Acute Respiratory Compromise Leading to Cardiac Arrest*
Observational studies using a large database are becoming more common as part of the medical literature, including within pediatrics, typically a field bereft of large volumes of patients with which to study. The advantages are inherent to their size, allowing us to study a large population of rare events or unusual diseases to determine significant factors relating to outcome. Advantages of these types of studies also include improved data access with query tools if the registry is well built, allowing for the data to be stratified into subgroups. This is countered, however, in that these databases are often complex, difficult, and reliant on training of programmers as well as users leading to significant heterogeneity of patient populations and missing data.
The specific database used here, Get with the Guidelines Registry, is the American Heart Association’s collaborative quality improvement program that collects data from 151 hospitals across the United States, including adult and pediatric hospitals as well as nonteaching and teaching hospitals. Current modules include acute respiratory compromise events, the focus of this study, as well as cardiopulmonary arrest events, medical emergency team events, and postcardiac arrest care events. In this study, acute respiratory compromise events were defined as absent, agonal, or inadequate respirations requiring emergency-assisted ventilation including noninvasive measures such as bag-valve-mask ventilation and pressure support and also eliciting a hospital-wide response (5). The study included events on the wards, emergency department, and PICU, as well as “other” locations making the study particularly heterogeneous in regards to acuity of illness, comorbidities, monitoring, emergency response time, and personnel skills. In addition, the statistical analysis relies on use of imputed data due to missing data from some of these various settings (6–8).
This study used the registry with two objectives: associated in-hospital mortality and cardiac arrest secondary to acute respiratory compromise. In total, the study looked at 1,952 patients, with 28% of the events ending with unassisted return of spontaneous ventilation, 88% requiring bag-valve-mask ventilation, and 62% requiring intubation. The results were not unexpected; a medical cardiac illness category, pre-existing hypotension, and septicemia were associated with increased mortality, consistent with previous studies (1). In addition, impending sepsis was likely underreported. Neonates and patients with an acute nonstroke neurologic event had decreased odds of cardiac arrest during the event, likely due to differences in etiology. This demonstrates issues raised earlier regarding the heterogeneity of the patient population. In regards to in-hospital mortality, overall 14.6% died before hospital discharge. Cardiac arrest occurred in 9.3% of total events. For these patients, the in-hospital mortality was 46.2% compared with 11.3% in those without an acute respiratory compromise-associated cardiac arrest. Given that the study was an observational study using a data registry, besides impending sepsis, many other factors could not be accounted for, leaving the reader to speculate about the indication for drugs, especially sedatives, causes of neurologic events, and if there were any prolonged emergency department stays possibly contributing to the frequency of cardiac arrest. In addition, the study did not compare hospitals by year of occurrence, which could lead to a bias.
In conclusion, the study does contribute to the literature, revealing intuitive findings regarding the association of comorbidities and acute respiratory in-hospital events.