Table 3.
Unadjusted mediating effects of BMI and WC on the association between HAART and blood pressure (N = 406)
| Mediator variable | Predictor variable | Outcome variable | Path A | Path B | Path C’ | Path AB | % mediated |
|---|---|---|---|---|---|---|---|
| Coefficient (95 % CI) | Coefficient (95 % CI) | Coefficient (95 % CI) | Coefficient (95 % CI) | ||||
| BMI | HAART | Systolic BP | 1.16 (0.16 to 2.16) | 1.11 (0.70 to 1.52) | 8.28 (4.15 to 12.41) | 1.29 (0.37 to 2.70) | 14 |
| Diastolic BP | 1.16 (0.16 to 2.16) | 0.62 (0.42 to 0.83) | 6.77 (4.65 to 8.88) | 0.73 (1.01 to 1.35) | 10 | ||
| WC | HAART | Systolic BP | 3.42 (0.92 to 5.92) | 0.39 (0.23 to 0.56) | 7.96 (3.86 to 12.06) | 1.35 (0.24 to 2.76) | 15 |
| Diastolic BP | 3.42 (0.92 to 5.92) | 0.26 (0.17 to 0.34) | 6.71 (4.60 to 8.81) | 0.87 (0.17 to 1.69) | 12 | ||
| Combined BMI/WC | HAART | Systolic BP | 4.36 (0.85 to 7.89) | 1.01 (0.24 to 1.78) | 8.15 (4.08 to 12.23) | 1.51 (0.28 to 3.15) | 16 |
| Diastolic BP | 4.36 (0.85 to 7.89) | 0.54 (0.15 to 0.94) | 6.87 (4.79 to 8.95) | 0.95 (0.19 to 2.24) | 12 |
Path A – regression coefficient for the independent variable when the mediator variable is regressed on the independent variable
Path B – regression coefficient for the mediator variable when the dependent variable is regressed on the mediator variable, adjusted for the independent variable
Path C – the direct effect of the independent variable on the dependent variable
Path AB – the indirect effect of the independent variable on the dependent variable through the mediator variable
% mediated – the amount (in percentage) of the total effect of the independent variable on the dependent variable that is attributed indirectly to the mediator variable
NS – not statistically significant at 5 % level; all other coefficients were statistically significant
95 % confidence intervals for indirect effects (Path AB) are bias-corrected
Abbreviations: BMI body mass index, CI confidence interval, HAART highly active antiretroviral therapy, WC waist circumference