The LIFE Child study is registered with ClinicalTrial.gov (NCT02550236) and was approved by the Ethical Committee of the Medical Faculty of the University of Leipzig (Reg. No. 264-10-19042010). The procedures were performed in accordance with the ethical standards of the Declaration of Helsinki. The serum samples were collected after receiving informed written consent from the parents and, for subjects age 12 and up, the subjects themselves. Data were collected between May 2011 and December 2014. Detailed information regarding the recruitment and the examinations is given in Poulain et al. and Quante et al. (13, 14).

The LIFE Child study is a longitudinal, prospective, and population-based cohort study and is part of the Leipzig Research Center for Civilization Diseases (LIFE) in the city of Leipzig (Saxony, Germany). Participants’ ages range from 3 months to 20 years. Most of the study participants originate from Leipzig or its immediate surroundings. They undergo an age-specific study program with yearly follow-up visits that include various medical, psychological, and sociodemographic assessments, as well as the collection of biological samples. The study was designed to examine the influence of genetic, metabolic, and environmental factors on children’s growth, development, and health. A total of 2,673 serum samples from 1,802 subjects were available ( Figure 1 ). Medication, diseases, or physiological circumstances that may affect DBP and vitamin D metabolism, such as concomitant kidney diseases, liver diseases, severely underweight (BMI-SDS < -4), glucocorticoid therapy, or pregnancy were excluded (2, 5, 15–17). In line with recent recommendations, we excluded subjects with C-reactive protein (CRP) > 20mg/l (18). Subjects older than 17 were excluded because this subgroup had such a small sample size. The final data set comprised 2,503 samples from 1,716 subjects between 0.25 and 17 years of age. Out of this study population three sub cohorts for additional analysis were established. For the reference cohort, children and adolescents using oral contraceptives or with BMI-SDS < -1.881 or > +1.881 were excluded. We did not exclude subjects with vitamin D supplementation because it is recommended for all infants in Germany and does not seem to affect DBP serum levels (19). 2,067 samples from 1,414 subjects were included to establish reference values. To assess the association between BMI-SDS and DBP, we established a sub cohort including only the individuals with BMI-SDS > 1.881 (317 samples from 253 individuals). In addition, to compare the DBP levels of women who reported (27 samples from 23 subjects) or were thought to be taking oral contraceptives with the corresponding population of women who were not taking oral contraceptives in total 73 samples from 65 subjects were included. This sub cohort contains, in line with Hörenz et al., all females with sexual hormone binding globulin (SHBG) levels ≥ 200 ng/ml and age >13 years (46 samples from 42 subjects) who did not report oral contraceptive use but were assumed to be taking oral contraceptives as well (20).

Height, weight, and pubertal status were assessed by trained study assistants. BMI was calculated and transformed into age- and sex-adjusted standard deviation scores (SDS) in accordance with the S3 guidelines of the German Obesity Society and the German Society of Pediatrics and Adolescent Medicine (21). Accordingly, obesity was defined as having a BMI-SDS > 1.881, overweight as a BMI-SDS of 1.28 to 1.881, normal weight as a BMI-SDS of -1.28 to < 1.28, underweight as a BMI-SDS of < -1.28 to -1.881, and extreme underweight as a BMI-SDS < -1.881. Pubertal status was measured as Tanner stages (22, 23).

Venous blood was drawn from the study subjects by trained personnel in the morning (with 90% collected between 07:30 and 10:00, 96% by 11:00, 99% by 12:00, and 100% by 16:00) (24). Parts of the samples were processed by trained staff from the Leipzig Medical Biobank and sent to the Institute of Laboratory Medicine, Leipzig, on the same day for analysis. The rest were aliquoted, frozen, and stored at -80°C or -150°C in the LIFE-Biobank for additional analytical procedures (14). To determine the DBP values, an automated immune turbidimetric assay was performed as a DAKO-derived free application for the Gc-globulin determination in serum using the c-module of the Cobas^© 8000 system from the Roche Diagnostics GmbH (Germany, 25, 26). The intra- and interassay coefficients of variation for a pool serum with a level between 266 and 271 mg/l were below 1.8% and below 2.1%, respectively. Albumin, SHBG, and CRP were analyzed with a Cobas^© 6000/8000 Clinical Chemistry Analyzer with test kits from Roche Diagnostics GmbH in accordance with the clinical chemistry routine. Cobas^© 601 and 801 from Roche were used to measure bone metabolism parameters (osteocalcin) and calcium homeostasis (parathormone, 25(OH)D₃ total). 1,25(OH)₂-D₃ was measured with enzyme immunoassays in the IDS-iSYS Multi-Discipline Automated System. Steroid hormones were analyzed with liquid chromatography tandem mass spectrometry, which was described in detail by Gaudl et al. (27).

Descriptive statistics are presented as counts (percentages) for categorical variables and means for continuous variables ( Table 1 ). We considered p < 0.05 to be statistically significant. All statistical analyses were computed in R (version 4.2.1) (28). Generalized additive models for location scale and shape were used to estimate age- and sex-adjusted reference intervals for DBP (29), assuming a Box-Cox-Cole-and-Green distribution. To avoid multicollinearity caused by correlated laboratory measures, laboratory measures were clustered using hierarchical clustering as a basis for the subsequent regression analyses. The number of clusters was determined by visual inspection ( Supplementary Figure S1 ). Only one representative per cluster was included in the multivariate models. Eight parameters were identified as representatives of their clusters. The final models were stratified by sex, as sex hormones were included as predictor variables. Non-significant variables were removed from the model in a stepwise fashion (backward deletion). Only parameters that were of particular interest for our research question were retained despite non-significance. Univariate regression analysis was applied to investigate the relationships between DBP and its metabolites. Differences in DBP serum levels between males and females, as well as between the reference and obesity groups, were estimated with linear mixed-effects models. Beta coefficients were used as measures of effect size to indicate relationship strength and direction. All models listed so far were adjusted for multiple measurements per subject by adding the subject as a random intercept. We computed t-tests to compare the mean values between Tanner stages.

Table 1 presents descriptive statistics for the sample. 1,716 subjects with 2,503 observations (49.3% females, 50.7% males) were included. The mean age (all measurements) was 10.9 ± 3.5 years (range: 0.94 – 17.98 years). In total, there were 73 samples from females taking oral contraceptives. 73% of all samples came from normal-weight subjects, 7.7% from underweight and extreme underweight subjects, 7.1% from overweight subjects, and 13% from subjects with obesity.

The 2.5th, 5th, 10th, 25th, 50th, 75th, 90th, 95th, and 97.5th percentiles for the reference cohort stratified by sex at different ages are shown in Table 2 . The DBP smoothed percentile curves ( Figures 2 , 3 ) showed a similar pattern in males and females. From infancy until late adolescence, DBP levels remained relatively stable. In the higher percentiles (P90-P97.5), the values were predominantly higher in females than males. The gap between females’ and males’ measurements was particularly large between the ages of 12 and17.98. The lower percentiles (P2.5-P10) did not differ markedly between males and females. For DBP, the 50th percentile remained relatively stable for males from 8 to 17.98 years of age, while it increased slightly but continuously in females. The increase was small (ß = 1.3 mg/l/year) but significant (p = 0.006) for girls. There was no similar trend for boys.

The mean DBP values differed between males (347 mg/l) and females (366 mg/l) (p < 0.001). Age had no significant association with the median DBP serum level ( Figure 2 ). DBP peaked at Tanner stage 4 in females (362 mg/l) and at Tanner stage 2 in males (356 mg/l) ( Figure 4 ). DBP levels were significantly lower in males than females during Tanner stages 3, 4, and 5 ( Supplementary Table S1 ). As presented in Table 1 , DBP levels were significantly lower in males during Tanner stages 3, 4, and 5 compared with pre-pubertal males (Tanner stage 1). In females, on the other hand, no significant relationship was observed between Tanner stage and DBP values.

In the study cohort (1716 subjects and 2503 samples) DBP was inversely associated with BMI-SDS, but the association was statistically significant only in males (p < 0.001). While an increase in BMI-SDS of +1 showed a decrease of 5.72 mg/l in the DBP serum concentration in males, it decreased by only 1.67 mg/l in females ( Table 3 , Figure 5 ). DBP values in the group with obesity were significantly lower than DBP in the reference group, age-adjusted and independent of sex (ß = -14.28, p < 0.001). These findings are illustrated in Figure 2 . Males with obesity (159 samples from 130 subjects) had significantly lower DBP levels than their reference weight peers (1081 samples from 742 subjects) in Tanner stages 1, 2, and 4. Similarly, females with obesity (158 samples from 123 subjects) had significantly lower levels in Tanner stages 1 and 5 than their reference group (986 samples from 672 subjects) ( Figure 4 ; Supplementary Table S2 ).

DBP values from subjects taking oral contraceptives are highlighted in Figure 3 (females). The mean DBP value (508 mg/l) was significantly higher (estimate = 141.38, b=0.69, p < 0.001) compared with age-BMI matched subjects in Tanner stage 4 and 5, independent of the oral contraceptive formulations. Furthermore, we found that estradiol serum levels in female subjects who were not taking oral contraceptives were not significantly associated with DBP levels.

In both sexes, significant positive associations were found between DBP and the vitamin D metabolites: 25(OH)D₃ (females: ß = 0.8, p < 0.001; males: ß=1.2, p<0.001) and 1,25(OH)₂-D₃ (females: ß = 0.3, p < 0.001; males: ß = 0.4, p < 0.001) ( Table 3A ). The analysis presented in Table 3 showed a significant inverse association between osteocalcin and DBP (females: ß = -0.1, p < 0.022; males: ß = -0.1, p = 0.027). A significant relationship was found between albumin and DBP (females: ß = 4.0, p < 0.001; males: ß = 3.2, p < 0.001). In males but not in females, there was also a significantly positive association between SHBG and DBP (ß = 0.1, p = 0.005). A statistically significant positive relationship between CRP levels (< 20mg/l) and the serum concentration of DBP (females: ß = 2.8, p = 0.001; males: ß = 5.1, p < 0.001) was observed. DBP did not show significant relationships with parathormone or either testosterone in males or estradiol in females.

With the current study, we were able to establish age- and sex-specific reference ranges for DBP serum concentrations for healthy children and adolescents. To the best of our knowledge, no other studies have used such a large cohort to publish reference values for DBP levels for healthy children and adolescents. DBP reference values are important for the accurate assessment of vitamin D status, which plays a central role in the development of children and adolescents (3). We estimated age- and sex-specific reference ranges for DBP using 2,067 samples from 1,414 children and adolescents between the ages of 0.25 to 17.98 years. Only a few studies have examined DBP levels in children and adolescents. They found mean values between 330 mg/l and 437 mg/l (30). In reviews by Bouillon et al. and Delanghe et al., DBP serum levels between 200 mg/l and 600 mg/l were reported across all age groups as comparative parameters (3, 15). Variability in DBP serum levels due to the lack of a standardized measurement method makes it challenging to assess DBP’s associations with age, sex, pubertal status, BMI, and other laboratory measurements (3).

Our results are in line with most of the published research, which also reported no significant association between DBP levels and age (12). Previous studies also reported higher DBP values in females (31, 32). In our study, DBP levels differed between pubertal stages, especially in male subjects. However, the differences were not substantial, thereby showing, in line with other studies, that DBP is not a good marker of pubertal status (33).

Due to the high prevalence of obesity worldwide and in all age groups, we investigated the relationship between BMI and DBP in different stages of life. We found that DBP values decreased as BMI-SDS values increased. Subsequently, the reference cohort subjects showed higher DBP values than the cohort with obesity. The liver is the primary site for the biosynthesis of DBP. Therefore, liver disease, like liver cirrhosis, can sometimes affect protein biosynthesis and result in reduced production of DBP (34). Obesity-related non-alcoholic fatty liver might have similar effects. In the long run, lower levels of DBP could lead to vitamin D deficiency, which is linked to the risk of infections and auto-immune disorders (35). Furthermore, a recent study observed an inverse association of DBP levels with insulin resistance and hyperinsulinemia (36), which can be based on higher BMI levels but also various other reasons such as type 2 diabetes mellitus reported in Moller et al. (37). Higher BMI and lower DBP values might be a risk factor for related pathologies. However, other studies have found no difference in DBP values concerning various BMI groups (12, 38). On the other hand, studies in adults samples have reported positive (10, 39) or negative (4) associations between DBP values and BMI. The size of our cohort may render it more representative than many other studies.

We found that oral contraceptives were significantly related to DBP serum levels. This finding lines up with the literature, where a positive association between DBP levels and oral contraceptives has often been reported (2, 6, 31). The DBP serum concentration may increase with estrogen exposure (7, 8, 40). The mechanism behind this effect is not yet fully understood. Oral contraceptives might have positive associations with levels of other binding proteins such as SHBG, trancortin, thyroxine-binding globulin (TBG) (6), and insulin-like growth factor-binding protein 3 (IGFBP-3) (19). These associations suggest an anabolic effect of estrogens on the regulation of hepatic globulin synthesis. Because we also found a small positive association between SHBG and DBP, there might be a common regulatory mechanism. However, oral contraceptives could increase the vitamin D reservoir and thus could prevent vitamin D deficiency. There is evidence that oral contraceptive therapy increases bone density in women with premature ovarian insufficiency (41). Oral contraceptive drugs should also be considered as hormone therapy to prevent diseases such as osteomalacia and osteoporosis. On the other hand, there is evidence that bioavailable vitamin D can be decreased under the use of specific oral contraceptive formulations in adults (42). Given the conflicting research results, it could be hypothesized that females who have used specific formulations of oral contraceptives could have a problem with bioavailable vitamin D, which is compensated for by higher DBP levels. More research is needed to find out.

We found that levels of 1,25(OH)₂-D_3, 25(OH)D₃, and albumin were positively associated with DBP levels, findings that were also previously supported (43, 44). Notably, although DBP increases the total amount of vitamin D metabolites in the bloodstream, it does not appear to affect the amount of the biologically active hormone that enters cells and tissues (45). It is important to note that different polymorphisms of DBP also have different affinities for vitamin D metabolites. The prevalence of the isoforms differs by ethnic origin (46), which is why the biological relevance of DBP-bound vitamin D metabolites versus the DBP-unbound or free fraction of vitamin D has not been definitively established (3, 47). Bone parameters’ associations with age, sex, pubertal status, and BMI were recently described by Geserick et al. (48). The associations between measured parameters and DBP found in our study were very small but still significant. Since bioavailable vitamin D induces the formation of osteocalcin (49), the inverse relationship between osteocalcin and DBP could indicate that DBP is part of a negative-feedback-based regulation of the total amount of vitamin D despite the finding that this effect was small. A positive association between DBP levels and acute phase reactants such as CRP or interleukin-6 has already been reported in various studies (5, 39, 50). It was speculated that higher CRP levels in people with obesity may be related to higher DBP serum levels, but our findings did not support this idea. Rather, our results support the hypothesis that DBP is part of the acute phase response in which liver protein synthesis of acute phase proteins is upregulated by proinflammatory cytokines (50, 51). However, Madden et al. reported lower DBP levels during severe infections (30). The association between CRP and DBP was positively significant, even at slightly elevated CRP levels (<20 mg/l), suggesting DBP may indicate minor inflammatory processes. The relationship between DBP and inflammatory cytokines may also depend on the severity and duration of inflammation, as DPB plays a major role in inflammation and immune cell modulation as an acute-phase reactant itself.

The main strength of the current study is the large number of subjects. Furthermore, all laboratory analyses were performed in only one facility. The subjects were mostly Caucasian and lived near or in the city of Leipzig. The results were based on the social distribution of the city of Leipzig, and therefore, the estimated reference ranges may be valid only for comparable populations. To include regional and social variety, other studies should be considered. It is essential to consider the limited amount of data available for children under 4.5 years of age when applying reference values from 0.25 years to 17.98 for the reference cohort.