Customize your JAMA Network experience by selecting one or more topics from the list below.
Mudd PA, Thottathil P, Giordano T, et al. Association Between Ibuprofen Use and Severity of Surgically Managed Posttonsillectomy Hemorrhage. JAMA Otolaryngol Head Neck Surg. 2017;143(7):712–717. doi:10.1001/jamaoto.2016.3839
Is the severity of posttonsillectomy hemorrhage requiring surgical control correlated with ibuprofen use?
In this retrospective cohort study of 8868 pediatric patients undergoing tonsillectomy, 222 (3.3%) required surgical intervention for posttonsillectomy hemorrhage, with no correlation between hemorrhage requiring surgical control and ibuprofen exposure. Fifteen children (0.2%) required transfusion; children using ibuprofen had an increased risk for transfusion when hemorrhage occurred.
Ibuprofen used in postoperative management of pain after tonsillectomy has not been shown to increase the overall risk for posttonsillectomy hemorrhage; the severity of bleeding is difficult to quantify but may be a more important outcome to measure.
Ibuprofen used in postoperative management of pain after tonsillectomy has not been shown to increase the overall risk for posttonsillectomy hemorrhage (PTH). The severity of bleeding is difficult to quantify but may be a more important outcome to measure.
To evaluate the association between ibuprofen use and severity of PTH using transfusion events as a marker of severity.
Design, Setting, and Participants
This retrospective cohort study identified 8868 patients who underwent tonsillectomy from January 20, 2011, through June 30, 2014, at the tertiary academic Children’s Hospital of Philadelphia. Of these patients, 6710 met the inclusion criteria. Data were collected using electronic database acquisition and query.
Main Outcomes and Measures
Multivariate analysis was performed to identify independent prognostic factors for PTH and receipt of transfusion.
Of the 6710 patients who met criteria for analysis (3454 male [51.5%] and 3256 female [48.5%]; median age, 5.4 years [interquartile range, 3.7-8.2 years]), 222 (3.3%) presented with PTH that required surgical control (sPTH). A total of 15 of the 8868 patients required transfusion for an overall risk for transfusion after tonsillectomy of 0.2%. Fifteen of 222 patients undergoing sPTH (6.8%) received transfusions. No significant independent increased risk for sPTH was associated with use of ibuprofen (adjusted odds ratio [OR], 0.90; 95% CI, 0.68-1.19). A significant independent association was found in the risk for sPTH in patients 12 years or older (adjusted OR, 2.74; 95% CI, 1.99-3.76) and in patients with a history of recurrent tonsillitis (adjusted OR, 1.52; 95% CI, 1.12-2.06). When using transfusion rates as a surrogate for severity of sPTH, transfusion increased by more than 3-fold among ibuprofen users compared with nonusers (adjusted OR, 3.16; 95% CI, 1.01-9.91), and the upper limit of the 95% CI suggests the difference could be nearly 10 times greater.
Conclusions and Relevance
The risk for sPTH is not increased with use of postoperative ibuprofen but is increased in patients 12 years or older and patients undergoing tonsillectomy with a history of recurrent tonsillitis. Hemorrhage severity is significantly increased with ibuprofen use when using transfusion rate as a surrogate marker for severity.
Tonsillectomy is one of the most commonly performed surgical procedures in children, with more than 500 000 procedures performed each year in the United States.1 Clinical practice guidelines have been established to provide evidence for high-quality perioperative care among patients undergoing tonsillectomy.2 Recommendations for corticosteroid use, antibiotic use, and pain management are included. The guidelines also recommend monitoring one of the common yet serious outcomes of tonsillectomy, posttonsillectomy hemorrhage (PTH). Although many studies have evaluated the risk associations of postoperative pain management and PTH, few studies have qualified the severity of PTH, with the exception of specifying the need for surgical control of PTH (sPTH).3,4
Ibuprofen is a nonsteroidal anti-inflammatory drug that acts by inhibiting cyclooxygenase and thereby inhibiting prostaglandin synthesis. Prostaglandins are local mediators of fever, inflammation, and pain. As a nonselective cyclooxygenase inhibitor, ibuprofen additionally inhibits platelet aggregation and increases bleeding time through inhibition of thromboxane A2.5 Therefore, ibuprofen has the potential to increase postoperative hemorrhage. Recent evidence, however, suggests that ibuprofen is not associated with an increased risk for PTH.6,7 The most recent guidelines, published in 2011, support the use of ibuprofen in postoperative pain management.2 Based on the literature and these guidelines, postoperative pain management protocols were modified for children undergoing tonsillectomy or adenotonsillectomy at Children’s Hospital of Philadelphia (CHOP) by routinely prescribing ibuprofen beginning January 1, 2013. The standard prescribed dose of ibuprofen was 7 mg/kg administered every 6 hours in addition to a combination of acetaminophen (10 mg/kg) and oxycodone hydrochloride (0.075 mg/kg). In May 2014, the use of ibuprofen was stopped because of clinical observations suggesting that the severity of PTH had increased, which may have had links to ibuprofen use. Therefore, a retrospective review of all patients undergoing tonsillectomy before and after ibuprofen implementation was performed to determine whether the clinical observation could be validated statistically. Emergency transfusion, defined as a transfusion in the emergency department or during sPTH, was used as a surrogate of severity of PTH.
The population of interest included patients undergoing tonsillectomy or adenotonsillectomy since implementation of the inpatient electronic health record (EHR) at CHOP in 2011, allowing inclusion of approximately 2 years of patient data before and 2 years after ibuprofen was included as part of standard postoperative pain control protocols. This retrospective cohort study included patients from January 1, 2011, through June 30, 2014. The patients were identified in the EHR (Epic Systems Corporation). Inclusion criteria consisted of all patients undergoing tonsillectomy (Current Procedural Terminology [CPT] code 42825/42826) or tonsillectomy and adenoidectomy (CPT code 42820/42821) at all surgical facilities (1 main operating room and 3 ambulatory surgical centers). Patients who underwent sPTH after the primary procedure was performed at an outside hospital and patients for whom the EHR data were not accessible for validation were excluded. All electronically derived data were deidentified. This study was approved by the internal review board of CHOP, which qualified the study as exempt from informed consent.
CHOP is a tertiary teaching children’s hospital with 13 pediatric otolaryngology–specialized attending staff. Resident- and fellow-level trainees are involved in surgical cases. The standard surgical method at CHOP is complete tonsillectomy using monopolar cautery. A single surgeon routinely used coblation for tonsillectomy in 2012 and then began to use monopolar cautery for tonsillectomy and coblation for adenoidectomy and tonsillar fossae hemostasis. Suture is not routinely used for tonsillectomy or for treatment of PTH. Cases could not be specifically examined for technique with our data set, however; previous studies have not shown a significant difference in PTH rates.8-10 Standard anesthetic regimen at CHOP includes sevoflurane induction with intermittent propofol and opioid (fentanyl citrate or morphine sulfate), or nondepolarizing muscle relaxants may be used as indicated throughout the procedure. Dexamethasone sodium phosphate is routinely administered. Dexmedetomidine hydrochloride was introduced in 2012 but was not used routinely during the study period. Most readmissions for PTH are managed surgically at CHOP.
The initial cohort was identified from hospital billing records through querying for patient encounters with the aforementioned CPT codes. The sPTH group was identified by querying control of oropharyngeal hemorrhage requiring secondary surgical intervention (CPT code 42962). Each case was specifically queried for age, preoperative diagnosis, and use of ibuprofen in postoperative pain management. Patients were assigned to the ibuprofen exposure group if they received ibuprofen before discharge from the hospital or if they were given an ibuprofen prescription for use after discharge. Patients who required secondary sPTH underwent screening for blood transfusions occurring in the 30-day period after tonsillectomy using an internal review board–approved study on evaluation of current surgical blood order practice at CHOP. The postoperative day of return for operative intervention indicated for PTH was also recorded.
Multivariate analysis was completed for patients who had complete information available in the clinical data warehouse, including age (younger than 12 years vs 12 years or older based on surgical CPT code that recognizes a difference between these age groups), preoperative diagnosis (sleep-disordered breathing and/or obstructive sleep apnea [SDB/OSA] vs tonsillitis), and whether an ibuprofen prescription was issued (yes or no). The preoperative diagnosis was retrieved from the surgical procedure notes via text mining. The surgical procedure notes used standard text descriptions that were mapped to SDB/OSA or tonsillitis. The EHR problem list was queried with International Classification of Diseases, Ninth Revision (ICD-9) codes 474.0 to 474.9 for tonsillar disease. Manual review of the preoperative assessment at the ear, nose, and throat clinic was performed for patients with PTH who did not have a preoperative diagnosis available for database extraction and for patients with other chronic diseases of the tonsils and adenoids (ICD-9 code 474.8) or unspecified chronic diseases of tonsils and adenoids (ICD-9 code 474.9).
Multivariate logistic regression analysis was applied to the dependent variables sPTH and sPTH with transfusion, with the following independent variables: (1) age categorized as younger than 12 years or 12 years or older, (2) preoperative diagnosis categorized as OSA (inclusive of SDB and obstructive tonsillar hypertrophy with or without adenoid hypertrophy) and recurrent tonsillitis (including a subgroup of patients with SDB/OSA and recurrent tonsillitis [<2%]), and (3) ibuprofen exposure categorized as yes or no. Additional variables described included the timing of occurrence of sPTH as early (postoperative day 0-1 [≤24 hours]) vs late (postoperative day 2 or longer [>24 hours]), administration of blood transfusion within 30 days after adenotonsillectomy, and the number of patients requiring more than 1 sPTH. Statistical analysis was conducted with STATA software (version 13; StataCorp) using multivariate logistic regression.
Databases queried included (1) the EHR (Epic Systems Corporation) to identify all patients undergoing tonsillectomy or adenotonsillectomy as well as sPTH and to extract medications administered in the hospital and prescriptions at discharge, (2) the anesthesia information management system (CompuRecord; Philips) to identify intraoperative transfusions, and (3) the blood bank database (Meditech) to identify all patients receiving blood products within 30 days after tonsillectomy or adenotonsillectomy. Data were validated using randomly selected medical records for manual review from an existing quality improvement database. The primary focus of the data review was to identify the frequency of patients requiring sPTH. A surrogate for severity was explored through capturing transfusion requirements in volume per body mass (milliliters per kilogram). A manual review of all patients who received transfusion in the postoperative period was completed for further validation.
The data query was unable to identify patients who presented with PTH and did not require sPTH. The primary outcome of this study was to determine the correlation of severity of sPTH with ibuprofen exposure.
We identified 8868 patients as having undergone tonsillectomy (or adenotonsillectomy) from January 20, 2011, through June 30, 2014. A total of 6710 patients (3454 male [51.5%] and 3256 female [48.5%]; median age, 5.4 years [interquartile range, 3.7-8.2 years]) met inclusion criteria for analysis. Two hundred thirty-five medical records (3.5%) required manual validation. All 6710 patients, including all 222 patients presenting with sPTH (3.3%), were included in the multivariate analysis (Table 1). None of the 2158 patients excluded from the study experienced the sPTH outcome. The mean and mode for age were 6.5 and 3.4 years. Most of the patients were younger than 12 years.
Twenty patients (0.3%) during the 4-year study period presented with subsequent recurrence of sPTH after initial sPTH. This group consisted of 7 patients exposed to ibuprofen and 13 not exposed to ibuprofen. Two patients in this group received transfusions, both of whom were younger than 12 years and were in the ibuprofen exposure group. For both patients, the transfusion was received during the second sPTH procedure.
The cumulative 4-year rate of sPTH in our multivariate cohort was 3.3%, with 20 patients (0.3%) requiring more than 1 surgical exploration for hemorrhage control. Fifteen patients (0.2%) required transfusion. A total of 2122 patients (31.6%) were exposed to ibuprofen, which was primarily confined to the years 2013 and 2014 after implementation of new postoperative tonsillectomy pain management protocols. Among the 2122 patients with ibuprofen exposure, 62 (2.9%) had sPTH compared with 160 of 4588 (3.5%) not exposed to ibuprofen. Ibuprofen exposure was not associated with sPTH (adjusted odds ratio [OR], 0.90; 95% CI, 0.68-1.19). Both age 12 years or older (adjusted OR, 2.74; 95% CI, 1.99-3.76) and preoperative diagnosis of recurrent tonsillitis (adjusted OR, 1.52; 95% CI, 1.12-2.06) were independently associated with sPTH (Table 2).
Transfusion associated with secondary surgical intervention was used as a marker for increased severity of hemorrhage. All patients requiring transfusion after tonsillectomy or adenotonsillectomy required sPTH. Of all patients undergoing tonsillectomy (or adenotonsillectomy), 15 (0.2%) received a transfusion, including 8 patients in the ibuprofen exposure group. Of the 222 patients undergoing sPTH, 15 underwent transfusion for a cumulative 4-year transfusion rate of 6.8%. The sPTH group consisted of 62 patients exposed to ibuprofen with 8 transfusions (12.9%) and 160 patients not exposed to ibuprofen with 7 transfusions (4.4%) for an absolute difference of 8.5% (95% CI, 0.1%-19.3%). Logistic regression analysis of sPTH with transfusion did not find association with age (adjusted OR, 0.28; 95% CI, 0.04-2.28) or preprocedure diagnosis (adjusted OR, 1.02; 95% CI, 0.26-3.99). Ibuprofen exposure was associated with a 3-fold higher risk for transfusion during sPTH (adjusted OR, 3.16; 95% CI, 1.01-9.91), suggesting a significant association of ibuprofen with more severe bleeding, with the upper limit of the 95% CI suggesting that the increased severity risk could be nearly 10 times greater (Table 3).
Ibuprofen use has become more acceptable and widespread in controlling postoperative pain in patients who have had tonsillectomy, likely relating to the recently published American Academy of Otolaryngology–Head and Neck Surgery Foundation clinical practice guidelines, which include reviews of critical papers suggesting that the hemorrhage rate is not elevated.2 The most recent prospective randomized trial comparing the use of ibuprofen with narcotic-based postoperative pain management in patients after tonsillectomy found no increased risk for hemorrhage with ibuprofen. However, a significant increase in postoperative hypoxemia was seen in patients receiving opioids.11 Conclusions from the study question the implication of using opioids at all in children aged 1 to 10 years because the risk outweighs the benefit of ibuprofen and acetaminophen, which have consistently been shown to have equivalent analgesia rates. Although analgesia was not examined in this study, we have shown that rate of sPTH is not significantly increased with use of ibuprofen.
Few studies have evaluated the severity of sPTH as a direct function of ibuprofen use likely in part because determining whether the quantity of blood loss relates to the hemorrhage severity or the timing of presentation to the emergency department is difficult. Attempts to examine trends in hemoglobin level nadir or cumulative change in hematocrit level were unsuccessful owing to the lack of standardization of when the hemoglobin and hematocrit results were obtained in association with timing of hemorrhage and because this information was recorded in multiple portions of the EHR at different times. In most cases, the hemorrhage starts when patients are outside the hospital and swallow blood, making it difficult to quantify the blood volume lost during this period. Therefore, in our study, transfusion rate was used as a marker of severity in sPTH. We found a statistically significant and clinically meaningful increased rate of transfusion with sPTH among patients with ibuprofen exposure. The 95% CI is very wide owing to the small number of outcome events (only 15 patients required transfusion). This 95% CI suggests that the association of ibuprofen with more severe bleeding that requires transfusion could be anywhere from a 1-fold to a nearly 10-fold increase (the upper limit of the 95% CI is 9.91). A larger study is needed to generate a more precise estimate of effect. We found no association between patient age and preoperative diagnosis of tonsillitis and ibuprofen use.
Few previous studies have evaluated transfusion risk in PTH. In a small study examining the risk for transfusion in 100 patients undergoing tonsillectomy during a 9-year period,12 the rate of hemorrhage was 6% and the rate of transfusion was 3%. Two of the 3 patients receiving a transfusion in that cohort had abnormal coagulation profiles. A postoperative pain management protocol was not mentioned in that study. Children with von Willebrand disease or hemophilia face an increased risk for PTH, with a rate of 15% reported in a UK study13 that included 500 high-risk patients. The rate of transfusion in that group was 2.4%, but the effect of postoperative pain management protocols was not mentioned.
In the literature, the reported rate of transfusion after PTH is low and is consistent with the findings in our large cohort. This study is the first, to our knowledge, to examine transfusion risk as a surrogate for severity of sPTH in a large cohort. The association of ibuprofen implementation and rate of transfusion is important to consider in counseling patients and may have specific implications in resuscitation management protocols for patients with severe PTH. Furthermore, transfusion rates in PTH require monitoring and reporting to determine whether the widespread increase in the severity of PTH is in fact associated with ibuprofen use. Last, ibuprofen dose standardization has not been examined to determine the minimal threshold for optimal analgesia or the dose-dependent risk of PTH.
The nature of the study using database acquisition relies on correct data entry. Furthermore, changes in postoperative pain management by a parent or outside clinician may have added ibuprofen, which cannot be captured and would lead to some patients in the nonexposure group having had exposure to ibuprofen and therefore a possible crossover effect that cannot be captured with our data set. Platelet function usually returns to normal within 12 hours after administration of ibuprofen,5 and the relative timing of the last dose of ibuprofen in association with the timing of hemorrhage cannot be captured and may be relevant. In addition, standardization in dosing regimen cannot be clarified and thus may also weaken the conclusion of this study. Based on these limitations, we believe it is prudent to consider standardized dosing and administration of ibuprofen with prospective follow-up of outcomes that include rates of transfusion. Finally, standardized documentation of readmission characteristics of patients with PTH can aid in understanding of the true effect of ibuprofen on sPTH and clinical course, specifically 24-hour medication reconciliation at the time of readmission and clear documentation regarding bleeding history, in addition to routine preresuscitation and postresuscitation hematocrit and hemoglobin levels.
Given that the severity of PTH is difficult to quantify and the incidence of the most severe complications is very small, our field must continue to evaluate these complications in multi-institutional outcome registries at regional or national levels. The PTH rate is low, and the rate of transfusion even lower; thus, to demonstrate definitely whether an association exists between ibuprofen and severe sPTH, a much larger patient sample would be required.
Multiple assumptions are made in this study. Patients who received a prescription for ibuprofen were considered as exposed to the medication. We cannot determine how many times the medication was administered from the existing databases nor the temporal association with ibuprofen use and the time of hemorrhage. Including only those patients who require sPTH limits the ability to make conclusions regarding all readmissions for PTH. The data retrieval techniques used limit our ability to examine patients individually to understand indications and variation in techniques. Posttonsillectomy hemorrhage is likely to be a multifactorial problem, and accounting for all variables that contribute to it may be difficult. Querying clinical data sets such as this one may pose additional challenges because often data are not recorded consistently in the same way over time.14 Future efforts to study patient outcomes will benefit from ongoing standardization of EHR at a national level and can build on frameworks, such as the one described in this study, when evaluating large patient populations.
Tonsillectomy is one of the most common surgical procedures performed in children. Monitoring critical outcomes of tonsillectomy aims to improve the quality of surgical care. Additional investigations of postoperative pain management protocols that optimize analgesia while achieving the maximal safety profile are ongoing. Ibuprofen is now widely used in postoperative management of patients undergoing tonsillectomy. The risk for PTH was not increased overall with the postoperative use of ibuprofen; however, a statistically significant and clinically meaningful independent increased risk was found among ibuprofen users 12 years or older who had preoperative diagnosis of recurrent tonsillitis. In addition, this cohort is, to our knowledge, the first and largest to specifically examine transfusion rates in sPTH as a surrogate marker for severity. The data suggest a statistically significant and clinically meaningful increase in the use of transfusion products among ibuprofen users. Additional studies are warranted to investigate this association before any definitive conclusions can be made regarding the potential risk for severe hemorrhage with routine prescription of nonsteroidal anti-inflammatory medications for posttonsillectomy analgesia. Prospective monitoring of hemorrhage inclusive of the severity of hemorrhage and the rates of transfusion should be considered for ongoing and future outcomes-based projects for adenotonsillectomy.
Corresponding Author: Pamela A. Mudd, MD, MBA, Division of Pediatric Otolaryngology, Children’s National Medical Center, 111 Michigan Ave NW, Washington, DC 20010 (firstname.lastname@example.org).
Accepted for Publication: January 28, 2017.
Published Online: May 4, 2017. doi:10.1001/jamaoto.2016.3839
Author Contributions: Drs Jawad and Ahumada had access to data for purposes of database and statistical analysis, respectively. Drs Mudd and Gálvez 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: Mudd, Wetmore, Jawad, Gálvez.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Mudd, Wetmore, Gálvez.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Thottathil, Jawad, Ahumada, Gálvez.
Administrative, technical, or material support: Gálvez.
Study supervision: Wetmore, Elden, Gálvez.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Meeting Presentation: This paper was presented at the Annual Meeting of the American Society of Pediatric Otolaryngology; April 25, 2015; Boston, Massachusetts.
Additional Contributions: David Friedman, MD, transfusion service at Children’s Hospital of Philadelphia (CHOP), contributed to the transfusion data, and Mohamed Rehman, MD, Department of Anesthesiology, CHOP, offered guidance and critical contributions to the study concept. Neither contributor was compensated for this work.
Create a personal account or sign in to: