Table 3.
Linear regression β-Coefficients or ordinal logistic regression OR with 95% confidence intervals for associations of hand grip strength and gait speed with neuroimaging biomarkers and cognitive function.
| N = 674 | Model 1 |
p-value | Model 2 |
p-value | p-value adjusteda |
|---|---|---|---|---|---|
| β (95% CI) | β (95% CI) | ||||
| Total mean cortical thicknessd | |||||
| Hand grip strengthb | 0.0002 (−0.0002, 0.0006) | 0.30 | 0.0002 (−0.0002, 0.0006) | 0.37 | 0.37 |
| Gait speedc | 0.039 (0.005, 0.074) | 0.026 | 0.038 (0.001, 0.075) | 0.047 | 0.094 |
| AD-signature cortical thicknessd | |||||
| Hand grip strengthb | 0.0003 (−0.0002, 0.0007) | 0.28 | 0.0003 (−0.0002, 0.0007) | 0.29 | 0.47 |
| Gait speedc | 0.042 (0.000, 0.084) | 0.052 | 0.046 (0.000, 0.092) | 0.052 | 0.16 |
| Hippocampal volumed | |||||
| Hand grip strengthb | 2.25 (0.65, 3.85) | 0.0060 | 2.00 (0.39, 3.61) | 0.015 | 0.044 |
| Gait speedc | 214.59 (71.10, 358.08) | 0.0034 | 174.61 (18.96, 330.27) | 0.028 | 0.084 |
| Predicted brain agef | |||||
| Hand grip strengthb | −0.010 (−0.019, −0.001) | 0.022 | −0.008 (−0.017, 0.001) | 0.073 | 0.15 |
| Gait speedc | −1.11 (−1.91, −0.31) | 0.0070 | −0.52 (−1.38, 0.35) | 0.24 | 0.36 |
| Global cognitive composite scoree | |||||
| Hand grip strengthb | 0.003 (0.000, 0.006) | 0.085 | 0.001 (−0.002, 0.004) | 0.36 | 0.43 |
| Gait speedc | 0.595 (0.327, 0.863) | <0.0001 | 0.341 (0.065, 0.618) | 0.016 | 0.019 |
| OR (95% CI) | p-value | OR (95% CI) | p-value | p-value adjusteda | |
|---|---|---|---|---|---|
| Cerebral small vessel disease scoreg | |||||
| Hand grip strengthb | 0.99 (0.98, 1.00) | 0.12 | 1.00 (0.99, 1.01) | 0.28 | 0.36 |
| Gait speedc | 0.20 (0.09, 0.47) | 0.00020 | 0.31 (0.12, 0.77) | 0.011 | 0.043 |
Note: Analyses were performed with linear or ordinal logistic regression models. Dependent variables in the linear regression models were total mean cortical thickness, mean cortical thickness in AD-signature regions, hippocampal volume, predicted brain age, and a global cognitive composite score. Dependent variable in the ordinal logistic regression model (cumulative logit model) was the cerebral small vessel disease (CSVD) score. Independent variables were hand grip strength and gait speed. Hand grip strength and gait speed were investigated in separate analyses. Model 1 was adjusted for sex, and model 2 was adjusted for sex, physical activity, education, and smoking. There were missing CSVD data for 4 participants, missing gait speed data for 10 participants (1.4%), and missing hand grip strength data for 1 participant (0.1%).
FDR (false discovery rate) adjusted p-values were added to explore the impact of conducting multiple comparisons (including the total number of analyses reported in Table 2, Table 3, Table 4) in model 2.
Handgrip strength (kPa), best value out of three on the dominant hand.
Gait speed self-selected 30 m indoor (meter/second).
Total mean cortical thickness is measured in mm. Alzheimer's disease (AD)-signature: Averaging cortical thicknesses of entorhinal, inferior temporal, middle temporal, and fusiform regions, adjusted for surface area, measured in mm. Hippocampal volume was adjusted for total intracranial volume, measured in mm3.
The global cognitive composite score (z-score) was derived from eight cognitive tests covering five cognitive domains: executive function, perceptual speed, verbal fluency, episodic memory, and visuospatial abilities.
Predicted brain age is investigated by the difference in years between chronological age and predicted biological age (brain age gap). A negative value indicates a biologically younger brain, and a positive value a biological older brain, compared to chronological age.
The CSVD score ranged from 0 (no cerebrovascular burden)—3 (high cerebrovascular burden).