点击显示 收起
1 From the US Department of Agriculture Agricultural Research Service Children’s Nutrition Research Center, Baylor College of Medicine, Houston (AL); the Department of Neonatology, Hôpital Edouard Herriot, Lyon, France (AL, BLS, and OC); and the Genetics Unit, Shriners Hospital for Children and McGill University, Montreal, (FHG)
2 The contents of this publication do not necessarily reflect the views or policies of the US Department of Agriculture, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government. 3 Supported by Wyeth Nutritionals International Inc, Philadelphia, and in part by the Institut Candia, Paris (to AL), and the Shriners of North America. 4 Address reprint requests to A Lapillonne, Department of Neonatology and Nutrition, Hopital Saint-Vincent de Paul, 82 Avenue Denfert Rochereau, 74014 Paris, France. E-mail: alexandre.lapillonne{at}svp.ap-hop-paris.fr.
ABSTRACT
Background: Because recent data on the effects of mineral concentrations in preterm infant formula on bone mineralization are lacking, recommendations for the mineral content of preterm infant formula differ greatly between committees.
Objective: The goal of the study was to assess the effects of an isocaloric, nutrient-enriched preterm formula, which was fed from the age when full enteral feedings were tolerated through expected term, on bone mineralization in preterm infants.
Design: We conducted a prospective, randomized, double-blind study in healthy, preterm infants (gestational age of 28–32 wk) who were fed either a control preterm formula (n = 20) or an isocaloric, nutrient-enriched preterm formula (n = 21) until 3 mo of age (ie, approximate expected term). Serum calcium indexes were taken throughout the study, and bone mass was determined by using dual-energy X-ray absorptiometry at hospital discharge and expected term.
Results: A total of 37 infants (experimental formula, n = 19; control formula, n = 18) completed the study. Compared with control subjects, infants fed the experimental formula had 25% and 40% higher intakes of calcium and phosphorus, respectively. Serum calcium, phosphorus, osteocalcin, and alkaline phosphatase concentrations and urinary collagen type I cross-linked N-telopetide concentrations were not significantly different between the groups at any time point. The bone mineral content of infants fed the experimental formula was 23% (P = 0.039) and 35% (P = 0.002) higher at hospital discharge and expected term, respectively.
Conclusions: Bone mineralization at hospital discharge and expected term was significantly higher in preterm infants fed the isocaloric, nutrient-enriched formula than in those fed control formula. Continuation of the experimental formula beyond hospital discharge, through expected term, further improved bone mineralization.
Key Words: Bone mineralization • body composition • bone mineral content • mineral intake
INTRODUCTION
Premature infants are known to be at risk of developing metabolic bone disease, and this risk is negatively associated with the birth weight of the infant (1). The skeletal mineral deficit accumulates between birth and term and leads to a lower bone mineral content (BMC) in premature infants at expected term than in term infants at birth (2–4).
Low dietary mineral intake is recognized as the major cause of metabolic bone disease in premature infants. Metabolic balance study results have shown that the major determinants of mineral retention for enterally fed premature infants, particularly for calcium, are mineral intake and mineral absorption (5, 6). On the basis of estimated intrauterine accretion rates or estimated requirement by factorial approach, which accounts for estimated intestinal absorption, urine and skin losses, and tissue increment, expert committees recommend routine supplementation with both calcium and phosphorus for premature infants fed human milk or the use of a high-mineral-content formula during the hospital stay (6–10). However, these recommendations vary greatly: from 70 to 192 mg/100 kcal for calcium (which is equivalent to 56–155 mg/100 mL formula at a caloric concentration of 81 kcal/100 mL), and from 50 to 117 mg/100 kcal for phosphorus (which is equivalent to 40–95 mg/100 mL formula at a caloric concentration of 81 kcal/100 mL).
Several randomized trials investigated the effects of mineral supplementation of mother’s milk on BMC (7–18), and other, nonrandomized studies compared the bone mass of preterm infants fed formula with that of preterm infants fed human milk (3, 14, 16, 17, 19, 20). The effects on bone mass in premature infants fed preterm formulas with high or low mineral content have surprisingly been investigated, to our knowledge, in only 2 randomized studies (21, 22). Both studies were, however, limited with respect to the number of infants studied and the method of assessing bone mass (ie, single-photon absorptiometry).
If one of the major goals of mineral intake is to achieve tissue accretion comparable to in utero accretion and, therefore, to achieve normal bone mass at expected term, the calcium and phosphorus requirements of premature infants involve not only the total mineral intake, but also the duration of high mineral intake. Because most premature infants are growth restricted and have low bone mass at hospital discharge, the desirability of promoting catch-up growth and bone mass in preterm infants after hospital discharge has been highlighted by several authors (23, 24). Several studies compared high-calorie, nutrient-enriched formula with term formula from hospital discharge through expected term or beyond and found some early advantages in growth or bone mass with the high-calorie, nutrient-enriched formula (25–28).
We hypothesized that providing an isocaloric, nutrient-enriched preterm formula increases bone mineralization during the hospital stay and that continuing this formula up to expected term would further improve bone mass in preterm infants. To address this issue, we performed a randomized, double-blind, longitudinal study in very-low-birth-weight infants and compared the effects of an isocaloric, nutrient-enriched preterm formula with those of a control preterm formula on BMC and serum indexes of calcium homeostasis as primary outcomes and on growth as a secondary outcome.
SUBJECTS AND METHODS
Subjects
We conducted a prospective, controlled, randomized study in very-low-birth-weight infants fed either experimental or control formula from the time full feeds were tolerated through 3 mo of age (ie, approximate expected term). The study had a parallel, double-blind design and was conducted in a single center with inpatients and outpatients. Preterm infants were recruited from the neonatal intensive care unit of the Edouard Herriot Hospital in Lyon, France, over a 12-mo period.
Infants (gestational age of 28–32 wk; birth weight of 1000–1600 g) whose size was appropriate for gestational age, who were medically stable, and whose mothers had decided against breastfeeding were eligible for inclusion, provided that full formula feeding (ie, no parenteral nutrition and >150 mL formula · kg–1 · d–1) was established by 3 wk of age. Gestational age was determined from the date of the mother’s last menstrual period and was confirmed by early ultrasound and clinical evaluation at birth (29). Eligible infants were assigned randomly in a doubly masked fashion to either the control or the experimental formula as the sole source of food until expected term.
Exclusion criteria included significant respiratory, neurologic, renal, genetic, cardiovascular, hepatic, or gastrointestinal disease; systemic infections; grade 3 or 4 intraventricular hemorrhage; use of corticosteroids or diuretics within 1 wk of enrollment; mechanical ventilation or oxygen requirement of >25%. Infants who did not adhere to the study requirements were withdrawn from the study and, regardless of the reason for withdrawal, not replaced. Reasons for withdrawal included lack of feeding of study formula for >48 consecutive hours or >96 cumulative hours and the development of any exclusion criteria, including development of sepsis, as documented by a positive blood culture.
The project was approved by the local ethics committee and the French Ministry of Health. All parents were informed of the details of the study, and written consent was obtained in all cases.
Diets
Enteral feeding of all infants was started during the first week of life with pooled, pasteurized, term breast milk (Lyon Lactarium, Lyon, France), as usual in the unit. Subsequently, those whose parents consented to enrollment were randomly assigned to one of the study formulas, which were fed exclusively until expected term. The formulas, which were especially prepared for the study, were ready to feed. Both contained 81 kcal/dL (339 kJ/dL) and had a common all-vegetable fat blend but differed in protein, mineral, and vitamin contents. The compositions of the formulas are shown in Table 1.
View this table:
TABLE 1. Compositions of the study formulas (per 100 mL)1
Infants received either experimental or control formula through nasogastric tubes, orogastric tubes, or nipple, as clinically indicated. Full enteral feeding (160 mL · kg–1 · d–1) of the assigned study formula was maintained while the infants were tube-fed during the inpatient phase of the study. Once the infants could nipple-feed, the formulas were fed ad libitum. The daily formula intake of each infant during the inpatient phase was recorded by the research nurse. All infants received 800 IU vitamin D/d (Sterogyl, Saint-Amant-Tallende, France) enterally until expected term. At discharge, parents were advised about feeding in a uniform manner. The goal was to ensure that each infant would be allowed to consume the amount that he or she desired. Formulas were provided to parents without restriction. During the outpatient phase, daily formula intake, which was determined by the difference in prefeeding and postfeeding volumes of each feeding unit offered, was recorded by the parents or guardians in an intake diary.
Data collection
A single research nurse monitored the infants enrolled during hospitalization. They were weighed daily with the use of a calibrated infant scale to the nearest 5 g. Crown-heel length was determined weekly by using an infant measuring board to the next succeeding 0.1 cm (Wyeth Nutritionals International Inc, Philadelphia). Head circumference was also measured weekly at the largest occipitofrontal circumference with the use of a nonstretchable paper tape to the next succeeding millimeter (Wyeth Nutritionals International Inc). After discharge, weight was measured weekly, and weight, length, and head circumference were measured at expected term. Anthropometric measurements were performed in duplicate by the same research nurse and then averaged.
Body composition was determined as previously described (30, 31) at hospital discharge and at expected term by using dual-energy X-ray absorptiometry (DXA) with a whole-body single-beam scanner (Hologic QDR 1000W, software 5.47; Hologic Inc, Waltham, MA). With this equipment, mean CVs for fat mass, lean mass, and whole-body BMC are <5% (30). Spontaneous sleep was obtained in all cases in the supine position without sedation. All infants were measured naked, and room temperature was maintained between 24 and 25°C; a heat lamp was used as needed to maintain a satisfactory environmental temperature. The average duration of each measurement was 10 min. BMC, lean mass, and fat mass were not measured at the start of the study because of the limitations of DXA for assessing body composition in small infants (32).
At enrollment, hospital discharge, and expected term, urine samples were collected by using adhesive bags, and blood samples were collected by venipuncture. A Hitachi 747 analyzer (Roche Molecular Biochemicals, Meylan, France) was used to measure plasma urea nitrogen, sodium, potassium, chloride, and albumin concentrations. Serum calcium, phosphorus, and alkaline phosphatase concentrations were measured by using a colorimetric technique and Monarch equipment (Instrumentation Laboratories, Lexington, MA). Serum concentrations of 25-hydroxyvitamin D, 1,25-dihydroxyvitamin D, and osteocalcin were measured as previously described (33) by using a radioligand assay, a specific receptor assay, and a radioimmunoassay, respectively. The urinary collagen type I cross-linked N-telopeptide (NTX) concentration in a single urine sample collected between 0800 and 1100 was determined as previously described, and urinary NTX excretion was expressed normalized to creatinine (nmol bone collagen equivalents/mmol creatinine) (34).
Statistical methods
Statistical analysis was performed by using MINITAB 13.32 (Minitab Inc, State College, PA). Sample sizes were estimated to allow detection of a 1-SD difference in BMC [ie, 18.3 g (30)] between the 2 formula-fed groups at term (power = 80%, P < 0.05). With the anticipation that 20% of those enrolled would not complete the study, 20 infants were enrolled in each group. The association between sex and group assignment was tested by using the chi-square test. Equality of variance for each continuous variable was tested by using the Bartlett test. All values are reported as means ± SDs throughout the text. Differences in demographic and biochemical variables were detected by using an unpaired t test. Other outcome variables were analyzed by using analysis of covariance with the general linear model procedure of MINITAB; study formula was the independent variable in all models, and other variables (ie, sex, postnatal age at each time point, and data at the start of the study) were covariates when appropriate. Only those that were statistically significant (P < 0.05) were included in the final model. Differences between groups at P < 0.05 were considered significant.
RESULTS
Clinical characteristics
Forty-one infants were randomly assigned to receive either the control (n = 20) or the experimental (n = 21) preterm formula. Four infants were excluded because of early food intolerance (control formula, n = 1; experimental formula, n = 1); worsening of bradycardia, apnea, and cyanosis unrelated to study feedings (experimental formula, n = 1); and family request (control formula, n = 1). One infant (experimental formula) who completed the study received one dose of furosemide (2 mg/kg) at 1 mo of age for edema; there were no other concerns with this infant, and his serum albumin was similar to the others; the data from this infant were included in the final data set because exclusion of these data from the statistical analyses did not affect the results. Thus, 18 and 19 infants fed the control formula and the experimental formula, respectively, completed the study through expected term.
The clinical characteristics of these infants are shown in Table 2. Intakes of energy, protein, calcium, and phosphorus were calculated from the volume of intake and the concentration of each in the formulas. Formula intake during the study did not differ significantly between the 2 groups, and as a result, the 2 groups did not differ significantly in energy intake. The mean volume intakes of infants fed the control and the experimental formulas did not differ significantly at hospital discharge (123 ± 36 and 124 ± 18 mL · kg–1 · d–1, respectively) or expected term (162 ± 27 and 156 ± 22 mL · kg–1 · d–1, respectively). The volume intakes at discharge were temporarily lower than 160 mL · kg–1 · d–1 because the policy of the unit is to favor hospital discharge as soon as the infants drink enough milk to maintain their body weight for 3 d. The infants fed the experimental formula had significantly higher mean protein intakes than did those fed the control formula at discharge (2.72 ± 0.39 compared with 2.46 ± 0.34 g · kg–1 · d–1; P = 0.033), but mean protein intakes at expected term did not differ significantly between the groups (3.42 ± 0.47 and 3.24 ± 0.54 g · kg–1 · d–1, respectively). Compared with the infants fed the control formula, those fed the experimental formula also had significantly higher mean calcium [discharge: 124 ± 18 compared with 98 ± 14 mg · kg–1 · d–1 (P < 0.001); expected term: 156 ± 22 compared with 130 ± 22 mg · kg–1 · d–1 (P = 0.001)] and phosphorus [discharge: 74.3 ± 10.7 compared with 52.2 ± 7.2 mg · kg–1 · d–1 (P < 0.001); expected term: 93 ± 13 compared with 69 ± 11 mg · kg–1 · d–1 (P < 0.001)] intakes.
View this table:
TABLE 2. Clinical characteristics of the infants1
Growth and bone mineral content
There were no significant differences in weight, length, or head circumference between the 2 groups at the start of the study, at hospital discharge, or at expected term, except for weight at expected term, which was 6% higher in the experimental group than in the control group (P = 0.007) (Table 3). Mean weight gain between entry into the study and expected term was significantly higher in the experimental formula group than in the control formula group [2727 ± 565 (95% CI: 2455, 2999) compared with 2337 ± 332 g (95% CI: 2172, 2502 g); P = 0.016], whereas the 2 groups did not differ significantly in mean gains in length [11.3 ± 2.2 (95% CI: 10.3, 12.4) and 10.7 ± 1.7 cm (95% CI: 19.8, 11.5 cm), respectively] or head circumference [9.4 ± 1.4 (95% CI: 8.7, 10.0) and 8.5 ± 1.7 cm (95% CI: 7.6, 9.3 cm), respectively].
View this table:
TABLE 3. Postnatal age and weight, length, and head circumference measurements of the infants at the start of the study, at discharge from the hospital, and at expected term1
Body-composition data are summarized in Table 4. Because body-composition measurements at entry into the study were not available and because BMC, fat mass, and lean mass are highly correlated with weight (30), differences between groups at hospital discharge and at expected term were assessed by using initial weight rather than initial BMC, fat mass, or lean mass as the covariate. Whole-body BMC, fat mass, and lean mass were expressed as absolute values and as g/kg body wt. Because the BMC, fat mass, and lean mass of neonates at hospital discharge and term are highly correlated with body weight (30), expressing the results as g/kg body wt addresses the effects of the formula independently of changes in body weight.
View this table:
TABLE 4. Bone mineral content (BMC) and body composition expressed as g and as g/kg body wt at discharge from the hospital and at expected term1
The absolute BMC of the experimental group was significantly higher than that of the control group both at discharge (P = 0.039) and at expected term (P = 0.002); on average, the infants fed the experimental formula had 23% and 35% higher BMC at hospital discharge and at expected term, respectively, than did those fed the control formula. The significantly higher BMC per kilogram of body weight at both times indicates higher bone mass in the infants fed the experimental formula, independent of the effect on weight (P < 0.05). Because the experimental group was slightly younger at the start of the study, we also tested differences in BMC both at hospital discharge and at expected term after adjusting for the number of days the infants were fed their assigned formula, and the differences between the 2 groups remained significant. The mean gain in BMC between hospital discharge and expected term was significantly higher in the experimental formula group than in the control formula group [53.6 ± 21.7 (95% CI: 45.8, 64.6) compared with 38.9 ± 17.7 g (95% CI: 29.4, 48.3 g); P = 0.037].
To compare the postnatal body composition of preterm infants with that of fetuses in utero, we compared body composition assessed at discharge and at expected term with normative values that were obtained in appropriate-for-gestational-age infants of similar weight and measured by using DXA within the first 48 h after birth (30). Fourteen infants with a mean birth weight of 2090 ± 113 g (95% CI: 2025, 2156 g) were used for determining normal body composition at hospital discharge. Their mean BMC, fat mass, and lean mass were 26.8 ± 7.7 (95% CI: 22.3, 33.2), 299 ± 55 (95% CI: 267, 331), and 1765 ± 91 g (95% CI: 1712, 1817 g), respectively. Twelve infants with a mean birth weight of 3649 ± 406 g (95% CI: 3391, 3908 g) were used for determining normal body composition at expected term. Their BMC, fat mass, and lean mass were 70.7 ± 21.8 (95% CI: 56.8, 84.5), 683 ± 164 (95% CI: 579, 787), and 2896 ± 263 g (95% CI: 2728, 3063 g), respectively. The body compositions of the 2 study groups expressed as percentages of normative values are shown in Figure 1. Low BMC was found in the control group and the experimental group (35% and 45% of the expected BMC, respectively) at hospital discharge, but low BMC was found only in the control group at expected term (ie, 70% of the expected BMC). Both groups had normal body composition at hospital discharge and expected term.
FIGURE 1.. Body composition [ie, bone mineral content (BMC), fat mass, and lean mass] assessed by using dual-energy X-ray absorptiometry at hospital discharge and at expected term of premature infants who were randomly assigned to receive either a control formula or an isocaloric, nutrient-enriched preterm formula (experimental formula) through expected term. The body-composition data for the 2 study groups were compared with previously reported normative values that were obtained by measuring appropriate-for-gestational-age infants within 48 h after birth (adapted from reference 30). *,**Significantly different from normative value (unpaired t test): *P < 0.001, **P = 0.01.
On the basis of the gain in BMC between hospital discharge and expected term and assuming a calcium concentration of 0.32 g/g BMC, we calculated that the mean estimated calcium accretion between hospital discharge and expected term was 16% higher in the experimental group than in the control group [69.8 ± 23.5 (95% CI: 58.2, 81.5) compared with 59.6 ± 29.0 mg · kg–1 · d–1 (95% CI: 44.1, 75.1 mg · kg–1 · d–1)], but this difference was not significant. On the basis of the estimated calcium accretion and the calcium intake deduced from the formula intake, we also estimated that the mean calcium retention of the experimental group was similar to that of the control group [46.6 ± 22.7 (95% CI: 34.5, 58.7) and 43.6 ± 14.7% (95% CI: 36.3, 50.9%), respectively]. These estimations of calcium accretion and retention suggest that differences in BMC are likely due to differences in calcium intake rather than differences in calcium absorption or retention.
The 2 groups did not differ significantly in either fat or lean mass at hospital discharge. At expected term, absolute fat mass and absolute lean mass in the experimental group were significantly higher than those in the control group (P = 0.014 and P < 0.001, respectively). However, the 2 groups did not differ significantly in either fat mass or lean mass expressed as g/kg body wt at either time. In addition, no significant difference in the gain in either fat mass or lean mass was found between the 2 groups (data not shown).
Biochemical variables
Laboratory variables are summarized in Table 5. The groups did not differ significantly in mean serum concentrations of calcium, phosphorus, 1,25-dihydroxyvitamin D, alkaline phosphatase, or osteocalcin at study entry, hospital discharge, or expected term. The groups also did not differ significantly in mean 25-hydroxyvitamin D concentrations at study entry or hospital discharge; however, mean 25-hydroxyvitamin D concentrations at expected term were significantly higher in the experimental formula group than in the control group (P = 0.002). There were no differences in socioeconomic class, ethnicity, or vitamin D supplement use to explain this difference in plasma 25-hydroxyvitamin D concentrations, and, therefore, this was attributed to the difference in the vitamin D content of the formulas. The groups also did not differ significantly in mean urinary NTX excretion at study entry, hospital discharge, or expected term.
View this table:
TABLE 5. Serum concentrations of calcium, phosphorus, 25-hydroxyvitamin D [25(OH)D], 1,25-dihydroxyvitamin D [1,25(OH)D2], alkaline phosphatase, and osteocalcin and urinary excretion of cross-linked N-telopeptides of type I collagen (NTX) at the start of the study, at discharge from the hospital, and at expected term1
DISCUSSION
The study reported here is the only randomized, controlled study of the effects of feeding very-low-birth-weight infants an isocaloric, nutrient-enriched preterm formula from full enteral feeding through term on bone mineralization, as assessed by using DXA. The data show that healthy preterm infants fed a formula containing 25% more calcium and 40% more phosphorus than in the control formula during hospitalization had significantly higher BMC at hospital discharge, but the formula did not prevent low bone mass at the time of hospital discharge. However, continuing this formula from hospital discharge to the equivalent of term resulted in further improvement in BMC, which led to normal bone mass at expected term.
The difference in bone mass between the 2 groups was probably due to the difference in mineral intake. Standard balance studies have shown that, within a few weeks after birth, premature infants can absorb and retain calcium and phosphorus in sufficient amounts to approximate estimated in utero accretion rates (6). These balance studies have also shown that net calcium absorption is a linear function of intake for a range of calcium intake varying from 40 to 130 mg · kg–1 · d–1 (5, 6). Calcium retention data in preterm infants obtained by using either standard balance or stable-isotope techniques indicate that the mineral preparations in bovine milk-based preterm infant formulas are usually well absorbed and that calcium absorption and retention are between 50 and 170 mg · kg–1 · d–1 and between 34% and 74%, respectively (6). The calcium retention calculated from the DXA results should be interpreted with caution. Interestingly, however, we observed that the estimated percentage retention values were not significantly different between the 2 groups, but the estimated retention value in mg · kg–1 · d–1 was 16% higher in the experimental group than in the control group. This suggests that the difference in bone mass was due to the higher mineral content of the experimental formula. The difference in bone mass was probably not due to the difference in vitamin D status, because vitamin D is not a major determinant of calcium absorption in preterm infants (5, 35). It is, however, feasible that nutrients other than calcium and phosphorus may have stimulated bone growth, and the higher protein intake of the supplemented infants may have had an additional beneficial effect on bone mineralization (36).
The mechanism by which the experimental formula resulted in higher BMC remains uncertain. Bone histomorphometry data suggest that increased bone resorption, not impaired formation, underlies the development of osteopenia in preterm infants (37). In the present study, however, the 2 groups did not differ significantly in either serum osteocalcin concentration, a marker of bone formation, or in urinary NTX excretion, a marker of bone resorption (38), at either hospital discharge or expected term. Thus, whether the higher BMC reflects reduced bone resorption or increased bone formation is unclear.
The effects on bone mass in premature infants of the mineral concentration of preterm formula fed during the hospital stay have been investigated by 2 randomized studies (21, 22). Chan et al (21) evaluated bone mass in premature infants fed either 1 of 3 formulas differing only in mineral content or their mother’s milk until the infants’ weight reached 1900 g. The human milk group had lower bone mass than did the 3 formula groups, but BMC did not differ significantly between the formula groups, despite the fact that the high-mineral formula contained 20% more calcium and 40% more phosphorus than did the low-mineral formula. Horsman et al (22), on the other hand, showed that mineral accretion in preterm infants was significantly higher in infants fed a supplemented formula containing 90% more calcium and 37% more phosphorus than in the standard formula. The absence of an effect of mineral supplementation on bone mass in the study by Chan et al (21) was probably due to the relatively low difference in mineral content between the formulas but may also have been due to the limited number of infants included and the method of assessing bone mass, which included only a small part of the skeleton.
Recommendations for the mineral content of preterm formula differ greatly depending on committees. The American Academy of Pediatrics Committee on Nutrition (8) recommends 175 mg Ca/100 kcal, which calculates to 141 mg/100 mL formula at a caloric concentration of 81 kcal/100 mL. The Committee on Nutrition of the Preterm Infant of the European Society of Pediatric Gastroenterology and Nutrition (9) recommends 70–140 mg Ca/100 kcal, which calculates to 56–113 mg/100 mL formula at a caloric concentration of 81 kcal/100 mL.
More recently, the Expert Panel of the Life Sciences Research Office of the American Society for Nutritional Sciences recommended that the calcium concentration of preterm infant formula range between 100 and 150 mg/100 mL (123–185 mg/100 kcal) (10). Interestingly, this panel also highlighted the need for further studies to confirm the minimum recommendation and suggested that a minimum recommendation of 80 mg Ca/100 mL (100 mg Ca/100 kcal) may be appropriate. The present study found a beneficial effect on bone mineralization of a formula with a calcium content of 100 mg/100 mL (ie, the experimental formula) compared with 80 mg/100 mL (ie, the control formula). Because no clinical or metabolic complications were observed, the results of this study support the actual minimum calcium content of preterm formulas recommended recently by the Life Sciences Research Office panel (10).
In our study, the 2 groups of infants did not differ significantly in growth at hospital discharge. This is somewhat surprising because feeding a protein-enriched formula positively influences growth and protein utilization in preterm infants during the hospitalization phase (39). The absence of a significant difference may have been due to the relatively small difference in protein content between the 2 study formulas. However, our finding that continuing a protein-enriched preterm formula after discharge from the hospital promotes early weight gain supports the concept that preterm infants respond positively to higher intakes of protein during the first weeks after discharge from the hospital. This is an important finding because most premature infants are growth restricted at hospital discharge (23).
The desirability of promoting catch-up growth in preterm infants after hospital discharge has been highlighted by several authors (23, 24). Because formulas designed to meet the nutritional needs of term infants do not adequately support continued catch-up growth, formulas designed specifically for feeding preterm infants after hospital discharge have been studied (25–28, 40, 41). These postdischarge formulas provide more protein, more energy, and more of several minerals and micronutrients than do standard term formulas. In most studies, infants fed these postdischarge formulas experience some early advantages in growth compared with infants fed a standard term formula. With regard to bone mass, these studies have shown conflicting results: some found a positive effect of postdischarge dietary practices on bone mineralization (19, 27, 41), whereas others found no effects (13, 40). Cooke et al (27) found that premature infants fed a preterm formula after hospital discharge have higher weight gain and higher BMC at expected term than do fed a term formula. On the other hand, De Curtis et al (40) found no significant difference in growth and bone mass between a group of preterm infants fed a postdischarge formula and a group fed a standard term formula. This apparent discrepancy was probably due to the difference in mineral content of the supplemented formula [108 mg Ca/100 mL and 54 mg P/100 mL in the study by Cooke et al (27) compared with 80 mg Ca/100 mL and 40 mg P/100 mL in the study by De Curtis et al (40)].
In our study, we showed that providing an enriched formula during the hospital stay and continuing this formula from hospital discharge to the equivalent of term resulted in higher bone mineral mass. Combined with the data of Wauben et al (19), which showed a positive effect of postdischarge dietary practices on bone mineralization at 6 mo corrected age, our findings support the concept that the calcium and phosphorus requirements of premature infants involve the amount of mineral intake and the duration of high mineral intake. Furthermore, the BMC of our experimental group was similar to that of normal term infants of the same weight at birth, which suggests that the isocaloric, nutrient-enriched preterm formula used during both the in- and outpatient phase of the study supports normal bone mineralization in preterm infants. Whether or not this formula will support normal bone mineralization in sicker or smaller premature infants needs further investigation.
Our study also differs from previous studies reporting the effects of postdischarge feeding practices on growth because the energy contents of the 2 formulas in our study were similar. Therefore, the difference in weight at expected term probably resulted from the difference in protein content, which suggests that protein intake is a limiting factor for growth between discharge and expected term. Interestingly, the preterm infants who were fed the experimental formula had absolute fat and lean masses at term that were significantly higher than those of the control group, but neither fat mass nor lean mass expressed as g/kg body wt differed significantly between the groups. This is somewhat surprising because the composition of weight gain is known to be dependent on the protein-to-energy ratio in the diet, with a higher ratio resulting in deposition of more lean mass than fat mass (39). However, the difference in protein intake between the groups in our study was probably not sufficient to result in significant changes in body composition.
In summary, in the present study, an isocaloric, nutrient-enriched preterm formula fed through hospital discharge significantly improved the bone mineralization of preterm infants, without any detectable complications. Furthermore, continuing this preterm formula beyond hospital discharge, through expected term, increased weight gain and further improved bone mineralization. These findings show that preterm infants are highly responsive to mineral intake during the early postdischarge period.
ACKNOWLEDGMENTS
We acknowledge Rose Travers (Shriners Hospital, Montreal) for help in assessing NTX concentrations, Bernadette Reygrobellet (Edouard Herriot Hospital) for help in collecting the data, Gyslaine Claris and Jean-Charles Picaud (Edouard Herriot Hospital) for their contributions in the follow-up of the infants, and Eric Lien (Wyeth Nutritionals International Inc, Philadelphia) and William C Heird (USDA/ARS Children’s Nutrition Research Center, Houston) for reviewing the manuscript.
AL and BLS designed and implemented the study. AL, BLS, and OC were responsible for recruiting the subjects, providing medical care during the study, evaluating the statistics, interpreting the data, and writing the manuscript. FHG was responsible for assaying calcium, phosphorus, 25-hydroxyvitamin D, 1,25-dihydroxyvitamin D, alkaline phosphatase, osteocalcin, and NTX; analyzing the laboratory data; and otherwise assisting in the preparation of the manuscript. None of the authors had any conflicts of interest.
REFERENCES