Difficult Intubation And Post Tonsillectomy Bleeding

A six year old boy with Down’s syndrome is on his way to the children’s hospital by ambulance with post-tonsillectomy bleeding.

He underwent adeno-tonsillectomy because of recurrent tonsillitis and enlarged adenoids under general anesthesia the day before, some 22 hours ago. Despite being overweight at 37 kilograms and enlarged adenoids he did not suffer from sleep apnea. Prior to his original surgery, the child was uncooperative necessitating an inhalation induction with some struggling. Venous access was difficult even post induction requiring several attempts, and finally being achieved in the left saphenous vein at the ankle. Because of possible atlanto-occipital instability associated with Down’s syndrome, laryngoscopy was performed with C-spine precautions. Direct laryngoscopy presented a Grade 3 view due to an enlarged tongue. Bag-mask-ventilation with an oro-pharyngeal airway was easy throughout the preintubation phase. Indirect laryngoscopy using the GlideScope® revealed a Grade 1 view followed by the placement of a styletted, uncuffed 5 mm ID oral RAE tube. Adeno-tonsillectomy was performed in the usual fashion and the child was discharged home after an uneventful 20 hour overnight observation period.

Apparently, while momentarily unattended at home, the boy ate a hard tea biscuit. The child immediately experienced a sharp pain and an intra oral bleeding started.

The emergency physician on duty is confronted with an overweight boy, sitting on a stretcher and spitting blood frequently into a kidney basin. The child is in moderate distress with the following vital signs (HR 152, BP 97/57). The child will not tolerate nasal prong oxygen and the pulse oximeter reading is 94% on room air. Auscultation of the chest is clear. Examination of the mouth reveals brisk bleeding in the right tonsillar bed. An attempt to start an intravenous line in the right saphenous vein is not successful, but blood is obtained for a CBC, coagulation parameters and a cross match for blood. The child is then transferred to the operating room.

45.2 INTRODUCTION:

45.2.1 What is the incidence, Morbidity and Mortality of pediatric post-tonsillectomy bleeding?

Tonsillectomy is one of the most frequently performed surgical procedures in children. Rates in children aged 0-14 vary considerably within and between countries. In 1998, they varied from 19 per 10,000 children in Canada to 118 per 10,000 in Northern Ireland, so a very common procedure in both countries (1).

The most common post-tonsillectomy complications include postoperative nausea and vomiting and pain. Dehydration may occur in children due to delayed and poor oral intake, nausea and fever. Delayed post-operative bleeding is the most significant complication and though uncommon, is not rare (2). Many estimates of the incidence of post-tonsillectomy bleeding exist in the literature varying widely from 0 to 11.5 percent (3). Typically, however, the rate ranges between 2.9 and 3.4% (4). Mortality rates are rarely reported in the literature. Two large studies reported 0 out of 15,996, and 1 out of 16,381 tonsillectomies in 1979 and 1970 (5). On the other hand, there are many published case reports.

Sixty-seven percent of post-adeno-tonsillectomy bleeding originates in the tonsillar fossa and 27% in the nasopharynx. There are two major time frames for post operative bleeding. Most often, the bleeding occurs within the first 24 hours after surgery (primary bleeding) (5). Primary bleeding is generally related to surgical technique, and incidence is declining. Twenty five percent of all post tonsillectomy hemorrhage occurs after 24 hours. This secondary bleeding is not related to surgical technique, is rare and of unchanged prevalence over the years (5). It is mainly observed between the 5th and 10th postoperative day, although it may occur at any time (6). Infection of the tonsillar bed with clot sloughing is believed to be the major cause of secondary bleeding. It tends to occur more commonly in older pediatric patients, because the indication for tonsillectomy in this age group is usually related to recurrent infections rather than airway obstruction, the most common indication in the younger pediatric age group.(5).

Since tonsillectomy is usually performed to improve the quality of life in otherwise healthy, young children, any death is unacceptable.

45.3 PATIENT EVALUATION:

45.3.1 What are the initial clinical steps one should take in the patient with post-tonsillectomy bleeding?

The diagnosis of post-tonsillectomy bleeding is usually made by a quick history. Parents or patients will mention right away the previous surgery. Differential diagnosis is blunt or sharp trauma to the oropharynx. Rare cases are bleeding tumors of the oropharynx, like hemangioma.

The child will present with fresh blood in the mouth and frequent swallowing of blood. Nausea with or without emesis of fresh blood is common. Newer and more potent antiemetic medications may mask or suppress vomiting. Therefore, the amount of blood swallowed may be underestimated. It is not uncommon for children to have been bleeding silently for a prolonged period of time with extensive blood loss.

The child is often restless, diaphoretic and pale. The vital signs may show an increased heart rate, because of pain and hypovolemia. In awake children, hypotension following blood loss is a very late sign and then indicates significant hypovolemia. Intravenous access must be established as soon as possible followed by initial volume resuscitation with crystalloid or colloid solutions or. A blood sample for baseline hematocrit or hemoglobin is necessary as well as for blood type and cross match.

An intra oral examination will show blood and blood clots. A bleeding source may be seen in the tonsillar bed.

Bleeding from the tonsillar bed may initially be controlled using pharyngeal packs and cautery. But children with post tonsillectomy bleeding should be taken back to the operating room for exploration and surgical hemostasis. Repeated attempts to stop bleeding on the ward or in the emergency department should be avoided, except if exsanguination is imminent.

A questionnaire of children undergoing tonsillectomy with or without postoperative bleeding showed an increased incidence of post-traumatic stress disorder if the children with bleeding were treated on the ward compared to children without bleeding or if the bleeding was treated in the operating room (7).

45.4 AIRWAY MANAGEMENT

45.4.1 How is the airway usually managed in post-tonsillectomy bleeding?

Large volumes of blood may be swallowed, and blood or blood clots are often present in the oral cavity of these children. Despite the fact that the aspiration of blood is not similar in severity to aspiration of gastric acid, it remains an undesirable occurrence.

In addition to hypovolemia, patients with post tonsillectomy bleeding present two major problems:

Aspiration: These patients must be considered to have a full stomach and are at an increased risk of aspiration.

Difficult airway: Blood and blood clots may impair visualization to the vocal cords. In addition swelling of the oropharynx may have occurred because of surgery or infection. This may lead to a changed laryngeal anatomy.

Because of the risk of aspiration, a mask induction maintaining spontaneous breathing is not desirable and a rapid sequence induction should be considered. The efficacy and use of cricoid pressure, especially in children is currently controversial. It is noteworthy that cricoid pressure can distort the laryngeal anatomy and worsen the view of the larynx. In addition, it can induce vomiting in the partially anesthetized patient.

The blood and blood clots in the oropharynx can impair vision during laryngoscopy or cause plugging of the endotracheal tube. A working suction apparatus is lifesaving and must be prepared in duplicate. One should be a large bore, rigid surgical suction and the other mounted with a flexible endotracheal suction catheter. If one becomes blocked with a blood clot, another is readily available. If large amounts of clot are present, it may be necessary during the initial laryngoscopy to manually remove them with a finger or gauze. A Magill forceps should be available to grab clots deeper in the pharynx, recognizing that these clots may be too fragile to be grasped and removed from the oral cavity using the Forceps.

A past history of difficult laryngoscopy is helpful, although this never precludes preparations for a difficult and failed airway. Different sized curved and straight blades as well as a flex tip blade (McCoy laryngoscope) should be readily available. Different sized cuffed endotracheal tubes, with one size up and down of the calculated size must be prepared. They should be preloaded with a well lubricated intubating stylet, as is standard for a rapid sequence induction.

A tracheal introducer (“Bougie”) may be helpful in the presence of a grade 3 view. If the epiglottis is visible, but no laryngeal entrance can be appreciated, a stroke of chest compression may help find the glottic opening by creating air bubbles.

The pediatric lightwand represents an elegant technique for intubation in the case of a glottic view obscured by secretions or blood. The extremely bright light can shine easily through blood and blood clots. However experience is necessary when using this device.

Indirect laryngoscopy using the video-laryngoscope (GlideScope® or the Airtraq®) can be difficult. Blood and secretions may block the optical lenses and impair the view to the vocal cords. The lens in the Airtraq® with its position between the lightsource on one side and the guide channel for the endotracheal tube on the other side might be more protected than the lens of the GlideScope®. Case reports or studies, however, have not been published in this regard.

The laryngeal mask plays an accepted role as an alternative airway device in managing the difficult pediatric airway (see Chapter 42). It is used frequently in primary adeno-tonsillectomies. It can be placed quickly and can be used as a conduit for a flexible bronchoscope to guide intubation if required. On the positive side, a laryngeal mask may briefly tamponade the bleeding site, and therefore protect the airway and the optical lens of the bronchoscope. Though, on the other hand, it may not provide sufficient airway protection in situations with increased risk of aspiration like post tonsillectomy bleeding. A case report recently described the successful use of a laryngeal mask for a failed intubation in a post tonsillectomy bleed (8).

The use of a flexible bronchoscope alone is not recommended in cases of oropharyngeal bleeding. Experts recommend that the practitioner should rely on the alternative techniques with which they have the most experience and skill.

Preparation for the unexpected is essential. An experienced otolaryngologist or other qualified rigid laryngoscopist/bronchoscopist should be in the OR for all of these cases. If direct laryngoscopy fails, a rigid device wielded by the otolaryngologist may just be successful. An appropriately sized, lubricated and tested rigid laryngoscope/bronchoscope connected to a light source and suction must be readily available at the head of the child. Preparation for a surgical airway is also essential (e.g. tracheotomy tray opened and ready).

To reduce the risk of postoperative nausea and vomiting the stomach content of the child should be suctioned using an orogastric tube at the end of the procedure, recognizing that this does not guarantee an empty stomach as much of the blood may be clotted.

45.4.2 What are the airway management options for this patient?

This patient presents several issues regarding anesthesia induction and airway management:

high risk of aspiration

difficult intravenous access

suspected atlanto-occipital instability

known difficult direct laryngoscopy with easy facemask ventilation

expected difficult view of the larynx due to blood and secretions

Several options for the anesthetic and airway management of this child need to be weighed and considered in light of their risks and benefits:

Intravenous induction versus inhalation induction without IV access:

This child is undergoing a second surgical procedure within 24 hours. Due to the frightening emergency situation, pain, bleeding and his mental impairment he is distressed and uncooperative. While a smooth inhalation induction with a facemask was preferred for his first surgery, a stomach potentially full of blood mitigates against this approach and for a rapid sequence induction to minimize the duration of an unprotected airway. One might even hope for a rapid venous access following a mask induction to permit medication administration but we know in this case that is not likely.

Anesthesia induction with cervical spine precautions versus no C-spine precautions:

Down’s syndrome is associated with atlanto-occipital instability in up to 20% of cases. It can occur in children as young as 4 years of age. The large head of the Down’s child coupled with this atlanto-occipital instability places them at increased risk for cervical spine injury during anesthetic induction. Radiographic findings of cervical spinal instability in Down’s syndrome remain controversial. Lateral radiographs of the neck in flexion and extension do not reliably detect atlanto-occipital instability. Due to impaired cognition and anxiety, positioning of the patient can be difficult. Old lateral neck radiographs are not available for this child. Due to the emergency situation, a current neck radiograph is not possible. Therefore cervical spine precautions should be performed. Extreme neck extension should be avoided in this child.

Awake tracheotomy versus anesthesia induction with attempted laryngoscopy:

The fact that this child has a known difficult direct laryngoscopy together with a documented Grade 3 view favors an awake tracheotomy under local anesthesia. This approach would maintain a protected airway at all times. Awake tracheotomy in adults and children are challenging. Optimal surgical positioning with neck extension is crucial for successful procedure. It is not expected, that this child will tolerate this procedure. This fact, together with the required c-spine precautions would exclude an awake tracheotomy as an option for this child.

The plan is to perform an intravenous rapid sequence induction employing indirect laryngoscopy to place an endotracheal tube. Preparations for rigid laryngoscopy are in place and the surgeon is prepared to embark immediately with a surgical airway (in this case a ‘triple set up’).

45.4.3 How should you prepare for this case?

Following the failed attempt to start an intravenous line in the emergency department, the child was brought to the operating room. As previously outlined, venous access is crucial for induction and fluid resuscitation. Placement of a central line in the awake child is a possible option. For internal jugular vein access, the head may need to be rotated with increased risk associated with the presumed atlanto-occipital instability. The subclavian approach has the risk of a pneumothorax. An ultrasound guided femoral vein approach is an alternative.

On the other hand, several studies have shown that an intraosseos cannula can be placed within 60 seconds and that this line provides an excellent access for the administration of medications and fluids. Because of the risks associated with central line placement, the child was prepared for an intraosseous cannula. The right leg was prepped with antiseptic solution, and local anesthetic injected at the tibial plateau. An intraosseous canula was placed without incident. A normal saline solution flowed freely permitting the administration of 20ml.kg-1 bodyweight.

Atropine 0.1 mg IV was administered to reduce additional secretions and mitigate vagal responses secondary to laryngoscopy. The usual monitors were applied (pulse oximetry, noninvasive blood pressure and EKG).

The surgeon was prepared as was his equipment; the rescue airway cart was in the room.

45.4.4 Management of this child

Concurrent with the placement of the intraosseous canula, the child was prepared for a rapid sequence induction. The child was denotrogenated with 100% FiO2 for 3 minutes employing a facemask that was reasonably tolerated with much cajoling. Considering the possibility of significant hypovolemia a 50/50 mix of Ketamine and Propofol (‘ketofol’) was selected for induction and Succinylcholine for neuromuscular blockade. Cricoid pressure during induction was not applied to avoid stimulating vomiting in the already agitated child. It was applied after the child was deeply anesthetized. As soon as the child was deeply asleep and paralyzed, the mouth was suctioned easily and several clots were removed with the Magill Forceps. Again the brisk bleeding from the right tonsillar bed was noted.

Since the previous direct laryngoscopy showed a Grade 3 view, a repeated direct laryngoscopy was not attempted. Because the oral cavity seemed to be free of clot, it was decided to proceed with indirect laryngoscopy with the Glidescope®. Unfortunately, blood obscured the lens and following a prolonged laryngoscopy the attempt to intubate was abandoned. Oxygen saturations fell from 100% to 94% and despite the risk of aspiration, mask ventilation was begun and cricoid pressure was maintained.The saturations recovered nicely.

At this point, faced with a failed intubation, rather than a failed airway it was decided to insert a ProSeal® laryngeal mask. The oral cavity was once again suctioned with a rigid catheter under direct laryngoscopy and a number 3.0 Proseal® laryngeal mask was easily placed. No air leak was noted and pressure controlled ventilation with a pressure limit of 15 cm H2O was started. A number 5.5 uncuffed endotracheal tube was loaded on a pediatric flexible bronchoscope. Using the LMA ProSeal® as a conduit, the bronchoscope was advanced into the trachea. Blood and secretions where present in the LMA and in the trachea but did not obscure the view through the bronchoscope. The ETT was placed easily. With a small air leak at 20 cm H2O airway pressure, it was decided not to change the ETT over a pediatric Cook airway exchanger to a cuffed ETT. Since the LMA did not obscure the surgeons view, it was decided to leave the LMA in place and remove it together with the endotracheal tube at the end of the procedure after emerge of anesthesia.

With a secured airway, the ENT surgeon cauterized the tonsillar bed, and the bleeding artery could be ligated.

At the end of the procedure, a large bore nasogastric tube was placed through the suction port of the Proseal® laryngeal mask and the stomach suctioned. The child was taken to the pediatric ICU where he was extubated fully awake one hour later.

45.5 OTHER CONSIDERATIONS

45.5 1 What is the current thinking with respect to the surgical management of post-tonsillectomy bleeding?

Life threatening post-tonsillectomy bleeding requires an aggressive approach to surgical management. Initially pressure on the bleeding tonsillar fossa with a clamped gauze or the index finger may give sufficient time to start an intravenous line for blood work and cross match, and to provide for fluid resuscitation or blood transfusion if indicated.

If intraoperative localization of the bleeding source is time consuming and local treatment is ineffective, ligation of the external carotid artery at an early stage may be required. Aberrant arterial blood supply to the tonsillar region deriving from the internal carotid artery or the carotid bulb may be present. In cases such as these, packing of the pharynx and angiographic embolization of the feeding artery may be necessary(9).

45.5.2 Are there specific measures that one ought to employ to reduce the postoperative morbidity and mortality of patients following tonsillectomy?

The focus on post-tonsillectomy bleeding is on preventive measures, both by the surgeon and the anesthesia practitioner.

Tonsillectomy technique:

In comparison to the cold knife technique, hot techniques employing bipolar diathermy or ‘coblation’ tonsillectomy are associated with an increased rate of secondary bleeding (4, 9). The duration, frequency and surgical extent of these techniques are linked to the amount of damage to the surrounding tissue. This damage leads to deeper zones of local necrosis which is vulnerable to bacteria- and enzyme-containing saliva, and therefore at increased risk of secondary bleeding (4, 5).

Effects of postoperative, nonsteroidal, antiinflammatory drugs:

Nonsteroidal anti-inflammatory drugs (NSAIDs) inhibit platelet cyclo-oxygenase (COX). A recent meta-analysis showed an increased risk of re-operation for hemostasis post-tonsillectomy if conventional NSAIDs such as ketorolac, ibuprofen or ketoprofen were used for postoperative pain control in children (10). On the other hand ketorolac has been proven to be an effective treatment for post tonsillectomy pain, and as a non-opioid delivers an intraoperative opioid sparing effect and leads to a reduction in postoperative respiratory depression, nausea and vomiting. A most recent meta-analysis did not find an altered number of perioperative bleeding events in patients given an NSAID (11). Still, the use of these drugs should be discussed with the surgeon and used with precautions.

Effects of dexamethasone for postoperative nausea and vomiting (PONV):

Postoperative nausea and vomiting increases the risk of primary hemorrhage and unexpected postoperative hospital admission. Dexamethasone has antiemetic properties in the perioperative setting. However, dexamethasone may inhibit wound healing, attenuate the inflammatory response to local infection and as a result perhaps increase the risk of postoperative bleeding. A recent study in children undergoing tonsillectomy and administered dexamethasone was prematurely terminated because of an increased bleeding rate (12). Similar to NSAIDs, the use of dexamethasone should be discussed with the surgeon und used with caution.

Tonsillectomy as outpatient surgery:

Traditionally, tonsillectomy has been associated with a hospital inpatient admission. Economic imperatives have pushed hospitals to perform tonsillectomies as outpatient day surgery procedures.

The evidence has shown that this can be safely performed with the following exceptions:

age under 3 years

medical disorders that increase anesthetic and surgical risk

craniofacial abnormalities

abnormal coagulation, with or without an identifiable bleeding disorder

obstructive sleep apnea

acute peritonsillar abscess

family conditions that prevent easy and rapid return to a medical facility

Patients should always be observed for a minimum of 6 hours. They should be able to tolerate oral fluids and be pain free of prior to discharge. As an alternative to hospital admission, a 23 hour overnight observation period can be considered.

45.6 SUMMARY

Post-tonsillectomy bleeding is a rare event, which occurs most often within 24 hours following tonsillectomy. However, it may be delayed for up to 14 days postoperatively. The amount and severity of bleeding along with the need to ensure patient comfort and a still surgical field most often make operative revision under general anesthesia necessary. The insidious and continuous nature of the bleeding may lead to significant hypovolemia which is often difficult to assess. Blood work and cross match as well as preoperative intravenous access with fluid resuscitation are crucial in the management of these children.

Aspiration and a difficult airway are ever-present risks during the induction of anesthesia in these patients. Rapid sequence induction with direct laryngoscopy and endotracheal intubation is the accepted first choice in the management of these children.

An array of pediatric airway management devices need to be immediately available. Blood and secretions can obscure the laryngeal view and can make some devices more useful than others. A surgeon experienced in rigid broncoscopy and establishment of a surgical airway must be present during anesthesia induction. Cautious use of nonsteroidal inflammatory drugs and steroids is advocated.

Any death related to a tonsillectomy is a unacceptable bleeding in otherwise healthy children.