Cambridge Cognition : The Predictive Utility of Cognitive Measures at Screening for Prodromal Alzheimer’s disease

This data analysis examined the use of cognitive tests such as CANTAB as a screening tool for Alzheimer's trials.

View the Poster

• Data were collected as part of the Roche SCarlet RoAD trial (NCT01224106; WN25203), a randomised control trial of gantenerumab, a fully human monoclonal antibody predominately targeting aggregated forms of amyloidβ.
• This analyses includes 1212 subjects (age range 50-85), all of whom had a deficit on initial cognitive screening tasks - Free and Cued Selective Reminding Test (FCSRT), Mini Mental State Examination (MMSE) and Clinical Dementia Rating (CDR), and had undergone CSF Aβ sampling.
• 799 subjects were classified as Aβ+ (cut-off
• In addition to the initial cognitive screening tasks, subjects were assessed on the CANTAB battery (measures of attention, working memory and episodic memory), the Alzheimer's Disease Assessment Scale (ADAS-cog) and the Functional Assessment Questionnaire (FAQ).

Data Analysis

• Data were analysed using logistic regression, adjusted for age and education. Between group differences were assessed by ANOVA. Data were analysed using SAS v9.0.
• Episodic memory tasks showed good utility for classifying subjects by CSF Aβ level. The FCSRT representing verbal episodic memory and used in initial screening, performed comparatively well across a range of measures; free recall; total recall and index of cueing. The CANTAB episodic memory task, Paired Associates Learning (PAL), also showed good predictive properties.
• The MMSE and CDR-SB both used in pre-screening showed slightly lower discrimination.
• CANTAB measures of attention and working memory also showed reasonable discrimination between Aβ+ and Aβ-, together with the ADAS-Cog11.
• By way of comparison, age and years of education without the addition of cognitive tasks showed lower discrimination.

Conclusions

• Whilst subjects were screened for prior cognitive impairment on a verbal episodic memory task (FCSRT), cognitive test performance still showed utility for classifying a subject's CSF Aβ status in a prodromal AD population.
• The analyses indicate comparative utility between the PAL and FCSRT for predicting CSF Aβ status in a population with a pre-defined memory impairment. As some of these measures were administered following initial selection for cognitive impairment, further research to establish their performance as first line screening tools is warranted.
• Whilst the predictive utility of these measures in a prodromal population were respectable, measuring cognition longitudinally (over the course of months or years) may improve the ability to assess suitability for a trial such as this, by identifying subjects who may progress more rapidly to a diagnosis of dementia.