Growth patterns for untreated individuals with MPS I: Report from the international MPS I registry

Abstract Mucopolysaccharidosis Type I (MPS I), caused by deficiency of α‐L‐iduronidase results in progressive, multisystemic disease with a broad phenotypic spectrum including patients with severe (Hurler syndrome) to attenuated (Hurler–Scheie and Scheie syndromes) disease. Disordered growth is common with either phenotype. The study objectives were to construct sex‐ and age‐specific estimated length/height and head circumference growth curves for untreated individuals with severe and attenuated disease and compare them with clinical reference standards. Untreated individuals in the MPS I Registry with at least one observation for length/height and/or head circumference and assigned phenotype as of May 2017 were included. Median growth for 463 untreated individuals with severe disease deviated from reference growth curves by ~6 months of age and fell below the third percentile by 4 years of age. Median head circumference was above reference curves from 3 to 4 months through 3 years of age. Among 207 individuals with untreated attenuated disease, median height fell below the third percentile by 9 years of age with divergence from reference curves by 2 years of age. MPS I‐specific growth curves will be useful in evaluation of long‐term outcomes of therapeutics interventions and will provide a foundation for understanding the pathogenesis of skeletal disease in MPS I.

type. The study objectives were to construct sex-and age-specific estimated length/height and head circumference growth curves for untreated individuals with severe and attenuated disease and compare them with clinical reference standards. Untreated individuals in the MPS I Registry with at least one observation for length/height and/or head circumference and assigned phenotype as of May 2017 were included. Median growth for 463 untreated individuals with severe disease deviated from reference growth curves by~6 months of age and fell below the third percentile by 4 years of age. Median head circumference was above reference curves from 3 to 4 months through 3 years of age. Among 207 individuals with untreated attenuated disease, median height fell below the third percentile by 9 years of age with divergence from reference curves by 2 years of age. MPS I-specific growth curves will be useful in evaluation of longterm outcomes of therapeutics interventions and will provide a foundation for understanding the pathogenesis of skeletal disease in MPS I.  (Montano, Tomatsu, Brusius, Smith, & Orii, 2008;Patel et al., 2014;Quartel et al., 2015) but have not been developed for MPS I. Describing the natural history of growth patterns in MPS I could prove helpful in the evaluation of treatment impact and provide insights into disease pathogenesis for this condition.
MPS I occurs in approximately 1/100,000 live births (Giugliani, Federhen, Michelin-Tirelli, Riegel, & Burin, 2017;Moore, Connock, Wraith, & Lavery, 2008) and is caused by deficiency of α-L-iduronidase (IDUA) resulting in altered catabolism of the glycosaminoglycans (GAGs), dermatan sulfate and heparan sulfate (Muenzer, 2012). These catabolic defects lead to GAG accumulation as well as to the alteration of many secondary cellular pathways. Through pathophysiologic processes that are not well understood, patients present with progressive, multisystemic disease with variable age of onset and rate of disease progression. Disease phenotypes fit within a continuous spectrum of severity ranging from severe (Hurler syndrome, OMIM 607014) to attenuated (Hurler-Scheie OMIM 607015 and Scheie syndromes, OMIM 607016). Key features that define the spectrum include early manifestations and progressive neurocognitive impairment in the severe form of the disease, Hurler syndrome, to little or no cognitive involvement and later onset of disease-specific manifestations in individuals with attenuated disease, Hurler-Scheie and Scheie syndromes.
Growth abnormalities are key manifestations of all of the mucopolysaccharidoses as well as many other skeletal dysplasias caused by genetic defects in glycosaminoglycan synthesis, such as hereditary multiple exostoses (Zak, Crawford, & Esko, 2002), Desbuquois dysplasia type 1 (Paganini, Monti, et al., 2019), among others (Mizumoto, Ikegawa, & Sugahara, 2013;. These clinical observations emphasize the importance of glycosaminoglycan homeostasis for normal growth (Clarke, 2011). The objective of this study was to construct sex-and age-specific estimated length/height and head circumference growth curves for untreated individuals with either severe or attenuated MPS I to provide a basis for the evaluation of therapeutic interventions and pathogenic mechanisms.

| Editorial policies and ethical considerations
The study was approved by institutional review boards or ethics committees as required. Written informed consent from patients and/or their parents/guardians was obtained in compliance with all local laws and regulation.

| Study design
The MPS I Registry (clinicalTrials.gov NCT00144794) is a voluntary, observational global longitudinal database established to capture standard of care clinical and biochemical assessments of MPS I and to evaluate clinical outcomes. Individuals are referred to the MPS I registry by their clinician and data are collected both retrospectively and prospectively as previously described (Pastores et al., 2007).

| Study population
Observations from untreated MPS I individuals in the Registry with at least one measure for length/height and/or head circumference and assigned phenotype as of May 2017 were eligible for inclusion in the analyses. For an observation to be included, data points must have reported units in an appropriate range (see below) and be linked with a date of diagnosis, date of height measure date of birth, and sex. Individuals with severe MPS I who had a natural history follow-up time of greater than 12 years were excluded from analyses (n = 49).

| Statistical analysis
Individuals were categorized with either "severe" (Hurler Syndrome") or Observations four standard deviations beyond the median in 3-month intervals (e.g., 0-3 months, 3-6 months) were excluded for those 0-24 months of age, and a 2 year interval was used for the 2-20 year range. This resulted in the exclusion of one observation from the 0-24 months range and two observations from the 2-20 years range. Head circumference measurements that were less than the preceding measurements were removed, as were measurements in excess of 58 cm (n = 1).
Descriptive analyses were performed on variables of interest to determine distribution and appropriate statistics. Continuous variables are expressed as medians and interquartile range (IQR: 25%, 75%), and as counts and frequency n (%). Age-and sex-stratified standardized growth and head circumference data were downloaded from the National Center for Health Statistics (NCHS) for the WHO and CDC growth and head circumference charts (DepartmentHHS, 2002).
The General Additive Models for Location Scale and Shape (GAMLSS) of the R package was used to estimate the median, 5% and 95% curves for stature and head circumference for the respective age, sex, and phenotype strata from the natural history data of the MPS I Registry (Flegal & Cole, 2013). The distribution of measurements as a function of age is summarized by three parameters: skewness power in the transformation (L), the median (M), and coefficient of variation (S).
The Lambda Mu Sigma (LMS) method was chosen because it allows equal weight per observation, regardless of the number of measurements per individual, and it provides results similar to those used by the NCHS (Flegal & Cole, 2013). Additionally, scatterplots were composed of the MPS I natural history data over the WHO and CDC reference charts using SAS v9.4 (Cary, NC). The head circumference chart was extended to 58 cm to accommodate the larger occipital frontal circumference (OFC) values seen among individuals with severe MPS I.

| Participants
Among all participants in the MPS I Registry from October 2003 to May 2017, 670 had usable length/height observations prior to treatment intervention. MPS I was diagnosed by enzyme level in 443 patients, by genotype in 20, and by both enzyme level and genotype in 122 patients. Eighty-five patients had no records available regarding method of confirmatory diagnosis; however, phenotype designation data were entered by health-care providers familiar with and actively treating individuals with MPS I. There were 463 (69%) individuals categorized with severe and 207 (31%) with attenuated disease. Table 1 shows the median natural history follow-up time and the median age at diagnosis for individuals with either severe or attenuated MPS I. Males and females were equally represented among individuals with the severe phenotype. There were more females than males (55 vs. 45%) among individuals with attenuated MPS I. The majority of individuals enrolled in the MPS I Registry were from either North America or Europe (50% and 43%, respectively). There was a 61.5% overlap among those with both length/ height and head circumference. For the subpopulation of individuals with head circumference data (n = 450), the geographic distribution was similar to the total population; however, due to the lower age of those used in the head circumference analysis (0-36 months), the median age of diagnosis was lower in those with attenuated disease.
3.2 | Growth curves for severe MPS I Figure 1a,b shows the estimated length versus age curves of untreated males (A) and females (B) from 0 to 24 months of age with severe MPS I. Growth patterns are similar between males and females and tend to follow the reference curves between 12 and 24 months. Increased length relative to the reference is observed for both untreated males and untreated females between 6 and 12 months of age. Growth curves through 12 years are shown for untreated males and females with severe disease in Figure 1c,d, respectively. By 4 years of age, estimated median height drops below the third percentile and remains below the reference curves. Figures S1 and S2 show WHO and CDC reference curves overlaid on scatter plots of natural history data for individuals with severe disease.

| Growth curves for attenuated MPS I
The median age of diagnosis of MPS I in the untreated attenuated cohort was 4.5 years in males and 4.8 years in females (Table 1).     Figures S7 and S8, respectively. The available data suggest that median head circumference in both sexes rises above the CDC median from 5 months through 3 years of age. and proliferation at the growth plate (Heppner, Zaucke, & Clarke, 2015).

| Head circumference
These studies have provided insight into potential causes of short stature in MPS I; however, the pathophysiology underlying the patterns of growth has yet to be elucidated.
The median head circumference for untreated individuals with severe MPS I is well above the CDC reference curves by 3-4 months of age through 3 years of age. Previous studies have shown that individuals with untreated severe MPS I have head circumferences larger than population norms, but less is known about the attenuated phenotype (Aldenhoven et al., 2015;Kiely, Kohler, Coletti, Poe, & Escolar, 2017;Tylki-Szymanska, Rozdzynska, Jurecka, Marucha, & Czartoryska, 2010).
An analysis of 55 individuals with severe MPS I from a single center reported that enlarged head circumference was apparent at a median age of 8 months (Kiely et al., 2017). Thus, higher-than-normal OFC can serve as a clue to the diagnosis of MPS I, similar to males with MPS II (Parini, Jones, Harmatz, Giugliani, & Mendelsohn, 2016). Several cranial defects in MPS I likely contribute to altered head circumference, but it is not clear which contributes most. Even though thickened calvaria are noted in MPS I, macrocephaly is attributed to intracranial processes associated with megalencephaly, including accumulation of GAGs and sphingolipids in neurons, enlarged perivascular spaces, chronic communicating hydrocephalus, and neuroinflammation (Pavone et al., 2017).
An initial report of a brain necropsy in MPS I identified swelling of most of the large cerebral cortical neurons with increased numbers of astrocytes, numerous perivascular aggregations of fibrous astrocytes throughout gray matter, and greatly enlarged perivascular spaces in the white matter filled with collagen and microglial phagocytes (Bishton, Norman, & Tingey, 1956). Brain assessments in MPS disorders also demonstrated a mesenchymal contribution to increased brain size that is due to mononuclear cells filled with GAGs in distended periadventitial spaces, GAG accumulation in leptomeninges, and, instead of GAGs, the neurons contained an excessive amount of glycolipid-like material (Dekaban & Constantopoulos, 1977) shown to be an accumulation of sphingolipids (Constantopoulos & Dekaban, 1978;Kreutz et al., 2013).
Brain MRIs of attenuated MPS I individuals demonstrate dilated perivascular spaces and enlargement of subarachnoid spaces in addition to increased size of supratentoral ventricles (Matheus et al., 2004).
There were several possible limitations of the study as it is a clinically based registry. Variation in length/height measurements due to interrater variability and complications inherent to MPS I (e.g., joint contracture) could contribute to errors. Some of this variation is compensated for by multiple observations from a single individual when possible. Due to the age at presentation of those with attenuated disease, usually over the age of 36 months, there were limited number of length and head circumference observations for this subset. The likelihood of the potential bias of joint contracture among those with attenuated disease in the natural history period is low as it tends to reflect the joints of the fingers and lower arms. This is because this symptom would likely merit the start of treatment. However, in the present analysis, the LMS method is used to account for this potential skewness by minimizing the sum of the squared errors (Cole, 1990

| CONCLUSIONS
Even though there are limitations, growth curve assembly for individuals with MPS I may be helpful for the evaluation of long-term outcomes of therapeutic interventions. While additional registry data continue to be collected in untreated individuals with MPS I, these growth curves could prove clinically beneficial in evaluating medical concerns unrelated to anticipated MPS I manifestations.

ACKNOWLEDGMENTS
This study was supported by funding from Sanofi Genzyme. Patrice

DATA AVAILABILITY STATEMENT
The data that support the findings of this study can be requested by

MPS I Registry participants through a MPS I Registry Data Analyses
Request form. The data are not publicly available due to privacy or ethical restrictions. For additional information, please contact rarediseaseregistries@sanofi.com.