Epidemiology and risk factors of hypovitaminosis D in a cohort of internationally adopted children: a retrospective study

Vitamin D (VD) is an essential nutrient with hormone-like activity, which regulates calcium and bone metabolism. Moreover, VD has been recently associated with immune status, autoimmune disorders, infectious diseases, decreased response to corticosteroids and psychiatric diseases [1‐5].

These associations highlighted the need for a definition of VD optimal values, and this is particularly true for the pediatric population, because of the importance of VD for the bone health and the immune status during the tricky phase represented by the growth period.

Serum 25-hydroxy-vitamin D (25(OH)D) concentration is currently the marker of choice, as it is the major circulating form of VD and its value reflects both the amount produced in the skin after sun exposure and that assumed with food [6].

Most studies consider values < 10 ng/ml or < 20 ng/ml as deficiency and values > 20 ng/ml or >  30 ng/ml as sufficiency, with a gray zone between 10 and 20 ng/ml and 20–30 ng/ml, respectively, indicated as insufficiency [7‐11]. When 25(OH)D serum concentration falls below 4 ng/ml, especially if associated with hypocalcaemia, hypophosphatemia and increased serum alkaline phosphatase (ALP), the risk of rickets and osteomalacia in children rises exponentially [12].

Relatively high rates of subclinical vitamin D deficiency (VDD) have been reported in otherwise healthy infants, children and adolescents in several studies, especially from low-income countries [13‐26].

Limited data on the prevalence of hypovitaminosis D among internationally adopted children, a population which could be considered “at risk” for geographical origin and a housing system not always subdued to an optimal surveillance, are available, and there is no study assessing parathyroid hormone (PTH) levels in these groups. The relationship between serum concentrations of 25(OH)D and PTH has been considered as a possible way to define VD status, since PTH increases when vitamin D and, consequently, calcium decrease. Studies performed in adults, and very few studies in pediatric subjects, have showed conflicting results about this inverse relationship [27].

Therefore, it would be of utmost importance to define VD levels under which PTH begins to increase, and if it could be considered a marker of bone health or immune status.

We hypothesized that the prevalence of biochemical rickets and hyperparathyroidism in a population of internationally adopted children was higher than in other children subpopulations. Hence, we evaluated the prevalence of hypovitaminosis D in internationally adopted children. Moreover, we tested if some of the population characteristics, such as age, BMI, sex, housing solution before adoption, season and month of observation, country of origin, could be considered risk factors for hypovitaminosis D in this particular cohort, aiming to find a model applicable to all our patients.

In order to define the prevalence of biochemical rickets, we considered the relationship existing between Vitamin D levels and serum calcium, serum phosphates and serum alkaline phosphatase levels. We also tried to define the relationship between VD status and PTH levels in internationally adopted children.

Characteristics of the population, season at blood drawing and mean values of the analytes investigated (25(OH)D, PTH, ALP, Ca, P) are resumed in Table 1. Detailed informations about post-hoc analyses are included in Additional file 1.

Migrant children are a population of great interest nowadays, and their health problems—infectious ones, but also not-infectious ones—are a source of interesting discussions for pediatricians all over the world. Internationally adopted children are a controlled sub-population of the migrant one, thus it could offer an insight on a larger population – though with a lot of differences regarding travelling, housing solutions and dietary habits and deprivations.

Vitamin D is an important pro-hormone, classically associated with bone health, but recently found to be involved in immune system and neuro-psychological disorders, and cardiovascular risk. It can be seen how a D hypovitaminosis could put at risk the entire well-being of a child [1‐5].

Our study, carried on for nine years, tried to test some of the well-known predictors for the Vitamin D deficiency in the internationally adopted children population and the relationships existing between vitamin D and other factors implied in the bone health, such as parathyroid hormone, alkaline phosphatase, free calcium and phosphates.

From January 2007 to December 2015, 29,866 children arrived in Italy through an international adoption, from Eastern Europe countries above all, followed by Latin America, Asia and Africa [30]. The cited report offers only details about 2014 and 2015, but it specifies that the percentages reported are similar to those previously observed. During this period, adopted male children were more than the female ones (58.3% vs 41.7%; ratio = 1.4:1), with a mean age at the arrival of 5.9 years (SD not specified).

The population we observed had a similar composition to that of the entire “internationally adopted children” Italian population: male to female ratio was 1.3:1, with a prevalence of children coming from Eastern Europe countries and a mean age at the arrival of 5.31 years (SD ± 2.92 years).

We did not find a statistical difference of mean 25(OH)D values between sexes. However, a slight statistical significance was found with regards to the Vitamin D status, with a slight higher percentage of females affected by VDD than males. Female children are, as a matter of fact, more interested by severe VDD and mild VDD than male children, while male children seem to be more affected by moderate VDD than females (Table 2 and Additional file 1).

A reason for this evidence could be a higher BMI in female children than males, a justification often found by other studies, highlighting how a higher BMI is directly related to the risk of developing VDD, as VD is a lipophilic vitamin and a higher fat mass increase its distribution volume [31‐35].

Our results, though, are opposite to those showed by other studies in which VDD was found to be more prevalent among males than in females, or it showed no significant relationship with sex at all [11, 36‐39].

In our population, prevalence of VDD is significantly related to the age at the arrival in Italy and to the age at first evaluation. Figure 1 shows the mean age per group of Vitamin D status, highlighting that in our population a higher mean age is related to a worst Vitamin D status.

The negative association of age and serum 25(OH)D in children has already been described in literature. According with Mansbach et al., this might be because of the difference in duration of sun exposure, while Gustafson et al. sustain that it is caused by the rapid growth spurt, or “catch-up growth” the children reach out in the first year after adoption [7, 40].

Catch-up growth is a process similar to the rapid growth experienced by infants and toddlers, when use of VD supplement is customary for the prevention of rickets [41]. Therefore, during this period internationally adopted need larger amounts of dietary or supplemental iron and VD than those a child of the same age would normally need.

Interestingly, in our population high 25(OH)D concentrations were observed in children aged < 1 year regardless of the area of origin (Fig. 1). This is most likely attributable to the use of fortified formula milk during the period spent in an institution and by pediatricians’ increasing awareness of the hypovitaminosis D issue once they arrive in Italy. As reported by other studies, this relation might lead to nullify the expected effects of an adoptee’s birth country on VD level [35, 40].

Our study showed a strong association between skin color and Vitamin D Status, that only remained significant for the value “olive-brown skin” after the multivariate analysis (logistic regression). Our data showed that children with olive to brown skin had a statistically significant 2-fold higher risk of having hypovitaminosis D than children with fair skin (OR 2.387, Table 8). These results agree with those reported by an Italian study by Franchi et al. evaluating 1.374 children living in the north of Italy. The study observed that ethnicity was a strong predictor of 25(OH)D levels, reporting a high prevalence of hypovitaminosis D in African, North African, and Indian children living in Italy (74.8, 81.2, 89.7, and 76.0%, respectively) [25]. Other studies reported the same results. A severe 25(OH)D deficiency associated to a proximal myopathy was observed in recent immigrants from Palestine, Pakistan and India to Northern Europe develop due to the limited effect of sunshine and a low dietary VD intake [37]. Hintzpeter [35] showed that Turkish and Arab-Islamic participants of both sexes as well as Asian and African girls were found to be at high risk for VDD, with the highest OR for African girls (7.8) [35]. Similarly, data from NHANES III showed that adolescents of African origin had an increased risk of VDD (OR 8.6) compared with white adolescents [15].

These consistent observations can lead us to conclude that a spontaneous recovery of VD status is unlikely in an adopted child arrived in Italy with VDD, independently from their ethnicity.

In our study macro area of origin was significantly associated with VD status (p < 0.001). Basing our observation on percentages alone, VDD seems to have been more common in Africa than in Latin America or Europe and Russian Federation. If statistically significant, our results would differ from those showed by Chiappini et al., whose study highlighted the highest prevalence of VDD in children coming from the Russian federation [26]. However, post hoc analyses highlighted that the real statistically significant differences are only between children coming from Asia and children coming from Africa. See Additional file 1 and the “results” section for a more detailed description of this difference. Thus, it is not possible to conclude that VDD is more common in children coming from Africa than in children coming from Latin America or Europe and Russian Federation.

As previously said with regards to skin color, it is thought that occasional exposure to sunlight provides most of the VD requirement of the human population [42]. However, skin synthesis of VD may not compensate for the low nutritional intake in Europe, because of the relationship occurring between its latitude and the rake angle of solar radiations. As a matter of fact, Webb et al. [43] demonstrated that the photosynthesis of pre-VD is nearly impossible during winter months at latitudes similar to those encasing Europe. A high prevalence of hypovitaminosis D combined with high PTH levels and lower bone mineral density has recently been reported in Finnish girls and Italian children and adolescents, suggesting that Italian children are at risk for this condition, as well as other European populations [44‐46]. Moreover, another Italian study by Vierucci et al., showed that no VD is produced in central Italy during late fall, winter and the beginning of spring because of sun exposure [38, 39].

The prevalence of VDD, even at southern latitudes, was remarkable, supporting other reports suggesting that living in sunny areas of the world does not necessarily protect children from VDD [39, 43, 47‐50].

Our work highlighted that the risk of developing hypovitaminosis D was significantly higher in children coming from foster families and orphanages, compared to those coming from foster families. This relationship was still statistically significant after multivariate analysis.

It is a well-demonstrated knownledge that children in foster care experience high rates of child abuse, emotional deprivation, and physical neglect than children living with their families [51‐53]. Moreover, it is plausible to think that children in orphanages likely do not spend much time outdoors and that they lack adequate sun exposure. Another reason for the association with VDD could be that as children grow up in institutional care, they shift from a diet of vitamin D–fortified formula milk to cooked food, which may not be fortified with vitamin D.

According to other studies, we observed how the season of blood sampling could be considered a predictor of the Vitamin D Status [26, 54]. As a matter of fact, it is only natural to find out that a pro-hormone which synthesis is related to sun-exposure has higher levels during summer than during winter. It has been shown that even minimal exposure of skin to sunlight, is adequate to reach VD sufficiency in US and Central Europe, and a number of studies demonstrated that both serum concentrations of 25(OH)D and other biomarkers of bone health status increase during summer only to reduce during winter, while PTH behavior is the exact opposite [55‐63].

Our data highlighted how drawing a blood sample during summer was a protective factor against hypovitaminosis D, when compared to spring (Fig. 1). OR for summer shows a reduction of the risk of approximately the 80%.

In our study, almost half the children had normal Ca, P, ALP and PTH levels. This data differs from other pediatric studies, which show a percentage way lower [23, 64]. Normal serum calcium works as an inhibitory signal for parathyroids, stopping PTH secretion, as described by Steingrimsdottir and Aggarwal [65, 66].

Our study differs from other studies on the same matter because of the interest taken in showing the relationship existing between vitamin D and parathyroid hormone. Only a very small percentage of our population was affected by hyperparathyroidism, but a significant data was that we could observe elevated PTH levels in children having 25(OH)D levels over 30 ng/mL and normal calcium, a finding deserving further studies: was it a primary hypersecretion or a signal of an altered bone metabolism?

We also tried to identify the PTH inflection point for this population. With this term, other studies indicated the mean 25(OH)D value at which PTH begins to increase. We did not succeed, probably because of the small number of children with an increased PTH included in our population and the extreme variability of values at which PTH was found to increase.

Biochemical rickets affected a significant percentage of our study population. Surprisingly, even children with only mild VDD were affected by it (Table 8). It is though that rickets is confined to children exposed to extreme undernourishment but, as a matter of fact, it can be also observed in western countries, because of the growing epidemia of childhood obesity and the previously discussed relationship of high body mass index and fat mass with VDD [67].

The study indicates a widespread global VD insufficiency compared with proposed threshold levels. This high prevalence implies that higher doses than currently recommended would be needed to reach country specific “desirable” levels.

Our results reinforce the importance of an early evaluation of serum 25(OH)D levels in internationally adopted children to start the adequate strategy of VD supplementation/treatment, avoiding further deterioration of VD status, particularly in older, dark skin children arrived in Italy during winter and spring period.