Journal of Nutrition and Metabolism

Amino Acids and Inherited Amino Acid-Related Disorders

Lead Guest Editor: Ina KnerrGuest Editors: Laurie Bernstein, Ellen Crushell, Siobhan O’Sullivan, and Jörn Oliver Sass

Amino Acids and Inherited AminoAcid-Related Disorders

Journal of Nutrition and Metabolism

Amino Acids and Inherited AminoAcid-Related Disorders

Lead Guest Editor: Ina KnerrGuest Editors: Laurie Bernstein, Ellen Crushell,Siobhan O’Sullivan, and Jörn Oliver Sass

Copyright © 2018 Hindawi. All rights reserved.

This is a special issue published in “Journal of Nutrition and Metabolism.” All articles are open access articles distributed under theCreative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided theoriginal work is properly cited.

Editorial Board

H. K. Biesalski, GermanyHeiner Boeing, GermanyChristopher L. Gentile, USATim Green, CanadaGiuseppe Grosso, ItalyJay R. Hoffman, USAJosé María Huerta, Spain

Phillip B. Hylemon, USAIris Iglesia, SpainC. S. Johnston, USAB. Koletzko, GermanyStan Kubow, CanadaM. Meydani, USAPedro Moreira, Portugal

Maurizio Muscaritoli, ItalyUte Nöthlings, GermanyA. Venketeshwer Rao, CanadaMohammed S. Razzaque, USALuigi Schiavo, ItalyNorman Temple, CanadaMichael B. Zemel, USA

Contents

Amino Acids and Inherited Amino Acid-Related DisordersIna Knerr , Laurie Bernstein , Ellen Crushell, Siobhan O’Sullivan, and Jörn Oliver SassEditorial (2 pages), Article ID 5629454, Volume 2018 (2018)

Finger Prick to Finger Tip: Use of Mobile Phone Technology to Send PKU Blood ResultsAnne Clark , Deirdre Deverell, Emma Corcoran, Margaret Macauley, Nicola Newcombe, Peter Branagan,Aoife Coughlan, Eimear Daly, Aoibhin Moore Heslin, Ellen Crushell, Joanne Hughes, Ina Knerr ,and Ahmad MonavariResearch Article (5 pages), Article ID 2178346, Volume 2018 (2018)

Age-Related Reference Intervals for Blood Amino Acids inThai Pediatric Population Measured byLiquid Chromatography TandemMass SpectrometryJaraspong Uaariyapanichkul, Sirinuch Chomtho , Kanya Suphapeetiporn , Vorasuk Shotelersuk,Santi Punnahitananda, Pannee Chinjarernpan, and Orapa SuteerojntrakoolResearch Article (10 pages), Article ID 5124035, Volume 2018 (2018)

Comparison of Glycomacropeptide with Phenylalanine Free-Synthetic Amino Acids in Test Meals toPKU Patients: No Significant Differences in Biomarkers, Including Plasma Phe LevelsKirsten K. Ahring , Allan M. Lund, Erik Jensen, Thomas G. Jensen, Karen Brøndum-Nielsen,Michael Pedersen, Allan Bardow, Jens Juul Holst , Jens F. Rehfeld, and Lisbeth B. MøllerClinical Study (11 pages), Article ID 6352919, Volume 2018 (2018)

Metabolomic Insights into the Nutritional Status of Adults and Adolescents with PhenylketonuriaConsuming a Low-Phenylalanine Diet in Combination with Amino Acid and GlycomacropeptideMedical FoodsBridget M. Stroup, Denise M. Ney, Sangita G. Murali, Frances Rohr, Sally T. Gleason, Sandra C. van Calcar,and Harvey L. LevyClinical Study (17 pages), Article ID 6859820, Volume 2017 (2018)

Growth Patterns in the Irish Pyridoxine Nonresponsive Homocystinuria Population and the Influenceof Metabolic Control and Protein IntakeOrla Purcell, Aoife Coughlan, Tim Grant, Jenny McNulty, Anne Clark, Deirdre Deverell, Philip Mayne,Joanne Hughes, Ahmad Monavari, Ina Knerr, and Ellen CrushellResearch Article (7 pages), Article ID 8570469, Volume 2017 (2018)

Multiclinic Observations on the Simplified Diet in PKULaurie Bernstein, Casey Burns, Melissa Sailer-Hammons, Angela Kurtz, and Frances RohrReview Article (5 pages), Article ID 4083293, Volume 2017 (2018)

EditorialAmino Acids and Inherited Amino Acid-Related Disorders

Ina Knerr ,1,2 Laurie Bernstein ,3 Ellen Crushell,1,2 Siobhan O’Sullivan,4

and Jorn Oliver Sass5

1National Centre for Inherited Metabolic Disorders, Temple Street Children’s University Hospital, Dublin, Ireland2University College Dublin, Belfield, Dublin, Ireland3Children’s Hospital Colorado, Aurora, USA4Royal Belfast Hospital for Sick Children, Belfast, UK5Bonn-Rhein Sieg University of Applied Sciences, Rheinbach, Germany

Correspondence should be addressed to Ina Knerr; ina.knerr@cuh.ie

Received 10 May 2018; Accepted 10 May 2018; Published 10 September 2018

Copyright © 2018 Ina Knerr et al. "is is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Amino acids perform multiple essential physiological rolesin humans, and accordingly, their importance to health hasbeen the subject of extensive attention. In this special issue ofthe Journal of Nutrition and Metabolism, we focus on thevarious inborn errors of amino acid metabolism, their di-agnostic challenges, new treatment approaches, and recentadvances in patient monitoring as well as clinical outcomes.

Inborn metabolic disorders affecting amino acid meta-bolism are generally rare but are, by their very nature,complex and challenging conditions. Amino acid-relatedmetabolic disorders comprise a very heterogeneous groupof disease entities with highly variable presentations. Clinicalseverity may range from occasional incidental findings insome cases to overwhelming illness, brain damage, ormultiorgan involvement in others. Some, but not all, of theseconditions are included in regular Newborn BloodspotScreening programmes. Analysis of amino acids, e.g., inplasma or serum, along with distinctive biochemical markerswhich may be identifiable by urinary organic acid analysis,depending on the underlying condition, is crucial in thediagnosis and care of patients with inborn errors of aminoacid metabolism called aminoacidopathies. Amino acid-related disorders are generally caused by an inborn ge-netic defect in the metabolic pathways of a particular aminoacid or a group of amino acids. Typical examples includephenylketonuria (PKU), maple syrup urine disease (MSUD),or classical homocystinuria (HCU, cystathionine-β-synthasedeficiency).

We here present exciting original work in the in-terdisciplinary field of amino acids and related inborn

metabolic disorders. It includes, for instance, studies onPKU which constitutes the most common inborn error ofamino acid metabolism in humans involving phenylalanine,or on HCU, a metabolic disorder in the metabolic pathwayof sulphur-containing amino acids. From a diagnosticviewpoint, age-specific reference intervals for amino acidsshould be used for a particular population and analysismethod, as presented here. We also demonstrate thatmetabolomic analyses in plasma and urine can serve in-formative functions in patients with inborn errors of aminoacid metabolism.

Overall, the treatment goal for affected individuals is tonormalise the striking metabolic imbalance, e.g., at a cellularlevel and in physiological fluids, as much as possible byimplementing, in particular, dietary treatment and medi-cation or cofactor supplementation as appropriate alongwith patient monitoring and emergency treatment as re-quired. In recent times, advances in diagnostic technology,including expanded Newborn Bloodspot Screening pro-grammes, as well as major advances in treatments, have ledto an exciting increase in the body of knowledge regardingamino acid-related disorders which will help to continuouslyimprove our patient outcomes.

In this special issue of the Journal of Nutrition andMetabolism, we aim at providing some new insights into thepathophysiological roles of amino acids, diagnostic ap-proaches, patient management, and new treatments of in-born errors of amino acid metabolism with a view tooptimising nutritional status and overall patient outcome.Early detection of these conditions in Newborn Bloodspot

HindawiJournal of Nutrition and MetabolismVolume 2018, Article ID 5629454, 2 pageshttps://doi.org/10.1155/2018/5629454

Screening programmes and early medical intervention anddietetic treatment are majormedical achievements; however,there are still ongoing challenges in treating these lifelongconditions which require not only insightful clinical man-agement but also more clinical research and fruitful col-laborative outcome studies.

We hope that topics addressed in this special issue of theJournal of Nutrition and Metabolism will lead to a betterunderstanding of these conditions as well as further researchstudies aimed at advancing our knowledge in the fields ofinherited amino acid-related disorders. We propose thatadvanced clinical research will not only offer new insightsinto the ongoing pathophysiology of these rare disorders butalso open up new therapeutic and diagnostic possibilitieswhich over time may lead to improved quality-of-life out-comes in affected individuals.

Conflicts of Interest

"e editors declare that they have no conflicts of interestregarding the publication of this special issue.

Ina KnerrLaurie Bernstein

Ellen CrushellSiobhan O’Sullivan

Jorn Oliver Sass

2 Journal of Nutrition and Metabolism

Research ArticleFinger Prick to Finger Tip: Use of Mobile Phone Technology toSend PKU Blood Results

Anne Clark ,1 Deirdre Deverell,2 Emma Corcoran,1 Margaret Macauley,1

Nicola Newcombe,3 Peter Branagan,3 Aoife Coughlan,4 Eimear Daly,5

Aoibhin Moore Heslin,5 Ellen Crushell,1 Joanne Hughes,1 Ina Knerr ,1

and Ahmad Monavari1

1National Centre for Inherited Metabolic Disorders, Temple Street Children’s University Hospital, Dublin, Ireland2Paediatric Laboratory Medicine, Temple Street Children’s University Hospital, Dublin, Ireland3Department of Information and Communication Technologies (ICT), Temple Street Children’s University Hospital,Dublin, Ireland4Department of Research, Temple Street Children’s University Hospital, Dublin, Ireland5University College Dublin, Belfield, Dublin, Ireland

Correspondence should be addressed to Anne Clark; anne.clark@cuh.ie

Received 12 October 2017; Revised 16 March 2018; Accepted 8 April 2018; Published 24 June 2018

Academic Editor: Jose Marıa Huerta

Copyright © 2018 Anne Clark et al. *is is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

*e Metabolic Dietetic Team in the National Centre for Inherited Metabolic Disorders (NCIMD) in Ireland deals with ap-proximately 120 weekly phenylalanine (Phe) levels for both adults and children. A review of 500 Phe levels highlighted that 52% ofthe results were within the target range. Collaboration between information and communication technologies (ICT) departments,metabolic laboratory, and metabolic dietitians enabled the development of the PKU texting system. Following a successful pilotstudy, the system was then offered to all PKU patients aged over 2 years. *e Phe is analysed and authorised on the laboratorysystem. *e demographics are matched with the patient mobile phone number. Text messages are then validated and sent by thedietitian via a web portal using the Defero SMS texting service. Approximately 290 patients/families currently use the textingsystem. In order to assess the effectiveness of this quality improvement initiative, a patient survey was carried out in 2017. *isshowed 87% rated the system as either very good or excellent. 94% agreed it was time saving. 84% felt there was no influence ondietary compliance. Analysis of financial implications on dietetic time over 21 months revealed savings of €3,275 and 580 hours ofdietetic time. *ere is no evidence, two years after implementation, that the system has had an effect on either the Phe levels interms of recommended range or frequency of sampling.

1. Introduction

Metabolic dietitians review approximately 120 phenylala-nine (Phe) levels weekly and discuss these results with pa-tients via telephone. A review carried out of 500 Phe levels in2012 highlighted that 52% of the results were within thetarget range. *is suggested that there were a large per-centage of patients who did not require further follow-up bya dietitian on a regular basis.

*ere are approximately 500 patients both adults andchildren with phenylketonuria (PKU) attending the

National Centre for Inherited Metabolic Disorders(NCIMD) in Temple Street Children’s University Hospital.Patients with PKU require regular monitoring of their bloodPhe levels. Regular monitoring of Phe levels in PKU patientsis important as it is an indicator of compliance.

A pilot texting system was introduced from January 2015to March 2015. *e pilot period was 3 months. During thisperiod, 28 parents received weekly text messages with theappropriate text messages for the Phe result via the Defero[1] texting service. Parents engaged with the dietitian asappropriate. Following this successful pilot period, patients

HindawiJournal of Nutrition and MetabolismVolume 2018, Article ID 2178346, 5 pageshttps://doi.org/10.1155/2018/2178346

with results within target range received a text withtheir/their child’s Phe level directing them to continue withcurrent management. *ose with levels outside of the targetrange received a text advising them to contact the NCIMDfor further advice and support.

Mobile phones are used by many people on daily basis. Asurvey completed by the Pew Research Center in 2012surveyed 21 countries. 75% of those surveyed sent andaccessed text messages. Mobile phone use was in fact mostcommon in Kenya and Indonesia. *ese were two of thepoorest nations who completed the survey [2]. 2016 was thefirst year that Ireland participated in Deloitte’s GlobalMobile Consumer Survey. 86% of Irish consumers eitherhave access to or own a smartphone. 26% do not use theirmobile phone to make phone calls on a regular basis. Usageis becoming more widespread using other channels such astexting. *e 2016 survey highlighted that 28% of Irishconsumers check text messages first after waking [3]. Jureckiet al. in 2017 studied adherence to clinic recommendationsin PKU patients and demonstrated need for further im-provement [4].

2. Aims and Objectives

*e aim of this study was to examine the PKU results textingsystem in terms of patient acceptability, safety, and any fi-nancial implications. *e objectives were to ensure thatpatients/parents received the correct advice via text ina timely matter without having a negative impact in terms ofboth metabolic control and the number of Phe samplesposted for analysis.

3. Patients and Methods

3.1. Application Process. An application outlining the pro-posed project was submitted to the ICT manager for con-sideration. Following approval, a collaborative team wasformed with representation from Dietetics, ICT, and Met-abolic Laboratory to develop this novel texting system. *eapplication process involved completion of a Project Initi-ation Document which was reviewed by both the In-formation and Communication Technologies Planning andReview and Strategy Committees. *is business case wasthen sighed off by the hospital management board. *is wasnecessary as it would impact on the ICT department. Inaddition hospital funding was required to develop thetechnology.

3.2. Data Protection Commission. In order to ensure patientconfidentiality and safety compliance, the Data ProtectionCommission was consulted. Data are collected in line withthe Data Protection Act 1988. In addition, it is in accordancewith Temple Street Children’s University Hospital DataProtection Policy.

3.3. Pilot Study. Commencing March 2015, a group of 27patients and their guardians participated in a pilot study forthree months. Assent was obtained from patients and

consent from their parents where applicable. As Phe levelresults were generated by the laboratory system, they werechecked firstly against a manual document, which is themethod used to generate paper Phe results for the dietitianand on a webpage to ensure accuracy. *ere were nodocumented errors during this time. *is meant that therewas complete reliability and validity between the textmessages sent to mobiles and the manual results generatedfrom the laboratory system.

Following the pilot project, this system is offered to theeligible patient group (all patients over 2 years of age ex-cluding maternal PKU) with no changes required from thepilot period.

3.4. Patients. *e total number of patients involved in thepilot study was 27. *e gender breakdown was 14 femalesand 13 males. *e age range was 18 months to 4 years. *eexclusion criteria were <2 years and pregnancy.

3.5. Routine Clinical Care in PKU Testing. Routine clinicalcare requires blood testing of all PKU patients. *is is doneby the patient at home. Drops of blood are spotted on toa sample card “dried blood spot” (DBS) by the patient or thepatient’s parent/guardian at home and posted to the De-partment of Paediatric Laboratory Medicine where the DBSPhe level is analysed. Prior to the pilot study, all results,normal or abnormal, were relayed to the patient by thedietitian by telephone.

Since the implementation of the texting system, once thetest has been performed and authorised on the LIMS(Laboratory Information Management System), the result isautomatically sent by the pathology server via a HL7 datainterface. *e interface server then matches the blood spottest results with the patient’s details on a patient manage-ment system.

*e interface server displays the blood Phe result alongwith the patient details and creates a web page viewable bythe dietitian for validation who can then send a text messageof the result to the patient or the patient’s guardians. *emobile then receives a text message informing of their/theirchild’s Phe result or to call the dietitian if the level is notwithin range.

3.6. Patient Survey. A survey was developed and circulatedto the dietetic team. *e signed off survey was then circu-lated to PKU patients/parents on our email database. *esurvey was available for completion for 3 weeks.

3.7. Statistical Analysis. Results were analysed using IBMSPSS Statistics 23.

4. Results

4.1. Financial Implications: 21 Months after Implementation.Table 1 highlights the financial savings which resulted fromthe implementation of the system. Text messages were costedat approximately €0.06 per text while the average phone call

2 Journal of Nutrition and Metabolism

was estimated at €1.00. Prior to the implementation of thetexting system, patients/parents contacted the dietitian for alllevels, that is, both those within and out of range. Table 1highlights savings of €3275which is a result of the reduced costof a text versus a phone call for levels within the recommendedrange. *is then impacted on dietetic time as Table 2.

4.2. Impact onDietetic Time: 21Months after Implementation.Table 2 highlights the impact on dietetic time. *e savingsallowed the dietitian to spend more time with those patientswho had levels out of range. Estimated phone call time perlevel averaged 10 minutes. Text messages averaged €0.06 pertext. Phone calls averaged €0.10 per minute (10 minuteslength of call� €1). *is resulted in a saving of 580.5 hours.

4.3. Retrospective Review of the Mean Phe Level before andafter Introduction (µmol/l). A retrospective review wascarried out 12 months after implementation in order toassess whether or not the texting system has any impact onPhe levels. Median phenylalanine levels for the year beforeand after implementation of the texting system were col-lected for 212 patients. *e number of samples beforeimplementation was 97 levels and after implementation was115 levels Table 3.

Cut-off points for the number of levels per year in eachage group are determined as follows:

(i) 2 years–4 years: 48 levels per year(ii) 4.01 years–10 years: 24 levels per year(iii) 10.01+ years: 12 levels per year

*e reference range used was 120–400µmol/l. Results weredeemed significant at p≤ 0.05. No significant difference wasfound before and after texting in the 2–4 years and 10+ years’ agegroups. *e 4–10-year-old group did show a statistical signifi-cance before and after texting implementation (370µmol/l±99versus 393µmol/l±129, respectively; p � 0.021).

4.4. Retrospective Review of the Frequency (Number of Levels)of Phe Level Sampling before and after Introduction of TextingSystem (TS) (µmol/l). No statistical difference was found inthe frequency of Phe levels before and after introduction of anygroup.*e frequency of Phe refers to the number of levels overthe period of 1 year.*is is taking into account of the age rangeof the patient and frequency of levels which varies fromweeklyto monthly in patients over 2 years of age.

*e percentage of texts within the normal range was 52%.*is is exactly the same percentage as the initial audit carriedout before the text system was implemented (Figure 1).

4.5. Survey Results January 2017. To date (August 2017), 290eligible patients (61.5%) are signed up to the texting system.16% (N � 46) responded to the questionnaire. *e mainfindings of the survey monkey were as follows:

(i) 87% found the system to be either very good orexcellent.

(ii) 71% found the system to have no impact on PKUblood levels.

(iii) 84% felt that the system had no influence on dietarycompliance.

(iv) 94% found the system to be a time saver.

*e complete survey monkey results are detailed inTables 4–7.

Table 1: Cost with and without use of the texting technology ineuros.

NormalPhe levels

AbnormalPhe levels Total

Cost without texting (€) 3,484 3,225 6,709Cost with texting (€) 209 3,225 3,434Saving (€) 3,275

TABLE 2: Saving on dietetic hours.

Numberof levels

Time spentwithouttexting(hours)

Time spentwith texting(hours)

Normal Phe levels 3,484 None 580.5Abnormal Phe levels 3,225 537.5 537.5Total number of levels 6,709 537.5 1118

TABLE 3: Retrospective review of the mean Phe level before and afterintroduction.

Mean before texting(±SD)

Mean after texting(±SD) p value

Whole group 406 (±160) µmol/l 416 (±188) µmol/l 0.1562–4 years 335 (±70) µmol/l 326 (±79) µmol/l 0.3314–10 years 370 (±99) µmol/l 393 (±129) µmol/l 0.02110+ years 483 (±207) µmol/l 491 (±245) µmol/l 0.555

40

35

30

25

20

15

10

50

0Whole group 0–4 yrs 4–10 yrs 10–99 yrs

p = 0.666

p = 0.060

p = 0.093

p = 0.159

Frequency before TSFrequency a�er TS

Figure 1

Journal of Nutrition and Metabolism 3

4.6. PatientComments. “I am all in favour of the texting. It isfast, efficient, and less time consuming. I think it is great.One can get the text no matter where they are.”

“Plus having the texting system reminds me of the levelsas some weeks it has totally slipped my mind to ring. Withthe texting service that does not happen as you get the texteither way.”

“I think texting is important as it saves time and also maybe a reminder to parents to check levels.”

5. Discussion

*emain objective of implementation of this system was asa quality improvement initiative in order to improve di-etetic management in patients with PKU. Prior to theintroduction of the texting system, the onus was on thepatient to contact the NCIMD for results within a specifiedtime period.

*ere was no detriment to the frequency or actual Phelevels after introduction of this system. *e 4–10-year-oldgroup did show a statistical significance before and aftertexting implementation, but this was not clinically significant.

In terms of the retrospective review of the frequency ofPhe level sampling before and after introduction of thetexting system, no statistical difference was found. *epercentage of texts within the normal range was 52%.

*is project required the integration of three disciplinesin the hospital: Dietetics, ICT, and Metabolic Laboratory.*e success of this project relied heavily on the collaborationand cooperation between these departments.

Feedback has confirmed that the system now in place issaving patients’ time (Table 7). *is frees up dietetic time forpatients/parents who require more intense input whenmetabolic control is not stable and for new parents. A benefitof the system which was not thought of at the planning stagewas that parents/patients now have a paperless record torefer to on their mobile. A study by Bilginsoy et al. in 2005highlighted record keeping as one of the obstacles to betteradherence in PKU [5]. Previously levels were recorded inbooks at home for reference.

Patient safety is always prioritised through the validationof results. A designated dietitian checks that all texts are sentand received once Phe results have been authorised by thelaboratory. It is the parent’s responsibility to contact thedietitian if levels are not within the range or the text messageis not received. *e text message sent is recorded on manuallevels sheets, so there is a record if patients do not respond toout of range text messages. Unfortunately, there is no ICTsystem which flags these to the dietitian.

*is project has huge potential to be applied in othertherapeutic conditions which require regular blood testmonitoring, for example, patients on warfarin or those withrenal conditions. *e main transferable feature of thisproject is the ability to link the laboratory system, fora variety of conditions, and the texting system by usingtechnology. McGillicuddy et al. studied mobile phone-basedprograms for the kidney transplant recipient. *ey founda participation and retention rate of 41/55 (75%) and 31/34(91%), respectively. *is was a 3-month proof-of-conceptrandomized controlled trial which was conducted in 20hypertensive kidney transplant patients [6]. Applebaumet al. carried out a survey in 20 patients aged 13–21 yearsfrom a paediatric university-based clinic. *ey identifieda preference for using text messaging for communication inthis cohort [7]. *e use of text messages by a GP in Londonto inform patients about results of routine tests freed upgreater than 600 appointments in the practice per year [8]. Astudy by Kerrison et al. in 2015 found that women who weresent a text message reminder before their first routine breastscreening appointment were more likely to attend [9]. Astudy by Hunt in 2015 found that technology was a supportto patients with diabetes in areas including taking medi-cation and monitoring for complications [10].

*e study period for the texting system was undertakenbefore the publication of the European Guidelines for Di-agnosis and Management of PKU [11]. *is study used theIrish reference rangewhich suggests a range of 120–400µmol/l.In order to accurately analyse results, the original referencerange was used for all analysis. From 1st July 2017, theranges have been amended to the European guidelines forPKU [11].

*e cost of implementing the system was minimal at€1000 for 2 licences. *is project was not about savingmoney. However, savings of €3,275 which were a result ofthe reduced cost of a text versus a phone call for levels withinthe recommended range was an advantage in Ireland inrecessionary times.

Patient satisfaction with the system was equally as im-portant as patient safety. In order to assess the effectiveness

Table 4: Overall how do you find the PKU texting system?

Poor Average Good Verygood Excellent

My rating of thetexting system 0% 0% 13% 34% 53%

Table 5: Do you think the system has had an impact on you/yourchild’s PKU blood levels?

Disimproved No change ImprovedPKU levels 0% 71% 29%

Table 6: Has the PKU texting system had an influence on you/yourchild in terms of dietary compliance?

Better thanbefore

Worse thanbefore

Nodifference

Dietarycompliance 16% 0% 84%

Table 7: Is the texting system saving you time?

Yes NoSaving you time 94% 6%

4 Journal of Nutrition and Metabolism

of this quality improvement initiative, a patient survey wascarried out in 2017. *is showed 87% rated the system aseither very good or excellent (Table 4). 94% agreed it wastime saving which was very positive. 84% felt there was noinfluence on dietary compliance (Table 6). *e system wasnot intended to impact dietary compliance in any way(Tables 3 and 5). Macdonald et al. cited technology as an aidin the reality of dietary compliance in PKU [12].

6. Conclusion

Text messaging Phe results is safe, time saving, and a moreefficient use of dietetic time. *is technology has the po-tential to be applied in other settings such as clinics andhospitals internationally.

Disclosure

*is manuscript was presented before as an abstract inthe 13th International Congress of Inborn Errors ofMetabolism-ICIEM 2017.

Conflicts of Interest

*e authors declared no potential conflicts of interest withrespect to the audit, authorship, and publication of thisarticle.

Acknowledgments

*e authors thank the 28 patients who piloted the system for3 months and the National Centre for Inherited MetabolicDisorders Multidisciplinary Team.

References

[1] Defero is an Enterprise Messaging Solution, http://www.grapevine.ie.

[2] Pews Research Center, “Global digital communication: text-ing, social networking popular worldwide,” Survey Report,Pews Research Center, Washington, DC, USA, 2012.

[3] Deloitte, Global Mobile Consumer Survey: <ere’s no place likephone, Deloitte, Ireland, 2016.

[4] E. R. Jurecki, S. Cederbaum, J. Kopesky et al., “Adherence toclinic recommendations among patients with phenylketon-uria in the United States,”Molecular Genetics andMetabolism,vol. 120, no. 3, pp. 190–197, 2017.

[5] C. Bilginsoy, N. Waitzman, C. O. Leonard, and S. L. Ernst,“Living with phenylketonuria: perspectives of patients andtheir families,” Journal of Inherited Metabolic Disorders,vol. 28, no. 5, pp. 639–649, 2005.

[6] J. W. McGillicuddy, M. J. Gregoski, A. K. Weiland et al.,“Mobile health medication adherence and blood pressurecontrol in renal transplant recipients: a proof-of-conceptrandomized controlled trial,” JMIR Research Protocols,vol. 2, no. 2, p. e32, 2013.

[7] M. A. Applebaum, E. F. Lawson, and E. Von Scheven,“Perception of transition readiness and preferences for use oftechnology in transition programs: teens’ ideas for the future,”International Journal of Adolescent Medicine and Health,vol. 25, no. 2, pp. 119–125, 2013.

[8] 2015, https://www.gponline.com/sms-service-saves-gp-practice-600-appointments-year/article/1329935.

[9] R. S. Kerrison, H. Shula, D. Cunningham, O. Oyebode, andE. Friedman, “Text- message reminders increase uptake ofroutine breast screening appointments: a randomised controltrial in a hard to reach population,” British Journal of Cancer,vol. 112, no. 6, pp. 1005–1010, 2015.

[10] C. W. Hunt, “Technology and diabetes self-management: anintegrative review,” World Journal of Diabetes, vol. 6, no. 2,pp. 225–233, 2015.

[11] F. J. Van Spronsen, A. M. J. Van Wegberg, K. Ahring et al.,“Key European guidelines for the diagnosis and managementof patients with phenylketonuria,” <e Lancet Diabetes En-docrinology, vol. 5, no. 9, pp. 743–756, 2017.

[12] A. MacDonald, H. Gokmen-Ozel, M. V. Rijn, and P. Bugard,“*e reality of dietary compliance in the management ofphenylketonuria,” Journal of Inherited Metabolic Disorders,vol. 33, no. 6, pp. 665–670, 2010.

Journal of Nutrition and Metabolism 5

Research ArticleAge-Related Reference Intervals for Blood Amino Acids inThai Pediatric Population Measured by LiquidChromatography Tandem Mass Spectrometry

Jaraspong Uaariyapanichkul,1,2 Sirinuch Chomtho ,1,2 Kanya Suphapeetiporn ,3,4

Vorasuk Shotelersuk,3,4 Santi Punnahitananda,5 Pannee Chinjarernpan,6

and Orapa Suteerojntrakool 2,7

1Division of Nutrition, Department of Pediatrics, King Chulalongkorn Memorial Hospital, �e �ai Red Cross Society,Bangkok 10330, �ailand2Pediatric Nutrition STAR (Special Task Force for Activating Research), Department of Pediatrics, Faculty of Medicine,Chulalongkorn University, Bangkok 10330, �ailand3Center of Excellence for Medical Genetics, Department of Pediatrics, Faculty of Medicine, Chulalongkorn University,Bangkok 10330, �ailand4Excellence Center for Medical Genetics, King Chulalongkorn Memorial Hospital, �e �ai Red Cross Society,Bangkok 10330, �ailand5Division of Neonatology, Department of Pediatrics, Faculty of Medicine, Chulalongkorn University, Bangkok 10330, �ailand6Center for Medical Diagnostic Laboratories, Faculty of Medicine, Chulalongkorn University, Bangkok 10330, �ailand7Division of Ambulatory, Department of Pediatrics, King Chulalongkorn Memorial Hospital, �e �ai Red Cross Society,Bangkok 10330, �ailand

Correspondence should be addressed to Orapa Suteerojntrakool; orapa.s@chula.ac.th

Received 13 October 2017; Revised 19 February 2018; Accepted 28 February 2018; Published 6 May 2018

Academic Editor: Ina Knerr

Copyright © 2018 Jaraspong Uaariyapanichkul et al. -is is an open access article distributed under the Creative CommonsAttribution License, which permits unrestricted use, distribution, and reproduction in anymedium, provided the original work isproperly cited.

Background. Age, race, and analytic method influence levels of blood amino acids, of which reference intervals are required for thediagnosis and management of inherited metabolic disorders. Objectives. To establish age-specific reference intervals for bloodamino acids in-ai pediatric population measured by liquid chromatography tandemmass spectrometry (LC-MS/MS).Methods.A cross-sectional study of 277 healthy children from birth to 12 years was conducted. Anthropometric, clinical, and dietaryinformation were recorded. Dried blood spots on a filtered paper were used for measurement by derivatized LC-MS/MS. Factorsthat might affect amino acids such as fasting time and dietary intake were analyzed using quantile regression analysis. Results.Levels of thirteen blood amino acids were reported as median and interval from 2.5th–97.5th percentiles. Compared with those ofCaucasian, most blood amino acid levels of -ai children were higher. Compared with a previous study using HPLC in -aichildren, many amino acid levels are different. Glycine, alanine, leucine/isoleucine, and glutamic acid sharply decreased afterbirth. Citrulline, arginine, and methionine stayed low from birth throughout childhood, whereas phenylalanine was at middlelevel and slightly increased during preadolescence. Conclusion. Reference intervals of age-specific blood amino acids using LC-MS/MS were established in the-ai pediatric population.-ey diverge from previous studies, substantiating the recommendationthat, for the optimal clinical practice, age-specific reference intervals of amino acids should be designated for the particularpopulation and analysis method.

HindawiJournal of Nutrition and MetabolismVolume 2018, Article ID 5124035, 10 pageshttps://doi.org/10.1155/2018/5124035

1. Introduction

Amino acids are the basic structural units of protein. -ediversity of amino acids provides variability in the functionand structure of proteins. Some amino acids cannot besynthesized by the body. -erefore, they are consideredessential constituents of the diet for maintenance of healthand growth [1].

Inborn errors of metabolism (IEM) are a complex andheterogeneous group of genetic disorders, caused by a defectin a metabolic pathway, leading to malfunctioning meta-bolism and/or the accumulation of toxic intermediate me-tabolites. IEM can be presented at any age with nonspecificclinical manifestations, complicating diagnostic pathways.Furthermore, consequences are severe, causing morbidityand mortality in clinical practice, especially in pediatrics [2].Although each disorder is individually rare, their cumulativeincidence is substantial; an incidence of 1 in 2500–5000 livebirths has been shown to be upwards of 1 in 800 [2, 3]. In-ailand, from a retrospective study, the estimated pediatricpatients with clinically suspected IEM are approximately2–4%. However, in high-risk infants and children, theprevalence rates of 11.2% were reported [4–6]. Nevertheless,no study has been carried out to identify the magnitude ofthe problem in a systematic way [7]. As the delay in thediagnosis can result in irreversible outcomes [8], mea-surement of the levels of amino acids in the plasma is usefulin the early diagnosis for the disorders to be treatedpromptly and for continuous monitoring of patients. Nu-tritional management includes limiting the substrates whichtrigger the symptoms while preventing their deficiencies.Patients will be assigned the type and amount of dietsthat can be taken without causing harm while maintainingadequate growth under the supervision of a nutritionspecialist [9].

Analysis of amino acids is considered to have an ex-tremely important role in the diagnosis and care of patients,detecting the type and amount of amino acids that areabnormal to identify the inborn errors of metabolisms,called aminoacidopathies. Originally, ion-exchange chro-matography coupled with postcolumn ninhydrin or liquidchromatography combined with postcolumn derivatizationusing ninhydrin and UV detection was used [10] but sufferedfrom a long analysis time, instability of reagents, and cross-reactivity with other constituents of the biological sample.Currently, there are several analytical methods, such ashigh-performance liquid chromatography (HPLC), gaschromatography (GC), and mass spectrometry (MS) [11].Some laboratories in-ailand can perform HPLC, but theremay be some drawbacks such as limited types of aminoacids, long separation time, and high interlaboratory var-iation [12].

-e use of two mass spectrometers in tandem (MS/MS)enables control of the formation of molecular and fragmentions.-e combination of molecular mass (or mass-to-chargeratio) and a specific fragment after collision-induced dis-sociation is achieved for compounds where there is noisobaric similar compound. -ere may be the chance thatsome very unstable compounds might dissociate in the ion

source and yield unspecific readings. However, this is highlyvariable across the different amino acids. Due to specificprecursors and product ions created by the tandem massspectrometers, this method achieves a sensitivity andspecificity up to 99% and 99.9%, respectively, for most aminoacid disorders, organic acidemias, and fatty acid oxidationdefects [2, 13]. Liquid chromatography tandem massspectrometry (LC-MS/MS) can analyze each sample quicklyin 2 minutes, can allow multiple analytes to be tested usingonly a small amount of blood, and can provide high-volumethroughput with rapid turnaround time. False positives areapproximately 0.05%, which is less than the HPLC method.-e accuracy is more than 90%, and the coefficient ofvariation (CV) is between 4 and 10% [14–16]. However, theaccuracy and CVs are highly variably across the differenttypes of amino acids [15]. -us, a level of expertise is neededfor preparing samples, operating the system, and inter-preting the data. Despite the high initial capital cost forequipment, reagent cost is relatively low [2].

Reference values are crucial for interpreting test resultsand for making diagnoses [11, 17, 18]. Pediatric populationrequires the use of references that reflect rapid physiologicchanges associated with growth but are often difficult toestablish because of the challenges related to obtainingsufficient numbers of samples from healthy children [19].-ere are no well-established reference intervals of aminoacids in our population, and data from international liter-atures are traditionally used [20]. Levels of amino acids inthe blood are influenced by several factors, such as age, sex,race, fasting time, nutritional status, illnesses, and medi-cations [20–23]. A previous reference study in -ailand in2001 was based on plasma amino acid analysis by the HPLC[24]. However, due to limited resources, the analysis methodwith plasma samples is not widely used in -ailand.Moreover, since the collected specimens must be trans-ported to academic or university hospitals, the shipping ofspecimens is more conveniently achieved as dried blood spot(DBS) samples. In addition, the changing of environmentand nutrients’ composition over time, in particular proteinintake, in young children might affect amino acid levels [22].Reference values derived by different analysis methodscannot be used interchangeably, and data obtained at a givenlaboratory are recommended to provide quality services[25]. -ere are emerging researches involved with referencevalues of amino acids around the world [22, 26–29].-e aimof this study was to establish the age-related reference in-tervals for amino acids by LC-MS/MS, which had never beendone in -ailand. Secondary objective was to assess factorsthat might affect amino acids.

2. Materials and Methods

2.1. Study Design. -is cross-sectional descriptive study wasconducted at the King Chulalongkorn Memorial Hospital(KCMH) from March 2016 to May 2017 with approval fromthe Institutional Review Board of the Faculty of Medicine,Chulalongkorn University, and informed consent was ob-tained from the parents. Study sites included outpatientdepartment, well-baby clinic, full-term neonatal ward, and

2 Journal of Nutrition and Metabolism

settings in accordance with ongoing pediatric departmentalactivities or research projects from the same locality(e.g., pediatric infectious unit). -e recruitment of healthysubjects was also carried by the social media of Chula Kids’Club (together with immunization program) and schoolhealth checkup programs.

2.2. Study Population. A total of 277 healthy infants andchildren from 0 to 12 years of age were included in thisstudy. -ey were divided into five age groups: <4 days old,6–12 months, 1–3 years, 3–6 years, and 6–12 years. -e ageranges were selected to represent the physiologic periods,such as newborn period and early and late childhood [24].We excluded the subjects with known inborn errors ofmetabolism, disorders for which variations of amino acidlevels have been described [17, 20, 26, 27], such as skindisease, neurological disease, gastrointestinal disease, liverdisease, kidney disease, endocrine disease, cancer, infection,known nutritional abnormalities, or were taking any med-ication. All infants were full term, appropriate for gestationalage, had normal birth weight, and were not born to diabeticmothers.

2.3. Data Collection. Data collection from the medical re-cords and physical examination were documented into thecase record forms with the details of age, sex, weight,height/length, fasting time before collecting blood samples,and 24 hr dietary recall.

2.4. Sample Collection. For <4 day-old infants, blood sam-ples were collected from the heel stick onto the filter paper(Whatman 903) as dried blood spots (DBSs). For childrenaged at least 6 months, blood samples were collected asheparinized blood 1-2ml and then pipetted 55 µl of bloodonto the filter paper. -en, the blood spots were left to dryfor 4 hours on a horizontal, nonhumid and nonabsorbentsurface at ambient temperature. -e specimens were pro-tected in dry plastic bags and were stored and frozen at−20°C until analysis within 1 month. After the DBS wascollected, the remaining whole blood samples were thencentrifuged, and the plasma were stored frozen at −80°C forfuture use.

2.5. Analysis of Amino Acids. Analysis was performed withWaters Acquity TQ-S, using MassLynx 4.1 and NeoLynx 4.1(Waters, Milford, MA, USA). Preparation of MassChrom®Amino Acids and Acylcarnitines from Dried Blood(Chromsystems, Munich/Germany) was in accordance withthe certified manufacturer’s specifications [14], as follows. A3.2mm dried blood spot disk was punched out of the filtercard into a 96-well microtiter plate. 200 µl of the recon-stituted internal standard was added. -e plate was sealedwith a protective sheet and agitated for 20min at 37°C and600 rpm. -e supernatant was transferred into a newmicrotiter plate. 150 µl of the internal standard for succi-nylacetone was added to the remaining blood spots. -e 96-well plate was sealed with a protective sheet and agitated at

600 rpm for 30min at 60°C. -e supernatants were pooledfrom after the addition of the internal standard for succi-nylacetone with the first supernatant and evaporated at 60°Cand 600 rpm to dryness. 60 µl derivatisation reagent (butan-1-ol, n-butyl acetate, and hydrogen chloride) was added, themicrotiter plate was sealed with a protective sheet, incubated15min at 60°C and 600 rpm, and evaporated at 60°C and600 rpm to dryness. 100 µl reconstitution buffer was addedand agitated for 1min at 600 rpm. 10 µl was injected into theLC-MS/MS system by an autosampler. -e operating con-ditions had 1.7-minute run time and mobile phase gradientwith a flow rate of 20–600 µl/min (Table 1).

2.6. Statistical Analysis. Data analysis was performed usingStata version 13.1 (Stata Corp., College Station, Texas). Fordescriptive analysis, the frequencies and percentages ofcategorical variables for the population characteristics werecalculated, while median, interquartile ranges (IQR), the2.5th, and 97.5th percentiles were calculated for continuousvariables. -e Kruskal–Wallis test was used to comparecontinuous variables between age groups. Levels of aminoacids in µmol/L were expressed as median since theywere not normally distributed according to tests for nor-mality (Kolmogorov–Smirnov and Shapiro–Wilk) and wereskewed to the right. -us, reference intervals were deter-mined nonparametrically and correspond to the 2.5th–97.5th percentiles of the distribution. Quantile regressionanalysis was used to determine the factors associated withamino acids in both univariate and multivariate models.Multivariate models were developed by including covariateswith p< 0.1 in univariate models. All p values reported aretwo-sided. Statistical significance was defined as p< 0.05.

3. Results

Table 2 shows the characteristics of subjects including de-mographic data and dietary intake, displayed as median andinterquartile range (IQR) and anthropometry as mean (SD).Regarding gender, 51.3% of 277 subjects were male. Allsubjects were in good nutritional state by WHO z-scorecriteria and had normal protein intake according to theirage. Of note, the energy and macronutrients intake, in-cluding caloric distribution, was in accordance with the 4th-ai National Health Examination Survey in 2008–2009(NHES IV) [30], with a slightly higher energy intake of the3–6 years age group. Fasting duration before collecting theblood samples was displayed in minutes. Children aged 6–12years had an overnight fast for their health checkups. Re-garding the <4 days old age group, their average gestationalage was 38.5 weeks with 3058 g birth weight. Dietary data ofinfants aged <4 days were not demonstrated, as all thesubjects of this group were newborns aged 48–72 hours

Table 1: Operating conditions with changing flow rate during therun time for the analysis (adapted from InstructionManual for LC-MS/MS Analysis MassChrom [14]).Time (min) 0 0.24 0.25 1.24 1.25 1.5 1.51Flow (µl/min) 200 200 20 20 600 200 200

Journal of Nutrition and Metabolism 3

(mean age 58.3 hours) who were predominantly breastfedand had their blood taken simultaneously with the -aiNational -yroid Screening Program.

3.1. Amino Acid Results. -irteen amino acid levels werereported as the median for each of the five age groups: <4days old, 6–12 months, 1–3 years, 3–6 years, and 6–12 years,as shown in Table 3. -e age-specific reference intervalscovered the central 95 percent of the test results, as repre-sented by the 2.5th percentile and 97.5th percentile, re-spectively, according to the CLSI EP28-A3c guideline [31].Median concentrations were significantly different for allamino acids when compared across age groups using theKruskal–Wallis test (mostly p � 0.0001). Since the levels ofeach amino acid showed no significant differences betweenmales and females, the data were considered together. -eisobaric analytes of leucine, isoleucine, alloisoleucine, andhydroxyproline could not be separated by this method. -ey

were summed up under the same mass-to-charge ratio andreported as “leucine/isoleucine” concentrations in this study.

3.2. Trends of Age-Specific Concentration of AminoAcids. Graphic forms showing trends of the age-specificdistribution of amino acid median concentrations are il-lustrated in Figure 1. Although each amino acid possessed itsown unique pattern of distribution, we can identify somecommon tendencies into two major different profiles.

-e first group consisted of seven amino acids (glycine,ornithine, alanine, leucine/isoleucine, tyrosine, asparticacid, and glutamic acid) that drop sharply from the zero-time point collected shortly after birth to the median ofthe 6–12 month age group. Especially glycine, alanine,leucine/isoleucine, and glutamic acid demonstrated a sharpdecrease in their concentrations approximately to half oftheir initial amounts. Afterwards, their concentrationsremained stable; except for glycine and leucine/isoleucine

Table 2: Characteristics of 277 subjects regarding demographic data, anthropometry, and dietary intake (BM� breast milk predominant;C� carbohydrate; P� protein; F� fat).

n <4 days old 6–12 months 1–3 years 3–6 years 6–12 years85 43 49 50 50

Demographic DataAge (month) 0.08∗ (0.07–0.09) 9 (7–11) 23 (18–28) 44 (40–57) 120 (108–138)Sex: male, n (%) 43 (51) 20 (47) 21 (43) 24 (48) 34 (68)Anthropometry, mean (SD)W/A z-score −0.5 (0.52) −0.26 (1.13) −0.17 (1.13) −0.33 (1.44) 0.35 (0.47)W/H z-score −0.72 (0.78) 0.06 (0.95) 0.11 (1.2) 0.13 (1.37) −0.19 (0.75)BMI z-score −0.73 (0.63) −0.02 (1) 0.21 (1.2) 0.08 (1.5) 0.2 (1.17)Dietary dataEnergy (Kcal/day) BM 973 (708–1121) 1049 (854–1402) 1304 (1144–1489) 1230 (1181–1382)Energy (Kcal/kg/day) BM 107 (86–130) 98 (80–124) 88 (74–100) 41 (33–48)Protein (g/day) BM 24 (15–32) 40 (35–52) 49 (42–56) 50 (40–65)Protein (g/kg/day) BM 3 (1.8–3.8) 3.9 (2.8–5.2) 3.2 (2.5–4) 1.7 (1.3–2.2)C : P : F BM 50 :13 : 37 50 :16 : 34 54 :16 : 30 50 :17 : 33Fasting time (min) — 120 (30–165) 135 (90–210) 140 (90–190) ≥720∗Age: <4 days old; mean age: 58.3 hours. Data were expressed as median (IQR) unless stated otherwise.

Table 3: Distribution of amino acid concentrations by age groups, displayed as median, 2.5th, and 97.5th percentiles.

Amino acid(µmol/L)

<4 days 6–12 months 1–3 years 3–6 years 6–12 years

Median P2.5

P97.5 Median P 2.5 P

97.5 Median P 2.5 P97.5 Median P 2.5 P

97.5 Median P 2.5 P97.5

Glycine 345 300 414 179.9 146.3 251.6 205.5 168.2 332.2 181.6 157.1 229 276.9 257.1 305.9Proline 122 97 150 124 101.2 155.1 132 107 183.4 104.5 88.9 143.7 124.3 111.6 155.1Ornithine 151 124 185 88.5 69.3 122.5 95.3 70.2 137.1 98.7 82.4 111.3 80.9 68.7 88.4Citrulline 16.9 13.8 23 20 18.5 25.6 27 21.9 31.8 26 22.1 31.8 27 24.4 28.8Arginine 8.4 5.6 14.7 17.2 14.2 23.5 17.4 12.7 25.1 18.4 13.8 25.8 17.2 15.4 20.7Alanine 543 424 633 290.3 253.5 355.8 316 258.1 406.2 332.2 293 386.9 256.9 221.6 307.8Valine 112 89.1 179.2 136 118 160.3 154 120.1 205.3 117 103 174.4 150 128.4 171.4Leucine/isoleucine 602 511 703 376 305.4 440.1 355 282.9 436.2 319 275.3 396.1 397 343 472.5Methionine 14.1 10.7 17.7 14.6 9.4 21.9 11.1 7 18.5 11.5 8.2 15.6 19.8 17.1 22.6Phenylalanine 59 50.3 70.1 50.5 40.5 62.7 63.2 49.3 72.3 53.9 47.1 84.2 59.2 53.3 66.2Tyrosine 91.1 74.1 114.4 73.5 52.6 85.5 78 58.3 93.3 61.1 50.1 91.3 72.1 62.3 84.3Aspartic acid 205.6 102 216 125.7 94 157.4 124 87.9 212.4 104.5 85.7 137.8 120.5 88.1 148.8Glutamic acid 592 485 687.4 237 194 306 246 201.2 283.8 239 214.8 262.3 232 211 258.7-e age-specific reference intervals were represented by the 2.5th and 97.5th percentiles according to the CLSI EP28-A3c guideline [31].

4 Journal of Nutrition and Metabolism

which increased throughout older childhood (after 6 years ofage) and alanine which peaked at 3–6 years and decreasedafterwards.

-e second group consisted of six others (citrulline, ar-ginine, methionine, proline, phenylalanine, and valine) whichshowed steady concentrations throughout infancy and child-hood. Citrulline, arginine, and methionine were at low levels(below 30µmol/L), while proline and phenylalanine were atmiddle levels of 50–130µmol/L and slightly increased duringpreadolescence. Lastly, the level of valine tended to slightlyincrease up to 3 years of age, followed by a period of modestdecrement and subsequently rise again after 6 years of age.

3.3. Factors Affecting Amino Acids. Factors that might affectamino acids such as demographic data, fasting duration, anddietary intake were analyzed through quantile regressionaimed at estimating the conditional median in both uni-variate and multivariate models. Initially, univariate analysisdemonstrated several potential factors related to amino acidconcentrations such as sex, energy intake, protein intake, andfasting time (data in Supplementary Table 1). Univariateanalysis of fasting time showed significant association withalanine, glycine, methionine, and citrulline. However, mul-tivariate models adjusting for these potential covariates withp< 0.1 in univariate analysis revealed that the remainingsignificant associated factor was only energy intake on gly-cine and phenylalanine. Subsequently, the acquired co-efficient values for each amino acid were interpreted todemonstrate the effects. For every 1Kcal/kg/day increasedenergy intake, glycine decreased by 0.7 µmol/L (p � 0.01)and phenylalanine decreased by 0.1 µmol/L (p � 0.03). Afterbeing adjusted in the multivariate models, protein intake andfasting time were not significant factors. In addition, we did

not find any further associated factors through quantileregression analysis aimed at estimating the higher condi-tional 90th percentiles (data not shown).

3.4.QualityControl. -e samples were run onmultiple assayplates. Each sample was run with its own internal standarduse as calibrators, and 2 controls with known concentrationsof the analyte for a high and a low value were included oneach plate for every 20 specimens. -e accuracies weredetermined by comparing the test results of 2 controls of theanalytes (a high and a low value) with the known targetvalues (provided from the test kit). Also, the test values ofthese controls had to fall within the given suggested rangesfrom the instruction manual. -en, we calculated the per-centage of the test values to the known target values for eachresult. Finally, the percentage ranges were listed for eachanalytes. -e precisions were determined by the coefficientof variation (CV). -e number (N) of values used to cal-culate the CVs is 14 each for high and low values.

-e results of the intra-assay CVs as shown in Table 4were satisfactory [32].-e variation found in the interassay ofmore than 10%, but not exceeding 15%, was observed forornithine, alanine, valine, and aspartic acid and was ac-ceptable [32] considering that methods with pipetting, such asthe one used for amino acid analysis, usually yield CVs be-tween 10% and 15% [20, 33]. -e accuracy for each aminoacid was acceptable, mostly at 85–105%, except for the un-derestimation of arginine and overestimation of valine at thelow target. Comparing the results obtained with the previousstudy with LC-MS/MS [15], it was observed that all of the CVsand accuracies found in this study were equivalent or better.

4. Discussion

4.1. Study Population andMethodology. We studied a groupof 277 -ai children from birth to 12 years of age, and thenumber of subjects included in this study yielded largersample size than that in most studies of pediatric referencesat other sites including Asia [20, 24, 27, 28]. Previous dataregarding amino acids in the -ai pediatric population werevery limited. Our study defined reference intervals for wholeblood amino acids in only healthy pediatric population anddid not include children admitted to the hospital with minorillnesses, contrary to some previous studies [20, 24]. Webelieve that our subjects who were all recruited in the urbanarea of Bangkok can sufficiently represent the -ai pediatricpopulation because geographical distribution does not di-rectly affect amino acids, and our results of dietary intake,especially protein, were in accordance with the 4th -aiNational Health Examination Survey (NHES IV) [30]. Al-though, an experimental study in Mexico showed that in-gestion of an urban diet induced a higher increase in theplasma concentration of some amino acids than ingestion ofa rural diet, the amount of protein consumed in the urbandiet was greater than in the rural diet [34].

-e recruitment of infants aged <4 days as the first agegroup was primarily intended for future implementation ofnewborn screening and could result in the selection bias.

0

100

200

300

400

500

600

700

<4 days 6–12 months 1–3 years 3–6 years 6–12 years

Am

ino

acid

conc

entr

atio

ns (µ

mol

/L)

Age

GlycineProlineOrnithineCitrullineArginineAlanineValine

Leucine/isoleucineMethioninePhenylalanineTyrosineAspartic acidGlutamic acid

Figure 1: Trends of the age-specific distribution of each amino acidmedian concentration.

Journal of Nutrition and Metabolism 5

However, the cultural tradition that healthy infants under 6months of age in-ailand rarely have their blood taken leadsto the limitation of finding sufficient numbers of participantsof this age group.-is suggests the room for improvement inthe future.

Analysis of amino acids in this study was performed byLC-MS/MS and demonstrated reliable precision and accu-racy comparable with previous reports of LC-MS/MS innewborns [15, 19] and other analytical methods such asHPLC while maintaining the advantage of short analysistime. Fingerhut et al. found that the accuracy and CVs werehighly variable across the different analytes, in a way thatbasic amino acids (CV 6.4–18.1%) and acidic amino acids(CV 7–13.7%) demonstrated higher CVs when comparedwith neutral amino acids (CV 3.3–5%) [15]. However, wecould not demonstrate the differences.

4.2. Trends ofAge-SpecificConcentration ofAminoAcids. -eage-related median values of seven amino acids (glycine,ornithine, alanine, leucine/isoleucine, tyrosine, aspartic acid,and glutamic acid) drop sharply from the zero-time pointcollected shortly after birth to the median of the 6–12 monthage group (Figure 1). Previous studies in -ailand have notbeen able to demonstrate the drop of amino acids during theage interval of 0–6 months [24]. It remains possible that thevalues drop sharply sooner than before 6 months of age ifgiven another age group between these two. -e gap ofknowledge during the 0–6 months of age period should befurther explored.

4.3. Comparison of Amino Acids with Previous Studies. Normalamino acid values in children vary considerably in differentreports [24]. Because there is no standard age-specificreference of amino acids for LC-MS/MS in -ai children,our results were compared with those of the study ofLepage et al. in Caucasian children by ion-exchangechromatography [20]. -ai reference was different in thatmost amino acid concentrations were higher, except for thesame ranges of citrulline, valine, proline, methionine, andlower arginine.

Most amino acids of -ai newborn in this study werecomparable with the recent study of LC-MS/MS on DBSsconducted only in newborns aged 0–4 days in the USA [19].However, some amino acid levels were still higher, notablyornithine, leucine/isoleucine, glutamic acid, and alanine.Because our method showed satisfactory accuracy withknown analytes, the reasons for these discrepancies couldresult from the difference in the ethnicity but might besecondary to other preanalytical errors such as technicalproblems regarding blood sample collection and storage,matrix effects [35], internal interferences (asparagine onornithine and hydroxyproline on leucine) [36], and differ-ences in extraction efficiency. Although hemolysis was foundto cause increases in ornithine, aspartic acid, glutamic acid,and decreased arginine [37], it should not affect whole bloodsamples as collected in this study.

Compared with the study using different analysismethod of HPLC in -ai children [24], six amino acids(glycine, ornithine, alanine, leucine/isoleucine, asparticacid and glutamic acid) were higher in our study. -ismeans that we should not directly use the reference in-tervals from different analytical methods. -e accuracy,intra-assay CVs, and interassay CVs of HPLC [28] and ourstudy were approximately the same, which suggested thatboth analytical methods are acceptable. -e primary ex-planations for these differences may lie within the methodof specimen collection as dried blood spots (DBSs), whichwas primarily designed for newborn screening, versusplasma specimens. Even with cautious handling, smallconcentration gradients of analytes might be presentduring spreading and drying of blood on the filter paper;combined with the effect of hematocrit, this can cause anoverall imprecision of approximately 10% [2]. Likewise,mildly elevated glutamate concentrations can result fromthe age of the specimens allowing for glutamine to gluta-mate conversion [23], even when stored at the recom-mended temperature. In order to overcome these problems,we propose more “transference studies” for comparison ofplasma and DBS samples [2].

-e discordance regarding common trends of amino acidpattern profiles was also observed. While arginine, methio-nine, proline, and phenylalanine demonstrated a decrease in

Table 4: Accuracy, intra-assay, and interassay CV (%) of each amino acid at two different concentrations.

Analyte (µmol/L) Lowtarget

Accuracy(%)

Intra-assay CV(%)

Interassay CV(%)

Hightarget

Accuracy(%)

Intra-assay CV(%)

Interassay CV(%)

Glycine 355 98–100 5.7 4.8 1018.5 99.5–101.6 5.7 8.3Proline 271.5 100–101 2.8 4.5 695 62.5–100 2.1 2.5Ornithine 219 64.9–82.4 2.5 13.1 519 65.8–85.6 1.2 11.7Citrulline 86 88.2–94.5 2.1 9.4 313 79.5–93 2.4 7.2Arginine 113 47.2–56.2 1.3 2.3 240 61.3–63.7 1.6 2.9Alanine 320 100–110.2 2.5 9 563 108–138.3 7.5 11.2Valine 159 124.5–128.4 7.1 8.2 361 85.5–106.4 9 12.2Leucine/isoleucine 224 110.3–137 5.5 7.6 537 97.7–115.4 4 4.5Methionine 46 99.3–100 5.2 8.6 172 100 2.3 4.5Phenylalanine 97 106.4–115.4 2.8 5.8 420 91.3–101.8 0.7 2Tyrosine 175 99.8–105 3.1 3.8 536 83.4–95.2 1.6 2.9Aspartic acid 161 102.1–106.6 2 10.5 282 107.5–112.3 6.8 7.1Glutamic acid 489 91.8–93.4 3.6 7.4 743 77.1–90.3 2.7 5.3

6 Journal of Nutrition and Metabolism

their concentrations during the first year of life in Caucasianchildren, they differently showed steady concentrations inour study. Nevertheless, aspartic acid and glutamic acidsimilarly demonstrated decreasing values. Valine alsoshowed the same trend throughout infancy and childhood.-e discrepancy between the previous studies and our workcould be due to differences in methodology, equipment,population, measurement conditions, or from dietary pat-terns which were not elucidated.

4.4. Factors Affecting Amino Acids. Age was confirmed to bean important factor because median concentrations for allamino acids across age groups were significantly different,similar to previous studies in-ailand and Spain [20, 24, 27].-is study showed no significant differences for amino acidsbetween both genders, in contrast with previous studies inTurkey and China [26, 28].

From the quantile regression analysis (SupplementaryTable 1), the only significant associated factor was energyintake on glycine and phenylalanine. For every 1Kcal/kg/day increased energy intake, glycine decreased by 0.7 µmol/L(p � 0.01) and phenylalanine decreased by 0.1 µmol/L(p � 0.03). However, the changes in amino acid concen-trations, acquired as coefficient values, were so small as toprobably have limited clinical significance. Glycine isa nonessential amino acid and is a biomarker for diagnosis ofnonketotic hyperglycinemia. While phenylalanine is anessential amino acid, the observed effect on the concen-trations seemed to be too minute to have important clinicalimpact. Further researches specially designed to study theeffect of these potential factors should be developed.

Protein intake was not a significant factor after beingadjusted in the multivariate models from quantile regressionanalysis. -is could result from inadequate power to detectminuscule differences between dietary protein intakes in thenormal ranges of our subjects. Our study cannot identify theexact protein intake of newborns because we have no way ofdetermining the protein content of the breast milk that theinfant study participants consumed. However, previous datademonstrated that high dietary protein intake affectedamino acids metabolism in infants aged 6 months [38] ormay even be earlier at 4 months of age [22].

Univariate analysis of fasting time showed significantassociation with alanine, glycine, methionine, and citrulline.For example, every 1 minute of increased fasting duration,alanine decreased by 0.1 µmol/L (p � 0.001) which was thelargest change of concentrations observed. -is could beexplained with gluconeogenesis from this glucogenic aminoacid during the postabsorptive state. However, it was also aninsignificant factor after being adjusted in the multivariatemodels. Moreover, the effect size on the change of aminoacid concentrations seemed not to be large enough to havesubstantial clinical impact. Previous studies tended to fastchildren after 2 years of age for 8–10 hours [20, 26, 28] beforeblood collection, while some studies did not mention aboutthe duration of fasting [24, 27]. However, no study clearlyelucidated the influence of fasting on amino acids in human.Hence, we suggested that with this method, fasting should

not always be obliged in children, especially in infants. -iscomplies with current practices which do not require pro-long fasting before blood collection in infants and youngerchildren for convenience, along with reducing distress andchances of hypoglycemia.

Despite some outliers of amino acid levels, all subjectswere confirmed to be doing well without any abnormalclinical status by tracing back from the case record forms andtelephone contact. Also, there was no report from thesubjects’ parents to the investigators or nutrition unit staffsregarding the subjects’ irregularity. -ese data were sug-gested to be resulted from in vitro or technical variations.

4.5. Clinical Applications. For the reason of the diagnosis ofinborn errors of metabolism, the abnormal levels of aminoacids should be significantly apparent. Although our subjectsof age <4 days were newborns aged 48–72 hours who werepredominantly breastfed, the patients with inborn errors ofmetabolism mostly have distinctly elevated metabolites andshould be detected with this reference, especially whencombined with suspected clinical presentations. Also, for thereason of biochemical follow-up, the trend of amino acidchanges over time along with the patients’ clinical profilesthat could guide us for the management. Valine at the lowerconcentration tended to be overestimated. While this shouldnot be problematic for the diagnosis of inborn errors ofmetabolism, it could result in underdetection of the de-ficiency during the monitoring period. From this study,amino acids such as phenylalanine, tyrosine, andmethioninewere reliable and within previous reported ranges [19, 24].Hence, the benefit was obtained for diseases, such as phe-nylketonuria and tyrosinemia, on the diagnosis and mon-itoring. Nevertheless, this study may have less impact on thediseases such as maple syrup urine disease (MSUD) becauseleucine and isoleucine levels cannot be separated. WhenMSUD patients were on nutritional management with di-etary restriction of BCAAs especially leucine, monitoring ofeach amino acid is important as to avoid excess and de-ficiency. -us, without the complete information of allcommon amino acids, defects of clinical decision may occur.-ere were some occasions of leucine excess with isoleucinedeficiency, and these scenarios cannot be detected by thismethod.

4.6. Considerations and Plans for Future Researches. To thebest of our knowledge, this is the first study in -ai childrenpopulation examining amino acids with LC-MS/MS. -estrengths of our study include larger sample size comparedwith most studies of pediatric references, information onnutritional status, and dietary intake of the subjects. Analysisof amino acids by LC-MS/MS demonstrated reliable pre-cision and accuracy while maintaining the advantage ofshort analysis time. -e specimen collection from the heelstick onto the filter paper as dried blood spots is currentlyutilized as a newborn screening tool in other countries and-ailand [2] because it can be done with ease and requiresless blood than plasma specimens. Moreover, the DBSsamples can be conveniently transported to academic

Journal of Nutrition and Metabolism 7

centers with laboratory facilities. We suggest that wholeblood amino acids might be collected at <4 days age for thepurpose of expanded neonatal screening simultaneouslywith the routine newborn screening in -ailand and fastingmay not always be necessary. -is study was also the first toexpand the application of the analysis onto different pedi-atric age groups. Nevertheless, the gap of knowledge duringthe 0–6 months of age period should be further explored.

Major limitations include incomplete information of allcommon amino acids and inability to separate leucine fromisoleucine and other isobaric analytes. -ere are rooms forimprovement on the analytical methods and proceduresfor the laboratory. For instance, a method of LC-MS/MScompatible with the current Waters Acquity TQ-S machineat the central laboratory, KCMH, using AbsoluteIDQ® p180test kit (Biocrates Life Sciences AG, Austria) is to be de-veloped in the near future. -is reliable test kit whichcovered 20 common amino acids was studied in many recentresearches on targeted metabolites and amino acids [39–41]but has never been used elsewhere in-ailand due to limitedaccess and the need for plasma samples. Furthermore, thereare published methods to overcome the problem regardingthe separation and quantification of BCAAs in patientswith MSUD from the initial extraction of the dried bloodsamples with a sensitive and rapid second-tier UPLC-MS/MSmethod [42].

We aim to develop more expertise regarding specimenpreparation, system operation, and data interpretation. Wesuggest further studies involving amino acids in patientswith inborn errors of metabolism. Also, data of acylcarni-tines and succinylacetone which can aid in the diagnosis ofinborn errors of metabolism should also be studied.

In the absence of established values for amino acids inthe -ai population measured by LC-MS/MS, the presentreference ranges may be used for the diagnosis and man-agement of patients, which may result in less expectedturnaround time and better patient care. We hope that thiswill be a reference for subsequent studies about amino acidsin -ai children with various conditions which affect aminoacids in the future.

5. Conclusion

LC-MS/MS is a rapid and reliable method for the mea-surement of amino acids. With different analysis methods,reference intervals for LC-MS/MS in the -ai pediatricpopulation diverge from previous studies. -us, for theoptimal clinical practice, age-specific reference intervals ofamino acids should be designated for the particular pop-ulation and analysis method. -e reference intervalsestablished in this study may guide us in the diagnosis andmanagement of inherited metabolic disorders, as well asother diseases that affect amino acid metabolism.

Conflicts of Interest

-e authors declare that there are no conflicts of interestregarding the publication of this paper.

Acknowledgments

-e authors are grateful to the infants and children whoparticipated in the study and staffs at the sampled schools(Yamjard Vichanusorn School and Wat Prayun WongsawatSchool).-e authors would like to give special thanks for thekind assistance in recruitment of subjects to Miss NaipapornChuenmeechow (also for dietary analysis), Miss NuanpanSiripen, pediatric nutrition unit staffs (Dr. Puthita Saengpanitand Miss Nuntaporn Charoenphol), KCMH pediatric IDunit staffs (Dr. Chitsanu Pancharoen and Dr. -anyaweePuthanakit), research facilitators, KCMH pediatric OPDand nursery staff, and Ramathibodi pediatric nutritionteam (Dr. Nitiroj Bongkotwilawan). -e authors are indeb-ted for laboratory analysis to Center for Medical DiagnosticLaboratories, Faculty of Medicine, Chulalongkorn University(Dr. Chintana Chirathaworn and Mr. Watchara Sirisuwan),and the company’s technicians (Miss Kanokwan Sae-iewand Miss Khunnalack Khitmoh) and for statistical analysisto Mrs. Jiratchaya Sophonphan. -is study was supportedby -ailand Research Fund (IRG5780015), Helena -aiLaboratories Co., Ltd., and the Chulalongkorn AcademicAdvancement into Its 2nd Century Project. Dr. JaraspongUaariyapanichkul is supported by the scholarship from “-e100th Anniversary Chulalongkorn University Fund forDoctoral Scholarship.”

Supplementary Materials

Supplementary Table 1: factors affecting blood amino acidsin both univariate and multivariate models. Description:factors that might affect blood amino acids such as age, sex,fasting duration, and dietary intake were analyzed throughquantile regression aimed at estimating the conditionalmedian in both univariate and multivariate models. Initially,univariate analysis demonstrated several potential covariateswith p< 0.1. However, the multivariate models revealed thatthe remaining significant associated factor was only energyintake on glycine and phenylalanine (p< 0.05).-e acquiredcoefficient values for each amino acid were then interpretedto demonstrate the effect size. (Supplementary Materials)

References

[1] V. R. Young, “Adult amino acid requirements: the case fora major revision in current recommendations,” Journal ofNutrition, vol. 124, no. 8, pp. 1517S–1523S, 1994.

[2] C. M. Mak, H. C. Lee, A. Y. Chan, and C. W. Lam, “Inbornerrors of metabolism and expanded newborn screening: re-view and update,” Critical Reviews in Clinical LaboratorySciences, vol. 50, no. 6, pp. 142–162, 2013.

[3] S. Sanderson, A. Green, M. A. Preece, and H. Burton, “-eincidence of inherited metabolic disorders in the WestMidlands, UK,” Archives of Disease in Childhood, vol. 91,no. 11, pp. 896–899, 2006.

[4] P. Wasant, E. W. Naylor, and S. Liammongkolkul, “Detectionof inherited metabolic disorders via tandem mass spec-trometry in -ai infants,” Southeast Asian Journal of TropicalMedicine and Public Health, vol. 30, no. 2, pp. 154–159, 1999.

8 Journal of Nutrition and Metabolism

[5] P. Wasant, J. Svasti, C. Srisomsap, S. Liammongkolkul,E. W. Naylor, and I. Matsumoto, “Inherited metabolic dis-orders in -ailand–Siriraj experience,” Southeast AsianJournal of Tropical Medicine and Public Health, vol. 30, no. 2,pp. 124–137, 1999.

[6] P. Wasant, J. Svasti, C. Srisomsap, and S. Liammongkolkul,“Inherited metabolic disorders in -ailand,” Journal of theMedical Association of �ailand, vol. 85, no. 2, pp. S700–S709,2002.

[7] K. -iboonboon, P. Leelahavarong, D. Wattanasirichaigoonet al., “An economic evaluation of neonatal screening forinborn errors of metabolism using tandemmass spectrometryin -ailand,” PLoS One, vol. 10, no. 8, article e0134782, 2015.

[8] N. Vatanavicharn, P. Ratanarak, S. Liammongkolkul,A. Sathienkijkanchai, and P. Wasant, “Amino acid disordersdetected by quantitative amino acid HPLC analysis in-ailand:an eight-year experience,” Clinica Chimica Acta, vol. 413,no. 13-14, pp. 1141–1144, 2012.

[9] F. R. Greer and R. E. Kleinman, Pediatric Nutrition, AmericanAcademy of Pediatrics, Elk Grove Village, IL, USA, 7thedition, 2013.

[10] P. Krumpochova, B. Bruyneel, D. Molenaar et al., “Aminoacid analysis using chromatography-mass spectrometry: aninter platform comparison study,” Journal of Pharmaceuticaland Biomedical Analysis, vol. 114, pp. 398–407, 2015.

[11] N. Lepage, D. Li, P. A. Kavsak et al., “Incomplete pediatricreference intervals for the management of patients with in-born errors of metabolism,” Clinical Biochemistry, vol. 39,no. 6, pp. 595–599, 2006.

[12] D. Fekkes, “State-of-the-art of high-performance liquidchromatographic analysis of amino acids in physiologicalsamples,” Journal of Chromatography B: Biomedical Sciencesand Applications, vol. 682, no. 1, pp. 3–22, 1996.

[13] A. Pandor, J. Eastham, C. Beverley, J. Chilcott, and S. Paisley,“Clinical effectiveness and cost-effectiveness of neonatalscreening for inborn errors of metabolism using tandemmassspectrometry: a systematic review,” Health Technology As-sessment, vol. 8, no. 12, pp. 1–121, 2004.

[14] MassChrom® Amino Acids and Acylcarnitines from DriedBlood-LC-MS/MS, 2017, https://www.chromsystems.com/products/newborn-screening/masschrom-amino-acids-and-acylcarnitines-from-dried-blood-lc-ms-ms-55000.html.

[15] R. Fingerhut, M. L. Silva Polanco, J. Silva Arevalo Gde, andM. A. Swiderska, “First experience with a fully automatedextraction system for simultaneous on-line direct tandemmass spectrometric analysis of amino acids and (acyl-)car-nitines in a newborn screening setting,” Rapid Communica-tions in Mass Spectrometry, vol. 28, no. 8, pp. 965–973, 2014.

[16] P. M. Jones and M. J. Bennett, “-e changing face of newbornscreening: diagnosis of inborn errors of metabolism by tan-dem mass spectrometry,” Clinica Chimica Acta, vol. 324,no. 1-2, pp. 121–128, 2002.

[17] D. M. Gregory, D. Sovetts, C. L. Clow, and C. R. Scriver,“Plasma free amino acid values in normal children and ad-olescents,” Metabolism, vol. 35, no. 10, pp. 967–969, 1986.

[18] B. Jung and K. Adeli, “Clinical laboratory reference intervalsin pediatrics: the CALIPER initiative,” Clinical Biochemistry,vol. 42, no. 16-17, pp. 1589–1595, 2009.

[19] D. J. Dietzen, M. J. Bennett, S. F. Lo, V. L. Grey, andP. M. Jones, “Dried blood spot reference intervals for steroidsand amino acids in a neonatal cohort of the national children’sstudy,” Clinical Chemistry, vol. 62, no. 12, pp. 1658–1667, 2016.

[20] N. Lepage, N. McDonald, L. Dallaire, and M. Lambert, “Age-specific distribution of plasma amino acid concentrations in

a healthy pediatric population,” Clinical Chemistry, vol. 43,no. 12, pp. 2397–2402, 1997.

[21] F. Hammarqvist, G. Angsten, S. Meurling, K. Andersson, andJ. Wernerman, “Age-related changes of muscle and plasmaamino acids in healthy children,” Amino Acids, vol. 39, no. 2,pp. 359–366, 2010.

[22] E. Haschke-Becher, A. Kainz, and C. Bachmann, “Referencevalues of amino acids and of common clinical chemistry inplasma of healthy infants aged 1 and 4 months,” Journal ofInherited Metabolic Disease, vol. 39, no. 1, pp. 25–37, 2016.

[23] O. O. Oladipo, A. L. Weindel, A. N. Saunders, andD. J. Dietzen, “Impact of premature birth and critical illnesson neonatal range of plasma amino acid concentrations de-termined by LC-MS/MS,” Molecular Genetics and Meta-bolism, vol. 104, no. 4, pp. 476–479, 2011.

[24] J. Svasti, P. Wasant, M. Tiensuwan et al., “Normal plasma freeamino acid levels in -ai children,” Journal of the MedicalAssociation of �ailand, vol. 84, no. 11, pp. 1558–1568, 2001.

[25] E. Grossi, R. Colombo, S. Cavuto, and C. Franzini, “-eREALAB project: a new method for the formulation of ref-erence intervals based on current data,” Clinical Chemistry,vol. 51, no. 7, pp. 1232–1240, 2005.

[26] E. Macit, M. Kizilgun, E. Cakir et al., “Pediatric referenceintervals for plasma and urine essential amino acids ina Turkish population,” Turkish Journal of Medical Sciences,vol. 44, no. 2, pp. 323–329, 2014.

[27] J. A. Mendez, A. L. Fernandez-Sanmamed, M. S. Gomez-Holgado, and F. Fernandez-Rodriguez, “Age-related referencevalues for plasma amino acids in a Spanish populationmeasured by gas chromatography-mass spectrometry,”Journal of Pediatric Endocrinology and Metabolism, vol. 26,no. 3-4, pp. 333–341, 2013.

[28] P. Yi, L. Liu, H. Mei, F. Zeng, Z. Huang, and H. Niu, “Es-tablishment of reference range of plasma amino acids foryounger Chinese children by reverse phase HPLC,” Journal ofPediatric Endocrinology and Metabolism, vol. 24, no. 9-10,pp. 733–738, 2011.

[29] H. Yamamoto, K. Kondo, T. Tanaka et al., “Reference intervalsfor plasma-free amino acid in a Japanese population,” Annalsof Clinical Biochemistry, vol. 53, no. 3, pp. 357–364, 2016.

[30] �ai National Health Examination Survey (2008-2009), 2011,http://www.hiso.or.th/hiso5/report/report6.php.

[31] G. L. Horowitz and S. Altaie, Defining, Establishing andVerifying Reference Intervals in the Clinical Laboratory; Ap-proved Guideline-�ird Edition: CLSI Document EP28–A3c,Clinical Laboratory Standards Institute, Wayne, PA, USA,2010.

[32] Influential Points, Coefficient of variation-Principle, http://influentialpoints.com/Training/coefficient_of_variation-principles-properties-assumptions.htm.

[33] R. A. Sacher, R. A. McPherson, J. M. Campos, andF. K. Widmann, Widmann’s Clinical Interpretation of Labo-ratory Tests, F. A. Davis, Philadelphia, PA, USA, 2000.

[34] A. M. Lopez, L. G. Noriega, M. Diaz, N. Torres, andA. R. Tovar, “Plasma branched-chain and aromatic aminoacid concentration after ingestion of an urban or rural diet inrural Mexican women,” BMC Obesity, vol. 2, no. 1, p. 8, 2015.

[35] P. Panuwet, R. E. Hunter Jr., P. E. D’Souza et al., “Biologicalmatrix effects in quantitative tandemmass spectrometry-basedanalytical methods: advancing biomonitoring,” Critical Re-views in Analytical Chemistry, vol. 46, no. 2, pp. 93–105, 2016.

[36] Siriraj Hospital FoM, Manual of Medical LaboratoryInvestigations-Comprehensive Metabolic Test, 2010, http://

Journal of Nutrition and Metabolism 9

www.si.mahidol.ac.th/th/manual/Document/a04/Comprehensive%20Metabolic%20test.htm.

[37] T. L. Perry and S. Hansen, “Technical pitfalls leading to errorsin the quantitation of plasma amino acids,” Clinica ChimicaActa, vol. 25, no. 1, pp. 53–58, 1969.

[38] F. F. Kirchberg, U. Harder, M. Weber et al., “Dietary proteinintake affects amino acid and acylcarnitine metabolism ininfants aged 6 months,” Journal of Clinical Endocrinology &Metabolism, vol. 100, no. 1, pp. 149–158, 2015.

[39] M. Breier, S. Wahl, C. Prehn et al., “Targeted metabolomicsidentifies reliable and stable metabolites in human serum andplasma samples,” PLoS One, vol. 9, no. 2, article e89728, 2014.

[40] S. Trabado, A. Al-Salameh, V. Croixmarie et al., “-e humanplasma-metabolome: reference values in 800 French healthyvolunteers; impact of cholesterol, gender and age,” PLoS One,vol. 12, no. 3, article e0173615, 2017.

[41] B. H. Walsh, D. I. Broadhurst, R. Mandal et al., “-emetabolomic profile of umbilical cord blood in neonatalhypoxic ischaemic encephalopathy,” PLoS One, vol. 7, no. 12,article e50520, 2012.

[42] R. Fingerhut, W. Roschinger, and M. Heck, “A rapid andsensitive UPLC-MS/MS-method for the separation andquantification of branched-chain amino acids from driedblood samples of patients with maple syrup urine disease(MSUD),” International Journal of Neonatal Screening, vol. 2,no. 2, p. 2, 2016.

10 Journal of Nutrition and Metabolism

Clinical StudyComparison of Glycomacropeptide with PhenylalanineFree-Synthetic Amino Acids in Test Meals to PKU Patients:No Significant Differences in Biomarkers, Including PlasmaPhe Levels

Kirsten K. Ahring ,1,2,3 Allan M. Lund,2,3 Erik Jensen,4 Thomas G. Jensen,5

Karen Brøndum-Nielsen,1 Michael Pedersen,6 Allan Bardow,7

Jens Juul Holst ,8 Jens F. Rehfeld,9 and Lisbeth B. Møller2

1 e PKU Clinic, Kennedy Centre, Centre for Paediatric and Adolescent Medicine, Copenhagen University Hospital, Rigshospitalet,Copenhagen, Denmark2Department of Clinical Genetics, Applied Human Molecular Genetics, Kennedy Center, Rigshospitalet, Denmark3Centre for Inherited Metabolic Diseases, Centre for Paediatric and Adolescent Medicine, Copenhagen University Hospital,Rigshospitalet, Copenhagen, Denmark4Arla Foods Ingredients Group P/S, Viby J, Denmark5Department of Biomedicine, Aarhus University, Aarhus, Denmark6Department of Clinical Medicine, Aarhus University Hospital, Aarhus, Denmark7Department of Odontology, Copenhagen University, Copenhagen, Denmark8Institute of Clinical Medicine, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark9Department of Clinical Biochemistry, University of Copenhagen, Rigshospitalet, Copenhagen, Denmark

Correspondence should be addressed to Kirsten K. Ahring; kka@kennedy.dk

Received 14 July 2017; Accepted 11 October 2017; Published 8 January 2018

Academic Editor: Ina Knerr

Copyright © 2018 Kirsten K. Ahring et al.(is is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Introduction. Management of phenylketonuria (PKU) is achieved through low-phenylalanine (Phe) diet, supplemented with low-protein food andmixture of free-synthetic (FS) amino acid (AA). Casein glycomacropeptide (CGMP) is a natural peptide releasedin whey during cheese-making and does not contain Phe. Lacprodan® CGMP-20 used in this study contained a small amountof Phe due to minor presence of other proteins/peptides. Objective. (e purpose of this study was to compare absorption ofCGMP-20 to FSAA with the aim of evaluating short-term e9ects on plasma AAs as well as biomarkers related to food intake.Methods. (is study included 8 patients, who had four visits and tested four drink mixtures (DM1–4), consisting of CGMP, FSAA,or a combination. Plasma blood samples were collected at baseline, 15, 30, 60, 120, and 240 minutes (min) after the meal. AApro?les and ghrelin were determined 6 times, while surrogate biomarkers were determined at baseline and 240min. A visualanalogue scale (VAS) was used for evaluation of taste and satiety. Results. (e surrogate biomarker concentrations and VASscores for satiety and taste were nonsigni?cant between the four DMs, and there were only few signi?cant results for AA pro?les(not Phe). Conclusion. CGMP and FSAA had the overall same nonsigni?cant short-term e9ect on biomarkers, including Phe.(is combination of FSAA and CGMP is a suitable supplement for PKU patients.

1. Introduction

Phenylketonuria (PKU) is an inborn error of metabolism. Ifleft untreated, severe brain damage will occur [1–3]. (eprimary aim of treatment of those su9ering from PKU is to

control the blood phenylalanine (Phe) concentration inorder to prevent neurological damage [2]. PKU treatment isbased on a low-protein (LP) diet in combination with free-synthetic (FS) amino acid (AA) supplements without Pheand enriched with vitamins, minerals, trace elements, and in

HindawiJournal of Nutrition and MetabolismVolume 2018, Article ID 6352919, 11 pageshttps://doi.org/10.1155/2018/6352919

some products also fat and carbohydrates [1, 4, 5]. DietaryAA supplements are administered to obtain optimal met-abolic control by ensuring adequate levels of essentialAAs. Diet for life is recommended [6]. Compliance becomesa challenge over time, especially in adolescence, and theyare often related to disagreeable taste and the current lim-itations in available dietary products [7, 8]. In order toachieve better compliance, the formulation of the AA sup-plement should satisfy the need for better taste and easiermanagement [9]. (erefore, alternatives to conventionaltreatment are investigated.

Casein glycomacropeptide (CGMP) is a 64-amino acidpeptide from cheese whey, which is rich in speci?c essentialAAs and is the only known natural protein free from Phe[10–14]. Hence, CGMP is an alternative source of proteinfor PKU patients. In this study, we tested the product Lac-prodan CGMP-20, a highly puri?ed CGMP product withminimum 95% CGMP and a low level of Phe (0.16 g/100 gAA) and above 78% protein. (e residual amount of Phe isdue to the presence of minor amounts of other proteins/peptides. However, to ensure the supplement is nutritionallyadequate, it requires supplementation of the following AAsto meet the standards of similar PKU supplements: tyrosine(Tyr), tryptophan (Trp), arginine (Arg), histidine (His), leucine(Leu), lysine (Lys), and methionine (Met) [15–17]. LacprodanCGMP-20 will be referred to as CGMP in this paper.

In this single-blinded, prospective, crossover interven-tion study, we investigated the utilization and metabolicshort-term e9ect on absorption of pure CGMP-20, FSAA,and a combination of both, all consumed with a standardizedmeal, by comparing selected relevant surrogate biomarkers.We also evaluated taste and satiety. A second aim was toinvestigate whether the small amount of Phe in CGMP-20a9ected the plasma Phe concentration signi?cantly.

2. Materials and Methods

2.1. Protocol. (e study was approved by the NationalCommittee on Health Research Ethics prior to the study (IDH-3-2014-115), and all participants gave written consentprior to start of the study.

2.2. Patient Recruitment. Patients were contacted by mailand invited to participate. (ey were recruited from theclinical PKU database, including all PKU patients diagnosedand living in Denmark. Patients received compensation forlost wages and travel expenses. Participants in the projectwere invited to participate if theymet the following inclusioncriteria: diagnosis of classical PKU con?rmed by mutationanalysis and a known phe tolerance (12–20mg/kg) [18–20],age≥ 15 years at inclusion, had received treatment witha protein-restricted diet since the neonatal period, and werewilling and able to visit the PKU clinic four times.(e periodbetween visits varied from 24 hours to 1 month, since weestimated this to be suLcient time for a wash-out period [21,22]. Exclusion criteria were (1) <15 years at inclusion, (2) hadnot followed the dietary treatment continuously, (3) hada second chronic disease or condition, which potentially

could inMuence the PKU treatment and outcome, (4) treatedwith BH4, or (5) pregnant, nursing, or planning to becomepregnant. Eight patients accepted to participate.

2.3. Study Design. (e primary purpose of this study was tocompare the absorption rate and absorbed amount ofpeptide-bound AAs (CGMP-20∗) (both in its pure form orsupplemented with selected FSAAs) with an almost identicalmixture of FSAAs with the aim of evaluating the short-terme9ect on Phe and other AAs as well as biomarkers relatedto food intake. All patients had four visits in the PKUclinic. (e patients consumed a di9erent drink mixture(DM1, DM2, DM3, and DM4) in random order at each visit(blinded to the participants). (e order of the DM wasrandomized by a doctor at the PKU clinic, who was nototherwise involved in the study. CGMP/AA supplements:four di9erent AA sources were tested. DM1: LacprodanCGMP-20. DM2: FSAA (equivalent AA pro?le as DM1).DM3: Lacprodan CGMP-20 and synthetic AA. DM4: FSAA(equivalent AA pro?le as DM3 but without Phe). DM2 hadthe same AA pro?le as DM1 consisting of pure CGMP inorder to evaluate this; DM3 consisted of CGMP supple-mented with FSAA to make it nutritionally adequate andsuitable for patients with PKU and had a similar AA pro?leas DM4 (though this was without the 0.16 g Phe/100 g AApresent in CGMP). (e crossover study design made itpossible to evaluate the e9ect of the Phe content in GMP.Patients arrived fasting to the clinic in the morning and thensubjected to the ?rst venous blood samples (4ml) at time 0.Subsequently, blood samples were drawn at 15, 30, 60, 120,and 240 minutes (min) after ?nishing the meal. Taste wasevaluated right after consumption. At the end of the visit, allparticipants were asked to evaluate satiety.

(e test meals consisted of a few selected food items(homemade LP bread, butter, and jam) with individuallycalculated amounts of fat and carbohydrates and onlya minimal content of protein and Phe from LP bread. (etest meal was consumed in combination with DM, andthe individual intake was designed to cover 25% of thedaily requirement. In eachmeal, the total content of proteinwas equivalent to 25% of 1 g per kg body weight per day(1 g/kg/d). (e composition of the meal was calculatedafter the Nordic Nutrition Recommendations (NNA) andmet the criteria for sex, age, and weight: 10–15% fromprotein, 30–35% from fat, and 50–60% from carbohydrate.After completing the trial, all participants would continuetheir usual diet and AA supplements. An example of a mealis shown in Table 1, and the content of DM1–4 is given inTable 2.

2.4. Blood Samples. All participants had the following bloodsamples drawn at start (time 0) before consuming themeal/DM and at the end of the study (240min): glucose,insulin, glucagon-like peptide-1(GLP-1), blood urea nitro-gen (BUN), peptide tyrosine-tyrosine (PYY), cholecystoki-nin (CCK), ghrelin, and AA pro?les. Furthermore, ghrelinand AA pro?les were (besides at time 0) also measured at15, 30, 60, 120, and 240min.

2 Journal of Nutrition and Metabolism

2.4.1. BUN, Glucose, Insulin, Ghrelin, GLP-1, PYY, andCCK. 2–4ml of blood was collected for each biomarker andhandled immediately according to protocols [23–28], placedon ice, spinned at 2500 g for 10min at 4°C, and transferredto speci?c glasses for further analyses. EDTA glasses for(1) GLP-1 were added 200µl DPP-IV inhibitor and for(2) ghrelin were added 100µl Pefabloc by using a 0.5ml syringe(e samples for the total plasma AA pro?le were frozen at−80°C and sent on dry ice for quantitative analysis of AA usingthe stable-isotope dilution technique and HPLC-MS/MS [29].

2.5. Visual Analogue Scale (VAS). (is scale (http://www.vastranslator.com) was presented to patients as a horizontalline, ranking from “very hungry” (“0”) to “very satis?ed”(“100”) and from “bad taste” (“0”) to “good taste” (“100”) asan application (APP) on iPad. (e patients were asked toevaluate the taste of the DM shortly after intake and after240min to determine the level of satiety.

2.6. Body Mass Index (BMI) and Fat Percentage. (ese weremeasured at visit 1 with a body fat monitor (BF306, OmronHealthcare, USA).

2.7. Diet Registration. (e patients had ?lled out a dietaryrecord covering the 24 hours before each visit. (ey wereinstructed to eat similar food items every time to ensure thatthe Phe level would be within the same range the day after.(e registrations were calculated with Dankost (http://dankost.dk/english).

2.8. Statistical Methods. All calculations were performedusing the software SPSS 22 or Microsoft Excel 2010 forwindows. Paired and unpaired t-tests and one- and two-wayANOVA performed at a signi?cance level of α� 0.05 wereused. (e statistical signi?cance level was at p< 0.05.

3. Results

3.1. Clinical Protocol. Eight patients (seven females, onemale), age 15–48 years (mean 33.25± SD 11.21), weight47–85 kg (mean 72.8 + SD 15.9), and height 162–179 cm(mean 170.1 + SD 5.4) were included in the study. Data foreach patient are presented in Table 3. Six out of eight (75%)patients completed the study. Patient ID#4 had diLcultieseating the meal within 15min at visit one, which subsequentlydelayed blood sampling up to 10min for each blood sample.Patient ID#4 completed day 1 and day 2 as planned, but onthe third day, blood sampling was impossible, and for thatreason, visit 4 was cancelled and the patient excluded. Patient

ID#5 completed visit 1 and visit 2 but was unable to completeday 3 and day 4 due to health problems.

3.2. Diet Registration. (e 24-hour diet record con?rmedsimilar intake for each patient prior to the four test days.

3.3. Test Meals and DM. Intake of protein from DM sup-plement was 25% of the daily recommendation of 1 gprotein/kg/day (mean volume 151.8 g (range 97.9–195.8)),and the test meal provided fat and carbohydrates (Table 4).All patients complied with the intake.

3.4. AA Pro?les. AA pro?les were compared in subgroups(DM1 versusDM2 andDM3 versusDM4), since the respectivegroups were identical concerning the AA composition. Sta-tistical results were calculated between these subgroups.

(e area under the curve (AUC) (adjusted for baseline)for total AA demonstrated insigni?cant di9erences betweenDM1 and DM2 (p � 0.852) as well as between DM3 andDM4 (p � 0.06).

We did ?nd signi?cant di9erences for the followingindividual AA for AUC: DM1 and DM2: Lys (p � 0.0287),Asn (p � 0.0210), and Asp (p � 0.0047) and DM3 andDM4: citrulline (p � 0.0162). Results for all the AUC foreach AA are presented in Table 5. (e highest value for theAUC for the individual AAwas Ala, Val, Ile, and Asp (DM1),Pro (DM2), and Leu (DM3).

3.5. Plasma Concentrations of AA. (e peak plasma con-centrations for AAs (given as mean values in percentage ofthe premeal level) were as follows: DM1: the peak serumconcentrations were reached after 30min for 19 of 21(90%)AAs. Glu reached a peak after 15min; Tyr only decreasedcompared to baseline. DM2: four AAs peaked after 30min,while 15 (71%) AAs peaked after 15 min. Citrullinepeaked after 60 min; Tyr decreased compared to time 0.DM3: majority of the AAs peaked after 30 min (67%)except for Asp, Met, and Gln, where the peaks werereached after 15 min. Phe, Glu, citrulline, and Gly alldecreased compared to baseline. DM4: ?fteen AAs (71%)peaked after 15 min, only ?ve after 30 min, while cit-rulline peaked after 60 min. All results are displayed inTable 6.

Since most of the AAs peaked after either 15 or 30min,comparison of the concentrations for each DM after 15minversus 30min was performed to test for signi?cant dif-ferences between these time points. (e following AAsshowed signi?cant changes between time 15 and time 30:DM1: Ala (0.0474), Pro (0.0174), Val (0.0299), and Ile(0.0294), DM2: Leu (0.0295), and DM4: Asp (0.0423). (erewere no signi?cant changes for any AA in DM3. Moredetails are provided in Supplemental Table A.

(e paired t-test was used for comparison of the con-centrations found in DM1 and DM2 and in DM3 and DM4,respectively, at each time point, to test for signi?cant dif-ferences between the four DMs. We found no signi?cantdi9erences at time 0. Plasma concentrations (µmol/l) after

Table 1: Example of a test breakfast meal including DM (calculatedindividually for each participant).DM (CGMP-20, AA, or CGMP+AA)2 slices (80 g) of LP breadButter (20 g)Jam (20 g)

Journal of Nutrition and Metabolism 3

Table 3: Patient data, all with classic PKU: age, mutations, height, weight, BMI, and usual AA supplement.

Patient ID# Age Mutation 1 Mutation 2 Phenotype Height (cm) Weight (kg) BMI Usual AA product1 48 c.1315+1G>A1 c.1315+1G>A1 Classic PKU 169 85 30 PreKUnil tablets2 27 c.1315+1G>A1 c.1222C>T2 Classic PKU 174 62 20 XPhe energy3 18 c.842C>T3 c.1315+1G>A1 Classic PKU 171 75 26 Avonil powder4 16 c.1222C>T2 c.1315+1G>A1 Classic PKU 171 47 16 Avonil tablets5 38 c.1315+1G>A1 c.1222C>T2 Classic PKU 179 79 25 PreKUnil tablets6 46 c.814G>T4 c.1222C>T2 Classic PKU 173 94 31 PreKUnil tablets7 34 c.473G>A5 c.1315+1G>A1 Classic PKU 162 53 20 PreKUnil tablets8 39 c.1222C>T2 c.1222C>T2 Classic PKU 162 87.5 33 Avonil tablets1(IVS12+1G>A); 2p.R408W; 3p.E280K; 4p.G272X; 5p.R158Q.

Table 2: Content of pure CGMP (g AA/100 g p) and DM1–4 (pr. 1000 g mixture) (CGMP and free-synthetic AA). DM1: 100% CGMP:158.46 g� the content of each AA displayed below. DM2: 100% FSAA (the total amount of AA is shown below). DM3: 119.04 g from CGMP(+AA� the additional amount of AA from FSAA). DM4: 100% FSAA.

CGMP-20 AA DM1: LacprodanCGMP-20: 158.46 g∗

DM2:FSAA

DM3: LacprodanCGMP-20: 119.04 g∗ +FSAA

DM4:FSAA

g AA/100 g protein Total amount of AA (g)6.4 Ala 8.57 8.40 6.44 6.310.3 Arg 0.44 0.43 4.85∗ 4.859.2 Asp 12.08 11.83 9.07 8.890.08 Cys 0.15 0.15 0.11 0.1121.1 Glu 26.53 25.99 19.93 19.521.2 Gly 1.53 1.50 1.15 1.130.2 His 0.16 0.16 3.0∗ 3.0411.5 Ile 14.50 14.21 10.89 10.672.5 Leu 3.12 3.06 12.11∗ 12.066.4 Lys 8.46 8.29 7.37∗ 7.242 Met 2.92 2.85 3.62∗ 3.570.2 Phe 0.20 0.20 0.15 012.6 Pro 16.45 16.11 12.36 12.108.5 Ser 10.46 10.25 7.86 7.7018.1 (r 23.62 23.14 17.74 17.380.04 Trp 0.00 — 2.44∗ 2.440.06 Tyr 0.05 0.05 10.81∗ 10.819.5 Val 11.76 11.52 8.83 8.65— Citric acid powder — — 14.40 1.40— Citric acid solution (50% w/w in water) 11.20 — — —— Tropical twist Mavour (IFF SC401962) 1.75 1.75 1.75 1.80— Sucrose 73.00 73.00 80.00 80.00— Water 755.59 787.12 751.95 780.17109.88 Total 1000.00 1000.00 1000.00 1000.00— DM 1 2 3 481 Protein equivalent (g/100 g) 12.00 12.00 12.00 12.00— Carbohydrate (g/100 g) 7.46 7.30 8.12 8.00— Fat (g/100 g) 0.03 0 0.02 0— Energy (kcal/100 g) 78 77 81 80∗Part of the AA content comes from CGMP and part comes from additional FSAA: Arg: 0.33 (CGMP) + 4.52 (AA)� 4.85, His 0.12 (CGMP) + 2.92 (AA)�3.04, Leu: 2.35 (CGMP) + 9.76 (AA)� 12.11, Lys: 6.36 (CGMP) + 1.01 (AA)� 7.37, Met: 2.19 (CGMP) + 1.43 (AA)� 3.62, Trp: 0.00 (CGMP) + 2.44 (AA)� 2.44,Tyr: 0.04 (CGMP) + 10.77 (AA)� 10.81.

4 Journal of Nutrition and Metabolism

intake of the DM and test meal for all AAs presented as %.Comparison between DM1+2 and DM3+4 gave the fol-lowing results: time 15: DM1+2: Tyr: 0.0228, Asp: 0.0136, andcitrulline: 0.0378; DM3+4: Pro: 0.0455, Val: 0.0204, and(r:0.0239; time 30: DM1+2: Leu: 0.0178, Ile: 0.0021, Asn: 0.0038,and Asp: 0.0362; DM3+4: Ser: 0.0031, Pro: 0.0486, (r:0.0027, His: 0.0108, Tyr: 0.0412, and citrulline: 0.0409; time60: DM1+2: Asp: 0.0060 and citrulline: 0.0100; DM3+4:Pro: 0.0003, Val: 0.0088, (r: 0.0270, and Asp: 0.0155; time

120: DM3+4: citrulline:0.0364; and time 240: DM1+2: Asp:0.0288. (ere were no signi?cant di9erences at time 0.

Compared to baseline values, ghrelin values demon-strated a signi?cant decrease at 30, 60, and 120min for DM1,at 30min for DM2, at 60 and 120min for DM3, and at 30, 60,and 120min for DM4, respectively. Final levels decreased 7%for DM1, increased 10% for DM2 and 5% for DM3, andremained unchanged in DM4, all compared to baseline.Table 7 displays results (% compared to baseline) and the

Table 4: Intake of DM (water +CGMP mixture powder� total volume) and standard meal in grams (g) and energy (kilojoule (kJ) andenergy % (E %)).

Patient ID 1 2 3 4 5 6 7 8 Mean SDPowder (g) 41 30 36 21 38 45 27 44 35 8Water (g) 136 99 120 77 126 151 83 139 116 25LP bread (g) 70 90 170 100 82 70 96 80 95 30Butter (g) 23 23 35 26 21 21 23 19 24 5Marmalade (g) 40 40 40 42 40 40 40 40 40 1Energy (kJ) 2324 2330 3445 2405 2324 2327 2316 2322 2474 368Protein (g) 22 16 20 13 20 24 14 23 19 4Protein E (%) 16 12 10 9 15 18 10 17 13 3Fat (g) 20 20 32 23 19 18 21 18 21 4Fat E (%) 32 33 34 35 30 29 33 28 32 2Carbohydrate (g) 70 75 111 78 74 71 76 74 79 12Carbohydrate E (%) 52 56 56 57 55 53 57 56 55 2

Table 5: Results for the AUC (µmol/l over time) for all AAs (adjusted for baseline (time 0)). Signi?cant di9erences for the following AAs forthe area under the curve (AUC) minus baseline: DM1 and 2: Lys (p � 0.0287), Asn (p � 0.0201), and Asp (p � 0.0046) and DM3 and 4:citrulline (p � 0.0162).

DM1 2 3 4

Mean SD Mean SD Mean SD Mean SDGly −1145 12734 11361 20714 −16590 15386 11709 9483Ala 38797 10277 29471 5292 8397 3693 20653 5239Ser 8021 2312 6262 680 −1264 2222 2180 1230Pro 28232 3802 33492 7623 9353 2444 15950 2204Val 46664 5818 33878 7511 14370 5722 24624 7244(r 38627 7843 35865 3144 13558 1338 20698 2662Leu −3102 499 −2979 786 6899 1798 5206 1564Ile 26638 2253 21551 1691 9666 1474 8429 1146Lys 16623 3356 5005 3255 3932 2076 7342 2542Asn 3036 688 910 409 1838 978 1129 652Met 1976 580 1848 230 842 429 3026 1961His −419 768 −1133 677 −1036 1151 1003 697Phe −6717 11135 −2778 6468 −24356 10869 −8944 9170Tyr −3376 592 −4000 649 6672 1012 6019 1622Glu 123 399 −305 541 −2302 1004 −988 782Gln 34247 7716 39614 10878 −4902 4414 9870 7528Asp 1399 413 −254 239 −682 490 −469 217Trp −1364 403 −1078 412 804 849 994 484Orn 1167 818 733 621 799 724 1288 304Arg −505 1292 −100 1136 −1121 1310 1047 941Cit −1115 634 790 647 −2193 549 305 652

Journal of Nutrition and Metabolism 5

Table 6: Plasma concentrations (µmol/l) after intake of the DM and test meal for all AAs presented as % compared to time 0 (�1).Comparison between DM1+2 and DM3+4 gave the following results: time 15: DM1+2: Tyr: 0.0228, Asp: 0.0136, citrulline: 0.0378; DM3+4:Pro: 0.0455, Val: 0.0204, (r: 0.0239; time 30: DM1+2: Leu: 0.0178, Ile: 0.0021, Asn: 0.0038, Asp: 0.0362; DM3+4: Ser: 0.0031, Pro: 0.0486,(r: 0.0027, His: 0.0108, Tyr: 0.0412, citrulline: 0.0409; time 60: DM1+2: Asp: 0.0060, citrulline: 0.0100; DM3+4: Pro: 0.0003, Val: 0.0088,(r:0.0270, Asp: 0.0155; time 120: DM3+4: citrulline: 0.0364; time 240: DM1+2: Asp: 0.0288. (ere were no signi?cant di9erences at time 0.

Time15 30 60 120 240

Mean SD Mean SD Mean SD Mean SD Mean SDDM1

Gly 0.961 0.339 1.062 0.616 0.944 0.500 1.011 0.355 0.963 0.717Ala 1.656 0.435 2.340 0.653 1.940 0.462 1.510 0.419 1.233 0.328Ser 1.669 0.434 2.189 0.615 1.535 0.434 1.024 0.172 1.194 0.440Pro 2.904 0.886 3.886 1.117 2.819 1.256 1.782 0.303 1.523 0.374Val 1.977 0.510 2.509 0.698 2.255 0.422 1.712 0.257 1.583 0.310(r 2.594 0.544 3.885 0.867 2.855 0.813 2.013 0.229 1.832 0.829Leu 1.365 0.152 1.425 0.155 0.995 0.094 0.582 0.162 0.828 0.143Ile 5.815 1.849 7.033 1.190 5.198 0.841 3.426 0.464 2.603 0.517Lys 2.195 0.794 2.258 0.500 1.828 0.690 1.468 0.653 1.263 0.734Asn 1.666 0.251 2.003 0.399 1.412 0.313 1.012 0.091 1.218 0.414Met 2.255 0.714 2.275 0.684 1.690 0.522 1.165 0.235 1.071 0.387His 1.000 0.116 1.153 0.245 1.034 0.231 0.842 0.118 1.060 0.339Phe 0.942 0.130 1.012 0.189 0.990 0.199 0.935 0.131 0.976 0.271Tyr 0.816 0.143 0.863 0.245 0.765 0.170 0.599 0.134 0.647 0.136Glu 1.302 0.307 1.243 0.149 1.061 0.180 0.915 0.127 0.961 0.173Gln 1.278 0.204 1.309 0.296 1.291 0.277 1.245 0.208 1.242 0.402Asp 1.783 0.495 2.225 0.760 1.592 0.327 1.076 0.245 1.104 0.316Trp 0.980 0.105 1.116 0.188 0.906 0.155 0.725 0.145 0.808 0.256Orn 1.288 0.340 1.520 0.501 1.287 0.366 0.989 0.175 1.006 0.325Arg 1.041 0.135 1.220 0.272 1.024 0.198 0.876 0.193 0.963 0.242Cit 0.889 0.251 1.075 0.316 0.809 0.228 0.744 0.244 0.939 0.467

DM2Gly 1.238 0.926 1.197 0.745 1.081 0.938 1.227 0.796 0.979 0.654Ala 1.603 0.354 1.664 0.483 1.620 0.447 1.312 0.335 1.163 0.379Ser 1.653 0.297 1.697 0.546 1.457 0.204 1.102 0.213 1.083 0.169Pro 3.177 0.867 3.052 1.017 2.846 1.185 2.291 1.110 1.617 0.435Val 1.647 0.579 1.856 0.730 1.719 0.397 1.544 0.560 1.420 0.497(r 2.555 0.512 3.182 0.977 2.808 0.691 2.254 0.735 1.957 0.447Leu 1.389 0.255 1.225 0.196 0.983 0.094 0.635 0.129 0.810 0.160Ile 4.892 2.668 4.640 2.055 4.192 1.679 3.112 1.420 2.572 0.857Lys 1.714 0.725 1.420 0.530 1.280 0.341 1.046 0.218 1.023 0.438Asn 1.518 0.172 1.314 0.359 1.155 0.124 0.954 0.197 1.046 0.270Met 1.992 0.542 1.830 0.641 1.570 0.328 1.220 0.173 1.085 0.184His 1.114 0.208 1.045 0.308 0.993 0.175 0.823 0.178 0.944 0.171Phe 1.055 0.227 0.952 0.195 0.956 0.109 0.964 0.159 1.036 0.130Tyr 0.951 0.231 0.812 0.187 0.692 0.132 0.586 0.145 0.606 0.125Glu 1.130 0.182 1.087 0.177 1.016 0.165 0.897 0.222 0.964 0.202Gln 1.537 0.427 1.508 0.350 1.478 0.531 1.199 0.393 1.271 0.390Asp 1.200 0.426 1.132 0.266 0.880 0.155 0.814 0.230 1.005 0.190Trp 1.075 0.304 0.996 0.361 0.871 0.263 0.768 0.171 0.884 0.171Orn 1.227 0.290 1.081 0.287 1.067 0.211 1.084 0.302 1.079 0.275Arg 1.207 0.286 1.111 0.395 1.084 0.215 0.927 0.215 0.931 0.258Cit 1.118 0.525 1.136 0.404 1.211 0.349 1.121 0.359 1.030 0.359

6 Journal of Nutrition and Metabolism

t-test for all levels and values (mean and SD) in Supple-mental Table B.

None of the biomarkers GLP-1, PYY, BUN, CCK, in-sulin, and glucose showed signi?cant changes from baseline

(premeal) to the end of the study period (240min after mealand DM). Results for biomarkers are presented in Figure 1(% compared to baseline) and values (mean +/− SD) inSupplemental Table C.

Table 6: Continued.

Time15 30 60 120 240

Mean SD Mean SD Mean SD Mean SD Mean SDDM3

Gly 0.809 0.355 0.913 0.337 0.822 0.621 0.772 0.318 0.804 0.679Ala 1.206 0.304 1.435 0.302 1.171 0.192 1.085 0.277 0.987 0.141Ser 1.210 0.211 1.248 0.262 0.941 0.194 0.862 0.292 0.900 0.200Pro 1.777 0.313 1.838 0.349 1.471 0.234 1.278 0.415 1.126 0.242Val 1.459 0.274 1.715 0.330 1.228 0.117 1.314 0.572 1.078 0.143(r 1.625 0.368 1.851 0.471 1.635 0.217 1.438 0.463 1.325 0.293Leu 2.282 0.435 2.328 0.442 1.658 0.366 1.091 0.617 0.887 0.141Ile 3.524 1.089 3.675 0.890 2.498 0.721 1.736 1.234 1.380 0.386Lys 1.632 0.560 1.662 0.712 1.389 0.599 0.943 0.117 0.874 0.189Asn 1.630 0.329 1.534 0.429 1.353 0.335 1.034 0.430 0.980 0.202Met 1.826 0.370 1.754 0.507 1.255 0.116 1.070 0.374 0.865 0.180His 1.049 0.153 1.153 0.240 0.915 0.137 0.900 0.264 0.895 0.129Phe 0.889 0.111 0.911 0.104 0.843 0.100 0.879 0.140 0.910 0.119Tyr 1.645 0.244 1.985 0.713 1.723 0.437 1.667 0.295 1.305 0.248Glu 0.950 0.144 0.951 0.152 0.801 0.189 0.756 0.221 0.774 0.111Gln 1.170 0.104 1.061 0.223 1.036 0.120 0.911 0.148 0.920 0.093Asp 0.953 0.291 1.111 0.138 0.998 0.158 0.755 0.221 0.724 0.170Trp 1.244 0.272 1.455 0.302 1.165 0.382 1.071 0.444 0.950 0.383Orn 1.259 0.247 1.374 0.230 1.147 0.222 1.064 0.335 0.977 0.195Arg 1.216 0.226 1.236 0.331 0.982 0.138 0.869 0.213 0.835 0.163Cit 0.815 0.158 0.802 0.173 0.649 0.163 0.642 0.141 0.842 0.211

DM4Gly 1.179 0.781 1.304 0.630 0.920 0.445 1.230 0.433 1.183 0.690Ala 1.546 0.395 1.552 0.278 1.463 0.406 1.304 0.157 1.003 0.378Ser 1.521 0.336 1.419 0.228 1.196 0.382 0.978 0.120 0.961 0.347Pro 2.499 1.154 2.357 0.774 2.003 0.704 1.509 0.287 1.135 0.429Val 1.861 0.438 1.865 0.450 1.528 0.297 1.511 0.913 1.263 0.491(r 2.067 1.025 2.389 0.795 2.069 0.687 1.654 0.224 1.380 0.525Leu 2.237 0.667 1.981 0.433 1.579 0.482 0.995 0.277 0.829 0.279Ile 3.681 1.813 3.072 1.203 2.667 1.251 1.481 0.582 1.132 0.457Lys 1.671 0.782 1.385 0.292 1.429 0.704 1.184 0.389 1.179 0.589Asn 1.588 0.591 1.502 0.439 1.291 0.342 1.008 0.191 0.851 0.231Met 2.527 1.706 2.337 1.887 1.956 2.094 1.503 1.681 1.035 0.763His 1.237 0.231 1.261 0.135 1.140 0.270 1.004 0.153 0.981 0.342Phe 1.054 0.126 0.981 0.109 0.947 0.141 0.942 0.108 0.921 0.293Tyr 1.548 0.594 1.524 0.559 1.606 0.723 1.667 0.568 1.347 0.525Glu 1.123 0.205 1.013 0.191 0.950 0.221 0.823 0.168 0.893 0.407Gln 1.236 0.246 1.150 0.160 1.120 0.211 1.107 0.196 0.911 0.319Asp 1.121 0.261 0.964 0.149 0.854 0.234 0.819 0.177 0.860 0.343Trp 1.505 0.436 1.403 0.272 1.252 0.345 1.031 0.147 1.010 0.355Orn 1.470 0.405 1.376 0.200 1.337 0.302 1.105 0.153 0.992 0.314Arg 1.477 0.315 1.344 0.238 1.138 0.218 0.967 0.116 0.940 0.314Cit 1.132 0.285 1.105 0.216 1.201 0.603 0.989 0.231 0.971 0.413

Journal of Nutrition and Metabolism 7

3.6. VAS Score. (e question “how satis?ed are you?” ob-tained the following nonsigni?cant results: DM1 (mean 36,SD 18), DM2 (mean 41, SD 16), DM3 (mean 28, SD 27), andDM4 (mean 35, SD 30). (e results for the other question“how does the DM taste?” were as follows: DM1 (mean 46,SD 31), DM2 (mean 44, SD 22), DM3 (mean 36, SD 28), andDM4 (mean 26, SD 22). All comparisons (DM1 and DM2,DM3 and DM4, but also DM3 compared to DM1 and DM2,resp.) were statistically insigni?cant.

4. Discussion

(is study evaluated the metabolic short-term e9ect ofCGMP compared to an almost identical combinationof FSAA by repeated measurements. One of the mostimportant ?ndings was that the residual content of Phein DM3 did not a9ect the plasma level signi?cantlycompared to DM4 at any time which support data fromNey et al. [11].

Over the last 8 years, CGMP has been investigatedin mice studies and a few human studies to evaluatesafety, acceptability, and eLcacy of CGMP as a nutritional

supplement for treatment of PKU [11–13, 15, 30, 31]. (epresent study supports the conclusions of these studies.

(e slower absorption of most of the AA in DM1 andDM3, which contained CGMP, compared to DM2 and DM4,which contained only FSAA, could be explained by the factthat CGMP delay the absorption in the gut. (e fact that Tyrincreased in DM3 and DM4, but only decreased without peakin DM1 and DM2, is assumed to be caused by the low contentin DM1 and DM2. Tyr and Trp both peaked at 30min forDM3 compared to 15min for DM4, suggesting that thecontent of Tyr and Trp in the CGMP mixtures were me-tabolized less rapidly than the FSAA. (e Phe/Tyr ratio de-creased over time with 30% in both DM3 and DM4, while itincreased (caused by the low-Tyr concentration in DM1 andDM2) with 50% for DM1 and 70% for DM2. (e Phe/Tyrratio is an important measure because a high ratio can havea long-term negative e9ect on executive functions [32].

(e AUC for “total AA” was not associated with ab-sorption rate. We also calculated the AUC for each AA andcompared with peak values to determine complete ab-sorption and absorption rate, and we did ?nd signi?cantdi9erences for Lys, Asn, and Asp for DM1 and DM2 and forcitrulline for DM3 and DM4. None of the LNAA (extra-added in DM3 and matched in DM4) was signi?cantlydi9erent.

Ala, Pro, Val, and Ile demonstrated a signi?cant increasefrom 15 to 30min for DM1, while only Leu in DM2 and Aspin DM4 decreased signi?cantly, which indicate a betterabsorption of pure CGMP. (e fact that none of the AA inDM3 changed signi?cantly could be the inMuence of theextra-added AA (LNAA and Lys). In contrast, we did seesigni?cant di9erences between several AAs by comparisonbetween DM1 and DM2 and between DM3 and DM4 at eachtime point. Especially His, Tyr, and Trp are noteworthy forDM3 and DM4, since they are FSAAs added to the pureCGMP. Also, (r, Ile, and Val are of certain interest, sincethe natural content of these three AAs in CGMP is high [33].(e high content of these AAs and addition of extra-LNAAto the CGMP o9er an additional positive e9ect, since LNAAcompetes for transport across the blood-brain barrier (BBB)via the L-type amino acid transporter (LAT1) [34, 35]. HighPhe in plasma diminishes uptake of Tyr and Trp into thebrain, and this results in reduced formation of neuro-transmitters [36]. (is imbalance is possibly the major causeof disturbed brain development in PKU patients [37]. Each

Table 7: Ghrelin levels over time, presented as % relative to start value (time 0) + SD and p value.

DM1 DM2 DM3 DM4

Time Relative(%)

SD(%) p value Relative

(%)SD(%) p value Relative

(%)SD(%) p value Relative

(%)SD(%) p value

0 100 — — 100 — — 100 — — 100 — —15 95 12 0.266 88 11 0.056 90 13 0.171 75 22 0.05030 82 5 0.006∗ 83 15 0.035∗ 78 19 0.055 80 10 0.015∗

60 81 13 0.030∗ 84 16 0.068 80 12 0.013∗ 75 11 0.006∗

120 81 10 0.020∗ 84 15 0.088 75 14 0.011∗ 74 13 0.010∗

240 93 18 0.454 110 20 0.180 105 25 0.473 100 34 0.986∗Signi?cant.

0%

50%

100%

150%

200%

250%

GLP-1(pmol/L)

PYY 3-36(pg/ml)

Urea(mmol/L)

CCK(pmol/L)

Insulin(pmol/L)

Glucose(mmol/L)

Drink 1

Drink 2

Drink 3

Drink 4

Figure 1: Results (mean +/− SD) for the following biomarkers:glucose, insulin, GLP-1, PYY, BUN, and CCK. None of themdemonstrated signi?cant changes from baseline (premeal) to theend of the study period (240min after meal and DM).

8 Journal of Nutrition and Metabolism

LNAA has individual aLnity relation to LAT1 [38]. Severalstudies have shown the positive blocking e9ect of LNAAwhich reduces Phe entering the brain [39–41]. Matalonet al. [17] demonstrated a decrease in Phe in the blood up to50% using LNAA tablets and emphasized that LNAA ina speci?c mixture inhibits Phe uptake already in the gut[16, 42]. Administrations of Val, Ile, and Leu have provedto reduce Phe concentrations in the cerebrospinal Muid ofhumans [43].

All biomarkers remained unchanged by comparingtime 0 and 240min, and there were no signi?cant changesin plasma Phe despite the residual amount of Phe in CGMPin line with ?ndings from Ney et al. [11, 13]. (is studydemonstrated that AA in CGMP is absorbed as eLcient as anidentical mixture of FSAA.

All the DMs demonstrated a decreasing e9ect on ghrelinafter the meal. Ghrelin values showed signi?cant decreasesat 30, 60, and 120min for DM1, at 30min for DM2, at 60and 120min for DM3, and at 30, 60, and 120min for DM4.Low-ghrelin concentrations are associated with a feeling ofsatiety [44, 45]. By only evaluating DM1 and DM2, it couldindicate that satiety is reached faster for CGMP. However,VAS scores for satiety did not show any signi?cant di9erencebetween DM1 and DM2, nor between DM3 and DM4.

BUN was nonsigni?cantly lower for DM1 and DM3compared to DM2 and DM4, which potentially could sug-gest a more eLcient utilization of GMP compared to AA, asfound by van Calcar et al. [13]. Similar ?ndings have beenreported by Ney et al. [46]. However, our present short-termstudy was not able to support these ?ndings.

GLP-1 promotes insulin secretion and reduces appetiteand reached the highest (nonsigni?cant) level with DM3after 240min (118%) which may indicate that satiety wasbetter obtained with DM3 compared with DM1, DM2, andDM4. (is ?nding concurs with previous studies, showingthat GMP promotes satiety [14]. PYY also reduces appetiteand reached the highest value for DM3 (111%). CCK isa peptide hormone in the gastrointestinal system responsiblefor stimulating the digestion of fat and protein. (e widevariation from a decrease of 18% in DM3 to an increase of33% in DM4 was unexpected, since CCK is expected to riseafter a mixed meal [23].

A limitation of this study was the small number ofpatients. However, it was important for the study design toselect as homogeneous a test population as possible (onlyearly-treated classical PKU con?rmed by mutation analysis),resulting in exclusion of a number of patients.

(ree patients had a BMI over 30 and were de?ned asobese, one slightly obese, three had a normal BMI, and onehad a BMI below normal.(e patients were receiving dosageafter their actual weight, which means that the patients thatwere categorized with normal BMI received less DM per kglean body mass.

Although this study demonstrated nonsigni?cantchanges for almost all biochemical markers, it is importantto notice that this study only replaced a single meal anda long-term e9ect could be di9erent. Based on the currentresults, we consider CGMP to be a safe alternative to FSAAbut should be supplemented with additional FSAA to make

it nutritionally adequate and potential also with other nu-trients as fat, carbohydrates, vitamins, and minerals to createan easy-to-use supplement for patients with PKU. If CGMPproducts substitute FSAA completely, it is necessary tocarefully monitor if the small content of Phe will have animpact on the blood level in the long term. If Phe levelsincrease, restrictions of the LP diet must be implemented tobalance and control the Phe intake and blood level. Furtherstudies are needed to evaluate the long-term impact andeLcacy of CGMP in the management of PKU.

5. Conclusion

Dietary management of PKU should be lifelong, and goodcompliance is crucial for a good outcome. CGMP did notchange any of the biomarkers signi?cantly compared to free-synthetic AA when consumed with a test meal in PKUpatients.(e residual amount of Phe in CGMP did not a9ectthe plasma Phe level signi?cantly. Based on these data, weconsider CGMP to be a suitable alternative as supplementfor PKU treatment. However, further research is needed todetermine the long-term e9ects and safety of CGMP. (isstudy demonstrates that CGMP has the same short-terme9ect as FSAA.

Abbreviations

CGMP: Casein glycomacropeptide, CGMP-20 (productname Lacprodan CGMP-20)

PKU: PhenylketonuriaAA: Amino acid(s)FSAA: Free-synthetic AABBB: Blood-brain barrierBUN: Blood urea nitrogenGLP-1: Glucagon-like peptide-1CCK: CholecystokininPYY: Peptide tyrosine-tyrosineDM: Drink mixAla: AlanineArg: ArginineAsn: AsparagineAsp: AspartateCys: CysteineGlu: GlutamineGly: GlycineHis: HistidineIle: IsoleucineLeu: LeucineLys: LysineMet: MethioninePhe: PhenylalaninePro: ProlineSer: Serine(r: (reonineTrp: TryptophanTyr: TyrosineVal: ValinePAH: Phenylalanine hydroxylaseVAS: Visual analogue score

Journal of Nutrition and Metabolism 9

LP: Low proteinmin: MinutesSD: Standard deviationkJ: Kilojoule.

Disclosure

Kirsten Kiær Ahring is an industrial PhD student sponsoredby AFI and (e Danish Agency for Science, Technologyand Innovation.

Conflicts of Interest

(e authors declare that they have no conMicts of interest.

Authors’ Contributions

All authors have made a substantial contribution to thedesign and interpretation of data for the work, revised itcritically, and been presented for the ?nal version to bepublished. Erik Jensen, Kirsten K. Ahring, and Lisbeth B.Møller were involved in designing the study. Kirsten K.Ahring was involved in collecting all the data and writing themain manuscript.

Acknowledgments

(e authors would like to thank Arla Foods Ingredients and(e Danish Agency for Science, Technology and Innovationfor funding, the PKU patients for participating in the study,nurse Signe Larsen (PKU clinic, Kennedy Centre) for per-forming the blood sampling, and lab technicians SørenAndresen and Marianne Falck (University of Copenhagen)for performing the blood analyses and assisting in handlingof blood samples.

Supplementary Materials

Table A. Results for all AA at time 0, 15, 30, 60, 120, and 240(Mean and SD). Table B. Ghrelin results for time 0, 15, 30, 60,120, and 240 (Mean and SD). Table C. Results for glucose,insulin, GLP-1, PYY, BUN, and CCK (Mean ± SD). None ofthem demonstrated signi?cant changes from baseline (pre-meal) to end of study period (240min after meal and DM).(Supplementary Materials)

References

[1] J. J. Mitchell, Y. J. Trakadis, and C. R. Scriver, “Phenylalaninehydroxylase de?ciency,” Genetics in Medicine, vol. 13, no. 8,pp. 697–707, 2011.

[2] N. Blau, “Genetics of phenylketonuria: then and now,”Human Mutation, vol. 37, no. 6, pp. 508–515, 2016.

[3] M. Giovannini, E. Verduci, E. Salvatici, S. Paci, and E. Riva,“Phenylketonuria: nutritional advances and challenges,”Nutrition and Metabolism, vol. 9, no. 1, p. 7, 2012.

[4] H. L. Levy, “Nutritional therapy for selected inborn errorsof metabolism,” Journal of the American College of Nutrition,vol. 8, no. 1, pp. 54S–60S, 1989.

[5] A. MacDonald, J. C. Rocha, M. van Rijn, and F. Feillet,“Nutrition in phenylketonuria,” Molecular Genetics andMetabolism, vol. 104, pp. S10–S18, 2011.

[6] R. A. Williams, C. D. S. Mamotte, and J. R. Burnett, “Phe-nylketonuria: an inborn error of phenylalanine metabolism,”Clinical Biochemist Reviews, vol. 29, no. 1, pp. 31–41, 2008.

[7] A. MacDonald, H. Gokmen-Ozel, M. van Rijn, andP. Burgard, “(e reality of dietary compliance in the man-agement of phenylketonuria,” Journal of Inherited MetabolicDisease, vol. 33, no. 6, pp. 665–670, 2010.

[8] A. Aguiar, K. Ahring, M. F. Almeida et al., “Practices inprescribing protein substitutes for PKU in Europe: no uni-formity of approach,” Molecular Genetics and Metabolism,vol. 115, no. 1, pp. 17–22, 2015.

[9] F. J. van Spronsen and P. Burgard, “(e truth of treatingpatients with phenylketonuria after childhood: the need fora new guideline,” Journal of Inherited Metabolic Disease,vol. 31, no. 6, pp. 673–679, 2008.

[10] B. Chabance, P. Marteau, J. C. Rambaud et al., “Casein peptiderelease and passage to the blood in humans during digestionof milk or yogurt,” Biochimie, vol. 80, no. 2, pp. 155–165, 1998.

[11] D. M. Ney, S. T. Gleason, S. C. van Calcar et al., “Nutritionalmanagement of PKU with glycomacropeptide from cheesewhey,” Journal of Inherited Metabolic Disease, vol. 32, no. 1,pp. 32–39, 2009.

[12] P. Solverson, S. G. Murali, A. S. Brinkman et al., “Glyco-macropeptide, a low-phenylalanine protein isolated fromcheese whey, supports growth and attenuates metabolic stressin the murine model of phenylketonuria,” American Journalof Physiology: Endocrinology and Metabolism, vol. 302, no. 7,pp. E885–E895, 2012.

[13] S. C. van Calcar, E. L.MacLeod, S. T. Gleason et al., “Improvednutritional management of phenylketonuria by using a dietcontaining glycomacropeptide compared with amino acids,”American Journal of Clinical Nutrition, vol. 89, no. 4,pp. 1068–1077, 2009.

[14] S. C. van Calcar and D. M. Ney, “Food products made withglycomacropeptide, a low-phenylalanine whey protein,provide a new alternative to amino acid-based medical foodsfor nutrition management of phenylketonuria,” Journal of theAcademy of Nutrition and Dietetics, vol. 112, no. 8, pp. 1201–1210, 2012.

[15] C. E. Laclair, D. M. Ney, E. L. MacLeod, and M. R. Etzel,“Puri?cation and use of glycomacropeptide for nutritionalmanagement of phenylketonuria,” Journal of Food Science,vol. 74, no. 4, pp. E199–E206, 2009.

[16] R. Matalon, K. Michals-Matalon, G. Bhatia et al., “Doubleblind placebo control trial of large neutral amino acids intreatment of PKU: e9ect on blood phenylalanine,” Journalof Inherited Metabolic Disease, vol. 30, no. 2, pp. 153–158,2007.

[17] R. Matalon, K. Michals-Matalon, G. Bhatia et al., “Largeneutral amino acids in the treatment of phenylketonuria(PKU),” Journal of Inherited Metabolic Disease, vol. 29, no. 6,pp. 732–738, 2006.

[18] P. Guldberg and F. Guttler, “Mutations in the phenylalaninehydroxylase gene: methods for their characterization,” ActaPaediatrica, vol. 83, no. 407, pp. 49–56, 1994.

[19] F. Guttler and H. Lou, “Phenylketonuria and hyper-phenylalaninemia,” in Inborn Metabolic Diseases: Diagnosisand Treatment, J. Fernandes, J.-M. Saudubray, and K. Tada,Eds., pp. 161–174, Springer, Berlin, Heidelberg, Germany, 1990.

[20] F. Guttler, “Hyperphenylalaninemia: diagnosis and classi?-cation of the various types of phenylalanine hydroxylase

10 Journal of Nutrition and Metabolism

de?ciency in childhood,” Acta paediatrica Scandinavica.Supplement, vol. 280, pp. 1–80, 1980.

[21] G. N. (ompson and D. Halliday, “Signi?cant phenylalaninehydroxylation in vivo in patients with classical phenylke-tonuria,” Journal of Clinical Investigation, vol. 86, no. 1,pp. 317–322, 1990.

[22] G. N. (ompson, P. J. Pacy, R. W. Watts, and D. Halliday,“Protein metabolism in phenylketonuria and Lesch–Nyhansyndrome,” Pediatric Research, vol. 28, no. 3, pp. 240–246,1990.

[23] J. F. Rehfeld, “Accurate measurement of cholecystokinin inplasma,” Clinical Chemistry, vol. 44, no. 5, pp. 991–1001, 1998.

[24] ABX H., Diagnostic test for glucose, 2008.[25] Diagnostic Test for Blood Urea Nitrogen (BUN), 2014, http://

www.horiba.com.[26] M. A. Nauck, N. Kleine, C. Ørskov, J. J. Holst, B. Willms, and

W. Creutzfeldt, “Normalization of fasting hyperglycaemia byexogenous glucagon-like peptide 1 (7-36 amide) in type 2(non-insulin-dependent) diabetic patients,” Diabetologia,vol. 36, no. 8, pp. 741–744, 1993.

[27] E. Naslund, J. Bogefors, S. Skogar et al., “GLP-1 slows solidgastric emptying and inhibits insulin, glucagon, and PYYrelease in humans,” American Journal of Physiology: Regu-latory, Integrative and Comparative Physiology, vol. 277, no. 3,pp. R910–R916, 1999.

[28] S. Andreasen, Ghrelin Elisa Kit, 2015.[29] Special Assays: Monitoring of Amino Acid on Request, 2016,

http://www.zfs-reutlingen.de.[30] O. K. Zaki, L. El-Wakeel, Y. Ebeid et al., “(e use of glyco-

macropeptide in dietary management of phenylketonuria,”Journal of Nutrition and Metabolism, vol. 2016, 5 pages, 2016.

[31] A. Pinto, M. F. Almeida, P. C. Ramos et al., “Nutritional statusin patients with phenylketonuria using glycomacropeptide astheir major protein source,” European Journal of ClinicalNutrition, vol. 71, no. 10, pp. 1230–1234, 2017.

[32] M. Luciana, J. Sullivan, and C. A. Nelson, “Associationsbetween phenylalanine-to-tyrosine ratios and performance ontests of neuropsychological function in adolescents treatedearly and continuously for phenylketonuria,” Child Devel-opment, vol. 72, no. 6, pp. 1637–1652, 2001.

[33] Neelima, R. Sharma, Y. S. Rajput, and B. Mann, “Chemicaland functional properties of glycomacropeptide (GMP) andits role in the detection of cheese whey adulteration in milk:a review,” Dairy Science and Technology, vol. 93, no. 1,pp. 21–43, 2013.

[34] J. Pietz, A. Rupp, F. Ebinger et al., “Cerebral energy meta-bolism in phenylketonuria: ?ndings by quantitative in vivo31P MR spectroscopy,” Pediatric Research, vol. 53, no. 4,pp. 654–662, 2003.

[35] J. Pietz, R. Kreis, A. Rupp et al., “Large neutral amino acidsblock phenylalanine transport into brain tissue in patientswith phenylketonuria,” Journal of Clinical Investigation,vol. 103, no. 8, pp. 1169–1178, 1999.

[36] V. Leuzzi, M. Tosetti, D. Montanaro et al., “(e pathogenesisof the white matter abnormalities in phenylketonuria. Amultimodal 3.0 tesla MRI and magnetic resonance spec-troscopy (1H MRS) study,” Journal of Inherited MetabolicDisease, vol. 30, no. 2, pp. 209–216, 2007.

[37] F. J. van Spronsen, M. J. de Groot, M. Hoeksma,D. J. Reijngoud, and M. van Rijn, “Large neutral amino acidsin the treatment of PKU: from theory to practice,” Journal ofInherited Metabolic Disease, vol. 33, no. 6, pp. 671–676, 2010.

[38] G. M. Knudsen, S. Hasselbalch, P. B. Toft, E. Christensen,O. B. Paulson, and H. Lou, “Blood-brain barrier transport of

amino acids in healthy controls and in patients with phe-nylketonuria,” Journal of Inherited Metabolic Disease, vol. 18,no. 6, pp. 653–664, 1995.

[39] S. Schindeler, S. Ghosh-Jerath, S.(ompson et al., “(e e9ectsof large neutral amino acid supplements in PKU: an MRS andneuropsychological study,” Molecular Genetics and Meta-bolism, vol. 91, no. 1, pp. 48–54, 2007.

[40] S. Yano, K. Moseley, and C. Azen, “Large neutral amino acidsupplementation increases melatonin synthesis in phenyl-ketonuria: a new biomarker,” Journal of Pediatrics, vol. 162,no. 5, pp. 999–1003, 2013.

[41] M. Bik-Multanowski and J. J. Pietrzyk, “LAT1 gene var-iants—potential factors inMuencing the clinical course ofphenylketonuria,” Journal of Inherited Metabolic Disease,vol. 29, no. 5, p. 684, 2006.

[42] R. Koch, “Large neutral amino acid therapy and phenylke-tonuria: a promising approach to treatment,” MolecularGenetics and Metabolism, vol. 79, no. 2, pp. 110–113, 2003.

[43] M. K. Jordan, R. L. Brunner, M. M. Hunt, and H. K. Berry,“Preliminary support for the oral administration of valine,isoleucine and leucine for phenylketonuria,” DevelopmentalMedicine and Child Neurology, vol. 27, no. 1, pp. 33–39, 1985.

[44] E. L.MacLeod,M. K. Clayton, S. C. van Calcar, and D.M. Ney,“Breakfast with glycomacropeptide compared with aminoacids suppresses plasma ghrelin levels in individuals withphenylketonuria,” Molecular Genetics and Metabolism,vol. 100, no. 4, pp. 303–308, 2010.

[45] K. H. Schulpis, I. Papassotiriou, M. Vounatsou,G. A. Karikas, S. Tsakiris, and G. P. Chrousos, “Morningpreprandial plasma ghrelin and catecholamine concentra-tions in patients with phenylketonuria and normal controls:evidence for catecholamine-mediated ghrelin regulation,”Journal of Clinical Endocrinology and Metabolism, vol. 89,no. 8, pp. 3983–3987, 2004.

[46] D. M. Ney, B. M. Stroup, M. K. Clayton et al., “Glyco-macropeptide for nutritional management of phenylketon-uria: a randomized, controlled, crossover trial,” AmericanJournal of Clinical Nutrition, vol. 104, no. 2, pp. 334–345, 2016.

Journal of Nutrition and Metabolism 11

Clinical StudyMetabolomic Insights into the Nutritional Status of Adults andAdolescents with Phenylketonuria Consuming a Low-Phenylalanine Diet in Combination with Amino Acid andGlycomacropeptide Medical Foods

Bridget M. Stroup,1 Denise M. Ney,1 Sangita G. Murali,1 Frances Rohr,2 Sally T. Gleason,1

Sandra C. van Calcar,3 and Harvey L. Levy2

1Department of Nutritional Sciences, University of Wisconsin-Madison, Madison, WI, USA2Boston Children’s Hospital, Harvard Medical School, Boston, MA, USA3Department of Molecular and Medical Genetics, School of Medicine, Oregon Health and Science University, Portland, OR, USA

Correspondence should be addressed to Denise M. Ney; ney@nutrisci.wisc.edu

Received 7 July 2017; Accepted 23 October 2017; Published 31 December 2017

Academic Editor: Ellen Crushell

Copyright © 2017 Bridget M. Stroup et al. *is is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Background. Nutrient status in phenylketonuria (PKU) requires surveillance due to the restrictive low-Phe diet in combinationwith amino acid medical foods (AA-MF) or glycomacropeptide medical foods (GMP-MF). Micronutrient pro8les of medicalfoods are diverse, and optimal micronutrient supplementation in PKU has not been established.Methods. In a crossover design,30 participants with PKU were randomized to consume AA-MF and Glytactin™ GMP-MF in combination with a low-Phe dietfor 3 weeks each. Fasting venipunctures, medical food logs, and 3-day food records were obtained. Metabolomic analyses werecompleted in plasma and urine by Metabolon, Inc. Results. *e low-Phe diets in combination with AA-MF and GMP-MF weregenerally adequate based on Dietary Reference Intakes, clinical measures, and metabolomics. Without micronutrient supple-mentation of medical foods, >70% of participants would have inadequate intakes for 11 micronutrients. Despite micronutrientsupplementation of medical foods, inadequate intakes of potassium in 93% of participants and choline in >40% and excessiveintakes of sodium in >63% of participants and folic acid in >27% were observed. Sugar intake was excessive and provided 27% ofenergy. Conclusions. Nutrient status was similar with AA-MF and Glytactin GMP-MF. More research related to micronutrientsupplementation of medical foods for the management of PKU is needed.

1. Introduction

Phenylketonuria (Online Mendelian Inheritance in Man261600) is an inherited metabolic disease that is characterizedby a loss of function of the hepatic phenylalanine hydroxylase(PAH8; Enzyme Commission number 1.14.16.1), thereby lim-iting the hydroxylation of phenylalanine to tyrosine [1].Diagnosis and initiation of treatment are required toprevent severe cognitive impairment caused by toxic ac-cumulations of Phe in the brain [2, 3]. Nutritional man-agement requires compliance with a lifelong low-Phe diet,which is central to the treatment of PKU. *e low-Phe diet

consists of a controlled amount of natural protein to provideminimum Phe requirements and consumption of elementalamino acid medical foods (AA-MF) or glycomacropeptidemedical foods (GMP-MF), distributed three times per day, toprovide the majority of dietary protein needs [4].

Glycomacropeptide is a 64-amino acid glycophosphopeptide(7–11 kDa) and is part of the κ-casein micelle in bovine milkthat is cleaved by chymosin during the cheese-makingprocess [5]. GMP is not a complete protein and requires sup-plementation with the following indispensable amino acids:Arg, His, Leu, Trp, and Tyr [6, 7]. Our randomized, controlledcrossover trial demonstrated that Glytactin GMP-MF are

HindawiJournal of Nutrition and MetabolismVolume 2017, Article ID 6859820, 17 pageshttps://doi.org/10.1155/2017/6859820

a safe and acceptable alternative with fewer reported gastro-intestinal side eJects compared to AA-MF, which may im-prove lifelong medical food compliance [8]. *is evidencesupports the new paradigm for use of primarily intact proteinfromGMP-MF for the nutritional management of PKU [4, 9].

*e focus on dietarymacro- andmicronutrient patterns hasprogressed from de8ciency prevention towards an emphasis ondisease prevention and wellness. Maintenance of adequatenutritional status, based on dietary intake and biomarkers, ischallenging for PKUdue to the small amount of natural proteinthat can be incorporated into the low-Phe diet [10]. To ac-commodate the restrictive low-Phe diet, PKU medical foodshave been supplemented with chemically derived vitamins andminerals within the last 25 years, and more recently, withessential fatty acids [10, 11]. *e bioavailability of both naturaland chemically derived vitamins and minerals is complex andimpacted by dietary constituents, such as 8ber and fat content[12–15]. For example, there is evidence of increased bio-availability of chemically derived folic acid and retinyl palmitate(vitamin A) and reduced bioavailability of sodium selenite,a common chemically derived form of selenium [12, 16, 17].

Research groups have reported clinical data from PKUsubjects indicating low or de8cient levels of several micro-nutrients, including iron, selenium, vitamin B-12, zinc, andfatty acids, such as total n-3 fatty acids, docosahexaenoic acid(DHA), and eicosapentaenoic acid (EPA) [10, 11, 18–20]. *is8nding is likely due, in part, to reduced bioavailability ofsynthetic nutrients, variable content of nutrients in medicalfoods, or low medical food compliance. At the same time,there is evidence of high serum concentrations of folate andselenium [18]. Regardless, prior studies provide nutritionalstatus assessment data based on mostly children with PKUusing AA-MF. Pinto et al. recently reported similar con-centrations of Phe, select micronutrients (iron, vitamins Dand B-12, folic acid, and zinc), and lipoproteins in individualswith PKU consuming a combination of AA-MF andGlytactinGMP-MF; micronutrient intake was not reported [21]. *eobjective of this study was to utilize metabolomics and tra-ditional dietary evaluation methods to assess the nutritionalstatus of adults and adolescents with PKU consuming AA-MFand Glytactin GMP-MF in combination with a low-Phe diet.Our comprehensive approach assessed dietary intake of

macronutrients and micronutrients from medical foods aswell as natural foods and modi8ed low-protein foods.

2. Methods

2.1. Study Design and Protocol. Assessment of macro- andmicronutrient intake of adults with PKU consuming a low-Phe diet in combination with AA-MF compared to GlytactinGMP-MF was an important secondary aim of our ran-domized, controlled, crossover clinical trial [8]. Participantscompleted the study protocol either at the Waisman Center(n� 19) or at Boston Children’s Hospital (n� 11) [8]. *eUniversity of Wisconsin-Madison Health Sciences Institu-tional Review Board approved the study protocol. All par-ticipants provided written informed consent. *e trial wasregistered at ClinicalTrials.gov as NCT01428258. As pre-viously reported, inclusion criteria included a diagnosis ofPKU that was early treated with medical food and a currentprescription providing greater than 50% of daily proteinneeds from medical foods [8]. *e study protocol includedbaseline (day 1) and 8nal (day 22) study visits for both AA-MFand GMP-MF treatments, where fasting blood samples and 3-day food records were obtained (Figure 1) [8].

Participants consumed a low-Phe diet and were ran-domized to consume AA-MF and GMP-MF for 3 weeks each.*e main change in the GMP-MF treatment was the sub-stitution of all protein equivalent (PE) intake from AA-MFwith GMP-MF. Participants consumed their usual Phe-freeAA-MF, as prescribed by their home metabolic clinics, whichresulted in the use of 15 diJerent AA-MF [8]. *e list of AA-and GMP-MF used by participants has been previously re-ported [8]. Cambrooke *erapeutics donated the GMP-MFfrom 2010 to 2015, which contained Glytactin, a patentedformulation of ∼70% glycomacropeptide (cGMP-20, ArlaFoods Ingredients) and ∼30% supplemental AAs (Arg, His,Leu, Trp, and Tyr). Consecutive 3-day food records andfasting venipunctures were collected prior to the start and theend of both dietary treatments. A daily medical food log waskept by subjects for each of the 3-week treatments and used toassess intake of PE from medical foods compared withprescribed intake of PE from medical foods (average medicalfood prescription in study: 0.85± 0.03 g PE/kg/d). Median

Randomized, controlled, crossover trial

Medical food log

Medical food log

3-day foodrecord

3-day foodrecord

Medical food log

Medical food log

3-day foodrecord

3-day foodrecord

Usualroutine

withAA-MF

GMP-MF, n = 15

AA-MF, n = 15

3 weeks4 weeks

3 weeks

Metabolomics

Metabolomics

Metabolomics

Metabolomics

AA-MF

GMP-MF⁎⁎ ⁎⁎

⁎⁎

⁎⁎

⁎ ⁎

⁎⁎⁎⁎

⁎ Baseline visitFinal visit

Figure 1: Experimental design. AA-MF, amino acid medical foods; GMP-MF, glycomacropeptide medical foods.

2 Journal of Nutrition and Metabolism

intake of PE from medical food was not diJerent betweenAA-MF and GMP-MF (medians, g PE/kg/d: AA-MF, 0.78;GMP-MF, 0.76; p � 0.94) and indicates that participantsobtained approximately 90% of the average medical foodprescription. Compliance with AA-MF and GMP-MF waspreviously reported [8].

2.2. Assessment of Medical Food and Natural FoodIntake. Nutrient intake distributions (10th, median, and 90thpercentiles) for macro- and micronutrients from the wholediet, medical foods, and natural foods, based on 3-day foodrecords at the end of AA-MF and GMP-MF treatments, arereported for the 8rst time. Food record analyses were con-ducted by a registered dietitian (RD) experienced in stan-dardized diet entry using Food Processor SQL (version 10.12.0;ESHA) and were surveilled by a second RD for quality as-surance. Nutrient intake of participants <18 y was combinedwith that of adult participants due to similar PE intake fromMFwith both treatments (means± SE, adults, 54± 3 g PE fromMF, n� 25; minors, 47± 6 g PE fromMF, n� 5; p � 0.38) andthe small sample size. Natural foods were de8ned as all foodand beverages that were not medical foods intended for thetreatment of PKU. Intake of modi8ed low-protein foods(MLPF) was investigated, andMLPFwere de8ned as foods thatwere “modi8ed to be low in protein and formulated for oralconsumption for individuals for whom a condition or disorderprevents typical food consumption.*is does not include foodsthat are naturally low in protein, such as some fruits or veg-etables” [22]. Because MLPF are intended to add variety to thelow-Phe PKU diet, MLPF were categorized as “natural food.”

2.3. Assessment of Inadequate and Excessive Intake of Macro-andMicronutrients. To investigate inadequate and excessiveintakes of macronutrients and micronutrients, we comparedthe average 3-day intake of each nutrient for AA- and GMP-MF at the end of the 3-week treatment period against thegender- and age-appropriate nutrient reference cutoJs foreach individual participant. *e United States’ acceptablemacronutrient distribution range (AMDR) was used toevaluate percent caloric intake from total protein, carbo-hydrate, and fat, and the Dietary Guidelines for Americans2015–2020 (DGA) were used to evaluate percent caloricintake from sugar and saturated fat [23, 24]. *e UnitedStates’ estimated average requirements (EAR) and adequateintake (AI) were used as reference cutoJs to evaluate in-adequate total micronutrient intakes, and the tolerableupper intake levels (UL) were used as reference cutoJs toevaluate excessive total micronutrient intakes [25, 26]. Ex-cessive intakes of vitamin A, vitamin E, niacin, folate, andmagnesium from medical foods, as opposed to the wholediet, were evaluated because the UL for those micronutrientsrepresent intakes from pharmacological sources and notnatural foods [25].

To estimate the percentage of participants with in-adequate or excessive intakes of vitamins A and E, as-sumptions related to vitamin form and the conversions usedto account for bioavailability were made. For vitamin A, wereported intake using international units (IUs) rather than

retinol activity equivalents (RAE) because American foodlabels generally report vitamin A content of foods as percentof daily value or IU and do not account for diJerencesrelated to bioavailability of preformed vitamin A (retinol)and provitamin A carotenoids (α-carotene, β-carotene, andβ-cryptoxanthin). To utilize the EAR and UL to evaluateinadequate and excessive intakes of vitamin A, vitamin Aintake from medical foods was evaluated as retinol, giventhat vitamin A is added to medical foods as retinyl palmitateor retinyl acetate and vitamin A intake from natural foodswas evaluated as β-carotene. We evaluated vitamin A intakefrom natural foods as β-carotene because our participantshad low intakes of forti8ed grain products, and many of thelow-Phe fruits and vegetables consumed by our participantsgenerally have high contents of β-carotene. To evaluateinadequate and excessive intake of vitamin E, vitamin E frommedical foods was evaluated as dl-alpha tocopherol, as this isthe most common form of vitamin E added tomedical foods.Vitamin E from natural foods was evaluated as l-alpha-tocopherol to account for diJerences in bioavailability.

2.4. Clinical Measurements. Nontargeted metabolomic an-alyses on fasting plasma samples from participants with PKU(n� 10) compared to fasting plasma samples from a gender-and age-matched control population (n� 15) and on aliquotsfrom 24-hour urine collections from participants with PKU(n� 9) were carried out by Metabolon, Inc. (Durham, NC,USA). *e plasma and urine samples used for the metab-olomic analyses were obtained from a subset of the 30participants enrolled in the previously reported clinical trial[8]. Methods for metabolomic analyses on plasma and urinesamples have been previously reported [27]. Hemoglobin,plasma ferritin, and serum zinc were analyzed using stan-dard techniques at the clinical laboratories located at theUniversity of Wisconsin Hospital and Clinics and BostonChildren’s Hospital. Serum methylmalonic acid (MMA)concentrations were analyzed using gas chromatography/mass spectrometry by Metabolite Laboratories, Inc.(Denver, CO, USA).

2.5. Statistical Analysis. All statistical analyses were per-formed using SAS version 9.4, and assumptions of normalityand equal variance were tested. Most analyses used PROCMIXED (SAS Institute Inc.). Subject characteristics andDXA scan data were analyzed using ANOVA with eJects forsex and genotype (classical or variant PKU). Nutrient andMLPF intake data were analyzed using ANOVA with eJectsfor diet (AA-MF or GMP-MF), genotype, and diet by ge-notype interactions. Two analyses were conducted formicronutrient intakes from medical foods using ANOVAwith eJects for diet, genotype, and diet by genotype in-teractions. *e 8rst analysis included all 30 subjects, whilethe second analysis only included subjects that consumedmicronutrients from both AA-MF and GMP-MF(n� 21–27). *e Kruskal–Wallis test was used to test fordiJerences due to diet or genotype, if data were skewed.Statistical analyses for the metabolomics were performed inArray Studio version 7.2 and R version 3.02. *e metabolites

Journal of Nutrition and Metabolism 3

in the metabolomic analyses for both plasma and urinesamples were rescaled to set the median to one, based on allsamples. Statistical analyses were conducted to detect dif-ferences in fold change. To detect diJerences in the metab-olomics of the plasma samples, paired t-tests were used tocompare diJerences between treatments and PKU genotypes.Welch’s two-sample t-tests were used to detect diJerencesbetween a dietary treatment and a PKU genotype againsta control population. To detect diJerences in the metab-olomics of the urine samples, paired t-tests were used to detectdiJerences between treatments and genotypes, and ANOVAwas used to test for treatment, genotype, and treatment bygenotype interactions. Data from urine samples were cor-rected for osmolality prior to statistical analysis; the dietarytreatments did not have a signi8cant eJect on urine osmo-lality. Statistical signi8cance was set at p< 0.05.

3. Results

3.1. Participants. *irty participants were enrolled in theclinical trial (18 females and 12males) and included 25 adults(18–49 y) and 5 minors (15–17 y) [8]. Participant charac-teristics are summarized in Table 1. Of the 20 participantscategorized with classical PKU, de8ned as having a PAHgenotype and/or inadequate response to sapropterin dihy-drochloride resulting in a severe PKU phenotype, eight weremale, ten were female, and two were minors. Of the 10participants categorized with variant PKU, de8ned as havinga mild PAH genotype and/or response to sapropterindihydrochloride resulting in a mild PKU phenotype, twowere male, 8ve were female, and three were minors. Fiveparticipants used sapropterin hydrsapropterin dihydro-chloride consistently throughout the study.

3.2.Macronutrient IntakePro3le of theWholeDiets. *e low-Phe diets in combination with AA-MF and GMP-MF weregenerally constant with similar total intakes (i.e., the wholediet) of energy, protein, carbohydrate, fat, Phe, and PE frommedical foods (medians, 53–59 g PE/d) (Table 2). Percentcalories from total protein, carbohydrate, sugar, fat, andsaturated fat as a percentage of total energy intake weresimilar for both AA-MF and GMP-MF whole diets and are

reported only for the low-Phe diet in combination withAA-MF (Figure 2). However, intakes of carbohydrate, fat,and saturated fat from medical foods were signi8cantlyhigher with GMP-MF.*e signi8cantly higher carbohydrateintake with GMP-MF was driven by the higher sugar intakewith GMP-MF (medians, 53 g/d sugar with GMP-MFcompared to 36 g/d sugar with AA-MF; p � 0.10). Partici-pants with variant PKU consumed more protein and Phefrom natural foods and a higher percentage of total energyintake from fat and saturated fat compared with participantswith classical PKU [28]. Additionally, signi8cant diJerenceswere found in the intake of 15 of 18 dietary amino acids,which have been previously reported [29].

3.3. Low Intake of Modi3ed Low-Protein Foods amongParticipants. Of 30 participants, only 13 consumed MLPFwith both AA- and GMP-MF treatments. Macronutrientand Phe intake from MLPF was similar with both diets(Table 3). Interestingly, only 211–257 kcal (medians), <1 g ofprotein, and 51–62 g carbohydrates from MLPF were con-sumed, which comprised ∼10% of median total calories and∼17% of median total carbohydrate intake with both diets(Table 3). Types of MLPF consumed most often during thisstudy included low-protein baking mixes, pasta, cereal, andcheese. Given that most of the MLPF consumed during thisstudy were comprised of mainly starch, micronutrients arenot reported. *e surprising low rates of MLPF intake andlow caloric contribution to the low-Phe diet are likely two-fold: (1) diJerences in public and private coverage of MLPFin the states of the home metabolic clinics within ourheterogeneous participant population and (2) the majorityof MLPF to which participants had access required somepreparation and may have acted as a barrier to excessiveconsumption. To the best of our knowledge, this study is the8rst to report macronutrient intake patterns from MLPF ofindividuals with PKU in the United States.

3.4. Micronutrient Intake Pro3le of the Diets. Vitaminand mineral intake pro8les of the low-Phe diets in combi-nation with AA-MF and GMP-MF are summarized in Tables4 and 5. *e greater intake of natural foods in the diet of

Table 1: Participant characteristics1.

Adult males Adult females Minors2p

Adult and minors Adults onlyAge, y 28± 9 (18–44) 30± 7 (23–49) 16± 1 (15–17) 0.003 0.64BMI, kg/m2 25.2± 3.6 26.9± 5.3 23.2± 2.4 0.29 0.34BMI percentile — — 69.6± 23.7 (30–93) — —Plasma Phe, µmol/L 770± 217 (367–1086) 658± 337 (201–1360) 706± 540 (224–1418) 0.63 0.52Classical PKU, n 8 10 2 — —Variant PKU, n 2 5 3 — —Kuvan™ user, n 1 2 2 — —1Data represent participant characteristic data collected at baseline (visit 1) for our clinical trial, n � 30 [8]. Values are means ± SD, and values inparenthesis represent the minimum and maximum. Subanalyses were done to compare adult male and female participants; 2of the 8ve participants whowere minors (<18 y), two were male and three were female; BMI, body mass index; PKU, phenylketonuria.

4 Journal of Nutrition and Metabolism

participants with variant PKU was associated with greaterintakes of vitamin D, vitamin B-12, iodine, selenium, andzinc compared to subjects with classical PKU [28]. Dietaryintakes of 17 of 25 vitamins andminerals from the whole dietand from medical foods were similar between the AA-MFand GMP-MF treatments. Dietary intakes from naturalfoods were not diJerent between AA-MF and GMP-MF,providing further support that the diets were generallyconstant except for the type of medical food consumed.Participants consumed signi8cantly more dietary thiaminand copper from the whole diet and medical foods withAA-MF compared to GMP-MF and consumed signi8cantlyless dietary choline and sodium from the whole diet andmedical foods with AA-MF (Tables 4 and 5).

Because several AA-MF and GMP-MF do not containmicronutrients, subanalyses were conducted to test whetherthere were diJerences in dietary intakes of micronutrientsfrom only the medical foods that did contain the micro-nutrient of interest (n � 21–27). Of the subanalyses con-ducted for the 25 dietary micronutrient intakes frommedical foods, only the 3 subanalyses, vitamin E, iron, andzinc, demonstrated diJerent outcomes from the fullanalysis (n � 30). *ere was no signi8cant diJerence indietary intake of zinc from medical food based on thesubanalysis (n � 21). However, the subanalyses indicatedthat AA-MF supplemented with vitamins and mineralstended to have higher contents of vitamin E and ironcompared to GMP-MF.

Table 2: Daily macronutrient intake pro8les in combination with AA-MF and GMP-MF1.

Median and percentile nutrient intakespAA-MF GMP-MF

10th Median 90th 10th Median 90thkcal 1471 2076 2711 1523 2148 3152 0.33kcal/kg 23 29 43 21 30 46 0.33kcal from MF 162 393 913 340 694 1007 0.002kcal from NF 969 1597 2247 905 1532 2467 0.53

g protein 61 76 107 56 81 109 0.81g protein/kg 0.85 1.10 1.50 0.80 1.10 1.70 1.00g PE from MF 33 59 75 30 53 73 0.98g PE from MF/kg 0.47 0.78 1.10 0.45 0.76 1.03 0.94% PE from MF 34% 77% 84% 40% 71% 86% 0.58g protein from NF 12 20 46 11 22 58 0.88

mg Phe 487 924 1973 544 1014 2592 0.25mg Phe from MF 0 0 0 50 85 137 0.0001†

mg Phe from NF 487 924 1973 419 929 2524 0.97g carbohydrate 203 294 406 195 347 473 0.10g carbohydrate from MF 1 36 125 42 91 166 0.0001†

g carbohydrate from NF 150 237 363 127 240 369 0.52g sugar/d 67 122 243 78 129 227 0.69g sugar from MF 0 36 102 22 53 109 0.10†

% energy from MF from sugar 0% 34% 51% 20% 30% 55% 0.18g sugar from NF 19 89 170 35 69 153 0.94

g 8ber 10 17 39 9 18 28 0.93g 8ber from MF 0 0 17 0 2 4 0.67g 8ber from NF 10 15 24 9 16 28 0.64

g fat 37 64 92 35 65 102 0.62g fat from MF 0 1 20 0 15 26 0.0004g fat from NF 31 58 82 23 53 93 0.63

g saturated fat 8 15 26 10 21 31 0.046g saturated fat from MF 0 0 1 0 4 11 0.005g saturated fat from NF 8 15 25 5 13 28 0.85

g trans fat2 0.1 0.4 1.3 0.03 0.4 1.8 0.66mg cholesterol2 9 30 118 4 40 136 0.761Nutrient intakes were based on 3-day food records (n� 30). Statistical analysis included ANOVA with eJects for treatment, genotype, and treatment-genotype interaction. *e p values in this table represent the treatment comparison; 2AA-MF and GMP-MF did not contain trans fat nor cholesterol;†Kruskal–Wallis test was used when data were skewed; AA-MF, amino acid medical food; GMP-MF, glycomacropeptide medical food; MF, medical food; NF,natural food; PE, protein equivalent.

Journal of Nutrition and Metabolism 5

3.5. Inadequate and Excessive Micronutrient Intakes per theDietary Reference Intakes. To assess the rates of inadequatemicronutrient intakes with AA-MF and GMP-MF treatments,dietary intakes from the whole diet were compared to the EARor AI, which were evaluated for each individual participantbased on age and gender (Tables 4 and 5). *ere were highrates of inadequate micronutrient intakes from the whole diet,expressed as percentages of participants below the EAR or AI

with AA-MF and GMP-MF treatments, for the followingmicronutrients: potassium (93% for both treatments), choline(>40%), vitamin D (33% for both treatments), and iodine(>23%). Although approximately one-third of participantsdid not meet the EAR for vitamin D, we have previouslyshown that our participants with PKU have no evidence ofvitamin D de8ciency based on concentrations of serum 1,25-dihydroxyvitamin D (means± SE, pg/mL: AA-MF, 65.4±3.39; GMP-MF, 71.9±4.10; p � 0.079) and 25-hydroxyvitaminD (means± SE, ng/mL: AA-MF, 33.6± 1.53; GMP-MF, 33.8±1.70; p � 0.797) [30]. Furthermore, there was some evidenceof inadequate micronutrient intakes from the whole diet withAA-MF and GMP-MF treatments for the following micro-nutrients: magnesium, vitamin E, biotin, zinc, selenium, pan-tothenate, vitamin K, vitamin C, copper, and thiamin.

In addition to rates of inadequate intakes, micronutrientintakes from either the whole diet or medical foods werecompared to the UL to investigate rates of excessive micro-nutrient intake. *ere were high rates of excessive micro-nutrient intakes frommedical foods, expressed as percentagesof participants above the UL with AA-MF and GMP-MFtreatments, for the following micronutrients: magnesium(>37%), folic acid (>27%), and niacin (20% with AA-MF)(Tables 4 and 5). Because the UL for magnesium, folate, andniacin applies to synthetic sources only, comparison of the ULagainst the intake from the whole diet was not evaluated forthose micronutrients. Not surprisingly, >63% of participantshad sodium intakes from the whole diet with AA-MF andGMP-MF that were above the UL (>2300mg/d). Participantsconsumed signi8cantly more sodium from the whole diet withGMP-MF due to the signi8cantly higher sodium intake fromGMP-MF compared to AA-MF (medians, 1140mg sodium/dfrom GMP-MF compared to 413mg sodium/d from AA-MF,

30%Other carbohydrate

27%Sugar

15%Total protein

7%Sat. fat

21%Other fat

(a)

Total proteinTotal carbohydrateSugarTotal fatSaturated fat

AMDR or DGA recommendation ranges as apercentage of total energy

Adults Minors, 4−18 y

10−35%45−65%

<10%20−35%

<10%

10−30%45−65%

<10%25−35%

<10%

(b)

Inadequate and excessive intakesas a percentage of subjects

Total proteinTotal carbohydrateSugarTotal fatSaturated fat

0%20%97%17%13%

7%10%—

17%—

<AMDR orDGA

>AMDR orDGA

(c)

Figure 2: Caloric contribution from carbohydrate, sugar, protein, fat,and saturated fat was calculated for the low-Phe diet in combinationwith AA-MF as a percentage of total energy intake (a). Macronutrientdistribution was similar for GMP-MF (data not shown). Caloric con-tribution from macronutrients was compared to the age-appropriateAMDR (protein, carbohydrate, and fat) or the DGA (sugar and sat-urated fat) (b) for each participant to determine the percentage ofsubjects that were belowor above theAMDRorDGA (c). Percentage ofparticipants with inadequate and excessive macronutrient distributionswere similar for GMP-MF (data not shown). AA-MF, amino acidmedical food; AMDR, acceptable macronutrient distribution ranges;DGA, Dietary Guidelines for Americans; GMP-MF, glycomacropep-tide medical foods; Sat. fat, saturated fat.

Table 3: Modi8ed low-protein daily food intake pro8les incombination with AA-MF and GMP-MF1.

Median and percentile nutrient intakespAA-MF GMP-MF

10th Median 90th 10th Median 90thEnergy, kcal/d 73 211 463 85 257 690 0.31Protein, g/d 0.1 0.6 1.3 0.2 0.4 1.6 0.71Phe, mg/d 3 13 60 3 20 93 0.32Carbohydrate,g/d 17 51 103 14 62 125 0.49‡

Fiber, g/d 0 0.3 5 0 1 6 0.13Fat, g/d 0 1 5 0.5 3 18 0.13‡

Saturated fat,g/d 0 0.2 2 0 0.5 4 0.25‡

1Nutrient intakes from MLPF were based on 3-day food records (n� 13;classical PKU, n� 10; variant PKU, n� 3). Participants that did not consumeMLPF at the end of both treatments were removed from the analysis. MLPFwere de8ned as foods modi8ed to be low in protein and speci8cally madefor individuals with disorders that have dietary protein restrictions. Sta-tistical analysis included ANOVA with eJects for treatment, genotype, andtreatment-genotype interaction. *e p values in this table represent thetreatment comparison. *ere were no signi8cant diJerences due to ge-notype or the treatment-genotype interaction (data not shown);‡Kruskal–Wallis test was used when data were skewed; MLPF, modi8edlow-protein foods.

6 Journal of Nutrition and Metabolism

Tabl

e4:Treatmentcomparisonofvitaminandadditivenutrientintakedistributionsincomparisonwith

referenceintakecutoJswith

AA-MFandGMP-MF1.

Vitamins

AA-MF

GMP-MF

pn

Percentilenutrientintakes

Inadequateand

excessiveintakes2

Percentilenutrientintakes

Inadequateand

excessiveintakes

10th

Median

90th

<EARorAI>UL

10th

Median

90th

<EARorAI>UL

Fat-solublevitamins

VitaminA,IU

303454

6996

17,572

0%—

2479

8089

16,320

3%—

0.39

VitaminAfrom

MF

300

2642

4037

—0%

02246

4008

—0%

0.15

VitaminAfrom

MF3

211543

2840

4063

——

1348

2295

4043

——

0.60

VitaminAfrom

NF4

301656

5061

14,604

7%—

983

5232

13,795

10%

—0.99

VitaminD,IU

30148

562

1941

33%

3%36

511

1243

33%

3%0.58†

VitaminDfrom

MF

300

396

960

——

0485

868

——

0.99†

VitaminDfrom

MF

24234

599

960

——

0500

929

——

0.50†

VitaminDfrom

NF

302

27644

90%

—3

30159

93%

—1.00

VitaminE,IU

3013

2955

13%

—7

2537

23%

—0.002

VitaminEfrom

MF

300

1828

—0%

015

23—

0%0.14†

VitaminEfrom

MF

2113

1827

——

815

25—

—0.01

VitaminEfrom

NF

306

1038

80%

—4

1020

87%

—0.16†

VitaminK,µg

30119

160

412

3%ND5

59190

393

23%

ND

0.27

VitaminKfrom

MF

300

75112

——

074

131

——

0.91†

VitaminKfrom

MF

2155

75109

——

4275

134

——

0.75

VitaminKfrom

NF

3029

103

311

47%

—20

102

308

53%

—0.39

Water-solublevitamins

VitaminC,mg

3078

161

276

3%0%

45164

296

13%

0%0.86

VitaminCfrom

MF

300

61130

——

074

129

——

0.46†

VitaminCfrom

MF

2142

62133

——

4481

131

——

0.19

VitaminCfrom

NF

3019

94212

37%

—27

72197

43%

—0.62

*iamin,m

g30

1.2

2.5

6.5

3%ND

0.9

2.0

2.7

13%

ND

0.03†

*iaminfrom

MF

300

1.3

3.9

——

01.0

1.7

——

0.01

*iaminfrom

MF

210.6

1.3

3.7

——

0.6

1.0

1.8

——

0.02

*iaminfrom

NF

300.5

1.0

3.0

33%

—0.4

0.9

1.7

63%

—0.15†

RiboUavin,mg

301.0

2.6

4.0

10%

ND

1.8

2.9

4.7

0%ND

0.33†

RiboUavinfrom

MF

300

1.5

2.8

——

1.11.8

2.9

——

0.12

RiboUavinfrom

MF

250.7

1.5

2.9

——

1.0

1.7

2.8

——

0.41

RiboUavinfrom

NF

300.5

1.0

2.1

47%

—0.4

0.8

1.9

63%

—0.31†

Niacin,mg

3016

3165

3%—

1632

423%

—0.87†

Niacinfrom

MF

300

1655

—20%

718

26—

0%0.52†

Niacinfrom

MF

255

2157

——

718

27—

—0.68†

Niacinfrom

NF

307

1326

37%

—6

1225

43%

—0.19

Journal of Nutrition and Metabolism 7

Tabl

e4:Continued.

Vitamins

AA-MF

GMP-MF

pn

Percentilenutrientintakes

Inadequateand

excessiveintakes2

Percentilenutrientintakes

Inadequateand

excessiveintakes

10th

Median

90th

<EARorAI>UL

10th

Median

90th

<EARorAI>UL

VitaminB-6,mg

301.9

3.1

5.7

3%0%

1.8

3.1

5.7

0%0%

0.65†

VitaminB-6from

MF

300

1.7

2.8

——

1.11.8

2.9

——

0.28

VitaminB-6from

MF

251

1.7

2.9

——

1.11.8

2.9

——

0.62

VitaminB-6from

NF

300.7

1.5

2.8

27%

—0.5

1.2

4.3

37%

—0.34†

FolateDFE,µg

30504

1159

1902

3%—

593

1193

1461

3%—

0.58†

FolateDFE

from

MF

300

743

1507

—37%

192

733

1126

—27%

0.80†

FolateDFE

from

MF

24382

851

1512

——

457

800

1141

——

0.06

FolateDFE

from

NF

30160

347

847

40%

—116

364

790

43%

—0.35

VitaminB-12,mg

302.3

6.5

11.8

10%

ND

4.1

7.1

13.9

0%ND

0.23†

VitaminB-12from

MF

300

5.1

6.8

——

2.4

4.8

9.4

——

0.14

VitaminB-12from

MF

252.2

5.4

7.0

——

2.4

4.5

10.8

——

0.71

VitaminB-12from

NF

300.1

1.5

8.1

53%

—0.2

1.18.7

63%

—0.71†

Choline,mg

3074

311

674

63%

0%71

512

868

40%

0%0.02

Cholinefrom

MF

300

277

628

——

0430

780

——

0.03

Cholinefrom

MF

21111

414

588

——

253

501

806

——

0.003

Cholinefrom

NF

3025

65109

100%

—36

63172

100%

—0.15

Pantothenate,mg

304

915

17%

ND

610

167%

ND

0.47†

Pantothenatefrom

MF

300

513

——

47

11—

—0.21

Pantothenatefrom

MF

253

513

——

47

12—

—0.96

Pantothenatefrom

NF

301

37

83%

—1

24

93%

—0.38†

Biotin,mg

3018

67242

23%

ND

27104

428

13%

ND

0.14

Biotinfrom

MF

300

42160

——

2367

359

——

0.06

Biotinfrom

MF

2520

97160

——

2254

403

——

0.69

Biotinfrom

NF

302

6126

83%

—1

413

93%

—0.32†

Other

Carnitinefrom

MF,g

300.030

0.057

0.100

——

<0.001

<0.001

<0.001

——

<0.0001†

Taurinefrom

MF,g

300.065

0.130

0.300

——

0.085

0.160

0.258

——

0.27

InositolfromMF,g

300

0.080

0.135

——

00.005

0.048

——

<0.0001†

1 Nutrientintakeswerebasedon3-dayfoodrecords(n�30);statisticalanalysisincludedANOVAwitheJectsfortreatment,genotype(classicalorvariantPKU),andtreatment-genotypeinteraction.*ep

valuesin

thistablerepresentthetreatmentcomparison;2 inadequateandexcessiveintakesareexpressedasapercentageof30subjects.UnitedStates’DietaryReferenceIntake(DRI)cutoJs,basedontheAMDR,EAR,AI,

andUL,werecomparedtonutrientintakesperthesexandageofeachindividualsubject[25,26].*

eUnitedStatesDietaryGuidelinesforAmericans2015–2020referencecutoJswereusedtoevaluatedietary

saturatedfatandsugarintakeofsubjects[24];3subanalysesformicronutrientintakesfrommedicalfoodswereconductedduetothehighratesofuseofAA-MFthatlackedmicronutrientsupplementation.*ese

subanalysesaimedtocomparemicronutrientintakesfrommedicalfoodsthatweresupplementedwithmicronutrients.Duetothediversemicronutrientsupplementationpro8lesofmedicalfoods,samplesizesfor

eachmicronutrientvary;4naturalfoodswerede8nedasallfoodsandbeveragesthatwerenotmedicalfoodsintendedforthetreatmentofPKU;5pertheDRI,aULforselectnutrientshasnotbeendetermineddueto

alackofscienti8cevidence;† Kruskal–W

allistestwasusedwhendatawereskewed;AMDR,acceptablemacronutrientdistributionrange;AI,adequateintake;AA-MF,aminoacidmedicalfood;DFE,dietaryfolate

equivalents;DGA,DietaryGuidelinesforAmericans;EAR,estim

atedaveragerequirement;GMP-MF,glycomacropeptidemedicalfood;M

F,medicalfood;ND,notdetermined;NF,naturalfood;PKU,

phenylketonuria;UL,uppertolerableintakelevels.

8 Journal of Nutrition and Metabolism

Tabl

e5:Treatmentcomparisonofmineralintakedistributionsincomparisonwith

referenceintakecutoJswith

AA-MFandGMP-MF1.

Minerals

AA-MF

GMP-MF

pn

Percentilenutrientintakes

Inadequateand

excessiveintakes2

Percentilenutrientintakes

Inadequateand

excessiveintakes

10th

Median

90th

<EARorAI>UL

10th

Median

90th

<EARorAI>UL

Calcium,mg

30839

1468

2402

10%

0%586

1529

2453

20%

7%1.00†

Calcium

from

MF

30106

1103

2077

——

631153

2225

——

0.99

Calcium

from

MF3

27552

1151

2129

——

601255

2391

——

0.62†

Calcium

from

NF4

30183

313

784

93%

—156

401

726

90%

—0.80†

Copper,µg

301.2

1.9

4.2

3%7%

0.6

1.5

2.7

13%

0%0.0006†

Copperfrom

MF

300

1.2

3.0

——

00.6

1.1—

—0.005

Copperfrom

MF

210.6

1.2

2.8

——

0.4

0.7

1.1—

—0.0001

Copperfrom

NF

300.5

0.9

3.4

33%

—0.4

0.8

1.9

40%

—0.27†

Iodine,µg

3013

140

235

23%

0%12

153

262

33%

0%0.66†

Iodinefrom

MF

300

132

223

——

0141

239

——

0.48†

Iodinefrom

MF

2177

137

223

——

78149

241

——

0.32

Iodinefrom

NF

302

942

93%

—0

848

97%

—0.26

Iron,m

g30

1425

430%

3%9

2331

3%0%

0.02

Iron

from

MF

300

1531

——

014

23—

—0.14†

Iron

from

MF

2110

1631

——

814

23—

—0.02

Iron

from

NF

305

822

37%

—4

817

33%

—0.15

Magnesium,mg

30281

501

746

13%

—196

514

782

23%

—0.78†

Magnesiumfrom

MF

3032

320

570

—37%

41353

578

—53%

0.77†

Magnesiumfrom

MF

27126

321

570

——

38354

600

——

0.80†

Magnesiumfrom

NF

30110

171

320

90%

—77

165

292

87%

—0.41

Manganese,mg

301.9

3.9

7.1

10%

0%1.8

3.8

5.3

13%

3%0.26

Manganesefrom

MF

300

1.6

3.4

——

01.6

3.0

——

0.84†

Manganesefrom

MF

210.9

1.7

3.4

——

0.6

1.9

3.1

——

0.41†

Manganesefrom

NF

301.1

2.2

6.3

40%

—0.9

1.9

3.7

50%

—0.10

Phosphorus,m

g30

952

1616

2431

7%0%

774

1536

2204

10%

0%0.60

Phosphorusfrom

MF

3081

972

2036

——

253

965

1708

——

0.71†

Phosphorusfrom

MF

27531

1032

2047

——

240

984

1800

——

0.26

Phosphorusfrom

NF

30329

526

1007

73%

—259

538

1064

67%

—0.80

Potassium,mg

301809

2810

4179

93%

ND5

2014

3350

4540

93%

ND

0.15

Potassiumfrom

MF

300

600

2558

——

718

1146

1781

——

0.08†

Potassiumfrom

MF

2560

880

2673

——

680

1120

1815

——

0.46†

Potassiumfrom

NF

30938

1910

3011

97%

—1077

1867

3204

100%

—1.00†

Journal of Nutrition and Metabolism 9

Tabl

e5:Continued.

Minerals

AA-MF

GMP-MF

pn

Percentilenutrientintakes

Inadequateand

excessiveintakes2

Percentilenutrientintakes

Inadequateand

excessiveintakes

10th

Median

90th

<EARorAI>UL

10th

Median

90th

<EARorAI>UL

Selenium,µg

3054

93133

7%0%

2588

142

17%

0%0.50†

Selenium

from

MF

309

5889

——

055

89—

—0.34†

Selenium

from

MF

2326

5989

——

2960

93—

—0.31

Selenium

from

NF

3012

2988

83%

—8

27103

73%

—0.95

Sodium,mg

301816

2637

3879

7%63%

2052

3261

4645

7%83%

0.048

Sodium

from

MF

300

413

1175

——

580

1140

2040

——

<0.0001†

Sodium

from

MF

2517

583

1198

——

480

1123

2280

——

0.0003

Sodium

from

NF

301363

2206

3138

17%

—1007

2041

3383

23%

—0.69

Zinc,mg

305

1641

13%

13%

513

3120%

10%

0.12

Zincfrom

MF

300

1230

——

09

17—

—0.05

Zincfrom

MF

216

1228

——

610

17—

—0.14†

Zincfrom

NF

303

417

77%

—2

419

83%

—0.64†

1 Nutrientintakeswerebasedon3-dayfoodrecords(n�30);statisticalanalysisincludedANOVAwitheJectsfortreatment,genotype(classicalorvariantPKU),andtreatment-genotypeinteraction.*ep

valuesin

thistablerepresentthetreatmentcomparison;

2 UnitedStates’DietaryReferenceIntake(DRI)cutoJs,basedontheAMDR,EAR,AI,andUL,werecomparedtonutrientintakesperthesexandageofeach

individualsubject[25,26];3sub-analysesformicronutrientintakesfrommedicalfoodswereconductedduetothehighratesofuseofAA-MFthatlackedmicronutrientmineralsupplementation.*esesubanalyses

aimedtocomparemicronutrientintakesfrom

medicalfoodsthatweresupplementedwith

micronutrients.Duetothediversemicronutrientsupplementationpro8lesofmedicalfoods,samplesizesforeach

micronutrientvary;4naturalfoodswerede8nedasallfoodsandbeveragesthatwerenotmedicalfoodsintendedforthetreatmentofPKU;5pertheDRI,aULforselectnutrientshasnotbeendetermineddueto

alackofscienti8cevidence;† Kruskal–W

allistestwasusedwhendatawereskewed;AMDR,acceptablemacronutrientdistributionrange;AI,adequateintake;AA-MF,aminoacidmedicalfood;DFE,dietaryfolate

equivalents;DGA,DietaryGuidelinesforAmericans;EAR,estim

atedaveragerequirement;GMP-MF,glycomacropeptidemedicalfood;M

F,medicalfood;ND,notdetermined;NF,naturalfood;PKU,

phenylketonuria;UL,uppertolerableintakelevels.

10 Journal of Nutrition and Metabolism

p< 0.0001, n� 30). Nonetheless, the majority of the dietarysodium intake was obtained from natural foods (medians,2041mg sodium/d with GMP-MF compared to 2206mgsodium/d with AA-MF, p � 0.69).

To test whether the vitamin and mineral content oflow-Phe diets would be adequate without intake of medicalfoods, we compared the micronutrient intake from naturalfoods to the gender- and age-appropriate EAR or AI foreach participant. Without medical foods, more than 70% ofparticipants would not meet the EAR or AI for the followingmicronutrients: choline (100%), potassium (>97%), iodine(>93%), calcium (>90%), vitamin D (>90%), magnesium(>87%), biotin (>83%), pantothenate (>83%), vitamin E(>80%), zinc (>77%), and selenium (>73%). Interestingly,without medical food intake, >90% of participants wouldbe able to obtain vitamin A from natural foods alone to meetor exceed the EAR due to high intakes of provitamin Acarotenoids from green leafy vegetables (i.e., spinach andlettuces), squashes, carrots, and tomato products, suchtomato-based pasta sauce (Table 4).

3.6. Metabolomics. Metabolomic analyses of Met meta-bolism, vitamins, and food components in plasma and urinesamples are reported in Supplemental Tables 1–3.

3.6.1. Vitamins A, C, and E. Plasma retinol concentrationswere similar between AA-MF and GMP-MF treatments and17% higher (p � 0.17) compared to controls. Given thatplasma retinol is homeostatically controlled, the somewhathigher retinol levels in PKU subjects are unexplained. Sig-ni8cantly higher plasma retinal concentrations (∼1.7-fold ↑)were found in participants with variant and classical PKUwho consumed AA-MF and GMP-MF compared to a con-trol population (GMP-MF versus controls, p � 0.0006;AA-MF versus controls, p � 0.0053). Although plasma retinalis not recognized as a biomarker of vitamin A status, thisagrees with the high dietary intakes of preformed vitamin Awith both AA-MF and GMP-MF in participants with PKU.Not surprisingly, given the low intake of citrus fruits,pumpkin, and red pepper, participants with variant andclassical PKU and participants who consumed AA-MF andGMP-MF showed 0.3-fold lower plasma β-cryptoxanthinconcentrations (p � 0.0064) compared to controls [12].Consistent with similar intakes of dietary vitamin C, therewere no diJerences in urinary ascorbate and dehydroascorbateconcentrations.

Plasma α-tocopherol concentrations were similar betweenAA-MF and GMP-MF treatments and between participantsconsuming AA-MF or GMP-MF compared to controls, despitesigni8cantly lower vitamin E intakes with GMP-MF. However,plasma α-tocopherol concentrations were 1.25-fold higher inparticipants with variant PKU compared to classical PKU(p � 0.03). Furthermore, plasma α-tocopherol concentrationswere 0.8-fold lower in participants with classical PKU comparedto controls (p � 0.054). Urinary excretion of alpha-CEHC(2,5,7,8-tetramethyl-2-(2′-carboxyethyl)-6-hydroxychroman),a major water-soluble α-tocopherol metabolite, was notsigni8cantly diJerent based on dietary treatment or PKU

genotype. Interestingly, plasma concentrations of gamma-CEHC (3-(2,7,8-trimethyl-3,4-dihydro-2H-chromen-2-yl)propanoate), a water-soluble c-tocopherol metabolite, wereapproximately twofold higher in participants consumingGMP-MF and AA-MF (GMP-MF versus controls, p � 0.03;AA-MF versus controls, p � 0.07) and participants withvariant PKU compared to the control population (VariantPKU versus controls, p � 0.002). Urinary excretion ofgamma-CEHC concentrations was 2.8-fold higher withGMP-MF compared to AA-MF (p � 0.028). *is is the 8rststudy to report dietary intake of vitamin E and metabolitesfor both α- and c-tocopherol in plasma and urine.

3.6.2. B-Vitamins. Compared to AA-MF, 0.48-fold less thi-amin was excreted in the urine with GMP-MF (p � 0.043),which agrees with the signi8cantly lower dietary thiaminintake with GMP-MF. No signi8cant treatment diJerenceswere found in urinary riboUavin concentrations, consistentwith similar dietary riboUavin intakes. Higher plasma nico-tinamide concentrations were observed in participantsconsuming AA-MF (1.42-fold ↑, p � 0.07) and GMP-MF(1.47-fold ↑, p � 0.06) compared to a control population.Plasma nicotinamide concentrations that were 1.66-foldhigher were observed in participants with variant PKUcompared to controls (p � 0.01). *ese results align withsimilar dietary niacin intakes between treatments and theobservation that 20% of participants exceeded the UL fordietary niacin intake with AA-MF.

Despite similar dietary pantothenate intakes and urinarypantothenate concentrations, plasma pantothenate concen-trations in participants with classical and variant PKU andparticipants who used GMP-MF were signi8cantly highercompared to controls (classical PKU versus controls, 1.34-fold↑, p � 0.02; variant PKU versus controls, 1.58-fold ↑, p � 0.03;GMP-MF versus controls, 1.48-fold ↑, p � 0.01). In agreementwith similar dietary vitamin B-6 intakes between treatmentsand PKU genotype, no signi8cant diJerences related to diet orgenotype were found in plasma concentrations of pyridoxaland pyridoxate or urine concentrations of pyridoxal, pyr-idoxate, and pyridoxamine. Additionally, plasma levels ofpyridoxal and pyridoxate in PKU subjects were not diJerentcompared with controls.

Although >40% of participants had dietary choline in-takes below the EAR and obtained higher intakes of cholinewith GMP-MF compared to AA-MF, there were no sig-ni8cant diJerences in plasma choline concentrations amongour participants related to diet or genotype and in com-parison with the control population. However, this 8nding isnot surprising, considering that plasma choline is homeo-statically controlled [31]. No diJerences in urinary excretionof choline-related metabolites, dimethylglycine and betaine,between dietary treatments nor PKU genotype were found.However, participants with variant PKU had 1.46-foldhigher plasma dimethylglycine compared to classical PKU(p � 0.013), and participants with classical PKU had 0.8-foldlower plasma betaine concentrations compared to controls(p � 0.042). Taken together, choline may be a nutrient ofconcern in PKU.

Journal of Nutrition and Metabolism 11

3.6.3. Taurine. Interestingly, despite similar dietary taurineintakes and plasma taurine concentrations, 0.2-fold lesstaurine was excreted in the urine with GMP-MF comparedto AA-MF (p � 0.002). Furthermore, urinary excretion oftaurine-related metabolites (hypotaurine, cysteine sul8nate,cysteine, and cystathionine) was signi8cantly higher withAA-MF compared to GMP-MF (Figure 3). We have pre-viously reported high urinary sulfate excretion with AA-MF(that also exceeded the reference range) compared to GMP-MF (urinary sulfate excretion (mEq/d, means± SE); AA-MF,34± 3, versus GMP-MF, 12± 3, p � 0.0008) [30]. Currentsulfur-containing amino acid content of AA-MF provides∼3 times the World Health Organization’s recommendationfor daily intakes of Met and Cys [32]. Moreover, higherurinary excretion of sulfate, taurine, and taurine-relatedmetabolites with AA-MF may be related to increasedneed to excrete sulfur with higher intake of sulfur-containing amino acids, Met and Cys, from AA-MF com-pared with GMP-MF [33]. Taken together, these data suggestthat the sulfur-containing amino acid content of AA-MFmight be higher than what is needed to support proteinsynthesis and one-carbon metabolism.

3.6.4. Sweeteners and Inositol. *e arti8cial sweetener ace-sulfame was added to both AA-MF and GMP-MF, andconsumption was higher in PKU subjects as reUected inthreefold higher plasma levels of acesulfame in PKU subjectscompared to controls (scaled intensity means, variant, 0.76;classical, 0.63; controls, 0.237; p � 0.067). Notably, urinaryacesulfame excretion was 6 times higher with GMP-MFcompared to AA-MF (p � 0.005). Plasma erythritol andurinary erythritol excretion were dramatically higher withGMP-MF compared to AA-MF (p � 0.019 to 0.003); plasmalevels were also higher with GMP-MF compared withcontrols (p � 0.006). Erythritol is a partially absorbed sugaralcohol that is approved by the FDA for use as a food ad-ditive. No signi8cant diJerences were found in plasma levelsof saccharin and in urinary excretion of saccharin andsucralose. Consistent with similar intake of dietary inositolfrom AA-MF and GMP-MF, no signi8cant diJerences werefound in plasma and urine concentrations of myo- andchiro-inositol.

3.7. Adequate Status of Iron, Vitamin B-12, and Zinc Based onHematological Measures. Average concentrations of he-moglobin, plasma ferritin, serum MMA, and zinc werewithin normal limits (Supplemental Table 4). Consistentwith adequate iron status, hemoglobin and ferritin levelswere within normal limits for >93% of participants, and>97% of participants met or exceeded the EAR for dietaryiron intake. Despite the fact that 13–20% of participantsconsuming AA-MF or GMP-MF, respectively, did not meetthe EAR for dietary zinc, serum zinc concentrations for 94%of participants were within normal limits for both treat-ments. Given that serum zinc is correlated with dietary zincintake and a good biomarker of zinc status [34], these datasuggest adequate zinc status. In line with adequate vitaminB-12 status, MMA concentrations were within normal limits

for >94% of participants, and >90% of participants met orexceeded the EAR for dietary vitamin B-12 intake.

4. Discussion

4.1. Summary of Intake. Individuals with PKU are at risk fornutrient de8ciencies and toxicities due to the restrictive low-Phe diet in combination with medical foods that are sup-plemented with chemically derived nutrients in variableamounts [35, 36]. We conducted an intervention trial thatinvestigated nutrient status using dietary intakes and bio-chemical measures in adults and adolescents with classicaland variant PKU who showed similar compliance whenconsuming AA-MF and Glytactin GMP-MF. A list of themain conclusions from this study is detailed in Table 6. Forthe majority of micronutrients, we found similar total in-takes when participants consumed AA-MF or GlytactinGMP-MF and no diJerences in intakes of micronutrientsfrom natural foods. *is indicates that diJerences inmicronutrient intake were driven by the diverse micro-nutrient supplementation pro8les of the medical foods usedby our participants. Although participants obtained ade-quate intakes of most micronutrients based on the EAR,inadequate intakes of potassium for 93% of participants andcholine for >40% of participants were observed. In contrastto inadequate intake, participants had excessive intakes(>UL) of chemically derived folic acid and magnesium frommedical foods, and >63% of participants had excessive in-takes of sodium driven by natural (likely processed) foodintake. Nonetheless, vitamin and mineral supplementationof medical foods is necessary in order to prevent nutrientde8ciency in PKU. Our data showed that at least 70% ofparticipants would obtain inadequate intakes for 11 vitaminsand minerals without supplemented medical foods (biotin,choline, pantothenate, vitamins D and E, potassium, cal-cium, iodine, magnesium, selenium, and zinc).

4.2. Common Nutrients of Concern in PKU. Adequate nu-trient status in PKU relies on compliance with medical foodsupplemented with vitamins and minerals due to the lownatural protein intake tolerated with the low-Phe diet [4].Micronutrient de8ciency, as evidenced by dietary intake andbiochemical measures, has been reported for iron, vitaminB-12, and zinc [35, 36] but can generally be avoided withcompliance with medical foods that are supplemented withthese micronutrients [35–38]. Consistent with adequate ironand vitamin B-12 status, Pinto et al. found similar levels offerritin, transferrin, hemoglobin, and vitamin B-12 that werewithin normal limits in 11 subjects with PKU consumingAA-MF in combination with Glytactin GMP-MF for ∼13months [21]. Our data also suggest that participants dem-onstrated adequate status of iron, vitamin B-12, and zinc asevidenced by dietary intakes that meet or exceed the EARand relevant biomarker concentrations within normal limitsfor the majority of participants.

Excessive folic acid intake supported by high levels offolate or folic acid biomarkers is noted to be of concern inPKU [18, 36, 39]. Evans et al. reported folate intakes that

12 Journal of Nutrition and Metabolism

Dietary methionine1.61.41.21.00.80.60.40.20.0

g M

et fr

om M

F/d

1.3 ± 0.2

1.2 ± 0.1

1.5 ± 0.4

1.9 ± 0.4

0.9 ± 0.1

0.9 ± 0.1

0.7 ± 0.2

0.5 ± 0.2

p = 0.047

p = 0.066

p = 0.005

p = 0.004

GMP-MFAA-MF

GMP-MFAA-MF

GMP-MFAA-MF

Urinary cysteine1.41.21.00.80.60.40.20.0

Scal

ed in

tens

itySc

aled

inte

nsity

Scal

ed in

tens

ity

O2Cysteine dioxygenase

Cystathioninelyase

,AA-MF

Urinary hypotaurine

½ O2Hypotaurine dehydrogenase

2.0

1.5

1.0

0.5

0.0

Urinary taurine2.5

2.0

1.5

1.0

0.5

0.0

Methionine

Homocysteine

Methioninesynthase

Cystathionine-β-synthase Serine

Diet

Cysteine sulfinate decarboxylase CO2

GMP-MFAA-MF

Dietary cysteine

2.5 ± 0.3

0.03 ± 0.007

p < 0.0001

GMP-MFAA-MF

3.02.52.01.51.00.50.0

g Cy

s fro

m M

F/d

,AA-MF

,AA-MF1.4 ± 0.3

0.7 ± 0.2

p = 0.02

Scal

ed in

tens

ity

Urinary cystathionine1.81.61.41.21.00.80.60.40.20.0

GMP-MFAA-MF,AA-MF

1.2 ± 0.2

0.8 ± 0.1

p = 0.03

GMP-MFAA-MF

Scal

ed in

tens

ity

Urinary cysteine sulfinate1.61.41.21.00.80.60.40.20.0

,AA-MF

,AA-MF

,AA-MF

Figure 3: Urine metabolomics of sulfur-containing amino acid (Met and Cys) metabolism, n� 9. Values are means±SE. Participants with PKUconsumed signi8cantly moreMet and Cys frommedical foods with AA-MF compared to GMP-MF, n� 8 (means±SE; Met, g/d: AA-MF, 1.3±0.2,versus GMP-MF, 0.9±0.1, p � 0.047; Cys, g/d: AA-MF, 2.5±0.3, versus GMP-MF, 0.3±0.007, p< 0.0001). Participants had higher urinaryexcretion of taurine and taurine-relatedmetabolites (hypotaurine, cysteine sulfonate, cysteine, and cystathionine), possibly related to higher intakes ofsulfur-containing amino acids and increased need to excrete sulfur. AA-MF, amino acidmedical foods; GMP-MF, glycomacropeptidemedical foods.

Journal of Nutrition and Metabolism 13

were 201–267% of the Recommended Nutrient Intake in 51children, aged 1–16 years, supported by high serum folateconcentrations that were above normal limits in 83% ofsubjects [18]. Crujeiras et al. showed high folate levels in 39%of 156 children and adults with PKU, aged 7 months–42years. We observed dietary folic acid intakes from AA-MFand Glytactin GMP-MF that exceeded the UL in >27% ofparticipants, but folate or folic acid concentrations were notmeasured. In contrast, Pinto et al. did not 8nd high folic acidconcentrations in subjects consuming AA-MF and GlytactinGMP-MF, but dietary intakes of folate and folic acid werenot reported [21]. Andrade et al. found altered methylationcapacity resulting in low homocysteine concentrations in 42subjects with PKU compared to 40 controls, which theyattribute to high intakes of folic acid and vitamin B-12 fromlow-Phe medical food products [40]. Dobrowolski et al.reported methylome repatterning in the brain demonstratedby aberrant DNAmethylation in brain tissue of subjects andmice with PKU [41, 42]. Given the role of folate in one-carbonmetabolism, it is possible that excessive intake of folicacid from both AA-MF and GMP-MF and excessive intakeof sulfur-containing amino acids from AA-MF could con-tribute to gene dysregulation in the brain secondary toaberrant DNA methylation [41, 42]. In short, excessive folicacid intake from medical foods might contribute to theneuropathology of PKU.

Selenium is one micronutrient where observations ofde8ciency and toxicity, supported by plasma seleniumand glutathione peroxidase activity, have been reported[18, 35, 36]. ConUicting reports of selenium de8ciency andtoxicity may be related to a combination of reduced bio-availability of chemically derived selenium and high sup-plementation of some medical foods with selenium.Nonetheless, >80% of participants in this study obtainedintakes of selenium that met the EAR, but biomarkers ofselenium status were not measured.

4.3. Metabolomics

4.3.1. Vitamin A. Most participants (>90%) were able toreach the EAR for vitamin A without supplementation frommedical foods, although status of our participants is unclear.Two studies have come to conUicting conclusions regardingvitamin A status in PKU based on reported dietary intake andutilization of diJerent biomarkers for vitamin A [43, 44].Colome et al. reported good vitamin A status based on plasmaretinol; however, retinol is homeostatically controlled and isonly reduced in cases of severe de8ciency [12, 43]. Schulpiset al. reported concerns of hypervitaminosis A based on highplasma β-carotene concentrations in 46 children with PKU,who consumed relatively large amounts of carotenoid-richlow-Phe fruits and vegetables compared to age-matchedcontrols [44]. Given the high consumption of carotenoid-rich fruits and vegetables in PKU and the potential for hy-pervitaminosis A with excess provitamin A intake, high plasmaconcentrations of retinal and β-carotene may be particularlyrelevant to assessment of vitamin A status in PKU [12, 43].

Most AA-MF and GMP-MF are supplemented withpreformed vitamin A in the form of retinyl palmitate or retinylacetate, which may increase the potential for excessive vitaminA intake. Hypervitaminosis A can increase risk for osteopo-rosis, liver dysfunction, and immune function alterations [12].*ough vitamin A intakes observed in our participants werehigh (median intakes, 6996–8089 IU/d), 64–72% of total vi-tamin A intake was obtained from natural foods. Because themajority of our participants’ vitamin A intake likely camefrom provitamin A carotenoids (β-carotene, α-carotene, andβ-cryptoxanthin), participants’ high vitamin A intake doesnot exceed the UL, which includes only preformed vitamin Aper current Institute of Medicine standards [25]. *ough weobserved ∼1.7-fold higher plasma retinal concentrations inPKU subjects, independent of diet and genotype, comparedto controls, no conclusions related to hypervitaminosis Acan be made, given the limited understanding of evaluationof vitamin A status in the vitamin A 8eld. Given the knowncomplications of skeletal fragility and inUammation inPKU, serum retinyl esters, a newly recognized biomarker ofhypervitaminosis A, should be measured to further investigatethe concerns of hypervitaminosis A [12]. Lastly, supplemen-tation of medical foods with provitamin A, rather thanpreformed vitamin A, may be prudent.

4.3.2. Choline. Choline is an essential nutrient that is im-portant for function of all cells, such as structural integrity

Table 6: Main study conclusions.

(1)

Similar total dietary intakes of most micronutrients whenparticipants consumed AA-MF or Glytactin GMP-MF andno diJerences in intakes of micronutrients from naturalfoods were observed. *us, diJerences in micronutrientintakes were driven by the diverse micronutrientsupplementation pro8les of the medical foods.

(2)

Participants obtained adequate intakes (≥EAR) of mostmicronutrients. However, inadequate intakes (i.e., <EAR) ofpotassium for 93% of participants and choline for >40% ofparticipants were observed.

(3)

Participants had excessive intakes (>UL) of chemicallyderived folic acid and magnesium from medical foods, and>63% of participants had excessive intakes of sodium drivenby natural (likely processed) food intake. Average sugarintake as a percentage of energy was 27% and was excessive(>DGA) for 97% of participants.

(4)

Without micronutrient supplementation of medical foods,>70% of participants would have inadequate intakes (≤EAR)for 11 micronutrients (biotin, choline, pantothenate,vitamins D and E, potassium, calcium, iodine, magnesium,selenium, and zinc). Greater than 90% of participants wouldobtain adequate intake (≥EAR) of vitamin A from naturalfoods alone due to high intakes of provitamin A carotenoidsfrom green leafy vegetables, squashes, carrots, and tomatoes.

(5)Of 30 participants, only 13 consumed MLPF with both AA-MF and GMP-MF treatments. MPLF comprisedapproximately 10% of median calories.

(6)

Increased urinary excretion of sulfate, taurine, and taurine-related metabolites with AA-MFmay be related to increasedneed to excrete sulfur with higher dietary intake of sulfur-containing amino acids, Met and Cys, from AA-MFcompared with Glytactin GMP-MF.

14 Journal of Nutrition and Metabolism

and signaling of cell membranes, and stem cell proliferationand apoptosis, which can impact brain structure andfunction, particularly during development [31]. Choline canbe obtained from high-protein dietary sources, such as meat,eggs, beans, and nuts, but can also be made via de novosynthesis from phosphatidylcholine in the liver [31]. Con-sidering that most sources of choline are in high-proteinfoods, individuals with PKU must rely on de novo synthesisand supplementation of medical foods with choline.

Choline intakes were below the EAR in >40% of par-ticipants, though participants obtained 1.55-fold greaterintakes of dietary choline with Glytactin GMP-MF com-pared to AA-MF. It was not surprising to 8nd no diJer-ences in plasma choline concentration, given its homeostaticcontrol. However, choline appears to be a nutrient of concernin participants with classical PKU as evidenced by the lowerplasma betaine concentrations compared to controls and lowerdimethylglycine concentrations compared to variant PKU.Given the challenge of evaluating choline status with plasmabiomarkers (choline, betaine, and dimethylglycine) and theunderestimated DRI for dietary choline, it is reasonable tosuggest that choline is a nutrient of concern in PKU based onthe low dietary choline intakes of our participants [31].

4.4. Strengths and Limitations. Strengths of this study in-clude a crossover study design, metabolomic analyses ofboth plasma and urine samples, and a comprehensive dietanalysis that consisted of macronutrients and 25 vitaminsand minerals. Furthermore, unlike the nutrition interven-tion studies with GMP-MF in patients with PKU, an addi-tional strength of this study is that GMP-MF contributed to95–100% of medical food intake [21, 45]. Limitations of thisstudy include a short dietary treatment with GMP-MF thatconsisted of 3 weeks and inclusion of 15 diJerent AA-MF toaccommodate participant preference.

5. Conclusions

*is is the 8rst study to utilizemetabolomic analysis of plasmaand urine samples and employ a comprehensive diet analysisthat estimated 25 vitamins and minerals from medical foodsand natural foods to investigate the nutrient status of adultsand adolescents with classical and variant PKU consumingAA-MF and Glytactin GMP-MF. Our results are pertinent toearly treated adults and adolescents with PKU who are ad-herent to the low-Phe diet in combination withmedical foods.We identi8ed several nutrients of concern related to in-adequate intakes (potassium and choline) or excessive intakes(sodium, magnesium, folic acid, and possibly vitamin A). Ourdata demonstrate that without vitamin and mineral supple-mentation of medical foods, more than 70% of participantswould have inadequate intakes for 11 vitamins and minerals(biotin, choline, pantothenate, vitamins D and E, potassium,calcium, iodine, magnesium, selenium, and zinc). *us,nutrient status in PKU relies on compliance with medicalfoods supplemented with vitamins and minerals. Regardless,there are ongoing challenges related to nutrient de8ciency andtoxicity in PKU due to the minimal intake of natural foods

that contain protein in the low-Phe diet, the bioavailability ofchemically derived micronutrients, and the diverse micro-nutrient supplementation pro8les of medical foods. Moreresearch is necessary to determine optimal supplementa-tion needs of chemically derived micronutrients forindividuals with PKU across the lifespan.

Abbreviations

AA: Amino acidAA-MF: Amino acid medical foodsAI: Adequate intakeAMDR: Acceptable macronutrient distribution rangesBMD: Bone mineral densityBMI: Body mass indexDGA: Dietary Guidelines for AmericansDHA: Docosahexaenoic acidDRI: Dietary Reference IntakeEAR: Estimated average requirementsEPA: Eicosapentaenoic acidGMP-MF: Glycomacropeptide medical foodsMF: Medical foodsMLPF: Modi8ed low-protein foodMMA: Methylmalonic acidPAH: Phenylalanine hydroxylasePE: Protein equivalentRD: Registered dietitianUL: Tolerable upper intake level.

Ethical Approval

*e University of Wisconsin-Madison Health Sciences re-view board approved the study protocol.

Consent

All subjects provided written informed consent.

Disclosure

Cambrooke *erapeutics, Inc., donated the GMP medicalfoods used in this study but was not involved in the design orconduct of the study or in the collection, analysis, or in-terpretation of the data.

Conflicts of Interest

Denise M. Ney is a coinventor on U.S. Patent 8,604,168 B2,“Glycomacropeptide Medical Foods for Nutritional Man-agement of Phenylketonuria and Other Metabolic Disor-ders,” which is held by the Wisconsin Alumni ResearchFoundation and licensed to Cambrooke *erapeutics, LLC.Denise M. Ney is a consultant to Arla Foods Ingredients andAgropur. Fran Rohr has received consulting fees fromCambrooke *erapeutics.

Authors’ Contributions

Bridget M. Stroup participated in data collection, analysis,interpretation, and manuscript preparation. Denise M. Ney

Journal of Nutrition and Metabolism 15

participated in study design, data interpretation, andmanuscript preparation. Sangita G. Murali participated indata collection and analysis. Fran Rohr and Sally T. Gleasonparticipated in data collection and interpretation. Sandy C.van Calcar and Harvey L. Levy participated in study designand data interpretation. All authors read the 8nal version ofthe manuscript.

Acknowledgments

*e authors would like to thank Nivi Nair for her contri-butions to the data collection of this study and Dr. KatarzynaBroniowska from Metabolon, Inc., for assistance in in-terpretation of metabolomic analyses. *is work was sup-ported by the Department of Health and Human Servicesgrant R01 FD003711 from the FDA OXce of OrphanProducts Development to Ney, P30-HD-03352, and bythe Clinical and Translational Science Award (CTSA) pro-gram, through the NIH National Center for AdvancingTranslational Sciences (NCATS) grant UL1TR000427. Cam-brooke*erapeutics, Inc., and Arla Foods Ingredients donatedfunds for the metabolomic analyses.

Supplementary Materials

Supplemental Table 1: Plasma metabolomics in subjects withPKU consuming AA-MF or GMP-MF compared to controls.Supplemental Table 2: Plasmametabolomics in subjects withclassical versus variant PKU consuming AA-MF and GMP-MF compared to controls. Supplemental Table 3: Urinemetabolomics in subjects with classical and variant PKUconsuming AA-MF and GMP-MF. Supplemental Table 4:Laboratory Measures. (Supplementary Materials)

References

[1] M. I. Flydal and A. Martinez, “Phenylalanine hydroxylase:function, structure, and regulation,” IUBMB Life, vol. 65,no. 4, pp. 341–349, 2013.

[2] J. Vockley, H. C. Andersson, K. M. Antshel et al., “Phenyl-alanine hydroxylase de8ciency: diagnosis and managementguideline,” Genetics in Medicine, vol. 16, no. 2, pp. 188–200,2014.

[3] D. A. Bilder, J. A. Kobori, J. L. Cohen-PfeJer, E. M. Johnson,E. R. Jurecki, and M. L. Grant, “Neuropsychiatric comor-bidities in adults with phenylketonuria: a retrospective cohortstudy,” Molecular Genetics and Metabolism, vol. 121, no. 1,pp. 1–8, 2017.

[4] R. H. Singh, F. Rohr, D. Frazier et al., “Recommendations forthe nutrition management of phenylalanine hydroxylase de-8ciency,” Genetics in Medicine, vol. 16, no. 2, pp. 121–131, 2014.

[5] M. R. Etzel, “Manufacture and use of dairy protein fractions,”Journal of Nutrition, vol. 134, no. 4, pp. 996S–1002S, 2004.

[6] C. E. Laclair, D. M. Ney, E. L. MacLeod, and M. R. Etzel,“Puri8cation and use of glycomacropeptide for nutritionalmanagement of phenylketonuria,” Journal of Food Science,vol. 74, no. 4, pp. E199–E206, 2009.

[7] S. C. van Calcar, E. L. MacLeod, S. T. Gleason et al., “Improvednutritional management of phenylketonuria by using a dietcontaining glycomacropeptide compared with amino acids,”American Journal of Clinical Nutrition, vol. 89, no. 4,pp. 1068–1077, 2009.

[8] D. M. Ney, B. M. Stroup, M. K. Clayton et al., “Glyco-macropeptide for nutritional management of phenylke-tonuria: a randomized, controlled, crossover trial,”American Journal of Clinical Nutrition, vol. 104, no. 2,pp. 334–345, 2016.

[9] R. H. Singh, A. C. Cunningham, S. Mo8di et al., “Updated,web-based nutrition management guideline for PKU: anevidence and consensus based approach,” Molecular Geneticsand Metabolism, vol. 118, no. 2, pp. 72–83, 2016.

[10] A. M. Lammardo, M. Robert, J. C. Rocha et al., “Main issuesin micronutrient supplementation in phenylketonuria,”Molecular Genetics and Metabolism, vol. 110, pp. S1–S5,2013.

[11] J. J. Jans, M. G. M. de Sain-van der Velden, P. M. van Hasseltet al., “Supplementation with a powdered blend ofPUFAs normalizes DHA and AA levels in patients withPKU,” Molecular Genetics and Metabolism, vol. 109, no. 2,pp. 121–124, 2013.

[12] S. A. Tanumihardjo, R. M. Russell, C. B. Stephensen et al.,“Biomarkers of nutrition for development (BOND)-vitaminA review,” Journal of Nutrition, vol. 146, no. 9, pp. 1816S–1848S, 2016.

[13] P. Borel, “Factors aJecting intestinal absorption of highlylipophilic food microconstituents (fat-soluble vitamins, ca-rotenoids and phytosterols),” Clinical Chemistry and Labo-ratory Medicine, vol. 41, no. 8, pp. 979–994, 2003.

[14] S. A. Jakeman, C. N. Henry, B. R. Martin et al., “Soluble corn8ber increases bone calcium retention in postmenopausalwomen in a dose-dependent manner: a randomized crossovertrial,” American Journal of Clinical Nutrition, vol. 104, no. 3,pp. 837–843, 2016.

[15] C. M. Whisner, B. R. Martin, C. H. Nakatsu et al., “Solublecorn 8ber increases calcium absorption associated with shiftsin the gut microbiome: a randomized dose-response trial infree-living pubertal females,” Journal of Nutrition, vol. 146,no. 7, pp. 1298–1306, 2016.

[16] S. J. Fairweather-Tait, R. Collings, and R. Hurst, “Seleniumbioavailability: current knowledge and future research re-quirements,” American Journal of Clinical Nutrition, vol. 91,no. 5, pp. 1484S–1491S, 2010.

[17] V. E. Ohrvik, B. E. Buttner, M. Rychlik, E. Lundin, andC. M. Witthoft, “Folate bioavailability from breads and a mealassessed with a human stable-isotope area under the curve andileostomy model,” American Journal of Clinical Nutrition,vol. 92, no. 3, pp. 532–538, 2010.

[18] S. Evans, A. Daly, J. MacDonald et al., “*e micronutrientstatus of patients with phenylketonuria on dietary treatment:an ongoing challenge,” Annals of Nutrition and Metabolism,vol. 65, no. 1, pp. 42–48, 2014.

[19] S. Demirdas, F. J. van Spronsen, C. E. M. Hollak et al.,“Micronutrients, essential fatty acids and bone health inphenylketonuria,” Annals of Nutrition and Metabolism,vol. 70, no. 2, pp. 111–121, 2017.

[20] I. Vugteveen, M. Hoeksma, A. L. BjorkeMonsen et al., “Serumvitamin B12 concentrations within reference values do notexclude functional vitamin B12 de8ciency in PKU patients ofvarious ages,” Molecular Genetics and Metabolism, vol. 102,no. 1, pp. 13–17, 2011.

[21] A. Pinto, M. F. Almeida, P. C. Ramos et al., “Nutritional statusin patients with phenylketonuria using glycomacropeptide astheir major protein source,” European Journal of ClinicalNutrition, vol. 71, no. 10, pp. 1230–1234, 2017.

[22] K. Wilson, P. Charmchi, and B. Dworetzsky, State Statutesand Regulations on Dietary Treatments of Disorders Identi3ed

16 Journal of Nutrition and Metabolism

?rough Newborn Screening, pp. 1–80, American College ofMedical Genetics & Genomics: Catalyst Center, Center forAdvancing Health Policy & Practice, Boston University,Bethesda, MD, USA, 2016.

[23] Institute of Medicine, Dietary Reference Intake for Energy,Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein andAmino Acids (Macronutrients), *e National AcademiesPress, Washington D.C., USA, 2005.

[24] U.S. Department of Health and Human Services and U.S. Department of Agriculture, Dietary Guidelines forAmericans 2015–2020, 8th edition, 2015, http://health.gov/dietaryguidelines/2015/guidelines/.

[25] Institute of Medicine, Dietary Reference Intakes: ?e EssentialGuide to Nutrient Requirements, *e National AcademiesPress, Washington, D.C., USA, 2006.

[26] Institute of Medicine, Dietary Reference Intakes for Calciumand Vitamin D, *e National Academies Press, WashingtonD.C., USA, 2011.

[27] D. M. Ney, S. G. Murali, B. M. Stroup et al., “Metabolomicchanges demonstrate reduced bioavailability of tyrosine andaltered metabolism of tryptophan via the kynurenine pathwaywith ingestion of medical foods in phenylketonuria,” Mo-lecular Genetics and Metabolism, vol. 121, no. 2, pp. 96–103,2017.

[28] B. M. Stroup, Nutritional management of phenylketonuriawith glycomacropeptide medical foods, ProQuest, order no.10604680, Ann Arbor, MI, USA, 2017.

[29] B. M. Stroup, S. G. Murali, N. Nair et al., “Dietary amino acidintakes associated with a low-phenylalanine diet combinedwith amino acid medical foods and glycomacropeptidemedical foods and neuropsychological outcomes in subjectswith phenylketonuria,” Data in Brief, vol. 13, pp. 377–384,2017.

[30] B. M. Stroup, E. A. Sawin, S. G. Murali, N. Binkley,K. E. Hansen, and D. M. Ney, “Amino acid medical foodsprovide a high dietary acid load and increase urinary ex-cretion of renal net acid, calcium, and magnesium comparedwith glycomacropeptide medical foods in phenylketonuria,”Journal of Nutrition and Metabolism, vol. 2017, Article ID1909101, 12 pages, 2017.

[31] S. H. Zeisel, “Choline: critical role during fetal developmentand dietary requirements in adults,” Annual Review of Nu-trition, vol. 26, pp. 229–250, 2006.

[32] WHO, “Protein and amino acid requirements in humannutrition,” Tech. Rep. 935, United Nations University, WorldHealth Organization, Geneva, Switzerland, 2007.

[33] J. E. Dominy Jr., J. Hwang, S. Guo, L. L. Hirschberger,S. Zhang, and M. H. Stipanuk, “Synthesis of amino acidcofactor in cysteine dioxygenase is regulated by substrate andrepresents a novel post-translational regulation of activity,”Journal of Biological Chemistry, vol. 283, no. 18, pp. 12188–12201, 2008.

[34] N. M. Lowe, K. Fekete, and T. Decsi, “Methods of assessmentof zinc status in humans: a systematic review,” AmericanJournal of Clinical Nutrition, vol. 89, no. 6, pp. 2040S–2051S,2009.

[35] M. Robert, J. C. Rocha, M. van Rijn et al., “Micronutrientstatus in phenylketonuria,” Molecular Genetics and Meta-bolism, vol. 110, pp. S6–S17, 2013.

[36] A. MacDonald, J. C. Rocha, M. van Rijn, and F. Feillet,“Nutrition in phenylketonuria,” Molecular Genetics andMetabolism, vol. 104, pp. S10–S18, 2011.

[37] R. M. Fisberg, M. E. da Silva-Femandes, M. Fisberg, andB. Jose Schmidt, “Plasma zinc, copper, and erythrocyte

superoxide dismutase in children with phenylketonuria,”Nutrition, vol. 15, no. 6, pp. 449–452, 1999.

[38] D. Dobbelaere, L. Michaud, A. Debrabander et al., “Evalua-tion of nutritional status and pathophysiology of growthretardation in patients with phenylketonuria,” Journal ofInherited Metabolic Disease, vol. 26, no. 1, pp. 1–11, 2003.

[39] V. Crujeiras, L. Aldamiz-Echevarria, J. Dalmau et al., “Vi-tamin and mineral status in patients with hyper-phenylalaninemia,” Molecular Genetics and Metabolism,vol. 115, no. 4, pp. 145–150, 2015.

[40] F. Andrade, O. Lopez-Suarez, M. Llarena, M. L. Couce, andL. Aldamiz-Echevarrıa, “InUuence of phenylketonuria’s dieton dimethylated arginines and methylation cycle,” Medicine,vol. 96, no. 27, p. e7392, 2017.

[41] S. F. Dobrowolski, J. Lyons-Weiler, K. Spridik et al., “AlteredDNA methylation in PAH de8cient phenylketonuria,”Molecular Genetics and Metabolism, vol. 115, no. 2-3,pp. 72–77, 2015.

[42] S. F. Dobrowolski, J. Lyons-Weiler, K. Spridik, J. Vockley,K. Skvorak, and A. Biery, “DNA methylation in the patho-physiology of hyperphenylalaninemia in the PAHenu2

mouse model of phenylketonuria,” Molecular Genetics andMetabolism, vol. 119, no. 1-2, pp. 1–7, 2016.

[43] C. Colome, R. Artuch, M. A. Vilaseca et al., “Lipophilic an-tioxidants in patients with phenylketonuria,” AmericanJournal of Clinical Nutrition, vol. 77, no. 1, pp. 185–188, 2003.

[44] K. H. Schulpis, S. Tsakiris, G. A. Karikas, M. Moukas, andP. Behrakis, “EJect of diet on plasma total antioxidant statusin phenylketonuric patients,” European Journal of ClinicalNutrition, vol. 57, no. 2, pp. 383–387, 2003.

[45] A. Daly, S. Evans, S. Chahal, S. Santra, and A. MacDonald,“Glycomacropeptide in children with phenylketonuria:does its phenylalanine content aJect blood phenylalaninecontrol?,” Journal of Human Nutrition and Dietetics, vol. 30,no. 4, pp. 515–523, 2017.

Journal of Nutrition and Metabolism 17

Research ArticleGrowth Patterns in the Irish Pyridoxine NonresponsiveHomocystinuria Population and the Influence of MetabolicControl and Protein Intake

Orla Purcell,1 Aoife Coughlan,2 Tim Grant,2 Jenny McNulty,1

Anne Clark,1 Deirdre Deverell,3 Philip Mayne,3 Joanne Hughes,1

AhmadMonavari,1 Ina Knerr,1 and Ellen Crushell1

1National Centre for Inherited Metabolic Disorders, Temple Street Children’s University Hospital, Dublin, Ireland2Department of Research, Temple Street Children’s University Hospital, Dublin, Ireland3Department of Laboratory Medicine, Temple Street Children’s University Hospital, Dublin, Ireland

Correspondence should be addressed to Orla Purcell; orla.purcell@cuh.ie

Received 13 June 2017; Accepted 11 September 2017; Published 15 November 2017

Academic Editor: Pedro Moreira

Copyright © 2017 Orla Purcell et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

A low methionine diet is the mainstay of treatment for pyridoxine nonresponsive homocystinuria (HCU). There are variousguidelines for recommended protein intakes for HCU and clinical practice varies. Poor growth has been associated with low cystinelevels.This retrospective reviewof 48 Irish pyridoxine nonresponsiveHCUpatients assessedweight, height, bodymass index (BMI),protein intake, and metabolic control up to 18 years at nine set time points. Patients diagnosed through newborn screening (NBS)were compared to late diagnosed (LD) patients. At 18 years the LD group (𝑛 = 12, mean age at diagnosis 5.09 years) were heavier(estimated effect +4.97Kg, 𝑃 = 0.0058) and taller (estimated effect +7.97 cm 𝑃 = 0.0204) than the NBS group (𝑛 = 36). Therewas no difference in growth rate between the groups after 10 years of age. The HCU population were heavier and taller than thegeneral population by one standard deviation with no difference in BMI. There was no association between intermittently lowcystine levels and height. Three protein intake guidelines were compared; there was no difference in adult height between thosewho met the lowest of the guidelines (Genetic Metabolic Dietitians International) and those with a higher protein intake.

1. Introduction

Classical homocystinuria (HCU) (OMIM: 236200) is a reces-sively inherited disorder of methionine metabolism causedby inactivating mutations in the gene encoding the enzymecystathionine 𝛽-synthase (CBS) (EC 4.2.1.22), leading todeficiency of the enzyme activity. CBS is necessary in thecatalysis of methionine to cysteine; this pathway requirespyridoxine (vitamin B6), vitamin B12, and folic acid ascofactors. Deficiency of CBS results in the accumulation ofhomocysteine and methionine along with a lack of cysteineand cystine, the oxidised dimer form. There are two typesof HCU: a “milder” form which responds to pyridoxine(pyridoxine-responsive HCU) and a more severe pyridoxinenonresponsive form.The incidence ofHCUvaries in differentpopulations; due to a high prevalence of the G307S mutation

in the CBS gene the incidence of pyridoxine nonresponsiveHCU in Ireland is approximately 1 in 65,000 [1–4]. Newbornscreening (NBS) for HCU was added to the Irish NationalNewborn Bloodspot screening programme in 1971.

The natural history of HCU was described by Muddand colleagues in 1985 [5]. They reported that untreatedpyridoxine nonresponsive HCU resulted in various ocular,vascular, skeletal, and central nervous system complications.They also described patients as being tall and lean [5]. Yapand Naughten [4] observed that maintaining lifetime plasmafree homocysteine (fHcy) ≤ 11 𝜇mol/L (which approximatesa total homocysteine (tHcy) of 100–120mmol/L) protectsagainst these well-recognised complications [6]. PlasmafHcy, tHcy, methionine (Met), and cystine (Cys) are moni-tored and tHcy and fHcy levels are lowered to within the rec-ommended therapeutic range by dietary restriction of natural

HindawiJournal of Nutrition and MetabolismVolume 2017, Article ID 8570469, 7 pageshttps://doi.org/10.1155/2017/8570469

2 Journal of Nutrition and Metabolism

protein/methionine. Supplementation with methionine-freecystine supplemented synthetic amino acid mixtures is partof the daily treatment. Lifelong adherence to this low proteindiet is recommended but is challenging and complianceissues are often encountered [7, 8].

While restriction of methionine is crucial to achievegood metabolic control, providing sufficient amounts of totalprotein for optimal growth and health is also needed andclose dietetic monitoring is required to achieve this [1–8].There are three widely referenced guidelines which outline arecommended total protein intake for patients with inheriteddisorders of protein metabolism. These guidelines include(1) the Great Ormond Street (GOS) guidelines for inbornerrors ofmetabolism [9], (2) theGeneticMetabolic DietitiansInternational (GMDI) nutrition guidelines [10] which arespecific to Phenylketonuria but are widely used for otherprotein disorders, and (3) the Ross Guidelines which arespecific to HCU [11].

Case reports suggesting low Cys levels can cause poorweight gain and growth despite adequate calories have led tothe hypothesis that maintaining Cys levels within the normalrange is essential for growth [12, 13]. It is also recognised thatelevated homocysteine concentrations are associated withlowCys levels. An increase inCys concentrationwas reportedto lead to a significant reduction in fHcy [14].

This retrospective study aims to examine the growth rateof Irish HCU patients from birth to 18 years, including thosewho were diagnosed through newborn bloodspot screening(NBS) and those who were late diagnosed (LD) before 18years of age.We aim to examine possible associations betweenmetabolic control (fHcy, Met, and Cys concentrations) andgrowth and the influence of varying protein intake recom-mendations on final adult height.

2. Patients and Methods

Ethical approval was received from the Temple Street Chil-dren’s University Hospital Research and Ethics Committee,reference number 16.020. Retrospective data were collectedfrom medical and dietetic notes and anonymised for furtheranalysis. Parameters for metabolic control included plasmalevels of fHcy, tHcy, Met, and Cys. Growth parametersincluded weight, height, body mass index (BMI), and mid-parental height. Dietetic assessments included the prescribedtotal protein intake including natural protein and prescribedsynthetic protein intake which was calculated from dieteticnotes.

2.1. Study Sample. All HCU patients who attend or had pre-viously attended the National Centre for Inherited MetabolicDisorders (NCIMD) in Temple Street Children’s UniversityHospital, Dublin, Ireland.TheNCIMD is the national referralcentre for children with HCU in Ireland. Patients who hadpyridoxine-responsive HCU (𝑛 = 3) and those who werediagnosed with pyridoxine nonresponsiveHCU after 18 yearsof age (𝑛 = 1) were excluded.

2.2. Newborn Screening. Blood samples are taken between 72and 120 hours of life. Since 2010, the cut-off for Met levels

using tandem mass spectrometry has been 50𝜇mol/L; priorto this it was 60 𝜇mol/L. Those not diagnosed through NBSare referred to as late diagnosed (LD).

2.3. Data Collection. Patients were categorised as those diag-nosed through NBS or those who were LD. There werenine set time points for data collection: three months, sixmonths, nine months, twelve months, two years, four years,ten years, fourteen years, and eighteen years of age. Thesetime points were selected as the best markers of growth asthey provided a continuous log or pattern of growth throughinfancy, childhood, and the teenage years with emphasison the rapid stages of growth which occur during infancy.The data for the LD group were collected at these specifictime points once diagnosis was made. Serial measurementsof fHcy, tHcy, Met, and Cys were recorded at each of thespecified time points. Visits are scheduled such that bloodsamples were taken 3-4 hours after their last meal, that is,usually preprandial. The prescribed total protein intake andmeasurements for weight, length/height, andBMIwere notedat each of the specified time points also. Data on medicationsprescribed such as Betaine were not included in the studydesign as Betaine is rarely used in our paediatric cohort asdietetic compliance is generally good and Betaine is reservedfor those with poor dietetic compliance, occasionally in lateadolescence.

The HCU population was compared to a non-HCUpopulation using standard deviation scores (SDS) calculatedusing the LMSgrowth© Excel add-in programme [15] andcompared with the British 1990 and UK-WHO growthreference data [16, 17]. Parental heights were either recordedfrom the patient records or taken at clinic visits and usedto calculate midparental height. Midparental height is theaverage of both parents’ heights, plotted on the appropriateheight centile chart at 18 years of age after adjustment for sex.

2.4. Analytical Methods for Blood Sampling. Amino acidanalysis of tHcy, fHcy, Met, and Cys was performed usingIon Exchange Column Chromatography with ninhydrindetection using an automatic amino acid analyser (theJEOL AminoTAC analyser, JEOL Croissy-sur-Seine, France).Lithium heparin is the preferred sample type. Separation ofplasma from red cells was performed within fifteen minutesof draw and an aliquot of plasma then deproteinised using10% by volume of 35% sulphosalicylic acid.The deproteinisedsupernatantwas used formeasurement of fHcy, Cys, andMet.Neat plasma was reduced by addition of 12% dithiothreitolprior to deproteinisation for measurement of tHcy. tHcy wasmeasured less frequently than the other markers as it was notroutinely measured in our laboratory until mid-1990s.

2.5. Reference Ranges and Guidelines Used. Biochemicalreferences ranges used to categorise metabolic control aredescribed in Tables S1 and S2 in Supplementary Materialavailable online at https://doi.org/10.1155/2017/8570469. Thetotal protein intake recommended by the GOS, GMDI, andRoss guidelines [9–11] is summarised in Table S3.

2.6. Statistical Analysis. Statistical analyses were conductedusing the IBM Statistical Package for the Social Sciences

Journal of Nutrition and Metabolism 3

(SPSS; Version 23, 2014, IBM, Armonk, NY). Independentcategorical variables were compared using Chi squaredanalysis and both independent sample T-tests were usedto compare continuous variables. A linear mixed effectsmodel was the primary analysis comparing the populationdiagnosed through NBS with the LD group. Age and diag-nosis were fixed factors in the model with between subjectmeasurements treated as a random effect to account for therepeated measures aspect of the study.These models are builtin R version 3-2-2 (R Core Team, 2013, Vienna, Austria). A 𝑃value < 0.05 was accepted for significance. For establishingthe impact of other covariates on the growth process, thevariables are added to the linear mixed effects model and, ifsignificant, it can be concluded that some of the variabilityin size is explained by the additional covariate. Fisher’s ExactTest was used as appropriate.

3. Results

3.1. Demographics. Data was collected on 48 pyridoxinenonresponsive HCU patients: 24 female (50%) and 24 male(50%). Thirty-six (75%) patients were diagnosed throughNBS with 12 LD patients (25%). Current mean age in thegroup of patients diagnosed through NBS was 23.70 years(range 5 months to 43 years).The current mean age in the LDgroup was 34.64 years (range 11.97–52.59 years) with a meanage at diagnosis of 5.09 years (range 1.33–11.79 years).

3.2. Growth within the HCU Cohort. Overall, HCU patientshad a similar birth weight compared to the general popu-lation. The mean birth weight of the male NBS cohort wasslightly higher at 3.71 kg (𝑛 = 14); the late diagnosed groupwas 3.37 kg (𝑛 = 4) compared to 3.55 Kg for the generalpopulation.There was no difference in the mean birth weightfor the female cohort which was 3.33 Kg for those diagnosedthrough NBS (𝑛 = 19), 3.36 Kg for the LD group (𝑛 = 3), and3.40Kg for the general population.

There was a significant difference in weight and heightreaching at 18 years between those diagnosed on NBS andthose LD. LD patients were both heavier and taller, with aweight estimated effect of +4.97 Kg (𝑃 = 0.0058) and heightestimated effect of +7.97 cm (𝑃 = 0.0204). However, thegrowth rate between both groups beyond 10 years of ageshowed no significant difference. Throughout the course ofthe study there was an annual average increase in heightof 5.95 cm for the NBS group compared to 6.85 cm for theLD group (𝑃 = 0.1651). Furthermore, there was an annualaverage increase in weight of 3.58 Kg from 10 to 18 years of agefor those diagnosed through NBS compared to 4.1 Kg for theLD group (𝑃 = 0.621).TheLDgroup had a slightly lower BMI(estimated effect = −0.36, 𝑃 = 0.4636) which also increasedat a slower rate from 10 years of age as illustrated in Figures1(a)–1(c).

The difference in growth between the two groupswas independent of metabolic control. When adjusted formetabolic control from ten years of age, there was nosignificant difference found (e.g., for Met 𝑃 = 0.2098, forfHcy 𝑃 = 0.2396, and for Cys 𝑃 = 0.2432) at the ninespecific age points.Therewas no significant association found

between intermittently low Cys levels with height SDS (𝑃 =0.9377).

3.3. Growth of the HCU Population Compared to the GeneralPopulation. Those diagnosed through NBS and those LDwere heavier and taller compared to the general populationat nearly all ages; however, the NBS group were closer to thegeneral population in terms of weight and height comparedto the LD group; these are described in Tables S4a andS4b in the Supplementary Material. There was no significantdifference in BMI between the groups suggesting a balancedincrease inweight and height.Thiswas observed in both sexes(Tables 2(a)–2(f), 4(a), and 4(b) and Figures 1(a)–1(f) in theSupplementary Material). The weight and height SDS graphsfor both the male and female populations are illustrated inFigures 2(a)–2(f).

3.4. Mid Parental Height and Predicted Adult Height.Seventy-seven percent of LD (𝑛 = 10) patients and 75% ofthose diagnosed through NBS (𝑛 = 25) grew within theirexpected midparental height range. Two patients (15.4%) inthe LD group and 6 (16.7%) of those diagnosed throughNBS grew taller than their midparental height range. Noneof those LD and 2 (5.6%) of those diagnosed through NBSwere smaller than their midparental height range at 18 years.Of those diagnosed throughNBS, 7 (39%) grew as expected, 7(39%) fell short, and 4 (24%) were taller.

3.5. Metabolic Control. Table 1 provides a summary of bio-chemical parameters taken at the defined time points inthe HCU cohort. There was no significant difference inmetabolic control achieved after diagnosis between the twoHCU groups. The LD group had a lower number of samplesrecorded and the number of samples per patient was naturallyskewed by their age at diagnosis.

3.6. Protein Requirements and Intake. Table 2 identifies thepercentage of patients who achieved their full estimatedprotein requirement as per dietetic assessment at the specifiedage points according to the three guidelines which werecompared. Using Fisher’s Exact Test for difference, significantdifferences were found to exist between the numbers ofpatients who were achieving their estimated requirements atthe various age points. However, for both the LD group andthose diagnosed through NBS, regardless of whether theywere meeting their estimated protein requirements or not,there was no significant difference in height at any age point(𝑃 = 0.5458). The highest percentage of patients met theGMDI recommended protein intake (which is lower thanGOS and Ross).

4. Discussion

This is the largest study to date examining the growth patternof children with HCU alongside their metabolic control.It highlights a significant difference in the way LD HCUpatients grow in comparison to those diagnosed throughNBS. At all ages, the LD patients were taller and heavier thanthose patients who had been diagnosed through NBS but this

4 Journal of Nutrition and Metabolism

Table 1: Metabolic control given as homocyst(e)ine, methionine, and cystine concentrations under treatment (𝜇mol/L).

Free Homocystine Total homocysteine Methionine CystineNBS LD NBS LD NBS LD NBS LD

Number of samples 234 25 104 8 237 25 223 24Mean 15.1 15.7 91.1 88.5 72.6 119.6 39 43Median 9.5 5 79 71.5 62.5 99.2 31.5 35.3Range 0–90 0–100 5–269 9–202 3–600 10–324 4–180 3–57P value P = 0.917 P = 0.933 P = 0.165 P = 0.660

0 50 100 150 200

0

20

40

60

80

Time (months)

Wei

ght (

kg)

LD groupNBS group

Estimated effect = 4.97 ＋Ａ

P = 0.0058

(a) Comparison of average weight gain (Kg) in NBS and LD HCUpatients from birth to 18 years

0 50 100 150 200

0

50

100

150

200

250

Leng

th (c

m)

Time (months)

LD groupNBS group

Estimated effect = 7.97 ＝Ｇ

P = 0.0204

(b) Comparison of average height gain (cm) in NBS andLD HCU patients from birth to 18 years

0 50 100 150 200

16

18

20

22

24

Time (months)

LD groupNBS group

Estimated effect = −0.36 KＡ/Ｇ2

P = 0.4636

BMI (

kg/－

2)

(c) Comparison of BMI (Kg/m2) in NBS and LD HCU patients frombirth to 18 years

Figure 1

accelerated growth occurred before 10 years suggesting itmaybe due to poor metabolic control in infancy and early child-hood prior to diagnosis. This correlates with descriptions ofuntreated or LDHCU patients as being taller and leaner witha “Marfanoid” appearance [1, 2, 18–20], reminiscent of thegenetic condition Marfan syndrome. However, our patients

are not lean; they have a similar BMI to the NBS group andthe general population indicating a balanced increase in bothheight and weight. The future adult height of a child who isgrowing normally with the correct nutritional intake can beapproximately predicted at 2 years of age [18, 21]. This wasnot observed within our NBS cohort where 39% (𝑛 = 7)

Journal of Nutrition and Metabolism 5

Table 2: Percentage of patients meeting the Great Ormond Street(GOS) and Ross and Genetic Metabolic Dietitians International(GMDI) protein recommendations.

Age GOS Ross GMDI3 months 70 70 856 months 52 56 859 months 54 75 8812 months 54 73 1002 years 88 96 1004 years 61 96 10010 years 50 100 10014 years 90 100 10018 years 96 100 100

P values using Fisher’s Exact Test for differenceAge GOS to Ross GOS to GMDI Ross to GMDI3 months — .004 .0046 months — .041 .0289 months .003 .082 .01012 months .001 — —2 years .120 — —4 years .393 — —

grew within their predicted range, 22% (𝑛 = 4) grew taller,and 39% (𝑛 = 7) grew less than predicted at 2 years of age.Of note, all of our LD patients are treated aggressively withdietetic treatment and biochemical targets are the same forboth NBS and LD groups. Following diagnosis, the metaboliccontrol achieved was comparable between both groups. Thishighlights the importance of encouraging LD patients toinitiate a protein restricted diet despite the challenges itpresents. The benefits of good long term metabolic controlare well established even when the diagnosis is made late inlife [3, 4].

It has been hypothesised that low Cys levels lead to poorgrowth independent of calorie and protein intake [12, 13].In this study we did not find any significant associationbetween intermittently low Cys levels and height, in eitherthe NBS or the LD groups. Gastrointestinal side effects area frequent clinical finding following supplementation with L-cystine, a sulphur amino acid. AlthoughCys supplementationis not routine practice internationally, it is sometimes usedfor patients when Cys concentrations are persistently lowdespite adequate protein supply. It has been our experiencethat compliance with Cys supplementation is difficult dueto the gastrointestinal side effects. If plasma homocysteineconcentrations remain well controlled and if protein storesand growth parameters are of no concern but cystine levelsare below the reference range, extra supplementation mayoffer little benefit.

As there are various guidelines for recommended proteinintake formetabolic patients, the approaches used in differentcentres can vary. Protein requirements for patients withdisorders of sulphur amino acid metabolism have not beenindividually studied so general recommendations are oftenused [22]. When the majority of protein sources are supplied

as L-amino acids, there is rapid absorption and catabolismof the amino acid with a possible decrease of biologicalvalue. A higher total protein intake beyond age specifiedrecommendations has therefore been recommended [23–25]. Some guidelines are condition specific; for example,Ross has published HCU specific guidelines [11], while otherguidelines used are common to different amino acid-relateddisorders, for example, GOS and GMDI guidelines [9, 10].These guidelines differ with GOS suggesting a much higherdaily total protein intake thanGMDI (Table 2).We found thatsignificantly fewer patients achieved the GOS requirementscompared to GMDI and Ross recommendations (wherethe suggested protein requirements are lower); however,growth was comparable irrespective of the guideline applied.Compliance with synthetic amino acid mixtures can bechallenging in different age groups. Our patients had amedian BMI SDS of up to 1.4 from two to eighteen yearsof age; therefore, using the lower requirements should beconsidered a reasonable option to help reduce the amount ofsynthetic protein prescribed and reduce unnecessary calorieintake. Synthetic protein intake in isolationmay not be impli-cated in excessive weight gain; however, parents/patientsoften struggle to achieve the recommended dose and, now,knowing that the lower of the recommended guidelines doesnot negatively alter growth is reassuring for these patients.Further prospective studies into this area are warranted andmay provide a change in practice if similar results are found.

In our NBS cohort it was more difficult to predict adultheight at two years of age compared with the general popu-lation. A possible combination of different factors includingenvironmental, genetic, hormonal, general nutrition related,and unidentified causes may be potential explanations. Thiswarrants further investigation.

Recently published international HCU guidelines byMorris et al. [6] support the findings of this study. Thereis no evidence to guide cystine supplementation dose andlittle evidence to support the benefit of supplementation,in terms of growth. There are also no specific recommen-dations for synthetic or total protein intake for the HCUpopulation. Although compliance with a protein restricteddiet and methionine free L-amino acid supplement remainsa lifelong challenge, good compliance and good metaboliccontrol are essential for the prevention of complications.These guidelines also highlight the need for international col-laborative studies to expand treatment options and improveour patients’ health and quality of life.

5. Conclusions

In conclusion, the LD HCU cohort was significantly heavierand taller than the NBS cohort due to accelerated earlygrowth but their rate of growth after 10 years of age was thesame. No association was found between intermittently lowCys levels and poor growth. We conclude that achieving ahigher total protein intake offers no added benefit in termsof adult height reached and patients may benefit from a lowerdose of synthetic protein in terms of compliance and calorieintake.

6 Journal of Nutrition and Metabolism

Additional Points

Suggestions. (1) If homocysteine control is good despite beingconsistently low Cys levels further supplementation withcystine may not be necessary. (2) Consider reference to morethan one guideline for protein requirements as followinga lower total protein requirement guideline may improvecompliance.

Disclosure

This work was presented in abbreviated form at the Interna-tional Congress of Inborn Errors of Metabolism (ICIEM) inSeptember 2017.

Conflicts of Interest

The authors declare that there are no conflicts of interestregarding the publication of this article but disclose the fol-lowing: Ellen Crushell and Orla Purcell received honoraria inthe past for lectures from Orphan Europe and Nutricia; OrlaPurcell, Anne Clark, and Jenny McNulty received travel andhotel expenses for attending relevant metabolic conferencesand meetings.

Acknowledgments

The authors wish to acknowledge the staff at the NCIMDand the Department of Laboratory Medicine, who paved theway for this research and, most importantly, the patients andfamilies of all the HCU patients in Ireland.

References

[1] S. Mudd, H. Levy, and F. Skovby, “Disorders of transulfuration,”inTheMetabolic and Molecular bases of Inherited Disease, C. R.Scriver, A. L. Baeudet,W. S. Sly, andD.Valle, Eds., pp. 1279–1327,McGraw-Hill, New York, NY, USA, 7th edition, 1995.

[2] H. Gan-Schreier, M. Kebbewar, J. Fang-Hoffmann et al., “New-born Population Screening for Classic Homocystinuria byDetermination of Total Homocysteine from Guthrie Cards,”Journal of Pediatrics, vol. 156, no. 3, pp. 427–432, 2010.

[3] S. Yap, Homocystinuria due to cystathionine 𝛽-synthase defi-ciency,” 2005, Orphanet encyclopaedia http://www.orpha.net/data/patho/BG/uk-CbS.pdf.

[4] S. Yap and E. Naughten, “Homocystinuria due to cystathionine𝛽-synthase deficiency in Ireland: 25 years’ experience of anewborn screened and treated population with reference toclinical outcome and biochemical control,” Journal of InheritedMetabolic Disease, vol. 21, no. 7, pp. 738–747, 1998.

[5] S. H. Mudd, F. Skovby, H. L. Levy et al., “The natural historyof homocystinura due to cystathionine 𝛽-synthase deficiency,”American Journal ofHumanGenetics, vol. 37, no. 1, pp. 1–31, 1985.

[6] A. A. M. Morris, V. Kozich, S. Santra et al., “Guidelines forthe diagnosis and management of cystathionine beta-synthasedeficiency,” Journal of Inherited Metabolic Disease, vol. 40, no. 1,pp. 49–74, 2017.

[7] S. Adam, M. F. Almeida, E. Carbasius Weber et al., “Dietarypractices in pyridoxine non-responsive homocystinuria: A

European survey,”Molecular Genetics and Metabolism, vol. 110,no. 4, pp. 454–459, 2013.

[8] A. Lawson-Yuen and H. L. Levy, “The use of betaine in thetreatment of elevated homocysteine,” Molecular Genetics andMetabolism, vol. 88, no. 3, pp. 201–207, 2006.

[9] Great Ormond Street Hospital for Children NHS FoundationTrust, Nutritional Requirements for Children in Health andDisease, 3rd edition, 2000.

[10] Southeast NBS &amp; Genetics Collaborative and GeneticMetabolic Dietitians International. PKU Nutrition Man-agement Guidelines, 1st edition, 2015. Retrieved from https://southeastgenetics.org/ngp/guidelines.php/90/PKU%20Nutrition%20Guidelines/Version%201.12.

[11] P. B. Acosta and S. Yannicelli, Nutrition Support Protocols: TheRoss Metabolic Formula System, Ross Products Division, AbbotLaboratories, Columbus, OH, 4th edition, 2001.

[12] T. L. Perry, H. G. Dunn, S. Hansen, L. MacDougall, and P.D. Warrington, “Early diagnosis and treatment of homocystin-uria.,” Pediatrics, vol. 37, no. 3, pp. 502–505, 1966.

[13] C. Sansaricq, S. Garg, P. M. Norton et al., “Cystine deficiencyduring dietotherapy of homocystinemia,” Acta Paediatrica, vol.64, pp. 215–218, 1975.

[14] P. J. Lee and A. Briddon, “A rationale for cystine supplementa-tion in severe homocystinuria,” Journal of Inherited MetabolicDisease, vol. 30, no. 1, pp. 35–38, 2007.

[15] H. Pan and T. J. Cole, 2012, LMS Growth, a MicrosoftExcel add-in to access growth references based on the LMSmethod. Version 2.77. http://www.healthforallchildren.com/product-category/shop/software/.

[16] J. V. Freeman, T. J. Cole, S. Chinn, P. R. M. Jones, E. M. White,and M. A. Preece, “Cross sectional stature and weight referencecurves for the UK, 1990,” Archives of Disease in Childhood, vol.73, no. 1, pp. 17–24, 1995.

[17] C. M. Wright, A. F. Williams, D. Elliman et al., “Using the newUK-WHO growth charts,” BMJ, vol. 340, no. mar15 1, pp. c1140–c1140, 2010.

[18] T. J. Cole, “Secular trends in growth,” Proceedings of theNutrition Society, vol. 59, no. 2, pp. 317–324, 2000.

[19] A. K. Elshorbagy, E. Nurk, C. G. Gjesdal et al., “Homo-cysteine, cysteine, and body composition in the HordalandHomocysteine Study: does cysteine link amino acid and lipidmetabolism?”American Journal of ClinicalNutrition, vol. 88, no.3, pp. 738–746, 2008.

[20] S. Yap, “Classical homocystinuria – newborn screening withearly treatment effectively prevents complications,” HamdanMedical Journal, vol. 5, no. 3, pp. 351–362, 2012.

[21] G. Cutberto, E. Borghi, A. W. Onyango et al., “Parental heightand child growth from birth to 2 years in theWHOMulticentreGrowth Reference Study,”Maternal and Child Nutrition, vol. 9,no. 2, pp. 58–68, 2013.

[22] P. B. Acosta, “Inherited Disorders of Sulfur Amino AcidMetabolism , in,” in Nutrition Management of Patients withInherited Metabolic Disorders.Jones and Barlett Publishers, LLC,Massachusetts, p. 252, 1st edition, 2010.

[23] S. S. Gropper and P. B. Acosta, “Effect of simultaneous ingestionof L-amino acids and whole protein on plasma amino acid andurea nitrogen concentrations in humans,” Journal of Parenteraland Enteral Nutrition, vol. 15, no. 1, pp. 48–53, 1991.

[24] S. S. Gropper, D. M. Gropper, and P. B. Acosta, “Plasma aminoacid response to ingestion of l-amino acids and whole protein,”

Journal of Nutrition and Metabolism 7

Journal of Pediatric Gastroenterology and Nutrition, vol. 16, no.2, pp. 143–150, 1993.

[25] J. L. Smith, C. Arteaga, and S. B. Heymsfield, “Increasedureagenesis and impaired nitrogen use during infusion of asynthetic amino acid formula. A controlled trial,” The NewEngland Journal of Medicine, vol. 306, no. 17, pp. 1013–1018, 1982.

Review ArticleMulticlinic Observations on the Simplified Diet in PKU

Laurie Bernstein,1 Casey Burns,1 Melissa Sailer-Hammons,1

Angela Kurtz,2 and Frances Rohr3

1 Inherited Metabolic Diseases Clinic, Children’s Hospital Colorado, Aurora, CO, USA2Metabolic Nutrition Program, Division of Medical Genetics, Icahn School of Medicine at Mount Sinai, New York, NY, USA3Department of Nutrition, Boston Children’s Hospital, Boston, MA, USA

Correspondence should be addressed to Casey Burns; casey.burns@childrenscolorado.org

Received 19 April 2017; Revised 10 July 2017; Accepted 25 July 2017; Published 13 September 2017

Academic Editor: Phillip B. Hylemon

Copyright © 2017 Laurie Bernstein et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Phenylketonuria is an inborn error of metabolism that historically has been treated with a strict phenylalanine-restricted dietwhere all foods are weighed and measured. This is cumbersome and difficult for patients and caregivers, especially patients withhigh phenylalanine blood concentrations who often have neurocognitive deficits.The SimplifiedDiet is an alternative approach thatallows for increased flexibility, promotes healthy food choices, and is easier to manage than a traditional diet for PKU. This paperdescribes the implementation of the Simplified Diet and outlines education, counseling strategies, and challenges encountered bythree metabolic clinics in the United States.

1. Introduction

Phenylketonuria (PKU) is an inborn error ofmetabolism thatcauses mutations in the phenylalanine hydroxylase (PAH)gene [1]. PAH catalyzes the conversion of phenylalanine intotyrosine with the help of the cofactor tetrahydrobiopterin(BH4) [2]. A deficiency in PAH causes the accumulation ofphenylalanine in the blood and other tissues, including thebrain, and results in intellectual disability if left untreated.The primary treatment for PKU is a lifelong diet thatconsists of a prescribed amount of phenylalanine and isaimed at keeping blood phenylalanine in the treatment rangeof 120–360 𝜇mol/L [3]. Patients with a severe or classicalform of PKU are prescribed approximately 250–300mg ofphenylalanine per day [4], which is equivalent to about 5-6 grams of protein from food. Allowed foods include fruits,vegetables, fats and oils, sugars, and modified low proteinfoods (e.g., low protein bread, pasta). High protein foodssuch as meat, fish, eggs, dairy products, nuts, and legumesmust be avoided. The majority of an individual’s proteinrequirements are met by medical foods designed for PKUthat contain amino acids other than phenylalanine, as well asvitamins and minerals. Adjunct treatment with sapropterin

dihydrochloride (synthetic BH4) causes a reduction in bloodphenylalanine in about 50% of individuals with PKU [5]and allows for 50% higher intake of protein-containing foodsin the diet, on average [6]; however, the vast majority ofthese individuals still require medical food and intact proteinrestriction to keep blood phenylalanine in the treatmentrange.

Many patients have difficulty adhering to the diet recom-mendations [7] due the extent of the dietary protein restric-tion, lack of access or acceptance of modified low proteinfoods, poor palatability of medical foods, and cost [8]. Fol-lowing the diet carefully requires the individual with PKU ortheir caretakers to strictlymonitor phenylalanine intake fromfood. Traditionally, the diet has been managed by weighingand measuring all foods eaten, looking up the phenylalaninecontent of each food, and keeping an exact record of dietaryintake. For individuals who have high blood phenylalanineconcentrations, following the diet is even more difficult dueto the neurocognitive problems associated with high bloodphenylalanine concentrations, including executive functiondeficits, anxiety, and slow processing speed [9], which caninterfere with choosing and measuring foods appropriately[8]. Following the diet can be even more difficult for

HindawiJournal of Nutrition and MetabolismVolume 2017, Article ID 4083293, 5 pageshttps://doi.org/10.1155/2017/4083293

2 Journal of Nutrition and Metabolism

individuals who have blood phenylalanine concentrationsabove treatment range, due to the association between highblood phenylalanine andneurocognitive problems. Executivefunction deficits, anxiety, and slow processing speed [9]can interfere with the skills needed to adhere to the diet,such as choosing and measuring foods appropriately [8]. Inaddition, counting and tracking phenylalanine intake add tothe burden of managing the diet for PKU [10].

The Simplified Diet method of managing dietary pheny-lalanine intake in patients with PKU has been studiedin Europe and Australia. In a 2003 study, 15 subjectswith PKU from the United Kingdom consumed differentamounts of “free” foods during 3 study phases. Free foodsin Phase 1 included fruits and vegetables containing lessthan 50mg phenylalanine/100 g; in Phase 2: fruits andvegetables containing less than 75mg phenylalanine/100 g;and in Phase 3, fruits and vegetables containing less than100mg phenylalanine/100 g of food. Blood phenylalanineand diet intake were monitored for 15 weeks and there wasno effect on plasma phenylalanine concentrations duringall three phases [11]. Another study, in which subjectsconsumed foods containing 100mg of phenylalanine/100 gwithout counting these foods, showed no negative effects onplasma phenylalanine concentrations [12]. However, duringthis study, fruits and vegetable consumption did not increase.Rohde et al. conducted three studies (lasting 4 weeks, 1year, and 3 years) where “free” foods were defined as thosecontaining less than 75mg of phenylalanine/100 grams offood. While total phenylalanine intake increased slightly,there was no significant difference in blood phenylalanineconcentrations when counting with this method [13, 14].In Australia, good metabolic control was associated with aphenylalanine-counting method where foods with less than50mg phenylalanine/100 g were considered free and otherfoods were counted using 0.5 g increments in protein [15].

The Simplified Diet is defined here as an approach tomanaging the PKU diet that allows individuals with PKU toconsume foods that contain lower amounts of phenylalaninewithout measuring or counting them, recognizing that nodiet for PKU is simple. It is designed to provide increasedflexibility, promote healthy food choices, and be easier dietmanagement than the traditional method of counting allphenylalanine consumed. While the counting method isdifferent in each clinic, the goal of the Simplified Diet is thesame: to maintain blood phenylalanine in the recommendedtreatment range of 120–360 umol/L. In many clinics in theUnited States, the Simplified Diet is a new approach tomanaging PKU, but in others, it has been used for yearsbut was called the Low Protein Diet. This paper aims todescribe research on the effectiveness of the Simplified Dietand outlines education, counseling strategies, and challengesencountered in implementing the diet by registered dietitiansin three US clinics.

2. Materials and Methods

This manuscript includes the experiences of five registereddietitians at three metabolic clinics in the US: one that beganusing the Simplified Diet in 2015, one that has been using this

approach since 1965, and one that has used this method withadults returning to diet since 1983, but only recently for allpatients with PKU.

3. Results and Discussion

Experiences regarding the effectiveness of the SimplifiedDiet,education, counseling strategies, and challenges encounteredin implementing the diet were collected and are described.While the foods counted or allowed as free vary slightlyfrom clinic to clinic (Figure 1 and Supplementary Materialsavailable online at https://doi.org/10.1155/2017/4083293), theconcept of only counting certain foods while allowing othersto be consumed freely is the same.Themainstay of diet treat-ment is medical food that contains little or no phenylalanineand provides the majority of protein to patients with PKU.

3.1. Transitioning to the Simplified Diet. In one clinic, allpatients who had counted in the traditional fashion weretransitioned to the Simplified Diet. Families and patientswere provided with a letter detailing the history of theSimplified Diet, how the trial period would work, and whatdiet changes would be made. The concept was first presentedin a group settingwith a pilot group of 15 patients and allowedfor parents and patients to openly discuss concerns andask questions for clarification. The main concern expressedby families was how fluctuating intake would affect bloodphenylalanine concentrations. All patients/families presentchose to implement the Simplified Diet following the clinicvisit.

Prior to starting the new counting method, blood pheny-lalanine was measured. Each patient’s phenylalanine pre-scription was reduced by 30%. For example, if a patient waspreviously prescribed 300mg of phenylalanine/day, whentransitioning to the Simplified Diet, they were counseled tocount 210mg of phenylalanine, but no longer to count ormeasure certain free fruits or vegetables that contain less than75mg of phenylalanine per 100 grams of food. Low proteinmodified foods containing less than 20mg of phenylalanineper serving were also considered free and did not need to becounted. Table 1 is an example of a diet containing 300mg ofPhe and compares what is counted using a traditionalmethodand the Simplified Diet method.

Families were provided with detailed handouts statingwhich fruits, vegetables, and low protein foods could be eatenfreely (Figure 1). Weekly blood phenylalanine was measuredand results were overall stable. Families reported increasedsatisfaction with the diet and more independence in foodchoices for the patient with PKU. Older children, teens, andadults who previously had to track intake of all foods oftenchose fewer fruits and vegetables because they wanted to usetheir phenylalanine “allowance” for foods such as potatoesand snack foods.

No diet changes were made during the 4-week trialperiod, even if blood phenylalanine concentrations fluctuatedslightly. This allowed for the “newness” of the diet to subside.When first allowed the option to eat certain foods freely,some patients ate more than the usual amount of some foodsbut after several weeks’ intake returned to typical volume.

Journal of Nutrition and Metabolism 3

Free fruits

Free vegetables

For any questions on specific items, please contact your metabolic dietitian.Developed by IMD Nutrition, Children’s Hospital ColoradoPlease remember your NO foods are still NO foods

Acorn squashBeetsBok choyButternut squashCabbageCarrotsCauliflowerCeleryChayote squashCucumber

Green beansEggplantJicamaLeeksLettuceOkraOnionsParsnipsPeppers (all varieties)Pumpkin

RadishesRutabagaSauerkrautSpaghetti squashSummer squash (zucchini and yellow)TomatoesTurnipsYucca (cassava root)

Apples—fresh and dryApricots—fresh and dryBananasBerries (all varieties)CherriesCranberries—fresh anddryDatesFigsGrapefruit

PapayaPeachesPears—dry and freshPersimmonPineapplePlantainsPlumsPomegranatesPrunesRaisins

Please remember your NO foods are still NO foods

Measure and count Phe/protein

All other fruits and vegetables (see separate sheet for specific list)

Low protein foods less than 20 mg of Phe per serving from low protein food companies

Do not measure andcount Phe/protein

ArtichokesArugulaAsparagusAvocadoBroccoli

Brussels sproutsCornKale, mustard greens, Swiss chardMushroomsPeas

PotatoesSeaweed/NoriSundried tomatoesSpinachYams/sweet potatoes

Limit orange juice to 1 cup per day

Dried fruit (except apples, apricots, craisins, pears, prunes, and raisins)

For any questions on specific items, please contact your metabolic dietitian.Developed by IMD Nutrition, Children’s Hospital Colorado

∗Limit orange juice to 1 cup per day

GrapesGuavaJackfruitKiwiLemonsLimesMangoMelon (all varieties)OlivesOranges∗

Figure 1: Simplified PKU diet; Inherited Metabolic Diseases Clinic, Children’s Hospital Colorado.

Following implementation of the diet with the pilot group,the letters and handouts specifying the free fruits, vegetables,and low protein foods were sent to all patients in the clinic.All families that are seen regularly in the clinic now followthe Simplified Diet, a transition that took approximately twoyears.

3.2. Implementing the Simplified Diet: Infants. When aninfant is 4–6 months of age, parents are instructed to offeronly free foods, at first, so that the child can become accus-tomed to eating solids before any foods must be counted. Asthe child accepts free fruits and vegetables, counted foods areintroduced, typically dry infant cereal or a counted vegetable,

such as spinach or sweet potato. In one clinic, parents arecounseled to start counting protein from dry cereal afterthe infant can consume two tablespoons of dry cereal perday. At that point, a one-gram protein diet is initiated,followed by increments in protein intake of 0.5 to 1 gramof protein, with a corresponding decrease of protein fromstandard infant formula. For clinics where phenylalanine iscounted, typically 15mg increments of phenylalanine fromfood are allowed. Concomitant with the introduction ofcounted foods, the dietitian must decrease the amount ofphenylalanine from standard infant formulas. To account forthe free foods that will be eaten, the dietitian must reducethe phenylalanine content of medical food/infant formula

4 Journal of Nutrition and Metabolism

Table 1: Example of counting phenylalanine using the traditional counting method and the Simplified Diet1.

Food eaten Traditional counting method Simplified dietApple Jacks� Cereal2, 1/2 cup 40mg 40Low protein bread, 2 slices 30mg FreeLow protein cheese, 2 slices 30mg 30mgPringles�2, 3 each 15mg 15mgApple, 1 10mg FreeFrench fries 120mg 120mgLow protein pasta, 1 cup 20mg FreeGreen beans, 1/2 cup 35mg FreeMedical food, 24 ounces 0 0Total 300 2051In this example, the patient’s phenylalanine tolerance is 300mg/day; the patient counts 210mg from food and is allowed unlimited “free foods” (defined asthose containing <75mg phenylalanine/100 g food as well as low protein foods with less than 20mg Phe per serving). 2Kellogg Company, Battle Creek, MI,USA.

recipe by more than 15mg of phenylalanine from the infant’sformula. For example, 20mg of phenylalanine would bereduced in the infant formula but the parents would countonly 15mg of solid food intake (30% of total phenylalanineintake accounted for by free foods). For breastfed infants, noadjustment is routinely made to the medical food. Intake ofbreast milk typically declines when food portions increaseand blood phenylalanine is closely monitored. If an increasein blood phenylalanine is seen, the volume of medical foodrecommended is increased. Breastfeeding is encouraged in away that does not interfere with mealtime or medical foodintake.

With the Simplified Diet, families of infants and toddlersreport being able to introduce new foods without worryingabout the phenylalanine that may not be consumed due tofinicky eating, food spills, or incomplete intake.

3.3. Adults Returning to Treatment. In one clinic, the simplerapproach to managing the diet was first used with adults andmaternal PKU patients returning to clinic who were treatedas children and taken off treatment at about 5 or 6 yearsof age. The simplified approach to counting phenylalaninewas born out of necessity as it was overly burdensome andunrealistic for patients who had been on an unrestricteddiet for decades to start to measure and record everythingthey ate. These adults had never learned to manage the dietthemselves as children, because they were taken off dietat a young age. Moreover, when returning to clinic, theadult patients often had the neurocognitive deficits associatedwith having high blood phenylalanine concentrations formany years that made learning and following the diet verydifficult. When this approach was first used, the target bloodphenylalanine was 120–600 𝜇mol/L. However, the approachwas also taught to women returning to the diet for pregnancywhere target blood phenylalanine was 120–360 𝜇mol/L. Forwomen with PKU who came to attention already pregnant,there was an inordinate amount of information to learn in ashort period of time in order to protect the fetus: choosingand preparing medical food, accessing medical food andlow protein food, cooking low protein foods, monitoring

blood phenylalanine, and tracking food intake. Simplifyingthe counting of phenylalanine was essential, practical, andeffective inmaintaining goodmetabolic control. Adults usingthe Simplified Diet have expressed that they enjoy freedom tochoose foodswhen eating outside of the home at school, whilesocializing with friends, or at work.

3.4. Challenges. Challenges have been minimal. While thesimplified method has been well accepted, implementing theSimplified Diet has been more of a challenge with parentswho have followed the diet carefully for many years and, atfirst, have been resistant to change. They sometimes havedifficulty “letting go” of counting and tracking all food intake.However, with time, most parents see the benefit of havingtheir child expand daily consumption of fruits and vegetableswhile maintaining blood phenylalanine in the treatmentrange. The ability to choose freely from the foods that arenot counted allows a greater sense of independence andmoreflexibility, especially as children age.

One challenge has been for patients who have a verylow tolerance to phenylalanine, less than 250mg/day. Forthese individuals, frequent intake of the free foods that aretoward the high end of phenylalanine content (those that areclose to the 75mg phenylalanine/100 g cut-off) may causeelevations in blood phenylalanine, and the ratio of freefoods to counted foods must be modified for these patients.Usually the phenylalanine prescription is divided into 30% ofphenylalanine from free foods and 70% from counted foods,but for the patient with very low phenylalanine tolerance,40% of the prescription must be “set aside” for free foods.If blood phenylalanine concentrations start to increase, dietrecords are kept and analyzed to determine if free food intakeexceeds 30%. Conversely, patients with a higher phenylala-nine tolerance, especially patients with mild or moderatePKU or those who respond to sapropterin dihydrochloride,can often choose several foods with relatively high pheny-lalanine content from the “free” foods on a particular dayand yet have little variation in blood phenylalanine. For theseindividuals, 20–25% of the prescription is set aside for freefoods.

Journal of Nutrition and Metabolism 5

For individuals that have been following this methodfor a long time, measuring of food portions may becometoo relaxed, the intake of higher protein vegetables maybecome excessive, and/or the quantity of low protein foodsmay increase, causing excessive protein and phenylalanineintake and elevated concentrations of blood phenylalanine.Regardless of the method used to count intake, propermanagement of the PKU diet requires close monitoring ofblood phenylalanine and adjusting the diet as necessary tokeep blood phenylalanine in the recommended treatmentrange.

4. Conclusion

This paper offers perspectives from three US clinics that haveimplemented the Simplified Diet. While the approaches toimplementing the diet vary slightly, all have observed thatthe Simplified Diet is easier to follow, encourages healthyfood choices, and can improve the quality of life for patientswith PKU as compared to the traditional counting method.Since adherence to the diet for PKU is poor, especially inolder teens and adults, strategies to simplify the diet shouldbe considered. Research on the long-term nutrient intake andmetabolic control of patients on the SimplifiedDiet is needed.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

The authors would like to acknowledge Ann Wessel, LeslieMartell, and Sommer Gaughan for their contributions topatient education materials developed by Boston Children’sHospital and Children’s Hospital Colorado.

References

[1] N. Al Hafid and J. Christodoulou, “Phenylketonuria: a reviewof current and future treatments,” Transl Pediatr, vol. 4.4, pp.304–317, 2015.

[2] D. Patel, J. Kopec, F. Fitzpatrick, T. J. Mccorvie, and W. W. Yue,“Structural basis for ligand-dependent dimerization of pheny-lalanine hydroxylase regulatory domain,” Scientific Reports, vol.6, Article ID 23748, 2016.

[3] J. Vockley,H. C. Andersson, K.M.Antshel et al., “Corrigendum:phenylalanine hydroxylase deficiency: diagnosis and manage-ment guideline,”Genetics inMedicine, vol. 16, no. 4, p. 356, 2014.

[4] K. M. Camp, M. A. Parisi, and P. B. Acosta, “Phenylketonuriascientific review conference: state of the science and futureresearch needs,”Molecular Genetics andMetabolism, vol. 112, pp.87–122, 2014.

[5] J. R. J. Utz, C. P. Lorentz, D. Markowitz et al., “START,a double blind, placebo-controlled pharmacogenetic test ofresponsiveness to sapropterin dihydrochloride in phenylke-tonuria patients,” Molecular Genetics and Metabolism, vol. 105,no. 2, pp. 193–197, 2012.

[6] N. Longo, G. L. Arnold, G. Pridjian et al., “Long-term safetyand efficacy of sapropterin: The PKUDOS registry experience,”

Molecular Genetics and Metabolism, vol. 114, no. 4, pp. 557–563,2015.

[7] E. R. Jurecki, S. Cederbaum, J. Kopesky et al., “Adherence clinicrecommendations among patients with phenylketonuria in theUnited States,”Molecular Genetics and Metabolism, vol. 120, no.3, pp. 190–197, 2017.

[8] A. MacDonald, M. Van Rijn, H. Gokmen-Ozel, and P. Burgard,“The reality of dietary compliance in the management ofphenylketonuria,” Journal of InheritedMetabolic Disease, vol. 33,no. 6, pp. 665–670, 2010.

[9] R. Jahja, S. C. Huijbregts, L. M. de Sonneville et al., “Cognitiveprofile and mental health in adult phenylketonuria: A PKU-COBESO study,” Neuropsychology, vol. 31, no. 4, pp. 437–447,2017.

[10] I. Eijgelshoven, S. Demirdas, T. A. Smith, J. M. T. Van Loon,S. Latour, and A. M. Bosch, “The time consuming natureof phenylketonuria: a cross-sectional study investigating timeburden and costs of phenylketonuria in the Netherlands,”Molecular Genetics and Metabolism, vol. 109, no. 3, pp. 237–242,2013.

[11] A.MacDonald,G. Rylance, P.Davies,D.Asplin, S. K.Hall, and I.W. Booth, “Free use of fruits and vegetables in phenylketonuria,”Journal of InheritedMetabolicDisease, vol. 26, no. 4, pp. 327–338,2003.

[12] M. Zimmermann, P. Jacobs, R. Fingerhut et al., “Positiveeffect of a simplified diet on blood phenylalanine control indifferent phenylketonuria variants, characterized by newbornBH 4 loading test and PAH analysis,” Molecular Genetics andMetabolism, vol. 106, no. 3, pp. 264–268, 2012.

[13] C. Rohde, U. Mutze, J. F. W. Weigel et al., “Unrestricted con-sumption of fruits and vegetables in phenylketonuria: no majorimpact on metabolic control,” European Journal of ClinicalNutrition, vol. 66, no. 5, pp. 633–638, 2012.

[14] C. Rohde, U. Mutze, S. Schulz et al., “Unrestricted fruits andvegetables in the PKU diet: A 1-year follow-up,” EuropeanJournal of Clinical Nutrition, vol. 68, no. 3, pp. 401–403, 2014.

[15] A. L. Sweeney, R. M. Roberts, and J. M. Fletcher, “Dietaryprotein counting as an alternative way ofmaintainingmetaboliccontrol in phenylketonuria,” JIMD Reports - Case and ResearchReports, vol. 3, pp. 131–139, 2012.