Coronary artery by-pass surgery in patients on dialysis: in-hospital outcomes from UK registry analysis

OBJECTIVES Chronic kidney disease requiring dialysis significantly increases the risks of coronary artery disease. However, there is limited data on this high-risk patient population requiring coronary artery bypass grafting. Using a UK national registry, we investigated the impact of preoperative dialysis on in-hospital mortality and early morbidity in patients undergoing CABG. METHODS A retrospective analysis of National Adult Cardiac Surgery Audit data between 02/01/1996 and 31/03/2019 identified patients who underwent first-time isolated CABG. Propensity matching was performed to balance the baseline characteristics between dialysis and non-dialysis patients, yielding 633 matched pairs. We evaluated trends in CABG among dialysis patients and EuroSCORE 2 performance in predicting in-hospital mortality (calibration, discrimination, and clinical utility). RESULTS There was a steep increase in CABG operations in dialysis patients after 2011. EuroSCORE 2 showed poor calibration, discrimination, and minimal clinical benefit in predicting mortality in dialysis cases. Dialysis patients exhibited a significantly higher in-hospital mortality rate (7.9% vs 2.1%, p < 0.001) than non-dialysis patients. The dialysis patients had longer median hospital stays (12 vs 9 days, p < 0.001) and a higher rate of return to the theatre for bleeding (5.5% vs 2.7%, P = 0.034). We found no difference in postoperative neurological deficit rates between the two cohorts. The odds ratio of in-hospital mortality for the dialysis versus non-dialysis patients was 4.62, P < 0.001, 95% (CI: 2.54–8.4). Significant predictors of mortality in the dialysis CABG cohort included advanced age (OR: 2.48), NYHA class IV (OR: 3.06), and pulmonary hypertension (OR: 11.91). CONCLUSIONS There has been an overall increase in coronary artery bypass operations performed in renal dialysis-dependent patients in the UK. Preoperative chronic dialysis is associated with considerable in-hospital mortality, return to theatre for bleeding and prolonged hospital stay. EuroScore 2 has poor predictive performance in this patient cohort.