Incidence of post-tonsillectomy hemorrhage in children and adults: a study of 4,848 patients. (Original Article).
We conducted a retrospective study of 4,848 patients to evaluate the age-specific incidence of post-tonsillectomy hemorrhage that required surgical treatment. We reviewed the charts of 2,567 patients younger than 15 years (pediatric group) and 2,281 patients aged 15 years and older (adult group) who had undergone tonsillectomy with or without adenoidectomy. We found that post-tonsillectomy hemorrhage occurred significantly more often in the adult group (3.9 vs 1.6%; p<0.001). Moreover, primary hemorrhage (<24 hr postoperatively) was also significantly more common in the adult group than in the pediatric group (82.9 vs 65.9%, p = 0.023). Analysis of other parameters revealed that post-tonsillectomy hemorrhage was significantly more common in males and in patients who had a history of chronic or recurrent throat infection. Awareness of these risk factors should help improve patient care and outcomes.
Postsurgical hemorrhage, the major complication of tonsillectomy and a potentially life-threatening occurrence, is classified as primary (<24 hr postoperatively) or secondary (>24 hr). In children younger than 3 years of age, surgery is also complicated by dehydration, poor oral intake of fluids and solids, and upper airway problems, especially in cases where adenotonsillectomy has been performed to resolve breathing problems. (1-3) Primary hemorrhage is considered to be more serious than secondary hemorrhage, but the experience in our clinic has been that secondary bleeding can also be brisk and massive and require treatment under general anesthesia.
Despite continuous specialization in otolaryngology, surgeons in many departments operate on patients of all ages. Many authors believe that age is associated with the risk of postoperative hemorrhage. However, there is no widely accepted international definition of age ranges. The American Medical Association classifies people as neonates (<1 mo of age), infants (1 mo to 2 yr), children (2 to 13 yr), adolescents (13 to 17 yr), and adults ([greater than or equal to]18 yr). On the other hand, many pediatric societies use 15 years as the threshold for adulthood. Therefore, reports of surgical results obtained only in "children" are of limited value in educating clinicians. These studies would be more useful if researchers could agree on a common classification of age when reporting results in populations that include a broad range of ages. Moreover, a common classification system would be beneficial from a medicolegal perspective because physicians could provide patients with more precise information regarding the risk of complications.
In this article, we report the results of our retrospective study of post-tonsillectomy hemorrhage, with or without adenoidectomy, in a large population with a broad range of ages. For our study, we classified patients younger than 15 years as children and patients 15 years and older as adults.
Patients and methods
We reviewed the charts of 4,848 patients who had undergone tonsillectomy, with or without adenoidectomy, at our clinic between Jan. 1, 1988, and Aug. 31, 2000. Indications for surgery included the management of upper airway obstruction, recurrent or chronic throat infection, and recurrent childhood ear disease. Patients who had risk factors for hemorrhage (e.g., hypertension, hyperthyroidism, and anticoagulation therapy) were managed preoperatively; those who did not respond did not undergo surgery.
We compiled data on all patients in whom postoperative bleeding from the tonsillar fossae required treatment under general anesthesia to prevent cardiovascular collapse. We used the Pearson chi-squared test (p<0.05) to compare the bleeding rates in the two groups. We excluded from our analysis patients who had peritonsillar abscess, malignant disease, or sleep apnea (confirmed by overnight monitoring). We also did not include patients who underwent adenoidectomy without tonsillectomy as well as those who had previously undergone adenoidectomy or tonsillectomy. We divided patients into two groups based on age and designated them as the pediatric group and the adult group:
Pediatric group. The pediatric group was made up of 2,567 patients aged 6 months to 14.9 years (mean: 7.3; median: 6; standard deviation [SD]: [+ or -]3.72); 1,364(53.1%) were boys (figure 1). Adenotonsillectomy was performed in 1,832 of these patients (71.4%) to treat upper airway obstruction caused by adenotonsillar hypertrophy; the remaining 735 patients (28.6%) underwent tonsillectomy alone for recurrent episodes of tonsillitis (more than five episodes per year or more than three episodes that required antibiotic treatment).
Adult group. The adult group was made up of 2,281 patients aged 15.0 to 93 years (mean: 31.5; median: 27; SD: [+ or -]14.4); 1,059 (46.4%) were men (figure 1). Of this group, 2,105 (92.3%) underwent tonsillectomy alone for recurrent episodes of tonsillitis (same criteria) and 176 (7.7%) underwent adenotonsillectomy for upper airway obstruction caused by adenotonsillar hypertrophy.
All procedures were performed with general anesthesia and oral intubation. The tonsils were removed via the dissection technique, and hemostasis was achieved by suture ligation; cautery was not used. Adenoidectomy was performed via curettage with an adenotome, and hemostasis achieved by temporarily packing the nasopharynx with sponges soaked in epinephrine. Patients refrained from eating and drinking for 6 hours postoperatively. Prophylactic antibiotics were administered to prevent endocarditis. The surgeon made at least two rounds per day in the ward, and the nursing staff provided continuous monitoring, which made early recognition of postoperative hemorrhage possible.
Patients in both groups who did not experience any complications were discharged 6 days after surgery in accordance with national guidelines. (The length of post-tonsillectomy hospital stay is a topic of debate in Germany.) Patients who did experience postoperative hemorrhage were sent home when their general condition, hemoglobin concentration, and state of wound healing allowed for safe discharge. Those patients who had undergone ligation of the external carotid artery because of massive blood loss were discharged only after their wounds had completely healed.
Hamorrhage rates according to age. The adult group had a significantly higher rate of post-tonsillectomy hemorrhage that required treatment than did the pediatric group (3.9 vs 1.6%; p<0.001).
Pediatric group. Of the 2,567 children, treatment of hemorrhage from the tonsillar fossae was necessary for 41 patients (1.6%) -- 20 boys and 21 girls. The youngest was 2 years of age and the oldest 14.9 years (mean: 8.7; median: 9; SD: [+ or -]4.0). Among these 41 patients were 23 of the 735 (3.1%) who underwent tonsillectomy alone and 18 of the 1,832 (0.98%) who underwent adenotonsillectomy. Among the latter group, hemorrhage occurred from the tonsillar fossae in 10 patients and at the adenoidectomy site in eight.
Adult group. Of the 2,281 adults, 88 (3.9%) required treatment under general anesthesia for hemorrhage from the tonsillar fossae -- 58 men and 30 women. The youngest patient was 16 years old and the oldest was 86 (mean: 33.7; median: 29; SD: [+ or -]16.8). Postoperative hemorrhage occurred in 84 of the 2,105 patients (4.0%) who underwent tonsillectomy alone and in four of the 176 patients (2.3%) who underwent adenotonsillectomy.
Rates according to procedure. Postoperative hemorrhage was more common among those patients who underwent tonsillectomy alone than in those who underwent adenotonsillectomy.
Adenotonsillectomy. A total of 2,008 patients underwent adenotonsillectomy -- 1,832 children and 176 adults. Hemorrhage from the tonsillar fossae occurred in 16 (0.87%) and six (3.4%) patients, respectively (figure 2). Sixteen of the 22 were primary bleeds. The latest onset of bleeding was observed 6 days following surgery.
Tonsillectomy alone. Tonsillectomy alone was performed for treatment of recurrent episodes of tonsillitis on 2,840 patients -- 735 children and 2,105 adults. Hemorrhage occurred in 23 (3.1%) and 84 (4.0%) patients, respectively (figure 3). The latest episode of bleeding occurred in an adult on postoperative day 18.
Rates according to sex. In the pediatric group, the rates of postoperative bleeding among boys and girls were comparable (1.5 and 1.7%, respectively; p = 0.47). In the adult group, men had a significantly higher rate than women (5.5 vs 2.5%; p<0.001).
Rates according to time of onset. Of the 41 hemorrhages in the pediatric group, 27 (65.9%) were primary. Primary hemorrhages accounted for 73 of the 88 hemorrhages (82.9%) in the adult group. The difference in these rates was statistically significant (p = 0.023).
Severe complications. Among the total population of 4,848 patients, only seven (0.14%) experienced severe complications that required a blood transfusion or ligation of the external carotid artery -- one in the pediatric group (0.04%) and six in the adult group (0.26%).
Pediatric group. The only severe complication among the children occurred in a 42-month-old boy, who underwent both a blood transfusion and ligation of the external carotid artery. Following the boy's tonsillectomy, we had recommended that this patient's hospital discharge be postponed because he had experienced recurrent hemorrhage following a previous tonsillectomy, but his mother insisted that he be sent home on postoperative day 6. Within hours after the boy returned home, he experienced sudden bleeding from his mouth. He was referred to a pediatric clinic, where a transfusion was administered. While at the clinic, the boy was diagnosed with signs of deep shock, and he was referred back to our clinic. We attempted to save his life by performing ligation of the external carotid artery, but because the boy had already lost so much blood, he died the next day. This case represents the only fatal outcome during our entire study.
Adult group. Transfusion was required for three of the 2,281 adults (0.13%). Two of these patients (one who had undergone an adenotonsillectomy and one who had undergone a tonsillectomy plus a uvulopalatoplasty) were transfused on the day of surgery. The other patient received a transfusion after he had experienced unusual bleeding on postoperative days 12 and 18.
Arterial ligation was performed on three adults (0.13%) as a result of massive blood loss -- a 55-year-old man on the day of surgery after two attempts to stop the bleeding by intraoral suture ligation had failed; a 22-year-old woman 5 days following surgery; and a 19-year-old man 5 days postoperatively who had experienced continuous bleeding following his surgery.
In general, post-tonsillectomy hemorrhage occurs in fewer than 10% of cases. (4) Most of these bleeds are primary. Secondary bleeds can occur at any time during the first 2 postoperative weeks. (4)
Indications for surgery appear to be associated with age. The percentage of adenotonsillectomies that are being performed to treat upper airway obstruction and sleep apnea caused by adenotonsillar hyperplasia has increased significantly during the past 15 years. (1-3) It is unclear whether this change is attributable to an increase in obstructive airway problems, an increase in diagnosis, or an increase in the number of children with adenotonsillar hyperplasia. In adults, the primary indication for surgery remains infection.
The importance of being able to provide patients/parents with precise information regarding the expected postoperative course is obvious. Because indications for surgery and the risks of postoperative complications appear to change depending on the patient's age, the use of a common system of age classification would be helpful. In Germany and some other countries, childhood legally ends and adulthood begins on the 15th birthday. (5) Therefore, we used this cutoff point in conducting our study.
According to our results, chronic or recurrent infection likely plays a role in the incidence of postoperative hemorrhage. We base this assertion on the facts that (1) bleeding was more common in the adult group and (2) 92.3% of the adult group underwent tonsillectomy alone for treatment of infection (compared with only 28.6% of the pediatric group). Moreover, the rate of hemorrhage was higher in adults who underwent tonsillectomy alone than in those who underwent adenotonsillectomy (4.0 and 2.3%, respectively). Likewise, the children who underwent tonsillectomy alone for infection had a higher rate of bleeding than did those who underwent adenotonsillectomy for hyperplasia (3.1 and 0.98%, respectively).
Studies of children. Previous studies of tonsillectomy in patients younger than 15 years of age have addressed the issue of the safety of postoperative outpatient management. Lee reviewed 3,240 patients younger than 14 years and found that 0.6% had to return to the operating room for treatment of hemorrhage. (6) The mean length of readmission stay was 7 days. All cases of primary bleeding in this study occurred within the first 3 hours. Episodes of secondary hemorrhage were recorded, but no treatment under general anesthesia was necessary for any of these patients. Lee concluded that outpatient tonsillectomy is a safe procedure provided that care is taken to ensure that patients meet proper selection criteria -- for example, an absence of systemic disease, a positive parental attitude, favorable logistics with regard to distance and transport, and a tractable patient.
Lannigan et al performed a retrospective study of 4,386 children younger than 13 years and found that 0.57% of these patients required treatment for episodes of primary bleeding, all of which occurred within 6 hours of surgery. (7) They did not mention the incidence of secondary hemorrhage.
Truy et al compared outcomes in 311 patients younger than 15 years who were scheduled for either inpatient or outpatient tonsillectomy.5 Inpatient surgery was planned for all patients who were in an unhealthy preoperative state, who had sleep apnea syndrome, whose social circumstances precluded outpatient surgery, and/or whose parents refused outpatient surgery. Among the outpatients, 23.8% had to be admitted as a result of postoperative complications; hemorrhages (most of them primary) had to be treated in 1%.
Lee and Sharp conducted a prospective questionnaire study of children younger than 14 years to ascertain the complication rate (excluding primary hemorrhage) during the first 5 days postoperatively.8 They found that hemorrhage was observed in 8.9% of these children, but none required surgical treatment. Of note is the fact that only 291 of the 419 questionnaires (69.5%) were suitable for analysis.
Gabalski et al prospectively evaluated the incidence of complications during the first postoperative week in 534 children younger than 15 years who had undergone outpatient tonsillectomy. (9) They found that bleeding occurred in only two patients (0.37%); both experienced primary hemorrhage and neither required treatment. These data were collected during an anonymous survey of surgeons who participated in the study voluntarily.
Rakover et al retrospectively evaluated risk factors for postoperative complications in 363 children younger than 15 years. (10) Eighteen patients (5.0%) experienced a total of 21 hemorrhages--14 primary and seven secondary. Surgical treatment was necessary for only five of the primary bleeds and four of the secondary bleeds. Rakover et al concluded that episodes of bleeding that occur in the recovery room were a significant risk factor for complications. The postoperative complication rate was not associated with age, indication for surgery, method of surgery, or intraoperative complications.
Studies of both children and adults. Kristensen and Tveteras published a retrospective analysis of 1,150 patients between 2 and 71 years of age." They reported that 32 tonsillectomies (2.8%) were complicated by hemorrhage--18 primary and 14 secondary. Male patients between 13 and 24 years old were at significant risk. Hemorrhage was associated with previous peritonsillar abscess but not with the indication for surgery (e.g., tonsillitis, current abscess, or gross hypertrophy).
Schroeder conducted a retrospective study of 756 tonsillectomy patients aged 2 to 83 years, including an unknown number who had peritonsillar abscess or unilateral tonsillar enlargement that suggested a neoplasm. (12) He found that only seven patients (0.93%; six of them female) experienced postoperative hemorrhage that required treatment.
Schmidt et al evaluated the association between hemorrhage and the method of hemostasis in 2,096 patients between the ages of 3 and 83 years (mean: 22.5), including those with malignant diseases, sleep apnea syndrome, infectious mononucleosis, and peritonsillar abscess. (13) They determined that the overall incidence of postoperative hemorrhage was 5.4%. Primary hemorrhage occurred more often when hemostasis was achieved by suture ligation than by electrocautery. Schmidt et al reported that an age greater than 10 years was a risk factor for bleeding; children younger than 10 years had a threefold lower incidence.
Study of adults. Moralee and Murray conducted a retrospective study of 2,157 patients older than 16 years who underwent inpatient tonsillectomy. (14) They found that bleeding occurred in only 42 patients (1.9%); 95% of these hemorrhages occurred within 6.8 hours. Moralee and Murray concluded that serious primary hemorrhage is uncommon. When it does occur, the authors were 95% confident that the latest onset of bleeding would occur in the range of 5.2 to 8.4 hours postoperatively.
Comparisons with other studies. Comparison of our results with those of other researchers is hampered because some reporting criteria were different and because relevant information was not always included in other reports. For example, Lee's study included patients who had been carefully selected, which was not the case in our study. (6) Lee and Sharp did not include cases of primary hemorrhage in their results. (8) Kristensen and Tveteras" and Schroeder (12) did not report the mean age of their patients, and Schroeder did not indicate the length of follow-up. Schmidt et al based their findings on different methods of hemostasis. (13) They also concluded that patients younger than 10 years had a low risk of hemorrhage, but they did not relate this finding to the particular indication for surgery. (13) Moralee and Murray did not definitively identify the age distribution of their patients, their indications for surgery, or the nature of their follow-up. (14) Finally, only Schmidt et al took into account the e xperience of the surgeons, something we did not do. (13)
Other authors have reported findings that are not in agreement with ours. (5-7,10-12,14) Some have reported substantially lower rates of hemorrhage than we did, but we believe the reason for this might be that their studies were smaller than ours. For example, the rate of bleeding among adults reported by Moralee and Murray (1.9%) is in sharp contrast to our findings (3.9%). (14) Also in contrast to our findings were those of Schroeder, (12) who found that most patients with postoperative hemorrhage were female, and Kristensen and Tveteras, (11) who found that girls younger than 12 years were AT particular risk for bleeding. On the other hand, our findings with respect to the rate of hemorrhage in adults was comparable to those of some other studies. (11,13,14)
Because of its retrospective nature, our study had some limitations. For example, we could not analyze variables of interest such as the length of operating time, the amount of blood loss, patients' body mass index, or the precise number of episodes of tonsillitis prior to surgery because these data were not available to us. Another drawback is that there was a great imbalance in the types of surgery performed in the two age groups, which might have made our findings susceptible to statistical bias. Therefore, a prospective study would be most helpful in determining whether the risk factors we identified for hemorrhage (male sex, age [greater than or equal to] 15 yr, and a history of chronic or recurrent tonsillitis) are valid.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
(1.) Rothschild MA, Catalano P. Biller HF. Ambulatory pediatric tonsillectomy and the identification of high-risk subgroups. Otolaryngol Head Neck Surg 1994;110:203- 10.
(2.) Tom LW, DeDio RM, Cohen DE, et al. Is outpatient tonsillectomy appropriate for young children? Laryngoscope 1992;102:277-80.
(3.) Wiatrak BJ, Myer CM III, Andrews TM. Complications of adenotonsillectomy in children under 3 years of age. Am J Otolaryngol 1991;12:170-2.
(4.) Randall DA, Hoffer ME. Complications of tonsillectomy and adenoidectomy. Otolaryngol Head Neck Surg 1998;1l8:61-8.
(5.) Truy E, Merad F, Robin P. et al. Failures in outpatient tonsillectomy policy in children: A retrospective study in 311 children. Int J Pediatr Otorhinolaryngol 1994;29:33-42.
(6.) Lee IN. Outpatient management of T and A procedure in children. J Otolaryngol 1985;14:176-8.
(7.) Lannigan FJ, Martin-Hirsch DP, Basey E. Clinical audit: Is daycase adenotonsillectomy safe? Br J Clin Pract 1993;47:254-5.
(8.) Lee WC, Sharp JF. Complications of paediatric tonsillectomy post-discharge. J Laryngol Otol 1996;110:136-40.
(9.) Gabalski EC, Mattucci KF, Setzen M, Moleski P. Ambulatory tonsillectomy and adenoidectomy. Laryngoscope 1996;106:77-80.
(10.) Rakover Y, Almog R, Rosen G. The risk of postoperative haemorrhage in tonsillectomy as an outpatient procedure in children. Int J Pediatr Otorhinolaryngol 1997;41:29-36.
(11.) Kristensen S, Tveteras K. Post-tonsillectomy hemorrhage. A retrospective study of 1150 operations. Clin Otolaryngol 1984;9:347-50.
(12.) Schroeder WA, Jr. Post tonsillectomy hemorrhage: A ten-year retrospective study. Mo Med 1995;92:592-5.
(13.) Schmidt H, Schmitz A, Stasehe N, Hormann K. [Surgically managed postoperative hemorrhage after tonsillectomy]. Laryngorhinootololgie 1996;75:447-54.
(14.) Moralee SJ, Murray JA. Would day-case adult tonsillectomy be safe? J Laryngol Otol 1995;109:l166-7.
From the Department of Otorhinolarynagology--Plastic Head and Neck Surgery, St. Anna Hospital, Duisburg, Germany (Dr. Windfuhr), and the Department of Otorhinolaryngology--Plastic Head and Neck Surgery, University of Aachen, Germany (Dr. Chen).
Reprint requests: Jochen P. Windfuhr, MD, Department of Otorhinolaryngology--Plastic Head and Neck Surgery; St. Anna Hospital, 47259 Duisburg, Albertus Magnus Str. 33, Germany. Phone: +49-203-755-1261; fax: +49-203-755-1266; e-mail: firstname.lastname@example.org.
|Printer friendly Cite/link Email Feedback|
|Publication:||Ear, Nose and Throat Journal|
|Date:||Sep 1, 2002|
|Previous Article:||Budgeting with projected payments. (Practice Management Clinic).|
|Next Article:||Intracranial complications of sinusitis: a 15-year review of 39 cases. (Original Article).|