- Email: [email protected]
Bacteriology of tonsil and adenoid and sampling techniques of adenoidal bacteriology
Respiratory
Medicine
(1993) 87,303-308
Bacteriology of tonsil and adenoid and sampling techniques of adenoidal bacteriology R. J. GAFFNEY*,
C. I. TIMON?, D. F. FREEMANS, M.
A. WALSH*
AND
M. T. CAFFERKEYt§
*Department of OtalaryngologylHead and Neck Surgery, Beaumont Hospital, TDepartment of Clinical Microbiology, Central Pathology Laboratory, St. James’s Hospital, and IDepartment of OtolaryngologylHead and Neck Surgery, Trinity College, Dublin, Ireland
The value of pernasal swabs and direct adenoid swabs in chronic adenoid and adenotonsillar disease was assessed in 175 patients. Prior to adenoidectomy (53 patients) or adenotonsillectomy (122 patients), pernasal and direct adenoid swabs were taken. Adenoid currettings and tonsil tissue were cultured. Haemophilus injuenzae was the bacterium most frequently isolated from adenoid currettings and from the centre (core) of the resected tonsil. There was a close relationship between the bacteriology of the pernasal swab and the adenoid tissue and tonsil core in 72 and 71% of patients, respectively. There was an identical profile of pathogens in 52 and 49%, respectively. We suggest that in children with adenoiditis or adenotonsillitis and hypertrophy of the adenoid, a pernasal swab should be used in preference to a throat swab in selecting appropriate antimicrobial therapy. Penicillin and ampicillin are not appropriate blind therapy in chronic adenoid and adenotonsillar infections because of the prevalance ofb-lactamase-producing aerobes (40%) in adenoid and tonsil core in these conditions.
Introduction Recurrent acute tonsillitis and adenotonsillitis are common disorders accounting for a substantial percentage of general practitioner visits and financial expenditure annually (1). Despite the ubiquity of the problem, the underlying pathogens are poorly understood. Treatment usually involves prescription of an antibiotic based on a superficial tonsillar swab, with surgical intervention in the event of failed medical treatment. The failure of superficial throat swabs to predict deep pathogens is well documented in the literature (2-S). To date the only way of determining the actual core microflora is by aseptically resecting the tonsil and sampling the centre or ‘core’ (2-5). Adenoiditis is often an integral part of adenotonsillitis, and methods of sampling the adenoid have varied from adenoidectomy to open biopsy (6-11). It is surprising that although the superficial tonsil swab is routinely used to sample the adenoid, few investigators have studied the feasibility of using the pernasal swab as an indicator of the adenoid pathogens. In an initial study involving 52 patients, we found that the pernasal swab taken at operation is a good predictor of adenoidal bacteriology. In addition, the operative pernasal swab was a better predictor of tonsil core Received 22 January
1992 and accepted in revised form 26 May 1992.
$70 whom correspondence
should be addressed.
0954-61 I l/93/040303+06
$08.00/O
pathogens than a superficial tonsil swab (12). The present study of a larger cohort was undertaken to confirm and extend our previous observations. Methods of sampling the adenoid, and the relationship between the bacteriology of the adenoid and tonsil, were studied in patients with adenoidal obstructive symptoms undergoing adenoidectomy or adenotonsillectomy. An additional feature of this study was the use of an operative direct adenoid swab as a predictor of adenotonsillar pathogens. The direct adenoid swab was compared with the pernasal swab to assess the importance of nasal contamination when taking a pernasal swab. Materials
and Methods
PATIENTS
One-hundred and seventy-five patients admitted for elective adenoidectomy (53 patients) or adenotonsillectomy (122 patients) between 1 March 1989 and 31 July 1990 were studied. The indication for adenoidectomy was obstructive nasal symptoms attributable to hypertrophy of the adenoid. This was confirmed by palpation under general anaesthesia prior to surgery. The indication for tonsillectomy was recurrent acute tonsillitis for at least 2 yr with five or more acute exacerbations per annum. Evidence of a significant infection included a history of pain and fever with redness or exudate of the tonsils and enlarged and tender @ I993 Baillibre Tindall
304
R. J. Gaffney et al.
cervial adenopathy. All patients included in the study population were free of acute respiratory symptoms for a minimum of 5 weeks prior to operation. Samples taken included pernasal swab, adenoid currettings, direct adenoid swab, superficial tonsil swab and tonsil. Swabs were taken following induction of anaesthesia. The nasopharynx was sampled by use of a cotton swab with fine metal shaft inserted through the nose. A superficial tonsil swab was taken if the patient was to undergo tonsillectomy. The uvula was retracted and the adenoid sampled under direct vision using a cotton swab with a fine metal shaft. Each swab was placed in transport medium (‘Transwab’, Medical Wire and Equipment Company, Corsham, Wiltshire). The adenoid was removed by currettage and placed in a sterile container. The tonsils were removed by dissection and placed in a sterile container. All specimens were dispatched without delay to the microbiology department. MICROBIOLOGY
Adenoid currettings were crushed using a sterile forceps. A sterile swab was throughly agitated in the specimen and inoculated on blood agar and chocolate blood agar. The methodology used for tonsil differed between patients included in the initial 4 months and those studied in the subsequent period. In the initial phase of the study, the tonsils were washed for three repeat rinsings in sterile saline. One tonsil of each pair was placed in a sterile Petri dish and held by a sterile forceps. The deep (cut) surface was cauterized using a flamed scalpel and the tonsil was bisected through the cauterized surface using a serile scalpel. The centre of the tonsil (referred to as ‘core’) was sampled using a sterile swab and cultured as for the adenoid. In the second phase of the study, one tonsil of each pair was disinfected by immersion in povidone iodine for 30 s. (The methodology was altered in order to further reduce the possibility of introduction of surface microorganisms into the core of the tonsil during sampling.) The tonsil was then thoroughly rinsed in sterile saline and the subsequent sampling was as in the initial period. The swab from the tonsil core and the pernasal, direct adenoid and superficial tonsil swabs were cultured on blood and chocolate blood agar. All plates were incubated in 5% CO, in a carbon dioxide incubator for 24-36 h. A blood agar plate containing 100 mg 1-l neomycin (to inhibit facultative organisms) was also included for the initial 80 patients and was incubated anaerobically for 48 h. After incubation, the following presumptive pathogens, Haemophilus influenzae, H. parainjluenzae, Staphylococcus aureus, Moraxella catarrhalis and /?-haemolytic streptococci of Lancefield’s groups Aa were identified by standard
Table
I
Specimens taken in 175 patients No. of patients
PNS+ PNS + PNS + PNS + PNS+
DASf DAS+
A+ A+ A+ A+
DAS+ DAS+
T T
T
A+
T
A+
T
49 53 22 19 12 14 6
PNS, pernasal swab. DAS, direct adenoid swab. A, adenoid currettings. T, tonsil.
methods as described elsewhere (5) using a semiquantitative method to record growth. Speciation of anaerobic isolates were not performed. /?-lactamase production in Haemophilus species was determined by use of a chromogenic cephalosporin. Finally, the culture plates from all sites in each individual patient were compared. Results PATIENTS
The age range was from 2-12 years with a mean of 5.6 years. There were 86 males and 89 females. Significant adenoid hypertrophy was confirmed by palpation in 163 of the 175 patients and adenoidectomy was performed. Tonsillectomy was performed in 134 patients including 12 in whom adenoidectomy was not performed. The specimens obtained are tabulated in Table 1. Adenoid currettings were obtained and tonsil core was sampled in 134 patients, allowing direct comparison of the bacteriology. A pernasal swab was obtained in 102 of these and, in addition, a direct adenoid swab was obtained in 53. ADENOID
BACTERIOLOGY
The bacteriology of adenoid currettings is summarized in Table 2. Pathogenic bacteria were isolated from all but three of the specimens. H. infuenzae (non-typable) was the bacterium most frequently encountered and was isolated from adenoid currettings in 125 patients (76.7%). Mixed pathogens were isolated in 61 (37.6%) and the most common mixture was H. influenzae with S. aureus. Strict anaerobes were rarely isolated and when present comprised less than 1% of the total bacterial content of the adenoid. One or more b-lactamase producer was present in 65 adenoid specimens (40%) including 16 isolates of
Bacteriology Table 2
Bacteriology
of adenoid
currettings
in 165 patients n
Haemophilus infruenzae Haemophilusparainjkenzae Staphylococcus aureus ,Shaemolytic streptococci Streptococcus pneumoniae Moraxella catarrhalis Normal flora n, Number isolated. [ 1, Number
ofpatients ofa-lactamase
from
whom
128
[16]
4’: 31 24 5 3
[45]
each individual
151
species was
producers.
H. injhenzae, (12.8%), 45 isolates of S. aweus (96%) and five isolates of hf. catarrhalis (100%). Both adenoid currettings and direct adenoid swabs were available in 77 patients. H. influenzae was the most common bacterial isolate in these specimens. Comparison revealed a close relationship in 71 cases (92.2%). This included 48 cases (62.3%) in whom the bacteriology was identical. In 14 patients the direct adenoid swab yielded one pathogen less, whilst in six there was one additional pathogen in the direct swab. In six patients (7.8%) there was no relationship between the results of these two sampling methods, including four in whom mixed commensal flora was obtained from the direct adenoid swab and pathogens were obtained from the currettings. There was a close correlation between the pernasal swab and the adenoid bacteriology in 103 (72%) of the 143 patients in whom both of these samples were obtained. The bacteriology was identical in 76 of these (52.1%). In five instances two pathogens were isolated from the pernasal swab with only one of these pathogens isolated from the adenoid currettings. In 2b patients one or more additional pathogen was isolated from the adenoid than predicted by the pernasal swab. In 40 cases there was no relationship between the pernasal swab and adenoid bacteriology; this included 23 patients in whom the pernasal swab yielded normal flora only, and three in whom there was no growth from the pernasal swab. Direct adenoid swabs were obtained from 90 patients. In Table 3. the direct adenoid bacteriology in patients with or without obstructive adenoid hypertrophy is compared. The direct adenoid swab was used for this comparison as it was the only adenoid specimen from patients in whom adenoidectomy was not performed. The prevalence of H. injluenzae is similar in the two groups. Direct adenoid swab and adenoid currettings were obtained from 77 patients. One or
of tonsil and adenoid
305
Table 3 Bacteriology of direct adenoid swab in 77 patients with adenoidal hypertrophy (A) and in 13 patients without significant post-nasal obstruction (B)
n (A)
n, Number isolated. [ 1, Number
ofpatients
from
of p-lactamase
Bacteriology
Table 4
[31
8
PI
2g2 [25] 23 18
3 6 1
131
6 VI
1
111
7
2
55
Haemophilw influenzae Haemophilus parainJIuenzae Staphylococcus aureus /?-haemolytic streptococci Streptococcuspneumoniae Moraxella catarrhalis Normal flora whom
n (B)
each individual
species was
producers.
of tonsil
in 122 patients n
Haemophilw influenzae Haemophilusparainj?uenzae Staphylococcus aureus /?-haemolytic streptococci Streptococcuspneumoniae Moraxella catarrhalis Normal flora n, Number isolated. [ 1, Number
of patients
from
of p-lactamase
whom
94
[17]
3; 24 9 0 3
[37]
each individual
species was
producers.
more pathogen was isolated from the adenoid currettings in all instances, and from 73 of the direct adenoid swabs. There was a close relationship between the results of these two sampling techniques in 71 patients, including identical bacteriology in 48 instances (62.3%). In 14 cases the adenoid currettings yielded an additional pathogen, and in six the direct adenoid swab yielded an additional pathogen. There was no relationship in six, including four in whom the direct adenoid swab yielded normal flora only. Comparison of the superficial tonsil swab cultures with the adenoid currettings showed that oropharyngeal contamination in collection of the adenoid currettings was not an appreciable problem. TONSIL
CORE BACTERIOLOGY
Details of the tonsil core bacteriology are shown in Table 4. Pathogens were obtained in 119 specimens
306
R. J. Gaffney et al.
(97.5%) and mixed pathogens in47 cases (38.5%). The predominant tonsil core isolate was non-typable H. infuenzae, and this compared with normal flora in the superficial swab in 57.8% of cases. Strict anaerobes, when detected, comprised less than 10% of the total bacterial flora, and were not deemed to be numerically significant. One or more /I-lactamase producer was found in 40.2% of tonsils. The use of povidone iodine for decontamination did not alter the qualitative bacteriology. The bacteriology of core tonsil and adenoid currettings was compared in 122 patients from whom both of these specimens were available. A common pathogen was found in tonsil and adenoid in 112 (91.8%). An identical picture was found in 69.7% and some relationship was found in 22.1%. In 15 instances, the adenoid yielded one or more pathogen in addition to those present in the tonsil, and in eight patients the tonsil yielded one additional pathogen. There was no relationship between the bacteriology of the tonsil and adenoid in ten patients (8.2%). Pernasal swab and tonsil core were compared in 102 patients with significant adenoid hypertrophy. There was a close relationship between the bacteriology in 81 patients (71%) including 43 in whom the pathogen profile was identical (44.7%). In six patients (5.3%) there was one or more additional pathogen in the pernasal swab and in 21 (18.4%) there was one or more additional pathogen in the tonsil. Three patients (2.6%) had one shared pathogen with an additional pathogen in each site. There was no relationship between the bacteriology of pernasal swab and tonsil in 32 patients (29%) including five in whom there was no growth from the pernasal swab, and the adequacy of the specimen may have been in question. Pernasal swab, adenoid currettings and tonsil were available for comparison in 102 patients. There was a close relationship evident in the bacteriology of all three specimens in 71.6% of cases. This included 39 patients in whom identical pathogens were isolated from all three specimens. In 29 patients (28.4%) there was no relationship. Discussion Tonsillectomy and/or adenoidectomy remain the most common surgical procedures in childhood. The principal indication for tonsillectomy at present is recurrent acute tonsilitis with five or more acute attacks per yr in the 2 yr preceding operation (13). Indications for adenoidectomy fall into three groups. Firstly, adenoidectomy may be performed with tonsillectomy in patients with recurrent acute tonsillitis. Secondly, adenoidectomy may be indicated for
symptoms of nasal obstruction, snoring and mouth breathing. Thirdly, adenoidectomy is frequently part of the treatment rationale in children with chronic middle ear disease (14). The tonsils and adenoid comprise the principal components of the pharyngeal lymphoid tissue. This lymphoid tissue forms a ring at the entrance to the aerodigestive tract (Waldeyer’s ring). There is widespread belief that infection of one component of the pharyngeal tissue may spread to other parts of the ring causing generalized infection. There is one study of 50 children in support of this theory (9). In the present study, we found a close correlation between the aerobic bacteriology of the tonsil and adenoid in 122 children with chronic or recurrent adenotonsillar disease in a quiescent phase of the infection. There was a shared pathogen in 91.8% with an identical profile of pathogens in 69.7%. The prevalence of mixed pathogens and /I-lactamase producers was similar. These findings confirm a close relationship between the bacteriology of the adenoid and tonsils. Strict anaerobes were not detected in numbers which we considered to be numerically significant from tonsils or adenoid. This is in contrast to the findings of Brook and his associates (7,15,16), but in agreement with those of other authors (6,8,9,17). There are several possible reasons for this difference. The first is that Brook and his co-workers used a fluid enrichment technique in addition to direct plating to detect anaerobes. The use of enrichment would facilitate recovery of small numbers of anaerobes present in the specimen. We considered that in view of the ubiquity of anaerobes in the oropharynx, it was necessary to use criteria which would allow a quantitative comparison with the aerobic component of the flora. In no instance in 80 patients in whom anaerobic culture was performed in this study did we find the anaerobic component of the tonsillar flora to exceed 10% of the total bacteria flora. It may be that these numerical criteria were too strict. Also, it is possible that use of one or more additional anaerobic culture media with prolonged incubation would have improved the yield of anaerobes. Non-typable H. influenzae are considered to be normal inhabitants of the nasopharynx. The possibility arises that the tonsil is the actual reservoir from which these organisms colonize the nasopharynx. In this study, we found that in children with adenoidal hypertrophy and recurrent acute tonsillitis, the tonsil yielded profuse growths of H. injluenzae. This has important implications in the management of the child with middle ear disease. Bacteria may spread from this reservoir - the tonsil ~ to the nasopharynx, resulting in colonization or infection of the adenoid. Adenoidectomy is often advocated with the intention of removing
Bacteriology
of tonsil and adenoid
307
a major reservoir of infection (11,14). It follows that that penicillinase-susceptiblepenicillins are not approtonsillectomy should be considered as an adjunct to priate treatment of chronic or recurrent adenoidal adenoidectomy in selected cases.The possibility that and/or tonsillar infections. The association of H. infruenzae with hypertrophy of the tonsil and not the nasopharynxper se is the actual reservoir of H. influenzae in other clinical situations, the adenoid and with the tonsil in the quiescant phase such as cystic fibrosis or chronic bronchitis, should be of recurrent acute tonsillitis is striking. The possibility that some strains of H. inzuenzae may have a addressed. In this study the bacteriology of the pernasal swab direct pathogenic role in these tissues remains to be and adenoid currettings showed a close relationship in addressed. The most popular theory of the pathogen72% of cases,with an identical profile of pathogens esisof middle ear diseaseis the so-called ‘inflammatory in 53%. This contrasts with our earlier pilot study in theory’ (11). The kernel of this theory is that histamine which we found that the pernasal swab accurately release in the inflamed adenoid results in Eustachian indentified core tonsil pathogens in 24 of 37 patients tube malfunction due to peri-tubal oedema. This (78%) (12). We believe that this descrepency may in theory is supported by the fact that the histamine conpart be due to varied methodology in sampling. In the tent is higher in the adenoid of children with bilateral initial study all samples were taken by one dedicated glue ear than in the adenoid in normal healthy controls otolaryngologist, whilst, in contrast, in this larger (21). It is generally believed that mast cells are the study the specimens were collected by several sur- source of tissue histamine, however, Davies and geons. This viewpoint is supported by the fact that in associates have proposed an alternative mechanism, the present study, when one individual surgeon was implicating certain bacterial strains in histamine proresponsible for specimen collection, the yield of duction (22,23). They found that sputum isolates of pathogens from the pernasal swab increased. it is of various respiratory pathogens including H. influenzae interest that although the value of the pernasal swab and H. parainjuenzae, from patients with cystic in predicting the adenoid pathogens varied with the fibrosis and chronic bronchitis, produced histamine experience of the operator, it was superior to the when grown in the presence of I-histidine. They throat swab in predicting tonsil core pathogens. developed the hypothesis that these bacteria contain There are two likely reasons for this difference. One histidine decarboxylase, and that histidine present in is related to the difference intrinsic in these two sam- sputum is converted to histamine by this enzyme. pling methods. A superficial swab is rubbed over the The population in this study were children underarea and does not sample the tonsil tissue. In con- going elective adenoidectomy who were free of acute trast, a pernasal swab is inserted until it hits the back infection. The long-term aim of our studies is to of the nasopharynx. In the presence of adenoidal develop methods of sampling which can be readily perhypertrophy a correctly taken pernasal swab is in formed in the outpatient clinic, which would give an fact a blunt biopsy of the adenoid. Finally, a super- accurate picture of the bacteriology to guide effective ficial tonsil swab is more likely to be contaminated antibiotic therapy, and thus perhaps avoid surgery. It with oropharyngeal bacteria. remains to be demonstrated if the findings are directly There was a close relationship between the bacteri- applicable to acute exacerbations of adenoid infection ology of the pernasal swab and the tonsil core in 71% or otitis media. It is, however, clear that in the presence of 102 patients. This contrasts with the well recognized of adenoidal hypertrophy a pernasal swab should be failure of a superficial tonsil swab to accurately used in preference to a throat swab in selecting approidentify core tonsil pathogens. priate antibiotic therapy. The pernasal swab gives a Forty per cent of adenoids and 40.2% of tonsils more reliable indication of the adenoid bacteriology contained one or more /I-lactamase producing bac- than a superficial throat swab, and is also more reliable terium, including Haemophilus species,S. aureus and than a superficial throat swab in predicting the tonsilM. catarrhalis. This probably reflects bacterial selec- lar microflora in children with hypertrophy of the tion induced by antibiotic administration. All patients adenoid and recurrent acute tonsillitis. in the study had received multiple courses of antibiotics. Repeated pencillin administration results in a shift in the oral microbial flora, with selection of Acknowlegements p-lactamase producing strains (18,19). p-lactamases protect penicillin-susceptible bacteria from the action This study was supported by a Fellowship grant of penicillin by inactivating the antibiotic (19,20). By from Eli Lilly and Company. We are grateful to this mechanism of indirect pathogenicity, p-lactamase the theatre staff for their assistance with specimen producing strains allow infection to persist. It follows collection.
308
R. J. Gaffney
et al.
References FM. Pediatric considerations of infections and 1. Palumbo inflammations of Waldeyer’s ring. Otolur~~ngol CZin North Am 1987; 20: 3 11-3 16. 2. Rosen G, Samuel J, Vered I. Surface tonsillar microflora versus deep tonsillar microflora in recurrent acute tonsillitis. JLuryngol 0~011977; 11: 911-913. 3. Surow JB. Handler SD. Telian SA. Fleisher GR. Baranak CC. Bacteriology of tonsil surface and core in children. Laryngoscope 1989; 99: 261-266. K, Bais AS. Tonsillar microflora, superficial 4. Uppal surface versus deep. J Laryngol Otoll989; 103: 175-I 77. 5. Timon CI, McAllister VA, Walsh MA, Cafferkey MT. Changes in tonsil bacteriology of recurrent acute tonsilitis: 1980 vs 1989. Respir Med 1990; 84: 395-400. 6. Maw RA, Speller DCE. Are the tonsils and adenoids a reservoir of infection in otitis media with effusion (glue ear)? CZZn Otolaryngoll985; 10: 265-269. 7. Brook I. Aerobic and anaerobic bacteriology of adenoids in children: a comparison between patients with chronic adenotonsillitis and adenoid hypertrophy. Laryngoscope 1981; 91: 377-382. 8. Ruokonen J, Sandelin K, Makinen J. Adenoids and otitis media with effusion. Ann Otol Rhino1 Laryngol1979; 88: 266-271. 9. Dedio RM, Tom LWC, McGowan KL, Wetmore RF, Handler SD, Potsic WP. Microbiology of the tonsils and adenoids in a pediatric population. Arch Otolaryngol Head Neck Surg 1988; 114: 763-765. 10. Kveton F, Pillsbury C, Sasaki CT. Nasal obstruction. Adenitis versus adenoid hypertrophy. Arch Otolaryngol Head Neck Sura 1982: 108: 3 15-3 18. 11. Osborne JE, Telford’ D, Barr G, Roberts C. Adenoid infection: its relationship to otitis media, glue ear and tonsillitis. Clin Otohzryngol 1987; 12: 261-264.
12. Timon CI, Cafferkey MT, Walsh MA. Infection of Waldever’s rine: value of retrooharvneeal swabs. J Otola&gol199?; 20: 279-282. a * y 13. Barr G. Further thoughts about tonsillectomy. Zr Med J 1989; 82: 142-143. 14. Stell PM. Adenoidectomy. Clin Otolaryngol198 1; 6: 1-3. 15. Brook I, Yokum P. Bacteriology of chronic tonsillitis in young adults. Arch Otolaryngol Head Neck Surg 1984; 110: 803-805. 16. Brook I, Foote P. Comparison of the microbiology of recurrent tonsillitis between children and adults. Laryngoscope 1986; 96: 1385-1387. 17. Teele DW, Healy GB, Tally FP. Persistent effusions of the middle ear. Culture for anaerobic bacteria. Ann Otol Rhino1 Laryngol1980; 89 (Suppl A): 102-103. 18. Scheifele DW, Fussell SJ. Ampicillin-resistant Huemophilus injluenzae colonising ambulatory children. Am J Dis Child 1981; 145: 4066409. 19. Kaplan SL, Mason EO, Kvernland SJ. Effect of prior antibiotics on the susceptibility of Haemophilus inzuenzae type b to ampicillin. Pediatrics 1981; 67: 269-271. 20. Bryant RE, Rashed AL, Mazza JA, Hammond D. /Ilactamase activity in human pus. J Znfect Dis 1980; 142: 594-601. 21. Collins MP, Church MF, Bakshi KN, Osborne JE. The role of histamine in adenoid tissue in secretory otitis media. J Laryngol Otol1985; 99: 685-69 1. 22. Devalia JL, Grady D, Harmanyeri Y, Tabaqchali S, Davies RJ. Histamine synthesis by respiratory tract microorganisms: a possible role in pathogenicity. J Clin Parholl989; 42: 516522. 23. Cundell DR, Devalia JL, Wilks M, Tabaqchali S, Davies RJ. Histidine decarboxylases from bacteria that colonise the human respiratory tract. J Med Microbial 1991; 35: 363-366.
Medicine
(1993) 87,303-308
Bacteriology of tonsil and adenoid and sampling techniques of adenoidal bacteriology R. J. GAFFNEY*,
C. I. TIMON?, D. F. FREEMANS, M.
A. WALSH*
AND
M. T. CAFFERKEYt§
*Department of OtalaryngologylHead and Neck Surgery, Beaumont Hospital, TDepartment of Clinical Microbiology, Central Pathology Laboratory, St. James’s Hospital, and IDepartment of OtolaryngologylHead and Neck Surgery, Trinity College, Dublin, Ireland
The value of pernasal swabs and direct adenoid swabs in chronic adenoid and adenotonsillar disease was assessed in 175 patients. Prior to adenoidectomy (53 patients) or adenotonsillectomy (122 patients), pernasal and direct adenoid swabs were taken. Adenoid currettings and tonsil tissue were cultured. Haemophilus injuenzae was the bacterium most frequently isolated from adenoid currettings and from the centre (core) of the resected tonsil. There was a close relationship between the bacteriology of the pernasal swab and the adenoid tissue and tonsil core in 72 and 71% of patients, respectively. There was an identical profile of pathogens in 52 and 49%, respectively. We suggest that in children with adenoiditis or adenotonsillitis and hypertrophy of the adenoid, a pernasal swab should be used in preference to a throat swab in selecting appropriate antimicrobial therapy. Penicillin and ampicillin are not appropriate blind therapy in chronic adenoid and adenotonsillar infections because of the prevalance ofb-lactamase-producing aerobes (40%) in adenoid and tonsil core in these conditions.
Introduction Recurrent acute tonsillitis and adenotonsillitis are common disorders accounting for a substantial percentage of general practitioner visits and financial expenditure annually (1). Despite the ubiquity of the problem, the underlying pathogens are poorly understood. Treatment usually involves prescription of an antibiotic based on a superficial tonsillar swab, with surgical intervention in the event of failed medical treatment. The failure of superficial throat swabs to predict deep pathogens is well documented in the literature (2-S). To date the only way of determining the actual core microflora is by aseptically resecting the tonsil and sampling the centre or ‘core’ (2-5). Adenoiditis is often an integral part of adenotonsillitis, and methods of sampling the adenoid have varied from adenoidectomy to open biopsy (6-11). It is surprising that although the superficial tonsil swab is routinely used to sample the adenoid, few investigators have studied the feasibility of using the pernasal swab as an indicator of the adenoid pathogens. In an initial study involving 52 patients, we found that the pernasal swab taken at operation is a good predictor of adenoidal bacteriology. In addition, the operative pernasal swab was a better predictor of tonsil core Received 22 January
1992 and accepted in revised form 26 May 1992.
$70 whom correspondence
should be addressed.
0954-61 I l/93/040303+06
$08.00/O
pathogens than a superficial tonsil swab (12). The present study of a larger cohort was undertaken to confirm and extend our previous observations. Methods of sampling the adenoid, and the relationship between the bacteriology of the adenoid and tonsil, were studied in patients with adenoidal obstructive symptoms undergoing adenoidectomy or adenotonsillectomy. An additional feature of this study was the use of an operative direct adenoid swab as a predictor of adenotonsillar pathogens. The direct adenoid swab was compared with the pernasal swab to assess the importance of nasal contamination when taking a pernasal swab. Materials
and Methods
PATIENTS
One-hundred and seventy-five patients admitted for elective adenoidectomy (53 patients) or adenotonsillectomy (122 patients) between 1 March 1989 and 31 July 1990 were studied. The indication for adenoidectomy was obstructive nasal symptoms attributable to hypertrophy of the adenoid. This was confirmed by palpation under general anaesthesia prior to surgery. The indication for tonsillectomy was recurrent acute tonsillitis for at least 2 yr with five or more acute exacerbations per annum. Evidence of a significant infection included a history of pain and fever with redness or exudate of the tonsils and enlarged and tender @ I993 Baillibre Tindall
304
R. J. Gaffney et al.
cervial adenopathy. All patients included in the study population were free of acute respiratory symptoms for a minimum of 5 weeks prior to operation. Samples taken included pernasal swab, adenoid currettings, direct adenoid swab, superficial tonsil swab and tonsil. Swabs were taken following induction of anaesthesia. The nasopharynx was sampled by use of a cotton swab with fine metal shaft inserted through the nose. A superficial tonsil swab was taken if the patient was to undergo tonsillectomy. The uvula was retracted and the adenoid sampled under direct vision using a cotton swab with a fine metal shaft. Each swab was placed in transport medium (‘Transwab’, Medical Wire and Equipment Company, Corsham, Wiltshire). The adenoid was removed by currettage and placed in a sterile container. The tonsils were removed by dissection and placed in a sterile container. All specimens were dispatched without delay to the microbiology department. MICROBIOLOGY
Adenoid currettings were crushed using a sterile forceps. A sterile swab was throughly agitated in the specimen and inoculated on blood agar and chocolate blood agar. The methodology used for tonsil differed between patients included in the initial 4 months and those studied in the subsequent period. In the initial phase of the study, the tonsils were washed for three repeat rinsings in sterile saline. One tonsil of each pair was placed in a sterile Petri dish and held by a sterile forceps. The deep (cut) surface was cauterized using a flamed scalpel and the tonsil was bisected through the cauterized surface using a serile scalpel. The centre of the tonsil (referred to as ‘core’) was sampled using a sterile swab and cultured as for the adenoid. In the second phase of the study, one tonsil of each pair was disinfected by immersion in povidone iodine for 30 s. (The methodology was altered in order to further reduce the possibility of introduction of surface microorganisms into the core of the tonsil during sampling.) The tonsil was then thoroughly rinsed in sterile saline and the subsequent sampling was as in the initial period. The swab from the tonsil core and the pernasal, direct adenoid and superficial tonsil swabs were cultured on blood and chocolate blood agar. All plates were incubated in 5% CO, in a carbon dioxide incubator for 24-36 h. A blood agar plate containing 100 mg 1-l neomycin (to inhibit facultative organisms) was also included for the initial 80 patients and was incubated anaerobically for 48 h. After incubation, the following presumptive pathogens, Haemophilus influenzae, H. parainjluenzae, Staphylococcus aureus, Moraxella catarrhalis and /?-haemolytic streptococci of Lancefield’s groups Aa were identified by standard
Table
I
Specimens taken in 175 patients No. of patients
PNS+ PNS + PNS + PNS + PNS+
DASf DAS+
A+ A+ A+ A+
DAS+ DAS+
T T
T
A+
T
A+
T
49 53 22 19 12 14 6
PNS, pernasal swab. DAS, direct adenoid swab. A, adenoid currettings. T, tonsil.
methods as described elsewhere (5) using a semiquantitative method to record growth. Speciation of anaerobic isolates were not performed. /?-lactamase production in Haemophilus species was determined by use of a chromogenic cephalosporin. Finally, the culture plates from all sites in each individual patient were compared. Results PATIENTS
The age range was from 2-12 years with a mean of 5.6 years. There were 86 males and 89 females. Significant adenoid hypertrophy was confirmed by palpation in 163 of the 175 patients and adenoidectomy was performed. Tonsillectomy was performed in 134 patients including 12 in whom adenoidectomy was not performed. The specimens obtained are tabulated in Table 1. Adenoid currettings were obtained and tonsil core was sampled in 134 patients, allowing direct comparison of the bacteriology. A pernasal swab was obtained in 102 of these and, in addition, a direct adenoid swab was obtained in 53. ADENOID
BACTERIOLOGY
The bacteriology of adenoid currettings is summarized in Table 2. Pathogenic bacteria were isolated from all but three of the specimens. H. infuenzae (non-typable) was the bacterium most frequently encountered and was isolated from adenoid currettings in 125 patients (76.7%). Mixed pathogens were isolated in 61 (37.6%) and the most common mixture was H. influenzae with S. aureus. Strict anaerobes were rarely isolated and when present comprised less than 1% of the total bacterial content of the adenoid. One or more b-lactamase producer was present in 65 adenoid specimens (40%) including 16 isolates of
Bacteriology Table 2
Bacteriology
of adenoid
currettings
in 165 patients n
Haemophilus infruenzae Haemophilusparainjkenzae Staphylococcus aureus ,Shaemolytic streptococci Streptococcus pneumoniae Moraxella catarrhalis Normal flora n, Number isolated. [ 1, Number
ofpatients ofa-lactamase
from
whom
128
[16]
4’: 31 24 5 3
[45]
each individual
151
species was
producers.
H. injhenzae, (12.8%), 45 isolates of S. aweus (96%) and five isolates of hf. catarrhalis (100%). Both adenoid currettings and direct adenoid swabs were available in 77 patients. H. influenzae was the most common bacterial isolate in these specimens. Comparison revealed a close relationship in 71 cases (92.2%). This included 48 cases (62.3%) in whom the bacteriology was identical. In 14 patients the direct adenoid swab yielded one pathogen less, whilst in six there was one additional pathogen in the direct swab. In six patients (7.8%) there was no relationship between the results of these two sampling methods, including four in whom mixed commensal flora was obtained from the direct adenoid swab and pathogens were obtained from the currettings. There was a close correlation between the pernasal swab and the adenoid bacteriology in 103 (72%) of the 143 patients in whom both of these samples were obtained. The bacteriology was identical in 76 of these (52.1%). In five instances two pathogens were isolated from the pernasal swab with only one of these pathogens isolated from the adenoid currettings. In 2b patients one or more additional pathogen was isolated from the adenoid than predicted by the pernasal swab. In 40 cases there was no relationship between the pernasal swab and adenoid bacteriology; this included 23 patients in whom the pernasal swab yielded normal flora only, and three in whom there was no growth from the pernasal swab. Direct adenoid swabs were obtained from 90 patients. In Table 3. the direct adenoid bacteriology in patients with or without obstructive adenoid hypertrophy is compared. The direct adenoid swab was used for this comparison as it was the only adenoid specimen from patients in whom adenoidectomy was not performed. The prevalence of H. injluenzae is similar in the two groups. Direct adenoid swab and adenoid currettings were obtained from 77 patients. One or
of tonsil and adenoid
305
Table 3 Bacteriology of direct adenoid swab in 77 patients with adenoidal hypertrophy (A) and in 13 patients without significant post-nasal obstruction (B)
n (A)
n, Number isolated. [ 1, Number
ofpatients
from
of p-lactamase
Bacteriology
Table 4
[31
8
PI
2g2 [25] 23 18
3 6 1
131
6 VI
1
111
7
2
55
Haemophilw influenzae Haemophilus parainJIuenzae Staphylococcus aureus /?-haemolytic streptococci Streptococcuspneumoniae Moraxella catarrhalis Normal flora whom
n (B)
each individual
species was
producers.
of tonsil
in 122 patients n
Haemophilw influenzae Haemophilusparainj?uenzae Staphylococcus aureus /?-haemolytic streptococci Streptococcuspneumoniae Moraxella catarrhalis Normal flora n, Number isolated. [ 1, Number
of patients
from
of p-lactamase
whom
94
[17]
3; 24 9 0 3
[37]
each individual
species was
producers.
more pathogen was isolated from the adenoid currettings in all instances, and from 73 of the direct adenoid swabs. There was a close relationship between the results of these two sampling techniques in 71 patients, including identical bacteriology in 48 instances (62.3%). In 14 cases the adenoid currettings yielded an additional pathogen, and in six the direct adenoid swab yielded an additional pathogen. There was no relationship in six, including four in whom the direct adenoid swab yielded normal flora only. Comparison of the superficial tonsil swab cultures with the adenoid currettings showed that oropharyngeal contamination in collection of the adenoid currettings was not an appreciable problem. TONSIL
CORE BACTERIOLOGY
Details of the tonsil core bacteriology are shown in Table 4. Pathogens were obtained in 119 specimens
306
R. J. Gaffney et al.
(97.5%) and mixed pathogens in47 cases (38.5%). The predominant tonsil core isolate was non-typable H. infuenzae, and this compared with normal flora in the superficial swab in 57.8% of cases. Strict anaerobes, when detected, comprised less than 10% of the total bacterial flora, and were not deemed to be numerically significant. One or more /I-lactamase producer was found in 40.2% of tonsils. The use of povidone iodine for decontamination did not alter the qualitative bacteriology. The bacteriology of core tonsil and adenoid currettings was compared in 122 patients from whom both of these specimens were available. A common pathogen was found in tonsil and adenoid in 112 (91.8%). An identical picture was found in 69.7% and some relationship was found in 22.1%. In 15 instances, the adenoid yielded one or more pathogen in addition to those present in the tonsil, and in eight patients the tonsil yielded one additional pathogen. There was no relationship between the bacteriology of the tonsil and adenoid in ten patients (8.2%). Pernasal swab and tonsil core were compared in 102 patients with significant adenoid hypertrophy. There was a close relationship between the bacteriology in 81 patients (71%) including 43 in whom the pathogen profile was identical (44.7%). In six patients (5.3%) there was one or more additional pathogen in the pernasal swab and in 21 (18.4%) there was one or more additional pathogen in the tonsil. Three patients (2.6%) had one shared pathogen with an additional pathogen in each site. There was no relationship between the bacteriology of pernasal swab and tonsil in 32 patients (29%) including five in whom there was no growth from the pernasal swab, and the adequacy of the specimen may have been in question. Pernasal swab, adenoid currettings and tonsil were available for comparison in 102 patients. There was a close relationship evident in the bacteriology of all three specimens in 71.6% of cases. This included 39 patients in whom identical pathogens were isolated from all three specimens. In 29 patients (28.4%) there was no relationship. Discussion Tonsillectomy and/or adenoidectomy remain the most common surgical procedures in childhood. The principal indication for tonsillectomy at present is recurrent acute tonsilitis with five or more acute attacks per yr in the 2 yr preceding operation (13). Indications for adenoidectomy fall into three groups. Firstly, adenoidectomy may be performed with tonsillectomy in patients with recurrent acute tonsillitis. Secondly, adenoidectomy may be indicated for
symptoms of nasal obstruction, snoring and mouth breathing. Thirdly, adenoidectomy is frequently part of the treatment rationale in children with chronic middle ear disease (14). The tonsils and adenoid comprise the principal components of the pharyngeal lymphoid tissue. This lymphoid tissue forms a ring at the entrance to the aerodigestive tract (Waldeyer’s ring). There is widespread belief that infection of one component of the pharyngeal tissue may spread to other parts of the ring causing generalized infection. There is one study of 50 children in support of this theory (9). In the present study, we found a close correlation between the aerobic bacteriology of the tonsil and adenoid in 122 children with chronic or recurrent adenotonsillar disease in a quiescent phase of the infection. There was a shared pathogen in 91.8% with an identical profile of pathogens in 69.7%. The prevalence of mixed pathogens and /I-lactamase producers was similar. These findings confirm a close relationship between the bacteriology of the adenoid and tonsils. Strict anaerobes were not detected in numbers which we considered to be numerically significant from tonsils or adenoid. This is in contrast to the findings of Brook and his associates (7,15,16), but in agreement with those of other authors (6,8,9,17). There are several possible reasons for this difference. The first is that Brook and his co-workers used a fluid enrichment technique in addition to direct plating to detect anaerobes. The use of enrichment would facilitate recovery of small numbers of anaerobes present in the specimen. We considered that in view of the ubiquity of anaerobes in the oropharynx, it was necessary to use criteria which would allow a quantitative comparison with the aerobic component of the flora. In no instance in 80 patients in whom anaerobic culture was performed in this study did we find the anaerobic component of the tonsillar flora to exceed 10% of the total bacteria flora. It may be that these numerical criteria were too strict. Also, it is possible that use of one or more additional anaerobic culture media with prolonged incubation would have improved the yield of anaerobes. Non-typable H. influenzae are considered to be normal inhabitants of the nasopharynx. The possibility arises that the tonsil is the actual reservoir from which these organisms colonize the nasopharynx. In this study, we found that in children with adenoidal hypertrophy and recurrent acute tonsillitis, the tonsil yielded profuse growths of H. injluenzae. This has important implications in the management of the child with middle ear disease. Bacteria may spread from this reservoir - the tonsil ~ to the nasopharynx, resulting in colonization or infection of the adenoid. Adenoidectomy is often advocated with the intention of removing
Bacteriology
of tonsil and adenoid
307
a major reservoir of infection (11,14). It follows that that penicillinase-susceptiblepenicillins are not approtonsillectomy should be considered as an adjunct to priate treatment of chronic or recurrent adenoidal adenoidectomy in selected cases.The possibility that and/or tonsillar infections. The association of H. infruenzae with hypertrophy of the tonsil and not the nasopharynxper se is the actual reservoir of H. influenzae in other clinical situations, the adenoid and with the tonsil in the quiescant phase such as cystic fibrosis or chronic bronchitis, should be of recurrent acute tonsillitis is striking. The possibility that some strains of H. inzuenzae may have a addressed. In this study the bacteriology of the pernasal swab direct pathogenic role in these tissues remains to be and adenoid currettings showed a close relationship in addressed. The most popular theory of the pathogen72% of cases,with an identical profile of pathogens esisof middle ear diseaseis the so-called ‘inflammatory in 53%. This contrasts with our earlier pilot study in theory’ (11). The kernel of this theory is that histamine which we found that the pernasal swab accurately release in the inflamed adenoid results in Eustachian indentified core tonsil pathogens in 24 of 37 patients tube malfunction due to peri-tubal oedema. This (78%) (12). We believe that this descrepency may in theory is supported by the fact that the histamine conpart be due to varied methodology in sampling. In the tent is higher in the adenoid of children with bilateral initial study all samples were taken by one dedicated glue ear than in the adenoid in normal healthy controls otolaryngologist, whilst, in contrast, in this larger (21). It is generally believed that mast cells are the study the specimens were collected by several sur- source of tissue histamine, however, Davies and geons. This viewpoint is supported by the fact that in associates have proposed an alternative mechanism, the present study, when one individual surgeon was implicating certain bacterial strains in histamine proresponsible for specimen collection, the yield of duction (22,23). They found that sputum isolates of pathogens from the pernasal swab increased. it is of various respiratory pathogens including H. influenzae interest that although the value of the pernasal swab and H. parainjuenzae, from patients with cystic in predicting the adenoid pathogens varied with the fibrosis and chronic bronchitis, produced histamine experience of the operator, it was superior to the when grown in the presence of I-histidine. They throat swab in predicting tonsil core pathogens. developed the hypothesis that these bacteria contain There are two likely reasons for this difference. One histidine decarboxylase, and that histidine present in is related to the difference intrinsic in these two sam- sputum is converted to histamine by this enzyme. pling methods. A superficial swab is rubbed over the The population in this study were children underarea and does not sample the tonsil tissue. In con- going elective adenoidectomy who were free of acute trast, a pernasal swab is inserted until it hits the back infection. The long-term aim of our studies is to of the nasopharynx. In the presence of adenoidal develop methods of sampling which can be readily perhypertrophy a correctly taken pernasal swab is in formed in the outpatient clinic, which would give an fact a blunt biopsy of the adenoid. Finally, a super- accurate picture of the bacteriology to guide effective ficial tonsil swab is more likely to be contaminated antibiotic therapy, and thus perhaps avoid surgery. It with oropharyngeal bacteria. remains to be demonstrated if the findings are directly There was a close relationship between the bacteri- applicable to acute exacerbations of adenoid infection ology of the pernasal swab and the tonsil core in 71% or otitis media. It is, however, clear that in the presence of 102 patients. This contrasts with the well recognized of adenoidal hypertrophy a pernasal swab should be failure of a superficial tonsil swab to accurately used in preference to a throat swab in selecting approidentify core tonsil pathogens. priate antibiotic therapy. The pernasal swab gives a Forty per cent of adenoids and 40.2% of tonsils more reliable indication of the adenoid bacteriology contained one or more /I-lactamase producing bac- than a superficial throat swab, and is also more reliable terium, including Haemophilus species,S. aureus and than a superficial throat swab in predicting the tonsilM. catarrhalis. This probably reflects bacterial selec- lar microflora in children with hypertrophy of the tion induced by antibiotic administration. All patients adenoid and recurrent acute tonsillitis. in the study had received multiple courses of antibiotics. Repeated pencillin administration results in a shift in the oral microbial flora, with selection of Acknowlegements p-lactamase producing strains (18,19). p-lactamases protect penicillin-susceptible bacteria from the action This study was supported by a Fellowship grant of penicillin by inactivating the antibiotic (19,20). By from Eli Lilly and Company. We are grateful to this mechanism of indirect pathogenicity, p-lactamase the theatre staff for their assistance with specimen producing strains allow infection to persist. It follows collection.
308
R. J. Gaffney
et al.
References FM. Pediatric considerations of infections and 1. Palumbo inflammations of Waldeyer’s ring. Otolur~~ngol CZin North Am 1987; 20: 3 11-3 16. 2. Rosen G, Samuel J, Vered I. Surface tonsillar microflora versus deep tonsillar microflora in recurrent acute tonsillitis. JLuryngol 0~011977; 11: 911-913. 3. Surow JB. Handler SD. Telian SA. Fleisher GR. Baranak CC. Bacteriology of tonsil surface and core in children. Laryngoscope 1989; 99: 261-266. K, Bais AS. Tonsillar microflora, superficial 4. Uppal surface versus deep. J Laryngol Otoll989; 103: 175-I 77. 5. Timon CI, McAllister VA, Walsh MA, Cafferkey MT. Changes in tonsil bacteriology of recurrent acute tonsilitis: 1980 vs 1989. Respir Med 1990; 84: 395-400. 6. Maw RA, Speller DCE. Are the tonsils and adenoids a reservoir of infection in otitis media with effusion (glue ear)? CZZn Otolaryngoll985; 10: 265-269. 7. Brook I. Aerobic and anaerobic bacteriology of adenoids in children: a comparison between patients with chronic adenotonsillitis and adenoid hypertrophy. Laryngoscope 1981; 91: 377-382. 8. Ruokonen J, Sandelin K, Makinen J. Adenoids and otitis media with effusion. Ann Otol Rhino1 Laryngol1979; 88: 266-271. 9. Dedio RM, Tom LWC, McGowan KL, Wetmore RF, Handler SD, Potsic WP. Microbiology of the tonsils and adenoids in a pediatric population. Arch Otolaryngol Head Neck Surg 1988; 114: 763-765. 10. Kveton F, Pillsbury C, Sasaki CT. Nasal obstruction. Adenitis versus adenoid hypertrophy. Arch Otolaryngol Head Neck Sura 1982: 108: 3 15-3 18. 11. Osborne JE, Telford’ D, Barr G, Roberts C. Adenoid infection: its relationship to otitis media, glue ear and tonsillitis. Clin Otohzryngol 1987; 12: 261-264.
12. Timon CI, Cafferkey MT, Walsh MA. Infection of Waldever’s rine: value of retrooharvneeal swabs. J Otola&gol199?; 20: 279-282. a * y 13. Barr G. Further thoughts about tonsillectomy. Zr Med J 1989; 82: 142-143. 14. Stell PM. Adenoidectomy. Clin Otolaryngol198 1; 6: 1-3. 15. Brook I, Yokum P. Bacteriology of chronic tonsillitis in young adults. Arch Otolaryngol Head Neck Surg 1984; 110: 803-805. 16. Brook I, Foote P. Comparison of the microbiology of recurrent tonsillitis between children and adults. Laryngoscope 1986; 96: 1385-1387. 17. Teele DW, Healy GB, Tally FP. Persistent effusions of the middle ear. Culture for anaerobic bacteria. Ann Otol Rhino1 Laryngol1980; 89 (Suppl A): 102-103. 18. Scheifele DW, Fussell SJ. Ampicillin-resistant Huemophilus injluenzae colonising ambulatory children. Am J Dis Child 1981; 145: 4066409. 19. Kaplan SL, Mason EO, Kvernland SJ. Effect of prior antibiotics on the susceptibility of Haemophilus inzuenzae type b to ampicillin. Pediatrics 1981; 67: 269-271. 20. Bryant RE, Rashed AL, Mazza JA, Hammond D. /Ilactamase activity in human pus. J Znfect Dis 1980; 142: 594-601. 21. Collins MP, Church MF, Bakshi KN, Osborne JE. The role of histamine in adenoid tissue in secretory otitis media. J Laryngol Otol1985; 99: 685-69 1. 22. Devalia JL, Grady D, Harmanyeri Y, Tabaqchali S, Davies RJ. Histamine synthesis by respiratory tract microorganisms: a possible role in pathogenicity. J Clin Parholl989; 42: 516522. 23. Cundell DR, Devalia JL, Wilks M, Tabaqchali S, Davies RJ. Histidine decarboxylases from bacteria that colonise the human respiratory tract. J Med Microbial 1991; 35: 363-366.