Background
Pompe disease, also known as acid maltase deficiency or glycogen storage disease type 2 (OMIM 232300), is a lysosomal storage disorder that presents as a progressive myopathy in which deficiency of the enzyme acid α-glucosidase (EC 3.2.1.20) causes glycogen to accumulate in lysosomes. Ultimately, this leads to cell destruction [
1‐
3]. In 1932, J.C. Pompe first described the classic-infantile form of the disease [
4]. Classic infantile patients characteristically present shortly after birth with generalized and severe muscle weakness and with hypertrophic cardiomyopathy. They do not reach major milestones like walking and usually die within their first year of life [
5,
6]. Later, other forms were reported and Pompe disease appeared to be a continuous spectrum of closely related phenotypes with the classic-infantile form at the most severe end of the spectrum. In the literature, milder phenotypes are referred to as childhood, juvenile, adult and late-onset. Patients with these non-classic variants of Pompe disease usually have no hypertrophic cardiomyopathy and present with a more slowly progressive limb-girdle muscle weakness, which eventually results in wheelchair dependency, respirator need, and shortened life expectancy [
3,
7‐
17].
While many publications have described the natural history of the disease in classic-infantile and adult Pompe patients [
5,
6,
12,
14‐
16], there is little information on presenting signs and symptoms in children who do not fulfill the criteria of classic-infantile Pompe disease [
18].
To gain knowledge of the presentation of Pompe disease in children and to describe their specific clinical characteristics, we set up an observational study to collect information on disease symptoms, the distribution and severity of muscle weakness, physical limitations, lung function, cardiac structure and function, and genotypes of 31 children diagnosed with a non-classical presentation of Pompe disease.
Methods
Subjects
This observational study included all patients under the age of 18 years who had been diagnosed at or referred to the Pompe Center at Erasmus MC University Medical Center between 1975 and 2012. These patients came from the Netherlands and abroad, and had been diagnosed by measurement of the acid α-glucosidase activity in cultured fibroblasts, leukocytes or muscle biopsy specimens, and by mutation analysis. Patients with classic-infantile Pompe disease were not included in this study.
Thirty-one patients participated in this cross-sectional study. They were evaluated as part of studies approved by the Institutional Review Board (n = 28) or as part of routine clinical evaluation (n = 3). Medical history was obtained at first visit. Per patient, the following data were collected: gender, current age, geographic origin, first symptoms, age at first symptoms, age at diagnosis, wheelchair use, respiratory support, specific clinical findings (e.g. facial muscle weakness, bulbar muscle weakness, scoliosis, contractures, or muscle atrophy); functional impairments, spirometry and weight and height. Low body weight was defined as weight corrected for height if under 2SD for peers. Blood tests included measurement of creatine kinase (CK), aspartate aminotransferase (AST), alanine aminotransferase (ALT), and lactate dehydrogenase (LDH). A pediatrician and child neurologist performed clinical and neurological examinations in all patients. Only natural course data were included. None of the patients had been treated with enzyme replacement therapy.
Muscle-strength testing
Muscle strength was assessed by hand-held dynamometry (HHD) (
n = 24) and manual muscle-strength testing (MMT) (
n = 24) [
19,
20]. The following muscle groups were tested with HHD: neck flexors, shoulder abductors, elbow flexors, wrist extensors, hip flexors, hip abductors, knee extensors, knee flexors, and foot dorsal flexors. HHD scores (Newton) were expressed as percentages of the reference values (50
th percentile) for healthy peers [
20]. All percentages were cumulated and divided by 9 to obtain a total HHD sum score expressed in percentage of normal.
MMT was performed according to the Medical Research Council guidelines [
21] for the following muscle groups: neck flexors, neck extensors, deltoid muscles, biceps, triceps, wrist extensors, hip flexors, hip extensors, hip abductors, hip adductors, knee flexors, knee extensors, and foot dorsal and plantar flexors.
Lung function testing
Lung function testing was performed by spirometry with the patients in upright-seated position (
n = 28), and supine position (
n = 23) according to ATS/ERS standards [
22]. The highest outcome of three reproducible tests was used for analysis. The results were expressed as percentage of predicted and as a z-score, due to the paradigm shift towards z-scores, based on reference values corrected for age, length, gender and race [
23]. A percentage lower than 80 % of predicted and a z-score below −1.64 were considered abnormal. Two patients were too young for reproducible spirometry to be reliable. Because of poor lung function six other patients were unable to perform testing in a supine position and one of them in either sitting or supine position.
Cardiac assessment
In all patients, conventional Doppler, and 2D M-mode tracings were performed by an experienced sonographer (JP) according to the recommendations of the American Society of Echocardiography (Sonos 5500 ultrasound system, Philips, Best, the Netherlands). Standard 12-lead electrocardiograms were also made and analyzed by a pediatric cardiologist.
Enzymatic and molecular assays
Acid α-glucosidase activity was measured in leukocytes [
24] and in cultured skin fibroblasts [
13] according to standard procedures, and was expressed in nmol/h/mg protein. The protein concentrations of cell homogenates was measured as described previously [
25].
Genomic DNA was isolated from blood or cultured fibroblasts, and mutation analysis of the
GAA gene was performed according to standard procedures [
13,
26]. The severity of the mutations were rated using the format of Kroos et al. [
27], and were examined for their effect on enzyme activity, quantity, and quality, in transfected cells after site directed mutagenesis [
26]. In case of splice site mutations, the effect of the mutation was examined using real-time PCR in mRNA, isolated from the patients’ fibroblasts [
27].
Statistics
Demographic and clinical data were summarized using descriptive statistics including mean, SD, median, ranges, and percentages. As the data were not normally distributed, differences between groups were analyzed using a Mann–Whitney test. P-values < 0.05 were considered statistically significant. All statistical analyses were performed using SPSS for Windows version 21.
Discussion
As little information is available on the clinical presentation of children with non-classic forms of Pompe disease, we evaluated 31 children’s clinical and molecular characteristics. Our findings highlight that non-classic Pompe disease can cause a significant burden in childhood and add to the understanding that Pompe disease presents as a broad spectrum of clinical phenotypes.
The presentation of Pompe disease can be variable. Children in our patient population typically presented with weakness of the limb-girdle muscles and/or delayed motor development; lung function was compromised in approximately half of them, and 2 patients whose muscle function deteriorated very rapid also had hypertrophic cardiomyopathy.
Nevertheless, the diagnosis of Pompe disease should also be considered in children whose symptoms are less typical, such as disproportional weakness of the neck flexors, unexplained fatigue, persistent diarrhea, and an elevation of transaminase levels. Independent of whether the children had symptoms or not, all patients participating in the study had elevated CK, ALT and AST values. It should be noted, however, that in rare cases CK, ALT and AST may be normal, as we very recently encountered in a childhood onset patient (personal communication) and as was earlier described for about 10 % of adult patients [
28].
Although respiratory problems did not precede proximal muscle weakness in any of the children, lung function was already significantly restricted in 48 % of the cases. In 26 % of the affected children, respiratory insufficiency either led to the need for ventilator support or resulted in death during childhood. This finding signifies the importance of early monitoring of lung volume by means of spirometry in children affected by Pompe disease as also advised in adults [
17,
29].
Lung function tests are best performed in both sitting and supine positions, as 9 patients had postural drops suggestive of diaphragm weakness. This is a well-known feature of Pompe disease, and contributes to the onset of nocturnal hypoventilation [
29,
30]. Since a recent study in children with neuromuscular disorders found that daytime lung function and nocturnal hypoxemia were poorly correlated, we suggest to regularly perform sleep studies as an additional tool for identifying children with nocturnal hypoventilation [
31].
Twenty-two of the 31 children that we investigated were male. Interestingly, among them were 18 with the c.-32–13T > G/‘null’ genotype that is most common among adult patients (against only 3 of the 9 female patients). In two studies that focused on disease variation among children and adults with the c.-32–13T > G/‘null’ genotype, the male-to-female distribution was equal (55 and 58 % males, respectively). However, neither study focused on any potential difference in age at onset between male and female patients [
12,
13]. The present study indicates the existence of such a gender difference. Our earlier analysis of 225 published case reports on children and adults with Pompe disease also showed a predominance of males (67 %) in patients under 18 years old [
11]. Previous studies reported that pulmonary function was more affected in males than in females [
17], and that more men than women had bulbar involvement and shoulder-girdle muscle weakness [
17]. A study comparing phenotypes in siblings with Pompe disease also confirmed that males were more severely affected than females [
32].
Since Pompe disease is inherited as an autosomal recessive trait, there has been no satisfactory explanation to date why males with the same
GAA genotype as females would present at an earlier age. This finding seems to suggest that the clinical expression of Pompe disease involves secondary gender-related factors. Gender differences have also been reported for other neuromuscular disorders such as facioscapulohumeral muscular dystrophy and some subtypes of limb-girdle muscular dystrophy [
33‐
35]. One muscle related difference between men and women found so far is that women with limb girdle muscular dystrophy type 2A and 2B showed less muscle fiber atrophy compared to males [
35]. This may also apply in Pompe disease. Though other causes have been suggested, such as differences in genetic and epigenetic factors, the exact mechanism remains elusive.
Our findings are fully consistent with the broad spectrum of clinical phenotypes associated with the c.-32–13T > G/‘null’
GAA genotype [
12,
13]. Looking at the genotype-phenotype correlation within our group of children shows a similar age of onset, age at diagnosis, and current age for all different genotypes. Several patients were wheelchair bound and/or ventilator dependent despite of having the c.-32–13T > G/‘null’ genotype, which is mostly associated with adult onset disease. The c.-32–13T > G, is a leaky splice site mutation that results in the formation of 10–20 % of normally processed alpha-glucosidase protein and activity, which explains the later onset non-classic phenotype in patients c.-32–13T > G/‘null’ genotype. In our group of patients the second mutation (severe, less severe or potentially less severe) did not seem to have an effect on the age of presentation. This is in line with earlier studies of Kroos et. al., Wens et.al. and Montalvo et. al. also showing a broad variation of phenotypes among patients with the c.-32–13T > G/null genotype. It was hypothesized that epigenetic factors and environmental factors influence the level of disease severity [
12,
13,
32]. More research is required.
At a group level, children in the non-IVS1 group seem to be more severely affected than patients with the 32-13T > G/‘null’ genotype. This is illustrated by two specific patients in the non-IVS1 group who had hypertrophic cardiomyopathy and became fully wheelchair bound; one at the age of 4 years and the other at the age of 6. The first child died from respiratory failure at the age of 10. The other child became completely ventilator dependent when she was 6 years old. These two patients expressed a phenotype that Slonim et al. previously called the “atypical infantile form” of Pompe disease [
18].
Overall, the genotypes identified in the non-IVS1 group were more severe than those in the IVS1 group (see Table
1 and
www.pompecenter.nl). Some of the genotypes have been described in the literature, such as the genotype c.1634C > T/c.2481 + 102_2646 + 31del (our patient 25). The c.1634C > T (ls) in combination with a very severe mutation was previously described in three patients. One of those patients presented at the age of one year, became dependent on respiratory support at the age of 20, and was wheelchair bound at the age of 23 [
15]. The second patient was diagnosed at 16 years of age, and began to use a walking stick at 19. Pulmonary function worsened at the age of 17–19 years and vital capacity dropped to 26 % of predicted and became respirator dependent at the age of 20 years [
36]. The third patient presented at the age of 13 years with pronounced limb girdle weakness and died at the age of 18 years [
37]. The similarities in clinical course of these previously reported cases are remarkable, and if the genotype-phenotype correlation holds for our patient, who was only 6 years old at time of examination, he carries a high probability to develop severe respiratory and mobility problems before adulthood. Such information may be relevant when it is time to decide when to start enzyme replacement therapy.
Children and adults share a wide variation of disease presentation and disease progression, and a similar involvement of respiratory and proximal skeletal muscles [
8,
9,
12‐
16]. Although the distribution of muscle weakness shows a limb-girdle pattern in both children and adults, there are also differences. While the neck flexors are by far the most severely affected muscle group in children, they are only mildly affected in adults [
8,
15]. A recent MRI study performed in 20 adult patients by Carlier et al. also showed relatively mild involvement of the neck flexors in adults [
38].
Another difference is the relative sparing of the quadriceps muscle in adult patients [
28,
38]. In the current population of 31 children with Pompe disease, the muscles of the thigh were affected more heterogeneously, and the quadriceps muscles were not spared. Neither did any of our patients have ptosis, despite a recent publication in which ptosis was present in 14.7 to 23 % of adult Pompe patients [
28,
39,
40]. It should be noted that ptosis has often been found in an early stage of the disease, even as a presenting symptom in adult Pompe patients. While van der Beek et al. found that patients had difficulties with speech, chewing or swallowing, which was suggestive of bulbar weakness in 28 % of their patients, we found bulbar weakness in only one patient. In contrast, 52 % of our children had scoliosis, compared to only 21 to 23 % of adult Pompe patients [
11,
28]. In several children the scoliosis was so severe that it interfered with their mobility and lung volume; four children needed surgical correction of the spine. A cross-sectional analysis of data from the Pompe Registry, a large multinational observational program, found scoliosis to be present in 57 % of patients with childhood disease onset [
41].
Our study had two main limitations. First, since the Pompe Center at Erasmus MC University Medical Center serves as a national and international referral center for Pompe disease, there may be selection bias, due to referral of patients who were more than average severely affected. Nonetheless, ten of the 31 patients had been diagnosed pre-symptomatically, 5 of whom were still symptom-free at time of evaluation. A second limitation is the fact that the study was cross-sectional. As all children manifesting significant symptoms of the disease, started to receive enzyme replacement therapy during follow-up, the approval of this therapy in 2006 interfered with the collection of longitudinal follow-up data.
Competing interests
Research on Pompe disease at Erasmus MC is financially supported by ZonMw- the Netherlands Organization for Health Research and Development [project no. 152001005]; the Dutch TI Pharma initiative “Sustainable Orphan Drug Development through Registries and Monitoring (T6-208); “EUCLYD-a European Consortium for Lysosomal Storage Diseases” (health F2/2008 grant agreement 201678); and the Prinses Beatrix Fonds [project no. OP07-08]; MEUSIX [FP7/2007-2013, grant agreement n 304999]; SSWO; Colciencias and Genzyme Corp. AvdP has provided consultancy services for various industries. RHL is supported by the National Institute for Health Research University College London Hospitals Biomedical Research Centre and had honoraria from Genzyme Corp.
Authors’ contributions
JCvdM and CIvC participated in the recruitment of patients, data collection, analyses, data interpretation, and drafting and revising the manuscript for important intellectual content. ATvdP participated in the coordination, recruitment of patients, data collection, data interpretation, and revising the manuscript for important intellectual content. The other authors participated in the recruitment of patients and/or data collection, and drafting and revising the manuscript for important intellectual content. All authors read and approved the final manuscript.