Comparative effectiveness of low-dose versus high-dose vitamin D on bone metabolic markers in preterm infants: a retrospective cohort study

Introduction

Vitamin D (VD) is a group of bioactive, fat-soluble secosteroids with central roles in human health (1). It regulates calcium and phosphorus homeostasis and supports bone mineralization and skeletal metabolic balance (2). Beyond this classical role, VD has been linked to neuroprotection (3,4), attenuation of lung injury risk (5,6), and immunomodulation (7,8). In newborns, VD stores depend largely on transplacental transfer during late gestation, particularly in the third trimester (9). Preterm birth interrupts this period of accumulation. Infants born at less than 32 weeks’ gestation therefore enter postnatal life with limited reserves, high metabolic demands, and substantial comorbidity burden, all of which increase vulnerability to VD deficiency and related metabolic disturbance (10,11).

Improved survival of very preterm and extremely preterm infants has shifted attention toward postnatal growth, metabolic adaptation, and long-term health (12). Metabolic bone disease of prematurity, impaired mineralization, and related abnormalities in bone metabolism remain major clinical problems. VD is a modifiable nutritional factor in early prevention and management strategies (13). Preterm infants undergo rapid growth, tissue repair, calcium-phosphorus deposition, bone remodeling, and endocrine adaptation. VD supplementation therefore requires a balance between adequacy and safety rather than dose escalation alone (14,15). This balance is unlikely to be uniform across gestational-age strata. Intestinal absorption, hepatic and renal hydroxylation, inflammatory stress, drug exposure, and nutritional support may alter VD bioavailability and metabolic response (16).

Recommendations for VD supplementation in preterm infants remain inconsistent. The American Academy of Pediatrics recommends 400 IU/day for exclusively or partially breastfed infants (17). The European Society for Paediatric Gastroenterology, Hepatology and Nutrition recommends 800–1,000 IU/day for preterm infants during the first months after birth (18). In China, the 2021 Expert Consensus on Clinical Management of Metabolic Bone Disease of Prematurity recommends 400–1,000 IU/day, with adjustment according to nutritional intake and clinical risk (19). These differences likely reflect variation in sunlight exposure, feeding and fortification practices, baseline VD status, genetic background, and neonatal intensive care unit management pathways. They also indicate persistent uncertainty regarding the most appropriate dose, the dose-response relationship, and the threshold beyond which additional supplementation provides limited benefit or adds metabolic burden (10,15,16).

Serum 25-hydroxyvitamin D [25(OH)D] is an important indicator of VD status, but target concentrations do not necessarily reflect optimal bone metabolic adaptation. In preterm infants, biochemical response may not be captured by 25(OH)D alone. Alkaline phosphatase (ALP), a marker of bone formation activity and mineralization burden, may provide complementary information under different supplementation regimens (14). Interpretation is further complicated by complex nutritional exposure. Prescribed VD dose is not the sole source of intake. Parenteral nutrition (PN), enteral feeds, and milk fortification may all contribute to background exposure in the neonatal intensive care unit.

This study included infants born preterm at less than 32 weeks’ gestation and compared two VD supplementation regimens commonly used in routine practice, 500 IU/day and 900 IU/day. We evaluated their associations with VD status, calcium-phosphorus metabolic indices, and bone metabolism-related biomarkers over the first four postnatal weeks, and examined whether these associations differed by gestational-age stratum. The objective was not to identify a single optimal dose, but to compare early observational outcomes associated with two commonly used supplementation pathways in a high-risk preterm population. The findings may inform the evidence base for VD supplementation in preterm infants and support a more stratified approach to neonatal intensive care unit practice. We present this article in accordance with the STROBE reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0188/rc).

Methods

Study design

This was a single-center, retrospective cohort study conducted in the Neonatal Intensive Care Unit (NICU) of the Women’s Hospital, Zhejiang University School of Medicine. Eligible participants were preterm infants born at less than 32 weeks’ gestation and discharged from the NICU between April 1, 2024 and May 31, 2025. The study protocol was approved by the Ethics Committee of Women’s Hospital, Zhejiang University School of Medicine (Approval No. IRB-20250350-R). Individual informed consent was waived because of the retrospective design. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study assessed biochemical outcomes during the first 4 postnatal weeks under routine NICU practice.

Data collection and covariate definitions

Maternal and perinatal variables included maternal age, gestational hypertension, gestational diabetes mellitus (GDM), preeclampsia, histologic chorioamnionitis and/or intrauterine infection, gestational hypothyroidism, and intrahepatic cholestasis of pregnancy (ICP). Maternal 25(OH)D concentrations were recorded when available. Maternal VD testing was not part of routine screening during the study period.

Neonatal baseline characteristics included gestational age (GA), birth weight, sex, small for gestational age (SGA) status, mode of delivery, multiple gestation, 1- and 5-minute Apgar scores, duration of PN, weight-for-age z score (WAZ) at 4 weeks of postnatal age, length of hospital stay, and human milk fortification status. PN duration and human milk fortification status were recorded as proxy indicators of background nutritional VD exposure.

Inclusion and exclusion criteria for study subjects

Infants were included if they met all of the following criteria: admission to the NICU within 24 hours after birth, no VD supplementation before enrollment, and complete maternal and neonatal clinical records. Exclusion criteria were chromosomal abnormalities or major congenital malformations, severe hepatic or renal dysfunction, biliary atresia, congenital metabolic disorders, failure to follow the prescribed VD supplementation protocol, incomplete 25(OH)D assessment at the scheduled 4-week time point, and severely incomplete clinical data. The participant flow diagram is shown in Figure S1.

Protocol non-adherence was defined as a change in VD dosage between initiation and the 4-week assessment period, or discontinuation for 7 days or longer. Severely incomplete key data were defined as absence of the primary outcome [4-week 25(OH)D concentration] or absence of at least one critical covariate, including GA, birth weight, PN duration, or SGA status.

Exposure definition and grouping

VD supplementation was initiated during routine NICU care at approximately 1 week after birth. The initiation window was defined as postnatal days 5 to 8 (day of life, DOL). Supplementation continued until the 4-week assessment. Infants were grouped according to the actual daily supplementation dose received.

The 500 IU group received one oral drop of a vitamin AD preparation daily (Shandong Dayin Marine Biological Pharmaceutical Co., Ltd., Rongcheng, China), providing 500 IU/day of VD. The 900 IU group received the same vitamin AD preparation plus one additional oral drop of VD (Guangzhou Pharmaceutical Co., Ltd., Guangzhou, China), yielding a total dose of 900 IU/day. The two exposure groups reflected the medication formulations available during the study period. Treatment allocation was non-randomized. The final regimen was determined by the treating physician during routine NICU care.

For infants discharged before the scheduled 4-week assessment, adherence was assessed from discharge prescriptions and follow-up records. Infants with unverifiable adherence were classified as having missing exposure data.

Nutritional management and background VD exposure

Nutritional support followed a standardized NICU protocol based on body weight, GA, and feeding tolerance. Lipid-soluble vitamins were routinely added during PN. Human milk fortifier was introduced after enteral feeding reached the predefined threshold and was advanced according to gastrointestinal tolerance. PN was discontinued after enteral feeds reached target volumes according to GA. Exact total VD intake from all nutritional sources was not reconstructed on an individual daily basis. PN duration and human milk fortification status were therefore used as proxy indicators of background nutritional VD exposure.

Outcome measures and definitions

The primary outcome was serum 25(OH)D concentration at the fourth week after birth. VD status was classified according to the 2016 Global Consensus Recommendations on Prevention and Management of Nutritional Rickets as deficiency (less than 12 ng/mL), insufficiency (12 to 20 ng/mL), and sufficiency (greater than 20 ng/mL) (20). Neonatal baseline 25(OH)D concentrations were not routinely measured during the first postnatal week. Initial neonatal VD status was therefore not directly comparable between groups.

Biochemical outcomes included serum calcium, phosphorus, and ALP concentrations at the fourth week after birth, together with baseline calcium, phosphorus, and ALP measurements obtained within the first week after birth. Clinical outcomes included late-onset sepsis (LOS), necrotizing enterocolitis (NEC), intraventricular hemorrhage (IVH), retinopathy of prematurity (ROP), and moderate-to-severe bronchopulmonary dysplasia (BPD), defined according to established NICU diagnostic criteria.

Blood sample collection, processing protocol, and laboratory analysis

Blood samples were collected at two predefined time points: within the first week after birth (DOL 5 to 8) and during the fourth week after birth (DOL 26 to 30). Sampling was scheduled preferentially between 08:00 and 10:00. The interval since the most recent VD administration was recorded at each sampling time point. After collection, blood samples were allowed to clot at room temperature for 30 minutes and were then centrifuged at 1,500 ×g for 10 minutes at room temperature (20 to 25 °C) to separate serum. Serum samples were stored at −80 °C until analysis. Freeze-thaw cycles were limited to no more than one.

Serum 25(OH)D concentrations were measured by liquid chromatography-tandem mass spectrometry (LC-MS/MS; CalQuant Technology, Hangzhou, China). The assay had a limit of detection of 1 ng/mL, a linear range of 1 to 100 ng/mL, an intra-assay coefficient of variation of 8%, and an inter-assay coefficient of variation of 10%. Low-, medium-, and high-level quality control samples were included in each analytical batch. Serum calcium and ALP concentrations were measured by colorimetric analysis using the AU5800 automated biochemical analyzer (Beckman Coulter, Brea, CA, USA). Serum phosphorus was measured by the molybdate ultraviolet method.

Prespecified subgroup analysis

Prespecified subgroup analysis was performed according to GA, with two strata defined as less than 28 weeks and 28⁺⁰ to 31⁺⁶ weeks. The stratification was based on clinical classification of preterm birth and differences in mineral deposition, enteral feeding progression, and risk of metabolic bone disease. Birth weight and SGA status were included as covariates in adjusted models. Interaction effects were explored when sample size allowed.

Statistical analysis

All analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 20.0. Analyses were conducted in a prespecified sequence. Baseline characteristics were compared first. Primary and secondary outcomes were then compared in unadjusted analyses. Adjusted analyses were subsequently performed for selected biochemical and clinical outcomes.

Continuous variables were assessed for normality using the Shapiro-Wilk test and for homogeneity of variance using the Levene test. Variables with normal distribution and homogeneous variance are presented as mean ± standard deviation (SD) and were compared using the independent-samples t-test. Welch’s t-test was used when variance homogeneity was not satisfied. Non-normally distributed variables are presented as median (interquartile range, IQR) and were compared using the Mann-Whitney U test. Categorical variables are presented as counts (percentages) and were compared using the chi-squared test or Fisher’s exact test, as appropriate. Intergroup differences and 95% confidence intervals (CIs) were reported for continuous variables. Odds ratios (ORs) and 95% CIs were reported for categorical variables. Binary outcomes were further analyzed using logistic regression.

Baseline-adjusted analyses were performed for calcium, phosphorus, and ALP using analysis of covariance (ANCOVA), with the 4-week measurement as the dependent variable, supplementation group as the primary independent variable, and the corresponding 1-week baseline value as a covariate. Homogeneity of regression slopes was assessed by testing the interaction between group and baseline value. Residual normality and homoscedasticity were assessed using residual-based diagnostic procedures. Change-from-baseline analyses were also performed as supplemental robustness analyses, with change scores defined as the 4-week value minus the corresponding baseline value.

For short-term neonatal complications, adjusted analyses were performed using multivariable logistic regression models. The models were adjusted for GA, birth weight, PN duration, and small-for-gestational-age status. All statistical tests were two-sided. A P value <0.05 was considered statistically significant.

Results

Selection of study subjects and comparison of baseline characteristics

During the study period, 499 preterm infants with a GA of less than 32 weeks were discharged. Based on the predefined exclusion criteria, 260 infants were excluded and 239 were included in the final analysis. The GA range was 24⁺³ to 31⁺⁶ weeks, and birth weight ranged from 570 to 2,350 g. According to the prescribed VD supplementation regimen, 124 infants were assigned to the 500 IU group and 115 to the 900 IU group.

Maternal complications were comparable between groups (Table 1). No significant between-group differences were observed in maternal age or pregnancy complications, including gestational hypertension, GDM, preeclampsia, histologic chorioamnionitis and/or intrauterine infection, gestational hypothyroidism, and ICP (all P>0.05).

Neonatal baseline characteristics were also comparable between groups (Table 2). No significant differences were observed in GA, birth weight, sex, SGA status, mode of delivery, multiple gestation, Apgar scores, PN duration, length of hospital stay, or baseline calcium, phosphorus, and ALP concentrations measured within the first week after birth (all P>0.05).

Maternal 25(OH)D data were available only in a subset of 61 mother-infant pairs because maternal VD testing was not routinely performed during the study period. Among the available cases, maternal 25(OH)D concentrations did not differ significantly between groups (Table S1). Human milk fortification status also did not differ significantly between groups (Table S1). Together with the comparable PN duration, these data provided supportive but indirect evidence regarding baseline comparability at the group level.

Comparison of serum 25(OH)D levels and VD status

At 4 weeks of postnatal age, serum 25(OH)D concentrations were similar between the 500 IU and 900 IU groups [median (IQR), 31.46 (26.18–39.52) ng/mL versus 32.52 (24.94–38.57) ng/mL; P=0.771; Table 3]. GA-stratified analyses also showed no significant between-group differences in serum 25(OH)D concentration or VD status distribution (Table 3).

In infants born at less than 28 weeks’ gestation (n=53), no case of VD deficiency [25(OH)D less than 12 ng/mL] was identified in either group. Concentrations of 12–20 ng/mL were observed in 0 and 1 infant (3.70%) in the 500 IU and 900 IU groups, respectively, whereas concentrations greater than 20 ng/mL were observed in 26 (100.00%) and 26 (96.30%) infants. VD status distribution did not differ between groups (P>0.999). Among infants born at 28⁺⁰ to 31⁺⁶ weeks’ gestation (n=186), no case of VD deficiency was identified in either group. Concentrations of 12–20 ng/mL were observed in 8 (8.16%) and 10 (11.36%) infants, whereas concentrations greater than 20 ng/mL were observed in 90 (91.84%) and 78 (88.64%) infants in the 500 IU and 900 IU groups, respectively. VD status distribution did not differ significantly between groups (P=0.625; Table 3).

No serum 25(OH)D concentration within the toxic range was detected during the study period. These findings represent 4-week end-point comparisons. Baseline neonatal 25(OH)D concentrations were not available.

Association between VD supplementation dose and ALP levels in preterm infants with GA-stratified findings

ALP concentrations at week 4 were higher than baseline values measured within the first week after birth in both groups (Table 4). In the overall cohort, the 900 IU group had higher ALP concentrations at week 4 than the 500 IU group after baseline adjustment (adjusted difference, 48.88 U/L; 95% CI, 6.09–91.67; P=0.025), with a small effect size (=0.144, η²=0.021).

In the subgroup born at less than 28 weeks’ gestation, ALP concentrations were similar between groups [492.00 (438.00–646.25) versus 546.00 (461.50–720.00) U/L; P=0.265], with a small effect size (r=0.157, η²=0.025). In the subgroup born at 28⁺⁰ to 31⁺⁶ weeks’ gestation, ALP concentrations were higher in the 900 IU group than in the 500 IU group [443.00 (373.00–556.50) versus 398.50 (323.25–514.25) U/L], with borderline statistical significance and a small effect size (P=0.051; r=0.144, η²=0.021). ANCOVA diagnostic checks did not indicate major violations of model assumptions.

Comparison of calcium and phosphorus metabolism between different VD supplementation doses in preterm infants with GA-stratified analysis

By week 4, serum calcium and phosphorus concentrations were higher than baseline values in both groups (baseline values in Table 2; week 4 values in Table 5). In the overall cohort, serum calcium concentration did not differ between the 500 IU and 900 IU groups [median (IQR), 2.45 (2.36–2.58) versus 2.44 (2.33–2.54) mmol/L; P=0.346]. No between-group difference in calcium concentration was observed in either the less than 28 weeks subgroup or the 28⁺⁰ to 31⁺⁶ weeks subgroup (P=0.859 and P=0.300, respectively; Table 5).

Serum phosphorus concentration in the overall cohort was higher in the 500 IU group than in the 900 IU group, although the difference did not reach statistical significance [1.79 (1.56–2.06) versus 1.69 (1.34–2.02) mmol/L; P=0.058; Table 5]. In the less than 28 weeks subgroup, phosphorus concentration did not differ between groups (1.61±0.33 versus 1.51±0.44 mmol/L; P=0.341). In the 28⁺⁰ to 31⁺⁶ weeks subgroup, phosphorus concentration was also higher in the 500 IU group, without statistical significance [1.82 (1.61–2.14) versus 1.72 (1.40–2.05) mmol/L; P=0.082; Table 5].

At week 4, calcium concentrations were similar between groups in both the overall cohort and GA-stratified analyses. Phosphorus showed a directional between-group difference in some comparisons, but no stable statistically significant difference was observed in the unadjusted analyses.

Comparison of short-term neonatal clinical complications

The incidences of LOS, NEC, IVH, ROP, and moderate-to-severe BPD were similar between groups. In unadjusted analyses, crude ORs showed no significant association between supplementation dose and any of these outcomes (all P>0.05; Figure 1, Table 6).

Figure 1 Association between VD supplementation dose and the risk of neonatal complications. ORs and corresponding 95% CIs for the risk of common neonatal complications (LOS, NEC, IVH, ROP, and moderate-to-severe BPD) comparing a daily VD dose of 900 IU with 500 IU. Blue circles represent unadjusted ORs, and orange squares represent aORs derived from multivariable logistic regression models; the horizontal axis is displayed on a logarithmic scale, and the vertical reference line indicates OR =1. The adjusted models control for GA, birth weight, PN duration, and SGA status. aOR, adjusted odds ratio; BPD, bronchopulmonary dysplasia; CI, confidence interval; GA, gestational age; IVH, intraventricular hemorrhage; LOS, late onset sepsis; NEC, necrotizing enterocolitis; OR, odds ratio; PN, parenteral nutrition; ROP, retinopathy of prematurity; SGA, small for gestational age; VD, vitamin D.

After adjustment for GA, birth weight, PN duration, and SGA status, no independent association was observed between supplementation dose and the risk of LOS, NEC, IVH, ROP, or moderate-to-severe BPD. The adjusted odds ratios were 0.75 (95% CI, 0.34–1.66) for LOS, 0.45 (95% CI, 0.04–4.58) for NEC, 1.11 (95% CI, 0.65–1.88) for IVH, 0.63 (95% CI, 0.33–1.18) for ROP, and 2.00 (95% CI, 0.74–5.44) for moderate-to-severe BPD (all P>0.05; Figure 1, Table 6).

The number of NEC events was low, and the corresponding confidence intervals were wide. These analyses were interpreted as exploratory.

Baseline correction and variability robustness analysis

ALP increased over time in both groups, and the mean value at week 4 was higher in the 900 IU group than in the 500 IU group (Figure 2A). Calcium also increased over time in both groups and converged by week 4 (Figure 2B). Phosphorus increased in both groups, with a higher mean value at week 4 in the 500 IU group than in the 900 IU group (Figure 2C).

Figure 2 Temporal changes in calcium, phosphorus, and ALP and baseline-adjusted and change-score effects of different VD supplementation doses. (A-C) Temporal trajectories of biochemical indicators at postnatal week 1 and week 4 in the two dose groups (mean ± standard error): (A) ALP, (B) calcium, and (C) phosphorus; (D,E) forest plots of baseline-adjusted dose effects at week 4, showing adjusted mean differences (point estimates) and 95% CIs (horizontal lines) comparing the 900 IU group with the 500 IU group: (D) ALP and (E) calcium and phosphorus; (F,G) forest plots of change-score (week 4 minus baseline) dose effects, showing point estimates and 95% CIs for between-group differences: (F) ΔALP and (G) Δcalcium and Δphosphorus. ALP, alkaline phosphatase; CI, confidence interval; VD, vitamin D.

In the baseline-adjusted model, the adjusted difference in ALP at week 4 between the 900 IU and 500 IU groups was 48.88 U/L (95% CI, 6.09–91.67; P=0.025) (Figure 2D). Calcium remained non-significant after baseline adjustment (adjusted difference, −0.024 mmol/L; 95% CI, −0.072 to 0.024; P=0.320), whereas phosphorus showed a borderline decrease (adjusted difference, −0.097 mmol/L; 95% CI, −0.195 to 0.001; P=0.053) (Figure 2E).

In the change-from-baseline analysis, the between-group difference in ΔALP was 47.29 U/L (95% CI, 4.27–90.31; P=0.031) (Figure 2F). Differences in Δcalcium (−0.058 mmol/L; P=0.130) and Δphosphorus (−0.077 mmol/L; P=0.233) were not statistically significant (P>0.05, Figure 2G).

These analyses assessed the consistency of the observed biochemical signals across model specifications. The ALP remained consistent across models, whereas the phosphorus result varied by analytic approach.

Discussion

In this retrospective cohort of preterm infants born at less than 32 weeks’ gestation, the 500 IU/day and 900 IU/day regimens yielded similar serum 25(OH)D concentrations at 4 weeks and broadly comparable calcium-phosphorus biochemical profiles. The principal between-group difference was observed in ALP, which was higher in the 900 IU/day group in the overall analysis and in some model specifications. Within the first postnatal month, the higher prescribed dose was not associated with a measurable increase in serum 25(OH)D at the 4-week assessment.

The 25(OH)D findings were broadly consistent with earlier studies in preterm infants. Previous trials reported that moderate supplementation often maintained biochemical sufficiency, whereas higher doses did not consistently produce greater short-term increases in serum 25(OH)D (21,22). More recent studies also suggested that higher-dose regimens may increase the proportion of infants reaching predefined targets, although the relation between dose escalation and broader metabolic benefit remained variable across populations and follow-up windows (15,23,24). The similar 25(OH)D concentrations observed at 4 weeks in the present cohort may have been related to the metabolic and nutritional context of early postnatal care, including immature intestinal function, dependence on PN, and concurrent nutritional support such as fortified human milk or formula feeding. The relatively narrow difference between the two prescribed doses may also have limited detectable separation in serum 25(OH)D concentrations.

Baseline VD status remained difficult to define because neonatal 25(OH)D concentrations were not routinely measured during the first postnatal week. Maternal data provided partial context. Neonatal VD stores depend largely on maternal transplacental transfer during late gestation, and maternal-neonatal associations have been reported in recent preterm cohorts (9,10,25). In the available maternal subset, no apparent between-group difference was observed. These data were compatible with broadly similar baseline exposure at the group level, but they did not replace direct neonatal baseline measurement.

ALP is a critical biochemical indicator of bone metabolism activity and turnover in preterm infants, with elevated levels typically associated with enhanced osteogenesis or increased bone metabolic load. The ALP result represented the clearest biochemical difference between regimens. In infants born at 28⁺⁰ to 31⁺⁶ weeks’ gestation, ALP tended to be higher in the 900 IU/day group, with borderline statistical significance and a small effect size. The direction of this difference was preserved across baseline-adjusted and change-from-baseline analyses. Recent studies and reviews continued to support ALP as a clinically used marker of bone turnover and mineralization burden in preterm infants, although its interpretation remained context dependent and threshold based (13,26,27). The present data were therefore more consistent with a dose-associated biochemical signal than with a defined difference in clinical bone outcome.

The ALP pattern was not accompanied by a corresponding increase in 25(OH)D, and serum calcium remained similar between groups. A higher ALP signal in the setting of similar 25(OH)D concentrations may have been related to differences in bone metabolic phenotype rather than to biochemical VD status alone. At the same time, ALP is influenced by growth, nutritional intake, inflammatory stress, and hepatobiliary factors. The present data did not distinguish among these contributors. The absence of a comparable signal in infants born at less than 28 weeks’ gestation may have been related to greater physiologic heterogeneity, more intensive concurrent treatment, or lower statistical power in that subgroup.

The calcium and phosphorus results were more neutral. Calcium did not differ between groups in either the overall or the stratified analyses. Phosphorus did not reach statistical significance in the primary unadjusted comparisons, although a modest signal appeared in some adjusted models. This pattern was compatible with the nutritional context of NICU care, in which calcium-phosphorus homeostasis is shaped by enteral advancement, PN, fortification, and illness-related stress in addition to prescribed VD dose (13,28). In the present cohort, PN duration and human milk fortification status were comparable between groups. Exact total VD intake from all nutritional sources, however, was not reconstructed for each infant. The phosphorus finding was therefore best interpreted cautiously, particularly because its direction was not uniformly stable across analytic approaches.

No clear between-group difference was observed for the major non-skeletal outcomes, including LOS, NEC, IVH, ROP, and moderate-to-severe BPD. This pattern was consistent with recent reviews indicating that evidence linking higher-dose VD supplementation with major preterm morbidities remained mixed, particularly when baseline status was partly corrected and follow-up was short (16,24,29). These outcomes arise from multifactorial pathways, and dose differences within the 500–900 IU/day range may have had limited detectable impact within the present sample and time frame. Modest differences in prescribed dose may also have been insufficient to generate measurable downstream effects once 25(OH)D concentrations had entered a relatively sufficient range.

From a clinical perspective, the data supported a restrained interpretation. Within the first month after birth, 500 and 900 IU/day were associated with similar 25(OH)D and calcium profiles, whereas the higher-dose regimen was associated with higher ALP in some analyses. This pattern supported continued use of 25(OH)D as a biochemical target, together with parallel attention to bone-metabolism markers when higher-dose supplementation was prescribed. The findings were more compatible with a stratified monitoring approach than with a uniform assumption that greater VD exposure necessarily improved short-term biochemical status in all preterm infants (13).

Several limitations constrained interpretation. The retrospective single-center design and non-randomized treatment allocation left room for residual confounding, including confounding by indication. Neonatal baseline 25(OH)D concentrations were unavailable. Maternal VD data were incomplete and available only in a subset, as maternal VD levels during pregnancy are not part of routine clinical screening. While there were no significant differences in maternal 25(OH)D concentrations across groups in the subset of available data, and while baseline comparability between mothers and neonates was confirmed, these indirect measures cannot serve as a substitute for direct assessment of neonatal baseline 25(OH)D levels. Total VD intake from PN, enteral feeds, and fortification was not quantified for each infant. The retrospective design and fixed available sample size may have limited the statistical power to detect small between-group differences in 4-week serum 25(OH)D concentrations. The follow-up window was limited to 4 weeks, whereas metabolic bone disease of prematurity often peaks later. Many infants in this cohort were born at 30–32 weeks’ gestation and had been discharged by 6–12 weeks after birth, making later follow-up incomplete. Some stratified and complication analyses were also based on small numbers and may have been underpowered for modest effects.

Future studies will require prospective multicenter designs, direct neonatal baseline 25(OH)D measurement, and follow-up beyond discharge. Integration of imaging-based bone mineralization, endocrine regulators such as parathyroid hormone, and more complete reconstruction of nutritional VD exposure would allow more detailed assessment of biochemical and clinical trajectories across gestational-age strata.

Conclusions

In this single-center retrospective cohort study of preterm infants with a GA of less than 32 weeks, VD supplementation with 500 and 900 IU/day was associated with similar serum 25(OH)D concentrations and broadly comparable calcium-phosphorus biochemical profiles at 4 weeks of postnatal age. No case of VD deficiency or toxicity was identified in either group during this observation window. In analyses of bone metabolic markers, the 900 IU/day regimen was associated with higher ALP in some models, particularly in the subgroup with a GA of 28⁺⁰ to 31⁺⁶ weeks. These findings support interpretation of the two regimens within the first month after birth, rather than beyond this early postnatal period. Larger prospective multicenter studies with direct baseline neonatal 25(OH)D assessment and longer follow-up are needed to further evaluate gestational-age-specific VD supplementation strategies and their relation to longer-term bone and developmental outcomes.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0188/rc

Data Sharing Statement: Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0188/dss

Peer Review File: Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0188/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0188/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study protocol was approved by the Ethics Committee of Women’s Hospital, Zhejiang University School of Medicine (Approval No. IRB-20250350-R). Individual informed consent was waived because of the retrospective design. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Christakos S, Dhawan P, Verstuyf A, et al. Vitamin D: Metabolism, Molecular Mechanism of Action, and Pleiotropic Effects. Physiol Rev 2016;96:365-408. [Crossref] [PubMed]
2. Kowalówka M, Główka AK, Karaźniewicz-Łada M, et al. Clinical Significance of Analysis of Vitamin D Status in Various Diseases. Nutrients 2020;12:2788. [Crossref] [PubMed]
3. Basta-Kaim A. The Vitamin D Receptor as a Potential Target for the Treatment of Age-Related Neurodegenerative Diseases Such as Alzheimer's and Parkinson's Diseases: A Narrative Review. Cells 2023;12:660.
4. Cui X, Eyles DW. Vitamin D and the Central Nervous System: Causative and Preventative Mechanisms in Brain Disorders. Nutrients 2022;14:4353. [Crossref] [PubMed]
5. Konstantinopoulou S, Tapia IE. Vitamin D and the lung. Paediatr Respir Rev 2017;24:39-43. [Crossref] [PubMed]
6. Weiss ST, Mirzakhani H, Carey VJ, et al. Prenatal vitamin D supplementation to prevent childhood asthma: 15-year results from the Vitamin D Antenatal Asthma Reduction Trial (VDAART). J Allergy Clin Immunol 2024;153:378-88. [Crossref] [PubMed]
7. Galoppin M, Kari S, Soldati S, et al. Full spectrum of vitamin D immunomodulation in multiple sclerosis: mechanisms and therapeutic implications. Brain Commun 2022;4:fcac171. [Crossref] [PubMed]
8. Zhao R, Zhang W, Ma C, et al. Immunomodulatory Function of Vitamin D and Its Role in Autoimmune Thyroid Disease. Front Immunol 2021;12:574967. [Crossref] [PubMed]
9. Taylor SN. Vitamin D for very preterm infants-determining the how, when, and why. Pediatr Res 2024; Epub ahead of print. [Crossref]
10. Radu IA, Cucerea M, Gheonea C, et al. Vitamin D Supplementation During Pregnancy and Maternal and Neonatal Vitamin D Status at ≤32 Weeks Gestation: Romanian Prospective Observational Cohort Study. Children (Basel) 2025;12:682. [Crossref] [PubMed]
11. Wu M, Zhao M, Jin X, et al. Vitamin D-related risk factors for preterm and full-term infants at birth: a retrospective study. BMC Pediatr 2025;25:417. [Crossref] [PubMed]
12. Kalyanam SM, Kuiper JR, Iwashyna TJ, et al. Center-specific variation in outcomes for extremely premature, extremely low birth weight neonates. Front Pediatr 2025;13:1570542. [Crossref] [PubMed]
13. Grover M, Ashraf AP, Bowden SA, et al. Invited Mini Review Metabolic Bone Disease of Prematurity: Overview and Practice Recommendations. Horm Res Paediatr 2025;98:40-50. [Crossref] [PubMed]
14. Matejek T, Pokorna V, Drahy V, et al. Urinary Calcium and Phosphorus Excretion in Vitamin D-Sufficient Preterm Infants: Establishing Age-Specific Normative Data. Acta Paediatr 2025;114:3267-78. [Crossref] [PubMed]
15. Rueang-Amnat N, Kittisakmontri K, Khuwuthyakorn V, et al. Improving Vitamin D Status in Preterm Newborns: A Randomized Trial of 800 vs. 400 IU/Day. Nutrients 2025;17:1888.
16. Radu IA, Ognean ML, Ștef L, et al. Vitamin D: What We Know and What We Still Do Not Know About Vitamin D in Preterm Infants-A Literature Review. Children (Basel) 2025;12:392. [Crossref] [PubMed]
17. Wagner CL, Greer FR, et al. Prevention of rickets and vitamin D deficiency in infants, children, and adolescents. Pediatrics 2008;122:1142-52. [Crossref] [PubMed]
18. Agostoni C, Buonocore G, Carnielli VP, et al. Enteral nutrient supply for preterm infants: commentary from the European Society of Paediatric Gastroenterology, Hepatology and Nutrition Committee on Nutrition. J Pediatr Gastroenterol Nutr 2010;50:85-91. [Crossref] [PubMed]
19. Yan-Mei C, Xin-Zhu L, Rong Z, et al. Expert consensus on clinical management of metabolic bone disease of prematurity (2021). Zhongguo Dang Dai Er Ke Za Zhi 2021;23:761-72. [Crossref] [PubMed]
20. Munns CF, Shaw N, Kiely M, et al. Global Consensus Recommendations on Prevention and Management of Nutritional Rickets. J Clin Endocrinol Metab 2016;101:394-415. [Crossref] [PubMed]
21. Backström MC, Mäki R, Kuusela AL, et al. Randomised controlled trial of vitamin D supplementation on bone density and biochemical indices in preterm infants. Arch Dis Child Fetal Neonatal Ed 1999;80:F161-6. [Crossref] [PubMed]
22. Pittard WB. How Much Vitamin D for Neonates? Arch Pediatr Adolesc Med 1991;145:1147.
23. Romero-Lopez M, Tyson JE, Naik M, et al. Randomized controlled trial of enteral vitamin D supplementation (ViDES) in infants <28 weeks gestational age or <1000 g birth weight: study protocol. Trials 2024;25:423. [Crossref] [PubMed]
24. Shin SH, Kim HJ, Heo JS. Short-term and long-term effects of vitamin D supplementation for preterm infants: a systematic review and meta-analysis. J Perinatol 2026;46:425-36. [Crossref] [PubMed]
25. Stoica AB, Săsăran MO, Mărginean C. Maternal Vitamin D Status and Its Association with Neonatal Health: Clinical Implications and Influencing Factors. Nutrients 2025;17:2761. [Crossref] [PubMed]
26. Levkovitz O, Khatib-Fadila S, Sirota GL, et al. Early serum alkaline phosphatase and other factors associated with enteral feeding in very low birth weight preterm infants. J Pediatr Gastroenterol Nutr 2025;81:113-24. [Crossref] [PubMed]
27. Alsaif S, Aljarbou T, Alrahili M, et al. Biochemical markers for metabolic bone disease in preterm infants: insights from a structured Bone Health Programme in the Middle East. BMJ Paediatr Open 2025;9:e003927. [Crossref] [PubMed]
28. Leonardi R, Mattia C, Decembrino N, et al. The Critical Role of Vitamin D Supplementation for Skeletal and Neurodevelopmental Outcomes in Preterm Neonates. Nutrients 2025;17:1381. [Crossref] [PubMed]
29. Salas AA, Argent T, Jeffcoat S, et al. Early Vitamin D Supplementation in Infants Born Extremely Preterm and Fed Human Milk: A Randomized Controlled Trial. J Pediatr 2025;287:114754. [Crossref] [PubMed]