Abstract
Objective:
To determine the relationship between frequency of tonsillitis and the risk of post-tonsillectomy hemorrhage (PTH) in pediatric patients undergoing tonsillectomy for recurrent tonsillitis.
Methods:
After obtaining IRB approval from Nationwide Children’s Hospital, charts for all patients who underwent a total tonsillectomy in 2017 for recurrent or chronic tonsillitis were retrospectively reviewed (n = 424). Patients were divided into 2 cohorts based on the frequency of tonsillitis prior to surgery: those meeting the 1-year criteria with 7 or more infections in the past year (n = 100), and those who did not meet criteria defined as those with fewer than 7 infections in the past year (n = 324). The primary outcome of interest was PTH. Comparison of cohorts and frequency of PTH were assessed using bivariate analyses. Kaplan-Meier curves were used to compare time to onset of hemorrhage between primary vs. secondary PTH. Generalized mixed and logistic regression models were used to evaluate risk of hemorrhage following tonsillectomy.
Results:
Among a total cohort of 424 patients undergoing tonsillectomy, 23.58% (n = 100) met criteria while 76.42% (n = 324) did not. A total of 8.73% (n = 37) patients experienced PTH. Compared to those who did not meet criteria, those who met criteria had a higher odds of developing PTH; however, this was not significant (OR: 1.42 [95% CI: 0.67, 2.98], P = .3582). Estimated probability of developing PTH for those who met criteria was 11% [95% CI: 6.19, 18.81] compared to 8.03% [95% CI: 5.52, 11.54] for those who did not meet criteria. Among all PTH cases, 5.41% (n = 2) were primary hemorrhage while 94.59% (n = 35) were secondary hemorrhage with 50% of those with secondary PTH having experienced hemorrhage within 6 days [95% CI: 5, 7] of tonsillectomy. Patients with neuromuscular conditions had significantly higher odds of PTH (OR: 4.75 [95% CI: 1.19, 18.97], P = .0276).
Conclusion:
Patients who met the 1-year criteria for tonsillectomy did not have a significantly higher odds of PTH. Further research is needed to better evaluate the relationship between infection frequency and risk of PTH.
Introduction
Tonsillectomy is one of the most frequently performed pediatric surgical procedures, with more than 530 000 children under 15 years old undergoing the procedure each year in the United States alone.1,2 The 2 most widely accepted indications for tonsillectomy are recurrent tonsillitis and sleep-disordered breathing (SDB). 3 Tonsillectomy is generally well-tolerated; however, complications such as post-tonsillectomy hemorrhage (PTH), while rare, can be life-threatening. 4 The overall rate of PTH has been reported to be around 4.5%. 2
Many studies have attempted to elicit risk factors for PTH; however, only a few variables have been consistently associated with an increased risk. One of the strongest risk factors is age, with older patients—specifically those between the ages of 11 and 17 years old—having a higher risk of post-operative bleeding.3 -5 Another well-documented risk factor is the indication for tonsillectomy, as multiple studies have demonstrated that patients undergoing the procedure for recurrent or chronic infection are more likely to experience PTH.5 -7
In its clinical practice guideline, The American Academy of Otolaryngology—Head and Neck Surgery (AAO-HNS) suggests that providers may recommend tonsillectomy for recurrent throat infections with a frequency of at least 7 episodes in the past year, at least 5 episodes per year for 2 years, or at least 3 episodes per year for 3 years. 8 Although an infectious indication for surgery may be a predictor of PTH, the relationship between the frequency of tonsillitis and the risk of post-operative bleeding remains unclear. The aim of this study is to assess if the frequency of recurrent tonsillitis affects the risk of PTH.
Methods
Study Participants
After obtaining approval from the Nationwide Children’s Hospital’s (NCH) Institutional Review Board (IRB), charts were retrospectively reviewed for all patients aged 2 to 18 who underwent total tonsillectomy in 2017 for indications of recurrent tonsillitis (RT), chronic tonsillitis (CT), or both recurrent and chronic tonsillitis (RCT). Patients having any indications of tonsillitis (RT, CT, or RCT) along with coexisting sleep disordered breathing (SDB) were identified and sleep disordered breathing was used as a covariate for analyses. Patients undergoing tonsillectomy for SDB with no other indications of RT, CT, or RCT were not included. Patients were divided into 2 cohorts based on the number of tonsillitis episodes documented in the pre-operative ENT clinic note prior to surgery: those meeting the 1-year criteria for tonsillectomy (7 or more infections in the past year), and those who did not meet the 1-year criteria (5 infections per year for 2 years or 3 infections per year for 3 years). Clinic notes include a template which captures number of tonsillitis episodes. All patients underwent surgery at NCH with all tonsillectomies performed by surgeons within NCH’s Otolaryngology department.
Data Elements
The primary outcome of interest for this study was the risk of post-tonsillectomy hemorrhage. For all cases of PTH, primary hemorrhage was defined as hemorrhage documented within an ED encounter 24 hours following tonsillectomy. Secondary hemorrhage was defined as hemorrhage documented within an ED encounter from 2 to 14 days post tonsillectomy. In addition to frequency of infection prior to surgery, other independent variables included demographic characteristics, co-morbidities such as presence of Down syndrome, neuromuscular conditions, bleeding disorders, and acute tonsillitis. Clinical characteristics included infection within 2 weeks of surgery, active infection at the time of surgery, and tonsillectomy technique.
Statistical Analysis
Patient and clinical characteristics of the sample are described using medians and interquartile ranges for continuous variables and counts and proportions for categorical variables. Normality of continuous variables were assessed using a combination of Shapiro-Wilk test, QQ plots, and histogram. Comparison of cohorts (those who met criteria vs those who did not) were done with Chi square tests or Fisher’s exact test for association among categorical variables. Wilcoxon Mann-Whitney test was used to compare differences among continuous variables. Kaplan-Meier curves were plotted to compare time to PTH between primary versus secondary hemorrhage with log-rank test used to assess for difference. Risk of PTH was assessed using both logistic and generalized mixed model regression. All P values <.02 in univariable regression analyses were used in multivariable regression modeling. Final models were chosen based off the lowest Akaike information criterion (AIC). All tests were 2 tailed with P value <.05 indicating significance. All figures and analyses were conducted using SAS Enterprise Guide 8.1 (Cary, NC).
Results
Among a total cohort of 424 tonsillectomy patients, a majority of patients were female (60%, n = 255) and non-Hispanic White (78.3%, n = 332) with median age of 8 years (IQR: 6, 10). Seventy-six percent (n = 324) had less than 7 infections in the prior year and did not meet 1-year criteria while 23.58% (n = 100) had 7 or more infections in the prior year and were considered to have met 1-year criteria. Both cohorts share similar demographic characteristics (P > .05) but differed significantly in proportion of acute tonsillitis. Those who met 1-year criteria had significantly more cases of acute tonsillitis compared to those who did not meet 1-year criteria (85% vs 3.7%, P < .0001). Median number of infections in the last year among those who met 1-year criteria was 7.5 (IQR: 7, 9) compared to 4 (IQR: 3, 5) for those who did not meet 1-year criteria (P < .0001). Indications of tonsillitis was also similar between both groups (P = .5552) as was presence of coexisting sleep disordered breathing (P = .7392). Quinsy tonsillectomy was performed only among 0.93% (n = 3) of those who met criteria (Table 1).
General Cohort Characteristics.
Note. Criteria of less than 7 infections or 7 or more infections is by year.
Abbreviations: D, Days; IQR, Interquartile range; y, year.
PTH was noted among 8.73% (37/424) of the entire cohort. Among all cases of PTH, 5.41% (n = 2) were defined as primary hemorrhage while 94.59% (n = 35) were defined as secondary hemorrhage. All primary hemorrhage cases occurred among those who did not meet 1-year criteria. Sixty-nine percent (n = 24) of secondary hemorrhage cases occurred among those who did not meet 1-year criteria while 31.43% (n = 11) occurred among those who met 1-year criteria. Overall, no association was found between proportion of hemorrhage cases and criteria type (P = .4215) (Table 2).
Breakdown of Post Tonsillectomy Hemorrhage by Infection Criteria.
Note. Infection criteria is by year.
Kaplan-Meier curves indicate that on average, 25% of tonsillectomy patients will experience onset of secondary hemorrhage 4 days [95% CI: 3, 5] following surgery. Fifty percent of tonsillectomy patients were estimated to experience secondary hemorrhage 6 days [95% CI: 5, 7] after tonsillectomy with 75% of the cohort experiencing secondary hemorrhage 9 days [95% CI: 7, 10] post tonsillectomy (Figure 1, Table 3).
Projected Onset of Post Tonsillectomy Hemorrhage.
Kaplan-Meier curve of time to post tonsillectomy hemorrhage following tonsillectomy.
Compared to those who did not meet criteria, those who did have a 1.42 odds ([95% CI: 0.67, 2.98], P = .3582) of having a PTH. This translates to an 11% [95% CI: 6.19, 18.81] predicted probability of developing a PTH compared to 8.03% [95% CI: 5.52, 11.54] among those did not meet criteria. Compared to males, females on average have a lower odds of developing PTH (OR: 0.97, [95% CI: 0.49, 1.93], P = .9291). Each year increase in age resulted in 1.05 odds [95% CI: 0.96, 1.15] of developing PTH (P = .3026). Those with neuromuscular conditions stand to have a significantly higher risk of PTH (OR: 4.18, [95% CI: 1.06, 16.49], P = .0411). Adjusting for sleep disordered breathing in multivariable analyses increases these odds to 4.75 ([95% CI: 1.19, 18.97], P = .0276) (Table 4).
Unadjusted and Adjusted Modeling of Risk for Post Tonsillectomy Hemorrhage.
Note. Tukey-Kramer adjustment used for multiple comparisons; P values <0.2 in univariable modeling used toward multivariable modeling.
Abbreviations: d, day; PTH, post tonsillectomy hemorrhage; y, year
Discussion
Although multiple studies have demonstrated that patients undergoing tonsillectomy for recurrent infections are more likely to experience hemorrhage, the relationship between infection frequency and risk of PTH remains unclear. Our overall hemorrhage rate was nearly 9%, which is higher than previous values reported in the literature. 2 This may be at least partially attributable to our inclusion of patients with recurrent infections. Moreover, the majority (95%) of PTH cases were secondary, with 75% occurring by post-operative day 9. It is important to note that there was no significant difference in the ratio of primary to secondary hemorrhage as a function of number of tonsillitis episodes (Tables 2 and 3). Our results align with previous literature stating that the majority of PTH occurs between 5 and 10 days after tonsillectomy. 9
Furthermore, the predicted probability of PTH for patients meeting the 1-year criteria was 11%, compared to 8% for those not meeting the 1-year criteria. This difference did not reach statistical significance (Table 2). This is concordant with a study published in 2020 by Aldrees et al which retrospectively analyzed risk factors for secondary PTH at a children’s hospital in Saudi Arabia. They found no significant differences in bleeding rates based on the number of yearly infections. 10 A study published in 2020 by Johnston et al found that children with infectious symptoms in the 2 weeks prior to tonsillectomy were more likely to experience PTH. 11 Our results did not replicate this finding. However, we did find that active infection at the time of tonsillectomy increased the risk of PTH, although this was not significant (Table 4).
The presence of a neuromuscular condition significantly increased the likelihood of experiencing PTH. The risk of PTH in this cohort increased even more if the patient had SDB in addition to recurrent and or chronic tonsillitis (Table 4). This is a very specific patient population, and the relationship between neuromuscular co-morbidities and increased risk of bleeding is unclear. Regardless, providers should be aware of this finding when performing tonsillectomies on this population.
This study is not without limitations—one of the largest being the single-institution, retrospective design, which limits the generalizability of the data beyond Nationwide Children’s Hospital. Additionally, it is possible that due to this design, we failed to capture every episode of infection for each patient, and the acuity of infection may be misclassified for some patients. Further, with a larger sample size, it is possible that our data may reach statistical significance.
Conclusion
Compared to patients who do not meet the 1-year criteria of recurrent/chronic tonsillitis, patients who did meet criteria did not have significantly higher odds of developing PTH. However, patients with a neuromuscular condition undergoing tonsillectomy for recurrent/chronic tonsillitis have a significantly increased risk of PTH, especially with coexisting SDB.
Footnotes
Authors’ Note
Accepted for podium presentation at the Society for Ear, Nose and Throat Advances in Children Annual Meeting, November 29-December 2, 2018.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
