Association of serum 25-hydroxyvitamin D status with bone mineral density in 0-7 year old children

Objective To describe the status of serum 25-hydoxyvitamin D [25(OH)D] concentrations and identify the relationship between 25(OH)D and bone mineral density (BMD). In an effort to explore the appropriate definition of vitamin D (VD) deficiency in 0–7 year old children. Results The median serum 25(OH)D concentrations was 62.9 nmol/L and 28.9% of the children had a low 25(OH)D (< 50 nmol/L). And a linear relation between 25(OH)D concentrations and BMD was surveyed (r = 0.144 , P < 0.001). After adjusting for the confounders, serum 25(OH)D was positively associated with BMD (β = 172.0, 95%CI = 142.8–201.2, P < 0.001), and low 25(OH)D (< 75 nmol/L) had a high stake for low BMD (OR = 1.424, 95%CI = 1.145–1.769, P = 0.001). Additionally, there was a nonlinear relation between 25(OH)D and low BMD, and a critical value for 25(OH)D of 75 nmol/L appeared for low BMD. The prevalence of low BMD was 14.1% in children with 25(OH)D ≥ 75 nmol/L, much lower than that of the concentrations between 50–75 nmol/L and < 50 nmol/L. Materials and Methods A total of 4,846 children 0–7 years old were recruited in Jiangsu Province, China. BMD and serum 25(OH)D concentrations were determined by quantitative ultrasound and enzyme-linked immunosorbent assay, respectively. Linear regression and logistic regression analyses were used to assess the association of 25(OH)D concentrations with BMD. Conclusions Serum 25(OH)D concentrations was related with BMD and 25(OH)D concentrations < 75 nmol/L might be a more appropriate definition of VD deficiency in 0–7 year old children.


INTRODUCTION
Vitamin D (VD) deficiency is common among children due to the combined effects of additional growth needs, inadequate intake and limited sunlight exposure, which has become an grave public health issue in developed and developing countries [1]. Nowadays VD deficiency is severe among children in China, with a reported prevalence of 30%-70% in the north [2] and 10%-40% in the south of China [3,4]. One of the classic roles of VD is to maintain the homeostasis of calcium and phosphorus by absorption in the intestines and mobilization from the skeleton [5,6]. Since the 1920s, VD was identified and its deficiency was linked to rickets [7] and more and more studies showed that VD deficiency was associated with low BMD [8][9][10][11]. With the extensive researches and global emphasis on the importance of supplement, nearly 30 years ago, the appearance of serious nutritional rickets was considered to vanish [12], however, low BMD remained prevalent [8,10,13]. On the basis of the international society for Clinical Densitometry States, low BMD can be one of the indicators to diagnose osteoporosis and low bone mass [14]. Therefore it is essential to maintain sustained circulating VD and bone mass from childhood to avoid these ricks. 25(OH)D is the storage form of VD, and serum 25(OH)D concentrations is an indicator of VD status. There is a lack of consensus on definition regarding VD deficiency in young children. It is supported by the Institute of Medicine (IOM) and the Endocrine Society's practice guidelines that VD deficiency considering bone health is defined as the serum free 25(OH)D concentrations < 50 nmol/L, while the VD insufficiency is defined as the 25(OH)D concentrations of 52-72 nmol/L [15,16]. A study by Vieth et al. suggested that the maximal inhibition of parathyroid hormone (PTH) by circulating 25(OH)D occurred at the concentrations of 25(OH)D > 80 nmol/L [17]. Some studies including our previous studies supported that when the concentrations > 75 nmol/L, the PTH levels plateaued [18]. But in late 2010, the IOM reported another statement that the concentrations of 25(OH)D > 50 nmol/L was sufficient for bone health.
Whether the definition of deficiency should be extent to 75 nmol/L or greater is less clear [7].
Although the serum free 25(OH)D concentrations can partly reflect the status of VD, in the clinic examination, BMD is a vital functional outcome of VD deficiency in children [4,19]. Quantitative ultrasound measure techniques are used wildly as an effective and convenient tool to measure BMD, particularly in the primary hospitals of China. According to the previous study, the measurement of speed of sound (SOS; m/s) is the common indicator of BMD [20].
At present, the relation between serum 25(OH)D concentrations and BMD in children remains confusion [21]. This study was designed to describe the status of 25(OH)D serum concentrations and determine the relation between serum 25(OH)D concentrations and BMD in 0-7 year old children, and to provide certain reference for the definition of VD deficiency of Chinese children.

RESULTS
A total of 4,846 serum samples were tested for 25(OH) D, with linked information related to the bone mineral density available for 4,622 (95.4%). We excluded 90 children who had liver, kidney or cardiovascular diseases, or who had 25(OH)D concentrations outside the assay detection limits. At last, a total of 4,532 (2380 boys and 2152 girls) children aged from 1 months to 7 years were included in the analysis. The mean (SD) age of children was 35.8 (20.9) months. 5.2% of the children were preterm infants (gestational age < 37 weeks), while the mean (SD) birth weight was 3.4 (0.5) kg. The percentile distributions of 25(OH)D concentrations and selected characteristics are presented in Table 1. The median concentrations was 62.9 nmol/L and 28.9% of the children had a low 25(OH)D (< 50 nmol/L). The serum free 25(OH)D concentrations was associated with gender, age, gestational season, milk intake during pregnancy and for children, VD or calcium supplementation during pregnancy and for children, and time of outdoor activity during pregnancy and for children (P < 0.05). With the increasing quartile of 25(OH)D, the BMD increased (P < 0.001). The lowest quartile of 25(OH)D had a highest prevalence (25.9%) of low BMD (P < 0.001). Compared with children with 25(OH)D > 50 nmol/L, children with low 25(OH)D (< 50 nmol/L) had a higher prevalence of low BMD (P < 0.001). BMD was also associated with age, gestational age and season, delivery mode, and time of outdoor activity during pregnancy and for children (P < 0.05). In addition, gestational season in autumn or winter, spontaneous delivery, milk intake for children < 250 ml/day, no VD or calcium supplementation during pregnancy and for children, and time of outdoor activity during pregnancy and for children < 2 h all increased the stake of low BMD (P < 0.05). www.impactjournals.com/oncotarget As shown in Figure 1, a linear relation between the 25(OH)D concentrations and BMD was discovered (r = 0.144 , P < 0.001). After adjusting the confounders of gender, age, body mass index, gestational age and season, delivery mode, birth weight, feeding patterns, milk intake during pregnancy and for children, maternal and children's VD or calcium supplementation, and time of outdoor activity during pregnancy and for children, the  Table 2).
Then we conducted a stratified analysis on the association between 25(OH)D concentrations and the stake of low BMD (Table 3)

DISCUSSION
There had been lack of large epidemiological studies of VD situation in Chinese young children. Our study made up these shortcomings. As one of the biggest provinces of China, the state of VD deficiency of young children in Jiangsu Province allowed of no optimism. The study indicated that nearly 30% of 0-7 year old children had serum 25(OH)D concentrations under 50 nmol/L, while if we improved the definition standard to 75 nmol/L, almost 70% of whom had VD deficiency. The situation was more severe than that of Shanghai and Sichuan Province of China [4,22]. The median 25(OH) D concentrations in this study was 62.9 nmol/L, which was much lower than that of US population based on the National Health and Nutrition Examination 2000-2004 survey (for 1-5 year old children, the mean serum 25(OH) D concentrations was 76.5 nmol/L) [23].
As is known to us, sunlight exposure plays an important role in serum VD concentrations. Considering the limited sunlight exposure nowadays, dietary intake www.impactjournals.com/oncotarget of VD is necessary. It is suggested by the Endocrine Society's practice guidelines that children under 1 year old should take 400-1000 IU/d and then 600-1000 IU/d until 18 years old [16]. According to the newest suggestion of the Pediatric Branch of Chinese Medical Association, Chinese children are recommended to take 400-800 IU/d of VD after birth to 2 years old. And for premature infants, the supplements are 800-1000 IU/d after birth to 3 months then 400-800 IU/d until 2 years old. However, due to the misunderstanding of VD, many parents dared about the intoxication of extended taken of VD and did not follow the advice of doctors strictly. Early in the 1920s, surveys of VD toxicity appeared. In fact, these children who died had received nearly 20,000,000 units of VD, which were pharmacological doses and not within the physiological range [24]. Nowadays, it was confirmed by many researchers including Vieth et al. that supplement of 10,000 IU/d to 5 months and longer was safe and without adverse events [25]. While another research involving divers without sunlight exposure for 6 months showed that even 600 IU/d VD could not maintain the circulating serum 25(OH)D concentrations [24]. Unfortunately, it was not clearly that what dosages of VD were physiologic and what were pharmacologic. Because of the misunderstandings, fear of causing VD intoxication in children has continued up to now. That is why the medical scientists are reluctant to improve vitamin D requirements and the standard of VD deficiency.
For the serum 25(OH)D measurement, the liquid chromatography-tandem mass spectrometry (LC-MS/MS) method is generally considered to be the gold standard. However, the enzyme immunoassay was used in our study. Although the immunoassay may result in a little negative biases and misclassification of participants for vitamin  D sufficiency when compared with the LC-MS/MS assay [26], it is more readily available and require small volume sampling, ideal for testing [27]. For assessment of BMD, the dual energy X-ray absorptiometry is the best way. However, it has only been used in major hospitals for its expensive and limited usage. Whereas quantitative ultrasound is of low cost, convenient and lack of radiation, which is used wildly to reflect the BMD of young children especially in clinical routine health exams [3,4]. The study involving the relation between quantitative ultrasound BMD and serum 25(OH)D were relatively rare, especially in China. A study of small sample size of 203 children in Shanghai, China supported that BMD increased with the increase of the 25(OH)D concentrations, and suggested a threshold for the 25(OH)D concentrations of 20 ng/ml existed for low BMD [4]. Another study of 6,838 children in Sichuan province of China reported that VD concentrations might be correlated with BMD within a certain range (serum 25(OH)D concentrations between 50-75 nmol/L) [22]. The limitations of the above studies  were lack of comprehensive design with large scale and multi-center or without adjusting for confounding factors.
In this study, we conducted a multi-center case-control study, and taken into account the potential confounding factors. It was confirmed in our study along with the fact that low BMD of children was possibly to decrease when the concentrations was over 50 nmol/L. We

MATERIALS AND METHODS
The study was approved by the institutional review board of the Jiangsu Women and Children Health Hospital, the First Affiliated Hospital with Nanjing Medical University. The methods were carried out according to the approved guidelines and conformed to the Declaration of Helsinki. All the participants or the guardians signed informed consent.

Participants and study design
A nested case-control study was conducted on the basis of a study population of more than 50,000 children (0-7 year old) who were attending health check-up between April 2014 and March 2015 in 13 Children's Health Care Centers of Jiangsu Province (between 30 and 36 degrees north latitude), China. The children with metabolic bone disease or abnormal PTH level were excluded. Data from banked serum and routinely collected structured questionnaire were used. Serum samples were collected and archived according to the time of collection in boxes containing 100 samples, and stored at -80°C. The cases of low BMD were defined as children with speed of sound under the 20th percentile, while the controls of standard BMD were defined as being greater than or equal to the 20th percentile according to the common standard separated by gender and age [4]. The cases were randomly selected from the screening population by using a computerized random number function, with the specific box of the cases identified. Controls were then sourced by using the remaining samples in each box. The full box was analyzed by the trained laboratory technicians without knowing the all subjects' case or control status. As a result, 4,846 samples consented to participate in the study. After written informed consent was obtained, a face to face investigation was conducted by the prepared doctors with the structured questionnaire to collect related details including gender, age, weight, height, gestational age and season, delivery mode, birth weight, feeding patterns, milk intake during pregnancy and for children, VD or calcium supplementation during pregnancy and for children, time of outdoor activity during pregnancy and for children, and so on.

Bone ultrasound and serum 25(OH)D measurement
The BMD were determined by using the quantitative ultrasound BMD scanner (Sunlight Omnisense TM 7000, Israel). The trained doctors measured SOS at the left midtibia of children and repeated once again. Serum samples were thawed once, and serum 25(OH)D concentrations was determined by using an in vitro diagnostic enzyme immunoassay kit OCTEIA 25-Hydroxy Vitamin D (Immunodiagnostic Systems, Boldon, United Kingdom) according to the instructions. The variations between intraand inter-assay coefficients were less than 9.0%, and the reported analytic sensitivity of the immunoassay was 6.8-380 nmol/L. Commonly used cutoffs to define 25(OH)D status were assigned at 50 and 75 nmol/L [4,15,16]. More than 10% results were selected to repeat for quality control at random.

Statistical analysis
The 25(OH)D concentrations and BMD by the selected characteristics were described as mean ± SD, and differences were calculated by the Student's t-test or One-Way ANOVA. Differences of the low BMD prevalence by 25(OH)D concentrations and selected characteristics were calculated by χ 2 test. As the distribution of serum 25(OH) D concentrations was obviously skewed toward the left, we converted it to the Log10 [Lg25(OH)D] for the next analysis. The associations of 25(OH)D concentrations with BMD and stakes of low BMD were estimated by using multiple linear regression and logistic regression analyses, respectively, in which the key potential confounders were taken into account. The relation between 25(OH)D and BMD, and between 25(OH)D and stakes of low BMD were explored by a smoothing plot. The relation between 25(OH)D and BMD was analyzed by using Pearson's test. All the statistical analyses were performed with R software (version 2.13.0), and P ≤ 0.05 in a two-sided test was interpreted as statistically significant.