Alimentación enteral completa temprana para lactantes prematuros o de bajo peso al nacer
Resumen
Antecedentes
La introducción y el avance de la alimentación enteral para los lactantes prematuros o de bajo peso al nacer se suele retrasar debido a la preocupación de que la alimentación enteral completa temprana no sea bien tolerada o pueda aumentar el riesgo de enterocolitis necrosante. Sin embargo, la alimentación enteral completa temprana podría aumentar la ingesta de nutrientes y las tasas de crecimiento, acelerar la transición posnatal fisiológica, metabólica y de la microbioma intestinal, y reducir el riesgo de complicaciones asociadas con los dispositivos intravasculares para la administración de líquidos.
Objetivos
Determinar la forma en la que la alimentación enteral completa temprana, en comparación con la introducción retardada o progresiva de la alimentación enteral, afecta el crecimiento y los eventos adversos como la enterocolitis necrosante, en los lactantes prematuros o de bajo peso al nacer.
Métodos de búsqueda
Se utilizó la estrategia de búsqueda estándar del Grupo Cochrane de Neonatología para buscar en el Registro Cochrane central de ensayos controlados (Cochrane Central Register of Controlled Trials); en MEDLINE Ovid, Embase Ovid, Maternity & Infant Care Database Ovid, el Cumulative Index to Nursing and Allied Health Literature, y en las bases de datos de ensayos clínicos, las actas de congresos y las listas de referencias de los artículos recuperados los ensayos controlados aleatorizados y cuasialeatorizados hasta octubre de 2020.
Criterios de selección
Ensayos controlados aleatorizados que compararon la alimentación enteral completa temprana con la introducción retardada o progresiva de la alimentación enteral en lactantes prematuros o de bajo peso al nacer.
Obtención y análisis de los datos
Se utilizaron los métodos estándar del Grupo Cochrane de Neonatología. Dos autores de la revisión por separado evaluaron la elegibilidad de los ensayos y su calidad, extrajeron los datos y resumieron las estimaciones del efecto mediante las razones de riesgos (RR), las diferencias de riesgos y las diferencias de medias (DM), con los intervalos de confianza (IC) del 95%. Se utilizaron los criterios GRADE para evaluar la certeza de la evidencia.
Resultados principales
Se incluyeron seis ensayos. Todos se realizaron en la década de 2010 en instalaciones de atención neonatal en la India. En total participaron 526 lactantes. En su mayoría fueron lactantes muy prematuros con un peso al nacer entre 1000 g y 1500 g. Los ensayos fueron de buena calidad metodológica, pero una posible fuente de sesgo fue que los padres, los médicos y los investigadores no estaban cegados a la intervención. Los ensayos compararon la alimentación completa temprana (60 ml/kg a 80 ml/kg el primer día después del parto) con la alimentación enteral mínima (habitualmente 20 ml/kg el primer día), complementada con líquidos intravenosos. Los volúmenes de alimentación aumentaron diariamente, según lo tolerado, de 20 ml/kg a 30 ml/kg de peso corporal, hasta un volumen objetivo en estado estable de 150 ml/kg a 180 ml/kg/día. Todos los lactantes participantes se alimentaron preferentemente con leche materna extraída, en dos ensayos los volúmenes insuficientes se complementaron con leche materna de donante y en cuatro se utilizó leche maternizada para prematuros.
Hubo pocos datos disponibles para evaluar los parámetros de crecimiento. Un ensayo (64 participantes) informó de una tasa de aumento de peso más lenta (diferencia mediana ‐3,0 g/kg/día), y otro (180 participantes) informó de una tasa de aumento de peso más rápida en el grupo de alimentación enteral completa temprana (DM 1,2 g/kg/día). No se realizaron metanálisis de estos datos (evidencia de certeza muy baja). Ninguno de los ensayos informó sobre la tasa de crecimiento del perímetro cefálico. Un ensayo informó que la puntuación z media del peso al alta hospitalaria fue mayor en el grupo de alimentación enteral completa temprana (DM 0,24; IC del 95%: 0,06 a 0,42; evidencia de certeza baja). Los metanálisis no mostraron evidencia de un efecto sobre la enterocolitis necrosante (RR 0,98; IC del 95%: 0,38 a 2,54; seis ensayos, 522 participantes; I² = 51%; evidencia de certeza muy baja).
Conclusiones de los autores
Los ensayos no proporcionaron datos suficientes para determinar con certeza la forma en la que la alimentación enteral completa temprana, en comparación con la introducción retrasada o progresiva de la alimentación enteral, afecta al crecimiento de los lactantes prematuros o de bajo peso al nacer. No se conoce con certeza si la alimentación enteral completa temprana afecta el riesgo de enterocolitis necrosante debido al riesgo de sesgo en los ensayos (a causa de la falta de enmascaramiento), la inconsistencia y la imprecisión.
PICO
Resumen en términos sencillos
Alimentos enterales completos tempranos para recién nacidos prematuros o de bajo peso al nacer
Pregunta de la revisión: ¿los recién nacidos prematuros o de bajo peso al nacer (lactantes nacidos prematuramente o pequeños) crecen más rápido y tienen menos problemas cuando reciben todos los nutrientes en forma de alimentos con leche desde poco después del parto (en comparación con la introducción gradual de los alimentos con leche mientras se administran líquidos o nutrientes a través de un goteo intravenoso [una infusión lenta en la sangre a través de una vena])?
Antecedentes: "alimentación enteral completa temprana" significa que los recién nacidos prematuros o de bajo peso al nacer reciben toda la nutrición en forma de alimentación con leche desde poco después del parto, y no reciben líquidos suplementarios ni nutrición por goteo intravenoso. La evaluación de si este enfoque es seguro y beneficioso es particularmente pertinente para la alimentación de los recién nacidos muy prematuros o de muy bajo peso al nacer (nacidos antes de las 32 semanas o con un peso inferior a 1500 g).
Características de los estudios: se incluyeron seis ensayos, todos realizados en unidades de atención neonatal en la India durante la década de 2010. Los ensayos fueron en general de buena calidad, aunque la mayoría fueron pequeños (en total 526 recién nacidos). Los participantes fueron recién nacidos prematuros con un peso al nacer entre 1000 y 1500 g.
La búsqueda está actualizada hasta julio de 2020.
Resultados clave: no hubo datos suficientes para demostrar si los recién nacidos que recibieron alimentación completa con leche desde el parto aumentaron de peso y crecieron más rápidamente que los lactantes en los que se introdujeron gradualmente los alimentos durante la primera o segunda semana después del parto. Los ensayos no informaron sobre los efectos que la alimentación completa temprana con leche podría tener sobre el desarrollo y el crecimiento posterior del recién nacido. Los ensayos incluidos no encontraron evidencia de otros posibles efectos beneficiosos o perjudiciales de la alimentación completa temprana, incluidos los efectos sobre la alimentación o los problemas intestinales.
Conclusión: no hay evidencia suficiente para determinar si la alimentación completa temprana con leche beneficia a los recién nacidos prematuros o de bajo peso al nacer. Se necesitarían nuevos ensayos para resolver esta incertidumbre.
Calidad de la evidencia: la evidencia se calificó como de calidad baja o muy baja porque los ensayos incluidos fueron pequeños, tuvieron algunas deficiencias metodológicas y sus hallazgos no fueron consistentes. Lo anterior significa que es muy probable que los estudios de investigación adicionales tengan un impacto importante sobre las estimaciones del efecto y en la confianza en los hallazgos.
Authors' conclusions
Summary of findings
| Early full compared to delayed/progressive enteral feeding for preterm or low birth weight infants | |||||
| Patient or population: preterm or low birth weight infants | |||||
| Outcomes | № of participants | Certainty of the evidence | Relative effect | Anticipated absolute effects* (95% CI) | |
|---|---|---|---|---|---|
| Risk with delayed/progressive enteral feeding | Risk difference with early full enteral feeding | ||||
| In hospital rate of weight gain (g/kg/day) until term equivalent | 236 | ⊕⊝⊝⊝ | — | Meta‐analysis not possible due to different outcome measures. | |
| In hospital rate of head circumference growth (cm/week) until term equivalent – not reported | — | — | — | — | — |
| Growth restriction (z‐score of weight) at hospital discharge | 46 | ⊕⊕⊝⊝ | — | The mean growth restriction (z‐score of weight) at hospital discharge was –1.09 | Mean 0.24 higher |
| Necrotising enterocolitis | 522 | ⊕⊝⊝⊝ | RR 0.98 | Study population | |
| 31 per 1000 | 1 fewer per 1000 | ||||
| *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). CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio. | |||||
| GRADE Working Group grades of evidence | |||||
| aDowngraded one level due to risk of surveillance and detection bias owing to lack of masking. | |||||
Background
This Cochrane Review appraised and synthesised data from randomised controlled trials (RCTs) that compared early full enteral feeding with delayed or progressive introduction of enteral feeds (combined with parenteral fluids or nutrition) for preterm or low birth weight (LBW; less than 2500 g) infants.
Description of the condition
Preterm or LBW infants, especially very preterm or very low birth weight (VLBW; less than 1500 g) infants, have few nutrient reserves at birth and are subject to physiological and metabolic stresses that increase their nutrient needs. Recommendations on nutrient requirements for preterm or LBW infants assume that the optimal rate of postnatal growth should be similar to that of uncompromised fetuses of an equivalent gestational age (Agostoni 2010; Tsang 1993). Such levels of nutrient input and growth are rarely achieved and most very preterm or VLBW infants accumulate nutrient deficits during their initial hospital stay (Embleton 2001; Horbar 2015). By the time they are ready to go home, many of these infants are growth restricted compared to their term‐born peers (Clark 2003; Dusick 2003). Growth deficits can persist through childhood and adolescence, and are associated with adverse neurodevelopmental, cognitive, and educational outcomes (Bracewell 2008; Cooke 2003; Doyle 2004; Farooqi 2006; Ford 2000; Hack 1991; Leppänen 2014).
Necrotising enterocolitis
Necrotising enterocolitis, a syndrome of acute intestinal necrosis of unknown aetiology, affects about 1 in 20 very preterm or VLBW infants (Gagliardi 2008; Holman 1997; Moro 2009). Infants who develop necrotising enterocolitis experience more infections, have lower levels of nutrient intake, grow more slowly, and have longer durations of intensive care and hospital stay than gestation‐comparable infants who do not develop necrotising enterocolitis (Bisquera 2002; Guthrie 2003). The associated mortality rate is greater than 20% (Fitzgibbons 2009). Compared with their peers, infants who develop necrotising enterocolitis have a higher incidence of long‐term neurological disability, which may be a consequence of infection and undernutrition during a critical period of brain development (Berrington 2012; Pike 2012; Rees 2007).
In addition to low gestational age at birth, the major risk factor for necrotising enterocolitis is intrauterine growth restriction, especially if it is associated with absent or reversed end diastolic flow velocities in Doppler studies of the fetal aorta or umbilical artery (Dorling 2005; Samuels 2017). Most very preterm or VLBW infants who develop necrotising enterocolitis have received enteral milk feeds (Ramani 2013). Feeding with artificial formula rather than human milk increases the risk of developing necrotising enterocolitis (Quigley 2019; Walsh 2019). However, the available data from RCTs do not provide evidence that delaying the introduction of progressive enteral feeds beyond four days after birth, or advancing feed volumes more slowly than 24 mL/kg/day, affects the risk of necrotising enterocolitis and associated morbidities or mortality in very preterm or VLBW infants (Morgan 2013; Morgan 2014; Oddie 2017).
Early enteral feeding strategies
Evidence exists that early enteral feeding strategies – particularly the timing of introduction and the rate of advancement of milk feeds – affect important outcomes in preterm or LBW infants, including nutrient intake, the risk of necrotising enterocolitis, and growth and development (Embleton 2013). Approaches to early enteral feeding vary by the gestational age and clinical condition of the infant (Hay 2008; Klingenberg 2012). Stable preterm infants born at or more than 32 weeks' gestation, or with a birth weight of 1500 g or greater, are generally treated similarly to well term infants; all fluid and nutrition is supplied enterally from birth, either orally or via an intragastric feeding tube. In contrast, newborn infants who are extremely preterm (born before 28 weeks' gestation), or of extremely low birth weight (ELBW; less than 1000 g) are commonly affected by respiratory distress, have delayed gastric emptying and inefficient intestinal motility, and are at high risk of developing necrotising enterocolitis. These infants tend to be supported with parenteral fluids and nutrition during the first few days after birth. Enteral milk feeds are then introduced in subnutritional volumes (trophic feeds) during the first week after birth, ideally as colostrum or expressed breast milk, and the volume of feeds is increased gradually over the next one to two weeks as the volume of parenteral nutrition is decreased.
There remains substantial variation in practice with regard to early enteral feeding strategies for those preterm infants born at gestations between approximately 28 and 32 weeks (or with birth weights between approximately 1000 g and 1500 g). This variation reflects ongoing uncertainty about whether these infants should be treated in the same way as those infants born at a later gestation (i.e. with full enteral feeds from birth), or more similarly to extremely preterm or ELBW infants (i.e. delayed introduction and gradual advancement of milk feeds while supporting nutritional needs with parenteral nutrition). In many high‐income countries, policy and practice has tended to favour the conservative approach to introducing and advancing enteral feeds for these infants because of concerns that early full enteral feeding might increase the risk of feed intolerance, gastro‐oesophageal reflux and aspiration of stomach contents, hypoglycaemia, and necrotising enterocolitis in very preterm or VLBW infants (de Waard 2018; Klingenberg 2012; Leaf 2013; Maas 2018). However, in low‐ and middle‐income countries with fewer resources for neonatal care, practice is more pragmatic and tends to favour early introduction and advancement of enteral feeds (sometimes facilitated by 'kangaroo' mother care) for stable infants born at 28 weeks' gestation or greater, or with a birth weight of 1000 g or more (Conde‐Agudelo 2016; Sankar 2008).
Description of the intervention
Early (sometimes termed 'immediate') full enteral feeding means that newborn infants receive all their prescribed nutrition as milk feeds (either human milk or formula) and do not receive any supplemental parenteral fluids or nutrition from birth (Nangia 2018). This approach for feeding preterm infants has been advocated since the earliest days of modern neonatology but has tended to be reserved for clinically stable preterm infants of gestational age at birth of more than approximately 32 weeks (Klingenberg 2012; Smallpeice 1964). In most neonatal care facilities, particularly in high‐income countries, the more common practice is to introduce enteral milk feeds for very preterm or VLBW infants at low volume (trophic feeds or minimal enteral nutrition) and to then advance the feed volume slowly during the next one to two weeks. During this time, infants receive most of their fluids and nutrition parenterally, usually in the form of commercially available solutions containing amino acids, glucose, minerals, vitamins, and fats (Klingenberg 2012).
There are potential disadvantages associated with conservative enteral feeding regimens (Flidel‐Rimon 2004; Flidel‐Rimon 2006). Because gastrointestinal hormone secretion and motility are stimulated by milk feeds, delaying full enteral feeding may diminish the functional adaptation of the gastrointestinal tract and disrupt the patterns of microbial colonisation (Embleton 2017). Intestinal dysmotility and dysbiosis might exacerbate feed intolerance and delay the establishment of enteral feeding independently of parenteral nutrition (Pammi 2017). Prolonging the duration of parenteral nutrition is associated with infectious and metabolic complications that increase mortality and morbidity, prolong hospital stay, and adversely affect growth and development (Embleton 2013). Due to cost and equipment implications, parenteral nutrition is less easily available in low‐ and middle‐income countries. In high‐income settings, earlier achievement of full enteral feeds and an associated reduction of length of hospital stay could be associated with considerable resource savings (Embleton 2014).
How the intervention might work
The aims of early full enteral feeding are to avoid the risks and costs associated with provision of parenteral nutrition and to accelerate gastrointestinal physiological, endocrine and metabolic maturity and so allow infants to attain nutrient intakes to optimise growth and development. Early full enteral feeding in this vulnerable population might reduce the risk of infection associated with intravascular devices used to deliver parenteral nutrition (Flidel‐Rimon 2004). Early provision of breast milk might promote successful expression and lactation and help establish maternal breast milk feeding as the primary source of infant nutrition (Hay 2008; Senterre 2014). However, there is some concern that exclusive enteral nutrition might not be sufficient to maintain normal blood glucose levels during the early metabolic transition phase in very preterm or VLBW infants, particularly in growth‐restricted infants with limited nutrient reserves at birth (Klingenberg 2012). Furthermore, any beneficial effects may be negated if early full enteral feeding increases the risk of feed intolerance or necrotising enterocolitis in very preterm or VLBW infants (Chauhan 2008; Samuels 2017).
Why it is important to do this review
This review focuses on the comparison of early full enteral feeding versus gradual introduction of progressive enteral feeding in combination with parenteral fluids. Other Cochrane Reviews have addressed the questions of whether delaying the introduction of any enteral milk feeding or restricting feed volumes to trophic levels (minimal enteral nutrition) affect the risk of necrotising enterocolitis in very preterm or VLBW infants (Morgan 2013; Morgan 2014). The findings of this review complement these other reviews of early enteral feeding strategies and might inform policy, practice, or research in this field (Chauhan 2008).
Objectives
To determine how early full enteral feeding, compared with delayed or progressive introduction of enteral feeds, affects growth and adverse events such as necrotising enterocolitis, in preterm or low birth weight infants.
Where data were available, we planned subgroup analyses of very preterm or VLBW infants (versus infants born after a longer gestation or with higher birth weight), infants 'small for gestational age' at birth (versus those 'appropriate for gestation'), infants fed with human milk only (versus formula‐fed infants), and trials set in low‐ or middle‐income countries (versus high‐income countries) (see: Subgroup analysis and investigation of heterogeneity).
Methods
Criteria for considering studies for this review
Types of studies
We included RCTs or quasi‐RCTs, including cluster‐RCTs.
Types of participants
We included infants born preterm (less than 37 weeks' gestation) or with LBW (less than 2500 g).
Types of interventions
Intervention
Full enteral feeds from birth without parenteral fluids or nutrition. Early full enteral feeding is defined as sufficient volumes of milk being fed orally or via an enteric feeding tube from soon after birth, without parenteral supplementation at any point.
Comparison
Any other feeding regimen, such as delayed initiation of full milk feeds and gradual advancement of feed volumes while receiving supplemental fluid or nutrients parenterally.
Types of outcome measures
Primary outcomes
-
In hospital rate of weight gain (g/kg/day) until term equivalent.
-
In hospital rate of head circumference growth (cm/week) until term equivalent.
-
Growth restriction: z‐score and proportion of infants who remain below the 10th percentile for the index population distribution of weight at term equivalent.
-
Necrotising enterocolitis, confirmed at surgery or autopsy or diagnosed by at least two of the following clinical features (Kliegman 1987):
-
abdominal radiograph showing pneumatosis intestinalis or gas in the portal venous system or free air in the abdomen;
-
abdominal distension with abdominal radiograph with gaseous distension or frothy appearance (or both) of bowel lumen;
-
blood in stool;
-
lethargy, hypotonia, or apnoea (or combination of these).
-
Secondary outcomes
-
Feed intolerance during the trial intervention period that results in cessation in enteral feeding for more than four hours.
-
Episodes of hypoglycaemia (investigator defined) requiring treatment with enteral supplement (including milk feed or buccal dextrose gel) or with intravenous fluids (including dextrose solution).
-
Invasive infection, as determined by culture of bacteria or fungus from blood, cerebrospinal fluid, or from a normally sterile body space.
-
Duration of birth hospitalisation (days).
-
All‐cause mortality up to 36 to 44 weeks' postmenstrual age and one‐year post‐term.
-
Growth parameters assessed beyond infancy: weight, height, or head circumference and proportion of infants who remain below the 10th percentile for the index population's distribution, and measures of body composition (lean/fat mass) and body mass index.
-
Severe neurodevelopmental disability, assessed beyond infancy until adulthood: non‐ambulant cerebral palsy, developmental quotient more than two standard deviations (SD) below the population mean and blindness (visual acuity less than 6/60) or deafness (any hearing impairment requiring or unimproved by amplification).
Search methods for identification of studies
We used the standard search strategy of Cochrane Neonatal.
Electronic searches
We conducted a comprehensive search including: Cochrane Central Register of Controlled Trials (CENTRAL; 2020, Issue 10) in the Cochrane Library; MEDLINE Ovid (1946 to October 2020), Embase Ovid (1974 to October 2020), Maternity & Infant Care Database Ovid (1971 to October 2020), and the Cumulative Index to Nursing and Allied Health Literature EBSCO (1982 to October 2020), using terms described in Appendix 1. We limited the search outputs with the relevant search filters for clinical trials as recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017). We applied no language restrictions.
We searched ClinicalTrials.gov and the World Health Organization's International Clinical Trials Registry Platform (www.who.int/ictrp/en/) for completed or ongoing trials.
Searching other resources
We examined the reference lists of studies identified as potentially relevant. We searched the abstracts from annual meetings of the Pediatric Academic Societies (1993 to 2019), the European Society for Paediatric Research (1995 to 2020), the UK Royal College of Paediatrics and Child Health (2000 to 2019), and the Perinatal Society of Australia and New Zealand (2000 to 2019). We considered trials reported only as abstracts to be eligible if sufficient information was available from the report, or from contact with study authors, to fulfil the inclusion criteria (see Dealing with missing data for further details).
Data collection and analysis
We used the standard methods of Cochrane Neonatal.
Selection of studies
Two review authors (JB, BC, or VW) independently screened the titles and abstracts of all studies and assess the full articles of all potentially relevant trials. We excluded studies that did not meet all the inclusion criteria and stated the reasons for exclusion in the Characteristics of excluded studies table. We discussed any disagreements until consensus was achieved.
Data extraction and management
Two review authors (JB, BC, or VW) used a form to independently extract data on design, methodology, participants, interventions, outcomes, and treatment effects from each included study. We discussed any disagreements until we reached a consensus. The data extraction form was based on forms used previously by the author team, with adaptations made to meet the requirements of this review.
Assessment of risk of bias in included studies
Two review authors (JB, BC, or VW) independently assessed the risk of bias (low, high, or unclear) of all included trials using the Cochrane 'Risk of bias' tool (Higgins 2017) for the following domains:
-
sequence generation (selection bias);
-
allocation concealment (selection bias);
-
blinding of participants and personnel (performance bias);
-
blinding of outcome assessment (detection bias);
-
incomplete outcome data (attrition bias);
-
selective reporting (reporting bias);
-
any other bias.
We resolved any disagreements by discussion. We did not exclude trials on the basis of risk of bias, but planned to conduct sensitivity analyses, if applicable, to explore the consequences of synthesising evidence of variable quality.
See Appendix 2 for a more detailed description of risk of bias for each domain.
Measures of treatment effect
We analysed the treatment effects in the individual trials and reported the risk ratio (RR) and risk difference (RD) for dichotomous data and the mean difference (MD) for continuous data, with 95% confidence intervals (CI). We planned to report number needed to treat for an additional beneficial outcome (NNTB), or number needed to treat for an additional harmful outcome (NNTH) for analyses with a statistically significant difference in the RD.
Unit of analysis issues
The unit of analysis used was the participating infant in individually randomised trials and the neonatal unit (or subunit) for cluster‐randomised trials.
For cluster‐randomised trials, we planned to undertake analyses at the level of the individual while accounting for clustering in the data using the methods recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017).
Dealing with missing data
Where data were missing without explanation, and could not be derived, we used the following approaches.
-
Where we had concerns about the extent of missing data or potential bias in missing data, we contacted the original study investigators to request the missing data for primary outcomes only, for studies published within the past 10 years where an email address was easily available from the published papers.
-
Where possible, we planned to impute missing SD using the coefficient of variation, or calculate this from other available statistics including standard errors, CIs, t values, and P values.
-
If the data were assumed to be missing at random, we planned to analyse the data without imputing any missing values.
-
If data could not be assumed to be missing at random, we planned to impute the missing outcomes with replacement values, assuming all to have a poor outcome, and conduct sensitivity analyses to assess any changes in the direction or magnitude of effect resulting from data imputation.
Assessment of heterogeneity
Two review authors (JB, VW, or WM) assessed clinical heterogeneity, and conducted a meta‐analysis when both authors agreed that study participants, interventions, and outcomes were sufficiently similar.
We examined the treatment effects of individual trials and heterogeneity between trial results by inspecting the forest plots. We calculated the I² statistic for each analysis to quantify inconsistency across studies and described the percentage of variability in effect estimates that may have been due to heterogeneity rather than to sampling error. If high levels of heterogeneity were detected (an I² value greater than 75%), we explored the possible sources (e.g. differences in study design, participants, interventions, or completeness of outcome assessments).
Assessment of reporting biases
If there were more than 10 trials in a meta‐analysis, we planned to examine a funnel plot for asymmetry.
Data synthesis
We used the fixed‐effect model in Review Manager 5 for meta‐analyses (Review Manager 2014).
Subgroup analysis and investigation of heterogeneity
We planned the following subgroup analyses for the primary outcomes.
-
Trials in which infants received human milk only (maternal expressed milk or donor human milk) versus those where formula could be given instead of, or as a supplement to, human milk.
-
Very preterm infants (born at less than 32 weeks' gestation) versus infants born at 32 weeks' gestation or greater.
-
Very low birth weight infants (less than 1500 g) versus infants with a birth weight of 1500 g or greater.
-
Infants with a birth weight below the 10th percentile for the reference population ('small for gestation') versus infants with a birth weight at or above the 10th percentile ('appropriate for gestation').
-
Trials set in low‐ or middle‐income countries versus trials set in high‐income countries (for classification, see datahelpdesk.worldbank.org/knowledgebase/articles/906519#High_income (accessed 18 April 2019)).
Sensitivity analysis
We planned sensitivity analyses to determine if the findings were affected by including only studies of adequate methodology (low risk of bias), defined as those studies with adequate randomisation and allocation concealment, blinding of intervention and measurement, and less than 10% loss to follow‐up assessment of the trial primary outcome.
Summary of findings and assessment of the certainty of the evidence
We used the GRADE approach, as outlined in the GRADE Handbook (Schünemann 2013), to assess the certainty of evidence of the following primary outcomes: in hospital rate of weight gain (g/kg/day) until term equivalent; in hospital rate of head circumference growth (cm/week) until term equivalent; growth restriction (z score of weight) at hospital discharge; necrotising enterocolitis.
Two review authors (VW and WM) independently assessed the certainty of the evidence for each of the outcomes above. We considered evidence from RCTs as high certainty but downgraded the evidence one level for serious (or two levels for very serious) limitations based upon the following: design (risk of bias), consistency across studies, directness of the evidence, precision of estimates, and presence of publication bias. We used GRADEpro GDT to create summary of findings Table 1 to report the certainty of the evidence.
The GRADE approach results in an assessment of the certainty of a body of evidence as one of four grades.
-
High certainty: further research is very unlikely to change our confidence in the estimate of effect.
-
Moderate certainty: further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
-
Low certainty: further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
-
Very low certainty: we are very uncertain about the estimate.
Results
Description of studies
We included six trials (Bora 2017; Chetry 2014; Jajoo 2017; Nangia 2019; Ramya 2014; Sanghvi 2013). Two reports were available as conference abstracts only (Chetry 2014; Jajoo 2017). One was a published dissertation (Ramya 2014).
Results of the search
The search identified 7791 records through database searching and two additional records from other sources. After removing duplicates, 5144 records remained. We excluded 5132 irrelevant records and assessed the full text of 12 records. We included six RCTs (Bora 2017; Chetry 2014; Jajoo 2017; Nangia 2019; Ramya 2014; Sanghvi 2013; Characteristics of included studies table). We excluded studies with reasons (Genzel‐Boroviczény 2002; Higgs 1974; Jain 2016; Yu 1979; Characteristics of excluded studies table), and found two ongoing studies (ISRCTN89654042; NCT03708068; Characteristics of ongoing studies table). See Figure 1.
Study flow diagram.
Included studies
All included trials were undertaken during the 2010s by investigators in neonatal care centres in India (see Characteristics of included studies table).
Participants
In total, 526 infants participated (range per trial 60 to 180). Most were medically stable VLBW infants (birth weight 1000 g to 1500 g) of gestational age at birth 28 to 34 weeks. None was extremely preterm or ELBW. Participants included infants born 'small‐for‐gestation' in all trials except Chetry 2014, which excluded infants with birth weight less than 3rd percentile for reference population. All trials excluded infants with congenital anomalies or gastrointestinal or neurological problems from participation. Two trials excluded infants in receipt of respiratory support beyond supplemental oxygen (Jajoo 2017; Nangia 2019). One trial excluded infants who needed mechanical ventilation (Sanghvi 2013).
Interventions
The trials compared early full feeding (60 mL/kg to 80 mL/kg on day one after birth) with minimal enteral feeding (typically 20 mL/kg to 30 mL/kg on day one) supplemented with intravenous fluids (typically 10% dextrose). Feed volumes were advanced daily as tolerated by 20 mL/kg to 30 mL/kg body weight to a target steady‐state volume of 150 mL/kg/day to 180 mL/kg/day.
All participating infants were fed preferentially with maternal expressed breast milk. Two trials used donor breast milk to supplement insufficient volumes of maternal expressed breast milk (Bora 2017; Ramya 2014). The other trials used preterm formula if maternal milk was insufficient. In three trials, human milk was enriched with a multi‐nutrient fortifier when volume of intake exceeded 100 mL/kg/day (Bora 2017; Nangia 2019; Sanghvi 2013).
Outcomes
Outcomes reported most commonly were feed intolerance (reported in various ways but often without accompanying numerical data), days to regain birth weight, duration of birth hospitalisation, necrotising enterocolitis, late‐onset infection, and mortality. None of the trials assessed any outcomes posthospitalisation, including growth and neurodevelopment.
Excluded studies
We excluded four studies after full‐text assessment (Genzel‐Boroviczény 2002; Higgs 1974; Jain 2016; Yu 1979; see Characteristics of excluded studies table).
Ongoing studies
We found two ongoing trials (ISRCTN89654042; NCT03708068; see Characteristics of ongoing studies table).
Risk of bias in included studies
Quality assessments are detailed in the Characteristics of included studies table and summarised in Figure 2.
Risk of bias summary.
Allocation
All trials were at low risk of selection bias as they reported adequate allocation concealment methods (sealed, numbered envelopes).
Blinding
None of the trials masked parents, carers, or clinicians or investigators assessing the trial outcomes (not reported in Sanghvi 2013).
Incomplete outcome data
All trials reported complete outcome assessments of the trial cohorts and were at low risk of attrition bias.
Selective reporting
We were unable to assess reliably whether selective reporting occurred in four of the trials as we identified no trial registrations, protocols, or other indicators of prespecified outcomes; these were at unclear risk of reporting bias (Bora 2017; Chetry 2014; Jajoo 2017; Ramya 2014). The other two trials listed outcomes of interest in their registrations (Nangia 2019; Sanghvi 2013). Both trials reported all prespecified outcomes, and additional, non‐prespecified data as secondary outcomes and were at low risk of reporting bias.
Other potential sources of bias
One trial received no funding and was at low risk of other bias (Sanghvi 2013). The others did not report whether they received trial funding and were at unclear risk of other bias.
Effects of interventions
Primary outcomes
In hospital rate of weight gain
Two studies reported in hospital rate of weight gain (Chetry 2014 (author correspondence); Nangia 2019).
Chetry 2014 reported a slower rate of weight gain in infants in the early full enteral feeding group (median 15.22 g/kg/day, interquartile range (IQR) 10.9 to 24.6) compared to the incremental feeding group (median 18.23 g/kg/day, IQR 11.36 to 27.3).
Nangia 2019 reported a faster rate of weight gain in the early full enteral feeding group compared with the control group (mean: 6.3 g/kg/day with early full enteral feeding compared to 5.06 g/kg/day with control; SDs not provided).
We did not meta‐analyse these data. We assessed the certainty of the evidence as very low, downgrading for risk of bias, inconsistency, and imprecision.
In hospital rate of head circumference growth
None of the trials reported in hospital rate of head circumference growth.
Growth restriction
One study reported z‐score for weight at hospital discharge (Sanghvi 2013).
The definition of growth restriction in the included trial (z‐score at discharge) differed from our prespecified definition (z‐score at term). We made a consensus post hoc decision to include the data based on the investigators' definition.
Analysis showed that infants in the early full enteral feeding group had higher z‐scores that those in the control group (MD 0.24, 95% CI 0.06 to 0.42).
We assessed the certainty of the evidence as low, downgrading for risk of bias and imprecision.
Necrotising enterocolitis
Six studies reported necrotising enterocolitis (Bora 2017; Chetry 2014; Jajoo 2017; Nangia 2019; Ramya 2014; Sanghvi 2013). Two trials (which were reported in conference abstracts) did not specify diagnostic criteria for necrotising enterocolitis (Chetry 2014; Jajoo 2017).
There was no evidence of a difference (RR 0.98, 95% CI 0.38 to 2.54; I² = 51%; RD –0.00, 95% CI –0.03 to 0.03; Figure 3).
Forest plot of comparison: 1 Early full versus delayed/progressive enteral nutrition, outcome: 1.2 Necrotising enterocolitis.
We assessed the certainty of the evidence as very low, downgrading for risk of bias, inconsistency (moderate heterogeneity), and imprecision.
Secondary outcomes
Feed intolerance
Four studies reported feed intolerance (Bora 2017; Chetry 2014; Nangia 2019; Sanghvi 2013). The definitions of feed intolerance used in the included trials did not specify the duration of enteral feed interruption. We made a consensus post hoc decision to include the data based on the investigators' definitions (see Characteristics of included studies table).
There was no evidence of a difference (RR 0.74, 95% CI 0.49 to 1.13; I² = 54%; RD –0.06, 95% CI –0.13 to 0.02; Figure 4).
Forest plot of comparison: 1 Early full versus delayed/progressive enteral nutrition, outcome: 1.3 Feed intolerance.
Episodes of hypoglycaemia
Two studies reported hypoglycaemia (Nangia 2019; Sanghvi 2013).
There was no evidence of a difference (RR 3.00, 95% CI 0.13 to 70.02; I² not estimable; RD 0.01, 95% CI –0.03 to 0.06; Figure 5).
Forest plot of comparison: 1 Early full versus delayed/progressive enteral nutrition, outcome: 1.4 Episodes of hypoglycaemia.
Invasive infection
Four studies reported invasive infection (Chetry 2014 (author correspondence); Nangia 2019; Ramya 2014; Sanghvi 2013).
There was no evidence of a difference (RR 0.72, 95% CI 0.36 to 1.46; I² = 45%; RD –0.03, 95% CI –0.08 to 0.03; Figure 6).
Forest plot of comparison: 1 Early full versus delayed/progressive enteral nutrition, outcome: 1.5 Invasive infection.
Duration of birth hospitalisation
Five studies reported days of hospitalisation (Bora 2017; Chetry 2014; Jajoo 2017; Nangia 2019; Sanghvi 2013).
Meta‐analysis showed that the early full enteral feeding group had a shorter duration of hospitalisation than infants in the control group (MD –3.07 days, 95% CI –4.13 to –2.02; I² = 90%).
All‐cause mortality
Six studies reported mortality (Bora 2017 (author correspondence); Chetry 2014 (author correspondence); Jajoo 2017; Nangia 2019; Ramya 2014; Sanghvi 2013).
There was no evidence of a difference (RR 0.78, 95% CI 0.36 to 1.70; I² = 0%; RD –0.01, 95% CI –0.05 to 0.03).
Long‐term growth
None of the studies reported long‐term growth.
Severe neurodevelopmental disability
None of the studies reported severe neurodevelopmental disability.
Additional post hoc outcome
Days to regain birth weight
Six studies reported days to regain birth weight (Bora 2017; Chetry 2014; Jajoo 2017; Nangia 2019; Ramya 2014; Sanghvi 2013).
Meta‐analysis showed that the early full enteral feeds group regained birth weight earlier than the control group (MD –3.42 days, 95% CI –4.31 to –2.53; I² = 89%).
Subgroup analyses
-
Trials in which infants received human milk only versus those where formula could be given instead of, or as a supplement to, human milk. The test for subgroup differences was statistically significant for:
-
feed intolerance (Chi² = 6.49, df = 1 (P = 0.01), I² = 84.6%); Figure 4):
-
exclusive human milk: RR 2.04, 95% CI 0.83 to 5.02; I² = not applicable; RD –0.12, 95% CI –0.03 to 0.27;
-
supplemental formula: RR 0.54, 95% CI 0.33 to 0.88; I² = 0%; RD –0.12, 95% CI –0.21 to ‐0.03
-
-
duration of hospitalisation (Chi² = 22.73, df = 1 (P < 0.00001), I² = 95.6%;):
-
exclusive human milk: MD –1.40 days, 95% CI –2.66 to –1.04 days; I² = not applicable;
-
supplemental formula: MD –7.03 days, 95% CI –8.96 to –5.09 days; I² = 82%;
-
-
days to regain birth weight (Chi² = 4.25, df = 1 (P = 0.04), I² = 76.5%):
-
exclusive human milk: MD –1.97, 95% CI –3.67 to –0.26 days; I² = 0%;
-
Supplemental formula: MD –4.02, 95% CI –4.98 to –3.07 days; I² = 91%.
-
-
-
Very preterm infants (born at less than 32 weeks' gestation) versus infants born at 32 weeks' gestation or greater.
-
Subgroup data not available (most trial participants were very preterm infant).
-
-
Very low birth weight infants (less than 1500 g) versus infants with a birth weight of 1500 g or greater.
-
All trial participants were very low birth weight infants (1000 g to 1500 g).
-
-
Infants with a birth weight below the 10th percentile for the reference population ('small for gestation') versus infants with a birth weight at or above the 10th percentile ('appropriate for gestation').
-
Subgroup data for infants born 'small for gestation' not available.
-
-
Trials set in low‐ or middle‐income countries versus trials set in high‐income countries.
-
All the trials were conducted in a middle‐income country (India).
-
Sensitivity analysis for risk of bias
We were unable to do planned sensitivity analyses to determine if the findings were affected by bias as the trials had a similar risk of bias with low risk for randomisation and allocation concealment, complete or near‐complete follow‐up, but high risk from inadequate masking of intervention and measurement.
Discussion
Summary of main results
Data from six trials, in which 526 VLBW infants participated, provided limited evidence about the effect of early full enteral feeding on growth and development. There are data from two trials on in hospital rates of weight gain, and these have inconsistent results (not suitable to be combined in meta‐analysis). One trial that assessed infant weight at the time of hospital discharge showed that early enteral feeding slightly increased the mean z‐score compared with delayed or progressive advancement of enteral feeds. There are no data on head growth or length at hospital discharge, or for growth parameters following discharge and developmental outcomes assessed beyond infancy.
The growth parameter most commonly reported was the time from birth to regain birth weight. We did not prespecify this outcome, but a post hoc meta‐analysis of data from six trials showed that infants who had early full enteral feeding regained birth weight about three days earlier than infants who had delayed introduction and progressive advancement of feed volumes. Given the paucity of data for other growth parameters, however, the importance of this effect is unclear.
Meta‐analysis of data from six trials found no effect of early full enteral feeding on risk of necrotising enterocolitis. The bounds of the 95% CI for the RD are consistent with either three extra or three fewer cases in every 100 infants who receive early full enteral feeding. We found no evidence of other possible benefits or harms, including insufficient data to determine whether early full enteral feeding is associated with the risk of feed intolerance, invasive infection, duration of birth hospitalisation, mortality, or episodes of hypoglycaemia during the early neonatal period.
Overall completeness and applicability of evidence
Most participants included in the six trials in this review were stable very preterm infants of birth weight between 1000 g and 1500 g. One trial excluded infants who were born 'small for gestation' from participation (Chetry 2014). Two trials excluded infants with evidence of intrauterine growth compromise indicated by the antenatal detection of absent or reversed end‐diastolic flow velocity in the umbilical vessels (Nangia 2019; Ramya 2014). The other trials did not report subgroup data for infants who were 'small for gestation' or growth restricted at birth. It is unclear, therefore, whether the review findings are applicable to infants who have both a high level of nutrient requirement and an elevated risk of developing necrotising enterocolitis (Dorling 2005; Samuels 2017).
In two trials, infants received only human milk (expressed own mother's milk or donor milk) (Bora 2017; Ramya 2014). The other trials used maternal milk preferentially, but used cow's milk formula to supplement any deficits (Chetry 2014; Jajoo 2017; Nangia 2019; Sanghvi 2013). The risk‐benefit balance of early enteral feeding strategies may differ between human milk‐fed and formula‐fed very preterm or VLBW infants (Quigley 2019). Prespecified subgroup analyses, however, found no differences on the risk of necrotising enterocolitis or death between infants who received human milk only versus those where formula could be given instead of, or as a supplement to, human milk.
All the trials included in this review were conducted in neonatal care facilities in India in the 2010s. The review findings may be applicable particularly to neonatal care facilities where staff and equipment costs limit the availability of parenteral fluids or nutrition, as well as to facilities without such resource constraints where early full enteral feeding could complement other infant and family centred care practices (Patel 2018).
Quality of the evidence
The certainty of evidence for the primary outcomes, assessed using GRADE methods, was low or very low, downgraded because of risk of bias in the included trials (lack of masking of investigators, parents, and carers), imprecision of estimates of effect, and inconsistency (heterogeneity) in meta‐analyses (summary of findings Table 1).
Risk of bias (lack of masking)
In all the included trials, staff and parents knew to which group the participating infant had been allocated (because it is difficult to conceal intravenous access devices and fluids). Lack of masking may have resulted in surveillance and detection biases. It is unclear, however, to what extent lack of masking might have biased effect estimates. It is plausible that clinicians who harboured concerns that early full enteral feeding could contribute to feed intolerance or necrotising enterocolitis, for example, might have undertaken more assessments or tests, or used a lower threshold for subjective criteria, to diagnose these conditions.
Inconsistency (heterogeneity) in meta‐analyses
The meta‐analyses in this review showed moderate to high levels of statistical heterogeneity that were not explained by major differences in trial design or conduct. All the trials were undertaken in secondary care facilities in India during the 2010s. Participants were similar (stable very preterm infants of birth weight between 1000 g and 1500 g). Trials used slightly different rates of advancement of feed volumes, but these were consistent with current practice (20 mL/kg to 30 mL/kg). Trials differed by the modality of intravenous fluids in the control group, with two trials providing parenteral nutrition and four providing 10% dextrose. Given the duration of use of intravenous fluid (up to one week), this may have affected growth outcomes.
The assessed outcomes were similar across the trials. For some outcomes, such as feed intolerance and necrotising enterocolitis, diagnosed using subjective criteria, variation in diagnostic thresholds might be a source of heterogeneity. As these criteria were not explicitly described in all the published reports, we were unable to explore the extent to which variation in outcome definition contributed to inconsistency in estimates of effect.
In prespecified subgroup comparisons of trials in which infants received human milk only versus those where formula could be given as a supplement to human milk, there was no evidence of a differential effect on the risk of necrotising enterocolitis. Small subgroup differences exist for the risk of feed intolerance and time taken to regain birth weight. The subgroup difference in the estimates of effect on the duration of hospitalisation favoured trials in which formula was used as a supplement when maternal milk was insufficient. This might be due to accelerated weight gain in infants receiving preterm formula with higher nutrient density than human milk. Since an attained body weight (ranging from 1300 g to 1600 g across trials) was a key discharge criterion for participants, infants reaching those targets earlier will have shorted durations of hospitalisation.
Imprecision
The estimates of effect were imprecise as indicated by the broad 95% CIs with upper and lower bounds encompassing substantial harm or benefit. This reflects the sample size of the studied population (526 infants in total), the rarity of the key outcomes such as necrotising enterocolitis (3% overall rate in the control groups), and the paucity of outcome data, especially of growth parameters, available for synthesis due to non‐reporting. Ongoing trials of this intervention should recruit sufficient participants to provide data to optimise the information size in future meta‐analyses.
Potential biases in the review process
The main concern is the possibility that findings of this review are subject to publication and other reporting biases. We attempted to minimise this threat by screening the reference lists of included trials and related reviews and searching the proceedings of major international perinatal conferences to identify trial reports that are not (yet) published in full form in academic journals. There were insufficient trials to examine for symmetry of funnel plots as a means of identifying possible publication or small‐study bias.
Agreements and disagreements with other studies or reviews
This review focused on the comparison of early full enteral feeding versus delayed introduction or progressive feed volume advancement. Other Cochrane Reviews have appraised and synthesised the trial evidence for the comparison of different rates of volume advancement and different timing of introduction of progressive enteral feeds. Consistent with the findings of this review, these show that conservative feeding strategies (delaying introduction of enteral feeds until beyond four days after birth, or advancing feed volumes more slowly than by 24 mL/kg daily) do not affect growth parameters or the risk of feed intolerance or necrotising enterocolitis in very preterm infants (Morgan 2014; Oddie 2017).
Study flow diagram.
Risk of bias summary.
Forest plot of comparison: 1 Early full versus delayed/progressive enteral nutrition, outcome: 1.2 Necrotising enterocolitis.
Forest plot of comparison: 1 Early full versus delayed/progressive enteral nutrition, outcome: 1.3 Feed intolerance.
Forest plot of comparison: 1 Early full versus delayed/progressive enteral nutrition, outcome: 1.4 Episodes of hypoglycaemia.
Forest plot of comparison: 1 Early full versus delayed/progressive enteral nutrition, outcome: 1.5 Invasive infection.
Comparison 1: Early full versus delayed/progressive enteral nutrition, Outcome 1: z‐score for weight at hospital discharge
Comparison 1: Early full versus delayed/progressive enteral nutrition, Outcome 2: Necrotising enterocolitis
Comparison 1: Early full versus delayed/progressive enteral nutrition, Outcome 3: Feed intolerance
Comparison 1: Early full versus delayed/progressive enteral nutrition, Outcome 4: Episodes of hypoglycaemia
Comparison 1: Early full versus delayed/progressive enteral nutrition, Outcome 5: Invasive infection
Comparison 1: Early full versus delayed/progressive enteral nutrition, Outcome 6: Duration of birth hospitalisation (days)
Comparison 1: Early full versus delayed/progressive enteral nutrition, Outcome 7: All‐cause mortality
Comparison 1: Early full versus delayed/progressive enteral nutrition, Outcome 8: Days to regain birth weight
| Early full compared to delayed/progressive enteral feeding for preterm or low birth weight infants | |||||
| Patient or population: preterm or low birth weight infants | |||||
| Outcomes | № of participants | Certainty of the evidence | Relative effect | Anticipated absolute effects* (95% CI) | |
|---|---|---|---|---|---|
| Risk with delayed/progressive enteral feeding | Risk difference with early full enteral feeding | ||||
| In hospital rate of weight gain (g/kg/day) until term equivalent | 236 | ⊕⊝⊝⊝ | — | Meta‐analysis not possible due to different outcome measures. | |
| In hospital rate of head circumference growth (cm/week) until term equivalent – not reported | — | — | — | — | — |
| Growth restriction (z‐score of weight) at hospital discharge | 46 | ⊕⊕⊝⊝ | — | The mean growth restriction (z‐score of weight) at hospital discharge was –1.09 | Mean 0.24 higher |
| Necrotising enterocolitis | 522 | ⊕⊝⊝⊝ | RR 0.98 | Study population | |
| 31 per 1000 | 1 fewer per 1000 | ||||
| *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). CI: confidence interval; RCT: randomised controlled trial; RR: risk ratio. | |||||
| GRADE Working Group grades of evidence | |||||
| aDowngraded one level due to risk of surveillance and detection bias owing to lack of masking. | |||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1.1 z‐score for weight at hospital discharge Show forest plot | 1 | 46 | Mean Difference (IV, Fixed, 95% CI) | 0.24 [0.06, 0.42] |
| 1.2 Necrotising enterocolitis Show forest plot | 6 | 522 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.98 [0.38, 2.54] |
| 1.2.1 Exclusive human milk | 2 | 172 | Risk Ratio (M‐H, Fixed, 95% CI) | 4.08 [0.47, 35.26] |
| 1.2.2 Supplemental formula | 4 | 350 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.55 [0.16, 1.82] |
| 1.3 Feed intolerance Show forest plot | 4 | 393 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.74 [0.49, 1.13] |
| 1.3.1 Exclusive human milk | 1 | 103 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.04 [0.83, 5.02] |
| 1.3.2 Supplemental formula | 3 | 290 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.54 [0.33, 0.88] |
| 1.4 Episodes of hypoglycaemia Show forest plot | 2 | 149 | Risk Ratio (M‐H, Fixed, 95% CI) | 3.00 [0.13, 70.02] |
| 1.5 Invasive infection Show forest plot | 4 | 359 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.72 [0.36, 1.46] |
| 1.5.1 Exclusive human milk | 1 | 69 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.06 [0.56, 7.58] |
| 1.5.2 Supplemental formula | 3 | 290 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.44 [0.18, 1.09] |
| 1.6 Duration of birth hospitalisation (days) Show forest plot | 5 | 436 | Mean Difference (IV, Fixed, 95% CI) | ‐3.07 [‐4.13, ‐2.02] |
| 1.6.1 Exclusive human milk | 1 | 103 | Mean Difference (IV, Fixed, 95% CI) | ‐1.40 [‐2.66, ‐0.14] |
| 1.6.2 Supplemental formula | 4 | 333 | Mean Difference (IV, Fixed, 95% CI) | ‐7.03 [‐8.96, ‐5.09] |
| 1.7 All‐cause mortality Show forest plot | 6 | 522 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.78 [0.36, 1.70] |
| 1.7.1 Exclusive human milk | 2 | 172 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.03 [0.29, 3.59] |
| 1.7.2 Supplemental formula | 4 | 350 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.66 [0.24, 1.80] |
| 1.8 Days to regain birth weight Show forest plot | 6 | 505 | Mean Difference (IV, Fixed, 95% CI) | ‐3.53 [‐4.37, ‐2.70] |
| 1.8.1 Exclusive human milk | 2 | 168 | Mean Difference (IV, Fixed, 95% CI) | ‐1.97 [‐3.67, ‐0.26] |
| 1.8.2 Supplemental formula | 4 | 337 | Mean Difference (IV, Fixed, 95% CI) | ‐4.02 [‐4.98, ‐3.07] |
