Evidence and use of less invasive coronary bypass surgery: A technology assessment and evaluation using national data

The problem. There is growing enthusiasm for less invasive coronary artery bypass grafting (CABG) performed off-pump (without cardiopulmonary-bypass). We determined the association of off-pump surgery with occurrence of adverse outcomes. In addition, we evaluated the use of less invasive CABG in the US.;Methods. Evidence: We searched the MEDLINE, EMBASE and the Cochrane Register 1980-2007(December). We also searched the reference lists of randomized clinical trials (RCT) and reviews. Study selection: RCTs comparing off-pump surgery to on-pump CABG. Data extraction: We independently searched for studies, read abstracts and abstracted all data. Data synthesis: Risk estimates were obtained using fixed or random effects meta-analyses. Evaluation of use: On- and off-pump CABG were identified in HCUP, NIS database including 250 hospitals. Procedures and diagnoses were based on ICD-9-CM coding. A multivariable model was constructed including variables that are clinically and statistically significant (GLIMMIX procedure in SAS).;Results. Evidence: There were 5,040 patients enrolled in 48 RCTs (mean age 62, 22% female). No study reported information on race. Off-pump CABG was associated with 55% reduction in stroke, 31% in atrial fibrillation and 48% in wound infection. This translated into avoidance of 12, 80 and 40 events per 1000 CABG accordingly. However, there were fewer grafts and more re-interventions (20-per-1000) with off-pump CABG.;Evaluation of use: There were 63,435 patients in 235 hospitals (mean age 65, 31% female). Off-pump use was 24.8% and variability range was 0% to 95%. Cluster level variables such as volume, size and hospital ownership were the strongest predictors of use. Large volume hospitals had 48% lower odds while private, investor owned hospitals had 2.66 times higher odds of using off-pump. Known characteristics explained very small portion of variation among hospitals.;Conclusions. Off-pump CABG is associated with reduced risk of stroke, atrial fibrillation and infections as compared to on-pump CABG. Evidence should be generalized taking into account RCT limitations, learning curve, propensity to perform fewer grafts and likely re-interventions following off-pump surgery. The variation in care among hospitals is not explained by patient characteristics and is only slightly explained by hospital characteristics. This needs careful reflection and policy level assessment/intervention.