Association of Vitamin D and Parathyroid Hormone with Insulin Sensitivity, Beta Cell Function and Gestational Diabetes in Pregnancy: A Cross-Sectional, Observational Study.

INTRODUCTION: Vitamin D (25OHD) and parathyroid hormone (PTH) are associated with dysglycemia, and we investigated them in gestational diabetes mellitus (GDM).
METHODS: In this cross-sectional, observational study, we included 75 pregnant women between 24 and 28 weeks of gestation. A fasting venous sample was collected for plasma glucose (FPG), insulin, PTH and 25OHD. Glucose and insulin samples were collected hourly after 75 g glucose load for 2 h. Insulin sensitivity was estimated by the Matsuda index (MI) and beta cell function by the insulin secretion sensitivity index (ISSI-2). The subjects were stratified into three groups and tertiles according to the 25OHD and PTH, respectively. Appropriate statistical tests were used to compare the MI, ISSI-2 and GDM among the groups.
RESULTS: GDM was seen in 14/75, and of these patients, 2 were 25OHD deficient, 7 insufficient and 5 had sufficient 25OHD. MI and ISSI-2, though not correlated with the 25OHD, decreased from the lower to higher PTH tertile (P < 0.001). FPG, AUCgluc (area under the curve glucose) and prevalence of GDM increased from the lower to higher PTH tertile (P < 0.001).
CONCLUSION: Increased PTH was associated with decreased insulin sensitivity, beta cell function and GDM in pregnancy, irrespective of the underlying 25OHD level.

Introduction

Pregnancy is a diabetogenic state due to changes in the insulin sensitivity and beta cell function in previously healthy women. The Hyperglycemia and Adverse Pregnancy Outcome (HAPO) Study showed a continuous association between the maternal glucose level and increased birth weight, contributed by fetal hyperinsulinemia [1]. Vitamin D deficiency (VDD) has been described as a risk factor for type 2 diabetes (T2D) and gestational diabetes mellitus (GDM) [2]. Vitamin D has an effect on beta cell function, insulin sensitivity and inflammatory markers both directly and indirectly [3]. The pathways include stimulating the expression of the insulin receptors, activating the PPAR-δ (peroxisome proliferator activator receptor) pathway, improving oxidative phosphorylation in the muscle, reducing proinflammatory cytokines and increasing the intracellular calcium level [3]. Vitamin D plays an important role in feto-maternal medicine, and the vitamin D requirement increases during pregnancy [4]. A limited number of studies have investigated the association between GDM and VDD in the context of insulin resistance and beta cell function [5, 6].

Vitamin D and parathyroid hormone (PTH) share an inverse relationship physiologically. PTH has an independent effect on dysglycemia and is a modifiable risk factor for the development of GDM in patients with VDD [7]. PTH facilitates glucose-induced insulin release from beta cells by increasing the cytosolic calcium and stimulating the production of cAMP (cyclic Adenosine Mono Phosphate) [8]. However, PTH also increases insulin resistance by downregulating the insulin receptors peripherally and has a direct effect on beta cells [9]. The net effect of PTH on glycemic metabolism is controversial and has not been defined clearly. The majority of the studies pertaining to VDD and GDM have not factored in the independent role of PTH on the glycemic level [10, 11]. Therefore, we conducted this study to evaluate the association of vitamin D and PTH levels on insulin sensitivity and beta cell function in patients with GDM.

Methods

Study Population

This cross-sectional, observational study was conducted at a tertiary level Armed Forces hospital between December 2016 and June 2017. Pregnant women (24–28 weeks of pregnancy) who reported for GDM screening were included in the study. All the study participants reside in India and their genetic background is the South Asian race. We excluded patients with twin pregnancy, pre-existing DM and past history of GDM. We excluded patients with any major medical ailment, glucocorticoid use of any dose in the preceding 3 months and past history of hyperparathyroidism. The local ethics committee approved the trial protocol, and informed consent was obtained from all patients for being included in the study. All procedures followed in the study were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964, as revised in 2013.

Study Measures

A detailed history regarding the medical, obstetric and family profile of diabetes was obtained from all the participants. A general physical examination was carried out with specific reference to the anthropometric measures and signs of insulin resistance. Body mass index (BMI) is defined as weight in kilograms divided by height in meters squared. A fasting venous blood sample was collected from each participant after an overnight fast of 10 h. This sample was analyzed for glucose, 25-hydroxy-vitamin D (25OHD), PTH, insulin and glycosylated hemoglobin (HbA1c). Thereafter, patients were given 75 g glucose solution orally, and venous samples were collected after 1 and 2 h for estimation of glucose and insulin. Glucose was estimated by the glucose oxidase method within 1 h of the sample collection and HbA1c by the cation exchange high-performance liquid chromatography method using the D10 hemoglobin testing system manufactured by Bio-Rad Laboratories. The levels of insulin, 25OHD and PTH were estimated with the IRMA kit (Beckman Coulter) using the Stratec SR300 automated radioimmunoassay system (STRATEC Biomedical Systems AG, Birkenfeld, Germany). The inter- and intraassay coefficient of variation for all the tests is < 8% in our laboratory.

Study Definitions

GDM was diagnosed according to the American Diabetes Association (ADA) criteria when any one of the following criteria was met after a 75 g OGTT (fasting > 92 mg/dl, 1 h > 180 and 2 h > 153 mg/dl) [12]. Insulin sensitivity was calculated using the Matsuda index [13]. We calculated the area under the curve (AUC) for insulin (pmol/l) and glucose (mmol/l) using the OGTT (oral glucose tolerance test) derived values and the trapezoidal rule. Beta cell function was estimated using the Insulin Secretion-Sensitivity Index-2 (ISSI-2), which is a validated measure derived from the OGTT [14]. It is a product of insulin secretion measured by the ratio of the AUC insulin curve to AUC glucose and insulin sensitivity. We stratified the patients, according to the levels of vitamin D and PTH. The patients were divided into three groups (D1, D2 and D3) based on 25OHD according to the Endocrine Society guidelines into deficiency (≤ 20 ng/ml), insufficiency (21–30) and sufficiency (25OHD > 30 ng/ml), respectively [15]. The patients were divided into three groups based on the PTH level as P1 (PTH ≤ 15 pg/ml), P2 (PTH 15.1–30) and P3 (PTH > 30 pg/ml).

Statistics

Continuous variables are presented as mean (SD), and categorical variables are presented as absolute numbers and percentage. Data were checked for normality before statistical analysis. One-way analysis of variance (ANOVA) was used to evaluate the significance of the differences among the three PTH and 25OHD groups. Multivariate linear regression analyses were done to evaluate independent associations of vitamin D and PTH groups with insulin sensitivity and beta cell function. P < 0.05 was taken as significant in all the tests. The statistical analysis was performed using the SPSS program for Windows, version 17.0 (SPSS, Chicago, IL, USA).

Results

The study participants comprised 75 pregnant women with a mean age of 27.5 (2.9) years, mean gestational age 26.7 (0.7) weeks, body weight 57.8 (6.5) kg and mean BMI 23.2 (1.9) kg/m2. Fourteen of 75 (18.7%) patients were diagnosed with GDM, and family history of DM was present in 24/75 (32%) women. The mean 25OHD level was 25.7 (7.6) ng/ml, PTH 26.7 (15) pg/ml, Matsuda index 4.4 (1.7) and mean ISSI-2 244.9 (68.1) in the entire study population. Patient characteristics based on their 25OHD level are shown in Table 1. 25OHD was deficient in 19 (25%), insufficient in 39 (52%) and sufficient in 17 (23%) participants. The prevalence of GDM in each 25OHD group was not significantly different (P = 0.0661), though the percentage of women with GDM increased with improving 25OHD level. Insulin sensitivity and beta cell function showed no difference according to the underlying 25OHD level. The FPG and HbA1c of women across all three groups of 25OHD were also similar.

Table 1

Comparison of parameters according to the 25OHD level

Features	D1 (25OHD ≤ 20)	D2 (25OHD 21–30)	D3 (25OHD > 30)	P value
Number (%)	19 (25.3%)	39 (52%)	17 (22.7%)
Age (years)	27.2 (3.3)*	27.5 (3.1)	27.6 (2.5)	0.9031
Family h/o DM (n)	5 (26.3%)	14 (35.5%)	5 (29.4%)	0.7382
Gestational age (weeks)	26.8 (0.96)	26.7 (0.96)	26.9 (1.1)	0.6943
Height (cm)	156.2 (4.6)	156.9 (5.8)	156.9 (5.5)	0.8794
Weight (kg)	56.6 (6.4)	57.7 (7.1)	57.2 (7.2)	0.8695
BMI (kg/m2)	23.2 (2.1)	23.4 (1.9)	23.2 (1.8)	0.9286
Weight gain (kg)	5.3 (0.67)	5.5 (0.86)	5.4 (0.75)	0.8627
Summer/winter	17/2	29/10	15/2	0.2718
Ca/Vit D use	2 (10.5%)	10 (26%)	6 (35.3%)	0.2089
25OHD (ng/ml)	12.2 (2.7)	25 (2.6)	36.7 (5.2)	< 0.001
PTH (pg/ml)	37.6 (12.2)	25.4 (13.3)	17.4 (14.5)	< 0.001
HbA1c (%)	5.2 (0.73)	5.2 (0.78)	5.3 (0.94)	0.9081
Matsuda index	4.1 (1.4)	4.5 (1.7)	4.4 (2.1)	0.7122
ISSI-2	244.7 (77.4)	242.9 (60.9)	250.1 (76.5)	0.9383
FPG (mmol/l)	4.6 (0.6)	4.7 (0.7)	4.9 (0.9)	0.8843
Glucose 1 h post OGTT	9 (1.1)	8.9 (1.2)	9 (1.5)	0.4432
Glucose 2 h post OGTT	7.6 (1.1)	7.3 (1.3)	7.2 (1.6)	0.6394
AUCgluc (mmol/l*1 h)	15.2 (1.9)	14.9 (2.1)	15.1 (2.7)	0.9375
GDM cases (%)	2 (10.5%)	7 (17.9%)	5 (29.4%)	0.0661

All significant P values are highlighted in bold

*Mean (SD)

The comparison of the patients based on the PTH groups is shown in Table 2. Twenty-three women were in the first group (≤ 15 pg/ml), 29 in the second and 23 in the third group (> 30 pg/ml). The patients in different PTH groups did not differ in age, BMI, gestational age or the season of blood collection, as shown in Table 2. GDM was more prevalent in the third group (higher PTH) compared with the others. Insulin sensitivity and beta cell function also decreased significantly from the first (low) to third (high) PTH group. The analysis of patients with and without GDM is shown in Table 3. Briefly, patients with GDM had higher weight gain and PTH levels and were insulin resistant with reduced beta cell function.

Table 2

Comparison of parameters according to the PTH value

Feature	P1 (PTH ≤ 15)	P2 (PTH 15.1–30)	P3 (PTH > 30)	P value
Number	23 (31%)	29 (38%)	23 (31%)
Age (years)	27.7 (2.8)*	27.2 (2.8)	27.6 (3.4)	0.7872
Family h/o DM (n)	5 (21.7%)	8 (27.6%)	11 (47.8%)	0.1342
Gestation at OGTT (weeks)	26.7 (1)	26.8 (1.01)	26.7 (0.96)	0.9704
BMI (kg/m2)	23.4 (1.6)	22.7 (2)	23.8 (1.9)	0.1262
Weight gain (kg)	5.3 (0.7)	5.3 (0.66)	5.7 (0.95)	0.1414
Summer/winter (n)	4/19	7/22	3/20	0.5827
Ca/Vit D use (n)	8 (34.8%)	6 (20.7%)	4 (17.4%)	0.3345
25OHD (ng/ml)	32.6 (7.4)	23.5 (3.9)	21.6 (6.7)	< 0.0001
PTH (pg/ml)	11.6 (1.8)	23.3 (4)	46.2 (8.9)	< 0.0001
HbA1c (%)	5 (0.78)	4.9 (0.54)	5.7 (0.9)	0.0026
Matsuda index	5.3 (1.4)	4.4 (1.2)	2.9 (1.4)	< 0.0001
ISSI-2	271.7 (59.7)	248.5 (48.3)	213.7 (85.3)	0.0124
FPG (mmol/l)	4.3 (0.4)	4.8 (0.7)	5.3 (0.3)	< 0.0001
Glucose 1 h post OGTT	8.4 (1)	8.6 (0.9)	9.6 (1.4)	0.0017
Glucose 2 h post OGTT	6.7 (1.1)	7.2 (1)	8.1 (1.5)	0.0016
AUCgluc (mmol/l*1 h)	14.1 (1.8)	14.7 (1.4)	16.4 (2.8)	0.0008
GDM cases (n)	3 (13%)	2 (6.9%)	9 (39.1%)	0.0088

All significant P values are highlighted in bold

*Mean (SD)

Table 3

Comparison between the patients with and without GDM

Feature	GDM absent	GDM present	P value
Number	61 (81%)	14 (19%)
Age (years)	27.4 (2.9)*	27.7 (3.3)	0.7350
Family h/o DM (n)	16 (26.2%)	8 (57.1%)	0.0528
Gestation at OGTT (weeks)	26.8 (0.9)	26.8 (1.1)	0.9146
BMI (kg/m2)	23.1 (1.9)	24.1 (1.9)	0.0886
Weight gain (kg)	5.2 (0.6)	6.4 (0.8)	< 0.0001
Summer/winter (n)	51/10	10/4	0.2818
Ca/Vit D use (n, %)	13 (21.3%)	5 (35.7%)	0.3028
PTH (pg/ml)	24.4 (13.4)	36.8 (17.9)	0.0043
25OHD (ng/ml)	25.3 (7.7)	27.3 (6.8)	0.3897
HbA1c (%)	4.9 (0.3)	6.7 (0.5)	< 0.0001
Matsuda index	5 (1.1)	1.6 (0.44)	< 0.0001
ISSI-2	267.5 (53.6)	147 (21.3)	< 0.0001
FPG (mmol/l)	4.5 (0.3)	6.2 (0.8)	< 0.0001
Glucose 1 h post OGTT	8.5 (0.6)	11.1 (0.8)	< 0.0001
Glucose 2 h post OGTT	6.9 (0.7)	9.4 (0.5)	< 0.0001
AUCgluc (mmol/l*1 h)	14.5 (1.3)	16.8 (2.2)	< 0.0001

All significant P values are highlighted in bold

*Mean (SD)

A simple univariate correlation analysis between the important clinical and biochemical parameters is shown in Table 4. Correlation analyses revealed that 25OHD is inversely related to the PTH in the entire study population (R = − 0.5162; P < 0.001). PTH correlated significantly with the insulin sensitivity (R = − 0.4889; P < 0.001) and beta cell function (R = − 0.2985; P = 0.0093), whereas 25OHD showed an insignificant correlation with the insulin sensitivity (R = 0.0934; P = 0.4253) and beta cell function (R = 0.1001; P = 0.3926), as shown in Fig. 1. The results of the multivariate analysis for the risk of GDM with the study parameters is shown in Table 5. Increased BMI and VDD and elevated PTH raise the risk of GDM independent of other parameters.

Table 4

Simple univariate correlation matrix

Parameter	Age	BMI	25OHD	PTH	FPG	HbA1c
Age	–	R = 0.0865 P = 0.4598	R = 0.1149 P = 0.3262	R = − 0.0604 P = 0.6065	R = − 0.1055 P = 0.3678	R = − 0.0834 P = 0.4767
BMI	R = 0.0865 P = 0.4598	–	R = 0.1455 P = 0.2130	R = 0.05916 P = 0.6141	R = 0.2096 P = 0.0711	R = 0.2952 P  = 0.0101
25OHD	R = 0.1149 P = 0.3262	R = 0.1455 P = 0.2130	–	R = − 0.5162 P  = < 0.0001	R = − 0.0307 P = 0.7937	R = − 0.0145 P = 0.9012
PTH	R = − 0.0604 P = 0.6065	R = 0.05916 P = 0.6141	R = − 0.5162 P  = < 0.0001	–	R = 0.4260 P  = 0.0001	R = 0.3694 P  = 0.0011
FPG	R = − 0.1055 P = 0.3678	R = 0.2096 P = 0.0711	R = − 0.0307 P = 0.7937	R = 0.4260 P  = 0.0001	–	R = 0.8686 P  = < 0.0001
HbA1c	R = − 0.0834 P = 0.4767	R = 0.2952 P  = 0.0101	R = − 0.0145 P = 0.9012	R = 0.3694 P  = 0.0011	R = 0.8686 P  = < 0.0001	–

All significant P values are highlighted in bold

Fig. 1

Correlation analyses among vitamin D, PTH with the Matsuda index and ISSI-2

Table 5

Logistic multivariate regression analysis for the risk of developing GDM

Parameter	Coefficient	Standard error	Odds ratio	95% CI	P value
Age	– 0.053942	0.13755	0.9475	0.7236–1.2407	0.6949
BMI	0.51948	0.23162	1.6812	1.0677–2.6471	0.0249
Family h/o DM	0.88737	0.76234	2.4287	0.5451–10.8217	0.2444
Gestational age	– 0.092310	0.36886	0.9118	0.4425–1.8788	0.8024
25OHD	0.13526	0.059754	1.1448	1.0183–1.2871	0.0236
PTH	0.090982	0.033457	1.0952	1.0257–1.1695	0.0065

All significant P values are highlighted in bold

= 0.0101

= < 0.0001

= < 0.0001

= 0.0001

= 0.0011

= 0.0001

= < 0.0001

= 0.0101

= 0.0011

= < 0.0001

Discussion

Our study showed that insulin sensitivity and beta cell function were related to the PTH level independent of the underlying 25OHD. The patients in the highest PTH group had a higher prevalence of GDM compared with the other two groups. Although previous studies have suggested a role of PTH in glucose homeostasis in the general population, few studies have addressed this issue in pregnancy. Whitelaw et al. showed that PTH was weakly associated with FPG but not with 2-h glucose or GDM [16]. In contrast, in a cross-sectional assessment in the late second trimester, Zuhur et al. found that PTH was higher in women with GDM than in their peers on unadjusted comparison [17]. Our study demonstrated a negative correlation of PTH with the Matsuda index and ISSI-2, as shown in Fig. 1. The inverse relation between the PTH level and ISSI-2 suggests beta cell dysfunction as a potential pathophysiologic mechanism through which increased PTH may be linked to dysglycemia in pregnancy. PTH may be an unrecognized biologic factor contributing to the insulin resistance in patients with VDD. PTH is a key indicator of 25OHD status, and interindividual variations exist in the level of 25OHD that provides maximal PTH suppression [18].

Our data showed the presence of VDD in a majority of the study participants. Previous studies have shown that VDD is a common problem in Indian women [19]. The literature is conflicting about the role of 25OHD in GDM. Our study showed an equal prevalence of VDD in patients with and without GDM. Similar to our study, Farrant et al. found that hypovitaminosis D at 30-week gestation is not associated with an increased risk of GDM [11]. In a cross-sectional study involving more than 1400 patients from the developed world, Whitelaw et al. demonstrated that 25OHD was not associated with theinsulin concentration or GDM [16]. Another study from India showed a strong association between the levels of 25OHD and GDM, though the 25OHD replacement did not reverse the gestational dysglycemia [20]. Zuhur et al. showed that GDM is associated with only severe VDD and not with mild or moderate deficiency [17]. The conflicting links between vitamin D and GDM could be explained by the differences in the study population, sun exposure, food fortification, VDD definition and measurement of 25OHD [21].

The best available index of insulin sensitivity during pregnancy is the NEFA (non-esterified fatty acid) index based on the insulin level and BMI [22]. This is cumbersome and is not widely available for performing in clinical practice. The Matsuda index is considered very useful for assessment of insulin secretion as it shows a stronger correlation with clamp-measured insulin sensitivity than other surrogate measures in pregnancy. The QUICKI and HOMA indices, which are based on a single blood sample, can also provide an easy but accurate measure of insulin sensitivity in pregnant women [23]. We used the Matsuda index for assessment of insulin sensitivity and ISSI-2 for the assessment of beta cell function. Lacroix et al. used HOMA-IR, beta cell function (HOMA-B), the Matsuda index and beta cell compensation (ISSI-2) [6]. Similar to our study, Kramer et al. used the Matsuda index and ISSI-2 in their study involving pregnant women [24].

Our study has many strengths, as explained below. First, we assessed the insulin sensitivity using the Matsuda index, which correlated strongly with insulin clamp-measured insulin sensitivity compared with other surrogate measures in pregnancy. Second, the beta cell function was assessed by ISSI-2, which is a validated OGTT-derived measure of beta cell function and has been directly validated against the disposition index from intravenous GTT. Third, we estimated the 25OHD and PTH simultaneously in relation to glycemia, unlike many previous studies. The limitations of our study include the small sample size, cross-sectional design and lack of estimation of active vitamin D and binding protein, which could provide more comprehensive information about this metabolic pathway. Pregnancy affects the assessment of total vitamin D, and few authors have suggested using the free form of vitamin D for accurate diagnosis of VDD [25]. The PTH assay methods using the immunometric technique measure the oxidized PTH, which is a biologically inactive molecule, thereby falsely elevating the PTH level [26, 27]. The measurement of total vitamin D instead of free vitamin D and the estimation of PTH by RIA assay are other methodologic limitations of our study.

Conclusion

To conclude, our study demonstrates that increased PTH was associated with decreased insulin sensitivity, beta cell dysfunction and dysglycemia in pregnancy. In contrast, the vitamin D level in pregnancy is not associated with GDM, insulin sensitivity and beta cell function. The PTH level may emerge as an underlying independent risk factor relevant to the interpretation of the links between 25OHD and GDM. Further prospective studies with large sample sizes are required to validate the findings observed in our study.