Annual severity increment score as a tool for stratifying patients with Niemann-Pick disease type C and for recruitment to clinical trials

Abstract

Background: Niemann-Pick disease type C (NPC) is a lysosomal storage disease with a heterogeneous neurodegenerative clinical course. Multiple therapies are in clinical trials and inclusion criteria are currently mainly based on age and neurological signs, not taking into consideration differential individual rates of disease progression.

Results: In this study, we have evaluated a simple metric, denoted annual severity increment score (ASIS), that measures rate of disease progression and could easily be used in clinical practice. We show that ASIS is stable over several years and can be used to stratify patients for clinical trials. It achieves greater homogeneity of the study cohort relative to age-based inclusion and provides an evidence-based approach for establishing inclusion/exclusion criteria. In addition, we show that ASIS has prognostic value and demonstrate that treatment with an experimental therapy - acetyl-DL-leucine - is associated with a reduction in ASIS scores.

Conclusion: ASIS has the potential to be a useful metric for clinical monitoring, trial recruitment, for prognosis and measuring response to therapy.

Keywords: ASIS; Acetyl-DL-leucine; Annual severity increment score; Clinical severity scale; Clinical trials; Experimental therapy; NPC; Niemann-Pick disease type C; Tanganil.

Fig. 1

Distribution of the study cohort with each patient represented once (first visit) plotting age against unadjusted clinical severity (minus hearing). a Patients fell into six main subgroups defined by lines a-f (latent class mixture regression model). b The same data are presented colour coded to lines a-f with arrows showing progression in severity score from first to last visit for each individual in the study

Fig. 2

Each of the main clinical subdomains (x-axis) were plotted against total severity score and the coefficient of variation (Spearman’s correlation coefficient) determined (see Additional file 1: Table S1). The dotted line represents the scatterplot’s Friedman’s adaptive smoother. Each patient appears once. Red dots are patients with seizures

Fig. 3

a ASIS scores (total severity/age for every time point) were plotted for each patient over time (years), colour coded to reflect the subgroups from Fig. 1a. b The same data were plotted colour coding patients with seizures in red. c Box and whisker plots demonstrating the greater variability of ASIS scores in the seizure versus non-seizure group. d Density plot of ASIS score variation around zero for seizure, non-seizure groups

Fig. 4

The inclusion criteria for three clinical trials were modeled on plots of age versus total severity score (minus hearing). a Miglustat trial, ages > 12; b Arimoclomol trial ages 2–18 and c 2-Hydroxy-beta-cyclodextrin ages 4–21. The plots depict the study cohort with these different inclusion criteria indicated with the dotted lines. Included patients are in black, excluded patients in white

Fig. 5

Inclusion criteria based on ASIS scores are plotted using three thresholds (ASIS 0.5 to 2; 0.75 to 2; 1 to 2) plotted in a1-a2 and a3 respectively. The ASIS bands plotted were extended to include 0.5–2.5, 0.75–2.5 and 1–2.5 (b1, b2 and b3 respectively). c The ASIS thresholds were plotted to demonstrate their influence on seizure/non-seizure cases (non-seizure cases in blue, seizure cases in red)

Fig. 6

Plots of total severity score minus hearing defined from either a single ASIS score (standardised to the first data point for each patient) or using a mean ASIS score (repeat measures). The four scatterplots represent patients with and without seizures, and with predicted trajectories calculated based on a single ASIS determination at first visit (a) (without seizures) and (c) (with seizures)) or where predicted trajectories are based on the average of all available ASIS scores for each patient (b and d) for patients with and without seizures, respectively. Each patient has been assigned an individual colour and each visit is represented by a data point (multiple points per patient)

Fig. 7

The mean of the maximum absolute deviations (MAD) was calculated for each individual in relation to the predicted trajectory, based on a single or average ASIS score relative to actual severity (panel (a) illustrates MAD in a single case). The black arrow shows the point of maximum deviation. The goodness-of-fit criterion, based on MAD, is depicted in the box and whisker plots for all patients (b) and patients with no seizures (c) and patients with seizures panel (d)). The effects of time on MAD determination depicted as box and whisker plots for patients without seizures (e) and with seizures (f). The analysis was performed for repeat measures spanning data over 0–2 years, 2–4 years and 4–6 years. Each patient only contributes once

Fig. 8

a Total Clinical Severity Score and b ASIS scores (total severity score/age at time of assessment) of NPC patients post-commencement of treatment with 5 g/day Acetyl-DL-Leucine. The initial severity score (and hence ASIS) of one patient was notably higher than that of the other nine. This patient’s data is therefore provided as a separate graph. Each assessment is represented by one data point