Vitamin D levels may be related to sunshine duration, and a standardized method for vitamin D supplementation could be beneficial for obese children and adolescents.

Vitamin D is essential for bone health in the growing skeleton. Vitamin D deficiency and/or low calcium intake can lead to nutritional rickets with defective mineralisation [1]. Moreover, vitamin D may play a role in the development of mental, immunological or cardiometabolic diseases [2]. The standard laboratory parameter for vitamin D supply is 25 hydroxyvitamin D (25-OH-D₃) [3].

Numerous cross-sectional studies have documented lower 25-OH-D₃ levels in children with obesity compared to normal weight control groups and hypothesised a negative effect on cardio-metabolic parameters [4, 5]. Although the bioavailability of total 25-OH-D₃ is reduced, free 25-OH-D₃ seems to be comparable in lean and overweight children [6, 7]. Levels of 25-OH-D₃ normalise with weight loss [8]. This supports the hypothesis that body composition and vitamin D storage in fat tissue are the main reasons for low 25-OH-D₃ levels in patients with obesity [9]. Large randomised controlled intervention studies in adults and smaller studies in children have shown no effect of vitamin D supplementation on weight, cardiovascular risk factors, or metabolic parameters [10 –13]. In children with 25-OH-D₃ levels above 50 nmol/l, vitamin supplementation does not improve bone health [14 –16].

Consequently, the current standard practice in the care of normal-weight children in Germany is to refrain from vitamin D testing and supplementation beyond infancy [17].

However, the long-term effects of persistently low 25-OH-D₃ levels on the bone health of children with obesity are still largely unknown, and guidelines differ on the recommendations for optimum 25-OH-D₃ serum levels and vitamin D supplementation [18]. A cross-sectional study found an increased fracture risk to be associated with 25-OH-D₃ levels below 50 nmol/l, which might justify vitamin D supplementation [19]. However, even in these cases, supplementing obese children with vitamin D might not always be advisable. Other studies suggest that it leads to increased cholesterol and triglyceride levels in adolescents without changing their BMI [20], warranting special caution as to possible adverse effects of vitamin D supplementation on metabolic comorbidities in patients with obesity. This leaves open questions for the paediatrician regarding when, and for whom, to start vitamin D supplementation in routine care for children with obesity.

For the outpatient program of the paediatric obesity clinic at Charité - Universitätsmedizin Berlin, we aimed to establish a pragmatic approach focusing narrowly on the effect of vitamin D on bone mineralisation. Calcipenic disorder of bone mineralisation is usually detected by low levels of 25-OH-D₃ in conjunction with elevated levels of parathyroid hormone (PTH) above 65 ng/l (normal range: 15,0–65,0 ng/l) and elevated alkaline phosphatase (AP) [21]. The goal of the present study is to evaluate follow-up data of unsupplemented children with obesity. We had the question about whether PTH levels remain in the normal range over time in children without vitamin D supplementation or whether a negative trend in 25-OH-D₃ and elevation PTH were apparent, suggesting a structurally deficient endogenous vitamin D production and prompting vitamin D supplementation to minimise the risk for development of rickets or osteomalacia.

The longitudinal analysis of vitamin D levels is considerably complicated by the presence of seasonal effects. Cutaneous synthesis induced by solar UV-B radiation of 290–315 nm is the principal source of vitamin D for humans [22]. Geographical and seasonal changes in UV radiation lead to a physiological variability of serum 25-OH-D₃ levels [23]. This seasonal (and geographical) effect on observed 25-OH-D₃ levels is a major confounder in vitamin D studies [24]. A recent analysis demonstrates that even at the same date and location, sunshine duration, and therewith endogenous vitamin D production, are highly variable between years [25]. 25-OH-D₃ levels are particularly volatile in spring and autumn. As a novel approach, we use meteorological data on the duration of sunshine at the time of vitamin D measurements to account for seasonal confounding. This contrasts with the standard approach to merely account for the month the measurements were taken that ignores differences present between years.

Our study adds to the scarce long-term follow-up data of 25-OH-D₃ levels in combination with PTH in children with obesity. To our knowledge, it is also the first study of children to incorporate actual sunshine duration to account for seasonal variation. In so doing, we hope to contribute to establishing an informed standard of care regarding vitamin D supplementation for children with obesity.

We included 543 patients with one 25-OH-D₃ measurement. Of these 543 patients, 229 patients had two, 115 patients had three, and 96 patients had four serial 25-OH-D₃ measurements. Table 1 gives details of the patients with up to four repeated 25-OH-D₃ measurements.

The interval between the first and the last measurement adds up to 120 weeks (2 ½ years). The mean age of the patients lies in the adolescent range (12 ± 3 years). Males and females are evenly distributed in the group. The mean BMI SDS and percentage of children with migration backgrounds decreased from the first to the fourth measurement of 25-OH-D₃.

At time T1, more than 50% of the patients, showed an unadjusted 25-OH-D₃ level below 50 nmol/l, and 20% showed levels below 30 nmol/l. These proportions increased to 66% and 33%, respectively, at the fourth measurement (T4). By contrast, the 25-OH-D₃ values adjusted for sun 30, BMI SDS, and sex remained stable over time (Fig. 1). The percentage of children with elevated PTH above the cut-off of 65 ng/dl did not rise between T1 and T4. The longitudinal PTH levels adjusted for 25-OH-D₃, sun 30, and BMI SDS remained stable over 2 ½ years (Fig. 2).

Influencing factors on 25-OH-D₃ levels are depicted in Table 2. We found a positive association of Sun 30, high triglycerides, high LDL and high total cholesterol with 25-OH-D₃ levels, respectively. Male sex and impaired fasting glucose had a negative association.

Sun 30 (ß = − 0.43; p = 0.03) and 25-OH-D₃ level (ß= − 0.26; p < 0.01) had a negative association with PTH levels.

In this study, we retrospectively evaluated longitudinal 25-OH-D₃ levels in a large cohort of children with obesity without vitamin D supplementation. We use a novel method to adjust 25-OH-D₃ data for sunshine duration. We show that the mean adjusted 25-OH-D₃ and PTH levels remain stable over 120 weeks (2 ½ years).

Based on this data we established the following standard of care for children with overweight in our obesity clinic:

Measurement of 25-oh-D levels only in combination with PTH, AP, serum calcium and phosphate. No vitamin D supplementation for low 25-OH-D3 levels if PTH and AP are normal. Instead, we advice regarding adequate calcium intake and recommend regular outdoor activity.

We prescribe vitamin D supplementation only if PTH is increased above 65 ng/dl (normal range: 15,0–65,0 ng/l) and AP is elevated. For low 25-OH-D₃ levels without elevated PTH or AP, we will recommend regular outdoor activity aiming to increase endogenous vitamin D production through sun exposure and to promote physical exercise as an important stimulus for bone remodelling [32]. These recommendations may have an additional positive impact on weight development and avoid the drawbacks of liberal vitamin D supplementation [33 –35]. Patients at special risk of vitamin D deficiency e.g. with dark or largely covered skin will receive an individualised follow-up on 25-OH-D₃ levels.

According to this approach 5% of our cohort with elevated PTH required treatment for biochemical signs of osteomalacia. In this study we used the upper limit of normal of 65 ng/l for PTH. Values above this cut off are observed in rickets and osteomalacia [36, 37]. Thresholds such as 50ng/l or 45 ng/l have been proposed for PTH in children [38 –40]. Low calcium intake elevates PTH and complicates the quest for a PTH threshold. Two studies analysed longitudinal 25-OH-D₃ and PTH levels in 32 German children [8] and 67 Spanish children [41] with obesity over 12 months. In both studies, 25-OH-D₃ levels remained stable, which is consistent with our findings. In the Spanish cohort, the number of children with 25-OH-D₃ levels below 20 ng/ml (50 nmol/l) increased from 32 to 41% over 12 months. Though the mean 25-OH-D₃ levels were higher in the Spanish group, the children showed higher mean PTH levels compared to our data [42]. The authors found no seasonal variance in PTH compared to lean children and argued that obesity might be a confounding factor. A causative relation between low 25-OH-D₃ and elevation of PTH in patients with obesity is still a matter of discussion. Several studies show that PTH suppression thresholds vary with body weight and may not reflect the actual vitamin D status [43]. From a pathophysiological view, raised PTH and low 25-OH-D₃ levels indicate the need for additional calcium or vitamin D intake to improve bone mineralisation. Our data show an inverse association of 25-OH-D₃ and PTH, accordingly. Longitudinal data in children with obesity with low 25-OH-D₃ and documented calcium intake are necessary to define PTH thresholds that require treatment with vitamin D and/or calcium.

Values for AP, serum calcium and phosphate were not available for analysis in our patients. There was no difference in AP, serum calcium and phosphate in the group of German children with obesity after 12 months [8]. Likewise, serum calcium and phosphate remained stable in the Spanish group but AP was not determined [41].

Interestingly, sunshine duration or other parameters of radiation are not included in the evaluation of 25-OH-D3 levels even in recent large association or treatment studies both in adults and children [44]. One recent study used day- and place specific UV radiation data from meteorological satellites to analyze the influence of radiation on 25-OH-D3 levels in a large population of healthy children in Europe. This study found that UV radiation and outdoor time were the most important determinants.

of 25-OH-D3 levels [45].

The existing studies in children with obesity re-evaluated the children after exactly one year to minimise the impact of seasonal variance [8, 41]. In children with obesity, the relevance of a seasonal influence is occasionally questioned as they are suspected to spend little time outdoors and to have reduced sun exposure [42]. Yet, according to our data, even the short periods of time these children spend outside notably influence their 25-OH-D₃ levels. This is in line with the recommendation that a daily outdoor activity of 15 min between 10:00 am and 3:00 pm is sufficient to promote endogenous vitamin D synthesis from April to September in Germany (geographic location between 48Æ and 54Æ Northern latitude) and may correspond to the way from school or running an errand in the afternoon [46, 47]. Future research should focus on the role and the use of sunshine duration or similar parameters of sun radiation in the interpretation of 25-OH-D₃ levels in clinical practice.

As to the influence of different metabolic parameters on the 25-OH-D₃ concentration, fasting insulin and fasting glucose do not seem to have a high explanatory potential. By contrast, elevated triglyceride and total cholesterol levels were strongly associated with higher 25-OH-D₃ levels. Given the nature of vitamin D as being fat-soluble and its association with lipid metabolism [10], this finding may not be surprising.

Although the effects of vitamin D supplementation on lipid levels in adults and children remain contradictory [5, 13, 48], previous studies suggest that it leads to increased cholesterol and triglyceride levels in adolescents [20]. These results should advise caution to the adverse effects of vitamin D supplementation on metabolic comorbidities in patients with obesity. All the more so as higher vitamin D doses are needed to achieve a normalisation of 25-OH-D₃ levels in overweight patients and daily and weekly dosing regimens differ [49]. A further drawback of liberal vitamin D supplementation is vitamin D toxicity. Cases of vitamin D intoxication are mainly reported in infants with mis-dosage of soluble vitamin D resulting in severe hypercalcemia, failure to thrive and nephrocalcinosis [34]. Possible harmful effects of chronic moderate vitamin D dosing are discussed [34]. High vitamin D doses might be required for the treatment of osteomalacia or rickets and should be accompanied by monitoring for urinary calcium excretion and nephrocalcinosis [33].

There are several limitations of our study. The data were analysed retrospectively. The statistical analysis involved multiple comparisons and have exploratory character. There is a probability that vitamin D intake was underreported in our patients since vitamin D is an easily available over-the-counter drug. Patients with multiple 25-OH-D₃ measurements adhered to the multimodal treatment in the obesity clinic. This might explain differences in BMI-SDS and the percentage of children with migration backgrounds between the time points. For patients that dropped out of treatment, no repeated measurements were available. Skin pigmentation and eating habits of different cultures have an impact on 25-OH-D₃ levels [50]. No data on ethnicity or skin pigmentation was available for analysis. Further data is needed to assess the role of ethnicity in children with obesity. One-third of our patients stem from parents with a migration background. This includes a high proportion of families of Turkish descent [51]. Vacation time in the Mediterranean area could have possibly elevated 25-OH-D₃ levels in this cohort. Although a cross-sectional analysis of 25-OH-D₃ levels in Turkish school children suggests that a stay in southern latitudes does not automatically optimise 25-OH-D₃ levels [52]. Pubertal status, AP, serum calcium and phosphate levels, daily outdoor activity, calcium intake or imaging were not available for analysis in this study. All these parameters could provide more detailed information on the effect of sun radiation and 25-OH-D₃ levels on bone mineralisation and structure. No longitudinal data exist on the bone mineral density (BMD) of vitamin D-deficient children or adolescents with obesity. Studies examining the effect of 25-OH-D₃ levels on BMD with Dual-energy X-ray absorptiometry (DXA) or peripheral quantitative CT (pQCT) report contradictory findings. While they do not show an association of 25-OH-D₃ levels with BMD before puberty a higher level of 25-OH-D₃ is associated with higher BMD in early adulthood [53]. Further studies that combine all these parameters in a longitudinal analysis are necessary to assess bone health in obese children with low 25-OH-D₃ levels. Especially the prospective analysis of the group of patients with low 25-OH-D₃ levels and elevated PTH and AP are important to define risk factors and understand the impact on their bone health. We hope to provide further data in the future.

We propose including parameters of sun radiation in all future studies concerning vitamin D in children with obesity to avoid the over-interpretation of physiologic fluctuations in 25-OH-D₃ levels. Sunshine duration in hours is a parameter that is easily obtainable from local weather stations and can be included in the interpretation.

We suggest the pragmatic strategy of limiting vitamin D supplementations to patients with biochemical signs of mineralisation disorders such as elevated PTH and AP. Starting vitamin D supplementation based on a single 25-OH-D3 value without the regard of season, radiation or AP and PTH level might be instead harmful and costly [35]. Instead we educate patients with low 25-OH-D₃ and normal PTH and AP levels on the necessity of regular outdoor activity and a calcium-rich diet. Further studies on the bone health of obese children with low 25-OH-D₃ and elevated PTH and AP are needed to evaluate this approach.