Serum 25-Hydroxyvitamin D Concentrations in Children Progressing to Autoimmunity and Clinical Type 1 Diabetes

Abstract

Context: The role of vitamin D in the development of type 1 diabetes (T1D) remains controversial.

Objective: The objective of the investigation was to study whether there are detectable differences in serum 25-hydroxyvitamin D (25[OH]D) concentrations between children who later progressed to T1D (cases) and matched children who remained nondiabetic and negative for islet autoantibodies (controls) when followed up from birth until disease onset.

Design: A total of 3702 prospective serum samples from 252 children were measured for 25(OH)D from the age of 3 months onward using an enzyme immunoassay. Differences between the groups were compared by the mixed-model analysis of variance.

Setting: T1D prediction and prevention study clinics in Turku, Oulu, and Tampere University Hospitals, Finland, participated in the study.

Participants: By the end of 2012, all 126 case children were diagnosed with T1D. The control children (n = 126) were matched for age, sex, study site, and human leukocyte antigen-HLA-DQ-conferred risk for T1D.

Main outcome measure: Median circulating 25(OH)D concentration (nanomoles per liter) was measured.

Results: The patterns of variation in circulating 25(OH)D concentrations were similar between cases and controls and did not correlate with the age at seroconversion to autoantibody positivity (P = .79) or disease onset (P = .13). The median concentration of all collected samples did not differ between case and control children (66.6 nmol/L [range 14.0-262.8] vs 67.4 nmol/L [range 19.9-213.0]) P = .56).

Conclusions: This study shows that serum 25(OH)D concentrations are not associated with the development of T1D in Finland.

Figure 1.

Monthly distribution of median 25(OH)D concentrations in cases (white) and controls (gray) of all ages in the data set matched additionally for the closest sample date, ie, restricted data, with number of samples for each group at the top for each box (first cases, then controls). Heavy lines indicate median concentrations and box ends represent quartiles, whereas whiskers indicate the lowest value still within 1.5 interquartile range of the lower quartile and the highest value still within 1.5 interquartile range of the upper quartile. Any data not included between the whiskers is plotted as an outlier (circles in cases, triangles in controls). The differences between cases and controls were not statistically significant at any month when adjusted for age, sex, and sample year.

Figure 2.

25(OH)D concentrations in serum samples of children of different ages. All analyzed samples are shown (full data set), with years 1994–2012 combined. Case samples are shown in gray (dots) and control samples in black (triangles), median 25(OH)D in cases (gray line), and controls (dashed line) first at 3-month, then 6-month, and finally 12-month intervals.

Figure 3.

Median monthly 25(OH)D concentrations before seroconversion (solid line), from seroconversion to T1D (line with shorter dashes), and controls (line with longer dashes) in the full data set.