Responsive feeding may improve health outcomes in preterm and low birth weight (LBW) infants. Our objective was to assess effects of responsive compared with scheduled feeding in preterm and LBW infants.
Data sources include PubMed, Scopus, Web of Science, CINAHL, LILACS, and MEDICUS. Randomized trials were screened. Primary outcomes were mortality, morbidity, growth, neurodevelopment. Secondary outcomes were feed intolerance and duration of hospitalization. Data were extracted and pooled with random-effects models.
Eleven eligible studies were identified, and data from 8 randomized control trials with 455 participants were pooled in the meta-analyses. At discharge, the mean difference in body weight between the intervention (responsive feeding) and comparison (scheduled feeding) was −2.80 g per day (95% CI −3.39 to −2.22, I2 = 0%, low certainty evidence, 4 trials, 213 participants); −0.99 g/kg per day (95% CI −2.45 to 0.46, I2 = 74%, very low certainty evidence, 5 trials, 372 participants); −22.21 g (95% CI −130.63 to 86.21, I2 = 41%, low certainty evidence, 3 trials, 183 participants). The mean difference in duration of hospitalization was −1.42 days (95% CI −5.43 to 2.59, I2 = 88%, very low certainty evidence, 5 trials, 342 participants). There were no trials assessing other growth outcomes (eg, length and head circumference) mortality, morbidity or neurodevelopment. Limitations include a high risk of bias, heterogeneity, and small sample size in included studies.
Overall, responsive feeding may decrease in-hospital weight gain. Although the evidence is very uncertain, responsive feeding may slightly decrease the duration of hospitalization. Evidence was insufficient to understand the effects of responsive compared with scheduled feeding on mortality, morbidity, linear growth, and neurodevelopmental outcomes in preterm and LBW infants.
One of the most important interventions to improve prognosis, growth, and development of preterm and low birth weight (LBW) infants is optimal feeding practices.1 There are broadly 2 major approaches to when to feed an infant scheduled feeding and responsive feeding. Scheduled feeding of infants involves feeding at regularly timed intervals, whereas responsive feeding involves a more dynamic call-and-response feeding style.2,3 The 3 main components of responsive feeding are: the infant exhibits a cue of hunger or satiety, the parent or caretaker engages with and responds to the infant’s cues, and the infant develops an understanding of the parental or caretaker response patterns according to the cue.4 Infant cues include but are not limited to crying, hand-mouth motions, suckling, and mindful awakeness.2,5
A 2016 Cochrane review reported that responsive feeding of preterm infants decreased the transition time from enteral tube to oral feeding, but it may have led to slower weight gain.6 A 2018 systematic review found that responsive feeding reduced time to reaching satiety, decreased the length of NICU hospitalization, and increased weight gain.7 No other outcomes have been reported, and there have been no systematic reviews of effects in nonpreterm low birth weight (<2.5 kg) infants.
Our objective was to evaluate the effect of responsive versus scheduled feeding on morbidity, mortality, growth, and neurodevelopment in preterm (<37 weeks gestation) and LBW infants.
Methods
Protocol and Registration
The protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) database (Registration number CRD42021240294). This study was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.8
Search Strategy
The full search strategy is in Appendix 1. We searched the following databases: Medline, Scopus, Web of Science, CINAHL, and Index Medicus on March 31, 2021.
Inclusion Criteria
We included randomized control trials (RCTs) comparing the effects of responsive versus scheduled feeding in preterm (<37 weeks) and LBW (<2.5 kg) infants. Our primary outcomes were mortality, morbidity, growth, and neurodevelopment at latest follow-up (Appendix 2). Secondary outcomes were feed intolerance and duration of hospitalization. Studies in any language were eligible.
Study Selection and Data Extraction
Two reviewers independently screened reports and extracted data, and conflicts were resolved by a third reviewer. Both title and abstract screening and full-text screening were managed on web-based software, Covidence.9
Risk of Bias
Two reviewers independently assessed the risk of bias of individual studies using the Cochrane Risk of Bias 2 tool.10 The Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach was adopted to assess the overall quality of the evidence and GRADEpro GDT software was used.11
Statistical Analysis
Data were pooled using random-effects models. For binary outcomes, risk ratios (RR) with 95% confidence intervals (CI) were used to summarize results. To include studies with 0 events in 1 arm, we applied a continuity correction in which 0.5 was added to each cell of the table. For continuous outcomes, we pooled the mean difference estimates between groups. We prespecified a gestational age and birth weight subgroup analyses to explore potential sources of heterogeneity. However, data were insufficient to perform this analysis. All analyses were done using Stata/IC 16.1 and Revman software.
Differences Between Protocol and Review
We included studies published only as abstract if the data had been published in the corresponding Cochrane review and used information from the Cochrane risk of bias assessment to complete our own assessment when information was not available from the published abstract.6
Results
Included Studies and Characteristics
We identified 11 eligible RCTs with 662 participants (see PRISMA flowchart, Appendix 3). Characteristics of included studies are presented in Appendix 4. The trials were conducted in the United States, Canada, Iran, and Israel. Only 1 study recruited exclusively preterm infants under 32 weeks gestation, 5 studies recruited moderately preterm infants born at 34 weeks or younger, and 3 studies included any preterm infant (<37 weeks gestation). Data from 8 RCTs with 455 participants were pooled in the meta-analyses.
Among included studies, the initiation of intervention began as early as 27 weeks postconceptual age to as late as 37 weeks postconceptual age. The duration of the interventions varied from 3 to 32 days (median 12 days, interquartile range 3–25 days). Most included studies continued the intervention until hospital discharge, however duration of hospitalization ranged from 10 to 115 days across studies. All intervention infants were fed in response to a standard protocol that included initiation in response to hunger cues (crying, sucking, rooting, or “unsettled”), and feeds ceased in response to any infant satiation cues that were unspecified. The comparator groups all received scheduled feeding every 3 to 4 hours.
Risk of Bias Assessment
We completed a risk of bias of assessment studies contributing to each outcome (Appendix 4). One study had some concerns, and 10 studies were at high risk of bias. Most bias was related to the randomization process because of the lack of concealment of allocation.
Primary Outcomes
Regarding weight gain velocity at discharge, the mean difference between intervention (responsive feeding) and comparison (scheduled feeding) was −2.80 g per day (95% CI −3.39 to −2.22, I2 = 0%, low certainty evidence, 4 trials, 213 participants), and the mean difference was −0.99 g/kg per day (95% CI −2.45 to 0.46, I2 = 74%, very low certainty evidence, 5 trials, 372 participants) when weight gain velocity was adjusted by bodyweight. The mean difference in final infant weight at discharge was −22.21 g (95% CI −130.63 to 86.21, I2 = 41%, low certainty evidence, 3 trials, 183 participants). There were no trials assessing length, head circumference, there were also no trials assessing mortality, morbidity, or neurodevelopment (Table 1).
Summary of Findings: Responsive Feeding Compared With Scheduled Feeding for Preterm and LBW Infants
| Outcomes . | No. of Participants (Studies) Follow-up . | Certainty of the Evidence (GRADE) . | Relative Effect (95% CI) . | Anticipated Absolute Effects . | |
|---|---|---|---|---|---|
| Risk With Scheduled Feeding . | Risk Difference With Responsive Feeding . | ||||
| Mortality | No trials | — | — | — | — |
| Morbidity | No trials | — | — | — | — |
| Weight, g, at discharge | 183 (3 RCTs) | ⨁⨁◯◯ Lowa,b | — | 2173.89 g | MD 22.21 gm less (130.63 fewer to 86.21 more) |
| Weight, g/day, at discharge | 213 (4 RCTs) | ⨁⨁◯◯ Low,a,b | — | 36.11 g/day | MD 2.80 gm/day less (3.39 fewer to 2.22 fewer) |
| Weight, g/kg per day, at discharge | 372 (5 RCTs) | ⨁⨁⨁◯ Very lowa,b,c | — | 9.93 g/kg per day | MD 0.99 gm less (2.45 fewer to 0.46 more) |
| Duration of hospitalization, days to discharge | 342 (5 RCTs) | ⨁⨁⨁◯ Very lowa,b,c | — | 32.49 d | MD 1.42 d fewer (5.43 fewer to 2.59 more) |
| Neurodevelopment | No trials | — | — | — | — |
| Outcomes . | No. of Participants (Studies) Follow-up . | Certainty of the Evidence (GRADE) . | Relative Effect (95% CI) . | Anticipated Absolute Effects . | |
|---|---|---|---|---|---|
| Risk With Scheduled Feeding . | Risk Difference With Responsive Feeding . | ||||
| Mortality | No trials | — | — | — | — |
| Morbidity | No trials | — | — | — | — |
| Weight, g, at discharge | 183 (3 RCTs) | ⨁⨁◯◯ Lowa,b | — | 2173.89 g | MD 22.21 gm less (130.63 fewer to 86.21 more) |
| Weight, g/day, at discharge | 213 (4 RCTs) | ⨁⨁◯◯ Low,a,b | — | 36.11 g/day | MD 2.80 gm/day less (3.39 fewer to 2.22 fewer) |
| Weight, g/kg per day, at discharge | 372 (5 RCTs) | ⨁⨁⨁◯ Very lowa,b,c | — | 9.93 g/kg per day | MD 0.99 gm less (2.45 fewer to 0.46 more) |
| Duration of hospitalization, days to discharge | 342 (5 RCTs) | ⨁⨁⨁◯ Very lowa,b,c | — | 32.49 d | MD 1.42 d fewer (5.43 fewer to 2.59 more) |
| Neurodevelopment | No trials | — | — | — | — |
Patient or population: preterm and LBW infants. Setting: hospitals. Intervention: early feeding. Comparison: delayed feeding. 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; and 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; g, gram; MD, mean difference; RR, risk ratio; —, not applicable.
Downgraded 1 level for some concerns of bias caused by the randomization process (allocation concealment, ie, not blinded) and selection of the reported result (no protocol).
Downgraded 1 level for imprecision caused by small sample size ie optimal information size not met, ie, the total cumulative study population is less than 400 participants for continuous outcomes and wide confidence interval crossing the line of no effect representing both appreciable benefit and appreciable harm.
Downgraded 1 level for serious unexplained heterogeneity as I squared > 70%.
Other Outcomes
The mean difference in duration of hospitalization comparing responsive to scheduled feeding was −1.42 days (95% CI −5.43 to 2.59, I2 = 88%, very low certainty evidence, 5 trials, 342 participants).
Subgroups
There was only 1 trial that restricted recruitment to infants under 32 weeks gestation, so it was not possible to perform subgroup analyses for very preterm or very LBW infants.
Discussion
Our systematic review comprised 11 trials with 662 total infants. Our systematic review suggests that responsive feeding may decrease in-hospital weight gain in preterm and low birth weight infants. Responsive feeding may slightly reduce the duration of hospitalization after birth, but the evidence is very uncertain. There were no studies that reported effects on mortality, morbidity, or neurodevelopment.
We included 8 studies with 455 participants that reported on weight gain during hospital admission.12–18 Four studies suggested low certainty evidence of a decrease in daily weight of 2.80 g per day during the hospital admission.15–18 Five studies suggested very low certainty evidence of a decrease in daily weight per kilogram body weight of 0.99 g per day during the hospital admission.13,14,16,17,19 Three other studies reported low certainty evidence of a decrease in weight at hospital discharge of 22 g.12–14 These findings are similar to a 2015 Cochrane review which found low quality evidence that infants randomized to responsive feeding had a lower daily weight gain rate (MD −1.36 g/kg per day, 95% CI −2.44 to −0.29 g/kg; 4 studies; 305 participants). The 2016 Cochrane review reported that responsive feeding decreased the transition time from enteral tube feeding to oral feeding by 5.5 days (MD −5.5 days, 95% CI −4.2 to −6.8 days, 2 studies; 167 participants).6 We also found very low certainty evidence from 5 studies of that responsive feeding may slightly reduce on the duration of hospitalization, an outcome that was not reported in the Cochrane review.12,14,16,17,19
Responsive feeding is a fundamental component of nurturing and responsive care of all infants regardless of cultural setting and education level.20,21 Responsive feeding requires family attention and sensitivity to visual and auditory cues or “leads” from infants. This sensitivity may be innate for some caretakers, but also can be learned and is essential in responding to an infant’s feeding needs.22 However, preterm infants are generally highly reflexive, autonomic, and less responsive to environmental stimuli in the first days after birth. Their arousal, breathing, and motor functions vary greatly and develop rapidly over time.23,24 The responses of preterm infants can be difficult to correctly perceive and understand for some parents.23,24–29 Careful training of health care providers and support of families is often needed to ensure responsive feeding meets an infant’s needs. However, none of the trials included in our review described the methods of training of health care providers or the support of families to provide responsive feeding. The environment in which responsive feeding takes place (eg, a noisy and brightly lit NICU) also can influence the degree to which responsive feeding can be successful.22 All of the trials included in our review were conducted in NICU settings.
There are limitations to our review. The studies were limited to the United States, Canada, Iran, and Israel. Sample sizes were small, and risk of bias was high because of methodological limitations in the studies, and heterogeneity was substantial in the pooled estimates. Further, there was variation across studies in the gestational age at birth, time of intervention initiation, and duration of the intervention; this makes it challenging to extrapolate the data to what is best for all preterm and LBW infants. However, we conducted a high-quality systematic search and were able to include 8 RCTs with over 400 infants in our meta-analyses. To our knowledge, this is the first systematic review on the subject of responsive and scheduled feeding to include LBW infants in its comprehensive search strategy.
Overall, responsive feeding may decrease in-hospital weight gain. Although the evidence is very uncertain, responsive feeding may slightly decrease the duration of hospitalization. Evidence was insufficient to understand the effects of responsive compared with scheduled feeding on mortality, morbidity, linear growth, and neurodevelopmental outcomes in preterm and LBW infants. Further studies are needed and should also investigate the best methods of training health care providers and supporting families in the provision of responsive feeding.
Ms Talej and Drs He and Ms Smith conceptualized and designed the study, conducted the initial analyses, and drafted the initial manuscript; Ms Lauria, Ms Chitale, and Ms Ferguson designed the data collection instruments and collected data; and all authors reviewed and revised the manuscript and approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
This trial is registered at PROSPERO 2021, https://www.crd.york.ac.uk/prospero/display_record.php? (identifier, CRD42021240294CRD42021240294).
FUNDING: All phases of this study were supported by a contract from the World Health Organization (WHO) Department of Maternal, Child, Adolescent Health and Aging to The George Washington University. The sponsor, WHO, commissioned the review for a guideline development group meeting for development of WHO recommendations on care of the preterm or low birth weight infant. The sponsor assisted in formulating the research question and provided inputs on the synthesis of the results and manuscript.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no conflicts of interest to disclose.
Comments