Enteral Nutrition in Preterm Infants (2022): A Position Paper From the ESPGHAN Committee on Nutrition and Invited Experts

Abstract

Objectives: 

To review the current literature and develop consensus conclusions and recommendations on nutrient intakes and nutritional practice in preterm infants with birthweight <1800 g.

Methods: 

The European Society of Pediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) Committee of Nutrition (CoN) led a process that included CoN members and invited experts. Invited experts with specific expertise were chosen to represent as broad a geographical spread as possible. A list of topics was developed, and individual leads were assigned to topics along with other members, who reviewed the current literature. A single face-to-face meeting was held in February 2020. Provisional conclusions and recommendations were developed between 2020 and 2021, and these were voted on electronically by all members of the working group between 2021 and 2022. Where >90% consensus was not achieved, online discussion meetings were held, along with further voting until agreement was reached.

Results: 

In general, there is a lack of strong evidence for most nutrients and topics. The summary paper is supported by additional supplementary digital content that provide a fuller explanation of the literature and relevant physiology: introduction and overview; human milk reference data; intakes of water, protein, energy, lipid, carbohydrate, electrolytes, minerals, trace elements, water soluble vitamins, and fat soluble vitamins; feeding mode including mineral enteral feeding, feed advancement, management of gastric residuals, gastric tube placement and bolus or continuous feeding; growth; breastmilk buccal colostrum, donor human milk, and risks of cytomegalovirus infection; hydrolyzed protein and osmolality; supplemental bionutrients; and use of breastmilk fortifier.

Conclusions: 

We provide updated ESPGHAN CoN consensus-based conclusions and recommendations on nutrient intakes and nutritional management for preterm infants

What Is Known

• Nutrient intakes and nutritional practices have a major impact on short-term morbidities and long-term outcomes.
• The available literature has expanded dramatically in the 12 years since the previous European Society of Paediatric Gastroenterology, Hepatology and Nutrition position paper was developed.

What Is New

• We provide an expert consensus on conclusions and recommendations for nutrient intakes and nutritional practices for preterm infants with a birthweight of <1800 g.
• We provide recommendations that can be used in clinical practice, but highlight the lack of strong evidence in several topic areas and the need for further high-quality research especially studies that assess long-term functional outcomes.

The Committee of Nutrition (CoN) of the European Society for Pediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) recognized the need to provide an update of the previous position paper on enteral nutrition (EN) for preterm infants and this was approved by the ESPGHAN council in 2019. The working group was coordinated by members of CoN but recognized the benefit of including additional experts. An initial planning meeting was held as in Oslo in March 2019, at which potential topics were discussed and a provisional list of sections were developed. A lead writer was assigned to each chapter to initiate literature reviews and write the first drafts. We met in Amsterdam in February 2020 at which broad consensus was achieved for most topics. Literature searching and review commenced in 2019 and continued until December 2020. Conclusions and recommendations were graded according to level of evidence (LOE) and grade of recommendation (GOR), and all received >90% consensus (^1,2) (see Supplementary digital content, Supplemental Digital Content 1, https://links.lww.com/MPG/C974 for further details on methods).

This paper provides evidenced informed conclusions and recommendations for clinicians, and EPSGHAN CoN consider this to be a position paper derived by expert consensus (See Table 1). The lack of strong and robust data in many areas imply we do not consider it to be a robust guideline to be adopted without consideration for local contexts and individual infants. We recognize multiple situations where variation in clinical practice is likely to be appropriate and strongly support the need for further research that might reasonably test or study nutrient intakes or nutritional strategies that differ from our current position. We strongly support the use of human milk (HM) and recognize that the variation in nutrient density and absorption make precise recommendations for supplements or fortifiers challenging. We also recognize the need to provide lactation support and hospital policies, guidelines and environments that enable the provision of mother’s own milk (MOM). We provide justification for our assumptions of the average content of human breastmilk in Supplementary digital content, Supplemental Digital Content 2, https://links.lww.com/MPG/C974. Neonatal nutrition research is extremely active, and it is likely that alternative approaches and recommendations may be preferable as our knowledge expands over the next 10 years. ESPGHAN is not responsible for the practices of physicians or other healthcare professionals and provides position papers as indicators of best practice only. Diagnosis and treatment is at the discretion of the healthcare provider.

TABLE 1. - ESPGHAN CoN recommendations for enteral nutrient intakes

	ESPGHAN 2010 recommendation	ESPGHAN 2022 recommendation
Fluid, mL/kg/d	135–200	150–180 (135–200)
Energy, kcal/kg/d	110–135	115–140 (−160)
Protein, g/kg/d	3.5–4.5	3.5–4.0 (−4.5)
Fat, g/kg/d	4.8–6.6	4.8–8.1
Linoleic acid, mg/kg/d	385–1540	385–1540
α-Linolenic acid, mg/kg/d	>55	≥55
DHA, mg/kg/d	12–30	30–65
ARA, mg/kg/d	18–42	30–100
EPA, mg/kg/d	–	<20
Carbohydrate, g/kg/d	11.6–13.2	11–15 (−17)
Sodium, mmol/kg/d	3.0–5.0	3.0–5.0 (−8.0)
Chloride, mmol/kg/d	3.0–5.0	3.0–5.0 (−8.0)
Potassium, mmol/kg/d	1.7–3.4	2.3–4.6
Calcium, mmol/kg/d	3.0–3.5	3.0–5.0
Phosphorus, mmol/kg/d	1.9–2.9	2.2–3.7
Magnesium, mmol/kg/d	0.3–0.6	0.4–0.5
Iron, mg/kg/d	2–3	2.0–3.0 (−6.0)
Zinc, mg/kg/d	1.1–2.0	2.0–3.0
Copper, µg/kg/d	100–132	120–230
Selenium, µg/kg/d	5–10	7–10
Manganese, µg/kg/d	<27.5	1–15
Iodine, µg/kg/d	11–55	11–55
Chromium, µg/kg/d	0.03–1.23	0.03–2.25
Molybdenum, µg/kg/d	0.3–5	0.3–5.0
Thiamine (B1), µg/kg/d	140–300	140–290
Pantothenic acid, mg/kg/d	0.33–2.1	0.6–2.2
Biotin, µg/kg/d	1.7–16.5	3.5–15
Niacin, µg/kg/d	380–5500	1100–5700
Ascorbic acid (vitamin C), mg/kg/d	11–46	17–43
Riboflavin (B2), µg/kg/d	200–400	200–430
Pyridoxine, µg/kg/d	45–300	70–290
Folic acid, µg/kg/d	35–100	23–100
Cobalamin (B12), µg/kg/d	0.1–0.77	0.1–0.6
Vitamin A, IU/kg/d	1333–3300 (400–1000 µg retinol ester/kg/d)	1333–3300 (400–1000 µg retinol ester/kg/d)
Vitamin D, IU/kg/d	800–1000 IU/d	400–700 IU/kg/d (<1000)
Vitamin E, mg/kg/d	2.2–11	2.2–11
Vitamin K, µg/kg/d	4.4–28	4.4–28

ARA = arachidonic acid; CoN = Committee of Nutrition; DHA = docosahexaenoic acid; EPA = Eicosapentaenoic acid, ESPGHAN = European Society of Pediatric Gastroenterology, Hepatology and Nutrition; IU = International units. Figures in brackets represent ranges or upper intakes that might occasionally be needed in routine clinical practice under certain conditions. See text for details.

WATER (see Supplementary digital content 3, https://links.lww.com/MPG/C974)

Water is the major constituent of the human body and a key component of EN, as an essential carrier for nutrients and metabolites. Preterm infants have high fluid requirements due to immature renal function, high water losses, higher surface area to body volume ratio, and because fluid needs are proportional to growth rates.

Determining water requirements is difficult, since a certain volume of water is needed both to maintain body homeostasis, cardiovascular and kidney function, and to provide adequate nutritional intake. These volumes may not be equal, depending on the individual clinical situations and dietary needs. The optimal water intake may also differ depending on macronutrient intakes as higher intakes of protein likely requiring higher fluid intakes. Furthermore, the composition of milk feeds (fortification/formula) affects milk osmolality and the renal load; the latter may be a further factor in determining fluid intakes in preterm infants with limited renal concentration ability and excretory capacity (³). Although prospective studies have demonstrated improved growth with feed volumes up to 200 mL/kg/d (^4,5) caution should be taken in high intake volumes, especially in infants with chronic lung disease or large patent arterial duct. Very few studies have explored outcomes in infancy.

Conclusions and Recommendations

C1: Water requirements show considerable inter- and intra-individual variation, especially in preterm infants. LOE 2++

C2: Water volume needed to maintain body homeostasis, cardiovascular, and kidney function may be different from the volume needed to provide adequate nutrient intakes. LOE 3

C3: In fully enterally fed preterm infants, water balance, hydration status, and renal function should be regularly assessed and considered for the administration of fluid intake. LOE 2++

R1: Most stable growing infants will require fluid intakes of 150–180 mL/kg/d to achieve appropriate nutrient intakes. GOR B

R2: If nutrition needs can be met, fluid intake as low as 135 mL/kg/d may be considered safe to maintain body homeostasis and avoid renal compromise. GPP

R3: In individual preterm infants, enteral fluid intakes up to 200 mL/kg/d may be appropriate and safe depending on current clinical status. GPP

ENERGY (Supplementary digital content 4, https://links.lww.com/MPG/C974)

Energy is required by all cells of the body. Energy supply needs to meet resting energy expenditure (REE), plus the requirements of any physical activity, diet induced thermogenesis, and importantly for preterm infants, tissue deposition (growth) (⁶). Since REE measurements in healthy growing preterm infants also include 1–1.2 kcal/kg weight gain (⁷), REE is directly related to growth rate. Accumulating evidence suggest that REE in preterm infants is around 35–60 kcal/kg/d when full enteral feeds are reached at around 2–4 weeks of age, rising with postnatal age up to 55–70 kcal/kg/d (^8–13). Meta-analysis of these data suggests a range for REE of 60–70 kcal/kg/d, depending on growth rate.

Energy for growth represents energy deposition, and this will vary according to the composition of weight gain, with protein and fat deposition representing 5.65 and 9.25 kcal, respectively per gram of tissue. The estimated average energy requirements for growth are ~3.6–4.7 kcal/g (^14,15), plus REE. Therefore, to achieve 17–20 g/kg/d weight gain, and assuming the composition of that weight gain is 13% protein and 20%–30% fat, the metabolizable energy needed for growth based on an REE of 60–70 kcal/kg/d would be 106–138 kcal/kg/d. Allowing for energy lost in stool (5%–10%) (^16,17), this equates to a total energy intake of approximately 115–160 kcal/kg/d, regardless of feed type. The upper limit is slightly higher than the 110–135 kcal/kg/d recommended in 2010 (¹⁸), though is extrapolated from higher rates of growth and fat deposition. A range of 115–140 kcal/kg/d is sufficient for adequate growth, and fits with data from randomized controlled trials (RCTs) of formula milk or fortifiers (^19–25) and cohort studies aimed at implementing the previous recommendations in clinical practice (^5,26–35).

A key challenge in determining energy requirements is the interdependence of the energy fractions provided by the respective macronutrients. Delivery of a protein:energy ratio (PER) which enables accretion of fat free mass (FFM) and fat mass in the appropriate proportions might have implications for long-term health (^36,37). Studies suggest that the optimal enteral PER for preterm infants is 2.8–3.6 g/100 kcal (^19,38), with PERs at the higher end of this range associated with improved weight gain and FFM accretion. The use of this ratio is however only meaningful if energy and protein intakes are within the recommended ranges. At equal protein and energy intakes, carbohydrate may result in higher nitrogen retention compared with fat (^39,40) although this may be due to differences in absorption rates. Hence, the relative proportion of the macronutrients in the diet also need to be considered.

Since recommendations for energy intake depend on growth targets, we base our recommendations on the aim of supporting growth, body composition, and nutrient retention like the in-utero fetus (¹⁸), while acknowledging that nutritional needs are different in the ex-utero environment. It is important to underline that these recommendations do not consider changes in energy needs related to acute illness or chronic disease states.

Conclusions and Recommendations

C1: The REE for healthy, growing very preterm infants is approximately 60–70 kcal/kg/d. LOE 1+

C2: Metabolizable energy intake needs to meet REE plus the energy needed for growth, adjusted for energy lost in stool. LOE 1-

C3: To promote optimal quality of growth and longer-term outcomes, energy intake recommendations also require consideration of the energy fractions provided by the respective macronutrients. LOE 1

R1: A reasonable range of total energy intake for most healthy growing preterm infants is 115–140 kcal/kg/d. GOR A

R2: Energy intakes >140 kcal/kg/d may be needed where growth is below the recommended range but should not be provided until protein and other nutrient sufficiency has been ensured and should not exceed 160 kcal/kg/d. GOR B

R3: Provided that energy and protein intakes are within the recommended ranges, a PER of 2.8–3.6 g/100 kcal is recommended. GOR B

PROTEIN (see Supplementary digital content, Supplemental Digital Content 5, https://links.lww.com/MPG/C974)

Amino acids are the building blocks for proteins, and selected amino acids have specific functions and are precursors for other metabolites (^41,42). Amino acids that are in excess for protein synthesis capacity are irreversibly oxidized to CO₂ and ammonia, which is detoxified into urea. Protein intake is the main driver of lean body mass growth provided sufficient energy intake. Protein quality is important (⁴³). HM contains approximately 25% nonprotein nitrogen, but its nutritional role remains unclear.

With the factorial approach it is estimated that protein accretion is ~2.5 g/kg/d in infants weighing 500 g and ~2.2 g/kg/d at a body weight of 1800 g (⁴⁴). Preterm infants have obligatory nitrogen losses of ~1 g protein/kg/d, and intestinal utilization of amino acids and suboptimal dietary protein absorption require an additional 0.5 g/kg/d. An extremely preterm neonate requires ~4 g/kg/d of enteral protein of optimal quality to achieve the intrauterine accretion rate (⁴⁵). The protein content of HM is variable with ~1 g/100 mL in mature breastmilk and 1.5–2.0 g/100 mL in colostrum (^46–48). This means that a typical intake of 150 to 180 mL/kg/d of unfortified HM in stable preterm infants will not meet protein requirements.

Several studies have explored optimal protein intakes showing higher growth rates [weight, length, and head circumference (HC)] with increased protein intakes; however, most are under-powered and, in most studies, actual protein intakes are often estimated rather than directly measured. There is no effect of protein intake on key neonatal morbidities or growth outcomes in infancy, and variation in study designs and reporting make meta-analyses challenging.

There are no easy methods to determine optimal protein intakes for individual infants. While there is a strong correlation between protein intake and plasma urea it is unclear whether plasma urea concentration provides information on protein synthesis.

In addition to being important for growth, individual amino acids may have selective functions. These include glutamine (immune function), arginine (gut health), and taurine (brain development), although extra administration of these individual amino acids has not been shown to result in clinical benefits.

Conclusions and Recommendations

C1: Protein content of HM decreases rapidly over time, from around 1.5–2.0 g/dL before 2 weeks of age to around 1.0–1.5 g/dL during the weeks thereafter. Donor milk contains around 0.9–1.0 g/dL of protein or less. LOE 1++

C2: Based on a factorial approach, an extremely preterm neonate would require around 4.0 g/kg/d of enteral protein of optimal quality to grow at a comparable rate as intrauterine. LOE 2+

C3: Several recent RCTs have been performed comparing higher versus more moderate protein intakes; however, most of these studies are underpowered and protein intake is measured imprecisely meaning it is difficult to draw firm conclusions. It seems that enteral protein intakes ranging from 3.5 up to 4.5 g/kg/d are justified to support somatic growth (including head growth), although data on functional outcomes are extremely limited. LOE 1-

C4: The evidence for cut-off values of plasma urea concentrations to guide protein intake is very limited and concentrations may be affected by immature glomerular filtration rate as well. Yet, elevated urea concentrations in the absence of fluid or renal derangements indicate that proteins are not fully used for protein synthesis but are oxidized instead. This thus hints at either optimizing concomitant nutritional intakes or decreasing protein intake. LOE 1-

C5: Separate supplementation of certain extra amino acids (eg, glutamine, arginine, or taurine) may reduce several neonatal morbidities but data are limited. While arginine appears promising in reducing necrotizing enterocolitis (NEC) rates, the number of infants studied remains limited. LOE 1-

R1: We strongly recommend very preterm infants are given at least 3.5–4.0 g protein/kg/d together with sufficient other macro- and micronutrients. Protein intake may be further increased up to 4.5 g/kg/d where growth is slow, provided protein quality is good, concomitant energy and other micronutrient intakes are optimal, and there are no other causes for suboptimal growth. GOR A

R2: We conditionally recommend monitoring plasma urea at regular intervals. Low urea concentrations after the first few weeks of life may indicate enteral protein intakes can be increased up to 4.5 g/kg/d. If urea concentrations are above 5.7 mmol/L (34 mg/dL; or 16 mg N/dL) in the absence of fluid or renal derangements, while providing sufficient concomitant energy, lowering of protein intake should be considered. GOR C

R3: No recommendation can be made regarding the use of additional supplementation with glutamine, arginine, or taurine to decrease neonatal morbidities. GOR B

LIPIDS (see Supplementary digital content, Supplemental Digital Content 6, https://links.lww.com/MPG/C974)

Dietary fats provide about 50% of the energy needs of preterm infants as well as essential polyunsaturated fatty acids (PUFAs), lipid soluble vitamins, and complex lipids. HM is a suspension of fat globules with a variable fat concentration of about 3.2–4 g/100 mL. The core of the milk fat globule (MFG) consists of 98%–99% triglycerides surrounded by a membrane of phospholipids, cholesterol, and other highly active bioactive components (⁴⁹). About 15%–20% of FAs in HM are PUFAs (^50,51). The balance of the essential and conditionally essential PUFAs to each other is important because these FAs compete for desaturases and elongases in the PUFA conversion pathways.

Arachidonic acid (ARA) and docosahexaenoic acid (DHA) are actively transferred through the placenta during the 3rd trimester of pregnancy and brain accumulation is considerable (^52,53). Many studies show a decrease in ARA and DHA levels in preterm infants after birth suggesting insufficient endogenous synthesis from the essential FAs linoleic acid (LA) and α-linolenic acid (ALA) (^54,55). ARA and DHA are thus considered conditionally essential in preterm infants. Reduced concentrations of both ARA and DHA are associated with increased risk of retinopathy of prematurity (ROP), septicemia, and severe bronchopulmonary dysplasia (⁴⁹). Data from meta-analysis and RCTs on the effect of DHA supplementation (with or without ARA) on neurodevelopmental and other clinical outcomes show inconsistent results but this may reflect variation in study design and methodology (see Supplementary digital content, Supplemental Digital Content 6, https://links.lww.com/MPG/C974 for further detail and references). Estimates for lipid needs based on fetal lipid accretion, losses due to fat malabsorption, unavoidable oxidation, and conversion of absorbed to tissue deposited triglyceride are 3.8–4.8 g/kg/d (¹⁸). Aiming for dietary fat to provide 45%–55% of the energy intake, a minimum supply of 4.8 g/kg/d is required to assure 96 kcal/kg/d of nonprotein calories. Mature breast milk fed at 160–180 mL/kg/d will provide a mean fat intake of up to 7 g/kg/d (^46,56) with an upper interquartile range of ~8.1 g/kg/d. These intakes appear safe even in extremely low-birth weight infants (⁵⁷).

Adding ARA and DHA to enteral feeds are thought to be a reliable way of ensuring adequate supplies of these PUFAs, but both the composition and the mode of administration (enteral or buccal) need to be considered. Providing >50 mg/kg/d of DHA seems sufficient to obtain DHA concentrations like fetal blood in utero. Determining appropriate ARA intakes are demanding because ARA requirements are less studied than that of DHA, and the endogenous synthesis of ARA is more efficient. Intakes based on the average value observed in HM of 0.5% of FAs would equal 30 mg/kg/d ARA. Considering a range of DHA intake of 30–65 mg/kg/d, and an ARA:DHA ratio ranging from 0.5 to 2, an ARA intake up to 100 mg/kg/d appears safe. Limited data are available to determine if there is any benefit for dietary eicosapentaenoic acid and because of concerns around toxicity.

Conclusions and Recommendations

C1: There is not enough new data to support a significant modification of previous recommendations for linoleic and linolenic acids nor medium chain triglycerides. LOE 2+

C2: DHA supplementation has modest and transient effects on neurodevelopment outcomes but may help achieve intakes close to intrauterine accretion rate. LOE 1+

C3: With a DHA intake range of 30–65 mg/kg/d and aiming for an ARA/DHA ratio ranging from 0.5 to 2, 15–100 mg/kg/d of ARA appear to be safe. LOE 1-

C4: Limited data are available to define if there is any benefit for including Eicosapentaenoic acid (EPA) in the diet of preterm infants. LOE 3

R1: A total fat intake of 4.8–8.1 g/kg/d is recommended although higher intakes may be safe. GOR B

R2: Amounts of medium chain triglycerides exceeding 40% of total fat are not recommended. GOR B

R3: A linoleic acid intake of 385–1540 mg/kg/d, a minimum linolenic acid intake of 55 mg/kg/d, and a linoleic acid to linolenic acid ratio of 5–15:1 (w/w) are considered acceptable. GOR B

R4: A DHA intake of 30–65 mg/kg/d is recommended assuming sufficient intake of ARA. GOR A

R5: An ARA intake of 30–100 mg/kg/d is recommended. GOR B

R6: EPA intake should be <20 mg/kg/d. GPP

CARBOHYDRATES (see Supplementary digital content, Supplemental Digital Content 7, https://links.lww.com/MPG/C974)

The carbohydrate concentration of HM is quite stable and increases from ~6.2 g/100 mL to 7.1 g/100 mL during the first month of life (^47,58). The predominant digestible carbohydrate is the disaccharide lactose (^48,59) but free glucose, galactose, and human milk oligosaccharides (HMOs) comprise about 15%–30% (⁴⁸). Incomplete digestion of lactose (HM) or lactose/glucose polymers (preterm formula) may limit the availability of energy from carbohydrate; however, lactose feeding increases intestinal lactase activity (⁶⁰). Undigested lactose and glucose polymers are salvaged by colonic bacteria (^59,61). Concerns that carbohydrate malabsorption increases the risk of NEC have been explored in studies, but replacing lactose with more readily digestible glucose polymers show inconsistent results in regard to feeding tolerance, weight gain, and calcium absorption (^62–66).

The low glycogen reserves and limited fat stores put preterm infants at risk of hypoglycemia (⁶⁷), but they are also at risk of hyperglycemia due to immature glucose regulatory mechanisms including persistent hepatic gluconeogenesis, decreased pancreatic beta-cell activation, and partial insulin resistance (^67–69). Hyperglycemia is associated with increased mortality, morbidity, and longer-term brain outcomes (^69,70).

Carbohydrates constitute 45%–50% of nonprotein calories in HM and standard preterm formulas. The relative contribution of carbohydrate to total nonprotein energy might be of importance (⁷¹). At equal protein and energy intakes, carbohydrate improves nitrogen retention compared with fat (^39,40). However, high-energy, high-carbohydrate intakes increase fat deposition (^39,40), and higher postnatal carbohydrate intakes and hyperglycemia have been associated with higher blood pressure at 6.5 years of age (⁷²). The balance of benefits and risks needs consideration when determining carbohydrate intake recommendations. Supplementation of HM exclusively by digestible carbohydrates (e.g. lactose, glucose, or glucose polymers) has not been studied in RCTs (⁷³). Therefore, the optimum enteral carbohydrate still is unknown but observational studies suggest that increasing energy together with protein by providing either fortified or non-fortified high HM volume regimens are safe and improve in-hospital growth (^4,74,75) and possibly language scores at 2 years of age (⁷⁴). Assuming energy needs of 115–140 kcal/kg/d, enteral protein intakes of 3.5–4.0 g/kg/d, and a carbohydrate composition contributing to 40%–50% of nonprotein energy intakes, a carbohydrate intake in the range of 11–15 g/kg/d seems reasonable.

Conclusions and Recommendations

C1: There are no new data from RCTs on the effects of exclusively increasing carbohydrate intakes on short- and long-term outcomes in preterm infants; therefore the optimal intake range is uncertain. LOE 2

C2: Observational data suggest that carbohydrate given as fortified HM at upper intake ranges is safe and improves in-hospital weight, length, and HC growth. LOE 2++

C3: Preterm infants fed formula may need lower carbohydrate intakes than infants fed fortified HM, due to higher absorption rates of glucose polymers compared to lactose. LOE 2

C4: The optimal lactose to total carbohydrate ratio in HM fortifiers or in preterm formulas is unknown. LOE 3

R1: In preterm infants, a carbohydrate intake of 11–15 g/kg/d is recommended. GOR B

R2: Higher carbohydrate supplies as part of higher multicomponent supplementation or higher HM intakes may be considered during a short period of time to cover cumulative deficits and facilitate catch-up growth if tolerated (euglycemia), but should also be tapered accordingly to avoid overnutrition. GPP

SODIUM, CHLORIDE, AND POTASSIUM (see Supplementary digital content, Supplemental Digital Content 8, https://links.lww.com/MPG/C974)

Sodium

Sodium (Na) is the principal cation in extracellular fluid and concentrations influence intravascular and interstitial volumes, and blood pressure. Na also has a role in bone mineralization, nerve conduction, nitrogen retention, and growth. Fetal accretion rates of Na are estimated at 1.6–2.1 mmol/kg/d at 27–34 weeks of gestation (⁷⁶). Gastrointestinal Na absorption is effective in preterm infants with typical fecal Na excretion rates <10% of intake, but is higher in early postnatal life and those born more preterm (⁷⁷). Preterm infants have limited renal capacity both to conserve Na when challenged by Na restriction and excrete Na when challenged by a Na load. Tubular Na loss is inversely associated with gestational age (⁷⁸), and it is increased during critical illness and by certain medications (⁷⁹). Urinary Na losses as high as 7 mmol/kg/d have been reported in preterm infants (⁸⁰). The Na concentration of HM declines rapidly over the first few postnatal days (⁸¹), and is also influenced by expression methods and maternal serum concentration (^82,83).

Few high quality RCTs of Na supplementation exist, although some show that higher Na intakes of 4–6 mmol/kg/d versus 3–4 mmol/kg/d increase weight gain. Furthermore, the addition of concentrated sodium chloride or sodium phosphate to expressed fortified HM may increase milk osmolality (⁸⁴).

Conclusions and Recommendations

C1: Na requirements show considerable inter- and intra-individual variation, especially in very low birthweight (VLBW) infants. LOE 2

C2: Breastmilk with added fortifiers may be insufficient to meet Na needs in preterm infants. LOE 2++

C3: The enteral administration of Na additives exposes the infant gut to higher osmolality. LOE 3

R1: A Na intake of 3–8 mmol/kg/d is recommended. The upper range of Na intake is slightly higher than in previous recommendations and should be considered in infants receiving high energy and protein intakes or with important sodium loss. GPP

R2: Na additives should be diluted with milk and divided between different feeds over 24 hours to maintain osmolality as low as possible. GOR C

Chloride

Chloride (Cl) is the most abundant anion in extracellular fluid and along with Na, helps maintain osmotic pressure and hydration. Cl is also involved in maintaining ionic neutrality. The difference between Na and Cl plasma concentration affects hydrogen and bicarbonate ion concentrations. The daily turnover of Cl is high, and renal tubular reabsorption rate is 60%–70%. Chloride content in HM is similar at different gestations (⁸⁵). Low Cl intakes may result in failure to thrive, slower growth, and delayed neurological development (^86,87).

In enterally fed preterm infants receiving oral salt supplementation, Cl intake parallels that of Na or K, so, where there are high intakes of Na or K there will also be high Cl intakes (⁸⁸). Cl losses and excretion can occur independently from Na. Medications may also be a source of additional Cl intake. Studies in infants receiving parenteral nutrition (PN) suggest that Cl intake should be slightly lower than the sum of Na and K intakes to avoid severe metabolic acidosis (⁸⁹).

Conclusions and Recommendations

C1: Cl intakes parallel that of Na where oral salt supplementation is used. LOE 2++

C2: High Cl intakes from additives and HM fortifiers with low strong ion difference may induce metabolic acidosis. LOE2++

R1: A Cl intake of 3–8 mmol/kg/d is recommended. GOR C

R2: Cl intake from HM fortifiers and preterm formula should be slightly lower than the sum of Na + K intakes to avoid metabolic acidosis. HM fortifiers should provide buffers to compensate for high renal acid load. GOR B

Potassium

Potassium (K) is the most abundant cation in the human body and the major intracellular ion. The K concentration gradient across cell membranes is crucial for maintaining contractility and neuronal function and is maintained by the tight balance of the influx or efflux of K from intra- to extracellular spaces. K is needed for somatic growth and the K pool correlates well with lean body mass. A growth rate of 15 g/kg/d results in a net storage of about 1.0–1.5 K mmol/kg/d (⁷⁷). The total body K content depends on the balance of K intake and excretion and is mainly dependent on renal regulation. After an enteral feed, 80% of the absorbed K enters the cells due to increased insulin concentrations stimulated by the contemporary absorption of glucose and amino acids (⁹⁰). Renal K excretion is increased by diuretics, and several factors increase gastrointestinal K losses including vomiting and diarrhea, changes in aldosterone, epinephrine, and prostaglandins (⁹⁰).

Several studies in parenterally fed preterm infants show an increased incidence of hypokalemia with higher protein and energy intakes (^91–93). In infants receiving amino acid intakes of 3 g/kg/d, the K balance remains positive with K intakes ≥2 mmol/kg/d. It seems likely that similar amounts are needed when enterally fed but the optimal K intake in infants receiving higher protein intakes (<4.5 g/kg/d) is not clear. K concentrations in extracellular fluid are tightly regulated, implying that intracellular K deficiency may still occur in the presence of a normal plasma K concentration.

Conclusions and Recommendations

C1: In enterally fed preterm infants there is a linear association between K needs and protein retention. LOE 3

R1: A K intake of 2.3–4.6 mmol/kg/d is recommended. The upper range of K intake should be considered in growing infants receiving the upper ranges of energy and protein intakes. GOR B

MINERAL INTAKES Ca, P, Mg (see Supplementary digital content, Supplemental Digital Content 9, https://links.lww.com/MPG/C974)

Bone mineral metabolism in early neonatal life is complex and determining optimal intakes is challenging. Ca accretion in the bone accounts for ~98% of total Ca-stores, whereas P stored in bone only represents ~80% of the total P accretion. The remaining P, about 20%, is involved in lean mass accretion, incorporated in nucleic acids and cell membranes, or used in the intra-cellular energy metabolism. This means that P intakes must be greater than that needed simply for bone mineral accretion (^94–96). Estimates of the fetal mineral accretion rate and estimates of the rate of mineral absorption by the preterm intestine (^{18,95,97–99}) suggest fetal accretion rates of calcium (Ca), phosphorus (P), and magnesium (Mg) of approximately 2.5–3.0 mmol/kg/d, 1.6–2.1 mmol/kg/d, and 0.12–0.21 mmol/kg/d, respectively (^76,100,101). Dietary mineral provision to achieve the estimated retention rates may be quite variable, as Ca and P absorption rates range between 30%–70% and 70%–90%, respectively (^102,103). Bioavailability is also dependent on the type and composition of the milk, the fatty acid components (eg, presence of beta-palmitate), and the form of mineral given (^102–110). Accretion rates in preterm infants are lower than in-utero estimates, resulting in a lower bone mineral content (BMC) at term corrected age compared to term-born peers.

The low mineral content of HM does not meet the needs of preterm infants, and studies show that current breastmilk fortifiers and preterm milk formula result in Ca and P retention rates of 2.3–2.8 and 2.2–2.6 mmol/kg/d, respectively, close to fetal accretion rates. Insufficient provision of Ca and/or P may lead to osteopenia and fractures. Improvements in nutritional care, including more appropriate use of breastmilk fortifiers, changes in milk formula composition, and better positioning of infants inside incubators, have resulted in a reduction in fractures, albeit osteopenic changes on X-ray are still common (^111,112). Unfortunately, there are no useful clinical techniques that directly measure or estimate BMC to guide clinical practice.

While adequate mineral intakes are clearly important, the only RCT with long-term outcomes showed no effect of different postnatal calcium and phosphorus intakes on adult BMC (¹¹³). Based on an intestinal absorption rate of 60% for Ca and 80%–90% for P, we estimate that healthy growing preterm infants will need approximately 4–4.5 mmol/kg/d of Ca and 3–3.5 mmol/kg/d of P when fed fortified HM. However, intakes up to 5 mmol/kg/d of Ca and 3.7 mmol/kg/d of P (or higher) may be needed if milk formulas with poor mineral absorption are provided.

Mg accretion during the last trimester of gestation is around 0.12–0.21 mmol/kg/d with around half being accreted in the bone, and the remainder in muscle and soft tissue. Mg absorption rates change depending on Mg intakes but are typically around 40%–50% (^102,114), and in preterm infants serum concentrations are higher than older infants with a range of 0.6–1.25 mmol/L (¹¹⁵). Studies in preterm infants on fortified HM providing 0.2–0.3 mmol/kg/d showed absorption rates of around 45%–50% leading to a Mg retention of 0.1 mmol/kg/d. In preterm infants fed formula, Mg intakes are approximately 0.4–0.5 mmol/kg/d which appears to be adequate. No RCTs determining effects on bone accretion have been conducted.

Conclusions and Recommendations

C1: The lack of a strong evidence base for determining mineral intakes that will optimize functional bone or other outcomes means that recommended reference ranges are wide. LOE 2+

C2: Inadequate mineral intakes postnatally result in osteopenia which increases the risks for bone fractures in preterm infants. However, there is no consensus regarding how best to assess BMC in clinical practice and there are few well-designed RCTs to determine optimal mineral intakes. LOE 3

C3: Targeting a Ca retention of 2.2–2.8 mmol (90–110 mg)/kg/d is appropriate to minimize mineral bone deficiency and the risk of fractures in preterm infants. The target for P retention is 2.2–2.6 mmol (70–80 mg)/kg/d and includes both the functional P requirements as well as the P requirement for bone- and soft tissue accretion. LOE 3

C4: Adequate phosphorus intakes are essential to accrete lean tissue (each gram of protein requires approximately 0.35 mmol of phosphorus). The provision of PN with low phosphate and unfortified HM increase the risk of both early and late hypophosphatemia. LOE 2+

R1: It is recommended to fortify HM early with phosphate followed by early introduction of multicomponent breastmilk fortifiers to optimize bone mineral outcomes. GOR C

R2: A Ca intake of 3.0–5.0 mmol (120–200 mg)/kg/d and a P intake of 2.2–3.7 mmol (70–115 mg P)/kg/d of P are recommended. GOR C

R3: The recommended molar calcium to phosphate ratio to ensure adequate Ca retention is ≤1.4 (≤1.8 in mass). GOR C

R4: Preterm infants fed artificial milk formula may require higher mineral intakes than those fed HM. GPP

R5: Regular monitoring of P and Ca status is recommended. We do not recommend the routine use of bone imaging or other direct assessments of BMC in clinical practice. GOR C

R6: In preterm infants fed fortified HM or preterm milk formula, a Mg intake of 0.4–0.5 mmol (9–12.5 mg)/kg/d is recommended. GOR C

TRACE ELEMENTS (see Supplementary digital content, Supplemental Digital Content 10, https://links.lww.com/MPG/C974)

Trace elements are essential for many functions in different organ systems, as well as for normal growth and development. While adequate dietary intakes are important for preterm infants to prevent deficiencies, important adverse effects of excessive intakes exist.

Iron

Systematic reviews clearly show that iron supplements effectively prevent iron deficiency anemia in preterm infants but there is no benefit in exceeding standard doses of iron (ie, 2–3 mg/kg/d) in VLBW infants (¹¹⁶). Overall, there is a lack of RCTs with long term neuro-developmental outcomes, but RCTs in late preterm infants have shown improved developmental outcomes in iron supplemented infants. Furthermore, starting at ~2–3 weeks versus later ~4–8 weeks of age is associated with a lower need for blood transfusions in VLBW infants (¹¹⁷). Delayed umbilical cord clamping increases neonatal iron stores and is associated with a lower mortality, lower risk of intraventricular hemorrhage and lower need for red cell transfusions in preterm infants (¹¹⁸). Erythropoietin may reduce the need for red blood cell transfusions but requires much higher iron intakes.

Ferritin is a useful biomarker of iron status, but reference intervals differ from older infants and children. Ferritin concentrations <35–40 µg/L indicate iron deficiency while concentrations >300–350 µg/L indicate iron overload (^119–121). Ferritin is not useful as a biomarker of iron status in patients with ongoing inflammation or liver disease.

Recommendations

• A daily iron intake of 2–3 mg/kg/d starting at 2 weeks of age is recommended for VLBW infants. LOE 1+, GOR A
• Infants who receive erythropoietin treatment need a higher dose (up to 6 mg/kg/d). LOE 1-, GOR B
• Since individual iron status in VLBW infants is highly variable, depending on the number of received blood transfusions and blood losses from phlebotomy, it is recommended to follow these infants with repeated measurements of serum ferritin. LOE 4, GOR GPP
• If ferritin is <35–70 µg/L, the iron dose may be increased up to 3–4 (or maximum 6) mg/kg/d for a limited period. LOE 4, GOR GPP
• Prolonged dietary iron intakes of >3 mg/kg/d should be avoided in most cases because of possible adverse effects. LOE 1-, GOR B
• If ferritin is >300 µg/L, which in the absence of ongoing inflammation and liver disease usually is the result of multiple blood transfusions, iron supplementation and fortification should be discontinued until serum ferritin falls below this level. LOE 4, GOR GPP
• Iron supplements or intake of iron-fortified formula in the recommended doses should be continued until 6–12 months of corrected age. LOE 4, GPP
• Like all infants, preterm infants should receive iron-rich complementary foods from 6 months of age. LOE 1+, GOR A
• Delayed umbilical cord clamping, whenever feasible, is recommended for all preterm infants. LOE 1++, GOR A

Zinc

Zinc is an essential trace element involved in growth and tissue differentiation. Zinc deficiency in preterm infants is associated with stunted growth, increased risk for infections, skin rash, and possibly poor neurodevelopment (¹²²). In contrast to iron and copper, zinc does not have a pro-oxidant effect and adverse effects of excess zinc intakes are rarely reported, except for a negative effect on copper absorption with high zinc intakes.

The factorial method combined with data from metabolic balance studies suggest enteral zinc intakes of at least 2.0–2.25 mg/kg/d are required (¹²³) and up to 3 mg/kg/d in extremely preterm infants due to faster growth rates (^88,124). A small number of studies suggest an intake of at least 1.4–2 mg/kg/d is needed to achieve optimal growth in preterm infants (^125,126) but higher enteral intakes appear safe and may be beneficial. Two recent meta-analyses suggests that zinc supplementation improves weight gain and linear growth in preterm infants and may decrease mortality (^127,128). Very preterm infants can develop symptomatic zinc deficiency with acrodermatitis enterohepatica and/or poor growth, especially those infants who have an enterostomy after NEC surgery (¹²⁹).

Recommendations

• We recommend an enteral zinc intake of 2–3mg/kg/d, based on the most recent RCT as well as on factorial calculations. LOE 2, GOR C
• Measurement of serum zinc should be considered in preterm infants with poor growth and low alkaline phosphatase level, especially if they have excessive GI fluid losses. LOE 4, GOR GPP

Other Trace Elements

Recommendations for copper, selenium, manganese, iodine, chromium, and molybdenum are covered only very briefly here, while the full background is reported in the supplementary material, Supplemental Digital Content, https://links.lww.com/MPG/C974. The recommended copper intake has been increased to 120–230 µg/kg/d to compensate for the higher recommended zinc intake (see above), since these 2 ions compete for intestinal absorption. We recommend an enteral selenium intake of 7–10 µg/kg/d, which has been shown to result in Se status like term infants and possibly a reduced risk of sepsis. Based on the average HM manganese content and the lower range of manganese in current preterm formulas, an enteral manganese intake of 1–15 µg/kg/d can be recommended. Despite a recent RCT, there is not enough conclusive evidence to change the previous recommendation, so an iodine intake of 11–55 µg/kg/d is recommended. The recommendations for chromium (0.03–2.25 µg/kg/d) and molybdenum (0.3–5 µg/kg/d) are unchanged.

WATER SOLUBLE VITAMINS (see Supplementary digital content, Supplemental Digital Content 11, https://links.lww.com/MPG/C974)

Water soluble vitamins are essential for whole body function and homeostasis. There is a major lack of studies to determine requirements for preterm infants. We used European Food Safety Authority (EFSA) recommendations for infants <6 months to calculate weight-based estimates, human breastmilk concentrations to estimate intakes in healthy infants, clinical studies where they exist and considered the amount provided by preterm infant formulas when fed at a minimum energy intake of 115 kcal/kg/d, acknowledging that this last approach may significantly overestimate needs and noting that excess intakes are unlikely to be beneficial. Wide ranges in intake recommendations for preterm infants do not represent the distribution of intakes in a population, and in the absence of robust evidence dietary recommendations may be inflated to ensure adequacy well in excess.

Our approach is described in the supplementary materials, Supplemental Digital Content, https://links.lww.com/MPG/C974. We acknowledge that estimates solely derived from EFSA recommendations or the concentration in breastmilk may underestimate the increased requirements of the rapidly growing preterm infant. However, it is likely that total daily recommendations from EFSA for term infants will be adequate for preterm infants as the weight difference results in an approximate 3–5-fold higher intake per kg body weight. The recommendations, which are largely similar (but not identical) to the previous ESPGHAN recommendations, are only briefly reported here and we refer to the full text in the supplementary documents, Supplemental Digital Content, https://links.lww.com/MPG/C974.

Thiamine: Considering the available evidence we propose an intake of 140–290 μg/kg/d based on the content of infant formula milk [LOE 3, GPP]. The B1 content in breastmilk or the EFSA daily recommendations for B1 (42 μg/100 kcal or 46 μg/kg) might also be adequate (¹³⁰).

Pantothenic acid: Considering the available evidence we propose an intake of 0.6 to 2.2 mg/kg based on the content of infant formula milk. LOE 3, GPP

Biotin: We recommend using the lowest concentration of biotin that would be provided using preterm formula and the upper level in the previous recommendations by ESPGHAN (¹⁸). We propose an intake of 3.5–15 μg/kg based on the content of infant formula milk. LOE 3, GPP

Niacin: Considering the available evidence we propose an intake of 1100–5700 μg/kg based on the content of infant formula milk. LOE 3, GPP

Ascorbic acid (vitamin C): Considering the available evidence we propose an intake of 17–43 mg/kg based on the content of infant formula milk. LOE 3, GPP

Riboflavin (B2): We recommend an intake of riboflavin of 200–430 μg/kg/d [LOE 3, GOR GPP]. The daily recommendations made by the EFSA panel may also be adequate considering the lack of evidence of differences in requirements between preterm and full-term infants.

Pyridoxine: We suggest a recommended intake in keeping with the concentration range that would be provided by commercially available preterm milk formula (70–290 μg/kg/d) [LOE 3, GOR GPP]. This is close to the previous guidelines of ESPGHAN (¹⁸), the EFSA daily recommendations, and the available evidence discussed above.

Folate: Considering the range of folic acid content in preterm infant formula (20–45 μg/100 kcal, or 23–52 μg/kg) we propose an intake of 23–100 μg/kg/d [LOE 3, GPP]. Based on the evidence available, a folate intake at the maximum of the recommended range may improve patients’ outcomes.

Cobalamin (B12): We propose dietary recommendations of 0.10–0.60 μg/kg based on the cobalamin content in preterm infant formulas [LOE 3, GPP]. An intake >0.6 μg might be associated with excessively elevated B12 levels in blood.

FAT SOLUBLE VITAMINS ADEK (see Supplementary digital content, Supplemental Digital Content 12, https://links.lww.com/MPG/C974)

Vitamin A

Vitamin A is essential for growth and tissue differentiation (^131,132), and especially important in lung maturation (¹³³). Preterm infants often have lower plasma concentrations of both retinol and retinol binding protein at birth compared with term infants reflecting low hepatic stores (¹³⁴). A plasma retinol concentration of ≥ 200 ng/mL is generally considered adequate (¹³⁵), but due to the complexity of vitamin A metabolism and organ immaturity in preterm infants, vitamin A supplementation may still not result in adequate concentrations in blood. The beneficial role of high dose intramuscular vitamin A for prevention of bronchopulmonary dysplasia (BPD) in preterm infants has been demonstrated (¹³⁶). However, due to discomfort, this form of supplementation is not common practice. Studies of higher enteral doses (5000 IU/d) produced inconsistent effects on BPD (^137–140).

Conclusions and Recommendations

C1: There are insufficient data to change the previous recommendation of daily vitamin A intake in preterm infants (¹⁸), but infants with hepatic impairment may need higher intakes, and those with renal impairment may require lower doses. LOE 2++

R1: Based on best current available data, we recommend a daily total intake of vitamin A of 1.333–3.300 IU/kg body weight (400–1000 µg retinol ester/kg/d). GOR B

Vitamin D

Vitamin D plays a critical role in multiple cellular processes especially bone metabolism and the immune system (¹⁴¹). Pathways of vitamin D absorption and metabolism are fully operative in babies <28 weeks GA (^142–144). Mineral bone deficiency in preterm infants is common and is primarily caused by suboptimal intakes of calcium and phosphate, but this can be compounded by vitamin D deficiency (¹⁴⁵). Even though there is no consensus regarding the definition of vitamin D deficiency in infants, the ESPGHAN Committee on Nutrition has previously recommended the pragmatic use of a serum 25-hydroxy vitamin D concentration >50 nmol/L to indicate sufficiency and a serum concentration <25 nmol/L to indicate severe deficiency, while noting that excessive vitamin D intakes resulting in concentrations >120nmol/L should also be avoided (^146,147).

Several studies have assessed the effect of vitamin D3 intake on circulating of 25(OH)D concentrations, but few studies assess the impact on bone mineral density after the immediate neonatal period and results are inconsistent (^{145,148–150}). Vitamin D intakes of 400–670 IU/kg/d in infants weighing 1500–2000 g and 500–1000 IU/kg/d for those weighing 1000–1500 g reduces the risk of deficiency (^148–150), but studies suggest that lower doses (200–300 IU/kg/d) may also be sufficient (^148,151). Conversely, another study suggested a daily vitamin D supplementation of 800 IU/d in extremely premature infants with a gestational age < 28 weeks (¹⁵²). Based on this, we recommend a daily vitamin D intake of 400–700 IU/kg/d (10–17.5 µg/kg/d) for stable preterm infants. This corresponds to 300–525 IU/d at 750 g body weight, 400–700 IU/d at 1000 g body weight, and 600–1000 IU/d at 1500 g body weight. The maximum recommended routine intake is 1000 IU/d, but preterm infants who are vitamin D deficient due to maternal vitamin D deficiency or cholestasis may temporarily need higher doses. Adequate vitamin D supplementation could be monitored by measuring serum 25(OH)D at 3–4 weeks of life and then every month until discharge to adapt vitamin D supplementation to each individual’s needs.

Conclusions and Recommendations

C1: Ensuring adequate vitamin D intakes in preterm infants is essential for bone health and may possibly have positive effects on immune function, even though this is not conclusively shown. LOE 2+

C2: There are few adequately powered controlled trials on which to base firm recommendations in preterm infants, and even fewer trials provide clinically relevant outcomes beyond vitamin D concentrations, for example, markers of bone health. LOE 4

R1: Based on currently available data, we recommend a daily vitamin D intake of 400–700 IU/kg/d (10–17.5 µg/kg/d) during the first months of life with a maximum dose of 1000 IU/d (25 µg/d). LOE 2++, GOR B

Vitamin E

Vitamin E comprises a group of 8 biologically active tocopherols which act as antioxidants to scavenge free radicals, potentially limiting lipid peroxidation which can lead to BPD, ROP, and hemolytic anemia (^153,154).

Low concentrations of vitamin E were found at birth and at discharge in preterm infants (¹⁵⁵) but serum concentrations may not reflect tissue concentrations (¹⁵⁶).

Clinical studies of supplementation are inconsistent, and although there may be benefits (^157,158), there are some data suggesting high intakes may increase risk of sepsis and NEC (¹⁵⁹).

Studies of routine enteral vitamin E supplementation in preterm infants suggest it may be prudent to maintain plasma vitamin E concentrations of 10–35 mg/L, and a ratio of serum-α-tocopherol of at least 1 mg to 1 g total lipids, which would suggest a minimum dose of 3.8 mg/kg/d. However, no clinical benefits have been seen, and the recommended daily intake of vitamin E in preterm infants is 2.2–11 mg/kg/d (^18,159,160). Infants with prolonged cholestasis may require higher intakes.

Conclusions and Recommendation

R1: Based on the current available data, we recommend a daily dose for vitamin E supplementation in preterm infants of 2.2–11 mg/kg/d. LOE 2++, GOR B

Vitamin K

Vitamin K is a group of lipophilic, hydrophobic vitamins necessary for the synthesis of coagulation factors [factors II (prothrombin), VII, IX, and X, and the anticoagulation proteins C and S in the liver]. Maternal transfer of vitamin K across the placenta is very low with cord blood concentrations of vitamin K often below the detection limit of 0.02 ng/mL in healthy newborns (¹⁶¹) and breastmilk also has very low levels of vitamin K (^162,163). Late onset vitamin K deficiency bleeding is primarily seen in exclusively breast-fed infants or in those with cholestatic disease (¹⁶⁴).

While preterm infants are at high-risk of vitamin-K deficiency bleeding most receive prophylactic vitamin K at birth, and additional intakes from PN, infant formula, and breast milk fortifiers. Vitamin K can be administered intramuscularly, intravenously, and orally with different recommended dosing regimen (¹⁶⁵). Thus, serum concentrations in preterm infants are usually higher than those found in term formula-fed infants (¹⁶⁶). There are no RCTs in preterm infants, and nutritional recommendations vary from 4.4 to 28 µg/kg/d, to up to 100 μg/kg/d (^18,163,167) although higher intakes may be needed where there is prolonged cholestasis.

Conclusions and Recommendation

R1: Based on the current based available data, we recommend a daily dose for vitamin K of 4.4–28 µg/kg. LOE 2++, GOR B

FEEDING MODE: MINIMAL ENTERAL FEEDS, FEED ADVANCEMENT, GASTRIC RESIDUALS, AND TIMELINE OF PARENTERAL AND ENTERAL NUTRITION (see Supplementary digital content, Supplemental Digital Content 13, https://links.lww.com/MPG/C974)

Minimal Enteral Feedings and Enteral Fasting During the First Days of Life

Minimal enteral feeding (MEF) is synonymous with gut priming, minimal EN, trophic feeding, or hypocaloric feeding and defined as small volumes of milk (typically 12–24 mL/kg/d) without advancement in feed volumes during the first 3–7 days (^168,169). Numerous studies and systematic reviews exist comparing MEF with no feeds, or earlier advancement, but many studies were conducted more than 20 years ago. Overall there are no consistent effects on NEC or all-cause mortality (^170,171) and this is supported by more recent studies (^172–174).

Conclusions and Recommendations

C1: MEFs are defined as nutritional insignificant small volumes of milk (typically 12–24 mL/kg/d) without advancing the feed volumes for a period of 3–7 days. LOE 1+

C2: There is no clear beneficial effect of enteral fasting or MEF of any duration compared to advancing feeds immediately after birth. LOE 1+

R1: Start small volume enteral feeds as soon as possible after birth in most preterm infants and advance feeds as clinically tolerated. GOR B

Advancement of Enteral Feeds

Systematic review of 10 RCTs including a total of 3753 infants show no impact of speed of advancement (¹⁷⁵) on NEC or sepsis supported by findings of the large Speed of Increasing milk Feeds Trial (SIFT) (¹⁷⁶) which compared daily increments of 18–30 mL/kg/d. The median enrolment age of 4 days in the SIFT trial may not adequately inform the relative safety of these increments at an earlier age. In addition, the actual daily increment was slower than targeted in both groups and other factors such as the approach and definition of feeding tolerance or gastric residuals (GR) might have influenced practice.

Conclusions and Recommendations

C1: After 4 days of life, faster feeding progression (30 mL/kg/d) of enteral feed volumes does not significantly increase the incidence of NEC or all-cause mortality compared to slower (15–20 mL/kg/d) feeding advancement. LOE 1+

C2: Meta-analysis showed that faster increment of enteral feed volumes positively reduces the time to full enteral feeding and the length of hospital stay as well as possibly the incidence of invasive infections. LOE 1+

R1: In stable preterm infants where the clinician considers that feed volume can be increased, a routine daily increment of 18–30 mL/kg/d is recommended, especially in breastmilk-fed infants. GOR A

Gastric Residuals

GR are commonly used to define feeding tolerance although there are few high-quality data and no adequately powered RCTs to determine the relationship with NEC. Gastric emptying is influenced by positioning of the infant and the type of enteral feed, with breastmilk emptied almost twice as fast as formula (^177,178) although this may differ following pasteurization (¹⁷⁹) or fortification (¹⁸⁰). There is no consistent, agreed definition of feeding intolerance, clinical practice varies widely, and RCTs have used different volume definitions. Evidence from relatively small studies suggest that routine monitoring of GR increases the risk of feed interruption episodes, the time taken to reach full enteral feeds and to regain birth weight, and PN days, but does not have an impact on NEC incidence (^181,182). There is no consensus on whether to re-feed or discard the aspirated GR.

Conclusions and Recommendations

C1: Positioning of the infant has an impact on gastric emptying with the prone position for the first half hour post feeding being quickest. LOE 2+

C2: The GR alone is neither a sensitive nor a specific indicator for bowel injury of the premature gut. LOE 2+

C3: Routine monitoring of GR increases the time taken to reach full enteral feeds and to regain birth weight and increases the number of PN days but does not have an impact on NEC incidence. LOE 2+

C4: There is no consensus on whether to re-feed or discard the gastric aspirate. LOE 3

R1: Routine monitoring of GR in the clinically stable infant is not recommended. GOR B

R2: Assessment of GR should be performed only when other clinical signs associated with feeding intolerance or NEC are present such as extreme abdominal distension, tenderness, emesis, bloody stools, apnea, and temperature instability. GOR B

Timeline of Parenteral and Enteral Nutrition

Most preterm infants receive PN, EN, and a transitional period in between influenced by local feeding practices, feeding intolerance, and metabolic intolerance (^183–185). The transition phase is a critical time period for poor growth (¹⁸⁴) although early progressive PN and EN strategies may lead to reductions in the cumulative energy and protein deficits that occur during the first weeks of life (^186,187). Standardized feeding guidelines and protocols, designed to maintain targeted intake throughout the transition phase (^188,189) can help to achieve nutritional goals (¹⁹⁰). Data from multiple observational studies suggest that the use of standardized feeding protocols allow preterm infants to achieve full enteral feeds faster, shorten the time on PN and hospital stay, decrease the rates of NEC, and improve growth and neurodevelopment (^5,191–198). A key challenge during the transition phase is to determine optimal intakes when both PN and EN are provided. The use of computerized software able to adapt to changing reference values may be helpful.

C1: Early progressive PN and EN strategies may reduce cumulative energy and protein deficits. LOE 2+

C2: The transition phase between PN and exclusive EN is a critical time period for cumulative nutrient deficits and for poor growth. LOE 2+

R1: To avoid nutrient deficits we recommend establishing a standardized feeding protocol in every Neonatal Intensive Care Unit (NICU) that defines the following parameters: duration of MEFs, daily advancement of milk feeds, definition and management of GR, definition, and approach to feeding intolerance, breast milk fortification strategy, and the definition of full enteral feedings. GOR B

FEEDING MODE: GASTRIC TUBE AND BOLUS OR CONTINUOUS FEEDING (see Supplementary digital content, Supplemental Digital Content 14, https://links.lww.com/MPG/C974)

Nasogastric Versus Orogastric Feeding Tubes

Orogastric (OG) and nasogastric (NG) feeding tubes are both used. NG tubes may increase nasal airway resistance especially in the smallest infants (¹⁹⁹), which may accelerate work of breathing and cause pharyngeal airway collapse (^200,201), although systematic reviews show no consistent effects on feed tolerance, incidence or frequency of apneas, desaturation episodes, or bradycardia (²⁰²). OG tubes may be more prone to vagal stimulation which may provoke bradycardia (^203,204) due to tube movements in the hypopharynx. Adverse effects of both NG and OG tube placement have been described, including tube misplacement (^205,206), nasal damage (²⁰⁷), and esophageal perforation.

Bolus Versus Continuous Feeding

Bolus feeds promote cyclical release of gastrointestinal tract hormones to stimulate gut maturation and motility (²⁰⁸) but marked variations in practice exist and many use continuous feeds (¹⁸³). Low-quality evidence suggests feeding 3-hourly is comparable to 2-hourly feeding although extremely low-birth-weight infants may reach full enteral feeds earlier when fed 2-hourly compared with 3-hourly (²⁰⁹). Bolus feeding increases splanchnic perfusion more than continuous feeding (²¹⁰). Energy expenditure may increase upon bolus feeding as compared to continuous feeding (²¹¹). Systematic reviews show longer time to reach full enteral feeding using continuous rather than intermittent feeding infants (²¹²) and fat loss may also be greater (^213,214) although there were no significant effects on growth (²¹²). Data on apnea are inconsistent (^215–219).

When to Start Oral (Breast)Feeding and Stop Tube Feeding

Infant oral feeding requires coordination of sucking, swallowing, breathing, and esophageal transport, but factors such as milk availability, NICU environment, and caregiver feeding approach are also important (²²⁰). Establishing oral feeding may be more challenging in high-risk infants, for example, those with BPD (²²¹) where micro-aspirations may compromise respiratory capacity further. Introducing oral feeds to infants on continuous positive airways pressure seems clinically safe based on 2 small studies and may decrease time to full oral feeds (^222,223), and studies also suggest feeding infants receiving high flow nasal cannula is possible (²²⁴) although swallowing dysfunction and risk of aspiration is important (²²⁵). There is no consistent evidence that early introduction and advancement of oral feeds based on the infant’s individualized cues, state, and behavior, rather than a predetermined feeding schedule, affects important outcomes for preterm infants or their families. Low quality evidence suggests preterm infants fed in response to feeding and satiation cues may achieve full oral feeding earlier (²²⁶). Meta-analyses provided evidence of low to moderate quality indicating that avoiding bottles increases breast feeding on discharge home.

Meta-analysis suggests that nonnutritive sucking reduces time to full oral feeding (²²⁷), and sensorimotor interventions may improve the sucking process (^228,229). Nipple shields may affect breastfeeding success, but reviews are contradictory (^230,231) and most do not advocate routine use (²³²).

Conclusions and Recommendations

C1: No preferential method to use either nasogastric or orogastric feeding tubes for preterm neonates can be determined. LOE 2

C2: Bolus feeding (2–3 hourly) might be slightly more preferential than continuous feeding in preterm infants, but more well-designed studies are needed for definitive advice. LOE 2+

C3: Establishment of nonnutritive sucking prior to the introduction of oral feeding may reduce time to reach full oral feeding and length of hospital stay. LOE 3

R1: Introducing oral feeding should be guided by the competence and stability of the preterm infant and may be started from 32 weeks postmenstrual age. GOR GPP

GROWTH (see Supplementary digital content, Supplemental Digital Content 15, https://links.lww.com/MPG/C974)

Growth is evaluated by measuring gain in weight, length, and HC but also using measures of body composition. Plotting on a growth chart enables clinicians to compare the growth trajectory of each infant to a reference group throughout infancy and childhood and is essential. Growth must be considered in the context of improving short- and long-term functional outcomes rather than simply promoting an increase in anthropometric values.

Growth standards for preterm infants are challenging to develop and many growth references are simply based on cross-sectional birth weight data. Using postnatal growth data is however challenging because many sick infants show altered growth and there are uncertainties of how best to define “healthy” preterm infants that might act as a standard. Growth velocity based on fetal ultrasound estimations can also be challenging, but fetal estimates can guide clinician’s evaluating growth in a stable preterm infant. Using, for example, World Health Organization (WHO) in-utero data, an average fetal weight gain of 20–23 g/kg/d during 23–25 weeks of gestation, decreasing to 17–20 g/kg/d during weeks 26–29, 13–17 g/kg/d during weeks 30–34, and 10–13 g/kg/d during weeks 35–37 can be guiding.

Growth faltering (GF) describes an infant whose growth slows and does not grow parallel to a centile during the period of established growth. GF is more common in sick infants and associations with poor neurodevelopmental outcomes may be confounded (²³³). Catch-up growth refers to accelerated rates of growth following a period of GF. However, there are concerns that rapid catch-up growth may increase the risk of cardiovascular and metabolic disease in later life especially when it is due to catch-up in weight without contemporaneous linear or head growth (²³⁴). There are no data that allow clinicians to determine the optimal degree or duration of catch-up growth in an individual infant.

During the first 3–4 days after birth, in appropriate for gestational age (AGA) infants, a weight loss is expected (7%–10%), mainly due to a 1-time irreversible contraction of extracellular water space (^235,236). Studies show small for gestational age (SGA) infants often lose less (4%–7%) (^237,238). A variety of approaches have been suggested to determine the optimal growth trajectory or target percentile for preterm born infants, for instance the pragmatic aim of not losing more than 1 Standard Deviation Score (SDS) in weight and HC from birth to discharge (²³⁹). However, calculating changes in SDS from birth to discharge is confounded by skewed reference data (²⁴⁰), and even more so for the most immature infants growing on lower centiles. It is not clear whether infants should regain their actual centile at birth, or whether a target centile should be based on weight at 1–3 weeks of age. Growth expectations of each infant have to be individualized since infant growth vary depending on genetic potential, intra uterine environment (eg, preeclampsia or poorly controlled diabetes), and morbidity in the NICU.

One concept is to avoid a large weight loss after birth, stabilize growth, avoid GF, approximately grow along a target centile, and adjust nutrient intakes gradually so that the preterm and term growth trajectories merge at around 44 weeks (²³⁶). Growth patterns in preterm SGA infants may differ (²⁴¹), and the optimal time frame and/or speed of catch-up growth are not known.

Conclusions and Recommendations

C1: Based on current evidence, the optimal growth velocity that optimizes outcomes in preterm infants remains unclear. LOE 2+

C2: According to WHO in-utero growth fetal weight gain decreases from slightly above 20 g/kg/d at 23–25 weeks of gestation to about 10 g/kg/d at term age. LOE2++

R1: Regular monitoring of weight, length, and HC growth is strongly recommended. Ideally, weight should be measured at least once or twice daily during the first 1–2 weeks, followed by measurements 2–3 times weekly in the stable growing phase. Length and HC should be measured once weekly unless clinical conditions (eg, hydrocephalus) indicate more frequent monitoring. GPP

R2: After a typical acceptable initial weight loss of 7%–10%, reaching a nadir at days 3–4, nutritional strategies should aim to regain birth weight by 7–10 days of age, followed by growth along a target centile and a gradual transition to the corresponding birth percentile on the WHO postnatal growth chart within the first weeks or months post term. GPP

R3: Nutritional management and growth assessment for infants born in-utero growth restriction and/or SGA should be the same as those born AGA, although initial weight loss is often less and acceptable up to 4%–7% of birth weight. GOR B

R4: Postnatal growth trajectories (weight, length, and HC) of each infant must be followed and evaluated to ensure nutrition is adequate; ideally using a growth chart based on a large robust dataset. GPP

R5: In infants experiencing postnatal GF, some catch-up growth should be allowed but rapid catch-up growth should be avoided. If catch-up growth is perceived as too rapid, ensure that nutrients are within recommended intake ranges and not excessive. GPP

R6: NICUs should adopt a standardized approach to the management of postnatal GF. If GF is recognized, ensure that nutrition is within recommended intake ranges. Careful consideration must be used in balancing the well-documented neurocognitive risks of nutrient deficiencies and slow growth in early life, against the theoretical risks from rapid catch-up growth and adverse metabolic programming in later life. GPP

BREAST MILK (BUCCAL COLOSTRUM, DONOR HUMAN MILK, AND PASTEURISATION OF MOTHER’S OWN MILK TO REDUCE CYTOMEGALOVIRUS TRANSMISSION) (see Supplementary digital content, Supplemental Digital Content 16, https://links.lww.com/MPG/C974)

Administration of buccal colostrum to preterm infants appears safe and theoretically attractive from both an emotional as well as immunological point of view, but no clear clinical benefits have consistently been proven in high-resource settings. There is therefore no current data to recommend routine administration of buccal colostrum to reduce morbidity or mortality, although there may be wider behavioral effects and other benefits (^242,243).

Fresh MOM contains higher amounts of macronutrients, and immunoactive and trophic factors than pasteurized MOM or donor human milk (DHM). Nevertheless, fortified pasteurized DHM instead of preterm formula may reduce NEC rates in preterm infants, whereas other neonatal morbidity and mortality rates are unaltered (^244,245). We strongly recommend MOM as the first choice of feeding in both preterm as well as term infants. In case of insufficient MOM availability, fortified DHM is conditionally recommended over preterm formula in preterm infants born <32 weeks’ gestation or with a birth weight <1500 g. When providing DHM, health care providers must continue to increase awareness of the benefits of MOM over both DHM and preterm formula and recognize the variable nutrient density of DHM.

The vast majority of cytomegalovirus (CMV) seropositive women undergo CMV reactivation during lactation and excrete CMV in their breast milk (^246,247). This leads to (sub)-clinical CMV transmission in approximately 15%–20% of very preterm infants, although rates may be higher in extremely preterm infants (^248,249). Symptomatic postnatal CMV infection, presenting as thrombocytopenia, cholestasis, or sepsis-like illness occurs in less than 5%–10% of infected infants (^248,249), whereas associations with BPD, NEC, and adverse neurological sequela are less clear (^250–252). While pasteurization will effectively eliminate CMV from breast milk, it also reduces or inactivates important bioactive factors. There is insufficient evidence to determine whether potential sequela of postnatal CMV transmission are more harmful than potential adverse effects arising from providing pasteurized MOM instead of fresh MOM (²⁵³). Thus, while we acknowledge the potential adverse consequences of postnatally acquired CMV, especially in the most immature infants, we do not recommend routinely pasteurizing MOM from CMV-positive women as this may reduce the beneficial effects of fresh MOM.

Conclusions and Recommendations

C1: While the administration of buccal colostrum to premature infants appears safe and theoretically attractive from both an emotional as well as immunological point of view, no clear clinical benefits for the infant have consistently been proven in high-resource settings. LOE 1-

C2: Fresh MOM contains higher amounts of macronutrients, and immunoactive and trophic factors than Holder-pasteurized DHM. LOE 1++

C3: Fortified pasteurized DHM instead of preterm formula milk reduces NEC rates in preterm infants, whereas other neonatal morbidity and mortality rates are similar. LOE 1+

C4: Most CMV-seropositive women undergo CMV reactivation in breast tissue during lactation and excrete CMV in their breast milk, which may cause (sub)-clinical CMV transmission in approximately 15%–20% of very preterm infants, although rates may be higher in extremely preterm infants. LOE 1+

C5: Symptomatic postnatal CMV infection, presenting as thrombocytopenia, cholestasis, or sepsis-like illness occurs in a minority of infected infants although associations with BPD, NEC, and adverse long-term neurological sequela are less clear. LOE 1-

C6: While Holder-pasteurization will effectively eliminate CMV from breast milk, it also reduces or destroys many beneficial and important bioactive factors. LOE 1++

C7: There is insufficient data to determine whether the potential sequela of postnatal CMV transmission are more harmful than potential adverse effects arising from providing pasteurized instead of fresh MOM. LOE 2+

R1: No recommendation can be made either for or against the use of buccal colostrum in preterm infants in order to reduce neonatal morbidities or mortality so parent preferences must be considered.

R2: We strongly recommend MOM as the first choice of feeding in both preterm as well as term infants. GOR A

R3: In case of insufficient MOM availability, fortified DHM is conditionally recommended over preterm formula milk in preterm infants born <32 weeks’ gestation or with a birth weight <1500 g. GOR B

R4: When providing DHM, health care providers must continue to increase awareness of the benefits of MOM over both DHM and preterm formula. Health care providers must support and facilitate mothers in order to promote higher rates and volumes of MOM provision (eg, through lactation consultants). GOR A

R5: While we acknowledge the potential adverse consequences of postnatally acquired CMV, especially in the most immature infants, there is insufficient evidence to recommend routine pasteurization of MOM from CMV-positive women as pasteurization simultaneously reduces the activity of many bioactive factors. GOR B

OSMOLARITY AND HYDROLYZED PROTEIN (see Supplementary digital content, Supplemental Digital Content 17, https://links.lww.com/MPG/C974)

Osmolality

More than 40 years ago an association between hypertonic infant formula and an increased incidence of NEC was noted (²⁵⁴) and the American Academy of Pediatrics recommended that the osmolarity of infant formula should not exceed 400 mOsm/L (approximately equivalent to an osmolality of 450 mOsm/kg) (²⁵⁵). However, recent systematic reviews have not found any consistent evidence that differences in feed osmolality in the range 300–500 mOsm/kg are associated with adverse gastrointestinal symptoms although study interpretation is challenging (^256,257).

Breast milk fortification increases the osmolality of the milk immediately after addition (up to more than 50%), followed by additional smaller increases in osmolality (up to 10%) after storage at 4°C for up to 24 hours (^{254,258–262}). Routine additives and medications can also significantly increase osmolality to levels that exceed guidelines from other professional bodies (^84,260,263). However, drugs as well as vitamin supplements often contain carrier molecules that can diffuse across membranes without increasing tonicity, and therefore are not likely to present a risk from their osmolar load. We consider it prudent to dilute additives in the largest possible volume of feed, to use multicomponent breast milk fortifiers in preference to multiple individual supplements, and to avoid simultaneous addition of multivitamins, electrolyte solutions, or other high osmolar substances where possible.

Hydrolyzed Protein

Hydrolyzed protein has increasingly been used in preterm infant milk formula and fortifiers. Protein hydrolysis alters amino acid kinetic, utilization by the gut, and may reduce nutrient utilization especially for nitrogen (¹⁰⁸) but these formulas are generally regarded as safe and no adverse effects on growth or development have been demonstrated in term infants. In preterm infants, RCTs show faster gastrointestinal transit with hydrolyzed protein formulas (^264–267) and improved nitrogen and mineral retention can be achieved with higher protein concentrations or due to other changes in formulation or production (²⁶⁸). Effects on growth (²⁶⁹), NEC, and feeding advancement vary in more recent studies (^270–272). Relatively high levels of “advanced glycation end-products” in most hydrolyzed formulas have been associated in other settings with development of chronic inflammatory, metabolic, or neurodegenerative diseases (^273,274) but the clinical relevance of these theoretical concerns is unknown. Finally, the more complex processing required to create safe hydrolyzed protein products adds substantially to costs.

Conclusions and Recommendations

C1: The available evidence does not allow the definition of an upper safety osmolality threshold for enteral feeding of preterm infants. LOE 2+

C2: Commercial ready-to-feed milk formula with an osmolality that is at the upper end of the intake range may create challenges for clinicians who want to use additional supplements (eg, iron, vitamins, sodium, etc) but avoid excess feed osmolality. LOE 4

C3: Commercial milk formula differ in the degree of protein hydrolysis (range of Dalton sizes) which may be associated with differing functional effects. LOE 3

C4: In preterm infants hydrolyzed protein formula accelerates gastrointestinal transit and enteral feeding advancement but there are no data to show routine use improves long-term outcome. LOE 1+

R1: Where supplements or other feed additives are given, these should be added to the largest possible volume of milk feed. GPP

R2: Where breastmilk fortification is required, multicomponent fortifiers should be used in preference to multiple individual nutrient supplements. GPP

R3: Hydrolyzed protein may be used for early enteral feeding in preterm infants if HM is not available. GOR B

SUPPLEMENTAL BIONUTRIENTS (see Supplementary digital content, Supplemental Digital Content 18, https://links.lww.com/MPG/C974)

Choline

Choline is a conditionally essential water-soluble nutrient with vitamin-like qualities and is found in a wide variety of foods including breastmilk and infant formula. Choline has multiple physiological functions including structural roles in cell membrane and myelin synthesis, cell signaling, neurotransmitter functions, and DNA methylation. Adults can produce choline in the liver, but de-novo synthesis in preterm infants may be limited. There are theoretical risks of toxicity as choline is metabolized to trimethylamine oxide, high levels of which may cause liver damage and are associated with cardiovascular disease in adults. EFSA recommends that infant formula contain a minimum choline concentration of 25 mg/100 kcal (EFSA 2014). No good studies exist, and no deficiency state has been described, but recent studies suggest higher intakes may be beneficial in preterm infants (^275–277).

Other Suggested Food Supplements for Preterm Infants

Lactoferrin, MFG membrane, nucleotides, inositol, HMOs, lutein, zeaxanthin, and bile salt-stimulated lipase have all been suggested to have health benefits for preterm infants but there is currently insufficient evidence to support the recommendation of any of these routinely as a food supplement for preterm infants (LOE varies 1++ to 4).

Conclusions and Recommendations

C1: Dietary intakes in preterm infants must include choline because de novo synthesis may be limited or compromised but defining minimum and maximum intakes is challenging due to a lack of RCTs in preterm infants. LOE 4

C2: There is no current evidence that preterm infants primarily fed with breastmilk benefit from routine choline supplements. LOE 4

R1: There are no data to support any change in the previous recommended daily intake of choline 8–55 mg/kg/d, although higher intakes appear safe. GPP

R2: Milk formula designed for preterm infants should include choline in a concentration designed to meet recommended intakes, but additional routine choline supplementation in preterm infants is not recommended. GPP

BREASTMILK FORTIFICATION (see Supplementary digital content, Supplemental Digital Content 19, https://links.lww.com/MPG/C974)

HM is the optimal source of nutrition, but both macro- and micronutrient density is insufficient to support optimal growth for most very preterm infants (see supplementary data, Supplemental Digital Content, https://links.lww.com/MPG/C974 for HM nutrient density). While data on nutrient accretion based on factorial methodology strongly support the use of breastmilk fortifiers there is only limited evidence from clinical trials (²⁷⁸). Multiple small studies have explored individual protein, fat, or carbohydrate supplements or multicomponent fortifier products. Most studies show slightly greater weight, length, and head gain, with no consistent adverse effects on NEC, but also no consistent data showing improvements in long-term neuro-developmental outcomes. Fortifiers will increase the osmolality of the milk feed (see earlier section) and risks bacterial contamination. It is not clear which subpopulations of preterm infants benefit most, or whether all very preterm infants should receive them routinely, and there are few data on the optimal time to commence fortifiers (²⁷⁹). In lower resource settings, a small number of studies have explored using milk formula powder rather than multicomponent fortifiers, but no studies are large enough to determine potential adverse effects including sepsis and NEC, and the nutrient composition of the resulting mixture will be suboptimal.

Most fortifiers provide approximately 1–1.1 g of additional protein per 100 mL although some provide more. Because protein content in HM decreases over the first 2–4 weeks, a standard regime of fortification may not be optimal, and DHM macronutrient concentrations may be lower than for MOM. Energy, PERs, and the proportion of energy provided as fat or carbohydrate also differ between studies making evidence synthesis challenging. While pooling of DHM will reduce macronutrient variability, this is not practical when using MOM. “Adjustable” fortification using serum urea concentrations (²²) may improve growth although methodological challenges exist and cut-off values for urea lack a robust evidence base. “Targeted” fortification describes various methods including analysis of macronutrient concentration at the cot-side (^280–282) but could also be utilized when donor milk banks provide data on nutrient concentrations of analyzed pooled milk. While earlier studies showed inconsistent or no effects on growth, more recent studies using better validated HM analyzers and individualized supplementation with protein, fat, and carbohydrate show promise (²⁸³).

There are strong observational data to show that the use of MOM is associated with a lower risk of NEC, and this also seems likely for DHM based on meta-analyses of RCTs, although no single adequately powered trial exists (²⁴⁴). It remains unclear whether DHM reduces NEC risk due to the presence of beneficial functional components (eg, HMOs) or whether protective mechanisms primarily involve decreasing exposure to bovine proteins or other components. HM-derived fortifiers are now commercially available either as concentrated liquids or lyophilized powders which make an exclusive HM diet possible. While a small number of studies suggesting benefit exist, no adequately powered trials to conclusively determine a reduction in NEC solely due to a human-milk derived fortifier have been performed, and some studies show slower growth in the human versus bovine-derived fortifier groups.

Conclusions and Recommendations

C1: The protein content of some fortifiers might be insufficient to increase protein concentrations to recommended intake levels if the volume of enteral feeds is limited. LOE 2

C2: The optimal time to start fortification is not clear, but early fortification seems to be as safe as delayed fortification, may reduce cumulative nutrient deficiencies, and positively influence bone metabolism. LOE 2+

C3: There is variation in the nutrient content of commercially available fortifiers and this may affect growth and health outcomes. LOE 2

C4: Adjustable and target fortification strategies may be employed to compensate for variation in HM macronutrient composition, but the optimal strategy is uncertain. DHM may require higher levels of fortification compared to MOM. LOE 2+

C5: Fortifiers derived from HM may reduce the risk of NEC but there are insufficient data from adequately powered studies to determine the optimal strategy. LOE 2+

R1: We recommend the use of multicomponent fortifier products to enhance the nutrient content of HM fed and to promote growth in preterm infants. GOR A

R2: We recommend starting fortifier when enteral intakes reach 40–100 mL/kg/d. GOR C

R3: Individualized fortification strategies including adjustable and targeted approaches may be appropriate. GOR A

R4: There is insufficient evidence to recommend the routine use of HM-derived fortifiers until further high-quality data is available. GOR C

Acknowledgment

We gratefully acknowledge the support of Maria Olsson, Pediatric Unit, Department of Clinical Sciences, Umea University, Sweden in manuscript preparation.