Jaryszak EM, Shah RK, Vanison CC, Lander L, Choi SS. Polysomnographic Variables Predictive of Adverse Respiratory Events After Pediatric Adenotonsillectomy. Arch Otolaryngol Head Neck Surg. 2011;137(1):15-18. doi:10.1001/archoto.2010.226
To determine polysomnographic (PSG) variables that may potentially predict adverse respiratory events after pediatric adenotonsillectomy.
Retrospective, case-control study.
Free-standing academic tertiary-care pediatric hospital.
The study included 1131 patients undergoing adenotonsillectomy by 2 attending surgeons. There were no exclusion criteria.
Main Outcome Measures
Variables from preoperative PSGs were analyzed to determine predictors of postoperative respiratory complications. Logistic regression analysis was performed.
A total of 151 patients (13.4%) underwent preoperative PSG. Twenty-three of these patients (15.2%) had adverse respiratory events. The primary adverse event was desaturation requiring supplemental oxygen therapy, with 1 case of postobstructive pulmonary edema. Patients with adverse events had a significantly higher apnea-hypopnea index) (31.8 vs 14.1; P = .001), higher hypopnea index (22.6 vs 8.9; P = .004), higher body mass index (z score, 1.43 vs 0.70; P = .02), and lower nadir oxygen saturation (72% vs 84%; P <.001). Patients with adverse events had a prolonged hospital course (odds ratio, 32.1; 95% confidence interval, 7.8-131.4). There were no differences in age or other PSG variables. There were no intubations or mortalities.
Polysomnography may be used to predict which patients are at higher risk for adverse respiratory events after adenotonsillectomy. Such knowledge is valuable for planning optimal postoperative management and intraoperative anesthesia. Predictors of increased respiratory complications include apnea-hypopnea index, hypopnea index, body mass index, and nadir oxygen saturation.
Pediatric adenotonsillectomy is a safe outpatient procedure1- 4; however, there is a subset of patients who do not meet the criteria for outpatient surgery. Specifically, the American Academy of Otolaryngology–Head and Neck Surgery issued guidelines for outpatient tonsillectomy in 19965: children should be healthy, with an American Society of Anesthesiologists classification I or II; have no evidence of obstructive sleep apnea-hypopnea syndrome (OSAHS); and be older than 3 years. These guidelines, however, have been challenged in the literature. In particular, the OSAHS criterion, which has become the most common indication for tonsillectomy,6 has been the focus of controversy.
In a systematic review of the literature, the effect of OSAHS on the rate of postoperative complications was reviewed.7 Initial analysis revealed a significant difference in the complication and unplanned admission rates in the OSAHS group; however, it was shown that this effect may have been confounded by the effect of age.7 In the subgroup of patients younger than 3 years, OSAHS was not an independent risk factor for complications or unplanned admissions. The study suggests that OSAHS may not be an independent modifier of operative risk, with the caveat that none of the source studies verified the presence or quantified the severity of OSAHS. Also, a primary study determined that OSAHS was not a criterion for a delayed discharge unless severe disease was suspected.2
The spectrum of respiratory difficulties after adenotonsillectomy varies widely. From desaturation caused by increased swelling and secretions to postobstructive pulmonary edema, respiratory complications can be quite severe and necessitate immediate, aggressive management, requiring further care in a pediatric intensive care unit. There are studies evaluating the utility of polysomnography (PSG) in predicting outcomes from adenotonsillectomy,8,9 but these studies do not evaluate the role of PSG in predicting which patients are at risk for adverse outcomes in the immediate postoperative period. There is evidence to suggest that younger patients,10,11 patients with multiple medical comorbidities,10- 12 patients with worse apnea-hypopnea indexes (AHIs),11,13 and patients with lower nadir oxygen saturations11,13,14 are at higher risk for respiratory complications. We hypothesized that PSG variables can be used to predict which patients are more likely to have a respiratory complication that will require treatment in the immediate postoperative period.
Institutional review board approval was obtained from the Children's National Medical Center, Washington, DC. A hospital database was searched via Current Procedural Terminology codes for patients who had undergone an adenotonsillectomy (codes 42820 and 42821) from July 2006 to December 2008. These patients were cross-referenced with patients who had undergone PSG during that same time frame. A total of 1131 patients met inclusion criteria, having undergone an adenotonsillectomy by 1 of 2 attending surgeons during that period. Of these 1131 patients, 151 underwent preoperative PSG and were included in our analysis. There were no rigid criteria for patients who were selected to undergo preoperative PSG, and similar percentages of patients from each attending surgeon underwent PSG. In general, children younger than 5 years, children with significant medical comorbidities, and children with significant obstructive sleep apnea are monitored overnight after adenotonsillectomy. In some instances, the patient underwent PSG before coming to our office.
A retrospective review of the medical records was performed to identify patients with postoperative respiratory complications. Data were of excellent quality, as all patients included were part of the electronic medical record. Furthermore, if outside forms were required for analysis, they had been scanned into the electronic medical record and were searchable. To be included in this study, patients had to have undergone PSG before surgery; therefore, no patients were missing data, so no patients were excluded for lack of data. Preoperative PSG data were collected for each patient. Patients were divided into 2 groups: those without postoperative respiratory complications and those with postoperative respiratory complications. Logistic regression analysis was performed to determine which variables differed between the 2 groups. Statistical significance was set at P ≤ .05.
Preoperative PSG was performed in 151 patients (13.4% of all patients undergoing adenotonsillectomy). The medical comorbidities of the patients who underwent PSG are listed in Table 1. Twenty-three of these patients (15.2% of patients with a PSG) had adverse respiratory events (Table 2). The primary adverse event was desaturation requiring supplemental oxygen therapy; there was 1 case of postobstructive pulmonary edema (0.7% of patients with a PSG). On PSG, patients with adverse events had a significantly higher AHI (31.8 vs 14.1; P = .001), higher hypopnea index (22.6 vs 8.9; P = .004), and lower nadir oxygen saturation (72% vs 84%; P <.001). Also, the complication group had a significantly higher body mass index (BMI) (z score, 1.43 vs 0.70; P = .02), with 47.8% having a BMI greater than the 95th percentile vs 29.7% in the noncomplication group. Age did not predict an increased risk of postoperative complications (5.8 years vs 5.8 years; P = .98). The remaining variables evaluated on PSG were not significantly different between the 2 groups (Table 3).
Overall, patients with adverse respiratory events (n = 23) spent an additional 22 days in the hospital beyond routine overnight observation for patients with OSAHS (mean [SD], 1.0 [0.3] days). Patients without adverse respiratory events (n = 128) spent an additional 4 days (0.03 [0.02] days) in the hospital beyond routine overnight observation (3 nights for poor oral intake and 1 night for planned admission for a 13-month-old patient). Patients with adverse events were at increased risk of having a prolonged hospital course beyond standard overnight observation (odds ratio, 32.1; 95% confidence interval, 7.8-131.4). There were no intubations or mortalities.
This study evaluated the potential for preoperative PSG variables to assist in predicting adverse respiratory events in the postoperative period after pediatric adenotonsillectomy. While it is intuitive that children with worse obstructive sleep apnea are more likely to have perioperative problems, the role of each of the components of PSG has not been previously evaluated, to our knowledge.
A series of 163 patients undergoing adenotonsillectomy who underwent preoperative PSG demonstrated that those with respiratory complications were younger and had associated medical conditions.15 Unfortunately, only 2 variables were analyzed: AHI and nadir oxygen saturation (a preoperative AHI >5 and a preoperative oxygen saturation nadir <80% increased postoperative respiratory complications).15 The study suggested that preoperative nocturnal oxygen saturation may be a useful tool with which to predict postoperative respiratory complications but did not fully evaluate the other PSG parameters. Similar to our rate of 15.2% for respiratory complications in the OSAHS cohort, the study reported a rate of 21%,15 which is significantly higher than the expected rate of 1.3% in patients without OSAHS.16
In addition to PSG parameters, patient characteristics can also affect outcomes after adenotonsillectomy for OSAHS. Even though age was previously shown to be a factor in the rate of complications after adenotonsillectomy,8,11,12 we did not see this effect in our analysis (5.8 years vs 5.8 years; P = .98). It may be that there was a selection bias, as patients were subjectively determined to undergo preoperative PSG based on the attending surgeons' practice patterns. While these practice patterns portend a cohort of sicker children receiving preoperative PSG, frequently in younger patients with obstructive symptoms and a history and physical examination findings that corroborate a diagnosis of OSAHS, a PSG will not be obtained. This approach is supported by a study that showed low complication rates after adenotonsillectomy, low postoperative morbidity, and significant costs for PSG.17 Only those patients who have a clinical picture that is not consistent with OSAHS, morbid obesity, failure to thrive, excessive daytime somnolence, or craniofacial and chromosomal abnormalities should undergo PSG. These conclusions are consistent with those reported by others.18
In our series, patients with respiratory complications had significantly higher BMI z scores. After adjustment for age and sex, 11 of 23 patients (47.8%) in the respiratory complication group were obese according to BMI criteria (>95th percentile). A series of 26 morbidly obese patients19 demonstrated a 46% rate of respiratory complications that required intervention. While most of these patients, as in our series, had oxygen desaturation that was alleviated with supplemental oxygen therapy, 2 patients required positive pressure therapy. The authors concluded that within this subgroup elective admission to the pediatric intensive care unit might not be necessary but that close monitoring and observation are required after adenotonsillectomy in obese patients.
While most of the adverse events in our series were oxygen desaturations that resolved with supplemental oxygen administration, 9 patients required admission and close monitoring, and 2 patients required bilevel positive airway pressure or oxygen via a nonrebreather mask. These rates of intervention are similar to those published in other studies.11,15 In our series, no patients required intubation, a complication that is also rare.11,13 Patients in the respiratory complication group spent a total of 22 additional days in the hospital, the economic and system effects of which can be significant, particularly if care is being administered in the pediatric intensive care unit.
The main limitation to our study is the inherent bias resulting from the small sample size: only 13.4% of patients received preoperative PSG. In general, sicker children with multiple medical comorbidities undergo PSG. Undoubtedly, there are patients who underwent adenotonsillectomy without PSG who had severe obstructive sleep apnea. Evidence suggests that we as clinicians are poor predictors of the severity of OSAHS based on history and physical examination findings alone.20 Therefore, there may be other patients who did not undergo PSG but who had postoperative respiratory events. This group of patients was not studied, a recognized limitation inherent in a retrospective review.
In conclusion, our study is the first (to our knowledge) to show the differences of each PSG variable in patients with and without respiratory complications after adenotonsillectomy. Polysomnographic data may potentially be used for predicting which patients are at higher risk for adverse respiratory events after adenotonsillectomy. Such knowledge is valuable in planning postoperative management and perhaps intraoperative anesthesia management. Predictors of increased respiratory complications include a higher AHI, hypopnea index, and BMI and a lower nadir oxygen saturation.
Correspondence: Sukgi S. Choi, MD, Division of Otolaryngology, Children's National Medical Center, 111 Michigan Ave NW, Washington, DC 20010 (email@example.com).
Submitted for Publication: March 15, 2010; final revision received July 27, 2010; accepted October 14, 2010.
Author Contributions: Drs Jaryszak, Shah, Lander, and Choi had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Jaryszak, Shah, and Choi. Acquisition of data: Jaryszak and Vanison. Analysis and interpretation of data: Jaryszak, Shah, Vanison, and Lander. Drafting of the manuscript: Jaryszak and Shah. Critical revision of the manuscript for important intellectual content: Jaryszak, Shah, Vanison, Lander, and Choi. Statistical analysis: Jaryszak and Lander. Administrative, technical, and material support: Jaryszak, Shah, and Vanison. Study supervision: Jaryszak, Shah, and Choi.
Financial Disclosure: None reported.
Previous Presentation: This study was presented at the American Society of Pediatric Otolaryngology Annual Meeting; May 2, 2010; Las Vegas, Nevada.