receptor (VDR) and exerts biological activity (13). Because mass and fragmentation patterns of 3-epi-25(OH)D3 equal those for 25(OH)D3, its presence results in an overestimation of 25(OH)D3 when both metabolites are not chromatographically resolved (14). None of the previous studies (3–10) took into account the 3-epi-25(OH)D3 form and thus likely overestimated 25(OH)D3 concentrations. To our knowledge, only 2 prospective studies assessed the association between vitamin D status and incidence of IR (4, 5). A small (n = 524) study that used the Ely population–based prospective data (15) showed that baseline 25(OH)D3 status was inversely associated with glycemic status, IR, and metabolic syndrome after 10 y of follow-up, independent of potential confounders such as age, sex, and BMI (4). The Australian Diabetes, Obesity, and Lifestyle Study included 6537 participants (5) and showed that at the 5-y follow-up, serum 25(OH)D3 concentrations were positively and independently associated with insulin sensitivity as evaluated by the homeostasis model assessment (HOMA) index (16). However, neither of these studies defined IR by using an HOMA index cutoff (4, 5). Finally, although most studies showed an inverse linear association between vitamin D and the risk of developing T2D, a recent prospective case-cohort study indicated that this association was no longer linear after adjusting for BMI (9). Thus, the aim of this study was to examine the association between serum concentrations of total 25(OH)D3 [i.e., 25(OH)D3 + 3-epi-25(OH)D3], 25(OH)D3, and 3-epi-25(OH)D3 and the 3-epi- to total 25(OH)D3 ratio with incident IR (defined by an HOMA index >2.6) at the 5.5-y follow-up by using data from a population-based prospective study [CoLaus (Cohorte Lausannoise) study].
with diabetes (with or without IR at baseline) were excluded from the analysis. Significance was set at P < 0.05.
Results
Associations between vitamin D and incidence of IR. After a follow-up of 5.5 y, 649 participants (16.9% of the included participants) developed IR. Characteristics of participants are shown in Table 1. Participants who developed IR were more frequently men, had a higher BMI, were more frequently sedentary, and presented more frequently with hypertension, dyslipidemia, or a personal history of CVD. Groups did not differ in terms of smoking status. Participants who developed IR had lower baseline concentrations of 25(OH)D3, total 25(OH)D3, and 3-epi-25(OH)D3 and a higher 3-epi- to total 25(OH)D3 ratio (Table 1). Similarly, participants in the lowest quartiles of vitamin D metabolites had higher incident IR (Table 2) and HOMA values at the 5.5-y follow-up, although this association was not significant after adjusting for potential confounders (Supplemental Table 2). Results of multivariable analysis of associations between serum vitamin D metabolites and incident IR are summarized in Table 3. After adjusting for month, age, and sex, a negative association was found between serum concentrations of 25(OH)D3, total 25(OH)D3, and 3-epi-25(OH)D3 and the 3-epi- to total 25(OH)D3 ratio with incident IR (Table 3). The association between total 25(OH)D3 (P-trend = 0.07), 25(OH)D3, and 3-epi-25(OH)D3 and IR was no longer significant after adjusting for BMI, sedentary status, and smoking, whereas the association with the 3-epi- to total 25(OH)D3 ratio remained significant (Table 3). There was no significant interaction between BMI and vitamin D. The use of vitamin D metabolites as log-transformed, continuous variables revealed similar findings [i.e., negative association with 25(OH)D3 and positive association with the 3-epi- to total 25(OH)D3 ratio], whereas the associations with the other metabolites were no longer significant after adjusting for BMI (Supplemental Table 3).
Discussion To our knowledge, this is one of the largest prospective cohorts analyzing the predictive effect of not only 25(OH)D3 and total 25(OH)D3 but also of 3-epi-25(OH)D3 and the 3-epi- to total 25(OH)D3 ratio on the development of IR (based on HOMA-IR). Our results suggest that in a homogenous healthy population from a single city, low concentrations of 25(OH)D3, total 25(OH)D3, and 3-epi-25(OH)D3 and a high 3-epi- to total 25(OH)D3 ratio are
n Women, n (%) Age, y BMI, kg/m2 BMI status, n (%) Normal Overweight Obese Smoking status, n (%) Never Former Current Physical activity, n (%) None Moderate High Hypertension, n (%) Dyslipidemia, n (%) History of cardiovascular disease, n (%) Serum vitamin D metabolites, nmol/L Total 25(OH)D3 25(OH)D3 3-epi-25(OH)D3 3-epi- to total 25(OH)D3 ratio
P
No IR
Incident IR
3192 1909 (59.8) 50.7 6 10.3 24.0 6 3.3
649 305 (47.0) 53.7 6 10.6 27.6 6 3.9
2078 (65.1) 983 (30.8) 131 (4.1)
148 (22.8) 350 (53.9) 151 (23.3)
1336 (41.9) 1011 (31.7) 845 (26.5)
288 (44.4) 214 (33.0) 147 (22.6)
953 (29.9) 335 (10.5) 1904 (59.7) 743 (23.3) 104 (3.3) 126 (4.0)
255 (39.3) 73 (11.3) 321 (49.5) 275 (42.4) 61 (9.4) 44 (6.8)
,0.001 ,0.001 0.001
6 6 6 6
,0.0012 ,0.0012 0.012 0.007
,0.001 ,0.001 ,0.001 ,0.001
0.13
,0.001
53.3 6 49.9 6 4.3 6 0.08 6
24.1 22.6 2.5 0.04
49.1 45.9 4.2 0.09
24.3 22.8 2.9 0.05
Values are means 6 SDs unless otherwise indicated. Statistical analysis was conducted by chi-square test for categorical variables or by StudentÕs t test for Gaussian-distributed continuous variables. IR, insulin resistance (defined as a homeostasis model assessment index .2.6); total 25(OH)D3, 25-hydroxyvitamin D3 + epi-25-hydroxyvitamin D3; 3-epi-25(OH)D3, 3-epi-25-hydroxyvitamin D3; 25(OH)D3, 25-hydroxyvitamin D3 (25-hydroxycholecalciferol). 2 Kruskal-Wallis test was used for skewed, non–Gaussian-distributed continuous variables. 1
associated with an increased risk of IR at the 5.5-y follow-up. However, these associations were no longer significant after adjusting for BMI. The molecular mechanisms by which vitamin D influences glucose metabolism are still unclear. VDRs are located in >30 tissues, including pancreatic b cells and insulin-sensitizing tissues (22). Furthermore, polymorphisms in the VDR gene have been associated with IR (23), suggesting that vitamin D likely contributes to glucose metabolism. Interestingly, although some studies on vitamin D supplementation suggest improved insulin sensitivity, the results are conflicting (24–27). Furthermore, a randomized clinical trial examining vitamin D supplementation and progression to T2D among subjects at risk is still ongoing (28); thus, the relation is yet unclear. Most studies assessing the association between vitamin D status and glucose metabolism used T2D as the primary endpoint, whereas we used IR (as defined by the HOMA index). We selected IR because it is a strong predictor of developing T2D (29, 30), although not all subjects with IR develop T2D. The use of the HOMA index to assess IR is valuable because it closely mirrors the gold-standard technique for assessing insulin sensitivity (hyperinsulinemic-euglycemic clamp) (19, 31). Contrary to most previous prospective studies (3–5), the inverse association between vitamin D metabolites and IR was no longer significant after adjusting for BMI. It should be noted that in the study by Forouhi et al. (4), only 54 of 524 (10%) subjects developed T2D and no HOMA index cutoff was used to define Vitamin D and insulin resistance risk
3 of 6
Downloaded from jn.nutrition.org at UNIVERSITY ALABAMA BIRMINGHAM on July 20, 2015
Characteristics of participants. Of the initial 6733 participants at baseline, 1626 (24.2%) had no follow-up data, 1132 (16.8%) were excluded because of diabetes or IR at baseline, 63 (0.9%) were excluded because of missing vitamin D at baseline, 36 (0.5%) were excluded because of other missing data at baseline, and 20 (0.3%) were excluded because of missing data for HOMA status at follow-up, leaving 3841 participants (57%) for analysis. Excluded participants were more often older, sedentary, overweight men who were current smokers (Supplemental Table 1). Participants excluded from the analysis were more frequently sampled in spring and summer than in autumn or winter. Baseline concentrations of 25(OH)D3, total 25(OH)D3, and 3-epi-25(OH)D3 were significantly lower in excluded subjects (Supplemental Table 1).
TABLE 1 Characteristics of healthy participants who developed and who remained free from IR at the 5.5-y follow-up1
TABLE 2 Distribution between quartiles of baseline serum concentrations of vitamin D metabolites in healthy participants who developed (n = 649) and who remained free (n = 3192) from IR at the 5.5-y follow-up1 Quartile First 25.0 725 192 23.0 767 195 1.8 745 172 0.05 766 150
(3.9–34.0) (23.9) (30.8) (4.3–31.6) (24.0) (30.1) (0.0–2.6) (24.5) (27.6) (0.0–0.06) (25.2) (24.0)
42.3 750 166 39.7 785 176 3.2 745 171 0.07 782 134
Third
(34.0–49.7) (24.7) (26.6) (31.6–46.9) (24.6) (27.1) (2.6–3.8) (24.5) (27.4) (0.06–0.08) (25.8) (21.5)
58.5 776 134 50.2 817 137 4.6 773 140 0.09 766 148
P
Fourth
(49.7–68.2) (25.6) (21.5) (46.9–64.2) (25.6) (21.1) (3.8–5.5) (25.5) (22.4) (0.08–0.10) (25.2) (23.7)
80.9 786 132 76.0 823 141 6.8 774 141 0.12 723 192
,0.001
(68.2–164) (25.9) (21.2) (64.2–149) (25.8) (21.7) (5.5–31.3) (25.5) (22.6) (0.10–0.69) (23.8) (30.8)
0.001
0.07
0.002
1 Values are medians (ranges) unless otherwise indicated. Statistical analysis was conducted by chi-square test; the P value corresponds to the chi-square comparing the distribution of participants with and without incident IR between quartiles of the different variables. IR, insulin resistance (defined as a homeostasis model assessment index .2.6); total 25(OH)D3, 25-hydroxyvitamin D3 + epi-25-hydroxyvitamin D3; 3-epi-25(OH)D3, 3-epi-25-hydroxyvitamin D3; 25(OH)D3, 25-hydroxyvitamin D3 (25-hydroxycholecalciferol). 2 n = 3037 with no IR and n = 624 with incident IR.
IR. Therefore, no clinical outcomes such as T2D or IR could be properly analyzed due to inadequate statistical power (4). A prospective study including 5200 Australian adult men and women reported a significant positive association between serum 25(OH)D3 and HOMA-insulin sensitivity (5). In this study, adjustment was performed for waist circumference and physical activity but not for BMI (although waist circumference and BMI are well correlated). Still, adjusting for waist circumference instead of BMI in our study led to similar conclusions (i.e., lack of linear association between quartiles of vitamin D metabolites and the incidence of IR; data not shown).
In our study, restricting the analysis to incident T2D led to conclusions similar to those for IR, although the inverse association between log-transformed vitamin D metabolite concentrations and incident T2D was still significant after adjusting for BMI. Adjusting for waist circumference instead of BMI also showed a significant protective effect of logtransformed 25(OH)D3 concentrations (OR: 0.79; 95% CI: 0.64, 0.97; P < 0.05) and a borderline significant effect of total 25(OH)D3 (OR: 0.82; 95% CI: 0.66, 1.01; P = 0.062). Although most studies argued for a potential link between vitamin D status and T2D risk, several points need to be considered. For
TABLE 3 Multivariable analysis of the association between quartiles of vitamin D metabolites and incident IR at the 5.5-y follow-up1 P
Quartile
Total 25(OH)D32 Incidence, cases/total n (%) Multivariable model 1 Multivariable model 2 25(OH)D3 Incidence, cases/total n (%) Multivariable model 1 Multivariable model 2 3-epi-25(OH)D3 Incidence, cases/total n (%) Multivariable model 1 Multivariable model 2 3-epi- to total 25(OH)D3 ratio Incidence, cases/total n (%) Multivariable model 1 Multivariable model 2
First
Second
Third
Fourth
Linear
Quadratic
Cubic
192/917 (20.9) 1 (ref) 1 (ref)
166/916 (18.1) 0.81 (0.63, 1.03) 0.88 (0.67, 1.14)
134/910 (14.7) 0.59 (0.45, 0.76) 0.74 (0.56, 0.99)
132/918 (14.4) 0.53 (0.40, 0.69) 0.78 (0.58, 1.06)
,0.001 0.07
0.56 0.56
0.45 0.21
195/962 (20.3) 1 (ref) 1 (ref)
176/961 (18.3) 0.86 (0.68, 1.09) 0.94 (0.72, 1.21)
137/954 (14.4) 0.60 (0.46, 0.78) 0.72 (0.54, 0.96)
141/964 (14.6) 0.57 (0.43, 0.75) 0.85 (0.63, 1.15)
,0.001 0.12
0.60 0.24
0.18 0.14
172/917 (18.8) 1 (ref) 1 (ref)
171/916 (18.7) 0.95 (0.75, 1.21) 1.14 (0.88, 1.49)
140/913 (15.3) 0.72 (0.56, 0.92) 0.95 (0.72, 1.25)
141/915 (15.4) 0.68 (0.53, 0.89) 1.02 (0.76, 1.37)
,0.001 0.80
0.99 0.75
0.25 0.18
165/916 (18.0) 1 (ref) 1 (ref)
142/916 (15.5) 0.78 (0.60, 1.00) 0.79 (0.60, 1.04)
160/914 (17.5) 0.86 (0.67, 1.10) 0.95 (0.73, 1.25)
211/915 (23.1) 1.24 (0.98, 1.57) 1.29 (0.99, 1.67)
0.06 0.03
0.03 0.03
0.96 0.60
1 Values are ORs (95% CIs) unless otherwise indicated. Statistical analysis was conducted by logistic regression. Multivariable model 1 adjusted for month, age, and sex; multivariable model 2 adjusted for month, age, sex, BMI, physical activity, and smoking status. IR, insulin resistance (defined as a homeostasis model assessment index .2.6); ref, reference; total 25(OH)D3, 25-hydroxyvitamin D3 + epi-25hydroxyvitamin D3; 3-epi-25(OH)D3, 3-epi-25-hydroxyvitamin D3; 25(OH)D3, 25-hydroxyvitamin D3 (25-hydroxycholecalciferol). 2 n = 3037 with no IR and n = 624 with incident IR.
4 of 6
Marques-Vidal et al.
Downloaded from jn.nutrition.org at UNIVERSITY ALABAMA BIRMINGHAM on July 20, 2015
Total 25(OH)D3,2 nmol/L No IR, n (%) Incident IR, n (%) 25(OH)D3, nmol/L No IR, n (%) Incident IR, n (%) 3-epi-25(OH)D3, nmol/L No IR, n (%) Incident IR, n (%) 3-epi to total 25(OH)D3 ratio No IR, n (%) Incident IR, n (%)
Second
results warrant further research in the assessment of a potential causal association between vitamin D and IR. Acknowledgments We thank Andrew A Dwyer for his careful reading of the manuscript. PM-V and FRJ analyzed the data and wrote part of the manuscript; PV performed the experiments, reviewed and edited the manuscript, and contributed to the discussion; IG reviewed and edited the manuscript; HH and OB performed vitamin D measurements and reviewed and edited the manuscript; GW performed the experiments, reviewed and edited the manuscript, and contributed to the discussion; and PM-V had full access to the data and is the guarantor of the study. All authors read and approved the final version of the manuscript.
References 1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care 2004;27:1047–53. Pittas AG, Chung M, Trikalinos T, Mitri J, Brendel M, Patel K, Lichtenstein AH, Lau J, Balk EM. Systematic review: vitamin D and cardiometabolic outcomes. Ann Intern Med 2010;152:307–14. Forouhi NG, Ye Z, Rickard AP, Khaw KT, Luben R, Langenberg C, Wareham NJ. 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 metaanalysis of prospective studies. Diabetologia 2012;55:2173–82. 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;57:2619–25. Gagnon C, Lu ZX, Magliano DJ, Dunstan DW, Shaw JE, Zimmet PZ, Sikaris K, Grantham N, Ebeling PR, Daly RM. Serum 25-hydroxyvitamin D, calcium intake, and risk of type 2 diabetes after 5 years: results from a national, population-based prospective study (the Australian Diabetes, Obesity and Lifestyle study). Diabetes Care 2011;34:1133–8. Knekt P, Laaksonen M, Mattila C, Harkanen T, Marniemi J, Heliovaara M, Rissanen H, Montonen J, Reunanen A. Serum vitamin D and subsequent occurrence of type 2 diabetes. Epidemiology 2008;19:666–71. Mattila C, Knekt P, Mannisto S, Rissanen H, Laaksonen MA, Montonen J, Reunanen A. Serum 25-hydroxyvitamin D concentration and subsequent risk of type 2 diabetes. Diabetes Care 2007;30:2569–70. Pittas AG, Sun Q, Manson JE, Dawson-Hughes B, Hu FB. Plasma 25hydroxyvitamin D concentration and risk of incident type 2 diabetes in women. Diabetes Care 2010;33:2021–3. Buijsse B, Boeing H, Hirche F, Weikert C, Schulze MB, Gottschald M, Kuhn T, Katzke VA, Teucher B, Dierkes J, et al. Plasma 25hydroxyvitamin D and its genetic determinants in relation to incident type 2 diabetes: a prospective case-cohort study. Eur J Epidemiol 2013;28:743–52. Scho¨ttker B, Herder C, Rothenbacher D, Perna L, Muller H, Brenner H. Serum 25-hydroxyvitamin D levels and incident diabetes mellitus type 2: a competing risk analysis in a large population-based cohort of older adults. Eur J Epidemiol 2013;28:267–75. Liu S, Song Y, Ford ES, Manson JE, Buring JE, Ridker PM. Dietary calcium, vitamin D, and the prevalence of metabolic syndrome in middle-aged and older U.S. women. Diabetes Care 2005;28:2926–32. Pittas AG, Dawson-Hughes B, Li T, Van Dam RM, Willett WC, Manson JE, Hu FB. Vitamin D and calcium intake in relation to type 2 diabetes in women. Diabetes Care 2006;29:650–6. Molnar ´ F, Sigueiro R, Sato Y, Araujo C, Schuster I, Antony P, Peluso J, Muller C, Mourino A, Moras D, et al. 1alpha,25(OH)2–3-epi-vitamin D3, a natural physiological metabolite of vitamin D3: its synthesis, biological activity and crystal structure with its receptor. PLoS One 2011;6:e18124. Singh RJ, Taylor RL, Reddy GS, Grebe SK. C-3 epimers can account for a significant proportion of total circulating 25-hydroxyvitamin D in infants, complicating accurate measurement and interpretation of vitamin D status. J Clin Endocrinol Metab 2006;91:3055–61.
Vitamin D and insulin resistance risk
5 of 6
Downloaded from jn.nutrition.org at UNIVERSITY ALABAMA BIRMINGHAM on July 20, 2015
instance, in a prospective study, an inverse association between 25(OH)D3 and T2D was reported in women but not in men (10). Similarly, an inverse association between baseline 25(OH)D3 and the risk of incident T2D was found in the NursesÕ Health Study (8), but T2D diagnosis was self-reported and not confirmed by blood sampling. A case-cohort study derived from the German arm of the EPIC (European Prospective Investigation into Cancer and Nutrition) study reported that, after adjusting for BMI and waist circumference, the inverse association between 25(OH)D3 and T2D was restricted to participants with 25(OH)D3 concentrations
