Oral Microbiol Immunol 2001: 16: 1–9 Copyright C Munksgaard 2001Printed in Denmark . All rights reserved

ISSN 0902-0055

W. K. Leung 1, L. J. Jin 1, W. C. Yam2,L. P. Samaranayake 1Oral colonization of aerobic and1Faculty of Dentistry and 2Department ofMicrobiology, Faculty of Medicine, Universityof Hong Kong, Hong Kong SAR, Chinafacultatively anaerobic

gram-negative rods and cocci inirradiated, dentate, xerostomicindividuals

Leung WK, Jin LJ, Yam WC, Samaranayake LP. Oral colonization of aerobicand facultatively anaerobic gram-negative rods and cocci in irradiated, dentate,xerostomic individuals.Oral Microbiol Immunol 2001: 16: 1–9. C Munksgaard, 2001.

This study aimed at investigating the oral colonization of aerobic and facultativelyanaerobic gram-negative rods and cocci in head- and neck-irradiated, dentate,xerostomic individuals. Subjects were recruited from a nasopharyngeal carcinomaclinic and were segregated into group A: ∞60 years (nΩ25, 48∫6 years, 5∫5 yearspost-irradiation) and group B: Ø60 years (nΩ8, 67∫4 years, 2∫2 years post-irradiation) and were compared with age- and sex-matched normal individuals,group C: ,60 years (nΩ20, 44∫12 years) and group D: Ø60 years (nΩ10, 70∫3years). Selective culture of the oral rinse samples was carried out to isolate,quantify and speciate (using API 20E kit) aerobic and facultatively anaerobicgram-negative rods and cocci recovery. All test subjects were put undercomprehensive oral and preventive care for 3 months, and 12 group A and 5group B subjects were recalled for reassessment of aerobic and facultativelyanaerobic gram-negative rods and cocci colonization. All identical isolates, pre-and post-hygienic care, were phenotypically (Vitek, Hazelwood, MA andantibiogram profile) and genotypically (pulsed-field gel electrophoresis)evaluated. The aerobic and facultatively anaerobic gram-negative rods andcocci isolated from the first round oral rinse samples included: Acinetobacter,Neisseria, Chryseomonas, Flavimonas, Pseudomonas, Citrobacter, Enterobacter,Escherichia, Klebsiella, Flavobacterium and Weeksella species. The aerobic andfacultatively anaerobic gram-negative rods and cocci isolation rate was high forirradiated individuals, and they were 64/25% and 100/80% for groups A/C andB/D, respectively. Recovery of aerobic and facultatively anaerobic gram-negativerods and cocci and Klebsiella pneumoniae subsp. pneumoniae in oral rinse sampleswere found to be significantly more prevalent in the irradiated subjects (groupsA and B). Enterobacteriaceae were more frequently isolated from oral rinsesamples of aged irradiated subjects (group B vs D, P,0.05), where the quantity ofCitrobacter freundii (colony-forming units/ml oral rinse) was also significantlyelevated. The isolation rate of aerobic and facultatively anaerobic gram-negative rods and cocci after hygienic care remained unchanged; 3 of 12 and 3 of5 of the recalled subjects from groups A and B, respectively, harbored same aerobicand facultatively anaerobic gram-negative rods and cocci species. However, only

Key words: microbiology; irradiation;two pairs of K. pneumoniae subsp. pneumoniae, sequentially isolated from same xerostomia; nasopharyngeal carcinoma;patients in group B, were found to be identical by pulsed-field gel electrophoresis. Chinese

2 Leung et al.

This may be due to reinfection of the microbes from the same source or W. Keung Leung, Faculty of Dentistry,permanent colonization. In conclusion, irradiation-induced xerostomia seems University of Hong Kong, Room 3B39,to favor frequent, repeated, transient intraoral colonization of aerobic and 34 Hospital Road, Hong Kong SAR, China

facultatively anaerobic gram-negative rods and cocci.Accepted for publication May 17, 2000

Nasopharyngeal carcinoma is prevalentin the southern Chinese population. Inthe southern Chinese city of HongKong, the age-standardized incidencerates of nasopharyngeal carcinomawere 23 per 100,000 for males and 9 per100,000 for females (10). The disease isranked third for males and eighth forfemales among the newly diagnosedmalignant tumors in Hong Kong (10).Radiotherapy is the treatment of choicefor this condition, as surgical resectionof the lesion is complicated and isusually reserved for the management ofrecurrent primary nasopharyngeal car-cinoma (40). The irradiation therapyfor the tumor, however, inevitably in-volves the stomatomasticatory complexand the major salivary glands. One ofthe major sequelae of the radiotherapyis prolonged xerostomia (5). In health,human saliva helps regulate the oral en-vironment via its moisturizing, lubricat-ing, buffering and various antimicrobialactivities (38). Once the salivary pro-duction is impaired concomitmentchanges in oropharyngeal physiologyand microbial ecology are likely to fol-low. In an earlier study, Makkonen etal. (16) reported an increase in the oralcolonization rate of gram-negative ba-cilli from 4% to 20% in head and necktumor patients pre- and immediatelypost-irradiation therapy, respectively.They also reported a more aggressivepresentation of oral mucositis associ-ated with such microbial changes (16,36). Six months after the irradiation,the Makkonen group observed a re-duced oral colonization rate of gram-negative bacilli (up to 5%), although apersistent 37% of the individuals ex-hibited erythema of the oral mucosa.

A series of recent studies by Sedgleyet al. (28, 29, 31, 33, 34) demonstratedthat out of the 820 non-irradiated pre-dominantly Hong Kong Chinese inves-tigated, approximately one-third ofadults and one-fifth of children carriedoral aerobic and facultatively anaerobicgram-negative rods and cocci, totaling43 bacterial species or subspecies. Suchan aerobic and facultatively anaerobicgram-negative rods and cocci isolationrate appeared to be relatively high com-

pared with data from Western countriesor an African group (18, 30). However,the reported point prevalence rate of or-opharyngeal aerobic and facultativelyanaerobic gram-negative rods in HongKong Chinese was comparable to thatof a Malaysian study (21), indicatingthat the oral prevalence of aerobic andfacultatively anaerobic gram-negativerods may be influenced by race, eth-nicity or yet other unknown factors. Itis therefore likely that the oral coloniza-tion rate of the aerobic and faculta-tively anaerobic gram-negative rods andcocci in the post-head and neck ir-radiated subjects in Hong Kong couldbe quite high. Thus the present studywas formulated to investigate the oralcolonization of aerobic and faculta-tively anaerobic gram-negative rods andcocci in a homogeneous cohort with ahistory of nasopharyngeal carcinomamanaged using a similar irradiationprotocol, at least 6 months following ir-radiation therapy. We also investigatedthe effect of standard hygienic therapyon the oral colonization of aerobic andfacultatively anaerobic gram-negativerods and cocci in such individuals.

Material and methodsStudy group

A total of 33 post-irradiated naso-pharyngeal carcinoma patients were re-cruited from the Department of Clin-ical Oncology, Queen Mary Hospital,University of Hong Kong as previouslyreported (27). All recruits were deemedxerostomic based on clinical signs andsymptoms, and salivary flow rates. Theirradiated subjects were segregated intotwo groups according to age: group A,60 years, group B Ø60 years. Age-and sex-matched non-irradiated indi-viduals were selected as follows. Tenhealthy males Ø60 years from a localsenior community center (group D –out of a total of more than 100 attend-ing senior citizens on the day of survey,of which 70 are Ø60 years) and the first20 age- and sex-matched individuals,60 years attending the Reception andPrimary Care Clinic of the Prince PhilipDental Hospital, Hong Kong (group

C), within a dedicated week, were re-cruited as the control cohorts.

Clinical examination, treatment and recall

A comprehensive clinical examinationof all subjects was carried out. The de-tailed clinical conditions of the post-ir-radiated nasopharyngeal carcinoma pa-tient have been reported previously(27). All subjects in groups A and Bwere given comprehensive oral healthcare including i) oral hygiene education,ii) daily home fluoride application, iii)scaling and polishing, iv) restoration ofcaries lesions, v) extraction of grosslycarious teeth beyond restoration, andvi) topical anti-fungal therapy, i.e.100,000 U/g nystatin ointment, threetimes a day (17) for patients with oralcandidiasis. Three months after thecomprehensive oral care, all group Aand B subjects were invited for a re-evaluation as well as repeated oral rinsesampling.

Sampling

Oral rinse sampling was carried out asper Samaranayake et al. (24) with slightmodifications. In brief, all subjects wereasked to rinse their mouth for 60 s with10 ml of sterile phosphate-buffered sa-line, pH 7.2 (PBS). Denture-wearingsubjects did not remove their prosthesis.After 60 s the subjects expectorated theoral rinse into a sterile universal con-tainer, which was then immediatelytaken to the laboratory for processing.Afterwards, stimulated whole saliva wascollected and the flow rate and pHvalues were measured (27).

Culture

All samples from groups C and D werecentrifuged at 1700¿g, 10 min. The pel-let was resuspended in 2.5 ml of PBSand vortexed for 30 s at maximum set-ting (Autovortex Mixer SA2, StuartScientific, London, UK). Samples fromgroups A and B were used as such with-out concentration because a pilot studyon similar individuals showed high con-centration of aerobic and facultatively

Oral coliforms, post-irradiation 3

anaerobic gram-negative rods and coccirendering the concentration step un-necessary. Fifty ml of each of the uncon-centrated (groups A and B, lowest de-tection limit: 2 colony-forming units/mloral rinse) and the resuspended (groupsC and D, lowest detection limit: 0.5 col-ony-forming units/ml oral rinse) oralrinses were spiral plated (Model DU,Spiral Systems, Cincinnati, OH) ontoduplicate MacConkey agar plates andincubated for 18 h at 37æC. Afterwardsall cultures were examined and the col-ony-forming units in each sample quan-tified. The aerobic and facultatively an-aerobic gram-negative rods and cocciwere then subcultured to obtain pureisolates.

Identification of isolates

Pure colonies of aerobic and faculta-tively anaerobic gram-negative rods andcocci were identified presumptively onthe following features: colony mor-phology, colonial pigmentation, cellmorphology, Gram-staining reaction,oxidase positivity. Isolates were furthercharacterized biochemically using theAPI 20E kit (Analytical Profile Index,Bio Merieux SA, France).

Biotyping

Individual aerobic and facultatively an-aerobic gram-negative rods and cocciisolated confirmed as biochemicallyidentical (API 20E) from oral rinse pre-and post-hygienic therapy (sequentialisolates) were biotyped as a single batchusing Vitek GNI Card (Gram-NegativeIdentification Card, Vitek System, BioMerieux Vitek, Hazelwood, MA). Alltests were done with the same standard-ized inoculum size, and temperatureand duration of incubation accordingto the manufacturer’s instructions. Re-peated biotyping was carried out to en-sure reproducibility of the Vitek codes.When the identification results indi-cated Æ80% confidence level, a con-firmatory biotyping test was conductedusing the Microscan system (22).

Antimicrobial susceptibility

The aerobic and facultatively anaerobicgram-negative rods and cocci sequentialisolates were also subjected to standardantimicrobial susceptibility testing asper the National Committee for Clin-ical Laboratory Standards (NCCLS)(20) protocol. Thus the isolates were

tested for susceptibility to cefotaxime,ceftazidime, cefuroxime, cephalothin,chloramphenicol, co-trimoxazole,gentamicin, imipenem, nalidixic acid,nitrofurantoin and ofloxacin using thedisk diffusion test (3, 20). NCCLS in-terpretive criteria were used for deter-mining resistance, intermediate suscep-tibility, or susceptibility.

Chromosomal analysis by pulsed-field gelelectrophoresis

The extraction of genomic DNA andthe conditions for pulsed-field gel elec-trophoresis were carried out as pre-viously described (4) with minor modi-fications. In brief, a young bacterial cellsuspension at suitably adjusted tur-bidity (optical density at 660 nmΩ0.6)was directly embedded in an equal vol-ume of 2% low-melting-temperatureagarose (BioRad Laboratories, Rich-mond, CA). Solidified agarose gel plugswere first treated with 1 mg/ml lyso-zyme plus 2 mg/ml RNase in 0.5% sar-kosyl – 0.2% sodium deoxycholate – 1M NaCl – 0.1 M EDTA – 10 mM Tris,pH 8.8 at 37æC, 1 h and then with lysisbuffer containing 1 mg/ml proteinase Kin 1% sarkosyl – 0.5 M EDTA, pH 0.8at 50æC, overnight. The plugs were thenrinsed and transferred to 1 mM phenyl-methylsulfonyl fluoride, 30 min, roomtemperature twice for inactivation ofproteinase K, cut to appropriate size,and then digested with 16 U of XbaI(Pharmacia Biotech, Uppsala, Sweden)at 37æC, 4 h. Pulsed-field gel electro-phoresis was performed with a 1% agar-ose gel by using a CHEF Mapper ap-paratus (BioRad Laboratories) or GeneNavigator Pulsed-field System (Pharm-

Table 1. Demographic and other clinical parameters of the study population

Groupa

A B C D

Age 48.1∫5.7b 66.5∫3.5 43.9∫11.8 69.5∫3.3Number 25 8 20 10% male 64 100 60 100% denture wearersc 8 25 10 80Years 0.5–19.2 0.5–3.4post-irradiation (4.9∫4.5)b (1.8∫1.1)Stimulated whole saliva

– flow rate (ml/min)d 0.05∫0.04b 0.04∫0.04 0.9∫0.6 1.1∫0.7– pHd 6.5∫0.5b 6.5∫0.8 7.2∫0.3 7.4∫0.3

a Group A,60 years, irradiated individuals; group BØ60 years, irradiated individuals; groupC,60 years, control individuals; group DØ60 years, control individuals.

b Mean∫SD.c PΩ0.06, Fisher’s exact test.d Significantly different (P,0.05, Bonferroni multiple comparison) of data value between:

groups A vs C, A vs D, B vs C, B vs D.

acia Biotech) in 2.5 mM EDTA – 89mM boric acid – 89 mM tris, pH 8.0at 13æC with appropriate settings. Afterpulsed-field gel electrophoresis, the gelswere stained with ethidium bromide(0.2 mg/ml) and photographed underultraviolet transillumination.

Statistics

Demographic and microbiological dataof the subjects were analyzed by Stat-view 4.5 for Macintosh computer (26).Differences between individual groupswere tested by Bonferroni multiplecomparison for nonparametric data, orFisher’s exact test as appropriate. Post-irradiation time and quantity of aerobicand facultatively anaerobic gram-nega-tive rods and cocci species isolationwere studied using Spearman rank cor-relation. Groups were regarded as sig-nificantly different from each other ifP,0.05.

Results

The demographic data, denture wearinghabits and oral physiological par-ameters including stimulated whole-sal-iva flow rate and salivary pH of the testand control subjects recruited in thefirst part of the study are shown inTable 1.

The individual bacterial families andtheir number and identity of aerobicand facultatively anaerobic gram-nega-tive rods and cocci retrievable from oralrinse samples of each individual groupsare recorded in Tables 2 and 3 respec-tively. A mean of 0.25 and 0.9 aerobicand facultatively anaerobic gram-nega-tive rods and cocci species were retriev-

4 Leung et al.

Table 2. Mean number (∫ SD) of species from different aerobic and facultatively anaerobicgram-negative rods and cocci families isolated from oral rinse samples

Groupa

A B C D

Neisseriaceae 0.16∫0.37 0.13∫0.35 0.05∫0.22 0.3∫0.483Pseudomonadaceae 0.32∫0.48 0.25∫0.46 0.05∫0.22 0.2∫0.422Enterobacteriaceaeb 0.40∫0.65 1.00∫1.07 0.05∫0.22 0.2∫0.422Aerobic and facultatively anaer-

obic gram-negative rods andcocci spp.c 0.88∫0.88 1.38∫0.74 0.25∫0.44 0.9∫0.568

a Group A,60 years, irradiated individuals; group B Ø 60 years, irradiated individuals; groupC,60 years, control individuals; group D Ø 60 years, control individuals.

b Significantly different (P,0.05, Bonferroni multiple comparison) data value between: groupsB vs C and B vs D.

c Significantly different (P,0.05, Bonferroni multiple comparison) data value between: groupsB vs C.

Table 3. Percentage frequency of aerobic and facultatively anaerobic gram-negative rods andcocci species isolated from oral rinse samples

Groupa

A B C D

Neisseriaceae 16.0 12.5 5.0 30.0Acinetobacter spp. 16.0 12.5 0 30.0Neisseria spp. 0 0 5.0 0

Pseudomonadaceae 32.0 25.0 5.0 20.0Chryseomonas luteola 16.0 0 0 0Flavimonas oryzihabitans 4.0 0 0 0Pseudomonas aeruginosa 8.0 25.0 0 10.0Pseudomonas fluorescenc/putida 0 0 5.0 0Stenotrophomonas maltophilia 4 0 0 10.0

Enterobacteriaceae 32.0b 62.5 5.0 30.0Citobacter freundii 0 25.0 0 0Enterobater cloacae 12.0 37.5 5.0 20.0Escherichia coli 4.0 0 0 10.0Klebsiella

pneumoniae subsp. pneumoniaec 24.0 37.5 0 0

OthersFlavobacterium indologenes 0 0 5.0 0Weeksella virosa 0 0 5.0 0

Total aerobic and facultatively anaerobicgram-negative rods and coccic 64.0b 100.0 25.0 80.0

a Group A,60 years, irradiated individuals; group BØ60 years, irradiated individuals; groupC,60 years, control individuals; group DØ60 years, control individuals.

b Total or sub-total value does not add up because some samples contain more than onespecies of the same family.

c Significantly higher prevalent of total aerobic and facultatively anaerobic gram-negative rodsand cocci and K. pneumoniae subsp. pneumoniae isolated in oral rinse of irradiated (groupsA plus B) than normal (groups C plus D) individual, Fisher’s exact test, P,0.05.

able from oral rinse samples of groupsC and D respectively. From the oralrinse samples of irradiated individuals,that is, groups A and B, a mean of 0.9and up to 1.4 aerobic and facultativelyanaerobic gram-negative rods and coccispecies per patient were recovered. Al-most all of the group B individuals(head and neck–irradiated subjects Ø60years) were found to be colonized by atleast one species of aerobic and faculta-

tively anaerobic gram-negative rods andcocci (Tables 2, 3). Enterobacteriaceaewere the most common bacterial familyisolated from irradiated individuals(Table 2) particularly in group B (Table3) with Enterobacter cloacae, Klebsiellapneumoniae subsp. pneumoniae and/orCitrobacter freundii together isolatedfrom more than half of the oral rinsesamples. Of the Pseudomonadaceae,Pseudomonas aeruginosa was the most

common and was found in one quarterof the group B samples (Table 3). A sig-nificantly higher prevalence of total aer-obic and facultatively anaerobic gram-negative rods and cocci and K. pneu-moniae subsp. pneumoniae was noted insamples of irradiated than normal indi-viduals (Table 3).

Further quantitative (colony-form-ing units/ml) and qualitative data (spe-cies of Ø15% prevalence) related to aer-obic and facultatively anaerobic gram-negative rods and cocci recovered fromthe oral rinse samples are summarizedin Table 4. A remarkable range in thequantities of aerobic and facultativelyanaerobic gram-negative rods and coccispecies recovered from the oral rinsesamples was noted. C. freundii wasfound to be significantly elevated in ir-radiated subjects compared with age-and sex-matched controls (group B vsD, Table 4). No correlation could befound between post-irradiation dur-ation and quantity of aerobic and fa-cultatively anaerobic gram-negativerods and cocci species isolated (data notshown).

A total of 12 (6 males, 6 females;45.9∫5.8 years, 5.0∫3.1 years post ir-radiation) and 5 (all males; 67.1∫3.4years, 2.0∫1.1 years post irradiation)subjects from groups A and B respec-tively participated in the recall reassess-ment of aerobic and facultatively anaer-obic gram-negative rods and coccicolonization. The remainder of thegroups A and B subjects either missedthe recall, attended their own generaldentists or succumbed to the disease.

The prevalence of aerobic and fa-cultatively anaerobic gram-negativerods and cocci species isolated fromoral rinses of the 17 individuals beforeand after hygienic therapy are summar-ized in Table 5. The effect of the hy-gienic therapy as indicated by thecolonization patterns of the aerobic andfacultatively anaerobic gram-negativerods and cocci is shown in Table 6. Theirradiated individuals that were fol-lowed up were found to be significantlyand persistently colonized by aerobicand facultatively anaerobic gram-nega-tive rods and cocci species (88% pre-hy-gienic therapy, 76% post-hygienic ther-apy, PΩ0.04, Table 6). Overall the hy-gienic therapy per se did not appear tobe effective in retarding the aerobic andfacultatively anaerobic gram-negativerods and cocci colonization, despite thefact that species such as Chryseomonasluteola, C. freundii, Flavimonas oryzih-

Oral coliforms, post-irradiation 5

Table 4. Quantity (colony-forming units/ml) of cultivable aerobic and facultatively anaerobicgram-negative rods and cocci from oral rinse samples

Groupa

A B C D

Neisseriaceaerange 0–105 0–894 0–61 0.231¿104

median 0 0 0 0Acinetobacter spp.b 0–105 0–894 0 0–2.31¿104

Pseudomonadaceaerange 0–1423 0–4.63¿104 0–61 0–2760median 0 0 0 0

Chryseomonas luteolab 0–224 0 0 0–254

Pseudomonas aeruginosab 0–1220 0–4.63¿104 0 0–2030

Enterobacteriaceaerange 0–1321 0–447 0–20 0–4.55¿104

median 0 122 0 0Citrobacter freundiib,c 0 0–203 0 0Enterobacter cloacaeb 0–203 0–81 0–20 0–305Klebsiella pneumoniae

subsp. pneumoniaeb 0–650 0–346 0 0–1.36¿104

Total aerobic and facultativelyanaerobic gram-negative rodsand coccid

range 0–105 41–4.63¿104 0–244 0–4.55¿104

median 20 325 0 5a Group A,60 years, irradiated individuals; group B Ø 60 years, irradiated individuals; group

C,60 years, control individuals; group D Ø 60 years control individuals.b Only species of prevalence Ø 15% shown; median valuesΩ0.c Significantly different (P,0.05, Bonferroni multiple comparison) of data value between:

group A vs B, B vs C, B vs D.d Reported normal range of combined Enterobacteriaceae, Pseudomonadacea and Acineto-

bacter speciesΩ10–1¿103 colony-forming units/ml (36).

Table 5. Percentage frequency of aerobic and facultatively anaerobic gram-negative rods andcocci species isolated from oral rinse samples pre- and post-hygienic phase in irradiated sub-jectsa

Group A Group B

Before After Before After

Neisseriaceae 16.7 33.3 0 60.0Acinetobacter spp. 16.7 25.0 0 60.0Neisseria spp. 0 8.3 0 0

Pseudomonadaceae 33.3 8.3 20.0 40.0Chryseomonas luteola 25.0 0 0 0Flavimonas oryzihabitans 8.3 0 0 0Pseudomonas aeruginosa 0 0 20.0 40.0Pseudomonas spp. 0 8.3 0 0

Enterobacteriaceae 33.3b 58.3b 80.0b 80.0b

Citrobacter freundii 0 0 40.0 0Enterobacter cloacae 8.3 41.7 40.0 40.0Escherichia coli 8.3 0 0 0Klebsiella pneumoniae subsp. pneumoniae 25.0 50.0 40.0 60.0Flavobacterium meningosepticoum 0 0 0 20.0

Total aerobic and facultatively anaerobicgram-negative rods and cocci 66.7b 83.3b 100.0 100.0b

a Group A, ∞60 years, irradiated individuals; nΩ12; group B Ø60 years, irradiated indi-viduals; nΩ5, only data from subjects who participated in both pre- and post-hygienictherapy are included.

b Total/sub-total values do not add up as some individuals were co-infected with more thanone species.

abitans and Escherichia coli were not re-coverable after the hygiene procedures(Table 5). K. pneumoniae subsp. pneu-moniae appeared to colonize and persistin the oral cavity of 24% of individualsstudied while colonization at least atany one of the two time points was 59%(Table 6). Both E. cloacae and Acineto-bacter spp. were not resident oral colon-izers of the post-irradiated individuals,although they exhibited transient car-riage to varying degrees.

The phenotypic and genotypiccharacteristics of the aerobic and fa-cultatively anaerobic gram-negativerods and cocci species that were sequen-tially isolated from the identical ir-radiated individual on two successiveoccasions were studied (Table 7). Fourisolate pairs of K. pneumoniae subsp.pneumoniae and a single isolate-pair ofP. aeruginosa were found to be pheno-typically identical as per API 20E char-acterization. However, upon genotypiccharacterization by pulsed-field gel elec-trophoresis, only two of the four pairsof K. pneumoniae were found to beidentical (Table 7, Fig. 1). The one iso-late pair of E. cloacae seemed to be dif-ferent based on the API biochemicalcharacterization, and so was the P.aeruginosa isolate pair that was foundto be phenotypically very different byVitek biotyping and to a lesser extentby antibiogram profile. The latter iso-late-pair was genotypically dissimilar aswell on pulsed-field gel electrophoresisprofiling.

Discussion

The present study evaluated the preva-lence, quantity and the possible strain di-versity of aerobic and facultatively an-aerobic gram-negative rods and cocci ina homogeneous group of nasopharynge-al carcinoma survivors subjected totherapeutic head and neck irradiation.To our knowledge this is the first studyto characterize aerobic and facultativelyanaerobic gram-negative rods and coccioral isolates from irradiated individualssuffering from solid tumors similar innature and subjected to the same thera-peutic irradiation protocol. A few re-ports are available on the increased oralcolonization of aerobic and facultativelyanaerobic gram-negative rods and cocciin mucositis either during or immedi-ately after irradiation therapy for whatwere usually a mixed collection ofpatients with various head and neckcancers (1, 16, 36). However only a few

6 Leung et al.

Table 6. Effect of hygienic therapy on oral colonization of aerobic and facultatively anaerobicgram-negative rods and cocci species in irradiated individualsa

IsolationIsolation after hygienicAerobic and facultatively anaerobic before

therapybgram-negative rods and cocci hygienicspecies therapyb π ª Pc

Any aerobic and facultativelyanaerobic gram-negative rods π 15 (88)d 13 (76) 2 (12)and cocci species ª 2 (12) 0 (0) 2 (12)

0.04

π 2 (12) 0 (0) 2 (12)Acinetobacter spp. ª 15 (88) 6 (35) 9 (52) NS

π 4 (24) 1 (6) 3 (18)Enterobacter cloacae ª 13 (76) 6 (35) 7 (41) NS

Klebisella pneumoniae subsp. π 5 (29) 4 (24) 1 (6)pneumoniae ª 12 (71) 5 (29) 7 (41) NS

π 1 (6) 1 (6) 0 (0)Pseudomonas aeruginosa ª 16 (94) 1 (6) 15 (88) NS

π 7 (41) 0 (0) 7 (41)Otherse ª 10 (59) 2 (12) 8 (47) NS

π 13 (76) 1 (6) 12 (71)Totalf ª 4 (24) 2 (12) 2 (12) NSa 17 subjects followed; please refer to text and Table 7 for details.b π: positive recovery; -: negative recovery.c Fisher’s extact test, NSΩnot significant.d actual count; data in parenthesesΩ%.e aerobic and facultatively anaerobic gram-negative rods and cocci other than E. cloacae,

K. pneumoniae subsp. pneumoniae, P. aeruginosa and Acinetobacter spp.f Only exactly identical isolation pattern recorded, i.e. same individual isolate(s) or group of

isolates before and after hygienic therapy.

Table 7. Phenotypic and genotypic characteristics of the sequential isolate-pairs of aerobic and facultatively anaerobic gram-negative rodsobtained from oral rinse samples

Vitek AntibiogramaPulsed-field gel

API 20E Confidence electrophoresisIsolates no. Profile Bionumber % CAZ CEF CHL CTX CXM GEN IPM NAL NIT OFX SXT profileb

Klebsiellapneumoniaesubsp.pneumoniae

1-ic 5215773 6674763631 99 S S S S S S S S S S S π1-ii 5215773 6674773631 99 S S S S S S S S S S S2-i 5215773 6664773631 96 S S S S S S S S I S S ª2-ii 5215773 6674771631 99 S S S S S S S S S S S3-i 5215773 6274773631 99 S S S S S S S S S S S π3-ii 5215773 6664773631 96 S S S S S S S S S S S4-i 5215773 6674773631 99 S S S S S S S S S S S ª4-ii 5215773 6674771631 99 S S S S S S S S S S S

Pseudomonasaeruginosa

5-i 2202000 61604000440 98 S R R I R S S R R S R ª5-ii 2202000 40600000440 59d S R R S R S S R R S S

Enterobactercloacae

6-ie 3305763 – – – – – – – – – – – – – ND6-ii 1305563 6424375232 86 S R S S S S S S S S S

a Abbreviations: CAZ, ceftazidime; CEF, cephalothin; CHL, chloramphenicol; CTX, cefotaxime; CXM, cefuroxime; GEN, gentamicin; IPM,imipenem; NAL, nalidixic acid; NIT, nitrofurantoin; OFX, ofloxacin; SXT, co-trimoxazole; R, resistance; I, intermediate; S, susceptible.

b Pulsed-field gel electrophoresis profile generated by digestion of genonic DNA of isolates with XbaI; -, not identical between isolates i andii; π, identical between isolates i and ii; ND, not determined.

c i, isolated before hygienic therapy, ii, isolated after hygienic therapy.d Identification confirmed by Microscan System (22).e Isolate lost upon subculturing.

have investigated the delayed coloniza-tion patterns of these organisms in headand neck–irradiated individuals, par-ticularly after resolution of irradiation-induced mucositis. The subjects in thecurrent report were at least 6 monthsafter their radiotherapy, and none hadclinical symptoms of mucositis. Mak-konen et al. (16) using a swab techniquefor sampling the oral flora, have re-ported a reduction in oral colonizationof gram-negative bacilli down to 5%, 6months after irradiation. On the con-trary, using the oral rinse samplingmethod with a detection level of 2 col-ony-forming units/ml oral rinse, a veryhigh aerobic and facultatively anaerobicgram-negative rods and cocci isolationfrequency was observed in irradiated in-dividuals in the current study (Table 3).This discrepancy may be due to tworeasons. First, the superior sensitivityof the oral rinse technique as comparedwith swab sampling (24), and second,the higher prevalence of oral aerobicand facultatively anaerobic gram-nega-tive rods and cocci in southern Chinesepopulation groups (31).

According to the current results, thedifference in isolation rate of aerobicand facultatively anaerobic gram-nega-

Oral coliforms, post-irradiation 7

Fig. 1. Pulsed-field gel electrophoresis of XbaI-digested genomic DNA from aerobic and fa-cultatively anaerobic gram-negative rods and cocci isolates. Electrophoresis under A) CHEFprotocol, B) Navigator protocol. Lanes 1, 8 and 13: molecular weight marker; lanes 2–7:isolates no. 3-i, 3-ii, 2-i, 2-ii, 4-i, 4-ii respectively; lanes 9–10: isolate no. 5-i and 5-ii; lanes 11–12: isolates no. 1-i and 1-ii (cf. Table 7).

tive rods and cocci was marked betweenthe two age groups less than 60 years(group A, 64% vs group C, 25%). Theaerobic and facultatively anaerobicgram-negative rods and cocci isolationrate of the group C subjects was akinto that of a similar cohort in a previousstudy from Hong Kong, i.e. 25% vs 31%(31). The quantity of aerobic and fa-cultatively anaerobic gram-negativerods and cocci isolated from group Cwas also within the range of a previousreport (36). These results therefore im-ply that head and neck irradiation hasa lasting effect on the oral aerobic andfacultatively anaerobic gram-negativerods and cocci carriage, possibly due todisturbances in the normal physiologi-cal attributes of the oral cavity. Con-trary to expectations, the aerobic andfacultatively anaerobic gram-negativerods and cocci isolation rates appearedto be similar in the two older agegroups (test group B vs control groupD). This may be due to the fact that80% of the healthy vs 25% of the ir-radiated individuals wore dentures(Table 1). Patients with dentures areknown to have higher aerobic and fa-cultatively anaerobic gram-negativerods and cocci colonization rates (9),and Sedgley & Samaranayake (31) havealso shown that in healthy Hong Kongdenture wearers of age Ø35 years, the

prevalence of these organisms is 58% vs41% in denture-free individuals. Thisfactor may have obfuscated the physio-logical effects of irradiation such asxerostomia in the older age groups,yielding this unexpected result.

The lowest detection level of aerobicand facultatively anaerobic gram-nega-tive rods and cocci by the oral rinsesampling protocol employed was 2 col-ony-forming units/ml and 0.5 colony-forming units/ml for groups A, B andC, D, respectively, implying that aerobicand facultatively anaerobic gram-nega-tive rods and cocci populations at alower density than the foregoing valueswould not be cultivable. Using the iden-tical detection protocol, Sedgley (28) re-covered higher viable counts of Enter-obacteriaceae from hospitalized thanoutpatient adults. Similarly, 66% ofaerobic and facultatively anaerobicgram-negative rods and cocci isolatedfrom burning mouth syndrome patientshad greater than 40 colony-formingunits/ml oral rinse, in contrast to 42%carriage in healthy controls (23). Al-though it has been reported that theoral prevalence of aerobic and faculta-tively anaerobic gram-negative rods andcocci increases with severity of the hostillness among the hospitalized patients(12) or with increasing age in the elderlycommunity-dwelling Hong Kong Chi-

nese (31), the biological significance oforal aerobic and facultatively anaerobicgram-negative rods and cocci carriageeither in health or disease in normal orcompromised hosts is yet to be deter-mined.

Improvements in oral hygiene as wellas provision of oral health care did notappear to significantly affect the colon-ization rate of the aerobic and faculta-tively anaerobic gram-negative rods andcocci in irradiated individuals in groupsA and B (Tables 5, 6). This also impliesthat the organisms in question are notresiding in the oral cavity by virtue ofpoor oral hygiene, and more fundamen-tal changes in the oral milieu due to ir-radiation have promoted their coloniza-tion.

Aerobic and facultatively anaerobicgram-negative rods and cocci are gener-ally regarded as transient inhabitants ofthe healthy human oral cavity (30). Pre-vious data from those undergoing headand neck radiotherapy indicate that theoccurrence and severity of the mucositisis related to the colonization of aerobicand facultatively anaerobic gram-nega-tive rods and cocci (25, 36). Indeed,some have shown that eliminating theaerobic and facultatively anaerobicgram-negative rods and cocci usingvarious antimicrobial agents often re-sults in reduction/prevention of the mu-cositis (25, 37). In one series of experi-ments, Spijkervet (36) reported admin-istering lozenges containing polymyxinE, tobramycin and amphotericin Bcould lead to reduction of Enterobacteri-aceae/Pseudomonadaceae/Acinetobacterspp. levels (,log 0.1 colony-formingunits/ml of sample) as well as mucos-itis incidence in cancer subjects under-going irradiation therapy. However, nodata were given regarding the normallevel of such bacterial species inhealthy individuals. The level of oralcarriage of aerobic and facultativelyanaerobic gram-negative rods and coc-ci in the present cohort was compar-able to that of hospitalized patients inHong Kong and the United States (34,39). The reason for such increased aer-obic and facultatively anaerobic gram-negative rods and cocci colonizationremains obscure. Johanson et al. (13)demonstrated that oropharyngeal epi-thelial cells from intensive care patientsalready colonized by aerobic and fa-cultatively anaerobic gram-negativerods and cocci were more receptive forthe microbes than un-colonized con-trols. It is likely that variations in ex-

8 Leung et al.

pression of intrinsic factors such asintegrin on mucosal epithelia cell sur-faces (7) might encourage colonizationof aerobic and facultatively anaerobicgram-negative rods and cocci. However,the reasons for the high colonizationrate of the aerobic and facultatively an-aerobic gram-negative rods and cocciwe witnessed in irradiated cohorts whono longer suffer from severe mucositis(27) is yet to be elucidated.

With regard to the individual aerobicand facultatively anaerobic gram-nega-tive rods and cocci species isolated fromirradiated individuals, the most pre-dominant isolate was K. pneumoniaesubsp. pneumoniae (Table 3). Oral Kleb-siella carriage is known to increase sig-nificantly during the period immedi-ately after irradiation (1). In anothersimilar study by Samaranayake et al.(25), K. pneumoniae was identified asthe most common coliform isolated de-spite the use of antiseptic mouthrinsesby the study participants. We reaffirmthese observations and demonstratefurther that it is still the case for sub-jects who are at least 6 months post-irradiation. This observation is interest-ing, as Ayars et al. (2) have demon-strated increased K. pneumoniae adher-ence in vitro to normal human buccalcells in a reduced pH environment andincreased affinity of the bacteria to buc-cal cells of xerostomic individuals.Further, using human buccal cells, tra-cheal cells and lung tissue cell sections,Hornick et al. (11) demonstrated thatK. pneumoniae adherence to epithelialcells could be mediated via the interac-tion between the bacterial type 3 fim-briae and peptide molecule(s) on theepithelial cell surface, whereas othershave reported that K. pneumoniae type1 fimbriae could also play a part in ad-hesion to respiratory epithelium (8).

A number of studies have indicatedthat a diminished salivary flow rateeither due to physiological or patholog-ical reasons is associated with increasedrisk for orophayngeal colonization byaerobic and facultatively anaerobicgram-negative rods and cocci (2, 16).Further, other data from southern Chi-nese indicate that prevalence of aerobicand facultatively anaerobic gram-nega-tive rods and cocci increases with age(28, 31). Hence the higher prevalence ofaerobic and facultatively anaerobicgram-negative rods and cocci in all ofour cohorts is not surprising in view oftheir age range. Nonetheless, 100% oralcarriage rate of aerobic and faculta-

tively anaerobic gram-negative rodsand cocci a significant period beyond ir-radiation therapy, in group B for in-stance, implies that the latter protocolin some manner provides conditionsthat, in tandem with age related physio-logical changes, foster aerobic and fa-cultatively anaerobic gram-negativerods and cocci colonization of the oralcavity.

A few workers have shown that cer-tain enteric rods and pseudomonadscolonizing the subgingival area, may beetiologically involved in refractory peri-odontitis (6, 35). Despite this, the naso-pharyngeal carcinoma patients weexamined were not found to be at in-creased risk for periodontitis (27). In-deed in a recent investigation of thesubgingival flora of the identical studygroup, we were unable to detect a sig-nificant degree of colonization by atypi-cal bacteria among the normal subgin-gival flora (14).

In the later part of the current studywe investigated whether the identicalaerobic and facultatively anaerobicgram-negative rods and cocci straincould consistently colonize the oral cav-ity of the irradiated subjects over aperiod of 3 months, despite hygienictherapy. A number of established bio-chemical and molecular methods wereused for this purpose (4, 19) to charac-terize six sequential isolates-pairs ofaerobic and facultatively anaerobicgram-negative rods. The phenotypic andthe genotypic characterization results ofthe isolates complemented each other.Other than the two sequential isolate-pairs of K. pneumoniae subsp. pneumon-iae, the rest of the aerobic and faculta-tively anaerobic gram-negative rodspairs were found to be unrelated transi-ent colonizers of the mouth (Table 7).

The antibiogram assay of the aerobicand facultatively anaerobic gram-nega-tive rods sequential isolate-pairs indi-cated that the vast majority except P.aeruginosa were sensitive to the anti-microbial agents tested (Table 7). Theseresults are comparable to the anti-microbial sensitivity patterns of oralcoliforms isolated from healthy HongKong adults (32). Thus our findings im-ply that aerobic and facultatively anaer-obic gram-negative rods and cocci, ex-cept perhaps P. aeruginosa, colonizingthe oral cavities of the irradiated naso-pharyngeal carcinoma patients mayhave limited pathogenic potential. Al-though oral colonization by P. aerugi-nosa was not significantly enhanced due

to irradiation-induced oral changes, thefact that two of the P. aeruginosasequential isolates were resistant to 5 ofthe 11 tested antibiotics underlines thepotential risk these bacteria may poseparticularly on access to the host air-way (15).

In conclusion, the current study dem-onstrates that aerobic and facultativelyanaerobic gram-negative rods and cocciare highly prevalent in head and neck–irradiated southern Chinese subjects.Nevertheless, it appears that the aerobicand facultatively anaerobic gram-nega-tive rods and cocci oral colonization inour study population is rather transi-ent. Perhaps as reported in previousstudies, the irradiation-induced xero-stomia and related oral changes favorthe initial oral colonization of aerobicand facultatively anaerobic gram-nega-tive rods and cocci, but most such mi-crobes, unlike the dental plaque flora,fail to establish themselves intraorallyin the longer term, at least in the cohortof irradiated individuals followed. Fi-nally, as the current cohort studied wasrelatively small, more comprehensive in-vestigations with larger populationgroups are required to confirm our pre-liminary findings.

Acknowledgments

We thank Jonathan S.T. Sham of Clin-ical Oncology, University of HongKong, for assistance in groups A and Bsubjects recruitment, Gordon K.C.Chiu of Reception and Primary CareClinic, Faculty of Dentistry, Universityof Hong Kong, for assistance in re-cruiting group C subjects, Grace Yungof the University of Hong Kong, Eug-ene Lam and Raymond Leung ofQueen Mary Hospital, Hong KongHospial Authority, for technical assist-ance and Nerissa Chan for help in pre-paring the manuscript. This project wassupported by CRCG from the Univer-sity of Hong Kong.

References

1. Abu Shara KA, Ghareeb MA, Zaher S,Mobacher A, Khalifa MC, Saleh SZ.Radiotherapeutic effect on oropharyn-geal flora in patients with head and neckcancer. J Laryngol Otol 1993: 107: 222–227.

2. Ayars GH, Altman LC, Fretwell MD.Effect of decreased salivation and pH onthe adherence of Klebsiella species to hu-man buccal epithelial cells. Infect Im-mun 1982: 38: 179–182.

Oral coliforms, post-irradiation 9

3. Bauer AW, Kirby WMM, Sherris JC,Turck M. Antibiotic susceptibility test-ing by a standardized single diskmethod. Am J Clin Pathol 1966: 45:493–496.

4. Cameron DN, Khambaty FM, Wachs-muth IK, Tauxe RV, Barrett TJ. Molecu-lar characterization of Vibrio choleraeO1 stains by pulsed-field gel electro-phoresis. J Clin Microbiol 1994: 32:1685–1690.

5. Chen KY, Fletcher GH. Malignant tu-mours of the nasopharynx. Radiology1971: 99: 165–171.

6. Colombo AP, Haffajee AD, DewhirstFE, Paster BJ, Smith CM, Cugini MA etal. Clinical and microbiological featuresof refractory periodontitis subjects. JClin Periodontol 1998: 25: 169–180.

7. de Bentzmann S, Plotkowski C, PuchelleE. Receptors in the Pseudomonas aerugi-nosa adherence to injured and repairingairway epithelium. Am J Respir CritCare Med 1996: 154: S155–S162.

8. Fader RC, Gondesen K, Tolley B, Ritch-ie DG, Moller P. Evidence that in vitroadherence of Klebsiella pneumoniae tociliated hamster tracheal cells is me-diated by type 1 fimbriae. Infect Immun1988: 56: 3011–3013.

9. Goldberg S, Cardash H, Browning HIII, Sahly H, Rosenberg M. Isolation ofEnterobacteriaceae from the mouth andpotential association with malodor. JDent Res 1997: 76: 1770–1775.

10. Hong Kong Cancer Registry. Annual re-port 1991. Hong Kong: Hospital Auth-ority, 1995: 12.

11. Hornick DB, Allen BL, Horn MA,Clegg S. Adherence to respiratory epi-thelia by recombinant Escherichia coliexpressing Klebsiella pneumoniae type 3fimbrial gene products. Infect Immun1992: 60: 1577–1588.

12. Johanson WG Jr, Pierce AK, Sanford JP.Changing pharyngeal bacterial flora ofhospitalized patients. Emergence ofgram-negative bacilli. N Engl J Med1969: 281: 1137–1140.

13. Johanson WG Jr, Woods DE, Chaud-huri T. Association of respiratory tractcolonization with adherence of gram-negative bacilli to epithelial cells. J InfectDis 1979: 139: 667–673.

14. Leung WK, Jin LJ, Samaranayake LP,Chiu KC. Subgingival microbiota ofshallow periodontal pockets in individ-uals after head and neck irradiation.Oral Microbiol Immunol 1998: 13: 1–10.

15. Lode H. Microbiological and clinical as-pects of aspiration pneumonia. J Anti-microb Chemother 1988: 21 (suppl C):83–90.

16. Makkonen TA, Borthen L, Heimdahl A,Joensuu H, Lehtonen OP, Nord CE. Or-opharyngeal colonization with fungi andgram-negative rods in patients treatedwith radiotherapy of the head and neck.Br J Oral Maxillofac Surg 1989: 27: 334–340.

17. Martin MV, Al-Tikriti U, Bramley PA.Yeast flora of the mouth and skin duringand after irradiation for oral and laryn-geal cancer. J Med Microbiol 1981: 14:457–467.

18. Mengistu Y, Gedebou M. Aerobic gram-negative pharyngeal bacilli of adult Ethi-opians: carrier rates and antibiograms. JHyg (Lond) 1986: 97: 247–253.

19. Miller JM, O’Hara CM. Substrate util-ization systems for the identification ofbacteria and yeasts. In: Murray PR,Baron EJ, Pfaller MA, Tenover FC,Yolken RH, ed. Manual of clinicalmicrobiology. Washington, DC: Am SocMicrobiol, 1995: 103–109.

20. National Committee for Clinical Lab-oratory Standards. Performance stan-dards for antimicrobial disk susceptibil-ity tests: approved standard M2-A5.Villanova, PA: NCCLS, 1993.

21. Philpot CR, McDonald PJ, Chai KH.Significance of enteric gram-negative ba-cilli in the throat. J Hyg (Lond) 1980:85: 205–210.

22. Rhoden DL, Smith PB, Baker CN,Schable B. AutoSCAN-4 system foridentification of gram-negative bacilli. JClin Microbiol 1985: 22: 915–918.

23. Samaranayake LP, Lamb AB, Lamey P-J, MacFarlane TW. Oral carriage ofCandida species and coliforms in pa-tients with burning mouth syndrome. JOral Pathol Med 1989: 18: 233–235.

24. Samaranayake LP, MacFarlane TW, La-mey P-J, Ferguson MM. A comparisonof oral rinse and imprint sampling tech-niques for the detection of yeast, coli-form and Staphylococcus aureus carriagein the oral cavity. J Oral Pathol 1986: 15:386–388.

25. Samaranayake LP, Robertson AG, Mac-Farlane TW, Hunter IP, MacFarlane G,Soutar DS et al. The effect of chlorhex-idine and benzydamine mouthwashes onmucositis induced by therapeutic ir-radiation. Clin Radiol 1988: 39: 291–294.

26. SAS Institute. StatViewA. 2nd edn.Cary, NC: SAS Institute Inc. 1998.

27. Schwarz E, Chiu GKC, Leung WK.Oral health status of southern Chinesefollowing head and neck irradiationtherapy for nasopharyngeal carcinoma.J Dent 1999: 27: 21–28.

28. Sedgley CM. Enterobacteriaceae in the

human oral cavity. Ph.D thesis. HongKong: University of Hong Kong, 1995.

29. Sedgley CM, Chu CS, Lo ECM, Sam-aranayake LP. The oral prevalence ofaerobic and facultative anaerobic gram-negative rods and yeasts in semi-reclusehuman vegetarians. Arch Oral Biol 1996:41: 307–309.

30. Sedgley CM, Samaranayake LP. Oraland oropharyngeal prevalence of Enter-obacteriaceae in humans: a review. J OralPathol Med 1994: 23: 104–113.

31. Sedgley CM, Samaranayake LP. Theoral prevalence of aerobic and faculta-tively anaerobic gram-negative rods andyeasts in Hong Kong Chinese. Arch OralBiol 1994: 39: 459–466.

32. Sedgley CM, Samaranayake LP. Anti-microbial susceptibility of oral isolatesof Enterobacter cloacae and Klebsiellapneumoniae from a southern Chinesepopulation. Oral Microbiol Immunol1998: 13: 315–321.

33. Sedgley CM, Samaranayake LP, ChanJCY, Wei SHY. A 4-year longitudinalstudy of the oral prevalence of entericgram-negative rods and yeasts in Chi-nese children. Oral Microbiol Immunol1997: 12: 183–188.

34. Sedgley CM, Samaranayake LP, HuWH, Lee MT. The oral prevalence ofaerobic and facultatively anaerobicgram-negative rods and yeasts in hospi-talised patients. Microb Ecol Health Dis1995: 8: 225–234.

35. Slots J, Feik D, Rams TE. Age and sexrelationships of superinfecting micro-organisms in periodontitis patients. OralMicrobiol Immunol 1990: 5: 305–308.

36. Spijkervet FKL. Irradiation mucositis.Prevention and treatment. Copenhagen:Munksgaard, 1991.

37. Spijkervet FKL, van Saene HK, van Sa-ene JJ, Panders AK, Vermey A, MehtaDM. Mucositis prevention by selectiveelimination of oral flora in irradiatedhead and neck cancer patients. J OralPathol Med 1990: 19: 486–489.

38. Tenovuo JO. Human saliva: clinicalchemistry and microbiology. Vol. I. BocaRaton, FL: CRC Press, 1989.

39. Wahlin YB, Holm A-K. Changes in theoral microflora in patients with acuteleukemia and related disorders duringthe period of induction therapy. OralSurg Oral Med Oral Pathol Oral RadiolEndod 1988: 65: 411–417.

40. Wei WI, Ho CM, Yuen PW, Fung CF,Sham JS, Lam KH. Maxillary swing ap-proach for resection of tumors in andaround the nasopharynx. Arch Oto-laryngol Head Neck Surg 1995: 12: 638–642.