Comparative Analysis of Coronary Artery Bypass Grafting Outcomes in Women Using Different Conduits in the National UK Data Set

BACKGROUND

There is limited report of outcomes in women undergoing isolated coronary artery bypass grafting (CABG) with left internal thoracic artery and different second conduits (saphenous vein graft [SVG], radial artery [RA], and right internal thoracic artery [RITA]).

METHODS

The National Adult Cardiac Surgery Audit database was queried for women undergoing isolated CABG with left internal thoracic artery graft in the United Kingdom from 1996 to 2019. Propensity score-based pairwise comparisons were performed between graft types. The primary outcome was in-hospital mortality.

RESULTS

The study included 58,063 women (SVG, n = 48,881 [84.2%]; RA, n = 6136 [10.6%]; RITA, n = 2445 [4.2%]). SVG use was stable over the years; RA and RITA use decreased. In-hospital mortality was similar between the RA and RITA grafts (2.3% vs 2.8%; odds ratio [OR], 0.80; 95% CI, 0.53-1.22; P = .39) and between the RA and SVG (2.3% vs 2.0%; OR, 1.20; 95% CI, 0.93-1.55; P = .17) but higher in the RITA group compared with the SVG (2.7% vs 1.4%; OR, 2.04; 95% CI, 1.27-3.36; P = .004). Women receiving the RITA graft were more likely to have sternal wound infection (SWI) compared with the RA (0.6% vs 0.06%; P = .004) and the SVG (0.6% vs 0.2%; P = .032). SWI was consistently associated with higher risk of in-hospital mortality.

CONCLUSIONS

Conduit selection may affect operative outcomes in women undergoing CABG. The RA shows similar mortality and risk of deep SWI as the SVG.

Current concepts in coronary artery revascularisation

Coronary artery revascularisation can be performed surgically or percutaneously. Surgery is associated with higher procedural risk and longer recovery than percutaneous interventions, but with long-term reduction of recurrent cardiac events. For many patients with obstructive coronary artery disease in need of revascularisation, surgical or percutaneous intervention is indicated on the basis of clinical and anatomical reasons or personal preferences. Medical therapy is a crucial accompaniment to coronary revascularisation, and data suggest that, in some subsets of patients, medical therapy alone might achieve similar results to coronary revascularisation. Most revascularisation data are based on prevalently White, non-elderly, male populations in high-income countries; robust data in women, older adults, and racial and other minorities, and from low-income and middle-income countries, are urgently needed.

Angiographic Outcomes for Arterial and Venous Conduits Used in CABG

Coronary artery bypass grafting is the most commonly performed cardiac surgical procedure. Conduit selection is crucial to achieving early optimal outcomes, with graft patency being likely the main driver to long-term survival. We present a review of current evidence on the patency of arterial and venous bypass conduits and of differences in angiographic outcomes.

Cardiac surgery risk prediction using ensemble machine learning to incorporate legacy risk scores: A benchmarking study

Objective

The introduction of new clinical risk scores (e.g. European System for Cardiac Operative Risk Evaluation (EuroSCORE) II) superseding original scores (e.g. EuroSCORE I) with different variable sets typically result in disparate datasets due to high levels of missingness for new score variables prior to time of adoption. Little is known about the use of ensemble learning to incorporate disparate data from legacy scores. We tested the hypothesised that Homogenenous and Heterogeneous Machine Learning (ML) ensembles will have better performance than ensembles of Dynamic Model Averaging (DMA) for combining knowledge from EuroSCORE I legacy data with EuroSCORE II data to predict cardiac surgery risk.

Methods

Using the National Adult Cardiac Surgery Audit dataset, we trained 12 different base learner models, based on two different variable sets from either EuroSCORE I (LogES) or EuroScore II (ES II), partitioned by the time of score adoption (1996-2016 or 2012-2016) and evaluated on holdout set (2017-2019). These base learner models were ensembled using nine different combinations of six ML algorithms to produce homogeneous or heterogeneous ensembles. Performance was assessed using a consensus metric.

Results

Xgboost homogenous ensemble (HE) was the highest performing model (clinical effectiveness metric (CEM) 0.725) with area under the curve (AUC) (0.8327; 95% confidence interval (CI) 0.8323-0.8329) followed by Random Forest HE (CEM 0.723; AUC 0.8325; 95%CI 0.8320-0.8326). Across different heterogenous ensembles, significantly better performance was obtained by combining siloed datasets across time (CEM 0.720) than building ensembles of either 1996-2011 (t-test adjusted, p = 1.67×10-6) or 2012-2019 (t-test adjusted, p = 1.35×10-193) datasets alone.

Conclusions

Both homogenous and heterogenous ML ensembles performed significantly better than DMA ensemble of Bayesian Update models. Time-dependent ensemble combination of variables, having differing qualities according to time of score adoption, enabled previously siloed data to be combined, leading to increased power, clinical interpretability of variables and usage of data.