Liang C, Ruiz AG, Jensen EL, Friedman NR. Indications, Clinical Course, and Postoperative Outcomes of Urgent Adenotonsillectomy in Children. JAMA Otolaryngol Head Neck Surg. 2015;141(3):236-244. doi:10.1001/jamaoto.2014.3341
Children undergoing urgent adenotonsillectomy have been poorly described in literature.
To characterize the clinical course and outcomes of patients who underwent urgent adenotonsillectomy.
Design, Setting, and Participants
A 7-year retrospective medical record review of patients undergoing urgent adenotonsillectomy (study group) at a tertiary care pediatric hospital was undertaken. Comparisons were made between the study group and the control group consisting of children undergoing adenotonsillectomy following diagnostic polysomnography.
Retrospective medical record review with no study interventions.
Main Outcomes and Measures
Demographics, hospital course, and clinical outcomes.
A total of 35 patients (21 boys [60%] and 14 girls [40%]; mean age, 3.8 years) were identified as having undergone urgent adenotonsillectomy defined as severe obstructive sleep apnea with associated hypoxemia unresponsive to oxygen. The control group included 301 patients who received a diagnostic polysomnogram prior to nonurgent adenotonsillectomy. Patients undergoing urgent adenotonsillectomy patients were more likely to be younger than 3 years (54%) than nonurgent patients (P < .001) and were characterized by elevated obstructive apnea-hypopnea indices (average, 39.4 events per hour). Persistent desaturation below 80% despite at least 0.5 L of supplemental oxygen was the most common admission indication (83%). Obesity was the most frequent comorbidity (9 patients [31%]). Two patients (6%) experienced a major postoperative complication requiring intervention. Fifteen patients (43%) were discharged with supplemental oxygen. Two patients (6%) were admitted to the hospital more than 72 hours after surgery.
Conclusions and Relevance
Patients requiring urgent adenotonsillectomy are marked by younger age, elevated obstructive apnea-hypopnea indices, and persistent desaturations below 80% unresponsive to supplemental oxygen. Following surgery, some children have a dramatic improvement in gas exchange and will no longer require supplemental oxygen.
Tonsillectomy without or with adenoidectomy (TA) remains one of the most frequently performed pediatric surgical procedures in the United States,.1,2 The primary indication has shifted from chronic throat infections to upper airway obstruction.3,4 Sleep-related breathing disorders consist of a spectrum of disorders marked by abnormal respiratory pattern during sleep that ranges from snoring to obstructive sleep apnea (OSA). Sleep-disordered breathing (SDB) is a clinical diagnosis made primarily based on history and physical examination.2,5 Polysomnography (PSG) is considered the gold standard for diagnosing OSA.6- 8 A consequence of obtaining a PSG is that some children will be identified who may have such severe gas exchange abnormalities that inpatient admission is prudent. Although no standardized criteria have been established, we define the need for urgent TA as severe OSA with associated hypoxemia unresponsive to oxygen.
There is a paucity of literature that characterizes the clinical course of patients who underwent urgent TA. Brouillette et al9 used overnight oximetry studies to determine the urgency of a tonsillectomy procedure. Children with 3 or more clusters of desaturation events below 80% were considered to have severe OSA and were recommended to undergo urgent TA within 48 hours of the oximetry study. A follow-up study has recommended urgent TA for patients with oxygen saturation (Sao2) nadir of below 80% and has focused largely on the risk factors of respiratory complications in children who underwent urgent TA.10 However, other clinical variables associated with an urgent TA, such as severity of OSA, comorbidities, and perioperative oxygen requirement, have not been extensively examined. This study seeks to characterize the preoperative evaluations, clinical course, and postoperative outcomes of pediatric patients undergoing an urgent TA.
All patients who underwent TA at Children’s Hospital Colorado (CHCO) over a 7-year period (December 1, 2003, to November 30, 2010) were identified using billing records for TA following institutional review board approval. Two groups of patients were identified, and a retrospective medical record review was performed. The group requiring urgent TA (urgent TA group) comprised patients who were admitted following their PSG with severe OSA and associated hypoxemia that was unresponsive to oxygen. The control group comprised a 1-year cohort of consecutive patients who underwent an elective TA following a PSG from July 1, 2009, to June 30, 2010.11
Data collected from the medical records for the urgent TA group included sex, race/ethnicity, age, diagnostic PSG results, indication for surgery, comorbidities, tonsil and adenoid size, major postoperative complications, length of postoperative hospital admission, and oxygen requirement on discharge. The tonsils and adenoid were graded 1 through 4 using established scales.12,13 Diagnostic PSG results consisted of obstructive apnea-hypopnea index (OAHI), average Sao2 level, Sao2 nadir, and percentage of total sleep time with Sao2 less than 90%. Data collected for the control group included demographics, preoperative clinical diagnosis, and comorbidities.
Indication for urgent TA consists of severe OSA, defined as severe hypoxemia unresponsive to oxygen. All of the patients had at least 1 of the following findings on oxygen: loss of respiratory drive, increased carbon dioxide level, worsening OSA, or persistent desaturation below 80% despite an oxygen concentration of at least 0.5 L. Comorbidities included obesity (body mass index–for-age percentile > 95% in children14), Down syndrome, congenital heart disease, cerebral palsy, asthma, other chromosomal abnormality, pulmonary hypertension, seizures, and prematurity.
Primary outcome measures included postoperative complications and oxygen need on discharge. Major complications included (1) airway obstruction that required jaw thrust, positive pressure ventilation, or reintubation; (2) respiratory depression requiring administration of naloxone; (3) admission to the pediatric intensive care unit; (4) postobstructive pulmonary edema; (5) need for high-flow oxygen; or (6) primary hemorrhage. Oxygen need of the patient on discharge is determined by whether the patient was sent home while receiving supplemental oxygen or room air.
Study data were collected and managed using Research Electronic Data Capture (REDCap) tools hosted at the University of Colorado Denver. REDCap is a Health Insurance Portability and Accountability Act–compliant, secure, web-based application designed to support data capture for research studies.15 Descriptive statistics were used to characterize the demographic variables and clinical course of patients requiring urgent TA. Age and sex distributions between the 1-year control group and the 7-year total tonsillectomy population were compared to verify that the control group was an adequate representation of the total tonsillectomy patient population at our institution.
Demographics and comorbidities were compared between the urgent TA cohort and the 1-year control group using χ2 tests to determine potential predictors for the need of urgent tonsillectomy on a patient after the diagnostic PSG. For this analysis, age was dichotomized into 2 categories: younger than 3 years and 3 years or older. All statistical comparisons were performed using SAS statistical software (version 9.2; SAS Institute Inc).
Table 1 displays the demographic characteristics. Of the 9023 patients who received TA at CHCO over a 7-year observation period, 35 patients (0.4%) were identified as having undergone urgent TA. All children receiving urgent TA were admitted following their PSG. Their ages ranged from 1 to 9 years (mean [SD] age, 3.8 [2.1] years). There were 21 boys (60%) and 14 girls (40%). Nineteen patients (54%) were younger than 3 years.
A total of 1527 patients who received TA over a 1-year period were screened for inclusion in the control group. The control group was determined to be an adequate representation of the total population of patients who had undergone tonsillectomy at CHCO based on the close resemblance of the age and sex distributions of both the 1-year control group and the 7-year patient population. Within the 1-year cohort, 301 patients (19.7%) were identified as having received a diagnostic PSG prior to receiving nonurgent TA and included in the control group. The control group consisted of 170 boys (56.5%) and 131 girls (43.5%) with a mean (SD) age of 6.6 (4.2) years. Sixty-five patients (21.6%) were younger than 3 years.
Statistically significant differences in age were found between urgent TA group and the control group. Those requiring urgent TA were, on average, younger (3.8 ± 2.1 years) and more likely to be younger than 3 years (54%) than the control group (average age, 6.6 ± 4.2 years; 22% were <3 years) (P < .001). No significant difference was found between sex (P = .69) and ethnicities (P = .89) (Table 1).
Complete diagnostic PSG results were available for only 31 patients. The OAHI was elevated at a mean (SD) number of 39.4 (23.8) events per hour. Two outliers with OAHI of 106.8 and 131.8 were found. Both of these patients had a short room air portion of the study owing to severe hypoxemia. Thus, an accurate OAHI was difficult to determine. Four patients with pulmonary hypertension received at least at least 0.5 L of oxygen prior to PSG. These patients were not included in the oxygen saturation calculation. All of these patients were admitted owing to either increased oxygen requirement and/or rising oxygen values on oxygen. The mean (SD) asleep Sao2 level was low in the cohort at 89.5% (4.5%). The average awake Sao2 level in the cohort was 93.4% (3.6) (Table 2). The most common admission indication was persistent desaturations below 80% despite receiving at least 0.5 L of oxygen (29 of 35 patients [83%]), followed by an increased carbon dioxide level after receiving supplemental oxygen (6 of 35 [17%]). One patient (3%) experienced worsening OSA after receiving oxygen, and 1 patient (3%) lost their respiratory drive (Table 2).
In our cohort, the patients were not having isolated desaturations. All but 1 patient spent 4.0% of the total sleep time with Sao2 level less than 90%. The 1 child had who had an Sao2 level lower than 90% for just 1.7% of the total sleep was having prolonged obstructive events with an average duration of 17 seconds. Some of the events were greater than 30 seconds. Two of the 35 patients (6%) were admitted for TA despite an OAHI of less than 10 events per hour (7.1 and 7.5 events per hour, respectively) owing to persistent desaturations below 80% despite receiving 0.5 L of oxygen.
Eighteen of the 35 patients (51%) had 1 or more associated comorbidities. Obesity (9 patients [31%]) was the most common comorbidity. Other comorbidities included Down syndrome (4 patients [11%]), congenital heart disease (3 [9%]), asthma (4 [11%]), other chromosomal abnormalities (4 [11%]), pulmonary hypertension (4 [11%]), seizures (3 [9%]), cerebral palsy (2 [6%]), and prematurity (2 [6%]) (Table 3).
χ2 Analysis showed no statistically significant differences in comorbidities between the urgent and nonurgent TA groups (Table 3). As a result, no univariate or multivariate logistic regressions that aim to further explore potential difference in comorbidities between the urgent and nonurgent groups were undertaken.
The average time to surgery was 0.5 days (range, 0-3.0 days), with only 2 children waiting more than 48 hours (patients 17 and 18). The delay in surgery was secondary to requiring additional medical clearance as well as surgeon and/or operating room availability. Thirty of 35 patients and 28 of 35 patients had tonsil and adenoid sizes recorded, respectively (Table 4). Size 3+ to 4+ tonsils occurred in most patients (28 of 30), and size 3 and 4 adenoids were present in more than half of the patients (17 of 28). These patients were divided into 2 groups: those with at least 1 comorbidity and those with no comorbidities. Notably, patients with no comorbidities had a higher proportion of size 4+ tonsils (11 of 15) than patients with 1 or more comorbidities, although the overall difference between the 2 groups were statistically significant (P = .07). Similarly, there were no significant differences in the adenoid sizes (P = .54). There were 2 children (patients 3 and 26) with smaller tonsils (≤2) of whom only 1 had associated comorbidities (patient 26: Down syndrome and seizures).
All patients continued as inpatients following surgery with a length of hospital stay ranging from 1 to 12 days and a mean (SD) of 2.1 (1.9) days. Two of the 35 patients experienced a major postoperative complication. Patient 26, a 2-year-old girl, developed respiratory distress with increased airway swelling. Within 8 hours after the surgery, she was transferred to the pediatric intensive care unit and intubated for 2 days. Patient 28, a 9-year-old girl, had a primary hemorrhage that was controlled by cauterization in the operating room.
Most of the patients (27 of 35 [77%]) were discharged within 48 hours. Six patients (17%) were discharged within 72 hours. Of the 2 patients who were admitted more than 72 hours postoperatively, patient 26 was admitted for 12 days after experiencing worsening airway obstruction postoperatively. Patient 2 had an asthma exacerbation and fevers postoperatively and remained hospitalized for 4 days.
The oxygen need for patients on discharge is shown in Table 5. Twenty of 35 patients (57%) did not require oxygen on discharge. Of the 15 patients who were discharged with supplemental oxygen, 8 were younger than 3 years and 9 had comorbidities. Down syndrome (3 of 4) and obese (7 of 9) patients were the mostly likely to have oxygen requirement on discharge. Follow-up data for oxygen use were available for only 12 patients, of whom 8 had stopped receiving oxygen within 3 months. Three other patients were still receiving oxygen for more than 1 year and another for 4 months.
Postoperative PSG to evaluate efficacy of surgery was requested for 28 of the 35 patients (80%) needing urgent TA during postoperative follow-ups. However only 9 of 28 postoperative PSGs were completed. Three of these patients had persistence of obstructive sleep symptoms, with OAHI greater than 5 events per hour (patients 17, 21, and 25). All 3 patients had multiple comorbidities and were sent home with supplemental oxygen on discharge after TA. Table 6 and Table 7 provide detailed supplemental information on the comorbidities, PSG findings, and postoperative course.
Sleep-related breathing disorders have emerged as the primary indication for TA in the past 35 years.4 They are a spectrum of sleep disorders ranging from primary snoring to OSA. Clinical guidelines from the American Academy of Pediatrics, American Academy of Sleep Medicine, and the American Academy of Otolaryngology (AAO) all advocate for a preoperative PSG.5,6,16 The AAO recommendations for preoperative is more selective than the other 2 societies. The AAO recommends a PSG prior to intervention under the following circumstances: obesity, Down syndrome, craniofacial abnormalities, neuromuscular disorders, sickle cell disease, mucopolysaccharidoses, uncertain need for surgery, or discordance between tonsil size and severity of SDB.5 With the increased use of PSG, more children will likely require urgent interventions. Previously, these children in these circumstances did not have objective testing and were electively scheduled for TA.
To date, indications for urgent TA in the pediatric population have not been clearly defined. The McGill University Sleep Group used oximetry results to determine the urgency of a TA.9 Children with 3 or more clusters of desaturation events lower than 80% were considered to have severe OSA and were subsequently recommended to undergo urgent TA within 48 hours. A follow-up series in children undergoing urgent TA defined severe OSA as a Sao2 nadir of less than 80%.10
Our criteria was less inclusive and mirror our hospital admission criteria for bronchiolitis. Children with bronchiolitis and hypoxia (defined as <88% on room air) with an Sao2 level of less than 90% on a 0.5 L/min nasal cannula are observed in the hospital.17
While all of our patients had an Sao2 nadir of less than 80% during the room air portion of the PSG, the decision to admit them for urgent TA is primarily due to persistent desaturations below 80% while receiving at least 0.5 L of oxygen. Owing to the retrospective nature of the study, it is possible that for younger children the on-call physician may have had a lower threshold to emergently admit and allowed older children to be seen urgently as outpatients. The stricter admission criteria of our institution have limited the incidence of urgent TA procedures. Other children who have severe OSA but do not meet our admission criteria are evaluated by the otolaryngology service urgently through the outpatient clinic and typically have surgical intervention in less than 3 weeks.
The understanding of OSA as a chronic rather than acute condition suggests that a timely rather than an urgent approach may be appropriate. In our institution, the decision to admit or discharge a child from the sleep laboratory relies on gas exchange unresponsive to low-flow supplemental oxygen. We prioritize gas exchange over the number of respiratory events. Intuitively, one should prioritize the care of a child who has unstable gas exchange. Hypoxemia is associated with bradycardia, which may be life threatening. Also, one cannot assume that the current condition is chronic rather than acute. Despite protocols to minimize performing a PSG in an acutely ill child, nothing is 100% effective. All children are admitted to the medical service for a comprehensive evaluation. The burden of proof is to rule out an acute medical condition and then perform timely intervention for the chronic condition. Of note, the policy of our sleep laboratory is delay a PSG if a child is acutely ill. Every procedure is confirmed with the child’s family 48 hours prior to the study, and acutely ill children are rescheduled. If a child presents to the sleep laboratory and is not in his or her usual state of health, the sleep technician will cancel the study.
Several unique demographic and clinical characteristics were identified in the urgent TA group. Not only was that cohort significantly younger on average than the control group (3.8 ± 2.1 years vs 6.6 ± 4.2 years, respectively), the age distribution in our series also differs significantly, with 54% of our patients being younger than 3 years compared with only 22% in the control group. This diverges significantly from the findings of Brown et al,10 in which patients requiring urgent TA were not statistically different in age than the control group (4.0 ± 2.4 years vs 3.5 ± 1.9 years, respectively) and only 36.0% of patients were younger than 3 years. Our data can only be generalized to children who have PSGs. Invariably, there are some children who did not have a PSG who had severe gas exchange abnormalities. We are discussing the treatment for those who underwent PSG.
Comparison of comorbidities between our cohort and the series reported by Brown et al10 also showed considerable differences. Table 3 demonstrates there was not an increased number of pulmonary comorbidities in our cohort. Almost half of our patients requiring urgent TA (17 of 35 [49%]) had at least 1 comorbidity, but no statistically significant differences in comorbidities were found between the urgent TA group and the control group. While the series reported by Brown et al10 had a similar proportion of patients requiring urgent TA with at least 1 comorbidity (45.5%), a greater proportion of their urgent cases had comorbidities compared with their control group (45.5% vs 13.6%; P = .01). Notably, asthma was a frequently associated comorbidity in both our series (11%) and the series reported by Brown et al10 (18.5%). In contrast to our cohort, in which only 1 child required reintubation, 6 patients (11.1%) in the series reported by Brown et al10 required reintubtation.10
Because there was no difference in the comorbidities between the urgent TA group and the control group, one may assume that those children had subclinical findings that made them more prone to developing hypoxemia. These subclinical findings may include lower tone, alveolar simplification, more nasal obstruction, and lower baseline partial oxygen pressure that is still high enough to be on the flat portion of the oxygen dissociation curve.
Besides comorbidities, one would expect that adenotonsillar hypertrophy plays an important role in the severity of OSA. Our findings, which showed a predominance of size 3+ and 4+ tonsils as well as size 3 and 4 adenoids in the urgent TA group, is in accordance with the observation by Friedman et al18 that tonsil hypertrophy remains a principle cause of SDB. These findings are intuitive. Since anatomy and neuromuscular tone both influence the size of the upper airway, one would expect that hypotonic children would be more likely to have OSA regardless of adenotonsillar size. Our results demonstrated no statistically significant differences between comorbidities and the size of tonsils and adenoids.
Although TA is considered a relatively safe procedure with low incidence of postoperative complications, one would assume that a child with severe gas exchange abnormalities preoperatively would have a more complicated postoperative course. Risk factors of postoperative complications for TA include craniofacial disorders, Down syndrome, cerebral palsy, major heart disease, or bleeding diatheses. Obesity and children younger than 3 years with PSG-proven OSA have also been shown to have higher risk of postoperative complications,19- 24 with airway compromise being the most common complication.25- 27 Despite the necessity of an urgent TA, patients in our study had a low incidence of postoperative complications (2 of 35 [6%]).
The postoperative course of our cohort is favorable. Most of the patients (27 of 35) were discharged within 48 hours. More than half of the patients (20 of 35 [57%]) did not require oxygen on discharge. Fifteen patients (15 of 35 [43%]) were discharged with supplemental oxygen. Those patients discharged with oxygen either were younger than 3 years or had comorbidities, suggesting that older healthy children have a dramatic improvement in their gas exchange following an urgent TA. Given the limited number of patients in our study, a study of a larger series of patients needing urgent TA is necessary to confirm our observations.
The low follow-up rate (26%) for postoperative PSGs may reflect both the evolution of follow-up criteria at our institution and reluctance of parents. All of these children qualified for postoperative PSG owing to their OSA severity. Recent studies have shown that the likelihood of cure is less for children with more severe OSA.28,29 In our institution, we have modified the electronic tonsillectomy order sets so that postoperative PSG is an option. Compliance with medical recommendations will always be an issue. Hopefully, effective communication with the parents and prompt scheduling will improve follow-up care.
One would expect that as more PSGs are performed, more children will be identified as having severe OSA with gas exchange abnormalities for whom it would be prudent to have expeditious care. Urgent TA is an effective and reasonable intervention after obtaining medical clearance. Our group of patients was characterized by an elevated OAHI and persistent desaturations below 80% despite supplemental oxygen. Younger age rather than associated comorbidities was associated with severe gas exchange abnormalities unresponsive to low flow oxygen that resulted in hospital admission. The postoperative outcome following an urgent TA is favorable, with most children being discharged home on room air less than 48 hours after surgery.
Corresponding Author: Norman R. Friedman, MD, DABSM, Department of Pediatric Otolaryngology, University of Colorado School of Medicine, 13123 E 16th Ave, B455, Aurora, CO 80045 (email@example.com).
Submitted for Publication: July 16, 2014; final revision received October 14, 2014; accepted November 12, 2014.
Published Online: January 2, 2015. doi:10.1001/jamaoto.2014.3341.
Author Contributions: Dr Friedman had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Liang, Friedman.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: Ruiz, Jensen, Friedman.
Statistical analysis: Liang, Ruiz, Jensen.
Administrative, technical, or material support: Liang, Ruiz, Jensen.
Study supervision: Friedman.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported in part by National Institutes of Health (NIH)/National Center For Research Resources, Colorado, Clinical and Translational Sciences Institute grant No. UL1 RR025780.
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The contents are the authors’ sole responsibility and do not necessarily represent the official view of the NIH.