Epilepsy Is Part of the CNS Phenotype in Classic Infantile Pompe Disease

Abstract

Background and Objectives

Enzyme replacement therapy has not only significantly improved motor outcome and survival in patients with classic infantile Pompe disease, but also revealed previously unrecognized central nervous system (CNS) involvement. In this international study, involving patients from the Netherlands, Italy, Argentina, Germany, the United Kingdom, and Taiwan, we investigated whether epilepsy should be considered part of the CNS phenotype.

Methods

We included patients with classic infantile Pompe disease, defined by the presence of hypertrophic cardiomyopathy, symptom onset < 6 months of age, complete acid α-glucosidase (GAA) deficiency, and/or 2 severe variants in the GAA gene, who developed epilepsy. Data on epilepsy characteristics, electroencephalogram (EEG), cognitive testing, serum neurofilament light chain (NfL), and brain magnetic resonance imaging (MRI) were retrospectively collected.

Results

Seventeen patients from 10 centers were identified. The median follow-up duration was 13.7 years (range 3.3–19). Seven patients had deceased at the time of analysis. The median age at first seizure was 11.5 years (range 2.5–17.5). Seizure semiology was variable: six patients experienced generalized tonic-clonic seizures and 3 focal seizures with impaired consciousness only; 6 had multiple seizure types, and 7 experienced seizures during fever or infection. Seizure frequency varied considerably (in 9 occasionally, 5 monthly, 2 weekly, 1 daily). The most common EEG findings were a slowed background activity and focal epileptiform discharges, not substantially activated by sleep. Levetiracetam was most frequently used as antiseizure medication. Overall, 70% of patients became seizure-free. Serum NfL was elevated in all 6 patients in whom it was measured, and 8 of 10 patients had an intelligence quotient ≤66 at onset of epilepsy. Although brain MRI was not always performed at the age of first seizure, 14 of 15 patients showed white matter abnormalities, which were extensive in 11 of 14 (score ≥7/12). Brain atrophy was present in 9 cases and calcifications in 4.

Discussion

Our findings suggest a potential increased frequency of seizures in classic infantile Pompe disease in comparison with unaffected children, occurring predominantly after the age of 7, and that epilepsy is part of the CNS phenotype. The risk of seizures should be evaluated during follow-up in long-term survivors with classic infantile Pompe disease.

Introduction

Pompe disease is a lysosomal storage disorder caused by a deficiency of the enzyme acid α-glucosidase (GAA),¹ leading to accumulation of glycogen, impairment of autophagy, and resultant pathologic cascades that ultimately result in muscle damage and weakness. The phenotypic spectrum ranges from the classic infantile form, characterized by onset within the first months of life and a hypertrophic cardiomyopathy at presentation, to the late onset form, which can manifest during childhood or adulthood and is typically not associated with cardiomyopathy at diagnosis. Classic infantile Pompe disease, if untreated, is a fatal disease which typically leads to death by 12 months of age.² Since implementation of enzyme replacement therapy (ERT) after the European Medicines Agency (EMA) and Food and Drug Administration (FDA) approval of alglucosidase alfa in 2006, the natural course of the disease has been transformed, enabling many patients to achieve independent ambulation and allowing some to survive into their twenties.^3,4 However, alglucosidase alfa does not cross the blood–brain barrier and, over the past 2 decades, it has become clear that CNS involvement is an increasingly common feature in classic infantile Pompe disease.^5-12

Although early postmortem studies in untreated infants demonstrated glycogen accumulation in neurons of the brainstem, thalamus, basal ganglia, hippocampus, cerebellum, and dentate nucleus,¹³ the first clinical data on CNS involvement, largely obtained in young ERT-treated children, were inconsistent. Although some case reports described periventricular white matter abnormalities^14-16 and a longitudinal study reported delayed myelination,¹⁷ 2 investigations assessing cognitive skills, with follow-up extending up to 12 years, found normal cognitive function although with reduced processing speed.^5,18-20 The description of patients with widespread white matter involvement and cognitive decline^15,20 prompted larger, long-term cohort studies, with follow-up extending to 17.5 years. This yielded variable results: although some patients maintained normal and stable cognitive skills, others developed learning disabilities or unequivocal cognitive decline.^21,22

With increasing patient age and cohort size, the long-term neurologic consequences have become increasingly evident. Characteristic white matter abnormalities have been described,^6,11,23 and longitudinal studies demonstrated significant progression over time, correlating with cognition.^6,9,11 In addition, a slow but significant decline of total intelligence quotient (IQ) and its subdomains^6,11 was observed. More recently, these cognitive and radiologic deteriorations were shown to be accompanied by a longitudinal increase of serum neurofilament light chain (NfL),^9,12 an accepted biomarker of neuroaxonal and astroglial damage and plasma glial fibrillary acidic protein (GFAP).²⁴ In addition, magnetic resonance spectroscopy has shown increased glycogen accumulation in the brain.²⁵

In some classic infantile Pompe patients epilepsy has been reported, but, so far, it is unknown whether this should be considered as part of the new evolving CNS phenotype of long-term ERT survivors.^9-11,26,27 Therefore, to investigate the occurrence and type of epilepsy associated with classic infantile Pompe disease, we collected, with an international group of expert physicians, data on patients with classic infantile Pompe disease diagnosed with epilepsy and analyzed seizure onset, frequency, semiology and outcome, as well as brain MRI findings, IQ testing results, and NfL values.

Methods

Data Collection and Analysis

We contacted international specialists treating patients with classic infantile Pompe known from former collaborations to identify patients with epilepsy. This resulted in 17 patients from 10 institutions in Germany, the United Kingdom, Argentina, Italy, Taiwan, and the Netherlands (eTable1). We used the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guideline²⁸ to draft this article, and the STROBE reporting checklist²⁹ when editing, included in eSAP.

Inclusion criteria were a confirmed diagnosis of classic infantile Pompe disease and epilepsy. The diagnosis was accepted when the following criteria were fulfilled: onset of clinical symptoms within 6 months after birth, presence of a hypertrophic cardiomyopathy, α-glucosidase deficiency in leukocytes and/or fibroblasts, and biallelic pathogenic variants in the GAA gene according to the Pompe Variant Database.^30,31 Epilepsy was diagnosed if patients had at least 2 unprovoked seizures or 1 unprovoked seizure and an abnormal electroencephalogram (EEG).

Data available from clinical charts, such as age at first symptoms, age at last follow-up, GAA variants, ERT (dose, frequency, age at start), motor outcome, nutritional and respiratory support, hearing and vision problems, epilepsy characteristics (age at onset, seizure frequency and semiology, treatment), EEG findings, cognitive testing results, brain MRI findings, and serum NfL values, were collected. All patient data were pseudonymized. Data collection was completed by April 28, 2025.

EEGs were performed according to the international 10–20 system, with electrodes placed on the scalp to record cortical electrical activity. Data were analyzed using high-resolution digital systems. Relevant EEGs were reevaluated and seizures were reclassified according to the latest update of the International League Against Epilepsy by an experienced pediatric epileptologist (AH).

MRIs (T1- and T2-weighted scans) were assessed for white matter abnormalities by using the 12-item Pompe Rotterdam brain MRI score.¹¹; the presence of calcifications and brain atrophy was also noted. MRIs were reevaluated by a child neurologist specialized in the analysis of brain MRI in classic infantile Pompe disease (J.M.P.H.).

NfL was measured in 6 patients using the Simoa method.¹²

Standard Protocol Approvals, Registrations, and Patient Consents

The study was approved by the ethics committee of the Erasmus Medical Center, Rotterdam, NL, study number MEC-2025-0279, and by The Institutional Review Board of National Taiwan University, study number 202506096RINC. The ethics committee determined that participant consent was not required. All other participating centers conducted the study in accordance with their local and national regulations governing the retrospective use of anonymized clinical data, which did not require additional ethical review or informed consent.

Authorization has been obtained for disclosure (consent to disclose) of any recognizable people (Video 1 and Video 2). Videos 1 and 2 show an individual experiencing a focal seizure with impaired consciousness.

Video 1

 Download Supplementary Video 1 ^{(1.9MB, mp4)} via http://dx.doi.org/10.1212/200432_Video_1

Video 2

Download Supplementary Video 1 ^{(1.1MB, mp4)} via http://dx.doi.org/10.1212/200432_Video_2

Data Availability

All original data are present in the article. Further enquiries can be addressed to the corresponding author.

Results

Patient Characteristics

Seventeen patients with classic infantile Pompe disease who developed epilepsy were identified across 10 centers in 6 different countries.

Our inquiry with international colleagues also identified 2 children with single fever associated seizures at the age of 1.1 and 3.3 years, respectively. No antiseizure medications (ASMs) were required, and no further seizures occurred. These patients were not included in this study because they experienced provoked seizures and did not meet inclusion criteria.

The cohort from the Erasmus Medical Center, which serves as the Dutch reference center and follows all patients with Pompe disease in the Netherlands, represented the entire Dutch population with classic infantile Pompe disease. Within this cohort, 4 of 28 (14%) ERT treated patients developed epilepsy. Data from other participating centers were not nation-based; therefore, the proportion of patients with epilepsy could not be reliably calculated.

The clinical and genetic data of the 17 patients included in the study are summarized in Table 1. In total, 24 different GAA variants were identified, either in compound heterozygosity or homozygosity. All variants are listed as associated with classic infantile Pompe disease.^30,31 The most frequent variants were c.525del and c.1935C > A.

Table 1.

Synopsis of Genetic and Clinical Data From 17 Patients With Classic Infantile Pompe Disease and Epilepsy

PT	Status	GAA variants	CRIM	ERT type and dose at start/end of follow-up	Best motor function	Motor function at last contact
1[11]	Deceased	c.del525/c.1933G>A	+	AGLU, 20 mg/kg/2 weeks/AGLU, 40 mg/kg/wk	Walking without support	—
2[11]	Deceased	c.925G>A/c.2608C>T	+	AGLU, 40 mg/kg/wk/AGLU, 40 mg/kg/wk	Climbing stairs	—
3	Alive	c.del525/c.2481 + 102_2646 + 31del	+	AGLU, 40 mg/kg/wk/AGLU, 40 mg/kg/wk	Climbing stairs	Sitting without support
4	Alive	c.del525/c.2481 + 102_2646 + 31del	+	AGLU, 20 mg/kg/2 weeks/AGLU, 40 mg/kg/wk	Standing with support	Sitting without support
5	Alive	c.2078insA/c.2078insA	+	AGLU, 20 mg/kg/2 weeks/AGLU, 40 mg/kg/wk	Climbing stairs	Sitting without support
6	Alive	c.del525/c.670C>T	+	AGLU, 20 mg/kg/wk/AGLU, 40 mg/kg/wk	Climbing stairs	Sitting without support
7	Alive	c.del525/c.2481 + 102_c.2646 + 31del	-*	AGLU, 20 mg/kg/2 weeks/AVA, 20 mg/kg/2 weeks	No motor milestone	No motor milestone
8	Deceased	c.1210G>A/c.2432del	+	AGLU, 20 mg/kg/2 weeks/AGLU, 20 mg/kg/wk	Standing with support	—
9	Deceased	c.del525/c.2237G>A	+	AGLU, 20 mg/kg/2 weeks/AGLU, 40 mg/kg/wk	Sitting without support	—
10	Deceased	c.1802C>G/c.2800-1G>C	+	AGLU, 20 mg/kg/wk/AGLU, 40 mg/kg/wk	Sitting without support	—
11[27]	Deceased	c.1465G>A/c.40_47del8	+	AGLU, 40 mg/kg/2 weeks/AGLU, 40 mg/kg/2 weeks	Climbing stairs	—
12	Alive	c.236_246del1/c.1927G>A	+	AGLU, 20 mg/kg/2 wk/AGLU, 30 mg/kg/wk	Standing with support	No motor milestone
13	Alive	c.896T>C/c.896T>C	+	AGLU, 40 mg/kg/wk/AGLU, 40 mg/kg/wk	Sit with support	No motor milestone
14	Deceased	c.1935C>A/c.2024_2026del	+	AGLU, 20 mg/kg/2 weeks/AGLU, 20 mg/kg/2 weeks	No motor milestone	—
15	Alive	c.1935C>A/c.1062C>G	+	AGLU, 20 mg/kg/wk/AVA 40 mg/kg/2 weeks	Climbing stairs	Sitting without support
16	Alive	c.1935C>A/c.1935C>A	+	AVA, 40 mg/kg/2 weeks/AVA, 40 mg/kg/2 weeks	Climbing stairs	Climbing stairs
17[9]	Alive	c.1935C>A/c.2841dup (p.L948Sfs*70)	+	AGLU, 20 mg/kg/2 weeks/AVA 40 mg/kg/2 weeks	Climbing stairs	Sitting without support

Abbreviations: AGLU = alglucosidase alfa; AVA = avalglucosidase alfa; CRIM = cross-reactive immunologic material; ERT = enzyme replacement therapy; PT = patient.

GAA variants were annotated according to RefSeq transcript NM_000152.5. *CRIM status, assessed by Western Blot on peripheral blood, gave a CRIM-negative result, despite deletion of exon 18 is predicted as CRIM positive in the Pompe Variant Database.^30,31 #c.2841dup (p.L948Sfs*70) variant is not present in the Pompe Variant Database but has been previously described.³² (NM_000152.3).

Fifteen were male participants (88%). The median duration of follow-up, from the time of initial contact until the most recent follow-up or death, was 13.7 years (range 3.3–19 years). Seven patients had already deceased by the time of analysis at a median age of 13.7 years (range 9–19). The median age at ERT initiation was 3.0 months (range 0.4–11 months); 3 of 4 Taiwanese patients were detected by newborn screening. Sixteen patients were initially treated with alglucosidase alfa and 1 with avalglucosidase alfa. ERT dose ranged from 20 mg/kg/every other week to 40 mg/kg/every week. In 13 cases, ERT dose and/or frequency were subsequently modified. ERT was switched to avalglucosidase alfa in 3 cases.

Nine patients learnt to walk. Among the 10 patients alive at last follow-up, 1 was able to climb stairs, 6 were capable to sit independently, and 3 could not perform any motor milestone. Across the entire cohort, 10 patients required assisted ventilation (5 noninvasive, 5 invasive) and 8 necessitated nutritional support (through percutaneous gastrostomy in 5, nasogastric tube in 2). Speech problems were present in 16, while hearing and visual impairments were reported in 11 and 10 patients, respectively. Six individuals with hearing deficits were provided with hearing aids and 7 wore corrective lenses.

Epilepsy Characteristics

The median age at seizure onset was 11.5 years (range 2.5–17.5 years). In 16 patients, the first seizure occurred after the age of 7 years. At the time of the first seizure, 6 patients were able to walk without support (age range 2.5–15.8 years) (Table 2).

Table 2.

Synopsis of Epilepsy Characteristics, MRI Findings, Neurocognitive Testing Results, and Serum NfL in 17 Classic Infantile Pompe Patients With Epilepsy

PT	Motor function at first seizure	Seizure phenotype	Seizure frequency	Fever/infection related	Mono-/poly-ASM therapy; ASM response (% decrease)	EEG findings	Serum NfL	Cognitive testing, age at testing, score	Brain MRI overall score, max = 12; (age at MRI in y)
1	Sitting without support	Focal impaired consciousness, “hemiconvulsion”	≥1 mo	No	Polytherapy, 75–99	1 + 2	Yes, abnormal	Vineland, 13.5 y all skills <4 y	11 (13 y)
2	Climbing stairs	Generalized tonic-clonic	<1 in 6 mo	No	Polytherapy, 100	1 + 2	Yes, abnormal	WISC-III, 9 y, IQ 53	9 (8 y)
3	Climbing stairs	Generalized tonic-clonic	≥1 mo	First seizure during IAR	Monotherapy, 100	1	Yes, abnormal	WISC-III, 6.4 y, IQ 99	7 (7 y)
4	Sitting without support	Unknown, only screaming was heard; probably generalized tonic-clonic	<1 in 6 mo	Yes	Polytherapy, 100%	1 + 2	Yes, abnormal	WISC-V, 17.1 y, IQ 63	10 (18 y)
5	Walking without support	Focal impaired and preserved consciousness	≥1 wk	Yes	Polytherapy, 100	1 + 2+4	Yes, abnormal	Not testable, IQ < 60	1 (2 y)a
6	Walking without support	Focal impaired consciousness, myoclonic, myoclonic-atonic, atonic	≥1 wk	Yes	Monotherapy, 100	1 + 2	Yes, abnormal	Culture fair test, 7.2 y, IQ 61	5 (7 y)
7	No motor milestone	Generalized tonic-clonic	Daily	No	Monotherapy, 50–74	1 + 2	—	NA	9 (12 y)
8	No motor milestone	Focal preserved consciousness, focal to bilateral tonic-clonic	<1 in 6 mo	No	Monotherapy, 100	2; background activity not assessable for artifacts	—	NA	Not performed
9	No motor milestone	Focal impaired consciousness, generalized tonic-clonic	≥1 mo	No	No ASM	Not performed	—	NA	Not performed
10	Sitting without support	Focal impaired consciousness; prolonged focal impaired consciousness	<1 in 6 mo	No	Monotherapy, 100	1 + 2+3; non convulsive status epilepticus	—	Not testable, IQ < 60	10 (12 y)
11	Walking with support	Focal impaired consciousness, generalized tonic-clonic; prolonged focal impaired consciousness	<1 in 6 mo	No	Monotherapy, 100	1 + 2; status epilepticus	—	WISC-III, 12.5 y, IQ 44	8 (14 y)
12	No motor milestone	Focal to bilateral tonic-clonic, focal preserved consciousness	≥1 mo	No	Monotherapy, 50–74	1 + 2	—	WISC-IV, 8.5 y, IQ 61	8 (13 y)
13	No motor milestone	Generalized tonic-clonic	<1 in 6 mo	Yes	Monotherapy, 100	1 + 2	—	KABC-II, 7.5 y, IQ 59	11 (10 y)
14	No motor milestone	Focal impaired consciousness	<1 in 6 mo	Yes	Monotherapy, 100	1 + 2+4	—	NA	11 (14 y)
15	Standing with support	Generalized tonic-clonic	<1 in 6 mo	No	No ASM	1 + 2+4	—	Leiter 3, 16 y, IQ 36	10 (15 y)
16	Climbing stairs	Focal preserved consciousness	≥1 mo	No	Monotherapy, 75–99	Normal	—	Leiter 3, 2.9 y, IQ 134	0 (2 y)
17	Sitting without support	Generalized tonic-clonic	<1 in 6 mo	Yes	No ASM	1 + 2	—	WISC-V, 16.7 y, IQ 66	4 (14 y)

Abbreviations: ASM = antiseizure medication; IAR = infusion-associated reaction; KABC = Kaufman Assessment Battery for Children (II edition); NA = not available data; NfL = neurofilament light; PT = patient; WISC = Wechsler Intelligence Scale For Children (III, IV or V edition), y = years.

EEG findings: 1 = abnormal background; 2 = focal epileptic discharges; 3 = generalized epileptic discharges; 4 = activation of epileptic discharges during sleep.

Four patients (patient 1, 2, 11, 17) had been previously briefly reported in literature.

^aThe MRI of patient 5 was performed 5 years before the onset of epilepsy, while for all other patients within 2 y of its onset.

Ten patients suffered from 1 type of seizure, including generalized tonic-clonic seizures (GTCS, n = 6), focal seizures with impaired consciousness (n = 3), or focal seizures with preserved consciousness (n = 1). Six patients experienced multiple seizure types. Three of them manifested focal seizures with impaired consciousness in combination with GTCS (n = 2) or focal seizures with preserved consciousness (n = 1) and 2 had a combination of focal to bilateral tonic-clonic and focal with preserved consciousness. One patient displayed a combination of focal impaired consciousness, myoclonic, myoclonic-atonic, and atonic seizures. In patient 4, no seizure was directly observed, making it impossible to define the exact seizure type. Videos 1 and 2 show an individual experiencing a focal seizure with impaired consciousness.

Two patients experienced status epilepticus (seizure duration >20 minutes). In both, prolonged focal seizures with impaired consciousness were reported as well. Facial involvement during seizures was documented in 6 individuals. In 7 patients, seizures provoked by fever occurred in addition to unprovoked epileptic seizures.

EEG recordings were available for nearly all individuals: 16 of the 17 patients had at least 1 awake EEG, and 7 of them also underwent EEG recording during sleep. The background activity was slowed in 14 patients. Focal epileptiform discharges were observed in 14 patients, while 1 patient exhibited both focal and generalized sharp waves. Among the 7 patients with EEGs during sleep, 3 displayed epileptiform discharges slightly activated by sleep. Only patient 16, who developed epilepsy early in life, had a normal interictal EEG.

Nine patients had occasional seizures, 5 monthly, 2 weekly, and 1 had daily seizures. Fourteen patients were treated with 1 (n = 10) or more (n = 4) ASMs; levetiracetam was most frequently prescribed (n = 12). Other ASMs included valproic acid, clobazam, lamotrigine, carbamazepine, topiramate, and phenobarbital. Ten of 14 individuals (71%) became seizure free, 2 showed a 75–99% reduction in seizure frequency, and 2 experienced a reduction of 50–74%.

CNS Involvement

To rule out other causes of cognitive decline or seizures, additional investigations were performed at the discretion of the treating physicians. Whole-exome/genome sequencing targeted for epilepsy (n = 1) or encephalopathy (n = 2) was conducted in 3 patients and yielded normal results. A lumbar puncture was performed in 6 patients; in all cases no signs of infection were found.

Other investigations aimed at assessing CNS involvement were variably performed across centers.

Serum NfL levels were measured in 6 patients around the time of first seizure and were elevated in all cases compared with age-matched controls.

Cognitive testing was performed in 11 patients with a variety of scales (Table 2). Formal IQ testing was performed in 10 patients, of whom 8 had an IQ ≤ 66. In 1 patient assessed with the Vineland Screener Scale, which evaluates adaptative functioning through caregiver interview, the developmental age corresponded to 2.2 years at a chronologic age of 13 years. Two further patients were considered untestable because they lacked instructional understanding, thus reflecting an IQ < 60.

A brain MRI was performed in 15 patients. In 14 patients, imaging was done within 2 years before or after seizure onset, while in 1 patient MRI was performed 5 years before epilepsy (patient 5). All patients displayed white matter abnormalities, except for the youngest who developed epilepsy at age 2.5 years (patient 16). The median MRI score was 9 of 12 (range 0–11), and white matter abnormalities were extensive (score ≥7/12) in 11 patients. Brain atrophy was present in 9 cases and calcifications in 4.

Discussion

In this study, we conducted a detailed analysis of 17 patients with classic infantile Pompe who developed epilepsy. Although the mode of data collection did not allow to determine the prevalence of epilepsy across this international cohort, analysis of the Dutch national cohort from the Erasmus Medical Center revealed a frequency (14%) among ERT treated patients which is more than 10 times higher than the rate determined for healthy children in Western Europe^33-36 and closely aligns with previous findings,¹⁰ which reported seizures in 15% of patients with infantile onset Pompe with severe CNS involvement. Together, these data suggest that epilepsy is a complication of classic infantile Pompe disease and that epilepsy may occur considerably more often in patients with the most severe form of Pompe disease.^10,33-36

Our cohort showed several characteristic features of epilepsy in classic infantile Pompe disease. The findings align closely with those described in 7 additionally reported cases from the literature (eTable 2).^10,27,37 Notably, to date, no patients with the late-onset form of Pompe disease and seizures have been reported, suggesting that the increased risk of epilepsy is confined to the classic infantile form. Most patients from our cohort were cross reactive immunological material (CRIM) positive, likely reflecting the historically poor survival of CRIM-negative patients before the introduction of immunomodulation protocols.

Epilepsy manifested in nearly all patients after the age of 7 years, i.e., later in disease course, and occurred in both walkers and nonwalkers, with no apparent correlation between seizure occurrence and motor function. We explored this relationship because patients with a poorer motor response to treatment, potentially reflecting a more severe disease phenotype, could have more pronounced CNS involvement. However, no clear association between seizure occurrence and motor outcome could be demonstrated in our cohort.

Focal seizures with impaired consciousness and presumably focal to bilateral tonic-clonic seizures were the prevailing seizure types. Facial involvement, particularly unilateral cheek twitching, was frequently observed. Seizures triggered by fever were documented in 7 patients, suggesting a potential lower threshold for seizures during fever in classic infantile Pompe disease.

Most EEGs were interictal recordings performed after the first seizure and revealed focal epileptiform discharges, mainly sharp waves located in the centrotemporal regions. These findings align with the predominantly focal seizure semiology. No or only a mild increase of epileptiform discharges during sleep was evident and generalization of sharp waves was rare. Background slow activity was frequently reported. The overall prognosis of epilepsy was favorable: despite initial seizure frequency ranging from occasional to daily, 3 quarter of our patients required no treatment or became seizure-free with 1 or 2 ASMs at the time of data analysis.

In contrast to the rest of the cohort, 1 patient (patient 16) developed epilepsy distinctly earlier, at age 2.5 years. His seizures were not fever related, interictal EEGs were normal, whole-genome sequencing did not reveal a genetic cause, and his brain MRI showed no white matter hyperintensity, brain atrophy, or other structural abnormalities. This patient represents a peculiar case requiring long-term follow-up. Another patient developed a more treatment-resistant epilepsy with multiple seizure types, resembling an epileptic encephalopathy. This mirrors 1 of the cases reported bin literature,¹⁰ suggesting that such patients represent the severe end of the epilepsy spectrum in classic infantile Pompe disease.

Epilepsy is a common manifestation of several lysosomal storage disorders, reflecting the relatively high CNS burden and the critical role of lysosomes in cellular homeostasis, particularly in neuronal and glial cells.³⁸ Although epilepsy can occur in classic infantile Pompe disease, it appears less common than in other LSDs, such as neuronal ceroid lipofuscinosis, for which seizures are one of the key symptoms of the disease.³⁹ In classic infantile Pompe disease, epileptogenesis may result from damage to different regions throughout the brain. Nevertheless, the boundaries between these structures are not always clear-cut because they are interconnected, with dysfunction in one region potentially also affecting others.

In our cohort, nearly all patients exhibited white matter abnormalities on MRI, often extensive, and many showed slowed background activity at EEG, consistent with diffuse white matter dysfunction and delayed neuronal signal transmission.⁴⁰ Cognitive impairment was noted in 8 of 10 patients (80%) with available assessments, a higher rate than previously reported.^9,11 For example, the Dutch group observed progressive cognitive decline, with IQ < 70 in 17% of participants at baseline, and 58% at last follow-up (median age 8.9 years, range 1.5–22.5),¹¹ while the Taiwanese one⁹ reported a median IQ score of 73 in 11 patients aged 10.1 years, with 18% presenting cognitive decline and an IQ < 70 at last assessment. These findings suggest that the risk of epilepsy may be higher in classic infantile patients with marked cognitive impairment. Both cognitive impairment and epilepsy may be related to disrupted white matter integrity. According to the Parieto-Frontal Integration Theory, efficient cognitive functioning depends on rapid, well-integrated communication between brain regions, mediated primarily by intact white matter tracts such as the superior longitudinal fasciculus. Diffusion tensor imaging (DTI) in a classic infantile Pompe cohort revealed abnormalities in this tract, indicating structural disruption.⁴¹ In addition, white matter damage may result in astrocyte dysfunction and accumulation of extra cellular glutamic acid and K+, thereby increasing neuronal excitability.⁴² Such extracellular imbalances and disrupted connectivity might lower the threshold for epileptogenesis in some patients, although it does not fully explain why other patients with severe white matter abnormalities remain seizure-free, suggesting that white matter abnormalities alone may be insufficient to explain seizure susceptibility. Indeed, 1 patient in our study developed epilepsy within the first years of life and presented a normal brain MRI, indicating that seizures may occur without white matter abnormalities.

Cognitive impairment and epilepsy may also be related to gray matter damage, which can lead to neuronal dysfunction and abnormal electrical discharges.⁴³ Indeed, cerebral atrophy was observed in 60% of this cohort. Supratentorial atrophy and subcortical gray matter (i.e., basal ganglia) involvement have been previously described in classic infantile Pompe disease, as well as in 1 previously reported patient with seizures.¹¹ In addition, glycogen accumulation in neurons in post mortem analyses has been documented.¹³ These findings suggest that both gray and white matter pathology may contribute to epileptogenesis in classic infantile Pompe disease.

This study has a few limitations. Its retrospective design limits data completeness and standardization. Only international specialists known through previous collaborations were involved, which may have introduced selection bias; however, the Dutch cohort is complete and includes all patients ever treated in the Netherlands. Sample sizes remain small, follow-up is limited, and international long-term follow-up studies will be essential to determine the true prevalence of severe CNS manifestations, including epilepsy. Finally, seizures were reclassified based on descriptions by physicians often not specialized in this field, and detailed epilepsy clarification and follow-up was not always performed, making an exact classification of seizures difficult.

As ERT-treated patients with classic infantile Pompe disease live longer and cohorts expand, our understanding of CNS involvement continues to evolve. An emerging clinical spectrum is becoming apparent, ranging from normal or mildly impaired cognitive function,^9,18 through progressive white matter abnormalities,^6,11 cognitive decline,^9,44 and increase of NfL,^9,12 to severe encephalopathy and seizures in some cases.¹⁰ Comprehensive assessments of CNS involvement, combining brain MRI, DTI, spectroscopy, and biomarkers, such as NfL, p-Tau, and GFAP, may help to elucidate the underlying mechanisms of seizure development. Moreover, autopsy studies of ERT treated long-term survivors are still lacking and would provide precious insight on the neuropathologic basis of CNS manifestations Figure.

Figure. EEG Recordings From 3 Different Patients.

(A) Interictal EEG of a patient presenting with focal seizures with impaired consciousness. Note left-sided temporo-occipital epileptiform discharges. (B + C) Interictal EEG of a patient presenting with several afebrile and febrile seizures. Notice right-sided temporal epileptiform discharges without activation during sleep. (D) Ictal recording of a patient during a focal impaired consciousness seizure.

Given these findings, regular evaluation of CNS involvement should be integrated into the systematic clinical follow-up of patients with classic infantile Pompe. This should comprise a combination of brain MRI, cognitive testing, EEG and, if needed, long-term EEG recordings. In patients with severe CNS involvement, such as extensive white matter abnormalities (brain MRI score ≥7/12¹¹), brain atrophy, cognitive impairment, and a substantial background slowing at EEG, the risk of seizures appears to be increased. Although only patients with classic infantile Pompe disease and epilepsy have been reported, monitoring CNS involvement including EEG should also be considered in children with atypical infantile and potentially even early juvenile-onset forms.

Conclusions

In this international case series, we describe in detail the characteristics of epilepsy as part of the broader CNS phenotype in classic infantile Pompe disease. Epilepsy typically developed after the age of 7 years, and most patients exhibited widespread white matter abnormalities and cognitive impairment. Seizures were predominantly focal with impaired consciousness or focal to generalized tonic-clonic and responded well to treatment. In ERT survivors with classic infantile Pompe disease, the possibility of epilepsy should be actively assessed during follow-up visits. These findings further delineate and expand the CNS phenotype of classic infantile Pompe disease.

Glossary

AEDKABC

Kaufman Assessment Battery for Children

AGLU

alglucosidase alfa

AR

infusion-associated reaction

ASM

antiseizure medication

AVA

avalglucosidase alfaI

DTI

diffusion tensor imaging

EMA

European Medicines Agency

ERT

enzyme replacement therapy

FDA

Food and Drug Administration

GAA

acid α-glucosidase

GFAP

glial fibrillary acidic protein

GTCS

generalized tonic-clonic seizure

MR

MRICRIM cross-reactive immunologic material

NA

not available data

NfL

neurofilament light chain

PT

patient

STROBE

STrengthening the Reporting of OBservational studies in Epidemiology

WISC

Wechsler Intelligence Scale For Children

Author Contributions

M.C. Faraguna: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. A. Broomfield: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. S. Gasperini: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. Yin-Hsiu H. Chien: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. Hui-An. Chen: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. N.M. Muschol: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. H.M. Amartino: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. G.Alessandra A. Kullmann: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. E. Procopio: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. F. Menni: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. F. Deodato: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. A.T. van der Ploeg: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. A. Hahn: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. J.M.P. Van Den Hout: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data.

Study Funding

Research on Pompe disease at the Erasmus MC is financially supported by the Prinses Beatrix Spierfonds (project number W.TR19-02), For Wis(h)dom Foundation, LSH-TKI (project number LSHM16008), and Stichting Ziekte van Pompe.

Disclosure

A.T. van der Ploeg and J.M.P. van den Hout received funding for research, clinical trials, and/or as advisors from various industries working on ERT or next-generation therapies in the field of Pompe disease under agreements with Erasmus MC University Medical Center. A. Broomfield participated on an advisory board for Sanofi and received travel and conference sponsorship. Y. Chien participated on an advisory board for Sanofi and received consultation/honoraria for Pompe disease related activities from various industries. N. Muschol has received travel support, consulting fees, and/or honoraria from Amicus and Sanofi. F. Deodato has participated in advisory boards and has received travel support and consulting fees from Sanofi. The remaining authors do not have declaration of interest to disclose. Go to Neurology.org/NG for full disclosures.