Background
Canavan disease (CD) is a genetic degenerative brain disorder caused by deficiency of the enzyme aspartoacylase (ASPA). The loss of ASPA activity results in an accumulation of N-acetylaspartic acid (NAA) in the brain and other parts of the body [
1]. NAA is suspected to function as a molecular water pump [
2], leading to fatal brain disease for which there is currently no effective treatment. CD is diagnosed by detection of elevated NAA in urine or blood or in brain by proton MR spectroscopy [
3], as well as by
ASPA mutation analysis [
4]. Histologically the disease is characterized by insufficient myelination and progressive spongy degeneration of the brain white matter [
5,
6]. The perinatal period is often uneventful and subtle changes during early infancy may escape attention. The evolution of symptoms distinguishes Canavan disease from static encephalopathies, such as cerebral palsy. To enable early diagnosis first symptoms and clinical signs specific to CD need to be recognized.
While CD is presently incurable, experimental therapies are in development [
7]. Their implementation requires an understanding of the optimal window of intervention. Detailed knowledge of early symptoms and the natural course of the disease provide a framework for trial design. Here, we identify early features of CD and propose a CD severity score based on a comprehensive study of the clinical course of 23 CD patients, including retrospective longitudinal data, in comparison with data of CD patients reported in the literature.
Discussion
This study describes the clinical course of 23 CD patients in detail and compares it with data of 90 patients reported in the literature. We further report a heterogeneity of ethnic background in CD patients that has not been published to date.
In our cohort, CD patients showed first symptoms at 2–4 months of age. At first onset most families reported a combination of up to twelve different symptoms, Fig.
1.
Only 3/23 patients acquired skills consistent with normal development in healthy children. These less severely affected patients learned to sit without support, developed basic fine motor skills (reaching for an object, transferring an item and scribbling) and acquired the ability to speak single words. Our data as well as data from the literature showed, that if patients did not acquire sitting by the age of 12 months, it was unlikely that they would ever gain it.
All patients, even those who gained a higher number of more advanced abilities, lost these skills during their course of disease. Our data show that fine motor skills were retained for a longer time than gross motor skills. Language comprehension was more commonly present in CD patients than expressive language and was retained for a longer time span. None of our patients was ever able to speak in full sentences.
While not all CD patients suffered from epilepsy during the first years of life, over the first decade seizure prevalence increased, with all CD patients suffering from seizures beyond 10 years of age. Management of neurological symptoms such as seizures and spasticity are often complicated by the need for multiple antispasmodic or antiepileptic drugs. There was no evidence that usage of experimental drugs showed any distinct benefit. Only lithium was reported to improve alertness and to contribute to mood stabilization. It had no effect upon the disease course or the number of abilities gained.
CD manifests in early childhood and leads to severe disability and shortened life expectancy. Recent reports state that NAA itself is not required for myelination [
13]. Rather, high levels of NAA, as found in CD, are thought to cause a toxic imbalance of brain water homeostasis [
14] and oxidative stress [
15,
16] which contributes to characteristic formation of vacuoles within the myelin, causing severe neurological symptoms. After an unremarkable perinatal course, affected children typically stagnate in their psychomotor milestones within the first six months of life. Only few CD patients progress beyond the developmental age of approximately 3 months.
Our survival analysis showed that about 73% of CD patients reach the age of ten years which extends survival beyond what had previously been reported [
17]. Lyon et al. had listed survival limited to three years with only rare exceptions to 10 years of life. We think prolonged survival may be due to improved care standards over time.
Our analysis of longitudinal data revealed details on the onset of macrocephaly during the first year of life (mean age for girls: 7 and for boys: 8.5 months) with a plateau occurring around the age of 18 months. Onset and frequency of macrocephaly documented for our study group was comparable to data from the literature. A larger cohort of CD patients is needed to determine whether gender impacts the age at onset of macrocephaly.
Macrocephaly in combination with early developmental stagnation is the hallmark of CD and allows clinical differentiation from many other infantile encephalopathies, such as cerebral palsy or other neurodegenerative disorders such as Krabbe disease. Within the group of neurodegenerative diseases with macrocephaly, developmental delay, affected brain white matter and visual problems, CD needs to be distinguished from genetic encephalopathies such as GM1 or GM2 gangliosidosis, PTEN (Phosphatase and tensin homolog protein diseases), megalencephalic leukoencephalopathy with subcortical cysts, Alexander disease or glutaric aciduria. Diagnosis can then easily be accomplished by measuring NAA levels in urine or on brain MRS.
To define hallmarks of the disease for endpoints of future clinical trials one first must consider that most CD patients never attain motor and language milestones. Next, potential endpoints should be chosen that can be compared across a large number of CD patients, such as the ability of visual tracking that 12/23 patients gained, and 7 CD patients subsequently lost again. Considering that only two patients were reported to suffer from optic nerve atrophy, we conclude that visual impairment is more commonly caused by white matter disease affecting the visual pathways.
Patients with CD suffer from several concomitant disabilities. The result that gastric tubes of literature patients [
10] being implanted significantly later (39 months versus 22 months within our study cohort) is likely related to improved standards of palliative care over time. We could not find any evidence that gaining early milestones in psychomotor development such as head control and visual tracking, was associated with prolonged survival.
All CD patients were reported by caregivers and medical records to be able to hear at birth and throughout the course of disease. This stands in contrast to findings of impaired hearing reported in the Canavan mouse model [
14]. Interestingly, acoustic startle was noted by more than half of the caregivers. This contradicts Lyon et al. [
17] who mentioned that CD patients show tonic spasms by acoustic stimuli or handling rather than a repetitive acoustic startle as seen in Tay-Sachs disease. Our clinical impression is that both tonic spasms as well as a startle response can be seen in children with CD.
Perhaps due to the small numbers of and the fact that patients with a milder phenotype are not coming to our attention, we could not demonstrate a genotype phenotype correlation within our study cohort. However, for less severely affected patients of our cohort we observed molecular genetic findings such as p.Glu285Ala;p.Ala305Glu/p.Ala305Glu;p.Ala305Glu/p.Ala287Thr;p.Ala305Glu. These patients achieved and sustained a higher number of abilities and more complex functions. We also observed similar courses of the disease among siblings. Individual reports of ambulatory patients without macrocephaly suggest that milder phenotypes may be more abundant [
18].
Limitations of our study were that cognitive testing by standard instruments was not possible due to the severity of disease in our study population. Additionally, there were only very limited MRI data due to the need for sedation that was not justified in a natural history without intervention.
To distinguish milder from more severe courses of the disease, we chose 11 pertinent symptoms that serve as a benchmark for disease progression and may be critical to future clinical trial design. Disease severity shall be ascertained retrospectively and prospectively. First assessments of the CDRS found an interrater variability of 96%. Further natural history studies are needed to validate the CDRS and optimize its performance in prospective studies.
Methods
Patients were recruited at our outpatient clinics (Department of Pediatrics, University Medical Center Hamburg Eppendorf, Hamburg, Germany, and the Department of Neurology, MGH, Harvard Medical School, Boston, USA, supported by cooperating partners and patient organizations as NTSAD, Myelin Project Germany and ELA Germany) over a 6-year period by sending out questionnaires to 30 caregivers of CD patients. Data on the natural course of 23 CD patients (11male,12female) were collected retrospectively. Data sources consisted of the patients’ medical records and standardized patients’ questionnaires. All data were anonymized, entered and processed within the web-based Hamburg Leukodystrophy Database. All entered data were cross-checked by two experienced data managers.
Only those patients were included, whose diagnosis was confirmed by urine analysis of NAA and/or by ASPA gene mutation analysis. Onset of CD was defined as time when first symptoms were observed by the parents and/or documented in the patients’ medical history records.
To describe acquisition and regression of psychomotor developmental milestones, we analyzed the median age of milestone acquisition, as well as median age of regression of skills for each psychomotor ability listed in the CD questionnaire. The categories were gross and fine motor skills, vision, hearing and language and other abilities listed in Table
1 and were based on clinical experience with CD and other early infantile neurodegenerative diseases such as Krabbe leukodystrophy. Whenever reasonable and available, psychomotor abilities were compared with control data from the Denver Developmental Screening Test II/DDST II [
8]. If available, psychomotor skills and neurological symptoms were compared with the published literature of CD patients.
In order to describe neurological symptoms that are typically associated with CD, we asked for age of onset of spasticity, abnormal eye movements, seizures, dystonic hyperkinetic movements and irregularities of the eye ground. Caregivers were asked to document medications necessary to control the above-mentioned symptoms.
To analyze CD patients’ seizure history, we collected longitudinal data and asked for annual seizure frequency. Types of seizures were not specified. Seizure frequency was classified into 4 categories: no seizure/year, 1–2 seizures/year, less than 12 seizures/year and more than 12 seizures/year based on experience with other infantile neurodegenerative diseases.
To evaluate onset and course of macrocephaly, longitudinal data of head circumference in CD patients was collected and compared with reference percentiles, Robert Koch Institute, KiGGS [
9]. Macrocephaly was defined as a head circumference larger than 97th percentile.
Review of literature/literature CD patients
We searched Pubmed literature from 1976 to 2016 with key words “Canavan” (192 results) and “Van Bogaert-Bertrand” (21 results) for articles reporting the clinical course of CD patients. Data of three articles [
10‐
12] reporting on larger patient cohorts (N > 10) were included. Frequency and median age of milestone acquisition and regression of single psychomotor skills, onset of disease and neurologic symptoms, as well as follow up/survival time was calculated for the literature group and compared with the data from our study group.
For analysis of disease onset, data of all three articles were used, summarizing up to 90 patients in the literature. For analysis of seizure onset, data of 74 patients reported by Ozand et al. [
11] and Traeger et al. [
10] were analysed and used for comparison with our study data. For analysis of psychomotor development, macrocephaly and gastric tube placement, data of 60 patients reported by Traeger et al. [
10] were included.
Statistical methods/survival analysis
We pooled data of our study cohort and a cohort of 60 CD patients reported by Traeger et al., summarizing a total number of 82 patients. We performed Kaplan Meier survival analysis to investigate differences in survival rates between the study cohort and literature group. A Log Rank Test was performed to test if there were significant differences between patient groups.
To investigate the influence of certain neurological symptoms, clinical complications or development of psychomotor skills upon survival, we selected those patients from both groups (literature and study group) who were observed for longer than one year. We analyzed whether development of seizures, the need for gastric tube placement or development of basic psychomotor skills (visual tracking & head control) within the first two years of life had an influence on survival of CD patients.
A pilot study to assess interrater variability of the CDRS was performed. Two independent physicians (one knowing the subjects, the other one not familiar with the patients) reviewed medical letters from different time points.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.