CARDIOPULMONARY BYPASS RE-INSTITUTION AND INTRAOPERATIVE REVISIONS ARE ASSOCIATED WITH INCREASED PERIOPERATIVE RISK IN CONGENITAL CARDIAC SURGERY

We aimed to examine the rationale and assess risk of additional cardiopulmonary bypass (CPB) episodes, and specifically, re-institution of CPB for reintervention, in congenital heart surgery.

We retrospectively reviewed all children undergoing cardiac surgery with CPB at one institution (2014-2016). Tetralogy of Fallot was excluded as the annulus-sparing surgical strategy involves planned intraoperative revision. The final cohort of 922 patients was stratified: single CPB run (CPB=1, n=771); >1CPB without revisions (CPB>1NR, n=77) for reperfusion, bleeding, or intractable arrhythmias; and >1CPB with revision (CPB>1R, n=74). Successful revisions were defined as: those resulting in at most mild residual lesions, did not induce new lesions, did not require postoperative revision, and did not result in mortality prior to discharge. Forced revisions were defined as revisions for unacceptable hemodynamics or failure to wean from CPB. Outcomes of interest were in-hospital mortality, complications, need for postoperative revision, and a composite (death, stroke, ECMO, arrest, and leaving OR with open chest). Predictors of the composite outcome were assessed by logistic regression. Mean age was 2.9±4.5 years, 54% were male and mean RACHS was 2.5±0.8. In CPB>1R, 18 (24%) required 2 or 3 revisions. Revisions were valvular in 34 (46%) and forced in 51 (69%) patients. Success with a single revision occurred in 22 (30%) and by the final attempt in 32 (43%) patients. One patient's lesion worsened as a result of the revision. When comparing CPB=1 vs CPB>1NR vs CPB>1R, outcomes with significant differences included: stroke, arrest, postoperative revision, open chest upon leaving OR, and the composite outcome (Table 1). In-hospital death, re-exploration, need for ECMO, and hospital stay were similar (Table 1). Among CPB>1R patients, 12 (16%) required postoperative revision. Logistic regression identified revision as a predictor of the composite outcome with greatest effect seen in 2 revisions: 1: OR=2.39 (95% CI: 1.35, 4.21) p=0.002, 2: OR=3.68 (95% CI: 1.26, 10.74) p=0.02, 3: OR=1.23 (95% CI: 0.25, 12.23) p=0.8. Additionally, CPB runs=3 was associated with the composite outcome: OR=2.25 (95% CI: 1.00, 5.05) p=0.048. A trend towards CPB runs = 4 was seen while CPB runs=2 was not correlated with the composite outcome.

Revisions are uncommon, and most are forced. Revisions frequently do not eliminate residual lesions, occasionally require additional postoperative revisions, but are not associated with increased early mortality. Revisions and >2 CPB runs were associated with greater risk of the composite outcome. Those for non-forced reasons should likely be considered with caution.