Extensive BITA Grafting in the Canary Islands: an IPTW Analysis

Overview

Coronary artery bypass grafting (CABG) remains the preferred revascularization strategy for patients with complex multivessel coronary artery disease. The left internal thoracic artery is universally accepted as the standard conduit for grafting the left anterior descending artery, but the long-term clinical benefit and safety of bilateral internal thoracic artery (BITA) grafting compared with single internal thoracic artery (SITA) grafting remain debated. This retrospective cohort study compares long-term outcomes after BITA versus SITA grafting in consecutive adult patients undergoing first-time isolated CABG with at least two coronary grafts between May 1, 2017 and October 31, 2025 at Hospital Universitario de Gran Canaria Dr. Negrin. The primary endpoint is all-cause mortality during follow-up. Secondary outcomes include cardiovascular mortality, non-cardiovascular mortality, deep sternal wound infection, stroke, myocardial infarction, repeat coronary revascularization, and major adverse cardiovascular events (MACE). Treatment groups will be balanced using stabilized inverse probability of treatment weighting derived from a multivariable propensity score model including clinically relevant baseline variables. Long-term outcomes will be analyzed using weighted Kaplan-Meier survival curves and Cox proportional hazards regression.

Background and scientific rationale: Coronary artery disease remains a leading cause of mortality worldwide. Coronary artery bypass grafting is recommended as a preferred revascularization strategy for selected patients with complex multivessel coronary artery disease and left main coronary artery disease because of durable long-term clinical benefits. Although the left internal thoracic artery graft to the left anterior descending artery has superior long-term patency and survival compared with venous conduits, routine use of bilateral internal thoracic arteries remains limited. Observational studies have suggested improved long-term survival following BITA grafting, whereas randomized evidence has not consistently demonstrated a statistically significant survival advantage in intention-to-treat analyses. Concerns regarding deep sternal wound infection, particularly in patients with diabetes mellitus, obesity, and chronic pulmonary disease, have also limited wider adoption. Study hypothesis: The use of bilateral internal thoracic artery grafting is associated with improved long-term survival compared with single internal thoracic artery grafting without increasing the adjusted risk of deep sternal wound infection. Any survival benefit may be primarily explained by a reduction in cardiovascular mortality rather than non-cardiovascular mortality. Primary objective: To compare long-term all-cause mortality between patients undergoing bilateral internal thoracic artery grafting and those receiving single internal thoracic artery grafting after adjustment using stabilized inverse probability of treatment weighting. Secondary objectives: To compare cardiovascular mortality, non-cardiovascular mortality, deep sternal wound infection, stroke, myocardial infarction, repeat coronary revascularization, major adverse cardiovascular events, early versus late mortality, temporal distribution of mortality, and predictors of deep sternal wound infection. Statistical analysis plan: Baseline demographic, clinical, and operative characteristics will be summarized according to arterial revascularization strategy. Continuous variables will be reported as mean with standard deviation or median with interquartile range, depending on distribution. Categorical variables will be reported as frequencies and percentages. Comparisons will use Student t-test or Mann-Whitney U test for continuous variables and Pearson chi-square or Fisher exact test for categorical variables, as appropriate. Propensity scores will be estimated using multivariable logistic regression representing the conditional probability of receiving BITA grafting. Covariates will include age, sex, body mass index, obesity, diabetes mellitus, chronic obstructive pulmonary disease, current smoking, hypertension, peripheral arterial disease, cerebrovascular disease, chronic dialysis, preoperative cardiac rhythm, serum creatinine greater than 200 micromol/L, left ventricular ejection fraction, left main coronary artery disease, and use of cardiopulmonary bypass. Stabilized inverse probability of treatment weights will be calculated using the marginal probability of treatment assignment. Balance will be assessed using standardized mean differences, with adequate balance defined as an absolute standardized mean difference less than 0.10. Overall survival will be estimated using Kaplan-Meier methodology. The primary comparison will be performed using IPTW-adjusted Cox proportional hazards regression and reported as hazard ratios with 95% confidence intervals and two-sided P values. Cause-specific mortality analyses will be performed separately for cardiovascular and non-cardiovascular mortality. Competing-risk regression using the Fine and Gray subdistribution hazards model will be performed when event numbers permit. Sensitivity analyses will include complete-case analysis, conventional multivariable regression, alternative propensity score specifications, exclusion of patients with very short follow-up, exclusion of deaths with undetermined cause, and multiple imputation analyses if appropriate. Exploratory subgroup analyses will be performed according to age 70 years or older, diabetes mellitus, obesity, chronic obstructive pulmonary disease, reduced left ventricular ejection fraction, left main coronary artery disease, and use versus non-use of cardiopulmonary bypass. Ethics and data protection: The study will be conducted in accordance with the Declaration of Helsinki, Good Clinical Practice principles applicable to observational research, European General Data Protection Regulation, Spanish Organic Law 3/2018 on Personal Data Protection, and Spanish Biomedical Research Act Law 14/2007. All patient information will be pseudonymized before analysis.