Statistical Analysis
Statistical analysis was performed using SPSS version 23 (IBM Corp., USA). Quantitative variables were expressed as means ± standard deviation (SD) or medians and Interquartile Range [IQR, (25th percentile, 75th percentile)]. Categorical variables were reported as numbers and percentages. Preoperative, intraoperative and postoperative patients’ characteristics were analyzed descriptively and compared among cohorts with between-cohort differences assess using chi-squared tests or Fisher’s exact test for categorical variables and Wilcoxon rank-sum tests for the continuous variables both before and after matching.
Propensity score matching was then used to mitigate underlying differences in covariates between the cohorts, and a multivariable logistic regression was used to build the propensity score model. Module of Propensity score matching for SPSS (Version 3.0.4) was used. All variables were retained in formulating the propensity score regardless of statistical significance because they were believed to be clinically relevant. Upon the estimation of propensity score, a greedy matching algorithm matched the LVD cohort to NLVF cohort 1:2. To ensure pairs were precisely matched on important covariates, the following variables were exact-matched: age groups (<60 years vs ≥60 years) and gender. Only patients with complete data sets were used in the creation of the propensity score.
Kaplan-Meier analysis was then used to estimate time to all cause death and cardio-death in LVD cohort versus NLVF cohort. To account for pairing, between-cohort differences were assed using partial likelihood ratio tests. Next, the number and proportion of patients who experienced major cardiovascular disease within the follow up period was estimated, and between-cohort differences were calculated using McNemar’s test.
Multivariable Cox proportional hazard regression analysis was used to quantify the relative risk of mid-term mortality in the LVD cohort versus the NLVF cohort, controlling for covariates and stratified by matched pairs. Statistical significance was accepted as a two-sided test with an alpha level of 0.05.