Improving resuscitation and Extracorporeal Membrane Oxygenation outcomes in critically ill pediatric cardiac patients: from big data, to bench, to bedside

Introduction: Despite an undeniable improvement in knowledge and care over time, resuscitation in cardiac patients remains one of the most relevant challenges for cardiologists and intensivists. Objectives: We aimed to provide insight into resuscitation and outcomes of critically ill pediatric cardiac patients, exploring different knowledge opportunities - from big data, to bench, to bedside. We performed 6 individual projects, aiming to define, predict, and treat resuscitation events and ultimately improve the associated outcome. Methods: We performed: 1. a systematic review and meta-analysis on the incidence, risk factors, and outcome of CA in pediatric cardiac critically ill patients; 2. a big data analysis to determine whether novel mathematically computed variables as shock index (SI), coronary perfusion pressure (CPP), and rate pressure product (RPP) may predict resuscitation events; 3. a retrospective analysis of ELSO Registry data on patients resuscitated with ECMO after failure to wean (FTW) from cardiopulmonary bypass (CPB); 4. a review of extracorporeal CPR (ECPR) events and their outcomes at our center (BCH), modeling prediction of severe functional impairment or death; 5. a propensity-weighted analysis to define the benefits of left atrial (LA) decompression in patients supported with ECMO; 6. a prospective Phase1 study for the safety evaluation of a new FDA-approved drug, the inhaled hydrogen (H2), which has shown potential in prevention/treatment of ischemia-reperfusion injury in animal models. Results: We have shown: 1. Among 126,087 critically ill cardiac patients, 5% (CI 4-7%) experienced CA, and21% (CI 15-28%) underwent ECPR. Overall, 35% of patients (CI 27-44%) did not reach ROSC, and 54% died before discharge (CI 47-62%); 2. 7% (296/4,161) of patients who underwent cardiac surgery had CPR/ECPR, need for ECMO/VAD, unplanned surgery, heart transplant, or death within 7 postoperative days. In a multivariable regression model adjusted for age, surgical complexity, inotropic and respiratory support, and organ dysfunction, SI>1.83 was significantly associated with the adverse outcome (OR 6.6 [CI 4.4-10.0]), and CPP>35mmHg was protective against the outcome (OR 0.5 [0.4-0.7]); 3. 55% of the 2,322 patients who FTW from CPB died before discharge. Non-cardiac congenital anomalies, comorbidities, pre-operative CA, pre-operative mechanical ventilation>24h, pre-operative bicarbonate administration, longer CPB time, complex surgical procedures, longer ECMO duration, and ECMO complications were all independently associated with in-hospital mortality. Age>26 days (OR 0.56 [CI 0.42-0.75]) reduced the odds of mortality; 4. 52% of the 182 patients who underwent ECPR at BCH died before discharge. The median Functional Status Scale (FSS) among survivors was 8 (IQR 6-8), and only one survivor had severe functional impairment. Predictive models identified FSS at admission, single ventricle physiology, ECMO duration, mean PELOD-2, and worst mASPECTS as independent predictors of severe functional outcome at discharge (AUC=0.931) and at 6 months (AUC=0.924); 5. 18% of the 1,508 cardiac patients with biventricular physiology supported with ECMO underwent LA decompression (LA+). Covariates were well-balanced after propensity-weighting. In-hospital adverse outcome rate was 47% in LA+ vs 51% in LA-. Propensity-weighted multivariable logistic regression showed LA decompression to be protective for in-hospital adverse outcome (OR 0.77 [CI 0.64-0.93]); 6. H2 inhalation is safe in adult healthy volunteers, with no significant adverse events. This lays the foundation of a future trial for the use of H2 for the prevention/treatment of ischemia-reperfusion injury. Conclusion: We have provided new insight into resuscitation and outcomes of critically ill cardiac patients, from big data, to bench, to bedside. Future steps will include a randomized trial on the use of H2 to improve neurologic outcomes in cardiac ECPR patients.