Table 4.
Univariate linear regression models for activity outcomes.
| Predictorsa | Lab-based outcomes | Accelerometer-based outcomes | ||||
| Gait speed | Gait speed | Low activity | ||||
| Coefficient (95% CI) | P value | Coefficient (95% CI) | P value | Coefficient (95% CI) | P value | |
| HIV status | −0.01 (−0.11 to 0.10) | .91 | −0.15 (−0.34 to 0.03) | .10 | 0.16 (0.02 to 0.29) | .02 |
| Body mass index | −0.32 (−0.63 to −0.02) | .04 | −0.44 (−0.98 to 0.097) | .11 | −0.06 (−0.50 to 0.38) | .78 |
| Total testosterone | 0.06 (−0.05 to 0.18) | .26 | 0.21 (0.02 to 0.40) | .04 | −0.02 (−0.19 to 0.14) | .77 |
| Free testosterone | 0.05 (−0.06 to 0.17) | .34 | 0.21 (0.01 to 0.40) | .04 | −0.05 (−0.22 to 0.13) | .60 |
| CD4/CD8 ratio | 0.01 (−0.06 to 0.08) | .81 | 0.09 (−0.03 to 0.20) | .14 | −0.10 (−0.18 to −0.02) | .02 |
aAll variables were natural log-transformed before analysis.