Introduction
Type 1 diabetes mellitus (T1DM) is a disease characterised by the destruction of pancreatic beta cells by an autoimmune process. The consequence is absolute insulin deficiency and constant insulin dependence [1–3]. Lymphocytic thyroiditis (up to 20-40%) and celiac disease (up to 10%) are among the most commonly coexisting diseases with T1DM [4, 5]. In recent years, there has been an increase in the prevalence of T1DM and associated autoimmune diseases worldwide, prompting research into factors that may modify autoimmune processes. This ongoing trend is primarily attributed to changes in the environment and lifestyle. Among the known immunomodulatory factors is vitamin D, which plays a role in regulating various biological processes, including immune system function [6, 7]. Alongside the increasing prevalence of T1DM, there is a growing trend of vitamin D deficiency across different age groups, including children and adolescents worldwide [8–10]. Scientists are increasingly interested in the extra-skeletal effects of vitamin D on the human body. Studies suggest that 25-hydroxy vitamin D (25(OH)D) deficiency carries potential health consequences, including the risk of cardiovascular, neoplastic, inflammatory, and autoimmune diseases [8, 11, 12]. Vitamin D deficiency affects populations worldwide and depends on various factors: genetic, ethnic, and environmental (indoor time, diet, head covering, supplementation) [9, 10]. 25(OH)D deficiency is also more common in patients with T1DM compared to the healthy population [13, 14]. The main source of vitamin D in the human body is endogenous synthesis in the skin (up to 80%). Cholecalciferol is synthesised in human skin under the influence of UVB radiation, which is why the lowest concentrations of vitamin D are observed in winter and spring, while the highest are in summer months. A small portion of vitamin D (up to 20%) is obtained from food (fatty fish, oils). Vitamin D undergoes hydroxylation processes in the liver to 25-hydroxyvitamin D (calcidiol), and then in the kidneys to the active form 1,25-dihydroxyvitamin D3 (calcitriol). Calcitriol acts on the body’s cells by binding to the nuclear vitamin D receptor, and it influences the transcription of several genes [15, 16]. The vitamin D receptor (VDR) is also present in immune cells, explaining the pleiotropic effect of vitamin D [17, 18].
The aim of the study was to assess vitamin D levels in patients with newly diagnosed T1DM, taking into account the most commonly coexisting autoimmune conditions associated with T1DM (positive antibodies against thyroperoxidase and/or thyroglobulin, positive antibodies against tissue transglutaminase IgA).
Material and methods
Study group
A cross-sectional study was conducted at an accredited, reference, regional diabetological centre [certificate International Society for Paediatric and Adolescent Diabetes (ISPAD), European Union grant acronym “Better control in Paediatric and Adolescent diabeteS: Working to crEate CEnTers of Reference” (SWEET) 2020–2026], at the Department of Children’s Diabetology and Paediatrics of the Upper Silesian Centre for Child Health in Katowice, Poland, during a 24-month observation period from January 2020 to December 2021. The database included all hospitalised patients suspected of newly diagnosed T1DM (455 patients). Patients with negative antibodies for T1DM (n = 81) and those lacking the result of the above-mentioned antibodies (n = 13) were excluded from the study group. Finally, the study cohort comprised 361 patients, with a mean age of 9.27 ± 4.1 years, 189 boys, and their results of routinely performed tests.
Anthropometry and biochemical measurements
The database included information on gender, date of birth, date of T1DM diagnosis, height and weight, and the results of routinely performed tests in patients with newly diagnosed T1DM: blood pH and bicarbonate (HCO3–) concentration on admission gasometry, glycated haemoglobin (HbA1c), levels of antibodies against glutamic acid decarboxylase (GAD), antibodies against tyrosine phosphatase (IA2), antibodies against zinc transporters (ZnT8), 25(OH)D concentration, anti-IgA tissue transglutaminase antibodies (TTG-IgA) concentration, total IgA concentration, and levels of antibodies against thyroperoxidase (TPOAb) and antibodies against thyroglobulin (TgAb).
Methodology
According to World Health Organisation (WHO) guidelines, vitamin D levels were assessed as optimal at 30–50 ng/mL, suboptimal at 20–30 ng/mL, and deficient below 20 ng/mL. Due to seasonal variability in vitamin D levels resulting from differences in sunlight exposure in different months, the calendar season was also taken into account in the analysis: spring 21.03–21.06, summer 22.06–23.09, autumn 24.09–20.12, and winter 21.12–20.03.
The body mass index (BMI), along with the date of birth, gender of the child, and date of examination, was then used to calculate the BMI z-score using a calculator [19] and to assess BMI in percentiles. The following ranges were adopted [19]: BMI z-score below -2 (< 3rd percentile), BMI z-score in the range of –2 to 1 (3rd–85th percentile), BMI z-score above 1 (> 85th percentile).
For children under 2 years of age (n = 11), WHO percentile charts were used to assess BMI due to the inability to use the BMI z-score calculator in this age group [20].
In accordance with the current Polish Diabetes Society (PTD) 2024 guidelines, diabetic ketoacidosis (DKA) was defined as venous blood pH < 7.3 or bicarbonate concentration < 18 mmol/L with hyperglycaemia > 200 mg/dL and accompanying ketonaemia or ketonuria [3].
In the accredited Upper Silesian Centre for Child Health (GCZD) Laboratory, blood pH and bicarbonate concentration were assessed in gasometry. HbA1c (%) levels in whole blood were measured using high-performance liquid chromatography (HPLC, Bio-Rad, USA). The concentration of 25-hydroxyvitamin D (ng/mL) in serum was measured by electrochemiluminescence (ECLIA, Vitamin D total, Cobas, Roche, USA), with a measurement range from 3.0 ng/mL to 70.0 ng/mL. TTG-IgA concentration (U/mL) in serum was determined by immunoenzymatic method (ELISA, BlueWell Transglutaminaze IgA, D-tek, Belgium). With normal total IgA concentration, TTG-IgA results above 50 U/mL were considered positive, while results equal to or below 50 U/mL were considered negative. Total IgA concentration (mg/dl) was quantitatively determined by immunoturbidimetry (Beckman Coulter, Ireland). TPOAb (IU/mL) and TgAb (IU/mL) antibodies were evaluated by chemiluminescent immunoenzymatic method (IMMULITE 2000 anti-TPOAb and anti-TgAb, UK). Results of TPOAb equal to or greater than 35 IU/mL and TgAb equal to or greater than 40 IU/mL were considered positive. GAD (U/mL), IA2 (U/mL), and ZnT8 antibody concentrations (U/mL) were determined by immunoenzymatic method (ELISA) in an accredited laboratory certified by the Department of Immunopathology and Genetics of the Medical University of Lodz, Poland.
Statistical analysis
Statistical analysis was performed retrospectively, using Statistica software (StatSoft, Tulsa, OK, USA). Mean values, medians, standard deviations, and ranges were used to present descriptive statistics of continuous variables. Normal distribution of data was verified by the Shapiro-Wilk test. Absolute values and percentages were provided for qualitative variables. Student’s t-test was uded for independent samples or the Mann-Whitney U test for comparative analyses of continuous variables, depending on the distribution of data. Analysis of variance (ANOVA) and subsequent verification with the least significant difference (LSD) test were conducted when comparing more than 2 subgroups. Correlation analyses were performed using Pearson’s method or Spearman’s test, depending on which was appropriate, according to the data distribution. A significance level of p < 0.05 was adopted for all statistical analyses.
Results
The general characteristics of the study group are presented in Table 1.
|
Parameters |
Number of records |
Mean value |
Median |
Minimum value |
Maximum value |
Standard deviation |
|
Age of onset |
361 |
9.27 |
9.44 |
1.01 |
17.99 |
4.11 |
|
Body weight [kg] |
361 |
33.52 |
31.00 |
7.60 |
112.80 |
16.84 |
|
Height [cm] |
361 |
136.86 |
139.00 |
72.00 |
187.00 |
25.02 |
|
BMI [kg/m2] |
361 |
17.00 |
15.93 |
11.53 |
35.57 |
3.95 |
|
BMI pc (%) |
361 |
37.6 |
26.50 |
1.00 |
99.90 |
34.29 |
|
BMI Z-score |
350* |
–0.61 |
-0.60 |
-5.56 |
3.15 |
1.58 |
|
HbA1c [%] |
361 |
12.39 |
12.4 |
5.4 |
20.1 |
2.63 |
|
pH |
361 |
7.27 |
7.32 |
6.9 |
7.52 |
0.16 |
|
25(OH)D [ng/mL] |
361 |
25.06 |
23.13 |
3.17 |
66.69 |
10.55 |
The analysis revealed a statistically significant difference in the concentration of 25(OH)D between 2020 and 2021. The mean vitamin D concentration among 175 hospitalised children in 2020 was 23.88 ng/ml, while among 186 with T1DM diagnosis in 2021, it was 26.16 ng/ml (p < 0.05). The lowest concentrations of 25(OH)D were observed in winter 2020 (p < 0.01) and spring 2020 (p < 0.05). The variability of vitamin D concentration according to seasons is presented in Table 2.
|
Season |
n |
Mean 25(OH)D concentration [ng/mL] |
SD |
p |
|
Winter 2020* |
37 |
19.09 |
7.75 |
< 0.01 |
|
Winter 2021 |
44 |
24.5 |
10.17 |
|
|
Spring 2020* |
38 |
20.1 |
10.24 |
< 0.05 |
|
Spring 2021 |
49 |
24.32 |
8.93 |
|
|
Summer 2020 |
54 |
29.89 |
11.55 |
NS |
|
Summer 2021 |
49 |
31.23 |
10.87 |
|
|
Autumn 2020 |
46 |
23.79 |
8.9 |
NS |
|
Autumn 2021 |
44 |
24.22 |
9.53 |
Vitamin D deficiency was observed in 35.5% of children, suboptimal levels in 37% of individuals, and 25% of results fell within the optimal concentration range. There was no statistically significant difference in 25(OH)D concentration between the boys and girls (24.96 vs. 25.16 ng/mL). The percentage distribution of vitamin D concentrations in specific groups is presented in pie charts.
In the conducted analysis, a negative correlation between 25(OH)D concentration and children’s age at diagnosis was demonstrated (p < 0.001). Statistical calculations did not show (p > 0.05) a relationship between 25(OH)D concentration and BMI Z-score.
In the ANOVA test, no association was found between the types of antibodies typical for T1DM and the 25(OH)D level at diagnosis. Furthermore, no relationship was found between the concentration of 25(OH)D and the presence of 2 or 3 classes of positive antibodies: GAD, IA2, and ZnT8.
However, the analysis revealed a positive correlation between the presence of ketoacidosis at diabetes diagnosis and 25(OH)D concentration (p < 0.01). The mean concentration of 25(OH)D in the group of children with pH below 7.3 was 23.84 ng/mL, while in the group with normal pH, it was 26.18 ng/ml. The study also demonstrated a relationship between vitamin D concentration and HbA1c value at diabetes diagnosis. Significantly higher 25(OH)D levels were observed in the subgroup of patients with HbA1c < 7.5% (p < 0.05).
Evaluation of the group of children with diabetes and positive antibodies against thyroid peroxidase and/or thyroglobulin (14%, n = 51), as well as with positive antibodies against tissue transglutaminase in the IgA class (5%, n = 18), did not reveal statistically significant differences in the concentration of 25(OH)D. Stepwise regression multivariable analysis demonstrated that in the conducted study, the following variables had the most significant impact on the value of 25(OH)D concentration: patient’s age at onset (p < 0.001) and blood pH value in gasometry at T1DM diagnosis (p < 0.001).
|
HbA1c cut-off point 1 — < 7.5%, 2 — ≥ 7.5% |
n |
Mean 25(OH)D concentration [ng/mL] |
SD |
p |
|
Grade 1 |
18 |
30.43 |
14.28 |
< 0.05 |
|
Grade 2 |
343 |
24.77 |
10.28 |
Discussion
There are numerous reports in the literature confirming the current increasing trend in the incidence of T1DM and vitamin D deficiencies in the paediatric population worldwide [13, 21–26]. The results of our study confirm the commonly occurring deficit of 25(OH)D in children with T1DM at disease onset. Only 25% of the studied population had normal 25(OH)D concentration at T1DM diagnosis. This has also been demonstrated in other clinical centres [8, 27, 28]. In the 2023 update of guidelines on the prevention and treatment of vitamin D deficiency in Poland, experts confirm the widespread deficiency of vitamin D among the general Polish population, including children and adolescents [29]. A multicentre study conducted in 6 representative geographical locations in Poland in 2011, involving 720 healthy children aged 9-13 years, clearly showed a common deficiency of 25(OH)D in the studied age group in early spring (March), when only 15.8% of the study group presented optimal vitamin D levels, whereas after the summer period in October, a significant improvement in 25(OH)D concentration was observed in the studied cohort — normal vitamin D levels were noted in as many as 74% of the participants [30]. In our study involving patients with newly diagnosed T1DM, we did not find significantly higher values of 25(OH)D in the summer and autumn. A study conducted in Piedmont (Italy) between 2008 and 2014, involving 141 patients with newly diagnosed T1DM, similarly demonstrated a generalised deficiency of vitamin D at every stage of this study, with the highest percentage of patients with 25(OH)D deficiency observed at the time of T1DM onset, reaching as high as 64% [8]. Publications addressing this topic present interesting hypotheses and therapeutic attempts. For example, a study conducted by Mohr et al. suggests a link between low UVB irradiation and a significantly higher frequency of T1DM occurrence in childhood [28]. Another international multicentre study showed that in patients with newly diagnosed T1DM, subcutaneous administration of glutamic acid decarboxylase, combined with oral vitamin D supplementation, allowed for maintaining higher C-peptide levels by modifying the course of the disease [31]. A systematic review (1980–2022) found that maintaining a normal, optimal level of 25(OH)D may reduce the risk of T1DM and delay the onset of complete insulin deficiency in patients with diabetes [32]. A meta-analysis (1965–2020) also demonstrated a statistically significant inverse proportional relationship between serum vitamin D levels and the risk of T1DM occurrence [33]. The assessment of seasonality comes from the DIAMOND observation (1990–1994), which illustrated a higher incidence of T1DM in geographical areas with lower UVB radiation [34]. Moreover, numerous studies have documented a seasonal pattern of T1DM incidence, peaking in winter, early spring, and late autumn months [35], when vitamin D deficiency is most commonly observed. It is worth noting that in our observation, the lowest concentrations of 25(OH)D were noted in winter and spring 2020, which may indicate a connection with the period of strict lockdown due to the COVID-19 pandemic at that time.
In our study, no significant differences in 25(OH)D concentrations were found between boys and girls, as in the aforementioned Polish study involving healthy children aged 9–13 years [30].
However, our study revealed a negative correlation between 25(OH)D levels and the age of patients at diagnosis (p < 0.001), which is most likely associated with compliance with the obligation to supplement vitamin D in infants and children up to 2 years of age. Many researchers also emphasise that the frequency of vitamin D deficiency increases with age in childhood, which probably results from increased demand for vitamin D during growth spurts, less adherence to vitamin D supplementation recommendations in the adolescent group, and reduced outdoor physical activity in this age group [36, 37].
In our study, a positive correlation was found between the concentration of 25(OH)D and the pH value in venous blood gas analysis at disease onset. Children diagnosed with T1DM who presented with diabetic ketoacidosis had significantly lower vitamin D levels compared to the group without acid-base imbalances. A previously mentioned Italian study also showed that vitamin D deficiency in the studied population was associated with a more severe form of diabetic ketoacidosis at diagnosis [8]. Other multicentre studies also confirm that lower 25(OH)D levels appear to be more common in patients with diabetic ketoacidosis at disease onset [38–40]. These data may suggest the role of vitamin D deficiency as a factor shaping the development and clinical course of T1DM. In our study, a negative correlation was also noted between the concentration of 25(OH)D and the initial value of HbA1c. In the group of patients with HbA1c above 7.5%, 25(OH)D levels were significantly lower, which may correspond to insulin deficiency and progressive decompensation of the disease, considering primarily the fact that the greatest influence on HbA1c value is exerted by the mean blood glucose values from the last 6 weeks [41, 42]. This observation was also confirmed in other previously mentioned studies [8].
However, in the conducted analysis, no significant relationship was found between vitamin D levels and BMI Z-score, probably due to the fact that the majority of patients in the study group had normal body weight.
There was also no association between the value of 25(OH)D and the type and number of positive antibodies characteristic of T1DM. An interesting observation is that among patients with 3 classes of positive antibodies: GAD, IA2, and ZnT8, no significantly lower values of 25(OH)D were observed. These observations confirm Danish studies based on the Danish Childhood and Adolescent Diabetes Register (DanDiabKids), which also did not show a relationship between 25(OH)D levels and the presence of antibodies against glutamic acid decarboxylase or tyrosine phosphatase [43].
The aim of our study was also to assess whether the coexistence of other autoimmune processes with diabetes is associated with higher hypovitaminosis D. In our study, positive thyroid antibodies were found in 14% of patients, while antibodies against tissue transglutaminase were found in 5%. In both groups, the presence of additional autoimmune disorders was not significantly associated with lower vitamin D levels. In the available literature, many studies have shown more frequent vitamin D deficiencies in patients with autoimmune thyroiditis compared to a healthy population of similar age [44–47]. The lack of statistical significance in our cohort study presumably results from the small number of patients with positive TPOAb and/or TgAb antibodies, and perhaps from the fact that these patients were “captured” in the euthyroid phase, at the beginning of the autoimmune process, before the development of metabolic disorders associated with thyroid hormone deficiency. Similarly, in celiac disease, literature data report that the disease is more often associated with hypovitaminosis D [48–50]. The lack of significance in our study may also result from the small number of patients with positive TTG-IgA antibodies and from the early stage of detection of autoimmune processes towards celiac disease.
There is growing evidence indicating a relationship between hypovitaminosis D and pathogenesis autoimmune diseases, such as T1DM [51–55], and its impact on potential complications in the subsequent course of diabetes. Given the above, the introduction of effective therapy for 25(OH)D deficiency as soon as possible is significant in children with T1DM.
The strength of our study is the fact that, to our knowledge, it is the first study conducted in the reference, regional diabetes centre, evaluating the concentration of vitamin D among a large number (361) of children with newly diagnosed T1DM. However, a limitation of our study is the relatively short 2-year observation period in one centre, which also included the strict lockdown period caused by the COVID-19 pandemic, which may have influenced the results obtained.
Conclusions
Deficiency of 25(OH)D is widely prevalent among children with newly diagnosed T1DM and often accompanies metabolic decompensation with ketoacidosis at onset. Determination of vitamin D levels should be a routine procedure in children with newly diagnosed T1DM.
Acknowledgments
The manuscript was supported by Joanna Zarębska, Ph.D. and Anna Dolbecka, M.A., who were involved in gathering information for the database. Thanks also to Mateusz Kciuk, M.A., for his help in preparing the tables and figures.
Author contributions
K.K.-K.: nature — conception, design, methods, investigations, writing of the manuscript; S.S.: nature — supervision, review; P.A.: nature — analysis of data; P.J.-Ch. — nature: conception, supervision, review.
Conflict of interests
The authors declare no conflict of interests.
Funding
There were no grants or other funding sources.
