Clinical and molecular characterization of hepatic glycogen storage disease in Saudi Arabia

TL;DR

This study characterizes the clinical and genetic features of hepatic glycogen storage diseases in Saudi children, highlighting the severity of most subtypes and identifying common genetic variants.

Contribution

The study provides the first detailed clinical and molecular analysis of hepatic GSD in Saudi Arabia, identifying novel and founder gene variants.

Findings

GSD Ia was the most common subtype, with severe clinical manifestations and complications such as liver transplantation and renal issues.

Eight novel genetic variants were identified, including a common founder variant in the G6PC1 gene prevalent in Aseer Province.

Most hepatic GSD subtypes in Saudi Arabia are associated with severe phenotypes, except for GSD VI which showed a mild phenotype.

Abstract

The paucity of data on glycogen storage diseases (GSDs) from Arabs prompted us to report on hepatic GSD to characterize its clinical and molecular features and outcomes among Saudi children and to evaluate genotype‒phenotype correlations.

We retrospectively reviewed the charts of 65 children (37 females) with genetically confirmed hepatic GSD who presented between 2008 and 2020 and were followed up for a median duration of 9 years (range: 0.4–21 years).

The most common hepatic GSD in our cohort was GSD Ia (37%), followed by GSD III (20%), GSD Ib (12.3%), and GSDVI (10.8%). Twenty-seven variants were identified (8 novel and 4 from the common ancestor, i.e., “founder in nature”). The most common founder variant is P.(Arg83Cys) in the G6PC1 gene (20% of the 65 GSD patients), clustering in Aseer Province. Six patients underwent liver transplantation (due to difficulty controlling hypoglycemia in 5 GSD Ia patients and severe portal hypertension in one GSD IV patient). One patient with GSD type 1b developed hepatic adenoma at the age of 17 years. A patient with GSD IXc developed portal hypertension at the age of 5 years, and one patient with GSD IXa developed cirrhosis. Renal complications developed in 18 patients. An echocardiogram was performed in 16 patients and revealed mild–moderate asymptomatic left ventricular hypertrophy in 5 patients. The majority of the hepatic GSD cases in our cohort manifested a severe phenotype (hepatomegaly, hypoglycemia, ± systemic involvement); only the 7 GSD VI patients manifested a mild phenotype (hepatomegaly without hypoglycemia). No “genotype‒phenotype correlations” could be observed when the two common G6PC1 gene variants [p.(Arg83Cys) versus p.(Gln20Arg)] were compared.

With the exception of GSD VI, all the hepatic GSD subtypes in Saudi Arabia are associated with a severe phenotype. Identification of the predominant gene variants and their geographic distribution in any population is likely to facilitate rapid molecular analysis by future targeting of that specific mutation.

Figures

Geographic distribution of the 65 cases of hepatic GSD in Saudi Arabia.The figure is similar but not identical to the original image and is therefore for illustrative purposes only.

Exon map of gene variants identified in GSD patients (A, forward strand; B, reverse strand) and conservation of mutated amino acids across species (C, novel variants and VUS only).(A) Detailed exon map of genes in the forward direction, with the solid bars representing exons (left to right) and arrows showing the variants at specific exons. The number beside the gene name represents the total number of exons in that particular gene. (B) Detailed exon map of genes in the reverse direction, with the solid bars representing exons (right to left) and arrows showing the variants at specific exons. The number beside the gene name represents the total number of exons in that particular gene. (C) Only VUS and novel variants are shown here. Part of the gene nucleotide sequence obtained from NCBI showing the conservation of the mutated amino acid (see arrows) across species.

Full text

Introduction

Glycogen storage diseases (GSDs) constitute a group of rare inherited inborn errors in glycogen metabolism caused by deficiencies in enzymes or transporters, which are essential for the synthesis and degradation of glycogen [1]. There are various clinical phenotypes of GSD that are classified on the basis of the deficient enzymes and affected tissues into the following: hepatic form, muscular form (skeletal and/or cardiac muscle), or both. Hepatic GSD includes several genetically heterogeneous disorders (GSD 0, GSD I, GSD III, GSD IV, GSD VI, GSD IX, and GSD XI) with overlapping clinical, biochemical, and histological features, which makes the identification of a specific hepatic subtype of GSD challenging for clinicians. Therefore, the definite diagnosis of a hepatic subtype of GSD requires molecular genetic analysis or the demonstration of specific enzyme deficiency in liver tissue. It is important to identify the exact diagnosis to provide personalized care and appropriate family counseling, as these subtypes of hepatic GSD vary in their treatment, natural course, and outcome.

Over the past two decades, more knowledge has been gained on the pathophysiology, disease course, treatment, and complications associated with GSDs. These data originate from different populations and ethnicities [2–9]. Data on GSD among Arabs are scarce and limited to two reports on GSD III from Egypt and Tunisia [10,11]. The current knowledge on GSD in Saudi Arabia derives mainly from case reports on a single PHKG2 mutation causing GSD IX [12] or small single-center case series on a single AGL mutation causing GSD III [13]. However, no reports have yet been published on other types of GSDs; particularly, data on long-term outcomes are lacking. The scarcity of data from Arabs prompted us to report 65 confirmed cases of various subtypes of hepatic GSD to characterize the epidemiological, clinical, laboratory, and molecular features and outcomes of these diseases among Saudi children and to evaluate genotype‒phenotype correlations.

Methods and patients

Study settings and design

This retrospective study was based on an analysis of data collected from patients’ electronic medical records. The study included 2 tertiary care centers in Riyadh city that receive patients with suspected GSD from different regions across Saudi Arabia.

Study population

We identified all patients with confirmed hepatic GSD who were referred to either hospital between 2008 and 2020. The inclusion criteria were as follows: 1) under 14 years of age at the time of diagnosis and 2) confirmed diagnosis of hepatic GSD on the basis of genetic testing. The exclusion criterion included suspected GSD that was not confirmed via molecular testing.

Study procedures

a) The medical records of the patients were accessed between 1/6/2020 and 31/12/2021 to collect the following data: (1) demographic and clinical characteristics; (2) laboratory data at presentation: total and direct bilirubin, alanine transaminase (ALT) [normal < 55 units/l], aspartate transaminase (AST) [normal < 34 units/l], gamma-glutamyl transferase [normal < 60 units/l], serum triglycerides (normal <1.7 mmol/L), serum cholesterol (normal <4.4 mmol/L), serum uric acid (normal <420 µmol/L), serum lactate (normal < 2.2 mmol/l), and creatine phosphokinase (normal <250 u/l); (3) imaging findings, including echocardiogram; (4) histopathological findings; (5) results of gene testing; (6) treatment provided; (7) complications that developed during the disease course; (8) date at last follow-up; and (9) final outcome. These data were compiled into an Excel spreadsheet.b) Molecular genetic Investigations

Throughout the study period, blood samples were examined by single-gene sequencing, targeted analysis of pathogenic variants, next-generation sequencing (GSD panel), or whole-exome sequencing (WES). The identified variants were classified and interpreted according to the ACMG guidelines [14]. Briefly, ACMG criteria that applied to the genetic variants in the current study included PVS1 (pathogenic very strong), PM2 (pathogenic moderate), PM3 (pathogenic very strong), PM4 (pathogenic moderate), PM5 (pathogenic moderate), PP2 (pathogenic supporting), PP3 (pathogenic moderate), PP5 (pathogenic supporting), and BP6 (benign supporting). All the variants were checked in the KFMC-in-house allele frequency (MAF) database (4500 WES cases containing 2.2 million variants) to rule out polymorphisms (MAF > 1%). For variants of unknown clinical significance (VUS) and novel variants, conservation across species was assessed as additional evidence for pathogenicity. Where available, parental blood DNA samples were analyzed for familial segregation by Sanger sequencing. All procedures were conducted with informed consent.

c) Histopathological studies

The liver samples were fixed in 10% buffered formaldehyde, paraffin-embedded, and stained with hematoxylin and eosin, Masson’s trichrome stain for fibrous tissue and Perls’ method for iron, reticulin, and periodic acid–Schiff diastase. The biopsy was evaluated for the presence of distension of hepatocytes with glycogen, steatosis, glycogenated nuclei, and portal fibrosis (stage 0: no fibrosis; stages I–II: mild fibrosis; stages III–IV: advanced fibrosis).

Study outcomes.

The primary outcomes included the following: 1) survival with native liver (defined as starting at birth and ending at death, liver transplantation (LT), or last follow-up in patients with native liver), 2) need for LT, and 3) death due to complications of the disease.The secondary outcomes included a) the development of complications during the disease course and b) the presence of genotype‒phenotype correlations.

Clinical phenotype

The clinical phenotype of the disease was categorized into the following groups on the basis of the clinical presentation, disease course, and outcome at the last follow-up:

a) “Mild phenotype” (defined as the presence of hepatomegaly without hypoglycemia, systemic complications, or progression to advanced liver disease)b) “Severe phenotype”, if any one of the following occurred:

• Hepatomegaly and difficulty in controlling hypoglycemic episodes that require continuous overnight feeding, cornstarch during sleeping, or LT.
• Progression of liver disease to cirrhosis, as evident by Grade III/IV fibrosis on liver biopsy; development of portal hypertension, as evident by splenomegaly and thrombocytopenia [platelet count < 150,000/ml]; or development of upper gastrointestinal bleeding due to esophageal or gastric varices)
• Hepatomegaly with mild hypoglycemic episodes and the development of a systemic complication (kidney, heart, or bone disease, neurodevelopmental delay, short stature). c)“Intermediate phenotype”, a status between mild and severe phenotypes characterized by hepatomegaly and mild hypoglycemia without the development of systemic complications or advanced liver disease.

Ethical consideration

The local review board of King Fahad Medical City approved the study (IRB Log number 17–093). Owing to the retrospective design of the study, the ethics committee waived the requirement for informed consent.

Statistical analysis

The data analyses were performed using the Statistical Package for Social Sciences, version 26 (SPSS, Armonk, NY: IBM Corp.). Descriptive statistics are presented as numbers, percentages, means and standard deviations. The relationships between the type of GSD and the different characteristics of the patients were determined using Fisher’s exact test for dichotomous data. Differences between continuous data were analyzed using the unpaired two-tailed t test for normally distributed continuous data and the Mann‒Whitney U test or Kruskal‒Wallis test for nonnormally distributed data. A P value <0.05 was considered to indicate statistical significance.

Results

General characteristics of the study cohort

During the study period, 81 patients were evaluated for the possibility of hepatic GSD. The diagnosis of hepatic GSD was confirmed in 65 patients (61 families; 62 were Saudi, two were Sudanese and one was Syrian; 37 females). The median age at first presentation was 10 months (range: 1 month–12 years). A diagnosis was made during the neonatal period in 3 patients [GSD 1(a), GSDIII, GSD IV)] on the basis of screening of an affected older sibling. A family history of GSD was positive in 28% of the 62 families, and consanguinity was observed in 75%. The most common type of hepatic GSD in our cohort was GSD Ia (37%), followed by GSD III (20%), GSD Ib (12.3%), and GSDVI (10.8%). GSD types IX (9.2%), IV (7.7%) and XI (3%) were less common in our cohort. Overall, the most common presenting clinical finding was hepatomegaly (84.5%), followed by hypoglycemia (39.7%). The median age at the time of the last follow-up was 9 years (0.4–21 years). Fig 1 shows the geographic distribution of hepatic GSD in the 13 provinces of Saudi Arabia. One-third of the cases (31%) originated from the southwestern region with prominent clustering of GSD1a, and 25% of the cases originated from the northern provinces without obvious clustering of a specific type. All 5 GSD IV cases were from 5 unrelated families from one tribe in Al Madinah AlMunawarah Province.

Geographic distribution of the 65 cases of hepatic GSD in Saudi Arabia.The figure is similar but not identical to the original image and is therefore for illustrative purposes only.

Clinical Phenotype

Tables 1,2 show a comparison of several clinical and laboratory variables related to hepatic GSD in our cohort. An earlier age of onset was more common in GSD Ia and GSD 1b patients, whereas a later age of onset was more common in GSD III and GSD VI patients (P < 0.001). The mean values of lactate (8.22 ± 4.84, p = 0.001), uric acid (386.1 ± 141.4, p = 0.031) and serum triglyceride (8.90 ± 9.28 p = 0.001) were significantly greater in GSD type Ia patients, whereas the mean values of ALT (525.9 ± 301.3, p < 0.001), AST (721.8 ± 788.4, p = 0.008) were significantly greater in GSD type III patients. The median serum CPK level was significantly greater in the GSD III group than in the GSD Ia, GSD Ib, and GSD VI groups [557 U/L, (60–2086), p < 0.001]. No statistically significant differences were observed in the comparisons of other variables. Types IV, IX, and XI were less common in our cohort (Table 2). The ALT value at presentation was normal in 16 of the 65 patients (25%) [GSD Ia = 6, GSD Ib = 4, GSD VI = 3, GSD IX = 1, GSD XI = 2]; 4 of the 16 patients presented elevated ALT levels at follow-up. Ten of the 49 patients with elevated ALT at presentation (21%) had normal ALT at the time of last follow-up.

Table 1: Clinical presentation of the most common hepatic GSDs.

Table 2: Clinical and laboratory features of the hepatic GSD IV, GSD IX, and XI.

Systemic involvement/complications

Table 3 summarizes the complications/comorbidities associated with the GSD patients in our cohort that developed over the follow-up period. A total of 6 patients (9.2%) underwent LT. Five patients had GSD1a [four harboring c.247C > T (p.Arg83Cys), one harboring c.620del (p.Lys207Argfs26)] associated with poor metabolic control, severe hypoglycemia and growth failure (median age of 5 years; range: 0.5–5 years), and one patient with GSD IV was transplanted at the age of 2 years because of worsening portal hypertension. One patient with GSD type 1b developed hepatic adenoma at the age of 17 years (S1 Fig) and needed surgical resection. Another patient with GSD IX [PHKG2: c.913dup (p.Arg305Profs84)] developed portal hypertension at the age of 5 years. Gallbladder stones were observed in one of the 7 patients with GSD VI.

Table 3: Systemic involvement and complications associated with GSD-subtypes over the follow up period.

Renal complications developed in 18 of the 65 hepatic GDS patients (28%): 13 GSD 1(a) patients developed large echogenic kidneys, including one patient with nephrocalcinosis and two with nephrolithiasis, and two patients developed chronic kidney disease resulting in renal failure. Renal involvement in the 2 patients with GSD XI was in the form of proximal renal tubular acidosis. An echocardiogram was performed in 16 patients and revealed abnormalities in 5 patients (31%) in the form of asymptomatic mild–moderate left ventricular hypertrophy not requiring medical therapy. Neurodevelopmental delay with intellectual disability was a feature in some patients with GSDs I, III, IV, or IX. Neutropenia was a unique common feature (65%) in GSD I(b) patients, leading to recurrent infections. Growth impairment was a common manifestation in GSD I(b) and occurred less frequently in other forms of hepatic GSD.

Liver histopathology

Liver biopsy was performed in 13 patients (median age at biopsy: 2.5 years, range: 8 months–8 years) (S1 Table). All the biopsies revealed hepatocyte distension with glycogen and variable degrees of glycogenated nuclei and steatosis. The stage of portal fibrosis was 0–1 in the GSD I (S2 Fig) and GSD VI patients (S3 Fig) who underwent liver biopsy. The most common findings were distension of hepatocytes with glycogen (93.75%), glycogenated nuclei (54%), fatty infiltration (38.5%), and portal fibrosis I/II (56.25%). Two patients with GSD III underwent liver biopsy, and the stage of fibrosis ranged between 3 and 4. Portal fibrosis ranged from stage 2–4 in the 3 patients with GSD IX (stage 2 in one patient with PHKB and stage 3–4 in PHKA2 [S4 Fig] and PHKG2, one each).

Molecular analysis

Fig 2 presents the exon map of the 27 homozygous mutations identified in the 9 genes causing hepatic GSD (8 are novel variants). S5 Fig shows the distribution of the GSD subtypes and variants among the 65 Saudi children. For further confirmation of the pathogenicity of the variants, parental testing was performed using targeted mutational analysis. The most common mutation type was missense (13 variants = 48%), followed by small deletions or duplications resulting in frameshifts (6 variants = 22%), splice site mutations (4 variants = 15%) predicted to be splice acceptor site mutations, nonsense mutations resulting in premature stop codons (3 variants = 11%), and only one in-frame deletion and insertion of an Alu repeat (Table 4). The most commonly encountered variants were c.247C > T p.(Arg83Cys) and c.59A > G, p.(Gln20Arg) in the G6PC1 gene, c.1768 + 1 G > A in the PYGL gene, and c.998A > T, p.(Glu333Val) in the GBE1 gene. These variants occurred in apparently unrelated families from several tribes in Saudi Arabia, indicating that these mutations probably came from common origins, i.e., “founder gene mutations”. The two founder G6PC1 variants originated from the southwestern region of Saudi Arabia; specifically, the variant p.(Arg83Cys) was the predominant variant in Aseer Province, and the variant p.(Gln20Arg) was the only variant in the GSD Ia cases from Najran Province.

Table 4: Hepatic GSD genetic variants (n = 65 patients, 61 families, 27 variants).

Exon map of gene variants identified in GSD patients (A, forward strand; B, reverse strand) and conservation of mutated amino acids across species (C, novel variants and VUS only).(A) Detailed exon map of genes in the forward direction, with the solid bars representing exons (left to right) and arrows showing the variants at specific exons. The number beside the gene name represents the total number of exons in that particular gene. (B) Detailed exon map of genes in the reverse direction, with the solid bars representing exons (right to left) and arrows showing the variants at specific exons. The number beside the gene name represents the total number of exons in that particular gene. (C) Only VUS and novel variants are shown here. Part of the gene nucleotide sequence obtained from NCBI showing the conservation of the mutated amino acid (see arrows) across species.

The vast majority of the hepatic GSD cases in our cohort manifested a severe phenotype, with few exceptions. All 7 unrelated GSD VI patients (harboring the founder PYGL variant c.1768 + 1 G > A from different regions in Saudi Arabia) manifested a mild phenotype characterized by hepatomegaly without hypoglycemia or significant complications/comorbidities (Table 4). Among the 7 variants causing GSD I(a), a missense variant [c.770C > T, p.(Pro257Leu)] led to a mild phenotype. Among the 8 variants causing GSD III, two missense variants [p.(Tyr495Asn) in one patient and p.(Trp1451Cys) in 3 unrelated patients] led to a mild (hepatomegaly and elevated liver enzymes alone without hypoglycemia over a 5-year follow-up period) and intermediate phenotype (hepatomegaly and mild hypoglycemia without the development of systemic complications or advanced liver disease over a follow-up period of 14–16 years), respectively (Table 4). Further subanalysis of patients with GSD Ia did not reveal a clear genotype‒phenotype correlation (S2 Table). Five of the GSD IX patients (PHKG2 = 4, *PHKA2 *= 1) manifested a severe phenotype, and one patient with a PHKB variant manifested an intermediate phenotype over a median 12-year follow-up period.

Treatment

The most common treatment modality used was uncooked cornstarch (76.5% of all the subjects, 94.4% of the subjects with GSD I), followed by frequent carbohydrate meals (67.6%). A high-protein diet and restriction of lactose and fructose were recommended for all patients with hypoglycemia. Allopurinol was used for hyperuricemia in 55.6% of the children with GSD Ia. Sodium or potassium acetate was used to treat the accompanying acidosis. Neutropenia was documented in 5 of the 8 patients with GSD Ib, and G-CSF was used for severe neutropenia in four patients.

Long-term survival of patients with hepatic GSD

In addition to the 6 patients (9.2%) who underwent LT, only patients with GSD III (1.5%) died at the age of 2 years due to sepsis. The remaining 58 patients (89%) were still alive with a native liver [median age at the time of the last follow-up was 9 years (range: 0.4–21 years)].

Discussion

Our study is the first study from the Arabian Peninsula that presents a comprehensive analysis of hepatic GSD in a large sample of Saudi patients. Our study highlights important observations. First, our molecular analysis confirmed that hepatic GSDs (particularly GSD Ia, GSD Ib, GSD III, GSD IX and GSD XI) are highly genetically heterogeneous disorders with a large spectrum of mutations in the same population. On the other hand, the 5 unrelated families with GSD IV and the 7 unrelated families with GSD VI harbored a single pathologic variant, p.(Glu333Val) and c.1768 + 1 G > A, respectively. Second, GSD Ia is the most common type of hepatic GSD in Saudi Arabia (37%), similar to the data from Japan. Brazil, and China [615], in contrast to Iran and Egypt, where GSD III is the most predominant GSD [9,11], and GSD Ib in the Serbian population [16]. Third, 19 of the 24 GSD Ia cases (80%) originated from the southwestern region of Saudi Arabia, and the missense homozygous variant p.(Arg83Cys) was the predominant cause in 13 of the 19 cases (68%) from 12 unrelated families, suggesting that this variant came from a common ancestor, i.e., “founder in nature”. Furthermore, we identified clustering of GDS IV cases only in Al Madinah AlMunawarah Province due to a missense variant p.(Glu333Val) affecting 5 unrelated families from a single tribe that has not previously been reported in other populations, strongly suggesting that this variant is a founder in nature. Identification of the predominant gene variants and their geographic distribution in any population is likely to facilitate rapid molecular analysis by future targeting of that specific mutation. Another important finding from our data is that a single patient with GSD IXa, which was previously recognized to cause mild liver disease, developed cirrhosis as early as 3 years of age.

In addition to the abovementioned founder variants, our study cohort also revealed another two likely founder variants in G6PC1 [c.59A > G, p.(Gln20Arg)] and PYGL [c.1768 + 1 G > A]. Recurrent founder mutations occur in populations with high consanguinity, such as Saudi Arabia, where the prevalence of consanguineous marriage is almost 60% [17], which presents a major risk factor for autosomal recessive diseases. Consanguinity was reported in 75% of our study cohort, and 28% reported a history of GSD in at least one close relative. The founder variants vary across different ethnic groups, where there is a high rate of consanguinity, such as the “4455delT” variant in the AGL gene (GSD III), which occurs with high frequency in the Tunisian population [10], and “W1327X” known as the disease-causing variant for GSD III in the Turkish population and German–Ukraine patients [18,19]. The most common founder variant in our study cohort [p.(Arg83Cys) affecting the G6PC1 gene] is not unique to the Saudi population and was found in several populations [7,20–22], suggesting that this variant is of common ancestral origin worldwide. The p.Arg83Cys variant constituted 50% of the G6PC disease-causing variants in our study and in French and Tunisian patients [23,24], 80% of Sicilian, and 100% of G6PC variants in Ashkenazi Jewish patients [25,26].

The correlation between specific pathogenic variants and the clinical phenotype of GSD has been investigated in few studies. There is increasing evidence in the literature linking variants in PHKG2 with more severe phenotypes than variants in the PHKA2 and PHKB genes, which are usually associated with milder phenotypes [5]. Several studies have shown that GSD III lacks clear links between the genotype and clinical phenotype [27–29]; however, an association between exon 3 mutations and GSD IIIb (only the liver is involved) has been reported previously [30,31]. According to our data, the two founder variants in G6PC [p.(Arg83Cys) and p.(Gln20Arg), constituting 75% of all mutations in all 24 GSD Ia patients], were associated with a severe phenotype. The p.(Arg83Cys) variant is located in the active center of the enzyme G6 Pase and presented no detectable activity in transient expression assays [32]. The second variant [c.59A > G, p. (Gln20Arg)] is a nonhelical mutation that also presented no detectable G6 Pase activity [32]. The complete absence of enzyme activity as a result of both variants might explain the resulting severe phenotype in all our 18 patients harboring the 2 founder variants. On the other hand, the variant c.770C > T, p.(Pro257Leu) led to a mild phenotype in 2 siblings (hepatomegaly, no hypoglycemia, normal uric acid, nearly normal lactate and no kidney involvement) because this mutation retained residual phosphohydrolase activity (6.1% of the wild type) [33].

According to our data, the 2 missense variants in the AGL gene led to a mild–intermediate phenotype, and the nonmissense variants resulting from splicing, frameshift, or nonsense modifications resulting in a truncated protein led to a severe phenotype. We hypothesize that missense mutations in the AGL gene cause a minor reduction in glycogen debranching enzyme activity, whereas nonmissense variants lead to a marked reduction in enzyme activity. All the variants in the G6PC1 gene led to a severe phenotype, except for the missense variant p.(Pro257Leu) in 2 siblings, which led to a mild phenotype over a 6- and 13-year follow-up period. There was no evidence of a genotype‒phenotype correlation within GSD I(a) patients when the two most common G6PC1 missense variants were compared (Supplementary S1 Table). However, long-term, natural history, large multicenter studies are needed to better understand the relationships between genotype and clinical presentation and outcome.

Compared with other types of hepatic GSD, GSD VI is usually mild with a benign course, and long-term complications, if they occur, are the exception. Consistent with the literature, all 7 unrelated cases of GSD VI in our study, which were from various regions in Saudi Arabia, manifested mild phenotypes, and none developed major complications over a median 10-year follow-up period. Only one patient developed mild asymptomatic cardiomyopathy, which has rarely been described [34]. However, in the literature, more severe cases with recurrent hypoglycemia, liver cirrhosis or developmental delay have been reported [35–37]. In a systematic literature review of 63 patients with GSD VI and 37 patients who underwent liver biopsy, fibrosis of various degrees was found in 32%, and early cirrhosis was diagnosed in approximately 10% of patients as early as preschool age [38]. Like in GSD VI, in the past, the clinical picture of GSD IXa was considered a benign and self-limited disorder; however, more recent reports linked GSD IXa with a greater prevalence of fibrotic or cirrhotic changes than did GSD VI [34,39,40]. Our single case of GSD IXa associated with early cirrhotic changes on liver biopsy at 3 years of age supports this observation; therefore, we highly recommend close monitoring for long-term liver complications in patients with GSD IXa. Unlike GSD IXa and GSD IXb, GSD IXc [due to deficiency of the γ subunit of phosphorylase kinase (PK) enzyme, encoded by the PHKG2 gene) is a more severe disorder with a high prevalence of liver fibrosis that can progress to cirrhosis during childhood and the need for LT [5]. Furthermore, there is a report of hepatocellular carcinoma development in a 27-year-old man with GSD IXc [5]. On the other hand, no patients with GSD VI, GSD IXa or GSD IXb were reported to have developed liver failure or hepatocellular carcinoma and received a liver graft. Two of our 4 patients with GSD IXc developed advanced liver fibrosis during early childhood, as evidenced by the cirrhotic changes in the liver biopsy in one patient at 3 years of age and the development of portal hypertension in the other patient by 5 years of age. These findings provide further evidence that GSD IXc patients with mutations in the γ subunit of the PK enzyme suffer from more severe liver disease at an earlier age than patients with mutations in the α and β subunits encoded by the PHKA2 and PHKB genes, respectively. A possible explanation for this disparity in severity between GSD IXc patients and GSD IXa and GSD IXb patients is that the γ subunit plays an important role in PK enzyme functionality because it houses the catalytic site, whereas the other subunits are involved in regulation [41].

Although our study provides valuable insights into the clinical and molecular characteristics of hepatic GSD in Saudi Arabia, it has several limitations. In addition to the retrospective design of the study, which has inherent limitations, the lack of consistency in assessment and monitoring for complications/comorbidities via standardized and systematic methods (e.g., regular performance of echocardiogram, neuromuscular/neurodevelopmental evaluation) precluded proper documentation of cardiac, neurodevelopmental, and muscular involvement. Histopathological examination of the liver was available for only 20% of the 65 patients, which might have led to an underestimation of the frequency of liver fibrosis reported in our study. Furthermore, the median follow-up time in our cohort was relatively short (≈ 9 years) and did not extend beyond 20 years of age. All these factors could have contributed to an underestimation of long-term complications. Another limitation is the limited availability of dietary management data, which has precluded assessment of the effects of dietary therapy on liver function tests and the occurrence of complications. Furthermore, the small sample size for some individual GSD subtypes did not allow for a meaningful comparative analysis to evaluate genotype‒phenotype correlations.

In conclusion, our results identified founder variants and provided geographic mapping of hepatic GSD in Saudi Arabia, which is likely to facilitate rapid molecular analysis by future targeting of that specific mutation. Longitudinal, prospective, multicenter studies are needed to better understand the natural history of hepatic GSD, perform regular surveillance for the development of hepatic and extrahepatic complications, evaluate “genotype‒phenotype correlations” among several variants leading to GSD, and examine the effects of various treatment approaches and proper metabolic management. The results from these studies will enable physicians to provide personalized care and appropriate family counseling.

Supporting information

S1 FigHepatic adenoma in a 17-year-old girl with GSD Ib.(A) Tissue biopsy from the resected hepatic adenoma (arrow) is sharply demarcated from the adjacent nontumor liver tissue (arrowhead) (hematoxylin‒eosin, magnification 40x). (B) Tissue biopsy from a hepatic adenoma consisting of benign hepatocytes without portal tracts (hematoxylin‒eosin, magnification 40x). (C) Focal hypoechoic image of a 3 cm × 4 cm lesion on ultrasound of the liver. (D) A hyperdense rounded lesion on CT of the liver.(JPG)

S2 FigLiver biopsy from a patient with GSD Ib performed at the age of 8 years.(A) Glycogenic hepatocyte distension accompanied by the presence of glycogenated nuclei (arrow) and mild macrovesicular steatosis (arrowhead) (hematoxylin‒eosin, magnification 100x). (B) No significant fibrosis (arrow) (Masson’s trichrome stain, magnification 100x).(JPG)

S3 FigLiver biopsy from a patient with GSD VI performed at 8 years of age.(A) Hepatocytes appear enlarged and pale with cytoplasmic clearing (arrow) secondary to marked accumulation of glycogen (hematoxylin‒eosin, magnification 100x). (B) Stage 1 fibrosis (arrow) [Masson’s trichrome stain] and macroscopic steatosis (arrowhead).(JPG)

S4 FigLiver biopsy from a patient with GSD XIa performed at 3 years of age (A) Enlarged hepatocytes with clear cytoplasm (arrow).Some hepatocytes have light eosinophilic haziness to their cytoplasm (arrowhead). Both appearances can occur during hepatocyte overglycogenation (hematoxylin‒eosin, magnification 100x). (B) Bridging fibrosis (arrow), Stage 3/4 (Masson’s trichrome stain).(JPG)

S5 FigDistribution of the GSD subtypes and variants among children in Saudi Arabia.(TIF)

S1 TableComparison of two G6PC1 gene variants.(DOCX)

S2 TableHistopathological features of 13 GSD patients.(DOCX)

S1 DataGDS Data Set (Plos One).(XLSM)