Efficacy of Tonsillectomy for Pediatric Patients With Dysphagia and Tonsillar Hypertrophy

Objective To determine the effectiveness of tonsillectomy for the treatment of dysphagia related to tonsillar hypertrophy.

Design Prospective cohort study.

Setting Tertiary care pediatric otolaryngology practice.

Participants Eighty-five children aged 2 to 14 years referred for tonsillectomy owing to dysphagia related to tonsillar hypertrophy (dysphagia cohort) or for other indications (control cohort).

Interventions Swallowing Quality of Life (SWAL-QOL) dysphagia questionnaires were administered at the initial clinic visit, on the day of surgery, and at 1 month and 6 months after surgery. Patients were weighed on the day of surgery and at 1 month after surgery.

Main Outcome Measures The primary outcome measure was the SWAL-QOL score. Secondary outcome measures were the type of diet consistency patients tolerated at home and the weight percentile for age.

Results Of 85 patients enrolled, 57 went on to have surgery, completed at least 1 postoperative questionnaire, and were included in the data analysis. At 1 month after tonsillectomy, the dysphagia cohort (n = 18) demonstrated improved SWAL-QOL scores (mean [SD], 58.4 [4.8] before surgery vs 82.4 [5.3] after surgery; P < .001), more patients tolerating a regular diet (12 of 37 patients [33.3%] before surgery vs 22 of 36 [60.0%] after surgery, P = .01), and increased weight percentile for age (mean [SD], 36.5 [10.7] before surgery vs 50.0 [10.6] after surgery; P = .01). Similarly, at 1 month after tonsillectomy, the control cohort (n = 39) demonstrated improved SWAL-QOL scores (mean [SD], 80.8 [2.6] before surgery vs 91.7 [1.8] after surgery; P < .001), more patients tolerating a regular diet (30 of 37 patients [81.1%] before surgery vs 34 of 36 patients [94.4%] after surgery, P = .04), and increased weight percentile for age (mean [SD], 62.8 [5.4] before surgery vs 70.4 [5.1] after surgery; P = .003).

Conclusions Dysphagia related to tonsillar hypertrophy is a significant problem not only among children with dysphagia with a primary complaint but also among a large subset of patients referred for tonsillectomy for other indications. Following tonsillectomy, both groups experience significant improvement in swallowing-related quality of life, ability to tolerate a regular diet, and weight percentile for age. Tonsillectomy is an effective treatment for the management of dysphagia related to tonsillar hypertrophy in children.

Swallowing is a complex process that involves oral, oropharyngeal, and esophageal phases to effectively prepare a food bolus and transfer this bolus down the gastrointestinal tract. Dysphagia is the result of dysfunction of 1 or more of these phases of swallowing. There are numerous potential causes of dysphagia in children, including craniofacial abnormalities, oromotor hypotonia, cerebral palsy, muscle dystonia, tonsillar hypertrophy, and esophageal web or stricture. Tonsillar hypertrophy is believed to present a physical barrier to bolus passage through the oropharynx, leading to dysphagia and feeding difficulties. Given this obvious physical limitation imposed on swallowing by tonsillar hypertrophy, tonsillectomy is often recommended for patients in whom dysphagia seems related to enlarged tonsils.

Tonsillectomy is the second most commonly performed procedure among children in the United States, with more than half a million cases performed annually.¹ Current indications for tonsillectomy include sleep-related breathing disorder and obstructive sleep apnea, recurrent tonsillitis, peritonsillar abscess, recurrent bleeding, and suspected malignant neoplasm. Less common indications include halitosis, dysarthria, dysphagia, and periodic fevers with aphthous stomatitis, pharyngitis, and adenitis.² Although the evidence is substantial to support the use of tonsillectomy in common indications, such as obstructive sleep apnea^3,4 and recurrent tonsillitis,⁵ few data support its use in the treatment of dysphagia. Only a handful of investigations have addressed the role of tonsillectomy in treating dysphagia, and these have focused on children with neurological impairment^6,7 or provided only cursory discussion of dysphagia and tonsillectomy in otherwise healthy children.⁸

The present study was designed to prospectively analyze the effect of tonsillectomy on subjective and objective measures of dysphagia. The following 2 cohorts of children who were referred for tonsillectomy were studied: (1) children with dysphagia-related tonsillar hypertrophy as a primary complaint and (2) children with other indications for tonsillectomy, primarily obstructive sleep apnea and recurrent tonsillitis. We hypothesized that tonsillectomy would lead to improvements in subjective and objective measures of dysphagia in these patients.

This prospective cohort study was approved by the institutional review board at Oregon Health and Science University, Portland. Parents of consecutive patients who had indications for tonsillectomy or adenotonsillectomy were asked to participate in the research study. Inclusion criteria for this study were patients aged 1 to 18 years who were to undergo tonsillectomy or adenotonsillectomy and whose parents provided consent for inclusion in the study. Exclusion criteria were complete dependence on tube feeds or a known neurological disorder causing dysphagia. Patients were divided into the following 2 cohorts: (1) those whose primary indication for surgery was tonsillar hypertrophy and oropharyngeal dysphagia diagnosed by radiographic swallow study or clinical swallow examination by a speech and language pathologist (the dysphagia group) and (2) those whose primary indication for surgery was another diagnosis, such as obstructive sleep apnea or chronic tonsillitis (the control group). Radiographic swallow study findings that led to a diagnosis of dysphagia included prolonged oral holding, bolus undercoating of tonsils, and decreased tongue base retraction, epiglottic deflection, and hyolaryngeal elevation. Before beginning the study, a power analysis was conducted. To observe an effect size of 0.50 in the dysphagia group with a power of 80% and α = .05, a total of 27 patients would be needed. Given the anticipated lower incidence of dysphagia in the control group, 52 patients would be needed to observe an effect size of 0.35.

To assess symptoms of dysphagia and dysphagia-related quality of life, the Swallowing Quality of Life (SWAL-QOL) questionnaire was adapted for use with pediatric patients. The SWAL-QOL is a 44-item validated questionnaire that tests swallowing-related symptoms and quality of life across 11 domains.^9-11 For use with our pediatric patients, the questions were rephrased to be answered from the perspective of the patient's parent (Table 1). In addition, the section investigating sleep symptoms was eliminated, as most members of our control group were expected to have obstructive sleep apnea as a primary diagnosis, rendering the relationship between sleep symptoms and dysphagia tenuous at best. Parents were asked to complete this questionnaire at the initial clinic visit, on the day of surgery, at 1 month after surgery, and at 4 to 6 months after surgery. Although the preoperative questionnaire was given directly to parents while in the clinic or hospital, not all patients followed up in the clinic at 1 month after surgery; among those in whom a telephone follow-up appointment was conducted, the questionnaire was administered over the telephone. Similarly, at 4 to 6 months after surgery, attempts were made to contact the parents via telephone to complete the questionnaire. If parents could not be contacted by telephone at 1 month or 4 to 6 months after surgery, questionnaires were mailed to parents, with instructions for them to complete and return them.

Scoring of the questionnaires followed the method described by the creators of the SWAL-QOL questionnaire¹¹ using a Likert-type method. Briefly, each response on the survey was linearly transformed to a metric ranging from 0 to 100 points, where 100 represents the most favorable state and 0 the least favorable state. Individual responses were equally weighted and were summed to generate an overall outcome score. Additional questions about patient diet were treated as categorical data and were not linearly transformed using the Likert-type method.

Patient weight was recorded on the day of surgery for all the children in the study. However, postoperative weights were available for only those patients who returned to the clinic at 1 month after surgery for follow-up care. To account for patient growth and age variability, weight was recorded as the weight percentile for age, determined using standard 2000 Centers for Disease Control and Prevention¹² growth charts for children 2 years and older and 2006 World Health Organization¹³ growth charts for children younger than 2 years.

Descriptive statistics are given as means (SDs). t Test for paired samples was used to compare preoperative and postoperative SWAL-QOL scores and weight percentiles for age within each cohort. χ² Analysis was used to compare changes in patient diet after surgery. A linear regression analysis was performed to assess change in weight after surgery vs change in SWAL-QOL score.

Eighty-five patients aged 2 to 14 years were enrolled in the study at their initial consult visit, and 77 went on to have surgery; 57 completed at least 1 postoperative questionnaire and were included in the data analysis (Table 2). Compared with the control cohort, the dysphagia cohort was younger and comprised more female patients, although the latter difference was not significant (P = .14).

The dysphagia cohort demonstrated significant improvement across all domains of the SWAL-QOL at 1 month after surgery, and these improvements persisted at 6 months after surgery (Table 3). Although the control cohort scored higher on the SWAL-QOL than the dysphagia cohort before surgery (P < .001), the control cohort demonstrated significant gains across almost all domains of the SWAL-QOL at 1 month after surgery, and these improvements remained up to 6 months after surgery. Although the primary indication for tonsillectomy in the control cohort was not dysphagia, it was striking that many control patients had significantly abnormal SWAL-QOL scores before surgery, suggesting that many of these patients had dysphagia in addition to their primary problem (Figure 1). After surgery, the distribution range of SWAL-QOL scores was much closer to the ideal score of 100.

Patients were also queried about their dietary habits; not surprisingly, most patients in the dysphagia cohort were unable to tolerate a regular diet and instead needed soft foods or purees. However, at 1 month after surgery, there was a significant improvement in the number of patients able to consume a regular diet (Figure 2) (P = .01). The control group demonstrated a smaller but significant change in the number of patients able to tolerate a regular diet after surgery (P = .04).

All the patients were weighed before surgery, and when they returned to the clinic for a 1-month follow-up appointment, weights were again obtained. To standardize them based on age, the weights were converted to a percentile per age based on World Health Organization and Centers for Disease Control and Prevention growth charts (Figure 3). The dysphagia cohort (n = 10) demonstrated a significant increase in weight percentile for age after surgery (P = .01). The control cohort (n = 23) exhibited a smaller but significant increase in weight percentile for age after surgery (P = .003). Finally, the overall change in weight percentile for age was plotted vs the overall change in SWAL-QOL score from the initial clinic visit to the visit at 1 month after surgery (Figure 4). This demonstrated a positive correlation (R² = 0.27) between change in SWAL-QOL score and change in weight percentile for age, indicating that the subjective changes in symptoms measured on the SWAL-QOL translated to measurable gains in patient health.

When parents are queried, dysphagia is a common complaint in their children and is especially prevalent among children requiring tonsillectomy or adenotonsillectomy. Pediatric dysphagia is often multifactorial. Mechanical obstruction of the oropharynx by tonsillar hypertrophy has long been recognized as a potential cause of dysphagia, and tonsillectomy is often recommended in these patients.² However, few studies have examined the role of tonsillectomy to treat dysphagia related to tonsillar hypertrophy; therefore, the present study was performed to address this gap in the literature. Our study confirmed that tonsillectomy is an effective treatment for dysphagia related to tonsillar hypertrophy in children.

The current literature about the use of tonsillectomy to treat dysphagia is sparse. Ahlqvist-Rastad et al⁸ described a cohort of 122 children in Sweden with symptoms suggesting tonsillar obstruction. They noted that 53% of these children had difficulty with eating by history, and after tonsillectomy the eating problems were completely ameliorated in all but 4. However, this study was primarily focused on airway obstruction and provided no quantitative data to demonstrate changes in dysphagia symptoms after tonsillectomy.

Studies have demonstrated the efficacy of tonsillectomy in children with neurological impairment. Conley et al⁷ showed that tonsillectomy is safe and may improve dysphagia based on oropharyngeal motility investigations in children with neurologically impaired deglutition. A 2009 study⁶ showed improvements in a similar patient population following tonsillectomy using videofluoroscopic swallow studies and a telephone survey of dysphagia symptoms. Although these studies provide good evidence for the use of tonsillectomy to treat dysphagia in patients with neurological impairment, the results are difficult to extrapolate to the general pediatric population.

A notable outcome in the present study was the improvement seen in both cohorts studied. Both study groups showed improvements in SWAL-QOL scores, ability to tolerate a regular diet, and weight gain following surgery. This was the expected outcome among the dysphagia cohort, as these patients were believed to have dysphagia primarily due to obstructive tonsillar hypertrophy. The improvement seen in the SWAL-QOL scores provides empirical evidence for the long-standing belief that tonsillectomy benefits this patient population. The improvement seen in the control cohort indicates that dysphagia may be prevalent among the patient population with adenotonsillar hypertrophy in general and that tonsillectomy improves swallowing among many of these patients in addition to treating their primary problem. This study was not designed to determine any prognostic tools that could be used clinically to define which patients may have improved swallowing following tonsillectomy. Patients in the dysphagia cohort were referred to the feeding clinic for specific problems with eating, such as failure to thrive, refusal to eat solids, and choking on feedings, and were evaluated by a team of specialists. Children who had minor swallowing difficulties or those who were not formally evaluated by a swallowing specialist were not included in the dysphagia cohort. Future studies could further define the subset of patients most at risk for dysphagia related to tonsillar hypertrophy and most likely to benefit from tonsillectomy, as well as validate clinical decision tools to help physicians identify these patients.

In both cohorts, significant improvements in QOL related to dysphagia and in the objective measurement of weight gain were noted after tonsillectomy, but the correlation between QOL and weight gain was weak. This is not unexpected, as previous investigations on the relationship between SWAL-QOL scores and bolus transit times have shown that there is a loose correlation between these subjective and objective measures of swallowing function.¹⁴ Several factors may account for this in our study, including the use of parents as a proxy for the child's experience via the questionnaire and the use of an indirect measure of dysphagia improvement (weight percentile for age) rather than a direct measure (videofluoroscopy). Furthermore, although evidence indicates that some children gain weight following tonsillectomy, factors other than improved swallowing function may have a role in this weight gain. Following tonsillectomy, children demonstrate increased levels of insulin-like growth factor and other hormones,¹⁵ decreased levels of inflammatory markers like tumor necrosis factor,¹⁶ and improved work of breathing,¹⁷ all of which could contribute to weight gain. Nevertheless, a correlation between SWAL-QOL scores and weight percentile changes exists, confirming that children experience an objective benefit following tonsillectomy.

Although the data in the present study are the first to measure the effect of tonsillectomy on dysphagia related to tonsillar hypertrophy in a general pediatric population, there are several limitations to this study. First, the fact that the study was performed among a small sample in a tertiary care academic center may decrease the ability to generalize the results to a community practice setting. Further research, possibly a multicenter trial that could accrue a large study population, may address this issue. Second, most patients in the control group had an obstructive sleep disturbance as an indication for surgery; although our results suggest that dysphagia may be a significant concern in this group, we cannot generalize this finding to patients undergoing tonsillectomy for other reasons, such as recurrent tonsillitis. Third, although it would be ideal to assess postoperative swallowing function by videofluoroscopic swallow studies as an objective measure of dysphagia, it would be ethically impossible to subject children to additional radiation exposure following tonsillectomy if their issues with dysphagia had resolved. Fourth, although the SWAL-QOL is a validated measure of dysphagia in adults, in modifying it for use in children, it can no longer be considered a validated tool, and we have introduced the additional layer of complexity of parent reporting. However, the score distributions we observed in our dysphagia and control cohorts agree with previous findings using the SWAL-QOL,¹⁸ and the consistency in score distribution between the initial clinic visit and preoperative time points, before any intervention, indicates that it seems to be a reliable tool. Fifth, although there was a high dropout rate of approximately one-third of patients, this is unlikely to have affected the results, as both cohorts were equally affected. Sixth, as in many surgical studies, it is impossible to perform a sham procedure to create a placebo group; therefore, we were unable to discriminate the observed benefit of tonsillectomy from a placebo effect.

In conclusion, our study provides the first quantifiable evidence to date that tonsillectomy is an effective treatment for dysphagia related to tonsillar hypertrophy. We show a benefit among children in whom dysphagia is their primary problem, as well as improvement in swallow function after tonsillectomy among patients with other obstructive issues related to tonsillar hypertrophy. Although further research is warranted to define the subset of patients most likely to benefit from tonsillectomy to treat dysphagia, these results may help the clinician counsel patients and families about the effects of tonsillectomy on dysphagia.

Correspondence: Daniel Clayburgh, MD, PhD, Department of Otolaryngology–Head and Neck Surgery, Oregon Health and Science University, 3181 SW Sam Jackson Park Rd, Mail Code PV-01, Portland, OR 97239 (clayburg@ohsu.edu).

Submitted for Publication: March 8, 2011; final revision received June 9, 2011; accepted July 31, 2011.

Author Contributions: Drs Clayburgh and MacArthur 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: Clayburgh, Gorsek, and MacArthur. Acquisition of data: Clayburgh, Milczuk, Gorsek, Sinden, Bowman, and MacArthur. Analysis and interpretation of data: Clayburgh, Milczuk, Gorsek, and MacArthur. Drafting of the manuscript: Clayburgh, Gorsek, and Bowman. Critical revision of the manuscript for important intellectual content: Clayburgh, Milczuk, Sinden, and MacArthur. Statistical analysis: Clayburgh. Administrative, technical, and material support: Clayburgh, Milczuk, Gorsek, Sinden, Bowman, and MacArthur. Study supervision: Clayburgh, Milczuk, and MacArthur.

Previous Presentation: This study was presented at the American Society of Pediatric Otolaryngology meeting; April 30, 2011; Chicago, Illinois.