- Email: [email protected]
The rising incidence of adult acute epiglottitis and epiglottic abscess
The Rising Incidence of Adult Acute Epiglottitis and Epiglottic Abscess Gilead Berger, MD,*‡ Tali Landau, MD,*‡ Sivan Berger, MD,*‡ Yehuda Finkelstein, MD,*‡ Joelle Bernheim, MD,†‡ and Dov Ophir, MD*‡ Objective: To examine preliminary observations that the incidence of adult acute epiglottitis has risen between 1986 and 2000. Materials and Methods: Demographics, annual and seasonal occurrences, clinical presentation, diagnostic procedures, treatment, airway management, and complications of 116 consecutive adult patients with laryngoscopically confirmed acute epiglottitis are presented. Results: The mean annual incidence of acute epiglottitis per 100,000 adults significantly increased from 0.88 (from 1986-1990) to 2.1 (from 1991-1995) and to 3.1 (from 1996-2000) (P ⬍ .001). This rise seems to be unrelated to Haemophilus influenzae type b infection but related to miscellaneous pathogenic bacteria. During these periods, the number of epiglottic abscesses increased concomitantly with the rise in the incidence of acute epiglottitis (from 4/14 episodes [29%], to 8/38 [21%], and to 16/66 [24%], respectively), showing a relatively constant ratio between both phenomena (P ⫽ .843). Twenty-five patients (21%) underwent airway intervention, 16 because of objective respiratory distress and 9 because of imminent respiratory obstruction. Stepwise logistic regression showed that drooling, diabetes mellitus, rapid onset of symptoms, and abscess formation were associated with airway obstruction. Diverse origins for the epiglottic abscess, either from coalescent epiglottic infection or from mucopyocele of the tongue base, are suggested. Conclusions: A rise in the incidence of acute epiglottitis and a concomitant rise in the number of epiglottic abscesses were established. Although the course of acute epiglottitis is often benign and can be safely treated with a conservative management approach, delayed airway obstruction may develop from a few hours to days after admission. (Am J Otolaryngol 2003;24:374-383. © 2003 Elsevier Inc. All rights reserved.)
Acute epiglottitis (AE) is a bacterial infection of the supraglottic structures that may result in a life-threatening upper airway obstruction. Because vaccination against Haemophilus influenzae type b (Hib) has become part of children’s immunization program, the incidence of AE in children has markedly dropped1-5 and attention shifted to its adult form. A review of the literature reveals a controversy over the incidence of the disease in adults. Several authors showed that the annual incidence remained relatively stable.1,3,6-9 Other indicated an increase2,4,10-20
From the Departments of *Otolaryngology-Head and Neck Surgery; and †Pathology, Meir Hospital, Sapir Medical Center, Kfar Saba; and ‡Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel. Address correspondence to: Gilead Berger, MD, Department of Otolaryngology-Head and Neck Surgery, Meir Hospital, Sapir Medical Center, Kfar Saba 44281, Israel. E-mail: [email protected] © 2003 Elsevier Inc. All rights reserved. 0196-0709/03/2406-0000$30.00/0 doi:10.1053/S0196-0709(03)00083-8 374
or a decrease5,21 in the incidence of AE in adults before or after Hib vaccination. Epiglottic abscess (EA) is considered an infrequent sequela of AE. Heeneman and Ward22 dealt comprehensively with the issue in 1977; since then only sporadic case reports have been published.23-25 The available data have shown considerable variations of incidence of EA among adults, ranging from nil9,21 to 29%.26 The study aims to verify preliminary observations that the incidence of AE and EA in adults has risen during the 15-year period from 1986 to 2000 and offers a possible mechanism that might trigger the phenomena. Special emphasize has been given to the annual and seasonal occurrence, presenting symptoms, clinical course, diagnostic procedures, airway management, and complications of this ailment. MATERIAL AND METHODS The study population consisted of 116 consecutive patients, 16 years old and above, who had a
American Journal of Otolaryngology, Vol 24, No 6 (November-December), 2003: pp 374-383
ADULT ACUTE EPIGLOTTITIS
diagnosis of AE or supraglottitis. Two female patients had recurrent AE, with a 2- and 3-year interval between each episode; accordingly, the total number of episodes was 118. Patients were admitted between 1986 and 2000 to the Department of Otolaryngology-Head and Neck Surgery, Meir Hospital, Sapir Medical Center, Kfar Saba, Isreal. Meir Hospital is a regional referral center, serving in the year 2000 a population of about 650,000 residents of the Sharon area and its vicinity (located north of Tel Aviv and part of Israel’s Central District), thus, making it an appropriate setting to accurately plot the incidence of AE in the area. To calculate the respective incidence rate of AE, the average annual number of adult cases over 3 successive periods each comprising 5 years (1986-1990, 1991-1995, and 1996-2000), served as numerators. Statistical data obtained from the Central Bureau of Statistics,27 calculating the average adult population in the Sharon area throughout the 3 periods, served as denominators. The yearly average number of adult population at risk in the aforementioned periods was 318,384, 364,565, and 423,054, respectively. Data were collected from hospital records, using a form exclusively designed for the purpose of this study, and included patients’ demographics and clinical presentation, annual and seasonal occurrence, diagnostic procedures, treatment, airway management, and complications of AE. The records of the National Center for Forensic Medicine were also studied to ascertain possible deaths linked to AE before hospital admission. Until 1991, the diagnosis of AE was established by indirect laryngoscopy; since then, transnasal flexible fiberoptic laryngoscopy has been added. The latter is accomplished by means of an Olympus ENF-P3 flexible fiberoptic endoscope (Olympus Opto-Electronics Co., Tokyo, Japan) and a Contec Medical LS-3525 halogen light source (Contec Medical Ltd., Ramat Hasharon, Israel). It is an indispensable tool for the diagnosis of patients with a hyperactive gag reflex and provides accurate visualization of the extent of epiglottic swelling; the existence of laryngeal narrowing; and the involvement of neighboring supraglottic structures such as the aryepiglottic folds, the arytenoids, and the false vocal cords; it also shows the formation of an EA. Both modalities were proven safe and did not precipitate airway obstruction, ultimately replacing lateral soft-tissue neck x-ray as the primary diagnostic tool.
Treatment Protocol According to the treatment protocol (Fig 1), a staff ear, nose, and throat (ENT) surgeon assesses the glottic aperture with a flexible fiberoptic laryngoscope and determines whether airway is patent or compromised. The decision to secure the airway depends on patients’ symptoms and airway appearance. In patients with objective respiratory difficulties (eg, stridor, cyanosis, sitting erect, and oxygen desaturation), airway intervention is immediately
375
Fig 1. Treatment protocol delineating the routine practice of the treatment of AE.
addressed; it is also addressed in those who lack symptoms and signs of airway obstruction but have a compromised airway (ie, the laryngeal inlet as observed with a fiberoptic laryngoscope is appreciably narrowed by at least 50%). The reasons for the latter policy are the fickle nature of the disease and apprehensions to handle airway problems in case of imminent respiratory distress. When feasible, patients are transferred to the operating room, where a staff anesthesiologist performs transnasal endotracheal intubation to maintain an open airway. Tracheotomy or cricothyrotomy are accomplished when repeated attempts at intubation fail. After airway is secured, patients are transferred to the intensive care unit. All other patients, whose laryngeal aperture appears patent, are admitted to the department and closely supervised by a staff ENT surgeon and in his/her absence by an ENT resident who is available at all times. This conservative treatment regimen includes intravenous antibiotics to treat the infection supplemented with fluids to increase hydration and 30% humidified oxygen administered via facemask to ease labored breathing. To decrease the perils of airway obstruction, patients with extensive swollen supraglottis and impending airway blockage receive corticosteroids. Arterial oxygen saturation is continuously monitored with finger pulse oxymetry.
Statistical Analysis The data were analyzed using the SPSS (Statistical Package for Social Sciences; SPSS, Inc, Chicago, IL) statistical program. To establish risk factors for airway intervention, demographics, clinical presentation, and diagnostic tests of patients requiring emergency intervention and those managed conservatively were compared, using Pearson chi-square test and Fisher exact test for categorical variables and independent t test for continuous variables. For variables significantly associated with airway intervention, stepwise logistic regression was performed to control the potential confounding factors
376
BERGER ET AL
Fig 3.
Mean seasonal occurrence of AE.
Fig 2. Yearly incidence of AE and EA between the years 1986 and 2000.
of these characteristics. The odds ratio (OR) estimates of this model and 95% confidence interval (CI) were calculated. Measurements were expressed as mean ⫾ SD; probability values lower than .05 were considered significant.
RESULTS Patients Of the 116 patients enrolled in the study, 63 were men (54%) and 53 women (46%). The ratio of men to women was 1.2:1. Patients varied in age from 16 to 88 years (mean ⫾ SD, 50.3 ⫾ 14.9 years). Altogether, 16 patients (14%) had background diseases; 10 were diabetics, 3 had malignancies, 1 had a combined hepatitis types B and C, 1 had rheumatoid arthritis, and 1 was a drug addict. The mean hospital stay was 5.1 ⫾ 4.2 days. Patients were discharged without documented sequelae; no deaths occurred in this series. The number of episodes of AE increased from 14 between 1986 and 1990, to 38 between 1991 and 1995, and to 66 between 1996 and 2000 (Fig 2). Furthermore, the mean annual incidence of AE per 100,000 residents increased from 0.88 between 1986 and 1990, to 2.1 between 1991 and 1995, and to 3.1 between 1996 and 2000. Data analysis confirmed that this rise is statistically significant (P ⬍ .001). A search conducted at the National Center for Forensic Medicine revealed that no deaths related to AE occurred over the 15-year period before hospital admission. In addition, the number of EAs during the 3 periods increased concomitantly with the rise in the incidence of AE (from 4/14 episodes [29%], to 8/38 episodes [21%], and to 16/66 [24%]),
showing a high incidence rate with insignificant variations among the periods (P ⫽ .843) (Fig 2). No seasonal variation in the incidence of the disease was encountered (Fig 3). Symptoms and Signs Table 1 shows that the duration of symptoms before hospital admission ranged from 8 to 168 hours (mean ⫾ SD, 63.6 ⫾ 44.8 hours). Only 12 patients were admitted between 8 to 16 hours after commencement of symptoms, whereas the majority had symptoms that lasted from 1 to 4 days. Before hospitalization, 37 patients (31%) were given antibiotics by family physicians. On admission, most patients complained of a sore throat (88%) and painful swallowing (83%). Forty-five patients (38%) complained of temperature elevation, 34% of respiratory difficulties, 25% of a muffled voice, and 7% of drooling (Table 2). Indirect laryngoscopy showed that all patients had a grossly swollen erythematous epiglottis, with extension of the inflammatory reaction to the arytenoids in 45 patients, the false vocal cords in 11, and the TABLE 1. Hospital
Duration of Symptoms Before Admission to
Duration
No. of Episodes
8-16 hours 1 day 2 days 3 days 4 days 5 days 6 days 7 days Not recorded Total
12 26 22 28 11 5 0 11 3 118
ADULT ACUTE EPIGLOTTITIS
377
TABLE 2. Symptoms on Admission to Hospital
Symptoms
No. of Episodes (%)*
Sore throat Odynophagia Temperature elevation Respiratory difficulties Muffled voice Drooling
104 (88) 98 (83) 45 (38) 40 (34) 30 (25) 8 (7)
*One hundred eighteen episodes.
aryepiglottic folds in 8. The mean body temperature was 37.6° ⫾ 0.7°C (range, 36°39.3°C); in 19 patients temperature exceeded 38°C. The mean pulse rate was 88 ⫾ 13 beats per minute. The mean white blood cell count was 12,817 ⫾ 4,581 cells/mm3 (range, 5,60025,900 cells/mm3). Blood cultures were obtained from 17 patients, of whom 16 yielded no growth of bacteria and only 1 (6%) had a positive culture for Clostridium sporogenes. Treatment Regimen The antibiotic regimen included intravenous amoxicillin and cloxacillin in the early years. Subsequently, patients were given either cefuroxime alone or in combination with metronidazole, or amoxicillin and potassium clavulanate. Fifty-eight patients (49%) were treated with intravenous corticosteroids. The mean hospital stay of patients receiving corticosteroids was 5.7 ⫾ 4.1 days compared with 4.7 ⫾ 4.3 days in patients not receiving these medications. The difference between the 2 groups was statistically insignificant (P ⫽ .2). Airway Management Ninety-nine patients whose laryngeal aperture was patent on arrival were admitted to the ENT department for close monitoring. The remaining 19 had clinical respiratory distress
with a compromised airway or imminent airway obstruction and were immediately transferred to the operating room to establish airway control (Table 3). Sixteen had successful endotracheal intubation; after failed attempts at intubation, airway control was accomplished by tracheotomy in 2 patients and cricothyrotomy in 1. Six other patients, who initially had no respiratory distress or airway compromise, developed from a few hours to 2 days after admission delayed respiratory distress; to secure their airway, 5 had successful endotracheal intubation and 1 required tracheotomy. Altogether, 25 patients had at some point respiratory obstruction that demanded airway intervention; they were transferred to the intensive care unit and kept there for 4.3 ⫾ 3.4 days until extubation. The mean hospital stay of patients with airway intervention was significantly longer than that of patients who did not experience respiratory distress (mean ⫾ SD, 9.9 ⫾ 6.4 days v 3.9 ⫾ 2 days; P ⬍ .001). Nine of the 25 patients (36%) lacked prominent respiratory symptoms and the decision to secure the airway was based on clinical observation, confirming that at least 50% of the laryngeal lumen was narrowed. Abscess Formation Epiglottic abscess was evident in 24% of the patients (28/118). Interestingly, their proportion was almost 2-fold higher among those who needed airway intervention (44%, 11/ 25). Of the latter, 7 had immediate respiratory obstruction and 4 had delayed obstruction; both went through incision and drainage of abscess under direct laryngoscopy at the time of airway control. Another 11 patients, who had an EA without airway obstruction, underwent incision and drainage of abscess under endotracheal general anesthesia (5 patients on the day of admission and 6 on the second, third, and fourth hospital day) and were extu-
TABLE 3. Airway Intervention With or Without an Abscess
Compromised airway
Uncompromised airway Total
With abscess Without abscess Total With abscess
Immediate Intervention
Delayed Intervention
Total
7 12 19 5 24
4 2 6 6 12
11 14 25 11 36
378
Fig 4. A section of a patient with acute epiglottitis, showing coalescence of an epiglottic abscess with abundance of neutrophils (N) located adjacent to the abscess cavity (C) and fibrin (F) located more peripherally (hematoxylin and eosin, original magnification ⴛ100).
bated after the drainage procedure (Table 3). Altogether, 22 patients required a surgical drainage, whereas the remaining 6 had a spontaneous drainage of the abscess and were conservatively managed with intravenous antibiotics and careful monitoring. Twenty-one abscesses were located along the lingual surface of the epiglottis, 1 toward its laryngeal surface and 6 had no predilection for either direction. Eight patients underwent computed tomography to delineate the confines of the abscess. Nine cultures were retrieved from the abscess cavity and yielded normal pharyngeal flora in 4, positive cultures for Escherichia coli, Enterobacter, Klebsiella pneumonia, and streptococcus in another 4 and no growth of bacteria in 1. At the time of drainage, a punch biopsy of the abscess wall was obtained from 6 patients. The material underwent routine histological processing and was stained with hematoxylin and eosin. Light microscopy confirmed that there were 5 abscesses involving the epiglottis itself, which originated from a coalescence of patchy tissue necroses forming an abscess cavity filled with necrotic white cells and debris and bounded by a wall composed of numerous neutrophils and a peripheral zone of fibrin (Fig 4). In the remaining biopsy, squamous epithelial cells were integrated within the inflamed walls of the abscess (Fig 5), indicating it is an infected mucus retention cyst (mucopyocele) of the base of the tongue that reached the lingual
BERGER ET AL
Fig 5. A section of a patient with AE, showing a mucopyocele with a squamous epithelial cell (arrow) integrated within the inflamed wall (hematoxylin and eosin, original magnification ⴛ100).
aspect of the epiglottis and caused bacterial infection of the supraglottic structure. A computed tomography scan image verified the different origin of the latter and showed an infected mucopyocele of the base of the tongue situated anteriorly to the swollen epiglottis (Fig 6). Predictors of Airway Obstruction Data analysis of patients undergoing airway intervention versus those who were managed conservatively confirmed that the 9 following factors were associated with increased likelihood to develop airway obstruction: (1) age (mean ⫾ SD, 55.8 ⫾ 13.8 years v 48.8 ⫾ 14.9
Fig 6. A computed tomography of a patient with AE injected with intravenous contrast material, showing an infected mucus retention cyst (mucopyocele) of the base of the tongue with enhanced walls (thick arrow) situated anteriorly to the swollen epiglottis (thin arrow).
ADULT ACUTE EPIGLOTTITIS
379
TABLE 4. Epidemiologic Features of Adult AE in Published Studies Author Before Hib vaccination Shih et al MayoSmith et al Carenfelt and Sobin Hugosson et al Kucera et al Murrage et al Wurtele Yardley Berg et al After Hib vaccination Alho et al Mayo-Smith et al Navarrete et al Solomon et al Ryan et al Frantz and Ragson Torkkeli et al Dort et al Takala et al Garpenholt et al Hebert et al Chung
Study Site
Period (Years)
No. of Patients
Incidence/100,000
Secular Trend
Florida* Rhode Island* Sweden Sweden Minnesota* Ontario† Quebec† UK Sweden
1963-1987 1975-1982 1975-1987 1975-1992 1976-1992 1977-1986 1982-1988 1985 1987-1989
48 56 543 93 20 26 12 7 502
NR 0.97 3.9 2.3 4-2.2 0.7 1 NR 2.5
Increasing Increasing Stable Stable Decreasing Increasing Stable Increasing Increasing
Finland Rhode Island* Spain Ontario† Pennsylvania* California* Finland Manitoba† Finland Sweden Ontario† Hong Kong
1967-1993 1975-1992 1978-1990 1979-1991 1979-1991 1980-1990 1981-1992 1982-1992 1985-1992 1987-1996 1989-1994 1998-2001
49 273 58 57 8 258 32 43 20 241 51 11
0.5-2.3 0.78-2.9 NR NR NR 1.8 1.5 NR 0.03 3.4 2.02 NR
Increasing Increasing Increasing Increasing Increasing Stable Increasing Increasing Stable Decreasing Stable Increasing
Abbreviations: NR, not recorded. *United States. † Canada.
years; P ⫽ .036), (2) diabetes mellitus (5 of 10 [50%] v 20 of 108 [19%]; P ⫽ .03), (3) rapid onset (ie, within 16 hours) (6 of 12 [50%] v 19 of 106 [18%]; P ⫽ .022), (4) voice changes (11 of 30 [37%] v 14 of 88 [16%]; P ⫽ .036), (5) drooling (5 of 8 [62.5%] v 20 of 110 [18%]; P ⫽ .011), (6) pulse rate (mean ⫾ SD, 94.2 ⫾ 15.5 beats per minute v 85.2 ⫾ 11.5 beats per minute; P ⫽ .042), (7) temperature (mean ⫾ SD, 37.8° ⫾ 0.9°C v 37.4° ⫾ 0.7°C; P ⫽ .035), (8) white blood cell count (14,625 ⫾ 5,210 cells v 12,310 ⫾ 4,287 cells; P ⫽ .032), and (9) abscess formation (11 of 28 [39%] v 14 of 90 [16%]; P ⫽ .015). Using the stepwise logistic regression analysis, 4 variables remained associated with airway obstruction: (1) drooling (OR, 8.1 [95% CI, 1.5-43.7]), (2) diabetes mellitus (OR, 6.9 [95% CI, 1.6-29.7]), (3) short onset (OR, 4.8 [95% CI, 1.2-19.8]), and (4) abscess (OR, 4.6 [95% CI, 1.5-13.5]). Based on previously determined criteria for evaluation, it was shown that the 4 following factors were not associated with airway obstruction: (1) gender (16 of 63 [25%] for men v 9 of 55 [16%] for women; P ⫽ .265), (2) duration of symptoms (mean ⫾ SD, 52.5 ⫾ 38.5
hours v 68.5 ⫾ 53.1 hours; P ⫽ .162), (3) complaints of respiratory difficulties (12 of 40 [30%] v 13 of 78 [17%]; P ⫽ .152), and (4) preadmission antibiotics (10 of 37 [27%] v 15 of 81 [19%]; P ⫽ .294). DISCUSSION The data show a significant increase in the incidence of the disease among adults in the Sharon area between 1986 and 2000 and agree with other studies on adult AE carried out elsewhere (Table 4). Notably, evidence of such a trend in adults was already apparent before the introduction of large-scale pediatric vaccination against Hib and continued after vaccination, despite the marked diminution of AE in children. This would imply that the vaccine has little effect on nonvaccinated adults. Evidently, further studies are needed to determine the possible effects of Hib vaccine on vaccinated adults. It should be born in mind that Hib, unlike in children, is not the exclusive pathogen in adults and other etiological agents also exist (eg, Streptococcus pneumoniae, Staphylococcus aureus, Kleb-
380
siella pneumoniae). A number of authors questioned whether the apparent increase in adult AE represents a genuine rise in the incidence of the disease and suggested that greater physician awareness and/or improved diagnosis are accountable for this change.10,13 Evidently, the introduction of the flexible fiberoptic nasopharyngoscope turned the visualization of the inflamed epiglottis and the assessment of a possibly compromised airway into a more accurate task than simply relying on soft tissue lateral x-ray of the neck. Yet, it is hardly conceivable that the concurrent increase in the incidence of adult AE in different areas of the world could be explained by the mere presence of better recognition and diagnosis of the disease.16 Our findings have shown that the number of adult patients with AE continued to increase between 1991 and 2000, when fiberoptic nasopharyngoscopy was routinely used, thus supporting the notion that there is a bona fide rise in the incidence of AE. Once established, there is a rise in the incidence of AE; it is of importance to consider the possible causes behind the phenomenon. The last 2 decades have witnessed a sharp increase in the number of human immunodeficiency virus–positive, acquired immunodeficiency syndrome, and other immunocompromised patients; yet except for case reports,28-30 larger series either did not report17 or only had isolated cases of immunocompromised patients with AE.19,31 We showed that 16 of 116 patients (14%) had background diseases of which 10 had diabetes mellitus; no human immunodeficiency virus– positive or acquired immunodeficiency syndrome patients were presented. Thus, so far, the data do not support the assertion that the spread of immunodeficiency states influenced the occurrence of adult AE. However, these conditions enhance the possibility to run a more severe course of the disease, as shown in our group of diabetics or require a longer hospital stay than the average.32 Mayo-Smith et al17 ascribed the rise in the incidence of EA mainly to increased numbers of patients who lack Hib bacteremia and are characterized by negative blood cultures, slow onset of symptoms, and reduced probability to develop airway obstruction. It is also suggested that similar to the significant increase in the incidence
BERGER ET AL
of other common upper airway infectious diseases such as acute otitis media and sinusitis,33 the rise in the incidence of EA may be related to indiscriminate use of antibiotics leading to the emergence and spread of antimicrobial resistance, which increasingly threatens the success of infectious disease treatment and prevention. The essence of AE management is to safeguard patients’ airway and prevent obstruction. Views vary between those advocating prophylactic airway intervention in all patients because of the unpredictable nature of the disease and its potentially swift course10 and those practicing conservative treatment.34 Nevertheless, most authors fall somewhere between these 2 extremes and adopt a selective approach. Thus, patients with severe airway symptoms and signs (eg, dyspnea, stridor, and cyanosis) require immediate establishment of an airway either by intubation or tracheotomy, whereas the less severe cases are conservatively treated in a monitored setting. Proponents of this practice find it safe and effective because it significantly lowers the proportion of patients undergoing airway intervention (Frantz et al,31 15%; Solomon et al,19 16%; Torkkeli et al,16 16%; Hebert et al,9 20%; Andreassen et al,35 22%; Cucera et al,21 25%; Deeb et al,36 30%, and Dort et al,15 33%) and avoids the increased risk of mortality rates. Two prominent questions remain: who will eventually progress to acute airway obstruction and at what point should airway be established? One must remember that in these intricate situations, decisions are often made under uncertainty. To guide the caring otolaryngologist when contemplating airway intervention, 4 predictors of impending airway obstruction are suggested: (1) drooling, (2) a history of diabetes mellitus, (3) rapid onset of symptoms, and (4) abscess formation. A number of authors have also suggested drooling, rapid onset, and EA as predictors of airway obstruction.12,17,19,36 In addition, our study has pointed out a significant relationship between diabetes mellitus and airway obstruction. An exception to the selective management approach pertains to a category of patients who lack prominent respiratory symptoms, yet their laryngeal inlet is narrowed by at least 50%.9,32 The condition should be regarded as
ADULT ACUTE EPIGLOTTITIS
an imminent airway obstruction that requires immediate intervention, thus avoiding the extremely difficult management during acute respiratory distress. The data show that in 9 out of 25 patients (36%) of the current study the decision to accomplish airway intervention was based on similar circumstances. Apparently, the use of the flexible fiberoptic nasopharyngoscopy allows an accurate estimate of the laryngeal airway and made it easier to detect these patients. On the other hand, the exclusive reliance on indirect laryngoscopy or soft-tissue lateral x-ray of the neck may provide incomplete view of the laryngeal inlet and patients’ condition can go unnoticed. It is estimated that patients, who in the early stages of the disease appeared to do well and later developed a severe or a sudden fatal airway obstruction,10 in fact, belong to this category. AE is commonly regarded as a self-limiting disease with an increased risk of developing a rapid and a fulminant course. According to MayoSmith et al,10 all deaths happened before or within 6 hours of admission. This finding supports other observations on patients, who were managed conservatively and did not subsequently required urgent airway intervention until bacterial infection had resolved.19,34,36-38 The present series, on the other hand, uncovered that 6 patients had delayed airway obstruction occurring up to 2 days of admission, of whom 4 were associated with an EA. Shih et al12 and Andreassen et al35 also found a considerable number of delayed obstructions, stressing the capricious nature of AE and the need for close and prolonged observation until swelling is significantly reduced. Corticosteroids are often administrated in an effort to reduce supraglottic inflammation, the likelihood of airway obstruction, complication rate, and length of hospital stay; yet, there is lack of controlled data demonstrating unequivocal evidence of the benefit of this treatment.9,13,15,17,31,32,34 Likewise, in the present series, the mean hospital stay of patients receiving corticosteroids and those not receiving corticosteroids did not differ significantly (5.7 ⫾ 4.1 days compared with 4.7 ⫾ 4.3 days, respectively; P ⫽ .2). Moreover, the average stay of the former exceeded the latter
381
by 1 day, probably reflecting the clinical selection of the more severe cases. Thirty-seven patients (31%) were examined by their family physicians and received antibiotic treatment before hospitalization. Other also reported a similar proportion of patients who had been seen by their primary care physician within 48 hours of diagnosis (32% and 29%, respectively).10,31 Moreover, the findings clearly show that the protective effect of preadmission antibiotics was inadequate (27% of the patients who received antibiotics had upper airway obstruction compared with 19% who did not receive antibiotic treatment before admission [P ⫽ .294]). This high proportion of missed diagnosis that can result in increased morbidity and even mortality highlights the need to provide specific guidance to family physicians and stresses the importance of immediate referral of patients with severe sore throat and odynophagia (especially when examination of the oropharynx is normal) to an otolaryngologist for further evaluation. The study revealed a high incidence of EA (28/118, 24%) among patients with AE. Several authors found similar rates of EA (Shih et al,12 21%; Wolf et al,34 26%; and Hawkins et al,26 29%); yet, others did not report the phenomenon9,21 or had much lesser rates (MayoSmith et al,10 2%; Stanley and Liang,37 2%; Solomon et al,19 2%; Hugosson et al,8 3%; MayoSmith et al,17 3%; Frantz et al,31 4%; Torkkeli et al,16 6%; and Murrage et al,13 8%). It is also shown that the number of EAs increased concomitantly with the rise in the incidence of AE (from 4/14 episodes [29%] between 1986 and 1990, to 8/38 [21%] between 1991 and 1995, and to 16/66 [24%] between 1996 and 2000), disclosing a relatively constant ratio between both phenomena and insignificant variations among the 3 periods (P ⫽ .843). This may lend support to the contingency that abscess formation considerably prevails in AE. Similar to Heeneman and Ward,22 we observed that the lingual surface of the epiglottis was the most commonly involved site of the EA. Concerning the origin the EA, our clinical and pathological data lend support to the Stack and Ridley24 assumptions that it originates either from coalescent epiglottic infection or from infection of the epiglottic mucocele. The former is linked to severe infection that causes patchy
382
necrosis with accumulation of numerous leucocytes in the area. The latter refers to an infected mucus retention cyst of the base of tongue or the vallecula that reaches the lingual aspect of the epiglottis and in contrast to the former has an epithelial wall. CONCLUSIONS Our 15-year experience from 1986 through 2000 suggests that (1) the incidence of adult AE increased significantly, a secular trend probably unrelated to Haemophilus influenzae type b infection, but to other pathogenic bacteria; (2) a selective management approach, which advocates airway intervention only in patients with severe airway symptoms or with imminent airway obstruction, yielded a relatively low rate of interventions (21%) and no fatalities; (3) although most patients requiring airway intervention demonstrated compromised airway on admission, a considerable number may develop delayed upper airway obstruction; (4) a significant proportion of the cases with compromised airway was related to EA formation; (5) the number of epiglottic abscesses increased concomitantly with the rise in the incidence of AE; and (6) 2 diverse pathophysiologic pathways of EA formation are proposed. ACKNOWLEDGMENT We thank Estela Derazne, MSc, from the Lady Sarah Cohen Unit for Family Centered Therapy and Healthcare, the Schneider Children’s Medical Center of Israel, Petah Tikva, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel, for assistance with statistical analysis; and Rachel Berger, BA, for writing and editing assistance. REFERENCES 1. Frantz TD, Rasgon BM: Acute epiglottitis: changing epidemiologic patterns. Otolaryngol Head Neck Surg 109: 457-460, 1993 2. Ryan M, Hunt M, Snowberger T: A changing pattern of epiglottitis. Clin Pediatr 31:532-535, 1992 3. Takala AK, Peltola H, Eskola J: Disappearance of epiglottitis during large-scale vaccination with Haemophilus influenzae type b conjugate vaccine among children in Finland. Laryngoscope 104:731-735, 1994 4. Alho OP, Jokinen K, Pirila T, et al: Acute epiglottitis and infant conjugate Haemophilus influenzae type b vac-
BERGER ET AL
cination in northern Finland. Arch Otolaryngol Head Neck Surg 121:898-902, 1995 ¨ , Hugosson S, Fredlund H, et al: Epi5. Garpenholt O glottitis in Sweden before and after introduction of vaccination against Haemophilus influenzae type b. Pediatr Infect Dis J 18:490-493, 1999 6. Carenfelt C, Sobin A: Acute infectious epiglottitis in children and adults: annual incidence and mortality. Clin Otolaryngol 14:489-493, 1989 7. Wurtele P: Acute epiglottitis in children and adults: A large-scale incidence study. Otolaryngol Head Neck Surg 103:902-908, 1990 8. Hugosson S, Olce´ n P, Ekedahl C: Acute epiglottitis— Aetiology, epidemiology and outcome in a population before large-scale Haemophilus influenzae type b vaccination. Clin Otolaryngol 19:441-445, 1994 9. Hebert PC, Ducic Y, Boisvert D, et al: Adult epiglottitis in a Canadian setting. Laryngoscope 108:64-69, 1998 10. MayoSmith MF, Hirsch PJ, Wodzinski SF, et al: Acute epiglottitis in adults. An eight-year experience in the state of Rhode Island. N Engl J Med 314:1133-1139, 1986 11. Yardley MH: Supraglottitis in adults. Practitioner 230:583-587, 1986 12. Shih L, Hawkins DB, Stanley RB Jr: Acute epiglottitis in adults. A review of 48 cases. Ann Otol Rhinol Laryngol 97:527-529, 1988 13. Murrage KJ, Janzen VD, Ruby RR: Epiglottitis: Adult and pediatric comparisons. J Otolaryngol 17:194198, 1988 14. Navarrete ML, Quesada P, Garcia M, et al: Acute epiglottitis in the adult. J Laryngol Otol 105:839-841, 1991 15. Dort JC, Frohlich AM, Tate RB: Acute epiglottitis in adults: diagnosis and treatment in 43 patients. J Otolaryngol 23:281-285, 1994 16. Torkkeli T, Ruoppi P, Nuutinen J, et al: A Changed clinical course and current treatment of acute epiglottitis in adults a 12-year experience. Laryngoscope 104:15031506, 1994 17. Mayo-Smith MF, Spinale JW, Donskey CJ, et al: Acute epiglottitis. An 18-year experience in Rhode Island. Chest 108:1640-1647, 1995 18. Berg S, Trollfors B, Nylen O, et al: Incidence, aetiology, and prognosis of acute epiglottitis in children and adults in Sweden. Scand J Infect Dis 28:261-264, 1996 19. Solomon P, Weisbrod M, Irish JC, et al: Adult epiglottitis: The Toronto Hospital experience. J. Otolaryngol 27:332-336, 1998 20. Chung CH: Case and literature review: Adult acute epiglottitis—Rising incidence or increasing awareness? Hong Kong J Emerg Med 8:227-231, 2001 21. Kucera CM, Silverstein MD, Jacobson RM, et al: Epiglottitis in adults and children in Olmsted County, Minnesota, 1976 through 1990. Mayo Clin Proc 71:11551161, 1996 22. Heeneman H, Ward KM: Epiglottic abscess: Its occurrence and management. J Otolaryngol 6:31-36, 1977 23. Ridgeway NA, Perlman PE, Verghese A, et al: Epiglottic abscess due to group B Streptococcus. Communication. Ann Otol Rhinol Laryngol 93:277-278, 1984 24. Stack BC, Ridley MB: Epiglottic abscess. Case report. Head Neck 17:263-265, 1995 25. Galli J, Corina L, Calo L, et al: Acute supraglottitis in adults: report of a case with development of abscess. Int J Surg Sci 6:59-61, 1999 26. Hawkins DB, Miller AH, Sachs GB, et al: Acute epiglottitis in adults. Laryngoscope 83:1211-1220, 1973 27. Central Bureau of Statistics: Statistical Abstract of Israel. No. 37, 1986; No. 42, 1991; No. 47, 1996; No. 52, 2001
ADULT ACUTE EPIGLOTTITIS
28. Rothstein SG, Persky MS, Edelman BA, et al: Epiglottitis in AIDS patients. Laryngoscope 99:389-392, 1989 29. Laing RB, Wardrop PJ, Welsby PD, et al: Stridor in patients with HIV infection. J Laryngol Otol 109:11971199, 1995 30. Colman MF: Epiglottitis in immunocompromised patients. Head Neck Surg 8:466-468, 1986 31. Frantz TD, Rasgon BM, Quesenberry CP Jr: Acute epiglottitis in adults. Analysis of 129 cases. JAMA 272: 1358-1360, 1994 32. Chan KO, Pang YT, Tan KK: Acute epiglottitis in the tropics: Is it an adult disease? J Laryngol Otol 115: 715-718, 2001 33. McCaig LF, Hughes JM: Trends in antimicrobial drug prescribing among office-based physicians in the United States. JAMA 273:214-219, 1995
383
34. Wolf M, Strauss B, Kronenberg J, et al: Conservative management of adult epiglottitis. Laryngoscope 100:183185, 1990 35. Andreassen UK, Baer S, Nielsen TG, et al: Acute epiglottitis—25 years experience with nasotracheal intubation, current management policy and future trends. J Laryngol Otol 106:1072-1075, 1992 36. Deeb ZE, Yenson AC, DeFries HO: Acute epiglottitis in the adult. Laryngoscope 95:289-291, 1985 37. Stanley RE, Liang TS: Acute epiglottitis in adults (the Singapore experience). J Laryngol Otol 102:10171021, 1988 38. Arndal H, Andreassen UK: Acute epiglottis in children and adults. Nasotracheal intubation, tracheostomy or careful observation? Current status in Scandinavia. J Laryngol Otol 102:1012-1016, 1988
Acute epiglottitis (AE) is a bacterial infection of the supraglottic structures that may result in a life-threatening upper airway obstruction. Because vaccination against Haemophilus influenzae type b (Hib) has become part of children’s immunization program, the incidence of AE in children has markedly dropped1-5 and attention shifted to its adult form. A review of the literature reveals a controversy over the incidence of the disease in adults. Several authors showed that the annual incidence remained relatively stable.1,3,6-9 Other indicated an increase2,4,10-20
From the Departments of *Otolaryngology-Head and Neck Surgery; and †Pathology, Meir Hospital, Sapir Medical Center, Kfar Saba; and ‡Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel. Address correspondence to: Gilead Berger, MD, Department of Otolaryngology-Head and Neck Surgery, Meir Hospital, Sapir Medical Center, Kfar Saba 44281, Israel. E-mail: [email protected] © 2003 Elsevier Inc. All rights reserved. 0196-0709/03/2406-0000$30.00/0 doi:10.1053/S0196-0709(03)00083-8 374
or a decrease5,21 in the incidence of AE in adults before or after Hib vaccination. Epiglottic abscess (EA) is considered an infrequent sequela of AE. Heeneman and Ward22 dealt comprehensively with the issue in 1977; since then only sporadic case reports have been published.23-25 The available data have shown considerable variations of incidence of EA among adults, ranging from nil9,21 to 29%.26 The study aims to verify preliminary observations that the incidence of AE and EA in adults has risen during the 15-year period from 1986 to 2000 and offers a possible mechanism that might trigger the phenomena. Special emphasize has been given to the annual and seasonal occurrence, presenting symptoms, clinical course, diagnostic procedures, airway management, and complications of this ailment. MATERIAL AND METHODS The study population consisted of 116 consecutive patients, 16 years old and above, who had a
American Journal of Otolaryngology, Vol 24, No 6 (November-December), 2003: pp 374-383
ADULT ACUTE EPIGLOTTITIS
diagnosis of AE or supraglottitis. Two female patients had recurrent AE, with a 2- and 3-year interval between each episode; accordingly, the total number of episodes was 118. Patients were admitted between 1986 and 2000 to the Department of Otolaryngology-Head and Neck Surgery, Meir Hospital, Sapir Medical Center, Kfar Saba, Isreal. Meir Hospital is a regional referral center, serving in the year 2000 a population of about 650,000 residents of the Sharon area and its vicinity (located north of Tel Aviv and part of Israel’s Central District), thus, making it an appropriate setting to accurately plot the incidence of AE in the area. To calculate the respective incidence rate of AE, the average annual number of adult cases over 3 successive periods each comprising 5 years (1986-1990, 1991-1995, and 1996-2000), served as numerators. Statistical data obtained from the Central Bureau of Statistics,27 calculating the average adult population in the Sharon area throughout the 3 periods, served as denominators. The yearly average number of adult population at risk in the aforementioned periods was 318,384, 364,565, and 423,054, respectively. Data were collected from hospital records, using a form exclusively designed for the purpose of this study, and included patients’ demographics and clinical presentation, annual and seasonal occurrence, diagnostic procedures, treatment, airway management, and complications of AE. The records of the National Center for Forensic Medicine were also studied to ascertain possible deaths linked to AE before hospital admission. Until 1991, the diagnosis of AE was established by indirect laryngoscopy; since then, transnasal flexible fiberoptic laryngoscopy has been added. The latter is accomplished by means of an Olympus ENF-P3 flexible fiberoptic endoscope (Olympus Opto-Electronics Co., Tokyo, Japan) and a Contec Medical LS-3525 halogen light source (Contec Medical Ltd., Ramat Hasharon, Israel). It is an indispensable tool for the diagnosis of patients with a hyperactive gag reflex and provides accurate visualization of the extent of epiglottic swelling; the existence of laryngeal narrowing; and the involvement of neighboring supraglottic structures such as the aryepiglottic folds, the arytenoids, and the false vocal cords; it also shows the formation of an EA. Both modalities were proven safe and did not precipitate airway obstruction, ultimately replacing lateral soft-tissue neck x-ray as the primary diagnostic tool.
Treatment Protocol According to the treatment protocol (Fig 1), a staff ear, nose, and throat (ENT) surgeon assesses the glottic aperture with a flexible fiberoptic laryngoscope and determines whether airway is patent or compromised. The decision to secure the airway depends on patients’ symptoms and airway appearance. In patients with objective respiratory difficulties (eg, stridor, cyanosis, sitting erect, and oxygen desaturation), airway intervention is immediately
375
Fig 1. Treatment protocol delineating the routine practice of the treatment of AE.
addressed; it is also addressed in those who lack symptoms and signs of airway obstruction but have a compromised airway (ie, the laryngeal inlet as observed with a fiberoptic laryngoscope is appreciably narrowed by at least 50%). The reasons for the latter policy are the fickle nature of the disease and apprehensions to handle airway problems in case of imminent respiratory distress. When feasible, patients are transferred to the operating room, where a staff anesthesiologist performs transnasal endotracheal intubation to maintain an open airway. Tracheotomy or cricothyrotomy are accomplished when repeated attempts at intubation fail. After airway is secured, patients are transferred to the intensive care unit. All other patients, whose laryngeal aperture appears patent, are admitted to the department and closely supervised by a staff ENT surgeon and in his/her absence by an ENT resident who is available at all times. This conservative treatment regimen includes intravenous antibiotics to treat the infection supplemented with fluids to increase hydration and 30% humidified oxygen administered via facemask to ease labored breathing. To decrease the perils of airway obstruction, patients with extensive swollen supraglottis and impending airway blockage receive corticosteroids. Arterial oxygen saturation is continuously monitored with finger pulse oxymetry.
Statistical Analysis The data were analyzed using the SPSS (Statistical Package for Social Sciences; SPSS, Inc, Chicago, IL) statistical program. To establish risk factors for airway intervention, demographics, clinical presentation, and diagnostic tests of patients requiring emergency intervention and those managed conservatively were compared, using Pearson chi-square test and Fisher exact test for categorical variables and independent t test for continuous variables. For variables significantly associated with airway intervention, stepwise logistic regression was performed to control the potential confounding factors
376
BERGER ET AL
Fig 3.
Mean seasonal occurrence of AE.
Fig 2. Yearly incidence of AE and EA between the years 1986 and 2000.
of these characteristics. The odds ratio (OR) estimates of this model and 95% confidence interval (CI) were calculated. Measurements were expressed as mean ⫾ SD; probability values lower than .05 were considered significant.
RESULTS Patients Of the 116 patients enrolled in the study, 63 were men (54%) and 53 women (46%). The ratio of men to women was 1.2:1. Patients varied in age from 16 to 88 years (mean ⫾ SD, 50.3 ⫾ 14.9 years). Altogether, 16 patients (14%) had background diseases; 10 were diabetics, 3 had malignancies, 1 had a combined hepatitis types B and C, 1 had rheumatoid arthritis, and 1 was a drug addict. The mean hospital stay was 5.1 ⫾ 4.2 days. Patients were discharged without documented sequelae; no deaths occurred in this series. The number of episodes of AE increased from 14 between 1986 and 1990, to 38 between 1991 and 1995, and to 66 between 1996 and 2000 (Fig 2). Furthermore, the mean annual incidence of AE per 100,000 residents increased from 0.88 between 1986 and 1990, to 2.1 between 1991 and 1995, and to 3.1 between 1996 and 2000. Data analysis confirmed that this rise is statistically significant (P ⬍ .001). A search conducted at the National Center for Forensic Medicine revealed that no deaths related to AE occurred over the 15-year period before hospital admission. In addition, the number of EAs during the 3 periods increased concomitantly with the rise in the incidence of AE (from 4/14 episodes [29%], to 8/38 episodes [21%], and to 16/66 [24%]),
showing a high incidence rate with insignificant variations among the periods (P ⫽ .843) (Fig 2). No seasonal variation in the incidence of the disease was encountered (Fig 3). Symptoms and Signs Table 1 shows that the duration of symptoms before hospital admission ranged from 8 to 168 hours (mean ⫾ SD, 63.6 ⫾ 44.8 hours). Only 12 patients were admitted between 8 to 16 hours after commencement of symptoms, whereas the majority had symptoms that lasted from 1 to 4 days. Before hospitalization, 37 patients (31%) were given antibiotics by family physicians. On admission, most patients complained of a sore throat (88%) and painful swallowing (83%). Forty-five patients (38%) complained of temperature elevation, 34% of respiratory difficulties, 25% of a muffled voice, and 7% of drooling (Table 2). Indirect laryngoscopy showed that all patients had a grossly swollen erythematous epiglottis, with extension of the inflammatory reaction to the arytenoids in 45 patients, the false vocal cords in 11, and the TABLE 1. Hospital
Duration of Symptoms Before Admission to
Duration
No. of Episodes
8-16 hours 1 day 2 days 3 days 4 days 5 days 6 days 7 days Not recorded Total
12 26 22 28 11 5 0 11 3 118
ADULT ACUTE EPIGLOTTITIS
377
TABLE 2. Symptoms on Admission to Hospital
Symptoms
No. of Episodes (%)*
Sore throat Odynophagia Temperature elevation Respiratory difficulties Muffled voice Drooling
104 (88) 98 (83) 45 (38) 40 (34) 30 (25) 8 (7)
*One hundred eighteen episodes.
aryepiglottic folds in 8. The mean body temperature was 37.6° ⫾ 0.7°C (range, 36°39.3°C); in 19 patients temperature exceeded 38°C. The mean pulse rate was 88 ⫾ 13 beats per minute. The mean white blood cell count was 12,817 ⫾ 4,581 cells/mm3 (range, 5,60025,900 cells/mm3). Blood cultures were obtained from 17 patients, of whom 16 yielded no growth of bacteria and only 1 (6%) had a positive culture for Clostridium sporogenes. Treatment Regimen The antibiotic regimen included intravenous amoxicillin and cloxacillin in the early years. Subsequently, patients were given either cefuroxime alone or in combination with metronidazole, or amoxicillin and potassium clavulanate. Fifty-eight patients (49%) were treated with intravenous corticosteroids. The mean hospital stay of patients receiving corticosteroids was 5.7 ⫾ 4.1 days compared with 4.7 ⫾ 4.3 days in patients not receiving these medications. The difference between the 2 groups was statistically insignificant (P ⫽ .2). Airway Management Ninety-nine patients whose laryngeal aperture was patent on arrival were admitted to the ENT department for close monitoring. The remaining 19 had clinical respiratory distress
with a compromised airway or imminent airway obstruction and were immediately transferred to the operating room to establish airway control (Table 3). Sixteen had successful endotracheal intubation; after failed attempts at intubation, airway control was accomplished by tracheotomy in 2 patients and cricothyrotomy in 1. Six other patients, who initially had no respiratory distress or airway compromise, developed from a few hours to 2 days after admission delayed respiratory distress; to secure their airway, 5 had successful endotracheal intubation and 1 required tracheotomy. Altogether, 25 patients had at some point respiratory obstruction that demanded airway intervention; they were transferred to the intensive care unit and kept there for 4.3 ⫾ 3.4 days until extubation. The mean hospital stay of patients with airway intervention was significantly longer than that of patients who did not experience respiratory distress (mean ⫾ SD, 9.9 ⫾ 6.4 days v 3.9 ⫾ 2 days; P ⬍ .001). Nine of the 25 patients (36%) lacked prominent respiratory symptoms and the decision to secure the airway was based on clinical observation, confirming that at least 50% of the laryngeal lumen was narrowed. Abscess Formation Epiglottic abscess was evident in 24% of the patients (28/118). Interestingly, their proportion was almost 2-fold higher among those who needed airway intervention (44%, 11/ 25). Of the latter, 7 had immediate respiratory obstruction and 4 had delayed obstruction; both went through incision and drainage of abscess under direct laryngoscopy at the time of airway control. Another 11 patients, who had an EA without airway obstruction, underwent incision and drainage of abscess under endotracheal general anesthesia (5 patients on the day of admission and 6 on the second, third, and fourth hospital day) and were extu-
TABLE 3. Airway Intervention With or Without an Abscess
Compromised airway
Uncompromised airway Total
With abscess Without abscess Total With abscess
Immediate Intervention
Delayed Intervention
Total
7 12 19 5 24
4 2 6 6 12
11 14 25 11 36
378
Fig 4. A section of a patient with acute epiglottitis, showing coalescence of an epiglottic abscess with abundance of neutrophils (N) located adjacent to the abscess cavity (C) and fibrin (F) located more peripherally (hematoxylin and eosin, original magnification ⴛ100).
bated after the drainage procedure (Table 3). Altogether, 22 patients required a surgical drainage, whereas the remaining 6 had a spontaneous drainage of the abscess and were conservatively managed with intravenous antibiotics and careful monitoring. Twenty-one abscesses were located along the lingual surface of the epiglottis, 1 toward its laryngeal surface and 6 had no predilection for either direction. Eight patients underwent computed tomography to delineate the confines of the abscess. Nine cultures were retrieved from the abscess cavity and yielded normal pharyngeal flora in 4, positive cultures for Escherichia coli, Enterobacter, Klebsiella pneumonia, and streptococcus in another 4 and no growth of bacteria in 1. At the time of drainage, a punch biopsy of the abscess wall was obtained from 6 patients. The material underwent routine histological processing and was stained with hematoxylin and eosin. Light microscopy confirmed that there were 5 abscesses involving the epiglottis itself, which originated from a coalescence of patchy tissue necroses forming an abscess cavity filled with necrotic white cells and debris and bounded by a wall composed of numerous neutrophils and a peripheral zone of fibrin (Fig 4). In the remaining biopsy, squamous epithelial cells were integrated within the inflamed walls of the abscess (Fig 5), indicating it is an infected mucus retention cyst (mucopyocele) of the base of the tongue that reached the lingual
BERGER ET AL
Fig 5. A section of a patient with AE, showing a mucopyocele with a squamous epithelial cell (arrow) integrated within the inflamed wall (hematoxylin and eosin, original magnification ⴛ100).
aspect of the epiglottis and caused bacterial infection of the supraglottic structure. A computed tomography scan image verified the different origin of the latter and showed an infected mucopyocele of the base of the tongue situated anteriorly to the swollen epiglottis (Fig 6). Predictors of Airway Obstruction Data analysis of patients undergoing airway intervention versus those who were managed conservatively confirmed that the 9 following factors were associated with increased likelihood to develop airway obstruction: (1) age (mean ⫾ SD, 55.8 ⫾ 13.8 years v 48.8 ⫾ 14.9
Fig 6. A computed tomography of a patient with AE injected with intravenous contrast material, showing an infected mucus retention cyst (mucopyocele) of the base of the tongue with enhanced walls (thick arrow) situated anteriorly to the swollen epiglottis (thin arrow).
ADULT ACUTE EPIGLOTTITIS
379
TABLE 4. Epidemiologic Features of Adult AE in Published Studies Author Before Hib vaccination Shih et al MayoSmith et al Carenfelt and Sobin Hugosson et al Kucera et al Murrage et al Wurtele Yardley Berg et al After Hib vaccination Alho et al Mayo-Smith et al Navarrete et al Solomon et al Ryan et al Frantz and Ragson Torkkeli et al Dort et al Takala et al Garpenholt et al Hebert et al Chung
Study Site
Period (Years)
No. of Patients
Incidence/100,000
Secular Trend
Florida* Rhode Island* Sweden Sweden Minnesota* Ontario† Quebec† UK Sweden
1963-1987 1975-1982 1975-1987 1975-1992 1976-1992 1977-1986 1982-1988 1985 1987-1989
48 56 543 93 20 26 12 7 502
NR 0.97 3.9 2.3 4-2.2 0.7 1 NR 2.5
Increasing Increasing Stable Stable Decreasing Increasing Stable Increasing Increasing
Finland Rhode Island* Spain Ontario† Pennsylvania* California* Finland Manitoba† Finland Sweden Ontario† Hong Kong
1967-1993 1975-1992 1978-1990 1979-1991 1979-1991 1980-1990 1981-1992 1982-1992 1985-1992 1987-1996 1989-1994 1998-2001
49 273 58 57 8 258 32 43 20 241 51 11
0.5-2.3 0.78-2.9 NR NR NR 1.8 1.5 NR 0.03 3.4 2.02 NR
Increasing Increasing Increasing Increasing Increasing Stable Increasing Increasing Stable Decreasing Stable Increasing
Abbreviations: NR, not recorded. *United States. † Canada.
years; P ⫽ .036), (2) diabetes mellitus (5 of 10 [50%] v 20 of 108 [19%]; P ⫽ .03), (3) rapid onset (ie, within 16 hours) (6 of 12 [50%] v 19 of 106 [18%]; P ⫽ .022), (4) voice changes (11 of 30 [37%] v 14 of 88 [16%]; P ⫽ .036), (5) drooling (5 of 8 [62.5%] v 20 of 110 [18%]; P ⫽ .011), (6) pulse rate (mean ⫾ SD, 94.2 ⫾ 15.5 beats per minute v 85.2 ⫾ 11.5 beats per minute; P ⫽ .042), (7) temperature (mean ⫾ SD, 37.8° ⫾ 0.9°C v 37.4° ⫾ 0.7°C; P ⫽ .035), (8) white blood cell count (14,625 ⫾ 5,210 cells v 12,310 ⫾ 4,287 cells; P ⫽ .032), and (9) abscess formation (11 of 28 [39%] v 14 of 90 [16%]; P ⫽ .015). Using the stepwise logistic regression analysis, 4 variables remained associated with airway obstruction: (1) drooling (OR, 8.1 [95% CI, 1.5-43.7]), (2) diabetes mellitus (OR, 6.9 [95% CI, 1.6-29.7]), (3) short onset (OR, 4.8 [95% CI, 1.2-19.8]), and (4) abscess (OR, 4.6 [95% CI, 1.5-13.5]). Based on previously determined criteria for evaluation, it was shown that the 4 following factors were not associated with airway obstruction: (1) gender (16 of 63 [25%] for men v 9 of 55 [16%] for women; P ⫽ .265), (2) duration of symptoms (mean ⫾ SD, 52.5 ⫾ 38.5
hours v 68.5 ⫾ 53.1 hours; P ⫽ .162), (3) complaints of respiratory difficulties (12 of 40 [30%] v 13 of 78 [17%]; P ⫽ .152), and (4) preadmission antibiotics (10 of 37 [27%] v 15 of 81 [19%]; P ⫽ .294). DISCUSSION The data show a significant increase in the incidence of the disease among adults in the Sharon area between 1986 and 2000 and agree with other studies on adult AE carried out elsewhere (Table 4). Notably, evidence of such a trend in adults was already apparent before the introduction of large-scale pediatric vaccination against Hib and continued after vaccination, despite the marked diminution of AE in children. This would imply that the vaccine has little effect on nonvaccinated adults. Evidently, further studies are needed to determine the possible effects of Hib vaccine on vaccinated adults. It should be born in mind that Hib, unlike in children, is not the exclusive pathogen in adults and other etiological agents also exist (eg, Streptococcus pneumoniae, Staphylococcus aureus, Kleb-
380
siella pneumoniae). A number of authors questioned whether the apparent increase in adult AE represents a genuine rise in the incidence of the disease and suggested that greater physician awareness and/or improved diagnosis are accountable for this change.10,13 Evidently, the introduction of the flexible fiberoptic nasopharyngoscope turned the visualization of the inflamed epiglottis and the assessment of a possibly compromised airway into a more accurate task than simply relying on soft tissue lateral x-ray of the neck. Yet, it is hardly conceivable that the concurrent increase in the incidence of adult AE in different areas of the world could be explained by the mere presence of better recognition and diagnosis of the disease.16 Our findings have shown that the number of adult patients with AE continued to increase between 1991 and 2000, when fiberoptic nasopharyngoscopy was routinely used, thus supporting the notion that there is a bona fide rise in the incidence of AE. Once established, there is a rise in the incidence of AE; it is of importance to consider the possible causes behind the phenomenon. The last 2 decades have witnessed a sharp increase in the number of human immunodeficiency virus–positive, acquired immunodeficiency syndrome, and other immunocompromised patients; yet except for case reports,28-30 larger series either did not report17 or only had isolated cases of immunocompromised patients with AE.19,31 We showed that 16 of 116 patients (14%) had background diseases of which 10 had diabetes mellitus; no human immunodeficiency virus– positive or acquired immunodeficiency syndrome patients were presented. Thus, so far, the data do not support the assertion that the spread of immunodeficiency states influenced the occurrence of adult AE. However, these conditions enhance the possibility to run a more severe course of the disease, as shown in our group of diabetics or require a longer hospital stay than the average.32 Mayo-Smith et al17 ascribed the rise in the incidence of EA mainly to increased numbers of patients who lack Hib bacteremia and are characterized by negative blood cultures, slow onset of symptoms, and reduced probability to develop airway obstruction. It is also suggested that similar to the significant increase in the incidence
BERGER ET AL
of other common upper airway infectious diseases such as acute otitis media and sinusitis,33 the rise in the incidence of EA may be related to indiscriminate use of antibiotics leading to the emergence and spread of antimicrobial resistance, which increasingly threatens the success of infectious disease treatment and prevention. The essence of AE management is to safeguard patients’ airway and prevent obstruction. Views vary between those advocating prophylactic airway intervention in all patients because of the unpredictable nature of the disease and its potentially swift course10 and those practicing conservative treatment.34 Nevertheless, most authors fall somewhere between these 2 extremes and adopt a selective approach. Thus, patients with severe airway symptoms and signs (eg, dyspnea, stridor, and cyanosis) require immediate establishment of an airway either by intubation or tracheotomy, whereas the less severe cases are conservatively treated in a monitored setting. Proponents of this practice find it safe and effective because it significantly lowers the proportion of patients undergoing airway intervention (Frantz et al,31 15%; Solomon et al,19 16%; Torkkeli et al,16 16%; Hebert et al,9 20%; Andreassen et al,35 22%; Cucera et al,21 25%; Deeb et al,36 30%, and Dort et al,15 33%) and avoids the increased risk of mortality rates. Two prominent questions remain: who will eventually progress to acute airway obstruction and at what point should airway be established? One must remember that in these intricate situations, decisions are often made under uncertainty. To guide the caring otolaryngologist when contemplating airway intervention, 4 predictors of impending airway obstruction are suggested: (1) drooling, (2) a history of diabetes mellitus, (3) rapid onset of symptoms, and (4) abscess formation. A number of authors have also suggested drooling, rapid onset, and EA as predictors of airway obstruction.12,17,19,36 In addition, our study has pointed out a significant relationship between diabetes mellitus and airway obstruction. An exception to the selective management approach pertains to a category of patients who lack prominent respiratory symptoms, yet their laryngeal inlet is narrowed by at least 50%.9,32 The condition should be regarded as
ADULT ACUTE EPIGLOTTITIS
an imminent airway obstruction that requires immediate intervention, thus avoiding the extremely difficult management during acute respiratory distress. The data show that in 9 out of 25 patients (36%) of the current study the decision to accomplish airway intervention was based on similar circumstances. Apparently, the use of the flexible fiberoptic nasopharyngoscopy allows an accurate estimate of the laryngeal airway and made it easier to detect these patients. On the other hand, the exclusive reliance on indirect laryngoscopy or soft-tissue lateral x-ray of the neck may provide incomplete view of the laryngeal inlet and patients’ condition can go unnoticed. It is estimated that patients, who in the early stages of the disease appeared to do well and later developed a severe or a sudden fatal airway obstruction,10 in fact, belong to this category. AE is commonly regarded as a self-limiting disease with an increased risk of developing a rapid and a fulminant course. According to MayoSmith et al,10 all deaths happened before or within 6 hours of admission. This finding supports other observations on patients, who were managed conservatively and did not subsequently required urgent airway intervention until bacterial infection had resolved.19,34,36-38 The present series, on the other hand, uncovered that 6 patients had delayed airway obstruction occurring up to 2 days of admission, of whom 4 were associated with an EA. Shih et al12 and Andreassen et al35 also found a considerable number of delayed obstructions, stressing the capricious nature of AE and the need for close and prolonged observation until swelling is significantly reduced. Corticosteroids are often administrated in an effort to reduce supraglottic inflammation, the likelihood of airway obstruction, complication rate, and length of hospital stay; yet, there is lack of controlled data demonstrating unequivocal evidence of the benefit of this treatment.9,13,15,17,31,32,34 Likewise, in the present series, the mean hospital stay of patients receiving corticosteroids and those not receiving corticosteroids did not differ significantly (5.7 ⫾ 4.1 days compared with 4.7 ⫾ 4.3 days, respectively; P ⫽ .2). Moreover, the average stay of the former exceeded the latter
381
by 1 day, probably reflecting the clinical selection of the more severe cases. Thirty-seven patients (31%) were examined by their family physicians and received antibiotic treatment before hospitalization. Other also reported a similar proportion of patients who had been seen by their primary care physician within 48 hours of diagnosis (32% and 29%, respectively).10,31 Moreover, the findings clearly show that the protective effect of preadmission antibiotics was inadequate (27% of the patients who received antibiotics had upper airway obstruction compared with 19% who did not receive antibiotic treatment before admission [P ⫽ .294]). This high proportion of missed diagnosis that can result in increased morbidity and even mortality highlights the need to provide specific guidance to family physicians and stresses the importance of immediate referral of patients with severe sore throat and odynophagia (especially when examination of the oropharynx is normal) to an otolaryngologist for further evaluation. The study revealed a high incidence of EA (28/118, 24%) among patients with AE. Several authors found similar rates of EA (Shih et al,12 21%; Wolf et al,34 26%; and Hawkins et al,26 29%); yet, others did not report the phenomenon9,21 or had much lesser rates (MayoSmith et al,10 2%; Stanley and Liang,37 2%; Solomon et al,19 2%; Hugosson et al,8 3%; MayoSmith et al,17 3%; Frantz et al,31 4%; Torkkeli et al,16 6%; and Murrage et al,13 8%). It is also shown that the number of EAs increased concomitantly with the rise in the incidence of AE (from 4/14 episodes [29%] between 1986 and 1990, to 8/38 [21%] between 1991 and 1995, and to 16/66 [24%] between 1996 and 2000), disclosing a relatively constant ratio between both phenomena and insignificant variations among the 3 periods (P ⫽ .843). This may lend support to the contingency that abscess formation considerably prevails in AE. Similar to Heeneman and Ward,22 we observed that the lingual surface of the epiglottis was the most commonly involved site of the EA. Concerning the origin the EA, our clinical and pathological data lend support to the Stack and Ridley24 assumptions that it originates either from coalescent epiglottic infection or from infection of the epiglottic mucocele. The former is linked to severe infection that causes patchy
382
necrosis with accumulation of numerous leucocytes in the area. The latter refers to an infected mucus retention cyst of the base of tongue or the vallecula that reaches the lingual aspect of the epiglottis and in contrast to the former has an epithelial wall. CONCLUSIONS Our 15-year experience from 1986 through 2000 suggests that (1) the incidence of adult AE increased significantly, a secular trend probably unrelated to Haemophilus influenzae type b infection, but to other pathogenic bacteria; (2) a selective management approach, which advocates airway intervention only in patients with severe airway symptoms or with imminent airway obstruction, yielded a relatively low rate of interventions (21%) and no fatalities; (3) although most patients requiring airway intervention demonstrated compromised airway on admission, a considerable number may develop delayed upper airway obstruction; (4) a significant proportion of the cases with compromised airway was related to EA formation; (5) the number of epiglottic abscesses increased concomitantly with the rise in the incidence of AE; and (6) 2 diverse pathophysiologic pathways of EA formation are proposed. ACKNOWLEDGMENT We thank Estela Derazne, MSc, from the Lady Sarah Cohen Unit for Family Centered Therapy and Healthcare, the Schneider Children’s Medical Center of Israel, Petah Tikva, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel, for assistance with statistical analysis; and Rachel Berger, BA, for writing and editing assistance. REFERENCES 1. Frantz TD, Rasgon BM: Acute epiglottitis: changing epidemiologic patterns. Otolaryngol Head Neck Surg 109: 457-460, 1993 2. Ryan M, Hunt M, Snowberger T: A changing pattern of epiglottitis. Clin Pediatr 31:532-535, 1992 3. Takala AK, Peltola H, Eskola J: Disappearance of epiglottitis during large-scale vaccination with Haemophilus influenzae type b conjugate vaccine among children in Finland. Laryngoscope 104:731-735, 1994 4. Alho OP, Jokinen K, Pirila T, et al: Acute epiglottitis and infant conjugate Haemophilus influenzae type b vac-
BERGER ET AL
cination in northern Finland. Arch Otolaryngol Head Neck Surg 121:898-902, 1995 ¨ , Hugosson S, Fredlund H, et al: Epi5. Garpenholt O glottitis in Sweden before and after introduction of vaccination against Haemophilus influenzae type b. Pediatr Infect Dis J 18:490-493, 1999 6. Carenfelt C, Sobin A: Acute infectious epiglottitis in children and adults: annual incidence and mortality. Clin Otolaryngol 14:489-493, 1989 7. Wurtele P: Acute epiglottitis in children and adults: A large-scale incidence study. Otolaryngol Head Neck Surg 103:902-908, 1990 8. Hugosson S, Olce´ n P, Ekedahl C: Acute epiglottitis— Aetiology, epidemiology and outcome in a population before large-scale Haemophilus influenzae type b vaccination. Clin Otolaryngol 19:441-445, 1994 9. Hebert PC, Ducic Y, Boisvert D, et al: Adult epiglottitis in a Canadian setting. Laryngoscope 108:64-69, 1998 10. MayoSmith MF, Hirsch PJ, Wodzinski SF, et al: Acute epiglottitis in adults. An eight-year experience in the state of Rhode Island. N Engl J Med 314:1133-1139, 1986 11. Yardley MH: Supraglottitis in adults. Practitioner 230:583-587, 1986 12. Shih L, Hawkins DB, Stanley RB Jr: Acute epiglottitis in adults. A review of 48 cases. Ann Otol Rhinol Laryngol 97:527-529, 1988 13. Murrage KJ, Janzen VD, Ruby RR: Epiglottitis: Adult and pediatric comparisons. J Otolaryngol 17:194198, 1988 14. Navarrete ML, Quesada P, Garcia M, et al: Acute epiglottitis in the adult. J Laryngol Otol 105:839-841, 1991 15. Dort JC, Frohlich AM, Tate RB: Acute epiglottitis in adults: diagnosis and treatment in 43 patients. J Otolaryngol 23:281-285, 1994 16. Torkkeli T, Ruoppi P, Nuutinen J, et al: A Changed clinical course and current treatment of acute epiglottitis in adults a 12-year experience. Laryngoscope 104:15031506, 1994 17. Mayo-Smith MF, Spinale JW, Donskey CJ, et al: Acute epiglottitis. An 18-year experience in Rhode Island. Chest 108:1640-1647, 1995 18. Berg S, Trollfors B, Nylen O, et al: Incidence, aetiology, and prognosis of acute epiglottitis in children and adults in Sweden. Scand J Infect Dis 28:261-264, 1996 19. Solomon P, Weisbrod M, Irish JC, et al: Adult epiglottitis: The Toronto Hospital experience. J. Otolaryngol 27:332-336, 1998 20. Chung CH: Case and literature review: Adult acute epiglottitis—Rising incidence or increasing awareness? Hong Kong J Emerg Med 8:227-231, 2001 21. Kucera CM, Silverstein MD, Jacobson RM, et al: Epiglottitis in adults and children in Olmsted County, Minnesota, 1976 through 1990. Mayo Clin Proc 71:11551161, 1996 22. Heeneman H, Ward KM: Epiglottic abscess: Its occurrence and management. J Otolaryngol 6:31-36, 1977 23. Ridgeway NA, Perlman PE, Verghese A, et al: Epiglottic abscess due to group B Streptococcus. Communication. Ann Otol Rhinol Laryngol 93:277-278, 1984 24. Stack BC, Ridley MB: Epiglottic abscess. Case report. Head Neck 17:263-265, 1995 25. Galli J, Corina L, Calo L, et al: Acute supraglottitis in adults: report of a case with development of abscess. Int J Surg Sci 6:59-61, 1999 26. Hawkins DB, Miller AH, Sachs GB, et al: Acute epiglottitis in adults. Laryngoscope 83:1211-1220, 1973 27. Central Bureau of Statistics: Statistical Abstract of Israel. No. 37, 1986; No. 42, 1991; No. 47, 1996; No. 52, 2001
ADULT ACUTE EPIGLOTTITIS
28. Rothstein SG, Persky MS, Edelman BA, et al: Epiglottitis in AIDS patients. Laryngoscope 99:389-392, 1989 29. Laing RB, Wardrop PJ, Welsby PD, et al: Stridor in patients with HIV infection. J Laryngol Otol 109:11971199, 1995 30. Colman MF: Epiglottitis in immunocompromised patients. Head Neck Surg 8:466-468, 1986 31. Frantz TD, Rasgon BM, Quesenberry CP Jr: Acute epiglottitis in adults. Analysis of 129 cases. JAMA 272: 1358-1360, 1994 32. Chan KO, Pang YT, Tan KK: Acute epiglottitis in the tropics: Is it an adult disease? J Laryngol Otol 115: 715-718, 2001 33. McCaig LF, Hughes JM: Trends in antimicrobial drug prescribing among office-based physicians in the United States. JAMA 273:214-219, 1995
383
34. Wolf M, Strauss B, Kronenberg J, et al: Conservative management of adult epiglottitis. Laryngoscope 100:183185, 1990 35. Andreassen UK, Baer S, Nielsen TG, et al: Acute epiglottitis—25 years experience with nasotracheal intubation, current management policy and future trends. J Laryngol Otol 106:1072-1075, 1992 36. Deeb ZE, Yenson AC, DeFries HO: Acute epiglottitis in the adult. Laryngoscope 95:289-291, 1985 37. Stanley RE, Liang TS: Acute epiglottitis in adults (the Singapore experience). J Laryngol Otol 102:10171021, 1988 38. Arndal H, Andreassen UK: Acute epiglottis in children and adults. Nasotracheal intubation, tracheostomy or careful observation? Current status in Scandinavia. J Laryngol Otol 102:1012-1016, 1988