Outcomes intensive care unit placement following pediatric adenotonsillectomy

Adenotonsillectomy (AT) is the most common surgical procedure for the treatment of sleep related breathing issues in children. While overnight observation in the hospital setting is utilized frequently in children after a AT, ICU setting is commonly used for patients with sleep apnea. This objective of this study is to examine factors associated with the preoperative decision to admit patients to PICU following AT as well as co-morbidities that may justify necessity for higher level of care.

This is a retrospective chart review from the years of 2009-2016. All patients who underwent AT for known sleep-related breathing issues at Nationwide Children's Hospital were eligible for inclusion. A complication was defined as an adverse event such as pulmonary edema, re-intubation, or a bleeding event. Respiratory support was defined as utilizing supplementary oxygen for more than one day, positive pressure ventilation, or intubation. Proportions and medians were used to describe the overall rate of complications/complexities in care, and bivariate statistics were used to evaluate the relationship between patient characteristics and outcomes. Similar methods were used to evaluate factors associated with preoperative referral to the PICU.

There were 180 patients admitted to hospital in non-ICU setting and 158 patients with a planned PICU stay. The patients with planned PICU stays had higher rates of technological dependence (13% vs. 3%; p = 0.0006), perioperative sleep studies (80% vs. 29%; p < 0.0001), and more severe classifications of OSA (p < 0.0001). Patients with planned ICU placement also had higher rates of apneas, hypopneas, respiratory disturbance indexes, apnea hypopnea indexes, lower oxygen saturation nadirs, and a longer time spent below 90% oxygenation in sleep studies (p < 0.0001). Nearly 45% of the patients with planned ICU stays required respiratory support compared to just 8% of non-PICU patients. Additionally, 32% of the patients with planned ICU stays experienced complications compared to just 8% of the floor population. Complications were associated with younger ages, gastrointestinal comorbidities, technological dependence, viral infections, and a history of reflux. Interestingly, there were no differences in the complication rate by sleep studies findings. Similarly, there were no population level differences between patients who required respiratory support in the ICU and those that did not. Unplanned PICU placement was a rare but significant adverse event (n = 24). None of the hypothesized risk factors were associated with unplanned PICU placement.

This study suggest that while our pre-operative referral program for PICU placement is effective in identifying patients needing higher levels of care, the program places many patients in the PICU who did not utilize respiratory support or suffer from complications. We observed some misalignment between characteristics associated with planned ICU stays and actual complications. This suggests that patients with specific clinical histories, not findings on their sleep studies, should be prepared to receive higher levels of care.