Revenaugh PC, Chmielewski LJ, Edwards T, Krishna J, Krakovitz P, Anne S. Utility of Preoperative Cardiac Evaluation in Pediatric Patients Undergoing Surgery for Obstructive Sleep Apnea. Arch Otolaryngol Head Neck Surg. 2011;137(12):1269-1275. doi:10.1001/archoto.2011.208
Author Affiliations: Head and Neck Institute (Drs Revenaugh, Krakovitz, and Anne), Department of Pediatric Cardiology, Children's Hospital (Dr Edwards), and Neurological Institute, Sleep Disorders Center (Dr Krishna), Cleveland Clinic, and Case Western Reserve University School of Medicine (Ms Chmielewski), Cleveland, Ohio.
Objectives To identify the prevalence of clinically relevant findings during cardiac evaluations of pediatric patients with obstructive sleep apnea (OSA) undergoing adenotonsillectomy (TA), and to determine the association between cardiac findings and postoperative respiratory complications.
Design Retrospective medical chart review.
Patients Pediatric patients aged 10 months to 15 years who underwent both echocardiography and polysomnography (PSG) within 6 months prior to TA for OSA from April 2007 through April 2011.
Main Outcome Measures Two pediatric cardiologists independently reviewed echocardiographic studies for evidence of cardiovascular disease. Patients were stratified based on apnea-hypopnea index (AHI) severity (1-5, >5-10, and >10). These groups were compared according to demographic, electrocardiographic (ECG), and echocardiographic values, and postoperative respiratory complications.
Results The medical charts of 57 of 900 patients identified were reviewed following exclusion of those with congenital cardiac abnormalities. The AHI groupings did not differ demographically. No clinically relevant abnormalities were identified on the echocardiogram of any patient. There was a statistically significant association between increased AHI and the appearance of postoperative respiratory complications (P < .05). Indicators of myocardial hypertrophy, such as left ventricular mass index, were not significantly related to AHI in contrast to previously published studies. No echocardiographic or ECG findings were identified that were associated with increased number of postoperative respiratory complications or OSA severity based on AHI.
Conclusions The lack of clinically relevant findings during preoperative cardiac evaluations suggests that aggressive cardiac workup in pediatric patients with OSA may not be indicated unless dictated by comorbidities. Consistent with results in prior studies, preoperative AHI can identify patients at risk for respiratory complications following TA.
Obstructive sleep apnea (OSA) in children is defined as a disorder of breathing during sleep characterized by prolonged partial upper airway obstruction and/or intermittent complete obstruction that disrupts normal ventilation during sleep and normal sleep patterns.1 In the pediatric population, OSA prevalence is estimated to be 2% to 3%, affecting approximately 2 million American children.2 In these children, adenotonsillar hypertrophy is often a major contributing factor in the development of sleep-disordered breathing.3 Obstructive sleep apnea in children is recognized as a major childhood health issue and has been linked to excessive daytime sleepiness,4 hyperactivity and cognitive deficits,5 and cardiovascular changes.6 Previous literature has associated childhood OSA with early cardiovascular changes, including right and left ventricular dysfunction,6 elevated blood pressure,7 autonomic dysregulation,8 and cor pulmonale.9 It is suggested that these cardiac effects may lead to systemic hypertension and right heart dysfunction observed in some children with OSA. The long-term significance and consequences of OSA and associated cardiovascular changes in children are not well studied.10
While adenotonsillectomy (TA) can be an effective treatment for OSA in children, the presence of OSA and associated cardiovascular changes have been linked to increased perioperative risk.11 There is currently no consensus regarding preoperative workup for pediatric patients with OSA undergoing TA. Specifically, it is uncertain when overnight polysomnography (PSG) or cardiac evaluation is necessary, if at all. Some physicians advocate preoperative cardiac evaluation in all pediatric patients with severe OSA to assess for right heart strain, ventricular hypertrophy, cor pulmonale, or other cardiac defects that could elevate the risk of postoperative complications.11 Echocardiography is the most sensitive test to identify these changes in the pediatric population and has been suggested to be of importance in the preoperative workup of patients with severe OSA.12 The aims of this study are to identify the prevalence of clinically relevant findings during cardiac evaluations of pediatric patients with OSA undergoing TA and to determine the association, if any, between cardiopulmonary findings and postoperative respiratory complications.
Medical charts were reviewed to identify patients younger than 18 years who underwent PSG and evaluation by pediatric otolaryngologists at our institution from April 2007 through April 2011. Of the 900 patients identified, only those patients who had OSA diagnosed on PSG and subsequent echocardiography and TA within 6 months were selected. Patients with congenital cardiac disease were excluded. This identified 57 patients for analysis. Patients were stratified into groups based on apnea-hypopnea index (AHI) for statistical analysis.
All patients underwent standard overnight PSG at our institution in an American Academy of Sleep Medicine–accredited laboratory. Studies were attended continuously by a sleep technologist and interpreted by a pediatric physician who is board certified or eligible in sleep medicine. The monitored parameters included single-lead electrocardiogram (ECG), left and right electrooculogram (EOG), central and occipital electroencephalogram (EEG), mental and submental electromyogram (EMG), intercostal EMG, left and right anterior tibialis EMG, snoring, continuous airflow with thermistor and nasal pressure transducer, chest and abdominal effort, oxygen saturation via pulse oximeter, transcutaneous partial pressure of carbon dioxide, and body position via video monitoring. Apnea was defined as the absence of airflow for a 2-breath duration with or without a fall in oxygen saturation. Hypopnea was defined by a 50% or greater reduction in the nasal pressure or thermistor signal lasting for a duration of at least 2 breaths and accompanied by a 3% or greater desaturation from preevent baseline, or an arousal. At least 90% of the event's duration must have met the amplitude reduction criteria for hypopnea. The apneas were classified as obstructive if there was continued evidence of respiratory efforts during the event. Obstructive, mixed, and central apneas were included in the AHI for each patient. Patients were excluded if most of the apneas were central or they were diagnosed as having central sleep apnea. Obstructive sleep apnea was identified by the AHI (rate of apneas + hypopneas per hour of sleep) and was used to stratify patients into 3 groups based on OSA severity: mild (AHI severity, 1-5), moderate (AHI severity, >>5-10), and severe (AHI severity, >>10).
Patients were referred for echocardiography based on clinical evaluation by either the treating otolaryngologist or pediatrician. Common reasons for referral included AHI greater than 10 or recommendation of the physician interpreting the PSG. Patients with comorbidities (eg, craniofacial anomalies, hypotonia, seizure disorder, or other cardiovascular concerns) were also referred for cardiac evaluation. Echocardiography was performed via a transthoracic approach using 2-dimensional and 2-dimensional M-mode recording by pediatric echocardiographers in an Intersocietal Commission for the Accreditation of Echocardiography Laboratories–accredited laboratory and interpreted by 2 board-certified pediatric cardiologists blinded to the magnitude of OSA. Echocardiographic measurements were made according to the recommendations of the American Society of Echocardiography.13 Special attention was given to changes in the right side of the heart, including right ventricular size, geometry, interventricular septal size (IVS) and orientation, presence and magnitude of tricuspid regurgitation (TR), evidence of pulmonary hypertension, or other previously unidentified cardiac defects. Left ventricular mass was calculated using the equation described by Devereux et al14 with left ventricular mass index (LVMI) calculated from the equation described by de Simone et al.15 Posterior wall and interventricular septal thicknesses were indexed to patient height, and relative wall thickness (RWT) was calculated as the sum of the posterior and septal wall divided by left ventricular diastolic diameter. Standard 12-lead ECG was also conducted and reviewed for evidence of ventricular hypertrophy or arrhythmia.
All patients underwent TA performed by a pediatric otolaryngologist. All patients were admitted postoperatively to the floor or pediatric intensive care unit (PICU) based on surgeon preference and/or severity of OSA. Perioperative clinical course was reviewed, and respiratory complications following TA were recorded. Complications were defined as oxygen desaturation below 90% requiring intervention such as oxygen supplementation for more than 8 hours, continuous positive airway pressure, nasopharyngeal airway, intubation, or return to the operating room. All patients were discharged once hospital criteria had been met.
Descriptive statistics were used for demographics and postoperative complications. Fisher exact test or χ2 test was used for comparison of categorical variables between groups as appropriate. Kruskal-Wallis test was used for comparison of qualitative and ordinal variables. Statistical significance was set at P < .05. For analysis of postoperative events, an odds ratio (OR) based on the logistic regression of the categorical variable and each continuous variable was calculated. The ORs for a continuous variable estimated the relative change in odds corresponding to a specified change in the variable, where the number of units of the specified change is defined.
Fifty-seven children were included in the study with groups stratified by OSA severity defined by AHI on PSG. The demographics, clinical measurements, and PSG data of the groups are shown in Table 1 and did not statistically differ among the groups. The patients' mean (SD) age was 5.5 (4.5) years). There was a predominance of male patients (67%). Comorbidities included asthma in 11 patients, trisomy 21 in 5 patients, seizures in 3 patients, and 5 other medical disorders (Pierre Robin sequence, fetal alcohol syndrome, retinoblastoma, myotonic dystrophy, and sickle cell anemia).
The PSG characteristics differed significantly across our AHI groupings (Table 2). As would be expected, there were lower minimum oxygen saturations and a higher percentage of time spent with oxygen saturation less 90% for those patients with higher AHI. In addition, the maximum partial pressure of carbon dioxide and time spent with partial pressure of carbon dioxide greater than 45 mm Hg was also correlated with increasing OSA severity.
There were a total of 12 postoperative respiratory complications, including desaturation requiring prolonged oxygen administration (>8 hours) in 8 patients, 3 instances of intubation and mechanical ventilation postoperatively, and 1 patient who received racemic epinephrine and intravenous steroids in addition to supplemental oxygen (Table 3). The presence of comorbidities was not found to be significantly correlated with the appearance of postoperative complications. In addition, there were no demographic or clinical measures, including weight and body mass index (BMI), that were significantly associated with postoperative respiratory complications (Table 4).
Cardiac parameters measured included left ventricular mass, wall thickness and indices, relative wall thickness, and TR. None of the cardiac measures were found to be significantly associated with AHI grouping (Table 5) or postoperative complications (Table 6). Furthermore, 2 independent blinded pediatric cardiologists reviewed the echocardiographic examinations and did not recommend additional treatment, follow-up, or perioperative management changes based on the echocardiographic results.
Several measurements from the PSG were significantly related to postoperative respiratory complications (Table 7). Total AHI was significantly related to complications (P = .003). The OR was 1.22 (95% CI, 1.07-1.40) for every 5-unit increase in AHI. Rapid eye movement (REM) AHI (P = .005) and non-REM AHI (P = .007) were also significantly associated with postoperative respiratory complications. Oxygen saturation measurements as well as partial pressure of carbon dioxide measurements were also significantly related to postoperative respiratory complications.
A multivariate logistic regression analysis to assess the significance of the association between AHI and the likelihood of a postoperative event, adjusting for minimum oxygen saturation and the percentage of oxygen saturation below 90% as covariates, obtained a likelihood ratio of P = .009. This indicates that the association between elevated AHI and the likelihood of a postoperative event remains statistically significant even when adjusting for oxygen saturation measurements.
Obstructive sleep apnea is now the primary indication for TA,16 and postoperative respiratory complications continue to represent a significant source of morbidity following this procedure. Identification of predictors of postoperative respiratory complications is paramount in reducing morbidity. We sought to delineate the clinical effectiveness of preoperative cardiac workup. Specifically, we investigated the association between preoperative cardiac findings on ECG or echocardiography and postoperative complications. The value of PSG in identifying patients at risk for complications was also evaluated. Based on our analysis, none of the echocardiographic parameters routinely recorded were related to clinical course, and no unexpected findings were discovered during echocardiography that effected a change in treatment planning.
As cardiac imaging techniques improve, subtle, preclinical cardiac changes become easier to detect. Prior research has identified global left ventricular dysfunction as measured by LVMI and myocardial performance index in pediatric patients with OSA.17 Several groups have observed cardiac dysfunction in children with OSA and correlation between global cardiac dysfunction and OSA severity.6,18 In the cases reported by Chan et al,6 they noted LVMI, interventricular and posterior wall thickness indices, and relative wall thickness as significantly related to severity of OSA. In our series, we did not observe statistical significance of any of these calculations. Obstructive sleep apnea has also been identified as an independent risk factor for RV and LV dysfunction in children.6 In contrast, our study finds that there were no statistically significant changes in LVMI or signs of pulmonary hypertension evidenced by severe TR in relation to OSA. The question remains as to whether there are other detectable changes and if these changes are clinically relevant in a pediatric presurgical population.
The development of right heart dysfunction and pulmonary hypertension due to chronic upper airway obstruction is complex and not well understood. In children, it is suggested that hypoxia and respiratory distress produce various neurohumoral factors that may mediate pulmonary vasoconstriction with resultant pulmonary hypertension, right ventricular dysfunction, and impairment of cardiac output.19 Severe cardiac impairment in the form of right heart failure and cor pulmonale has been identified in prior pediatric studies.9 Although previous studies have revealed occurrence of this complication, our study did not identify any patients with cor pulmonale or echocardiographic findings suggestive of right heart dysfunction even in the most severe OSA (top 5% with an average AHI of 93.6). However, owing to the potentially devastating consequences it is important to consider pulmonary hypertension as a possibility in high-risk patients.
The benefits of TA in a pediatric population with OSA are many, including significant improvement in OSA. Mitchell20 reported on 79 pediatric patients who had statistically significant reduction in AHI at mean interval of 7.2 months after TA. Cardiac dysfunction parameters have also been reported to improve after TA. Constantin et al21 retrospectively identified improvement in pulse rate and pulse rate variability following TA in a pediatric population. Improvement in right ventricular dimension, left ventricular end-diastolic diameter, and IVS measurement as assessed by echocardiography following TA has also been reported.22
With the recognition of cardiac dysfunction in pediatric patients with OSA coupled with the need for TA to treat the obstruction, the question remains as how to best evaluate and treat these patients. At some centers, patients with severe OSA are referred for cardiac workup to rule out clinically significant cardiac changes. Based on our results, even in the most severe category of patients with OSA, there were no clinically relevant findings on cardiac echocardiography. We suggest the use of cardiac echocardiography only in the patients with congenital heart disease or in patients considered to be at high risk for having cardiac complications.
Amin et al12 suggested that the presence of LV hypertrophy may be useful in determining a subset of patients for whom more aggressive management of OSA is indicated. In a community-based study, Chan et al6 noted a significant difference in LVMI, relative wall thickness, and IVS in relation to increased AHI, which we did not observe. Furthermore, early and clinically significant changes in OSA could include pulmonary hypertension, resultant TR, and eventual changes in septal and ventricular geometry that were not observed in our study. In addition, we did not observe any significant relation of the presence or magnitude of TR to the appearance of respiratory complications.
Kalra et al11 go further to suggest that clinicians should consider use of echocardiography as a diagnostic tool for patients undergoing TA for OSA. In their case-control study, they identified an increased prevalence of structural cardiac changes (LVMI and left ventricular hypertrophy) in those patients who had postoperative respiratory complications following TA. Unfortunately, the severity or even presence of OSA was not objectively measured in their study. With the absence of clinically relevant anomalies identified in our study and the additional health care cost of echocardiography, it would not be financially prudent to recommend its regular use. In addition, ECG, regarded as less sensitive than echocardiography for detection of these changes,19 did not identify any statistically or clinically significant findings to recommend its use in preoperative workup.
Although cardiac assessment was not predictive of postoperative complications, PSG did indicate a statistically significant relationship. Few studies have reported this relation. McColley et al23 correlated respiratory compromise with severity of OSA based on AHI. In our study, as the total AHI increased, so did the OR for potential complications. Each increase in AHI by a factor of 5 increased the OR by 1.17. In support of this relationship, the American Thoracic Society has recommended that PSG be used to evaluate OSA prior to surgery.1 However, otolaryngologists rarely request PSG prior to TA for OSA in children, and there is no consensus recommendation for its use.24 Owing to the enormous cost in evaluation of every child seen for sleep-disordered breathing, PSG is currently not recommended routinely in all patients. This decision still remains clinician and family based. However, PSG may be considered when there are risk factors for perioperative complications. Cost analyses would be more relevant in this particular instance, but are confounded by definition and incidence of OSA.20
As a retrospective case series, there are several limitations of this study. Inclusion of a control group would have helped to stratify changes related to OSA. Although patients with congenital cardiac disease were excluded, patients with trisomy 21, Pierre Robin sequence, and myotonic dystrophy were included. These patients tend to have higher AHI owing to the effects of their condition on airway anatomy and function. Our study did not involve calculations of right or left heart performance or specific measurements of right heart size. This was deliberate in that we sought to study echocardiographic parameters that are consistently measured and reported in routine evaluations. This leads to the potential for sampling error, because the patients suspected of having severe OSA or those considered to be at increased perioperative risk were generally referred for cardiac evaluation. Indeed, almost 65% of our patients were found to have severe sleep apnea. However, this reinforces our results because patients with severe OSA did not have clinically significant echocardiographic findings. Further research in the form of prospective case-control studies should help to indicate which patients may benefit from preoperative cardiac workup or PSG.
In conclusion, a paucity of clinically relevant findings on echocardiography and lack of correlation between cardiac findings and the presence of postoperative complications suggest that aggressive cardiac workup is not indicated in the evaluation of pediatric patients undergoing TA, regardless of PSG findings. Rather, cardiac workup should be left to the discretion of the clinician for patients with known or suspected cardiac disease. Our study does support the role of PSG in preoperative evaluation of patients deemed to be at high risk for severe OSA. Increased AHI, and to a lesser extent overnight pulse oximetry nadir, was correlated with patients at risk for postoperative respiratory complications and can also be used to stratify patients requiring close postoperative monitoring, such as PICU admission.
Correspondence: Samantha Anne, MD, MS, Head and Neck Institute, Department of Otolaryngology–Head and Neck Surgery, Cleveland Clinic, 9500 Euclid Ave, Desk A71, Cleveland, OH 44195 (firstname.lastname@example.org).
Submitted for Publication: April 28, 2011; final revision received July 11, 2011; accepted September 28, 2011.
Author Contributions: Drs Revenaugh, Edwards, Krishna, and Anne and Ms Chmielewski 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: Revenaugh, Chmielewski, Krishna, Krakovitz, and Anne. Acquisition of data: Revenaugh, Chmielewski, and Edwards. Analysis and interpretation of data: Revenaugh, Edwards, Krishna, Krakovitz, and Anne. Drafting of the manuscript: Revenaugh, Chmielewski, Krishna, and Anne. Critical revision of the manuscript for important intellectual content: Revenaugh, Edwards, Krishna, Krakovitz, and Anne. Administrative, technical, and material support: Revenaugh, Edwards, and Krakovitz. Study supervision: Revenaugh, Krishna, Krakovitz, and Anne.
Financial Disclosure: Dr Krakovitz was a paid consultant for Gyrus/Olympus and Otosonics, and Dr Krishna was a paid consultant for Takeda.
Previous Presentation: This study was presented at the American Society of Pediatric Otolaryngology meeting; May 1, 2011; Chicago, Illinois.