**A Three-Year Longitudinal Study Comparing Bone Mass, Density, and Geometry Measured by DXA, pQCT, and Bone Turnover Markers in Children with PKU Taking L-Amino Acid or Glycomacropeptide Protein Substitutes**

**Anne Daly 1,\*, Wolfgang Högler 2, Nicola Crabtree 1, Nick Shaw 1, Sharon Evans 1, Alex Pinto 1, Richard Jackson 3, Catherine Ashmore 1, Júlio C. Rocha 4,5, Boyd J. Strauss 6,7, Gisela Wilcox 6,8, William D. Fraser 9, Jonathan C. Y. Tang 9,10 and Anita MacDonald <sup>1</sup>**


**Abstract:** In patients with phenylketonuria (PKU), treated by diet therapy only, evidence suggests that areal bone mineral density (BMDa) is within the normal clinical reference range but is below the population norm. Aims: To study longitudinal bone density, mass, and geometry over 36 months in children with PKU taking either amino acid (L-AA) or casein glycomacropeptide substitutes (CGMP-AA) as their main protein source. Methodology: A total of 48 subjects completed the study, 19 subjects in the L-AA group (median age 11.1, range 5–16 years) and 29 subjects in the CGMP-AA group (median age 8.3, range 5–16 years). The CGMP-AA was further divided into two groups, CGMP100 (median age 9.2, range 5–16 years) (*n* = 13), children taking CGMP-AA only and CGMP50 (median age 7.3, range 5–15 years) (*n* = 16), children taking a combination of CGMP-AA and L-AA. Dual X-ray absorptiometry (DXA) was measured at enrolment and 36 months, peripheral quantitative computer tomography (pQCT) at 36 months only, and serum blood and urine bone turnover markers (BTM) and blood bone biochemistry at enrolment, 6, 12, and 36 months. Results: No statistically significant differences were found between the three groups for DXA outcome parameters, i.e., BMDa (L2–L4 BMDa g/cm2), bone mineral apparent density (L2–L4 BMAD g/cm3) and total body less head BMDa (TBLH g/cm2). All blood biochemistry markers were within the reference ranges, and BTM showed active bone turnover with a trend for BTM to decrease with increasing age. Conclusions: Bone density was clinically normal, although the median z scores were below the population mean. BTM showed active bone turnover and blood biochemistry was within the reference ranges. There appeared to be no advantage to bone density, mass, or geometry from taking a macropeptide-based protein substitute as compared with L-AAs.

**Citation:** Daly, A.; Högler, W.; Crabtree, N.; Shaw, N.; Evans, S.; Pinto, A.; Jackson, R.; Ashmore, C.; Rocha, J.C.; Strauss, B.J.; et al. A Three-Year Longitudinal Study Comparing Bone Mass, Density, and Geometry Measured by DXA, pQCT, and Bone Turnover Markers in Children with PKU Taking L-Amino Acid or Glycomacropeptide Protein Substitutes. *Nutrients* **2021**, *13*, 2075. https://doi.org/10.3390/nu13062075

Academic Editor: Roberto Iacone

Received: 30 May 2021 Accepted: 9 June 2021 Published: 17 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Keywords:** PKU; bone mineral density; bone turnover markers; osteoporosis; blood biochemistry; casein glycomacropeptide; amino acid protein substitute

#### **1. Introduction**

Optimal bone mass is key to preventing the risk of fractures later in life, and many factors influence peak bone mass accretion including genetics, physical activity, body composition, and quality of diet. Severe dietary restriction may be problematic in conditions such as phenylketonuria (PKU) which require rigorous exclusion of many natural foods [1]. In children with classical PKU, the majority of protein is provided by a low phenylalanine, semisynthetic protein (protein substitute), with some limited dietary phenylalanine given from natural foods according to individual metabolic tolerance and disorder severity. Dependency on a synthetic protein may compromise both peak bone mass attainment and bone geometry [2,3].

Protein substitutes, are traditionally derived from essential and non-essential amino acids and are usually supplemented with added vitamins, minerals, and trace minerals aimed at achieving optimal growth, bone mass, and body composition. Protein substitutes are necessary lifelong, but long-term adherence is difficult to sustain particularly during adolescence [4,5], which is a vulnerable time for maximising bone mass, density, mineralization, and growth potential. Amino acids (AAs) contribute to the structural components of bone in addition to those of growth and tissue maintenance [2,6,7].

Protein has a positive effect on bone [6,7], and protein intake promotes peripubertal bone growth and delays bone loss [8,9]. Several long-term prospective observational studies [10,11] have shown significant positive associations between protein intake and bone mineral content, periosteal circumference, cortical area, and an index of strength strain. These studies reinforce that a moderate to high protein diet promotes bone accretion. The acid ash theory suggests that a high protein intake including protein substitutes based on amino acids are detrimental to bone accretion [8,12]. Protein substitutes are acidic, producing sulphuric acid from sulphur containing amino acids. The hypothesis suggests that calcium stored primarily in bones is slowly excreted to buffer the acidic pH, and this process leads to a decreased bone mineral density [13–16]. However, systematic and meta-analysis studies have dismissed this theory [17,18]. Although the urine pH is lower when taking a protein rich diet, the pH of the extracellular fluid is undisturbed due to regulatory control by the kidneys [8].

The use of casein glycomacropeptide supplemented with amino acids (CGMP-AA) has been associated with improved bone mass in PKU animal models [19], but CGMP (a bioactive peptide) compared with AAs and their influence on bone mass, density, and geometry has not been studied in children with PKU.

In this longitudinal prospective controlled study over 36 months, we investigated the efficacy of CGMP-AA as compared with L-AA protein substitutes on bone mass, density, geometry, and turnover markers in children with PKU.

#### **2. Materials and Methods**

#### *2.1. Methods*

The inclusion criteria included the following: children with PKU diagnosed by newborn screening, children aged 5–16 years and not treated with sapropterin dihydrochloride, known adherence with protein substitutes, and maintenance of 70% of blood phenylalanine concentrations within the European PKU target therapeutic range for 6 months prior to study enrolment [20]. Target blood phenylalanine ranges for children aged 5–12 years were from 120 to ≤360 μmol/L, and for children 12 years and older they were from 120 to ≤600 μmol/L.

#### 2.1.1. Ethical Approval

This study was registered by the Health Research Authority and was given a favourable ethical opinion by the South Birmingham Research Ethical Committee (referenced 13/WM/0435 and IRAS (integrated research application system) number 129497). Written informed consent was given by at least one caregiver with parental responsibility and written consent was obtained from the subjects if appropriate for their age and level of understanding.

#### 2.1.2. CGMP-AA and L-AA Protein Substitutes

The CGMP-AA (a test product by Vitaflo International Ltd., Liverpool, UK) was a flavoured powder. Each 35 g sachet contained 20 g protein equivalent, and 36 mg phenylalanine, mixed with 120 mL of water. The flavoured L-AA was either a powder mixed with water or a ready-prepared liquid that provided 10, 15, or 20 g of protein equivalent. The CGMP-AA and L-AA products both had a similar nutritional and AA profile, except CGMP-AA contained residual phenylalanine and higher amounts of threonine and leucine.

#### 2.1.3. Selection into the CGMP Group or L-AA Group

The children chose the product they preferred, depending on their taste preference, i.e., the CGMP-AA group or L-AA group. They remained on this formula for the duration of the study.

#### *2.2. Study Design*

The primary aim of this 3-year longitudinal study was to compare bone mass, density and geometry of children with PKU taking CGMP-AA or L-AA as their primary protein source. The following examinations were conducted: dual-energy X-ray absorptiometry (DXA), together with blood bone biochemistry and blood and urine bone turnover markers. Peripheral quantitative computed tomography (pQCT) of the forearm was performed at 36 months only (Figure 1 and Table 1).

**Figure 1.** Diagram of the scheme for study methodology, from enrolment to 36 months. Legend: CGMP, casein glycomacropeptide; CGMP100, children taking all their protein substitute as casein glycomacropeptide; CGMP50, children taking a combination of casein glycomacropeptide and amino acids; L-AA, amino acids.

A previous pilot study [21] demonstrated that the residual phenylalanine in the CGMP-AA group led to compromised phenylalanine control in some children. Therefore, the CGMP-AA group was subdivided into: (1) CGMP100 group, in which the children took the entire protein substitute as CGMP-AA and (2) CGMP50 group, in which children took a combination of L-AA and CGMP-AA. There was also a third group of children who remained on their usual L-AA only (L-AA group).


**Table 1.** Frequency of nutritional blood biochemistry, bone blood and urine markers, DXA and pQCT scans, over study duration from enrolment to 36 months.

**Blood phenylalanine: weekly**

Legend: CGMP, casein glycomacropeptide; CGMP100, children taking all their protein substitute as casein glycomacropeptide; CGMP50, children taking a combination of casein glycomacropeptide and amino acids; L-AA, amino acids; DXA, dual-energy x-ray absorptiometry; pQCT, peripheral quantitative computerised tomography.

> 2.2.1. Dual-Energy X-ray Absorptiometry (DXA) and Peripheral Quantitative Computed Tomography (pQCT)

> A GE Lunar iDXA and Encore™ software version 13.1 g (GE Healthcare, Madison, WI, USA) was used to measure bone density at enrolment and at the end of 36 months. Trunk thickness and body weight were used as a guide for scanning each child in the most appropriate acquisition mode. Children lay supine on a bed, while the DXA scan was completed. The following measurements were performed: lumbar spine (L2–L4) areal bone mineral density (L2–L4 BMDa) in g/cm2, lumbar spine (L2–L4) bone mineral content (L2–L4 BMC) in g, total body mineral content (BMC) in g, total body less head BMDa (TBLH) in g/cm2, and size corrected outcome measures included lumbar spine bone mineral apparent density (L2–L4 BMAD) in g/cm3. At 36 months, in addition to the DXA assessment, pQCT was also performed.

#### 2.2.2. pQCT

The pQCT (Stratec XCT 2000 L, Pfozheim, Germany) measurements were taken at the 4% and 66% region of the non-dominant forearm, evaluating volumetric bone mineral density, together with muscle and bone geometry, size, and strength. At the 4% site, trabecular and total cross-sectional area were measured, while at the 66% site, cortical density, as well as muscle, bone, and fat area were measured. The pQCT also measured the strength strain index as a surrogate marker of bone strength.

#### 2.2.3. Serum Blood and Urine Bone Turnover Markers

Fasting, early morning, venous blood samples were collected at enrolment, 6, 12, and 36 months for the following serum bone markers: procollagen type 1 N-terminal propeptide (P1NP), type 1 collagen β crosslinked C-telopeptide (β-CTX), and bone alkaline phosphatase (bone ALP). A urine sample, the second sample of the day, was collected at enrolment, 6, 12, and 36 months for urine creatinine adjusted free urine pyridinoline (fPYD/Ur Cr) and urine free deoxypyridinoline crosslinks (fDPD/Ur Cr), and urinary calcium/creatinine ratio (Ur Ca/Cr). Urine samples were collected in containers, which were wrapped in tin foil and put into an envelope to shield them from any light. All urine samples were taken immediately to the laboratory for processing and stored at −80 ◦C. β-CTX and P1NP were analysed using an electro-chemiluminescence immunoassay (ECLIA) on a COBAS e601 analyser (Roche Diagnostics, Mannheim, Germany). The inter-assay coefficient of variation (CV) for β-CTX was <3% across the analytical range, between 0.01 and 6.0 μg/L, with a sensitivity of 0.01 μg/L. The inter-assay CV for P1NP was <3%, between 5 and 1200 μg/L, with a sensitivity of 5 μg/L. The serum bone ALP was determined

by MicroVue™ enzyme-linked immunosorbent assay ELISA kit (Quidel Corporation, San Diego, USA). The inter-assay CV for bone ALP was <5.8%, between 0.5 and 150 U/L, with a detection limit of 0.7 U/L.

The analyses for urinary fPYD and fDPD were performed using the liquid chromatography tandem mass spectrometry (LC-MS/MS) method, as described by Tang et al [22]. In brief, 0.5 mL of urine sample/calibration/quality control materials pretreated with 0.5 mL hydrochloric acid (40% concentrate) was extracted using a solid phase extraction (SPE) column packed with cellulose slurry. Pyridinium crosslinks were eluted from the SPE columns and analysed by LC–MS/MS coupled with an electrospray ionisation (ESI) source operated in positive mode. The inter-assay CVs were ≤10.3% for PYD in the concentration range of 5–2000 nmol/L and ≤13.1% for DPD between 2 and 1000 nmol/L. The lower limit of quantification was 6 nmol/L for fPYD and 2.5 nmol/L for fDPD.

Urine creatinine was measured to obtain the fPYD/ and fDPD/urine creatinine ratios and the urine calcium/creatine ratio. Samples were analysed using Roche kinetic colorimetric assays performed on a COBAS® C501 analyser (Roche, Burgess Hill, UK), according to the manufacturer's instructions. The inter-assay CV ranged from 1.3 to 2.1% across the assay working range for Ur Ca of 0.20–7.5 mmol/L and Ur creatinine of 0.355 mmol/L.

#### 2.2.4. Blood Biochemistry Markers

Overnight fasting blood samples for serum calcium, magnesium, phosphate, vitamin D, and parathyroid hormone were collected at enrolment, 6, 12, and 36 months.

#### 2.2.5. Blood Phenylalanine/Tyrosine Monitoring

Throughout the 36-month study, trained caregivers collected weekly overnight fasting morning blood spots at home for phenylalanine and tyrosine. Blood specimens were sent via the post to the Birmingham Women's and Children's Hospital Laboratory. The blood spot filter cards used were Perkin Elmer 226 UK standard NBS (Perkin Elmer, Waltham, MA, USA). All the cards had a standard thickness, and the blood phenylalanine and tyrosine concentrations were calculated on a 3.2 mm punch by tandem mass spectrometry.

#### 2.2.6. Pubertal Status

A general medical examination and pubertal status was measured at enrolment using the Tanner picture index. Stages 1 and 2 are classified as pre-pubertal, and Stages 3, 4, and 5 are classified as pubertal.

#### 2.2.7. Anthropometric Measurements

Weight and height were measured once every 3 months by one of two metabolic dietitians. Height was measured using a Harpenden stadiometer (Holtain Ltd., Crymych, Wales, UK).

#### *2.3. Statistical Methods*

Continuous data are presented as median and interquartile ranges and categorical data are presented as frequencies of counts with associated percentages. Longitudinal data are presented graphically using profile plots to show the average change over time. Correlations between continuous covariates were evaluated using Pearson's correlation coefficient. Comparisons between treatment groups were performed using analysis of covariance (ANCOVA) techniques, to analyse the follow-up data, while including baseline measures as adjusting covariates. Models also included covariates for patients' gender, age, and puberty status (supplementary data are provided for these parameters). A *p*-value of 0.05 was used throughout to determine statistical significance. All analyses were performed using R (Version 3).

#### **3. Results**

#### *3.1. Subjects*

Fifty children (28 boys and 22 girls) with PKU were recruited. Forty-seven children were of European origin and three children were of Asian origin. Forty-eight children completed the study, 29 children in the CGMP-AA group and 19 children in the L-AA group. At enrolment, the median age (range) in the CGMP100 group was 9.2 years (5–16 years) (*n* = 13); in the CGMP50 group, the median age was 7.3 years (5–15 years) (*n* = 16), and in the L-AA group, the median age was 11.1 years (5–16 years) (*n* = 19). Only six children were able to tolerate >10 g/day of natural protein (CGMP100 *n* = 2, CGMP50 *n* = 1, and L-AA *n* = 3), all the others received <10 g/day of natural protein.

#### 3.1.1. Subject Drop Out

One boy and one girl (both aged 12 years) in the CGMP-AA group were excluded from the study as both failed to comply with the study protocol. One failed to return blood phenylalanine samples and both had poor adherence to the low phenylalanine diet.

#### 3.1.2. Pubertal Status

The number of children prepubertal (Stages 1 and 2) at enrolment were: CGMP100 group, 62% (*n* = 8/13); CGMP50 group, 69% (*n* = 11/16); and L-AA group, 32% (*n* = 6/19).

The number of children in puberty (Stages 3 to 5) were: CGMP100 group, 38% (*n* = 5/13); CGMP50 group, 31% (*n* = 5/16); and L-AA group, 68% (*n* = 13/19).

#### 3.1.3. Median DXA Z Score Measurements for CGMP100, CGMP50, and L-AA Groups

Overall, there were no significant differences among the groups for any of the measured DXA parameters. Bone density was on the lower side of normal but within a normal reference range (Table 2).

**Table 2.** Median z scores (range) for L2–L4 bone mineral density (BMDa), lumbar spine bone mineral apparent density (L2–L4 BMAD), and total body less head BMDa (TBLH). Other parameters measured include median (range) L2–L4 bone mineral content and total bone mineral content for CGMP100, CGMP50, and L-AA groups, at enrolment and 36 months.



Legend: CGMP, casein glycomacropeptide; CGMP100, children taking all their protein substitute as casein glycomacropeptide; CGMP50, children taking a combination of casein glycomacropeptide and amino acids; L-AA, amino acids; L2–L4 BMD, bone mineral density lumbar vertebrae 2 to 4; BMAD, bone mineral apparent density; TBLH, total body less head; L2–L4 BMC, bone mineral content lumbar vertebrae 2 to 4; TBMC, total bone mineral content.

3.1.4. Median pQCT Z Score Measurements at 36 Months for CGMP100, CGMP50, and L-AA Groups

Similar to the DXA z score measurements, overall, there were no significant differences among the groups, but cortical density at the 66% site was statistically significantly different between the CGMP100 and L-AA groups (Table 3).

**Table 3.** Results from the pQCT scan measuring median z scores (range) for trabecular, cortical, and total densities at the 4% site; bone, muscle, and fat areas; strength strain index; and bone area/muscle area at 36 months in the CGMP100, CGMP50, and L-AA groups.




\* CGMP100 as compared with L-AA (*p* = 0.05). Legend: CGMP, casein glycomacropeptide; CGMP100, children taking all their protein substitute as casein glycomacropeptide; CGMP50, children taking a combination of casein glycomacropeptide and amino acids; L-AA, amino acids.

#### *3.2. Nutritional Bone Biochemistry Markers*

Median concentrations for all the biochemistry markers (calcium, phosphate, magnesium, vitamin D, and parathyroid hormone) were within normal reference ranges for all the groups over the 36-month study period (Table 4). There were no statistically significant differences within or among the groups.



Normal reference ranges (references from Birmingham Children's Hospital Clinical Chemistry Laboratory): Calcium 2.2–2.7 mmol/L, phosphate 0.8–1.9 mmol/L, magnesium 0.7–1.0 mmol/L, 25 (OH) vitamin D ≥50 nmol/L; parathyroid hormone (PTH) 15–60 ng/. Legend: CGMP, casein glycomacropeptide; CGMP100, children taking all their protein substitute as casein glycomacropeptide; CGMP50, children taking a combination of casein glycomacropeptide and amino acids; L-AA, amino acids.

#### Measurement for Bone Formation Markers and Urine Calcium

The urine calcium/creatinine ratio (Ur Ca/Cr) a measure of renal acid excretion was normal with no indication of excess calcium excretion (Table 5). Similarly, serum and urine BTM showed a physiological decrease with age, and no evidence of a disturbance between formation and resorption.


**Table 5.** Median (range) serum bone and urine turnover markers calculated from enrolment, 12, 24, and 36 months for CGMP100, CGMP50 and L-AA groups in girls and boys.

Legend: M, males; F, females; CGMP, casein glycomacropeptide; CGMP100, children taking all their protein substitute as casein glycomacropeptide; CGMP50, children taking a combination of casein glycomacropeptide and amino acids; L-AA, amino acids; β-CTX, type 1 collagen β crosslinked C-telopeptide; bone ALP, bone alkaline phosphatase; P1NP, procollagen type 1 N-terminal propeptide; fDPD, urine free deoxypyridinoline; fDPD/Ur Cr, deoxypyridinoline (free)/creatinine ratio; fPYD, urine free pyridinoline; fPYD/Ur Cr, pyridinoline (free)/creatinine ratio; Ur Ca/Cr, urine calcium/creatinine ratio; Ur Cr, urine creatinine. Standard references for children are not available.

> A strong positive correlation was observed between PN1P and β-CTX at 36 months (*r* = 0.82) (Figure 2). The ANCOVA analysis performed on PN1P indicated that the level of PN1P was somewhat dependent on age, with older subjects having a lower PN1P level. Furthermore, there was evidence of an increase in PN1P at 36 months associated with CGMP100 as compared with L-AA (*p* = 0.041) (Figure 3). There was no difference between the CGMP50 and L-AA groups (*p* = 0.80).

**Figure 2.** Correlation of β-CTX with PINP for CGMP100, CGMP50, and L-AA, at 36 months ( CGMP100, glycomacropeptide only; CGMP50, combination of CGMP and L-AA; and L-AA only).

**Figure 3.** Graphs showing serum and urine bone turnover markers at enrolment, 6, 12, and 36 months separated by gender for CGMP100, CGMP50, and L-AA groups.

#### *3.3. Anthropometry*

We have previously reported height, weight, and body mass index in this group of children [23]. At 36 months, all groups had a median positive height z score: L-AA, 0.2 (range 0 to 0.5); for CGMP50, 0.3 (range −0.1 to 0.7); and for CGMP100, 0.6 (range 0.1 to 0.7). Median weight for height z scores and BMI z scores were above the ideal reference mean, indicating an overweight group of children (Table 6).


**Table 6.** Median z scores (range) for height, weight, and BMI in the L-AA, CGMP100, and CGMP50 groups, measured annually from enrolment to 36 months in PKU children taking either L-AA, CGMP50, or CGMP100.

#### *3.4. Blood Phenylalanine Concentrations*

The median phenylalanine concentrations for this study have been previously reported. Median phenylalanine concentrations were within recommended target reference ranges for children aged ≤11 and ≥12 years old [23].

The median daily dose of protein equivalent from protein substitute was 60 g/day (range 40–80 g), and the median amount of prescribed natural protein was 5.5 g protein/day (range 3–30 g) or 275 mg/day of phenylalanine (range 150–1500 mg), in all three groups. Eighty-eight percent (*n* = 42) of the children tolerated ≤10 g/day natural protein and 12% (*n* = 6) >10 g/day (CGMP100, *n* = 2; CGMP50, *n* = 1; and L-AA, *n* = 3).

#### **4. Discussion**

In this 36-month longitudinal study in children with PKU, bone mass, density, and geometry were comprehensively examined by DXA and pQCT, in addition to serum BTM and blood biochemistry. With the exception of cortical density at the 66% site, none of the other bone measurements showed any benefit of CGMP100 over L-AA or CGMP50, suggesting that CGMP-AA had no advantage over L-AA for bone development. Similarly, there was no evidence to suggest any differences in bone mass, density, or geometry by gender, age, or puberty (Supplementary Tables S1 and S2).

A strong positive correlation between β-CTX and P1NP was observed in all three study groups, with P1NP being lower in the older age subjects, and an increased P1NP being evident in the CGMP100 group. This synergy between bone formation and resorption shows active bone turnover and reflects appropriate bone growth, since these markers derive from physiological processes. Our results contrast with those reported by Casto et al. [24], which suggested a trend towards increased bone resorption in subjects with PKU. This controlled study, was the first to monitor bone mass and density using two separate imaging technologies (DXA and pQCT), and holistically assesses serum bone, urine, and blood biochemistry parameters in PKU. Similar to findings from two systematic reviews [24,25], the overall bone density values for the groups were below the population mean but within the normal reference values. Imaging results met the International Society

for Clinical Densitometry (ISCD) recommendations (ISCD 2013) [26]. There were no differences in biochemical or BTM among the groups, suggesting no changes in bone metabolism attributed to the type of protein substitute. Naturally, BTM concentrations decreased in older adolescents towards those of lower adult levels, as a physiological phenomenon expected in a healthy population [27].

Unlike the findings of Schwahn et al., Mc Murry et al., and Fernandez et al. [28–30], we found no evidence to suggest that mineralization defects began in childhood, and then became more evident in adolescents. In this study, the groups of children were overweight. The relationship between overweight, obesity and bone is contentious.

Evidence [31] suggests that in early childhood obesity confers a structural advantage on the developing skeleton, but with age this relationship is reversed and becomes detrimental to skeletal development. Clarke et al [32] reported a positive relationship between adiposity and bone mass accrual in 3082 healthy children, while others [33,34] have reported opposite findings. Lean body mass has been shown to be the strongest predictor of bone mineral content [35,36] and relates to bone mass and skeletal development in children. Our previous study [37] indicated a trend towards improved lean body mass in the CGMP100 group; however, there was no evidence to suggest a similar beneficial effect on bone density in this group.

In PKU mouse models, CGMP as compared with L-AA has been shown to increase bone strength measured by biochemical mechanisms. Solverson [19] gave PKU and wild type mice different dietary regimens, i.e., a normal diet or a low phenylalanine diet supplemented with L-AA or CGMP protein substitutes. The PKU mice, regardless of protein substitute type, had lower bone density as compared with wild type mice, and those taking L-AA had inferior bone strength as compared with the CGMP protein substitute group. The authors proposed that the peptide structure of CGMP could possibly account for the positive influence on bone radial size improving biochemical performance. Alternatively, the high acid load due to L-AA could decrease bone strength via excreting higher amounts of calcium. However, both these suggestions were conjecture, as they did not measure net acid excretion, bone collagen, and markers of bone biomechanical performance. The results from our study in our cohort of children would suggest that neither of these mechanisms are active. BTM monitoring collagen were physiologically normal and there was no evidence of net acid excretion with a normal calcium/creatinine ratio.

Although many studies have identified lower BMD in PKU [38–41], not all of these studies included a size correlation for DXA output and there has been little agreement about lower BMD pathophysiology. Dobrowolski et al. [42] studied bone mineralization in PKU mice and showed phenylalanine toxicity inhibited bone mineralization. However, in human studies, there is a discord on the link between hyperphenylalaninemia and bone mass, with some studies showing a correlation and others not [38,40,43,44].

Within the three groups (CGMP100, CGMP50, and L-AA) there were expected physiological changes in the concentrations of BTM. In adults, BTM mainly represent bone remodelling; in children, BTM are released during bone remodelling, modelling, and perpendicular growth. Millet et al. [44] measured urine DPD and bone ALP in patients with PKU and compared these with a healthy paediatric group; bone remodelling was active in children with PKU aged 7–14 years, and bone ALP, as expected, was found to be significantly lower in the oldest group of patients (aged >18 years), although significantly higher DPD concentrations independent of age were reported. In our study, bone resorption and formation markers were consistently lowest in the L-AA group, particularly noticeable in the L-AA girls who had reached late puberty with a median age of 17 (8–18 years) at 36 months [27,43,45,46]. In contrast, the youngest group of CGMP50 boys showed an increase in BTM over the 36 months.

The interpretation of BTM is difficult and their concentrations vary widely in children, affected by a multitude of factors including age, gender, puberty, growth velocity, the rate of mineral accrual, hormonal regulation, nutritional status, circadian, and even day-to-day changes [47]. Paediatric reference data are available for some BTM [48–51], although UK

specific data are lacking, which hampers appropriate interpretation. Specificity for bone tissue as well as sensitivity and specificity of the measurement assays lead to variations, rendering comparisons among study groups difficult [50,52]. Despite these challenges, in our study in which children were followed for 36 months, BTM followed the expected variations for age with no differences between the groups. These children had an active bone turnover profile, supportive of a normal bone mineral density. The reason why their bone mineral densities were below the population median was unclear, but these groups were not at any increased clinical risk of fractures.

There are limitations to this study. Patient numbers in each group were small which reduced the power of this study. An extended follow-up period of >3 years may be needed for any differences to emerge between protein substitute sources, as noted, P1NP was increased in the CGMP100 group. We also did not have a healthy control group, which would have been beneficial to compare differences with the children with PKU. The ages of the children were significantly different in all three groups, and CGMP was given at two different concentrations making any absolute differences difficult to recognize, although statistical modelling was used to account for this variable. Age influences bone changes and children entered puberty over the study period. In children, no bone marker is specific for any of the three different biological processes of modelling, remodelling, and changes in endochondral ossification. However, our findings were consistent, i.e., all measurements were taken via DXA or pQCT and showed a below average bone density, with no significant differences among the groups taking CGMP-AA or L-AA. Bone markers appeared to follow a similar pattern to that in healthy children. We did not measure exercise activity in these groups of children, but a high proportion (60%) participated in regular activities such as football, dancing, and gymnastics.

#### **5. Conclusions**

In this detailed and comprehensive study measuring global bone development, using both two- and three-dimensional imaging in addition to serum BTM and blood biochemistry, a complete assessment of bone mass, density, geometry, and bone turnover was conducted. There were no statistical differences in the groups of children, who had good metabolic control when taking either L-AA or CGMP-AA protein substitutes. Bone density was normal and similar to the findings from systematic reviews, which suggests it was lower than the population norm but carried no increased osteoporotic risk. Bone remodelling processes appear to be active in children with PKU, with both L-AA and CGMP-AA protein substitutes supporting normal bone growth.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/nu13062075/s1. Table S1: Differentiation of median DXA z scores (range): lumbar vertebrae L2–L4 areal bone mineral density (L2–L4 aBMD); lumbar vertebrae L2–L4 bone mineral apparent density (L2–L4 BMAD) and total body less head areal bone mineral density (TBLH BMDa) by gender. Median value (range) for total body bone mineral content (BMC) by gender. Table S2: Differentiation of peripheral quantitative computerised tomography(pQCT) z scores (range) by gender at 36 months.

**Author Contributions:** Conceptualization, N.S., A.M. and A.D.; methodology, A.M. and A.D.; software, R.J. and A.D.; validation, A.M., A.D., W.H., N.C., S.E., A.P., C.A., J.C.R., B.J.S., G.W., W.D.F. and J.C.Y.T.; formal analysis, A.D. and R.J.; investigation, A.D. and A.M.; writing—original draft preparation, A.D. and A.M.; writing—review and editing, A.M., A.D., W.H., N.C., S.E., A.P., C.A., J.C.R., B.J.S., G.W., W.D.F. and J.C.Y.T.; supervision, A.M. Funding was part of Ph.D. project from Vitaflo International. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Vitaflo International as part of Ph.D. grant.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** Please contact the corresponding author.

**Conflicts of Interest:** A.D. research funding from Vitaflo International, financial support from Nutricia, Mevalia & Vitaflo International to attend study days & conferences. S.E. research funding from Nutricia, financial support from Nutricia & Vitaflo International to attend study days & conferences. A.P. has received an educational grant from Cambrooke Therapeutics and grants from Vitaflo International, Nutricia, Merck Serono, Biomarin and Mevalia to attend scientific meetings. C.A. received honoraria from Nutricia and Vitaflo International to attend study days and conferences. J.C.R. member of the European Nutritionist Expert Panel (Biomarin), the Advisory Board for Applied Pharma Research and Nutricia, and received honoraria as a speaker from APR, Merck Serono, Biomarin, Nutricia, Vitaflo, Cambrooke, PIAM and Lifediet. A.M. research funding & honoraria from Nutricia, Vitaflo International & Merck Serono, Member of European Nutrition Expert Panel (Merck Serono international), member of Sapropterin Advisory Board (Merck Serono international), member of the Advisory Board Element (Danone-Nutricia). The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


## *Article* **Current Practices and Challenges in the Diagnosis and Management of PKU in Latin America: A Multicenter Survey**

**Soraia Poloni 1,\*, Bruna Bento dos Santos 1,2, Ana Chiesa 3, Norma Specola 4, Marcela Pereyra 5, Manuel Saborío-Rocafort 6, María Florencia Salazar 7, María Jesús Leal-Witt 7, Gabriela Castro 7, Felipe Peñaloza 7, Sunling Palma Wong 8, Ramsés Badilla Porras 9, Lourdes Ortiz Paranza 10, Marta Cristina Sanabria 11, Marcela Vela Amieva 12, Marco Morales 13, Amanda Rocío Caro Naranjo 14, Antonieta Mahfoud 15, Ana Rosa Colmenares 16, Aida Lemes 17, José Fernando Sotillo-Lindo 18, Ceila Perez 19, Laritza Martínez Rey 20, Georgina María Zayas Torriente 21, Lilia Farret Refosco 1, Ida Vanessa Doederlein Schwartz 1,2 and Veronica Cornejo <sup>7</sup>**

	- gcastro@inta.uchile.cl (G.C.); felipe.penaloza@inta.uchile.cl (F.P.); vcornejo@inta.uchile.cl (V.C.)

**Abstract:** This study aimed to describe the current practices in the diagnosis and dietary management of phenylketonuria (PKU) in Latin America, as well as the main barriers to treatment. We developed a 44-item online survey aimed at health professionals. After a pilot test, the final version was sent to 25 practitioners working with inborn errors of metabolism (IEM) in 14 countries. Our results include 22 centers in 13 countries. Most countries (12/13) screened newborns for PKU. Phenylalanine (Phe) targets at different ages were very heterogeneous among centers, with greater consistency at the 0–1 year age group (14/22 sought 120–240 μmol/L) and the lowest at >12 years (10 targets reported).

**Citation:** Poloni, S.; dos Santos, B.B.; Chiesa, A.; Specola, N.; Pereyra, M.; Saborío-Rocafort, M.; Salazar, M.F.; Leal-Witt, M.J.; Castro, G.; Peñaloza, F.; et al. Current Practices and Challenges in the Diagnosis and Management of PKU in Latin America: A Multicenter Survey. *Nutrients* **2021**, *13*, 2566. https:// doi.org/10.3390/nu13082566

Academic Editor: Shanon L. Casperson

Received: 30 May 2021 Accepted: 8 July 2021 Published: 27 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Most countries had only unflavored powdered amino acid substitutes (10/13) and did not have low-protein foods (8/13). Only 3/13 countries had regional databases of the Phe content of foods, and only 4/22 centers had nutrient analysis software. The perceived obstacles to treatment were: low purchasing power (62%), limited/insufficient availability of low-protein foods (60%), poor adherence, and lack of technical resources to manage the diet (50% each). We observed a heterogeneous scenario in the dietary management of PKU, and most countries experienced a lack of dietary resources for both patients and health professionals.

**Keywords:** phenylketonuria; PKU; low-protein diet; newborn screening

#### **1. Introduction**

Latin America comprises 20 countries and has an ethnically diverse population of over 650 million people. With a complex political and economic background, these countries face many challenges in the diagnosis and care of patients with inborn errors of metabolism (IEM), such as phenylketonuria (PKU, OMIM #261600). The success of early diagnosis and dietary treatment of PKU has been well described since the 1960s. Since then, and until the mid-1970s, most developed countries have initiated national newborn screening (NBS) programs for PKU [1]. In Latin America, the first organized NBS programs were only started in 1986 in Cuba, followed by Costa Rica (1990), and Chile (1992) [2]. At present, 16 countries have national or regional NBS programs, but only 6 have coverage ≥90% [3].

Similarly to NBS programs, PKU management faces many challenges in Latin America. Despite significant health system reforms in the 1980s, inequality and impaired access to health care remains a major problem in the region [4]. A recent report by the Organization for Economic Co-operation and Development (OECD) showed that government and compulsory health insurance represented only 54.3% of the current expenditure on health in Latin America, with 34% of all health spending being paid out-of-pocket. Nearly 8% of the Latin American population spends more than 10% of their household consumption or income on health care services. Latin American countries also have a much lower availability of medical technologies and health professionals when compared to other countries [5].

Treatment of PKU inflicts a substantial time and cost burden on patients and their families [6], and this can be a significant obstacle to encouraging patients to remain on a restricted diet [7]. Moreover, a trained health care team is needed to manage the extremely restrictive diet and to educate patients and families, and frequent laboratory tests are required to guide the treatment. The current situation of PKU diagnosis and management in Latin America is unknown, since only sparse and country-based reports have been published [8–12]. Therefore, the aim of this study was to map the current practices in the diagnosis and dietary management of PKU in Latin America, as well as the main barriers to treatment perceived by health care providers.

#### **2. Materials and Methods**

#### *2.1. Study Design*

A questionnaire containing 44 questions on the diagnosis and management of PKU was developed by a team of experts from Brazil (S.P., B.B.S., I.V.D.S., and L.F.R.) and Chile (M.J.L., F.S., G.C., and V.C.). These were experienced metabolic dietitians and geneticists, all co-authors of this paper. The survey had multiple choice and short answer questions and was aimed at health care professionals following patients with PKU. Five main issues were addressed: features and professional training of the health care team, newborn screening, treatment goals and dietary practices, availability of alternative treatments, and perceived barriers to treatment.

After the first Portuguese and Spanish versions of the questionnaire were finished, a pilot study was performed with 6 PKU experts (3 Portuguese and 3 Spanish speakers)

to identify possible flaws or misinterpretations of the questions. Only minor adaptations were made, and the final version was then shared on an online platform.

To disseminate the survey, practitioners of IEM were searched for (through public archives of the Sociedad Latinoamericana de Errores Innatos del Metabolismo y Pesquisa Neonatal in all Latin American countries, and were found in 14 of them. The coordinator team designated 1 responsible person in each country to distribute the survey to other centers nationally. The aim was to distribute the survey to as many centers as possible in each country. The only exception was Brazil, the largest country with the most PKU treatment centers (>20); to avoid overrepresentation, we chose 1 center from each region of the country. The invitation and distribution of the survey was performed by e-mail from July to November 2020. The final version of the questionnaire is available by request.

#### *2.2. Ethical Aspects*

The study was approved by the Research Ethics Committee of Hospital de Clínicas de Porto Alegre, Brazil (CAAE 36401120.6.0000.5327), and the survey was initiated only after the participants agreed with the online informed consent form.

#### **3. Results**

#### *3.1. Participants*

Out of the 14 contacted country representatives, 13 were following patients with PKU, and all of them agreed to participate in the study. In total, 22 treatment centers were enrolled from the following countries: Brazil (*n* = 5), Argentina (*n* = 4), Colombia (*n* = 2), Venezuela (*n* = 2), Costa Rica (*n* = 1), Chile (*n* = 1), Mexico (*n* = 1), Paraguay (*n* = 1), Peru (*n* = 1), Dominican Republic (*n* = 1), Panama (*n* = 1), Uruguay (*n* = 1), and Cuba (*n* = 1).

The respondents were mostly female (91%), were aged ≥45 years (61%), and had worked with PKU for over 10 years (70%). Physicians represented 59% of the respondents, with the remaining respondents being dietitians. Regarding professional training, 45% (*n* = 10/22) stated that they had a specialization course in the field, 41% (*n* = 9/22) had only short-term courses, and 9% (*n* = 2) had no formal training. The number of patients with PKU who were followed up by the professionals varied considerably: 18.2% (*n* = 4/22) had <10 patients, 18.2% (*n* = 4/22) had 10–25 patients, 18.2% (*n* = 4/22) had 26–50 patients, 13.6% (*n* = 3/22) had 51–75 patients, and 32% (*n* = 7/22) had >75 patients.

#### *3.2. Newborn Screening and Phenylalanine (Phe) Monitoring*

Regarding NBS, all countries but one (Dominican Republic) had a national NBS program, the most recent one being in Colombia (2019). When inquired on the Phe cutoff level used to start dietary treatment, 13/22 centers (59%) responded ≥360 μmol/L (≥6 mg/dL), 5/22 (23%) responded ≥600 μmol/L (≥10 mg/dL), 2/22 (9%) responded <360 μmol/L (<6 mg/dL), and 1 (4.5%) responded ≥480 μmol/L (≥8 mg/dL). In most centers (19/22), blood Phe was measured in dried blood spots. The most used method to analyze blood Phe was the fluorometric assay (12/22), followed by tandem mass spectrometry (5/22). Figure 1 shows the recommended frequency of Phe and tyrosine (Tyr) monitoring in the studied centers.

**Figure 1.** Frequencies of blood phenylalanine (Phe, **a**) and tyrosine (Tyr, **b**) monitoring for each age group as adopted by the Latin American centers included in the study (*n* = 22). Numbers within columns represent relative percentages.

#### *3.3. Treatment Targets and Dietary Practices*

Figure 2 shows Phe target levels at different ages in the studied centers. Dietary guidance was most frequently performed through the simplified method of high/medium/low Phe content of foods (10/22), followed by individualized meal plans (8/22), and protein counting (3/22). A 24 h dietary recall (or similar) was performed at every appointment in most centers (17/22). Total protein prescriptions are described in Figure 3.

All but two respondents reported that the maintenance of partial breastfeeding was encouraged in classical PKU patients. Most respondents (80%) said that they instructed mothers to offer the protein substitute right before breastfeeding to control Phe intake.

Regarding nutritional monitoring, all centers reported weight and height measurements at every appointment, and 19/22 always assessed head circumference. The evaluation of body composition was less frequent; 12/22 (54%) did not assess skinfolds and none performed bioelectrical impedance analyses on a regular basis. The blood tests that were performed at least once a year were: a complete blood count (95%), fasting glucose (91%), total protein (91%), creatinine (91%), urea (77%), lipoproteins and triglycerides (77%), albumin (73%), vitamins B12 and D (60%), and ferritin (60%). A complete amino acid profile

was requested in 10/22 centers (45%), and only 1 center evaluated essential fatty acids on a regular basis. Bone densitometry was routinely performed in 11/22 centers (50%).

**Figure 3.** Total protein (natural + protein substitute) prescriptions, in different age groups, in the studied centers (*n* = 22) Some centers did not fully answer this question; therefore, the sample size varies in different age groups.

#### *3.4. Nutritional Resources*

Out of the 13 included countries, 9 reported having national guidelines for PKU management, and 12/22 (54%) centers had local management protocols. The theoretical background most commonly used by the respondents was: international guidelines (61%), scientific papers (56%), and national guidelines (48%). Regarding dietary resources, only 18% of the centers (4/22) reported having an adequate regional database of the Phe content of foods, and 33% stated that only an incomplete database was available. The remaining (49%) centers utilized a variety of international databases or considered only the protein content of foods for guiding the diet. Food recalls were usually calculated manually (48%) or through a customized spreadsheet developed by the center (3%). Only 3/22 centers reported having specialized nutrition software.

Except for one country, none of the participant countries had the Phe content available on food labels. Regarding protein substitutes, 11/13 countries had only unflavored powdered amino acid formulas; only 1 country (Argentina) had several options, such as gels and tablets, available. In 10/13 countries, the protein substitute was fully subsidized by the government. Specific low-protein foods for PKU were not available in 8/13 countries; even when these were available, 58% of the centers stated that they were not affordable. These products were subsidized by the government in only 2/13 countries.

#### *3.5. Alternative Treatments and Challenges*

Six countries had no alternative treatments available. Among those that had them, sapropterin (BH4) was the most frequent (six countries—Argentina, Brazil, Costa Rica, Dominican Republic and Mexico; approximately 60 patients in total); large neutral amino acids (LNAA) were available in two countries (Argentina and Peru), and glycomacropeptide (GMP) was available in one country (Argentina). Argentina was the only country that had all three options available, also with the most patients using them (>20 patients on BH4 and GMP and nearly 10 patients on LNAA).

Participants were asked to provide a score from 0 to 100 on how much they believed each category had contributed to hampering therapeutic success. Median scores are depicted in Figure 4. In addition to the aspects shown in Figure 4, other cited barriers to treatment were: low accessibility due to geographic location, limited access to alternative treatments, high cost of treatments, and long periods of time for samples to arrive at the laboratory.

**Figure 4.** Barriers to treatment most commonly perceived by the respondents (*n* = 22). Values represent the median scores assigned by the respondents. \* Educational level of patients and caregivers; \*\* Technical resources required or desirable to manage the diet, such as a local database of the Phe content of foods and specialized nutrition software.

#### **4. Discussion**

This study reports a broad and unprecedented characterization of the current state of diagnosis and management of PKU across Latin America. Data on NBS, laboratory tests, professional training, treatment targets, dietary practices, and resources, among other aspects, were compiled from 13 different countries and 22 treatment centers. These countries represent 87% of the Latin American population. Respondents were physicians and dietitians, most of whom were experienced in PKU treatments and were following a variable amount of PKU patients of all ages.

NBS for PKU began mostly after the 1990s in Latin America, nearly 30 years after the USA and some European countries had initiated their screening programs [1,2]. Nevertheless, most Latin American countries currently have wide-coverage national NBS programs for PKU and multidisciplinary reference centers for the follow-up of these patients, as shown in our study. Although the need for NBS and early treatment of PKU was generally agreed upon, other practices were not. Whereas both American and European guidelines [13,14] recommend that treatment should be started when Phe levels are ≥360 μmol/L, 35% (*n* = 8) of the centers in our sample employed different cutoffs, with most of them (*n* = 6) using higher levels. Higher cutoffs could miss mild PKU patients and raise concern due to the detrimental effect of high Phe levels in early life. A meta-analysis showed that each 100 μmol/L increase in Phe in early life predicted a 1.3- to 3.9-point decrease in intelligence quotient (IQ) over a Phe range of 394 to 750 μmol/L [15]. However, the exact cutoff at which treatment should begin is still debatable. There is a consensus that individuals with Phe levels >600 μmol/L should be treated, but the evidence regarding the initiation of treatment with blood Phe concentrations between 360 and 600 μmol/L is inconsistent. Given the risk of neurocognitive consequences, most guidelines recommend initiating treatment when blood Phe concentrations are >360 μmol/L [8,13,14].

The frequency of Phe and Tyr monitoring was highly heterogeneous among centers. The highest agreement (65%) found was in respect to measuring Phe once a week or more in infants younger than 1 year of age (Figure 1), which is in line with both American and European recommendations [13,14,16]. An even greater disagreement was observed for Tyr measurements, in all age groups. This probably reflects the lower availability of Tyr analyses in several centers: more than 20% of them rarely or never measured Tyr, regardless of the patient's age group or condition. Tyr monitoring is critical in PKU, since this amino acid cannot be synthetized properly due to the metabolic blockage, and a decreased availability of Tyr in the brain likely contributes to the cognitive impairment found in untreated patients [17]. American Genetic Metabolic Dietitians International (GMDI) guidelines recommend that Tyr measurements be performed as frequently as Phe measurements [16].

A similar heterogeneity was observed for Phe target values throughout life (Figure 2). The highest agreement (69%) was for children aged 2–12 years, where the Phe target was 120–360 μmol/L; this was in agreement with both international guidelines [13,14]. For infants younger than 1 year of age, most (61%) centers aimed for Phe levels to be between 120 and 240 μmol/L, a goal that differed from the American and European guidelines, which recommend a Phe target of 120–360 μmol/L [13,14]. Chilean guidelines support the 120–240 μmol/L target at this age, since in this period many factors interfere with the Phe level, such as growth, teething, infections, and frequent vaccinations [8]. The age group with the highest heterogeneity was >12 years, with 11 different targets reported. Eight (35%) centers agreed with the American target (120–360 μmol/L), and two (9%) agreed with the European values (120–600 μmol/L). A trend towards more restrictive targets was observed in all age groups and in pregnant patients. However, the theoretical basis for some of the reported targets was not clear; in some cases they were unique and diverged within the same country, even when national guidelines were available.

Greater consistency was found in dietary practices. The simplified method was the most frequently used approach to manage dietary intake (in 48% of the centers). The simplified diet approach has been shown to be easier to follow, encourages healthy food choices, and can improve the quality of life and adherence of patients with PKU [18,19]. Breastfeeding was encouraged in most centers, reflecting its clear evidence-based benefits in PKU [20,21]. Regarding nutritional monitoring, basic measurements such as weight, height, head circumference, and food recalls were performed in most centers at all appointments, meeting international recommendations [13,14]. Blood tests for nutritional monitoring were usually performed on a regular basis. It is noteworthy, however, that

pivotal examinations such as amino acid profiles and bone densitometry were not regularly performed in most centers (56%). These assessments are required in the follow-up of patients with PKU who are being treated, since they are at risk of amino acid deficiencies and osteopenia [13,14,17,22]. A likely explanation for this is that these two technologies are less available due to their high costs and need for specialized facilities. There are substantial differences in the availability of technologies across Latin American countries [5]. The total protein prescriptions showed some heterogeneity among centers (Figure 3). However, the well-established recommendation that a higher protein intake is necessary for patients with PKU [23] has been mostly followed.

Nutritional resources to support patients, families, and professionals were scarce. Although most countries had national guidelines, most respondents reached for international guidelines (61%) as theoretical background. This might be due to outdated or incomplete local guidelines. While most professionals used Phe intake for managing the diet, most of them (78%) did not have a suitable regional database of the Phe content of foods and had to rely on international databases. However, nutrient contents of foods can vary due to environmental factors, production, and processing, and might differ between countries [24]. Health professionals also face difficulties calculating the diets: only four (17%) centers had specialized nutrition software.

For patients, unflavored powdered amino acid formulas were the only protein substitute available in most countries. Specific low-protein foods for PKU were unavailable in 61% of the countries. Even when available, they were usually not affordable, since these were rarely subsidized by the government. Specially designed low-protein products are important for satiety and diet variety [25], and were also proven valuable in improving metabolic control and growth in patients with PKU [26]. However, they inflict a significant financial burden to the PKU diet: in an American study, low-protein foods represented the highest annual out-of-pocket costs (child = US\$1651.00; adult = US\$967.00) when compared to other categories of care [6]. Considering the gross national income per capita in 2019 [27], this would be equivalent to 20% of the income of a Latin American citizen. The average expenditure on food of a Brazilian citizen, for instance, is USD 866.00 per year. Therefore, it is completely unreasonable to expect that Latin American patients with PKU would be able to afford low-protein foods without subsidy.

Alternative therapies are also a reality for a few in Latin America. BH4 was the most common therapy, available in 7/13 countries. However, even when approved, this therapy was only used in a few patients. LNAA and GMP were even rarer, despite several products being available in Europe and the USA for years [28]. Alternative treatments are highly relevant in PKU since most patients struggle to follow the restrictive diet and to take the protein substitute [29,30]. As a consequence of suboptimal metabolic control and restrictive dietary management, psychiatric illness is common in adult PKU patients. The advent of new treatments that do not require such a restricted diet might improve metabolic control, mental health, and cognitive functioning in these patients [31].

Finally, we asked the respondents to score the topics they considered the greatest barriers to the adequate treatment of PKU in their realities. The answers largely reflected the major gaps found throughout the study: lack of nutritional resources for patients and professionals and the high cost of therapies. One of the highest assigned scores was for "poor adherence", which may also be an outcome of the difficulties mentioned above. Another barrier cited by the respondents was low accessibility due to geographic location. In Latin America, most of the sophisticated technologies that are required for the follow-up of patients with IEM are geographically concentrated in larger and wealthier urban areas, contributing to health inequalities in this population [5].

#### **5. Conclusions**

In conclusion, here we have reported the first compilation of the status of PKU care in Latin America. Despite most countries having national NBS programs and guidelines, we found a highly heterogeneous scenario considering practices across countries and

even within the same country. The struggles, however, were similar. Most countries experienced a lack of resources for both patients and health care professionals, which may be impairing treatment outcomes. Together, these results indicate an urgent need for a comprehensive Latin American guideline that must be able to integrate the latest evidence-based recommendations with the challenges and possibilities faced by Latin American countries.

**Author Contributions:** Conceptualization, data collection and analysis, writing—review and editing: S.P., B.B.d.S., I.V.D.S., V.C. The review of the project, data collection and proofreading of the manuscript: M.F.S., M.J.L.-W. and G.C. Data collection and proofreading of the manuscript: A.C., N.S., M.P., M.S.-R., F.P., S.P.W., R.B.P., L.O.P., M.C.S., M.V.A., M.M., A.R.C.N., A.M., A.R.C., A.L., J.F.S.-L., C.P., L.M.R., G.M.Z.T. and L.F.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by FIPE-HCPA, grant number 2020-0439, Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – Brasil (CAPES) and Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul (FAPERGS), grant number: 88887.192047/2018-00.

**Institutional Review Board Statement:** The study was approved by the Research Ethics Committee of Hospital de Clínicas de Porto Alegre, Brazil (CAAE 36401120.6.0000.5327).

**Informed Consent Statement:** Informed consent was obtained from all participants.

**Data Availability Statement:** Data available on request due to restrictions eg privacy or ethical. The data presented in this study are available on request from the corresponding author. The data are not publicly available due to containing information that could compromise the privacy of research participants.

**Acknowledgments:** The authors thank Dévora Random, Laura Fernández and Mariana Castro for their assistance with data collection. The authors also thank all Latin American PKU patients and their families.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

