Table 4:
Multivariate Linear Regression Model Showing Determinants of LVMI Regression
| Variables | β coefficient ± SE | p |
|---|---|---|
|
| ||
| Age, per 5y increment | −0.28±0.14 | 0.02 |
| Male sex | 0.18±0.24 | 0.1 |
| Stage 1 hypertension at Y1* | −0.31±0.18 | 0.02 |
| Stage 2 hypertension at Y1* | −0.38±0.21 | 0.009 |
| Doppler COA mean gradient | 0.16±0.43 | 0.5 |
| Doppler aortic valve mean gradient | −0.08±0.14 | 0.1 |
| Transcatheter stent therapy | 0.23±0.35 | 0.3 |
COA: Coarctation of aorta; SE: standard error; LVMI: Left ventricular mass index
Severity of hypertension at Y1 modeled as categorical variable using the normal/elevated BP group as the reference group.
A diagnosis of hypertension pre-intervention was also independently associated with regression of LVMI (β coefficient ± SE: −0.49±0.23, p0.005). However, we did not input both pre-intervention hypertension and post-intervention hypertension in the same model because of collinearity