Risk factors for prolonged mechanical ventilation following total cavopulmonary connection surgery: 8 years’ experience at Fuwai Hospital

Abstract Objective Prolonged mechanical ventilation (PMV) is closely associated with higher morbidity and mortality after total cavopulmonary (TCPC) surgery. We aim to identify the clinical risk factors for PMV. Design A retrospective case-control study. Setting Fuwai Hospital. Participants 504 patients with TCPC surgery from 2010 to 2018. Interventions None. Measurements and Main Results The definition of PMV was derived from the Cox regression model for predicting postoperative length of hospital stay. Lasso regression, logistic regression and Cox regression were applied to identify predictors for PMV. Patients with mechanical ventilation time > 9 hours were identified as having PMV. Independent predictors of PMV included: age, intraoperative maximum vasoactive-inotropic score (VIS), minimal temperature during CPB, postoperative prothrombin time, alkaline phosphatase and total bilirubin levels, and postoperative fluid balance. These predictors were also achieved in the Cox regression for predicting the duration of mechanical ventilation. Patients with PMV were associated with increased blood transfusions, more consumption of vasopressin and anti-pulmonary hypertension medication, higher incidence of re-intubation, more renal replacement treatment, longer ICU stay, higher hospitalization costs and more specialist visits. Conclusions Based on our experience, age at operation, maximal VIS and minimal temperature during CPB, postoperative prothrombin time, alkaline phosphatase and total bilirubin levels, and postoperative fluid balance were independent predictors of PMV. Adopting a comprehensive strategy of perioperative management that targets the identified risk factors might significantly lower the risk of PMV and improve in-hospital outcomes and patients with PMV might need more specialist visits.