The vitamin D status is negatively correlated to glucose metabolism in Chinese patients with gestational diabetes

Background: Vitamin D deciency plays a role in the development of patients with gestational diabetes (GDM). However, the effects and mechanism of vitamin D on fasting blood glucose are conicting. This research aimed to explore the status of vitamin D in patients with GDM, its effect on fasting blood glucose and HbA1c, and to clarify the possible mechanism of vitamin D on glucose metabolism of patients with GDM. Methods: From January 1, 2017, to March 31, 2018, we included all the hospitalized patients with GDM. The main indices were fasting blood glucose (FBG) and 25-hydroxy vitamin D [25(OH) D], glycosylated hemoglobin (HbA1c), fasting insulin (FINS), fasting peptide (FPC), parathyroid hormone (iPTH), serum creatinine (SCr), blood neutrophil count, lymphocyte count, blood calcium (Ca), and blood phosphorus (P), neutrophil count (Ne), lymphocyte count (Lym), neutrophil-to-lymphocyte ratio (NLR). Results The mean 25(OH) D concentration of 33 patients with GDM was 21.40±10.68 ng/mL, among them, the patients with vitamin D insuciency was 30.30% (10 cases), the patients with vitamin D deciency was 60.61%(20 cases), the patients with normal vitamin D level was 9.1%(3 cases). Adjusted by age, gestational age, and glomerular ltration rate (eGFR), 25(OH)D concentration were negatively correlated to and FBG ( β = -0.587, p =0.001), HbA1c ( β = -0.408, p =0.018 ), respectively. There was a negative correlation between 25(OH)D concentrations and NLR ( β = -0.389, p = 0.035), adjusted by eGFR, gestational age, and age. Conclusions: The status of vitamin D deciency in patients with GDM is common. With the decrease of


Background
The prevalence of vitamin D insu ciency or de ciency is high in women with pregnancy [1]. Homles et al. [2] found that both pregnant and nonpregnant women have vitamin D de ciency. Li found the pregnant women with vitamin D de ciency (< 75 nmol/L) was 62.34%, and the proportion of severe de ciency (< 25 nmol/L) was 0.25% [3]. Vitamin D de ciency, de ned as serum 25(OH) D concentration above 50 nmol/L, is found to be common globally in pregnancy with a prevalence of 26 to-98% [4]. Some researches in the Chinese population have found that the status of vitamin D in patients with GDM is lower than that in pregnant women with normal blood glucose [5]. GDM has been recognized as a risk factor for many complications of pregnant women and fetus. In the short term, it increases the risk of Csection [6] [7], gestational hypertensive disorders [8], preterm birth [9], and macrosomia [6]. In the long term, it also increases the risk of type 2 diabetes and obesity for both mothers and the offspring in the future [10,11].
Vitamin D de ciency may affect islet function and blood glucose control in patients with type 2 diabetes.
Little information is found about the effect of vitamin D on blood glucose of patients with GDM. Meabh Walsh suggested in the rst trimester of pregnancy, 25OHD was negatively associated with blood glucose, avoiding maternal vitamin D de ciency in early pregnancy is associated with lower blood glucose in early pregnancy and throughout pregnancy [6]. However, the relationship between vitamin D and blood glucose in patients with GDM is con icting. Naseh found there is no relationship between maternal serum vitamin D levels and maternal or neonatal serum glucose or insulin levels [6]. Maryam Akbari did not found any bene cial effect of vitamin D supplementation on fasting plasma glucose (FPG), insulin, HbA1c [7]. Moreover, the effect and mechanism of vitamin D on the islet function in gestational diabetes is not clear. Research [5] showed that the neutrophil lymphocyte ratio (NLR) was signi cantly associated with GDM occurrence in patients with GDM, NLR could suggest GDM development and in ammatory response. Many research suggested vitamin D could ameliorate the in ammation. The pathogenesis of GDM is mainly related to insulin resistance during pregnancy, such as relative reduction of insulin secretion, tissue sensitivity to insulin, and autoimmune-mediated in ammation of islet cells [8]. Base on the researches above, we hypothesis that vitamin D may affect fasting blood glucose or HbA1c in patients with GDM via in ammation response. This study aimed to explore the status of vitamin D in patients with GDM and whether it may affect fasting blood glucose and or HbA1c in patients with GDM via in ammation response.

Materials and Methods.
We identi ed 33 patients with GDM who were treated in the inpatient Department of Endocrinology and Metabolism in the Second Hospital a liated Xiamen medical college from January 2017 to March 2018.
According to the GDM diagnostic criteria of the American Diabetes Association (ADA) in 2011, the normal values of fasting, 1-hour, 2-hour venous blood glucose were, respectively, 5.1 mmo/L 10.0mmoL/L 8.5 mmol/L, any blood glucose value greater than the normal value can be diagnosed as GDM. According to Endocrine Society Clinical Practice Guidelines on vitamin D de ciency, serum circulating 25hydroxyvitamin D [25(OH)D] concentrations were measured to evaluate vitamin D status [9]. Five of them refused to test the concentration of 25(OH) D. A total of 33 subjects were nally enrolled. The following criteria were used to include subjects: 1. con rmed GDM, 2.serum iPTH at 15.0-65.0 pg/mL, 3.blood calcium less than 2.45 mmol/L, 4. normal liver function, serum creatinine, urea nitrogen, electrolytes, 5.without medication of insulin, insulin analogs, vitamin D and vitamin D analogs. The following criteria were used to exclude subjects:1.Women with a history of diabetes, 2.metabolic bone diseases,3.liver dysfunction or impaired kidney function 4.non-pregnant diabetes 5.patients with diabetic ketoacidosis, diabetic ketosis, diabetic hyperosmotic state, 6.blood phosphorus > 1.60 mmol/L, 7. acute infection, 8. tumor, 9. refused to participate in the group.

Anthropometric and Biochemical Analysis
(1) Data were gathered on age, sex. Participants' weights and heights were recorded. As all the patients were pregnant, bodymass index (BMI) was not calculated. The body weights and heights were measured while the subjects wore light clothing without shoes. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured while sitting, after a 5-minute rest, and again after a10 minute intervals, and the mean values were recorded. Hypertension was also diagnosed with a history of antihypertensive medication.
Serum calcium (calcium kits, Ningbo Medical System; Siemens ADVIA 2400 automatic biochemical analyzer) and serum phosphorus (phosphorus kits, Siemens China; Siemens ADVIA 2400 automatic biochemical analyzer) were measured by the ion selective electrode method.
HOMA-IR was calculated from fasting insulin and fasting glucose. HOMA-IR had the formula of fasting glucose (mmol/L) × fasting insulin (pmol/L)/22.5 [23]. eGFR was calculated by the MDRD GFR equation [24]. As this was a retrospective study, and the data were anonymously analyzed, informed consent was unnecessary. The formula for calculating the product of calcium and phosphorus [(mg/dL) 2 ] is blood calcium (mmol/L) multiplication and blood phosphorus (mmol/L) multiplication and 12.4. The neutrophil ratio is the neutrophil count divided by the lymphocyte count.
Statistics Processing SPSS19.0 software was used for statistical analysis. The continuous variables were expressed by mean ± standard deviation. Least-Signi cant Difference test (LSD-t) tests were used to compare the status of glucose metabolism and β cell function and NHL according to different concentrations of 25(OH) D. The statistical analysis was performed using SPSS19.0 software. The continuous variables were represented by mean ± standard deviation. LSD-t tests were performed to compare 25(OH) D,NHL and islet function in GDM patients with different vitamin D concentrations. Multiple linear regression analysis was used to examine the association between serum 25(OH)D concentration and HOMA-IR, HOMA-β, respectively analyzed as the dependent variable with the other signi cantly associated variables [25(OH)D, eGFR, BMI, and age] as independent variables, and value < 0.05was considered signi cant. This study was a retrospective study, informed consent was unnecessary.

Basic characteristics of the subjects
A total of 33 subjects were enrolled, with an average age of 31.52 ± 5.26 years old. The mean 25(OH) D concentration of 33 patients with gestational diabetes was 21.40 ± 10.68 ng/mL, all blood biochemical and glucose metabolism-related indices were within the normal range (showed in Table 1). The patients with vitamin D de ciency accounted for 30.30% (10 cases), the patients with vitamin D de ciency accounted for 60.61%(20 cases), the patients with normal vitamin D status accounted for 9.1%(3 cases), the average serum PTH was 43.01 ± 17.06 ng/mL, fasting blood glucose 5.79 ± 0.78 mmol/L, HbA1c 5.95 ± 0.93%.   Table 2; the mean HOMA-IR was signi cantly different between all the three groups, as shown in Table 2, and both HOMA-IR and FBG were lower with higher mean 25(OH)D concentrations. There was no signi cant difference of NLR between the three groups (P > 0.05) (showed in Table 2.).

Relationship between serum 25(OH)D concentration and glucose metabolism and NLR.
In the multiple linear regression analysis (shown in Table 3), 25(OH)D concentration is negatively correlated to FBG ( β = -0.587, p = 0.001), adjusted to eGFR, gestational age, and age. In the multiple linear regression analysis (shown in Table 4 and Fig. 1), 25(OH)D concentration is negatively correlated to HbA1c ( β = -0.408, p = 0.018 ), adjusted to eGFR, gestational age, and age. In the multiple linear regression analysis, vitamin D status was a predictor of HOMA-IR (β = -0.481, p = 0.007), as the dependent variable, but not eGFR, gestational age or age, as independent variables(shown in Table 5). In the 33 subjects, eGFR, gestational age and age, HOMA β is independent variable, 25(OH) D concentration as dependent variables, serum 25(OH)D concentration was objectively correlated to HOMA β ( β = 0.240, p = 0.206), however, the relationship is not signi cant (shown in Table 6.). 25(OH)D concentration and NRL were negatively related ( β = -0.389, p = 0.035), adjusted to eGFR, gestational age, and age(shown in Table 7. and Fig. 2.).

Discussion
Vitamin D is a fat-soluble vitamin, regulating calcium and phosphorus metabolism in vivo. The serum 25hydroxyvitamin concentrations are the best index to measure the nutritional status of vitamin D in vivo [9]. Vitamin D, regulating bone metabolism and calcium management in bone health processes, is widely con rmed, however, its extraskeletal effects such as the immune system, regulation of cell proliferation and differentiation, and glucose metabolism got more attention either [5].
The research abroad has reported that the serum 25-hydroxyvitamin D concentration in gestational diabetes patients is lower than that of normal pregnant women, and there is an maternal elevated risk of GDM patients with vitamin D de ciency at the rst trimester of pregnancy [11]. Interestingly the demand of vitamin D could be higher in pregnant women than in that without pregnant [12], which may aggravate the relative de ciency of vitamin D. In fact, 40-100% pregnant women in both the developing or developed countries, still suffer from vitamin D de ciency [11]. Some research showed vitamin D may be a potential candidate for the prevention of gestational diabetes mellitus (GDM), despite con icting current opinions [11].
The pathogenesis of gestational diabetes,referring to the fact of vitamin D de ciency, has attracted increasingly attention. The researches show the relationship between low vitamin D status and the risk of GDM. A meta-analysis of 12 studies with 5,615 patients found a moderate correlation between 25(OH)D concentrations below 50 nmol/L of pregnant women and an increased risk of GDM (OR 1.38, 95% CI: 1.12-1.70) [13]. Moreover, Zhang found similar conclusions in a meta-analysis of 20 studies conducted in Europe, Australia, North America, and Asia with different study design including cross-sectional, casecontrol, nested case-control, and cohort studies including 9209 participants [1]. Vitamin D supplementation for patients with GDM seems to ameliorate different metabolic markers including blood glucose levels, insulin resistance, and in ammatory biomarkers [14] [7]. Supplementation with 50,000 International Units (IU) twice monthly has been shown to improve insulin resistance signi cantly, while 5000 IU daily has failed to improve blood glucose in another trial [15]. Low vitamin D status is recognized to be correlated with fasting blood glucose in prediabetes patients [16]. However, as far as we know the information on the effect and mechanism of vitamin D on fasting glucose is limited.
In our study, 60.61% of the subjects showed vitamin D de ciency, accounting for more than half of the subjects, indicating that vitamin D de ciency was prevalent in patients with gestational diabetes. In the multiple linear regression analysis, 25(OH)D concentration is negatively correlated to FBG ( β = -0.587, p = 0.001), adjusted to eGFR, gestational age, and age. In the multiple linear regression analysis, 25(OH)D concentration is negatively correlated to HbA1c ( β = -0.408, p = 0.018 ), adjusted to eGFR, gestational age, and age. As 25(OH) D concentration decreased, fasting blood glucose and HbA1c increased, and 25(OH) D concentration was negatively correlated with fasting blood glucose and HbA1c. Vitamin D is probably a protective factor of glucose metabolism in Chinese women with gestational diabetes.
Vitamin D binds to the corresponding receptor of human islet β cells to promote insulin secretion and plays an important role in inhibiting the progression of diabetes [4]. The vitamin D status may affect blood glucose and islet function in patients with obesity [17]. Vitamin D can bind to the corresponding receptors of islet β cells or act directly on β cells to promote insulin transcription and synthesis. Consider that vitamin D de ciency may affect insulin secretion and cause insulin resistance, leading to increased fasting blood glucose and HbA1c. Vitamin D receptors (VDRs), expressed in different extra-bones peripheral tissues, were found the action on insulin receptor that promotes insulin sensitivity and insulin secretion [18] .
In the multiple linear regression analysis) of our study, vitamin D status was a predictor of HOMA-IR (β = -0.481, p = 0.007), as the dependent variable, but not eGFR, gestational age or age, as independent variable. In the 33 subjects, eGFR, gestational age, and age, HOMA β is independent variables, 25(OH) D concentration as dependent variable, serum 25(OH)D concentration was objectively correlated to HOMA β ( β = 0.235, p = 0.206), however, the relationship is not signi cant.
There is growing evidence that in ammation plays an important role in the pathogenesis of gestational diabetes. Vitamin D may also play an important role in the in ammatory reaction, autoimmune injury, insulin secretion, and insulin resistance of islet β cells and then affect the body's glucose metabolism [19]. In ammation in the form of immune cells in ltrating among glandular cells can result in functional pancreatic alterations [20]. Vitamin D that can recover the physiological insulin secretion by exerting antiin ammatory properties [21].The neutrophil to lymphocyte ratio (neutrophil-to-lymphocyte ratio, NLR) as a new in ammatory index, the value in the diagnosis and treatment of coronary heart disease, tumor, endometriosis [22]. And other diseases has attracted attention. NLR is a new in ammatory marker proposed in recent years, which is the ratio of neutrophils to lymphocytes in blood routine and can re ect the in ammatory state of the body. At present, a large number of studies have been carried out abroad on the ratio of neutrophils to lymphocytes, mainly aimed at the prognosis of tumor[23], coronary heart disease grade [22] and so on, it is found that the ratio of neutrophils to lymphocytes has a signi cant prognosis for the above diseases. Yilmaz, et al. [22]found a signi cant association with gestational diabetes. As multiple linear regression analysis can only prove the negative correlation, not further prove its causality.
Our study found that with the decrease of 25(OH) D concentrations, the NLR level increased, suggesting that with the decrease of vitamin D status may probably lead to the enhancement of in ammatory response, which may increase the fasting blood glucose level in patients with GDM. 25(OH)D concentration and NRL were negatively related ( β = -0.389, p = 0.035), adjusted to eGFR, gestational age, and age. It is interesting that the negative relationship is found between 25(OH) D concentration and HOMA-IR, but not between 25(OH) D concentration and HOMA-β, which may lead to a hypothesis of vitamin D affecting glucose metabolism via insulin sensitivity, not islet function. Further research is needed to give more evidence and proof. In all, vitamin D may affect blood glucose in patients with gestational diabetes by participating in the in ammatory response.

Conclusion
The vitamin D de ciency in patients with GDM was obvious, with a decrease of 25(OH) D level, fasting blood glucose and HbA1c increased. Vitamin D may affect fasting blood glucose and HbA1c in patients with gestational diabetes by participating in the regulation of NLR in ammatory response.