All premature infants <35 weeks of pregnancy and most ill term infants require full or partial PN whilst enteral nutrition is gradually introduced (IV).

There are a variety of reasons why premature infants (<35 weeks gestation) and seriously ill full-term infants may be unable to receive adequate enteral nutrition after birth, such as gastrointestinal tract immaturity with threat of developing necrotising enterocolitis, muscular and neurological immaturity, illnesses etc., and, hence, will require immediate PN in most cases.

The choice of nutrition administered (oral, enteral, partial PN or total PN) should be made on an individual basis according to medical indication and based on the principle of such nutrition being “as non-invasive as possible”. This process promotes low complication rates [18], [19], [20], [21] and is the reason why the highest possible percentage of nutrition should be administered orally, or enterally wherever possible. If this is not possible, (partial) parenteral nutrition should be used to supplement the enteral intake to provide adequate nutrition [22].

Premature infants are born with low food reserves (low subcutaneous fatty tissue, low glycogen reserves in the liver) in comparison to full-term newborns. There is specific risk of hypoglycaemia in connection with their high nutrient requirements. Premature infants <35 full weeks of pregnancy should therefore be prophylactically (re)introduced to solid food whilst administering (partial) PN. The “optimum supply” with various different nutritional components in premature and ill newborns is still being discussed. Principally it makes sense to differentiate between the recommendations for the adaptation and stabilisation phase and the recommendations for the phase of stable growth, due to differing nutritional requirements.

Published recommendations (e.g. [11], [15], [23], [24], [25], [26]) do not always take into consideration the specific conditions of the adaptation and stabilisation phase after birth (5–7 days after birth), which precede the phase of stable growth.

This means that the stunted growth occurring during that phase (compared to intrauterine percentiles) is often not regained by the pregnancy due date [27]. There are discussions about this target stating that the majority of eutrophic premature infants should have once again attained their birth percentile by the pregnancy due date. This target seems necessary, because growth retardation could be a risk factor for long-term neurological development [27], [28], [29]. The potential long-term risks of growth retardation and the side-effects of increased substrate intake in the newborn period have to be weighed up as there is no evidence-based data on long-term development when considering the different growth rates of premature infants. The ideal target for premature infants is usually seen as growth corresponding with intrauterine growth percentiles. It is, therefore, necessary to adapt the nutrient supply to the respective needs of the individual premature infant. This is also the case in the phase of continual growth (see below), where the target is not only to attain a positive nitrogen balance, as in the adaptation and stabilisation phase, but also to promote catch-up growth wherever possible right up to the birth percentile.

If the energy and nutrient demands of a patient cannot be met through enteral nutrition during pre-school or school age, then partial or total PN is to be considered within 7 days at the latest in dependance on the nutritional state and clinical conditions (C).

The beginning of (partial) parenteral nutrition in patients above the period of infancy should be specific to the individual circumstances, age and illness of the child or teenager. In contrast to infants, a period of inadequate nutrition and depletion of the body’s store can be tolerated in pre-school and school children with positive nutritional status for up to seven days subject to their clinical conditions.

Energy requirements are age-related (see Table 1 (Tab. 1)) and influenced by illness and therapy. Tables can provide sufficient help in typical clinical situations providing a rough approach to calculating the estimated requirements of individual patients, with the appropriate additions and deductions (e.g. in cases of pyrexia/respiratory therapy). The estimate should account for the fact that all details are based on data taken from healthy patients. Individual circumstances (e.g. reduced physical activity in case of bed rest, infections, inflammatory processes, energy losses from the stomata etc.) result in differences between the actual energy requirements and the calculated (estimated) energy requirements.

The actual requirements of the treated patient can be narrowed down further (e.g. weight chart) by means of monitoring measures (cf. chapter “Complications and monitoring” (http://www.egms.de/en/journals/gms/2009-7/000076.shtml)).

If the desired therapy effect (e.g. percentile parallel growth) is not attained with the estimated energy requirements and monitoring offers no solid references regarding an adequate energy intake, then various formulas can be helpful when calculating the evaluation. The equations (WHO 1985, Schofield 1985, and Harris-Benedict 1919 [30], [31], [32]) have, however, also been established for healthy children and have to be corrected for individual patients as specified in the figures published in tables. Measuring the energy expenditure is rarely carried out routinely during daily clinical routines and is indicated, if, despite estimation aids and monitoring, there are still doubts regarding an adequate energy intake.

We refer to the “Guidelines on Paediatric Parenteral Nutrition” of ESPGHAN and ESPEN for a more detailed overview of aspects regarding the energy requirements of children of various age groups and in case of varying clinical pictures [1].

The endogenous glucose production varies from approx. 2 mg/kg/min (3 g/kg/day) in adults to approx. 8 mg/kg/min (11.5 g/kg/day) glucose in premature infants (II).

Glucose is the sugar used for PN and usually contributes substantially to osmolarity in the PN solution. The osmolarity of a glucose solution rises significantly with increasing concentration from 255 mosm/l in a 5% glucose solution to 1020 mosm/l in a 20% glucose solution. Based on experience, glucose concentrations of up to 12.5% are well tolerated through a peripheral vein cannula as long as no other osmolarity-increasing agents are added.

Glucose can be directly metabolised by the central nervous system. The endogenous glucose production varies from approx. 2 mg/kg/min (3 g/kg/day) in adults to approx. 8 mg/kg/min (11.5 g/kg/day) glucose in premature infants [33], [34], [35], [36]. The maximum glucose intake should not exceed the glucose oxidation rate in PN. Maximum glucose oxidation is approx. 7 mg/kg/min (10 g/kg/day) [37], [38] in premature infants and approx. 12 mg/kg/min (18 g/kg/day) in full-term newborns and infants under long-term PN [39], [40], [41]. In critically ill children with burns, maximum glucose oxidation of 5 mg/kg/min has been described [41].

Excessive glucose intake results in net lipogenesis and subsequent fat deposition [42], [43]. Excessive intake can lead to hepatic steatosis with an impairment of the liver function [44], [45].

Higher incidence of hyperglycaemia with increasing immaturity (lower gestational age) (II).

There are often fluctuations in blood sugar levels in premature infants during the adaptation and stabilisation phase which can be influenced by low substrate reserves (hypoglycaemia) or insulin resistance (hyperglycaemia) [46], [47], [48]. Common definitions of hypo- or hyperglycaemia, respectively, are a blood glucose level of 50 mg/dl (2.75 mmol/l) or 150 mg/dl (8.3 mmol/l), respectively, although these levels are not based on short and long term outcome studies. The incidence of hyperglycaemia increases with diminishing gestational age [49], [50]. An early supply of intravenous insulin in hyperglycaemia has been proposed in order to more rapidly attain the targeted energy intake and a positive nitrogen balance [51], but this advantage should be weighed against potential complications. There are no controlled studies regarding the advantages and disadvantages of an early insulin therapy in premature infants. It has also been suggested that an added amino acid supply of 2–3 g/kg body weight and day should be started from the very first day to reduce hyperglycaemia during the adaptation and stabilisation phase, since it may stimulate endogenous insulin secretion and hence reduce the frequency and extent of neonatal hyperglycaemia [52]. This approach should be evaluated further in controlled studies.

The requirements for essential amino acids are higher per kg body weight in infants, and particularly in premature infants, than in older children or adults (II).

Crystalline amino acid solutions in concentrations of 3.5–15% (osmolarity 450–1450 mosmol/l) are used in PN. Some specifics are to be taken into consideration when using amino acid solutions in children compared to adults due to the special amino acid metabolism and growth-specific requirements.

The composition of amino acid solutions which are suitable for infants and toddlers, have to be adapted to the demands of metabolic immaturity and requirements of physical growth. Apart from the eight classic essential amino acids (Phe, Thr, Val, Leu, Ile, Tyr, Ser, Met), cysteine, tyrosine, histidine, taurine, glutamine and arginine are regarded as essential or conditionally essential especially in premature infants (see “Information on the selection and production of amino acid solutions” in the attachment).

Some amino acids regarded as non-essential in older children and adults are regarded as conditionally essential amino acids in newborns (II).

The need for essential amino acids is higher in premature infants than in older children or adults [11].

Various metabolic pathways for metabolising amino acids are immature in newborns (phenylalanine hydroxylase, tyrosine-aminotransferase, cystathionase [53], [54]). As a consequence amino acids regarded as non essential in adults i.e. cysteine, tyrosine, histidine, taurine, glutamine and arginine, become “conditionally” essential amino acids in newborns [55], [56]. Other amino acids like methionine quickly reach high levels because key enzymes are immature. Amino acid imbalances develop quicker in newborns than in adults or older children due to the immaturity of the neonatal metabolism. Such imbalances could have potential negative effects on organ development [57].

The plasma amino acid levels in infants receiving parenteral nutrition differ to those in breast-fed infants despite extensive endeavours to create optimised amino acid solutions for infants [58], [59], [60]. This is partly due to the poor solubility or stability of various amino acids (e.g. glutamine, tyrosine, cysteine) such that not all admixtures can be used. The nitrogen balance is not significantly influenced by the diverse composition of amino acid solutions [61], [62].

The maximum intake is basically influenced by two given facts:

Physiological rate of protein synthesis subject to age:

According to data by Pohlandt et al. [63]and Micheli [64] premature infants, who are physiologically in approx. week 30 of pregnancy, have a maximum rate of protein synthesis with protein requirements of approx. 2.7 g/kg body weight/day, which returns to <2.0 g/kg body weight/day until week 40 of pregnancy. A positive nitrogen balance can usually be attained in premature infants with an amino acid intake of 2.5 g/kg body weight/day and 60–90 kcal/kg body weight/day [65]. In individual cases an amino acid intake of up to 3.5–4 g/kg body weight/day may be necessary in order to achieve protein synthesis according to the intrauterine ratio [64].

Nitrogen balance studies in premature infants receiving parenteral nutrition show that approx. 380 mg (70%) are retained from an intake of 530 mg nitrogen, similar to the ratios found in enteral nutrition [66], [67]. Amino acid imbalances in PN are also discussed as a factor for developing cholestasis in young infants (a frequent side effect of long-term PN; up to 50% of long-term ELBW-premature infants receiving parenteral nutrition develop cholestasis [68]) [18], [69], [70].

References are made to the recommendations for enteral intake of amino acids due to insufficient data regarding the parenteral intake of amino acids in older children [71] (see Table 2 (Tab. 2)).

Lipid emulsions are an integral component of longer-term PN in children (C).

Lipid emulsions are used in PN in paediatric patients as they provide an energy source with low osmolarity and high energy content per volume unit. In addition, they also safeguard the essential fatty acid supply. The CO₂ production is lowered compared to PN with a high proportion of carbohydrates [72], [73], [74]. The nitrogen metabolism can be improved by adding lipid emulsions to PN [75], [76], [77].

Lipid oxidation depends on the overall energy intake and consumption, intake of carbohydrates and triglycerides and the carbohydrate/lipid ratio [72], [73]. Lipid oxidation decreases as the carbohydrate intake increases and is replaced by lipid storage. Lipogenesis takes place in infants with a carbohydrate intake of 18 g/kg/day and above [39], [72]. In older children the level of carbohydrate intake, above which net lipid deposition takes place, is lower. Lipid oxidation reaches its maximum at a parenteral lipid intake of 40% of non-protein energy intake in neonates [74] or 50% in infants [72]. Generally a lipid intake of 25–40 % of non-protein calories is recommended.

A deficiency of essential fatty acids can be biochemically shown in parenterally fed premature infants only after a few days with continuous glucose but not lipid infusion [78], [79], [80]. At least 0.25 g/kg body weight/day of linoleic acid should be administered to premature infants in order to prevent a deficiency of essential fatty acids [80], [81]. An intake of 0.1g/kg body weight/day is probably sufficient in full-term infants and older children. The different linoleic acid content in varying lipid emulsions has to be taken into consideration in the calculation for lipid intake. It is difficult to define the minimum requirements of alpha linolenic acid. The majority of data on this subject has been gained from animal experiments [82]. In children there is only one case study on alpha linolenic acid deficiency [83]. All lipid emulsions used in Germany contain alpha linolenic acid.

It is also difficult to define the upper limit for lipid intake. In premature infants an intake of 3 g/kg body weight/day is well tolerated as a continuous infusion according to the concentration of plasma triglycerides and cholesterol and also the ratio of non-esterified fatty acids/albumin [84], [85], [86]. In premature infants it may be desirable to achieve a lipid intake surpassing the oxidation capacity in order to attain weight gain and lipid deposition. Caution is, however, recommended in premature infants weighing less than 1000 g as tolerance to intravenous lipid emulsions may be more limited [87].

Maximum lipid oxidation of 4 g/kg body weight/day is attained in full-term infants with a glucose intake below 18 g/kg per day [73], [75].

It is important to monitor plasma triglycerides because lipid utilisation varies depending on age, the severity of the illness and various other factors.

An increase in the concentration of plasma triglycerides is to be expected if the infusion speed of the lipid emulsions exceeds the speed of hydrolosis of the triglycerides.

In premature infants the gradual increase in lipid intake compared to immediate administration of the target amount did not result in raised lipid tolerance [88]. If a gradual increase of 0.5 to 1 g/kg per day is carried out, this can serve the purpose of monitoring with measurements of plasma triglycerides concentrations.

In premature infants tolerance of a lipid infusion is raised by continuous versus intermittent administration [84], [86], [88]. In stable patients intermittent administration can also be carried out within the framework of cyclic home PN.

The speed of triglyceride hydrolysis depends on lipoprotein lipase activity. The activity of the post heparin lipoprotein lipase can be raised by administering heparin but this does not increase lipid utilisation [89], [90]. Raised lipoprotein lipase activity results in increased non-esterified fatty acids, which are not necessarily metabolised at the same speed [90], [91].

Lipid supply may result in enhanced lipid peroxidation and the formation of free radicals [92], [93], [94]. An increased lipid utilisation by reducing the carbohydrate/lipid ratio, at a constant lipid intake and therefore reduced energy intake, results in a reduction of lipid peroxidation and free radical formation [92]. PN should be supplemented with multi-vitamin preparations including both vitamin C and vitamin E (alpha-tocopherol) which have anti-oxidative effects [95], [96], [97].

PN with 20% lipid emulsions results in more physiological phospholipid and cholesterol levels, due to its low phospholipid content compared with PN containing classic 10% lipid emulsions [98]. At the same time there is, however, a lower proportion of long-chain polyunsaturated fatty acids (LC-PUFA) contributed by the phospholipid emulsifier. LC-PUFAs are “conditionally” essential for premature infants. Deficiency could result in malformations of the retina and central nervous system [99], [100].

No age differentiation is indicated in literature due to the limited data situation.

Vitamin supplementation should be given during parenteral nutrition (C).

The optimum time to begin vitamin supplementation in newborns and premature infants is not clear. It should be taken into consideration, that water soluble vitamins, with the exception of Vitamin B12, are inadequately stored. Insufficient thiamine intake can result in severe lactic acidosis within only a few days in children on PN [101]. It is, therefore, recommended that vitamin substitution is started early with PN.

Vitamin intake should be administered daily. All clinical studies have been carried out with commercially available vitamin preparations, therefore existing intake recommendations [11], [23], [102] are based on the composition of available preparations. At this time there is no commercially produced vitamin supplement available on the market, which completely meets the current intake recommendations for premature infants as a supplement for long-term PN.

In the practical administration of PN it should be considered that vitamins can be degraded by oxygen, light and heat. Degradation reactions can be accelerated by catalytically active trace elements like copper and iron. Fat-soluble vitamins can be adsorbed on specific synthetic materials (infusion needles) in rare cases. Thus, the administered dose is uncontrolled and significantly reduced [103]. In Europe, combined preparations are available which, by dissolving the vitamins in a lipid emulsion, reduce absorption on plastic. Furthermore, the formation of peroxides in lipid emulsions can be reduced by adding multi-vitamin preparations [104]. Vitamin preparations should, if possible, be administered together with the lipid emulsion. The clinical impact of free radicals, which can develop in intravenous multi-vitamin emulsions when exposed to light, is being currently discussed. They can result in an increase in peroxide excretion in urine of newborns [105]. Feed lines with light protection should be used when administering parenteral solutions with vitamins until the clinical significance of these results has been determined [106].

Vitamin D rickets prophylaxis: Oral vitamin D prophylaxis with 1000 IU Vitamin D/day per os from day seven of life is sufficient as a prophylaxis in extremely premature infants [107].

Vitamin K prophylaxis: Premature infants (<week 35 of pregnancy) and ill full-term infants should receive their first two doses of Vitamin K prophylaxis intramuscularly, subcutaneously or intravenously due to the unclear resorption from the gastrointestinal tract [108], [109]. The dose of intravenous Vitamin K prophylaxis is still being discussed at present. Vitamin K levels after one administration of 1 mg Vit K intramuscularly on the first day of life were higher in premature infants than in full-term infants [110].

Detailed specifications can be found in the “Guidelines on Paediatric Parenteral Nutrition” from ESPGHAN and ESPEN [1].

No age differentiation is indicated in literature on account of the limited data situation.

The optimum time to begin with trace element supplementation in premature infants <1500 g birth weight is not clear. We recommend beginning supplementation to coincide with an increase in body weight (5^th day of life) (C).

Trace element requirements and optimum time to begin supplementation in premature and full-term infants, as well as infants and children, are not completely clear. The enteral absorption of trace elements in partial-parenteral nutrition mainly depends on the existing compound and the composition of the nutrition [111] such that it is extremely difficult to evaluate the remaining parenteral needs. Requirements for some trace elements have at least been narrowed down by controlled studies [102], [111], [112], [113], [114]. There is no trace element supplement for long-term PN which meets the current recommendations [102]. In long-term PN (>7 days), where enteral nutrition provides less than 50% of the energy intake, we recommend administering trace element supplements according to manufacturer instructions.

Premature and ill full-term infants have an increased risk of developing a trace element deficiency. Premature infants are born with lower trace element stores as the laying down of these stores occurs during the last trimester of the pregnancy. Rapid growth at unidentified requirement levels and variable resorption are further risk factors contributing to the development of a trace element deficiency in premature infants [115]. In ill full-term infants the requirements and intake vary according to the basic illness. The optimum time to begin supplementation is disputed.

Zinc: Zinc deficiency should be considered in newborns where the reason for unclear, poor development (especially linear growth) and/or skin efflorescences (acra, mechanically burdened parts of the body) are not clear [116], [117]. A multitude of case studies have been published on zinc deficiency in neonates [116]. Rare genetically-related trace element metabolic disorders are also to be considered in the case of unclear clinical symptoms [117].

There is no documentation indicating a benefit from parenteral supplementation of glutamine or arginine in children (C).

Tests on arginine supplementation [118] in adult intensive care patients as well as some studies on enteral or parenteral glutamine administration in newborns show potential positive effects [119], [120] (II), although these are controversially discussed. There is no evidence from meta analyses to document the benefit of the parenteral supplementation of glutamine or arginine in premature and newborn infants [121].

Vitamin A is necessary for physiological lung growth (lung epithelial cells). A meta-analysis, which is based on seven randomised, controlled studies, documents the effectiveness of a supranutritive Vitamin A supplementation on reducing oxygen requirements and improving survival at one month of age in premature infants with a birth weight <1000 g. There is a tendency towards lower values regarding the incidence of retinopathy in premature infants [122], [123] (I). The authors of the systematic analysis of the Cochrane database feel there is a need to carry out further studies on the effects of intravenous Vitamin A supplementation.

Carnitine is necessary for the transportation of long-chain fatty acids via the mitochondrial membrane and its oxidative metabolism. Carnitine is found in breast milk and baby foods, but is not usually added to conventional PN. Low carnitine levels have been measured in body tissues in premature infants on PN [124]. The clinical significance of this has not been determined. An impairment in fatty acid oxidation is only to be anticipated where a massive drop in serum carnitine concentrations occurs. For the metabolic availability of carnitine its free nonesterified percentage, is essential. The availability of free carnitine is reflected in the concentration and the acyl-carnitine/free carnitine (AC/FC) ratio. This ratio is in dynamic equilibrium with the intramitochondrial acyl-CoA/free CoA ratio. Lower intramitochondrial CoA-availability is assumed with an AC/FC ratio >0.4 (>0.7 when fasting) Carnitine supplementation results in a release of intramitochondrial CoA and should be considered in case of an explicit AC/FC ratio. A meta analysis (based on 14 randomised, controlled studies) showed there to be no effect of carnitine supplementation on the metabolism of lipids, lipogenesis or weight gain [125].

The utilisation of an adapted glucose/electrolyte solution (usually 10% glucose) with potassium and sodium supplementation in short-term (<48 h) intravenous intake is recommended in well-nourished toddlers and school children without specific metabolic or nutritional risks (IV).

If intravenous intake is required in well-nourished toddlers and school children without specific metabolic or nutritional risks, the following situations should be considered when anticipating duration:

I: short-term parenteral intake for less than 48 h. II: medium-term parenteral nutrition for 2 to 7 days.III: long-term PN for >7 days.

It makes sense to differentiate between short, medium and long-term parenteral intake, because the procedure varies according to the duration of PN

There should be a written concept on the provision of PN in order to minimise errors (C).

Due to the heterogeneity in pathophysiology and the differing maturity of patients it may be necessary to vary procedures during the neonatal period (e.g. fluid volumes, electrolyte substitution etc.). Classification according to, for example, birth weight is recommended for the calculation of PN in neonates:

premature infants <1500 g,

This differentiation makes sense with regards to variations in the administration of different nutrients. The described protocol can result in the procedures in (partial) parenteral nutrition being structured during the daily clinical routine and errors minimised.

A standardised questionnaire or electronic programme, which takes into account partial parenteral nutrition and enteral nutrition (see Table 3 (Tab. 3) for an example of a procedure questionnaire), should be used for prescribing PN to neonates.

The prescription of nutrition for neonates must be calculated and not estimated in order to ensure the nutrient intake in neonates and infants is adapted to their specific needs.

Peripheral commercial vein cannulae (PCVC) have a lower complication rate in infants (infection, thrombosis) than central venous catheters (CVC) [20]. Peripheral commercial vein cannulae (PCVC) can be used in partial or total parenteral nutrition if the osmolarity and state of the veins allow for this (risk of paravasatation/skin necrosis!).

Long-term total parenteral nutrition often cannot be applied safely in older children or teenagers without central venous catheters (CVC) due to the osmolarity. An individual decision is required on the choice of access while taking into consideration the underlying illness, therapy, osmolarity of the nutritional solution used or drugs and the expected duration.

Two meta-analyses showed no positive effect of heparin supply on central venous catheter lifetime or the formation of thromboses in neonates with percutanous central venous catheters [136], [137].

After requirements have been calculated, standard solutions (e.g. produced by the hospital pharmacy), which are adapted to suit the specific nutrient needs of the respective age group, can be administered in short-term PN (B).

Using ready-made standard solutions requires fewer personnel and allows for less risk regarding dosage errors or microbial contamination. Individually mixed infusions can be adapted to individual characteristics. Standard solutions are suitable for short-term total and partial PN [138].

A standardised procedure should be followed and the individual steps systematically documented in order to minimise errors. Computer programmes, which enable the fast and exact calculation of enteral and parenteral intake, are recommended (and in part commercially available). The fluid, glucose and electrolyte intakes and additional intakes with drugs can be calculated.

The following aspects should be considered: Estimation of the required duration of PN, enteral nutrition, fluid intake, protein and lipid intake, parenteral vs enteral/oral intakes, electrolyte/vitamin and trace element supply, concentration of the glucose solution, infusion speed, monitoring, and validation (see Table 3 (Tab. 3) “Example of a procedure questionnaire”).

A practical example is shown in Table 4 (Tab. 4), which illustrates the (re)establishing of enteral food in newborns during the adaptation and stabilisation phase.

Due to the low blood volume in infants, staff at facilities in which infants receive medium and long-term parenteral nutrition must have access to a special laboratory with micromethods (IV).

Daily clinical tests, to include the monitoring of the fluid balance, checks of the acid base status, electrolytes and blood sugar, are usually required in the initial phase of PN depending on the age and maturity of the children and their underlying illness. In medium and long-term PN, routine clinical tests should also be accompanied by the following: documentation of weight, length and head circumference development (in percentile questionnaires), weekly evaluation of acid base status, blood sugar, electrolytes, haematocrit, urea, creatinine, at least one transaminase, Y-GT, urine osmolarity or specific weight (alkaline phosphatase every two weeks).

The concentration of plasma triglycerides at which undesirable effects occur is not known [145]. Average triglyceride concentrations of 150 to 200 mg/dl and above are often specified in infants fed with breast milk or infant formula [86], [146]. Values of 250 mg/dl in healthy infants are not unusual. In older children higher concentrations of up to 300 and 400 mg/dl can also be acceptable as the lipoprotein lipase doesn’t become saturated until approx. 400 mg/dl [147]. When gradually introduced, plasma triglycerides should be initially checked on a weekly basis for every 1 g/kg lipid intake increase and after reaching maximum intake levels.

Monitoring the fluid balance in premature infants and sick newborns is extremely important due to their high fluid volume, high body water content in comparison to older patients and immature regulatory mechanisms (see above). Monitoring must take the special physiological features of neonates into consideration to be efficient:

The measurement of the specific weight or osmolarity of urine are only indicative in premature infants and newborns in the first weeks of life when high values are measured. Low (normal) values can be due to low renal concentrating ability in premature infants and newborns because of immature kidneys.

Several metabolic pathways involved in the synthesis of amino acids are still immature in newborn and premature infants. As a result, some amino acids, which are regarded as non-essential for adults, may become conditionally essential (e.g. cysteine, tyrosine, histidine, taurine, and glutamine). In addition, individual plasma amino acids reach clearly increased levels because significant degradation enzymes are still immature. Amino acid imbalances occur more rapidly in comparison to adults (I).

Increased phenylalanine concentrations are toxic for the central nervous system and can result in serious developmental disorders. Both hepatic phenylalanine hydroxylase activity and the enzymatic system of the metabolism of tyrosine are still immature in premature infants. Premature infants (<week 30 of pregnancy) with high amino acid intakes therefore have a tendency to develop hyperphenylalaninaemia and hypertyrosinaemia.

The enzyme activities in full-term infants result, however, in a rapid amino acid metabolism, causing low phenylalanine and tyrosine concentrations. Attempts are made to increase the low plasma tyrosine concentrations in infants receiving parenteral nutrition by supplementing with N-acetyl tyrosine, which is more soluble. Newborns and premature infants do still, however, have a reduced deacetylation capacity such that tyrosine may not be sufficiently utilised in this more soluble form [148], [149], [150], [151].

There is an interaction between the branched-chain and aromatic amino acids: Leucine supports the phenylalanine and tyrosine metabolism [152].

Methionine is a sulphurous, essential amino acid and is regarded as the precursor to cysteine and taurine. Methionine is the major methyl group donator in the hepatic intermediary metabolism. It is converted to homocysteine through demethylation, which remethylates to methionine when tetrahydrofolic acid is introduced or which is finally converted to cysteine via a transsulfuration pathway forming cystathionine.

Cysteine is a largely essential amino acid in both premature infants and full-term newborns due to its immature cystathionase activity. Plasma cysteine concentration is already a function of methionine intake in premature infants. It is often observed that normal plasma cysteine concentrations only occur if plasma methionine concentrations are raised. As a result cysteine supplementation of amino acid solutions is recommended in young infants [153], [154].

Supplementation with cysteine-HCl is not usually applied because of its low solubility and risk of acidosis. The available N-acetyl cysteine may not be effectively metabolised due to limited deacetylation and immature metabolic rates in premature infants. Procysteine, a preliminary stage of glutathione, has also been tested [155].

Taurine is formed through decarboxylation of cysteine sulfinic acid, an enzyme stage that is still immature in premature infants. The foetus accumulates approx. 50–60 µmol taurine/day during the last trimester of pregnancy [156].

Taurine has many functions, not all of which are completely understood, although it is of importance to the central nervous system, retina and gastrointestinal tract. The proven relevance of the maturing of acoustically evoked potentials points to the clinical significance of taurine [157], [158]. Young infants on PN are, therefore, dependent on exogenous taurine intake in order to maintain plasma concentrations [157], [159].

Threonine: The plasma threonine concentration is directly proportional to the intake as well as inversely proportional to the gestational age in PN for premature infants [160]. A toxicity of raised plasma threonine concentrations has not yet been proven, although it is kown that concentration levels in the brain increase proportionally to plasma concentrations.

Lysine: Of all amino acids the materno-foetal gradient is most prominent in lysine. Plasma lysine concentrations in the foetus are approx. four times higher than those of the mother. Both lysine deficiency and surplus have negative effects on growth, brain DNA concentration and the dynamics of urea synthesis [161].

Leucine influences the speed of protein synthesis in the muscles. The demand for branched-chain amino acids is greater in premature infants than in full-term infants, and thus they easily tolerate amino acid solutions with a higher concentration of branched-chain amino acids.

Histidine: Of all essential amino acids, the plasma concentration of histidine is least influenced by the intake. Histidine is therefore regarded as a conditionally essential amino acid in young infants.

Arginine: The arginine concentration in an amino acid solution contributes to preventing hyperammonaemia. Premature infants with asymptomatic hyperammonaemia show lower plasma concentrations of urea cycle intermediates, which can be normalised by raising the arginine intake. Part of the arginine requirement can also be met by ornithine, which is taken into consideration in some amino acid solutions [162].

Based on recent data, it is assumed that N-acetyl amino acids are only metabolised to a limited extent in humans and therefore are only of limited significance as alternative amino acid sources in clinical nutrition (IV).

There are no documented benefits of additionally administering N-acetyl cysteine due to limited deacetylation and immature metabolic rates in premature infants.

PE with a low phospholipid triglyceride ratio (e.g. in 20% lipid emulsions) results in less elevated phospholipid and cholesterol levels (II).

Apart from the long established use of lipid emulsions based on soy bean oil, a new emulsion based on an olive oil and soy bean oil mixture has been available for some time and studies with this new mixture have already shown promising results in children, infants and premature infants [163], [164]. Postulated benefits of the olive oil-based lipid emulsions are reduced lipid peroxidation, lower PUFA intake and higher intake of antioxidative substances. Currently data are insufficient to propose a recommendation for olive oil-based emulsions [165].

Soy bean and olive oil-based lipid emulsions contain long chain triglycerides. There are also lipid emulsions with the equal proportions of LCT and MCT from coconut oil. They contain less PUFA, and the MCT proportion is oxidised more rapidly [166]. A further potential benefit is that oxidation of MCT is less dependent on carnitine that of LCT. The energy content of MCT per g fat is, however, approx. 16% lower than that of LCT. Studies in adults and children show a higher lipid oxidation, less influence on the parameters of the liver function, improved leukocyte function and less influence on pulmonary haemodynamics and the gas exchange than with LCT emulsions [167], [168], [169], [170], [171]. There were no significant differences in regard to the plasma lipids and fatty acids [167], [172], [173], [174]. There is differing data concerning a nitrogen-saving effect when administering a MCT/LCT mixture. One study describes increased nitrogen retention [175], [176], while another study describes a more unfavourable leucine metabolism than when on LCT intake [177]. The current studies in children and newborns [125], [167], [173], [178], [179] do not justify any generally favoured utilisation of MCT/LCT emulsions compared to pure LCT emulsions at present [165], [174].

The ratio between phospholipids (PL) and triglycerides (TG) is lower in 20% lipid emulsions than in standard 10% lipid emulsions [180]. Lipid emulsions with a low PL/TG ratio as in the standard 20% emulsions, should be favoured in PN [98], [125], [181], [182]. Higher amounts of phospholipids impair the clearance of triglycerides from the plasma, resulting in an increase in the plasma triglyceride concentration and of cholesterol and phospholipids in Low-Density-Lipoproteins [98]. The emulsion characteristics (dispersity, stability) and various incompatibilities are of great significance in the parenteral intake of lipids.