Association of Longitudinal β-Amyloid Accumulation Determined by Positron Emission Tomography With Clinical and Cognitive Decline in Adults With Probable Lewy Body Dementia.

Importance: In patients with probable dementia with Lewy bodies (DLB), overlapping Alzheimer disease pathology is frequent and is associated with faster decline and shorter survival. More than half of patients with DLB have elevated β-amyloid levels on carbon-11 labeled Pittsburgh compound B (PiB) positron emission tomography, but the trajectory of longitudinal β-amyloid accumulation and its associations with clinical and cognitive decline in DLB are not known. Objectives: To determine the trajectory of β-amyloid accumulation in patients with probable DLB and to investigate the associations of β-amyloid accumulation with measures of clinical and cognitive decline over time in DLB. Design, Setting, and Participants: This cohort study included 35 consecutive patients with probable DLB from the Mayo Clinic Alzheimer Disease Research Center and matched them by age, sex, and apolipoprotein e4 status with 140 cognitively unimpaired participants from the population-based Mayo Clinic Study of Aging. Participants were observed from April 2010 to September 2017. Data analysis was conducted from January 2018 to January 2019. Exposure: Baseline and follow-up PiB positron emission tomography and comprehensive clinical evaluations. Main Outcomes and Measures: Rate of change in PiB standardized uptake value ratios (SUVRs) by PiB SUVR and time in years; the associations between baseline PiB SUVR, change in PiB SUVR, and change in several measures of clinical and cognitive decline. Results: A total of 175 participants were evaluated (35 [20.0%] with probable DLB; mean [SD] age, 69.6 [7.3] years; 16 [45.7%] apolipoprotein e4 carriers; 31 [88.6%] men; and 140 [80.0%] cognitively unimpaired adults; mean [SD] age, 69.7 [7.2] years; 64 [45.7%] apolipoprotein e4 carriers; 124 [88.6%] men). In both groups, the rates of change in PiB SUVR showed an initial acceleration at lower baseline PiB SUVR followed by a deceleration at higher baseline PiB SUVR, thus forming an inverted-U shape. The trajectories of the rates of change in PiB SUVR did not differ between participants with probable DLB and cognitively unimpaired participants in terms of shape (P = .59) or vertical shift (coefficient [SE] 0.007 [0.006]; P = .22). The integral association of cumulative PiB SUVR with time in years showed a sigmoid-shaped functional form in both groups. In participants with probable DLB, higher baseline PiB SUVR and change in PiB SUVR were associated with more rapid clinical decline, as measured by the Clinical Dementia Rating, sum of boxes (baseline PiB SUVR: regression coefficient [SE], 1.90 [0.63]; P = .005; R2 = 0.215; change in PiB SUVR, regression coefficient [SE], 16.17 [7.47]; P = .04; R2 = 0.124) and the Auditory Verbal Learning Test, delayed recall (baseline PiB SUVR, regression coefficient [SE], -2.09 [0.95]; P = .04; R2 = 0.182; change in PiB SUVR, regression coefficient [SE], -25.05 [10.04]; P = .02; R2 = 0.221). Conclusions and Relevance: In this study, the rate of change in PiB SUVR among participants with probable DLB increased, peaked, and then decreased, which was similar to the trajectory in cognitively unimpaired participants and the Alzheimer disease dementia continuum. Higher baseline PiB SUVR and change in PiB SUVR were associated with more rapid clinical and cognitive decline over time. Measuring the change in PiB SUVR has implications for designing anti-β-amyloid randomized clinical trials for individuals with probable DLB.