Effect of serum 25-hydroxyvitamin D3 on insulin resistance and β-cell function in newly diagnosed type 2 diabetes patients.

AIMS/
INTRODUCTION: To evaluate serum 25-hydroxyvitamin D3 (25(OH)D3) in newly diagnosed type 2 diabetes patients and to explore the associations of 25(OH)D3 with insulin resistance and β-cell function.
MATERIALS AND METHODS: A total of 97 newly diagnosed type 2 diabetes patients and 69 healthy controls were recruited. Serum 25(OH)D3 was determined using high-pressure liquid chromatography. Insulin resistance was measured using a homeostasis model assessment of insulin resistance (HOMA-IR). β-Cell function was determined using the HOMA β-cell function index (HOMA-β), early-phase insulin secretion index (ΔI30/ΔG30) and area under the insulin curve (AUCins). Correlation analysis was carried out using Pearson's correlation and multiple stepwise regression analysis.
RESULTS: Serum 25(OH)D3 was much lower in patients with newly diagnosed type 2 diabetes (t = -13.00, P < 0.01), and the prevalence of hypovitaminosis 25(OH)D3 was 62.9% (61/97) in diabetic patients. Among the diabetic patients, patients with hypovitaminosis 25(OH)D3 showed higher glycosylated hemoglobin and AUCglu (P < 0.01) as well as lower HOMA-β, ΔI30/ΔG30 and AUCins. Serum 25(OH)D3 was independently positively correlated with ΔI30/ΔG30 and AUCins (P < 0.05), but was not significantly correlated with either HOMA-IR or HOMA-β. Only triglycerides, glycosylated hemoglobin and ΔI30/ΔG30 emerged as independent factors associated with serum 25(OH)D3 in both diabetes patients and the health control group.
CONCLUSIONS: The present results further showed a low serum 25(OH)D3 concentration in patients with newly diagnosed type 2 diabetes. 25(OH)D3 deficiency is associated with disturbances in glucose metabolism and lipid metabolism. Serum 25(OH)D3 is not correlated with basal insulin resistance or β-cell function, but is significantly positively correlated with glucose-stimulated insulin secretion and β-cell function.

Introduction

Vitamin D deficiency or insufficiency is a risk factor in the development of diabetes1. Among many non‐calcemic functions, vitamin D acts as a necessary cofactor for insulin secretion2. The main circulating form of vitamin D, 25‐hydroxyvitamin D3 (25(OH)D3), is considered to be the best indicator of vitamin D status3. The 25(OH)D3 concentration was lower in individuals with type 2 diabetes and impaired glucose tolerance than in those with normal glucose tolerance4. Abnormal glucose tolerance has been reported to adversely affect insulin sensitivity and β‐cell function5, and the loss of β‐cell function and insulin sensitivity is known to contribute to the development of diabetes6. Therefore, vitamin D, measured as 25(OH)D3, might play a significant role in the pathogenesis of type 2 diabetes7.

The mechanisms whereby low 25(OH)D3 concentrations increase the risk of type 2 diabetes are still not well understood; the association of low serum 25(OH)D3 concentrations with type 2 diabetes might be mediated through glucose homeostasis effects. In particular, a direct effect of vitamin D on glucose homeostasis is insulin resistance and/or β‐cell dysfunction, the main pathophysiological disorders underlying type 2 diabetes8. Several studies have suggested that low vitamin D status contributed to insulin resistance9 and was associated with markers of impaired glucose metabolism, such as glycosylated hemoglobin (HbA1c)10. Vitamin D could play a role in the pathogenesis of type 2 diabetes by affecting insulin resistance, β‐cell function, or both11; there is a direct role for 25(OH)D, the major circulating metabolite precursor of the hormonally active form (25(OH)D3), in pancreatic β‐cell function and insulin sensitivity12. Cross‐sectional studies have reported that 25(OH)D is associated with insulin resistance and β‐cell function in healthy, glucose‐tolerant subjects11, and in subjects who are at risk for type 2 diabetes9. However, the effect of serum 25(OH)D3 level on insulin resistance and β‐cell function has not been well documented in diabetes patients.

The homeostasis model assessment (HOMA) of β‐cell function (HOMA‐β) and insulin resistance (HOMA‐IR) from basal plasma glucose and insulin concentrations is a widely used clinical and epidemiologic tool13 that, however, only shows what is occurring with glucose homeostasis in the fasting state14. Vitamin D is well known to play a role in the development and progression of type 2 diabetes, and a high vitamin D status provides protection against type 2 diabetes15. However, the exact mechanisms through which vitamin D affects diabetes development and progression are not yet fully understood, and the interaction of 25(OH)D3 with insulin resistance and β‐cell function has not been explored in newly diagnosed type 2 diabetes patients. Therefore, to further elucidate the role of vitamin D, measured as 25(OH)D3, on the potential mechanisms of type 2 diabetes development, the present study was designed to examine the correlation of 25(OH)D3 with insulin resistance and β‐cell function in a cohort of Chinese patients with newly diagnosed type 2 diabetes, with the combined use of HOMA and early‐phase insulin secretion index (ΔI30/ΔG30), as well as area under the insulin curve.

Materials and Methods

Patients

We prospectively evaluated 97 consecutive patients admitted to the Sichuan Academy of Medical Sciences & Sichuan Provincial People's Hospital between April and July 2008. The patients included 57 men and 40 women, aged 33–70 years (median age 52 ± 10 years), who were newly diagnosed with type 2 diabetes mellitus according to the World Health Organization criteria from 199916. All patients underwent careful clinical examination to exclude the presence of acute diabetic complications, cancer or any other complications that could influence the calcium phosphorus metabolism. Patients who reported stress (trauma, surgery or mental stimulation) within the past 3 months, who had a severe infection, chronic renal disease or alcohol use, or who were taking any antidiabetic medications or other medicine (vitamin, mineral supplements or glucocorticoids) that could influence vitamin D metabolism were also excluded from the present study. The 69 sex‐ and age‐matched healthy controls (40 men, 29 women; age 33–70 years, mean 50 ± 11 years) were recruited into this study undergoing oral glucose tolerance test and physical examination during the same period. Smoking, chronic alcoholism, past or present drug intake, abnormal blood glucose level (>140 mg/dL) and any medical conditions that could influence the calcium phosphorus metabolism constituted the exclusion criteria. The study was approved by the institutional review board for human subjects of Sichuan Academy of Medical Sciences, Sichuan Provincial People's Hospital. Written informed consent was obtained from each participant before participating in the study.

Variable Assessment

Data on baseline characteristics, including height, body mass, waist‐to‐hip ratio, systolic pressure and diastolic pressure, were measured after overnight fasting at the time of enrolment. Fasting venous blood samples were collected to measure the concentration of HbA1c, blood glucose, triglycerides, total cholesterol, low density lipoprotein cholesterol (LDL‐C), high density lipoprotein cholesterol (HDL‐C), insulin and serum 25(OH)D3. HbA1c values were measured with the Bio‐Rad glycosylated hemoglobin column assay (Bio‐Rad, Hercules, CA, USA). Blood glucose was measured by a glucose oxidase method using a UV–Vis spectrophotometer (UV751GD; Shanghai Analytical Instrument Co., Shanghai, China). Insulin concentrations were measured by electrochemiluminescence immunoassay (Roche Diagnostics, Indianapolis, IN, USA). Total cholesterol, HDL‐C, LDL‐C and triglyceride levels were determined from the blood specimen collected after an overnight fast, and measured on the Olympus AU 5400/AU2700 autoanalyzer (Olympus Co., Tokyo, Japan). The 25(OH)D3 serum levels were measured by high‐pressure liquid chromatography (Shimadzu Corporation, Kyoto, Japan) using an Aichrombond‐AQ C18 column (250 × 4 mm, 5 μm; Agilent, Los Angeles, CA, USA), as previously described17. The standard 25(OH)D3 was purchased from Dr Ehrenstorfer (Augsburg, Germany). The calibration curve was linear over the 0.5–120‐ng/mL range (r = 0.9993). The extraction recoveries of 25(OH)D3 ranged from 81.7 to 87.9%, and the inter‐ and intraday relative standard deviations were <7.7 and <9.9%, respectively. Serum levels of blood glucose and insulin were also determined at 0.5 and 2 h after the oral administration of 75 g glucose.

Assessment of Insulin Resistance and β‐Cell Function

The β‐cell function was estimated by HOMA‐β, early‐phase insulin secretion index (ΔI30/ΔG30) and area under the insulin curve (AUCins); insulin resistance was estimated using HOMA‐IR. HOMA‐β and HOMA‐IR were calculated using the following formulas: β‐cell function (%) = 20 × insulin / (glucose – 3.5); and insulin resistance = fasting serum insulin (μU/mL) × fasting plasma glucose (mmol/L) / 22.5, respectively13. The total glucose response and insulin secretion were measured by the areas under the glucose concentration curve (AUCglu) and AUCins, and AUC was calculated by the trapezoidal method. The early‐phase insulin secretion index was calculated according to the ratio of insulin change to glucose change during the first 30 min after a meal (ΔI30/ΔG30). Deficiency of 25(OH)D3 was defined as a 25(OH)D3 concentration <37.5 nmol/L18.

Statistical Analysis

All statistical analyses were carried out with SPSS, version 13.0 (SPSS Inc., Chicago, IL, USA). Data were expressed as the mean value ± standard deviation. Measurement data with a normal distribution between the groups were compared using an independent sample t‐test or one‐way analysis of variance (anova) as appropriate. Enumeration data with a normal distribution between groups were compared with a χ2‐test. Data with a non‐normal distribution were analyzed after a logarithmic transformation or square root transformation. The correlation between 25(OH)D3 and HOMA‐β, ΔI30/ΔG30 or AUCins was calculated using Pearson's correlation analysis and multiple stepwise regression analysis. A P‐value of <0.05 was considered statistically significant.

Results

Serum 25(OH)D3 Levels Between Diabetic and Control Group

In total, 166 participants were enrolled as previously described in the inclusion criteria of the present study, and there was no significant difference in age and sex distribution between the diabetic and control groups (P > 0.05). The serum 25(OH)D3 levels were significantly lower in patients with newly diagnosed diabetes (36 ± 19 nmol/L) compared with those of the healthy control group (80 ± 26 nmol/L; t = −13.00, P < 0.01). The prevalence of 25(OH)D3 deficiency was 62.89% (61/97) in the diabetic group, which was significantly higher than that of the control group, in which no one was identified as being deficient of 25(OH)D3 (χ2 = 68.60, P < 0.01).

Comparison of Clinical Data Among the Different 25(OH)D3 Level Groups

According to the distribution of 25(OH)D3 levels, patients in the diabetic group were further divided into two groups; that is, the 25(OH)D3 deficiency group and the non‐deficiency group, with a cut‐off value of 37.5 nmol/L7. Comparisons of the clinical data among the patients with or without hypovitaminosis 25(OH)D3 and healthy controls are listed in Table 1. No significant difference in age and sex distributions, body mass index (BMI), systolic blood pressure, diastolic blood pressure, Ca2+ or serum level of phosphorus was found among the three groups. There were several statistically significant differences between the patients with and without hypovitaminosis 25(OH)D3. The values of HbA1c, fasting blood glucose, 0.5‐h blood glucose, 2‐h blood glucose, AUCglu and creatinine clearance rate were significantly higher in the hypovitaminosis 25(OH)D3 group, whereas fasting insulin, 0.5‐ and 2‐h insulin, AUCins, HOMA‐β, ΔI30/ΔG30, HOMA‐IR and serum creatinine were comparatively lower (P < 0.05) compared with patients without hypovitaminosis 25(OH)D3. However, after adjustment for age, sex, BMI, waist‐to‐hip ratio, triglycerides, total cholesterol, LDL‐C and HDL‐C, differences in AUCglu, AUCins, HOMA‐β, ΔI30/ΔG30 and HOMA‐IR were found to be statistically insignificant between the groups (F = 1.65, P > 0.05). No significant difference in BMI, waist‐to‐hip ratio, systolic pressure, diastolic pressure, triglycerides, total cholesterol, LDL‐C or HDL‐C was observed between the groups. The serum 25(OH)D3 level in patients without hypovitaminosis 25(OH)D3 was comparatively lower than that of the healthy control group (t = −5.60, P < 0.01).

Table 1

Comparison of clinical characteristics among individuals in the different groups

	Control group	Diabetic group		χ2/F‐value	P‐value
		Hypovitaminosis 25(OH)D3	Non‐hypovitaminosis 25(OH)D3
n (case)	69	61	36
Sex (M/F)	40/29	41/20	16/20	4.84	>0.05
Age (years)	50 ± 11	52 ± 10	53 ± 10	1.23	>0.05
BMI (kg/m2)	24 ± 3	25 ± 4	25 ± 4	2.16	>0.05
WHR	0.90 ± 0.06	0.96 ± 0.07*	0.94 ± 0.05*	8.54	<0.01
SBP (mmHg)	125 ± 15	128 ± 16	126 ± 15	0.65	>0.05
DBP (mmHg)	74 ± 12	80 ± 11	77 ± 10	3.39	<0.05
TC (mmol/L)	4.7 ± 1.0	5.5 ± 1.3*	5.2 ± 0.7*	9.76	<0.01
TG (mmol/L)	1.3 ± 0.8	3.6 ± 3.5*, &	2.4 ± 1.6*	26.41	<0.01
LDL‐C (mmol/L)	2.9 ± 0.8	3.5 ± 1.0*	3.2 ± 0.7*	9.62	<0.01
HDL‐C (mmol/L)	1.3 ± 0.3	1.1 ± 0.3*, &	1.2 ± 0.3	9.62	<0.01
Ca2+ (mmol/L)	2.47 ± 0.11	2.43 ± 0.11	2.44 ± 0.11	1.61	>0.05
25(OH)D (nmol/L)	80 ± 26	26 ± 6*, &	52 ± 22*, &	127.66	<0.01
HbA1c (%)	5.2 ± 0.4	10.1 ± 3.0*, &	7.7 ± 2.6*	120.32	<0.01
FBG (mmol/L)	5.1 ± 0.5	10.2 ± 3.2*, &	7.9 ± 2.9*	120.78	<0.01
0.5‐h BG (mmol/L)	8.7 ± 1.7	16.0 ± 3.8*, &	13.2 ± 3.5*	113.51	<0.01
2‐h BG (mmol/L)	6.6 ± 0.8	20.9 ± 5.0*, &	16.2 ± 5.0*	470.17	<0.01
FI (mU/L)	7 ± 5	6 ± 4&	11 ± 7*	12.32	<0.01
0.5‐h Insulin (mU/L)	52 ± 33	16 ± 12*, &	50 ± 40	46.35	<0.01
2‐h Insulin (mU/L)	38 ± 22	22 ± 17*, &	96 ± 79*	42.10	<0.01
AUCglu (h/(mU/L)	14.6 ± 2.0	31.5 ± 7.3*, &	25.2 ± 6.8*	195.69	<0.01
AUCins (h/(mU/L)	74 ± 40	30 ± 21*, &	104 ± 80*	42.80	<0.01
ΔI30/ΔG30	14.5 ± 9.9	1.9 ± 1.9*, &	8.3 ± 7.7*	103.81	<0.01
HOMA‐β	110 ± 156	21 ± 16*, &	75 ± 64*	50.61	<0.01
HOMA‐IR	1.5 ± 1.1	2.6 ± 1.8*, &	3.5 ± 2.1*	23.52	<0.01
P (mmol/L)	1.31 ± 0.1	1.33 ± 0.1	1.32 ± 0.1	2.26	>0.05
Crea (μmol/L)	88.4 ± 10.3	80.1 ± 16.8*	80.9 ± 15.6*	6.47	<0.01
Ccr (mL/min)	74.2 ± 20.7	89.9 ± 31.2*, &	76.7 ± 19.3	5.58	<0.01

ΔI30/ΔG30, early‐phase insulin secretion index; AUCglu, area under the glucose curve; AUCins, area under the insulin curve; BG, blood glucose; BMI, body mass index; Ccr, creatinine clearance rate; Crea, serum creatinine; DBP, diastolic blood pressure; F, female; FBG, fasting blood glucose; FI, fasting insulin; HbA1c, glycosylated hemoglobin; HDL‐C, high‐density lipoprotein cholesterol; HOMA‐β, homeostasis model assessment of β‐cell function; HOMA‐IR, homeostasis model assessment of insulin resistance; LDL‐C, low‐density lipoprotein cholesterol; M, male; SBP, systolic blood pressure; TC, total cholesterol; TG, triglyceride; WHR, waist‐to‐hip ratio. *P < 0.05 vs control group; &P < 0.05 vs non‐hypovitaminosis 25‐hydroxyvitamin D3 (25(OH)D3) group.

Comparison of 25(OH)D3 Levels Among Patients With Different HbA1c Levels

According to the distribution of the HbA1c levels, 97 diabetic patients were divided into three groups: (i) HbA1c levels <8%; (ii) HbA1c levels 8–10%; and (iii) HbA1c levels ≥10%. The serum 25(OH)D3 levels in patients with HbA1c levels <8% (44 ± 22 nmol/L) were significantly higher than those of patients with higher HbA1c levels (HbA1c levels 8–10% and HbA1c levels ≥10% for 28 ± 7 nmol/L and 30 ± 10 nmol/L, respectively; F = 16.67, P < 0.01). The distributions of age and sex were found to be statistically insignificant (P > 0.05).

Correlation Analysis of Serum 25(OH)D3 Levels and Clinical Features

The effects of 25(OH)D3 concentration on relevant clinical features of diabetes patients were investigated (Table 2). Univariate analysis showed negative correlations between the serum 25(OH)D3 level and concentrations of triglycerides, total cholesterol, LDL‐C, and HbA1c. However, after adjustment for age, sex, BMI, waist‐to‐hip ratio, systolic and diastolic blood pressure, triglycerides, total cholesterol, LDL‐C, HDL‐C and HbA1c, multiple stepwise regression analysis showed that only triglycerides (r = −0.15, t = −2.94, P < 0.05) and HbA1c (r = −0.58, t = −3.98, P < 0.05) were independently negatively correlated with serum 25(OH)D3 concentration.

Table 2

Correlation analysis of the effect of 25‐hydroxyvitamin D3 on the clinical characteristics of diabetes patients

	Univariate analysis
	r	P
Systolic blood pressure	0.07	>0.05
Diastolic blood pressure	0.02	>0.05
BMI	0.20	>0.05
Waist‐to‐hip ratio	−0.10	>0.05
Lipid profile
Total cholesterol	−0.21	<0.05
Triglyceride	−0.31	<0.01
LDL‐C	−0.29	<0.05
HbA1c	−0.49	<0.01

Covariates considered were sex, age, body mass index (BMI), waist‐to‐hip ratio, systolic and diastolic blood pressure, glycosylated hemoglobin (HbA1c), blood glucose, triglyceride, total cholesterol, low‐density lipoprotein cholesterol (LDL‐C) and high‐density lipoprotein cholesterol.

Correlation Analysis of Serum 25(OH)D3 Levels and Insulin Resistance

With HOMA‐IR as a dependent variable, other factors, including serum 25(OH)D3 levels, age, sex, BMI, waist‐to‐hip ratio, systolic and diastolic blood pressure, triglycerides, total cholesterol, LDL‐C, HDL‐C, and HbA1c as independent variables, multiple stepwise regression analysis showed that there was no significant difference between the HOMA‐IR and serum 25(OH)D3 levels.

Correlation Analysis of Serum 25(OH)D3 Levels and β‐Cell Function

Multiple stepwise regression analysis included the HOMA‐β, ΔI30/ΔG30 and AUCins individually as dependent variables, and the serum 25(OH)D3 levels, age, sex, BMI, waist‐to‐hip ratio, systolic and diastolic blood pressure, triglyceride, total cholesterol, LDL‐C, HDL‐C, HbA1c, and HOMA‐IR as independent variables. The results showed that HOMA‐β was independently associated with HbA1c and HOMA‐IR (P < 0.01; Table 3), whereas serum 25(OH)D3 level had no independent effect on HOMA‐β. Both ΔI30/ΔG30 and AUCins were associated with sex, HbA1c, serum 25(OH)D3 levels, and HOMA‐IR (P < 0.05), respectively. Serum 25(OH)D3 level was independently positively correlated with ΔI30/ΔG30 and AUCins (Table 3).

Table 3

Multivariate analysis of the effect of 25‐hydroxyvitamin D3 on β‴cell function

Dependent variable and covariate	r	t	P
HOMA‐β
HbA1c	−2.04	−8.85	<0.01
HOMA‐IR	0.68	6.93	<0.01
ΔI30/ΔG30
Sex	0.19	2.04	<0.05
HbA1c	−2.23	−5.63	<0.01
25(OH)D3	0.71	2.21	<0.05
HOMA‐IR	0.35	2.32	<0.05
AUCins
Sex	0.99	2.23	<0.05
HbA1c	−8.58	−4.57	<0.01
25(OH)D3	7.07	4.67	<0.01
HOMA‐IR	3.96	5.62	<0.01

ΔI30/ΔG30, early‐phase insulin secretion index; 25(OH)D3, 25‐hydroxyvitamin D3 AUCins, area under the insulin curve; HbA1c, glycosylated hemoglobin; HOMA‐β, homeostasis model assessment of β‐cell function; HOMA‐IR, homeostasis model assessment of insulin resistance.

Correlation Analysis of Serum 25(OH)D3 Level and Clinical Features, Insulin Resistance, and β‐Cell Function Between Diabetic and Control Groups

Univariate analysis showed a significant correlation between serum 25(OH)D3 level and waist‐to‐hip ratio, triglycerides, total cholesterol, LDL‐C, HDL‐C, HbA1c, HOMA‐IR, HOMA‐β, ΔI30/ΔG30, AUCins, serum creatinine, and creatinine clearance rate (P < 0.01 or P < 0.05; Table 4). However, by multiple stepwise regression analysis, only triglycerides (r = −0.19, t = −3.57, P < 0.01), HbA1c (r = −0.52, t = −3.12, P < 0.01) and ΔI30/ΔG30 (r = 0.16, t = 3.92, P < 0.01) emerged as independent factors associated with serum 25(OH)D3 level in diabetic patients and the healthy control group.

Table 4

Correlation of 25‐hydroxyvitamin D3 with clinical characteristics, insulin resistance and β‐cell function of participants

	Univariate analysis
	r	P
Waist‐to‐hip ratio	−0.29	<0.01
Total cholesterol	−0.31	<0.01
Triglyceride	−0.52	<0.01
LDL‐C	−0.32	<0.01
HDL‐C	0.35	<0.01
HbA1c	−0.74	<0.01
HOMA‐IR	−0.22	<0.01
HOMA‐β	0.63	<0.01
ΔI30/ΔG30	0.73	<0.01
AUCins	0.52	<0.01
Crea	0.25	<0.01
Ccr	−0.20	<0.05

ΔI30/ΔG30, early‐phase insulin secretion index; AUCins, area under the insulin curve; Ccr, creatinine clearance rate; Crea, serum creatinine; HOMA‐β, homeostasis model assessment of β‐cell function; HOMA‐IR, homeostasis model assessment of insulin resistance. Covariates considered were gender, age, body mass index, waist‐to‐hip ratio, systolic and diastolic blood pressure, glycosylated hemoglobin (HbA1c), blood glucose, triglyceride, total cholesterol, low‐density lipoprotein cholesterol (LDL‐C) and high‐density lipoprotein cholesterol (HDL‐C).

Discussion

The current data showed that serum 25(OH)D3 level was significantly lower in patients with newly diagnosed type 2 diabetes compared with that of the healthy control group, and the prevalence of hypovitaminosis 25(OH)D3 was 62.9% in the diabetic patients, corroborating the results of the previous studies, which showed that the baseline 25(OH)D3 concentration was lower in diabetes patients and was inversely associated with the incidence of type 2 diabetes19, 20. The vitamin D status of <50 nmol/L has been previously reported to double the risk of newly diagnosed type 2 diabetes after adjustment for clinical covariables21. In further support of this association, we observed, in newly diagnosed type 2 diabetes patients with decreasing 25(OH)D3 concentration, increasing concentrations of HbA1c, blood glucose and AUCglu, with decreasing insulin levels, AUCins, HOMA‐β, ΔI30/ΔG30 and HOMA‐IR. These differences were statistically insignificant after adjustment for clinical parameters, including age, sex, BMI, waist‐to‐hip ratio, triglycerides, total cholesterol, LDL‐C and HDL‐C (F = 1.65, P > 0.05). However, hypovitaminosis D was positively associated with insulin resistance and negatively correlated with β‐cell dysfunction in type 2 diabetes11. HbA1c, a global measure of glucose homeostasis, has also been reported to inversely associate with vitamin D status, as measured by serum 25(OH)D levels10, 21, 22. Elevated HbA1c levels might be considered for screening people with vitamin D insufficiency10. Abnormalities could be a result of changes in insulin secretion and sensitivity21, and potential mechanisms contributing to vitamin D regulation of glucose homeostasis and insulin sensitivity include enhanced insulin secretion and synthesis23, and reduced inflammatory processes that reduce the functional capacity of pancreatic β‐cells expressing the vitamin D receptor and the enzyme 1‐a hydroxylase24.

HOMA is a widely used clinical and epidemiological tool for assessing β‐cell function and insulin resistance from basal (fasting) glucose and insulin concentrations25; when used appropriately, it can closely mirror the glucose clamp technique in the assessment of insulin sensitivity26. Therefore, HOMA was used to assess insulin resistance and β‐cell function. However, because HOMA was reported to only show what is occurring with glucose homeostasis in the fasting state14, AUCglu, ΔI30/ΔG30 and AUCins were also used in the present study.

Ample evidence has suggested that vitamin D activity is essential for pancreatic β‐cell function, and a higher baseline vitamin D level independently predicted better β‐cell function in subjects who were at risk for type 2 diabetes27, 28, 29. Insulin resistance has also been reported to associate with vitamin D insufficiency30, and data on both associations in a single study have been reported. The concentration of 25(OH)D was found to be independently associated with both insulin sensitivity and β‐cell function among individuals at risk of type 2 diabetes9, 11, though another study showed that serum 25(OH)D was not associated with insulin resistance or β‐cell function in the Canadian Cree31. Such controversy might a result of the variability of the method used or ethnic differences. In the present study, our multiple stepwise regression analysis was carried out to investigate the association of 25(OH)D3 levels with the IR and β‐cell function. Our results showed no independent effect of serum 25(OH)D3 level on the insulin resistance index (HOMA‐IR) derived in the basal state, suggesting that serum 25(OH)D3 might not associate with basal or hepatic insulin sensitivity in newly diagnosed type 2 diabetes patients. β‐Cell function was determined using HOMA‐β, ΔI30/ΔG30 and AUCins. The results showed that there was no significant correlation between serum 25(OH)D3 and HOMA‐β, a marker of basal insulin secretion of pancreatic β‐cells, and the latter was found to be independently associated with HbA1c and HOMA‐IR, corroborating the previous results that insulin sensitivity and secretion are mutually related, and that HOMA‐β alone might not accurately assess β‐cell function32. Previous studies have shown that vitamin D affects β‐cells by increasing the insulin response to glucose stimulation and does not affect basal insulin secretion33. HOMA‐β can be considered one aspect of β‐cell function, with the ability to increase basal insulin secretion to meet the elevation of basal glycemia34. Therefore, a correlation analysis of serum 25(OH)D3 concentration and ΔI30/ΔG30 as well as AUCins was carried out. The results showed that serum 25(OH)D3 concentration was significantly positively correlated with ΔI30/ΔG30 and AUCins; serum 25(OH)D3 was also shown to correlate with sex, HbA1c and HOMA‐IR. These results showed that the serum 25(OH)D3 concentration was less correlated with the basal insulin resistance and basal β‐cell function, and was significantly correlated with glucose‐stimulated insulin secretion.

Abnormal fat accumulation in the liver, muscle and pancreatic islets plays a crucial role in the development of both β‐cell dysfunction and insulin resistance35. Our findings showed that triacylglycerol emerged as an independent factor that is negatively associated with serum 25(OH)D3 level. A previous study by Liu et al.36 reported that vitamin D, as measured by plasma 25(OH)D, was inversely associated with plasma triacylglycerol in non‐diabetic adults after adjusting for age and sex, but these associations were no longer significant after further adjustment for BMI, waist circumference and current smoking status. Further study is, therefore, still required to clarify the correlation of serum 25(OH)D3 level and triglycerides, and the potential mechanism of their effects on β‐cell dysfunction and insulin resistance.

The present study also had some limitations that should be considered when interpreting the results. Most importantly, it was a cross‐sectional study; some parameters and confounding factors, including different, non‐standardized meals taken by our patients and the degree of proteinuria or preferable urinary albumin excretion, were not considered in the present study. Thus, we can only show the existence of a correlation, but not a causal nexus between hypovitaminosis D and glucose tolerance in type 2 diabetes. In addition, the present study was restricted to one site and took place in a relatively limited number of patients, indicating that interpretation of the results needs to be done with caution. A large cohort of prospective studies should be further carried out to establish the link between hypovitaminosis D and development of type 2 diabetes in newly diagnosed type 2 diabetes patients.

In conclusion, the present findings support earlier results that showed a low serum 25(OH)D3 concentration in patients with newly diagnosed type 2 diabetes. Vitamin D deficiency is associated with disturbances in glucose and lipid metabolism. Serum 25(OH)D3 concentration is not correlated with basal insulin resistance or basal β‐cell function, but is significantly positively correlated with glucose‐stimulated insulin secretion and β‐cell function.