Está en la página 1de 17

Frederick A.

Schmitt University of Kentucky

Wien Center for Alzheimer's Disease and Memory Disorders, Mount Sinai Medical Center 12th Mild Cognitive Impairment Symposium January, 2014


Nothing to disclose

Prodromal / Preclinical AD

Memory complaints by patient or family Progressive onset Normal or mildly impaired complex ADLs Amnestic syndrome of the hippocampal type Persistence of memory impairment at a subsequent assessment Absence of fully developed syndrome of dementia Exclusion of other disorders that may cause MCI (e.g., neuroimaging & biomarkers)

MCI, Cognition & Brain Pathology

Delayed Recall (max = 10)
9 8 7 6 5 4 3 2 1 0 0 1 2 3 4 5 6
A -mean recall visit 5< v1,2,3 B -mean recall visit 4< v2

Annual Visit

Schmitt, et al. Neurology, 2000, 55:370-376; Markesbery, et al. Arch. Neurol. 2006, 63:38-46.

Frequency of memory problem: Varies by classification

60 50 40 30 20 10 0 Memory Complaints AAMI MCI CIND MCIa

22 - 56%

DeCarli C. Lancet Neurology 2003;2:15-21

How to Better Predict Transitions?

NIA: P30 AG028383

NIA: R01 AG038651

Competing Risks

Events that alter the probability of experiencing (or observing) the outcome (Satagopan et al, 2004). Traditional survival analysis assumes censored observations are non-informative and ignorable.
Death alters the probability of observing dementia.

Markov Chain Multistate Models

Markov chains - model the outcome of interest and competing events as mutually exclusive absorbing states to identify risk factors.

Probability models describing the random movements of a person among various finite states.
Movement between states depends only on the (immediate) prior state (1st order chain)

Markov Chain Multistate Models

Time between successive observations is assumed to be constant. At each observation, persons may remain where they are or move to a new state until they are absorbed. Calculate the probability of remaining in or moving to a state vs. remaining in or returning to a state.

MCI Transition Risks

increases risk for aMCI & mMCI

risk increased for aMCI

lower education increases risk

for mMCI

Kryscio, et al., Neurology, 2006

no increased risk

Family Hx
no increased risk

Test-based MCI

At next visit:

aMCITB: Memory tests show reduced learning and/or delayed recall (-1.5 ) mMCITB: Praxis, executive, and/or language tests are also 1.5 below average MCICC: Consensus clinical criteria for are met for MCI

40.7% show no back transitions from aMCITB 43.8% show no back transitions from mMCITB 4.4% transition from aMCITB to MCICC or dementia 7.1% transition from aMCITB to MCICC or dementia
Abner, et al., Int. J. Alz. Dis., 2012


Hypertension (1.5X) increases mortality risk mMCITB (4.9X) increases risk for dementia and mortality (2.7X)

aMCITB & mMCITB: Biggest risk is Age Low education predicts transition to mMCITB Family history and female gender are protective for mMCITB MCICC: risks include ApoE4, age, low education aMCITB more than doubles the risk for MCICC mMCITB more than quadruples the risk for MCICC

SMC: Subjective Memory Complaint

Have you noticed a change in your memory since the last visit?
YES response implies the participant enters a state called SMC
55.7% answered YES on average 8.3 yrs. after study


Determine risk factors associated with transitions into impaired states Assess neuropathology examinations for the 243 participants who came to autopsy

SMC & Clinical Transitions

Neuropathology, SMC & Conversions to MCI / Dementia

SMC? No Yes Yes MCI or dementia? No No Yes Number of Autopsies 56 120 50 Neuritic Plaques Low Moderate* High** NFT Low Moderate* High**



SMC+ Cog+


SMC+ Cog-


SMC+ Cog+

Implications of SMC

Over half of volunteers have Subjective Memory Complaints during follow-up SMC accounts for most MCI/dementia transitions (OR = 2.8) SMC is no guarantee MCI/dementia will occur Identified risks affect the time and probability for transitions to occur

Autopsy results indicate a subset of individuals (but not all) have AD type pathology
Can inform the design of future prevention trials

Research Team:
Richard J. Kryscio & Frederick A. Schmitt (PI s) Erin Abner, Ph.D., Peter Nelson, M.D., Ph.D, Gregory Jicha, M.D., Ph.D., David Fardo, Ph.D.