Examining multimorbidity contributors to dementia over time

Abstract

Introduction: Multimorbidity is associated with increased risk of dementia, but previous estimation of the joint contribution of constituent conditions to dementia incidence did not model additive contributions or temporal proximity in the sequential onset of conditions.

Methods: Data were analyzed from 9944 Health and Retirement Study participants and Medicare fee-for-service beneficiaries, ages 68-99, without Alzheimer's disease and related dementias (ADRD) at baseline, from 1998-2016. ADRD and chronic condition were encoded using validated claims algorithms. We estimated the absolute contribution of eight conditions to ADRD with the longitudinal extension of the average attributable fraction (LE-AAF).

Results: Hypertension, acute myocardial infarction, atrial fibrillation, diabetes, heart failure, ischemic heart disease, stroke, and arthritis additively accounted for 71.8% (95% confidence interval [CI]: 62.9%-79.1%) of ADRD incident cases based on LE-AAF.

Discussion: Our findings suggest that multimorbidity plays a pivotal role in ADRD incidence. Targeting constituents of a cardiovascular path to dementia may contribute most to lowering dementia risk.

Highlights: Most dementia cases (71.8%) were attributable to eight chronic conditions. Hypertension was the largest contributor to dementia risk. Confidence intervals were smallest for constituents of a cardiovascular path to dementia. Longitudinal extension of the average attributable fractions (LE-AAFs) explicitly consider longitudinal patterns of comorbidities. Acute myocardial infarction did not contribute significantly to dementia incidence.

Keywords: ADRD; Alzheimer's disease and related dementias; longitudinal extension of the average attributable fraction ( LE‐AAF); multimorbidity; population attributable fraction ( PAF); population attributable risk.

FIGURE 1

Flowchart of inclusion. *CCW algorithms cannot be applied to participants enrolled in managed care or without Medicare Parts A and B coverage, resulting in missing chronic condition status. ^†Of observations with missing data on HRS covariates, 99% had missing BMI. ADRD, Alzheimer's disease and related dementias; BMI, body mass index; CCW, Chronic Conditions Data Warehouse; FFS, fee‐for‐service; HRS, Health and Retirement Study.

FIGURE 2

Prevalence of cardiometabolic multimorbidity combinations at baseline. Note that prevalence combinations ocurring in less than 1% of the sample are not depicted. AFIB, atrial fibrillation; AMI, acute myocardial infarction; HF, heart failure; HTN, hypertension; IHD, ischemic heart disease; RA/OA, rheumatoid arthritis or osteoarthritis; STIA, stroke or transient ischemic attack.

FIGURE 3

(A) Odds ratios with 95% CIs from the pooled logistic regression. (B) LE‐AAF estimates with 95% BCa CIs (black circles and error bars), PAFs based on PAF_ORs (red diamonds) and PAF_RRs (blue circles). AFIB, atrial fibrillation; AMI, acute myocardial infarction; HF, heart failure; HTN, hypertension; IHD, ischemic heart disease; PAFs, population attributable fractions; PAF_ORs, adjusted odds ratios; PAF_RRs, unadjusted risk ratios; RA/OA, rheumatoid arthritis or osteoarthritis; STIA, stroke or transient ischemic attack.