Article Text

Download PDFPDF

Original Article
Hormonal, metabolic and skeletal phenotype of Schaaf-Yang syndrome: a comparison to Prader-Willi syndrome
  1. John M McCarthy1,
  2. Bonnie M McCann-Crosby2,
  3. Megan E Rech1,
  4. Jiani Yin1,3,
  5. Chun-An Chen1,3,
  6. May A Ali1,
  7. HaiThuy N Nguyen4,
  8. Jennifer L Miller5,
  9. Christian P Schaaf1,3
  1. 1 Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston, Texas, USA
  2. 2 Division of Endocrinology, Department of Pediatrics, Texas Children’s Hospital, Houston, Texas, USA
  3. 3 Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, Texas, USA
  4. 4 Division of Radiology, Texas Children’s Hospital, Houston, Texas, USA
  5. 5 Division of Endocrinology, Department of Pediatrics, University of Florida, Gainesville, Florida, USA
  1. Correspondence to Dr Christian P Schaaf, Jan and Dan Duncan Neurological Research Institute at Texas Children’s Hospital, Houston, TX 77030, USA; schaaf{at}


Background Nonsense and frameshift mutations in the maternally imprinted, paternally expressed gene MAGEL2, located in the Prader-Willi critical region 15q11-15q13, have been reported to cause Schaaf-Yang syndrome (SYS), a genetic disorder that manifests as developmental delay/intellectual disability, hypotonia, feeding difficulties and autism spectrum disorder. Prader-Willi syndrome (PWS) is a genetic disorder characterised by severe infantile hypotonia, hypogonadotrophic hypogonadism, early childhood onset obesity/hyperphagia, developmental delay/intellectual disability and short stature. Scoliosis and growth hormone insufficiency are also prevalent in PWS.

There is extensive documentation of the endocrine and metabolic phenotypes for PWS, but not for SYS. This study served to investigate the hormonal, metabolic and body composition phenotype of SYS and its potential overlap with PWS.

Methods In nine individuals with SYS (5 female/4 male; aged 5–17 years), we measured serum ghrelin, glucose, insulin-like growth factor 1 (IGF-1), insulin-like growth factor binding protein 3, follicle-stimulating hormone, luteinising hormone, thyroid-stimulating hormone, free T4, uric acid and testosterone, and performed a comprehensive lipid panel. Patients also underwent X-ray and dual-energy X-ray absorptiometry analyses to assess for scoliosis and bone mineral density.

Results Low IGF-1 levels despite normal weight/adequate nutrition were observed in six patients, suggesting growth hormone deficiency similar to PWS. Fasting ghrelin levels were elevated, as seen in individuals with PWS. X-rays revealed scoliosis >10° in three patients, and abnormal bone mineral density in six patients, indicated by Z-scores of below −2 SDs.

Conclusion This is the first analysis of the hormonal, metabolic and body composition phenotype of SYS. Our findings suggest that there is marked, but not complete overlap between PWS and SYS.

  • MAGEL2
  • IGF-1
  • ghrelin
  • scoliosis
  • bone mineral density

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Schaaf-Yang syndrome (SYS (OMIM 615547)) is a genetic disorder caused by nonsense and frameshift pathogenic variants in the maternally imprinted, paternally expressed MAGEL2 gene. MAGEL2 is an intronless gene in the Prader-Willi domain on chromosome 15q11-15q13 that encodes a protein important for endosomal protein trafficking.1 SYS is diagnosed through either whole exome sequencing or single-gene sequencing of MAGEL2. SYS was originally considered ‘Prader-Willi-like syndrome’, due to the location of the gene in the Prader-Willi domain and the phenotypic overlap between individuals with MAGEL2 loss-of-function and those with Prader-Willi syndrome (PWS (OMIM 176270)).2 Phenotypic overlap between SYS and PWS includes neonatal hypotonia, feeding difficulties, weight gain, developmental delay/intellectual disability and hypogonadism.3

While there are many shared characteristics, particularly during infancy, some clinical features of SYS highlight a clinical profile distinct from PWS, leading to the subsequent renaming of the disorder from ‘Prader-Willi-like syndrome’ to Schaaf-Yang syndrome. Specifically, interphalangeal joint contractures are not typically seen in PWS, yet are present in the majority of individuals with nonsense and frameshift MAGEL2 mutations (82%). Additionally, the prevalence of autism spectrum disorder was recently reported as 27% in individuals with PWS versus 77% in individuals with SYS. The manifestations of SYS further diverge from PWS during later childhood and adolescence, as individuals with PWS typically develop hyperphagia, potentially leading to life-threatening obesity, while only a minority of individuals with SYS manifest abnormally increased appetites and morbid obesity.3 4

The endocrine and metabolic phenotypes of PWS have been documented extensively, with studies highlighting the prevalence of hyperphagia/obesity, hypogonadism, short stature/growth hormone (GH) deficiency, hypothyroidism and impaired glucose tolerance/diabetes mellitus,5 as well as high levels of total cholesterol and low-density lipoprotein (LDL) cholesterol.6

Here, we report the endocrine and metabolic phenotypes of nine individuals with SYS from seven unrelated families (patients 3 and 5 are biological siblings, and patients 1 and 8 are cousins). Our study provides evidence that there is some endocrinological overlap between PWS and SYS, particularly as it relates to ghrelin and insulin-like growth factor 1 (IGF-1) levels. We also found skeletal similarities, through the measurement of scoliosis and bone mineral density (BMD), and overlap in body composition. However, our study also highlights important differences in the metabolic and endocrine phenotypes of individuals with SYS and PWS, demonstrating that while there are striking similarities among the two disorders, they remain phenotypically distinct.

Subjects and methods


Eighty two individuals with nonsense and frameshift mutations in MAGEL2 have been identified through whole-exome sequencing, and reported to Baylor College of Medicine in Houston, Texas. We recruited nine patients from this cohort, aged 5–17 years. All families known to the investigator who had children within the appropriate age range were notified of the research study, and sent the consent form and contact information of the investigators. The first nine patients whose families expressed interest and met study criteria were consented and enrolled. Inclusion criteria were based on a previously identified nonsense or frameshift mutation in the paternal allele of MAGEL2, and an age of 5–17 years. Exclusion criteria consisted of dependence on mechanical ventilation, an inability to travel to Texas Children’s Hospital and/or an inability to tolerate fasting for up to 7 hours. All participants or their legal guardians provided informed consent. Subjects travelled to Texas Children’s Hospital in Houston, Texas, where they completed a 2-day study at the Texas Children’s Wallace Tower and the Texas Children’s Autism Center. One of the two days was reserved for psychological testing, which will be reported separately. Proband information is summarised in table 1. Patients 1, 7 and 8 of this study had been reported previously as patients 2, 3, and 1, respectively, by Fountain et al.4 Patients 3 and 5 of this study had been reported previously as patients CMH383 and CMH382 by Soden et al.7 Patients 6 and 9 of this study had been reported previously as patients 3 and 1 by Schaaf et al.2 Patients 2 and 4 of this study had not been reported previously.

Table 1

Baseline characteristics of participants


We investigated circulating levels of ghrelin, glucose, IGF-1, insulin-like growth factor binding protein 3 (IGFBP-3), follicle-stimulating hormone (FSH), luteinising hormone (LH), thyroid-stimulating hormone (TSH), free T4 and uric acid, and performed a comprehensive lipid panel of each patient, following overnight (>6 hours) fasting. In addition, testosterone was measured in male subjects. Blood was drawn at the Texas Children’s Hospital outpatient pathology laboratory through simple venipuncture, and all measurements were analysed by Texas Children’s Pathology and Quest Diagnostics, with the exception of ghrelin levels, which were analysed by Inter Science Institute, Inglewood California, USA. Immediately following the initial fasting blood draw, patients were given a dose of dextrose (1.75 g of dextrose per kilogram body weight, with a total maximum dose of 75 g). Dextrose was administered through oral solution, or through the use of a G-button for patients who could not tolerate oral feeds. After 2 hours, two additional blood samples were collected to assess glucose tolerance and postprandial ghrelin levels.

On the same morning of the blood draws, patients were seen in the Texas Children’s Pediatric Radiology Outpatient Clinic to undergo spinal X-ray and dual-energy X-ray absorptiometry (DEXA) scans. X-ray imaging was performed using EOS low-dose three-dimensional X-ray technology. Posteroanterior and lateral radiographs of the spine were obtained to assess for scoliosis. DEXA scans were performed using the Hologic Horizon DXA System to assess patients’ BMD in the full body, lumbar spine, left hip and right hip. Age-specific, sex-specific and ethnicity-specific BMD Z-scores were generated for each patient from available Hologic database information. All imaging performed was interpreted by clinical paediatric radiologists.


Insulin-like growth factor and insulin-like growth factor binding protein 3

IGF-1 and IGFBP-3 were measured and compared with the normal range for age and gender. Z-scores were calculated based on SD with a reference range of −2.0±2.0. Z-scores outside of this range were considered abnormal. Two patients (patient six and seven) were on treatment with recombinant GH at the time of the study. Six out of the seven individuals not on GH treatment demonstrated abnormally low IGF-1 levels, showing Z-scores below −2. The mean IGF-1 Z-score for all nine patients was −2.1±0.9. Two out of the seven patients not on GH showed IGFBP-3 values with Z-scores lower than −2; these two patients also demonstrated low IGF-1 levels, further confirming GH deficiency. The mean IGFBP-3 Z-score for all nine patients was −1.4±0.9 These data are presented in table 2 and figure 1, panel A.

Table 2

Glucose tolerance, ghrelin and growth factor results

Figure 1

(A) Serum levels of insulin-like growth factor-1 (IGF-1) and insulin-like growth factor binding protein 3 (IGFBP-3) in nine individuals with Schaaf-Yang syndrome. Normal range based on age/gender is represented by vertical grey-shaded boxes. Individual patient levels of IGF-1 and IGFBP-3 are represented by black squares. (B) Serum levels of fasting and 2-hour glucose in nine individuals with Schaaf-Yang syndrome. Normal range of blood glucose levels is represented by the grey-shaded trapezoid. Individual patient values are represented by black dots, with a black line connecting the fasting and 2-hour glucose levels for each individual patient. (C) Serum levels of fasting and postprandial ghrelin in nine individuals with Schaaf-Yang syndrome. Individual fasting and postprandial ghrelin levels are represented by black dots, with black lines connecting the values for each individual patient. (D) Serum levels of luteinising hormone (LH) and follicle-stimulating hormone (FSH) in nine individuals with Schaaf-Yang syndrome. FSH and LH were measured in all patients, and serum values were compared with normal ranges, based on age, gender, Tanner stage and menstrual cycle stage. Patients 8 and 9 were postpubertal at the time of assessment, all other patients were prepubertal, based on physical examination. Normal ranges for each patient are represented by grey-shaded vertical boxes. Individual values are represented for each patient by black boxes.

Fasting and 2-hour glucose tolerance

All nine subjects tested for serum glucose values within the normal range for overnight fasting glucose of 60–99 mg/dL, with a mean level of 75.1±6.6 mg/dL.

Glucose tolerance was measured 2 hours after administration of the dextrose solution. Six out of nine patients demonstrated postprandial serum glucose levels in the normal range of <140 mg/dL, with a mean level for these six patients of 109.5±32.9 mg/dL. Two of nine patients (patients 3 and 8) tested in the prediabetic range for glucose tolerance (140–199 mg/dL), with a mean serum glucose value of 168.5±17.7 mg/dL. Finally, one out of nine patients, patient 6, tested in the range consistent with diabetes (>200 mg/dL), although he was not previously diagnosed. These data are presented in table 2 and figure 1, panel B.

Fasting and postprandial ghrelin

All nine subjects’ fasting ghrelin levels were elevated compared with reference ranges of 520–700 pg/mL for lean patients and 340–450 for obese patients (BMI >30 kg/m2); the mean level was 1615±899 pg/mL. Postprandial ghrelin levels were measured 2 hours after administration of dextrose solution, and the mean level was 1024±367 pg/mL. The postprandial ghrelin level of patient 5 was elevated compared with the reference range of <420 pg/mL, but no appropriate reference range for lean individuals exists, so comparisons were not made. These data are presented in table 2 and figure 1, panel C.

Lipid panel

A comprehensive lipid panel was performed for all patients, measuring total cholesterol, triglycerides, high-density lipoprotein (HDL) cholesterol and LDL cholesterol. Two of nine patients manifested high total cholesterol levels, 1/9 manifested high triglyceride levels, 2/9 manifested high LDL cholesterol levels and 1/9 manifested high HDL cholesterol levels. These data are shown in table 3.

Table 3

Lipid panel results

Follicle-stimulating hormone and luteinising hormone

All nine patients showed LH values within the normal range for sex and pubertal status, and five out of nine patients showed LH values below the normal range for sex and pubertal status. These data are presented in table 4 and figure 1, panel D.

Table 4

Gonadotropins and testosterone results

Other assessments

Seven of nine patients had TSH levels within the normal range of 0.5–4.0 uIU/mL for individuals aged between 6 months and 18 years. Patients six and eight demonstrated slightly elevated TSH levels of 4.4 and 4.37 uIU/mL, respectively. Both patients were overweight (defined by the Centers for Disease Control and Prevention as BMI at or above the 85th percentile and below the 95th percentile), which may have contributed to their high TSH levels. All nine patients displayed free T4 values within the normal range of 0.8–2.0 ng/dL for individuals aged between 6 months and 18 years.

Because elevated uric acid levels can be seen in PWS,8 9 uric acid was measured in all patients and compared with the normal range of 2–6.2 mg/dL. One out of nine patients, patient nine, fell outside this range, with a level measured to be 6.8 mg/dL.

All four male patients were found to have normal serum testosterone levels for their age range and pubertal status.

X-ray and DEXA scan

All patients received spinal X-rays, which were interpreted by clinical radiologists at Texas Children’s Radiology Clinic. Three out of nine patients showed significant scoliosis of >10°. Patient two had a 65° dextroconvex scoliosis in the thoracolumbar spine. Patient four showed a 19° right-sided curvature in the mid to lower thoracic spine and a 52° left-sided curvature in the upper lumbar spine. Patient five showed a 16° levoconvex scoliosis of the thoracolumbar spine. X-ray images of patient four and patient two are presented in figure 2, panel B.

Figure 2

(A) X-ray image of patient 4 (on the left), taken while sitting, showing a 19° right-sided curvature in the mid to lower thoracic spine and a 52° left-sided curvature in the upper lumbar spine. X-ray image of patient 2 (on the right), taken while sitting, showing a 65° dextroconvex scoliosis in the thoracolumbar spine. (B) Results of DEXA scan measuring for bone density in nine individuals with Schaaf-Yang syndrome. Three measurements are represented: bone mineral density in the lumbar spine, right hip and left hip. A normal range between +2 and −2, based on SD is represented by the grey-shaded rectangle. Individual values based on Z-scores, using SD, are represented by black squares for each patient.

DEXA scans were performed to determine BMD in the lumbar spine, hips and total body. Due to the behavioural phenotype of SYS, many of the participants could not cooperate, and therefore did not complete all scans. All nine patients were able to receive lumbar scans, and 7/9 patients were able to complete right/left hip scans. Scans for patients two, three and six were not weight-bearing. Of the patients who could participate, BMDs were considerably low, with average lumbar Z-scores of −0.7±1.2; right hip Z-scores of −3.8±2.6 and left hip Z-scores of −3.0±4.1. Hip Z-scores are presented in figure 2, panel A.

As total body scans required the most time sitting still, only participants 3, 5, 7 and 9 were able to complete these scans. These four participants showed an average total body BMD Z-score of −2.6±1.6. Body composition measurements were also determined for these same four individuals, revealing total body fat Z-scores of 1.75, 2.33, 1.13 and 1.88, respectively; fat mass index (FMI) Z-scores of 0.23, 2.1, 0.81 and 1.3, respectively; and lean index (LMI) Z-scores of −2.3, –0.18, −1.5 and −0.64, respectively. Body composition measurements are presented in online supplementary table 1.

Supplementary file 1


Here, we present the clinical characterisation of nine individuals with SYS, as it pertains to endocrine/metabolic function, scoliosis, body composition and BMD. All nine individuals have molecularly confirmed nonsense or frameshift mutations in the MAGEL2 gene, a protein-coding gene in the Prader-Willi domain. Endocrine/metabolic function and body composition have not previously been studied in SYS, and this is the first effort to compare these findings with PWS.

Ghrelin levels and hyperphagia

Hyperphagia is a primary clinical feature of PWS, typically beginning around age 8 years and persisting through adulthood.10 Individuals’ excessive appetite, combined with delayed gastric emptying and decreased activity levels, often leads to life-threatening obesity in patients with PWS.5 11 Fasting and postprandial hyperghrelinaemia have been found in individuals with PWS, and have been considered a contributing factor in their hyperphagia.12 13 Ghrelin, known as the ‘hunger hormone’, is a potent circulating orexigenic hormone produced mainly in the stomach, with levels that rise during fasting and are suppressed following food intake. As plasma ghrelin levels are dependent on BMI, levels are generally reduced in obese individuals.14 In children with PWS, overnight fasting ghrelin has been measured to be consistently higher than lean and obese controls, although a range of values has been reported, from 898.0±238.2 pg/mL15 to 1458.6±1271.6 pg/mL16.

Interestingly, patients with SYS showed an even further elevated average overnight fasting ghrelin level of 1615±899 pg/mL, and while a decrease was seen in average postprandial ghrelin levels, these values were still high compared with reference. Of note, ghrelin reference ranges were established by Inter Science Institute based on studies of obese17 and lean18 adults and by in-house comparison to normal; however, all nine patients’ levels are also elevated compared with results from a study of prepubertal children, which reported average overnight fasting ghrelin levels of 563.4±81.5 pg/mL for lean individuals and 281.3±29.5 pg/mL for obese individuals.19 Although no appropriate reference range for postprandial ghrelin was provided for lean individuals, patient five did exhibit elevated levels compared with obese controls.

These results seem surprising, given that individuals with SYS rarely develop hyperphagia, and do not typically engage in the food-seeking behaviours that are seen among individuals with PWS; furthermore, several families report that their children with SYS do not have an appreciation of satiety. Nevertheless, several potential explanations may account for these data. First, circulating ghrelin is found in both acylated and unacylated forms, and while acylated ghrelin is associated with hyperphagia and is found to be elevated in PWS,20 individuals with SYS may instead have high levels of unacylated ghrelin. Second, it is possible that the feeding difficulties commonly experienced in infants with SYS lead to ghrelin resistance/unresponsiveness as they get older, as is suggested to be the case in infants with failure to thrive, who have high circulating acylated and total ghrelin levels but paradoxically manifest decreased feeding and malnutrition. The report proposes that the elevated ghrelin may be an adaptive attempt to signal the hypothalamus to increase appetite, but the infants may have become desensitised to ghrelin’s orexigenic effects.21 Third, three individuals in our sample (patients 1, 2 and 4) have IQs <25, indicating a profound intellectual disability, which may make it difficult for them even to express feelings like hunger to their caretakers.

It is also possible that elevated ghrelin levels do not directly lead to hyperphagia, as several groups have now shown that pharmacological reduction of ghrelin to normal levels in PWS did not affect the weight, appetite or eating behaviour of the individuals.22–24 The pathophysiology of appetite in PWS is still unclear, and future studies should aim at clarifying the role of ghrelin in patients with PWS and SYS.

Moreover, the precise physiological function of ghrelin remains to be defined. Traditionally, ghrelin was believed to control appetite and facilitate excessive weight gain in response to a high-fat diet, but recent findings question these conclusions. Studies in congenic C57BL/6J ghrelin knockout (KO) and ghrelin-receptor KO mice showed food intake is independent of ghrelin signalling, and that the absence of ghrelin fails to protect mice from diet-induced obesity.25 Acute stimulation of food intake in ghrelin-cell ablated mice requires doses of exogenous ghrelin that produce plasma ghrelin concentrations many-fold higher than the endogenous concentrations found in wild-type mice, suggesting endogenous ghrelin is not a critical regulator of food intake.26

Ultimately, it is premature to speculate the exact cause of elevated ghrelin, but the data challenge the simplistic idea that elevated ghrelin levels lead to hyperphagia in PWS.

Glucose tolerance and diabetes mellitus

It has been estimated that up to 25% of adults with PWS have non-insulin-dependent diabetes mellitus, with a mean age of onset of 20 years.27 Additionally, a study of 142 French children with PWS found impaired glucose tolerance in 4% of individuals (mean age: 10.2 years), but did not find any cases of diabetes mellitus in the cohort, reaffirming a non-juvenile onset of diabetes.28

Although the sample size was small, our study found an increased prevalence of both glucose intolerance and diabetes mellitus among patients with SYS compared with PWS. In the 2-hour glucose tolerance test, 2/9 patients with SYS tested in the prediabetic range, and 1/9 tested in the diabetic range. As all study participants are children or adolescents, glucose intolerance may manifest earlier in development in SYS than PWS. Therefore, ongoing screening for diabetes in patients with SYS is recommended throughout the lifetime. Ultimately, it appears too early to draw any conclusions about whether or not the increased glucose intolerance seen among patients with SYS in this small study is pathogenetically related to the disease.

IGF-1, IGFBP-3 and GH deficiency

IGF-1 is a small peptide, produced primarily by the liver, which plays a large role in tissue development, including the muscles and bones. IGF-1 deficiency is known to cause underdevelopment of the facial bones, small hands/feet, weakness of the muscular system and short stature.29 IGF binding proteins regulate the availability of free IGF-1 to the tissues. In humans, almost 80% of circulating IGF-1 is bound to IGFBP-3.29 30 Both IGF-1 and IGFBP-3 production are stimulated by GH, and therefore serve as indicators of proper GH function.

Short stature and GH deficiency are frequently seen in PWS, and data from at least 15 studies involving more than 300 children demonstrate reduced GH secretion.31 It has been found that in most children with PWS, serum IGF-1 levels are reduced, whereas obese controls have normal or slightly elevated IGF-1 levels. Low IGF-1 and GH levels are seen both in the severely obese patients with PWS and in those of normal weight. Additionally, in contrast to healthy obese children, decreased levels of IGFBP-3 have been reported in PWS.31

In a study of 19 patients with PWS, IGF-1 Z-scores were measured to be −0.7±0.8. Another study of 11 patients found the Z-scores to be −1.9±1.3.31 Our study found IGF-1 Z-scores of −2.1±0.9, demonstrating that patients with SYS express depressed levels of IGF-1, similar to PWS. We also found that patients with  SYS share the phenotypic trait of low levels of IGFBP-3, as the mean IGFBP-3 Z-score for all nine patients was −1.4±0.9.

As seen in table 1, 8/9 patients with SYS displayed height Z-scores of less than −1, while their weight Z-scores were much higher. The combination of short stature, elevated fat mass/reduced lean mass, low bone density and low IGF-1/IGFBP-3 levels suggest GH deficiency similar to that seen in PWS. Among the nine individuals in this study, two (patients 6 and 7) were on GH treatment at the time of the study. Both had been diagnosed with GH deficiency by a GH stimulation test. Clinically, GH administration improved longitudinal growth in both cases, although improvement in individual 7 was more modest given that treatment was started more recently. The height of patient 6 progressed from −4 SD when GH treatment was started in January 2010 to approximately −1.28 SD as of August 2017, and the height of patient 7 progressed from −4.2 SD at the time of GH treatment initiation in April of 2016 to −3.7 SD as of November 2016.

Current best practice involves early intervention of GH therapy for PWS. GH therapy has proven to decrease fat mass and increase muscle mass. It is also thought to cause improvements in height, BMI, gross motor skills, language acquisition and cognitive scores.5 Therefore, a clinical trial of GH therapy should be considered in SYS to explore potential benefits in this patient population.

Thyroid hormones and sex hormones

It has been reported that individuals with PWS are more likely to develop hypothyroidism than controls.32 Studies report that up to 25% of individuals with PWS manifest central hypothyroidism, with inappropriately normal TSH in the setting of low T4 levels, with a mean age of diagnosis of 2 years.28 32 Of the SYS cohort, none displayed evidence of central hypothyroidism; two out of nine patients displayed slightly elevated levels of TSH, while all patients had T4 measurements within the normal range. For these two patients, the levels represent mild subclinical hypothyroidism that does not require treatment. As it is unclear how the phenotype progresses from childhood to adulthood, thyroid levels in patients with SYS should continue to be monitored by an endocrinologist throughout development.

Hypogonadotrophic hypogonadism is also a major feature of PWS,33 and has been previously reported in studies of SYS,2 motivating testing for FSH, LH and testosterone. Although five of nine individuals (three females/two males) showed LH values below the reference ranges provided by the performing laboratory, these five individuals were all young and prepubertal, so a lower-end LH level would be expected. Moreover, the two males who had low LH also had testosterone levels appropriate for prepubertal individuals, confirming that they were still in Tanner stage I–II, further justifying the low LH. In fact, several other paediatric reference intervals show all nine individuals’ LH values to be within the normal range for sex and pubertal status.34–36 In addition, the males aged 15 and 17 years did show evidence of pubertal changes. Therefore, the results of this study do not seem to provide sufficient evidence of hypogonadotrophic hypogonadism; however, it would be important to follow how the LH and FSH levels of prepubertal individuals develop as they approach the expected onset of puberty.

One caveat worth noting is that three of the four male individuals in this study had a history of undescended testicles. However, given that all males had normal levels of testosterone, and a pattern of hypogonadotrophic hypogonadism is not otherwise observed, this finding may be more related to prenatal hypotonia than to hypogonadotrophic hypogonadism. Because fetal abdominal movements sporadically increase intra-abdominal pressure which promotes testicular descent, a lack of fetal movement due to prenatal hypotonia in SYS would inhibit this process.37 Further investigations will be needed to assess gonadal function in individuals with SYS over time.


Scoliosis is defined as a spinal curve with a Cobb angle of >10° on a standing posteroanterior radiograph. Scoliosis data for individuals with PWS are wide-ranging, showing a prevalence of 40%–80%, and varying in age of onset and severity.5 38 It has been hypothesised through the use of Magel2 KO mice that MAGEL2 could be the scoliosis-determining gene in PWS.39 In our study, 33% of patients with SYS showed significant scoliosis of >10°, with curvatures of 16°, 52° and 65°. Studies in PWS show that scoliosis is a progressive deformity, so we suggest that patients with SYS continue to monitor scoliosis throughout childhood and adolescence.38

Bone mineral density and body composition

Individuals with PWS display markedly low BMD which can result in a high risk of fractures and osteoporosis.31 40 DEXA scans in PWS have shown BMD Z-scores of 0.2 spinal BMD, −0.66 total hip BMD and −0.65 total body BMD,41 and values in the SYS cohort were even lower.

PWS is also associated with obesity, abnormally high fat mass and low lean body mass. A study of 25 prepubertal PWS children reported an average total fat mass Z-score of 1.2 and fat percentage Z-score of 2.1, both adjusted for sex and age, and an average lean body mass Z-score of −1.7, adjusted for sex and height.42 The four individuals in the SYS cohort who completed total body scans each demonstrated relatively similar trends in body composition, showing high fat mass and low lean mass. However, while the PWS children showed an average BMI Z-score of 1.2, reflecting a high prevalence of overweight/obesity, scores of the participants in our cohort were lower (with an average Z-score of −0.05 for the four individuals who completed total body scans, and of 0.61 for all nine individuals).

Although the pattern of high fat mass paired with comparatively lower BMI may initially seem incongruous, a similar phenotype has been recapitulated in mouse models. In a 2007 study by Bischof et al, the percentage fat mass for Magel2-null mice was found to be elevated compared with wild-type littermates (mean SDs=2.8), and the percentage lean mass was found to be reduced (mean SDs=−3.2), but no significant difference was found in body weight,43 further supporting our findings.

It is thought that low bone density and abnormal body composition in PWS may be attributable to GH deficiency,44 and data generally support the use of GH therapy in PWS. The positive effect of GH treatment on body composition in PWS populations is well-established; however, little conclusive data on the effect of GH treatment on bone density exist. One study on 46 prepubertal patients with PWS measured the effects of 24 months of GH therapy and found that there was no effect on BMD45; conversely, other authors have observed that GH therapy does exert a beneficial effect on BMD and mineralisation content in PWS.46 In the SYS cohort, the two individuals on GH treatment did not consistently show markedly higher BMD measurements than average; likewise, of the four individuals who completed the total body scan, the one individual on GH treatment did not show a FMI or LMI that differed substantially from the others. However, it would be premature to draw any conclusions from such a small sample size.

The option for GH therapy in patients with SYS should be carefully considered, based on benefits, risks and potential costs. In particular, GH therapy has been thought to worsen obstructive sleep apnoea in PWS due to IGF-I-mediated tonsillar/adenoid hypertrophy, which may have contributed to spontaneous fatalities of individuals with PWS soon after they had begun GH treatment.47 Nevertheless, the success of using GH in the treatment of both physical and behavioural characteristics of PWS, and the inherent similarities between the disorders, provide a rationale for considering clinical trials to investigate whether or not GH therapy could be beneficial in SYS. Limited mobility, which is a challenge for some individuals with SYS, may affect BMD, highlighting the importance of combining pharmacological interventions with physical therapy.

Comparison of human and mouse phenotypes

A comparison of human subjects and Magel2 mouse models provides additional insight. Magel2-null mice have shown decreased weight gain and slower growth in early development (5–6 weeks of age), followed by a period of increased weight gain when compared with non-mutant littermates.48 Similarly, all nine patients with SYS reported difficulty gaining weight and feeding difficulties in early childhood development; after further development, however, patients with SYS demonstrated increased weight gain, with 5/9 patients in our cohort falling in the overweight BMI range, and 1/9 falling in the obese range. Interestingly, it has been shown that Magel2-null mice do not become morbidly obese, and loss of Magel2 has actually been found to cause hypophagia, in contrast to the hyperphagia that is characteristic of PWS.43 Likewise, no individuals in this cohort of patients with SYS reported hyperphagia. Murine models have suggested a modestly impaired intraperitoneal glucose tolerance test (IPGTT) in male Magel2 null mice and normal IPGTT in female Magel2 null mice.48 Our study found elevated postprandial glucose in 33% of patients (two males/one female). Magel2-null female mice have lower circulating IGF-1 levels than Magel2-null male mice, which was recapitulated in both male and female human subjects with MAGEL2 mutations in this study.48


This article is dedicated to patient 1, Kaylee Mitchell, who passed away unexpectedly at the age of 8 years. With Kaylee’s help, we gained more insight into Schaaf-Yang syndrome, allowing us to provide better care and guidance for generations to come. The authors thank the individuals and the families with Schaaf-Yang syndrome, who have provided clinical data and overwhelming support for our efforts. The authors also thank the phlebotomists, X-ray technicians and clinical radiologists for their help in acquiring and interpreting the data.



  • Contributors JMMcC performed the study and data analysis, and wrote the initial version of the manuscript. BMMcC-C participated in conceiving the study, and edited the manuscript. MER edited the manuscript and re-analysed the data. JY analysed data and generated figures. C-AC participated in performing the study. MAA participated in performing the study. HTNN analysed data. JLM provided guidance in study design, interpreted data and edited the manuscript. CPS conceived the study, supervised the study, acquired funding and edited the manuscript.

  • Funding This study was supported by the Foundation for Prader-Willi Research, and NIH grant U54HD08302 (IDDRC, Clinical Translational Core).

  • Competing interests None declared.

  • Ethics approval Baylor College of Medicine, IRB.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement There are no additional, unpublished data from the study.