Fast feed advancement may reduce hospital stay and infection but may increase adverse outcomes in preterm and low birth weight infants. The objective of this study was to assess effects of fast feed advancement (≥30 ml/kg per day) compared with slow feed advancement (<30 ml/kg per day) in preterm and low birth weight infants.
Data sources include Medline, Scopus, Web of Science, CINAHL, and Index Medicus through June 30, 2021. Randomized trials were selected. Primary outcomes were mortality, morbidity, growth, and neurodevelopment. Data were extracted and pooled using random-effects models. The Cochrane Risk of Bias 2 tool was used.
A total of 12 RCTs with 4291 participants were included. At discharge, there was moderate certainty evidence that fast advancement likely slightly reduces the risk of: mortality (relative risk [RR] 0.93, 95% confidence interval [95% CI] 0.73 to 1.18, I2 = 18%, 11 trials, 4132 participants); necrotizing enterocolitis (RR 0.89, 95% CI 0.68 to 1.15, I2 = 0%, 12 trials, 4291 participants); sepsis (RR 0.92, 95% CI 0.83 to 1.03, I2 = 0%, 9 trials, 3648 participants); and feed intolerance (RR 0.92, 95% CI 0.77 to 1.10, I2 = 0%, 8 trials, 1114 participants). Fast feed advancement may also reduce the risk of apnea (RR 0.72, 95% CI 0.47 to 1.12, I2 = 0%, low certainty, 2 trials, 153 participants). Fast feed advancement decreases time to regain birth weight (mean difference [MD] −3.69 days, 95% CI −4.44 to −2.95, I2 = 70%, high certainty, 6 trials, 993 participants,) and likely reduces the duration of hospitalization (MD −3.08 days, 95% CI −4.34 to −1.81, I2 = 77%, moderate certainty, 7 trials, 3864 participants). Limitations include heterogeneity between studies and small sample sizes.
Fast feed advancement reduces time to regain birth weight and likely reduces the length of hospital stay; it also likely reduces the risk of neonatal morbidity and mortality slightly. However, it may increase the risk of neurodevelopmental disability slightly. More studies are needed to understand the long-term effects of fast feed advancement.
Feed advancement is a critical and controversial issue in the clinical care of hospitalized preterm and low birth weight (LBW) infants. Slow advancement of enteral milk volumes has traditionally been used because of concerns from clinicians about feed intolerance and necrotizing enterocolitis.1 Early observational studies indicated associations between slow feed advancement rates and lower risks of necrotizing enterocolitis.2–5 However, slow feed advancement has potential disadvantages. It prolongs the usage of parenteral nutrition and thus increases the risks of parenteral nutrition-associated liver disease and central line-associated bloodstream infections, both of which are associated with severe adverse consequences and high mortality.6
Current World Health Organization (WHO) recommendations are that feed volumes can be increased by up to 30 ml/kg per day under careful monitoring for feed intolerance in very low birth weight infants (VLBW) (<1.5 kg) who need to be fed by an alternative oral feeding method or nasogastric or orogastric tubes.7 To determine an optimal but also safe feed advancement regimen, a number of randomized controlled trials (RCTs) have compared the effects of different speeds of increasing milk feed volumes.8–22 A 2021 Cochrane systematic review compared slow feed increment (≤24 ml/kg per day) with fast increment (>24 ml/kg per day) in VLBW infants and included 13 trials. The results demonstrated no difference between slower versus faster feed advancement on the risks of mortality and necrotizing enterocolitis. However, the review did report that slower feed advancement may slightly increase risks of feeding intolerance and invasive infection.23 There have been no systemic reviews of feed advancement rates for “larger” LBW infants of 1.5 to 2.4 kg, and no systematic reviews have examined the 30 ml/kg per day increments included in the current WHO guidelines.7
Thus, our primary objective was to evaluate the effects of fast feed advancement (≥30 ml/kg per day) compared with slow feed advancement (<30 ml/kg per day) on mortality, morbidity, growth, and neurodevelopmental outcomes in preterm (<37 weeks) and LBW infants (<2.5 kg) infants.
Methods
Inclusion Criteria
We included RCTs comparing the effects of fast (≥30 ml/kg per day) versus slow (<30 ml/kg per day) feed advancement rates in preterm (<37 weeks) and LBW (<2.5 kg) infants. Our primary outcomes were mortality, morbidity, growth, and neurodevelopment at the latest follow-up (Appendix 1). Additional outcomes included feed intolerance and duration of hospitalization. We only included studies with available full text. Studies in any language were eligible.
Search Strategy
The full search strategy is in Appendix 2. We searched the following databases: Medline, Scopus, Web of Science, CINAHL, and Index Medicus through June 30, 2021.
Study Selection and Data Extraction
Three reviewers independently screened titles and abstracts and extracted data for eligible studies using a checklist based on inclusion and exclusion criteria. Both title and abstract screening and full-text screening were managed on the web-based software, Covidence.26 Detailed study information including authors, publication year, location, number of participants, gestational age, and birth body weight of participants, inclusion and exclusion criteria, and feed advancement rates in intervention and comparison groups were collected.
Risk of Bias
Two reviewers independently assessed the risk of bias of individual studies using the Cochrane Risk of Bias 2 tool.27 Disagreements were assessed by a third reviewer. The Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach was used to evaluate the overall quality of the evidence.28 GRADEpro GDT software was used.29
Statistical Analysis
Data were pooled using DerSimonian and Laird random-effects models.30 For binary outcomes, risk ratios (RR) with 95% confidence intervals (CI) were used to summarize results. To include studies with 0 events in either group, we applied a continuity correction in which 0.5 was added to each cell of the corresponding 2 × 2 table.31 For continuous outcomes, we pooled the mean difference estimates between groups. For studies that reported the median and interquartile range or range rather than mean and standard deviation (SD), we used the formulas developed by Luo et al and Wan et al to convert the data to estimated mean and SD.32–34
We prespecified 2 subgroup analyses to explore potential sources of heterogeneity: (1) gestational age and birth weight and (2) milk type (human milk or preterm formula). We performed an analysis of trials with any LBW (<2.5 kg) or preterm (<37 weeks) infants compared with trials with only VLBW (<1.5 kg) or very preterm (<32 weeks) infants as there were insufficient data to calculate the mean (SD) or median (interquartile range) of the gestational age or birth weight of the infants. Also, we were unable to perform subgroup analysis by milk type as there was only 1 trial that used infant formula; all other studies used human milk. All analyses were performed using Stata/IC 16.1 and RevMan software.
Results
Included Studies and Characteristics
We included a total of 12 RCTs with 4291 participants (see PRISMA flowchart, Appendix 3). Characteristics of included studies are presented in Appendix 4. The trials were conducted in the United States (2), Ireland (1), Iran (1), Turkey (1), India (4), Columbia (1), Bangladesh (1), and South Africa (1). The Ireland Speed of Increasing Milk Feeds trial (SIFT) was the largest trial (n = 2973).19,35 Most studies included clinically stable infants and excluded those with perinatal asphyxia or hemodynamic instability. The infants were typically randomized on day 1 to 3 of life. Intervention (fast advancement) feed increments ranged from 30 to 40 ml/kg per day; comparison (slow advancement) feed increments ranged from 15 to 30 ml/kg per day. Three studies with both intervention and comparison increments ≤30 ml/kg per day were excluded.8,16,22 Target full volumes ranged from 120 to 200 ml/kg per day from the included trials.
All studies included infants <1.5 kg or <32 weeks gestation. In addition, 5 studies also included larger infants who weighed 1.5 to 2.0 kg at birth (Caple 2004, Jain 2016, Hasshemi 2018, Saha 2019 Montealegre-Pomar 2020),10,14,17,18,21 and 7 studies enrolled infants with gestational age 32 to 36 weeks (Rayyis 1999, Caple 2004, Krishnamurthy 2010, Jain 2016, Hasshemi 2018, Saha 2019, Montealegre-Pomar 2020).9,10,12,14,17,18,21 Raban 2016 had a factorial design, so we presented data as Raban 2016a (initial volume of 4 ml/kg per day) and Raban 2016b (initial vol 24 ml/kg per day) but we considered it as 1 trial.15 Saha 2019 presented data separately for infants 1.2 to <1.5 kg (Saha 2019a) and infants 1.5 to <2.5 kg (Saha 2019b). We presented these data separately for continuous outcomes, but the data were pooled for dichotomous outcomes.18
Risk of Bias
No studies were classified as at “high risk.” However, the overall judgment was some concerns in the risk of bias assessment for all trials (Appendix 5) as none of the trials masked parents, caregivers, or investigators.
Primary Outcomes
Data are summarized in Table 1 and Appendix 6. At discharge, when comparing the intervention (fast advancement) to the control (slow advancement), the relative risk for all-cause mortality was 0.93 (95% confidence interval [95% CI] 0.73 to 1.18, I2 = 18%, moderate certainty evidence, 11 trials, 4132 participants). The relative risk for necrotizing enterocolitis was 0.89 (95% CI 0.68 to 1.15, I2 = 0%, moderate certainty evidence, 12 trials, 4291 participants) and for sepsis was 0.92 (95% CI 0.83 to 1.03, I2 = 0%, moderate certainty evidence, 9 trials, 3648 participants). The relative risk for feed intolerance was 0.92 (95% CI 0.77 to 1.10, I2 = 0%, moderate certainty evidence, 8 trials, 1114 participants) and for apnea was 0.72 (95% CI 0.47 to 1.12, I2 = 0%, low certainty evidence, 2 trials, 153 participants). The relative risk for neurodevelopmental disability at 24 months in the only 1 trial that reported this outcome was 1.12 (95% CI 0.98 to 1.27, I2 = NA, low certainty evidence, 1 trial, 2325 participants).19,34
Summary of Findings: Fast Enteral Feeding Advancement Compared With Slow Enteral Feeding Advancement for Preterm and Low Birth Weight Infants
Outcomes . | No. of Participants (Studies) . | Certainty of the Evidence (GRADE) . | Relative Effect (95% CI) . | Anticipated Absolute Effects . | |
---|---|---|---|---|---|
Risk With Slow Advancement . | Risk Difference With Fast Advancement . | ||||
Mortality (cases) at discharge | 4132 (11 RCTs) | ⨁⨁⨁◯ Moderatea | RR 0.93 (0.73 to 1.18) | 78 per 1000 | 5 fewer per 1000 (21 fewer to 14 more) |
Necrotizing enterocolitis (cases) during admission | 4291 (12 RCTs) | ⨁⨁⨁◯ Moderatea | RR 0.89 (0.68 to 1.15) | 55 per 1000 | 6 fewer per 1000 (18 fewer to 8 more) |
Sepsis (cases) during admission | 3648 (9 RCTs) | ⨁⨁⨁◯ Moderatea | RR 0.92 (0.83 to 1.03) | 275 per 1000 | 22 fewer per 1000 (47 fewer to 8 more) |
Apnea (cases) during admission | 153 (2 RCTs) | ⨁⨁◯◯ Lowb,c | RR 0.72 (0.47 to 1.12) | 382 per 1000 | 107 fewer per 1000 (202 fewer to 46 more) |
Feed intolerance (cases) during admission | 1114 (8 RCTs) | ⨁⨁⨁◯ Moderatea | RR 0.92 (0.77 to 1.10) | 306 per 1000 | 24 fewer per 1000 (70 fewer to 31 more) |
Time to regain birth weight (days) during admission | 993 (6 RCTs) | ⨁⨁⨁⨁ High | — | 14.7 d | MD 3.69 d fewer (4.44 fewer to 2.95 fewer) |
Wt (z score) at discharge | 2793 (1 RCT) | ⨁⨁◯◯ Lowa,d | — | −1.5 z score | MD 0.0 z score (0.08 less to 0.08 more) |
Wt (g/kg per day) at discharge | 131 (1 RCT) | ⨁⨁◯◯ Lowa,d | — | 12 g/kg per day | MD 0.5 g/kg per day more (1.2 less to 2.2 more) |
Wt (gm) at discharge | 100 (1 RCT) | ⨁⨁◯◯ Lowa,d | 1225 g | MD 29 gm less (75 less to 17 more) | |
Length | No studies | ||||
Head circumference (z score) at discharge | 2793 (1 RCT) | ⨁⨁◯◯ Lowa,d | — | −0.7 z score | MD 0.1 z score less (0.22 less to 0.02 more) |
Neurodevelopmental disability (cases) at 24 mo corrected age | 2325 (1 RCT) | ⨁⨁◯◯ Lowa,d | RR 1.12 (0.98 to 1.27) | 275 per 1000 | 33 more per 1000 (5 fewer to 74 more) |
Duration of hospitalization (days to discharge) | 3864 (7 RCTs) | ⨁⨁⨁◯ Moderatee | — | 27.8 d | MD 3.08 d fewer (4.34 fewer to 1.81 fewer) |
Outcomes . | No. of Participants (Studies) . | Certainty of the Evidence (GRADE) . | Relative Effect (95% CI) . | Anticipated Absolute Effects . | |
---|---|---|---|---|---|
Risk With Slow Advancement . | Risk Difference With Fast Advancement . | ||||
Mortality (cases) at discharge | 4132 (11 RCTs) | ⨁⨁⨁◯ Moderatea | RR 0.93 (0.73 to 1.18) | 78 per 1000 | 5 fewer per 1000 (21 fewer to 14 more) |
Necrotizing enterocolitis (cases) during admission | 4291 (12 RCTs) | ⨁⨁⨁◯ Moderatea | RR 0.89 (0.68 to 1.15) | 55 per 1000 | 6 fewer per 1000 (18 fewer to 8 more) |
Sepsis (cases) during admission | 3648 (9 RCTs) | ⨁⨁⨁◯ Moderatea | RR 0.92 (0.83 to 1.03) | 275 per 1000 | 22 fewer per 1000 (47 fewer to 8 more) |
Apnea (cases) during admission | 153 (2 RCTs) | ⨁⨁◯◯ Lowb,c | RR 0.72 (0.47 to 1.12) | 382 per 1000 | 107 fewer per 1000 (202 fewer to 46 more) |
Feed intolerance (cases) during admission | 1114 (8 RCTs) | ⨁⨁⨁◯ Moderatea | RR 0.92 (0.77 to 1.10) | 306 per 1000 | 24 fewer per 1000 (70 fewer to 31 more) |
Time to regain birth weight (days) during admission | 993 (6 RCTs) | ⨁⨁⨁⨁ High | — | 14.7 d | MD 3.69 d fewer (4.44 fewer to 2.95 fewer) |
Wt (z score) at discharge | 2793 (1 RCT) | ⨁⨁◯◯ Lowa,d | — | −1.5 z score | MD 0.0 z score (0.08 less to 0.08 more) |
Wt (g/kg per day) at discharge | 131 (1 RCT) | ⨁⨁◯◯ Lowa,d | — | 12 g/kg per day | MD 0.5 g/kg per day more (1.2 less to 2.2 more) |
Wt (gm) at discharge | 100 (1 RCT) | ⨁⨁◯◯ Lowa,d | 1225 g | MD 29 gm less (75 less to 17 more) | |
Length | No studies | ||||
Head circumference (z score) at discharge | 2793 (1 RCT) | ⨁⨁◯◯ Lowa,d | — | −0.7 z score | MD 0.1 z score less (0.22 less to 0.02 more) |
Neurodevelopmental disability (cases) at 24 mo corrected age | 2325 (1 RCT) | ⨁⨁◯◯ Lowa,d | RR 1.12 (0.98 to 1.27) | 275 per 1000 | 33 more per 1000 (5 fewer to 74 more) |
Duration of hospitalization (days to discharge) | 3864 (7 RCTs) | ⨁⨁⨁◯ Moderatee | — | 27.8 d | MD 3.08 d fewer (4.34 fewer to 1.81 fewer) |
Patient or population: preterm or low birth weight infants. Setting: hospitals. Intervention: fast advancement. Comparison: slow advancement. The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). GRADE Working Group grades of evidence: high certainty, we are very confident that the true effect lies close to that of the estimate of the effect; moderate certainty, we are moderately confident in the effect estimate and the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different; low certainty, our confidence in the effect estimate is limited and the true effect may be substantially different from the estimate of the effect; very low certainty, we have very little confidence in the effect estimate and the true effect is likely to be substantially different from the estimate of effect. CI, confidence interval; MD, mean difference; RR, risk ratio; —, not applicable.
Downgraded 1 level for imprecision caused by wide confidence interval crossing the line of no effect representing both appreciable benefit and appreciable harm.
Downgraded 1 level for imprecision because of small sample size, ie, optimal information size not met, ie, total cumulative study population is less than 300 participants for dichotomous outcomes and wide confidence interval crossing the line of no effect representing both appreciable benefit and appreciable harm.
Downgraded 1 level for some concerns of bias caused by the randomization process (allocation concealment, ie, not blinded).
Downgraded 1 level for heterogeneity as only 1 study and could not assess inconsistency.
Downgraded 1 level for serious unexplained heterogeneity as I squared = 79%.
On average, infants in the fast advancement group took 3.69 days fewer (95% CI −4.44 to −2.95, I2 = 70%, high certainty evidence, 6 trials, 993 participants) to regain birth weight than infants in slow advancement groups. The mean difference in weight z-score at discharge was 0 (95% CI −0.08 to 0.08, I2 = NA, low certainty evidence, 1 trial, 2793 participants); the mean difference in daily weight gain rate was 0.5 g/kg per day (95% CI −1.2 to 2.2, I2 = NA, low certainty evidence, 1 trial, 131 participants); the mean difference in weight at discharge was −29 g (95% CI −75 to 17, I2 = NA, low certainty evidence, 1 trial, 100 participants); the mean difference in head circumference z-score was −0.1 (95% CI −0.22 to 0.02, I2 = NA, low certainty evidence, 1 trial, 2793 participants). There were no trials assessing body length.
Also, infants randomized to fast feed advancement groups spent 3.08 fewer days in hospital on average than those in slow feed advancement groups. (95% CI −4.34 to −1.81, I2 = 77%, 7 trials, 3864 participants, moderate quality evidence) (Appendix 6).
Subgroups
In the subgroup analysis by gestational age and birth weight, there were no apparent differences in any outcomes in trials with only VLBW (<1.5 kg) or very preterm (<32 weeks) infants compared with trials with any LBW (<2.5 kg) or preterm (<37 weeks) infants (Appendix 6).
Discussion
In our systematic review of 12 RCTs and 4291 preterm or LBW infants, we found that fast feed advancement rates of more than 30 ml/kg per day reduced the time to regain birth weight by almost 4 days and the duration of hospitalization by 3 days when compared with slow advancement rates. We also found that fast feed advancement likely slightly reduces the risk of mortality, necrotizing enterocolitis, sepsis, and feed intolerance. In addition, fast feed advancement had little or no effect on weight or head circumference outcomes at discharge but there were no data on length or linear growth outcomes.
A recent 2021 Cochrane systematic review of 13 trials and 4033 VLBW infants compared slow advancement rates (≤24 ml/kg per day) and fast rates (>24 ml/kg per day), and reported similar findings on mortality, necrotizing enterocolitis, invasive infection, and feeding intolerance. There was 1 trial that was included in the Cochrane review but was not included in our review as both the intervention and the comparator groups had rates under 30 ml/kg per day.22 The Cochrane review did not assess growth or hospitalization outcomes.23 Based on our results, we consider that the effects of fast advancement on time to regain birth weight and duration of hospitalization are clinically important for preterm and LBW infants and their families. The possible reductions in mortality, necrotizing enterocolitis, and sepsis in fast feed advancement groups that we reported are biologically plausible; however, all lower bounds of the confidence intervals were consistent with no effect, so results must be viewed with caution.
We also found low certainty evidence of an increased risk of neurodevelopmental impairment at 24 months in the 1 trial (the SIFT trial) that reported this outcome (RR 1.12, 95% CI 0.98 to 1.27)19,34 Additionally, the SIFT trial found an increased risk of moderate to severe motor impairment in the fast feed advancement group.19,34 However, these findings need to be interpreted with extra caution for several reasons. The authors suggest such a relationship is biologically plausible if fast feed advancement, combined with poor ability to absorb enteral nutrition, increased the risk of cardiorespiratory events caused by pressure on the diaphragm. However, the biological association between feed advancement rates and neurodevelopment disability might be mediated by other factors, such as NEC. Second, both fast and slow groups in the SIFT trial had about 200 participants that were lost to follow-up and did not complete the neurodevelopment assessment. As the neurodevelopment status of this substantial portion of participants was unavailable, we should be careful in interpreting the finding and drawing conclusions. Importantly, these findings are from a single trial and may have occurred by chance as it was 1 of many secondary outcomes assessed and was not a prespecified trial outcome.19,34
Preterm infants are generally at increased risk for adverse health outcomes, and the risk increases at earlier gestational age and lower birth weight.36,37 The incidence of necrotizing enterocolitis increases from about 10% in infants with birth weight under 1.5 kg to 15% in infants with birth weight under 1.0 kg.38 Our analysis showed no evidence of a differential in effect in trials with only VLBW (<1.5 kg) or very preterm (<32 weeks) infants and trials with any LBW (<2.5 kg) or preterm (<37 weeks) infants. While we were unable to assess if there were differences in effect among extremely low birth weight infant infants (ELBW, <1.0 kg), the recent Cochrane review reported that there was no difference in effects of slow compared with fast advancement rates among ELBW infants compared with infants of 1.0 to 1.4 kg.23
There were other limitations to our review. Firstly, most RCTs excluded unstable infants with perinatal asphyxia or hemodynamic instability requiring vasopressors.39,40 Thus our evidence may not be generalizable to babies who are unstable or seriously unwell at the time of initiating enteral feeds. Secondly, most studies assessed outcomes while newborns were still hospitalized or at the time of discharge, and there was limited information on long-term outcomes. Thirdly, we were not able to assess the impact of higher feeding rate increments (eg, if 40 ml/kg per day is better than 30 ml/kg per day) or the impact of lower feeding rate increments (eg, if 10 ml/kg per day is better than 20 ml/kg per day). Subgroup analysis by milk type was not possible as only 1 studied formula exclusively. Further studies are needed to understand if a slower advancement rate is needed for exclusively formula-fed infants.
However, strengths of our study include the high to moderate certainty evidence from over 4000 participants in 12 trials, including studies from both high-income and middle-income countries. Also, studies started enteral feeds from 1 to 3 days of life and adhered to many standard recommendations for NICU enteral feeding, including target volumes of 135 to 200 ml/kg per day.41 Thus, results should be generalizable to infants in hospitals receiving these standard feeding regimes. Our results are also directly relevant to current WHO feeding guidelines.
In conclusion, we found that fast feed advancement rates at or above 30 ml/kg per day reduced the time to regain birth weight and duration of hospitalization. Also, fast feed advancement likely slightly reduces the risks of mortality, necrotizing enterocolitis, and feed intolerance. Fast feed advancement may result in little or no difference in weight gain or head circumference at discharge. More studies are needed to understand the long-term effects of fast feed advancement.
Drs Yang and Smith conceptualized and designed the study, conducted the initial analyses, drafted the initial manuscript, and reviewed and revised the manuscript; Ms Fogel, Ms Lauria, and Ms Ferguson designed the data collection instruments, collected data, and reviewed and revised the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
This trial is registered at Prospero CRD42021241656 https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=241656.
FUNDING: All phases of this study were supported by a contract from the World Health Organization Department of Maternal, Child, Adolescent Health and Aging to The George Washington University (ERS). The World Health Organization (Karen Edmond) assisted in the conceptualization and design of the study, conducted additional analyses, and assisted in the review and revision of the manuscript.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no conflicts of interest to disclose.
Comments