The goal of neonatal nutrition in the preterm infant is to achieve postnatal growth and body composition approximating that of a normal fetus of the same postmenstrual age( Reference Kleinman 1 ), and to obtain a functional outcome comparable with infants born at term( Reference Agostoni, Buonocore and Carnielli 2 ). Neonatal units (NU) attempt to achieve this by implementing nutrition policies incorporating growth assessment, but this has its challenges.
Firstly, the exact nutritional requirements of preterm infants (born < 35 weeks completed gestation) are not yet fully known, and current published nutrition recommendations( Reference Agostoni, Buonocore and Carnielli 2 – Reference Koletzko, Goulet and Hunt 5 ) are based on limited evidence and depend largely on expert opinion. Thus, there is an ongoing debate as to the validity of these recommendations. In recent years, mounting evidence proposes a more ‘aggressive’ approach to the nutritional management of preterm infants, with the aim of reducing nutrient deficits and postnatal growth failure( Reference Ehrenkranz 6 , Reference Poindexter, Langer and Dusick 7 ). However, aggressive nutrition and accelerated growth in infancy have been associated with the development later in life of an increased and aberrant adiposity, which is a marker of morbidity risk( Reference Euser, Finken and Keijzer-Veen 8 ).
Secondly, the current assessment of growth is predominately based on anthropometry, the measurement of weight, length and head circumference, with insufficient attention given to the quality of growth, in terms of fat mass and fat-free mass. Research to date has informed us that when preterm infants were assessed at term corrected age, they had an altered body composition when compared with term infants( Reference Roggero, Gianni and Amato 9 – Reference Cooke and Griffin 11 ). Furthermore, changes in an infant's growth pattern( Reference Lucas, Fewtrell and Cole 12 – Reference Singhal, Fewtrell and Cole 14 ) and body composition( Reference Dulloo, Jacquet and Seydoux 15 ) in early life may exert programming effects on disease risk in later life. The accurate and non-invasive measurement of body composition has been shown to be useful in assessing quality of growth.
To date, there is insufficient data assessing the adequacy of nutrient intake on growth and subsequent body composition, to provide clear, evidence-based nutrition guidelines for this vulnerable patient group. Studies assessing the adequacy of nutrient intakes after the implementation of nutrition guidelines, still focus on the rate rather than the quality of growth( Reference Senterre and Rigo 16 , Reference Senterre and Rigo 17 ). The assessment of the pattern of growth and changes in body composition in early infancy will enhance the knowledge of the nutritional requirements of preterm infants and provide evidence to inform future nutrition recommendations. The present paper provides a review of the current literature on the nutritional management and assessment of growth in preterm infants. It also explores several elements that may be essential for optimising nutrient intakes in preterm infants, such as measurement of body composition, collection of nutrient intake data in real-time and personalising nutritional support.
Nutrient requirements and recommendations
Nutrient requirements of preterm infants have been determined by two methods, the factorial method and the empirical method. The former derived requirements from accretion rates of nutrients derived from the analysis of fetal body composition at different stages of gestation( Reference Ziegler 18 ). The empirical method involved the manipulation of nutrient intakes and observation of the growth response, comparing actual energy/protein intakes with actual growth( Reference Ziegler 19 ).
Several expert groups have formulated international consensus guidelines for the nutritional management of preterm infants (Tables 1 and 2)( Reference Agostoni, Buonocore and Carnielli 2 – Reference Koletzko, Goulet and Hunt 5 ) that have allowed NU to introduce and develop nutrition policies to improve standards of nutritional care. The first set of recommendations on nutrition of the preterm infant was published by the European Society for Paediatric Gastroenterology, Hepatology and Nutrition in 1987, and provided guidance on feeding the preterm infant( 20 ). Published international nutrition guidelines are available in the book Nutrition of the Preterm Infant: Scientific Basis and Practical Guidelines, edited by Tsang et al.( Reference Tsang, Uauy, Koletzko, Tsang, Uauy, Koletzko and Zlotkin 4 ). These recommendations have recently been updated by Koletzko et al.( Reference Koletzko, Poindexter, Uauy and Koletzco 3 ). In Europe, the European Society for Paediatric Gastroenterology, Hepatology and Nutrition released guidelines on parenteral nutrition (PN) in 2005( Reference Koletzko, Goulet and Hunt 5 ), but unlike Tsang et al.( Reference Tsang, Uauy, Koletzko, Tsang, Uauy, Koletzko and Zlotkin 4 ), these guidelines give broad recommendations about PN requirements. They provide neither the specific guidance as to what daily prescriptions should be, nor the increment of PN in early postnatal life. In 2010, the European Society for Paediatric Gastroenterology, Hepatology and Nutrition published enteral nutrition guidelines( Reference Agostoni, Buonocore and Carnielli 2 ), which are consistent with, but not identical to, the recommendations from Tsang et al.( Reference Tsang, Uauy, Koletzko, Tsang, Uauy, Koletzko and Zlotkin 4 ). These guidelines propose advisable ranges for nutrient intakes for stable-growing preterm infants up to a weight of 1800 g. There are no specific recommendations for infants weighing <1000 g because data are lacking for most nutrients in this group; protein is the exception.
ELBW, extremely low birth weight infant (<1000 g); VLBW, very low birth weight infant (1000–1500 g).
* Reflects European recommendations.
† Reflects global recommendations.
‡ Days 2–7 indicate the period of metabolic and physiologic instability after birth and may last for up to 7 d.
ELBW, extremely low birth weight infant (<1000 g); VLBW, very low birth weight infant (1000–1500 g); BW, birth weight.
* Reflects European recommendations.
† Reflects global recommendations for infants with a BW up to 1500 g.
Although much progress has been made in the field of neonatal nutrition over the past few decades, the nutritional requirements of preterm infants are still not yet fully known and there are limitations to the current recommendations( Reference Agostoni, Buonocore and Carnielli 2 , Reference Tsang, Uauy, Koletzko, Tsang, Uauy, Koletzko and Zlotkin 4 , Reference Koletzko, Goulet and Hunt 5 ). Firstly, they are based on limited evidence and largely depend on expert opinion. Secondly, they are primarily birth weight (BW) based, and do not account for gestational age. Preterm infants are a heterogeneous population in terms of their nutritional and growth status, with those infants born early likely to have different nutritional needs than those born late, related to their more immature physiological development. Nutrient requirements cannot be consistent throughout gestation; thus, recommendations should take this into consideration. And thirdly, these recommendations are based on the needs for maintenance and growth and do not take into account the need for catch-up growth. The nutrient requirements of preterm infants born early have not been extensively examined, and there are no published studies stratifying infants by both BW and gestational age. More research is required to determine if recommended intakes should consider both gestational age and the need for catch-up growth, and not just BW.
The European Society for Paediatric Gastroenterology, Hepatology and Nutrition Committee on Nutrition recommends the implementation of multidisciplinary paediatric nutrition support teams in hospitals to screen patients for nutritional risk, identify patients who require nutritional support, and provide adequate nutritional management( Reference Agostoni, Axelson and Colomb 21 ). It has been shown that implementation of a multidisciplinary team that includes a registered dietitian improves nutritional outcomes of preterm infants in the NU( Reference Sneve, Kattelmann and Ren 22 ). In particular, involvement of registered dietitians in NU increases the intensity of important aspects of nutritional care( Reference Olsen, Richardson and Schmid 23 ). There is substantial evidence to support the role of nutrition guidelines in clinical practice with standardised feeding regimens suggested to be the single most important global tool to prevent necrotising enterocolitis in preterm infants( Reference Patole 24 ). In addition, improvements in nutrient intakes and growth are observed after the implementation of evidence-based nutrition support practices( Reference Senterre and Rigo 16 , Reference Rochow, Fusch and Muhlinghaus 25 , Reference Hanson, Sundermeier and Dugick 26 ).
Nutrient intake in preterm infants is divided into parenteral and enteral routes. Preterm infants are initially dependent on receiving nutrition parenterally due to immaturity of the gastrointestinal tract precluding the digestion and absorption of adequate nutrients, followed by the subsequent slow initiation and advancement of enteral nutrition until full enteral feeds are established.
Evidence base for parenteral nutrition guidelines
Conventional PN consists of an aqueous solution containing glucose, amino acids (AA) and electrolytes (± vitamins and trace elements) and a lipid emulsion (± vitamins) that are infused separately. PN can be prescribed on an individual basis (individualised PN) typically every 24 h, whereby nutrients (± acetate) are individually prescribed specific to each infant's requirements. Alternatively, standardised PN can be used, containing a fixed amount of nutrients that cannot be altered. More recently, some units have started to use concentrated standardised PN (fixed amount of nutrients in a low volume), which prevents nutrient intakes being compromised when fluid is restricted or while enteral feeds are introduced and advanced. It has been shown to be effective in optimising nutrient intakes in the PN-dependent period( Reference Morgan, McGowan and Herwitker 27 – Reference Mahaveer, Grime and Morgan 29 ), and also has the added advantage of being cheaper than formulating individual solutions( Reference Morgan, Badhawi and Grime 30 , Reference Yeung, Smyth and Maheshwari 31 ).
A minimum AA supply of 1·5–2 g/kg per d( Reference Tsang, Uauy, Koletzko, Tsang, Uauy, Koletzko and Zlotkin 4 , Reference Koletzko, Goulet and Hunt 5 ) on the first day of life should be provided to avoid catabolism, establish anabolism and promote linear growth. AA are generally advanced in a stepwise manner, and a maximum intake of 4 g/kg per d is recommended( Reference Tsang, Uauy, Koletzko, Tsang, Uauy, Koletzko and Zlotkin 4 , Reference Koletzko, Goulet and Hunt 5 ). Protein-to-energy ratios are important, and most authorities suggest 104·6–167·36 kJ (25–40 kcal) of non-protein energy is required per gram AA to promote lean mass accretion( Reference Cauderay, Schutz and Micheli 32 ). Recently, evidence has supported a more ‘aggressive’ approach for early AA initiation in preterm infants( Reference Ehrenkranz 6 , Reference Denne and Poindexter 33 ), with the initiation of 2–2·5 g/kg per d( Reference te Braake, van den Akker and Riedijk 34 , Reference Vlaardingerbroek, van Goudoever and van den Akker 35 ) immediately after birth suggested. This more aggressive approach not only prevents catabolism, but may also promote improved growth and neurodevelopmental outcomes( Reference Ehrenkranz 6 , Reference Poindexter, Langer and Dusick 7 ).
Lipids constitute not only an important source of energy due to their high-energy density, but also a source of essential fatty acids and long-chain PUFA. In a recent systematic review and meta-analysis by Vlaardingerbroek et al.( Reference Vlaardingerbroek, Veldhorst and Spronk 36 ), the initiation of lipids within the first 2d of life in preterm infants appeared to be safe and well tolerated. More recently, Vlaardingerbroek et al. ( Reference Vlaardingerbroek, Vermeulen and Rook 37 ) demonstrated that preterm infants tolerated 2–3 g/kg per d lipid administration starting at birth, with no increased incidence of adverse events in the short-term but possible long-term effects remain unknown.
Glucose is the major energy source and the most widely used intravenous carbohydrate for neonates because it is readily available to the brain. Intravenous glucose must commence as soon as possible after birth, with an initial minimum glucose infusion of 4–8 mg/kg per min to prevent hypoglycaemia( Reference Koletzko, Goulet and Hunt 5 ). Maximal glucose oxidation in preterm infants is 8·3 mg/kg per min (12 g/kg per d) after birth( Reference Koletzko, Goulet and Hunt 5 ).
Evidence base for enteral nutrition guidelines
The American Academy of Pediatrics( 45 ) recommends the use of mother's own milk, fresh or frozen, as the first choice in preterm infant feeding, and if mother's own milk is unavailable or is contraindicated, pasteurised donor milk is the recommended alternative. When neither mother's own milk or donor milk is available, preterm formula should be used( Reference Arslanoglu, Corpeleijn and Moro 38 ). There are several significant short- and long-term beneficial effects of feeding preterm infants human milk (HM), including lower incidence of sepsis and necrotising enterocolitis ( Reference Meinzen-Derr, Poindexter and Wrage 39 – Reference Sullivan, Schanler and Kim 41 ), improved feeding tolerance, and the faster achievement of full enteral feeds( Reference Vohr, Poindexter and Dusick 42 , Reference Vohr, Poindexter and Dusick 43 ).
It is generally acknowledged that HM cannot adequately support growth of preterm infants because it does not meet the requirements for many nutrients, most notably protein, calcium and phosphorus, and fortification is therefore required( Reference Schanler 44 , 45 ). In general, commercially available fortifiers contain protein, carbohydrate and/or fat, electrolytes, vitamins and minerals. Recently, it has been shown that the initiation of fortification with the first feed was well tolerated( Reference Tillman, Brandon and Silva 46 ). The most widely used fortification method involves adding a standard amount of fortifier to HM. However, there is now growing interest in individualising the nutrient fortification of HM to address each preterm infant's unique nutritional requirements and differences in HM composition( Reference Arslanoglu, Moro and Ziegler 47 , Reference Polberger 48 ). There are two models of individualisation: targeted fortification( Reference Polberger, Raiha and Juvonen 49 ) and adjustable fortification( Reference Arslanoglu, Moro and Ziegler 50 ). The concept of targeted fortification is that the HM is analysed periodically and a target nutrient intake, for instance, protein, is chosen according to the predefined requirements of preterm infants. The amount of fortifier added to reach the target intake is dependent on the protein content of the milk. The adjustable fortification method does not make any assumptions regarding an infant's protein requirements; protein intake is adjusted on the basis of the infant's metabolic response, evaluated through periodic determinations of blood urea nitrogen.
Enteral feeds are generally initiated within 24–72 h after birth. Minimal enteral feeds (<24 ml/kg per d) may be given for the first few days of life to promote gastrointestinal maturation and to reduce mucosal atrophy( Reference Neu 51 ). A recent systematic review demonstrated that slower feed advancement (<24 ml/kg per d) did not reduce the incidence of necrotising enterocolitis in preterm infants compared with faster rates of 25–35 ml/kg per d( Reference Morgan, Young and McGuire 52 ). Protocols in vitamin, trace element and mineral supplementation vary considerably amongst NU.
Nutritional concerns arising from current nutritional management
Provision of nutrition in the NU is often overlooked, as the effects of under- or overnutrition are not immediate. In addition, the response to inappropriate nutrient intakes is delayed due to limited nutritional feedback at the cot-side, with more acute issues such as cardiovascular and respiratory justifiably taking precedence.
Implementation of nutrition guidelines is challenging and gaps between nutrition guidelines and clinical practice have been extensively reported( Reference Lapillonne, Carnielli and Embleton 53 – Reference Klingenberg, Embleton and Jacobs 57 ), leading to cumulative nutrient deficits( Reference Embleton, Pang and Cooke 58 – Reference Dinerstein, Nieto and Solana 60 ) and inadequate growth( Reference De Curtis and Rigo 61 – Reference Hulst, Joosten and Zimmermann 64 ). A large discrepancy often exists between prescribed and actual nutrient intakes( Reference Turpin, Liu and Prinz 54 , Reference Lapillonne, Fellous and Mokthari 55 , Reference Grover, Khashu and Mukherjee 59 ). The causes of suboptimal nutrient intakes are multifactorial and partly iatrogenic. Reasons include ineffective PN prescribing practices due to fear of metabolic intolerance of PN constituents, nutritionally suboptimal PN weaning protocols, restricted fluid volumes to minimise morbidities related to fluid overload such as patent ductus arteriosus, evolving neonatal chronic lung disease, and feeding intolerance associated with immaturity, sepsis and necrotising enterocolitis ( Reference Corpeleijn, Vermeulen and van den Akker 65 ). In addition, most nutritional studies do not analyse the macronutrient content of HM, and published values( 66 ) are used to calculate intakes, leading to possible inaccuracies in the estimation of nutrient intakes arising from the HM component of the total nutrient supply. The analysis of HM should be a prerequisite for future nutritional studies.
More recent observations have revealed adverse effects from the enhanced nutritional management of preterm infants, especially to extremely low BW infants( Reference Moltu, Strommen and Blakstad 67 , Reference Blanco, Gong and Schoolfield 68 ). Early and high-dose (4 g/kg per d) AA in the first week of life have been reported to impact negatively on growth and neurodevelopment( Reference Blanco, Gong and Schoolfield 68 ), and increase the incidence of electrolyte disturbances, that is, hypophosphataemia and hypokalaemia( Reference Senterre and Rigo 17 , Reference Moltu, Strommen and Blakstad 67 , Reference Rigo, Marlowe and Bonnot 69 , Reference Bonsante, Iacobelli and Chantegret 70 ). This is possibly due to high AA intakes inducing a progressive depletion of phosphate and potassium from accelerated protein synthesis( Reference Jamin, D'Inca and Le Floc'h 71 ). These findings emphasise the need to undertake preliminary analysis and testing of novel nutritional strategies to optimise nutrient intakes in preterm infants prior to their implementation in intervention studies, due to the risk of unintended adverse effects. Furthermore, real-time nutrient data collection in the NU could play an important role in allowing nutrient deficits or excesses to be promptly identified and responded to in real-time, at the cot-side, to avoid these undesirable effects. The focus of future research should be to develop a software tool that will collect real-time nutritional data at the cot-side to enable the assessment and monitoring of nutrient intakes in preterm infants.
Assessment of growth
It is clear that the goal of nutritional management of preterm infants should be to optimise quantitative and qualitative rates of growth to limit long-term morbidity and enhance long-term outcomes. The adequacy of nutrient intakes among infants is currently monitored by changes in weight gain, length and head circumference. Serial measurements of length and head circumference are important as they are better indicators of true growth, rather than weight alone, which may fluctuate due to changes in fluid balance rather than adipose or lean tissue mass. Whilst these measurements provide an important tool for assessing growth of infants, they do not provide information on the quality of growth achieved. The accurate and non-invasive measurement of infant body composition has been shown to be useful in assessing the quality of growth. Over the past two decades, the applicability of air displacement plethysmography for the assessment of human body composition has been developed( Reference Dempster and Aitkens 72 , Reference Fields, Gunatilake and Kalaitzoglou 73 ), and is now the preferred method for paediatric measurements( 74 ).
The infant should be weighed nude, preferably at the same time of day, on a regularly calibrated electronic scale which is recorded to the nearest 10 g. Weight may need to be measured daily to assist fluid and electrolyte management, and to provide an index to daily growth. Measurements should be plotted weekly on an appropriate growth chart. After birth, contraction of the extracellular fluid results in postnatal weight loss reported to be between 7 and 20 % of BW during the first 3–5 d( Reference Ehrenkranz 6 , Reference Fusch, Jochum, Tsang, Uauy, Koletzko and Zlotkin 75 ). This weight loss can be further contributed to by catabolism of endogenous glycogen, fat stores and lean tissue if adequate nutrition is not provided. The smallest infants tend to have the largest loss related to their higher body water composition and thinner epidermis. BW should be regained by 14–21 d of life( Reference Ehrenkranz, Younes and Lemons 62 , Reference Shaffer, Quimiro and Anderson 76 , Reference Wright, Dawson and Fallis 77 ). More recent studies evaluating the impact of optimisation of early postnatal nutrition in very low BW infants have demonstrated that BW can be regained as early as 7 d (n 102)( Reference Senterre and Rigo 16 ) and 12 d (n 123)( Reference Rochow, Fusch and Muhlinghaus 25 ). It has been proposed that the earlier recovery from initial weight loss during the first days of life appears to be key for optimising growth in extremely preterm infants, as later catch-up requires a higher growth rate that would be difficult to achieve in most infants( Reference Rochow, Fusch and Muhlinghaus 25 ).
Length measurement compared with weight measurement more accurately reflects lean tissue mass accretion, and is not influenced by fluid status and is therefore, a better indicator of long-term growth. Length should be monitored weekly and plotted on an appropriate growth chart. Accurate length measurements require two examiners, one holding the infant's head, and the other holding the infant's legs, and the average of two measurements taken. To obtain the measurement, the infant should be placed on a flat surface in a supine, fully extended position with knees straightened, and feet at right angles to the body. Plastic, recumbent length boards, for instance, the Leicester Incubator Measure (Harlow, UK) allows body length to be measured in the incubator, to the nearest 1 mm, thereby increasing the accuracy of measurements compared with the use of a measuring tape. An incremental gain in crown to heel length of approximately 0·9–1·1 cm/week should be expected( Reference Ehrenkranz, Younes and Lemons 62 , Reference Lubchenco, Hansman and Boyd 78 ).
Head circumference is measured to the nearest 1 mm with a non-stretch measuring tape at the maximal occipitofrontal circumference. Head circumference should be measured weekly, and the average of two measurements taken, and plotted on an appropriate growth chart. More frequent measurements may be indicated for infants with micro- or macrocephaly or suspected abnormal increases in head circumference. Head growth may remain normal despite inadequate postnatal nutrition( Reference Ehrenkranz, Younes and Lemons 62 ). During the first postnatal week, head circumference may decrease by about 0·5 cm due to extracellular fluid space contraction. A growth rate of approximately 0·9 cm/week is the goal for head circumference( Reference Ehrenkranz, Younes and Lemons 62 ).
Anthropometric measurements should be plotted on an appropriate growth chart. They provide the basis for growth and nutritional assessment of infants by presenting a comparison of an infant's actual size and growth trajectory with reference data. In the absence of a prescriptive growth chart depicting the growth of preterm infants under optimal conditions, monitoring postnatal growth of preterm infants is complicated, and there is a lack of global consensus on what is the most appropriate growth reference to use. BW growth charts are the mainstay for monitoring growth in preterm infants( Reference Tudehope, Gibbons and Cormack 79 ), and they include the WHO( Reference de Onis, Garza and Onyango 80 ) and UK-WHO growth chart( Reference Cole, Williams and Wright 81 ), the CDC (Centers for disease control and prevention) growth chart( Reference Ogden, Kuczmarski and Flegal 82 ), and more recently, the Fenton growth chart( Reference Fenton and Kim 83 ). Establishing a consensus regarding the most appropriate growth reference to use would be an important component in the standardisation of care for preterm infants, and would allow comparisons to be made between institutions and studies.
Body composition measurement
Quality of growth can be assessed by measuring an infant's body composition, which is calculated from body density (body density = body mass/body volume). The air displacement plethysmography methodology is used to obtain a measurement of the infant's body volume, which is used with body weight to determine total body density. This, in turn, is used with the basic two-compartment model of fat mass and fat-free mass to calculate the body's percentage of fat. This technique uses commercial equipment such as a device called the PEA POD (COSMED USA, Inc., Concord, CA, USA) (Fig. 1), which has been validated for use in infants 1000–8000 g( Reference Ellis, Yao and Shypailo 84 – Reference Urlando, Dempster and Aitkens 86 ). The description and operation of the PEA POD are reported elsewhere( Reference Ellis, Yao and Shypailo 84 , Reference Ma, Yao and Liu 85 ). The PEA POD is a portable device that can be used at the infant's bed side. The test chamber is temperature-controlled and a complete analysis takes about 5 min. Validation of the PEA POD has been performed against the deuterium dilution method and a reference four-compartment model for the estimation of infant body composition( Reference Ellis, Yao and Shypailo 84 ). It was found to be accurate and precise, with excellent within-day and between-day reliability( Reference Ellis, Yao and Shypailo 84 ).
To ensure optimal growth and body composition is achieved in preterm infants, their nutritional management should be personalised to meet their individual needs according to their gestational age, BW and their need for catch-up growth. The development and implementation of responsive, personalised nutritional support in preterm infants is required. This should utilise real-time nutrient intake data collection, with ongoing nutritional assessments that includes the measurement of body composition.
I would like to acknowledge the contribution of Ms. Sarah Fenton, senior pharmacist for her advice.
Conflicts of Interest
A. M. B. wrote the manuscript and A. M. B. , B. P. M. and M. E. K. approved the final content.