Statistical analysis
Results were analysed and presented as means and standard deviations.
Pre-operative covariates were assessed for normal distribution using the
Shapiro-Wilk test. Between group characteristics were assessed for
statistical differences using the Student T test or Wilcoxon Rank Test
for non-parametric variables. Multivariate logistic regression models
were constructed to assess the influence of a variety of covariates on
short term outcomes. Linear regression was used to assess the influence
of covariates on parametric outcomes, namely hospital stay. Adjusted
odds ratio with 95% confidence interval (CI) of binary outcomes were
calculated. Crude survival curves were estimated using the nonparametric
Kaplan-Meier
method, and log rank tests were used to compare the survival
distribution among groups. Cox proportional hazard regression was
conducted to calculate the adjusted hazards ratio with 95% CI. Stepwise
selection was performed using age, sex, COPD, Euroscore, NYHA class and
LV function in the model. Statistical analyses were conducted using the
Stata 13.0 software (Stata Corp., College Station, TX, USA).