Table 3.
Term | Coefficient | Standard error | t-value | p-value |
---|---|---|---|---|
Intercept | 3.877795 | 0.06891898 | 56.26600 | <0.001 |
Sex effect | 0.223058 | 0.10668516 | 2.09081 | 0.037 |
Aβ42 effect | −0.007031 | 0.00100517 | −6.99508 | <0.001 |
Baseline rate | 0.023720 | 0.00184203 | 12.87688 | <0.001 |
Baseline curvature | 0.000177 | 0.00002078 | 8.52360 | <0.001 |
Age effect | 0.002999 | 0.00680578 | 0.44068 | 0.660 |
Education effect | 0.000530 | 0.01816607 | 0.02918 | 0.977 |
Baseline cognition effect | 0.139866 | 0.01143890 | 12.22720 | <0.001 |
Female effect on slope | 0.006420 | 0.00290663 | 2.20866 | 0.027 |
Female effect on curvature | −0.000006 | 0.00003607 | −0.15359 | 0.878 |
Baseline Aβ42 effect on slope | −0.000216 | 0.00002632 | −8.21396 | <0.001 |
Baseline Aβ42 effect on curvature | −0.000001 | 0.00000028 | −3.52631 | <0.001 |
Education effect on slope | 0.000301 | 0.00048418 | 0.62259 | 0.534 |
Age effect on slope | −0.000006 | 0.00018281 | −0.03192 | 0.975 |
Baseline cognition effect on slope | 0.000378 | 0.00030937 | 1.22315 | 0.221 |
Baseline cognition indicates ADAS-Cog 11. Bold p-values are statistically significant. Abbreviations: MCI (mild cognitive impairment), ADAS-Cog11 (Alzheimer’s disease assessment scale- cognitive subscale), and Aβ42 (amyloid-beta1–42). In this model, the follow-up time (month) was centered with the median follow-up time (36 months) and covariates were centered i.e. a 75 years old with 16 years of education, ADAS-Cog11 of 11, and Aβ42 of 147. Table depicts that the effect of sex on ADAS-Cog11 slope was significant with females declining faster than males. Baseline Aβ42 also had a significant effect on both slope and curvature. Age and education did not have significant effects on slope. The intercept is a term to get the correct estimate of the outcome when time = 0. The baseline rate is the reference population rate of change in the outcome starting at time zero, and the baseline curvature is the “acceleration” of that rate of change at time zero. The upper half of the table shows the effect of specific variables on ADAS-Cog11 and the bottom half shows their effects on ADAS-Cog11 slope and curvature.