No new eligible trials were identified
in searches updated in June 2004. One of the included trials (Caple 2004)
has been published in full since the previous review.
Rapid versus slow rate of advancement of feedings for promoting growth and preventing necrotizing enterocolitis in parenterally fed low birthweight babies
Very premature babies have poor coordination of sucking and swallowing so they must be fed through a tube. This means the babies cannot control the amount of milk they receive. The caregiver must decide on the amount. Slowly increasing the amount of feeding can mean the baby does not get enough nutrition. There is some concern that feeding too much too soon could cause necrotizing enterocolitis (NEC), a life-threatening disease affecting the bowel. The review of trials shows some evidence that there are advantages with more rapid increases in feeding amounts but more research is needed.
Very premature infants fed by gavage are unable to regulate their own enteral intake. Therefore the rate at which feedings are advanced must be determined by caregivers. While advancing feedings too rapidly may increase the risk of necrotizing enterocolitis, advancing feedings too slowly might result in undernutrition or prolonged hospital stay.
For low birth weight or premature infants receiving parenteral nutrition, to assess the effect of different rates of advancement of enteral feedings beginning at the same age on measures of feeding tolerance and neonatal outcome.
Searches were performed of the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 3, 2004), MEDLINE (1966-June 2004), CINAHL (1982-June 2004), abstracts and conference proceedings, references from relevant publications in the English language, and studies identified by personal communication.
Only randomized or quasi-randomized trials were considered. Trials were included if premature or low birth weight infants were studied and if the strategies being compared were different rates of advancement of feedings (accomplished by either differences in volume or concentration) with the onset of feedings at the same postnatal age in each group.
The two reviewers reached consensus for inclusion of trials. Data regarding clinical outcomes were extracted and evaluated by the two reviewers independently of each other. Authors were contacted as needed and feasible to clarify or provide missing data. The specific data that were needed were requested in writing.
Among infants randomized to more rapid rates of advancement of feedings, there was an overall reduction in days to full enteral feeding and days to regain birth weight. There was no significant effect on necrotizing enterocolitis (relative risk = 0.97, 95% confidence interval = 0.50, 1.87).
There are suggested advantages of more rapid rates of advancing feedings in premature low-birth-weight infants (shorter time to regain birth weight and shorter time to achieve full feedings). It is unclear whether this strategy should be adopted as routine practice because of limited information regarding safety (broad confidence intervals for the incidence of necrotizing enterocolitis) and the effect on length of hospital stay (broad confidence intervals). Because different birth weight ranges and different rates of advancement were used in each of these studies, the ideal rate of advancement remains unclear, particularly for extremely low birth weight (< 1000 g) infants.
Because of poor coordination of sucking and swallowing, very premature infants are fed by gavage and are thus unable to regulate their own enteral intake. Therefore the rate at which feedings are advanced must be determined by caregivers. Feeding intolerance manifested by gastric residuals and abdominal distention is a common problem associated with advancing enteral feedings in very premature infants. Based largely on retrospective studies, there is concern that advancing feedings too rapidly may increase the risk of necrotizing enterocolitis (Uauy 1991; Brown 1978). On the other hand, advancing feedings too slowly might result in undernutrition, prolonged exposure to parenteral nutrition, and/or prolonged hospital stay.
This review updates the existing review of 'Rapid versus slow rate of advancement of feedings for promoting growth and preventing necrotizing enterocolitis in parenterally fed low birthweight infants' which was published in The Cochrane Library, Issue 4, 1998 and updated in Issue 1, 2000. No new eligible trials have been identified.
For low birth weight or premature infants receiving parenteral nutrition, to assess the effect of different rates of advancement of enteral feedings beginning at the same postnatal age on measures of feeding tolerance and neonatal outcome.
Randomized or quasi-randomized trials when the unit of randomization was the patient. Randomized cross-over trials were not included.
Infants at high risk of necrotizing enterocolitis and feeding problems on the basis of prematurity or low birth weight. Studies were included only if the participants were receiving supplemental parenteral nutrition (fluid and glucose with or without amino acids and lipid) while enteral feedings were being advanced.
Different rates of progression in the advancement of feedings (differences in enteral caloric intake accomplished by either differences in volume or concentration) with the onset of feedings at the same postnatal age in each group.
The following outcomes measures (listed in decreasing order of clinical significance) were selected:
Necrotizing enterocolitis > Bell's stage 1
Necrotizing enterocolitis with perforation
Length of hospital stay
Days to reach full feeding
Days feedings held
Days to regain birth weight
Electronic searches of the MEDLINE (1966 - June 2004) and CINAHL (1982
- June 2004) databases were performed using the following strategy:
1) enteral nutrition OR enteral feeding (as MeSH terms or text word)
2) AND (infant, premature OR infant, low birth weight) (as MeSH terms or text word)
3) limited to humans
Additional searches were performed of the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 3, 2004), references in review articles and relevant chapters of textbooks on neonatal nutrition, trials identified in a previous systematic review (Steer PA, Lucas A, Sinclair JC. Feeding the low birthweight infant. In: Sinclair JC, Bracken MB, eds, Effective Care of the Newborn Infant. New York: Oxford University Press, 1992: 130-132), review of recent American Pediatric Society - Society for Pediatric Research abstracts (1990-2004), personal files, and personal communication.
Both authors participated in identifying trials for consideration and in describing the main results and conclusions. Criteria for inclusion of studies were discussed prior to searching. Each author evaluated whether the trial met eligibility criteria for inclusion in the meta-analysis. Each author independently extracted data from the trials and evaluated their methodologic quality.
Details of the studies are included in the table, Characteristics of Included Studies. Entry criteria and rates of advancement were somewhat different among the three included studies. In Book 1976, infants < 1200 g were included (mean birth weight of enrolled patients was 1060-1065 g) and feedings were advanced at 10 cc/kg/day and 20 cc/kg/day in the slow and rapid groups respectively. Reported outcomes included bloody stools, abdominal distention, diarrhea, frequent emesis, and necrotizing enterocolitis. In Rayyis 1999, infants 501-1500 g were included (mean birth weight of enrolled patients was 1080-1116 g) and feedings were advanced at 15 cc/kg/day and 35 cc/kg/day in the slow and rapid groups respectively. Reported outcomes included the age at which full feedings (160 ml/kg/day) were achieved, the age at which birth weight was regained, the age at discharge, and necrotizing enterocolitis. In Caple 2004, infants 1000-2000 g were included (mean birth weight of enrolled patients was 1446-1495 g) and feedings were advanced at 20 cc/kg/day and 30 cc/kg/day in the slow and rapid groups respectively. Reported outcomes included age at which full feedings (150 ml/kg/day) were achieved, age at which birth weight was regained, length of stay, and necrotizing enterocolitis. There were relatively few patients < 1000 g (33 patients randomized to the rapid feeding regimen in Rayyis 1999) in the latter two studies in which rapid rate of advancement was defined as 30-35 cc/kg/day.
Quality assessments are included in the table, Characteristics of Included Studies. Overall, the trials were of good methodologic quality. Group assignment at randomization was concealed for at least two of the studies. Caregivers were not blinded to treatment group in any trial -- a problem that probably cannot be avoided. In two of the three studies, the radiologists who confirmed the diagnosis of necrotizing enterocolitis were blinded. An important strength of these studies is that almost all randomized infants were included in the analyses for all of the studies.
There was an overall reduction in mean days to full enteral feeding (weighted mean difference [WMD] = -3.2 days) and days to regain birth weight (WMD = -2.1 days) among infants in the rapid rate of feeding advancement group. There was no significant effect on necrotizing enterocolitis (relative risk [RR] = 0.97; 95% confidence interval [95% CI] = 0.50, 1.87) or necrotizing enterocolitis with perforation (RR = 0.88; 95% CI = 0.22, 3.49). (A relative risk <1.00 favors the rapid feeding group.) There was also no significant effect on length of hospital stay (WMD = -3.8 days; 95% CI = -9.1, 1.5). There were no data available for the number of days feedings were held or for the proportion of infants who were able to achieve the desired intakes prescribed in the study regimens. Results were consistent among the three studies, given the wide confidence intervals observed for length of hospital stay and for necrotizing enterocolitis. See Discussion.
In current practice, premature low birth weight infants are routinely supplemented with intravenous glucose with or without amino acids and lipids while enteral feedings are being established. There have been two randomized trials of slow vs. rapid advancement of feeding without supplemental intravenous fluid or glucose published in 1966 (Russell 1966) and 1976 (Weber 1976). These trials have been previously reviewed (Steer 1992) and are not included in this review because the risk/benefit ratio of different enteral feeding schedules is likely to be different if parenteral fluids and nutrients are provided as in current practice than if they are not provided.
There have been a number of clinical trials comparing a variety of feeding regimens in parenterally fed premature infants. These have included a) variations in the age of onset of feeding with either a steady progression in each group or constant low volume enteral intake in the early group (trophic feedings), b) variations in the composition of the feedings, and c) variations in the rate of progression of enteral caloric intake with either different but steady rates of progression in each group or with one group held at a fixed intake for at least several days (trophic feedings). Some of these trials have also dictated differences in parenteral caloric intake as part of the study regimen. For this review, we chose to include only studies in which the age of onset of feeding was the same in each group and different steady rates of progression of enteral caloric intake were used in each of the study groups. Trophic feedings (Tyson 1997) and early vs. late onset of feeding (Kennedy 2005) are the subjects of separate reviews.
The three studies which met our criteria for review used different birth weight ranges for entry criteria and different definitions for slow and rapid rates of feeding advancement. There are no other data on which to base an a priori decision regarding the definition of slow and rapid feeding and the same rate of feeding advancement might be considered rapid for a 600 g infant and slow for an 1800 g infant. The slow and rapid regimens used are all within the scope of current clinical practice. Therefore we decided to include them a single review with the recognition that an ideal feeding regimen could not be defined using the existing data.
The evidence of the advantages and safety of faster rates of advancement of feeding should be interpreted cautiously for several reasons: a) There remains uncertainty about the effect of different rates of feeding advancement on necrotizing enterocolitis. Although there were a total of 371 infants evaluated for this outcome in these studies, the confidence interval for this outcome is broad because of the low rate of occurrence. b) It would be quite difficult, if not impossible, to blind the caregivers to infant feeding. Therefore the diagnosis of necrotizing enterocolitis is subject to diagnostic surveillance bias because caregivers might order more x-rays in one feeding group than in the other and their interpretation of radiologic and clinical findings might be influenced by knowledge of feeding group assignment. c) These data might not be generalizable to extremely-low-birth-weight infants (< 1000 g). Only one of the two large trials (Rayyis 1999) included these infants so there were relatively few of these infants contributing to the data in this review.
Despite the suggested advantages of more rapid rates of advancing feedings, it is unclear whether this strategy should be adopted as routine practice because of limited information regarding both safety and effectiveness, particularly for extremely low birth weight infants.
A large multi-center trial comparing different rates of advancement of feedings among infants less than 1000g birth weight is needed. A convincing comparison of feeding strategies is likely to require a trial sufficiently large to identify an important effect on one or more of the following variables: 1) death before discharge home 2) death or necrotizing enterocolitis requiring surgical resection before discharge home 3) death or very prolonged hospital stay (an outcome that is associated with major long-term morbidity and that would address not only necrotizing enterocolitis but sepsis or other major complications of prolonged parenteral nutrition) or 4) death or major morbidity (including severe developmental delay and short bowel syndrome) at follow-up. To be feasible, such a large trial would require a simple protocol and a well organized group of participating centers.
None
| Study | Methods | Participants | Interventions | Outcomes | Notes | Allocation concealment |
| Book 1976 | Method of randomization - not stated Blinding of caregivers - no Blinding of assessors - no Almost or all subjects analyzed - can't tell | Premature infants with birth weights <1200g who survived at least one week. 29 infants were enrolled. Feedings were initiated when the respiratory status was stable, at a mean postnatal age of 2-3 days. | 20-24 kcal/oz cow milk formula was used. Slow feeding: increased at rate of 10ml/kg/day to 140ml/kg/day Rapid feeding: increased at rate of 20 ml/kg/day to 140ml/kg/day | Reported outcomes included bloody stools, abdominal distention, diarrhea, frequent emesis, and necrotizing enterocolitis. | There were no significant differences between the groups in the proportion of infants who developed bloody stools, abdominal distention, diarrhea, or frequent emesis. | B |
| Caple 2004 | Method of randomization - "blinded, random draw, numerical assignment" (numbers drawn from an envelope) Blinding of caregivers - no Blinding of assessors - yes for radiologists, no for clinicians Almost or all subjects analyzed - yes (5 of 160 enrolled infants were excluded after enrollment because of ineligibility or protocol violations. See Notes.) | Inborn AGA infants with birth weight 1000-2000g and less than 36 weeks gestation. Infants with congenital anomalies or exposure to illegal drugs were excluded. 160 infants were enrolled. 184 infants who met the birth weight criteria were not enrolled because of failure to meet the other criteria or inability to obtain parental consent. Feedings were initiated at a mean postnatal age of 2.7 days (when the clinician deemed the infant to be ready and 24 hours after umbilical catheters had been removed). | Full strength
commercial formula or human milk was used initially; 24 kcal/oz premature
formula or fortified human milk was used when the volume was advanced to
150cc/kg/day. Slow feeding: feedings were begun at 20cc/kg/day and advanced at a rate of 20cc/kg/day Rapid feeding: feedings were begun at 30cc/kg/day and advanced at a rate of 30cc/kg/day | Reported outcomes included age at which full feedings were achieved, age at which birth weight was regained, length of stay, and necrotizing enterocolitis. Full feeding volume was defined as 150cc/kg/day. | In the published report, 5 infants were excluded after enrollment (3 in the rapid group, one of whom developed NEC, and 2 in the slow group). Three of these infants (including the 1 with NEC) were excluded because of protocol violations and have been included in the meta-analysis of the NEC outcomes. Two of these infants (1 in each group) were excluded because they were determined not to be eligible for enrollment (because of an in utero gastrointestinal perforation and fetal alcohol syndrome); these infants were not included in the meta-analysis. | A |
| Rayyis 1999 | Method of randomization - sealed envelopes in variable block sizes Blinding of caregivers - no Blinding of assessors - yes for radiologists, no for clinicians Almost or all subjects analyzed - yes | Infants with birth weight 501-1500g and gestational age less than or equal to 34 weeks and surviving until initiation of feedings. 241 infants were excluded because of one of the following exclusion criteria: human milk feedings (153), major congenital anomalies (93), low Apgar scores (26), or inability to obtain consent (36). Infants were stratified into four birth weight groups: 501-750g (n=22), 751-1000g (n=54), 1001-1250g (n=47), 1251-1500g (n=62). Feedings were initiated at a mean of 4-5 days postnatal age, according to the discretion of the clinicians. | Similac Special Care 20 was used. Slow feeding: feedings begun at 20cc/kg/day and advanced by 15cc/kg/day. Rapid feeding: feedings begun at 35cc/kg/day and advanced by 35cc/kg/day. Feedings were not advanced if gastric residuals exceeded 20% of the feeding. | Reported outcomes included the age at which full feedings were achieved, the age at which birth weight was regained, the age at discharge, and necrotizing enterocolitis. Full feedings were defined as 160cc/kg/day. | A |
| Study | Reason for exclusion |
| Berseth 2003 | Infants in the slow "trophic" feeding group had a stable (not progressively increased) feeding volume over the first 10 study days. (This trial was included in the Cochrane review of trophic feedings for promoting tolerance and preventing morbidity in parenterally fed infants.) |
| Currao 1988 | Infants were randomized to receive feedings of different concentrations, but the enteral caloric intake was the same in both groups. |
| Drew 1974 | It is not clear whether group allocation was accomplished in a randomized or quasi-randomized fashion. Term and preterm infants were studied. One group received full-strength feedings initially; the other group had the caloric density of the feedings gradually advanced. Only laboratory findings and vomiting, diarrhea, and constipation were reported as outcomes. |
| Glass 1984 | Infants were quasi-randomized (by alternating assignment) into a transpyloric feeding group and a parenteral nutrition group. The feeding group had feedings initiated at an average age of 9.9 hrs and advanced to full feedings by an average age of 4.9 days (over 4 days). The parenteral nutrition group had feedings initiated at an average age of 6.7 days and advanced to full feedings by an average age of 20.7 days (over 14 days). |
| Jadcherla 1995 | Infants were randomized to begin feedings with either 20kcal/oz term formula or 24 kcal/oz preterm formula. The feedings were not advanced for the next 10 days in either group. |
| Kuschel 2000 | Feedings were advanced at the same rate in each group with different final target feeding volumes. |
| Melnick 1988 | Low-birth-weight infants were randomized to have the caloric density of the feedings increased to 24 kcal/oz either 2 days or 9 days after the initiation of feedings. Within each group, the composition (two types) of the 24 kcal/oz formula was also randomly assigned. The only outcome reported was intolerance which included >2cc gastric residual, abdominal distention, diarrhea, vomiting, or oral intake <110kcal/kg/day 10 days after initiation of feedings. |
| Mileur 1991 | This study was a three-phase randomized trial. Infants in each of the first two phases were randomized to 2 of 3 different rates of advancement of feedings. Infants in phase 3 were randomized to different ages for the onset of feedings as well as different rates of advancement. Data from either of the first two phases might have been appropriate for inclusion, but we were unable to obtain complete information for each phase of this study. |
| Siegel 1984 | Each individual infant received feedings of different caloric density. The order of the feedings was randomly assigned. Gastric emptying was measured after each feeding. No other outcomes were reported. |
| Wilson 1997 | Feedings were begun at different times. The study was designed so that full feedings would be reached at the same age in each group. |
Book LS, Herbst JJ, Jung AL. Comparison of fast- and slow-feeding rate schedules to the development of necrotizing enterocolitis. Journal of Pediatrics 1976;89:463-6.
Caple 2004 {published and unpublished data}
Caple J, Armentrout D, Huseby V, Halbardier B, Garcia J, Sparks JW et al. Randomized controlled trial of slow versus rapid feeding volume advancement in preterm infants. Pediatrics 2004;114:1597-1600.
Rayyis 1999 {published and unpublished data}
Rayyis SF, Ambalavanan N, Wright L, Carlo WA. Randomized trial of "slow" versus "fast" feed advancements on the incidence of necrotizing enterocolitis in very low birth weight infants. Journal of Pediatrics 1999;134:293-7.
Berseth CL, Bisquera JA, Paje VU. Prolonging small feeding volumes early in life decreases the incidence of necrotizing enterocolitis in very low birth weight infants. Pediatrics 2003;111:529-34.
Currao 1988 {published data only}
Currao WJ, Cox C, Shapiro DL. Diluted formula for beginning the feeding of premature infants. American Journal of Diseases of Children 1988;142:730-1.
Drew 1974 {published data only}
Drew JH, Breheny JE, Gleeson M. Evaluation of non-graded 20 kilocalorie per 30 millilitre feedings to newborn infants. Medical Journal of Australia 1974;1:879-81.
Glass 1984 {published data only}
Glass EJ, Hume R, Lang MA, Forfar JO. Parenteral nutrition compared with transpyloric feeding. Archives of Disease in Childhood 1984;59:131-5.
Jadcherla 1995 {published data only}
Jadcherla SR, Berseth CL. Acute and chronic intestinal motor activity responses to two infant formulas. Pediatrics 1995;96:331-5.
Kuschel 2000 {published data only}
Kuschel CA, Evans N, Askie L, Bredemeyer S, Nash J, Polverino J. A randomized trial of enteral feeding volumes in infants born before 30 weeks' gestation. Journal of Paediatrics and Child Health 2000;36:581-6.
Melnick 1988 {published data only}
Melnick G, Crouch JB, Cakackkas et al. Tolerance of LBW infants to early and late introduction of two 24 kcal formulas. Pediatric Research 1988;23:488A.
Mileur 1991 {published and unpublished data}
Mileur LM, Chan GM, Kimura RE. Effect of early, low volume feedings on very low birthweight (VLBW) infants. Pediatric Research 1991;29:300A.
Siegel 1984 {published data only}
Siegel M, Lebenthal E, Krantz B. Effect of caloric density on gastric emptying in premature infants. Journal of Pediatrics 1984;104:118-22.
Wilson 1997 {published data only}
Wilson DC, Cairns P, Halliday HL, Reid M, McClure G, Dodge JA. Randomised controlled trial of an aggressive nitritional regimen in sick very low birthweight infants. Archives of Disease in Childhood 1997;77:F4-11.
* indicates the primary reference for the study
Brown EG, Sweet AY. Preventing necrotizing enterocolitis in neonates. JAMA 1978;240:2452-4.
Kennedy KA, Tyson JE. Early vs delayed initiation of progressive enteral feedings for parenterally fed low birth weight or preterm infants. In: The Cochrane Database of Systematic Reviews, Issue 1, 2005.
Russell G, McKay E. Blood glucose concentration in the perinatal period. Archives of Disease in Childhood 1966;41:496-502.
Steer PA, Lucas A, Sinclair JC. Feeding the low birthweight infant. In: Sinclair JC, Bracken MB, editor(s). Effective Care of the Newborn Infant. New York: Oxford University Press, 1992:130-132.
Tyson JE, Kennedy KA. Minimal enteral feeding in parenterally fed neonates. In: The Cochrane Database of Systematic Reviews, Issue 4, 1997.
Uauy RD, Fanaroff AA, Korones SB, Phillips EA, Philips JB, Wright LL, for members of the National Institute of Child Health and Human Development Neonatal Research Network. Necrotizing enterocolitis in very low birth weight infants: biodemographic and clinical correlates. Journal of Pediatrics 1991;119:630-8.
Weber HP, Kowalewski S, Gilje A, Mollering M, Schnaufer I. Unterschiedliche calorienzufuhr bei 75 'low birth weights': Einfluss auf gewichtszunahme, serumeiweiss, blutzucker und serumbilirubin. European Journal of Pediatrics 1976;122:207-16.
Kennedy KA, Tyson JE. Rapid versus slow rate of advancement of feedings in parenterally fed low-birth-weight infants. In: The Cochrane Database of Systematic Reviews, Issue 4, 1998.
Kennedy KA, Tyson JE, Chamnanvanakij S. Rapid versus slow rate of advancement of feedings for promoting growth and preventing necrotizing enterocolitis in parenterally fed low-birth-weight infants. In: The Cochrane Database of Systematic Reviews, Issue 1, 2000.
| Comparison or outcome | Studies | Participants | Statistical method | Effect size |
|---|---|---|---|---|
| 01 Rate of advancement of feedings | ||||
| 01 Necrotizing enterocolitis (> Bell stage 1) | 3 | 372 | RR (fixed), 95% CI | 0.97 [0.50, 1.87] |
| 02 Necrotizing enterocolitis with perforation | 2 | 343 | RR (fixed), 95% CI | 0.88 [0.22, 3.49] |
| 03 Length of hospital stay | 2 | 340 | WMD (fixed), 95% CI | -3.81 [-9.13, 1.51] |
| 04 Age at full feeding | 2 | 340 | WMD (fixed), 95% CI | -3.25 [-4.87, -1.62] |
| 05 Age to regain birth weight | 2 | 340 | WMD (fixed), 95% CI | -2.11 [-3.26, -0.96] |
Old title: Rapid versus slow rate of advancement of feedings for parenterally fed low birth weight or preterm infants
| The review is published as a Cochrane review in The
Cochrane Library, Issue 2, 2005 (see http://www.thecochranelibrary.com for
information). Cochrane reviews are regularly updated as new evidence emerges
and in response to comments and criticisms, and The Cochrane Library should
be consulted for the most recent version of the Review. |