Clinical Nutrition 33 (2014) 982e990

Contents lists available at ScienceDirect

Clinical Nutrition
journal homepage: http://www.elsevier.com/locate/clnu

Randomized control trials

Adaptive regulation of amino acid metabolism on early parenteral

lipid and high-dose amino acid administration in VLBW infants e A
randomized, controlled trial
Hester Vlaardingerbroek a, Jorine A. Roelants a, Denise Rook a, Kristien Dorst a,
Henk Schierbeek a, b, c, Andras Vermes d, Marijn J. Vermeulen a,
Johannes B. van Goudoever a, b, c, *, Chris H.P. van den Akker a
a
Department of Pediatrics, Division of Neonatology, Erasmus MC e Sophia Childrens Hospital, c/o Room SP3433, P.O. Box 2060,
3000 CB Rotterdam, The Netherlands
b
Department of Pediatrics, Emma Childrens Hospital e AMC, c/o Room H7-282, P.O. Box 22660, 1100 DD Amsterdam, The Netherlands
c
Department of Pediatrics, VU University Medical Center, c/o Room ZH 9D11, P.O. Box 7057, 1007 MB Amsterdam, The Netherlands
d
Hospital Pharmacy, Erasmus MC, P.O. Box 2060, 3000 CB Rotterdam, The Netherlands

a r t i c l e i n f o

s u m m a r y

Article history:
Received 25 June 2013
Accepted 3 January 2014

Background & aims: An anabolic state can be achieved upon intravenous amino acid administration
during the immediate postnatal phase despite a low energy intake. The optimal dosing of amino acid and
energy intake has yet to be established. The aim was to quantify the efcacy of early initiation of
parenteral lipids and increased amounts of amino acids on metabolism and protein accretion in very low
birth weight infants.
Methods: 28 very low birth weight infants were randomized to receive parenteral nutrition with glucose
and either 2.4 g amino acids/(kg$d) (control group), 2.4 g amino acids/(kg$d) plus 2e3 g lipid/(kg$d)
(AA lipid group), or 3.6 g amino acids/(kg$d) plus 2e3 g lipid/(kg$d) (high AA lipid group) from birth
onward. On postnatal day 2, we performed a stable isotope study with [1-13C]phenylalanine, [ring-D4]
tyrosine, [U-13C6,15N]leucine, and [methyl-D3]a-ketoisocaproic acid to quantify intermediate amino acid
metabolism.
Results: The addition of lipids only had no effect on phenylalanine metabolism, whereas the addition of
both lipids and additional amino acids increased the amount of phenylalanine used for protein synthesis.
In addition, high amino acid intake signicantly increased the rate of hydroxylation of phenylalanine to
tyrosine, increasing the availability of tyrosine for protein synthesis. However, it also increased urea
concentrations. Increasing energy intake from 40 to 60 kcal/(kg$d) did not increase protein efciency as
measured by phenylalanine kinetics. The leucine data were difcult to interpret due to the wide range of
results and inconsistency in the data between the phenylalanine and leucine models.
Conclusions: High amino acid and energy intakes from birth onwards result in a more anabolic state in
very low birth weight infants, but at the expense of higher urea concentrations, which reects a higher
amino acid oxidation. Long-term outcome data should reveal whether this policy deserves routine
implementation.
This trial was registered at www.trialregister.nl, trial number NTR1445, name Nutritional Intervention for
Preterm Infants-2.
2014 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

Keywords:
Metabolism
Preterm infant
Parenteral nutrition
Growth
Stable isotopes

1. Introduction

* Corresponding author. Emma Childrens Hospital e AMC, c/o Room H7-282, P.O.
Box 22660, 1100 DD Amsterdam, The Netherlands. Tel.: 31 20 5667729; fax: 31
20 5669683.
E-mail addresses: h.vangoudoever@amc.uva.nl, h.vangoudoever@vumc.nl
(J.B. van Goudoever).

The neonatal period is characterized by protein accretion and

rapid growth. In response to protein administration, adults1,2 and
term infants3,4 decrease proteolysis and increase their endogenous
glucose production rate.5 However, several studies have shown that
fetuses and preterm infants increase protein synthesis rather than

0261-5614/$ e see front matter 2014 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
http://dx.doi.org/10.1016/j.clnu.2014.01.002

H. Vlaardingerbroek et al. / Clinical Nutrition 33 (2014) 982e990

Abbreviations
AA
CRIB
ECF
GC/MS
GMP
KIC
Leu
MPE
Phe
Q
RP
S
TG
TTR
VLBW

amino acid
Critical Risk Index for Babies
ethyl chloroformate
gas chromatography mass spectrometer
good manufacturing practice
a-ketoisocaproic acid
leucine
mole percent excess
phenylalanine
ux
rate of amino acid release from protein
rate of utilization of amino acid for protein synthesis
triacylglycerol
tracer-tracee ratio
very low birth weight

suppress protein breakdown in response to protein administration.3,6e8 Because protein synthesis is an energy-demanding process, sufcient energy should be provided to optimize this process.
The relationship between energy supply and protein synthesis
appears to be curvilinear, with the greater benecial effect of energy on protein accretion occurring at intakes of <50e60 kcal/
(kg$d). At higher intakes, the amount of supplied amino acids is
more highly correlated with anabolism.9 The optimal glucose and
lipid intakes that maximize protein accretion and growth in the
early neonatal phase have not yet been determined.
In this study, we analyzed the effect of the administration of
lipids (as an energy source) from birth onward, in combination with
various levels of amino acid administration, on amino acid metabolism in very low birth weight (VLBW; birth weight < 1500 g)
infants by means of stable isotope modeling using [1-13C]phenylalanine and [ring-D4]tyrosine and a model using [U-13C6,15N]
leucine and [methyl-D3]a-ketoisocaproic acid. The major routes of
disposal of phenylalanine are protein synthesis and hydroxylation
to tyrosine. The main sites of hydroxylation are the liver and kidneys. Tyrosine can be synthesized into proteins, oxidized to carbon
dioxide, or metabolized into several neurotransmitters and the skin
pigment melanin. Leucine has a unique role in protein synthesis in
skeletal muscle, can act independently as a nutrient signal, and
stimulates protein synthesis via the activation of translational
initiation factors. In addition, leucine stimulates insulin release and
tissue sensitivity. Infants were randomized to either a control group
(no lipid during the rst two days), an early lipid administration
(AA lipid) group, or an early lipid administration plus high amino
acid intake (high AA lipid) group. Because of the provision of
additional energy, early lipid administration may increase protein
synthesis and/or decrease protein breakdown, leading to enhanced
net protein accretion. We hypothesized that additional amino acids
in combination with the administration of lipids from birth onward
would augment protein accretion further.

2. Materials and methods

983

present study were a subset of those included in our larger study10

to determine the safety and efcacy of early lipid initiation with or
without additional amino acids from birth onward. The subjects
were inborn VLBW infants (birth weight < 1500 g) with a central
venous and arterial catheter inserted for clinical purposes and
written informed consent from the parents. Infants were excluded
in cases of congenital anomalies (including chromosome defects),
metabolic diseases, or endocrine, renal or hepatic disorders. The
study protocol was approved by the institutional medical ethical
review board.
Within 6 h after birth, the attending physician included infants
in the study by opening a sealed opaque randomization envelope
stratied by weight (<1000 g versus 1000e1499 g) and gender. The
envelopes were created by a research pharmacist who was not
involved in clinical care and were based on a computer-generated
block randomization list with variable block sizes that was provided by a statistician. The study group randomization was open
after inclusion for logistical reasons; however, all of the technicians
were blinded for study group randomization throughout the study
and the analyses.
2.2. Nutritional intervention
As soon as intravenous access was obtained after birth, the infants received glucose (at least 4.0 mg/(kg$min)) and 2.4 g/(kg$d) of
amino acids (always on stock on the ward) as part of standard
clinical care. Immediately after randomization to one of the three
study groups, the experimental parenteral nutrition was
substituted for the infants in the AA lipid and high AA lipid
groups.
2.2.1. Control group
The infants in the control group continued to receive only
glucose and amino acids (2.4 g/(kg$d)) during the rst two days of
life.
2.2.2. AA lipid group
The infants in the AA lipid group received glucose and amino
acids similarly to those in the control group (2.4 g/(kg$d)); however, lipids were started immediately after randomization (always
within 6 h after birth) (starting dose 2 g/(kg$d), increased next day
to 3 g/(kg$d)).
2.2.3. High AA lipid group
The infants in the high AA lipid group received, in addition to
glucose from birth onward, high-dose amino acids (3.6 g/(kg$d)
from birth onward) and lipids (starting dose 2 g/(kg$d) from birth
onward, increased next day to 3 g/(kg$d)).
The three groups received the same amino acid solution: Primene 10% (Baxter, The Netherlands). The infants in the control
group received Intralipid 20%. The infants in the intervention
groups were randomized to receive Intralipid 20% or SMOFlipid 20%
(both Fresenius Kabi, Germany). Because the type of lipid had no
effect on amino acid kinetics, the lipid source was ignored in the
nal analyses. Whenever possible, minimal enteral feeding was
initiated on the day of birth and feeding progressed to full enteral
nutrition in the following days according to local protocol.

2.1. Subjects
2.3. Study design and tracer protocol
We performed a randomized controlled trial on the initiation of
parenteral nutrition to preterm infants in the early postnatal phase
from June 2009 to January 2012. The study was conducted at the
neonatal intensive care unit of the Erasmus MC e Sophia Childrens
Hospital, Rotterdam, the Netherlands. The infants included in the

The parameters recorded at baseline were gender, birth weight

and birth weight z-score,11 gestational age based on best obstetric
measurement (ultrasound in early pregnancy or last menstrual
period), number of prenatal steroid doses (0, 1, or 2), and severity of

984

H. Vlaardingerbroek et al. / Clinical Nutrition 33 (2014) 982e990

illness at entry in the study based on the Apgar score at 5 min and
the Critical Risk Index for Babies (CRIB) score.12
The actual daily nutritional intake (parenteral and enteral) was
recorded throughout the study.
A stable isotope study was performed on the second day of life.
Good manufacturing practice (GMP) tested L-[1-13C]phenylalanine,
L-[ring-D4]tyrosine, L-[U-13C6,15N]leucine, and [methyl-D3]aketoisocaproic acid ([D3]KIC) (all 94% atom enriched and tested
for sterility and pyrogenicity) were from Cambridge Isotope Laboratories (Andover, MA). The amino acid isotopes were dissolved in
0.9% saline by the hospital pharmacy, and the solution was ltered
(0.2 mm lter) and sterilized according to GMP guidelines. Separate
vials with a precisely weighed dry powder form of [D3]KIC were
aseptically produced by the hospital pharmacy according to GMP
guidelines. Tests were performed to ensure the correct identity,
concentration, and sterile and pyrogen-free status of the product.
Immediately before infusion, the [D3]KIC powder was dissolved in
the amino acid tracer solution. The infants received a primed
continuous infusion of [1-13C]phenylalanine (6 mmol/(kg$h)), [ringD4]tyrosine (2 mmol/(kg$h)), [U-13C6,15N]leucine (12 mmol/(kg$h)),
and [D3]KIC (6 mmol/(kg$h)) for 8 h using a Perfusor fm infusion
pump (BjBraun Medical B.V., Oss, the Netherlands). The priming
doses were equivalent to an hourly dose. After 6, 7, and 8 h of
infusion, blood (0.5 mL) was sampled from the arterial catheter. The
total amount of blood sampled did not exceed 5% of the patients
estimated total blood volume (75 mL/kg). The samples were
collected in EDTA-containing tubes and immediately placed on
melting ice and centrifuged (10 min, 3500 g). The plasma was
stored at 80  C until analysis.
2.4. Analytical methods
Plasma amino acid enrichments were analyzed as follows. A 50-

mL aliquot of plasma was acidied with 20 mL of hydrochloride

(pH < 3). To isolate the amino acids, 0.2 mL of prewashed Dowex
suspension (Ag 50W-X8 H, 200e400 mesh) was added to the
acidied samples, thoroughly shaken, and centrifuged at 4000 rpm
for 1 min. The supernatant was discarded and the pellet was washed
twice with 1 mL of H2O. After centrifugation, the amino acids were
extracted from the Dowex pellet with 0.5 mL of ammonia 6N and
transferred to a new vial. The original vial was rinsed with 0.2 mL of
ammonia 6N, and 0.2 mL of the supernatant was added to the new
vial and evaporated using a SpeedVac (GeneVac miVac, GeneVac
Ltd., Ipswich, England). The samples were redissolved in 200 mL of
H2O and derivatized with ethyl chloroformate (ECF) by the addition
of 140 mL of ethanol/pyridine (4:1) and 20 mL of ECF. The samples
were left at room temperature for 5 min and then extracted with
400 mL of hexane/dichloromethane/ECF (50:50:1). After centrifugation, 200 mL of the supernatant was transferred to a vial. The
extraction step was repeated; the second time, 400 mL of the supernatant was added to the rst portion. The combined solutions
were evaporated under an N2 stream at room temperature, redissolved in 50 mL of ethyl acetate, and analyzed with an MSD 5975C
Agilent gas chromatography mass spectrometer (GC/MS) (Agilent
Technologies, Amstelveen, The Netherlands) on a VF-17ms,
30 m  0.25 mm ID capillary column (Varian Inc., Middelburg,
The Netherlands). KIC enrichments were measured as described
previously.13 Briey, samples were derivatized with silylquinoxalinol, phenylenediamine, and N-methyl-N-(tert-butyldimethylsilyl)triuoroacetamide and analyzed by GC/MS.
2.5. Calculations
For the calculation of whole body phenylalanine and tyrosine
kinetics, including rates of hydroxylation to tyrosine, we used the

model developed by Clarke and Bier14 with the enhancement

proposed by Thompson et al.15
The following calculations were used; all in (mmol/(kg$h)):
Qphe i$phe * ((E infusate phe/E plasma phe)  1)
Qtyr i$tyr * ((E infusate tyr/E plasma tyr)  1)
where Q indicates ux in mmol/(kg$h), i is the tracer infusion rate in
mmol/(kg$h), and E is the tracer enrichment.
Hydroxylation Qtyr * (E plasma tyr 13C/E plasma phe) * (Qphe/(i
phe Q phe))
Non-oxidative phenylalanine disposal (reecting synthesis) Q
phe  hydroxylation
Phenylalanine release from protein (reecting
breakdown) Qphe  phe intake
Tyrosine release from protein (reecting breakdown) Qtyr  tyr
intake  hydroxylation
Phe and tyr intake were calculated from the molar content of
parenteral and enteral intake (the latter was only a very small
amount).
Phenylalanine balance non-oxidative phenylalanine
disposal  phenylalanine release from protein
Leucine calculations are presented in the Addendum.
We made the following assumptions regarding the tracer
model: 1) during the course of the experiment, the release of label
from body protein is small in relation to the rate of isotope infusion;
therefore, the recycling of label is negligible in comparison with the
total amount of unlabeled substrate entering the mixing pool; 2) a
labeled molecule is not distinguished from an unlabeled molecule.
2.6. Statistics
A power calculation based on leucine oxidation showed that,
given an expected decrease in leucine oxidation of 15 mmol/(kg$h)
and a standard deviation of 10 mmol/(kg$h), 21 infants (7 in each
group) were needed to detect a signicant difference with a of 0.05
and power of 0.80. For additional power, we included 7e12 infants
per group.
The results are expressed as median (IQR). All statistical analyses
were performed using SPSS Version 20.0 (IBM SPSS Statistics,
Somers, NY). The signicance level was set to p < 0.05. The differences between groups were analyzed using the c2-test, Kruskale
Wallis test, or ManneWhitney U test, as appropriate. Spearman
correlations were calculated between the two isotope models.
Linear regression analyses were used to correct for the potential
effects of gestational age, birth weight, and birth weight z-score.
3. Results
3.1. Demographic and clinical characteristics
We included 28 infants (7e12 infants per nutritional group) in
the nal analyses (Table 1). The CONSORT ow is depicted in Fig. S1
(supplemental data). Not all of the infants were eligible for further
analysis following the randomization to the three nutritional regimens. The major reasons were adjustments of nutrition according
to the local protocol (i.e., plasma urea concentration 10 mmol/L or

H. Vlaardingerbroek et al. / Clinical Nutrition 33 (2014) 982e990

985

Table 1
Clinical characteristics.

N (male/female)
Gestational age,
weeks
Birth weight, g
Birth weight
z-score (13)
Prenatal steroids
(% 0/1/2 doses)
Apgar score at 5 min
CRIB score (14)

Control
group

AA lipid
group

High AA lipid
group

7 (6/1)
29.4 (27.4e29.7)

9 (4/5)
26.9 (25.1e27.6)

12 (7/5)
26.7 (25.3e28.0)

980 (670e1140)
3.4
(4.5 to 1.2)
0/0/100

800 (669e918)
0.7
(3.2 to 0.01)
0/22/78

755 (642e975)
1.9
(3.5 to 0.5)
0/33/67

8 (4e9)
4 (1e8)

9 (7e9)
4 (3e8)

8 (6e9)
5 (2e9)

Data are presented as median; IQR in parenthesis, unless otherwise indicated. AA,
amino acid; CRIB, Critical Risk Index for Babies.

triacylglycerol (TG) concentration 3 mmol/L) before the start or

during the isotope study and clinical contra-indications for the
isotope study (e.g., uid restriction, restriction of blood
withdrawals).
Per denition, the actual intake on day 2 of life was in accordance with the intended study intakes (Table 2).
3.2. Phenylalanine and tyrosine kinetics
Phenylalanine and tyrosine enrichments at steady state are
presented in Table 3 (Supplementary data). In the high AA lipid
group, phenylalanine hydroxylation to tyrosine and utilization of
phenylalanine for protein synthesis (Fig. 1(A)) were signicantly
higher than in the groups with the standard amino acid intake,
whereas the release of phenylalanine from protein (reecting
proteolysis) did not differ among the groups. Phenylalanine balance
was signicantly higher in the high AA lipid group (Fig. 1(B)).
Phenylalanine metabolism did not differ between the control group
and the AA lipid group.
Phenylalanine efciency was calculated as the net phenylalanine balance divided by the phenylalanine intake and was not

Table 2
Actual nutritional intake on day two of life.

N
Parenteral glucose intake,
mg/(kg$min)
Parenteral amino acid
intake, g/(kg$d)
Parenteral Enteral
amino acid and
protein intake,
g/(kg$d)
Total phenylalanine
intake, mmol/(kg$h)
Total tyrosine
intake, mmol/(kg$h)
Total leucine
intake, mmol/(kg$h)
Parenteral lipid
intake, g/(kg$d)
Enteral intake, ml/(kg$d)
Parenteral Enteral
non-protein
energy intake,
kcal/(kg$d)

Control
group

AA lipid
group

High AA
lipid group

7
5.1 (4.4e8.3)

9
4.6 (4.2e5.7)

2.4 (2.3e2.5)

2.3 (2.2e2.4)

2.5 (2.4e2.6)

2.5 (2.4e3.0)

12
4.7
(4.0e5.3)
3.6
(3.5e3.7)a,b
3.8
(3.6e3.8)a,b

25.9
(25.4e27.2)
3.2 (2.5e3.9)

27.0
(25.5e35.7)a
4.6
(3.3e12.7)
78.1 (76.9e81.7) 81.8
(76.8e111.8)
0.0 (0.0e0.0)
2.9 (2.8e3.2)a

40.3
(39.1e40.6)a,b
5.5 (4.6e6.0)
102
(117.6e122.4)
3.1 (2.6e3.4)a

5.3 (4.3e11.3)
9.4 (6.3e28.8) 6.4 (4.7e8.6)
39.7 (30.8e50.4) 62.7
59.6 (54.4e61.6)a
(57.7e75.6)a

Data are presented as median; IQR in parenthesis. AA, amino acid.

a
Signicantly different from control group (ManneWhitney U test).
b
Signicantly different from AA lipid group (ManneWhitney U test).

Fig. 1. Phenylalanine (phe) and tyrosine (tyr) kinetics on day two of life. Amino acid
kinetics were measured by primed continuous infusion of L-[1-13C]phenylalanine and
L-[ring-D4]tyrosine. Tracer enrichments were analyzed in plasma by means of GC/MS.
(A) Phe intake, hydroxylation (Hydr), phe utilization for protein synthesis (Sphe), phe
release from protein (RPphe), and tyr release from protein (RPtyr); (B) phe balance.
Boxes and whiskers indicate the medians, IQR, and 2.5th and 97.5th percentiles.

different between groups: control group (82% (56e82%));

AA lipid (68% (65e75%)); high AA lipids (67% (62e77%)) group.
This might be due to the relative low numbers.
Regression analyses showed a signicant effect of birth weight
for gestational age z-score on phenylalanine hydroxylation to
tyrosine; the hydroxylation rate was higher for infants with a
higher z-score. Correction for gestational age and birth weight did
not affect the differences in phenylalanine metabolism.
3.3. Leucine and a-KIC kinetics
The leucine model resulted in a wide range of outcomes and
these data were difcult to interpret due to various reasons (see
Discussion). Therefore, the leucine and a-KIC kinetics and comparisons between the phenylalanine and leucine model are presented in the Addendum.
4. Discussion
Our results demonstrated that a high amino acid intake in
combination with a high energy intake (lipids) increased phenylalanine hydroxylation and the use of phenylalanine for protein
synthesis in preterm infants. The release of phenylalanine from
proteins (reecting proteolysis) did not change after higher amino
acid intake, nor after early lipid administration. One may speculate
that infants in this study received sufcient energy to facilitate the
energy cost of protein synthesis. On the other hand, phenylalanine
hydroxylation increased in the high amino acid group. Phenylalanine balances were highest in the high AA lipid group (3.6 g
amino acids/(kg$d)) compared to the control group and the
AA lipid group (2.4 g amino acids/(kg$d)). The addition of only
lipids to standard-dose amino acids did not affect metabolism.
4.1. The phenylalanine model versus the leucine model
In this complicated stable isotope study we tried to combine
several aspects of intermediate metabolism. The phenylalanine and
tyrosine tracers indicate whole body metabolism as all are

986

H. Vlaardingerbroek et al. / Clinical Nutrition 33 (2014) 982e990

measured in steady state. There is no other way to dispose

phenylalanine than through protein synthesis or hydroxylation to
tyrosine. Using phenylalanine as an indicator of protein metabolism
has been validated numerous times. Using leucine as a tracer of
amino acid metabolism can be more problematic as it is in need CO2
sampling which can be problematic. In addition, one needs to
correct for CO2 retention in the body. We anticipated that the
majority of infants included in our study were ventilated with noninvasive methods, such that collection of expired air without
mixing with surrounding air was not possible. In order to measure
leucine oxidation and transamination without the use of breath
samples or assumptions about CO2 xation, we used the open twopool leucine model of Helland et al.16 However, this model was
developed in an animal study and was performed using radioactive
tracers instead of stable isotopes. The validity might be questioned
in our patient group, since many assumptions were made and this
model has not been used in published studies ever since. This is also
illustrated by the wide range of results in the leucine model and the
inconsistency in the data between the phenylalanine and leucine
models. Therefore, we presented the leucine data in the Addendum
and based the conclusions of our manuscript on the phenylalanine
data only. However, we also must realize that results from a single
amino acid do not necessarily pertain to another amino acid or to
protein metabolism in general.
4.2. Effects of nutrition on metabolism in preterm infants
In agreement with the ndings of several previous studies,6e8
our results demonstrate that preterm infants are resistant to the
suppression of proteolysis in response to parenteral amino acids.
Term infants generally increase protein synthesis rates in response
to the administration of additional energy as well.17,18 In older
children and adults,1,2 an additional energy supply results in
reduced proteolysis. It has been suggested that high rates of proteolysis in combination with even higher rates of protein synthesis
are required to support the high remodeling and growth rates in
preterm infants8 and that the primary mechanism for protein accretion in parenterally fed preterm infants is the stimulation of
protein synthesis rather than the suppression of proteolysis.7
Despite the difference in metabolic response between preterm
infants and more mature infants, the regression analyses in this
study did not show an effect of gestational age on amino acid
metabolism, although the median gestational ages were somewhat
lower in the intervention groups than in the control group. The
transition point from preterm metabolism to mature metabolism
most likely lies closer to the term age. Of course, it is of interest
whether the differences observed on the second day of life expand
to a longer period. This tracer study was primarily designed to
quantify the adaptive mechanism in intermediate metabolism due
to different nutritional intakes, and to determine the rate of
anabolism shortly following birth. Infants were deliberately studied
on postnatal day two, since nutritional differences between groups
were largest at that day. From the third day of life onwards, infants
in the control groups received lipids as well, aiming at similar intakes on day four of life. Furthermore, the situation gets more
complex by the increasing amount of enteral nutrition which can be
human milk or formula with different amino acid intakes. Further
studies should aim at measuring intermediate metabolism at later
life.
Disposal of phenylalanine by hydroxylation was signicantly
higher in the high AA lipid group (13.4 mmol/(kg$h)) than in the
AA lipid group (9.1 mmol/(kg$h)) or the control group (5.0 mmol/
(kg$h)). The higher hydroxylation rates in response to higher amino
acid intake are in agreement with those of previous studies.3,4,8
These rates are also comparable to the hydroxylation rates in

healthy human fetuses at term gestation.19 Still, the plasma tyrosine

concentrations in these infants10 were lower than the concentrations observed in healthy term breast-fed infants20 but were
comparable to previously reported ranges in preterm infants7,21,22
and in second and third trimester human fetuses.23 Sufcient hydroxylation rates are important because parenteral amino acid
solutions have a low tyrosine content due to low solubility. In fact,
the endogenous tyrosine synthesis rates (phenylalanine hydroxylation) were twice as high as tyrosine intake. However, the total
tyrosine availability (intake plus synthesis from hydroxylation) was
approximately two-thirds of the molar phenylalanine tissue
deposition (phenylalanine balance). This value is equivalent to the
molar ratio of 0.64 extracted in tissue proteins of deceased fetuses.24 These ndings suggest that hydroxylation rates are
marginally sufcient to cope with neonatal demands for protein
synthesis. It remains to be determined whether hydroxylation rates
are not increased further because of impaired enzyme activity or
insufcient phenylalanine intake.
4.3. Tracer data versus nitrogen balance data
This isotope study was part of a larger trial, which included a
total of 144 infants (47e49 infants per group), to determine the
safety and efcacy of early lipid initiation with normal or high-dose
amino acids from birth onward.10 In our study urea concentrations
were measured routinely on day 2, 4 and 6. In the larger trial, the
incidence of hyperuremia was signicantly higher in the high
AA lipid group than in the control and AA lipid groups,
resulting in lowering the amino acid dose as per protocol in 81% of
infants in the high AA lipid group, in 48% of infants in the control
group, and 39% of infants in the AA lipid group. However, the urea
thresholds set in our study were arbitrary and the validity of urea as
a monitoring tool for amino acid tolerance is questionable and is
discussed in more detail in our previous study.10 On the second day
of life, isotope studies were started early in the morning, before
urea concentrations were available. If adaptations in amino acid
dosage were necessary based on urea concentration above
10 mmol/L, these adaptations were postponed several hours until
the end of the tracer study, if possible. The nal decision to postpone adaptations to amino acid dose was made by the attending
physician. In some infants, physicians measured urea concentrations at day 1, occasionally resulting in lowering the amino acid
dose before start of the isotope study. Indeed, in those infants
metabolism might be different, or most likely is different resulting
in higher oxidation rates and thus higher urea synthesis rates at
high protein intake. However, these infants were not included in
this substudy. Demographics at baseline were not different between infants in the isotope study and the larger study. However,
the applicability of our results to the whole population of VLBW
infants might be slightly hampered by the few hours-difference in
postponement of adjustments to the amino acid dose. Hence, the
isotope study gives the effects on metabolism of administration of
exact 3.6 versus 2.4 g amino acids/kg while the larger study gives
the effects on overall outcomes of the adjusted intakes, e.g. in
practice an intake of 2.0e3.6 g amino acids/kg per day. In the
isotope study, the number of infants in whom amino acid intake
was reduced after the isotope measurements were nished was 2
in the control group, 1 in the high AA lipid group, and 7 in the
high AA lipid group. This was comparable to the relative number
of adjustments in the larger study. As stated before, the preset
guideline to lower amino acid intake based on urea concentrations
is rather arbitrary and besides, concentrations are for large part
inuenced by hydration status and kidney function, rather than
solely by amino acid oxidation.10 However, we cannot exclude that
the higher urea concentrations in the high AA lipid group may be

H. Vlaardingerbroek et al. / Clinical Nutrition 33 (2014) 982e990

due to increased oxidation of amino acids, because of any supposed

energy decit to nance the cost of protein synthesis as discussed
before, or a consequence of suboptimal amino acid composition of
the currently available amino acid solutions. The search for an
optimal composition of amino acid solutions specially designed for
preterm infants is therefore of crucial importance. Future longterm follow-up studies should demonstrate if 3.6 g amino acids/
(kg$d) can be safely implemented in clinical practice.
Although the nitrogen balances did not differ between the
AA lipid group and the high AA lipid group in our larger study,10
the present tracer study demonstrates a benecial effects of high
amino acid administration in combination with early lipids on
whole body metabolism, resulting in an improved amino acid
balance. These amino acid balances can be converted from molar
rates to grams of protein and grams of tissue based on the assumptions that 1 g of protein contains 246 mmol phenylalanine24
(Fig. S2, Supplementary data). The addition of another 1.2 g
amino acids/(kg$d) resulted in a protein balance that is approximately 1 g/(kg$d) higher, illustrating increased protein accretion
available for tissue growth. The discrepancy in results using the
nitrogen balance method and the stable isotope study can be
explained in three ways. First, it could be coincidence, because the
infants participating in the stable isotope study are a subset from
those included in the larger study on nitrogen balance. Second,
isotope studies are generally considered a more specic method to
analyze the effects of amino acid administration on metabolism.
Stable isotope techniques track the metabolic processes in amino
acid metabolism and can give information about several rates of
intermediate metabolism. The disadvantage of the tracer data is the
necessity to extrapolate data from one single amino acid to whole
protein metabolism. A single amino acid may not reect accurately
what is happening on whole body level. Furthermore, stable
isotope techniques require highly sophisticated machines and
much more lab work compared to nitrogen balance techniques.
Nitrogen balance techniques are based on measurement of end
product concentrations. The nitrogen balance is less precise than
tracer data as intakes may quite easily be overestimated, while
nitrogen excretion is per denition underestimated, since specic
N-losses (hair, skin, feces, phlebotomy) are not taken into account.
In addition, shifts in body pools of N containing substrates as
ammonia, amino acid levels are also not taken into account. Third,
in the isotope study, both enteral and parenteral nutritional intakes
were kept stable during the 8 h study period, while the intake was
changed in the majority of infants during the nitrogen balance
study. Our conclusion is that the tracer data are more reliable than
the nitrogen balance data, suggesting that higher-dose amino acid
administration to VLBW infants results in a more anabolic state
during the rst postnatal phase. However, this does not indicate
that nitrogen balances are useless. Nitrogen balances are noninvasive and can give an easy to obtain, general indication of
anabolism, especially during stable nutritional intake. Therefore,
we still recommend measurement of nitrogen balances as a basic
measurement during nutritional intervention studies.

987

comparing a 50% MCT-50% LCT lipid emulsion and a pure LCT

emulsion.25
In conclusion, anabolism is of great clinical importance in VLBW
infants, although it is often difcult to achieve in the rst days of
life. To date, not enough studies are available to determine the
upper safe and most effective level of amino acid and lipid
administration in very low birth weight infants. This study adds to
the body of evidence that amino acid supplementation up to a level
of 3.6 g/(kg$d) seems to result in a more anabolic state in this group
of infants, which is usually considered as benecial; although it also
resulted in higher urea concentrations, which reects a higher
amino acid oxidation. The higher urea concentrations observed in
several studies, including our own, reect a higher amino acid
oxidation rate, which has not shown to be correlated with any
detrimental outcome in this population. The value of measuring
urea concentrations in a routine fashion remains questionable. That
higher protein synthesis rates require higher amounts of other
substrates, such as micronutrients and energy is underscored in the
present study and in others.26 Our planned follow-up study at two
years of age will reveal whether high amino acid and energy intake
in the rst few years has any benets on the long run.
Source of funding
Funding was not received for this project.
Statement of authorship
HV designed and conducted the research, analyzed the data, and
drafted the manuscript. JR conducted the research, and drafted the
manuscript. DR conducted the research, and drafted the manuscript. KD and HS analyzed the data. AV manufactured the stable
isotope preparations. MV drafted the manuscript. JvG designed the
research, and drafted the manuscript. CvdA designed the research,
and drafted the manuscript.
All of the authors read and approved the nal manuscript.
Conict of interest
None of the authors has a conict of interest with regard to the
manuscript.
Acknowledgments
We would like to thank Aimon Niklasson for calculation of the
growth SD scores.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.clnu.2014.01.002.
Addendum. Leucine model

4.4. Potential effects of type of lipid

Methods
Infants in the early lipid groups were randomized to two
different lipid types, a pure soybean oil emulsion or a multicomponent emulsion containing soybean oil, medium-chain triacylglycerol, olive oil, and sh oil. Type of lipid had no effect on
amino acid kinetics. However, due to the small number of infants in
the isotope study, possible effects might have been missed,
although in the larger study10 type of lipid had no effect on nitrogen
balances and urea kinetics either. Studies on the effect of lipid type
on protein metabolism are lacking, except for a single study

Calculations
Leucine is reversibly transaminated within the cell to its ketoacid, a-ketoisocaproic acid (KIC). The plasma enrichment of
[U-13C6]KIC is therefore very close to the intracellular [U-13C6]
leucine enrichment.27 In our model, various leucine isotopomers
were available in plasma. To adjust for the contribution of the
available isotopomers, the plasma enrichments of leucine were
calculated utilizing all of the measured molecular forms of leucine,

988

H. Vlaardingerbroek et al. / Clinical Nutrition 33 (2014) 982e990

xleucine enrichment

xleucine
U  12 C6 leucine U  13 C6 ; 15 Nleucine U  13 C6 leucine 15 Nleucine D3 leucine

by the following formula28:where [x]leucine can be each of the

specic tracer enrichments.
To measure amino acid oxidation, most models measure labeled
carbon dioxide in expired air. However, we used two carbon-

Oxidationmmol=kg$h

Utilization of leucine for protein synthesis (Sleu; mmol/(kg$h))

iU13 C6 ;15 Nleucine  TTRD3 KIC  iD3 KIC  TTRU13 C6 mean



TTRU13 C6 mean  TTRD3 KIC  TTRD3 leucine

iD3 KIC  TTRU13 C6 mean  iU13 C6 ;15 Nleucine  TTRD3 leucine



TTRU13 C6 mean  TTRD3 KIC  TTRD3 leucine

Deaminationmmol=kg$h

iD3 KIC  TTRU13 C6 mean



TTRU13 C6 mean  TTRD3 KIC  TTRD3 leucine

Reaminationmmol=kg$h

iU13 C6 ;15 Nleucine  TTRD3 leucine



TTRU13 C6 mean  TTRD3 KIC  TTRD3 leucine

labeled amino acids simultaneously. Most of the infants in our

study were ventilated with non-invasive methods on day 2 such
that the collection of expired air without mixture with surrounding
air was not possible. We therefore used the alternative open twopool model (dual-isotope model) of Helland et al.,16 which
allowed us to measure leucine oxidation and transamination
without the use of breath samples or assumptions about CO2 xation. This model is a modied version of the model proposed by
Shipley and Clark,29 adapted to leucine metabolism.28,30 Because
these models use radioactively labeled isotopes instead of stable
isotopes, enrichments in the following equations are expressed as
tracer-tracee ratios (TTR, %) instead of mole percent excess.
Release of leucine from protein (RPleu; reecting proteolysis;
mmol/(kg$h))

iU13 C6 ;15 Nleucine

TTRU13 C6 mean

Fluxes (Q) were calculated as follows:

Q leu mmol=kg$h RPleu reamination intake

Sleu deamination
Q KIC mmol=kg$h deamination
oxidation reamination
Leucine balancemmol=kg$h Sleu  RPleu
The leucine balance was converted from molar rates to grams of
protein under the assumption that 1 g protein contains 590 mmol
leucine.31

!
 intakeleucine

where i is the tracer infusion rate in mmol/(kg$h) and

TTRU13 C6 mean is the mean TTR of [U-13C6]KIC and the total sum of
[U-13C6,15N]leucine plus [U-13C6]leucine. This calculation is based
on the assumption that the only source of KIC is from the deamination of leucine within the twopool system. As such, these three
values should be equal, but discrepancies are averaged to improve
model and measurement accuracy, thereby omitting a non-existing
ux.

Results
Leucine and a-KIC kinetics
Leucine and a-KIC enrichments at steady state are presented in
Table 3. In the high AA lipid group, the utilization of leucine for
protein synthesis (Sleu) was signicantly higher than in the
AA lipid group (Fig. A1(A)), but its use was not signicantly
higher than in the control group. The release of leucine from protein (RPleu; reecting proteolysis), oxidation, deamination, reamination, and whole-body uxes (Fig. A1(B)) did not differ among the
groups. The leucine balance was signicantly higher in the high

H. Vlaardingerbroek et al. / Clinical Nutrition 33 (2014) 982e990

AA lipid group than in the AA lipid group, but was not significantly different from the value in the control group (Fig. A1(C)).

Fig. A1. Leucine kinetics on day two of life. Amino acid kinetics were measured by primed
continuous infusion of L-[U-13C6,15N]leucine, and L-[methyl-D3]a-ketoisocaproic acid.
Tracer enrichments were analyzed in plasma by means of GC/MS. (A) Leucine utilization for
protein synthesis (Sleu), leucine release from protein (RPleu), and leucine oxidation (oxid);
(B) leucine deamination (deam) and a-ketoisocaproic acid reamination (ream); (C) leucine
balance. Boxes and whiskers indicate the medians, IQR, and 2.5th and 97.5th percentiles.

Correlations and regression analyses

Positive correlations were observed between the rates of
phenylalanine and leucine utilization for protein synthesis
(R 0.586, P 0.001), the rates of phenylalanine and leucine
release from protein (R 0.853, P < 0.001), and the amino acid
balances based on phenylalanine and leucine data (Fig. A2;
R 0.476, P 0.010).
Control group
AA+lipid group
High AA+lipid group

150

R2 Linear = 0.181

Leucine balance, mol/(kgh)

100

50

0
10

15

20

25

30

35

Phenylalanine balance, mol/(kgh)

-50

-100

Fig. A2. Comparison between the phenylalanine and leucine balance.

989

References
1. Giordano M, Castellino P, DeFronzo RA. Differential responsiveness of protein
synthesis and degradation to amino acid availability in humans. Diabetes 1996
Apr;45(4):393e9.
2. Melville S, McNurlan MA, McHardy KC, Broom J, Milne E, Calder AG, et al. The
role of degradation in the acute control of protein balance in adult man: failure
of feeding to stimulate protein synthesis as assessed by L-[1-13C]leucin infusion. Metabolism Clin Exp 1989 Mar;38(3):248e55.
3. Denne SC, Karn CA, Ahlrichs JA, Dorotheo AR, Wang J, Liechty EA. Proteolysis
and phenylalanine hydroxylation in response to parenteral nutrition in
extremely premature and normal newborns. J Clin Invest 1996 Feb 1;97(3):
746e54.
4. Poindexter BB, Karn CA, Ahlrichs JA, Wang J, Leitch CA, Liechty EA, et al. Amino
acids suppress proteolysis independent of insulin throughout the neonatal
period. Am J Physiol 1997 Apr;272(4 Pt 1):E592e9.
5. Tappy L, Acheson K, Normand S, Schneeberger D, Thelin A, Pachiaudi C, et al.
Effects of infused amino acids on glucose production and utilization in healthy
human subjects. Am J Physiol 1992 Jun;262(6 Pt 1):E826e33.
6. van den Akker CH, te Braake FW, Wattimena DJ, Voortman G, Schierbeek H,
Vermes A, et al. Effects of early amino acid administration on leucine and
glucose kinetics in premature infants. Pediatr Res 2006 May;59(5):732e5.
7. Thureen PJ, Melara D, Fennessey PV, Hay Jr WW. Effect of low versus high
intravenous amino acid intake on very low birth weight infants in the early
neonatal period. Pediatr Res 2003 Jan;53(1):24e32.
8. Poindexter BB, Karn CA, Leitch CA, Liechty EA, Denne SC. Amino acids do not
suppress proteolysis in premature neonates. Am J Physiol Endocrinol Metab
2001 Sep;281(3):E472e8.
9. Thureen PJ, Hay Jr WW. Intravenous nutrition and postnatal growth of the
micropremie. Clin Perinatol 2000 Mar;27(1):197e219.
10. Vlaardingerbroek H, Vermeulen MJ, Rook D, van den Akker CH, Dorst K,
Wattimena JL, et al. Safety and efcacy of early parenteral lipid and high-dose
amino acid administration to very low birth weight infants. J Pediatr
2013;163(3):638e644.e5.
11. Niklasson A, Albertsson-Wikland K. Continuous growth reference from 24th
week of gestation to 24 months by gender. BMC Pediatr 2008;8:8.
12. The CRIB (clinical risk index for babies) score: a tool for assessing initial
neonatal risk and comparing performance of neonatal intensive care units. The
International Neonatal Network. Lancet 1993 Jul 24;342(8865):193e8.
13. van den Akker CH, Schierbeek H, Minderman G, Vermes A,
Schoonderwaldt EM, Duvekot JJ, et al. Amino acid metabolism in the human
fetus at term: leucine, valine, and methionine kinetics. Pediatr Res 2011
Dec;70(6):566e71.
14. Clarke JT, Bier DM. The conversion of phenylalanine to tyrosine in man. Direct
measurement by continuous intravenous tracer infusions of L-[ring-2H5]
phenylalanine and L-[1-13C] tyrosine in the postabsorptive state. Metabolism
Clin Exp 1982 Oct;31(10):999e1005.
15. Thompson GN, Pacy PJ, Merritt H, Ford GC, Read MA, Cheng KN, et al. Rapid
measurement of whole body and forearm protein turnover using a [2H5]
phenylalanine model. Am J Physiol 1989 May;256(5 Pt 1):E631e9.
16. Helland SJ, Ewan RC, Trenkle A, Nissen S. In vivo metabolism of leucine and
alpha-ketoisocaproate in the pig: inuence of dietary glucose or sucrose. J Nutr
1986 Oct;116(10):1902e9.
17. Pencharz P, Beesley J, Sauer P, Van Aerde J, Canagarayar U, Renner J, et al. Totalbody protein turnover in parenterally fed neonates: effects of energy source
studied by using [15N]glycine and [1-13C]leucine. Am J Clin Nutr 1989;50(6):
1395e400.
18. Van Aerde JE, Sauer PJ, Pencharz PB, Smith JM, Heim T, Swyer PR. Metabolic
consequences of increasing energy intake by adding lipid to parenteral nutrition in full-term infants. Am J Clin Nutr 1994 Mar;59(3):659e62.
19. van den Akker CH, Schierbeek H, Dorst KY, Schoonderwaldt EM, Vermes A,
Duvekot JJ, et al. Human fetal amino acid metabolism at term gestation. Am J
Clin Nutr 2009 Jan;89(1):153e60.
20. Scott PH, Sandham S, Balmer SE, Wharton BA. Diet-related reference values for
plasma amino acids in newborns measured by reversed-phase HPLC. Clin Chem
1990;36(11):1922e7.
21. Clark RH, Chace DH, Spitzer AR. Effects of two different doses of amino acid
supplementation on growth and blood amino acid levels in premature neonates admitted to the neonatal intensive care unit: a randomized, controlled
trial. Pediatrics 2007 Dec;120(6):1286e96.
22. Blanco CL, Gong AK, Green BK, Falck A, Schooleld J, Liechty EA. Early
changes in plasma amino acid concentrations during aggressive nutritional
therapy in extremely low birth weight infants. J Pediatr 2011 Dec 1;158:
543e548.e1.
23. Cetin I, Corbetta C, Sereni LP, Marconi AM, Bozzetti P, Pardi G, et al. Umbilical amino acid concentrations in normal and growth-retarded fetuses
sampled in utero by cordocentesis. Am J Obstet Gynecol 1990 Jan;162(1):
253e61.
24. Widdowson EM. Chemical composition and nutritional needs of the fetus at
different stages of gestation. In: Aebi H, Whitehead R, editors. Maternal nutrition
during pregnancy and lactation: a Nestl foundation workshop, Lutry/Lausanne,
April 26th and 27th 1979. Bern, Switzerland: Hans Huber; 1980. pp. 39e48.
25. Liet JM, Piloquet H, Marchini JS, Maugere P, Bobin C, Roze JC, et al. Leucine
metabolism in preterm infants receiving parenteral nutrition with medium-

990

H. Vlaardingerbroek et al. / Clinical Nutrition 33 (2014) 982e990

chain compared with long-chain triacylglycerol emulsions. Am J Clin Nutr 1999

Mar;69(3):539e43.
26. Moltu SJ, Strommen K, Blakstad EW, Almaas AN, Westerberg AC, Braekke K,
et al. Enhanced feeding in very-low-birth-weight infants may cause electrolyte
disturbances and septicemia e a randomized, controlled trial. Clin Nutr 2013
Sep 21;32(2):207e12.
27. Schwenk WF, Beaufrere B, Haymond MW. Use of reciprocal pool specic activities to model leucine metabolism in humans. Am J Physiol 1985 Dec;249(6
Pt 1):E646e50.

28. Tessari P, Tsalikian E, Schwenk WF, Nissen SL, Haymond MW. Effects of [15N]
leucine infused at low rates on leucine metabolism in humans. Am J Physiol
1985 Jul;249(1 Pt 1):E121e30.
29. Shipley RA, Clark RE. Tracer methods for in vivo kinetics: theory and applications.
New York: Academic Press Inc; 1972.
30. Nissen S, Haymond MW. Effects of fasting on ux and interconversion of
leucine and alpha-ketoisocaproate in vivo. Am J Physiol 1981 Jul;241(1):E72e5.
31. Waterlow JC, Garlick PJ, Milward DJ. Protein turnover in mammalian tissues and
in the whole body. Amsterdam: North Holland Publishing; 1978. pp. 225e300.