Background
Niemann-Pick disease type C (NP-C) is a rare autosomal recessive neurovisceral lipid storage disease characterized in a majority of cases by progressive neurological deterioration leading to premature death, and (hepato)splenomegaly; its estimated incidence appears close to 1:100,000 live births [
1‐
3]. NP-C is caused by bi-allelic pathogenic variants in either the
NPC1 (in 95% of cases) or the
NPC2 gene (in around 5% of cases) [
1], leading to impaired intracellular lipid trafficking and accumulation of cholesterol and sphingolipids in the brain and other tissues [
4].
Visceral symptoms include neonatal cholestasis, hepatosplenomegaly and pulmonary dysfunction. Liver failure causes rapid death (before 3–6 months of age) in 5–10% of neonates presenting with a cholestatic icterus [
1,
5]); severe pulmonary insufficiency is lethal in a few others. Besides this rare systemic perinatal rapidly fatal form, four main clinical forms of NP-C are described according to the age of onset of neurological symptoms: early-infantile (EI) (< 2 years), late-infantile (2–5 years), juvenile (6–14 years), and adolescent-adult (≥ 15 years) [
1,
6,
7]. A few additional patients with systemic disease only [
8,
9] are temporarily unclassifiable since most of them—if not all—will later on enter one of the neurological forms. Typical neurological findings in paediatric NP-C include delay in psychomotor milestones, vertical supranuclear gaze palsy (VSGP), ataxia, dystonia, dysarthria, dysphagia, gelastic cataplexy, seizures and psychomotor or cognitive deterioration [
7,
10]. These manifestations have a continuous, unbroken progression, however consistently faster in patients with onset in early childhood compared with those showing a later onset form [
6].
At the time of our study, miglustat was the only disease-specific therapy available for the treatment of progressive neurological manifestations in patients with NP-C; to-date, it remains the only approved drug. This small iminosugar molecule able to partially cross the blood–brain barrier is a competitive inhibitor of glucosylceramide synthase, which catalyses the first committed step in the synthesis of glycosphingolipids accumulating in NP-C, but its mode of action appears more complex [
11]. Particularly, concomitant inhibition of the GBA2 glucosylceramidase might be involved [
12].
The approval of miglustat was based on data from a randomized clinical trial [
13], long-term extension studies [
14,
15] and a retrospective observational cohort study [
16], demonstrating stabilization of key neurological manifestations. A comprehensive review of published studies of miglustat in NP-C has been published [
11], as well as a consensus clinical management guideline [
3].
While miglustat can stabilize neurological manifestations in later onset forms of NP-C, its efficacy in the EI form has not been demonstrated [
17‐
19] 11/07/2023 21:09:00. Most published studies on miglustat efficacy included only a small number of EI-NP-C patients, with no long-term evaluation of therapy (duration of treatment < 2 years). The present observational retrospective multicentric study on patients with NP-C from the French cohort is the first one to specifically focus on the EI-NP-C form. We aimed to compare long-term neurodevelopmental outcome and survival between a miglustat-treated group and an untreated group.
Discussion
In the global French NP-C cohort diagnosed between 1990 and 2013, the EI-NP-C form constituted globally 21% of cases, a proportion very similar to that (22%) reported for Italy, but higher than for Czech Republic and above all the UK, with 15% and 5%, respectively [
8,
26‐
28]. Among the 10 patients with
NPC2 mutations diagnosed during this period, only 2 developed an EI-NP-C form.
By reference to a recent literature survey [
26], the present French series of 26 well characterized patients constitutes the largest EI-NP-C cohort with data on psychomotor achievements, neurological outcome and survival. It is also the first study to evaluate miglustat efficacy on EI patients by comparison to an untreated group.
Cell biology studies in cultured fibroblasts, primarily used to confirm the diagnosis, disclosed in all cases (with either
NPC1 or
NPC2 mutations) a profound alteration of LDL-cholesterol cellular trafficking, with massive accumulation of unesterified cholesterol in lysosomes, and in all cases studied, a severe block of LDL-induced cholesteryl ester formation (classical biochemical pattern) [
20]. This profile is in good accordance with results in EI-NP-C patients published from Czech Republic [
28], Italy [
27], and Spain [
29,
30]. A correlation between two null alleles and the EI-NP-C form is usual [
3,
26]. Our data confirm that the missense
NPC1 variants p.(Arg518Gln) and p.(Thr1036Met) are severe [
26], and further pinpoint p.(Ala172Pro) and p.(Gly1240Arg) as very deleterious, correlating with an EI-NP-C form when homozygous or associated with another severe allele. Of note, the c.1553G>A variant leading to the p.(Arg518Gln) missense is known to also cause a splicing error with exon 9 skipping [
31]. Previous functional studies on the
NPC2 p.(Ser67Pro) and p.(Cys99Arg) variants had shown that these mutated proteins were misfolded and retained in the endoplasmic reticulum [
32].
Since miglustat was only approved in 2009, treated patients were born and diagnosed in a more recent period than those in the untreated group. This may explain some observed differences between groups not linked to the treatment (Table
2). Under-reporting of neonatal cholestasis in the untreated group cannot be excluded, as a retrospective history of neonatal cholestasis, if not severe, could have been neglected in older medical reports. On the other hand, in recent years NP-C testing in neonatal cholestasis has been performed in France nearly systematically after exclusion of the most common causes [
5], a factor which could also contribute to the younger age at diagnosis observed in the treated group. Furthermore, availability of a specific treatment has led to a better awareness of NP-C disease and to a closer clinical monitoring. The observation that all patients were diagnosed before neurological onset in the treated group (100%), compared to only 37.5% in the untreated group (p = 0.003) is likely a consequence of these combined factors. Of note, the strategy for laboratory testing had remained similar for both groups, since plasma biomarkers [
3,
33] were not routinely introduced in France before 2015.
Inclusion criteria about age of neurological onset, steps of psychomotor development and neurological progression/outcome were clearly defined. Importantly, we chose to exclude isolated hypotonia as initial neurological symptom when it was observed in infants with a poor general state, i.e. life threatening hepatic or respiratory disease, because in our experience it does not necessarily reflects the subsequent neurological form. Categorization of a patient as EI-NP-C form on this sole criterion likely explains rare outliers with survival into adult age in NP-C registries [
9].
Our clinical observations are consistent with the comprehensive review on EI-NP-C published by Seker Yilmaz et al. [
26]. Neonatal cholestasis of variable length had been present in 54% of the cases, hepatosplenomegaly was nearly constant, and specific pulmonary dysfunction had occurred in about half of them. All patients in the present EI-NP-C series began neurological symptoms before 18 months of age (median global age of 9 months, slightly earlier than in [
26]; only few patients temporarily acquired autonomous walking, and none did acquire a fully normal gait. Hypotonia and motor development delay were the first neurological sign in 85% of the patients, ataxia, and psychomotor regression in the remaining 15%.
No record of VSGP was found for 11 of the 16 untreated patients, but this sign was noticed in all patients but one in the more closely monitored treated group, indicating that it also belongs to the EI-NP-C form, although not an early sign (observed at a median age of 2.3–2.5 years). During the later course of disease, about half of the patients developed seizures, and approximately one fourth gelastic cataplexy. Limb hypertonia reported during the course or EI-NP-C contrasted with early hypotonia and combined spasticity with pyramidal signs and rigidity due to dystonia. Lower limbs were more severely involved than upper limbs, and medications were poorly efficient using antispastic, antidystonic and antalgic drugs in more than 75% of patients. As for later onset NP-C forms, swallowing troubles were very frequent in EI-NP-C patients, with a need of nutritional support for 88% of untreated patients and 100% of miglustat treated patients, at a similar median age of 2.75 and 2.92 years respectively.
In our study, the median age at death was 4.4 years in the untreated group (4.7 years for all patients), compared to 4.0 years in the survey by Seker-Yilmaz et al. [
26], with also a narrower range (2.8–7.4 years) in our cohort.
Because of the retrospective nature of the study, information bias (data not reported in medical files, subjective interpretation of data) might have involved even data on neurological manifestations. It must be underscored that for many parameters, the small size of samples and the amount of unknown items prevented statistical treatment. Finally, we did not attempt to evaluate the progression of functional disability scores in our individual patients. Several NP-C severity scales, in which a composite score is calculated after assessment of 4 or 5 key neurological domains (ambulation, fine motor skills, language, swallowing, cognition) or more, have been developed [
17,
34‐
37] and used in natural history studies and evaluation of therapies [
10,
15,
17]. In this study, they could not be used for retrospective untreated patients. Furthermore, except for the more recent one [
37], these validated scales do not include milestones before ambulation, and therefore—although they were used in some studies involving patients with the EI form [
10,
17‐
19]—, none of them was at the time of our study fully suitable for precise evaluation of toddlers, as would have been needed here.
Notwithstanding the above limitations of this retrospective study, assessment of key parameters of neurological disease progression indicated no significant long-term improvement for patients with EI neurological onset who had received miglustat therapy for a median period of 30 months. Among others, the present study completes, 3.9 years later, the follow-up for 6 of the early-infantile NP-C patients included in our 2012 early study [
10] (see method section), in which the median duration of miglustat therapy was only 1.3 years. Data at that time already indicated a global poor response to miglustat. Héron et al
. [
10], however, suggested that a short interval between neurological involvement onset and the start of miglustat therapy and/or young age at treatment start could be associated with a better outcome of therapy, at least initially, as observed in case #2 (same as #T5 in our study) with the shortest delay (4 months) between neurological onset and miglustat initiation. This better initial outcome was confirmed and slightly extended in the present study, but it was not maintained, and the patient died at 5 years 5 months of age. Several other publications reported a transient stabilization or improvement of EI patients during the first months (or years) of miglustat therapy, especially when the treatment was started early, but these studies either described no change in long-term outcome [
38,
39], or the follow-up was too short to conclude on long-term outcome [
17,
40,
41]. The results of the retrospective observational chart review study by Pineda et al. [
42] cannot be compared to our study, because early infantile and late infantile patients were analysed together (mean age at first neurological symptom was 2.28 years), and the control group was much smaller than ours. Finally, initial results of an international NP-C registry [
18] showed improved or stable neurological disease in 33% of patients with the EI form who had received continuous miglustat therapy for an average follow-up period of 1.8 years, compared with 50% in the late-infantile group and 79% in the juvenile group. In the final report of this prospective registry [
19], results in the 19 treated EI-NP-C patients confirmed the lack of significant stabilization of the composite disability score, with a mean at enrolment of 0.59 and 0.70 at last follow-up. Our present results are thus globally in line with most published studies, including our own [
10], showing that while miglustat can stabilize neurological manifestations in later onset forms of NP-C, its long-term efficacy in the EI form has so far not been clearly demonstrated.
Regarding the specific question of survival, our results suggested a trend towards treatment efficacy, which, however, was not significant. A role of potential better global care in the treated group (covering the more recent period 2007–2017) cannot be fully excluded, as oxygen or non-invasive ventilation was more frequently used in our miglustat treated group (78%)—with date of birth after 2006 for all patients—, than in the historic non treated group (29%)—date of birth before 2006 for all but one—. Similarly, use of nutritional support for 100% of EI patients in the miglustat-treated group reflects earlier indication of nutritional support in more recent patients, to improve general state and comfort. A recent extensive survival study on 333 patients with NP-C (eventual therapy unknown) who had died between 1981 and 2018 [
43] has concluded that supportive medical care had not impacted survival in the recent past. However, since no stratification by clinical form could be done, a slight change such as that observed by us could hardly have been captured.
In any case, treated or not by miglustat, no EI-NP-C patient in our series was still alive beyond 7 years 5 months. Our observation on survival is consistent with the first comparative international survival study recently published by Patterson et al. [
23], which among others includes data from the French cohort. A statistically significant reduction in risk of mortality for miglustat-treated patients was observed in the overall analysis. In clinical subgroups analysis, however, a significant level was only achieved for the late infantile patients, and the smallest improvement of survival was found in the miglustat-treated EI-NP-C group.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.