The prevalence and determinants of vitamin D deficiency in Indonesian infants at birth and six months of age

Background Vitamin D deficiency in infants has been associated with an increased risk of a number of diseases but there are limited data on the prevalence and determinants of vitamin D deficiency from tropical settings with high infant morbidity and mortality. Objective To determine the prevalence and determinants of vitamin D deficiency in infants at birth and at six months of age in Yogyakarta province, Indonesia. Design Serum vitamin D of eligible infants was measured in cord blood at birth and at six months of age. Factors associated with vitamin D deficiency (serum 25-hydroxyvitamin D <50 nmol/L) were collected prospectively monthly from birth and concentrations measured by liquid chromatography-tandem mass spectrometry. Independent risk factors were identified by multiple logistic regression. Results Between December 2015 to December 2017, 350 maternal-newborn participants were recruited and followed up. Vitamin D deficiency was detected in 90% (308/344) of cord blood samples and 13% (33/255) of venous blood samples at six months. Longer time outdoors (≥2 hours per day) and maternal multivitamin intake containing vitamin D during pregnancy were protective against vitamin D deficiency at birth (AOR: 0.10, 95% CI: 0.01–0.90 and AOR: 0.21, 95% CI: 0.06–0.68, respectively). Risk factors for vitamin D deficiency at six months included lower cumulative skin-sun exposure score (AOR: 1.12, 95% CI: 1.04–1.20), severe vitamin D deficiency at birth (AOR: 7.73, 95% CI: 1.20–49.60) and exclusive breastfeeding (AOR: 2.64, 95% CI: 1.07–6.49) until six months. Among exclusively breast fed (EBF) infants, a higher skin-sun exposure score was associated with reduced vitamin D deficiency risk. Conclusion In equatorial regions, the role of ‘safe’ morning sun exposure in infants and mothers in populations with medium to dark brown skin pigmentation and effective interventions to prevent vitamin D deficiency in newborns and EBF infants, need further consideration and evaluation.

Introduction E) and Kulon Progo regencies (7.83˚S, 110.2˚E), Yogyakarta province, Indonesia. Inclusion criteria included expected delivery to be in one of the study sites without family intention to leave the area within 12 months. Expectant mothers were approached in the third trimester to consent to the collection of 5 mL of cord blood at delivery and 2.5 mL of infant venous blood at six months of age at the Primary Healthcare Center or private clinic for vitamin D testing. Newborns were recruited at birth and followed monthly for 12 months after written consent had been obtained from both parents.
Soon after birth, mothers completed a structured questionnaire interview on their diet, time spent outdoors exposed to direct sunlight, the area of skin that was exposed to direct sunlight, and any sun protection used during pregnancy. Similar data on infant sun exposure and behavior as well as infant feeding practices were collected prospectively every month by faceto-face or phone interviews. The mother's and infant's skin types were determined at the first face-to-face visit for mothers and at the six-month visit for infants with the Fitzpatrick visual skin chart, a reliable instrument if spectophotometry is not available [19]. Skin type was measured in the area that was least exposed to sunlight, such as in the inner arms [19].

Study assays and measurement for vitamin D
We measured total serum 25-hydroxyvitamin D3 and 25-hydroxyvitamin D2 concentrations using liquid chromatography-tandem mass spectrometry (LC-MS/MS), the current gold standard for vitamin D measurement. The LC-MS/MS analysis was performed by Metabolomics Australia, at the University of Western Australia. The analytical sensitivity (or limit of detection) and the functional sensitivity (or limit of quantitation) of the assay for 25-hydroxyvitamin D3 is 0.5 nM and 2 nmol/L respectively. The center has been certified by the Center for Disease Control and Prevention for standardized vitamin D measurement (r 2 = 0.99) [20]. Vitamin D deficiency was defined as a concentration of 25-hydroxyvitamin D of less than 50 nmol/L. The vitamin D concentrations at birth and at six months of age were also categorized by severity using the categories suggested by Paxton et al.: mild deficiency (30-49 nmol/L); moderate deficiency (12.5-29 nmol/L) and severe deficiency (<12.5 nmol/L) [6].

Study definitions
Composite estimate UVR exposure and skin-sun exposure/protection score were measured in mother and infant. Composite estimate UVR exposure dose variables were created by multiplying the time spent outdoors and the ambient UVB for the appropriate month and presented as a monthly or cumulative value [21]. Skin-sun exposure score was defined as the frequency of uncovered/exposed arms and legs (1 = never, 2 = less than 50% of the time, 3 = more than 50% of the time, 4 = all of the time) [21]. The score for skin exposure was the summation of each category (each month score: 2 = minimum sun exposure, 8 = maximum sun exposure; and 6 month cumulative score: 12 = minimum sun exposure, 48 = maximum sun exposure) [21]. Exclusive breastfeeding (EBF) infants were defined as infants who received only breast milk until six months of age.

Data analysis
We entered de-identified data into REDCap database (Research Electronic Data Capture) and exported to STATA version 12 (Stata Corporation, College Station, Texas) for analysis. All variables were presented as mean ± SD (or median and interquartile ranges) and proportion (%), as appropriate. Correlation between cord blood and venous blood vitamin D was tested using the Pearson correlation test. Univariate and multivariate logistic regression was performed to explore the strength of the association between vitamin D status as a binary outcome and other categorical (i.e. monthly family income or exclusive breastfeeding status) or continuous variables (i.e. skin-sun exposure score or cumulative composite UV estimate) that served as risk factors for vitamin D deficiency at birth or at six months of age and are presented as odds ratios with 95% confidence intervals. Confounding was determined if the difference between the OR in the univariate and multivariate models changed by more than 10% after the inclusion of suspected confounding variables or if the variables have been commonly associated with the exposures or the outcomes in the existing literature. A p-value of <0.05 was considered statistically significant.

Results
We recruited 350 participants to the study. Of these, 344 (98%) participants had cord blood samples, 255 (73%) participants provided a venous blood sample at six months of age and 249 (71%) participants had both cord and venous blood samples (Fig 1). The demographic characteristics for rural, urban, and all participants are presented in Table 1. Mothers in rural Kulon Progo were less educated, had lower family income and lower EBF rates, but household size (7 people) was greater in urban Kota Yogyakarta (21% vs 6%).

Vitamin D status of Indonesian infants at birth and at six months of age
The prevalence of vitamin D deficiency was 90% in 344 participants at birth and 13% in 255 participants at six months (± 4 weeks) (Fig 2). The mean (± SD) of vitamin D concentrations at birth and six months were 30 (± 14) nmol/L and 77 (± 26) nmol/L respectively. There was a positive correlation between vitamin D concentration in cord blood and infants' venous blood at six months of age (Pearson's rho = 0.23, p-value = 0.03, Fig 3). Of those with vitamin D deficiency at birth, 87% (192/222) had normal vitamin D status by six months of age. Thirty-five percent (6/17) of participants who were severely deficient (<12.5 nmol /L) at birth had a vitamin D concentration below 50 nmol/L at six months of age (Table 2).
Seven percent of mothers (23/350) took a multivitamin containing vitamin D during pregnancy and 23% (78/350) of infants had formula feeding by the age of six months old (Table 3). Eighty four percent of the infants (280/334) sunbathed on a daily basis before 10 am with uncovered arms and legs before the age of six months.

Factors associated with vitamin D status at birth
A longer time spent outdoors exposed to direct sunlight during pregnancy was a protective factor against vitamin D deficiency at birth (Table 4). Mothers who spent two hours or more per day outside were less likely to have a newborn with vitamin D deficiency at birth compared to those who spent only 15 minutes or less in the sun. Taking into account the monthly ambient UVB, by multiplying the ambient UVB with the average time spent outdoor during pregnancy, a higher cumulative composite estimate UVR exposure for nine months was protective against vitamin D deficiency at birth. Maternal intake of a multivitamin containing vitamin D during pregnancy was independently associated with a reduced risk of vitamin D deficiency at birth.

Factors associated with vitamin D status at six months of age
Infants who were severely vitamin D deficient at birth were seven times more likely to be vitamin D deficient at six months of age (OR: 7.73, 95% CI 1.20-49.60) compared to those who were not deficient at birth (Table 5). EBF infants were twice as likely to have vitamin D deficiency at six months of age (OR: 2.64, 95% CI: 1.07-6.49) even after adjustment for sex, family income, skin type, and the month of blood collection. Lower skin-sun exposure score was an independent risk factor for vitamin D deficiency at six months of age (Table 5). In stratified analysis by EBF status, a higher magnitude of association between skin-sun exposure score and risk of vitamin D deficiency at six months of age was demonstrated only in EBF infants particularly on weekday exposure (AOR: 1.13,95% CI: 1.04-1.24, not shown in the table).

Discussion
This study found a high prevalence of vitamin D deficiency in Indonesian infants at birth that decreased markedly by six months of age. Increased maternal and postnatal infant sun exposure reduced the risk of vitamin D deficiency at birth and six months respectively. Other risk factors for vitamin D deficiency at six months were severe vitamin D deficiency at birth and EBF until six months of age. Among EBF infants, greater skin surface exposure to the sun was associated with a reduced risk of vitamin D deficiency.
High prevalence of vitamin D deficiency at birth has been previously documented with no clear pattern associated with latitude or ethnicity [22][23][24]. In Poland, New Zealand and the USA, the reported prevalence of vitamin D deficiency (<50 nmol/L) in cord blood was between 29% and 57%, particularly in low UV ambience seasons and dark-skinned people, while in Nigeria, the prevalence was 30% with an indoor lifestyle and maternal veils as the common risk factors [25][26][27][28]. We are not aware of previous studies that have evaluated vitamin D status in Indonesian newborns, though Yani et al. reported that of 168 Indonesian children aged less than five years old, one-third were vitamin D deficient [29]. Thai investigators recently reported a similar prevalence of cord blood vitamin D deficiency (89%) to our study but with a smaller sample size (N = 94) [8]. The limited data suggest that vitamin D deficiency in pregnant women may be common in South East Asian populations, as cord blood concentrations would reflect transplacental vitamin D supply [29,30]. As such, there are recommendations for pregnant women relating to sun exposure in Indonesia, i.e. continuous sun exposure between 10 am to 1 pm for 30 minutes (or 60 minutes for veiled women) per day [31]. The findings here support the recommendations that pregnant women have adequate sun exposure, as higher sun exposure reduced the risk of vitamin D deficiency at birth.
There is a complex interplay between multiple factors, such as genetics, environment, and lifestyle that influence the epidemiology of vitamin D deficiency [32]. Despite a very high prevalence of vitamin D deficiency at birth, only a minority of infants in our study had persistent vitamin D deficiency at six months of age. Routine cultural practices of Indonesian mothers exposing their uncovered newborns directly to the morning sun may have contributed to cutaneous vitamin D production [14,15]. Although we measured the infant skin-sun exposure score for consecutively for six months, using reflection on habitual sun behavior rather than temporal variations, our finding in favour of sun exposure needs careful interpretation as infant skin is vulnerable to sun damage [33,34]. Experts highlight the need for a balance between endogenous vitamin D production and the risk of skin malignancy with respect to direct sunlight [32,34,35]. There is no universal agreement as to the sun-exposure required to be equivalent to the recommended dietary intake of at least 400 IU of vitamin D daily. To reach a minimal erythema dose or a slight pinkness in the skin (10,000-25,000 IU), brown to dark-skinned infants need five to ten times the duration of sun exposure than fair-skinned infants who need approximately 30 minutes per week of sun exposure wearing only a diaper [13,34,35]. The optimum time for cutaneous synthesis of vitamin D is between 10 am to 3 pm when the UV index is high, but the risk of skin cancer is increased [32]. Most of our infants were sunbathed with uncovered arms and legs before 10 am and their vitamin D concentrations improved by six months. This practice is perhaps sufficient for typical Indonesian skin and the environment but needs further confirmatory studies. Future research should determine the optimum duration and daytime for safe sun exposure for Indonesian infants. Prenatal and postnatal vitamin D supplementation is not routinely recommended in our setting. Further, there are limited selections of food with a high content of vitamin D in the daily intake of Indonesian people [36]. Only 5% of our study infants had egg yolk in the first six months, mainly those who were living in the rural area, and even less had had fish or red meats. A minority (7%) of mothers reported taking a multivitamin supplement that contained vitamin D during pregnancy with 50% reporting a daily vitamin D intake below 400 IU, the minimum recommended daily dose in pregnancy [6]. None of the participating infants received vitamin D supplementation. Future determination and education on the optimum dose of prenatal and postnatal vitamin D intake or supplementation for the high-risk population in our setting may be warranted.

Vitamin D status of Indonesian infants
Our finding of an increased risk of vitamin D deficiency in EBF infants is consistent with previous studies [37,38]. Many studies from a range of settings have consistently reported that breast milk has poor vitamin D content [39,40]. Even settings where lifetime sunlight exposure was abundant or maternal vitamin D deficiency was uncommon, EBF infants only received 20% of their daily vitamin D requirement ie of 400 IU [39,41]. Given the many nutritional and health benefits of EBF in early infancy, especially in populations in resource-limited settings [42,43], the implications of breast milk being a poor source of vitamin D intake for infants requires careful consideration. Our findings indicate that higher postnatal skin-sun exposure protects against vitamin D deficiency at six months of age in this group. Thus, adequate postnatal sunbathing may be particularly important to prevent vitamin D deficiency among exclusively breastfed infants.
International organizations such as UNICEF, national guidelines such as NICE and vitamin D experts have formulated recommendations for EBF practices in relation to ensuring the potential health benefits of vitamin, including vitamin D supplementation up to 400 IU for EBF infants or breastfeeding mothers [6,30,44,45]. Experts agree that 50 nmol/L is the minimum 25-hydroxyvitamin D concentration for bone homeostasis. In addition to the prevention of rickets, there is some evidence that vitamin D protects against infections [2,46,47] and neurodevelopmental delay [48]. Vitamin D plays a direct role in regulating the expression of human antimicrobial peptides (cathelicidin and human beta-defensin 2) for pathogen elimination and reduction of viral replication [49]. However, findings from clinical studies are inconsistent and potentially biased by confounders with a lack of high-quality clinical trials, and so there is no consensus yet on minimum vitamin D requirements for "non-skeletal" benefits [1,45,50].
Some strategies have been proposed to prevent vitamin D deficiency in infants, but implementation is challenging. Indirect supplementation of breastfeeding women is impractical as the vitamin D content in supplemented breast milk does not always provide the infants with the amount currently recommended of 400 IU per day. To achieve this amount, high doses are often required of 4000-6400 IU per day with a potential increased risk of hypercalcemia and nephrocalcinosis [30,44,45,51]. Therefore, direct infant supplementation is now preferentially recommended [30,44,45,51]. In Indonesia, affordability is an issue as currently available  1 Skin-sun exposure score: the frequency of uncovered/exposed arms and legs (1 = never, 2 = less than 50% of the time, 3 = more than 50% of the time, 4 = all the time).
The score for skin exposure was the summation of each category (2 = min. sun exposure, 8 = max. sun exposure). 2 Cumulative composite estimate UVR exposure: the product of time spent outdoors multiplied by the ambient UVB during the period of time. 3 Adjusted for sex, socio-economic status, the month of birth and mother's skin types. 4 Adjusted for sex, socio-economic status, the month of birth, skin-sun exposure score, and mother's skin types. 5 Adjusted for sex, socio-economic status, the month of birth, time spent outdoor and mother's skin types. 6 Adjusted for sex, socio-economic status, the month of birth, mother's skin types, and cumulative composite estimate UVR exposure in weekdays. 7 Multivitamin containing vitamin D. https://doi.org/10.1371/journal.pone.0239603.t004

PLOS ONE
Vitamin D status of Indonesian infants Skin-sun exposure score: the frequency of uncovered/exposed arms and legs (1 = never, 2 = less than 50% of the time, 3 = more than 50% of the time, 4 = all the time).
The score for skin exposure was the summation of each category (2 = min. sun exposure, 8 = max. sun exposure), The cumulative score for skin exposure in 6 months was minimum 12 and maximum 48. Cumulative composite estimate UVR exposure: the product of time spent outdoors multiplied by the ambient UVB during the period of time.
n/a: not applicable. 1 adjusted for sex and family income, EBF, skin types, the month of birth and the month of blood collection. 2 adjusted for sex, family income, skin types and the month of blood collection.
formulations (400 IU/5 mL) of vitamin D for infants are expensive, not covered by universal health coverage and so are beyond the reach of most low-income families. Our findings indicate the role of 'safe' sun exposure for the mother in pregnancy, the infant postnatally, and particularly among EBF infants requires further consideration. These findings should be taken into account for recommendations aimed to avoid early life vitamin D deficiency.
The current study has several strengths. We performed gold standard vitamin D assay (LC-MS/MS) and measured a broad range of sun-related variables such as sun-phenotypes, sun behavior, environmental factors, infant feeding practice, and dietary intake in both mothers during pregnancy (retrospectively) and infants (for 12 months, prospectively). A prospective design allowed us to measure sun behavior cumulatively for six months prior to venous blood collection. The infant sun exposure measures used have previously been shown to related to vitamin D status [21]. The current study has some potential limitations. The measurement of sun exposure behavior was based on parental recall at four-weekly intervals that are not as accurate as a daily diary record for sun exposure time or by using a device such as a dosimeter to record the actual sun exposure time [13]. Maternal vitamin D status was not measured to examine the correlation with the concentration of cord blood vitamin D. Vitamin D deficiency has been reported in 20% of 140 Indonesian women in the first trimester of pregnancy but Indonesian data that correlate vitamin D status during pregnancy with cord blood are unavailable [48].

Conclusions
The high prevalence of vitamin D deficiency at birth in Indonesian infants had improved in the majority of infants by six months of age most likely due to the current cultural practice of infant sunbathing. Risk factors for vitamin D deficiency at six months of age were less infant skin area exposed to direct sunlight, severe vitamin D deficiency at birth, and EBF until six months of age. Among EBF infants, higher skin-sun exposure mitigated vitamin D deficiency risk. The role of 'safe' morning sun exposure in infants and mothers to prevent early life vitamin D deficiency in populations with medium to dark brown skin pigmentation requires further consideration. Effective interventions to improve vitamin D status in EBF infants that are locally appropriate need further consideration and evaluation.