Epiglottitis: Avoiding Pitfalls in Diagnosis and Management
By Glenn C. Freas, MD, JD, FACEP, Associate Program Director for Emergency medicine Residency, Allegheny University Hospitals, Philadelphia, PA
Editor’s Note: Acute epiglottitis remains a serious airway emergency despite two encouraging points: first, the incidence of the disease has decreased in children in the last 10-12 years;1 and second, epiglottitis in adults does not always result in upper airway obstruction and respiratory failure.2 Nonetheless, the potential for catastrophic results in children and adults (death or anoxic encephalopathy) continues to be very real. When these catastrophes lead to malpractice litigation, the awards for damages can be in the millions, with one recent case resulting in damages in excess of $11,000,000.3
Emergency physicians are frequently challenged with deciding which of the many patients we see with sore throats may have supraglottic airway inflammation. Less frequently, we may be called upon to intervene when epiglottitis causes complete or impending airway obstruction. As with all problems in emergency medicine, we rely on the medical literature to guide us in the management of patients who may have epiglottitis. Prior to the mid 1980s, the medical literature on epiglottitis consisted mostly of case reports or small case series that did not reflect today’s diagnostic techniques like fiberoptic nasopharyngoscopy. Furthermore, these older case reports frequently discussed the use of "emergency" tracheostomy in patients with airway obstruction and respiratory failure, but did not discuss the more accepted emergent surgical airway techniques currently employed by emergency physicians like cricothyrotomy, retrograde intubation, or transtracheal jet ventilation.4 In addition, some references tried to extrapolate the pediatric experience with acute epiglottitis to the diagnosis and treatment of adults with epiglottitis. In the last decade, more extensive reviews involving larger case series have yielded meaningful recommendations on the ED diagnosis and treatment of acute epiglottitis in both the pediatric and adult populations. While controversies still exist, there are some specific features of the disease that have significant implications for successfully identifying which patients may have epiglottitis and selecting those patients who are more likely to require urgent airway control.
This article begins with an overview of the recent medical literature concerning epiglottitis. The malpractice cases presented thereafter underscore areas of consensus as well as areas of controversy about the diagnosis and management of patients with acute epiglottitis. Important questions will be addressed, including: 1) Is an oropharyngeal exam with a tongue depressor contraindicated in all patients with suspected epiglottitis? 2) Are there identifiable subsets of patients with epiglottitis who do not need to have urgent airway control? 3) If a patient has a bad sore throat but has evidence of pharyngitis or tonsillitis, can the emergency physician rest comfortably that those findings are the only cause of the sore throat, or must they still worry about epiglottitis? 4) What is the best way to control the airway of a patient who has had complete obstruction secondary to epiglottitis? Finally, some of these cases serve as reminders about risk management strategies that can be broadly applied to other emergent conditions in addition to epiglottitis. Despite the potentially lethal nature of this disease process and the challenge of emergently controlling a distorted airway, a heightened awareness of the pitfalls in the evaluation and stabilization of these patients should decrease the morbidity and mortality associated with epiglottitis.
Overview of the Medical Literature
Incidence, Pathophysiology, and Bacteriology
Prior to the 1940s, acute epiglottitis was considered an adult disease. From the 1940s through the 1970s, epiglottitis was thought to be almost exclusively a disease of children. Today the incidence of epiglottitis in adults is now higher than that in children. One study showed the incidence in adults staying constant at 1.8 per 100,000 from 1980-1990, while the incidence in children dropped from 3.47 per 100,000 to 0.6 per 100,000 during the same time period.1 The decreased incidence in children is widely attributed to the introduction of the vaccine against Haemophilus influenzae type b (Hib); however, the decrease in the incidence of epiglottitis in children apparently started as early as 1984. The vaccine was not available until 1985 and was not given to infants beginning at 2 months of age until 1990.5 The mean age of adult patients with epiglottitis is close to 40 years old,6 while the mean age in children since the widespread availability of the Hib vaccine appears to have increased from 3 years to over 6 years of age.5 It should be emphasized that these are mean ages, and the potential for epiglottitis in a wide range of ages is illustrated by some of the malpractice cases below.
The pathophysiology of epiglottitis is best described by the suggestion that the disease be renamed supraglottitis. Any or all of the supraglottic airway can be involved, including the uvula, base of the tongue and lingual tonsils, aryepiglottic folds, epiglottis, and false vocal cords. The variability in presenting signs and symptoms and degree of respiratory distress may depend on the anatomic areas involved in the inflammatory process. It has been asserted that the pathophysiology in adults is different than that in children, in that adults are more likely to have diffuse supraglottic involvement while children are more likely to have discrete involvement of the epiglottis.7 Adults are less likely than children to progress to complete airway obstruction because the adult supraglottic airway is larger, the surrounding structures are more rigid, and there is less reactive lymphoid tissue. Despite these more favorable anatomic features, adults can precipitously occlude their airway just as children can.
The mechanism of airway obstruction is unclear. There are clearly cases of progressive airway obstruction manifest by classic signs of impending respiratory failure: increased retractions, stridor, tachycardia, changes in mental status, cyanosis, and, finally, apnea.7 However, there are case reports of sudden respiratory arrest with a paucity of prodromal symptoms or signs of respiratory distress.6 This has led commentators to speculate that the swollen epiglottis can get suddenly lodged in the glottic opening, like a "ball-valve" mechanism.8 This proposed mechanism explains the general caution about stimulating patients, particularly children, with epiglottitis and the specific caution about tongue depressor exams. The rationale is that making the patient gag or inhale suddenly may precipitate this ball-valve type obstruction. Another theory about the cause of sudden airway obstruction is that patients may aspirate pooled secretions (because of pain and inability to swallow them), causing sudden laryngospasm. The same caution about stimulating patients would theoretically apply to this proposed mechanism as well.
The bacteriology of the disease is poorly studied in adults, but somewhat better described in children. Since the introduction of the Hib vaccine, the proportion of pediatric cases of epiglottitis caused by H. influenzae has dramatically decreased. Pathogens in children include H. influenzae and group A beta-hemolytic Streptococcus. A significant percentage of pediatric patients have negative cultures of blood and the pharynx.5 In adults, pharyngeal cultures do not uniformly demonstrate a pathogen. Blood cultures are obtained in only about one-third of cases; most are negative. While some commentators maintain that H. influenzae is the most commonly identified pathogen in adults,6 others dispute this.7 It appears that when a pathogen is identified in adults, it is most likely to be H. influenzae, group A beta-hemolytic Streptococcus, non-group A hemolytic Streptococcus, and Streptococcus pneumoniae.1,9 When H. influenzae is the pathogen in adults, particularly when H. influenzae bacteremia is documented, a more fulminant course can be expected.6,10
Clinical Features and Diagnosis
Almost all patients with epiglottitis complain of sore throat, odynophagia, and dysphagia; however, these are very nonspecific symptoms. Other more specific signs and symptoms occur less commonly in adults: drooling, 30%; maintaining an erect sniffing position, 15%; stridor, 12%; and dysphonia, 50%.1 It is estimated that 25-30% of adult patients with epiglottitis will present with some signs or symptoms of respiratory distress: dyspnea or orthopnea, inability to handle secretions, stridor, tachypnea, retractions, maintenance of an erect sitting position, and/or cyanosis.2 As discussed below, the presence of any signs or symptoms of respiratory distress in adults has prognostic significance. Specific signs and symptoms of epiglottitis occur much more frequently in children, and the diagnosis should be relatively easy when these are present: respiratory distress, 96%; drooling, 60%; stridor, 82%; and toxic appearance, 60%.5
Diagnostic options include indirect laryngoscopy, soft tissue radiographs of the neck, and fiberoptic nasopharyngoscopy. Controversies center around the reliability of plain radiographs and the advisability of indirect or direct visualization of the upper airway in patients with suspected epiglottitis. The threshold question should be whether the patient needs any diagnostic studies at all. As suggested above, the diagnosis of epiglottitis should be more straightforward in children based upon the history and physical. Children are more likely to present toxic and in respiratory distress. There is a concern that unnecessary stimulation (IVs, rectal temperatures, tongue depressor exam) may lead to airway obstruction, and children almost always require tracheal intubation when they have epiglottitis.5 For these reasons, it appears more prudent to control the pediatric patient’s airway without diagnostic studies when the diagnosis is highly suspected.
Adults with suspected epiglottitis are much more likely to be stable for diagnostic studies. Plain film radiography is widely recommended for stable adults and for children in whom the diagnosis is less clear and there is no respiratory distress or toxic appearance. It is dogma in emergency medicine to send these patients to radiology only if they are accompanied by the proper airway equipment and personnel. Objective criteria have been proposed to help diagnose epiglottitis radiographically,11 but they have not been prospectively shown to be superior to subjective criteria, which are: ballooning of the hypopharynx, "thumb print" epiglottis, increased size of aryepiglottic folds, a narrowed tracheal air column, and obliteration of the vallecula.12 There are those who oppose the use of radiographs because there is an unacceptably high number of false negatives and because indirect or direct laryngoscopy are widely available and superior in diagnostic accuracy.6,10
Indirect visualization, fiberoptic nasopharyngoscopy, and tongue depressor exams all appear safe in adults who are not in severe respiratory distress.6 Assuming adequate visualization of supraglottic structures with indirect or direct techniques, it appears reasonable to expect close to 100% diagnostic accuracy.2
Treatment
The overwhelming majority of children in whom epiglottitis is highly suspected or confirmed will need tracheal intubation. A coordinated approach including transport to the operating room (OR) for orotracheal intubation with surgical airway standby is preferred, when time allows. When a child has acute airway obstruction in the ED, it is possible to use bag-valve-mask ventilation to temporarily ventilate. Orotracheal intubation with a tube one size smaller than ordinarily used is advised. Intravenous access can be established after the airway is controlled. If the child is septic, fluid resuscitation may be necessary. A third-generation cephalosporin should be adequate for initial antibiotic coverage.
Treatment of adults is somewhat controversial, although treatment of patients at either end of the spectrum of stability is more clear-cut. Stable patients with no respiratory complaints or signs of respiratory distress can be safely admitted to the intensive care unit without intubation.14 Patients with clear signs of respiratory distressrespiratory rate greater than 30/min, stridor, retractions, cyanosis, or altered level of consciousnessrequire immediate intubation. If time allows and the proper personnel are available, transport to the OR, induction with inhalation anesthesia (to preserve respiratory reflexes), and orotracheal intubation with bronchoscopy and surgical airway standby is preferred.2 If the patient cannot be transported to the OR, the technique for controlling the airway in the ED depends on the patient’s status. If he or she is still breathing and able to cooperate, nasotracheal intubation, with or without the aid of a nasopharyngoscope, is an option.6 Although not discussed in the literature, there is a theoretical concern that administration of neuromuscular blockade agents (NMBs) may create a completely occluded airway where preoxygenation may be impossible. That is, preservation of the rigid support structures of the adult airway may be sacrificed with NMBs. This theoretical concern has not been studied. If patients are apneic, orotracheal intubation will be necessary. Surgical airway techniques are discussed in the context of case #1 and case #6 below.
Treatment of adults with minimal signs of respiratory distress (respiratory rate 20-30 with no retractions, stridor, or maintenance of an upright sniffing position) but with some respiratory complaints (dyspnea or orthopnea) is more controversial. Factors to consider include the duration of symptoms at the time of presentation and some specific signs thought to be predictive of airway obstruction. Treatment of this controversial group of patients will be discussed in the context of case #5 below.
All adult patients who are admitted with epiglottitis who do not require urgent airway control should be admitted to an intensive care unit and given cool humidified oxygen, intravenous antibiotics (a third-generation cephalosporin), and adequate analgesia. Patients who are observed in this manner who exhibit any increasing respiratory distress should have their airway controlled.
Malpractice Cases Involving Epiglottitis
Case #1: Rapid Deterioration from "Exudative Pharyngitis"
A 55-year-old man went to an ED with complaints of sore throat for four days with a muffled voice for one day. On exam he had a temperature of 102°F and a red pharynx with exudate. Despite the dysphonia, there is no evidence that the patient had any signs of respiratory distress. He was given an injection of penicillin. Before he could leave the ED, he developed respiratory distress. The emergency physician was unable to intubate due to extensive supraglottic swelling. By the time an emergency tracheostomy was performed, the patient sustained extensive anoxic encephalopathic damage. The allergist consulted in the case said that there was evidence of infectious epiglottitis and that it was less likely that the respiratory arrest was due to an allergic reaction to the penicillin. The emergency physician was sued for failure to diagnose epiglottitis and failure to control the patient’s airway in a timely way. The defense position was that once the emergency physician found evidence of pharyngitis, he was under no obligation to pursue the diagnosis of epiglottitis. The jury found for the plaintiff and awarded $11,400,000. The court adjusted the award to $14,534,908 after prejudgment interest was added.3
This case raises three important questions. First, does evidence of tonsillitis or pharyngitis preclude the diagnosis of epiglottitis? Second, what are the options for controlling the airway in a patient with epiglottitis who is in respiratory arrest? Third, how commonly do patients with epiglottitis progress from no symptoms or signs of respiratory distress to complete airway obstruction without warning?
Classic teaching is that when an adult has a severe sore throat but has a paucity of pharyngeal findings, epiglottitis should be suspected. That begs the more pressing concern that epiglottitis can be present even when there are findings of pharyngitis or tonsillitis. In one series, 17 of 21 adult patients with epiglottitis had evidence of mild-to-moderate pharyngitis when they were examined with a tongue depressor.2 In another series, 44% of adult patients with epiglottitis had evidence of pharyngitis.1 It is clear that when a patient has enough presenting signs or symptoms consistent with epiglottitis, evidence of pharyngitis should not prevent the emergency physician from further pursuing the diagnosis of epiglottitis. In this case, dysphonia was the only other evidence of supraglottic inflammation. Uncomplicated pharyngitis does not commonly cause a muffled voice. (The importance of this sign will be discussed in the context of case #6.
The difficulty of endotracheal intubation in patients with epiglottitis is well-described.6 Even using a tube one size smaller than normal will not be helpful when the distortion of the airway from severe supraglottic edema precludes visualization of the glottic opening. Surgical techniques are frequently necessary. In this case, by the time an emergency tracheostomy was performed, the patient suffered irreversible anoxic brain damage. Even in skilled surgical hands, a tracheostomy is a technically difficult procedure that usually takes more time than it takes to develop anoxic brain damage. Emergency physicians should be skilled in other, less time-consuming surgical airway techniques. It depends on the experience and preference of the individual emergency physician which surgical technique is employed to manage the airway of someone with complete obstruction. Cricothyrotomy is widely recommended as a quicker alternative to tracheostomy in the ED.15 Retrograde intubation and transtracheal jet ventilation are other options for emergent airway control when endotracheal intubation is impossible. It is commonly taught that it is possible to bag-valve-mask ventilate patients with upper airway obstruction from epiglottitis prior to and during surgical airway attempts.16 This preoxygenation may be the difference between preservation of neurologic function and anoxic encephalopathy. In this case, if an alternative to tracheostomy were selected, perhaps it would have taken less time to restore oxygenation.
If the facts of this case are accurate, this patient presented with a normal respiratory rate and no signs of respiratory distress and precipitously developed a complete airway obstruction. How commonly does this occur? If it is a well-described scenario, then the recommendation that adult patients with no signs or symptoms of respiratory distress can be managed without airway intervention may be suspect. A careful review of all available case series reveals only rare cases where there was not at least some objective or subjective evidence of respiratory distress prior to complete airway obstruction. In these rare cases, there appeared to be inadequate observation, including admission to a regular floor bed.6 In this case, the temporal relation between the parenteral dose of penicillin and the patient’s respiratory arrest was approximately 5-10 minutes. Despite the testimony at trial and the decision in this case, it appears much more likely that the precipitous development of airway edema was related to anaphylaxis to the penicillin than it was to the underlying disease.
Case #2: From Sore Throat to Airway Obstruction in Hours
A 35-year-old man presented to an ED with complaints of sore throat and inability to swallow. His symptoms were of short duration (hours). He was diagnosed with pharyngitis and discharged. Within two hours, he returned to the ED in respiratory failure and died while in the ED. The defense maintained that symptoms of epiglottitis were not apparent on the first presentation. The defense settled out of court (for an undisclosed amount) prior to trial on the issue of failure to control the patient’s airway on the second visit. The jury returned a verdict for the defense on the issue of misdiagnosis on the first visit.17
This "mixed verdict" is hardly comforting. The bottom line is that there was money paid out in this case for the combined allegation of failure to properly diagnose and treat epiglottitis. There is a subtle but critical point illustrated by this case. This patient went from the onset of symptoms to death in a matter of hours. It appears that there are fulminant cases of epiglottitis in adults that are distinct from cases where there is a more prolonged prodrome of symptoms and gradual progression to signs and symptoms of supraglottic inflammation over days, not hours. These more fulminant cases are thought to be caused by H. influenzae and are thought to be associated with bacteremia and a toxic appearance; when these more fulminant cases are identified, early airway control is mandatory.6
In this case, a perfectly healthy man sought care in an ED for a severe sore throat and inability to handle his own secretions. His symptoms were of very short duration. This should have served as a "red flag." We are sometimes quick to criticize patients for seeking care in an ED for a seemingly trivial complaint like a sore throat. Perhaps instead we should be asking why patients’ symptoms are so severe that they are coming to the ED. In retrospect, this patient had subtle clues that there was something going on beyond a simple sore throat. Typical pharyngitis has a gradual onset, not sudden. Most patients with pharyngitis do not come to the ED within a couple of hours of onset of symptoms. Most patients with uncomplicated pharyngitis do not develop inability to handle their own secretions within hours of the onset of symptoms, if at all. In this case, the severity and rapidity of symptoms were not consistent with pharyngitis. Had this been recognized, further diagnostic studies and/or a period of observation might have prevented the rapid deterioration of this patient.
Case #3: "You Are Not an Emergency"
This case further illustrates the point about what constitutes an emergency. A 32-year-old man developed a sore throat and hoarseness in the morning. By 5 p.m. the patient had severe pain on swallowing and had to use a handkerchief to soak up his saliva. His wife drove him to the ED of a U.S. Navy hospital. They were told it would be a three-hour wait. He was not seen by a doctor and was told, "You are not an emergency." Since he was unable to speak, the patient wrote back: "Your system stinks." At 9:20 p.m. he went back to the ED, waited over two hours, and was then examined and treated for a total of 11 minutes. There was testimony that the examining physician did not look at the patient’s throat with a dental mirror and light source (indirect laryngoscopy), because the light was in a different examination room. He was diagnosed with pharyngitis and given prescriptions for penicillin and a throat spray.
The patient was apparently unable to sleep all night, and at 8:30 a.m. his wife found him choking. He wrote on a piece of paper: "ambulance." He was taken to another ED, where he died shortly after arrival. The coroner reported the cause of death as epiglottitis. The patient’s spouse sued the government for the failure of the Navy hospital to diagnose epiglottitis. The defense contended that the decedent’s symptoms were not made known to the emergency room nursing staff or doctor. The jury returned a verdict for the plaintiff for $1,122,900.18
This case vividly illustrates that when patients or their loved ones perceive they have been treated badly, those perceptions can result in damaging testimony in malpractice suits. Telling patients that they do not have an emergency, subjecting them to long waits without reassurances or explanations, and spending little time with them and creating the perception that "nothing was done" can all come back to haunt us in court. Triage and registration personnel should be instructed never to judge patients on the face value of their chief complaint. Patients should be told the reasons for their waits. And even though many ED visits are necessarily brief, patients should be afforded the opportunity to ask questions or express concerns. The damaging testimony detailed above undoubtedly influenced the jury to be skeptical about the defense’s contention that the patient did not make his symptoms known to the nurses and doctors. It seems difficult to believe that a man in obvious distress who was spitting his saliva into a rag and unable to talk did not convey his symptoms to the treating nurses and doctors.
This case also raises two other important points. First, this patient had two signs that needed to be carefully addressedhe was unable to handle his own secretions and he had a hoarse voice, both in the context of a severe sore throat. Second, the testimony indicated that indirect laryngoscopy was readily available in this ED. These points deserve some consideration.
Over half of adult patients with epiglottitis will have dysphonia (hoarseness or muffled voice) and at least 30% will have inability to handle their own secretions.1 While not uniformly predictive of epiglottitis, these symptoms should alert the examiner that there is a process more complicated than simple pharyngitis going on. Whether it is a peritonsillar abscess or some other more serious entity like supraglottic inflammation causing dysphonia and inability to handle secretions, these symptoms demand a more careful search for their cause. As stated in the overview, maintenance of an erect sniffing position and stridor are also ominous signs that are not found with simple pharyngitis.
In this case, indirect laryngoscopy was clearly an accessible diagnostic technique. One concern was raised in the editor’s note: Do indirect or direct visualization techniques induce airway obstruction in adults? As long as the patient is not in respiratory extremis, tongue depressor exam, indirect mirror technique, and direct fiberoptic nasopharyngoscopy are all safe. In seven separate series, 295 adult patients with proven epiglottitis had either indirect or direct visualization, some with signs or symptoms of mild-to-moderate respiratory distress; none had any complications from the visualization techniques.1,2,6-8,14,15 Had this patient’s symptoms been taken more seriously, and had he had one of these relatively simple visualization procedures done, the poor outcome in this case may have been prevented.
Case #4: An Ill-Advised Trip from an ED to a Pediatrician
A 14-month-old presented to an ED with unspecified signs and symptoms of bacteremia. The emergency physician conferred with the child’s pediatrician by telephone, and they agreed to delay instituting treatment until the pediatrician had the opportunity to see the patient. On the way to the pediatrician’s office, the child suffered a respiratory arrest and died at the hospital four days later. The allegation against the emergency physician was failure to diagnose and treat epiglottitis. The defense argued that the child was stable in the ED and that there was no reason to suspect epiglottitis. The jury decided in favor of the defense.19
This unfortunate case raises some troubling questions, despite the outcome in favor of the defense. Why transport the patient out of the ED? Why defer treatment decisions if there was concern about a serious bacterial illness? Did the child really look well enough to leave the ED, as the defense contends, despite the fact that they were worried about bacteremia? We can only speculate on the answers and learn some lessons from this case.
Emergency physicians should rarely, if ever, defer to someone else’s judgment about the management of a seriously ill patient in the ED unless that someone else is at the bedside, has some special expertise, and/or is going to assume the continuing inpatient care of the patient. Some malpractice cases against emergency physicians involve, to some extent, bad advice over the telephone from a consultant. It is understandable to want to maintain good relationships with our colleagues in other specialties; however, we can never compromise patient safety or care. The level of knowledge of the emergency physician who is at the bedside is superior to that of the consultant on the phone; the emergency physician has actually examined the patient. There is never an excuse to transport a seriously ill patient from the safety of the ED to a facility with lesser capabilities. This is indefensible from a standard-of-care point of view, and it probably violates patient transfer laws and policies. Had this child remained in the ED, the proper personnel and equipment would have been more readily available when the child arrested.
This case also illustrates a point made in the overview. Epiglottitis can affect patients of almost any age. In an earlier case, the victim of epiglottitis was a 55-year-old. In this case, the patient was 14 months of age. Although there are clusters of cases at the ages of 6-7 and 40, emergency physicians should suspect the disease in any patient with sore throat, respiratory distress, and a toxic appearance. The diagnosis in the very young child may be difficult, but attention to simple details, like carefully counting a respiratory rate and undressing the child to look for retractions, can make a true difference. It is unlikely that the child in this case did not manifest some signs of respiratory distress. Subtle signs of early respiratory distress in infants and young toddlers include an increased respiratory rate, minimal retractions, and decreased feeding. Any of these signs, in combination with s
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content