Introduction
It is estimated that 10% of all births are preterm, and this incidence is increasing worldwide.1 Every year, over 1 million infants die because of complications associated with preterm birth, which is the leading cause of neonatal mortality worldwide and the second leading cause of death among children under 5 years of age.1 Preterm babies are also frequently born small-for-gestational-age, and there is a large body of epidemiological evidence indicating that reduced size at birth is linked to increased risk of developing a number of diseases in later life. These include hypertension,Reference Law, Shiell and Newsome2 impaired glucose tolerance,Reference Newsome, Shiell and Fall3 obesityReference Kensara, Wootton and Phillips4, Reference Ong5 and coronary heart disease.Reference Roseboom, van der Meulen and Osmond6, Reference Eriksson, Forsen, Tuomilehto, Osmond and Barker7
Earlier studies focused on reduced size at birth in babies born at term, which was assumed to result from impaired foetal growth. However, more recent studies suggest that preterm birth, independent of foetal growth rate and even at late preterm gestations, is itself associated with increased blood pressure,Reference Crump, Winkleby, Sundquist and Sundquist8 altered blood pressure regulation,Reference Bayrakci, Schaefer, Duzova, Yigit and Bakkaloglu9 diabetes and insulin resistance,Reference Lawlor, Davey Smith, Clark and Leon10–Reference Rotteveel, van Weissenbruch, Twisk and Delemarre-Van de Waal14 as well as ischaemic heart disease and strokeReference Koupil, Leon and Lithell15 in later life. The adverse short- and long-term neurological and social outcomes of preterm birth are well known.Reference Saigal and Doyle16, Reference Moster, Lie and Markestad17 Nonetheless, it is becoming increasingly apparent that the adverse metabolic and neurodevelopmental outcomes following preterm birth not only affect those born at the extremes of viability but appear to be present as a graded response across preterm gestations, right up to and including late preterm births.Reference Crump, Winkleby, Sundquist and Sundquist8, Reference MacKay, Smith, Dobbie and Pell18
Both infants born preterm and those who experience intrauterine growth restriction (IUGR) are exposed to early environmental stress and suboptimal nutrition. This insult becomes apparent clinically during the later stages of pregnancy in IUGR infants, whereas for those born preterm this nutritional restriction occurs early in postnatal life, which is comparable to the third trimester of pregnancy.Reference Rowe, Derraik, Robinson, Cutfield and Hofman19 However, preterm infants are also much more likely to experience poor growth in utero, and those infants born extremely preterm (24–32 weeks) appear to grow poorly as early as the first trimester of pregnancy.Reference Smith, Smith, McNay and Fleming20 As a result, the rate of IUGR is much greater among infants born preterm than those born at term,Reference Cooke21 and it has been estimated that 40% of preterm babies have also experienced IUGR.Reference Cooke22 The origins of a significant proportion of IUGR, such as that secondary to placental vascular disease, also lie in early pregnancy.
The causes of preterm birth are multi-factorial;Reference Bloomfield23 however, irrespective of the underlying causes, most infants born preterm also experience suboptimal postnatal growth. Further, although this suboptimal growth is often most profound in the immediate postnatal period,Reference Ehrenkranz, Dusick and Vohr24 it may also persist through childhood and adolescence. After the neonatal period, this suboptimal growth may be in part a result of alterations in the endocrine regulation of growth, which appears to be disrupted in preterm children and adolescents, so that growth is no longer associated with its normal hormone regulators.Reference Rowe, Derraik, Robinson, Cutfield and Hofman19 However, a major contributor to postnatal faltering growth in infants born preterm is nutritional disruption due to inadequate nutrient intake.Reference Embleton, Pang and Cooke25 It is common practice in neonatal units worldwide to increase intravenous nutrition incrementally in preterm infants over the 1st week of life and to introduce enteral feeds cautiously. When compared with estimated placental nutrient transfer in utero,Reference Embleton, Pang and Cooke25 there is an inevitable and substantial nutrient deficit, particularly for nitrogen, by the end of the 1st week.Reference Thureen, Melara, Fennessey and Hay26 This deficit occurs even if the latest recommended nutritional intakes are met, and contributes to early postnatal growth faltering. There are potentially serious consequences for the preterm infant, as postnatal malnutrition not only hinders somatic growth but may also adversely affect organ growth, structure and function.Reference Cooke22
Poor early postnatal growth is likely to be associated with adverse neurodevelopmental outcomes,Reference Ehrenkranz, Dusick and Vohr24, Reference Latal-Hajnal, von Siebenthal, Kovari, Bucher and Largo27 but this relationship may be modifiable by other factors such as breastfeeding. In a recent large population study, infants born very preterm who were breastfed at the time of hospital discharge had significantly better neurodevelopmental outcomes than those who did not receive mothers’ milk, despite an increased risk for poor early postnatal weight gain.Reference Rozé, Darmaun and Boquien28 Thus, both early diet and early growth appear to be important mediators of later developmental outcomes.Reference Rozé, Darmaun and Boquien28 However, it is important to appreciate that, in studies on breastfeeding, there are often confounding factors; in these studies, <20% of babies were breastfed at discharge, and maternal demographics were very different among breastfed and non-breastfed babies.
Cause and effect
Although preterm birth itself is associated with adverse short- and long-term health outcomes, it is not yet clear whether this relationship is causal. Rather, there is evidence that reduced foetal growth, preterm birth and the long-term health effects of both of these may all arise from a suboptimal intrauterine environment. In addition, there is some evidence that decreased foetal growth and shorter gestation length may also be linked to common antecedents, suggesting that further research into this area is warranted. In humans, for example, poor maternal nutrition before pregnancy or in the periconceptional period is associated with an increased risk for preterm birth.Reference Bloomfield23 Mothers who are underweight before pregnancy are also more likely to deliver preterm.Reference Han, Mulla, Beyene, Liao and McDonald29, Reference Smith, Shah, Pell, Crossley and Dobbie30 Twins are born both early and small,Reference Muhlhausler, Hancock, Bloomfield and Harding31 outcomes that seem to have their origin in early pregnancyReference Hancock, Oliver, McLean, Jaquiery and Bloomfield32 and which appear to be linked.Reference Hediger, Luke and Gonzalez-Quintero33 Further, poor maternal nutrition is also associated with poor long-term health in the offspring.Reference Painter, Roseboom and Bleker34
Animal studies have provided more conclusive evidence of the causal relationships involved. In sheep, maternal undernutrition around the time of conception led to preterm delivery in a proportion of animals.Reference Bloomfield, Oliver and Hawkins35 Maternal undernutrition before conception also impaired the normal development of insulin resistance during pregnancy, which in turn was related to impaired foetal growth.Reference Jaquiery, Oliver, Rumball, Bloomfield and Harding36 However, following periconceptional undernutrition, even the lambs born at term had evidence of disrupted regulation of early postnatal growth,Reference Jaquiery, Oliver, Bloomfield and Harding37 and in adulthood had altered body composition,Reference Jaquiery, Oliver, Honeyfield-Ross, Harding and Bloomfield109 impaired glucose toleranceReference Todd, Oliver, Jaquiery, Bloomfield and Harding38 and blunted adrenal response to corticotrophic stimulation.Reference Oliver, Bloomfield and Jaquiery39 These data suggest that nutritional factors before and around the time of conception, rather than preterm birth itself, may be the primary cause of subsequent adverse outcomes.
Interventions
At an equivalent gestational age, a preterm infant has a very different nutritional intake from that supplied by the placenta to a foetus. During the second and the beginning of the third trimester of pregnancy, the estimated placental transfer to the foetus is 3.6–4.8 g/kg.d of protein and 12–14 g/kg.d of glucose.Reference Hay, Lucas and Heird40In utero, protein uptake by the foetus exceeds the amount needed for protein accretion, and the excess is oxidised to produce energy. The total foetal requirement for fatty acids at mid-gestation is ∼1 g/kg.d, suggesting that oxidation of fatty acids for energy is relatively unimportant in foetal life.Reference Hay, Lucas and Heird40 In comparison, guidelines for intravenous nutrition in preterm infants recommend high intakes for both fat (3.0–4.0 g/kg.d) and carbohydrate (up to 18 g/kg.d), with a much lower protein content (1.5–4.0 g/kg.d) than the estimated placental transfer.Reference Tsang, Uauy, Koletzko and Zlotkin41, Reference Koletzko, Goulet and Hunt42 However, in clinical practice, these nutrient intakes are difficult to meet with current intravenous nutrition solutions and delivery systems. In reality, many preterm infants receive very low protein intakes in the first few days after birth,Reference Lapillonne, Fellous, Mokthari and Kermorvant-Duchemin43, Reference Kiefer, Wickremasinghe and Johnson44 resulting in large deficits in energy and protein.Reference Ahmed, Irwin and Tuthill45, Reference Stewart, Mason, Smith, Protopapa and Mason46 Further, the amount and ratio of macronutrients given intravenously to preterm infants differ from the estimated in utero supply. These differences may promote the accretion of fat mass rather than lean body mass and may explain some of the observed differences in the body composition between preterm and term babies.Reference Kashyap, Forsyth and Zucker47–Reference Schulze, Stefanski and Masterson49
Higher protein and energy intakes in preterm infants during the first few days after birth reduce growth faltering at discharge.Reference Wilson, Cairns and Halliday50–Reference Rochow, Fusch and Muhlinghaus55 Other studies report favourable effects of earlier and/or higher total protein intake on the incidence of chronic lung disease,Reference Radmacher, Rafail and Adamkin56–Reference Ehrenkranz, Das and Wrage58 although most studies found no differences in clinical outcomes.Reference Wilson, Cairns and Halliday50, Reference Ibrahim, Jeroudi, Baier, Dhanireddy and Krouskop59–Reference Tan and Cooke61 Nonetheless, a nutritional approach that matches the estimated foetal macronutrient intakes has yet to be developed. Furthermore, protein intake is not the only determinant of growth, as an adequate balance of macronutrients and micronutrients is crucial for optimal growth and development in early postnatal life.Reference Wahlqvist and Tienboon62
Breast milk is widely recognised to be the best enteral feed, but it is inadequate to sustain postnatal growth of preterm babies at intrauterine growth rates. Importantly, the concentration of protein in expressed breast milk appears to be even lower than previously estimated.Reference de Boo and Harding63, Reference Arslanoglu, Moro and Ziegler64 Consequently, commercial breast milk fortifiers are commonly added to supplement the infant diet with additional quantities of protein and several other nutrients.Reference Tsang, Uauy, Koletzko and Zlotkin41, Reference Agostoni, Buonocore and Carnielli65 However, even with breast milk fortification, recommended enteral protein intakes are seldom achieved.Reference Henriksen, Westerberg and Ronnestad66–Reference Cormack and Bloomfield68 In addition, consensus recommendations for enteral protein intake were recently increased to 4.0–4.5 g/kg.d, resulting in the reformulation of several commercial breast milk fortifiers to increase the supplementary amount of added protein to 1.0–1.2 g per 100 ml breast milk and 2.5 g per 100 ml in some preterm formulae.Reference Agostoni, Buonocore and Carnielli65
There is growing evidence that manipulation of the macronutrient composition of milk leads to changes in weight gain and proportionality in the infant.Reference Grote, von Kries and Closa-Monasterolo69, Reference Mennella, Ventura and Beauchamp70 Systematic reviews report that both protein and multi-component breast milk fortification increase short-term growth parameters.Reference Kuschel and Harding71, Reference Kuschel and Harding72 However, the level of protein fortification in the studies included in these reviews was lower than that supplied by newly available fortifiers, and the long-term effects of higher protein fortifiers and formulas are yet to be determined.
There is evidence that postnatal faltering growth is associated with adverse neurodevelopmental outcomes,Reference Ehrenkranz, Dusick and Vohr24, Reference Latal-Hajnal, von Siebenthal, Kovari, Bucher and Largo27 and improved growth over the first few months after birth has been shown to be associated with better outcomes at 18 months of age.Reference Lucas, Morley and Cole73 There is also evidence that nutrient-enriched formula provided to preterm infants after discharge may lead to improved later development.Reference Lucas, Morley and Cole73–Reference Jeon, Jung and Koh75 A randomised trial of nearly 300 small-for-gestational-age infants born at term also showed significant gains in length and head circumference on an enriched formula.Reference Fewtrell, Morley and Abbott76 However, this trial reported adverse developmental outcomes at 9 months of age, although these seem to have disappeared at 18 months.Reference Morley, Fewtrell and Abbott77
Further investigation is also required into the long-term effects of other dietary components in preterm neonates. For example, supplementation of breast milk with fat alone has little effect on postnatal growth.Reference Kuschel and Harding78 However, additional long-chain polyunsaturated acids provided in breast milk via maternal supplementation are associated with improved problem-solving subscores (but not overall developmental scores) at the age of 6 months,Reference Henriksen, Haugholt and Lindgren79 and in the Bayley mental development index at the age of 18 months in girls but not in boys.Reference Makrides, Gibson and McPhee80 A subgroup of children enrolled in this trial underwent an assessment of language skills, behaviour and temperament in early childhood (2–5 years of age); there were no significant differences between those who received docosahexaenoic acid supplementation and those who did not.Reference Smithers, Collins and Simmonds81 Data on possible later cardiovascular effects of these interventions are not yet available.
Similarly, some neonatal units routinely prescribe calcium supplementation for very preterm babies to facilitate appropriate bone growth. There are few good data to support this;Reference Kuschel, Harding and Kumaran110 small trials have not found a benefit in bone mineral density by hospital discharge.Reference Wauben, Atkinson, Grad, Shah and Paes111 Supplementation of breast milk with a multi-nutrient fortifier for 12 weeks after discharge did result in higher calcium and phosphate intakes and increased skeletal growth but not bone density.Reference Aimone, Rovet and Ward82 However, supplementation reduced milk intake in these infants, so that protein and energy intakes and body composition at the end of the intervention period were comparable between supplemented and unsupplemented infants. Nonetheless, it is conceivable that neonatal calcium intake may improve later cardiovascular outcomes, as maternal calcium supplementation during pregnancy has been shown to be associated with lowered blood pressure in the offspring.Reference Belizan, Villar and Bergel83
In lambs, ewe milk fortification (with a nutritional supplement analogous to breast milk fortifiers used clinically) for the first 2 weeks of life led to gestation- and sex-specific effects on early growth. For example, supplements altered growth in male but not female lambs, and resulted in increased weight-for-length (and by inference adiposity) at weaning in lambs born preterm but decreased adiposity in lambs born at term.Reference Berry84 Some studies in humans have also shown differential effects in girls and boys following nutritional supplementation early in life.Reference Lucas, Fewtrell and Morley85, Reference Waber, Vuori-Christiansen and Ortiz86
Importantly, nutritional supplementation given in the neonatal period may also affect later endocrine function. In sheep, for example, early nutritional supplementation altered glucose–insulin axis function in females, independent of growth.Reference Berry84 Further, although supplementation reduced the insulin-secretory response to both glucose and arginine in female lambs born preterm, the response was increased in females born at term. In humans, diet in the neonatal period has also been shown to influence later glucose–insulin axis function in babies. When term infants of diabetic women were fed on either maternal milk or banked non-diabetic breast milk, those with the greatest exposure to ‘diabetic’ milk in the 1st (but not 2nd) week after birth had the greatest risk of glucose intolerance at 2 years,Reference Plagemann, Harder, Franke and Kohlhoff87 irrespective of the total duration of breastfeeding.Reference Rodekamp, Harder and Kohlhoff88 The enteroinsular axis is known to be active at birthReference Gentz, Persson, Kellum, Bengtsson and Thorell89 and to be modified by both maturityReference Padidela, Patterson, Sharief, Ghatei and Hussain90 and diet.Reference Lucas, Boyes, Bloom and Aynsley-Green91 It is therefore plausible that exposure to specific dietary components may modify pancreatic development in the newborn period. In addition, preterm infants have immature gastrointestinal tract function and experience delayed initiation of enteral feeding, as well as considerable variation in the type, composition and method of administration of milk feeds in the early neonatal period,Reference Hans, Pylipow, Long, Thureen and Georgieff92 all of which may have significant implications for later metabolic function.
Apart from nutritional interventions, common neonatal morbidities and their treatment can also affect nutritional intake and growth. For example, in a prospective birth cohort from the United States, delivery by Caesarean section at term was associated with a twofold increased risk of obesity at 3 years of age.Reference Huh, Rifas-Shiman and Zera93 In preterm infants, corticosteroids are used to treat ventilator-dependent chronic lung disease but can significantly impair growth in preterm babies.Reference Bloomfield, Knight and Harding94 Hyperglycaemia may affect over 50% of extremely low-birth-weight infantsReference Hays, Smith and Sunehag95 and is associated with increased mortality and morbidity.Reference Hays, Smith and Sunehag95, Reference Alexandrou, Skiöld and Karlén96 Hyperglycaemia in these babies is often managed by reducing glucose intake,Reference Alsweiler, Kuschel and Bloomfield97 which may further impair growth. A recent clinical trial found that tight glycaemic control with insulin in hyperglycaemic preterm babies improved weight gain and mitigated the expected reduction in head circumference, suggesting a possible improvement in brain growth.Reference Alsweiler, Harding and Bloomfield98 However, treatment reduced linear growth, suggesting an increase in fat mass. In addition, tight glycaemic control increased the risk for hypoglycaemia,Reference Alsweiler, Harding and Bloomfield98 which is associated with poor neurodevelopmental outcomes in preterm babies.Reference Lucas, Morley and Cole99
Potential risks and benefits
The majority of infants born preterm undergo accelerated growth after discharge, so that 85% reach the normal height range by 2 years of age.Reference Knops, Sneeuw and Brand100 Nonetheless, nutritional interventions are commonly introduced to further accelerate their growth, even though the long-term benefits and costs of such interventions are still unclear. In fact, a number of studies have suggested that accelerated growth in early childhood may actually be detrimental to health in the long-term.
For example, in a cohort of very preterm babies, those with growth rates in the highest tertile over the first 2 years of life had lower insulin sensitivity and higher blood pressure as young adults, compared with those in the lowest tertile.Reference Rotteveel, van Weissenbruch, Twisk and Delemarre-Van de Waal14 Further, babies randomised to receive an enriched formula who exhibited accelerated neonatal growth had higher systolic blood pressure at 6–8 years of age than did babies fed a standard formula.Reference Singhal, Cole and Fewtrell101 Accelerated neonatal growth for as little as 2 weeks after birth was associated with higher levels of a marker of insulin resistance in adolescence,Reference Singhal, Fewtrell, Cole and Lucas102 as well as with alterations in vasodilatation suggestive of impaired cardiovascular health.Reference Singhal, Cole, Fewtrell, Deanfield and Lucas103 The US Collaborative Perinatal Project studied nearly 56,000 pregnancies and showed that an increase in growth percentile during the first 7 years was associated with increased blood pressure.Reference Hemachandra, Howards, Furth and Klebanoff104 A review of 21 studies identified that early accelerated growth (involving an increase of at least 0.67 weight SDS) was associated with a twofold to threefold increase in risk of overweight or obesity.Reference Ong and Loos105, Reference Ong106 In addition, accelerated growth was associated with a greater likelihood of developing insulin resistance and subsequently the metabolic syndrome.Reference Ong106 Thus, these findings have led some authors to question the wisdom of encouraging rapid weight gain in infants born preterm or IUGR.Reference Hemachandra, Howards, Furth and Klebanoff104
In contrast, other studies have observed beneficial effects of accelerated growth on neurodevelopmental outcomes. Preterm babies randomised to an enriched formula within 48 h of birth and fed this formula for an average of <6 weeks showed a one-third reduction in the risk for cerebral palsy in childhood.Reference Morley, Fewtrell and Abbott77 Further, breastfed preterm babies supplemented with human milk fortifier for 6 weeks had accelerated growth with no adverse effects on neurodevelopment.Reference Lucas, Fewtrell and Morley107 Observational studies have also reported that preterm babies who have higher growth rates in the neonatal period have higher neurodevelopmental scores in later life.Reference Ehrenkranz, Dusick and Vohr24, Reference Franz, Pohlandt and Bode108
The relationship between accelerated postnatal growth and long-term cognitive and metabolic health is difficult to tease out from the underlying causes of accelerated growth itself. It is also important to take into account the end result of accelerated growth for an individual, such as the relationship between final and starting growth (expressed as a Z-score or centile). Accelerated growth usually follows a period of faltering growth, whether this is before birth in IUGR babies, after birth in babies with postnatal growth faltering as for most very preterm babies, or both before and after birth as may occur in preterm babies born IUGR. There are currently no proven interventions for improving intrauterine growth in IUGR foetuses, suggesting that attention should be focussed on preventing the faltering postnatal growth that is so prevalent in babies born preterm.
Conclusions
Preterm birth is known to be associated with numerous short- and long-term adverse health outcomes. However, it is unclear whether these outcomes are a direct consequence of preterm birth itself or of associated factors, such as perturbed antenatal and/or postnatal growth and nutrition. There is evidence that both prenatal and postnatal nutritional restriction may contribute to adverse outcomes, and a number of interventions in the neonatal period have been shown to improve postnatal growth and development, at least in the short-term. Nonetheless, evidence for the long-term consequences of these interventions, particularly with respect to metabolic and cardiovascular health, is lacking, and most evidence on the long-term outcomes of neonatal nutrition and growth is epidemiological. High-quality prospective research in this area has lagged behind research in other aspects of neonatal care, and this deficit needs to be addressed. Future studies should attempt to dissect out the consequences of accelerated growth from its antecedents and whether the long-term risks and benefits of neonatal nutrition and growth, especially accelerated growth, may differ according to the outcome studied. This applies particularly to neurodevelopmental v. metabolic/cardiovascular outcomes, as targeting improvements in one aspect may involve a trade-off with respect to another.
Acknowledgment
The authors would like to thank all members of the LiFEPATH Group for their contributions to the referenced studies. This work was supported by the Health Research Council of New Zealand, the New Zealand Lotteries Grants Board, and the Auckland Medical Research Foundation.
