Association of serum vitamin D status with gestational diabetes mellitus and other laboratory parameters in early pregnant women

This was a prospective cohort study for a grant supported by the Foshan Science and Technology Bureau Project (2018AB000251) of Guangdong, China. During the study, a total of 241,651 pregnant women who visited our hospital for scheduled check-ups during the prenatal period at Foshan Women and Children’s Hospital of Southern Medical University, a specialized tier-3, grade-A hospital in South China, were examined. Of them, 12,536 women were giving birth, of which after excluding those with medical conditions such as diabetes and hypertension as described below, and also on voluntary enrollment basis, a total of 1,516 women, ages 18 to 46, were recruited from August 1, 2018 to August 31, 2019. The relevant information of those early pregnant women who met the criteria of the study was collected and serum 25 (OH) D and other laboratory testing were performed during early pregnancy. Pregnant women continued regular antenatal examination and follow-up until the termination of pregnancy. Women who had pregestational diabetes mellitus (PGDM), malignant diseases or missing OGTT data, or pregestational hypertension were excluded from the study. Exclusion criteria included pre-existing diabetes, hypertension or other chronic diseases, abnormal child-bearing history, a history of drug use, and vitamin supplements—including vitamin D and calcium—that may affect metabolism or unavailable data. Pregnant women with singleton pregnancies and conceiving naturally were included. All pregnant women exerted a normal daily life without special sunlight exposure or quarantines. All study subjects were living in the Foshan, Guangdong Province of South China. Ethics approval was obtained by the Institutional Review Board (IRB) and Ethics Committee of Foshan Women and Children’s Hospital of Southern Medical University. All methods were carried out in accordance with relevant guidelines and regulations. Methods and laboratory testing followed standard operating procedures required for medical practice in Guangdong Province, including in this hospital. All participants provided written informed consent.

Serum total 25(OH)D concentration was analyzed at gestational weeks 11 to 14 using a Mokosensor-A300 immune colloidal gold analyzer (MedicalSystems, Ningbo, China) following the manufacturer’s instructions. Briefly, 2 ml of blood was sampled from pregnant women at weeks 11–14 of gestation using an EDTA anti-coagulant blood tube. The blood sample was delivered to clinical laboratory immediately after sampling. Twenty microliters of whole blood were used for the analysis. The accuracy of the assay was from 85–115%, while the precision of the assay was ≤ 15% (coefficient of variation, CV). According to the Vitamin D Standardization-Certification Program (VDSCP) guidelines, a CV of < 15% is close to acceptable in the analysis [20‐22].

The statuses of maternal serum 25(OH)D were categorized into three groups: 25(OH)D deficiency (< 50 nmol/L), 25(OH)D insufficiency (50–74 nmol/L), and 25(OH)D sufficiency (≥ 75 nmol/L).

Other blood chemistry parameters also came from the samples collected during weeks 11 to 14 of gestation. Specifically, alkaline phosphatase, pre-albumin, creatinine, cystatin C, calcium, and magnesium were analyzed on a Beckman-Coulter Automatic Chemistry Analyzer platform (Beckman-Coulter, Brea, CA, USA) following the manufacturer’s instructions and laboratory operating procedures. The collection of blood samples, including the selection of a collecting tube, followed routine clinical laboratory procedures.

A 75 g oral glucose tolerance test (75 g OGTT) was performed between 24 and 28 weeks of gestation. Pregnant women who had a fasting glucose level < 5.1 before pregnancy were enrolled in the study. The diagnosis of GDM was defined according to the American Diabetes Association [23] using the following protocol: A 75 g oral glucose tolerance test (75 g OGTT) was performed between weeks 24 and 28 of gestation. Either of the following criteria met was diagnosed as GDM: fasting serum glucose ≥ 5.1 mmol/L, or 1 h of OGTT serum glucose ≥ 10 mmol/L, or 2 h of OGTT serum glucose ≥ 8.5 mmol/L).

The blood coagulation test was performed on a Stago STA-R Evolution (Diagnostica Stago, Asnières sur Seine Cedex, France) platform following the manufacturer’s instructions using a heparin anti-coagulant blood collection tube collected during weeks 11 to 14 of gestation. The coagulation parameters include thrombin time (TT), activated partial thromboplastin time (aPTT), fibrinogen (Fib), and D-dimer.

Bone density was measured using a Hong Yang BMD-1000 Ultrasound Bone Sonometer (Baoding, Hebei, China) by a specialist in our hospital following the protocol provided by the manufacturer. Pregnant women at weeks 11–14 of gestation were assigned for bone density measurement of the distal radius for 2 min. In brief, the Z value is obtained by comparing the sonographic value of the bone of a patient with a reference (same age, same sex). Thus, bone density was expressed as Z-scores.

Statistical analysis was performed using the SPSS (version 24). Continuous variables are reported as mean ± standard deviation or median (interquartile), while categorical variables are presented as frequencies and percentages. For continuous variables, hypothesis testing for significant differences was performed using One-Way ANOVA and Nonparametric Tests for normal and non-normal distribution data, respectively. Pearson’s chi-square was used for categorical variables. Post Hoc Multiple Comparison was used to further identify significant differences among groups. Possible confounding factors with p < 0.20 were entered into multiple linear regression analysis to explore independent influencing factors of 25(OH)D. Spearman coefficient was used to determine the correlation between 25(OH)D status and pregnant outcomes. Unadjusted and adjusted logistic regression analysis were used to calculate the crude or adjusted odds ratios (OR), and their 95% confidence intervals (CI) were used to evaluate the relationship between 25(OH)D and GDM and other parameters (gestational adverse events). Variables with p < 0.10 in unadjusted analysis and possible confounding factors were entered into multivariable logistic regression procedure. The model’s adequacy was assessed by predicting correct percentage. A p < 0.05 were considered statistically significant.

In this observation, the average 25(OH)D concentration was 59.1 ± 12.6 nmol/L; none of the women had 25(OH)D < 25 nmol/L. Four hundred and thirty-four (434, 28.6%) women had 25(OH)D < 50 nmol/L, 882 women (58.2%) had 25(OH)D insufficiency (50–74 nmol/L), and 200 (13.2%) women had 25(OH)D ≥ 75 nmol/L. Pre-adjusted analysis using the Chi-square test, One-way ANOVA, and nonparametric test results that the maternal and clinical characteristics are listed in Table 1 on the basis of 25(OH)D status. Table 1 also shows that serum 25(OH)D varies significantly in regard to season at sampling. Serum alkaline phosphatase (ALP), pre-albumin, creatinine, and thrombin time (TT) also differed upon different serum 25(OH)D status (p < 0.05 for all). However, the categories of pre-albumin and creatinine, as well as the rest of variables in the list, do not show difference based on 25(OH)D status.

Tables 2 and 3 show the association between maternal serum 25(OH)D concentrations and parity, season at sampling, serum alkaline phosphatase. Women going into birth ≥ 2 times had a higher 25(OH)D concentration than those with 1 or 0 paritys (p = 0.018 and p = 0.044 respectively). The same result was found in adjusted linear analysis (p = 0.010 and p = 0.032, respectively). There was a significant difference in 25(OH)D concentrations among samples collected in different seasons; the 25(OH)D concentrations assended in the order of spring, winter, summer, and autumn. There was a significant difference between serum 25(OH)D concentration and alkaline phosphatase activities, of which a higher serum 25(OH)D concentration tended to have higher alkaline phosphatase activities (p = 0.005 and 0.004, respectively).

As shown in Table 4, results from spearman correlation analysis indicate that there is no correlation between vitamin D status and adverse outcomes in early gestating women (p > 0.05 for all). Interestingly, the rate of 1 h GTT positive (r = -0.037) and 2 h GTT positive (r = -0.037) was inversely correlated with the increament of serum 25(OH)D status, but with no significance (p = 0.145 and 0.534, respectively).

Logistic regression analysis was performed to assess the association between serum 25(OH)D and gestational diabetes mellitus. In this study, we found that there was no association between 25(OH)D and the risk of GDM, as shown in Table 5. In unadjusted logistic regression analysis, no significant difference was observed between maternal 25(OH)D and GDM (p = 0.627). After adjusted by age, parity, and season at sampling, women with serum 25(OH)D ≥ 75 nmol/L (vitamin D sufficient) were grouped as the reference group, while women whose serum 25(OH)D between 50–74 nmol/L (vitamin D insufficient), or < 50 nmol/L (vitamin D defficient) groups were compared with the reference group. The results indicated that there was no association between serum 25(OH)D and GDM, with an odds ratio of 1.250 (95%CI 0.763–2.046, P = 0.376) and 1.224 (95%CI 0.794–1.888, P = 0.360), respectively. GDM women were more likely to be older than those without GDM.

Since serum 25(OH)D concentrations is affected by several factors, such as seasons, geographical location, race, diet/lifestyle, medical conditions, etc., it is important to develop a more rigourous protocol when performing future vitamin D related studies. In this study, patients enrollment, diet and sun exposure, maternal nutrition supplement, etc. could be more accurately controlled. The analyzer used in 25(OH)D assay could be improved if the budget allowed, such as an LC–MS/MS spectrometer. Otherwise, more accurate methods and standardization of the assay should be performed in a future study and in clinical testing. For nstance, we would re-measure 100–150 samples based on a VDSCP assay along with standard reference materials (SRM) and applying Deming regression to standardize their values in the full sample. Socioeconomic status, lifestyle, the eating habits of pregnant woman, the monitoring and classification of gestational weight gain of the pregnancy women, and their exercise levels are factors of limitation in this study and should be considered in the future study. The study population mainly composed of Han Chinese women, which ensured data homogeneity, but it may under-represent population groups with different lifestyles and socioecomic situations, leading to biased results.