Longitudinal assessment of reflexive and volitional saccades in Niemann-Pick Type C disease during treatment with miglustat

Abstract

Background: Niemann-Pick Type C disease (NPC), is an autosomal recessive neurovisceral disorder of lipid metabolism. One characteristic feature of NPC is a vertical supranuclear gaze palsy particularly affecting saccades. However, horizontal saccades are also impaired and as a consequence a parameter related to horizontal peak saccadic velocity was used as an outcome measure in the clinical trial of miglustat, the first drug approved in several jurisdictions for the treatment of NPC. As NPC-related neuropathology is widespread in the brain we examined a wider range of horizontal saccade parameters and to determine whether these showed treatment-related improvement and, if so, if this was maintained over time.

Methods: Nine adult NPC patients participated in the study; 8 were treated with miglustat for periods between 33 and 61 months. Data were available for 2 patients before their treatment commenced and 1 patient was untreated. Tasks included reflexive saccades, antisaccades and self-paced saccades, with eye movements recorded by an infrared reflectance eye tracker. Parameters analysed were reflexive saccade gain and latency, asymptotic peak saccadic velocity, HSEM-α (the slope of the peak duration-amplitude regression line), antisaccade error percentage, self-paced saccade count and time between refixations on the self-paced task. Data were analysed by plotting the change from baseline as a proportion of the baseline value at each test time and, where multiple data values were available at each session, by linear mixed effects (LME) analysis.

Results: Examination of change plots suggested some modest sustained improvement in gain, no consistent changes in asymptotic peak velocity or HSEM-α, deterioration in the already poor antisaccade error rate and sustained improvement in self-paced saccade rate. LME analysis showed statistically significant improvement in gain and the interval between self-paced saccades, with differences over time between treated and untreated patients.

Conclusions: Both qualitative examination of change scores and statistical evaluation with LME analysis support the idea that some saccadic parameters are robust indicators of efficacy, and that the variability observed across measures may indicate locally different effects of neurodegeneration and of drug actions.

Fig. 1

Change scores for (a) gain, (b) HSEM-α, (c) Vmax and (d) latency

Fig. 2

Change scores for (a) antisaccade error rate and (b) self-paced saccade count

Fig. 3

Individual patient ISIs across time, indicated in months since treatment onset. In this and the following two figures, plots with red triangles are treated patients, blue triangles are reported in months since first visit of the untreated patient (UT5) and the two patients recorded before miglustat intervention (UT1 and UT2). The grey lines are the treated and untreated model fits, with intercept varying across participant). The black lines are individually fit regression models with blue shaded ± SE (allowing comparison of individual performance to the overall LME model fit across all participants in that group). Patient numbering is as in Figs. 1 and 2

Fig. 4

Individual patient gains across time

Fig. 5

Individual patient latencies across time

Fig. 6

Individual Iturriaga scores across time