Epiglottitis in Canada : A multiregional review

Epiglottitis is an acute. life threatening infection usually caused by Haemophilus infiuenzae type b. Although antibiotic therapy is an important part of management. the optimal route and duration is unknown. A multicentre retrospective review of 305 children wHh epiglottitis was carried out in order to relate antibiotic therapy to hospital course and outcome. as well as to examine regional variation in patient demographics. clinical presentation and course of disease. A standardized form was used to extract data from hospital records. Although management varied significanlly among the sLx centres in terms of mean duration of intubation (46 to 81 h). intravenous antibiotic therapy (3.8to 5. 7 days) and hospital stay (5.3 to 8.4 days). there were no s ignificant centre-related differences in epidemiology. clinical course or outcome of cpiglollitis. An EPIGLOTI'ITIS IS AN ACUTE. LIFE THREATENING ILLness caused almost exclusively by Haemophilus injluenzae type b . The majority of cases are in children younger than five years of age, with a peak incidence at age two to four years ( 11 7). Respiratory distress with impending acute obWinnipeg Children's Hospital. Department of Pediatrics and Child Health. Uniuersity of Manitoba. Winnipeg. Manitoba: Montreal Children ·s Hospital. Department qf Pediatrics. McGill Uniuersity. Montreal. Quebec: Jzaac Walton Killam Jlospital for Children. Department of Pediatrics. Dalhousie Uniuersity. Halifax. Noua ScaLia: Children ·s Hospital of East em Ontario. Department of Pediatrics. Uniuersity of Ottawa. Ottawa. Ontario; Uniuersity Hospital. Department of Pediatrics. Uniuersity of Saskatchewan. Saskatoon. Saskatchewan: and Charles A Janeway Child Health Centre. Department qf Pediatrics. Memorial University. StJohn ·s. Newfottndland Correspondence and reprints: Dr BJ Law. Department of Medical Microbiology. Uniuersity of Manitoba. Room 530. 730 William Auenue. Winnipeg. Manitoba R3E OW3. Telephone (204) 788·6629 Receiuedfor publication February 13. 1990. Accepted April 3. /990 CAN J INFECT DIS V OL 1 No 1 MAY 1990 exlraepiglotlic focus of infection was present in 15% of patients and included three with septic arthrttis and one with meningitis. The duration of fever in hospital and maximum recorded temperature in hospital were s ignificanlly g reater for children with extraepiglollic infection compared to those \vilh epiglottitis alone. The data presented in this review suggest that most children with epiglottitis have an uncomplicated course and respond ra pidly to antimicrobial therapy follo\ving airway securement. A short period of intravenous and oral antibiotic therapy is likely adequate for most children \'lith epiglottitis. A well designed multicenlre prospective Liial is still needed to determine the optimal duration of antibiotic therapy. Can J Infect Dis 1990;1(1):15-22

exlraepiglotlic foc us of infection was present in 15% of patients and included three with septic arthrttis and one with meningitis.The duration of fever in hospital and maximum recorded temperature in hospital were s ignificanlly g reater for children with extraepiglollic infection compared to those \vilh epiglottitis alone.The data presented in this review s uggest that most children with epiglottitis have an uncomplicated course and respond ra pidly to antimicrobial therapy follo\ving airway securement.A short period of intravenous and oral antibiotic therapy is likely adequate for most children \'lith epiglottitis.A well designed multicenlre prospective Liial is still needed to determine the optimal duration of antibiotic therapy.Can J Infect Dis 1990;1(1):15-22 Key Words: Epiglottitis.Haemophilus influenzae type b struction is the most notable clinical presentation, but evidence of systemic infection \vith fever and leukocytosis is also present in the majority of cases.Although securing an airway is essential in the immediate management of epiglottitis.the use of parenteral antibiotics is also a mainstay of therapy since bacteremia is documented in 60 to 90% of cases.
The appropriate duration of total as well as parenteral antibiotic therapy for epiglottitis has not been addressed in controlled randomized prospective studies.A search of the literature revealed a variety of suggested therapeutic regimens providing a total antibiotic course of seven to 10 days (5.8.12,13).Specific recommendations for duration of intravenous therapy varied from 24 h after e ither admission (18) or extubation (19).to seven days (20)(21)(22).Data supporting these approaches were not found.Prior to the development of a prospective therapeutic trial.a retrospective study in six different Canadian centres was con-

PATIENTS AND METHODS
The hospital records of children with a discharge diagnosis of epiglottitis were reviewed in the following cen trcs for the years as specified: Charles A Janeway Child Health Centre Data were extracted from each record using a standardized reporting form which included demographic data.presenting symptoms includ-ing initial management prior to arrival at the admitting hospital.management and course while in hospital.laboratory investigations.subsequent management and follow-up a fte r discharge from hospital.
Definite epigloltitis was defined as a characteristic clinical presentation with direct visual evidence of acute inflammation of the epiglottis a nd /or a positive blood culture for H influenzae type b.Probable epiglottitis was defined as typical c linical symptoms with eviden ce of an enlarged epiglottis or arycpigloltis on radiograph of the lateral neck in the absence of both direct visualization of th e epiglottis and a positive blood culture.Supraglotlitis was defined as a typical clinical syndrome in which.on direct visual inspection.the epiglottis was normal but the aryepiglollis and surrounding area were inflamed and edematous.irrespective of culture results.In order to determine duration of antibiotics given in hospital by intravenous as well as oral routes.the actual number of doses of each antibiotic was counted and indicated on the data form along with the dosing interval.The duration for each drug by a specific route was defined as the number of doses multiplied by one twenty-fourth of the dosing interval.The duration of antibiotics following discharge was estimated from the interval specified on discharge orders.
Standard statistical methods were used to analyze differences between centres with respect to presentation of illness.results oflaboratory investigations and management in hospital.

RESULTS
Patient population/ demographic variables: A total of 311 charts were re,'iewed: 31 from St John•s.61 from I lalifax.66 from Montreal.47 from Ottawa.79 from Winnipeg and 27 from Saskatoon.Six cases were eliminated on the basis of age ( 16 years or older) or diagnosis: three adu lts from Saskatoon (the only centre which admitted adulls).and three cases in which the diagnosis of epiglottitis was not confirmed on chart review.TI1ese included one case with croup (llalifax).one with retropharyngeal abscess (Winnipeg) and one wilh suspected epiglottitis who was hospitalized for less than 24 hand in whom chest and lateral neck radiographs were nonnal and blood cultures negative (Saskatoon) .Of the remaining 305 cases.291 were drfined as definite epiglottitis.sLx as probable epiglottitis and eight as supraglotlitis.
The mean +standard deviation fo r patient age was 44.8:±-26.6 months with a range of four to 176 months.The peak age of occurrence was from 25 to 48 months (48% of cases).The age distribution did not vary significantly between centres.The proportion of cases potentially preventable by administration of II injluenzae type b conjugate vaccine at 18 months of age was 92% overall with a range of 8 1 to lOO% between centres.
Therapy for epiglottitis was initiated at a location other than the admitting hospital for 1 18 (39%) children .including another h ospital for 109.a physician•s office for six and •not stated• in three.Among the 109 patients first seen at a referring hospital.airway management prior to transfer included none in 39%.intubation in 51 0/o and tracheotomy in 5%: 12% were reported to have had respiratory atTest.Thr proportion of referred patients varied signifieanlly between centres: Montreal 18% : Ottawa 38°1o: St John•s 39°1o: l lalifax 40%: Winnipeg 50%: and Saskatoon 560;o (x 2 =35: P<O.OOOl).The initial assessment was performed at the admitting hospital for 187 (61 °1o) patients.Respiratory arrest occurred in 2%.In the majority of instances.the patients were transferred immediately to th e intensive ca re unit or operati ng room.a nd th us there was little infommlion as to status when seen in the emergency dcpariment.Fever (temperature greater than 38oC) was documented in 60%.Airway management i n hospital: Overall.93% of cases were intubated.3% had a tracheostomy and 4% were observed.None of the patients who were initially observed required intubation.The majority of int ubations were done in the ope rating room.primarily by an anesthetist.Accidental extubation occutT<.'cl in nin e cases (3%).seven of which were rc-intubated.No complications secondary to a ccidental extubation were reported .
The centre-specific lime to extubation is shown in Figure 2. The mean duration of intubation was Post extubation complications included: stridor in 21% .atelectasis in 4%, pneumonia in 2% and wheeze in 1% .Post extubation s tridor was reported more frequenlly in children who wer e reported to have had a two stage oral then nasal intubation (30% of 122) vers us children with a single orolracheaJ (9% of 11) or nasotracheal ( 12% of 115) intubation procedure (x 2 =12.75;P=0.002).Stridor was mild.however.and responded in all cases to medical manage me nt consisting of mist l11erapy and/or racemic adrenaline.Other complications included one vallecular cyst.one laryngeal ulcer and one bruised larynx.

Investigations in hospital:
A white blood cell (WBC) count was determined shortly after admiss ion for 280 (92%) patients.Leuk ocytosis defined as total WBC greater than 10.0 x 10 9 /L was present in 89% of patients.The elevation in total WBC was moderate to marked in most patients: 15 to 20 x 10 9 /Lin 27% : 21 to 25 in 22%: 25 to 30 in 13% : and more than 30 in 7%.The absolute band count was less than 0.5 x 10 9 / Lin 15%.0.5 to less than 1.0 in 9%.1.0 to 5.0 in 53% and more than 5.0 in 23%.A plate let count was recorded for 71 % of cases and was nom1al in 99%.The mean hemoglobin was 120 ±10 g/L.A value of less t han 110 was noted for 15%.
Blood cultures were obtained for 292 (96%) pa tients and were posit ive in 82%.All but three of the 241 isolates were H injluenzae (99%).The exceptions (grou p B beta-hemolytic streptococcus: Haemophilus parai'1fluenzae: p eptococcus) a ll occurred as single b lood isolates in three Montreal patients.All but eight of the H injluenzae isolates were typed , and of t hese, a ll were capsular type b.Receipt of antibiotics prior to admission s ignificantly decreased the likelihood that a blood culture would be positive.Only 56% (40 of 7 1) of patients given antibiotics prior to admission had positive b lood cultures versu s 91% (200 of220) of those who had not received prior treatment (x 2 =42.0:P<O.OOOl).Urine for rapid antigen detection was obtained from only 28 patients and was positive for H injluenzae type b in 14.Test results s u ggested an e tiology of H injluenzae in four pa tients with negative blood cultures.Among H i'1fluenzae type b isolates.24% were res istant to ampicillin.There were no significant centre-specific differences in the prevalence of ampicillin resistance: Halifax 20% (nine of 46).St John's 36% (eight of22).Monlreal 24% (14 of 55), Ottawa 28% (10 of 36).Saskatoon 20% (one of five).and Winnipeg 23% (16 of 69).
Cultures of the epiglottis wer e obtained in 207 pa tients.H injlLLenzae was recovered from 62% .H parainjluenzae from 8%.Staphylococcus aureus from 11%. group A beta-hem olytic streptococcus from 3%. group B beta-hemolytic streptococcus from 0.5% and mixed normal pharyngeal flora from 14%.There was no growth in 15%.Beta-lactamase activity was determined for paired b lood and epiglottis isolates in 58 of the 1 18 cases in which H injluenzae was isolated from both sites.Concordant results were obtained for 57 of 58 (98%).In one case.the blood isolate was beta-lactamase positive.whereas the epiglottis isolate was beta-lactamase negative.Most epiglottis H injluenzae isolates were not serotyped.and thus correlation with blood culture results could not be assessed.
Radiographs of the lateral n eck were obtained for 243 patients (80%) .Abnormalities were noted in 99% with findings typical of epiglottitis in 91%.aryepiglottic swelling only in 4%.subglottic swelling in 1 o/o and uninterpretable changes in 3%.Antibio tic therapy: Al l but one patient received at least one dose of antibiotic intravenously while in hospital.Empiric therapy on admission consisted of ampicillin and chloramphenicol in 8 1%.chloramphenicol alone in 10%.ampicillin a lone in 1%.cefurox:ime in 5%, cefamandole in 1 o/o and miscellaneous in 1% (one ampicillin/cefotax:ime. one ampicillin/cloxacillin.and one chloramphenicol/cloxacillin).A s ingle patient was started on oral amoxicillin.This patient was observed in hospital without intubation and had no blood culture taken.The diagnosis was based on a lateral n eck radiograph which showed swollen aryepiglottic folds and a slightly swollen epiglottis.
Eight patients were excluded from the analysis of duration of antibiotic therapy: four who died early in the hospital course and four with H injluenzae bactere mia in whom the presence of meningitis (one patien t) or septic arthritis (three patients) determined the total du ration of therapy.
For the group as a whole.the mean ± SO duration of intravenous antibiotics was 4.8 ± 2.2 days (range 1.3 to 18.5 days): hospital-monitored antibiotics (intravenous + oral) was 6.0 ± 2.3 days (range two to 19.5 days); and of estimated total antibiotics (hospital + p lanned discharge medications) was 9.9 ± 2.8 days (range two to 19.8 days).Oral antibiotics were started in hospital for 69% of patients; these included amoxicillin in 62%.chloramphenicol in 27%.ampicillin in 7%.cefaclor in 3% and an unspecified antistaphylococcal penicillin in 1%.Orders for oral antibiotics following discharge were written for 73% of patients.The Epiglottitis in Canada duration was specified for 186 of 209 patients (89%) and averaged 5.2 ± 2 .8days.The drugs u sed were similar to those started in hospital.
Centre variation in duration of antibiotic therapy was analyzed for hospital-monitored therapy only.as compliance with discharge medication could not be determined.The duration (mean ± SO) of intravenous antibiotics was significantly shorter in Winnipeg (3 .8± 1.4 days) t han in Montreal (5.7 ± 2.2 days) .Ottawa (5.0 ± 1.5 days) and St John's (5.3 ± 2. 7 days) (P<0.05 by Tukey's studenlized range test).The durations in Halifax (4.8 ± 2.6 days) and Saskatoon (4.7 ± 2.0 days) were not significantly different from the other centres.The centre-specific cumulative duration of intravenous antibiotics is shown in Figure 2 .The mean duration of intravenous plus oral antibiotics administered in hospital was also significantly shorter for patients from Winn ipeg (5.4 ± 1.5 days) compared with all other centres (6.6 ± 2.9; 6 .7 ± 2.8: 6.6 ± 2.1: and 6.8 ± 1.4 days) except Saskatoon (5.7 ± 1.8 days).The presence or absence of documented bacteremia had no s ignificant effect on the duration of either intravenous (4.9 ± 2 .2 and 4.2 ± 1.8 days.respectively) or total hospital monitored antibiotics (6.1 ± 2.3 and 5.7 ± 2.2.respectively).Hospital c ourse : During hospitalization.fever (temperature greater than 38°C irrespective of site of measurement) was documented in 88% of patients.The maximum recorded temperature in hospital occurred on the first day for 84% of patients.the second day for 10% and the third day or later for 6%.The mean duration of fever (defined as first hospital day afebrile minus one) was 1.9 ± 1.3 days overall.
The discharge diagnosis indicated an extraepiglottic focus of infection in 46 patients (15%) including 25 pneumonia: 13 otitis media; three septic arthritis (one with associated osteomyelitis): three cervical adenitis: one pharyngitis; one mediastinitis: one submandibular cellulitis; and one meningitis.Clinical data for the children with septic arthritis and meningitis are summariZed in Table 2.All four children had bacteremia with beta-lactamase negative H i'1fluenzae. and none had received antibiotics prior to admission.The duration of fever was significanUy longer (ANOVA: P<0.0001) in th e children with a second site of infection (2. 9 ± 1. 7 days) than in children with epiglollitis only (1.9 ± 1.1 days).The maximum recorded temperature in hospital equalled or exceeded 40°C in 33% of children with an extraepiglottic s ite of infection versus 13% of children with epiglottitis only (x 2 test.P=0.002).The admitting leukocyte count was not helpful in identifying children with a second focus of infection.Four ch ildren died.for a case fatality rate of 1.3%.In all four cases.cardiorespiratory arrest occurred prior to transfer and the triage s ite was 120 km or more from lhe admitting hospita l.Respiratory arrest occurred in two of the children at time of presentation a nd.despi te resuscitation.death occurred within 5 a nd 18 h of admission to hospital.Post mortem information was not available.One chi ld had a la rge tens ion pneumothorax and subcutaneous emphysema of the neck.The other two had been managed as croup at a referring hospital for severa l hours prior to acute deterioration and transfer.Bolh of these children died two days after tra nsfer.Post mortem examina tion confirmed the diagnosis of epiglottitis a nd provided evidence of hypoxic enceph a lopathy a nd pulmonary edema.
Complications noted among survival-s included one bleeding stress ulcer.one feb1ile seizure.and on e case of cerebral irritability related to a postopera tive m ed ication .

DISCUSSION
Despite the geographic separation a nd regiona l differences between the six centres included in this study.patient demographics.c linical presentation .hospital course a nd outcome of acute epiglottitis.admitting la bora tory data includ in g complete blood count a nd radiography.a nd etiology including preva le nce of ampici llin -resis tant H in.fluenzae type b were remarkably s imila r.A recent review from Toronto's l lospita l for S ick C hildren presented a s imilar clinical picture (12) .These results were n ot unexpected given the uniformity of presentation and course reported in review a rticles from severa l different countries.including the United States (2 -4.9-1 I. l 7. I9).England (J 8 ).Austra lia ( I4) .New Zeala nd (7).Switzerland (8} and Sweden (15) .
Th e primary focu s of this review was to document the duration of intravenou s antibiotic therapy currently practised in several Canadian centres a nd to determine whether or not a very short course of intravenous antibiotic was assoc ia ted with recurrent infection.Signifi cant centre -20 re la ted differences in management were a pparent.with the s hortest course of both intraven ou s a nd hospita l-monitored a ntibiotic therapy employed in Winnipeg.Among 78 patients treated there.63 and 69%, respectively received intravenous antib iotics for at most three days and s tayed in hos pital for no more than five d ays.Th e cor responding figu res for the entire series were 36 and 3 1%.respective ly.
Recurrences of epiglottitis or other s ites of infection were n ot documented for a ny of the children in this series.Minor infeclions treated a t a physician's office or more serious infections managed in a differe nt hos pita l cou ld.however.have been missed.Rctu m visits a fter hospitalization for epiglottitis were recorded for 46% of the c hildre n reviewed.Proble ms potentially re lated to the acute illness were noted for only 2% of this group.a nd none of these were for recurrence of infection.A review of the literature revealed only one ins ta nce of read mission for recurrent ep iglottitis.This was one of 349 cases reported from Australia (1 4).The recurrence occurred in a ch ild who ha d been in hospital for 24 h only.
The optimal duration of parenteral a nd hos pital-monitored a ntibiolic therapy is unknown.Factors which may contribute to complications include the presence a nd m agnitude of lhe bacteremia.as well as lhc presen ce or a bsence of a second focus of infection.Nearly a ll of the c hildren reviewed ha d b lood cu ltures drawn.and 82% had documented bact eremia.For children who received no oral a ntibiotics prior to admission.lhe prevalence of bacteremia was even highe r a t 9 1%.This is in agreement with the currenlly held view that most cases of clinical epiglottitis in chi ldren a re due to I-I iqfluenzae type b. and that bacteremia is probably universa lly present a t some time during the course of the illness.The magnitude of bacteremia in epiglottitis is 10 to I 000 colony forming u nits per mL.which is I to 10% of tha t found in I-I injluenzae type b meningitis (23) .This may expla in lhe reported range of 50 to 9 0% for positive blood c ultures in epiglottitis.as well as the present observation of a significantly lower prevalence of documented bacteremia (56%) among the 71 ch ildren in this series who received oral antibiotics prior to admission.J<'urthennore. the rarity of reported concurrent sites of invasive infection -including meningitis.septic arthritis and cellulitis (24.25) -support the concept that bacteremia is not a major part of the paU1ogenesis of epiglotlitis but probably develops secondarily to the marked degree of inflammation of the epiglottis and surrounding structures in conjunction with mucosal colonization.This line of reasoning would support a relatively short duration not only of parenteral but also of total antibiotic therapy.The majority of the 233 (76%) children in this series who had no extraepiglottic sites of infection identified during the hospital course had a very rapid defervescence over two lo four days.In this subgroup.one or two daily doses of an antibiotic with a long half-life.such as ceftria.xone.given for the duration of endotracheal tube placement and followed by a few days of oral antibiotics.may be sufncicnt.
For children with a second focus of infection.the length and type of antibiotic therapy would depend upon the specific site of ex.1.raepigiotticinfection.As noted above.most reviews of epiglottitis stress the low frequency of secondary sites of infcclion.The most extensive review to date was done by Molteni (24) who reported a 25% incidence of both pneumonia and cervical lymphadenitis.8% pharyngiUs and 5% otitis media among 72 patients from his own institution.The reported incidence of concurrent pneumonia ranges from I to 50%.Unfortunately. the validity of a clinical diagnosis of pneumonia versus atelectasis is impossible to establish in a retrospective review.For the 305 patients in th is review.a diagnosis of pneumonia was based on a slated discharge diagnosis as well as x-ray findings.and thus the observed rate of 8% may have underestimated the true incidence.
There have been at least one dozen reported cases of concurrent epiglottitis and meningitis.To the authors• knowledge.septic arthritis and osteomyelitis have not previously been reported as accompanying epiglottitis.In one of the three children in the present series.the diagnosis was only suspected after six days in hospital.at which lime the patient was afebrile: the hospital course had been uncomplicated.The other two children with septic arthritis and the one with meningitis had unusual hospital courses with fever which was either slow to resolve or recurrent after an afebrile period.These cases.although infrequent.must be considered In determining the optimal duration and route of antibiotic therapy and the period of observation in hospital.Investigation to nile out a Epiglottitis in Canada second focus of in fcclion shou ld be carefully considered in children with fever exceeding 40°C or persisting four days or more after admission to hospital.
The data gathered in th is review raise questions about issues other than rate and duration of a ntibiotic therapy.As shown in Table 1. the reported prevalence of p resenting symptoms for epiglotlilis varies over a wide range even for the classically recognized tetrad of dyspnea.drool.dysphagia and altered voice.This finding likely follows from the unreliable nature of retrospective reviews.In most reports.it is impossible to tell whether or not the absence of a given symptom relates to absence of recorded data or absence of a clinical finding.A recent report of four cases of 'atypical' epiglottitis stressed that children may present with a more gradual illness evolving over several days with symptoms.such as cough and hoarse voice.that arc more usually associated \vith croup (26).Although acute onset of respiratory distress occurred in most children in this ser•ies.several had been ill for two to five days and 24% of the group had been seen by a physician prior to presentation at hospital with an illness that was not diagnosed as cpiglotutis.Furthermore, one of the four deaths occun•ed in a chi ld who had a sudden respiratory arrest while being managed as croup at a referring hospital.A prospective study of epiglottitis would give a much clearer p icture of the natural h istory immediately preceding the acute onset of respiratory distress.Nevertheless.the data presented here serves to emphasize that an •atypical' presentation of epiglottitis evolving over a few days is perhaps more common than previously acknowledged.
Most of the data from this series regarding airway management.including the lack of significant com plical ions associated with intubation. is consistent with previous reports.The increased incidence of post extubation stt;dor associated with a two stage orotracheal to nasotracheal intubation was unexpected.The retrospective nature of the study with lack of an objective definition of stridor makes any conclusion to be drawn from this finding tenuous.Nevertheless.given the degree of inflammation and edema which accompanies acute epiglottitis.the possibility of greater damage with a two stage intubation procedure deserves further study.
The wide variety of bacterial isolates recovered from cullure swabs of the epiglottis as well as the one instance in which an an1picillin-sensitive H injluenzae was isolated from the epiglottis whereas an ampicillin-resistant 11 i'1fluenzae was recovered from blood.underscores the lack of reliability of epiglottis cultures for determining antibiotic c hoice.In the absence of a blood isolate.management with a beta-lactamase stable antibiotic should be continued, irrespective of the sensitivity of isolates from the epiglottis.
In conclusion.this review provides baseline data on the clinical presentation and course of epiglottitis in children from several regions of Canada.As with other retrospective reviews of epiglottitis.it raises many questions.Despite a multitude of reviews of epiglottitis over the past four decades involving nearly 4500 children.there remain many unresolved issues surrou nding the pathogenesis.natural history and management of epiglottitis.These could potentially be addressed in a multicentre prospective therapeutic u;al com-paring a short course of intravenous antibiotics alone to sequential intravenous-oral combined therapy for seven to 10 days.Of note is that 92% of cases in this series were potentially preventable with the administration of conjugate H injluenzae type b vaccine at 18 months of age.This vaccine was not in use during the period covered by this review.Thus. it is possible that the incidence of epiglottitis in Canada will be substantially reduced by the introduction of this vaccine, making a large therapeutic study difficult.Perhaps a more important issue than optimal duration of therapy is the need for early recognition and appropriate emergency management in the hopes of preventing deaths secondary to epiglottitis.

TABlE 1 Prevalence of presenting symptoms in Canadian children with epigloHitis
INFECt D1s VoL 1 No 1 M Av 1990 CAN J

TABLE 2
Clinical data for four c hildren with a second invasive focus of infection accompanying epiglottitis