Maternal and Newborn Vitamin D–Binding Protein, Vitamin D Levels, Vitamin D Receptor Genotype, and Childhood Type 1 Diabetes

Epidemiology/Health Services Research

¹Department of Chronic Diseases and Ageing, Norwegian Institute of Public Health, Oslo, Norway
²Hormone Laboratory, Department of Medical Biochemistry, Oslo University Hospital, Aker, Oslo, Norway
³Barbara Davis Center for Diabetes, University of Colorado, Anschutz Medical Campus, Aurora, CO
⁴Division of Paediatric and Adolescent Medicine, Oslo University Hospital, Oslo, Norway
⁵Department of Immunology, Rikshospitalet, Oslo University Hospital, Oslo, Norway
⁶Department of Medical Genetics, University of Oslo, Oslo, Norway
⁷Department of Pediatric and Adolescent Medicine, Haukeland University Hospital, Bergen, Norway
⁸KG Jebsen Center for Diabetes Research, Department of Clinical Science, University of Bergen, Bergen, Norway
⁹Institute of Clinical Medicine, University of Oslo, Oslo, Norway
¹⁰Department of Congenital Disorders, Statens Serum Institut, Copenhagen, Denmark
¹¹Department of Pediatrics, Østfold Hospital Trust, Grålum, Norway

Diabetes Care 2019 Apr; 42(4): 553-559. https://doi.org/10.2337/dc18-2176

Article Figures & Tables

Download figure
Open in new tab
Download powerpoint
Figure 1
Formation of the analysis sample. *148 had three blood samples, 38 had two blood samples, and 3 had one blood sample available for 25(OH)D and DBP testing. There were 174 midpregnancy, 174 postpartum, and 175 cord blood samples. †456 had three blood samples, 111 had two blood samples, and 9 had one blood sample available for 25(OH)D and DBP testing. There were 532 midpregnancy, 525 postpartum, and 542 cord blood samples.
Download figure
Open in new tab
Download powerpoint
Figure 2
Distribution of vitamin DBP concentrations in maternal and cord blood plasma samples from randomly selected control subjects (n = 576) in MoBa. The maternal delivery (postpartum) sample was collected at median 1 day (IQR 0–3) after delivery.

Table 1

Characteristics of case subjects with childhood T1D and randomly selected control subjects in MoBa

	Control subjects (n = 576)	Case subjects (n = 189)
Age at end of follow-up (years)^*	11.7 (7.7–17.1)	12.7 (8.0–16.6)^†
Female sex	285 (49.5)	93 (49.2)
Maternal T1D	0 (0.0)	7 (3.7)
Preterm birth	19 (3.3)	10 (5.3)
Missing data	1 (0.2)	1 (0.5)
Birth weight (g)
<2,500	8 (1.4)	7 (3.7)
2,500–4,500	539 (93.6)	174 (92.1)
>4,500	29 (5.0)	8 (4.2)
Parity
No previous births	248 (43.1)	93 (49.2)
One birth	215 (37.3)	57 (30.2)
Two or more births	113 (19.6)	39 (20.6)
Maternal age (years)	30 (17–42)	30 (19–42)
Maternal non-Norwegian ethnicity	30 (5.2)	10 (5.3)
Maternal smoking during pregnancy
Nonsmoker at end of pregnancy^‡	469 (81.4)	161 (85.2)
Smoked at end of pregnancy	76 (13.2)	20 (10.6)
Missing data	31 (5.4)	8 (4.2)
Maternal prepregnancy BMI (kg/m²)
<25	379 (65.8)	97 (51.3)
25–30	109 (18.9)	49 (25.9)
>30	40 (6.9)	28 (14.8)
Missing data	48 (8.3)	15 (7.9)
Child’s HLA^§ genotype
Protective (DQ6)	168 (29.2)	3 (1.6)
Neutral (any other HLA not mentioned)	111 (19.3)	5 (2.6)
Increased risk (≥1 copy of either DQ8 or DQ2)	212 (36.8)	93 (49.2)
High risk (DQ8/DQ2 heterozygote)	30 (5.2)	71 (37.6)
Missing data	55 (9.5)	17 (9.0)
Child’s non-HLA T1D GRS^‖	61.2 (45.7–76.3)	63.2 (45.4–78.6)
Missing data	19 (3.5)	12 (6.8)
Child’s 25(OH)D GRS	3 (0–8)	3 (0–8)
Missing data	20 (3.7)	13 (7.4)
Cesarean section^¶	59 (10.2)	36 (19.0)

Data are median (range) or n (%).
↵*Diagnosis date of the last case subject included: 3 February 2014.
↵†Median age at diagnosis of T1D case subjects was 5.7 years (range 0.7–12.7).
↵‡Including those who quit smoking shortly before or during pregnancy.
↵§Groups defined as protective (carrying at least one copy of HLA DQA1*01:02-DQB1*06:02-DRB1*15:01 [DQ6-DR15]), increased risk (at least one copy of HLA DQA1*03-DQB1*03:02-DRB1*04 [DQ8-DR4] or DQA1*05:01-DQB1*02:01-DRB1*03:01 [DQ2-DR3] but not both haplotypes), high risk (HLA DQ2-DR3/DQ8-DR4), or neutral (any other genotype).
↵‖Weighted score, calculated by multiplying the number of risk alleles in 51 non-HLA SNPs with their reported risk per allele.
↵¶Includes unknown (n = 1), emergency (n = 55), and elective (n = 39) cesarean section.

Table 2

Association between exposures and T1D

	OR (95% CI)	aOR (95% CI)^*	P
DBP, per 1 μmol/L increase
Predicted maternal DBP^†	0.74 (0.39–1.22)	0.49 (0.18–1.02)	NS^‡
Midpregnancy	1.03 (0.91–1.16)	0.96 (0.79–1.16)	0.65
Cord blood	0.98 (0.81–1.20)	0.87 (0.67–1.14)	0.32
Postpartum	0.86 (0.74–0.98)	0.80 (0.67–0.95)	0.01
25(OH)D-to-DBP ratio
Midpregnancy	1.00 (0.96–1.04)	1.04 (0.98–1.09)	0.17
Cord blood	0.99 (0.97–1.01)	1.00 (0.97–1.04)	0.81
Postpartum	1.01 (0.98–1.04)	1.05 (1.00–1.10)	0.049
25(OH)D, per 10 nmol/L increase, stratified by VDR rs11568820^§
Cord blood, AA/AG	1.18 (1.00–1.40)	1.17 (0.95–1.44)	0.15
Cord blood, GG	0.87 (0.77–0.98)	0.85 (0.72–1.00)	0.047

P value shown for adjusted analysis. aOR, adjusted OR.
↵*Adjusted for child’s HLA genotype and sex and maternal ethnicity, age, prepregnancy BMI, cesarean section, and smoking.
↵†Using maternal (midpregnancy and postpartum) samples in a mixed model to predict maternal DBP values at gestational week 36. Owing to the reduction of the sampling variation in prediction of maternal DBP, the predicted values have a lower SD of 0.35 (while, e.g., DBP in the postpartum samples has an SD of 1.52). This results in a greater observed estimate, as an increase per unit is roughly equivalent to 3 SD in this analysis. We used bootstrapping (10,000 replications) to obtain unbiased CIs and present bias-corrected CIs.
↵‡As these results arise from bootstrapping estimations, a P value is not provided.
↵§There was a statistically significant interaction (P_interaction = 0.01) between rs11568820 and 25(OH)D (Supplementary Table 3).