Background Data on enteral feeding management of premature infants are limited and often not the subject of randomized clinical trials. Several small studies suggest benefits from the early initiation of feeding, but do not assess the combined effects of time of initiation of feeding, tube-feeding method, and type of milk used. Either singly or in combination, these treatments may affect growth, bone mineralization, biochemical measures of nutritional status, and feeding tolerance, and, ultimately, the duration of hospitalization.
Methods. A total of 171 premature infants, stratified by gestational age (26 to 30 weeks) and diet (human milk or preterm formula) were assigned randomly among four treatment combinations in a balanced two-way design comparing the presence or absence of gastrointestinal (GI) priming for 10 days and continuous infusion versus intermittent bolus tube-feeding.
Results. The major outcome, time required for infants to attain full oral feeding, was similar among treatments. GI priming was not associated with any measured adverse effect and was associated with better calcium and phosphorus retention, higher serum calcium and alkaline phosphatase activity, and shorter intestinal transit times. The bolus tube-feeding method was associated with significantly less feeding intolerance and greater rate of weight gain than the continuous method. In addition, the greater the quantity of human milk fed, the lower the morbidity.
Conclusions. Early GI priming with human milk, using the bolus tube-feeding method, may provide the best advantage for the premature infant.
The provision of adequate enteral nutrition for premature infants is one of the major clinical challenges facing neonatologists throughout the world. A number of feeding strategies are practiced, although few have been subjected to experimental validation through randomized controlled trials.1–3 The wisdom of the practice of withholding enteral nutrition from the premature infant has been questioned.4 Improved tolerance of feedings, less need of parenteral nutrition, more mature intestinal motility patterns, and improved mineral homeostasis have been reported in small, randomized trials of the early initiation of low volumes of milk for several days (gastrointestinal [GI] priming) versus parenteral nutrition only.5–11 These beneficial effects, in turn, could be associated with a significant economic advantage if they reduce the duration of hospitalization.
Because premature infants are unable to coordinate sucking, swallowing, and breathing, orogastric tube-feeding is necessary. The most common methods used are continuous milk infusion and intermittent, or bolus, milk delivery (usually every 3 hours).12 Data from studies in older infants with intestinal disease suggest that the continuous tube-feeding method is associated with better nutrient absorption than the bolus method.13 However, recent studies have questioned these results for growing premature infants.14Moreover, bolus feeding is thought to promote more “normal” feed-fasting hormonal concentrations that potentially might benefit intestinal development and nutrient partitioning.15,,16
Interpretation of previous studies is limited in that they have not considered potential interactions with other treatments (eg, method of tube-feeding and/or type of milk used). For example, it is unclear whether early initiation of feeding is equally beneficial whether it involves continuous or bolus tube-feeding, or whether human milk or formula is used.
The objective of this study was to evaluate the effects of three treatments common to tube-feeding. GI priming (yes, no), method of tube-feeding (continuous, bolus), and type of milk (human milk, preterm formula). The specific aim was to determine if the combination of treatments to which an infant was assigned affected the time required for the infant to attain full oral feeding. We chose that endpoint because it generally represents the combined effects of improved rates of growth, bone mineralization, nutrient retention, biochemical measures of nutritional status, feeding tolerance, and morbidity, and is a key determinant in the duration of hospitalization. The secondary objective was to determine the effects of combinations of treatments on growth, bone mineralization, nutrient retention, biochemical measures of nutritional status, feeding tolerance, and morbidity.
Study infants were enrolled within 96 hours of birth, stratified by gestational age (26–27 vs 28–30 weeks) and by diet (human milk versus preterm formula), and assigned randomly among four treatment combinations in a balanced two-way design where the two treatments were the use of GI priming versus “non per os,” or no enteral intake (NPO) from day 4 through 14 after birth and the method of tube-feeding, continuous infusion versus bolus. The four treatment combinations were designated prime-continuous, prime-bolus, NPO-continuous, and NPO-bolus. The GI priming interval was selected because, before beginning the study, our nursery initiated any milk feedings in this population at approximately 17 days. To ensure the objective assessment of the major outcome variable, the time required by the infant to attain full oral feeding (8 breast and/or bottle feedings per day), oral-motor function was assessed serially, using a method designed specially for this study.17,,18
The sample size, 30 infants in each of four treatment combinations, was chosen to permit detection of a 2-week difference between treatment combinations in the time required to attain full oral feeding, with a type I error of 0.05 and a power of 0.90. Randomization was performed using sealed opaque envelopes that were grouped, in an uneven blocked design, by stratification variables (gestational age, intent to feed human milk). The study was approved by the Baylor College of Medicine Institutional Review Board for Human Subject Research. Informed written consent was obtained from parents before enrollment.
A total of 171 premature infants from the nurseries of Texas Children's Hospital were enrolled between March 1992 and April 1996 based on the following criteria: 26 to 30 weeks gestation (as determined by a combination of maternal dates and early antenatal ultrasound), gestational age agreement between the two methods ≤2 weeks, appropriate weight for gestational age, postnatal age ≤96 hours, absence of major congenital malformations, fraction of inspired oxygen <0.60 by 72 hours, and informed written consent obtained from parents. Infants were removed from the treatment protocol if they were unable to adhere to the feeding protocol for >1 week. Infants removed from the study were monitored for as many outcomes as their clinical status permitted and their data were included in the analysis.
A feeding schedule was maintained at each infant's bedside and monitored daily by a nutrition support team so that milk advancement and use of parenteral nutrition were consistent for all study infants. The nutrition protocol was standardized to provide similar intakes of fluid, energy, protein, calcium, and phosphorus among groups. Previously, we quantitated nutrient losses attributable to continuous tube-feeding of human milk (eg, fat adherent to feeding syringes and tubes) and developed protocols to ensure optimal nutrient delivery.19 Bolus feedings were given every 3 hours over 20 minutes. The study was implemented in three phases. The assigned orogastric/nasogastric tube-feeding method (continuous vs bolus) was maintained throughout the three phases.
In phase I, from day 4 through 14 after birth, the GI priming group (prime-continuous, prime-bolus) received 20 mL·kg−1·day−1 of milk and the remaining fluids as parenteral nutrition, whereas the NPO group (NPO-continuous, NPO-bolus) received only parenteral nutrition. Infants received either human milk (from their own mother) or ½-strength Enfamil Premature Formula (Mead Johnson Nutritional Division, Evansville, IN) during this phase. The composition of parenteral nutrition was standardized.20 In phase II, 15 to 22 days after birth, all infants received 20 mL·kg−1·day−1 and increased by that amount daily until complete tube-feeding was achieved (150 mL·kg−1·day−1). Enfamil Human Milk Fortifier (Mead Johnson Nutritional Division) was added to human milk when the intake reached 100 mL·kg−1·day−1. Formula-fed infants received the full-strength concentration during this phase. Phase III began the day complete tube-feeding was achieved, and the respective milk formulations were continued until full oral feeding was attained. Volumes were adjusted (to a maximum of 180 mL·kg−1·day−1) to enable a weight gain of ∼15 g·kg−1·day−1.
The rate of progression of the number of oral feedings was determined based on the following criteria: adequate body weight gain, maintenance of thermal stability, favorable oral motor assessment,17and absence of exacerbations of feeding intolerance and/or abnormal medical conditions (eg, increased apnea or oxygen need). The criteria for hospital discharge were uniform among attending physicians: satisfactory weight gain while receiving full oral feeding, maintenance of thermal stability, and resolution of acute medical conditions.
Body weight was measured at the same time each day using electronic scales.21 Fronto-occipital circumference, crown-heel length, and bilateral skinfold thicknesses at biceps, triceps, subscapular, suprailiac, and lateral thigh sites were measured every 2 weeks using methods published previously.21 Growth rate was computed by linear regression for each subject, using daily body weight from the time of minimum weight to hospital discharge.
Nutritional balance studies were conducted for 72 hours at 6 and 9 weeks postnatally to determine absolute and incremental changes in metabolizable energy intake and the absorption and retention of nitrogen, fat, calcium, phosphorus, magnesium, zinc, and copper. The methods for the continuous collections of urine and feces and the quantitative measurement of milk intake have been reported previously.21–23 Carmine red (100 mg·kg−1) was given with the initial and final milk feeding of the 72-hour study to mark the duration of the fecal collection. All stools passed between the appearance of the markers were collected directly into polyethylene bags attached to the buttocks with Compound Benzoin Tincture (Clinipad Corp, Guilford, CT) and Coloplast Skin Barrier (Coloplast, Inc, Tampa, FL). The methods for the determination of nitrogen, calcium, phosphorus, magnesium, zinc, and copper in milk, urine, and feces, and fat and energy in milk and feces have been reported previously.21–23 The balance (net retention) of each nutrient was calculated as the difference between intake and the sum of urine and fecal losses during the 72-hour interval. Net absorption was defined as the difference between intake and fecal losses.
Bone mineral content of the distal one-third of the radius of the left arm was measured at the beginning (end of the first week of the study) and the end of the study (near hospital discharge) using portable single photon absorptiometry (model SP-2 scanner, Lunar Radiation Corp, Madison, WI).
Serum indices of protein and mineral nutritional status were measured every 2 weeks. Serum calcium, phosphorus, magnesium, alkaline phosphatase activity, and albumin were analyzed by automated techniques (Cobas Fara, Roche Diagnostics Systems, Montclair, NJ). Serum 25-hydroxyvitamin D (Nichols Institute, San Juan Capistrano, CA) and osteocalcin (Incstar, Stillwater, MN) were measured by radioimmunoassay.
Feeding tolerance was assessed daily by the following characteristics: gastric residual volume (GRV, determined by aspiration of gastric contents from the indwelling orogastric tube every 3 hours in all infants), emesis, abdominal distention and/or tenderness, stool number, hematochezia, and the number of hours feedings were stopped. GRV was tabulated as =0, >0, ≥2 mL·kg−1, or >50% of a 3-hour feeding volume. The color of the GRV was assessed as clear (mucous), milk-colored, green (bilious), or bloodstained. An algorithm based on excess GRV was used to assist in the uniform assessment and management of feeding tolerance.20 Necrotizing enterocolitis (NEC) was defined as clinical signs with the presence of pneumatosis intestinalis on abdominal radiograph (Bell stage II).24
The primary data analysis was by intent to treat, ie, all subjects were included as randomized. Two-way analysis of variance and covariance was used to determine the effect of specific treatments (GI priming vs NPO, continuous vs bolus tube-feeding method, average human milk intake throughout hospitalization) and the effect of treatment combinations on the attainment of full oral feeding, growth, bone mineralization, nutrient retention, biochemical measures of nutritional status, feeding tolerance, and morbidity. Survival analysis techniques also were used to analyze time to full oral feeding. χ2and logistic regression analyses were used to assess treatments with respect to dichotomous outcomes. Statistical significance was set at 0.05. The data were expressed as mean ± SD.
There were no statistically significant differences in the clinical characteristics of the 171 infants among groups (Table 1). Because of the range of differences among groups for certain potentially relevant variables, Apgar score and antenatal steroid exposure were tested as covariates on major outcome variables. These variables did not impact the outcomes reported. The racial/ethnic distribution (47% white, 29% African-American, 23% Hispanic, 1% Asian) and maternal characteristics were similar among groups: number of days hospitalized before delivery (5 ± 8 days), outborn delivery (13%), vaginal delivery (36%), and substance (2%) or alcohol (2%) abuse. To ensure equivalent lactation support in all groups for mothers intending to breast feed, the initiation of lactation was similar: the first milk was available at 66 ± 29 hours (range 17–199 hours) and volume was 17 ± 17 mL (range 1–120 mL).
There were no differences among groups for the number of infants removed from the treatment protocol for death or medical conditions such as, NEC, gastrointestinal surgery, progressive hydrocephalus secondary to intraventricular hemorrhage, and septic arthritis of the hip. However, 11 infants were switched to the opposite tube-feeding method because their intolerance to feeding precluded adherence to the feeding protocol; 4 infants were in the prime-continuous and 6 infants were in the NPO-continuous groups (Table 1). Because of feeding intolerance, one infant in the bolus group was switched to continuous tube-feeding. Overall, protocol dropouts were significantly greater in continuous versus bolus groups (P = .029). No infants were removed because of priming.
Fluid, energy, and nutrient intakes were similar among groups throughout hospitalization. The average human milk intake throughout hospitalization, however, was significantly greater in the prime than the NPO groups, by ∼15 mL·kg−1·day−1(P = .02). Because of this difference, human milk intake was used as a covariate in the analyses.
The distribution of clinical morbidities, including patent ductus arteriosus, late-onset sepsis, intraventricular hemorrhage, mechanical ventilation, oxygen use, apnea and bradycardia, and medication use, was similar among groups. Although the incidence of NEC did not differ significantly among groups, it was inversely related to the quantity of human milk fed (P = .003).
There was no difference in the duration of umbilical artery or vein catheterization among groups. Enteral feeding occurred while umbilical artery catheters were in place for 1 to 4 days in approximately one third of the infants in the GI prime group. There were no associations between the use of umbilical artery catheters, concomitant feeding, and either the time to achievement of nutritional milestones, feeding tolerance, or NEC.
Nutritional milestones were tabulated among groups (Table 2). The continuous tube-feeding method delayed the time required to attain complete tube-feeding (150 mL·kg−1·day−1 by tube-feeding) in the most immature infants (gestational age 26–27 weeks). There were no differences among groups in the time required to attain full oral feeding (8 oral feedings per day), regardless of whether analysis of variance or survival analysis techniques were used. The first successful oral feeding and full oral feeding were achieved at 35 ± 2 weeks and 37 ± 2 weeks postmenstrual age, respectively. The duration of hospitalization was correlated positively with the age at attainment of complete tube-feeding (r = 0.41,P < .001) as well as full oral feeding (r = 0.92, P < .001) but was unrelated to study group designation.
Feeding intolerance, as assessed by a significantly greater number of GRVs, was observed in the continuous versus bolus groups throughout the hospitalization (Table 3), and these differences were noted as early as Phases I and II. There were no differences among groups, however, in the number of hours feedings were stopped, the incidence of abdominal distention, or in the number of abdominal radiographs. Differences in feeding tolerance among groups were unrelated to the type of milk fed.
Energy and nutrient intakes were similar among groups during the nutritional balance studies conducted at 6 weeks (43 ± 11 days) and 9 weeks (61 ± 14 days) postnatal age, 4 to 7 weeks beyond phase I. The carmine red intestinal transit time was significantly shorter in the GI prime versus NPO group (32 ± 11 vs 41 ± 17 hours, respectively, P = .005). Significant differences were observed in calcium, phosphorus, and copper retentions between the GI prime and NPO groups (Table 4). Significant increments from 6 to 9 weeks in percent absorption of energy, fat, calcium, and magnesium were observed, such that by 9 weeks there were no differences in nutrient retention between GI prime and NPO groups (data not shown). There were no differences between the continuous versus bolus groups in intestinal transit times, energy and nutrient intakes, or absorption and retention.
Biochemical indices of nutritional status, phosphorus, albumin, osteocalcin, and 25-hydroxyvitamin D did not differ among groups, but changed over time. Serum Ca and alkaline phosphatase activity significantly increased over time (P < .001), and both were greater in the prime versus NPO groups (P = .02). There were no differences in the number of infants with a conjugated bilirubin >2.0 mg/dL (>34 μmol/L) in the prime (15%) versus NPO (11%) group (P = .50).
The rate of weight gain was significantly slower in the continuous versus bolus tube-feeding group, 20 ± 4 vs 22 ± 5 g·kg−1·day−1 (P = .02). There were no differences among groups in other anthropometric indices or in measures of bone mineralization.
We report the largest, randomized, intent to treat, premature infant feeding study that also is the first study to evaluate simultaneously and prospectively the potential impact of the most commonly used feeding strategies: GI priming versus NPO, continuous versus bolus tube-feeding method, and human milk versus preterm formula. Although no treatment combination affected the time to attain full oral feeding, there were important differences in clinical responses to particular feeding strategies.
Importantly, we observed that GI priming was not associated with any measured adverse effect, eg, no increased deaths, prolonged hospitalization, or NEC in the GI prime group compared with the NPO group. The incidence of NEC was similar to that reported (∼12%) in other studies of GI priming.25 Our study differed from those reported previously in its larger size, differences in birth weight and gestational age, and time to attainment of certain nutritional milestones.5,,8,10,11
Contrary to a previous report on a small number of infants, we found that GI priming resulted in greater serum calcium and alkaline phosphatase activity, but did not affect the incidence of conjugated hyperbilirubinemia.5 We also observed that GI priming was associated with greater mineral absorption and retention, at least 4 weeks beyond phase I of the study. The net retentions in the GI prime group approximated the intrauterine accretion rate for calcium and phosphorus, suggesting that the absence of GI priming may have delayed the early programming necessary for mineral absorption, which was not achieved in the NPO group until 9 weeks.26 The differences in biochemical indices of nutritional status between studies may be explained by newer nutritional regimens, closer nutritional monitoring, and more rapid rates of growth during hospitalization than observed in earlier years. The elevated alkaline phosphatase activity, in conjunction with higher serum calcium and mineral retentions may be a function of rapid postnatal growth, commonly observed in premature infants.27
Differences in gross gastrointestinal function, at least 4 weeks beyond Phase I of the study, also were observed in the prime versus NPO groups. GI primed feeding was associated with shorter intestinal transit times, more frequent stool output, and fewer bilious gastric residuals. These data are consistent with those of Berseth8,,11 who demonstrated more mature gastroduodenal motor function in the group receiving GI priming versus NPO only.
We observed marked differences in feeding tolerance and growth between continuous versus bolus tube-feeding methods. Continuous feeding was associated with significantly more feeding intolerance, such that more protocol violations occurred that precluded the use of that tube-feeding method. Therefore, more infants were switched to bolus feeding. The greater number of GRVs in the continuous group was associated with an increase in the time to achieve complete tube-feeding. Importantly, throughout hospitalization the continuous feeding method was associated with slower growth compared with the bolus group. These data confirm and expand on recently reported observations in premature infants.14,,28Akintorin28 observed that feeding intolerance was more common when the continuous tube-feeding method was used but, distinct from our larger study, no overall effects on growth were reported. Furthermore, our data do not support earlier studies in larger infants with gastrointestinal disease that suggested better weight gain and nutrient absorption with the continuous feeding method.13Thus, for premature infants with relatively healthy gastrointestinal tracts, bolus feeding is more advantageous than continuous infusion.
In a previously reported small feeding study, bolus tube-feeding was associated with adverse effects on arterial oxygenation, chest wall compliance, and the incidence of apnea.29 However, we did not observe significant differences either in the duration of mechanical ventilation and oxygen therapy or the incidence of apnea and bradycardia episodes between the continuous and bolus group. Thus, in this study, feeding method did not affect respiratory status.
Our controlled study established projected intakes in a feeding protocol that was assessed daily by a neonatal nutrition team. Any deviations from the protocol were noted and adjustments made within study guidelines to achieve the projected daily weight gain of 15 g·kg−1·day−1. Contrary to other reports in which the GI prime group received greater energy and/or nutrient intakes, energy and nutrient intakes in this study did not differ significantly among groups throughout hospitalization.5,,6,9,10 The assurance of meeting projected fluid/milk intakes allowed us to make specific comparisons of GI priming, feeding method, and type of milk.
Although not differing among study groups, we did observe that full oral feeding, as well as complete tube-feeding, were important determinants of the duration of hospitalization. These relationships suggest that a method that hastens the attainment of complete tube-feeding and full oral feeding will decrease the duration of hospitalization. Once tube-feeding was initiated and tolerated, advancement in volume and transition to oral feeding appeared to follow in sequence. We believe that these relationships verify the use of our primary outcome variable.
There was a greater intake of human milk in the GI prime group. Although the initial lactation experience was similar among groups, we speculate that the early use of human milk in the GI prime group had subtle psychological effects that promoted maternal milk production. We used human milk intake as a covariate in the analyses. Independent of treatment group assignment, those infants receiving the most human milk had a significantly lower incidence of NEC. These results are in agreement with other reports indicating a protective effect of human milk against NEC.30
Our data support the practice of starting GI priming early, without adding to the complications of neonatal intensive care. Further studies are needed to determine if early feeding can be advanced in volume so that the use of parenteral nutrition can be reduced. Bolus feeding results in better feeding tolerance and growth than the continuous tube-feeding method and also obviates the need for costly infusion pumps and support care. The use of human milk, however, independent of study group assignment may have the most profound effects because of its association with a decrease in morbidity.
This study was supported by the National Institute of Child Health and Human Development, Grant No. RO-1-HD-28140 and the General Clinical Research Center, Baylor College of Medicine/Texas Children's Hospital Clinical Research Center, Grant No. MO-1-RR-00188, National Institutes of Health. Partial funding also has been provided from the USDA/ARS under Cooperative Agreement No. 58–6250-1–003.
We thank Pamela Burns, RN, Christina Valentine, RD, CNSD, Leanne Renfro, RN, and Ellen Newton-Lovato, RN; the nursery staff of the Neonatal General Clinical Research Center and neonatal nurseries at Texas Children's Hospital; and Nancy Hurst, RN, and the staff of the Lactation Program at Texas Children's Hospital for their expertise. We thank Charles Imo for laboratory assistance, J. Kennard Fraley for database management, Idelle Tapper for secretarial skills, and Leslie Loddeke for editorial advice.
- Received May 7, 1998.
- Accepted July 8, 1998.
Reprint requests to (R.J.S.) 1100 Bates St, Houston, TX 77030-2600.
This work is a publication of the USDA/ARS Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine and Texas Children's Hospital, Houston, TX. The contents of this publication do not necessarily reflect the views or policies of the USDA, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government.
- GI =
- gastrointestinal •
- NPO =
- “non per os,” no enteral intake •
- GRV =
- gastric residual volume •
- NEC =
- necrotizing enterocolitis
- Churella HR,
- Bachhuber WL,
- MacLean WC
- Ostertag SG,
- LaGamma EF,
- Reisen CE,
- Ferrentino FL
- Mashako LMN,
- Bernard C,
- Cezard JP,
- Chayvialle JA,
- Navarro J
- ↵Schanler RJ. Special methods in feeding the preterm infant. In: Tsang RC, Nichols BL, eds. Nutrition During Infancy. Philadelphia, PA: Hanley & Belfus; 1988:314–325
- ↵Schanler RJ. The low birth weight infant: perinatal nutrition. In: Walker WA, Watkins JB, eds. Nutrition in Pediatrics: Basic Science and Clinical Applications. Hamilton, Ontario, Canada: BC Decker Inc; 1996:387–407
- ↵Tyson JE, Kennedy KA. Minimal enteral nutrition in parenterally fed neonates. In: Sinclair JC, Bracken MB, Soll RF, Horbar JD, eds. Neonatal Module of The Cochrane Database of Systematic Reviews [updated 02 December 1997]. Available in The Cochrane Library [database on disk and CD-ROM]. The Cochrane Collaboration; Issue 1. Oxford: Update Software; 1998. Updated quarterly
- ↵Akintorin SM, Kamat M, Pildes RS, et al. A prospective randomized trial of feeding methods in very low birth weight infants. Pediatrics. 1997;100(4). URL: http://www.pediatrics.org/cgi/content/full/100/4/e4
- Copyright © 1999 American Academy of Pediatrics