Longitudinal Data Analysis with stratification:
Stratification by the sex, type of transposition, balloon atrial
septostomy and the operating surgeon did not show significant
differences. The most notable finding in the stratified longitudinal
data display was the rate of drop of the pressure gradient between those
who underwent reintervention and those who did not (Figure 8).
Longitudinal data analysis with mixed effect modeling showed that the
significant variables in relation to the peak pressure gradient
following arterial switch operation included: younger age at surgery,
complex transposition, and longer bypass time.
As the rate of change of pressure gradient immediately after surgery was
the most notable difference between patients based on reintervention.
Further testing with proportional hazard modeling for all variables with
time-to-reintervention as an outcome with stepwise selection methodology
showed that the only predictive preoperative variable for reintervention
was the total peak gradient (MPA+RPA+LPA) (p<0.0001). Further
analysis with the area under the receiver operating characteristics
curve showed that a total peak gradient 55 mmHg or more was a predictor
for reintervention with a hazard ratio of 12.