Post-Kidney Transplant Dementia and Alzheimer's Disease Prediction Leveraging the Pre-Transplant Evaluation Data
Funded Grant
Overview
Affiliation
View All
Overview
description
ABSTRACT Of the >225,000 older kidney transplant (KT) recipients in the US, 19% experience post-KT cognitive decline, which elevates their dementia and Alzheimer’s disease (AD) risk. Compared to a 10-year risk for community- dwelling older adults of <1%, the risk of post-KT dementia/AD diagnosis is 7.2-17.0%, and this even underestimates the problem, as only half of those who meet diagnostic criteria receive a diagnosis. Transplant evaluation is the ideal setting for post-KT dementia/AD risk prediction because there is ample time to intervene with cognitive prehabilitation/rehabilitation, routine monitoring for cognitive changes, caregiver medication training, immunosuppression tailoring, and/or enhanced nursing interventions. For example, cognitive prehabilitation improves pre-KT cognition in just 3 months. Yet, predicting post-KT dementia/AD risk with pre- KT data is not part of KT evaluation. Existing models from the general geriatrics population have poor predictive ability in KT and are no better than those with only age at KT. They fail to generalize to this unique population because KT is a major surgery that can lead to delirium and cognitive decline, both are exacerbated by immunosuppression. It is likely that in KT, the upper bound of prediction (best general population models) is 0.86, but will require burdensome testing that is not part of KT evaluation. Pre-KT geriatric-specific risk factors as well as tabular and multimodal data will likely be required for further prediction improvement. We hypothesize that post-KT dementia/AD risk can be predicted using tabular/multimodal data during pre-KT evaluation and updated with new data at the time of KT. We will use Cosmos data on >170,000 older (≥50) KT candidates to predict post-KT dementia/AD risk during KT evaluation. Then we will use our existing NIA-funded cohort of 3,900 older candidates (FAIR) to establish a new study (FAIR-COG) with up to 14 years of follow-up for internal validation, recruiting an additional 500 older candidates for external validation. We seek to: 1) design an assessment to predict post-KT dementia/AD risk during KT evaluation using tabular data; 2) evaluate whether multimodal data from FAIR-COG improves dementia/AD risk prediction; and 3) study validity, implementation, and generalizability of the best assessment from Aims 1 and 2. We will improve post-KT dementia/AD risk prediction for the 21,000 older candidates evaluated each year. This will allow for the identification of patients who would benefit from pre- and post-KT interventions, like cognitive prehabilitation/rehabilitation and immunosuppression tailoring. Ultimately, these methods for developing a valid and implementable risk prediction will be generalized to other surgical settings.
