Table 2.
Association of brain copper levels with cognitive decline (global cognition and different domains) among deceased Memory and Aging Project participants.
| N | Composite Brain copper levels (four regions)a | |||||
|---|---|---|---|---|---|---|
| Continuous | Tertile comparison | |||||
| T1 | T2 | T3 | P -trend | |||
| Change in cognition, global and the domains | ||||||
| Global Cognition β (SE, p-value) | 621 | 0.028 (0.01,0.001) | Ref | 0.021 (0.011,0.054) | 0.030 (0.011,0.009) | 0.009 |
| Episodic memory β (SE, p-value) | 619 | 0.027 (0.01, 0.006) | Ref | 0.015 (0.012,0.216) | 0.028 (0.013, 0.029) | 0.03 |
| Semantic memory β (SE, p-value) | 609 | 0.035 (0.01, 0.0008) | Ref | 0.023 (0.013,0.074) | 0.030 (0.013, 0.025) | 0.03 |
| Visuospatial ability β (SE, p-value) | 601 | 0.013 (0.01, 0.183) | Ref | 0.010 (0.012,0.401) | 0.007 (0.012, 0.559) | 0.60 |
| Perceptual speed β (SE, p-value) | 600 | 0.024 (0.01, 0.007) | Ref | 0.030 (0.011, 0.006) | 0.028 (0.011, 0.015) | 0.02 |
| Working memory β (SE, p-value) | 621 | 0.026 (0.01, 0.003) | Ref | 0.017 (0.011, 0.111) | 0.030 (0.011, 0.007) | 0.007 |
| Alzheimer’s disease (AD) pathology outcome | ||||||
| Global AD pathology, β (SE, p-value) | 657 | −0.069 (0.019,0.0004) | Ref | −0.118 (0.035,0.0008) | −0.095 (0.035,0.007) | 0.006 |
| Amyloid load, β (SE, p-value) | 657 | −0.132 (0.060,0.027) | Ref | −0.315 (0.11, 0.003) | −0.206 (0.11, 0.058) | 0.092 |
| Phosphorylated -tau tangles, β (SE, p-value) | 657 | −0.289 (0.068,0.00003) | Ref | −0.370 (0.12, 0.003) | −0.358 (0.12, 0.004) | 0.001 |
Composite brain copper levels: Brain copper in all four regions combined (inferior temporal, mid-frontal, anterior cingulate, and cerebellum).
Statistical models for cognitive outcomes were linear mixed-effects regression models; estimates are coefficients of the interaction of brain copper composite with time before death. The linear mixed-effect models were controlled for age at death, sex, education, APO-ε4 status, time and time * all covariates. Statistical models for AD pathology outcomes were linear regression models; estimates are coefficients of brain copper composite. All the models were controlled for age at death, sex, education, and APO-ε4 status.