Invited Mini Review Metabolic Bone Disease of Prematurity: Overview and Practice Recommendations Open Access

Metabolic bone disease of prematurity (MBDP) is defined as undermineralization of the preterm infant skeleton arising from inadequate prenatal and postnatal calcium (Ca) and phosphate (PO4) accretion. Severe MBDP is associated with radiologic evidence of rickets and fractures. Preterm infants should be monitored closely for early biochemical signs of MBDP to prevent more advanced disease. Although many other terms have been used, at times interchangeably, for MBDP, including neonatal metabolic bone disease, osteopenia of prematurity, osteoporosis of prematurity, rickets of prematurity, and bone mineral deficiency of prematurity, the term MBDP is preferred for the overarching diagnosis. Throughout this article, we have also employed the term “phosphate” or PO4 to refer to the molecule that is relevant for biological processes and reserved the term phosphorus for laboratory assays (as these results, expressed as mg/dL or mmol/L, represent phosphorus) [1].

MBDP’s true incidence and prevalence are unknown. Radiographic evidence of bone undermineralization is found in up to 40% of infants born at <28 weeks’ gestation, with peak occurrence between 4 and 8 weeks of postnatal age [2]. MBDP total incidence may be higher in preterm infants with bronchopulmonary dysplasia [3, 4]. Fractures are reported in 1–10% of preterm babies with MBDP [2, 4‒6]. Long-term outcomes of MBDP are largely unknown [7].

A recent national survey indicated significant variability and knowledge gaps in the management of MBDP among neonatologists and pediatric endocrinologists [8]. This highlights the need for updated screening and treatment recommendations, as well as continued education in MBDP evaluation and management. Below, we outline a pragmatic data-driven expert consensus approach to MBDP, starting with 2 case vignettes. (These are hypothetical cases based on our usual consultations; as such, consent to publish is not required.)

At 84 days of life, a male infant born at 25 weeks’ gestation (birth weight 650 g) had a routine chest radiograph which revealed several rib fractures at various stages of healing and a left distal humerus fracture with rachitic changes. He had chronic lung disease, apnea of prematurity, two episodes of sepsis, medical necrotizing enterocolitis with multiple enteral feed interruptions, and long-term total parenteral nutrition (PN) dependency. He was being treated with furosemide and caffeine. Laboratories while on total PN revealed a serum Ca of 10.3 mg/dL (2.57 mmol/L) (normal range 9–11 mg/dL; 2.25–2.75 mmol/L), serum phosphorus of 3.6 mg/dL (1.16 mmol/L) (normal range 4–8 mg/dL; 1.29–2.58 mmol/L), alkaline phosphatase (ALP) 1112 U/L (normal range 150–420 U/L), 25 hydroxy vitamin D (25OHD) of 40 ng/mL (100 nmol/L) (normal range 20–80 ng/mL; 62–200 nmol/L), intact parathyroid hormone (PTH) of 57 pg/mL (6 pmol/L) (normal range 15–65 pg/mL; 1.6–6.9 pmol/L), and a tubular resorption of phosphate (TRP) of 98% (normal range 78–91%).

A female infant with a history of intrauterine growth restriction was born small for gestational age at 28 weeks’ gestation. She required continuous positive airway pressure. She was fed unfortified expressed breast milk exclusively. At 56 days of age, she was noted to have persistently decreased left forearm movement. Radiographs showed rachitic fraying and cupping of the distal radial and ulnar metaphyses. Laboratories showed a serum Ca 9.5 mg/dL (2.37 mmol/L) (normal range 9–11 mg/dL; 2.25–2.75 mmol/L), serum phosphorus 3.6 mg/dL (1.16 mmol/L) (normal range 4–8 mg/dL; 1.29–2.58 mmol/L), ALP 950 U/L (normal range 150–420 U/L), 25OHD 35 ng/mL (87.5 nmol/L) (normal range 20–80 ng/mL; 62–200 nmol/L), intact PTH of 110 pg/mL (11.6 pmol/L) (normal range 15–65 pg/mL; 1.6–6.9 pmol/L), and TRP of 75% (normal range 78–91%). Both infants show biochemical and radiographic evidence of MBDP with predisposing risk factors which will be discussed below.

Bone volume and mineralization increase with gestational age peaking between 32 and 36 weeks of gestation. Up to 80% of skeletal mineral accrual (100–120 mg/kg/day of Ca and 50–65 mg/kg/day of PO4) occurs during the third trimester [9]. As a result, infants born prematurely (especially prior to 28 weeks’ gestation) are deprived of intrauterine mineral accrual. The placenta is critical for fetal skeletal mineralization prenatally. It transfers Ca against the maternal-fetal gradient [10] and converts 25OHD to 1,25 dihydroxy vitamin D (1,25(OH)₂D), which is essential for transplacental PO4 transport [11]. Chronic placental insufficiency (such as that associated with intrauterine growth restriction or infection) can alter PO4 transport and increase MBDP risk [12]. Maternal dietary Ca intake may also play a role in fetal skeletal mineralization [13], but in utero skeletal mineralization appears to be independent of maternal vitamin D status [14].

Postnatally, enteral and parenteral Ca and PO4 intake are critical; inadequate intake and/or absorption can contribute to MBDP. Many infants achieve “full feeds” slowly, if at all. Normal enteral feeds of unfortified human breast milk and full-term infant formulas do not provide adequate Ca and PO4 for preterm infants. Formulas designed for premature infants have higher Ca and PO4 content, but bioavailability may differ. Human milk fortifiers also contain additional Ca and PO4 but may not be well tolerated. PO4 deficiency can result in relative Ca excess, leading to hypercalcemia and hypercalciuria. PN provides at best 60–70% of mineral requirements due to solubility constraints [15]. Aluminum exposure from long-term PN (including formulations of parenteral Ca and PO4 supplements) can reduce bone mineralization and bone formation [16, 17]. Medications commonly used in premature infants, including caffeine, diuretics, and steroids, can also contribute to MBDP due to their adverse effect on mineral homeostasis [18‒20].

Although maternal 25OHD status determines infant 25OHD status at birth, vitamin D deficiency is not a primary cause of MBDP and does not predict rickets or fractures in very low birth weight (VLBW, <1,500 g) infants [14, 21]. In low birth weight (LBW, <2,500 g) infants, hypocalcemia may be caused by the rapid skeletal accretion of Ca, reduced absorption of Ca in the intestinal tract, and relative resistance of bone to 1,25(OH)₂D [22]. Early hypocalcemia may not be responsive to calcitriol administration as passive saturable intestinal Ca absorption mechanisms predominate [23]. As newborns mature, 1,25(OH)₂D-dependent active Ca transport mechanisms become increasingly important, and 1,25(OH)₂D concentrations increase in premature infants over the first several weeks of life, not otherwise related to hypophosphatemia or hyperparathyroidism [24, 25]. Table 1 summarizes risk factors associated with MBDP development.

MBDP screening should be determined by prevalent risk. American Academy of Pediatrics (AAP) guidelines suggest MBDP screening begin around 4–6 weeks after birth as clinical rickets is uncommon in the first few weeks of life [14]. Screening may also be initiated if there is radiographic evidence of rickets and/or fractures [26].

Our consensus approach to screening is outlined in Figure 1. Published screening algorithms for MBDP most commonly include initial testing of serum Ca, phosphorus, and ALP to provide a mineral homeostasis overview. Some also recommend that infants with a birth weight <600 g be screened for MBDP by wrist or knee X-ray at 6 weeks of life [26]. Though ultrasound and dual-energy X-ray absorptiometry have been used in research studies to diagnose MBDP, they are not currently used clinically.

Isolated serum Ca measurements are not useful as MBDP screening tests, given that PTH can usually maintain eucalcemia at the expense of increased bone turnover in the setting of increased mineral needs. However, hypocalcemia may result from decreased Ca intake/absorption or increased urinary Ca losses, and hypercalcemia can accompany hypophosphatemia.

Infants with hypophosphatemia are at increased risk for MBDP, but the exact serum phosphorus concentration at which intervention is required is unknown. Although some literature suggests infants with serum phosphorus levels <5.5 mg/dL have increased MBDP risk [27], these infants received enteral feeds with mineral content far below that of current recommendations. The 2013 AAP guidelines suggest further evaluation of phosphorus values <4 mg/dL, especially if persistent for more than 1–2 weeks since transient hypophosphatemia can be present during an acute illness [14].

In general, rachitic changes are more often observed in infants with higher ALP concentrations (>800 IU/L) but can occur even at lower concentrations (<600 IU/L) [26]. As total ALP may be elevated in liver disease, measurement of the bone-specific isoform can be useful if the etiology is unclear but does not improve diagnosis in routine MBDP screening [28]. ALP interpretation may be complicated by states of undernutrition (e.g., zinc deficiency) and concurrent glucocorticoid use, both of which can decrease bone formation and thereby offer false reassurance by lowering ALP levels.

A small study in VLBW neonates suggests a physiologic reference range for PTH in preterm infants (9.4–66 pg/mL) like that of adults [29]. PTH concentrations >100 pg/mL, therefore, also raise concern for MBDP with Ca deficiency [30]. Low/normal PTH combined with hypophosphatemia in the setting of MBDP may reflect PO4 deficiency. Of note, PTH can be elevated in infants with chronic kidney disease and must be interpreted with caution in that population.

Some advocate using TRP as well [31, 32] as part of an initial screen or instead of PTH in children where PTH measurement may be difficult due to blood volume or laboratory constraints. TRP, calculated as: {1-[Urinary phosphorus/Urinary creatinine (Cr) × Serum Cr/Serum phosphorus]} × 100, quantifies the degree of PO4 wasting [33]. Normal TRP in premature infants ranges from 78 to 91% [34]. TRP >95% suggests almost complete renal PO4 reabsorption secondary to insufficient intake. TRP may be useful in lieu of PTH or as a confirmatory test; a low TRP may suggest underlying Ca deficiency, while a high TRP suggests PO4 deficiency (Fig. 2).

Vitamin D status should be monitored exclusively by 25OHD. Maternal vitamin D deficiency, conditions adversely affecting vitamin D absorption, and hepatic insufficiency may result in low 25OHD. Conversely, infants receiving fortified feedings and supplementation risk vitamin D toxicity (often defined as a 25OHD concentration of >150 ng/mL [35]). There are insufficient data to use 1,25(OH)₂D measurements to guide therapy in MBDP.

Once screening is initiated, we recommend repeat monitoring every 1–2 weeks based on initial results and baseline risk [36] (Table 2, Fig. 1). There is no defined ALP cutoff for MBDP; various studies identify values above 500–900 IU/L as associated with MBDP [26, 28, 37]. The initial ALP may not reflect the severity of demineralization since ALP is a measure of bone formation [38], but persistently elevated or upward trending ALP after 6 weeks of life suggests increased MBDP risk.

The differential diagnosis of fractures in the early newborn period includes osteogenesis imperfecta (OI), hypophosphatasia, or rarer syndromes such as Cole-Carpenter syndrome or juvenile Paget’s disease. Bone deformities and Wormian bones are important clues for OI. While commonly seen in OI, blue sclera can be a normal finding in infants less than 6 months due to the thinness of their sclera. Cole-Carpenter syndrome shares similarities with OI, but it also presents with craniosynostosis, hydrocephalus, and ocular proptosis [39]. In infancy, juvenile Paget’s disease may manifest with markedly elevated ALP levels (4–13 fold) and skeletal deformities [40]. Another genetic disorder to consider is X-linked hypophosphatemic rickets, which may exhibit a biochemical profile resembling that of MBDP with hypophosphatemia and elevated ALP, although rickets is not apparent in early infancy in X-linked hypophosphatemic rickets. Low serum ALP level with rickets should prompt a consideration of hypophosphatasia [41].

Rib fractures in premature infants are much less common (0.3%) than previously reported likely secondary to improved nutritional management [6]. In hospitalized preterm neonates, they can be secondary to chest compression or chest physiotherapy [41, 42]. Non-accidental injury should also be considered as a cause of fractures in preterm infants, especially for posterior rib fractures detected posthospital discharge [43, 44].

Multiple fractures have been reported in preterm neonates after prolonged in utero exposure to magnesium sulfate for tocolysis of maternal premature labor [45, 46]. Nutritional deficiencies (e.g., zinc) should also be considered in infants who have faltering growth and evidence of significant bone disease. Neonatal severe hyperparathyroidism from inactivating Ca-sensing receptor mutations can present in preterm infants with early-onset hypercalcemia, inappropriately elevated PTH, and radiographic findings similar to MBDP [47].

The initial management goal is to prevent MBDP by providing adequate Ca and PO4 to mimic physiologic in utero mineral accretion (Fig. 2). Recommendations for Ca and PO4 intake should be tailored to route of administration. For PN, recommendations on goal intake/ratios may differ based on regional and national availability of different Ca and PO4 salts, and each NICU should evaluate their solubility limits and ability to maximize mineral delivery [42]. For instance, despite achieving normal serum levels with PN, mineral intake may fail to reach even 50% of rates of in utero mineral retention due to the poor solubility of available mineral salts in the USA. Adequate PO4 intake is important, particularly in the first few days of life, to prevent hypophosphatemia and hypercalcemia. More aggressive feeding strategies aimed at promoting growth, such as earlier introduction of protein supplementation in PN, have become commonplace in NICUs, but in VLBW preterm infants, high amino acid intake in PN with inadequate concurrent electrolyte and mineral provisions in first few days after birth can lead to hypophosphatemia secondary to increased transport of potassium and PO4 into the cells, resulting in “refeeding-like syndrome” [48]. Updated recommendations promote earlier provision of PO4 in PN and the use of lower Ca:PO4 ratios (0.8–1:1 mmol:mmol) until days 4–7 for VLBW infants to prevent hypercalcemia and hypophosphatemia (Table 3 [49, 50]). Decreased ratios (<0.8:1) increase the risk of hypocalcemia, leading to secondary hyperparathyroidism and eventual hypophosphatemia due to renal losses [48]. After the first few days of life, upward adjustment of phosphorus given as PO4 and Ca should ensue to optimize bone mineralization with a targeted Ca:PO4 ratio of 1–1.3:1 (Table 3 [50]). When PN is required, close biochemical monitoring and adjustments are critical.

Once infants are transitioned to enteral feedings, fortified breast milk or preterm formulas should be used for optimal mineral intake (Table 3) with most products having 1.8:1 Ca-PO4 ratio on mg-to-mg basis. By comparison, despite 60% Ca and 80% PO4 absorption by the newborn, unfortified human milk (in volumes of 180–200 mL/day) provides only 1/3rd of the in utero mineral accretion. While a 2:1 ratio matching the ratio seen in human milk has been thought to help mineral absorption, no optimal ratio has been identified for Ca:PO4. In fact, some have suggested a lower ratio of 1.5–1.7:1 for preterm infants to provide more PO4 and improve mineral retention [14]. Despite reaching higher mineral intake, some infants still develop MBDP and may require targeted supplementation with Ca and/or PO4.

Data on optimal dosing for vitamin D supplements and target 25OHD concentrations in preterm and LBW infants are limited [51, 52]. Recommendations range from 160 IU/day to 1,000 IU/day [36, 51, 53]. The AAP Committee on Nutrition has endorsed providing vitamin D at 200–400 IU/day for enterally fed preterm infants with birth weight >1,500 g with an upper tolerable intake of 1,000 IU/day for otherwise healthy infants [14]. For VLBW infants, the AAP recommends similar intakes with the added recommendation that vitamin D supplementation be increased to 400 IU once the infant weighs >1,500 g and is tolerating full enteral nutrition. The increase to 400 IU may be postponed until infants are >2,000 g if there are concerns about increasing the formula osmolarity. Overall, it seems prudent to keep 25OHD concentrations >20 ng/mL while avoiding vitamin D toxicity; supplementation may be discontinued if 25OHD levels are >75 ng/mL.

Calcitriol is another potential MBDP treatment; however, clear recommendations and guidelines for use of calcitriol are lacking. Chen et al. [54] reported two infants (one with rickets and hepatic insufficiency and the other with increased ALP and gamma-glutamyl transferase) who received calcitriol 0.01 μg/kg/day for 37–40 days (until discharged) with improvement in radiographic markers and decrease in ALP level. Rustico et al. [30] reported a retrospective case series of 35 infants with MBDP, 25OHD sufficiency, and elevated PTH (most of whom had insufficient Ca/PO4 intake and were receiving medications with adverse bone effects) who were given additional enteral Ca and started on calcitriol 0.05 μg/kg/day divided twice a day. Calcitriol supplementation was discontinued at varying timepoints for each infant in the retrospective study. Infants receiving calcitriol normalized PTH by 38 days with nadir on average at 61 days. Overall, this study posited that calcitriol use is safe in babies with PTH >100 pg/mL and recommended using calcitriol (and potentially supplemental Ca) to normalize the PTH.

Calcitriol may be, therefore, an adjunct to therapy for MBDP in select infants such as those with secondary hyperparathyroidism, intolerance to enteral Ca, renal or hepatic insufficiency. We suggest using calcitriol (starting dose of 0.05 μg/kg/day up to 0.2 μg/kg/day, in 1 or 2 divided doses) to improve enteral Ca and PO4 absorption in infants with secondary hyperparathyroidism. Practically speaking, the small doses required by preterm infants may require dilution or flushing if given enterally via tubing to ensure delivery of the entire dose, and it can take several weeks to see maximal effect. High doses of calcitriol can cause hypercalcemia and hypercalciuria. Serum Ca, PTH, and urine Ca/Cr must be monitored; however, we do not suggest routine monitoring of 1,25(OH)₂D for patients on calcitriol. The calcitriol dose can be increased if serum Ca or urine Ca/Cr is not elevated and PTH remains high.

Many premature infants have LBW, low nutrient stores, and high energy expenditure; therefore, optimal bone health also requires adequate protein and energy through enteral and PN. The recommended energy intake for premature neonates receiving PN is approximately 90–120 kcal/kg/day [55] and 110–135 kcal/kg/day during enteral nutrition [53].

Fracture prevention is a cornerstone of management for infants at risk. Bedside signage alerting members of the medical team to safe handling of at-risk infants should be displayed [56]. Therapeutic positioning can promote opportunities for the infant to move and build bone and muscle strength. Physical therapy programs with 5–15 min/day of activity can promote body weight, length, muscle mass, bone strength, bone mineral content (BMC), and mineralization in VLBW infants [38, 57]. Massage may also be used as an adjunct to physical therapy and may positively affect bone formation markers [58]. Both therapies can be taught to parents to practice with infants at the bedside and at home and can improve bonding [59].

Both infants show biochemical and radiological evidence of MBDP. The first infant has hypophosphatemia, elevated TRP, hyperphosphatasia, normal PTH, and normal 25OHD suggestive of PO4 deficiency. PO4 supplementation should be initiated as first-line therapy for this infant (Fig. 1).

The second infant has hypophosphatemia, eucalcemia, low TRP, hyperphosphatasia, elevated PTH, and normal 25OHD, suggestive of Ca deficiency. Ca supplementation (Fig. 1) is required to normalize elevated PTH and thereby reduce resultant bone resorption and hypophosphatemia. Despite having hypophosphatemia, PO4 supplementation as first-line therapy may result in worsening hyperparathyroidism, leading to exacerbation of MBDP (Fig. 2) [32].

In some infants, metabolic abnormalities may persist despite initial treatment with Ca (persistent hypophosphatemia despite improved TRP, PTH) or PO4 (triggering or worsening secondary hyperparathyroidism), and both Ca and PO4 may be required. When prescribed together, Ca and PO4 should be administered at least 2 h apart to avoid precipitation with poor absorption. Additionally, review of the infants’ medications with consideration of minimizing agents that would be detrimental to bone health, decreasing interventions that increase risk of fracture by promoting safe handling techniques, optimizing protein and energy intake, and increasing the integrity of the bone-muscle unit by therapeutic handling and massage should also be considered.

Close monitoring of response to treatment should include serum Ca, phosphorus, ALP, and PTH every 1–2 weeks. Repeat radiographs may be indicated if concerning for worsening MBDP. Mineral supplementation may be discontinued when all biochemical parameters have normalized (and ALP remains <600 U/L and trending downward), at which time, repeat radiographs should also be obtained to document resolution of MBDP [60]. Serum Ca, phosphorus, ALP, and PTH should be repeated 4 weeks after discontinuation of Ca/PO4 supplementation and/or 2–4 weeks after hospital discharge.

Studies on long-term bone health outcomes in prepubertal children affected by MBDP offer diverse findings. Most show that preterm children remained shorter and lighter than their term peers during early or late childhood with significantly lower bone mass and BMD even when adjusted for height than those born at term [61‒64], while one study reported appropriate BMC for their size [7]. Many adult studies also indicate lower BMC and BMD in VLBW adults in comparison to controls, when adjusted for height [65, 66]. Moreover, lower peak bone mass (achieved in mid-20s, regarded as the most important determinant of osteoporosis) and higher frequency of osteopenia/osteoporosis (with vertebral compression fractures) were found in adults born with VLBW, implying increased future fracture risk [67]. However, Kuitunen et al. showed a lower overall childhood fracture incidence for children born very premature or with ELBW as compared to children born full term with normal birth weight [68]. These findings may be due to improvement in neonatal intensive care and early nutrition. Overall available data highlight that this population may be at increased risk of impaired bone accrual, but that there are insufficient data to recommend routine monitoring of bone density by dual-energy X-ray absorptiometry but to counsel on prevention of osteoporosis [67].

Despite advances in neonatal nutrition, MBDP remains prevalent in premature infants due to inadequate mineral accretion ex utero. Preventing and treating MBDP can prevent subsequent rickets and pathologic fractures. First-tier laboratory screening with serum Ca, phosphorus, and ALP helps identify infants at risk. If these show abnormalities, second-tier laboratories can include PTH and/or concurrent serum and urine phosphorus and creatinine for determination of TRP and differentiate between Ca or PO4 deficiency as a primary etiology to guide appropriate treatment. Additional research into optimal mineral supplementation for prevention and treatment with goals to improve long-term bone health outcomes is needed to improve the evidence base for future treatment guidelines.

The authors would like to thank Ms. Akane Hawpe Gamage for administrative support with the formatting of the paper and Meg Begany, RD, for careful review of the paper.

Jennifer L. Miller’s spouse is majority owner of Element Bars, Inc., a snack food company. Jennifer L Miller has served on an advisory board for Ipsen Biopharmaceuticals. The remaining authors have no conflicts of interest to declare.

There is no relevant funding source for this manuscript.

M.G., A.P.A., S.A.B., A.C., A.D.-T., S.K., J.L.M., M.-E.R., and L.A.D. contributed to the conception, wrote first drafts of sections, edited subsequent drafts, reviewed the manuscript, and approved the final version for submission.