Table 2.
Npatients | Nsites | Estimate [95%CI] | p | Marginal means [95%CI] | ||
---|---|---|---|---|---|---|
Aβ-negative | Aβ-positive | |||||
| ||||||
Age at symptom onset | 644 | 28 | −1.322 [−3.130, 0.477] | 0.15 | 60.4 [58.4, 62.4] | 59.1 [57.8, 60.4] |
Age at diagnosis | 671 | 30 | −1.415 [−3.18, 0.348] | 0.12 | 63.9 [62.0, 65.8] | 62.5 [61.4, 63.6] |
MMSE at diagnosis | 596 | 28 | −0.059 [−1.350, 1.290] | 0.93 | 20.8 [19.5, 22.2] | 20.7 [19.9, 21.6] |
CDR at diagnosis (%CDR≥1) | 413 | 21 | 0.118 [−0.573, 0.818] | 0.74 | 61% [42, 77] | 58% [46, 70] |
Sex (% Women) | 689 | 30 | 0.406 [−0.058, 0.867] | 0.085 | 53% [42, 63] | 63% [59, 67] |
APOE (% e4 carriers) | 342 | 22 | 0.089 [−0.639, 0.818] | 0.81 | 45% [29, 62] | 47.5% [40, 55] |
PCA diagnosis (% PCA pure) | 625 | 29 | 1.506 [0.769, 2.26] | <0.0001 | 81% [56, 94] | 95% [86, 98] |
MRI (% posterior atrophy) | 543 | 27 | 0.226 [−0.605, 1.060] | 0.59 | 92% [78, 97] | 94% [85, 97] |
FDG-PET (% post. hypometabolism) | 324 | 22 | 0.578 [−0.896, 2.05] | 0.44 | 97% [84, 99] | 98% [94, 99] |
For each variable, a separate linear mixed model (continuous outcomes) or generalized mixed effect model (binary outcomes) was run, using amyloid status as a fixed effect and site as a random effect. Estimate 95%CI were obtained using the likelihood profile method. Number of participants and sites with available data varies from one variable to the other. Aβ-positive patients included those who received a positive result on CSFAβ42 and/or amyloid-PET, and Aβ-negative patients received negative results from CSF or PET. This sample included 390 with CSF only, 153 with PET only, and 146 with both; in this subgroup 27 patients had discrepant results (3 CSF+/PET-, 24 CSF-/PET+) and were included in the amyloid-positive group.