Summary. Introduction

Transcription

1

2 Clinical Endocrinology (2013) 79, doi: /cen ORIGINAL ARTICLE Association of 25-hydroxyvitamin D with type 2 diabetes among patients undergoing coronary angiography: cross-sectional findings from the LUdwigshafen Risk and Cardiovascular Health (LURIC) Study Bríain O Hartaigh*,, G. Neil Thomas*,, Günther Silbernagel, Jos A. Bosch,,, Stefan Pilz, **, Adrian Loerbroks, Marcus E. Kleber, Tanja B. Grammer, Bernhard O. Böhm and Winfried März,, *Public Health, Epidemiology and Biostatistics, University of Birmingham, School of Sport and Exercise Sciences, University of Birmingham, Birmingham, UK, Social and Preventive Medicine, Mannheim Medical Faculty, Institute of Public Health, University of Heidelberg, Mannheim, Division of Endocrinology, Diabetology, Nephrology, Vascular Disease, and Clinical Chemistry, Department of Internal Medicine, Eberhard-Karls-University, Tübingen, Germany, Department of Clinical Psychology, University of Amsterdam, Amsterdam, The Netherlands, **Division of Endocrinology and Metabolism, Department of Internal Medicine, Medical University of Graz, Graz, Austria, Division of Endocrinology and Diabetes, Department of Internal Medicine I, Ulm University, Ulm, Germany, Clinical Institute of Medical and Chemical Laboratory Diagnostics, Medical University of Graz, Graz, Austria and Synlab Academy, Synlab Services GmbH, Mannheim, Germany Summary Objective Evidence suggests that vitamin D may protect against the onset of diabetes. However, the mechanisms underlying the role of vitamin D on glycaemic status are unclear and warrant further investigation. We sought to determine the relationship between serum 25-hydroxyvitamin D (25[OH]D) and glycaemic status among intermediate-to-high-risk patients scheduled for coronary angiography. Methods Participants were 3316 male and female patients (mean ± SD age, 627 ± 106 years). Four categories were formed according to serum 25[OH]D levels. The association between serum 25[OH]D and diabetes was assessed using multivariable logistic regression. Results Fasting and 2 h post-load glucose, HbA1c and the HOMA-IR indices diminished with increasing serum 25[OH]D levels (P < 0001). However, no associations were observed between insulin, pro-insulin or C-peptide and serum 25[OH]D concentrations. The pro-inflammatory markers IL-6 and hs-crp also decreased considerably with higher vitamin D levels (P < 0001). After full adjustment, those with optimal serum 25 [OH]D levels had a reduced odds for fasting diabetes (OR = 063; 95% CI, ; P trend = 001), 2 h post-load diabetes (OR = 046; 95% CI, ; P trend = 0004), both fasting/2 h post-load diabetes (OR = 061; 95% CI, ; Correspondence: G. Neil Thomas, Public Health, Epidemiology and Biostatistics, University of Birmingham, Birmingham, B15 2TT, UK. Tel.: ; Fax: ; gneilthomas@yahoo.co.uk P trend = 0001) and all of the combined hyperglycaemic states (OR = 068; 95% CI, ; P trend = 001). Conclusions Higher serum 25[OH]D levels were associated with better glycaemic status and lower inflammation. Should these observations be confirmed in future studies, vitamin D supplementation may prove a useful adjunct in attenuating the onset of diabetes. (Received 18 May 2012; returned for revision 6 August 2012; finally revised 9 August 2012; accepted 20 August 2012) Introduction Worldwide, diabetes is rapidly emerging as a significant threat to health and a significant challenge to health care. It has been estimated that the total population with diabetes will exceed 366 million by The ever-increasing prevalence and incidence of diabetes underlines the need for novel approaches in managing and preventing the disease. Mounting evidence supports the notion that vitamin D may be influential in modifying the risk of diabetes. 2,3 Earlier studies have linked vitamin D deficiency with impaired glucose tolerance (IGT) and type 2 diabetes. 4 8 These observations reported that vitamin D deficiency was associated with impaired insulin secretion of pancreatic b-cells and increased insulin resistance, 4,7 which are key factors in the pathogenesis of type 2 diabetes. In this regard, vitamin D supplementation has been shown to ameliorate diminished insulin activity. For example, restoration of vitamin D levels among insulin resistant and diabetic patients significantly improved glucose metabolism. 5,9 192

3 Vitamin D and type 2 diabetes 193 However, while some studies yielded an association between vitamin D and better glycaemic status, others demonstrated no effect, or even worsening. 15 In one small study, replacement of vitamin D among British Asians with diabetes resulted in increased insulin resistance and deterioration of glycaemic control. 15 Thus far, only a few large studies have assessed the association of vitamin D with insulin secretion and sensitivity, 16 whereas most other investigations have generally been small and therefore underpowered, which is maybe responsible for the heterogeneity of observations. In the present study, we aimed to extend the current limited knowledge by examining the association of vitamin D status, as determined by serum 25 hydroxyvitamin D (25[OH]D), with components of glycaemic status in a large sample of German patients scheduled for coronary angiography. Methods Patients and setting Data reported in this study come from the Ludwigshafen Risk and Cardiovascular Health (LURIC) study, which is an ongoing prospective cohort study designed to investigate environmental and genetic risk factors for cardiovascular diseases. 17 A total of 3316 patients (2309 men and 1007 women), aged years were enrolled in the LURIC study. Baseline examination was performed between July 1997 and January 2000 in a single tertiary care medical centre in South-West Germany (Herzzentrum, Ludwigshafen). Inclusion criteria were availability of a coronary angiogram, Caucasians of German ancestry to limit genetic heterogeneity and clinical stability, with the exception of acute coronary syndromes (ACS). ACS presentations included a stable pattern of chest pain (i.e. stable angina pectoris) or an unstable pattern of chest pain, suggesting unstable angina pectoris. Those with a history of malignancy within the past 5 years, any acute illness other than ACS or any predominant noncardiac disease were excluded from the study. The LURIC study was approved by the institutional review board of the ethics committee at the Landesärztekammer Rheinland- Pfalz (Mainz, Germany) and written informed consent was obtained from each patient. Baseline examination Detailed descriptions of the LURIC baseline examination have been described previously. 17 In brief, patients received a standardised individual and family history questionnaire. Current alcohol consumption was determined in patients who answered yes to the following question, Do you currently consume any alcoholic beverages. Daily physical activity was recorded using a nonvalidated 11-point scale ranging from bedridden to extremely active. Key points on the scale are as follows: 1, bed rest; 2, mostly supine; 3, not very active; 6, usual office work; 9, heavy work or sports; and 11, extremely active sport. 18 Physical activity level was then categorised into below average (not very active = 1 3), average (usual office work = 4 8) and above average (heavy work or sports = 9 11). All body measures were obtained and recorded by trained nurses belonging to the LURIC study team. Blood pressure was measured with an automated oscillometric device (Omron MX4; Omron Healthcare GmbH, Hamburg, Germany) while in the supine position for at least 10 min. Five consecutive measures of systolic and diastolic pressure were taken 30 s apart with the average determined from the last two. Body mass index (BMI) was estimated as weight (kg) divided by height squared (m 2 ). Waist circumference was measured in cm, horizontally around the smallest circumference between the ribs and iliac crest or at the navel if no natural waistline was present. Measurements Sampling of fasting venous blood was collected in the early morning and was always performed in the supine position. Routine laboratory parameters were immediately measured on a daily basis as previously mentioned. 17 Remaining blood samples were snap-frozen and stored at 80 C for further analyses. Glucose concentrations were measured by enzymatic hexokinase/ glucose 6-phosphate dehydrogenase (Roche, Mannheim, Germany). Insulin was measured by immunoenzymometric assay (Eurogenetics, Eschborn, Germany). Glycated haemoglobin (HbA1c) was measured with immunoassay (UNIMATE; Roche, Grenzach-Whylen, Germany). C-Peptide was measured by enzyme immunoassay (Eurogenetics). Interleukin-6 (IL-6) was measured with a high-sensitivity enzyme immunoassay (R&D, Wiesbaden, Germany). High-sensitivity C-reactive protein (hs- CRP) concentration was determined by immunonephelometry (Dade Behring, Marburg, Germany). Lipoproteins (LDL/HDLcholesterol and triglycerides) were measured with a ß-quantification method and enzymatic reagents from WAKO (Neuss, Germany) on a WAKO 30R analyser. 17 Cystatin C was determined by immunonephelometry (Dade Behring). Serum 25[OH] D was measured by radioimmunoassay (DiaSorin, Antony, France and Stillwater, MN, USA) with intra- and interassay coefficients of variation of 86% and 92%, respectively. Classifications All pre-diabetic and diabetic states were based solely in accordance with the revised American Diabetes Association (ADA) criteria. 19 Impaired fasting glycaemia (IFG) was diagnosed if plasma glucose concentrations were between 56 and 69 mm, and fasting type 2 diabetes mellitus was diagnosed if plasma glucose concentrations were 70 mm or HbA1c levels were 65%. Based on a 2 h post-oral glucose tolerance test (OGTT), impaired glucose tolerance (IGT) was diagnosed if plasma glucose concentrations were between 78 and 110mM, and 2 h post-load type 2 diabetes mellitus was diagnosed if plasma glucose concentrations were 111 mm. Individuals who required antidiabetic medication (i.e. oral antidiabetic and/or insulin use for control of glycaemia) were also defined as diabetic.

4 194 B. O'Hartaigh et al. Statistical analyses Based on serum 25[OH]D concentrations, four vitamin D categories were formed according to widely used cut-off values. 20,21 Severe vitamin D deficiency, <250 nm (<100 ng/ml); moderate vitamin D deficiency, nm ( ng/ml); vitamin D insufficiency, nm ( ng/ml); and vitamin D optimal range, at least 750 nm ( 300 ng/ml). We employed the homoeostatic model assessment-insulin resistance (HOMA-IR) from Matthews et al. 22 using a noncomplex equation to determine a surrogate index of insulin resistance. The HOMA-IR derived index has been shown to correlate well with the results of the euglycaemic hyperinsulinaemic clamp in populationbased studies. The equation simplifies to: HOMA-IR = (FPI * FPG/225), where FPI is fasting plasma insulin (mu/l) and FPG is fasting plasma glucose (mm). All continuous parameters following a non-normal distribution were logarithmically (base-10 logarithm) transformed before being used in statistical procedures. Categorical data are reported as percentages and, depending on their distribution, continuous data are presented as means with standard deviation (SD) values (normal distribution) or as geometric means with 95% confidence interval (95% CI) values (skewed distribution). Comparisons between groups were performed by analysis of variance (ANOVA), with a Bonferroni post hoc test for continuous parameters and chi-squared test with P for linear-by-linear test for categorical variables. Multivariable logistic regression showing odds ratios (OR) with 95% CI were employed to examine the association between vitamin D and type 2 diabetes mellitus. We used the category with the lowest vitamin D concentration as the reference group and adjusted for a range of potential confounders including age, gender, BMI, waist circumference, tobacco and alcohol use, physical activity, cystatin C, IL-6, hs-crp and seasonal change in serum 25[OH]D concentrations. To better understand the role of vitamin D in type 2 diabetes, the latter analyses were also performed using multiple linear regression where the relationship between vitamin D, fasting and 2 h post-load glucose levels were examined using continuous parameters. In this analysis, standardized regression coefficients (b) were reported as a measure of association. A two-tailed P-value of <005 was considered to be statistically significant. Data analyses were performed using SPSS 18.0 statistical package (SPSS Inc., Chicago, IL, USA). Results Of the 3316 patients who entered the study at baseline, 17 (05%) were excluded because of missing data for a number of the variables examined, leaving a total of 3299 (995%) available for analysis. Table 1 presents clinical and laboratory parameters at baseline according to vitamin D groups. Compared to the highest vitamin D category, patients in the lower groups were older, included more women, had a higher waist circumference, and a higher BMI. A larger number reported being active smokers in the lower categories. In contrast, those with the highest vitamin D concentrations contained a higher proportion of current alcohol drinkers. This category also reported participating in more physical activity. Diastolic blood pressure was significantly lower in the lowest vitamin D category compared with the others. Of the lipids, HDL-cholesterol was significantly lower among those in the lowest compared with the highest vitamin D category. As we expected, cystatin C concentrations were significantly lower in the optimal compared with the severely deficient vitamin D category. Table 1. Baseline characteristics according to 25(OH)D 25(OH)D Categories Severe deficiency (<25 nm) n = 802 (242%) Moderate deficiency ( nm) n = 1,362 (413%) Insufficient ( nm) n = 796 (242%) Optimal ( 75 nm) n = 339 (103%) P-value* Age 648 ± ± ± ± 96 <0001 Female (%) <0001 Waist circumference (cm) 991 ± ± ± ± 111 <0001 Body mass index (kg/m 2 ) 275 ± ± ± ± Active smokers (%) Active drinkers (%) <0001 Physical activity level (%) Below average <0001 Average Above average Systolic blood pressure (mmhg) 141 ± ± ± ± Diastolic blood pressure (mmhg) 79 ± ± ± ± 10 <0001 Cystatin C (mg/l) 107 (104, 111) 098 (096, 100) 095 (093, 098) 092 (090, 094) < (OH)D, 25-Hydroxyvitamin D; LDL, low density lipoprotein; HDL, high density lipoprotein. Continuous data shown as mean ± SD or Geometric mean (95% confidence interval) and categorical data shown as percentages. *ANOVA with P for trend and chi-squared test was used.

5 Vitamin D and type 2 diabetes 195 Table 2. Biochemical characteristics according to 25(OH)D status 25(OH)D Categories Severe deficiency (<25 nm) n = 802 (242%) Moderate deficiency ( nm) n = 1362 (413%) Insufficient ( nm) n = 796 (242%) Optimal ( 75 nm) n = 339 (103%) P-value* IFG/fasting T2DM (%) 369/ / / /127 <0001 IGT/2 h post-load T2DM (%) 235/ / / /138 <0001 IFG & IGT/fasting & 2 h 324/ / / /209 <0001 post-load T2DM (%) HbA1c (%) 65 ± ± ± ± 10 <0001 Fasting plasma glucose (mm) 617 (611, 630) 605 (593, 611) 587 (575, 593) 569 (558, 581) < h post-load plasma glucose (mm) 850 (816, 886) 810 (788, 834) 778 (753, 805) 735 (702, 770) <0001 Fasting insulin (mu/l) 980 (934, 1020) 965 (931, 1000) 920 (879, 963) 921 (856, 991) 018 Fasting pro-insulin (ng/l) 809 (766, 854) 841 (806, 877) 765 (724, 808) 801 (738, 808) 036 Fasting C-peptide (lg/l) 208 (197, 221) 210 (201, 219) 201 (191, 212) 196 (179, 214) 006 HOMA-IR index 280 (255, 306) 287 (266, 309) 264 (235, 293) 252 (218, 286) h post-load HOMA-IR index 3457 (3107, 3807) 2957 (2752, 3162) 2742 (2475, 3009) 2360 (2009, 2710) <0001 LDL-cholesterol (mm) 298 (291, 304) 301 (297, 306) 302 (297, 306) 306 (297, 315) 052 HDL-cholesterol (mm) 095 (093, 097) 099 (097, 100) 103 (101, 105) 106 (102, 109) <0001 Triglycerides (mm) 171 (166, 177) 171 (167, 177) 170 (164, 176) 169 (161, 178) 098 High-sensitivity C-reactive 404 (370, 448) 273 (255, 293) 269 (244, 296) 238 (205, 275) <0001 protein (mg/l) Interleukin-6 (ng/l) 551 (524, 580) 458 (442, 475) 456 (434, 479) 387 (365, 411) <0001 Vitamin supplementation (%) <0001 Antidiabetic treatment (%) <0001 Insulin therapy (%) < (OH)D, 25-Hydroxyvitamin D; IFG, impaired fasting glycaemia; T2DM, type 2 diabetes mellitus; IGT, impaired glucose tolerance; HbA1c, glycosylated haemoglobina1c; HOMA, homoeostatic model assessment for insulin resistance. Continuous data shown as mean ± SD or geometric mean (95% confidence interval). Categorical data are shown as percentages. *ANOVA with P for trend and chi-squared test was used. In Table 2, lower concentrations of HbA1c, fasting and 2 h post-load plasma glucose were significantly associated with increasing serum 25(OH)D levels. Fasting HOMA-IR attenuated marginally with increasing vitamin D concentrations. However, 2 h post-load HOMA-IR strongly declined with increasing serum 25(OH)D concentrations. In contrast, fasting insulin, pro-insulin and C-peptide were not associated with vitamin D. The proportion of diabetic patients defined using fasting or 2 h post-load glucose levels, as well as the combined states, significantly decreased with increasing levels of vitamin D status. We also found that the inflammatory markers IL-6 and hs-crp were significantly lower in the highest vitamin D category as compared to the others. Unsurprisingly, the use of vitamin supplementation was highest in patients with optimal vitamin D concentrations. Further, the proportion of patients who reported the use of antidiabetic agents increased among those with severely deficient vitamin D levels. In Table 3, ORs (with 95% CI) for fasting and 2 h post-load diabetes in the optimal versus the lowest vitamin D category were 063 (046, 086) and 046 (029, 074), respectively. For both combined, the OR (with 95% CI) for the highest versus lowest category of vitamin D concentrations was 061 (042, 074). Lastly, the OR (with 95% CI) for all combined hyperglycaemic states in the highest versus the lowest vitamin D category was 068 (052, 080). These findings remained robust even after adjusting for a range of potential confounders. As patients in the highest category of vitamin D reported a significantly greater use of vitamin supplementation, the protective effects of vitamin D against diabetes in the highest vitamin D category may have been overestimated. Subsequently, as a form of sensitivity check, we removed those who reported the use of these supplementations (n = 81). We found the current study findings did not materially differ. Further, patients who reported the use of antidiabetic treatment (n = 278) may also have influenced the current study findings. Despite this conjecture, after removing these patients, we found the present observations were not affected (data not shown). In multiple regression analyses, fasting glucose appeared to be inversely related (b = 007, P < 0001) with serum 25(OH) D, which remained robust after controlling for a number of potential confounders. Likewise, 2 h post-load glucose yielded a modest, yet significant inverse association (b = 005, P = 002) with vitamin D status following adjustment. Overall, each unit increase in serum 25(OH)D levels significantly lowered fasting and 2 h post-load glucose concentrations by 007 and 005 mm, respectively (data not shown). All analyses remained essentially unchanged when gender categories were examined separately.

6 196 B. O'Hartaigh et al. Table 3. Odds ratios for fasting and 2 h post-load diabetes and all of the combined hyperglycaemic states according to 25(OH)D status 25(OH)D Categories Severe deficiency (<25 nm) OR (95% CI) Moderate deficiency ( nm) OR (95% CI) Insufficient ( nm) OR (95% CI) Optimal ( 75 nm) OR (95% CI) P trend Model 1 T2DM (fasting) 1 (reference) 077 (062, 097) 048 (037, 063) 038 (026, 055) <0001 T2DM (2 h post-load) (051, 084) 043 (032, 056) 023 (015, 035) <0001 T2DM (fasting/2 h post-load) (066, 098) 050 (039, 063) 035 (025, 049) <0001 Combined IFG/IGT/T2DM (065, 099) 058 (046, 073) 045 (033, 060) <0001 (fasting/2 h post-load) Model 2 T2DM (fasting) (062, 100) 052 (039, 070) 049 (033, 073) <0001 T2DM (2 h post-load) (052, 088) 049 (036, 068) 030 (019, 047) <0001 T2DM (fasting/2 h post-load) (067, 103) 055 (043, 071) 043 (033, 061) <0001 Combined IFG/IGT/T2DM (066, 102) 062 (049, 080) 050 (037, 069) <0001 (fasting/2 h post-load) Model 3 T2DM (fasting) (069, 116) 065 (042, 099) 063 (046, 086) 001 T2DM (2 h post-load) (064, 114) 065 (046, 091) 046 (029, 074) 0004 T2DM (fasting/2 h post-load) (077, 121) 067 (051, 087) 061 (042, 087) 0001 Combined IFG/IGT/T2DM (fasting/2 h post-load) (078, 120) 089 (073, 116) 068 (052, 080) (OH)D, 25-Hydroxyvitamin D; OR, odds ratio; CI, confidence interval; T2DM, type 2 diabetes mellitus; IFG, impaired fasting glucose; IGT, impaired glucose tolerance. Model 1, unadjusted; model 2 adjusted for age, gender and BMI; model 3 also adjusted for smoking, alcohol, physical activity, waist circumference, cystatin C, interleukin-6, high-sensitivity C-reactive protein and seasonal change in 25(OH)D concentrations. Discussion Among these German patients, higher concentrations of vitamin D were associated with an 11 37% reduced odds of having type 2 diabetes mellitus. Likewise, several markers of glycaemic status including baseline and 2 h post-load glucose, HbA1c and HOMA-IR indices were inversely related with serum 25(OH)D levels. Our study significantly adds to the current literature because it is, to our knowledge, one of the largest to address the association of serum 25(OH)D with glucose metabolism, assessed by OGTT. Our results mainly confirm previous cross-sectional, prospective and clinical studies, which have shown patients with type 2 diabetes or impaired glucose tolerance to have lower serum 25[OH]D concentrations compared with matched control subjects, as well as the inverse relationship between indices of glycaemia and serum 25[OH]D concentrations. 3,6,23 26 Need and colleagues 27 demonstrated that low serum 25[OH]D was associated with higher fasting glucose and was most marked when concentrations were <40 nm. In the British Birth Cohort Study, elevated HbA1c was inversely associated with serum 25 [OH]D levels. 28 Further, it emerged from the Framingham Offspring Study that participants in the upper tertile for serum 25[OH]D had a 127% lower HOMA-IR score compared with those in the lowest tertile. 29 Vitamin D has been reported to improve insulin sensitivity and plays an important role in regulating insulin release. 30,31 Both in vitro and in vivo models have described vitamin D as an essential modulator for normal insulin release in response to glucose. For instance, vitamin D repletion in the early stages of experimental dietary vitamin D deficiency 32 or in subjects with hypovitaminosis D improved insulin secretion in response to glucose. 7 In line with these findings, short-term supplementation with cholecalciferol (2000 IU once daily for 16 weeks) improved b-cell function and marginally attenuated the rise in HbA1c among adults at high risk of diabetes. 11 Most studies have described a beneficial relationship between insulin and vitamin D, 3 though not all. 5,10 In a study of 510 nondiabetic Cree Indians, the observed inverse association between serum 25[OH]D, and insulin sensitivity was no longer present after correcting for BMI. 10 Infusion of vitamin D in a diabetic patient who was deficient for vitamin D over 5 months improved glucose tolerance and b-cell function, but not insulin sensitivity. 5 In this study, we showed a significant decrease in the insulin resistance indices with increasing serum 25[OH]D levels, but this association seemed mainly driven by modulation of glycaemia, with insulin, pro-insulin and C-peptide concentrations not being associated with serum 25[OH]D levels. The precise reasons why the insulinogenic parameters were not associated with serum 25[OH]D levels in the current study remain unclear, although other vitamin D-related mechanisms might be involved in mediating the onset of diabetes. To this end, low-grade systemic inflammation may contribute to the development of diabetes. Increased adipose tissue mass can lead to a significant invasion of immune cells (i.e. macrophages) resulting in an up-regulation of pro-inflammatory

7 Vitamin D and type 2 diabetes 197 cytokines such as TNF-a and IL-6, which are thought to induce key features of diabetes such as hyperglycaemia and impaired fasting glucose. 33,34 Additionally, the link between glycaemia and inflammation is likely reciprocal as persistent elevated glucose concentrations are thought to promote oxidative stress, thereby inducing inflammation at the cellular and molecular level. 33 Indeed, a large literature supports the potential role of the acute-phase protein hs-crp in diabetogenesis Elevated concentrations in response to inflammation are also strongly recognised as a useful proxy for risk of vascular diseases. 38,39 In this study, IL-6 and hs-crp attenuated considerably in response to higher serum 25[OH]D levels. It should, however, be noted that previous data on inflammation and vitamin D are partially inconsistent, but there exists evidence from interventional trials that vitamin D supplementation may reduce inflammation and may exert antiautoimmunological actions. 40,41 In the present study, several inherent limitations bear mention. Foremost, given its cross-sectional nature, future investigations employing a longitudinal design are necessary to identify a causative role for vitamin D in reducing the risk of type 2 diabetes. The LURIC cohort consists of German ancestral Caucasian participants from a single geographical area who are already at intermediate-to-high risk of coronary artery disease; therefore, caution is warranted when extrapolating the current study findings to other ethnic groups, persons from different latitudes or those belonging to the general population. Despite this, inclusion only of Caucasian participants may be considered a strength of the present work, as this may have reduced the potential heterogeneity of seasonal vitamin D variations, which are apparent among different ethnic backgrounds. 42,43 Moreover, each participant only received a single blood draw, which may not have provided an accurate long-term estimate of vitamin D status in each person. Another limitation is the use of the OR in statistical analyses, which will have overestimated the risk of diabetes associated with vitamin D status. In general, the OR is calculated in cross-sectional studies based on the understanding that it is an approximation of the prevalence ratio (PR), provided the prevalence of disease is low. However, in this study, as the prevalence of type 2 diabetes was common (>10%), the OR will have yielded an over-estimate of the true PR. Nonetheless, in spite of these limitations, the present study included a large sample size; objective measurement of vitamin D status, as opposed to relying on self-reported vitamin D intake or exposure to sun light; a comprehensive range of quantitative metabolic markers including the derived insulin/glucose product (HOMA-IR) at baseline and 2 h; and corrected for a robust range of potential confounders. In summary, it appears the role of vitamin D in the pathogenesis of type 2 diabetes is complex, involving various mechanisms that are not yet fully elucidated. Our data further highlight the need for well-conducted randomised, controlled trials with adequate vitamin D doses, to effectively assess whether this vitamin can prevent the onset of diabetes. If such studies prove the efficacy of vitamin D in the prevention of diabetes, the potential public health benefits from strategies to increase population vitamin D levels would likely be very large. Acknowledgements We thank the LURIC study team either temporarily or permanently involved in patient recruitment and sample and data handling; the laboratory staff at the Ludwigshafen General Hospital and the Universities of Freiburg, Ulm and Graz; and the German registration offices and local public health departments for their assistance. Each of the authors contributed to the manuscript consistent with the authorship criteria set by the journal, including significant input in conception, design and acquisition of data (WM, BB), analyses and interpretation of data (BόH, GNT, GS, JB, AL, SP, WM), drafting and critical revision of manuscript (all). WM and BB had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. LURIC has received funding trough the 6th Framework Program (integrated project Bloodomics, grant LSHM-CT ) and the 7th Framework Program (integrated project Atheroremo, Grant Agreement number ) of the European Union. BόH is funded by a BBSRC studentship grant to JB and GNT. References 1 Wild, S., Roglic, G., Green, A., et al. (2004) Global prevalence of diabetes: estimates for the year 2000 and projections for Diabetes Care, 27, Danescu, L.G., Levy, S. & Levy, J. (2009) Vitamin D and diabetes mellitus. Endocrine, 35, Pittas, A.G., Lau, J., Hu, F.B., et al. (2007) The role of vitamin D and calcium in type 2 diabetes. A systematic review and metaanalysis. Journal of Clinical Endocrinology and Metabolism, 92, Norman, A.W., Frankel, J.B., Heldt, A.M., et al. (1980) Vitamin D deficiency inhibits pancreatic secretion of insulin. Science, 209, Kumar, S., Davies, M., Zakaria, Y., et al. (1994) Improvement in glucose tolerance and beta-cell function in a patient with vitamin D deficiency during treatment with vitamin D. Postgraduate Medical Journal, 70, Scragg, R., Holdaway, I., Singh, V., et al. (1995) Serum 25-hydroxyvitamin D3 levels decreased in impaired glucose tolerance and diabetes mellitus. Diabetes Research and Clinical Practice, 27, Chiu, K.C., Chu, A., Go, V.L., et al. (2004) Hypovitaminosis D is associated with insulin resistance and beta cell dysfunction. American Journal of Clinical Nutrition, 79, Reis, A.F., Hauache, O.M. & Velho, G. (2005) Vitamin D endocrine system and the genetic susceptibility to diabetes, obesity and vascular disease. A review of evidence. Diabetes & Metabolism, 31, von Hurst, P.R., Stonehouse, W. & Coad, J. (2010) 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. British Journal of Nutrition, 103, Del Gobbo, L.C., Song, Y., Dannenbaum, D.A., et al. (2011) Serum 25-hydroxyvitamin D is not associated with insulin resistance or beta cell function in Canadian Cree. Journal of Nutrition, 141,

8 198 B. O'Hartaigh et al. 11 Mitri, J., Dawson-Hughes, B., Hu, F.B., et al. (2011) 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. American Journal of Clinical Nutrition, 94, Avenell, A., Cook, J.A., MacLennan, G.S., et al. (2009) Vitamin D supplementation and type 2 diabetes: a substudy of a randomised placebo-controlled trial in older people (RECORD trial, ISRCTN ). Age and Ageing, 38, de Boer, I.H., Tinker, L.F., Connelly, S., et al. (2008) Calcium plus vitamin D supplementation and the risk of incident diabetes in the Women s Health Initiative. Diabetes Care, 31, Orwoll, E., Riddle, M. & Prince, M. (1994) Effects of vitamin D on insulin and glucagon secretion in non-insulin-dependent diabetes mellitus. American Journal of Clinical Nutrition, 59, Taylor, A.V. & Wise, P.H. (1998) Vitamin D replacement in Asians with diabetes may increase insulin resistance. Postgraduate Medical Journal, 74, Scragg, R., Sowers, M., Bell, C., et al. (2004) Serum 25-hydroxyvitamin D, diabetes, and ethnicity in the Third National Health and Nutrition Examination Survey. Diabetes Care, 27, Winkelmann, B.R., Marz, W., Boehm, B.O., et al. (2001) Rationale and design of the LURIC study a resource for functional genomics, pharmacogenomics and long-term prognosis of cardiovascular disease. Pharmacogenomics, 2, O Hartaigh, B., Bosch, J.A., Pilz, S., et al. (2012) Influence of Resting Heart Rate on Mortality in Patients Undergoing Coronary Angiography (from the Ludwigshafen Risk and Cardiovascular Health [LURIC] Study). American Journal of Cardiology, 110, American Diabetes Association. (2010) Summary of revisions for the 2010 Clinical Practice Recommendations. Diabetes Care, 33, Holick, M.F. (2007) Vitamin D deficiency. New England Journal of Medicine, 357, Mosekilde, L. (2005) Vitamin D and the elderly. Clinical Endocrinology (Oxford), 62, Matthews, D.R., Hosker, J.P., Rudenski, A.S., et al. (1985) Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia, 28, Liu, E., McKeown, N.M., Pittas, A.G., et al. (2012) Predicted 25-hydroxyvitamin D score and change in fasting plasma glucose in the Framingham offspring study. European Journal of Clinical Nutrition, 66, Hutchinson, M.S., Figenschau, Y., Njolstad, I., et al. (2011) Serum 25-hydroxyvitamin D levels are inversely associated with glycated haemoglobin (HbA(1c)). The Tromso Study. Scandinavian Journal of Clinical and Laboratory Investigation, 71, Hutchinson, M.S., Figenschau, Y., Almas, B., et al. (2011) Serum 25-hydroxyvitamin D levels in subjects with reduced glucose tolerance and type 2 diabetes - the Tromso OGTT-study. International Journal for Vitamin and Nutrition Research, 81, Mattila, C., Knekt, P., Mannisto, S., et al. (2007) Serum 25-hydroxyvitamin D concentration and subsequent risk of type 2 diabetes. Diabetes Care, 30, Need, A.G., O Loughlin, P.D., Horowitz, M., et al. (2005) Relationship between fasting serum glucose, age, body mass index and serum 25 hydroxyvitamin D in postmenopausal women. Clinical Endocrinology (Oxford), 62, Hypponen, E. & Power, C. (2006) Vitamin D status and glucose homeostasis in the 1958 British birth cohort: the role of obesity. Diabetes Care, 29, Liu, E., Meigs, J.B., Pittas, A.G., et al. (2009) Plasma 25-hydroxyvitamin d is associated with markers of the insulin resistant phenotype in nondiabetic adults. Journal of Nutrition, 139, Chowdhury, T.A., Boucher, B.J. & Hitman, G.A. (2009) Vitamin D and type 2 diabetes: is there a link? Primary Care Diabetes, 3, Palomer, X., Gonzalez-Clemente, J.M., Blanco-Vaca, F., et al. (2008) Role of vitamin D in the pathogenesis of type 2 diabetes mellitus. Diabetes, Obesity & Metabolism, 10, Chertow, B.S., Sivitz, W.I., Baranetsky, N.G., et al. (1983) Cellular mechanisms of insulin release: the effects of vitamin D deficiency and repletion on rat insulin secretion. Endocrinology, 113, Kempf, K., Rose, B., Herder, C., et al. (2006) Inflammation in metabolic syndrome and type 2 diabetes: Impact of dietary glucose. Annals of the New York Academy of Sciences, 1084, Weisberg, S.P., McCann, D., Desai, M., et al. (2003) Obesity is associated with macrophage accumulation in adipose tissue. The Journal of Clinical Investigation, 112, Barzilay, J.I., Abraham, L., Heckbert, S.R., et al. (2001) The relation of markers of inflammation to the development of glucose disorders in the elderly: the Cardiovascular Health Study. Diabetes, 50, Freeman, D.J., Norrie, J., Caslake, M.J., et al. (2002) C-reactive protein is an independent predictor of risk for the development of diabetes in the West of Scotland Coronary Prevention Study. Diabetes, 51, Pradhan, A.D., Manson, J.E., Rifai, N., et al. (2001) C-reactive protein, interleukin 6, and risk of developing type 2 diabetes mellitus. JAMA, 286, Koenig, W., Sund, M., Frohlich, M., et al. (1999) C-Reactive protein, a sensitive marker of inflammation, predicts future risk of coronary heart disease in initially healthy middle-aged men: results from the MONICA (Monitoring Trends and Determinants in Cardiovascular Disease) Augsburg Cohort Study, 1984 to Circulation, 99, Ridker, P.M., Hennekens, C.H., Buring, J.E., et al. (2000) C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. New England Journal of Medicine, 342, Schleithoff, S.S., Zittermann, A., Tenderich, G., et al. (2006) Vitamin D supplementation improves cytokine profiles in patients with congestive heart failure: a double-blind, randomized, placebo-controlled trial. American Journal of Clinical Nutrition, 83, Prietl, B., Pilz, S., Wolf, M., et al. (2010) Vitamin D supplementation and regulatory T cells in apparently healthy subjects: vitamin D treatment for autoimmune diseases? The Israel Medical Association Journal, 12, Harris, S.S. & Dawson-Hughes, B. (1998) Seasonal changes in plasma 25-hydroxyvitamin D concentrations of young American black and white women. American Journal of Clinical Nutrition, 67, Levis, S., Gomez, A., Jimenez, C., et al. (2005) Vitamin d deficiency and seasonal variation in an adult South Florida population. Journal of Clinical Endocrinology and Metabolism, 90,