Invited critical review
The role of metabolomics in neonatal and pediatric laboratory medicine
Michele Mussapa, Roberto Antonuccib, Antonio Notoc, Vassilios Fanosc,d,⁎
aLaboratory Medicine Service, IRCCS AOU San Martino-IST, University-Hospital, National Institute for Cancer Research, Genova, Italy
bDivision of Neonatology and Pediatrics “Nostra Signora di Bonaria” Hospital, San Gavino Monreale, Cagliari, Italy
cNeonatal Intensive Care Unit, Puericulture Institute and Neonatal Section, Azienda Ospedaliera Universitaria, Cagliari 09131, Italy
dDepartment of Surgical Sciences, University of Cagliari, Cagliari 09131, Italy
a b s t r a c ta r t i c l ei n f o
Received 2 July 2013
Received in revised form 26 August 2013
Accepted 26 August 2013
Available online 11 September 2013
Proton nuclear magnetic resonance
Metabolomics consists of the quantitative analysis of a large number of low molecular mass metabolites involv-
ing substratesor products inmetabolic pathways existing inall living systems. The analysis ofthemetabolic pro-
file detectable in a human biological fluid allows to instantly identify changes in the composition of endogenous
and exogenous metabolites caused by the interaction between specific physiopathological states, gene expres-
sion, and environment. In pediatrics and neonatology, metabolomics offers new encouraging perspectives for
the improvement of critically ill patient outcome, for the early recognition of metabolic profiles associated
with the development of diseases in the adult life, and for delivery of individualized medicine. In this view,
nutrimetabolomics, based on the recognition of specific cluster of metabolites associated with nutrition and
due to therapeutic treatment may open new frontiers in the prevention and in the treatment of pediatric and
neonatal diseases. This review summarizes the most relevant results published in the literature on the applica-
tion of metabolomics inpediatric and neonatalclinicalsettings. However,thereisthe urgentneedtostandardize
physiological and preanalytical variables, analytical methods, data processing, and result presentation, before
establishing the definitive clinical value of results.
© 2013 Elsevier B.V. All rights reserved.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Metabolomics: historical background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Metabolomics workflow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.Targeted and non-targeted approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.Current metabolomics technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3. Biological samples for searching the metabolic profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.Metabolomic data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Clinical metabolomics in neonatology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.Intrauterine growth restricted (IUGR) and small for gestational age (SGA) neonates . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.Prematurity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.Mode of delivery and newborn twins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.4.Perinatal asphyxia and hypoxic ischemic encephalopathy (HIE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Clinica Chimica Acta 426 (2013) 127–138
ney disease; AKI, acute kidney injury; AMP, adenosine monophosphate; BALF, bronchoalveolar lavage fluid; BMI,bodymass index; CE, capillary electrophoresis; CKD, chronic kidney dis-
infrared spectroscopy; GC–MS, gas chromatography–mass spectrometry; HIE, hypoxic ischemic encephalopathy; HMDB, human metabolome data base; HVA-SO4, homovanillic acid sul-
fate; ICU, intensive care unit; IQ, intelligence quotient; IUGR, intrauterine growth restricted; LBW, low birth weight; LC–MS, liquid chromatography–mass spectrometry; MDA,
malondialdehyde; MMA, methylmalonic acidemia; MS, mass spectrometry; OPLS-DA, orthogonal projections to latent structures for discriminant analysis; PA, propionic acidemia;
PCA, principal component analysis; PDA, patent ductus arteriosus (Botallo's duct); PLS-DA, projection to latent structures discriminant analysis; RCDs, respiratory chain deficiencies;
RDS, Respiratory Distress Syndrome (hyaline membrane disease); SGA, small for gestational age; TMAO, trimethylamine-N-oxide; VLBW, very low birth weight; VOCs, volatile organic
⁎ Corresponding author. Tel.: +39 70 6093403; fax: +39 70 6093495.
E-mail address: firstname.lastname@example.org (V. Fanos).
0009-8981/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
Clinica Chimica Acta
journal homepage: www.elsevier.com/locate/clinchim
Clinical metabolomics in children
5.1. Pediatric age and metabolomics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2. Respiratory tract diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3. Neurological diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4.Kidney diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.5.Diabetes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.6. Celiac disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.7. Cystic fibrosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.8. Obesity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.9. Inborn errors of metabolism. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Nutrimetabolomics in childhood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Pharmametabolomics in childhood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Respiratory Distress Syndrome (RDS) and surfactant therapy
Patent ductus arteriosus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Over recent decades, the contributions of laboratory medicine in
improving patient care have increasingly become essential, because of
the advent of new generation of laboratory diagnostics consistingof so-
phisticated tests withpotentially profoundimplications for the delivery
of personalized health care. The “genomic revolution” started with the
mappingof theentire sequenceof human genoma[1,2]and accelerated
the development of system biology studies based on various disciplines
like genomics, transcriptomics and proteomics. These “omics” may be
considered the most relevant driver in changing the face of laboratory
medicine, opening new challenges for patients, clinicians, and clinical
pathologists. These perspectives seem to be of extreme importance in
neonatal care to improve diagnosis, prognosis, and clinical outcome
in critically ill newborns by the routine application of emerging molec-
ular technologies. In addition, such disciplines seem to open new hopes
of them is specifically associated with a severe pathological condition
(sepsis, acute kidney injury, etc.). The identification of these “clusters”
of substances could lead to the development of very low-cost devices
like a dipstick, easily usable in low-income countries, representing an
excellent example of how to convert translational research results in
low-cost medical device .
2. Metabolomics: historical background
The presence of a metabolic pattern extremely variable between
individuals, but relatively constant for a given individual, was firstly re-
ported by Roger Wlilliams in 1951 . By examining data from over
200,000 paper chromatograms belonging to a variety of subjects, in-
cluding alcoholics, schizophrenics, and residents of mental hospitals,
Williams found characteristic metabolic patterns associated with each
of these groups. Despite these results giving rise to the idea that the
metabolic pattern analysis mighthave clinical utility,nofurther investi-
gation was performed until the '70s, when a boost came by the rapid
increase of technology progresses in gas chromatography (GC), liquid
chromatography (LC), and mass spectrometry (MS) methods. In 1971,
theterm “metabolic profile” wasintroduced,with theintentto describe
chromatographic patterns observed in biological fluids characterizing
both normal and pathologic states and potentially useful for studies of
drug metabolism and for human developmental studies . Simul-
taneously, Linus Pauling realized that information-rich data reflecting
thefunctionalstatus ofa complex biologicalsystemresides inthequan-
titative and qualitative pattern of metabolites in body fluids . This
emerging approach to the quantitative metabolic profiling of large
numbers of small molecules in biofluids was experimented by several
research groups  and ultimately was termed metabonomics by
Nicholson et al. in 1999  and metabolomics by Fiehn in 2002 .
Metabolic profiles include endogenous and exogenous chemical enti-
ties such as peptides, amino acids, nucleic acids, carbohydrates, fatty
acids, organic acids, vitamins, hormones, drugs, food additives, phyto-
chemicals, toxins, and other chemicals ingested or synthesized by a
cell or organism. This heterogeneous multitude of molecules was called
metabolome by Oliveret al. : the metabolome is theholistic quanti-
tative set of low molecular-weight compounds (b1000 Da). The first
draft of the human metabolome was completed on January 23, 2007
and consisted of a database of about 2500 metabolites . Despite
the very high overall number of endogenous metabolites (~100,000),
the number of major metabolites relevant for clinical diagnostics
and drug development has been estimated at 1400–3000 molecules
, which means less data to manipulate and interpret, being genes
(~25,000), transcripts (~85,000), and proteins (N10,000,000) greatly
outnumbered. Most endogenous metabolites are tied to specific bio-
chemical pathways such as glycolysis, Krebs' cycle, lipid or amino acid
metabolism, signaling pathways such as transmitters and hormones
and specific pathobiochemical processes like oxidative stress. Thus,
changes in specific metabolite patterns reflect changes in pathways
and processes; it is reasonable to argue that metabolome is typically
more closely associated with a disease process or drug effect than pro-
teins, mRNA or genes .
3. Metabolomics workflow
3.1. Targeted and non-targeted approach
The metabolomics workflow includes sample preparation, analysis
using various instruments, data processing and data analysis, as con-
densed in Fig. 1. The power of metabolomics lies on the acquisition of
analytical data in which metabolites in a cellular system are quantified,
various data analysis tool. Two strategies configure metabolomics stud-
ies: the targeted and the non-targeted approach . The latter may
be defined asa “nonspecific approach”, investigatingall the metabolites
(both endogenous and exogenous) detectable in a fluid or tissue; this
analysis is focused on capture as much information as possible, provid-
the body. The former is focused on the investigation of several well de-
fined compounds (for example those discovered in a new metabolic
pathway); it is only used when the target of a drug or disease process
is at least partially understood. Metabolic fingerprinting describes the
unbiased analysis of the metabolome by examination of metabolite
patterns in different experimental groups with the subsequent classi-
fication of these patterns into a fingerprint. Samples can be classified
clustering. Metabolite identification relies on public databases : the
human metabolome Data Base (HMDB) is the metabolomic equivalent
M. Mussap et al. / Clinica Chimica Acta 426 (2013) 127–138
of GenBank. It is an open access database (http://www.hmdb.ca)
providing reference to nuclear magnetic resonance (NMR) and mass
spectra, metabolite disease associations, metabolic pathway data and
referenceto metabolite concentrationsfor hundreds of human metabo-
lites from several biofluids .
3.2. Current metabolomics technologies
The most common used platforms for the detection and measure-
ment of metabolites involve their separation by GC, LC or capillary elec-
trophoresis (CE) coupled with subsequent MS, enabling to analyze a
large number of metabolites simultaneously . MS is still considered
on the metabolite, the sensitivity of MS is in the picomolar and nano-
molar range. Basically, liquid chromatography coupled with mass spec-
trometry (LC/MS) is rapidly evolving as a method of choicefor chemical
phenotyping of biological systems . Proton nuclear magnetic reso-
nance (1H NMR) spectroscopy is also considered a major tool to analyze
a large number of metabolites (up to 20–50) simultaneously [19,20].
A comparison between1H NMR and MS techniques has been reported
in Table 1. Other technologies less commonly used for metabolomics
are Raman and infrared spectroscopy [21,22]; each of them has serious
drawbacks, such that neither by itself is ideal. Although the technology
is highly sophisticated and sensitive, some bottlenecks remain in
metabolomics. Due to the huge diversity of chemical structures and
the large differences in abundance, there is no single technology avail-
able to analyze the entire metabolome. Therefore, a number of comple-
mentary approaches have to be established for extraction, detection,
quantification, and identification of as many metabolites as possible. In
definitive, the selection of the most appropriate methodology can be
considered as a compromise between the chemical selectivity, sensitiv-
ity and speed of the different techniques .
3.3. Biological samples for searching the metabolic profile
In general, biological fluids are considered highly adequate for
metabolomics, because they closely represent quantitative and qualita-
tive variations of phenotypic molecular markers such as metabolites.
In neonatal and pediatric nephrology, however, urine is considered
the ideal sample, since it is a so-called “proximal matrix”, being closer
to (or in direct contact with) the kidney, that is the site of disease or
drug effect under investigation . This means that urinemetabolome
better reflects kidney pathophysiological changes, while metabolome
in whole blood, plasma, and serum better reflects systemic changes.
Fig. 1. The cycle of the metabolomics workflow (CSF = cerebro-spinal fluid).
Comparison between the approaches commonly used in metabolomics studies.
Technique BenefitsDrawbacks Cost
1H NMR1. Results in a single experiment
3. Availability to study various nuclei (1H,13C,19F,31P)
4. Consistent with liquid and solid matrices
6. Sample can be recovered after analysis
1. Very high accuracy and repeatability of results
2. Very small amount of sample is required
3. High discrimination ability between molecules with a
very similar structure
4. Very high sensitivity (b1 μmol/L)
1. Suitable for measuring lipids, di- and tripeptides, and
2. Very high sensitivity (b1 μmol/L)
1. The presence of very skilled technicians and
statisticians is essential
2. Limitations in sensitivity (1–10 μmol/L)
1. NMR spectrometers are very expensive
2. Very low cost-per-test
(absence of any reagent)
1. Extensive sample preparation, including
1. GC–MS is moderately expensive
2. Low cost-per-test
1. Extensive sample preparation, including
2. Long analytical time (20–60 min per sample)
3. Limitations to volatile compounds
1. LC–MS is moderately expensive
2. Low cost-per-test
Abbreviations:1H NMR = proton nuclear magnetic resonance spectroscopy; GC–MS = gas chromatography–mass spectrometry; LC–MS = liquid chromatography–mass spectrometry.
M. Mussap et al. / Clinica Chimica Acta 426 (2013) 127–138
Furthermore, urine represents an “open system” by which the body
through the elimination of water, ions, metabolic degradation, harmful
or toxic substances, regulates important balance, maintaining homeo-
stasis. It is also of importance that the urine metabolome includes the
intermediate metabolites, which reflects specific metabolic processes.
Finally, urine can be collected easily (a spot sample is adequate) and
non-invasively: these aspects are of extreme importance in neonatolo-
gy, especially for preterm low birth weight (LBW) babies. Two condi-
tions are essential to perform metabolomics studies on urine samples:
bacteria metabolism significantly interferes on the urine metabolome.
Secondly, urine samples must be frozen at −80 °C immediately after
collection, until analysis . A valid alternative to urine sample might
be dried blood spot sample. Currently, the analysis of biological fluids
for metabolomics is most often a process centered on solvent precip-
itation for protein removal . The use of dried blood spot samples,
collected on a suitable inert paper matrix, is an alternative method of
sample collection; it has been suggested as a convenient method for
metabolic profiling allowing an easier collection, storage and transport
of samples . Interestingly, dried blood spot sample does not require
a number of preanalytical steps [sample preparation, storage, freezing
and then transportation on dry ice], offering huge benefits for the
simplification of study conduct . Several studies have shown the
potential utility of blood spots for global metabolic profiling with the
metabolite profiles obtained from blood spots being similar to those of
protein-precipitated plasma [29,30].
3.4. Metabolomic data analysis
Typically, metabolomics produces large amounts of data. Handling
and analyzing each data sets have a great impact on the quality of the
identification and quantification of putative low-mass regulators, and
therefore to the resulting biological interpretation. Metabolomic data
are analyzed by two major approaches: chemometric approaches and
quantitative approaches . Chemometric approaches can be applied
to data acquired by NMR, Fourier transform infrared spectroscopy
(FTIR) and direct injection mass spectrometry (DIMS). Quantitative
metabolomics (or targeted profiling) aims to formally identify and
quantify all detectable metabolites from the spectra, prior to subse-
quent data analysis. This approach requires that the compounds of
interest to be known a priori. With the availability of several compre-
hensive metabolomic databases and metabolome libraries, quantitative
metabolomics is becoming increasingly common . The choice of the
data analysis strategy depends heavily on the questions that are asked
. Metabolomic commercial solutions include built-in statistical
packages for data analysis. Common chemometric tools such as princi-
pal component analysis (PCA) are generally proposed for display and
exploratory analysis purposes, whereas univariate statistical tests such
as the Student's t-test are used to identify the relevant variables .
Table 2 is a scheme of the metabolomics data mining.
4. Clinical metabolomics in neonatology
4.1. Intrauterine growth restricted (IUGR) and small for gestational age
LBW, defined as a birth weight of less than 2500 g, is a well-
documented risk factor for perinatal mortality and morbidity . Con-
cisely, birthweightis a critical indicator of prenataldevelopmentalcon-
ditions, being related to long-term risk of cardiovascular disease, type 2
diabetes, and metabolic syndrome . In 1973, Forsdahl initiated the
theory on the influence of the fetal environment in the etiology of car-
diovascular disease in adulthood . Later, this theory was confirmed
by experimental studies  and the concept that developmental pro-
gramming of adult disease occurs in response to an imbalance during
fetal life between fetal demands and nutrient supply was developed,
resulting in fetal undernutrition . In other words, a lifetime risk for
obesity, diabetes, high blood pressure, heart disease, liver and kidney
function is programmed before birth. Impairment in fetal development,
which can be marked by intrauterine growth restriction (IUGR) and
LBW, results from these fetal adaptations to an adverse fetal environ-
ment leading to molecular and physiological adaptive changes .
On the basis of the ‘fetal programming theory’, these diseases
originate vascular, metabolic or endocrine adaptations resulting from
malnutrition. Although these adaptive changes improve the chance of
survival of the fetus, their persistence throughout postnatal life may
predispose to adult diseases. Clinical metabolomics has been recently
proposed as a tool to investigate individual phenotypes and to advance
hypotheses on thegenesis of diseases.Severalexperimental and clinical
studies [41–45] investigated the influence of LBW originating from
IUGR on metabolomic profiling in blood from umbilical cord, in mater-
nal venous peripheral blood, and in blood from animal models, as sum-
marized in Table 3. Starting at birth, metabolic abnormalities provide
information on the impact of intrauterine life, suggesting that prenatal
development plays a major role in assessing the risk of metabolic dis-
placental amino acid transport and IUGR. Metabolomics has the ability
to discriminate different populations without overlapping results: in a
group of very low birth weight (VLBW) babies we investigated their
urinemetabolic profile comparingresults between babies wholater de-
veloped bronchopulmonary dysplasia (BPD) with controls (Fig. 2); the
urine metabolite cluster of the earlier was totally different from that of
the latter, confirming no false positive as well as false negative result
Premature birth is associated with metabolic adaptations. Meta-
bolomics can offer a sensitive means to identify the specific biological
pathways affected, and may lead to the identification of biomarkers
for interventional studies to ameliorate the consequences of preterm
birth. For example, screening body fluids at birth using metabolomic
Metabolomics data mining.
Group or Class:
Multivariate statistical analysis:
People, phenomenon, process
Set of objects having similar features
Set of variables describing the object
Unsupervised classification (i.e., PCA) It can be used if no information about classes/groups are known
It can be used if the classes in the data set are known but not
what class and object belong to
It can be used if we know the classes and objects
It looks for analogies in objects
It looks for analogies in objects
Supervised classification (i.e., PLS-DA, OPLS-DA) It can calibrate a classification model using the variance
which maximizes the separation between classes
Abbreviations: PCA = principal component analysis; PLS-DA = projection to latent structures for discriminant analysis; OPLS-DA = orthogonal projections to latent structures for dis-
M. Mussap et al. / Clinica Chimica Acta 426 (2013) 127–138
analysis can differentiate urinary metabolic profiles in neonates with
different gestational ages, identifying any individual discriminating me-
tabolite. The clinically most relevant studies on changes in metabolic
profile associated with prematurity [46–48] are showed in Table 4. An
increase in the N-acetyl signals from the glycoproteins in the blood of
preterm mothers has been demonstrated: these signals, arising from
acute phase glycoproteins , may reflect inflammatory status. Impor-
low birth weight (VLBW) plasma compared with full term plasma. It is
well known that amino acid concentrations are significantly higher in
fetal than in maternal blood, reflecting the presence of active transport
systems within the placenta . In VLBW, placental amino acid trans-
port can be reduced due to either impaired fetal and placental growth
or to decreased transporter concentrations. Alterations in fetal energy,
antioxidant defense, and polyamines and purines flux have been identi-
fied as a metabolic fingerprint of premature birth .
4.3. Mode of delivery and newborn twins
Over the last 30 years, the number of babies born by elective cesar-
ean section increased dramatically . Currently, there is a great de-
bate about the relationship between the mode of delivery and the
development of certain diseases in adulthood. Our preliminary analysis
of1H NMR spectra in urine at birth identified different metabolic pro-
files between newborns born spontaneously and those born by elective
cesarean section (unpublished data), suggesting that the metabolite
pact of fetal stress duringlabor is reflected by changes in themetabolite
profile of the umbilical cord blood . Eighty-one metabolites were
detected in theserumof umbilical cord belongingto 41infants grouped
according to the method of delivery. In babies born by vaginal delivery
without medication, 3-hydroxybutyric acid and isoleucine were found
statistically significant lower than in babies born by elective cesarean
section, while fructose, mannose, glucose, allose, glucuronic acid, inosi-
tol, and cystine were found statistically significant higher. In addition,
fructose, mannose, glucose, allose, and glucuronic acid were found sta-
tiparous mothers. Therefore, the stress associated with labor seems to
be involved in alterations in the levels of metabolites, particularly
saccharides, in umbilical cord blood . Little is known about meta-
bolic profiles in the newborn twins. Preliminary studies suggest that
ic profile at birth. The most relevant urinarymetabolites are galactitol, N-
dimethylformamide, and 5-hydroxyindol-3-acetate . The heritability
of metabolic profiles in newborns has been determined in 107
Main studies investigating metabolomics profile changes in IUGR LBW babies.
Clinical condition Biolog. sampleMethod Most relevant findings YearRef.
LBW (b10th percentile)Urine
1H NMRUrine myo-inositol significantly increased in IUGR babies within 34 weeks
of gestation compared with controls (p b 0.05)
A positive relationship between plasma glucose level and birth weight and
a negative relationship between myo-inositol and D-chiroinositol plasma
levels on the one hand and birth weight on the other hand in the naturally
occurring pig model of IUGR
Significant differences in relative levels of essential amino acids (PHE, TRP, MET)
between IUGR and AGA; PHE and TRP cut-off levels discriminate IUGR from AGA
Metabolomic analysis may provide a predictive early screening test for SGA.
Multivariate analysis by cross-validated PLS-DA of all 3 studies showed a
comprehensive and similar disruption of metabolism. A 19 named metabolomic
signature of presymptomatic SGA has been uncovered. The final panel of
metabolites proved effective at discriminating SGA plasma from controls in both
the presymptomatic week-15 and the venous cord plasma data, giving an OR for
developing SGA of 44, with an area under the ROC curve of 0.9. A number of
sphingolipids were among this panel of metabolites; phospholipids also showed
significant disruption. A number of the putatively identified metabolites in both
cord and RUPP plasma demonstrated disruption in carnitine metabolism
PRO, ALA, GLN, choline, and glucose were reduced in LBW babies as compared to
those in controls. LBW newborns exhibited an increase of 15.7% in PHE level and
of 20.2% in CIT level.
The metabolomic profiling in maternal blood showed no significant differences
between the two groups of mothers
Experimental animal model: IUGR in the pigUmbilical cord
IUGR compared with AGA babiesUmbilical cord
1. Time-of-disease biomarker discovery.
Normal and SGA babies
2. Biomarker validation. Rat model of
3. Validation of biomarkers in a presymptomatic
clinical setting. Women delivering SGA babies
LBW babies and their mothersUmbilical cord
Abbreviations: LBW = low birth weight; IUGR = intrauterine growth restricted;1H NMR = proton nuclear magnetic resonance spectroscopy; GC–MS = gas chromatography–mass
spectrometry; LC–HRMS = liquid chromatography–high resolution mass spectrometry; AGA = appropriate for gestational age; UPLC–MS = ultra‐high‐performance liquid
chromatography–mass spectrometry; PLS-DA = Partial Least Squares Discriminant Analysis; OR = Odds Ratio; ROC = Receiver Operator Characteristic; RUPP = reduced uterine per-
fusion pressure; PHE = phenylalanine; TRP = tryptophan; MET = methionine; CIT = citrulline; GLN = glutamine; ALA = alanine.
Abbreviations: VLBW = very low birth weight; BDP = bronchopulmonary dysplasia; PCA = principal component analysis.
M. Mussap et al. / Clinica Chimica Acta 426 (2013) 127–138
monozygotic and 274 dizygotic twin pairs included in the Iowa
Neonatal Metabolic Screening Program . C4-DC of the short
chain acylcarnitines had the highest heritability; other short chain
acylcarnitines with high heritability included C2, C3, C4, C4-OH, and
C5. Free carnitine C0 also had a significant heritability. The only amino
acid significant after correction for multiple testing was glutamate. Of
considerable clinical value is the role of C4-DC, which in addition to
being elevated in diabetics, was also associated with poor glycemic
control, implicating this metabolite as a strong biomarker for gluco-
and lipotoxicity in type 2 diabetes. Furthermore, the short-chain
acylcarnitines, C4-DC and C4-OH, were reported to be significantly as-
sociated with adverse outcomes after coronary artery bypass grafting
dictperformanceafter coronary arterybypassgrafting.The levels of po-
tentially important biomarkers of adult diseases are heritable at birth.
Therefore, identifying genetic variants associated with these metabo-
lites at birth may provide important screening tools to identify diseases
developed later in life.
4.4. Perinatal asphyxia and hypoxic ischemic encephalopathy (HIE)
Perinatal asphyxia is a major cause of neonatal death especially in
developing countries and is defined as the inability of the newborn to
initiateand sustain adequaterespiration afterdelivery. The internation-
larger rates in developing countries. Perinatal asphyxia is the most im-
portant cause of brain injury in full-term infants, leading to correlated
neurological sequelae. The most severe neurological manifestation is
tal retardation and epilepsy are frequent in children and adolescents
who suffered from perinatal asphyxia . Currently, a limited range
of biochemical tests for hypoxia are easily available for routine clinical
purpose. Such biomarkers seem to be of particular importance in cases
of asphyxiated and resuscitated newborns who do not suffer from con-
ventionally diagnosed HIE, as these newborns have an increased risk of
low intelligence quotient (IQ) score at eight years . The classical use
of single biomarkers to predict disease does not allow the adaptability
required in such a multi-faceted disease entity, rather the use of a
multi-factorial model is required to appropriately define the injury.
Adenosine diphosphate (ADP) and adenosine monophosphate (AMP)
accumulate during asphyxia, generating accumulation of adenosine,
inosine and hypoxanthine. These substrates are channeled to purine
catabolism leading to generation of uric acid, with a consequent in-
crease in urine excretion of uric acid. In the re-oxygenation period,
free radicals are produced concomitantly to uric acid formation, which
is claimed as an indicator of the severity of perinatal asphyxia .
Reperfusion in perinatal asphyxia is associated with oxidative stress,
which leads to carbonylation, fragmentation, cross-linking, and loss of
thiol groups, and nitration of proteins. Banupriya et al. demonstrated
that the excretion rate of urinary proteins, urinary malondialdehyde
(MDA) and urinary uric acid increases with the severity of perinatal
asphyxia and associated brain damage . These markers have poten-
tial as severity evaluation and prognostic markers. Urinary protein at
cut-off level of 9.04 mg/mg of creatinine, urinary uric acid at cut-off
level of 2.59 mg/mg of creatinine and urinary MDA at cut-off level of
2.195 μg/mg of creatinine might be used to predict impending death
in cases with perinatal asphyxia. In 2006, Chu et al. tested the applica-
tion of bioinformatics methods in studying the urine metabolomic
profiles in newborns with severe birth asphyxia and subsequentneuro-
developmental handicap . The clinical outcomes of 256 new-
borns were examined. Urinary metabolite profiles were measured
using MS and then analyzed by bioinformatic methods. As a result, the
metabolomic discriminators between good neonatal outcome and
poor neonatal outcome were identified: urine levels of eight organic
acids were increased and significantly associated with the prognosis.
In particular, ethylmalonate, 3-hydroxy-3-methylglutarate, 2-hydroxy-
glutarate and 2-oxo-glutarate were found to be associated with good
neonatal outcome; conversely, glutarate, methylmalonate, 3-hydroxy-
butyrate and orotate wereassociatedwithpoor outcome. The met-
abolic profile of umbilical cord blood in a group of term infants with
hypoxic ischemic encephalopathy (HIE) has been evaluated and com-
pared with those of both a group of babies with asphyxia and a group
of matched controls . Targeted metabolomic analysis revealed a
significant increase in 29 out of the 148 measured metabolites in the
umbilical cord blood of infants with either asphyxia or HIE compared
to matched healthy controls. Three distinct metabolite classes were
disrupted: amino acids, acylcarnitines, and phosphatidylcholines. Of
these, a group of 8 aminoacids were significantly increased in neonates
with HIE relative to matched controls, but not in the asphyxia group,
while a group of 13 acylcarnitines were significantly increased in both
study groups. However, for the acylcarnitines the increase was more
pronounced in the HIE population. These results support the impor-
sight into the pathogenesis of perinatal asphyxia and to distinguish
Metabolomics in prematurity.
AimBiolog. sample Method Most relevant findings Year Ref.
To assess the change in the urinary metabolome
with post-natal age in preterm neonates
1H NMR The urinary metabolome of preterm infants altered over time.
In preterm babies, alanine, formate, and citrate were found increased
when compared with full-term newborns, while creatinine, creatine,
and dimethylglycine were decreased
Distinct urinary metabolic profiles in neonates of different gestational
ages (term, 23–32 ws, 33–36 ws). Discriminating metabolites: hippurate,
tryptophan, phenylalanine, malate, tyrosine, hydroxybutyrate,
N-acetyl-glutamate, and proline
Arterial cord blood of VLBW: significant decreases of HDL, LDL, VLDL,
glucose, pyruvate, acetone, albumin-lysyl, alanine, tyrosine, valine,
isoleucine, leucine, threonine, and 3-methyl-histidine associated with an
increase in glutamine levels (no difference between arterial and venous
umbilical cord plasma).
Maternal plasma: mothers delivering VLBW infants have significantly
higher levels of lipids, pyruvate, glutamine, valine and threonine, and
significantly lower levels of acetate and isoleucine
In VLBW: a significant elevation in the levels and maternal–fetal gradients
of butyryl-, isovaleryl-, hexanoyl- and octanoyl-carnitines
Significant decrease in glutamine-glutamate in preterm arterial cord blood
Increase in both the circulating levels and maternal–fetal gradients of
several polyamines in their acetylated form
To identify gestational age-related metabolic
To assess the metabolic adaptations associated
with premature birth
Arterial and venous
cord blood, maternal
To assess the global effect of preterm birth on
fetal metabolism and maternal–fetal nutrient
Arterial and venous
cord blood, maternal
Abbreviations:1H NMR = proton nuclear magnetic resonance spectroscopy; LC–HRMS = liquid chromatography–high resolution mass spectrometry; VLBW = very low birth weight
time-of-flight mass spectrometry.
M. Mussap et al. / Clinica Chimica Acta 426 (2013) 127–138
injury severity, and therefore the possibility of directing the need for
treatment . From the practical standpoint, combining experimental
model of perinatal asphyxia , and in superfused neonatal rat brain
slices  with clinical data , metabolomics may be determinant
urinary ratio N3.5: sensitivity and specificity about 90%), predicting
outcome (high serum level of glutarate, MDA, 3-hydroxy-butyrate,
and orotate correlate with positive outcome), and validating therapeu-
tic treatments like hypothermia or drugs. An in-depth understandingof
metabolic flows can provide useful elements for specific personalized
treatment to avoid or reduce the side effects of oxygen.
4.5. Respiratory Distress Syndrome (RDS) and surfactant therapy
Respiratory Distress Syndrome (RDS) is still a major cause of neona-
tal mortality and morbidity in VLBW infants. To date, little is known
about the metabolic status of these infants with RDS, since a limited
number of metabolites have been investigated in biological fluids by
conventional techniques. The bronchoalveolar lavage fluid (BALF) is
the preferred biofluid for assessing neonatal lung diseases. Currently,
metabolomic technology is able to analyze the BALF to obtain a metab-
olite profile providing valuable information on the lung status. In a
group of preterm newborns, it was found that the metabolite profile
of BALF is different between the pre-surfactant and post-surfactant/
mechanical ventilation time-points. GC–MS-based metabolomic analy-
sis revealed a total of 25 metabolites ; ten out 25 metabolites had
a known molecular structure (undecane, decanoic acid, dodecanoic
acid, hexadecanoic acid, octadecanoic acid, hexadecanoic acid methyl
ester, 9-octadecanoic acid, tetracosanoic acid, myristic acid, phosphate)
and were overexpressed in BALF collected during mechanical ventila-
tion following surfactant administration . These preliminary results
encourage the hope that the metabolomic approach can substantially
contribute to identify novel biomarkers of RDS in preterm infants.
4.6. Patent ductus arteriosus
Patent ductus arteriosus (PDA) is one of the most common congen-
ital abnormalities found in preterm infants. Neonates with PDA were
found to have a risk of death eight times higher than that of neonates
with closed ductus arteriosus. A persistent ductus arteriosus is diag-
nosed when it fails to close after 72 h. A preliminary study was carried
out to assess the relevance of1H NMR-based metabolomic urine analy-
sis in anticipation of persistent PDA in term and preterm infants .
PLS-DA was able to discriminate the term infant group, preterm infant
group with PDA and preterm infant group without PDA on the basis of
different urine metabolic patterns at birth . These findings suggest
that metabolomics is a promising tool for predicting and monitoring
persistent PDA, thus avoiding unnecessary drug prophylaxis .
5. Clinical metabolomics in children
5.1. Pediatric age and metabolomics
Metabolomics seems to have promising applications in Pediatrics,
although it has been used only in a limited number of clinical applica-
tion so far . The metabolomic approach was applied to analyze
age-related metabolic changes in children aged 12 years and below by
using1H NMR-based spectroscopy analysis of urine . Unsupervised
PCA analysis showed a distinct age-dependent clustering, indicating
the effect of age on the urinary metabolite profile. Further statistical
analysis led to the identification of age-related metabolic profiles.
increased with age, while creatine, glycine, betaine/trimethylamine-N-
oxide (TMAO), citrate, succinate, and acetone decreased. This investi-
gation elucidated that metabolomic approach has the potential to be
useful in assessingthe biological age of younghumansaswell as in pro-
viding more information about the confounding factors in the clinical
application of metabolomics.
5.2. Respiratory tract diseases
Asthma is one of the most common chronic childhood diseases. An
improved definition of asthma phenotypes would be invaluable for un-
derstanding the pathogenic mechanisms of this disease and the correct
treatment. In Table 5 we have reported the most relevant results pub-
lished in the literature [73–78]. Finally, preliminary unpublished data
with Chronic Lung Disease (CLD) from controls at 3 years of age and
that it is possible to anticipate the diagnosis of CLD at birth. A number
of discriminating metabolites are identical at birth and at 3 years.
5.3. Neurological diseases
A characteristic metabolomic profile in the cerebral spinal fluid
(CSF) of children with influenza-associated encephalopathy has been
reported, suggesting that it might be possible to identify biomarkers
suitable for the early diagnosis of this disease . Autism is an early
onset developmental disorder which has a severe life-long impact on
behavior and social functioning, with associated metabolic and gastro-
intestinal abnormalities of unclear etiology. The urinary metabolic phe-
notypes in 3 groups of children aged 3–9 years (39 affected by autism,
28 by non-autistic siblings, and 34 age-matched healthy volunteers)
were characterized using1HNMRspectroscopy andpatternrecognition
methods . Autistic children showed increased urinary excretion of
N-methyl-2-pyridone-5-carboxamide, N-methyl nicotinic acid, and N-
methyl nicotinamide, suggesting a perturbation in the tryptophan-
nicotinic acid metabolic pathway. Urinary patterns of glutamate and
taurine were found to be significantly different between groups; higher
levels of urinary taurine and lower levels of urinary glutamate were
found in autistic children, indicating perturbation in sulfur and amino
Metabolomics in pediatric respiratory tract diseases.
DiseaseBiolog. sampleMethod Most relevant findingsYear Ref.
Presence of acetylate compounds, showing new metabolic pathways in asthma
Stable asthma shows a unique pattern of metabolites compared with absence of asthma or
Reduction in the urocanic acid and methyl‐imidazoleacetic acid contents may be correlated with
the modulation of immunity in asthma
Asthma shows increased ADMA levels, supporting a role for this mediator in asthma pathogenesis
Uric acid, hypoxanthine, glutamic acid increased; L-tryptophan and ADP decreased
L-histidine and uric acid decreased
Alterations in creatinine, betaine and glycine
1H NMR 2011 
chromatography–mass spectrometry; TOF-MS = time-of-flight mass spectrometry; ADMA = asymmetric dimethylarginine; ADP = adenosine diphosphate.
1H NMR = proton nuclear magnetic resonance spectroscopy; LC–MS = liquid chromatography-mass spectrometry; UPLC–MS = ultra‐high‐performance liquid
M. Mussap et al. / Clinica Chimica Acta 426 (2013) 127–138
acid metabolism. Additionally, differences in urinary mammalian-
microbial co-metabolites including dimethylamine, hippurate, and
phenylacetylglutamine were found between autistic and control chil-
dren. The biochemical changes found in autistic children are consistent
withsomeabnormalitiesof gutmicrobiota, andhavethepotentialtobe
used to monitor therapeutic interventions .1H NMR spectroscopy-
based metabolomic profiling was applied to characterize the cerebral
metabolic status of 59 children undergoing NMR imaging during anes-
thesia with either sevoflurane or propofol . NMR scans were ac-
quired in the parietal cortex after approximately 60 min of anesthesia.
The results demonstrated higher glucose and lactate with sevoflurane
in the human brain compared with propofol: this could reflect greater
neuronal activity with sevoflurane resulting in enhanced glutamate-
neurotransmitter cycling, increased glycolysis, and lactate shuttling
from astrocytes to neurons or mitochondrial dysfunction. Further, the
association between emergence delirium and lactate suggests that
anesthesia-induced enhanced cortical activity in the unconscious state
may interfere with rapid return to “coherent” brain connectivity pat-
terns required for normal cognition upon emergence of anesthesia
. Epilepsy is major neurological sequelae in extremely low birth
weight (ELBW) newborns with a range of incidence between 4.1% and
18.7%. The occurrence of seizures in epileptic children reflects the vul-
nerability of their immature brain due to neuropathological processes
thattake place in that periodof life. A preliminary studywasperformed
to explore the metabolic differences between a group of children born
ELBW who developed epilepsy and a group of children born ELBW
without epilepsy, selected as control group . Urine samples were
collected from 7 children with epilepsy and 9 controls. The mathemat-
ical model wasable todiscriminate between thegroup of epileptic chil-
cine, malonic acid, creatinine, and sugars, which were decreased in ep-
ileptics compared to controls.
5.4. Kidney diseases
Kidney disease is a major cause of illness and death among infants,
children, and adolescent. Little is known about the epidemiology of
chronic kidney disease (CKD) in the pediatric population, and specific
problems occur in children, such as impaired growth and psychosocial
adjustment, all of which severely impact upon the quality of life .
The leading causes of CKD are cystic/hereditary and congenital disor-
ders, including congenital anomalies of the kidney and urinary tract
(48% in USA, about 59% in Europe) and hereditary nephropathies (10%
in USA, 15–19% in Europe), while glomerular diseases account for 14%
in USA and 5–7% in Europe [84,85]. Among children developing CKD,
more than 35% progress to end stage renal disease (ESRD), requiring
renal replacement therapy . Various severe clinical conditions af-
fecting children and newborns admitted to the intensive care unit
(ICU), like ischemia–reperfusion, sepsis, surgery, and fluid imbalance,
easily lead to a rapid deterioration of kidney function characterized by
oliguria/anuria. This clinical condition has been defined acute kidney
disease and disorder (AKD) and it can lead, in turn, to the development
of acute kidney injury (AKI), mostly resulting from concomitant sys-
temic diseases or their related therapeutic treatments (e.g., sepsis and
nephrotoxic medications) . In critically ill patients, fluid overload
and AKI are associated with adverse outcomes, especially in the pediat-
ric setting, with significant morbidity and mortality. Despite a number
of experimental animal model-based studies as well as clinical studies
have investigated the metabolite profile in CKD , very few clinical
studies based on the application of metabolomics approach in pediatric
nephrology have been published . In 2008, Beger et al., performed
LC–MS–based metabolomic analysis of urine samples from 40 children
that underwent cardiac surgery for congenital heart defects . Data
analysis from urine samples collected 4 h and 12 h after surgery
showed that samples of children developing AKI clustered separately
from those who did not. Further analysis by tandem MS leads to the
identificationofhomovanillic acidsulfate(HVA-SO4), ametaboliteorig-
inating from dopamine. These findings suggest that urinary HVA-SO4
may be a promising, sensitive, and predictive biomarker of AKI after
pediatric cardiac surgery. The metabolite patterns associated with
nephrourological diseases were investigated by a1H NMR-based meta-
bolic profiling of urine samples collected both in 21 children with
nephrouropathies (renal dysplasia, vesico-ureteral reflux, urinary tract
infection, etc.) and in 19 healthy controls . Linear discriminant
analysis-based classification of the spectral data demonstrated high ac-
curacy (95%) in the separation of the two groups of samples. In that
model, the urine metabolite profiles correlated with nephrourological
disorders, confirming the importance of metabolomics in improving
the early diagnosis and the monitoring of pediatric nephrouropathies.
Over the last decade, the rapidly growing research field of meta-
bolomics has introduced new insights into the pathology of diabetes
as well as methods to predict disease onset by discovering new bio-
markers. Inadults, theyquantified 140 metabolitesfor4297fasting
serum samples belonging to normal glucose tolerance and pre-diabetic
individuals (impaired fasting glucose and/or impaired glucose toler-
ance, isolated impaired fasting glucose, newly diagnosed type 2 diabe-
tes) . Three novel metabolites, glycine, lysophosphatidylcholine
(18:2), and acetylcarnitine were identified as pre-diabetes-specific
markers. Their changes might precede other branched-chain and
aromatic amino acids markers in the progression of type 2 diabetes.
Combined levels of glycine, lysophosphatidylcholine(18:2), and acetyl-
carnitine can predict risk not only for impaired glucose tolerance but
also for type 2 diabetes. Serum metabolic alterations precede the auto-
immunity that is characteristically observed before the development
pared between 56 children who progressed to type 1 diabetes and 73
controls who remained non-diabetic and permanently autoantibody
negative, by using a GC–MS technique . Results strongly suggested
that metabolic dysregulationprecedes overt autoimmunity intype 1di-
abetes. Children who developed type 1 diabetes had reduced serum
levels of succinic acid and phosphatidylcholine at birth and reduced
up. The appearance of insulin and glutamic acid decarboxylase auto-
antibodies was preceded by decrease in ketoleucine and increase in
glutamic acid. The metabolic profile was partially normalized after the
seroconversion . In a subsequent study on a murine model, it was
foundthat femalenon-obeseprediabetic micewholaterprogresstoau-
toimmune diabetes exhibit the same lipidomic pattern as prediabetic
children . These metabolic changes are accompanied by enhanced
glucose-stimulated insulin secretion, normoglycemia, upregulation
of insulinotropic amino acids in islets, elevated plasma leptin and
adiponectin, and diminished gut microbial diversity of the Clostridium
leptum group. Thus, autoimmune diabetes is preceded by a state of
increased metabolic demands on the islets resulting in elevated insulin
secretion, suggesting alternative metabolic related pathways as thera-
peutic targets to prevent diabetes.
5.6. Celiac disease
factors (HLA class II genes DQ2 and/or DQ8) and the environmental
opment. Sellitto et al. characterized the longitudinal changes in the mi-
crobial communities that colonize infants from birth to 24 months, the
impact of two patterns of gluten introduction (early vs. late) on the gut
microbiota and metabolome, and the switch from gluten tolerance to
immune response, including onset of celiac disease autoimmunity
M. Mussap et al. / Clinica Chimica Acta 426 (2013) 127–138
. Data suggest significant differences between the developing
microbiota of infants with a genetic predisposition for celiac disease
and those from infants with a non-selected genetic background. The
retrospective analyses of the gastro intestinal microbiota and meta-
bolomic data also suggest that potential specific biomarkers might be
identified; these biomarkers would be predictive for autoimmune de-
velopment in subjects genetically at risk, possibly leading to the devel-
opment of potential interventions during the pre-clinical phase of the
the onset of celiac disease autoimmunity.
5.7. Cystic fibrosis
In cystic fibrosis, airway inflammation leads to an increased produc-
tion of reactive oxygen species, resulting in the degradation of cell
membranes and the generation of volatile organic compounds (VOCs).
By the1H NMR-based metabolomic approach, the metabolic profiles
of BALF from 11 pediatric patients with cystic fibrosis and various de-
grees of inflammation were detected . The most obvious finding
was the stark contrast between the number of metabolites present in
the spectra of the subjects with high inflammation compared with the
subjects with low inflammation. Indeed, a large number of metabolites
were present in detectable concentrations exclusively within the high-
inflammation group. However, despite the presence of a metabolic
signature related to inflammation, it was impossible to distinguish
whether or not the identified metabolites were from host or from
breath was able to discriminate between 48 patients affected by cystic
Pseudomonas aeruginosa colonization . Analysis revealed that 1099
VOCs exhibited a prevalence of at least 7%. A 100% correct identification
of cystic fibrosis and healthy controls was obtained by using 22 VOCs.
Therefore, metabolomic analysis of VOCs in exhaled breath appears to
be a reproducible technique, and is able to discriminate not only be-
tween cystic fibrosis patients and controls but also patients with or
without Pseudomonas colonization.
Only few studies have been devoted to childhood obesity. Serum
samples of 80 obese and 40 normal-weight children between 6 and
15 years of age were analyzed using a MS-based metabolomics ap-
proach targeting 163 metabolites . Metabolite concentrations and
metabolite ratios were compared between obese and normal-weight
olite profiles of obese children were clearly distinguishable from those
of normal-weight children. In particular, 14 metabolites (glutamine,
methionine, proline, nine phospholipids, and two acylcarnitines) and
69 metabolite ratios were significantly altered in obese children. The
identified metabolite markers are indicative of oxidative stress and of
changes in sphingomyelin metabolism, in β-oxidation,and in pathways
associated with energy expenditure. The altered metabolites might be
on the biological mechanisms behind obesity. Mihalik et al. compared
and plasma amino acids in normal weight (n = 39), obese (n = 64),
and type 2 diabetic (n = 17) adolescents . Plasma very long- and
long-chain acylcarnitine (C18:2-CN to C14:0-CN), derived primarily
from the entry of fatty acids into β-oxidation, were similar in all three
groups. Plasma short- and medium-chain acylcarnitine species were
lower in youth with type 2 diabetes compared with normal weight,
with a trend toward lower values in the obese subjects. In addition,
both C3-CN and C5-CN, intermediates derived from branched-chain
amino acids oxidation, were significantly lower in the diabetic group
compared with normal weight. Similarly, C4-CN, a short chain species
that derives from both the latter stages of fatty acid oxidation and
from the branched-chain amino acids valine, was lower in the group
with type 2 diabetes relative to normal weight. Lastly, free carnitine
(free CN) was lower in adolescents with type 2 diabetes compared
with normal weight, with a similar trend in obese. Plasma amino acids
were lowest in adolescents with type 2 diabetes compared with both
obese and normal weight adolescents. Fasting fat oxidation was nega-
ever, these relationships disappeared after adjusting for body mass
index (BMI), Tanner stage, and sex. Insulin sensitivity was positively
associated with free carnitine, C6-CN and C10:1-CN. The relationships
between insulin sensitivity and acylcarnitine species disappeared after
adjustment for adiposity indices, Tanner stage, and sex. All amino
acids, except for phenylalanine, methionine, tyrosine, and alanine,
were positively associated with insulin sensitivity. Such observations
are consistent with early adaptive metabolic plasticity in youth over
time, with continued obesity and aging, becoming dysfunctional, as
observed in adults.
5.9. Inborn errors of metabolism
Metabolic disorders are rare events affecting about 1 in every 5000
babies born. Many of the metabolic disorders have been characterized
with well-defined clinical pathology, and specific “metabonomic” bio-
markers have been identified . Indeed, several pediatric hospitals
screen newborn babies using MS techniques with the known appropri-
ate metabonomic biomarkers . Inherited metabolic disorders in-
volve defects in the metabolism of enzymes catalyzing carbohydrate,
amino acid, fatty acid, nucleic acid, and urea cycle. Additional inherited
disorders are linked to defects in cytochrome P450 enzymes catalyzing
cholesterol, bile acid, steroid, and vitamin D synthesis and metabolism.
Untargeted MS-based metabolomics was applied to the most common
disorders of propionate metabolism, specifically methylmalonic acide-
mia (MMA) and propionic acidemia (PA). Metabolomic analysis was
able to reveal metabolites that differentiated between disease and nor-
mal patients, and between MMA and PA. Plasma metabolites including
unsaturated acylcarnitines, isovaleryl carnitine, and ©-butyrobetaine
were associated with MMA and PA. Propionyl carnitine was identified
as the best biomarker of disease . A subsequent study investigated
39 children with defined respiratory chain deficiencies (RCDs) using
untargeted MS-based metabolomic analysis of the urinary organic
acids. The analysis revealed the presence of 255 endogenous and 46 ex-
ogenous substances. Statistical analysis showed 24 metabolites that
were highly and statistically significant for the combined and clinically
related group of respiratory chain deficiencies. This study has revealed
ture for RCDs, thus enabling the development of a non-invasive screen-
ing method for RCD disorders . Last but not least, the available
metabonomics biomarkers from the IMD database should help toxicol-
ogists to characterize the mechanisms of drug-induced toxicities .
6. Nutrimetabolomics in childhood
In the pediatric age, inappropriate choices of diets lead to metabolic
imbalance, which in turn increase the riskof devastatingdiseases of the
adult (obesity, intolerances, food allergies, diabetes, hypertension,
atherosclerosis, etc.). Food-induced metabolic reaction varies greatly
among children due to key factors such as age, stress, environment,
and especially gut microbiota. Thus, dietary recommendations should
be customized on the basis of the individual nutritional phenotype in
order to provide a comprehensive characterization of nutritional status
 and to minimize the risk of future correlated diseases . No
single biomarker can be considered adequate in distinguishing the nu-
tritional phenotype. By contrast, metabolomics assures the characteri-
zation of metabolic fingerprints that can be associated with individual
phenotypes, whichencompass dietary or diseasestatus. Theemergence
M. Mussap et al. / Clinica Chimica Acta 426 (2013) 127–138
of “nutrimetabolomics” illustrates the mutual link among nutrition and
metabolomics research to study the effects of specific ingredients and
food components elucidating the molecular mechanisms behind indi-
vidual metabolic responses . The effects of diet on metabo-
lism were investigated by a metabolomic approach based on1H NMR
spectroscopy. Children receiving a diet rich in milk (n = 12) were
compared with children receiving a diet consisting of meat proteins
(n = 12). At baseline and after 7 days, urine and serum samples were
collected and the corresponding1H NMR spectra were acquired. The
milk diet was found to reduce the urinary excretion of hippurate,
while the diet rich in meat proteins was found to increase the urinary
excretion of creatine, histidine and urea. The diet rich in meat proteins
exhibited no effect on the serum metabolic profile.
metabolomic technologyidentified thebiochemicaleffects ofconsump-
the metabolic profile of human breast milk compared with that of com-
a dominance of signals in the carbohydrate region. In addition to these
signals, well-defined signals in the aliphatic region of the spectra were
present, while a few peaks emerged in the aromatic portion. Human
breast milk contains relatively higher contents of lactose compared
with formula milk, that is richer in maltose; oleic and linoleic acids
appeared to be higher in the artificial formulas than in human breast
milk. A correlation was also observed between gestational age and
human breast milk . Metabolomics was successful in identifying
themetabolites specific totheintestinalcontents ofbreastmilkandfor-
mula milk in both animal model(youngpiglets) and premature infants.
By supplying the juvenile gut with an array of specific substrates avail-
able for bacterial consumption, breast milk and infant formula clearly
serve as distinct bacterial growth environments and select on different
bacterial communities. Sugars, amino-sugars, fatty acids (namely un-
saturated fatty acids), and sterols were among the most important dis-
criminating metabolites between breast fed and formula fed groups.
Furthermore, the metabolomics approach focused on intestinal ecosys-
dence of microbe-mediated complications in premature infants fed
with formula milk . The addition of novel ingredients to formula
for infant health; therefore, it is essential that the evaluation of such in-
rigorous scientific approach to assess their safety and their potential
toxicity. In this view, metabolomics may become an extraordinary tool
veloping organ system .
7. Pharmametabolomics in childhood
Metabolomics has the potential to predict drug efficacy and drug-
induced side effects. This approach, called pharmacometabolomics,
has the capability to explore the “tailored drug therapy field”, through
fine responders and non-responders. Recently, pharmacometabolomics
ing the metabolic modification due to drug treatments, metabolomics
gives a number of information on the pharmacokinetic and pharmaco-
dynamic response of an individual to the pharmacological therapy
. This is of great importance in pediatrics and neonatology, since
pharmacokinetic of almost all the drugs is completely different from
adults, especially in the neonatal age and over the first four years of
life. The simple linear reduction of the adult dose cannot be considered
a reliable approach toobtain a safe andeffective pediatric dose, because
of the contribution of several, independent factors such as liver enzyme
maturation, gastrointestinal function, abrupt changes of the body com-
position, and volume distribution. Most drugs are not specifically ap-
proved for pediatric use and this makes an approximation of the
can improve the knowledge by accurately defining the drug response
and the patient's metabolic phenotype (sometimes referred as a
metabotype). The application of metabolomics for understanding the
effects of drug interaction in a high variable context such as in new-
borns, will allow to discover typical features, representing changes
(currently unclear) that occur after drug treatment. A comprehensive
review on this topic has been recently published .
The mounting speed of knowledge in basic biological research to-
gether with the rapid development of high-technology diagnostic sys-
tems (e.g., nanotechnologies) significantly contributes to gains in
neonatal life expectancy. Metabolomics represents one of the most
stimulating challenges for laboratory medicine and specifically in pedi-
atrics and neonatology. Identifying the heritable components of the
metabolome, particularly in the neonatal period, may pave the way
for future genetic association studies that may provide insight into the
physiology of different cellular processes, as well as the biology of com-
plex neonatal and adulthood diseases . Cord plasma appears to be
particularly suitable for metabolomic analysis and provides new infor-
mation on the materno-fetal nutrient exchange in preterm infants.
However, new efforts should be done in reducing the growing gap be-
tween research and practice. Progresses in translational research could
lead to the improvement of outcome in critically ill newborns and chil-
dren by identifying early changes in metabolic profile of various biolog-
ical fluids. The identification of anindividual metabolic profile featuring
a specific pathological condition encourages the application of a more
individualized medicine by predicting the course of the disease and
the response to the therapeutic treatment. In addition, metabolomics
permits to identify new biomarkers potentially useful in various clinical
conditions, facilitatingtheir translation into routinetools for diagnosing
and monitoring specific diseases. Finally, the negligible false-positive
and false-negative rate of metabolomics results may confer to this ap-
proach the role of “gold standard” in selecting and assessing the patient
group category for the clinical evaluation of new biomarkers. However,
there is the urgent need to standardize physiological (age, sex, lifestyle,
diet, etc.) and preanalytical (sample collection and storage, etc.) vari-
ables, analytical methods (by developing standard protocols), data pro-
of the method by which results have been obtained (sensitivity, inter-
ferences, etc.). The Metabolomics Standard Initiative (MSI) was created
with the intentto generate recommendationson standard protocols for
use in all the phases of the metabolomic research . Finally, an ade-
quate monitoringof the analytical quality is mandatory before to estab-
lish the definitive clinical value of a cluster of metabolites featuring a
well-defined physiopathological condition.
 Lander ES, Linton LM, Birren B, et al. Initial sequencing and analysis of the human
genome. Nature 2001;409(6822):860–921.
 Venter JC, Adams MD, Myers EW, et al. The sequence of the human genoma.
 Mussap M, Fanos V. Reducing neonatal mortality and expenditure in the era of
 Williams RJ. Introduction, general discussion and tentative conclusions. University
Texas publication no. 5109. Biochemical Institute studies IV. Individual metabolic
patterns and human disease: an exploratory study utilizing predominantly paper
chromatographic methods. Austin: The University of Texas; 1951. p. 7–21.
 Horning EC, Horning MG. Human metabolic profiles obtained by GC and GC/MS.
J Chromatogr Sci 1971;9:129–40.
 Pauling L, Robinson AB, Teranishi R, Cary P. Quantitative analysis of urine vapor
and breath by gas-liquid partition chromatography. Proc Natl Acad Sci U S A
 Gates SC, Sweeley CC. Quantitative metabolic profiling based on gas chromatogra-
phy. Clin Chem 1978;24:1663–73.
 Nicholson JK, Lindon JC, Holmes E. ‘Metabonomics’: understanding the metabolic
responsesof livingsystems to pathophysiological stimuli via multivariate statistical
analysis of biological NMR spectroscopic data. Xenobiotica 1999;29:1181–9.
M. Mussap et al. / Clinica Chimica Acta 426 (2013) 127–138
 Fiehn O. Metabolomics — the link between genotypes and phenotypes. Plant Mol
 Oliver SG, Winson MK, Kell DB, Baganz F. Systematic functional analysis of the
yeast genome. Trends Biotechnol 1998;16:373–8.
 Wishart DS, Tzur D, Knox C, et al. HMDB: the Human Metabolome Database.
Nucleic Acids Res 2007;35:D521–6 [Database issue].
 Xu EY, Schaefer WH, Xu Q. Metabolomics in pharmaceutical research and develop-
ment: metabolites, mechanisms and pathways. Curr Opin Drug Discov Devel
 Schnackenberg LK. Global metabolic profiling and its role in systems biology to
advance personalized medicine in the 21st century. Expert Rev Mol Diagn
 Oldiges M, Lütz S, Pflug S, Schroer K, Stein N, Wiendahl C. Metabolomics: cur-
rent state and evolving methodologies and tools. Appl Microbiol Biotechnol
 Griffiths WJ, Wang Y. Mass spectrometry: from proteomics to metabolomics and
lipidomics. Chem Soc Rev 2009;38:1882–96.
 Wishart DS, Knox C, Guo AC, et al. HMDB: a knowledgebase for the human
metabolome. Nucleic Acids Res 2009;37:D603–10 [Database issue].
 Roessner U, Beckles DM. Metabolite measurements. In: Schwender J, editor. Plant
metabolic networks. New York: Springer; 2009. p. 39–69.
 Brown M, Wedge DC, Goodacre R, et al. Automated workflows for accurate mass-
based putative metabolite identification in LC/MS-derived metabolomic datasets.
 Fonville JM, Maher AD, Coen M, Holmes E, Lindon JC, Nicholson JK. Evaluation of
full-resolution J-resolved1H NMR projections of biofluids for metabonomics infor-
mation retrieval and biomarker identification. Anal Chem 2010;82:1811–21.
 Pan Z, Raftery D. Comparing and combining NMR spectroscopy and mass spec-
trometry in metabolomics. Anal Bioanal Chem 2007;387:525–7.
 Ellis DI, Goodacre R. Metabolic fingerprinting in disease diagnosis: biomedical ap-
plications of infrared and Raman spectroscopy. Analyst 2006;131:875–85.
 Krafft C, Steiner G, Beleites C, Salzer R. Disease recognition by infrared and Raman
spectroscopy. J Biophotonics 2009;2:13–28.
 Brown M, Dunn WB, Ellis DI, et al. A metabolome pipeline: from concept to data to
knowledge. Metabolomics 2005;1:39–51.
 Rifai N, Gillette MA, Carr SA. Protein biomarker discovery and validation: the long
and uncertain path to clinical utility. Nat Biotechnol 2006;24:971–83.
liquid chromatography–mass spectrometry fingerprinting of human urine: sample
stability underdifferent handlingandstorage conditionsfor metabonomics studies.
J Chromatogr A 2008;1189:314–22.
 Dunn WB, Broadhurst D, Begley P, et al. Human Serum Metabolome [HUSERMET]
Consortium. Procedures for large-scale metabolic profiling of serum and plasma
using gas chromatography and liquid chromatography coupled to mass spectrom-
etry. Nat Protoc 2011;6:1060–83.
 Michopoulos F, Theodoridis G, Smith CJ, Wilson ID. Metabolite profiles from dried
biofluid spots for metabonomic studies using UPLC combined with oaToF-MS.
J Proteome Res 2010;9:3328–34.
 KongST, Lin HS,Ching J,HoPC. Evaluation of dried blood spots as sample matrix for
gas chromatography/mass spectrometry based metabolomic profiling. Anal Chem
 Michopoulos F, Theodoridis G, Smith CJ, Wilson ID. Metabolite profiles from dried
blood spots for metabonomic studies using UPLC combined with orthogonal accel-
eration ToF-MS: effects of different papers and sample storage stability. Bioanalysis
 Wilson I. Global metabolic profiling [metabonomics/metabolomics] using dried
blood spots: advantages and pitfalls. Bioanalysis 2011;3:2255–7.
 Wishart DS. Quantitative metabolomics using NMR. Trends Anal Chem
 Xia J, Psychogios N, Young N, Wishart DS. MetaboAnalyst: a web server for meta-
bolomic data analysis and interpretation. Nucleic Acids Res 2009;37:W652–60
[Web Server issue].
 Boccard J,Veuthey JL, Rudaz S. Knowledge discovery inmetabolomics:an overview
of MS data handling. J Sep Sci 2010;33:290–304.
 van den Berg RA, Hoefsloot HC, Westerhuis JA, Smilde AK, van der WerfMJ. Center-
ing, scaling, and transformations: improving the biological information content of
metabolomics data. BMC Genomics 2006;7:142.
 McIntire DD, Bloom SL, Casey BM, Leveno KJ. Birth weight in relation to morbidity
and mortality among newborn infants. N Engl J Med 1999;340:1234–8.
 Brooks AA, Johnson MR, Steer PJ, Pawson ME, Abdalla HI. Birth weight: nature or
nurture? Early Hum Dev 1995;42:29–35.
 Forsdahl A. Points which enlighten the high mortality rate in the county of
Finnmark. Can the high mortality rate today be a consequence of bad conditions
of life in childhood and adolescence? Tidsskr Nor Laegeforen 1973;93:661–7.
 Barker DJ, Winter PD, Osmond C, Margetts B, Simmonds SJ. Weight in infancy and
death from ischaemic heart disease. Lancet 1989;2:577–80.
 Barker DJ, Osmond C. Low birth weight and hypertension. BMJ 1988;297:134–5.
 Dessì A,Atzori L, Noto A,et al. Metabolomics innewborns with intrauterine growth
retardation (IUGR): urine reveals markers of metabolic syndrome. J Matern Fetal
Neonatal Med 2011;24(Suppl. 2):35–9.
 Nissen PM, Nebel C,Oksbjerg N, Bertram HC. Metabolomics revealsrelationship be-
tween plasma inositols and birth weight: possible markers for fetal programming
of type 2 diabetes. J Biomed Biotechnol 2011;2011:1–8.
 Favretto D, Cosmi E, Ragazzi E, et al. Cord blood metabolomic profiling in intrauter-
ine growth restriction. Anal Bioanal Chem 2012;402:1109–21.
 Horgan RP, Broadhurst DI, Walsh SK, et al. Metabolic profiling uncovers a pheno-
typic signature of small for gestational age in early pregnancy. J Proteome Res
 Ivorra C, García-Vicent C, Chaves FJ, Monleón D, Morales JM, Lurbe E. Metabolomic
profiling in blood from umbilical cords of low birth weight newborns. J Transl Med
Jul 9 2012;10:142.
 Hyde MJ, Beckkonert OP, Yap IKS, Booms C, Gale CRK, Logan K, et al. The effect of pre-
term delivery on the urinary metabolome. Neonatal Society 2010 Summer Meeting,
UK; 2010 http://www.neonatalsociety.ac.uk/meetings/2010_summer_program.shtml.
from preterm and term neonates. Front Biosci (Elite Ed) Jun 1 2011;3:1005–12.
 Tea I, Le Gall G, Küster A, et al.1H-NMR-based metabolic profiling of maternal and
umbilical cord blood indicates altered materno-foetal nutrient exchange in pre-
term infants. PLoS One 2012;7(1):e29947.
 Bell JD, Brown JC, Nicholson JK, Sadler PJ. Assignment of resonances for ‘acute-
phase’ glycoproteins in high resolution proton NMR spectra of human blood plas-
ma. FEBS Lett 1987;215:311–5.
port and utilization of amino acids in IUGR — a review. Placenta 2005;26(Suppl. A):
 Alexandre-Gouabau MC, Courant F, Moyon T, et al. Maternal and cord blood LC–
HRMS metabolomics reveal alterations in energy and polyamine metabolism, and
oxidative stress in very-low birth weightinfants. J Proteome Res 2013;12:2764–78.
 Althabe F, Belizan JF. Caesarean section: the paradox. Lancet 2006;368:1472–3.
 Hashimoto F, Nishiumi S, Miyake O, et al. Metabolomics analysis of umbilical cord
blood clarifies changes in saccharides associated with delivery method. Early Hum
 PaladiniP,PaladiniA,Noto A,et al. Urinary metabolomicsintwins at birth. J Pediatr
Neonat Individual Med 2012;1:113.
 Alul FY, Cook DE, Shchelochkov OA, et al. The heritability of metabolic profiles in
newborn twins. Heredity 2013;110:253–8.
 Shah AA, Craig DM, Sebek JK, et al. Metabolic profiles predict adverse events after
coronary artery bypass grafting. J Thorac Cardiovasc Surg 2012;143:873–8.
 Lawn JE, Cousens S, Zupan J. 4 million neonatal deaths: when? Where? Why? Lan-
 de Hann M, Wyatt JS, Roth S, Vargha-Khadem F, Gadian D, Mishki M. Brain
and cognitive-behavioural development after asphyxia at term birth. Dev Sci
 Odd DE, Lewis G, Whitelaw A, Gunnell D. Resuscitation at birth and cognition at
8 years of age: a cohort study. Lancet 2009;373:1615–22.
 Fellman V, Raivio KO. Reperfusion injury as the mechanism of brain damage after
perinatal asphyxia. Pediatr Res 1997;41:599–606.
 Banupriya C, Ratnakar P, Doureradjou N, Mondal N, Vishnu B, Koner BC. Can uri-
nary excretion rate of malondialdehyde, uric acid and protein predict the severity
and impending death in perinatal asphyxia? Clin Biochem 2008;41:968–73.
 Chu CY, Xiao X, Zhou XG, et al. Metabolomic and bioinformatic analyses in asphyx-
iated neonates. Clin Biochem 2006;39:203–9.
 Walsh BH,Broadhurst DI, MandalR,etal.Themetabolomicprofileof umbilical cord
blood in neonatal hypoxic ischaemic encephalopathy. PLoS One 2012;7:e50520.
 Solberg R, Enot D, Deigner HP, et al. Metabolomic analyses of plasma reveals new
insights into asphyxia and resuscitation in pigs. PLoS One 2010;5:e9606.
 Beckstrom AC, Humston EM, Snyder LR, Synovec RE, Juul SE. Application of com-
prehensive two-dimensional gas chromatography with time-of-flight mass spec-
trometry method to identify potential biomarkers of perinatal asphyxia in a non-
human primate model. J Chromatogr A 2011;1218:1899–906.
 Liu J, Litt L, Segal MR, Kelly MJ, Yoshihara HA, James TL. Outcome-related
metabolomic patterns from1H/31P NMR after mild hypothermia treatments of
oxygen-glucose deprivation in a neonatal brain slice model of asphyxia. J Cereb
Blood Flow Metab 2011;31:547–59.
 Atzori L, Xanthos T, Barberini L, et al. A metabolomic approach in an experimen-
tal model of hypoxia-reoxygenation in newborn piglets: urine predicts outcome.
J Matern Fetal Neonatal Med 2010;23(S3):134–7.
 Fabiano A, Gazzolo D, Zimmermann LJ, et al. Metabolomic analysis of bronchoalve-
olar lavage fluid in preterm infants complicated by respiratory distress syndrome:
preliminary results. J Matern Fetal Neonatal Med 2011;24(Suppl. 2):55–8.
 Atzori L, Barberini L, Lussu M, et al. Metabolomics & patent ductus arteriosus diag-
nosis: is1H-NMR [Nuclear Magnetic Resonance] spectroscopy of urine at birth pre-
dictive as ultrasound?J Matern Fetal Neonatal Med 2011;24(Suppl. 2) [I].
 Fanos V, Pusceddu M, Dessì A, Marcialis MA. Should we definitively abandon pro-
phylaxis for patent ductus arteriosus in preterm newborns? Clinics (Sao Paulo)
 Moco S, Collino S, Rezzi S, Martin FP. Metabolomics perspectives in pediatric re-
search. Pediatr Res 2013;73:570–6.
 Gu H, Pan Z, Xi B, et al.1H NMR metabolomics study of age profiling in children.
NMR Biomed 2009;22:826–33.
 Carraro S, Rezzi S, Reniero F, et al. Metabolomics applied to exhaled breath conden-
sate in childhood asthma. Am J Respir Crit Care Med 2007;175:986–90.
 Saude EJ, Skappak CD, Regush S, et al. Metabolomic profiling of asthma: diagnostic
utility of urine nuclear magnetic resonance spectroscopy. J Allergy Clin Immunol
 Mattarucchi E,Baraldi E, Guillou C. Metabolomicsapplied to urine samples inchild-
hood asthma; differentiation between asthma phenotypesand identification of rel-
evant metabolites. Biomed Chromatogr 2012;26:89–94.
 Carraro S, Giordano G, Piacentini G, et al. Asymmetric dimethylarginine [ADMA] in
exhaled breath condensate and serum of asthmatic children. Chest Feb 14 2013.
http://dx.doi.org/10.1378/chest.12-2379 [Epub ahead of print].
M. Mussap et al. / Clinica Chimica Acta 426 (2013) 127–138
 Laiakis EC, Morris GA, Fornace AJ, Howie SR. Metabolomic analysis in severe child- Download full-text
hood pneumonia in the Gambia, West Africa: findings from a pilot study. PLoS One
 Atzei A, Atzori L, Moretti C, et al. Metabolomics in paediatric respiratory diseases
and bronchiolitis. J Matern Fetal Neonatal Med 2011;24(Suppl. 2):59–62.
 Kawashima H, Oguchi M, Ioi H, et al. Primary biomarkers in cerebral spinal fluid
obtained from patients with influenza-associated encephalopathy analyzed by
metabolomics. Int J Neurosci 2006;116:927–36.
 YapIK, AngleyM,VeselkovKA,HolmesE,LindonJC, Nicholson JK. Urinary metabol-
ic phenotyping differentiates children with autism from their unaffected siblings
and age-matched controls. J Proteome Res 2010;9:2996–3004.
 Jacob Z, Li H, Makaryus R, et al. Metabolomic profiling of children's brains undergoing
 Serci C, Murgia F, Pruna P, et al. A non-invasive approach to characterize epileptic
children born ELBW compared to non epileptic: a metabolomic point of view.
J Pediatr Neonat Individual Med 2012;1:118.
 Harambat J,van Stralen KJ, Kim JJ, TizardEJ. Epidemiologyofchronic kidneydisease
in children. Pediatr Nephrol 2012;27:363–73.
 North American Pediatric Renal Transplant Cooperative Study (NAPRTCS). 2008
Annual report. Rockville, MD: The EMMES Corporation; 2008.
 Ardissino G, Daccò V, Testa S, et al. ItalKid project epidemiology of chronic renal
failure in children: data from the ItalKid project. Pediatrics 2003;111:e382–7.
 Staples AO, Greenbaum LA, Smith JM, et al. Association between clinical risk factors
and progression of chronic kidney disease in children. Clin J Am Soc Nephrol
 Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney Injury Work
Group. KDIGO clinical practice guideline for acute kidney injury. Kidney Int Suppl
 Zhao YY. Metabolomics in chronic kidney disease. Clin Chim Acta 2013;422:59–69.
 Fanos V, Fanni C, Ottonello G, Noto A, Dessì A, Mussap M. Metabolomics in adult
and pediatric nephrology. Molecules 2013;18:4844–57.
 Beger RD, Holland RD, Sun J, et al. Metabonomics of acute kidney injury in children
after cardiac surgery. Pediatr Nephrol 2008;23:977–84.
 Atzori L, Antonucci R, Barberini L, et al.1H NMR-based metabolic profiling of urine
from children with nephrouropathies. Front Biosci (Elite Ed) 2010;2:725–32.
 Friedrich N. Metabolomics in diabetes research. J Endocrinol 2012;215:29–42.
metabolomics. Mol Syst Biol 2012;8:615. http://dx.doi.org/10.1038/msb.2012.43.
 Orešič M, Simell S, Sysi-Aho M, et al. Dysregulation of lipid and amino acid metab-
olism precedes islet autoimmunity in children who later progress to type 1 diabe-
tes. J Exp Med 2008;205:2975–84.
 Sysi-Aho M, Ermolov A, Gopalacharyulu PV, et al. Metabolic regulation in progres-
sion to autoimmune diabetes. PLoS Comput Biol 2011;7:e1002257.
 Sellitto M, Bai G, Serena G, et al. Proof of concept of microbiome-metabolome anal-
ysis and delayed gluten exposure on celiac disease autoimmunity in genetically at-
risk infants. PLoS One 2012;7:e33387.
 Wolak JE, Esther Jr CR, O'Connell TM. Metabolomic analysis of bronchoalveolar la-
vage fluid from cystic fibrosis patients. Biomarkers 2009;14:55–60.
 Robroeks CM, van Berkel JJ, Dallinga JW, et al. Metabolomics of volatile organic
compounds in cystic fibrosis patients and controls. Pediatr Res 2010;68:75–80.
 Wahl S, Yu Z, Kleber M, et al. Childhood obesity is associated with changes in the
serum metabolite profile. Obes Facts 2012;5:660–70.
 Mihalik SJ, Michaliszyn SF, de las Heras J, et al. Metabolomic profiling of fatty acid
and aminoacidmetabolism inyouth with obesity andtype 2 diabetes: evidencefor
enhanced mitochondrial oxidation. Diabetes Care 2012;35:605–11.
 Baumgartner C, Baumgartner D. Biomarker discovery, disease classification, and
similarity query processing on high-throughput MS/MS data of inborn errors of
metabolism. J Biomol Screen 2006;11:90–9.
 Want EJ, Cravatt BF, Siuzdak G. The expanding role of mass spectrometry in metab-
olite profiling and characterization. Chembiochem 2005;6:1941–51.
 Wikoff WR, Gangoiti JA, Barshop BA, Siuzdak G. Metabolomics identifies perturba-
tions in human disorders of propionate metabolism. Clin Chem 2007;53:2169–76.
 Reinecke CJ, Koekemoer G, van der Westhuizen FH, et al. Metabolomics of uri-
nary organic acids in respiratory chain deficiencies in children. Metabolomics
 Vangala S, Tonelli A. Biomarkers, metabonomics, and drug development: can
inborn errors of metabolism help in understanding drug toxicity? AAPS J 2007;9:
 German JB, Bauman DE, Burrin DG, et al. Metabolomics in the opening decade
of the 21st century: building the roads to individualized health. J Nutr
 Zeisel SH, Freake HC, Bauman DE, et al. The nutritional phenotype in the age of
metabolomics. J Nutr 2005;135:1613–6.
 Bertram HC, Hoppe C, Petersen BO, Duus JØ, Mølgaard C, Michaelsen KF. An NMR-
basedmetabonomic investigation oneffects ofmilkand meatprotein diets given to
8-year-old boys. Br J Nutr 2007;97:758–63.
 Marincola FC, Reali A, Barberini L, et al. NMR metabolic target analysis of the
aqueous human breast extract. J Matern Fetal Neonatal Med 2011;25(Suppl.
 Poroyko V, Morowitz M, Bell T, et al. Diet creates metabolic niches in the “imma-
ture gut” that shape microbial communities. Nutr Hosp 2011;26:1283–95.
 Lonnerdal B. Preclinical assessment of infant formula. Ann Nutr Metab 2012;60:196–9.
 Johnson CH, Patterson AD, Idle JR, Gonzalez FJ. Xenobiotic metabolomics: major
impact on the metabolome. Annu Rev Pharmacol Toxicol 2012;52:37–56.
 Clayton TA, Lindon JC, Cloarec O, et al. Pharmaco-metabonomic phenotyping and
personalized drug treatment. Nature 2006;440:1073–7.
 Fanos V, Barberini L, Antonucci R, Atzori L. Pharma-metabolomics in neonatology:
is it a dream or a fact? Curr Pharm Des 2012;18:2996–3006.
 Carraro S, Giordano G, Reniero F, Perilongo G, Baraldi E. Metabolomics: a new fron-
tier for research in pediatrics. J Pediatr 2009;154:638–44.
 MSI Board Members, Sansone SA, Fan T, et al. The metabolomics standards initia-
tive (MSI). Nat Biotechnol 2007;25:846–8.
M. Mussap et al. / Clinica Chimica Acta 426 (2013) 127–138