Objective. At hospital discharge, preterm infants may have low body stores of nutrients, deficient bone mineralization, and an accumulated energy deficit. This double-blind, randomized study evaluated the growth of premature infants with birth weights <1800 g who were fed a 22 kcal/fl oz nutrient-enriched postdischarge formula (PDF) or a 20 kcal/fl oz term-infant formula (TF) from hospital discharge to 12 months' corrected age (CA).
Methods. Infants were randomized to PDF or TF a few days before hospital discharge with stratification by gender and birth weight (<1250 g or ≥1250 g). The formulas were fed to 12 months' CA. Growth was evaluated using analysis of variance controlling for site, feeding, gender, and birth weight group. Interaction effects were also assessed. Secondary analyses included a repeated measures analysis and growth modeling.
Results. One hundred twenty-five infants were randomized; 74 completed to 6 months' CA and 53 to 12 months' CA. PDF-fed infants weighed more than TF-fed infants at 1 and 2 months' CA, gained more weight from study day 1 to 1 and 2 months' CA, and were longer at 3 months' CA. There were significant interactions between feeding and birth weight group—among infants with birth weights <1250 g, those fed PDF weighed more at 6 months' CA, were longer at 6 months' CA, had larger head circumferences at term 1, 3, 6, and 12 months' CA, and gained more in head circumference from study day 1 to term and to 1 month CA. The repeated measures and growth modeling analyses confirmed the analysis of variance results. The PDF formula seemed to be of particular benefit for the growth of male infants. Infants fed the PDF consumed less formula and had higher protein intakes at several time points. Energy intakes, however, were not different.
Conclusions. Growth was improved in preterm infants fed a nutrient-enriched postdischarge formula after hospital discharge to 12 months' CA. Beneficial effects were most evident among infants with birth weights <1250 g, particularly for head circumference measurements.
- VLBW =
- very low birth weight •
- CA =
- corrected age •
- PDF =
- postdischarge formula •
- TF =
- term formula
Although adequate nutrition is critical in the management of small preterm infants, no feeding standard exists that is comparable to the breast milk standard for term infants. Present recommendations are designed to provide nutrients to approximate the rate of growth and composition of weight gain for a normal fetus of the same postconceptional age without inducing metabolic stress.1However, many preterm infants who retain nutrients at these levels remain growth retarded relative to a fetus or newborn of comparable postconceptional age. The delay in catch-up growth is attributable in part to the slow initial growth during hospital stay and to the initial postnatal weight loss, which may be at least 10% of body weight2 and which may not be regained for 2 weeks or more in infants with very low birth weights (VLBW).2,3
Typical hospital discharge weights for preterm infants in the United States are 1800 to 2000 g. At discharge, preterm infants may have low body stores of nutrients,4 deficient bone mineralization,5 and an accumulated energy deficit of 3780 to 5460 kJ.2 At 40 weeks' postmenstrual age, preterm infants are generally smaller than term infants, and they are projected to have higher nutrient requirements. They may remain growth retarded for several years,6–11 have higher rates of childhood morbidity and mortality,9,12 have undermineralized bones,13 have failure to thrive,14 and often experience delays in neurodevelopment15–17 that can persist to school age.18–19
Aggressive nutritional management during hospitalization promotes earlier onset of and more rapid rates of postnatal growth of preterm infants,16,20,21 which may be associated with later developmental advantages.19,22,23 Studies by Kashyap et al24–26 have demonstrated that hospitalized preterm infants can effectively use, without metabolic stress, higher concentrations of protein and energy than are typically fed. Less is known about the nutritional needs of the postdischarge preterm infant relative to the term infant. Preterm infants may be discharged from the hospital on breast milk, formula designed for term infants, or postdischarge formula designed for preterm infants. The potential benefit of postdischarge formulas is being increasingly recognized.1 Improved growth27–29 and bone mineralization5,30 have been reported in preterm infants fed nutrient-enriched formulas after hospital discharge.
The purpose of the present study was to determine if feeding a nutrient-enriched formula designed for the nutritional needs of preterm infants after hospital discharge would result in improved growth compared with infants fed a standard formula designed for term infants.
METHODS AND MATERIALS
A controlled, double-blind, randomized, parallel study was designed to evaluate the growth of premature infants fed 1 of 2 different formulas from hospital discharge to 12 months' corrected age (CA). Infants were randomized to 22 kcal/fl oz postdischarge formula (PDF) or 20 kcal/fl oz term formula (TF) 2 to 4 days before hospital discharge, at which time body weight was expected to be 1800 to 2300 g. The first day of study formula feeding was designated as study day 1. Infants were measured at study visits that corresponded to term (40 weeks after last menstrual period), and to 1, 2, 3, 6, 9, and 12 months' CA. Blood samples were obtained at term, 3 months', and 9 months' CA. Dietary intakes and measures of formula tolerance were recorded for 3 days before each study visit in booklets provided to parents.
This protocol was approved by the Institutional Review Board for the Protection of Human Subjects at each institution.
Preterm infants were recruited from hospitals in 6 cities (Los Angeles, CA, Tampa, FL, St Petersburg, FL, Kansas City, MO, Iowa City, IA and Ft Lauderdale, FL). Inclusion criteria included birth weight ≤1800 g, gestational age <37 weeks, and previous parental decision not to provide breast milk for their infant. Infants with severe bronchopulmonary dysplasia, cardiac, respiratory, gastrointestinal, or other systemic diseases were excluded. At the time study feedings were initiated, infant weights were 1635 to 2715 g (PDF), and 1640 to 2800 g (TF), and their postnatal ages were >1 week. There were no restrictions on the type of feeding before study entry. Parental informed consent was obtained before infants entered the study.
The randomization schedule, prepared by the Biostatistics Department of Ross Products Division, was stratified for gender and birth weight group (<1250 g, ≥1250 g). Participants who did not complete the study were not replaced.
Study formulas were the 22 kcal/fl oz nutrient-enriched PDF or a commercially available 20 kcal/fl oz TF. Both formulas were provided by Ross Products Division, Abbott Laboratories, Columbus, Ohio. The PDF provided higher quantities of protein, most minerals, and vitamins per 100 kcal than TF (Table 1). The ratio of whey to casein proteins in PDF was 50:50, and the fat blend consisted of 45% soy oil, 30% coconut oil, and 25% medium-chain triglyceride oil. The TF had a whey to casein ratio of 20:80, and its fat blend consisted of 60% soy oil and 40% coconut oil. Clinically labeled ready-to-feed formulas were provided in 4-fl oz bottles for inhospital use and in 32-fl oz cans for use after hospital discharge.
At the time of hospital discharge, parents were given a supply of formula and were instructed to feed only the study formula to 4 months' CA. From 4 to 12 months' CA, infants continued to receive the assigned formula, and parents were instructed to conform to the American Academy of Pediatrics recommendations regarding the inclusion of other foods.1 Clinic visits were scheduled to correspond to term, and to 1, 2, 3, 6, 9, and 12 months' CA.
At each clinic visit, infants were weighed nude and head circumference and length were measured. Each site was requested to have the same examiners obtain anthropometric measurements at each visit; measurements were made in duplicate. Measurements that did not meet these criteria were not included in the analyses. All centers were provided with O'Leary recumbent length boards (Ellard Instrumentation, Inc, Seattle, WA), head circumference measuring tapes (Similac Inser-Tape, Associated Visual Communications, Canton, OH) and weights for calibrating scales.
Blood Sampling and Analyses
At the term, 3, and 9 months' CA visits, blood samples (∼1 mL) were obtained using the Tenderfoot Premie heel incision device (International Technidyne Corp, Edison, NJ) or from a peripheral vein by venipuncture. Only nonhemolyzed samples were analyzed. Hemoglobin concentrations in blood obtained by heel incision were determined using the HemoCue AB (HemoCue, Inc, Mission Viejo, CA) while those obtained by venipuncture were determined using a Coulter STKS Automated CBC Spectrophotometer analyzer or a Model M430 Coulter Counter (Coulter Electronics, Hialeah, FL). Remaining blood was centrifuged and serum was transferred to polypropylene tubes and frozen. The samples were shipped on dry ice to Ross Products Division where serum levels of albumin, prealbumin, and retinol-binding protein were determined using the Behring Nephelometer 100 (Hoechst-AG, Frankfurt, Germany) and serum urea nitrogen was determined using the Abbott Spectrum EPX Clinical Chemistry analyzer (Abbot Laboratories, Irving, TX).
Growth data were analyzed at each time point using a 3-way analysis of variance controlling for site, feeding group, gender, birth weight group, and interactions between feeding group and gender, feeding group and birth weight group, gender and birth weight group, and feeding, gender, and birth weight group. When significant interactions were found (P <.05), least squares means adjusted for the other variables in the model were used for comparisons between feeding groups. Bonferroni adjustments to P values were made for multiple comparisons. A posthoc analysis that included age at each study visit was also done at each time point. Longitudinal analysis of the growth data were also done using a mixed model repeated measures analysis controlling for site, birth weight group, feeding group, gender, and visit (PROC MIXED, SAS v6.09 Enhanced, Cary, NC). A growth modeling approach was applied to control for variability in the timing of clinic visits and for missed visits. A 3-parameter Gompertz31 model was used to fit a curve for each subject, using all available data (PROC NLIN, SAS v6.09 Enhanced, Cary, NC). Values for the exact, protocol-specified anthropometric values were estimated from the individual curves and analyzed using the same mixed model repeated measures analysis as described above.
Formula intake data (volume, protein, and energy per kg body weight per day) were analyzed at each time point using 3-way analysis of variance controlling for site, birth weight group, gender, and interactions as described for growth. Blood biochemistries were analyzed at each time point using 1-way analysis of variance controlling for site.
In general, continuous data were evaluated using analysis of variance and categorical data using χ2 tests of association or exact tests. The null hypothesis was that growth of infants fed nutrient-enriched formula would be no greater than that of infants fed control formula. Hypothesis tests for growth were 1-sided with α<0.05 considered statistically significant. All other hypothesis tests were 2-sided 0.05 level tests.
One hundred twenty-five infants were randomized; 1 participant who was randomized never received study formula and 1 was not included in the analyses because of the diagnosis of a genetic disorder before hospital discharge. Birth anthropometrics, gestational age, percentage of infants with intrauterine growth restriction, ethnicity, gender distributions, and reasons for early exit from the study did not differ between infants fed PDF and those fed TF (Table 2). The postnatal ages of infants at study day 1, term, 1, 2, 3, 6, 9 and 12 months' CA were 41 ± 3, 70 ± 3, 103 ± 3, 132 ± 4, 164 ± 4, 247 ± 4, 329 ± 4, 438 ± 5 days (PDF) and 43 ± 4, 68 ± 3, 99 ± 3, 131 ± 3, 161 ± 4, 240 ± 3, 330 ± 4, 437 ± 5 days (TF) (mean ± standard error of mean). The postnatal ages were not different, except at the 6 months' CA visit (P = .04). However, including postnatal age at the study visits as an additional covariate in the analysis of variances did not significantly affect the results. Sample size varied at individual timepoints because of dropouts and missed clinic appointments.
Anthropometric measurements at study day 1 did not differ between infants fed PDF or TF. The PDF-fed infants had greater weight, length, head circumferences, and gains in these measures at several of the subsequent visits (Tables 3 and4), with confirmatory longitudinal analyses using repeated measures and growth modeling.
PDF-fed infants weighed significantly more at 1 and 2 months' CA, and gained more weight (g/d) from study day 1 to each of these visits. There were significant interactions between feeding and birth weight group; among infants with birth weights <1250 g, PDF-fed infants weighed more than TF-fed infants at 6 months' CA, and they gained more weight from study day 1 to 12 months' CA. There were also significant interactions between feeding, birth weight group, and gender; among male infants with birth weights <1250 g, those fed PDF weighed more at 12 months' CA than those fed TF. From study day 1 to term, there was a significant feeding by gender interaction; among male infants, those fed PDF gained more weight than those fed TF. The repeated measures analysis similarly demonstrated a significant interaction between feeding and birth weight group (P = .02); among infants with birth weights <1250 g, those fed PDF weighed significantly more (P < .001). The growth modeling analysis showed a significant overall feeding effect, with PDF-fed infants weighing significantly more than TF-fed infants (P = .044).
Infants fed PDF were significantly longer than those fed TF at 3 months' CA. At 6 months' CA, there was a significant interaction between feeding and birth weight group; among infants with birth weights <1250 g, those fed PDF were longer than those fed TF. The repeated measures analysis similarly demonstrated an overall greater length for PDF-fed infants (P = .004), and the growth modeling analysis showed a significant overall feeding effect, with PDF-fed infants having greater lengths than TF-fed infants (P = .008).
For head circumference, there were significant interactions between feeding and birth weight group at term, 1, 3, 6, and 12 months' CA; among infants with birth weights <1250 g, those fed PDF had larger circumferences than those fed TF (Fig 1). There were also interactions between feeding and birth weight group for head circumference gains (mm/d); among infants with birth weights <1250 g, gains in head circumference from study day 1 to term and to 1 month CA were greater for PDF-fed infants than for TF-fed infants. From study day 1 to 12 months' CA, there was a significant feeding by gender by birth weight group interaction; among male infants with birth weight <1250, those fed PDF had greater gains in head circumference than those fed TF. The repeated measures analysis demonstrated greater head circumference growth among PDF-fed infants with birth weights <1250 g (P = .015), and the growth modeling analysis demonstrated a significant feeding by birth weight interaction (P = .003), with PDF-fed infants with birth weights <1250 g having greater head circumferences than those fed TF (P = .004).
After hospital discharge, energy intakes from formula were not different among infants fed PDF versus TF (Table 5). However, the volume of formula intake was lower (P < .05) at 1 month CA and protein intake from formula was higher (P < .05) at term, 2, and 3 months' CA for the PDF-fed infants.
Serum levels of albumin and blood levels of hemoglobin did not differ between diet groups (Table 6). Serum levels of prealbumin and retinol-binding protein were higher in infants fed PDF; these differences were statistically significant at term and 9 months' CA for prealbumin and at term for retinol- binding protein. Serum urea nitrogen levels were consistently significantly higher (P ≤ .01) for PDF-fed infants.
Both formulas were well-tolerated. Stools of PDF-fed infants were darker than those of TF-fed infants at several time points.
During hospitalization, preterm infants, particularly those born with very low birth weights, are fed fortified human milk or preterm infant formulas to try to achieve weight gain and bone mineralization that approximate that of the reference fetus.1 These infants often are fed unfortified breast milk or term formulas after hospital discharge. This practice is associated with poor somatic growth, developmental delays, poor bone mineralization,30,32–34 and insufficient nutrient stores.35–37 Strategies that are often used to fortify feedings for formula-fed infants after hospital discharge may be inappropriate. Fortification with energy alone, for example, may not be sufficient because postdischarge preterm infants may need additional intakes of specific nutrients and higher nutrient to energy ratios to achieve optimal growth and bone mineralization. Concentrating term formulas to levels that would provide adequate levels of protein and calcium may result in an unacceptable increase in osmolality or in the concentrations of other nutrients.38
In the present study, preterm infants were fed either PDF, a nutrient-enriched 22 kcal/fl oz formula designed for preterm infants after discharge, or TF, a 20 kcal/fl oz standard formula for term infants. Infants fed PDF had improved growth compared with those fed TF, with the most significant beneficial effects seen among infants with birth weights <1250 g. Infants fed PDF also had greater gains in weight and head circumference compared with those fed TF, especially within the first 1 to 2 months after discharge. The early differences in gains suggest a particularly important effect of nutrient-enriched feedings during the early postdischarge period. Although a limitation of the present study is the significant loss to follow-up among infants who exited the study early, the improvements in growth were found using 3 separate statistical approaches. Other investigators have reported enhanced growth in preterm infants fed nutrient-enriched formula after hospital discharge,27–29,39,40 or after a weight of 1850 g was achieved.42 Chan et al,41 however, reported no advantage to feeding nutrient-enriched formula to 8 weeks of age.
In agreement with other investigators, infants fed the less calorically dense TF had higher volumes of formula intake than those fed PDF.28,29 This led to consumption of similar energy intakes despite a 10% difference in caloric density. These data suggest some ability of the infants to compensate for differences in energy density. Serum levels of prealbumin, which are reported to correlate with weight and length in VLBW infants,43 were significantly higher among infants fed PDF at term and 9 months' CA. Serum urea nitrogen levels were higher among PDF-fed infants at each time point, but mean values were within reference ranges.
Poor somatic growth during the first year is associated with delays in neurodevelopment of preterm infants.17,44,45 Small head circumference measurements, in particular, may have long-term prognostic significance for later neurodevelopment in infants born prematurely. Hack et al15 reported that subnormal head size at 8 months of age predicted the 20 month Mental Development Index Score (Bayley's Scales of Infant Development) in VLBW infants, while Gross et al46 reported that head circumference at birth and postnatal head circumference, when taken together, were strong predictors of early developmental outcome in VLBW infants. Hack and colleagues44 have suggested that the first year of life provides an important opportunity for human somatic and brain growth to compensate for earlier deprivation. An effect of nutrition during this vulnerable period of brain development on long-term developmental outcomes is suggested by studies in which the feeding of nutrient-enriched formulas to preterm infants during hospitalization was associated with improved neurodevelopmental outcomes at 9 months,22 18 months,23 and 7.5 to 8 years old.19 In the present study, the most significant effect seen with feeding PDF versus TF was larger head circumference measurements among infants born with birth weights <1250 g.
Provision of additional nutrients after hospital discharge may be particularly beneficial for infants with conditions that contribute to a relative increase in morbidity, including the presence of chronic disease, in utero growth retardation, or extremely low birth weight. A particular growth benefit of nutrient-enriched formulas is indicated by the present study for infants with birth weights <1250 g. In addition, an effect of gender on postdischarge responsiveness to nutrient supplementation of preterm infants is suggested by our data and those of Cooke et al,29,41 in which the feeding of nutrient-enriched formulas after hospital discharge was of particular benefit for male infants.
It is increasingly recognized that formulas designed to meet the nutritional needs of term infants are unlikely to provide the most appropriate postdischarge nutrition for preterm infants. The results of the present study demonstrated that the feeding of a nutrient-enriched formula to preterm infants after hospital discharge resulted in improved growth and growth gains to 12 months' CA compared with feeding infants a formula designed for term infants. The beneficial effects were most evident for infants with birth weights <1250 g.
This study was supported in part by Ross Products Division, Abbott Laboratories, Columbus, Ohio, and by NIH General Clinical Research Center Grant RR00059.
The assistance of Robin Carroll, MS, RD, LD; Michelle Atwood, BSN; Mary Reilly, RN; and the General Clinical Research Center nurses, Karen J. Johnson and Gretchen A. Cress, is gratefully acknowledged.
- Received March 24, 2000.
- Accepted September 7, 2000.
Reprint requests to (J.C.) University of South Florida College of Medicine, Department of Pediatrics, Division of Neonatology, 17 Davis Blvd, Suite 200, Tampa, FL 33606. E-mail:
- American Academy of Pediatrics, Committee on Nutrition. Nutritional needs of preterm infants. In Kleinman RE, ed. Pediatric Nutrition Handbook. Elk Grove, IL: American Academy of Pediatrics; 1998:55–79
- Fenton TR,
- McMillan DD,
- Sauve RS
- Hack M,
- Horbar JD,
- Malloy MH,
- Tyson JE,
- Wright E
- Widdowson EM. Changes in body proportions and composition during growth. In Davis JA, Dobbing J, eds. Scientific Foundations of Paediatrics. Philadelphia, PA: WB Saunders Co; 1974:153–163
- Bishop NJ,
- King FJ,
- Lucas A
- Kelleher KJ,
- Casey PH,
- Bradley RH,
- et al.
- Lucas A,
- Morley R,
- Cole TJ
- Lucas A,
- Gor SM,
- Cole TJ,
- et al.
- Lucas A,
- Morley R,
- Cole TJ,
- et al.
- Kashyap S,
- Schulze KF,
- Forsyth M,
- Dell RB,
- Ramakrishnan R,
- Heird WC
- Lucas A,
- Bishop NJ,
- King FJ,
- Cole TJ
- Zhang S,
- Kuznetsova OM,
- Jacobs JR
- Zimmerman AW,
- Hambidge KM,
- Lepow ML,
- et al.
- Lucas A, Hay W. Posthospital nutrition in the preterm infant: general conclusions. In: Report of the 106th Ross Conference on Pediatric Research. Irving, TX: Ross Products Division, Abbott Laboratories; 1996:135–137
- Chan GM,
- Borschel MW,
- Jacobs JR
- Polberger SK,
- Fex GH,
- Axelsson IE,
- Raiha NC
- Gross SJ,
- Oehler JM,
- Eckerman CO
- Copyright © 2001 American Academy of Pediatrics