S2 Otolaryngology���Head and Neck Surgery 144(1S) Tproceduresthe onsillectomy is one of the most common surgical pro- cedures in United States, with more than 530 000 performed annually in children younger than 15 years.1 Indications for surgery include recurrent throat infections and sleep-disordered breathing (SDB),2 both of which can substantially affect child health status and quality of life (QoL). Although there are benefits of tonsillectomy, complications of surgery may include throat pain, postopera- tive nausea and vomiting, delayed feeding, voice changes, hemorrhage, and rarely death. The frequency of tonsillectomy, associated morbidity, and availability of hundreds of random- ized clinical trials assessing interventions create a pressing need for evidence-based guidance to aid clinicians. This document is intended for all clinicians who diagnose or manage patients aged 1 to 18 years for whom tonsillec- tomy is being considered for indications of recurrent throat infection and/or SDB as defined as follows: ��� Tonsillectomy is defined as a surgical procedure per- formed with or without adenoidectomy that completely removes the tonsil, including its capsule, by dissecting the peritonsillar space between the tonsil capsule and the muscular wall. Depending on the context in which it is used, it may indicate tonsil- lectomy with adenoidectomy, especially in relation to SDB. ��� Throat infection is defined as sore throat caused by viral or bacterial infection of the pharynx, palatine tonsils, or both, which may or may not be culture positive for group A streptococcus. This includes the terms strep throat and acute tonsillitis, pharyngitis, adenotonsillitis, or tonsillopharyngitis. ��� Sleep-disordered breathing is characterized by abnor- malities of respiratory pattern or the adequacy of ventilation during sleep, which include snoring, mouth breathing, and pauses in breathing. SDB encompasses a spectrum of obstructive disorders that increases in severity from primary snoring to obstructive sleep apnea (OSA). Daytime symptoms associated with SDB may include excessive sleepiness, inattention, poor concentration, and hyperactivity. ��� Caregiver is used throughout the document to refer to parents, guardians, or other adults providing care to patients undergoing tonsillectomy. The importance of tonsillectomy as an intervention relates to its documented benefit on child QoL. For example, when compared with healthy children, children with recurrent throat infections have more bodily pain and poorer general health and physical functioning.3 Tonsillectomy may improve QoL by reducing throat infections, health care provider visits, and the need for antibiotic therapy.4 Similarly, SDB is associated with cognitive and behavioral impairment in children that usually improves after tonsillectomy5-8 along with QoL,6,9-12 sleep disturbance,5,10,13,14 and vocal quality.15,16 Wide variations in tonsillectomy rates have been reported across the world, including Japan,17 Canada,18 the United King- dom,19 and the United States.20 Such variations are usually ascribed to heterogeneity in clinical practice and training rather than to differences in clinical need.21 The current lack of consen- sus in the United States on surgical indications and perioperative management further supports the need for an evidence-based clinical practice guideline to highlight best practices. Guideline Scope and Purpose This guideline is intended for all clinicians in any setting who interact with children aged 1 to 18 years who may be candi- dates for tonsillectomy. The guideline does not apply to tonsillotomy, intracapsular surgery, or other partial removal techniques of the tonsil because of the relatively sparse high- quality published evidence on these techniques and limited long-term follow-up. Similarly, the guideline does not apply to populations of children excluded from most tonsillectomy research studies, including those with diabetes mellitus, car- diopulmonary disease, craniofacial disorders, congenital anomalies of the head and neck region, sickle cell disease, and other coagulopathies or immunodeficiency disorders. The primary purpose of this guideline is to provide clinicians with evidence-based guidance in identifying children who are the best candidates for tonsillectomy. Secondary objectives are to optimize the perioperative management of children undergo- ing tonsillectomy, emphasize the need for evaluation and intervention in special populations, improve counseling and education of families of children who are considering tonsillec- tomy for their child, highlight the management options for patients with modifying factors, and reduce inappropriate or unnecessary variations in care. Lastly, we sought to identify gaps in knowledge that would guide future research. This guideline predominantly addresses indications for tonsillectomy based on 1Department of Surgery, University of Toledo Medical Center, Toledo, Ohio, USA 2Division of Otolaryngology-Head & Neck Surgery, University of Kentucky Chandler Medical Center, Lexington, Kentucky, USA 3Cardinal Glennon Children���s Medical Center, Saint Louis University School of Medicine, St Louis, Missouri, USA 4Department of Otolaryngology, SUNY Downstate Medical Center and Long Island College Hospital, Brooklyn, New York, USA 5Department of Pediatrics, Cincinnati Children���s Hospital Medical Center, Cincinnati, Ohio, USA 6Department of Pediatrics, Baystate Children���s Hospital, Springfield, Massachusetts, USA 7Department of Otolaryngology, Eastern Virginia Medical School, Norfolk, Virginia, USA 8Division of Otolaryngology, The Children���s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA 9Department of Anesthesiology and Critical Care, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, USA 10Sleep Apnea Surgery Center, East Palo Alto, California, USA 11James���s Project, Wayne, Pennsylvania, USA 12Advanced Pediatrics, Vienna, Virginia, USA 13Albany ENT & Allergy Services, PC, Albany, New York, USA 14Department of Pediatrics, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin, USA 15Qualis Health, Seattle, Washington, USA 16The Cochrane ENT Disorders Group, Oxford, United Kingdom 17American Academy of Otolaryngology���Head and Neck Surgery Foundation, Alexandria, Virginia, USA Corresponding Author: Reginald F. Baugh, MD, Professor and Chief, Otolaryngology-Head and Neck Surgery, Dowling Hall 2124 Mail Stop 1095, 3000 Arlington Avenue, Toledo, OH 43614 Email: reginald.baugh@utoledo.edu by guest on January 4, 2011 oto.sagepub.com Downloaded from

Baugh et al S3 obstructive and infectious causes. Other indications for surgery, including orthodontic concerns, tonsiliths, halitosis, and chronic tonsillitis, are not extensively discussed. The evidence in these areas is limited and generally of lesser quality, and a role for shared decision making is present. This guideline is intended to focus on quality improvement opportunities judged most important by the working group. It is not intended to be a comprehensive, general guide for man- aging patients undergoing tonsillectomy. In this context, the purpose is to define useful actions for clinicians, regardless of discipline, to improve the quality of care. Conversely, the statements in this guideline are not intended to limit or restrict care provided by clinicians based on the assessment of indi- vidual patients. Although there is evidence to guide management of certain aspects of tonsillectomy, there is no evidence-based clinical practice guideline relevant to all specialties managing such patients in the United States. A guideline is warranted because of documented practice variations in the care of patients who undergo tonsillectomy and the morbidity and mortality associ- ated with this surgical intervention. Health Care Burden Incidence of Tonsillectomy Tonsillectomy is the second most common ambulatory surgi- cal procedure performed on children in the United States.1 In 2006, there were 530 000 tonsillectomies performed in chil- dren younger than 15 years, constituting 16% of all ambulatory surgery in this age group. The only procedure with greater fre- quency was myringotomy with insertion of tube, for which 667 000 procedures were reported the same year. Between 1915 and the 1960s, tonsillectomy was the most frequently performed surgical procedure in the United States. Data in 1993 from the National Hospital Discharge Survey, however, noted a decrease of more than 50% in inpatient ton- sillectomy rates from 1977 to 1989.22 Similar reports from 1978 to 1986 showed that the rate of tonsillectomy for treat- ment of throat infections declined however, the frequency of SDB as the primary indication for the procedure increased.2 A recent study reported that the overall incidence rates of tonsil- lectomy have significantly increased in the past 35 years, with SDB being the primary indication for surgery.23 Indications for Surgery The 2 most common indications for tonsillectomy are recur- rent throat infections and SDB. Throat infections are a common reason to see a primary care physician and often result in antibiotic treatment.24 The cost of outpatient visits and the medications prescribed for sore throats including anti- biotics are substantial. Indirect costs associated with throat infections and SDB are substantial due to missed school and loss of time from work for caregivers. Treatment of SDB is associated with an increase in health care utilization and cost. Children with SDB, compared with controls, have a significantly higher rate of antibiotic use, 40% more hospital visits, and an overall elevation of 215% in health care usage mostly from increased respiratory tract infections.25 Children with tonsillar disease, including chil- dren with throat infections and SDB, also showed significantly lower scores on several QoL subscales including general health, physical functioning, behavior, bodily pain, and care- giver impact when compared with healthy children.3 SDB represents a spectrum of disorders ranging in severity from primary snoring to OSA. The prevalence of OSA in the pediatric population is 1% to 4%26 as many as 10% of chil- dren have primary snoring.27 Up to 30% to 40% of children with clinically diagnosed SDB exhibit behavioral problems that include enuresis,28 hyperactivity, aggression, anxiety, depression, and somatization.29 OSA is also associated with poor school performance and a decrease in QoL.8 The QoL of children with OSA is similar to children with chronic condi- tions such as asthma and juvenile rheumatoid arthritis.30 Controversy persists about the actual benefits of tonsillec- tomy as compared with observation and medical treatment of throat infections. Although tonsillectomy for recurrent throat infections in severely affected children has been shown in a ran- domized controlled trial to reduce the frequency and severity of infections in the 2 years following surgery,31 the same cannot be shown for less severe cases or for a period greater than 2 years after surgery.20,31 Observational studies, however, show improved disease-specific and global QoL after tonsillectomy for recurrent or chronic sore throat, as measured by validated instruments.4 These children suffered fewer infections after sur- gery, resulting in fewer antibiotics and physician visits. A growing body of evidence indicates that tonsillectomy is an effective treatment for SDB,32 based on the idea that tonsil- lar hypertrophy is a principal cause. A meta-analysis of case series33 and a recent study34 showed that tonsillectomy was effective at improving or resolving SDB in most children. There is also evidence that behavioral parameters, school per- formance, and QoL improve after resolution of this sleep disorder.8 Harms and Adverse Events of Tonsillectomy Tonsillectomy is a surgical procedure with an associated mor- bidity that includes possible hospitalization, risks of anesthesia, prolonged throat pain, and financial costs. A common complication of tonsillectomy is bleeding during or after the surgery. In published reports, the rate of primary hemorrhage (within 24 hours of surgery) has ranged from 0.2% to 2.2% and the rate of secondary hemorrhage (more than 24 hours after surgery) from 0.1% to 3%.35 Hemorrhage after tonsillectomy may result in readmission for observation or in further surgery to control bleeding. Other complications of tonsillectomy are diverse and have been well described.36 Operative complications include trauma to the teeth, larynx, pharyngeal wall, or soft palate difficult intubation laryngospasm laryngeal edema aspira- tion respiratory compromise endotracheal tube ignition and cardiac arrest. Injury to nearby structures has been reported, including lip burn, eye injury, and fracture of the mandibular condyle. Postoperative complications include nausea, vomiting, pain, dehydration, referred otalgia, postobstructive pulmonary by guest on January 4, 2011 oto.sagepub.com Downloaded from

S4 Otolaryngology���Head and Neck Surgery 144(1S) edema, velopharyngeal insufficiency, and nasopharyngeal stenosis. Complications are more common in patients with cra- niofacial disorders, Down syndrome, cerebral palsy, major heart disease, or bleeding diatheses and in children younger than 3 years with polysomnography (PSG)���proven OSA.37-41 After tonsillectomy, about 1.3% of patients experience delayed discharge during the initial hospital stay, and up to 3.9% have secondary complications requiring readmission.42 The primary reasons for readmission or prolonged initial stay included pain, vomiting, fever, and tonsillar hemorrhage. In addition to these common causes of morbidity, many unusual and rare complications of tonsillectomy have also been described.43 Among these are reports of vascular injury, sub- cutaneous emphysema, jugular vein thrombosis, atlantoaxial subluxation (Grisel syndrome), taste disorders (hypogeusia, ageusia, dysgeusia, and phantogeusia), and persistent neck pain (Eagle syndrome). Mortality rates for tonsillectomy have been estimated at between 1 in 16 000 to 1 in 35 000, based on data from the 1970s.44 There are no current estimates of tonsillectomy mor- tality, but a prospective audit reported only 1 postoperative death after 33 921 procedures in England and Northern Ire- land.42 About one-third of deaths are attributable to bleeding, while the remainder are related to aspiration, cardiopulmonary failure, electrolyte imbalance, or anesthetic complications.35,45 Similarly, airway compromise is the major cause of death or major injury in malpractice claims after tonsillectomy.46 Structure and Function of the Tonsils The palatine tonsils are lymphoepithelial organs located at the junction of the oral cavity and the oropharynx. They are stra- tegically positioned to serve as secondary lymphoid organs, initiating immune responses against antigens entering the body through the mouth or nose. The greatest immunological activity of the tonsils is found between the ages of 3 and 10 years.47 As a result, the tonsils are most prominent during this period of childhood and subsequently demonstrate age-dependent involution.48 The epithelium of the tonsils is cryptic and reticulated and contains a system of specialized channels lined by ���M��� cells.49 These cells take up antigens into vesicles and transport them to the extrafollicular region or the lymphoid follicles. In the extrafollicular region, interdigitating dendritic cells and mac- rophages process the antigens and present them to helper T lymphocytes. These lymphocytes stimulate proliferation of follicular B lymphocytes and their development into either antibody-expressing B memory cells capable of migration to the nasopharynx and other sites or plasma cells that produce antibodies and release them into the lumen of the crypt.49 While all 5 immunoglobulin (Ig) isotypes are produced in the palatine tonsils, IgA is arguably the most important prod- uct of the tonsillar immune system. In its dimeric form, IgA may attach to the transmembrane secretory component to form secretory IgA, a critical component of the mucosal immune system of the upper airway. Although the secretory component is produced only in the extratonsillar epithelium, the tonsils do produce immunocytes bearing the J (joining) chain carbohydrate.50 This component is necessary for bind- ing of IgA monomers to each other and to the secretory component and is an important product of B-cell activity in the follicles of the tonsil. Effects of Tonsillitis and Tonsillectomy on Immunity With chronic or recurrent tonsillitis, the controlled process of antigen transport and presentation is altered due to shedding of the M cells from the tonsil epithelium.49 The direct influx of antigens disproportionately expands the population of mature B-cell clones and, as a result, fewer early memory B cells go on to become J-chain���positive IgA immunocytes. In addition, the tonsillar lymphocytes can become so overwhelmed with persistent antigenic stimulation that they may be unable to respond to other antigens. Once this immunological impair- ment occurs, the tonsil is no longer able to function adequately in local protection, nor can it appropriately reinforce the secre- tory immune system of the upper respiratory tract. There would therefore appear to be a therapeutic advantage to removing recurrently or chronically diseased tonsils. On the other hand, some studies demonstrate minor alterations of Ig concentrations in the serum and adjacent tissues following tonsillectomy.51-54 Nevertheless, there are no studies to date that demonstrate a significant clinical impact of tonsillectomy on the immune system.51-53 Methods and Literature Search The guideline was developed using an explicit and transparent a priori protocol for creating actionable statements based on supporting evidence and the associated balance of benefit and harm.55 The guideline panel was chosen to represent fields of sleep medicine, advanced practice nursing, anesthesiology, infectious disease, family medicine, otolaryngology���head and neck surgery, pediatrics, and consumers. Several group mem- bers had prior experience in developing clinical practice guidelines. The systematic literature search was divided into 2 stages and aimed to identify clinical practice guidelines, systematic reviews, or meta-analyses (stage I) and randomized controlled trials (stage II) using key biomedical literature databases (Table 1). The search was based on the string tonsillectom*, adenotonsillectom*, tonsillotom*, posttonsillectom*, (tonsil* OR adenotonsil*) AND (surg* OR operat* OR remov* OR preop* OR periop* OR postop*). Results were screened to remove duplicates and citations that were not pertinent. 1. Published and unpublished consensus- and evidence- based clinical practice guidelines less than 10 years old in English met inclusion criteria. The final data set included 17 guidelines, and of those, 2 guidelines56,57 met quality criteria of having been produced under the auspices of a medical associa- tion or organization and having an explicit, a priori, method for ranking evidence and linking evidence to recommendations. 2. Systematic reviews less than 15 years and meta- analyses with a systematic review in English met by guest on January 4, 2011 oto.sagepub.com Downloaded from

Baugh et al S5 inclusion criteria. The search filter used to identify systematic reviews in PubMed was devised based on the search strategy used by the National Health Service Evidence���Cancer.58 Reviews that met a rating of adequate required a clear objective, explicit search strategy, and valid data extraction. The final data set included 36 systematic reviews (including 9 Cochrane systematic reviews). 3. Randomized controlled trials published in English with no age restrictions were identified using an adaptation of the Cochrane Highly Sensitive Search Strategy.59 Published or unpublished completed trials with a definite or possible randomized con- trolled design met inclusion criteria. The final data set yielded 705 studies that were grouped into the following broad topics: analgesia (193), technique (125), anesthesia (67), nausea/vomiting (62), hemostasis (45), recovery (35), steroids (26), surgi- cal indications (24), antibiotics (16), outcomes assessment (15), surgical complications (2), periop- erative care (2), and other (93). Results of the literature searches were distributed to guideline panel members at the first meeting, including electronic listings with abstracts (if available) of the searches for guidelines, randomized controlled trials, and systematic reviews. This material was supplemented, as needed, with targeted systematic searches to address specific needs identified in developing the guideline through April 11, 2010. In a series of conference calls, the working group defined the scope and objectives of the proposed guideline. During the 9 months devoted to guideline development ending in 2010, the group met twice with interval electronic review and feed- back on each guideline draft to ensure accuracy of content and consistency with standardized criteria for reporting clinical practice guidelines.60 The Guideline Implementability Appraisal and Extractor tool61 was used to appraise adherence of the draft guideline to method- ological standards, to improve clarity of recommendations, and to predict potential obstacles to implementation. Guideline panel members received summary appraisals in May 2010 and modi- fied an advanced draft of the guideline. The final draft guideline was distributed to a multidisciplinary group of 44 external reviewers, representing the target audience, for feedback and comment. Responses were compiled, reviewed by a subgroup of the panel, and incorporated into the guideline. The document was then submitted to the journal���s peer-review process before publi- cation. A scheduled review process will occur at 5 years from publication or sooner if new compelling evidence warrants ear- lier consideration. Classification of Evidence-Based Statements Guidelines are intended to reduce inappropriate variations in clinical care, to produce optimal health outcomes for patients, and to minimize harm. The evidence-based approach to guide- line development requires that the evidence supporting a policy be identified, appraised, and summarized and that an explicit link between evidence and statements be defined. Evidence-based statements reflect both the quality of evidence and the balance of benefit and harm that is anticipated when the statement is followed. The definitions for evidence-based statements62 are listed in Tables 2 and 3. Guidelines are never intended to supersede professional judgment rather, they may be viewed as a relative constraint on individual clinician discretion in a particular clinical cir- cumstance. Less frequent variation in practice is expected for a ���strong recommendation��� than might be expected with a ���recommendation.��� ���Options��� offer the most opportunity for practice variability.63 Clinicians should always act and decide in a way that they believe will best serve their patients��� interests and needs, regardless of guideline recommendations. Guide- lines represent the best judgment of a team of experienced clinicians and methodologists addressing the scientific evi- dence for a particular topic.62 Making recommendations about health practices involves value judgments on the desirability of various outcomes associ- ated with management options. Values applied by the guideline Table 1. Databases Housing Guidelines, Systematic Reviews, and Randomized Controlled Trials Guidelines Systematic Reviews Randomized Controlled Trials National Guidelines Clearinghouse CMA Infobase (Canada) NHS Evidence ENT & Audiology (UK) National Library of Guidelines (UK) NICE (UK) SIGN (Scotland) New Zealand Guidelines Group Australian National Health and Medical Research Council TRIP Database PubMed GIN NHS Evidence ENT & Audiology (UK) Cochrane Library (Cochrane Database of Systematic Reviews, DARE, HTA Database, NHS EED) TRIP Database PubMed EMBASE CINAHL AMED BIOSIS Previews CAB Abstracts ISI Web of Science AHRQ HSTAT Cochrane ENT Disorders Group Trials Register CENTRAL (Cochrane Library) HTA Database (Cochrane Library) PubMed EMBASE CINAHL AMED BIOSIS Previews CAB Abstracts ISI Web of Science clinicaltrials.gov mRCT ICTRP TRIP Database by guest on January 4, 2011 oto.sagepub.com Downloaded from

S6 Otolaryngology���Head and Neck Surgery 144(1S) panel sought to minimize harm and diminish unnecessary and inappropriate therapy. A major goal of the committee was to maintain transparency and be explicit about how values were applied and to document the process. Financial Disclosure and Conflicts of Interest The cost of developing this guideline, including travel expenses of all panel members, was covered in full by the American Academy of Otolaryngology���Head and Neck Surgery (AAO- HNS) Foundation. Potential conflicts of interest for all panel members were compiled and distributed before the first confer- ence call. After review and discussion of these disclosures,64 the panel concluded that individuals with potential conflicts could remain on the panel if they: (1) reminded the panel of potential conflicts before any related discussion, (2) recused themselves from a related discussion if asked by the panel, and (3) agreed not to discuss any aspect of the guideline with indus- try before publication. Lastly, panelists were reminded that conflicts of interest extend beyond financial relationships and may include personal experiences, how a participant earns a living, and the participant���s previously established ���stake��� in an issue.65 Guideline Key Action Statements Each evidence-based statement is organized in a similar fashion: an evidence-based statement in bold, followed by a strength of the recommendation in italic. Several para- graphs subsequently discuss the evidence base supporting the statement, concluding with an ���evidence profile��� of aggregate evidence quality, benefit-harm assessment, and statement of costs. Lastly, there is an explicit statement of the value judgments, the role of patient preferences, clarifi- cation of any intentional vagueness by the panel, and a repeat statement of the strength of the recommendation. An Table 2. Guideline Definitions for Evidence-Based Statements Statement Definition Implication Strong recommendation A strong recommendation means the benefits of the recommended approach clearly exceed the harms (or that the harms clearly exceed the benefits in the case of a strong negative recommendation) and that the quality of the supporting evidence is excellent (Grade A or B).a In some clearly identified circumstances, strong recommendations may be made based on lesser evidence when high-quality evidence is impossible to obtain and the anticipated benefits strongly outweigh the harms. Clinicians should follow a strong recommendation unless a clear and compelling rationale for an alternative approach is present. Recommendation A recommendation means the benefits exceed the harms (or that the harms exceed the benefits in the case of a negative recommendation), but the quality of evidence is not as strong (Grade B or C).a In some clearly identified circumstances, recommendations may be made based on lesser evidence when high- quality evidence is impossible to obtain and the anticipated benefits outweigh the harms. Clinicians should generally follow a recommendation but should remain alert to new information and sensitive to patient preferences. Option An option means that either the quality of evidence that exists is suspect (Grade D)a or that well-done studies (Grade A, B, or C)a show little clear advantage to one approach versus another. Clinicians should be flexible in their decision making regarding appropriate practice, although they may set bounds on alternatives patient preference should have a substantial influencing role. No recommendation No recommendation means there is both a lack of pertinent evidence (Grade D)a and an unclear balance between benefits and harms. Clinicians should feel little constraint in their decision making and be alert to new published evidence that clarifies the balance of benefit versus harm patient preference should have a substantial influencing role. aSee Table 3 for definition of evidence grades. Table 3. Evidence Quality for Grades of Evidence Grade Evidence Quality A Well-designed randomized controlled trials or diagnostic studies performed on a population similar to the guideline���s target population B Randomized controlled trials or diagnostic studies with minor limitations overwhelmingly consistent evidence from observational studies C Observational studies (case control and cohort design) D Case reports, reasoning from first principles (bench research or animal studies) X Exceptional situations in which validating studies cannot be performed and there is a clear preponderance of benefit over harm by guest on January 4, 2011 oto.sagepub.com Downloaded from