Statistical analysis
Statistical analyses were performed using statistical software STATISTICA(TIBCO Software Inc.2017,version-13.Palo Alto-USA). Continuous variables were expressed as means±SD, while categorical variables as numbers and percentages. For continuous data Student’s t-test was used for in between groups comparisons, while categorical variables were compared with Pearson-χ2-test. Cox regression models were used to find the univariable and multivariable predictors of late all-cause mortality. Multivariable model for the whole cohort was built using preoperative variables presented in Table-1 in addition to type of surgery: On-pump-CABG(ONCAB) or off-pump-CABG(OPCAB) and completeness of revascularization. The multivariable model considered all univariate significant variables(p <0.1) with using the stepwise method(backward and forward methods resulted in the same model).
To further confirm the results and to reduce the risk of selection bias inherent to a retrospective, observational studies, a propensity score(PS) matching was used to match patients between the groups(SITA+Vein group vs 2nd-arterial group)and(RITA group vs RA group). Propensity scores were generated from a multivariable logistic regression model based on 18-preoperative variables as mentioned in Table-1. Patients were then matched in 1:1 fashion using a caliper matching method without replacement with a caliper width of 0.2-standard deviation of the logit of the PS18-20. The balance of the covariates was tested using standardized mean difference(SMD). Statistical guidelines suggest a meaningful covariate balance of the variables used to generate the PS between the two groups to be between -0.1≤SMD≤ 0.118.Matched data were analyzed using procedures for matched analyses. McNemar’s test was used for binary outcomes.
Survival in the unadjusted and PS adjusted populations were estimated using the Kaplan-Meier method and were expressed as percentages. Log-rank test was used to compare the data. Statistical significance was defined as p- value of<0.05.