Starting and Increasing Feeds, Milk Tolerance and Monitoring of Gut Health in Significantly Preterm Infants

Abstract

Approaches to enteral feeding significantly preterm infants' impact short-term outcomes including survival, late-onset sepsis (LOS), and necrotizing enterocolitis (NEC), and neurode- velopmental and later health outcomes. Clinical practice and trial data are dominated by short-term outcomes (NEC and LOS) with limited longer-term outcomes. Strategies maximizing early maternal breast milk (MOM) exposure and duration of MOM use are key given global health benefits of MOM, but few feeding trials use these as outcomes. Current data support colostrum receipt, early introduction, and progression of volumes between 18 and 30 mL/kg/day, without adverse impact on NEC, LOS, or mortality. Little evidence supports choosing between route of gastric tube placement, bolus, or continuous feed delivery. Individual infants may have specific features that require individualized feed management, such as combinations of growth restriction, antenatal blood flow concerns, intensive supportive needs (including inotropes), and large open patent ductus arteriosus, currently poorly represented in feeding trials. Infant tolerance monitoring includes clinical observations (stool- ing, abdominal size, vomiting) but routine gastric aspiration appears unhelpful. Infants should be monitored biochemically, anthropometrically, and in the future through bedside microbiomics or metabolomics. Units and networks should audit and compare their rates of mortality, NEC, LOS, neurodevelopment, and growth achieved.  

Enteral nutrition is required for growth and development of a significantly preterm infant (born at <32 weeks' gestation) if they are to be independent of parenteral nutrition (PN) and associated risks and receive maternal breast milk (MOM), with the associated benefits to immune and gastrointestinal (GI) development and function, and reduction in associated risks of necrotizing enterocolitis (NEC), late-onset sepsis (LOS), and neonatal mortality [1]. How and when enteral feeding is established influences these early preterm health complications, and minimizing complications alongside achieving optimal growth and body composition are key goals of preterm neonatal feeding strategies. However, standardizing definitions of NEC and LOS across studies is challenging and agreeing optimal growth trajectories and body composition even more so. Additionally, given the crucial developmental stage of the preterm brain developing ex-utero through the period of what should have been the third trimester, feeding practices directly (via constituents) and indirectly (via the gut-brain axis) [2] impact on neurodevelopmental outcome after early birth. Through metabolic [3] and microbiomic mechanisms, later health outcomes are also influenced [4]. Enteral feeding approaches are therefore a crucial element of neonatal care. Data from infants born <32 weeks' gestation from 13 neonatal intensive care units (NICUs) on 5 continents identified surprisingly large variations in median times to full enteral feeds (range 8-33 days), illustrating wide variations in feeding practices, suggesting cultural as well as scientific influence on feeding practices. Wide variations in clinical outcomes including weight gain achieved (range 5-14.6 g/kg/day), changes in weight z-scores (-0.54 to 1.64), NEC incidence (1-13%) and mortality (1-18%) were shown, but no clear associations between these clinical outcomes and time to full feeds (TFF) [5].

Research Base

Feeding practices in preterm infants are extensively studied: the Cochrane database currently identifies >1,900 trials and 300 reviews on preterm feeding. Interventions are often presented as a simple comparison of one practice against another, but in reality feeding practices are often complex interventions. For example, studies that include an intervention of early milk delivery, or rapid progression will also potentially unwittingly include a decision around whether to wait for sufficient maternal milk (MOM) to implement this strategy, or using an alternative to MOM to allow earlier implementation, and if so, which alternative should be used. Feeding trials are also almost always non-blinded, leading

to potential bias, and definitions of key outcomes hard to standardize, and not always allied to parental and clinician priorities [6]. Multicenter neonatal feeding trials often leave these options to attending clinicians, allowing recruitment in units with differing approaches, and broad applicability of the results. Single center studies are often limited to the one approach in use in that unit, making extrapolation of study findings difficult. This chapter focuses on the practical aspects of enteral feeding from birth to the point of established full feeds (TFF) and does not specifically address the issues of what to do if insufficient MOM is available. Table 1 summarizes the key feeding issues, and the current evidence base on which recommendations are made.



Starting Enteral Feeds

Exposure to Buccal Colostrum

Exposure to buccal colostrum sets the scene for receipt of MOM, allowing early exposure of the infant to small volumes of MOM in the first day(s) of lactation without waste. Several small studies have examined the impact of receipt of colostrum, then subject to meta-analysis [7]. Meta-analysis showed no impact on rates of NEC, LOS, or mortality. Individually some of these studies were very small (n = 12-200) and some focused on microbiomic [8] or immunological outcomes [9]. Potentially beneficial changes in urinary lactoferrin, sIgA, and inflammatory responses have been shown [9, 10], but to date without associated demonstrable improvement in short-term clinical measures. Colostrum receipt has been shown to be associated with increased receipt of MOM in infants of median gestational age 28 weeks, at both 6 weeks of age (67 vs. 55%, p = 0.03) and discharge (53 vs. 32%, p = 0.07), in a retrospective case-controlled study after implementation of an oral colostrum protocol [11].

Minimal Enteral Nutrition versus Fasting

Minimal enteral nutrition (MEN) refers to the practice of giving deliberately small volumes enterally for a period of time after birth - also sometimes called trophic feeding. Whether preterm infants should deliberately spend time receiving only MEN before increasing feeds, compared to remaining nil by mouth, has been studied in randomized controlled trials (RCTs) and subject to meta-anal- ysis. Comparison of the incidence of key short-term outcomes discussed above did not identify additional risk from MEN with odds ratios for NEC of 1.07 (95% CI 0.67-1.7), mortality of 0.66 (95% CI 0.41-1.07), and LOS of 1.06 (95% CI 0.72-1.56) [12]. There were 9 trials identified, with the most recent included in the meta-analysis (MA) published in 2008 [13], with 754 included infants. In the included trials, infants deliberately spent at least a week receiving <24 mL/kg/ day after introduction of milk in comparison to a comparable fasting period. The number of infants who were <28 weeks gestation or <1,000 g was small, and the applicability to the most immature of today's infants is uncertain.

Delaying Introduction

The ADEPT study (Abnormal Dopplers Enteral Prescription Trial) [14] explored delaying introduction of first milk until day 6 in comparison to starting on day 2 but did not offer MEN. This group of <35 weeks of gestation infants was high risk by virtue of having both abnormal antenatal Dopplers (absent or reversed end diastolic flow, AREDF) and birth weight <10th centile. Infants who commenced enteral feeds on day 2 achieved full feeds at median age 18 days  (IQR 15-24) in comparison with those waiting until day 6 who were fully fed at median age 21 days (IQR 19-27) with no impact on NEC or LOS demonstrated. This lack of impact on NEC was confirmed in meta-analysis of 8 trials totaling 1,092 infants [15]. There are however data showing clinical benefit of MOM in a dose-dependent manner even in the first 10 days of life, with more MOM positively impacting on survival free of NEC and LOS. This Dutch study took place in a unit without donor human milk using hydrolyzed preterm formula for shortfall in MOM [16]. It has also been shown that the number of days in the first month of life with the proportion of enteral milk that is MOM >50% positively impacts on deep nuclear gray matter volume on MRI (0.15 cm3/day, 95% CI 0.05-0.25), IQ (0.5 points, 95% CI 0.2-0.8), and motor skills at 7 years of age in infants <30 weeks gestation [17]. Delaying the introduction of enteral feeds with MOM cannot be recommended, but the balance of risks and benefits of feeding strategies in the absence of MOM may differ.

Increasing Enteral Feeds

Rates of Progression

Once first milk has been tolerated, the next clinical challenge is how quickly to attempt to advance milk. Given current understanding of etiology/mechanisms of LOS and NEC, these might function as competing outcomes for slower or faster feed increases: intent to increase faster may result in more NEC, intent to increase slower may result in more LOS from longer indwelling catheter duration. The largest single trial to address this is the UK Speed of Increasing Feeds Trial (SIFT), recruiting 2,804 infants <32 weeks gestation or 1,500 g birth weight [18]. This RCT compared intention to feed at increases of18 mL/kg/day with intention to feed at increases of 30 mL/kg/day. Importantly, infants were randomized once the clinician was comfortable increasing feeds and median age at increases commencing was 4 days (IQR 3-6). The primary outcome was survival without moderate or severe neurological impairment at 2 years of corrected gestational age, but predischarge morbidities including NEC and LOS were secondary outcomes, alongside a preplanned health economic evaluation. There were no statistically significant differences in predischarge morbidities or mortality. NEC occurred in 5% of the faster arm and 5.6% of the slower arm, culture positive LOS in 19.9% of the faster arm and 17.9% of the slower arm, and all LOS (including culture negative) in 30.1% of the faster arm and 27.5% of the slower arm. There was a difference of 2 days' duration of PN: median 7 days (IQR 7-10) for the faster arm, and 10 days (IQR 9-13) for the slower arm. This difference was statistically significant with an effect measure of 1.7 (95% CI 1.52-1.89) favoring faster feeds. At 2 years' corrected gestational age, the primary was not statistically different: 65.5% (faster arm) and 68.1% (slower arm), effect measure 0.96 (95% CI 0.91-1.02). Subgroup analyses confirmed no impact of gestation on these findings, but motor impairment alone was different across the two arms, with more affected infants in the faster arm (7.5 vs. 5%), statistically significant at the 99% confidence level with a risk ratio of 1.48 (99% CI 1.02-2.14). There was no impact on the diagnosis of cerebral palsy, cognitive impairment, visual impairment, or hearing impairment. The possible mechanism behind motor impairment differences is unclear, and this effect may still be by chance. Overall, this very large study suggests that pragmatically approaching intent to increase “slower” or “faster,” within the range 18-30 mL/ kg/day does not impact on clinical outcome measures for most infants, but faster intent may shorten duration of PN. The accompanying health economic analysis suggested that faster feeding was more expensive by an average of GBP 267 per baby [19], resulting from the higher rates of adverse outcome even in the absence of statistical significance, arising from the stochastic approach to economic analysis as used by health decision maker groups such as the National Institute for Health and Clinical Excellence [20].

The 2017 Cochrane meta-analysis of slow versus fast feed increments [21] identified 9 studies of 3,576 infants (2,804 from SIFT). No difference was seen in mortality (OR 1.15 [95% CI 0.93-1.42]), NEC (OR 1.07 [95% CI 0.83-1.39]), or LOS (OR 1.15 [95% CI 1.0-1.32]). Clinical decision makers may feel reassured to feed at increments of 30 mL/kg/day in a baby tolerating this well, but also assured that for a less tolerant baby advancing more slowly does not cause obvious harm.

Route of Gastric Tube Placement

Feeding tubes may be placed into the stomach (gastric G) or positioned across the pylorus (jejunal J) and sited nasally (N) or orally (O), meaning there are at least four possible placements of feeding tubes in preterm infants. Data are limited on route of placement (<200 infants studied) but suggest no clear differences in growth or adverse events for NG or OG placement and that gastric feeding is preferable [22], and more physiological in terms of GI tract hormone promotion, gastric emptying, and achieving normal GI motility [23].
Intermittent or Continuous Feeding Strategies
Feed can be delivered by small boluses to mimic normal feeding patterns, or continuously. Studies show no differences in time to achieve full feeds, growth, or NEC rates [24], but only included small numbers of infants weighing <1,000 g. Studies undertaken in the subsequent 10 years have not changed current understanding of risks and benefits of intermittent or continuous strategies.

Milk Tolerance

When to Slow Down, Pause or Stop Enteral Feeds

Individual infants may have specific clinical factors that clinicians feel add uncertainty about whether to attempt to increase feed as for infants without such factors, or whether there should be deliberate intent to advance more slowly. These include a significant patent ductus arteriosus (PDA), inotrope receipt, blood transfusion, being small for gestational age (SGA), or the presence of abnormal dopplers (AREDF) antenatally.

Infant Factors

The ADEPT trial addressed AREDF and SGA from a “when to start” perspective, reassuring clinicians that delaying feeding was not necessary [14]. SIFT, comparing rates of increasing, recruited all infants regardless of growth restriction (18% of recruits <19th centile) or AREDF (15% recruits) [18]. Primary outcome subgroup analysis of the growth-restricted infants still showed no impact of speed of increment (OR 1.04 [95% CI 0.92-1.17]), suggesting that a similar approach in these infants as for other infants is sensible, although subgroup analyses lack power compared to whole study analysis. Meta-analysis of this subgroup of infants with either growth restriction, AREDF, or both was also undertaken [15] confirming no change in NEC risks with early or late introduction of feeds (OR for NEC 0.87 [95% CI 0.54-1.41]). There are less clear data for other infant factors: the number of infants needed for clinical end points (NEC or LOS) is often prohibitively large when subgroups of infants are considered. Some specific issues are considered below.

Blood Transfusion

Pausing feeding during packed red cell transfusion is advocated by some, given the reported association between transfusion and NEC and the observations of altered mesenteric flow during transfusion [25]. A pilot trial demonstrated feasibility of recruitment of infants to an electronic care record-based trial [26] and further data are awaited from ongoing trials.

Patent Ductus Arteriosus

Like PDA management itself, optimal feeding strategies for infants with a hemo- dynamically significant PDA are controversial, and opinions vary. Data on which to base decisions are limited: definitions of “significant” PDA are variable, as are definitions of feed tolerance, and full feeds. Current literature is reviewed in Martini et al. [27] but concludes with a caution around interpreting the current evidence as endorsing any one approach over another, rather encouraging an individualized hemodynamic based approach and calling for higher quality larger RCTs.

Feed Intolerance

Clinical thresholds for labelling an infant “‘feed intolerant” are highly variable, often unit specific, and embedded in cultural practices, as evidenced by the large variation in TTF across the world [5]. Units may use a combination of stooling patterns, abdominal examination, abdominal girth, and gastric aspirates/residu- als to label an infant “feed intolerant.” Definitions of these issues are variable and non-standardized, making conclusions very difficult to draw.

Gastric Aspirate/Residuals

Gastric contents consist of milk (partially digested or undigested) and gastric secretions consisting of hormones, acid, and enzymes which aid digestion, promote GI motility, and intestinal maturation. Whilst discarding residual milk may not be harmful as long as volumes are replaced, discarding the associated components may be physiologically harmful [23], delay achievement of full enteral feeds, and reduce growth. Nonetheless, it is often practiced routinely in NICUs and difficult to stop. Studies assessing the impact of routine gastric aspiration found no differences in rates of NEC, but delayed time to regain birth weight was seen in one study of 80 infants with gastric aspiration in comparison to abdominal circumference measurement [28]. Larger studies are needed to definitively address this, with clinically relevant end points important to parents and infants.

Where gastric aspiration is undertaken, a further question is raised - what to do with the residual. A single trial has compared refeeding the residual to discarding and refeeding with fresh milk in infants between 23 and 28 weeks' gestation being fed by bolus feeds every 3 h [29], suggesting this appeared safe, but only 72 infants participated.

Monitoring

Unit Practices

Given the importance of the approach to early initiation and establishment of full enteral feeds, individual units are likely to benefit from agreed guidance to help staff share a similar approach. In addition, standardized feeding guidance has been shown to reduce the rates of NEC. As such, units would benefit from monitoring through audit their own adherence to their feeding guidance, their actual TTF, and their use of PN.

Clinical Outcomes

Feed-associated outcomes are also important for individual units to understand their own data for, including rates of NEC, LOS, mortality, and 2 years' neuro- developmental outcomes. These offer an ability to monitor outcomes over time and compare outcomes to similar units through systems like the Vermont Oxford Network (VON) or the UK National Neonatal Audit Programme (NNAP).

Individual Infant Monitoring

For individual infants, monitoring of the current feeding strategy is undertaken by a combination of biochemical and anthropometric measures. The importance of calcium and phosphorous in early feeding, the requirement for most preterm infants to be supplemented with sodium, vitamins, and minerals to adequately accrete tissue, and the immaturity of the preterm kidney makes regular biochemical monitoring essential. This should be supplemented with anthropometric monitoring to give a picture of body growth that is being achieved, al-though standards for this are subject to much discussion.

Mechanistic Monitoring

Given that the mechanisms linking feed approaches to outcomes such as NEC, LOS, and neurodevelopmental delay are incompletely understood, it is difficult to identify the best monitoring strategies for these elements. However, in the future it is possible that bedside microbial, metabolic, or inflammatory measures may be both much more routinely available, and more clearly mechanistically linked to outcomes, such that real time adjustments could be made. This may also apply to measures around breast milk content, such as routine monitoring of human milk oligosaccharide (HMO) content, IgA content, and potential modification of these elements, mechanistically linked to NEC [30]. Table 2 summarizes suggested monitoring strategies.



Summary and Suggestions for Future Research

This article summarizes the key practical steps in starting and increasing milk feeds, and a suggested approach to monitoring. As can be seen for some decisions, there are data on large numbers of infants supporting practice, but for others data are lacking or from small numbers of infants only. Moderate certainty exists around when to start, and a safe range in which to aim to progress milk feeds. Outcome measures in these studies are however usually very short term, and the emphasis is on rates of LOS and NEC, with the exception of SIFT, where 2-year outcome was the primary end point. Few studies focus on increasing receipt of maternal milk despite the clear benefits of this, and future studies should place more emphasis on the importance of receipt of maximal amounts of maternal milk. Mechanistic data obtained during feeding studies is also limited and may inform future study design and clinical practice.

Conflict of Interest Statement

J.E. Berrington declares research funding to her institution from Prolacta Biosciences and Danone Early Life Nutrition.

References 

1    Granger CL, Embleton ND, Palmer JM, et al: Ma-ternal breast milk, infant gut microbiome, and the impact on preterm infant health. Acta Paedi- atr. 2021;110:450-57.
2    Sherman MP, Zaghouani H and Niklas V: Gut microbiota, the immune system, and diet influence the neonatal gut-brain axis. Pediatr Res. 2015;77:127-35.
3    Bardanzellu F, Fanos V: How could metabolo- mics change pediatric health? Ital J Pediatr. 2020;46:37.
4    Wang M, Ivanov I, Davidson LA, et al: Infant nu-trition and the microbiome: a systems biology approach to uncovering host-microbe interactions. In: Kussmann M, Stover P, editors. Nutrig- enomics and Proteomics in Health and Disease: Towards a Systems Level Understanding of Gene Diet Interactions. London: Wylie and Sons; 2017.
5    de Waard M, Li Y, Zhu Y, et al: Time to full enteral feeding for very low-birth-weight infants varies markedly among hospitals worldwide but may not be associated with incidence of necrotizing enterocolitis: The NEOMUNE-NeoNutriNet Cohort Study. J Parenter Enter Nutr.
2019;43:658-67.
6    Webbe J, Brunton G, Ali S, et al: Parent, patient and clinician perceptions of outcomes during and following neonatal care: a systematic review of qualitative research. BMJ Paediatr Open. 2018;2:e000343.
7    Tao J, Mao J, Yang J, et al: Effects of oropharyngeal administration of colostrum on the incidence of necrotizing enterocolitis, late-onset sepsis, and death in preterm infants: a meta-analysis of RCTs. Eur J Clin Nutr. 2020;74:1122-131.
8    Sohn K, Kalanetra KM, Mills DA, et al: Buccal administration of human colostrum: impact on the oral microbiota of premature infants. J Peri- natol. 2016;36:106-11.
9    Maffei D, Brewer M, Codipilly C, et al: Early oral colostrum administration in preterm infants. J Perinatol. 2020 Feb;40(2):284-87.
10    Lee J, Kim H-S, Jung YH, et al: Oropharyngeal colostrum administration in extremely premature infants: an RCT. Pediatrics. 2015;135:e357-66.
11    Snyder R, Herdt A, Mejias-Cepeda N, et al: Early provision of oropharyngeal colostrum leads to sustained breast milk feedings in preterm infants. Pediatr Neonatol. 2017;58:534-40.
12    Morgan J, Bombell S, Mcguire W: Early trophic feeding versus enteral fasting for very preterm or very low birth weight infants. Cochrane Database Syst Rev. 2013 Mar 28;(3):CD000504.
13    Mosqueda E, Sapiegiene L, Glynn L, et al: The early use of minimal enteral nutrition in extremely low birth weight newborns. J Perinatol. 2008;28:264-69.
14    Leaf A, Dorling J, Kempley S, et al: Early or de-layed enteral feeding for preterm growth-restricted infants: a randomized trial. Pediatrics.
2012;129:e1260-e1268.
15    Morgan J, Young L, Mcguire W: Delayed intro-duction of progressive enteral feeds to prevent necrotising enterocolitis in very low birth weight infants. Cochrane Database Syst Rev. 2014;2014(12):CD001970.
16    Corpeleijn WE, Kouwenhoven SMP, Paap MC, et al: Intake of own mother's milk during the first days of life is associated with decreased morbidity and mortality in very low birth weight infants during the first 60 days of life. Neonatology. 2012;102:276-281.
17    Belfort MB, Anderson PJ, Nowak VA, et al: Breast milk feeding, brain development, and neurocog- nitive outcomes: a 7-year longitudinal study in infants born at less than 30 weeks' gestation. J Pediatr. 2016;177:133-139.e1.
18    Dorling J, Abbott J, Berrington J, et al: Controlled trial of two incremental milk-feeding rates in preterm infants. NEJM. 2019;381:1434-43.
19    Tahir W, Monahan M, Dorling J, et al: Economic evaluation alongside the Speed of Increasing milk Feeds Trial (SIFT). Arch Dis Child Fetal Neonatal Ed. 2020 Nov;105(6):587-92.
20    The National Institute for Health and Care: Guide to the Methods of Technology Appraisal. London: NICE; 2018. pp 1-93.
21    Oddie SJ, Young L, Mcguire W: Slow advancement of enteral feed volumes to prevent necrotis- ing enterocolitis in very low birth weight infants. Cochrane Database Syst Rev. 2021 Aug;8(8):CD001241.
22    Watson J, Mcguire W: Transpyloric versus gastric tube feeding for preterm infants. Cochrane Database Syst Rev. 2013 Feb 28;2013(2):CD003487.
23    Parker L, Torrazza R, Li, Y, et al: Aspiration and evaluation of gastric residuals in the NICU: state of the science. J Perinat Neonatal Nurs. 2015;29:51-9.
24    Premji SS, Chessell L: Continuous nasogastric milk feeding versus intermittent bolus milk feeding for premature infants less than 1,500 g. Cochrane Database Syst Rev. 2011;11:CD001819.
25    Krimmel GA, Baker R, Yanowitz TD: Blood transfusion alters the superior mesenteric artery blood flow velocity response to feeding in prema-ture infants. Am J Perinatol. 2009;26:99-106.
26    Gale C, Modi N, Jawad S, et al: The WHEAT pilot trial - With Holding Enteral feeds around packed red cell Transfusion to prevent necrotising en-terocolitis in preterm neonates: a multicentre, electronic patient record (EPR), randomised con-trolled point-of-care pilot trial. BMJ Open. 2019;9:1-7. 
21 Martini S, Aceti A, Galletti S, et al: To feed or not to feed: a critical overview of enteral feeding management and gastrointestinal complications in preterm neonates with a patent ductus arteriosus. Nutrients. 2020;12:83.
28    Kaur A, Kler N, Saluja S, et al: Abdominal cir-cumference or gastric residual volume as measure of feed intolerance in VLBW infants. J Pedi- atr Gastroenterol Nutr. 2015;60:259-63.
29    Salas AA, Cuna A, Bhat R, et al: A randomised trial of re-feeding gastric residuals in preterm infants. Arch Dis Child Fetal Neonat Ed. 2015;100:F224-28.
30    Masi AC, Embleton ND, Lamb CA, et al: Human milk oligosaccharide DSLNT and gut microbiome in preterm infants predicts necrotising enterocolitis. Gut. 2020 Dec 16;gutjnl-2020-322111.
doi: 10.1136/gutjnl-2020-322111.