Vitamin D (vD) is a prohormone that is synthesized in humans following skin exposure to ultraviolet B radiation in the range of 280–320 nm. In comparison to sunlight, diet provides less than 10% of the body’s vD requirement in unsupplemented individuals1,2. Vitamin D is necessary to maintain calcium homeostasis and bone health, and there are increasing reports of its role in innate and autoimmune functions3. Vitamin D deficiency is detrimental to the health of mothers and children because of increased risk of osteomalacia in adults and rickets and delayed growth in infants and children4,5,6. In addition, vD deficiency or low vD intake has been associated with increased risk of autoimmune diseases in adults and children7,8 and lower respiratory tract infection in children9,10,11,12.
Recently, vD deficiency has been reported to be a public health problem worldwide despite abundant sunshine in many countries and the demonstration of the efficacy of vD supplements to prevent vD deficiency7. A review of recent studies suggests that vD deficiency is a global problem during pregnancy13. The prevalence of serum 25(OH)D levels <50 nmol/L, considered vD deficiency14, ranges between 33% in the USA15 and 42% in Canada16 to 75%–77% in the UK17 and Finland18 and from 74% to 98% in India19, New Zealand20 and the United Arab Emirates21. The high prevalence of vD deficiency in pregnancy raises concern about increased risk of low vD status in mothers and infants after birth and especially the segments that breastfeed, because of the low vD content of breast milk. Furthermore, in view of the drive to increase the prevalence and duration of exclusive breastfeeding and the reported increased risk of rickets in breastfeeding infants5, the vD status of lactating women and their infants should be of global health concern. The reported few studies appear to indicate that vitamin D deficiency is common in breastfeeding infants, and rickets may represent the tip of the iceberg. Some authors have suggested that vD deficiency may be an under-diagnosed public health problem in breastfeeding mothers and their infants in many countries5,9, but standardized comparisons of global prevalence are lacking. For example, the cutoff values for vD deficiency, age of study, season of study and methods of assessment of risk factors varied among the studies9. The objectives of this study were to compare the prevalence and risk factors for vD deficiency in a cohort of breastfeeding mothers and infants in Shanghai, China, Cincinnati, Ohio, and Mexico City, Mexico, using the same study design and serum vD level measurement in a single center. The mothers and infant dyads were enrolled in the longitudinal Global Exploration of Human Milk study, which was designed to explore the effects of different environments and diets on human milk composition, infant nutrition and health.
Table 1 shows the comparison of the maternal baseline characteristics by site. Cincinnati mothers were older (p < 0.001), included mothers with higher prevalence of obesity (p = 0.0001), four-year college education (p = 0.0001) and vD supplementation rate (p = 0.0001). The mean sun index was negligible in Shanghai (10.0) and Mexican mothers (0.87) compared with Cincinnati mothers (239.0), p < 0.001.
At four weeks postpartum, maternal mean serum 25(OH)D concentrations differed (p < 0.001) by site (Figure 1). The mean values of 48.6 nmol/L in Shanghai and 48.2 nmol/L in Mexico City were lower than the value of 70.2 nmol/L in Cincinnati. Vitamin D deficiency (serum 25(OH)D <50 nmol/L) in order of magnitude was found in 62%, 52% and 17% of Mexican, Shanghai and Cincinnati mothers, respectively (p < 0.001).
Figure 2 displays the maternal serum 25(OH)D concentrations and sun indices by sites and season. The sun index varies by season among Cincinnati mothers, but there was lack of seasonal variation among Shanghai and Mexican mothers. Maternal vD status as measured by serum 25(OH)D concentrations showed seasonal variation, with the highest values among Cincinnati mothers within each season.
Comparing infants seen during the fall/winter seasons at 26 weeks of age at the three sites, the mean serum 25(OH)D concentrations were lower in Mexico City (44.0 nmol/L) and Cincinnati (68.3 nmol/L) than Shanghai (95.3 nmol/L), p < 0.001 (Figure 3). Vitamin D deficiency in order of magnitude was found in 62%, 28% and 6% of Mexican, Cincinnati and Shanghai infants, respectively (p < 0.001). Sunlight exposures during the fall/winter seasons were different by sites and were higher in Shanghai infants at 26 weeks than in Cincinnati and Mexican infants. The duration of sun exposure in hours per week (median, IQR) were 3.0 (1.1, 4.4) vs. 2.0 (0.5, 5.3) vs. 1.0 (0.3, 1.7) in Shanghai, Cincinnati and Mexican infants, respectively. The median (IQR) percent of BSAs exposed were 26 (14, 26) vs. 14 (7, 26) vs. 19 (14, 24) in Shanghai, Cincinnati and Mexican infants, respectively. Median (IQR) sun index (BSA . duration of sun exposure) values were 55 (22,104) vs. 27 (3.8, 65) vs. 21 (7.1, 29.1) in Shanghai, Cincinnati and Mexican infants, respectively. In spring/summer seasons, the infant sun index was higher in the Cincinnati cohort followed by Shanghai and lowest in the Mexico City cohort, but we did not have enough blood for infant serum 25(OH)D measurement in the Shanghai and Mexico City cohorts. Serum 25(OH)D in infants are associated with their sun index, including all infants measured in the fall/winter season (r = 0.39, p < 0.001) across the three sites. This is also true for infants in the Cincinnati site in the fall/winter season (r = 0.37, p = 0.01), but not for the other two sites independently.
Categories of vD Deficiency in Mothers and Infants
The different categories of low vD status, moderate (serum 25(OH)D 30 to <50 nmol/L) and severe (serum 25(OH)D <30 nmol/L) deficiency, in mothers and infants differed by site and are shown in Table 2. Overall, less than 4% of the mothers were severely deficient, and the percent of mothers with severe deficiency was over two-fold higher in Mexican mothers than Shanghai and Cincinnati mothers. About half of the Shanghai and Mexican mothers had moderate deficiency compared with 14% of Cincinnati mothers. Overall, 12.5% of the infants evaluated during the fall/winter season at 26 weeks of age were severely deficient. In contrast to maternal findings, the percent of infants with severe and moderate deficiencies were lower in Shanghai compared with Cincinnati and Mexico City cohorts.
In this prospective study of urban population in North America, Latin America and China using the same study design, we found vD deficiency to be common in breastfeeding mothers from Shanghai and Mexico City and less common in mothers in Cincinnati. The mean serum 25(OH)D concentration in lactating mothers in Mexico City (48.2 nmol/L) and in Shanghai (48.6 nmol/L) was lower than the recommended target value of >50 nmol/L22. At four weeks postpartum, over 60% of the mothers in Mexico City and half of the mothers in Shanghai were vD deficient ()serum 25(OH)D concentration <50 nmol/L) compared with 17% of mothers in Cincinnati. Most of the vD deficient mothers had moderate vD deficiency (serum 25(OH)D levels of 30 to <50 nmol/L) while 2% of the mothers in Shanghai and Cincinnati and 6.2% of Mexican mothers had severe deficiency (serum 25(OH)D levels <30 nmol/L), which is associated with increased risk of osteomalacia. As expected, there was an inverse relationship between 25(OH)D levels and serum PTH levels, indicating an inadequate vD status in the mothers, which is associated with elevated levels of serum PTH. The degree of severe vD deficiency is significantly lower than previously reported from other countries. Based on previous studies, 61% of breastfeeding mothers from the United Arab Emirates23, 48% of mothers in India24 and 46% of breastfeeding mothers from Turkey25 had severe deficiency (serum 25(OH)D concentration <25 nmol/L), which is associated with increased risk of osteomalacia22. Such a low vD status would also predispose breastfeeding infants without vD supplementation to vD deficiency. It thus appears that moderate to severe vD deficiency in early postpartum in breastfeeding mothers is a common problem in many countries and may be under recognized.
In this study, the higher vD status in mothers from Cincinnati was associated with higher sun index and vD supplementation intake compared with mothers in Shanghai and Mexico City. It is of note that the restricted exposure to sunlight during postpartum among Shanghai and Mexican mothers is related to cultural practices in which mothers are restricted from outdoor activities and are cared for by family members. This practice of restricting mothers from outdoor exposure in the immediate postpartum convalescent period (known as “doing the month”) in China has been associated with increased risk of low vD status and rickets in some rural Chinese communities26. Recognition of the possible impact of these cultural practices on vD nutrition in postpartum mothers should heighten attention to the need for vD supplementation in such settings. In all of the cohorts from the three sites, vD intake, obesity and season were independent predictors of maternal vD status in multivariate regression analysis. In individual site analysis, high maternal sun index (>500) was associated with high maternal vD status only in the Cincinnati cohort, which attained such a high index. Therefore, differences in the prevalence of these potential risk factors among populations will impact maternal global prevalence, as well as the degree of maternal vD deficiency. For example, in the study from the United Arab Emirates, where severe vD deficiency is more common than in this study, it was found that mothers were more severely sunshine deprived during lactation and had a lower rate of vD supplement intake23,27.
The vitamin D status of breastfeeding infants who were seen at 26 weeks of age during the fall and winter seasons were lower in the Mexican infants (mean serum 25(OH)D 44 nmol/L) and Cincinnati infants (68.3 nmol/L) than in Shanghai infants (95.3 nmol/L). Vitamin D deficiency, defined as serum 25(OH)D <50 nmol/L, was ten-fold higher in Mexican infants and almost five-fold higher in Cincinnati infants than in Shanghai (62% vs. 28% vs. 6%, respectively). In addition, severe deficiency (serum 25(OH)D <30 nmol/L) was over two-fold higher in Mexican than in Shanghai and Cincinnati infants. This degree of severe deficiency could theoretically predispose Mexican infants to increased risk of rickets 23,22.