Risk Factors for Peri-Intubation Cardiac Arrest in Pediatric Cardiac Intensive Care Patients: A Multicenter Study

Risk Factors for Peri-Intubation Cardiac Arrest in Pediatric Cardiac Intensive Care Patients: A Multicenter Study.

Esangbedo ID, Byrnes J, Brandewie K, Ebraheem M, Yu P, Zhang S, Raymond T.Pediatr Crit Care Med. 2020 Dec;21(12):e1126-e1133. doi: 10.1097/PCC.0000000000002472.PMID: 32740187


Take-Home Points:

  • The frequency of peri-intubation cardiac arrest (PICA) in the cardiac intensive care unit (ICU) is high among critically ill children with congenital and acquired heart disease.
  • In the cohort of children with cardiac disease, moderate to severe systolic dysfunction of the systemic ventricle, pre-intubation hypotension, lactic acidosis > 10 mmol/L, and pH< 7.0 are found to be risk factors associated with PICA.


Commentary from Venu Amula MD, Pediatric Cardiac Intensivist at Primary Children’s Hospital/University of Utah SLC, Utah:  The transition from spontaneous to positive pressure ventilation during endotracheal intubation is challenging for critically ill infants and children. The acute change in intrathoracic pressure and ventricular loading conditions, coupled with the pharmacological effects of medications used during intubation, can result in adverse hemodynamic consequences. Patients with low hemodynamic reserve owing to cardiac disease are more vulnerable, given that they sometimes operate at the extremes of physiologic compensation.

Peri-intubation cardiac arrest (PICA), defined as cardiac arrest requiring chest compressions for > 1 min and occurring during and within 30 minutes of the procedure, has adverse outcomes in the adult population with high immediate and 28-day mortality. Pre-intubation hypoxia and hypotension are identified as risk factors associated with PICA in adults. In children, studies using National Emergency Airway Registry for Children (NEAR4Kids) show a PICA rate of 1.7% among all pediatric intensive care patients with higher rates in those with a cardiac diagnosis than without (2.8% vs. 1.8%). Studies evaluating intubations in dedicated cardiac ICUs are lacking.

The authors of this multicenter retrospective cohort study aimed to evaluate the characteristics and frequency of PICA in critically ill children with cardiac disease cared for in three specialized cardiac ICUs. They also sought to identify risk factors associated with PICA. The primary outcome of interest is the frequency of PICA. Pediatric patients (0-18 yrs.) with congenital and acquired heart disease undergoing endotracheal intubation between January 2015 and December 2017 were retrospectively reviewed in this study. Patient, provider, and procedural characteristics of intubation events that resulted in cardiac arrest were compared with those that did not (comparison groups) using the two-sample t-test and chi-square test. Patients were excluded for the following reasons: intubation outside the cardiac ICU, prior cardiac arrest, intubation as part of tube exchange, current ECMO requirement, elective intubation performed by an anesthesiologist, and incomplete data.

A total of 186 intubation events were identified that occurred in 151 pts over the study period. The median age of the patients was 3.1 months, and the median weight was 4.3 kg. Seventy-one of the 186 events occurred in those with associated non-cardiac morbidity, and 28 occurred in those with a genetic syndrome. The majority of the intubations were presurgical, while 40% occurred within one month of cardiac surgery. Immediate indications for intubation were (in the order of frequency): hypoxemia, work of breathing, hypercarbia, shock, and metabolic acidosis. The indications did not differ between the groups. Etiologies of arrest as recorded were: prolonged hypoxia, worsened acidosis, medications, and misplaced endotracheal tube.

The primary outcome of PICA occurred in 13/186 cases with a rate of 7%. This rate was 21% among the subgroup with moderate to severe systolic dysfunction of the systemic ventricle. The most common rhythms at arrest were bradycardia and pulseless electrical activity. The median duration of cardiopulmonary resuscitation was 10 min (IQR 4.5- 21 min). Among the 13 patients who had a cardiac arrest episode, 7 had a return of circulation without extracorporeal membrane oxygenation (ECMO), 3 had a return of circulation with ECMO, and 3 died.

Upon comparing arrest and the non-arrest events, pre-intubation hypotension (60 vs 18.4%, p=0.007), moderate to severe left ventricular dysfunction (23 vs 8.5%, p=0.048), pre-intubation lactate > 10 mmol/L (25 vs 6% ,p=0.018) and pH < 7.0 (23 vs 4.1 % , p=0.036) were significantly higher in the PICA group. Procedural and provider characteristics, including the number of attempts, video laryngoscopy, clinician performing the procedure, were not significantly different between the two groups. The type of medications used also did not differ.

What does this mean for us? 

Few studies from large registries show that cardiac arrest associated with intubation is more likely to occur in critically ill children with cardiac disease. These studies provide limited details of patient, provider, and procedural characteristics. The authors in the current study attempted to characterize PICA in children with acquired and congenital heart disease cared for in dedicated cardiac ICUs. They found a higher frequency of PICA in the study population when compared to published literature, and they identified patient characteristics significantly associated with the cardiac arrest.

While the study emphasizes the possibility of adverse events associated with intubation in children with cardiac disease, the conclusions need to be read cautiously. Though multicentric, the study is conducted only in three cardiac ICUs and thus lacks generalizability. As acknowledged by the authors, the other major limitation of the study is the small sample size. The retrospective and observational nature of this study lend to selection bias in both directions, and the usual strategies to adjust for confounding effects of variables is limited by the rare occurrence of the primary outcome event (cardiac arrest) in relation to the predictor variables, i.e., few events per predictor variable. It is difficult to reliably assess the risk of cardiac arrest with the study variables though statistical association exists in unadjusted models.

Regardless of its limitations, the study brings two points to light concerning intubation in children with cardiac disease. Both focus on anticipatory management.

  1. Early recognition of hemodynamic and respiratory perturbations (before the anaerobic threshold for oxygen delivery sets in) is critical in children with heart disease, particularly those with significant systolic dysfunction. A timely decision to intubate while some respiratory and hemodynamic reserve remains may mitigate adverse events.
  2. If a child in the cardiac ICU needs intubation and the multiple risk factors identified in this study are already established, thoughtful consideration of pharmacologic and personnel management is vital. The strategy may include calling for anesthesia backup to establish airway quickly and having the team prepared for advanced cardiopulmonary resuscitation if need arises.

The authors should be congratulated for bringing these focus points to the forefront, and future studies with different centers may be worthwhile.