Post-tonsillectomy hemorrhage: Results of a 3-month follow-up. (Original Article).
Episodes of post-tonsillectomy hemorrhage are unpredictable and potentially life-threatening. Primary post-tonsillectomy hemorrhage (<24 hr postoperatively) is generally considered to be more common and more serious than secondary hemorrhage (>24 hr). Therefore, recent studies have focused on the control of primary hemorrhage in order to determine the appropriate length of postoperative observation. The issue of follow-up is becoming more important in light of the increasing popularity of outpatient tonsillectomy. We undertook a prospective study to evaluate the incidence of post-tonsillectomy hemorrhage over the short and long term (3 mo). We studied 602 patients, aged 23 months to 89 years (mean: 20.6 yr), who had undergone inpatient tonsillectomy in 1999 and who had been hospitalized for at least 5 days. Our goal was to ascertain the number of episodes of postoperative hemorrhage that required surgical treatment under general anesthesia. We were able to contact 601 of these patients (or their parents) by t elephone 3 months postoperatively to inquire about any instances of delayed secondary bleeding. In all, 16 patients (2.7%) had experienced post-tonsillectomy bleeding that required surgically achieved hemostasis under general anesthesia. Of this group, 11 patients (68.8%) had experienced primary hemorrhage and were treated immediately, and five (31.3%) experienced secondary hemorrhage. One patient in the latter group experienced excessive bleeding 38 days postoperatively, which we believe is the latest episode of secondary bleeding reported to date. Based on the findings of this study, we believe that a postoperative follow-up period of 10 days is sufficient to identify all but the most rare cases of post-tonsillectomy hemorrhage.
In light of the ever-intensifying emphasis on cost control, increasing numbers of tonsillectomies throughout North America and most of Europe are being performed as outpatient procedures. (1) Exceptions are often made for patients who have important underlying conditions. (2)
Postoperative hemorrhage remains the most serious complication of tonsillectomy, and its incidence approaches 10%. (3) Post-tonsillectomy bleeding events are generally classified as either primary (<24 hr) or secondary (>24 hr). Unfortunately, there is no uniform system of classifying the intensity of post-tonsillectomy hemorrhage. Some authors consider all bleeding events, while others include only those events that require subsequent treatment under general anesthesia. For this study, we included in our results only those patients who met the latter criterion.
Among children, other common short-term complications of tonsillectomy are nausea, vomiting, fever, and an inability to eat or drink. These complications can become serious if they lead to dehydration, which can require postoperative infusion. Even so, some authors consider these complications to be controllable and therefore not a contraindication to outpatient tonsillectomy. (4-6) Nevertheless, post-tonsillectomy hemorrhage remains an important concern. For example, between 1988 and 1998, we treated 65 patients with delayed postoperative hemorrhage in our emergency department after they had undergone tonsillectomy elsewhere. (7) Most of these patients had experienced their hemorrhage more than 1 week following their tonsillectomy.
In this article, we describe our study of the incidence of post-tonsillectomy hemorrhage, with particular attention on delayed hemorrhage. Our goal was to determine the appropriate length of postoperative follow-up, which should help in the postoperative monitoring of patients who undergo tonsillectomy as an outpatient procedure.
Patients and methods
Our study population was made up of 602 patients, aged 23 months to 89 years (mean: 20.6 yr; median: 17.0; 52% female), who had undergone inpatient tonsillectomy at our clinic between January 1999 and January 2000 (figure 1). The indications for tonsillectomy, with or without adenoidectomy, included chronic or recurrent tonsillitis, peritonsillar abscess, removal of tonsil rests or malignant tumors, infectious mononucleosis, and upper airway obstruction caused by hyperplasia of the tonsils and/or neck (figure 2).
All 602 patients underwent surgery under general anesthesia and with oral intubation. A Mclvor gag with a blade was installed, and the tonsils were exposed. The tonsils were grasped with a forceps, the anterior pillars and superior poles were incised with scissors, and the surrounding mucosa was carefully dissected. The plane lateral to the tonsils was dissected, and the forceps were reapplied to grasp the medial and lateral surfaces of the tonsils. The tonsils were bluntly dissected, and the inferior poles were snared.
Any bleeding that occurred after pressure had been applied with gauze was controlled by suture ligation (catgut or Dexon 2-0). No electrosurgical means of achieving hemostasis was required. Patients were instructed to refrain from eating and drinking for 6 hours following surgery. Antibiotics were administered to prevent endocarditis in accordance with current national guidelines in Germany. Following each operation, the surgeon made at least two rounds on the ward per day, and the nursing staff provided continual monitoring, which made possible the early recognition of postoperative hemorrhage.
In accordance with national recommendations, all of these tonsillectomies were performed on an inpatient basis, and all patients remained hospitalized for at least 5 days postoperatively. Patients who experienced postoperative hemorrhage were discharged only after their general condition, hemoglobin concentration, and state of wound healing had stabilized. Two patients who required ligation of the external carotid artery because of a profuse blood loss were discharged only after their wounds had completely healed.
Three months following each surgery, each patient (or parent) was interviewed by telephone to ascertain whether any new bleeding had occurred following discharge and, if so, what kind of hemostasis had been necessary to control it.
In all, 16 patients (2.7%), aged 14 to 58 years (mean: 28.9; median: 28.5; 75% male) experienced postoperative hemorrhage that required surgical treatment under general anesthesia. The episodes occurred between 0 and 38 days postoperatively. Another patient, a 22-year-old man, had a coagulum removed 24 hours after surgery; he experienced no subsequent bleeding. None of the 16 patients required a repeat intervention to achieve hemostasis. An enoral suture was sufficient for 14 patients; 11 patients received their suture on the day of the operation (primary hemorrhage) and one patient each on postoperative day 5, 7, and 9 (secondary hemorrhage). The patient who experienced bleeding on postoperative day 7 had undergone an abscess tonsillectomy and had been readmitted 1 day earlier. The patient who hemorrhaged on day 9 had undergone a tumor tonsillectomy and a radical neck dissection; she had underlying hepatic cirrhosis and had not yet been discharged. The remaining two patients had required an external carotid artery ligation to control their bleeding. One of these patients, who had also undergone radio- and chemotherapy, had experienced profuse bleeding 38 days following his tumor tonsillectomy, and he had referred himself to our emergency department. The other patient had experienced hemorrhage on postoperative day 5; this patient had experienced a significant amount of bleeding during surgery and had required an unusually large number of sutures to control it. Three months after each operation, we attempted to interview each patient (or parent) by telephone, and we were successful in 601 of the 602 cases (99.8%). Using a standard questionnaire, we asked about any incidence of delayed bleeding that we were unaware of, regardless of its intensity. Only two patients reported a new episode of postoperative hemorrhage, and both were minor. A 28-year-old woman had noted blood-tinged sputum on postoperative days 6 and 8, and a 10-year-old boy had experienced some minor bleeding from his nose 10 days after his adenotons illectomy. Both patients were referred to their otolaryngologist, but neither required any further treatment. In all, our telephone interviews uncovered no new case of postoperative hemorrhage that met our criteria for a serious bleed.
The reported incidence of post-tonsillectomy hemorrhage ranges from 0.38 to 6% (8,9) Hemostasis is generally achieved via enoral ligation or electrocautery, but the appropriate management strategy is still controversial. (10-12) Confounding this issue is the variance in published reports regarding specific definitions of postoperative hemorrhage and the process of patient selection. Even though a classification of primary (<24 hr) and secondary (>24 hr) bleeding has been widely established in the literature, there is still no uniform method of quantifying its seriousness. As a result, some authors count all incidents of bleeding and therefore report higher rates of hemorrhage than do authors who include only hemorrhages that require surgical intervention.
Several other limitations in published reports contribute to the lack of a consensus on this issue:
* The incidence of post-tonsillectomy hemorrhage is usually determined in retrospective studies of large numbers of patients. However, some of these published reports did not include adequate information about the study's design, materials, and methods, and therefore they should not be considered for evaluation. (13,14) In other reports, the authors did not specify the age distribution of the study population, and some operations were performed with general anesthesia with ether, which has an influence on vascular tonus and which is no longer used. (15)
* At some clinics, surgeons and/or adult patients prefer that tonsillectomy be performed with local anesthesia. (16) The results of these procedures are not comparable with those reported by pediatric otolaryngologists.
* An impressively low rate of postoperative hemorrhage following the use of bismuth subgallate to achieve hemostasis was reported by Maniglia et al. (17) However, the use of bismuth subgallate in daily practice is not common.
* Many studies were published by experienced surgeons, who are likely to have lower-than-average complication rates. Accordingly, some authors (18) recognized the relationship between surgical experience and the rate of postoperative bleeding, while others (19,20) ignored any such influence.
* An increasing number of reports is being published by pediatric otolaryngologists, and the limited range of their patients' ages means that their results cannot be extrapolated to all patients. (12-26)
* Because most authors today set out to prove that outpatient tonsillectomy is safe, their patient selection criteria are very specific and their study populations are not always reflective of the general patient population. (18,22,24)
* Finally, data on follow-up are frequently missing. (6,8,10,14,16,18,20,21,23,24,26,27,-30) As a result, these reports do not reveal whether patients might have undergone surgical treatment for post-tonsillectomy hemorrhage elsewhere.
Our study was undertaken to assess the incidence of postoperative hemorrhage in nonselected patients who underwent tonsillectomy (performed with only a scissors and snare) under general anesthesia and who were operated on by surgeons with various degrees of experience. When it occurred, post-tonsillectomy hemorrhage was treated exclusively by suture ligation of the bleeding vessels. Only those patients whose postoperative bleeding required surgical treatment under general anesthesia were included in our results. As such, our data seem to be more applicable to the standard practice of general otolaryngology.
Another aim of our study was to analyze the incidence of secondary bleeding, which has not received as much attention in the literature as primary bleeding. The reason for the increased emphasis on primary bleeding is attributable to the belief that it is more common and more serious. (5,14,20,23,30,31) We agree only partly with this opinion; other studies have shown that episodes of delayed bleeding--some as long as 3 weeks following surgery--have also required surgical treatment under general anesthesia. (7,32)
Irani and Berkowitz reported that most cases of secondary hemorrhage they studied occurred within 10 days. (33) Carmody et al reported that 80% of the secondary bleeds they found occurred within 7 days. (11) For our study, we lengthened the duration of follow-up to 3 months. In our opinion, such a long follow-up would certainly reveal all incidents of postoperative hemorrhage. We decided against mailing questionnaires because response rates are frequently poor. For example, Pratt (27) and Pratt and Gallagher (34) attempted to ascertain post-tonsillectomy mortality and morbidity rates by mail, but only 40% of their questionnaires were returned by the selected hospitals. Other reports of response rates range from 31 to 86%. (35) Pringle et al had to mail repeat questionnaires in order to achieve an 80% response. (36) Rosbe et al recently conducted a prospective study of the efficacy of follow-up telephone calls placed 3 to 4 weeks postoperatively and concluded that this method of monitoring is cost-effective. (31)
In our study, the rate of post-tonsillectomy hemorrhage that required treatment under general anesthesia was 2.7% (16/602). Of these patients, 75% (12/16) were male and 75% were older than 20 years of age. These findings are similar to those reported by Roberts et al (10) and by Kristensen and Tveteras, (20) but they are dissimilar to those reported by Carmody et al, (11) who found a significant predominance of post-tonsillectomy hemorrhage among females aged 10 to 19 years.
The rate of primary bleeding in our study--68.8% (11/ 16)-was consistent with several other reports. (12,14,15,20,37) On the other hand, Carmody et al (11) found no difference in primary and secondary bleeding rates; others have even reported that their rates of secondary bleeding were actually higher than the rates of primary bleeding. (8,28,29,38) It should be noted that the latter studies included all episodes of bleeding, including those that did not require surgical treatment under general anesthesia. Moreover, they also included rare episodes of bleeding from the nasopharynx following adenotomy or adenotonsillectomy. Ten patients in our study underwent tonsillectomy to remove a malignant tumor, and two of them experienced secondary bleeding after surgery; if we had excluded these patients from our results, the rate of overall post-tonsillectomy hemorrhage would have fallen to 2.4% (14/592), and the rate of primary hemorrhage would have risen to 78.6% (11/14).
The onset of primary bleeding in our study usually occurred within 6 hours (mean: 5 hr. 46 min). This finding is consistent with those of other studies. (24,28,38,39) Some authors suggest that a duration of postoperative observation of 6 to 8 hours should be safe, and others recommend an even shorter duration, (17,30)
In our study, the onset of secondary bleeding in some patients was both abrupt and excessive, and in two patients it required ligation of the external carotid artery (5 and 38 days postoperatively). There were no episodes of recurrent postoperative hemorrhage in our study.
Although the effort to contact our patients was difficult in some cases and time-consuming overall, we were able to speak to all but one patient or parent (response rate: 99.8%); the sole patient who was not contacted had left the country. We found that only two patients-the 28-year-old woman and the 10-year-old boy-had experienced a minor episode of delayed bleeding that we had not been aware of, and neither had required intervention. Both of these patients had been discharged on postoperative day 6.
During the course of our study, we also treated eight other patients who had come to our emergency department with delayed postoperative hemorrhage after they had undergone tonsillectomy elsewhere. Four of these patients required immediate ligation of the external carotid artery, which was performed 5, 6, 7, and 9 days following their tonsillectomy. With the increasing popularity of outpatient tonsillectomy, emergency departments can expect to see an increase in the number of patients with delayed hemorrhage.
The probability and intensity of post-tonsillectomy hemorrhage appear to be unpredictable. A study of a large number of these cases might better help determine how often delayed bleeding occurs. Most studies feature a short follow-up period and focus on an assessment of risk factors.
To supplement the findings of our 1-year study, we expanded the scope of our population and conducted a companion analysis of the incidence of hemorrhage that required surgical treatment under general anesthesia among 6,076 patients who had undergone tonsillectomy at our institution between January 1988 and January 2000 (figure 3). We found that such bleeding had occurred in 168 of these patients (2.8 %)-virtually the same rate that we found in our 1-year study. Secondary bleeding that had occurred beyond postoperative day 10 was rare. Between 1988 and 1998, only two such cases were observed, both of which occurred in a single patient-once on postoperative day 12 and again on day 18. Overall, eight patients experienced a recurrence of postoperative bleeding, all prior to January 1999.
The patient in our original study who experienced hemorrhage 38 days postoperatively is unique. Before now, the latest recorded occurrence of postoperative hemorrhage in the literature was 31 days. (40)
Because most cases of postoperative bleeding occur within 6 hours of surgery, tonsillectomies should be performed in the early morning. Such a schedule allows the surgeon to observe and treat most, cases of post-tonsillectomy hemorrhage during the day, when the full complement of hospital staff is on duty. Coagulation therapy must be administered with great care. Field cauterization creates deep and extensive zones of necrosis and exposes more and larger vessels to bleeding when sloughing of the eschar occurs.
We conclude that a follow-up period of 10 days appears to provide sufficient time to evaluate the incidence of post-tonsillectomy hemorrhage. Bleeding episodes that occur beyond 10 days are rare. If delayed bleeding should occur with any regularity at all, an analysis of surgeon-specific practices and characteristics is indicated.
(1.) Raymond CA. Study questions safety, economic benefits of outpatient tonsil/adenoid surgery. JAMA 1986;256:311-2.
(2.) Tonsillectomy and adenoidectomy inpatient guidelines. Recommendations of the AAO--HNS Pediatric Otolaryngology Committee. St. Louis: American Academy of Otolaryngology--Head and Neck Surgery, 1996:1-4.
(3.) Randall DA, Hoffer ME. Complications of tonsillectomy and adenoidectomy. Otolaryngol Head Neck Surg 1998;118:61-8.
(4.) Crysdale WS, Russel D. Complications of tonsillectomy and adenoidectomy in 9409 children observed overnight. CMAJ 1986;135:1139-42.
(5.) Gabalski EC, Mattucci KF, Setzen M, Moleski P. Ambulatory tonsillectomy and adenoidectomy. Laryngoscope 1996;106:77-80.
(6.) Mitchell RB, Pereira KD, Friedman NR, Lazar RH. Outpatient adenotonsillectomy. Is it safe in children younger than 3 years? Arch Otolaryngol Head Neck Surg 1997;123:681-3.
(7.) Windfuhr J, Sesterhenn K. Indications and complications of tonsillectomy: Current opinion and clinical study. Presented at the 4th European Congress of Oto-Rhino-Laryngology Head and Neck Surgery; Berlin; May 13-18, 2000.
(8.) Chowdhury K, Tewfik TL, Schloss MD. Post-tonsillectomy and adenoidectomy hemorrhage. J Otolaryngol 1988;17:46-9.
(9.) Kang J, Brodsky L, Danziger I, et al. Coagulation profile as a predictor for post-tonsillectomy and adenoidectomy (T+A) hemorrhage. Int J Pediatr Otorhinolaryngol 1994;28:157-65.
(10.) Roberts C, Jayaramachandran S, Raine CH. A prospective study of factors which may predispose to post-operative tonsillar fossa hnemorrhage. Clin Otolaryngol 1992;17:13-7.
(11.) Carmody D, Vamadevan T, Cooper SM. Post tonsillectomy haemorrhage. J Laryngol Otol 1982;96:635-8.
(12.) Watson MG, Dawes PJ, Samuel PR, et al. A study of haemostasis following tonsillectomy comparing ligatures with diathermy. J Laryngol Otol 1993;107:711-5.
(13.) Chiang TM, Sukis AE, Ross DE. Tonsillectomy performed on an outpatient basis. Report of a series of 40,000 cases performed without a death. Arch Otolaryngol 1968;88:307-l0.
(14.) Lannigan FJ, Martin-Hirsch DP, Basey E. Clinical audit: Is day-case adenotonsillectomy safe? Br J Clin Pract 1993;47:254-5.
(15.) Jakse R. [For prevention of complications, especially postoperative bleeding, in tonsillectomy adenoidectomy]. Laryngol Rhinol Otol (Stuttg) 1981;60:345-50.
(16.) Schmidt H, Schmiz A, Stasche N, Hormann K. [Surgically managed postoperative hemorrhage after tonsillectomy]. Laryngorhinootologie 1996;75:447-54.
(17.) Maniglia AJ, Kushner H, Cozzi L. Adenotonsillectomy. A safe outpatient procedure. Arch Otolaryngol Head Neck Surg 1989;115:92-4.
(18.) Rothschild MA, Catalano P, Biller HF. Ambulatory pediatric tonsillectomy and the identification of high-risk subgroups. Otolaryngol Head Neck Surg 1994;l10:203-l0.
(19.) Breson K, Diepeveen J. Dissection tonsillectomy--complications and follow-up. J Laryngol Otol 1969;83:601-8.
(20.) Kristensen S, Tveteras K. Post-tonsillectomy, haemorrhage. A retrospective study of 1150 operations. Clin Otolaryngol 1984;9:347-50.
(21.) Nicklaus PJ, Herzon FS, Steinle EW IV. Short-stay outpatient tonsillectomy. Arch Otolaryngol Head Neck Surg 1995;121:521-4.
(22.) Richmond KH, Wetmore RF, Baranak CC. Postoperative complications following tonsillectomy and adenoidectomy--who is at risk? Int J Pediatr Otorhinolaryngol 1987;13:117-24.
(23.) 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.
(24.) Shott SR, Myer CM III, Cotton RT. Efficacy of tonsillectomy and adenoidectomy as an outpatient procedure: A preliminary report. Int J Pediatr Otorhinolaryngol 1987;13:157-63.
(25.) Tom LW, DeDio RM, Cohen DE, et at. Is outpatient tonsillectomy appropriate for young children? Laryngoscope 1992;102:277-80.
(26.) Berkowitz RG, Zalzal GH. Tonsillectomy in children under 3 years of age. Arch Otolaryngol Head Neck Surg 1990;116:685-6.
(27.) Pratt LW. Tonsillectomy and adenoidectomy: Mortality and morbidity. Trans Am Acad Ophthalmol Otolaryngol 1970;74:1146-54.
(28.) Carithers JS, Gebhardt DE, Williams JA. Postoperative risks of pediatric tonsilloadenoidectomy. Laryngoscope 1987;97:422-9.
(29.) Colclasure JB, Graham SS. Complications of outpatient tonsillectomy and adenoidectomy: A review of 3,340 cases. Ear Nose Throat J 1990;69:155-60.
(30.) Tewary AK, Curry AR. Same-day tonsillectomy. J Laryngol Otol 1993;107:706-8.
(31.) Rosbe KW, Jones D, Jalisi S, Bray MA. Efficacy of postoperative follow-up telephone calls for patients who underwent adenotonsillectomy. Arch Otolaryngol Head Neck Surg 2000;126:718-21.
(32.) Steketee KG, Reisdorff EJ. Emergency care for posttonsillectomy and postadenoidectomy hemorrhage. Am J Emerg Med 1995;13:518-23.
(33.) Irani DB, Berkowitz RG. Management of secondary hemorrhage following pediatric adenotonsillectomy. Int J Pediatr Otorhinolaryngol 1997;40:115-24.
(34.) Pratt LW, Gallagher RA. Tonsillectomy and adenoidectomy: Incidence and mortality, 1968-72. Otolaryngol Head Neck Surg 1979;87:159-66.
(35.) Livesey JR. Are haematological tests warranted prior to tonsillectomy? J Laryngol Otolaryngol 1993;107:205-7.
(36.) Pringle MB, Cosford E, Beasley P, Brightwell AP. Day-case tonsillectomy--is it appropriate? Clin Otol 1996;21:504-11.
(37.) Peeters A, Van Rompaey D, Schmelzer B, et at. Tonsillectomy and adenotomy as a one day procedure? Acta Otorhinolaryngol Belg 1999;53:91-7.
(38.) Helmus C, Grin M, Westfall R. Same-day-stay adenotonsillectomy. Laryngoscope 1990;100:593-6.
(39.) Reiner SA, Sawyer WP, Clark KF, Wood MW. Safety of outpatient tonsillectomy and adenoidectomy. Otolaryngol Head Neck Surg 1990;102:161-8.
(40.) Dey FL. Late hemorrhage following tonsillectomy. Arch Otolaryngol 1968;87:558.
[Figure 1 omitted]
Figure 2. Chart illustrates the number (%) of each type of procedure that was performed in study of the 602 patients. Most patients were treated for chronic or recurrent tonsillitis, symptoms of upper airway obstruction, or peritonsilar abscess. Less common indications were the removal of a malignant tumor of the tonsils or tonsil rests and the treatment of infectious mononucleosis. Tx of Infectious mononucleosis n = 3 (0.5%) Removal of tonsil rests n = 15 (2.5%) Tumor tonsillectomy n = 10 (1.7%) Immediate abscess tonsillectonmy n = 142 (23.6%) Tonsillectomy n = 246 (40.9%) Adenotonsillectomy n = 186 (30.9%) Note: Table made from pie chart
[Figure 3 omitted]