The role of maternal vitamin D supplementation in the prevention of infantile rickets is unknown, particularly in low- and middle-income countries without routine infant vitamin D supplementation. Through secondary analysis of a randomized, placebo-controlled trial in Bangladesh, we examined the dose-ranging effects of maternal vitamin D supplementation on the risk of biochemical rickets at 6 to 12 months of age.
Pregnant women (n = 1300) were randomized into 5 groups: placebo, or vitamin D 4200 IU/week, 16 800 IU/week, or 28 000 IU/week from second trimester to delivery and placebo until 6 months postpartum; or 28 000 IU/week prenatally and until 6 months postpartum. Infants underwent biochemical rickets screening from 6 to 12 months of age (n = 790). Relative risks (RR) and 95% confidence intervals (95% CI) of biochemical rickets were estimated for each group versus placebo.
Overall, 39/790 (4.9%) infants had biochemical rickets. Prevalence was highest in the placebo group (7.8%), and the risk was significantly lower among infants whose mothers received combined prenatal and postpartum vitamin D at 28 000 IU/week (1.3%; RR, 0.16; 95% CI, 0.03–0.72). Risks among infants whose mothers received only prenatal supplementation (4200 IU, 16 800 IU, 28 000 IU weekly) were not significantly different from placebo: 3.8% (RR, 0.48; 95% CI, 0.19–1.22), 5.8% (RR, 0.74; 95% CI, 0.33–1.69), and 5.7% (RR, 0.73; 95% CI, 0.32–1.65), respectively.
Maternal vitamin D supplementation (28 000 IU/week) during the third trimester of pregnancy until 6 months postpartum reduced the risk of infantile biochemical rickets. Further research is needed to define optimal postpartum supplementation dosing during lactation.
Maternal vitamin D supplementation during pregnancy and lactation modifies infant vitamin D status, but its effects on the risk of infantile rickets have not previously been established.
High-dose maternal vitamin D supplementation during the third trimester of pregnancy and up to 6-months postpartum reduced the risk of infantile rickets in Bangladesh. Maternal postpartum vitamin D supplementation may be an alternative to direct infant supplementation for rickets prevention.
Nutritional rickets is one of the most common causes of pediatric bone disease globally.1 Biochemical abnormalities are detectable at an early stage of rickets across all age groups and have an important role in screening and diagnosis.2 –6 Young infants with rickets often have a more subtle bony phenotype compared with older children given their lack of substantial weight bearing and may remain undiagnosed until later stages of the disease. However, the high metabolic demand for calcium resulting from rapid growth in infancy can lead to acute presentations of rickets with hypocalcemia before the emergence of other clinical or radiologic signs.7 –9 Compared with older children, there may be substantial morbidity associated with infantile rickets given sequelae such as hypocalcemic seizures and, in rare cases, cardiomyopathy.10 –12
Vitamin D deficiency is the predominant cause of nutritional rickets worldwide, particularly in infants. Maternal prenatal vitamin D status is the primary determinant of newborn vitamin D status.13 –17 The major circulating metabolite of vitamin D, 25-hydroxyvitamin D (25(OH)D), crosses the placenta such that cord blood concentrations are highly correlated with maternal values at term.18 However, the influence of maternal prenatal vitamin D status on infant vitamin D stores diminishes by 2 months of age and infants become dependent on other vitamin D sources.19 In the Maternal Vitamin D for Infant Growth (MDIG) trial,20 there was a dose-response effect of prenatal vitamin D supplementation on cord blood and infant vitamin D blood concentrations up to 3 months of age, as has been observed in other prenatal vitamin D supplementation trials.21 ,22 Therefore, although deficiency in the early postnatal period may be caused primarily by maternal prenatal vitamin D deficiency,23 vitamin D deficiency later in infancy is attributable to other risk factors. Because breast milk is a poor source of vitamin D if a lactating mother has inadequate vitamin D intake/status, prolonged breast feeding without vitamin D supplementation is an important cause of vitamin D deficiency in infants. However, adequate maternal intake of vitamin D during lactation can support vitamin D sufficiency in the breastfed infant.24 For example, the MDIG trial demonstrated that continued maternal postpartum supplementation (28 000 IU/week) maintained infant 25(OH)D concentrations at or above 30 nmol/L up to 6 months of age.20
The role of vitamin D in fetal calcium homeostasis is uncertain; whereas animal studies suggest transplacental transfer may be independent of prenatal maternal vitamin D status, some human studies have provided evidence that maternal prenatal vitamin D status affects fetal calcium accrual.25 Immediately after delivery, vitamin D is required as an essential regulator of infant intestinal calcium absorption and bone mineral metabolism, similar to older children.26 Therefore, it is plausible that maternal vitamin D supplementation in the prenatal and postpartum period would reduce the risk of infantile rickets by supporting fetal calcium accrual, neonatal vitamin D endowment, and infant vitamin D intake via breastmilk.
Although there is limited evidence establishing the effect of postpartum vitamin D supplementation in breastfeeding women on the risk of infantile rickets,27 ,28 there have not been published trials examining prenatal supplementation alone or in combination with postpartum supplementation. Such evidence would be particularly relevant to many low- and middle-income countries such as Bangladesh, where there is a high burden of vitamin D deficiency among both women of child-bearing age and newborns and vitamin D supplementation in infants is not a routine practice.17 ,29 –31 In this substudy of a randomized controlled trial, we aimed to estimate the effect of a range of doses of maternal vitamin D supplementation during pregnancy and continued supplementation during lactation, compared with placebo, on the risk of infantile biochemical rickets at 6 to 12 months of age in Dhaka, Bangladesh.
Methods
Study Design
This study was based on secondary analyses of data from the MDIG trial, conducted in Dhaka, Bangladesh, from 2014 to 2018. This was a randomized double-blinded, placebo-controlled, dose-ranging trial of maternal vitamin D supplementation (from mid-gestation up to 6 months postpartum) for which the primary outcome was infant growth.20 ,32 Briefly, 1300 generally healthy females 18 years of age or older were enrolled in the second trimester of pregnancy and randomized into 1 of 5 intervention groups: (1) placebo in prenatal and postpartum; (2) prenatal vitamin D3 (4200 IU/week) and placebo postpartum; (3) prenatal vitamin D3 (16 800 IU/week) and placebo postpartum; (4) prenatal vitamin D3 (28 000 IU/week) and placebo postpartum; or (5) vitamin D3 (28 000 IU/week) prenatal and to 6 months postpartum. Supplementation was administered weekly under direct supervision by trained study personnel either in the participant’s home or in the clinic. Participants in all groups were provided daily calcium (500 mg) and iron–folic acid supplements. Ethics approval for secondary use of the trial data for this sub-study was provided by the Research Ethics Board at the Hospital for Sick Children in Canada (REB #1000061259).
Participants
Individuals were excluded from the MDIG if there was history of medical conditions with altered vitamin D metabolism and/or hypercalcemia, were having a high-risk pregnancy, were unwilling to stop taking nonstudy vitamin D or calcium supplements or multivitamins containing calcium and/or vitamin D, or were currently being prescribed vitamin D supplements as part of a physician’s treatment plan for vitamin D deficiency. Infants in the MDIG cohort were eligible for biochemical screening at or after 6 months of age; those included in this substudy had at least 1 measurement of serum alkaline phosphatase (ALP) between 6 and 12 months of age (Supplemental Fig 2). Infants with known disorders that affect calcium homeostasis or known skeletal dysplasia would have been excluded from the study, yet no such cases were identified.
Rickets Screening Procedures
Infants in the MDIG were born between June 2014 and February 2016. Systematic screening for rickets at 6-month follow-up visits was launched in May 2016. The biochemical screening panel included serum concentrations of ALP, calcium, and phosphate. Any of the initial parameters found to be outside of established reference ranges prompted a physician referral for assessment and treatment, facilitation of radiographs of wrists and/or knees and an extended laboratory panel (including parathyroid hormone [PTH] and 25(OH)D) that were managed according to the treating physician.
Laboratory Methods and Radiographic Interpretations
Infant serum calcium, phosphate, and ALP concentrations were measured using quantitative colorimetric assays (Beckman Coulter OSR60117, OSR6122, and OSR6104) at the Clinical Biochemistry Laboratory in Dhaka (icddr,b). Serum 25(OH)D concentrations were measured at the Analytical Facility for Bioactive Molecules (AFBM) in Toronto using high-performance liquid chromatography-tandem mass spectrometry, as previously described.33 Infant intact PTH concentrations were quantified using a sandwich enzyme-linked immunosorbent assay kit (Immunotopic 60-3100) at AFBM. Clinical management by physicians in Dhaka was informed by local radiologist interpretations of wrist and/or knee radiographs, where available. However, if possible, wrist and/or knee radiographs obtained from children who screened positive for biochemical rickets were further reviewed using a standardized approach by a pediatric radiologist who was blinded to the clinical and laboratory data, as previously described.20
Case Definition
Biochemical rickets is marked by an elevated ALP level, which is indicative of increased bone turnover; this is a nearly universal feature of rickets and usually the earliest biochemical abnormality.34 A common compensatory response to hypocalcemia is an elevation in PTH, which promotes the mobilization of calcium from bones. The development of hypocalcemia and hypophosphatemia may occur as the disease progresses or in the presence of an inadequate PTH response.35 ,36 However, there are no standardized cutoff points for these biochemical markers that define onset or stages of progression of rickets. Age-specific reference ranges must be used for these biochemical markers; ALP in particular is highly dependent on age and rate of bone growth. Here, we defined “biochemical rickets” as (1) ALP 450 U/L or (2) ALP 350 U/L plus at least 1 of the following: calcium 2.2 mmol/L or phosphate 1.6 mmol/L or PTH 6.9 pmol/L. The cutoffs for this definition were consensus-based among investigators. This definition used for analytical purposes differed slightly from the definition used to prompt clinical referral during the MDIG study because PTH was not available in real time as part of the initial screening panel.
Left skewing of ALP was noted with a higher-than-expected proportion of low values; of 790 infants in this substudy, 132 (17%) had ALP <90 U/L. These low values were distributed throughout the study period. Following an extensive review, no preanalytical factors were identified that might have artifactually lowered ALP. The distribution of other biochemical markers analyzed in the same samples were similarly distributed in the low ALP and non–low ALP groups (data not shown), ruling out overdilution as an explanation. Hypercalcemia was not observed in the infants with low ALP, making hereditary hypophosphatasia less likely. Malnutrition is known to decrease ALP production,37 although we did not find differences in anthropometric parameters (weight for age z-score and height for age z-score at 6 months of age) between the low ALP and non–low ALP groups (data not shown). A set of serum samples (n = 244) from infants in the MDIG across a wider age range than included in this study was tested at the AFBM laboratory at The Hospital for Sick Children using a different colorimetric assay (Alkaline Phosphatase Colorimetric Assay Kit; ab83369); 8.2% (20/244) were found to have ALP <90 IU/L compared with a frequency of 12% among all samples tested at the Clinical Biochemistry Laboratory (135/1085), suggesting that the high proportion of low values in this cohort was a reproducible finding.
Statistical Analysis
Participant characteristics and biomarker concentrations were expressed as mean ± SD, median (25th and 75th percentiles), or frequencies and percentages. PTH was log-transformed because of right-skewing. Participant demographics across the 5 maternal vitamin D treatment arms were compared using analysis of variance for normally distributed continuous variables, Kruskal-Wallis for nonnormally distributed continuous variables, and χ-squared tests for categorical variables. To estimate the relative risk (RR) of infantile rickets in each prenatal and postnatal maternal vitamin D supplementation group, versus placebo, we used a modified Poisson regression with robust error variance.38 Planned subgroup analyses included unadjusted regression models stratified by child sex, maternal vitamin D status at randomization (25(OH)D ≥30 nmol/L vs <30 nmol/L), and gestational age (term 37 weeks), respectively. All point estimates were presented with 95% confidence intervals (95% CI) and P values (α < 0.05 considered statistically significant). Data were analyzed using Stata version 16.1 (StataCorp 2019).
Results
Maternal and Infant Characteristics
Characteristics of participants included in this substudy were similar across the 5 intervention groups (Table 1), as previously reported for the MDIG trial.20
Demographics and Characteristics of Participants, Stratified by Vitamin D Treatment Group
. | Treatment Groupa . | |||||
---|---|---|---|---|---|---|
. | Placebo n = 165 . | 4200; 0 IU/wk n = 160 . | 16 800; 0 IU/wk n = 154 . | 28 000; 0 IU/wk n = 157 . | 28 000; 28 000 IU/wk n = 154 . | Pb . |
Maternal age (y), median (min, max) | 22 (18, 38) | 23 (18, 35) | 22 (18, 35) | 23 (18, 38) | 23 (18, 38) | 0.46 |
Maternal level of education, no. (%) | ||||||
No education | 5 (3%) | 8 (5%) | 9 (6%) | 8 (5%) | 6 (4%) | 0.94 |
Primary school incomplete | 40 (24%) | 32 (20%) | 30 (20%) | 34 (22%) | 38 (25%) | |
Primary school complete | 26 (16%) | 21 (13%) | 17 (11%) | 22 (14%) | 25 (16%) | |
Previous live births, median (min, max) | 61 (37%) | 58 (36%) | 65 (42%) | 57 (36%) | 54 (35%) | |
Secondary school incomplete | 1 (0, 5) | 1 (0, 4) | 1 (0, 4) | 1 (0, 4) | 1 (0, 3) | 0.49 |
Secondary school complete | 33 (20%) | 41 (26%) | 33 (21%) | 36 (23%) | 31 (20%) | |
Asset Index Quintile, no. (%) | ||||||
First | 35 (21%) | 34 (21%) | 23 (15%) | 36 (23%) | 33 (21%) | 0.51 |
Second | 27 (16%) | 35 (22%) | 43 (28%) | 30 (19%) | 26 (17%) | |
Third | 39 (24%) | 25 (16%) | 26 (17%) | 28 (18%) | 29 (19%) | |
Fourth | 36 (22%) | 30 (19%) | 32 (21%) | 34 (22%) | 37 (24%) | |
Fifth | 28 (17%) | 35 (22%) | 30 (19%) | 29 (18%) | 29 (19%) | |
Maternal baseline 25(OH)D (nmol/L), mean (SD) | 27.1 (13.5) | 27.1 (13.5) | 27.3 (13.9) | 27.7 (14.8) | 26.4 (13.8) | 0.95 |
Month of birth, no. (%) | ||||||
March–May | 21 (13%) | 19 (12%) | 20 (13%) | 14 (9%) | 22 (14%) | 0.65 |
June–August | 39 (23%) | 40 (25%) | 31 (20%) | 44 (28%) | 34 (22%) | |
September–November | 64 (39%) | 51 (32%) | 62 (40%) | 53 (34%) | 48 (31%) | |
December–February | 41 (25%) | 50 (31%) | 41 (27%) | 46 (29%) | 50 (33%) | |
Gestational age at birth (wk), median (25th, 75th percentiles) | 39.1 (38.3, 40.1) | 39.1 (38.3, 40.1) | 39.1 (38.0, 40.0) | 39.4 (38.1, 40.3) | 39.2 (38.6, 40.0) | 0.68 |
Infant sex, no. (%) | ||||||
Female | 88 (53%) | 71 (44%) | 75 (49%) | 78 (50%) | 67 (44%) | 0.38 |
Male | 77 (47%) | 89 (56%) | 79 (51%) | 79 (50%) | 87 (56%) | |
Infant birth WAZ, mean (SD)c | −1.13 (0.82) | −1.29 (0.92) | −1.17 (0.91) | −1.36 (0.84) | −1.14 (0.93) | 0.19 |
Infant birth LAZ, mean (SD)d | −0.80 (1.0) | −0.91 (1.1) | −0.92 (1.1) | −1.0 (0.94) | −0.83 (0.95) | 0.54 |
Infant vitamin D use, no. (%)e | 29 (17%) | 13 (8%) | 22 (14%) | 24 (15%) | 24 (16%) | 0.15 |
Duration of infant vitamin D use (weeks), median (25th, 75th percentiles)f | 1 (1, 1) | 2 (1, 3) | 1 (1, 2) | 1 (1, 2) | 1 (1, 2) | 0.15 |
. | Treatment Groupa . | |||||
---|---|---|---|---|---|---|
. | Placebo n = 165 . | 4200; 0 IU/wk n = 160 . | 16 800; 0 IU/wk n = 154 . | 28 000; 0 IU/wk n = 157 . | 28 000; 28 000 IU/wk n = 154 . | Pb . |
Maternal age (y), median (min, max) | 22 (18, 38) | 23 (18, 35) | 22 (18, 35) | 23 (18, 38) | 23 (18, 38) | 0.46 |
Maternal level of education, no. (%) | ||||||
No education | 5 (3%) | 8 (5%) | 9 (6%) | 8 (5%) | 6 (4%) | 0.94 |
Primary school incomplete | 40 (24%) | 32 (20%) | 30 (20%) | 34 (22%) | 38 (25%) | |
Primary school complete | 26 (16%) | 21 (13%) | 17 (11%) | 22 (14%) | 25 (16%) | |
Previous live births, median (min, max) | 61 (37%) | 58 (36%) | 65 (42%) | 57 (36%) | 54 (35%) | |
Secondary school incomplete | 1 (0, 5) | 1 (0, 4) | 1 (0, 4) | 1 (0, 4) | 1 (0, 3) | 0.49 |
Secondary school complete | 33 (20%) | 41 (26%) | 33 (21%) | 36 (23%) | 31 (20%) | |
Asset Index Quintile, no. (%) | ||||||
First | 35 (21%) | 34 (21%) | 23 (15%) | 36 (23%) | 33 (21%) | 0.51 |
Second | 27 (16%) | 35 (22%) | 43 (28%) | 30 (19%) | 26 (17%) | |
Third | 39 (24%) | 25 (16%) | 26 (17%) | 28 (18%) | 29 (19%) | |
Fourth | 36 (22%) | 30 (19%) | 32 (21%) | 34 (22%) | 37 (24%) | |
Fifth | 28 (17%) | 35 (22%) | 30 (19%) | 29 (18%) | 29 (19%) | |
Maternal baseline 25(OH)D (nmol/L), mean (SD) | 27.1 (13.5) | 27.1 (13.5) | 27.3 (13.9) | 27.7 (14.8) | 26.4 (13.8) | 0.95 |
Month of birth, no. (%) | ||||||
March–May | 21 (13%) | 19 (12%) | 20 (13%) | 14 (9%) | 22 (14%) | 0.65 |
June–August | 39 (23%) | 40 (25%) | 31 (20%) | 44 (28%) | 34 (22%) | |
September–November | 64 (39%) | 51 (32%) | 62 (40%) | 53 (34%) | 48 (31%) | |
December–February | 41 (25%) | 50 (31%) | 41 (27%) | 46 (29%) | 50 (33%) | |
Gestational age at birth (wk), median (25th, 75th percentiles) | 39.1 (38.3, 40.1) | 39.1 (38.3, 40.1) | 39.1 (38.0, 40.0) | 39.4 (38.1, 40.3) | 39.2 (38.6, 40.0) | 0.68 |
Infant sex, no. (%) | ||||||
Female | 88 (53%) | 71 (44%) | 75 (49%) | 78 (50%) | 67 (44%) | 0.38 |
Male | 77 (47%) | 89 (56%) | 79 (51%) | 79 (50%) | 87 (56%) | |
Infant birth WAZ, mean (SD)c | −1.13 (0.82) | −1.29 (0.92) | −1.17 (0.91) | −1.36 (0.84) | −1.14 (0.93) | 0.19 |
Infant birth LAZ, mean (SD)d | −0.80 (1.0) | −0.91 (1.1) | −0.92 (1.1) | −1.0 (0.94) | −0.83 (0.95) | 0.54 |
Infant vitamin D use, no. (%)e | 29 (17%) | 13 (8%) | 22 (14%) | 24 (15%) | 24 (16%) | 0.15 |
Duration of infant vitamin D use (weeks), median (25th, 75th percentiles)f | 1 (1, 1) | 2 (1, 3) | 1 (1, 2) | 1 (1, 2) | 1 (1, 2) | 0.15 |
LAZ, length for age z-score; WAZ, weight for age z-score
Maternal prenatal vitamin D supplementation (second trimester to delivery); postnatal maternal supplementation (0–6 mo).
p value for Kruskal Wallis, Pearson χ2, or analysis of variance test.
Based on Intergrowth-21 growth standards, by gestational age, within first 48 h of life, n = 566.
Based on Intergrowth-21st growth standards, by gestational age, within first 48 h of life, n = 550.
Ever consumed a vitamin/supplement containing or possibly containing vitamin D from birth to 1 y.
Number of weeks a supplement containing or possibly containing vitamin D was consumed among infants with at least 1 wk of reported consumption from birth to 6 mo of age, median (interquartile range).
Effect of Maternal Vitamin D Supplementation on Biochemical Rickets
A total of 39 cases of biochemical rickets were identified among 790 infants who underwent biochemical screening. Of these 39 cases, 10 met the criteria based on ALP ≥450 U/L alone, 12 had ALP ≥350 U/L and phosphate ≤1.6 mmol/L as the only abnormalities, 14 had ALP ≥350 U/L and intact PTH ≥6.9 pmol/L as the only abnormalities, and 3 had more than 2 abnormalities.
The highest prevalence of rickets (7.9%) was found in the placebo group (Table 2). The lowest prevalence (1.3%) was in the high-dose supplementation group in which mothers received 28 000 IU prenatally and up to 6 months postpartum; this corresponded to a significantly reduced risk of infantile biochemical rickets compared with placebo (Table 2). High-dose vitamin D during the prenatal period alone (4200 IU/week, 16 800 IU/week, and 28 000 IU/week) did not have a significant effect on the risk of rickets, although there were fewer rickets cases identified in each of these groups compared with placebo (Table 2).
RR of Rickets in Each Treatment Arm Compared With Placebo
Treatment Group (Prenatal; Postpartum IU/week) . | n (Total 790) . | Proportion of Infants With Biochemical Rickets, No. (%) . | RRa (95% CI) . | P value . |
---|---|---|---|---|
Placebo | 165 | 13 (7.9) | ||
4200; 0 | 160 | 6 (3.8) | 0.48 (0.19–1.22) | 0.12 |
16 800; 0 | 154 | 9 (5.8) | 0.74 (0.33–1.69) | 0.48 |
28 000; 0 | 157 | 9 (5.7) | 0.73 (0.32–1.65) | 0.45 |
28 000; 28 000 | 154 | 2 (1.3) | 0.16 (0.04–0.72) | 0.02 |
Treatment Group (Prenatal; Postpartum IU/week) . | n (Total 790) . | Proportion of Infants With Biochemical Rickets, No. (%) . | RRa (95% CI) . | P value . |
---|---|---|---|---|
Placebo | 165 | 13 (7.9) | ||
4200; 0 | 160 | 6 (3.8) | 0.48 (0.19–1.22) | 0.12 |
16 800; 0 | 154 | 9 (5.8) | 0.74 (0.33–1.69) | 0.48 |
28 000; 0 | 157 | 9 (5.7) | 0.73 (0.32–1.65) | 0.45 |
28 000; 28 000 | 154 | 2 (1.3) | 0.16 (0.04–0.72) | 0.02 |
RR, relative risk.
Poisson regression model with robust error variance used to obtain RR.
Subgroup Analyses
In an analysis restricted to infants born to women with baseline 25(OH)D <30 nmol/L during the second trimester of pregnancy (n = 507), inferences were unchanged (Fig 1). Inferences also remained the same in stratified analysis by sex (males or females), albeit more male than female infants were affected by rickets overall. Inferences remained the same when analysis was restricted to infants born at term (37 weeks’ gestation) (Supplemental Tables 3–5).
The relative risk of biochemical rickets among varying doses of maternal prenatal and postpartum vitamin D supplementation compared with placebo using modified Poisson regression (blue bars). Subgroup analysis assessing the effect of maternal vitamin D supplementation on infantile rickets among women with vitamin D deficiency (25(OH)D <30 nmol) at baseline (n = 507). The circles represent the effect estimates, with 95% confidence interval (CI) bars.
The relative risk of biochemical rickets among varying doses of maternal prenatal and postpartum vitamin D supplementation compared with placebo using modified Poisson regression (blue bars). Subgroup analysis assessing the effect of maternal vitamin D supplementation on infantile rickets among women with vitamin D deficiency (25(OH)D <30 nmol) at baseline (n = 507). The circles represent the effect estimates, with 95% confidence interval (CI) bars.
Infant Bone Biomarkers
Serum calcium concentrations were highest in the combined supplementation group and lowest in the placebo group; however, these differences were not statistically significant (Supplemental Fig 4). Phosphate concentrations were significantly higher and ALP concentrations were significantly lower in the combined supplementation group compared with placebo (Supplemental Fig 4).
Radiographically Confirmed Rickets
Of the 39 infants with biochemical rickets, 16 had radiographs of the wrist and/or knee available for review by the SickKids radiologist, of whom 4 were found to have radiographic findings of rickets, as previously reported.20 Three of the 4 infants were in the placebo group, and the fourth was in the group administered 4200 IU/week prenatally. Mean ALP was higher at presentation for these infants, at 705 U/L, compared with mean 439 U/L for the other infants with biochemical rickets. All 4 infants were hypophosphatemic (serum phosphate <1.56 mmol/), and 1 was hypocalcemic (serum calcium <2.1 mmol/L). Radiographs were not available for all infants with biochemical rickets. In large part, this was because infants who met criteria of ALP 350 U/L and PTH 6.9 pmol/L were not flagged for imaging because PTH was not available in real time as part of the initial screening panels.
Discussion
Combined prenatal and postpartum maternal supplementation (28 000; 28 000 IU/week) decreased the risk of biochemical rickets compared with placebo among infants 6 to 12 months of age. However, maternal prenatal supplementation alone at any dose, without postpartum continuation, did not significantly decrease the risk of biochemical rickets. Prenatal maternal vitamin D supplementation influences early postnatal infant 25(OH)D, but postpartum continuation was required to maintain 25(OH)D ≥30 nmol/L up to 6 months of age, as previously reported in the MDIG trial (Supplemental Fig 3).20 Therefore, the present findings strongly support the hypothesis that vitamin D deficiency (marked by inadequate circulating 25(OH)D), is an important cause of biochemical rickets in this infant population. As previously reported, all the cases of radiographically confirmed rickets were in the placebo and lowest-dose prenatal supplementation (4200 IU weekly prenatally) groups, further supporting the potential role of vitamin D in rickets prevention. However, we cannot rule out other causes of rickets in this setting; moreover, most infants with 25(OH)D <30 nmol/L did not have biochemical rickets, indicating that other contributing factors act in concert with vitamin D deficiency.
There were relatively more male infants affected by biochemical rickets in our study. It has been speculated that rickets may occur more frequently in boys because of greater linear bone growth and increased skeletal demands during times of rapid growth. Although not seen consistently, this phenomenon has been noted in several studies evaluating rickets in infancy.39 –41 The present findings are consistent with evidence from 2 smaller randomized trials in India that previously found that there were fewer cases of biochemical rickets among infants of mothers who received postpartum supplementation.27 ,28 Although it has been well established that infant 25(OH)D status can be influenced by maternal supplementation during lactation, the dose-response relationship remains uncertain.24 ,42 Human milk is considered a poor source of vitamin D3 unless the lactating woman has high amount of vitamin D intake.43 The transfer of the vitamin D parent compound (vitamin D3) is favored over 25(OH)D in the mammary gland, suggesting that the vitamin D concentration of breast milk is primarily affected by maternal vitamin D intake or cutaneous synthesis rather than maternal vitamin D status (ie, circulating 25(OH)D).44 ,45 This distinction is important because the short half-life of vitamin D3 (12–24 hours) implies that an analogous dose of vitamin D is consumed by the infant soon after the corresponding maternal ingestion.46 However, low daily doses of maternal vitamin D supplementation may not achieve sufficiently high circulating levels of vitamin D in breast milk to impact infant 25(OH)D, even if they prevent maternal vitamin D deficiency.47 High-dose maternal supplementation, often greater than the Institute of Medicine–recommended upper limit of 4000 IU/day,48 has been previously shown to have similar effects on breastfeeding infant 25(OH)D as daily infant vitamin D supplementation.42 ,49 ,50 Further research involving direct comparison of various doses, including daily maternal dosing compared with intermittent weekly or bolus dosing regimens, is required to determine the minimum effective maternal postpartum dose to maintain 25(OH)D sufficiency in infants and in turn minimize the risk of rickets.
A strength of this study is that the randomized, dose-ranging, placebo-controlled design of the MDIG trial and the lack of routine infant supplementation permitted causal inferences regarding the effects of maternal vitamin D supplementation on the risk of biochemical rickets. However, several limitations of the study should be acknowledged. This is a substudy of a previous trial; the mother and infant pairs included were selected from the existing MDIG cohort based on data availability, which may have compromised the generalizability of the findings. Although the participants in this substudy were similar to the remainder of the MDIG cohort, it is possible that this cohort was not fully representative of the mothers and infants in the MDIG trial or of the general population in Dhaka. The biochemical case definition was useful for identifying early disease because infants with rickets may present without skeletal abnormalities; however, we lacked complete radiographic information for all the infants who met biochemical rickets criteria, and the longer term clinical significance of infantile biochemical rickets is uncertain. Because the diagnosis of biochemical rickets was based on cross-sectional biochemical evaluation starting at 6 months of age, we were unable to determine the precise age of onset of the abnormalities. Furthermore, a greater number of infants screened late in infancy or at older ages might have enabled us to describe the natural history of this process in the absence of routine supplementation or vitamin D treatment of those who screened positive in early infancy.
Conclusions
High-dose maternal postpartum vitamin D supplementation may serve as a viable public health strategy for rickets prevention by effectively increasing infant 25(OH)D status in conjunction with efforts to promote breastfeeding. Other low- and middle-income countries in South Asia that have similar burdens of maternal and infant vitamin D deficiency and do not have vitamin D supplementation programs could benefit from this strategy. Future studies should include comparisons of different doses of maternal postpartum supplementation and longer term follow-up including radiologic assessments and clinical outcomes.
Acknowledgments
We thank Huma Qamar of The Global Centre for Child Health, The Hospital for Sick Children, for her assistance with data organization and Talia Wolfe, former summer student at The Global Centre for Child Health, The Hospital for Sick Children, for her work on the initial data analysis.
Dr Roth is the principal investigator, conceptualized, designed, and supervised the study, drafted the initial manuscript, and critically reviewed and revised the manuscript; Dr Lautatzis designed the study, performed statistical analysis, drafted the initial manuscript, and critically reviewed and revised the manuscript; Dr Al Mahmud supervised data collection and field study activities in Dhaka and critically reviewed and revised the manuscript; Drs Ahmed and Keya contributed to local implementation of the study and data collection in Dhaka, and critically reviewed and revised the manuscript; Ms Tariq contributed to study design, performed statistical analysis, drafted the initial manuscript, and critically reviewed and revised the manuscript; Dr Harrington, Dr Zlotkin, Dr Lam, and Dr Morris contributed to study design, and critically reviewed and revised the manuscript; Dr Stimec provided expert review of radiographic data and critically reviewed and revised the manuscript; and all authors read and approved the final manuscript and agree to be accountable for all aspects of the work. The authors report no conflicts of interest or financial relationships relevant to this article to disclose.
Clinical Trial Registration: This trial has been registered at www.clinicaltrials.gov (identifier NCT01924013).
FUNDING: This work was supported in part by the Bill & Melinda Gates Foundation (OPP1066764). Under the grant conditions of the Foundation, a Creative Commons Attribution 4.0 Generic License has already been assigned to the Author Accepted Manuscript version that might arise from this submission. Dr Lautatzis received salary support from the Canadian Pediatric Endocrine Group Fellowship Program and CIHR Canada Graduate Scholarship. The funding agencies were not involved in the design, implementation, analysis, or interpretation of the data.
CONFLICT OF INTEREST DISCLOSURES: There are no conflicts of interest to disclose.
Comments