Predictors of cognitive decline in memory clinic patients

Abstract

Background: The major challenge in memory clinics is to predict development as well as non-development of dementia among the heterogeneous group of patients with cognitive complaints. Since the advent of pharmacological treatment possibilities for patients with Alzheimer s disease (AD), much research has been focused on predictors for dementia since initiation of treatment in the preclinical phase may prevent development of dementia. At present, there are no tests specific enough to alone predict cognitive progression or stability. There are also no biochemical markers available to monitor the neurodegenerative disease process in AD. Thus there is a need for valid and clinically easy-to-use methods to differentiate patients who are at high risk of cognitive decline from those who will not progress. Objectives: The overall aim of this thesis was to examine which clinical methods that can be used and in what combinations to differentiate between patients at high and low risk for cognitive decline and dementia. Material & Methods: The thesis includes four retrospective studies that were performed among patients admitted to the Memory Clinic at Karolinska University Hospital in Huddinge. All patients were non-demented at baseline. They were clinically followed up during approximately 3 years because suspicion of a progressive cognitive disorder could not be ruled out at baseline. Study I investigated the efficacy of a number of neuropsychological tests for prediction of subsequent cognitive decline and conversion to dementia. Study II investigated the relationship between episodic memory function, APOE å4 allele status and levels of cerebrospinal fluid (CSF) biomarkers: total-tau (T-tau), hyperphosphorylated tau (P-tau) and the 42 amino acid form of beta amyloid (Aâ42). Study III investigated the relationship between longitudinal changes in CSF biomarkers and cognitive function. Study IV investigated what combination of clinical methods at baseline that was most predictive of cognitive decline and conversion to dementia. Results: Rey Auditory Verbal Learning Test (RAVLT) was the most efficient neuropsychological test for prediction of memory impairment and dementia. Cutoff levels in baseline RAVLT were defined and could identify patients at high and low risk for subsequent cognitive decline. Among moderately memory impaired patients, low CSF Aâ42 values differentiated those who did decline from those who did not. The relationship between episodic memory and CSF biomarkers was affected by the APOE å4 allele, where APOE å4 carriers who declined cognitively had pathological CSF T-tau, P-tau and Aâ42. Severely memory impaired patients showed significantly increasing P-tau levels during cognitive decline and progression to AD while patients with normal or moderately impaired memory showed unchanging P-tau levels and remained cognitively stable. CSF T-tau and Aâ42 did not change during follow-up.

Conclusion: RAVLT is a useful test for identifying patients at high and low risk for cognitive decline. Adding CSF Aâ42 values increase the ability to differentiate those who will decline from those who will not, especially among moderately memory impaired patients. The APOE å4 allele may affect the CSF biomarker levels among patients with memory impairment. Increasing P-tau levels during cognitive decline suggest that P-tau may be useful as a longitudinal marker of the early neurodegenerative process in AD.

Key Words: mild cognitive impairment, dementia, episodic memory, Rey Auditory Verbal Learning Test, neuropsychological assessment, cerebrospinal fluid, total tau, hyperphosphorylated tau, beta-amyloid (Aâ42), apolipoprotein E (APOE) genotype