Distinguishing neurocognitive deficits in adult patients with NP-C from early onset Alzheimer’s dementia

Niemann-Pick disease type C (NP-C) is a rare, progressive neurodegenerative disease caused by mutations in the NPC1 or the NPC2 gene, which lead to impaired cholesterol metabolism [1, 2]. Neurocognitive and neuropsychiatric deficits are commonly reported in patients NP-C, particularly among patients with the adolescent/adult-onset form [1, 3‐5]. Early-onset cognitive decline (EOCD) is characterised by presentation of cognitive impairment before the age of 65 years, and clinical diagnoses of EOCD are increasingly being reported. Most cases are related to early Alzheimer’s disease (eAD). However, compared to dementia in patients aged > 65 years, there is a much wider range of differential diagnoses, including underlying inherited neurodegenerative aetiologies such as NP-C.

NP-C shares a number of clinical and neuropathological features in common with AD and other dementias, including eAD, frontotemporal dementia (FTD), and Lewy body dementia [6‐10]. Patients with EOCD and eAD have therefore been suggested as a potential clinical niche for the identification of new, as yet undetected cases of NP-C [11]. In particular, patients with dementia-plus syndromes featuring concomitant psychiatric symptoms, movement disorders such as degenerative ataxia and/or vertical supranuclear saccade palsy (VSSP) are also considered to have an increased likelihood of having NP-C [11‐13].

Although the burden of clinical symptoms of adult NP-C patients impacts on activities of daily living, cerebral magnetic resonance imaging (MRI) findings are usually normal or show non-specific minor cerebral atrophy [2, 5]. Thus, cognitive deterioration among patients with adolescent/adult-onset NP-C can easily be mistaken for primary psychiatric disorders or other neurodegenerative dementias [12‐14]. It is therefore important to be able to distinguish at the initial clinical level between the cognitive profile of patients with NP-C and those with common dementia subtypes such as eAD. However, there are few published studies on valid neurocognitive tests that could be used to distinguish adult NP-C patients from those with other neurodegenerative dementia aetiologies [15].

A wide range of neurocognitive symptoms are encountered among NP-C patients, but published data on the specific neurocognitive profile of NP-C are based on studies adopting a range of methodologies, resulting in inconsistent or incomplete findings [12, 16‐19]. Early signs of cognitive impairment have been reported to comprise reduced executive function, processing speed, and verbal memory due to frontal-subcortical neural dysfunction [3, 17, 19]. Continued disease progression leads to further general cognitive decline, progressive deterioration in abilities to perform daily tasks, highly diminished memory, and behavioural impairment [5, 17, 20, 21].

Based on findings from a pilot study in 10 patients with NP-C, Klarner et al. [17] reported impairments in fine motor skills, language, attention, working memory, and visuospatial functions. A number of deficits similar to those seen in eAD were observed, including verbal episodic memory impairment. Appropriate neurocognitive tests incorporating disease staging were recommended to detect NP-C: the Trail Making tests (TMT) A & B and verbal fluency tests were judged as most useful in patients with mild disease, and The Mini-Mental Status Examination (MMSE), Corsi Block-tapping, Find Similarities, and Clock Drawing Tests were considered more appropriate in patients with more advanced disease. Other, separate studies have reported frontal impairments and decreases in executive functions and attention [5, 16]. A systematic literature review of case reports based on 23 separate patients has also reported executive dysfunction as the most frequent cognitive deficit in adult NP-C [15], possibly due to neuropathology in cerebellar regions and deep grey matter nuclei (e.g., the thalamus and striatum). Overall, the current literature points towards attentional and executive dysfunctions as the hallmark of neurocognitive impairment in patients with NP-C. However, it remains unclear whether the previously recommended tests assessing these cognitive domains could help distinguish NP-C patients from other dementia subtypes such as eAD.

The recognition of NP-C as early as possible in the disease course is important as a current treatment (miglustat) is available in Europe and many other countries that may help stabilise the progression of neurological and neurocognitive symptoms. Intrathecal hydroxypropyl-beta-cyclodextrin is a promising future alternative treatment that is currently undergoing clinical trials. Encouraging data with this agent were reported by Ory et al. in 2017, and further studies are ongoing [14, 16, 22‐25].

The advent of biochemical laboratory markers and the increasing availability of next-generation gene sequencing methods are major advances for the identification of new cases of NP-C. However, these methods are only applied in patients or patient groups considered at high risk of NP-C. Clinical methods of defining the level of risk or suspicion of NP-C are necessary to direct such further diagnostic investigations. The characterization and delineation of the neurocognitive profile of NP-C may help clinicians to achieve earlier diagnoses, particularly when combined with approaches that help distinguish potential cases from another primary degenerative disease such as eAD.

In this study, we used a detailed neurocognitive test battery to assess whether there is a recognizable, specific profile of deficits in adult patients with genetically proven NP-C. We were particularly interested in whether neurocognitive tests from attentional and executive domains identified by previous studies as useful in the detection of NP-C (e.g., TMT-A and B, verbal fluency tests) were able to differentiate NP-C from eAD. Importantly and in contrast to previous studies [17], the neurocognitive test battery used here was geared towards the typical age span of patients with adult NP-C and eAD instead of senile dementia, in which ceiling effects must be expected. Thus, we conducted the first direct comparison of the observed NP-C neurocognitive profile with a reference group of adults with confirmed eAD, the most frequent aetiology of EOCD.

This is the first study to directly compare the neurocognitive symptom profile of patients with NP-C with that seen in a ‘reference’ group of patients with deficits due to eAD – the most common primary neurodegenerative dementia syndrome before the age of 65 years. The detailed battery of neurocognitive tests applied in this study identified specific, statistically significant differences between the NP-C and eAD groups, which may aid in the differential diagnosis of patients with early cognitive impairment and further neurological deficits (i.e., dementia-plus syndrome) suggestive of possible NP-C.

NP-C and AD share a number of pathophysiological similarities such as increased levels of brain tau protein, amyloid deposition, the presence of neurofibrillary tangles, and the influence of apolipoprotein E ε4 genotype [6‐9, 35]. Other neurological commonalities include basal forebrain cholinergic system alterations and chronic neuroinflammation [36‐38]. However, there are distinct differences in the localisation of neuropathology between the two conditions. Purkinje cells in the cerebellum are the most affected neurons in NP-C, with NFTs mainly found in subcortical structures, while lesions in AD are mainly seen in the neocortical and medial temporal lobes [6, 15, 39, 40]. It therefore follows that certain clinical neurocognitive differences exist between NP-C and eAD patients [16, 17].

MRI and diffusion tensor imaging (DTI) studies in NP-C patients have suggested that cerebral atrophy in key deep grey matter regions including the hippocampus, thalamus, cerebellum, and striatum, as well as major white-matter tracts, may account for global impairments in cognitive function in NP-C [41‐43]. More specifically, positron emission tomography (PET) studies in adult NP-C patients have indicated that frontal-lobe hypometabolism may contribute to frontal-executive deficits [44, 45].

Overall, taking patient age, gender, and education into account, eAD patients showed wider, more generalised impairments in affected cognitive domains compared with the profile seen in NP-C patients in the current study. Deficits were also more frequent and greater in magnitude in eAD on most of the neurocognitive tests that we employed. Bergeron et al. reported that in contrast to NP-C, the general cognitive profile in AD is characterized predominantly by memory dysfunction [15]. In the current study, the key neurocognitive difference between NP-C and eAD was also observed in terms of verbal memory performance, as measured by the RAVLT. Most RAVLT subscores (excluding forgetfulness after interference) were statistically significantly more impaired in eAD patients. In particular, the RAVLT7 subscale (delayed verbal recall), an indicator of verbal memory that has been shown to correlate with hippocampal pathology [46, 47], showed the greatest effect size between the two groups. Our ROC analyses suggested that delayed verbal recall also showed good sensitivity and specificity in distinguishing between NP-C and eAD patients. Relatively preserved verbal memory has recently also been reported in a French cohort of 21 patients with mild adult NP-C [16].

With regard to alternative tests for verbal episodic memory, the Consortium to Establish a Registry for Alzheimer’s disease (CERAD) word list recall test also addresses this neurocognitive domain through delayed recall of a learned word list and is faster to apply than the RAVLT. However, the disadvantage of this test is that it is geared toward dementia patients of higher age, and normative data for younger patients are not available [17]. The Free and Cued Selective Reminding Test (FCSRT) is an alternative verbal memory test that, similar to the RAVLT, provides normative data versus younger adults, and might therefore also be suited to cognitive assessments in NP-C [16].

A number of previous studies in patients with NP-C have described impaired executive functions (e.g. set-shifting and word fluency) and attention as the primary neurocognitive deficit in NP-C [5, 15‐17, 19]. Our findings are in line with previous published evidence, in that executive functions as measured by the TMT-A/B and word fluency as measured by the RWT, were the most frequently impaired domains in NP-C patients [17]. However, the frequency and magnitude of these impairments were actually lower among NP-C patients on the majority of neurocognitive tests assessing attention and executive functions versus eAD. As a result, these tests cannot be expected to function as differential cognitive markers on their own.

Interestingly, the RWT word fluency (letters) task was the only neurocognitive measure on which NP-C patients were more impaired than eAD patients at the raw score level. However, the NP-C versus eAD effect-size seen with this parameter was small and non-significant, and the frequency of impaired cases (compared with age-matched normative data) on this test was similar in both groups.

While global dementia screening tools such as the MMSE, or more specific tools such as the FAB may not be powerful enough to distinguish specific neurocognitive deficits between NP-C and eAD, the majority of NP-C patients in our study presented with impaired performance on both of these scales. In particular, FAB may prove useful for this purpose in clinical practice as it effectively assesses executive dysfunctions in a time-economic way. Overall, due to lack of specificity and the lack of normative age-matched control data in younger patients, these screening tests cannot substitute an extensive neurocognitive assessment, particularly in addressing memory and executive function.

The small sample size, particularly of the NP-C group is a weakness of this study. However, this is due to the extreme rarity of NP-C and the single-centre nature of this study. The between-group differences in age in the NP-C and eAD patients introduces a further degree of uncertainty, with patient age acting as a potential confounding factor in our statistical analyses. Age has an important influence on the range and severity of cognitive, neurological, visceral and psychiatric symptoms in NP-C [1, 2, 15]. It is notable that patients with eAD in this study were significantly older than those with NP-C. While this could be considered as a study limitation, it should also be noted that, in general, cognitive impairment in NP-C has a lower age at onset compared with eAD. While cognitive deterioration is observed in most patients with adolescent/adult-onset NP-C [1, 2, 15], EOCD is most usually recognized among patients with childhood-onset NP-C in the form of intellectual developmental disorder, poor school performance, and learning disabilities [4, 15, 48]. One difficulty in this respect is that not all neurocognitive tests are similarly applicable in different age-groups, particularly in patients with the juvenile-onset or early adolescent-onset forms: this constitutes a problem. For instance, Klarner et al. used the CERAD battery to determine the cognitive profile of patients with NP-C [17]. However, this test battery is not associated with normative data for patients aged < 50 years and frequently produces ceiling effects in younger patients. In contrast, the RCFT, RAVLT, TMT, RWT and WMS-digit span tests included in the current study all provide normative data for both children, as well as older age groups. The current test battery may therefore be more suited for use in NP-C than classical senile dementia test batteries like the CERAD or dementia screenings.

Finally, structural neuroimaging data to investigate potential links between memory scores and medial temporal lobe atrophy in eAD versus NP-C, and the possible relationship between frontal executive functions and frontal/subcortical atrophy with executive dysfunction, are a subject for potential, future studies. In addition, more specific cognitive and behavioural functions associated with frontal lobe integrity (e.g., social cognition tests, abnormal behaviour) may also function as differential markers for NP-C. However, these aspects have not yet been tested in detail in this rare patient group.

In conclusion, based on our findings, we recommend that general dementia screening assessments (e.g., MMSE, FAB) should not be used in isolation to evaluate cognitive performance in patients with suspected NP-C, as these patients may present with milder cognitive deficits than patients with eAD. Such screening studies may therefore prevent the detection of important differences in the neurocognitive profiles between eAD and NP-C. Detailed, yet easy-to-apply neurocognitive tests such as the RAVLT assessing verbal episodic memory, the TMT-B assessing set-shifting and a verbal letter fluency task, with normative data available for age-groups typically associated with NP-C, appear more appropriate. Further studies in larger numbers of patients are warranted. Parallel studies of such neurocognitive scales alongside imaging parameters (e.g., based on MRI volumetry or DTI) would allow further insight through the assessment of underlying neuroanatomical correlates of cognitive dysfunction in NP-C.