Multimodal prediction of dementia with up to 10 years follow up: the Gothenburg MCI study.

Abstract

Neuropsychological tests, CSF Aβ42, T-tau, P-tau181, hippocampal volume, and white matter lesions have been shown to predict conversion to dementia in patients with mild cognitive impairment (MCI). To examine the predictive value of combinations of these markers and to examine if the absence of pathological markers provides a lasting reduction of conversion rates. The Gothenburg MCI study is a clinically based study. Seventy-three MCI patients were included in the present sub-study and followed for a maximum of ten years. Thirty-four patients converted to dementia (18 to AD) and 39 remained stable. At inclusion, patients were classified into positive or negative risk groups according to results from neuropsychological testing (Rey auditory verbal learning test, Boston naming test, Trail making test B), CSF biomarkers (amyloid β42, T-tau, and P-tau181), and MRI scans (hippocampal volume, white matter lesions). Trail making test B (TMT-B) was the best single predictor for the prediction of dementia (AUC 0.89, HR 25), and T-tau was the best predictor of AD (AUC 0.97, HR 41). The combination of hippocampal volume and TMT-B was the best combination for the prediction of dementia (HR 25), and the combination of hippocampal volume and T-tau was the best combination for the prediction of AD (HR 37). Neuropsychological tests, CSF markers, and hippocampal volume predicted conversion from MCI to AD and general dementia. The absence of pathological markers provided a long-time protection from dementia.