Fusobacterium necrophorum detection and

identification on a selective agar

STEFFEN BANK, HANNEMERETE NIELSEN, BORIS HOYERMATHIASEN,DORTE CHRISTIANSEN LETH, LENA HAGELSKJR KRISTENSEN and JRGEN PRAG

Department of Clinical Microbiology, Viborg Hospital, Heibergs Alle 4, Viborg, Denmark

Bank S, Nielsen HM, Hoyer Mathiasen B, Christiansen Leth D, Hagelskjr Kristensen L, Prag J.Fusobacterium necrophorum detection and identification on a selective agar. APMIS 2010; 118: 9949.

Within the last decade, Fusobacterium necrophorum subsp. funduliforme has been considered a clinicallyimportant pathogen causing pharyngitis especially in adolescents and young adults. F. necrophorumpharyngitis can progress into Lemierres syndrome, which is a severe and life-threatening infection.However, throat swabs are not cultured anaerobically in the routine and even if cultured anaerobically,it can be difficult to identify F. necrophorum from the normal flora of the throat. F. necrophorum istherefore often overlooked as the cause of pharyngitis. In our laboratory, a F. necrophorum selectiveagar has been developed containing vancomycin and nalidixin, which inhibit the growth of mostGram-positive and many Gram-negative bacteria, respectively. b-haemolysis of horse blood can bedetected, which further facilitates the detection and identification of F. necrophorum. The F. necropho-rum selective agar was evaluated against a quantitative real-time polymerase chain reaction assay andshown to have a significantly higher sensitivity for detecting F. necrophorum than the anaerobic agarcommonly used in Denmark. Furthermore, the F. necrophorum selective agar does not require experi-enced laboratory technicians, require fewer subcultures, is probably less expensive and is faster toperform than other culture methods.

Key words: Fusobacterium necrophorum; culture media; pharyngitis; tonsillitis.

Steffen Bank, Department of Clinical Microbiology, Viborg Hospital, Heibergs Alle 4, DK-8800Viborg, Denmark. e-mail: stb@mb.au.dk

Fusobacterium necrophorum has been recentlyreported to be a cause of pharyngitis especiallyin adolescents and young adults, where it mayaccount for more than 20% of acute pharyngitis(13). F. necrophorum has also been consideredresponsible for up to 20% of recurrent, persis-tent or chronic sore throats (2). Besides,F. necrophorum also causes otitis media in chil-dren, peritonsillar abscesses in adolescents andyoung adults and sinusitis in adults aged 3050 years (4). Furthermore, F. necrophorumpharyngitis can, although rarely, develop intoLemierres syndrome, a severe and life-threaten-ing infection, which has been estimated to have

a higher incidence, mortality and morbiditythan rheumatic fever in adolescents and youngadults in the Western world (5).F. necrophorum is a slow growing, kanamycin-

and metronidazole-sensitive Gram-negativeobligate anaerobe pleomorphic rod (6).F. necrophorum is divided into subsp. necropho-rum and subsp. funduliforme, the former mostlyfound as pathogen in animals and the lattercausing infections in humans (6). F. necropho-rum is probably a commensal in the humanpharynx (3). Isolation and identification of obli-gate anaerobic bacteria such as F. necrophorumfrom the normal flora of the throat can bedifficult and time-consuming (7).A selective anaerobe agar for the isolation

of Gram-negative anaerobes exists, however,Received 4 June 2010. Accepted 14 August 2010

APMIS 118: 994999 2010 The AuthorsAPMIS 2010 APMIS

DOI 10.1111/j.1600-0463.2010.02683.x

994

although the agar has been used to isolateF. necrophorum, studies indicate that prolongedincubation period is required (2, 8, 9). A selec-tive medium could increase the detection rateand reduce the time required to identifyF. necrophorum.

MATERIALS AND METHODS

Sample collection

All throat swabs sent to our laboratory during Octo-ber and November 2009 were included. The labora-tory services 150 general practitioners with acatchment population of 230 000 persons.

Handling of the throat swabs

The throat swabs were transported in Stuarts Trans-port Medium (Statens Serum Institute or SSI, Copen-hagen, Denmark) and inoculated anaerobicallywithin 24 h on three different anaerobic agar mediafrom SSI, described in Table 1. By turning the swabs120, one-third of the bacteria from the surface of theswab were inoculated on the F. necrophorum selectiveagar, another third on the anaerobic vancomycin andnalidixin (AVN) agar and the last third on the SSIanaerobe agar. The three anaerobic agar mediawere inoculated by triangular streaking with a1000 lg kanamycin disc (Oxoid Ltd., Basingstoke,

Hampshire, England) on the first streak and a 5 lgmetronidazole disc (Oxoid) placed between the firstand second streaks as shown in Fig. 1.The swabs were then inoculated aerobic on a 5%

Colombia sheep blood agar (Becton Dickinson, Hei-delberg, Germany) and examined for b-haemolyticstreptococci and other aerobic pathogens. Finally, theswab was transferred to 1 mL saline and stored for24 h before DNA extraction.

Incubation of the agar plates

The three anaerobic agar media were incubated in5% H2, 10% CO2 and 85% NO2 v v for 2 days at35 C before the first examination and for another2 days before final examination. Subcultures ofF. necrophorum on anaerobe agar (SSI) or 5% horseblood agar (SSI) were incubated anaerobically at35 C for 12 days, unless otherwise noted.

Handling of suspected F. necrophorum colonies

From all the three agar media, suspected F. necropho-rum colonies were Gram-stained using BiomerieuxPrevi Colour Gram (BioMerieux, Marcy lEtoile,France) according to the manufacturers instructions.The bacteria were then subcultured anaerobically onan anaerobe agar (SSI) with kanamycin and metroni-dazole discs (Oxoid) and aerobically on a secondanaerobe agar (SSI) where no growth was expected.Antibiotic susceptibility was performed on a thirdanaerobe agar (SSI). In addition, from the SSI anaer-obe agar and the AVN agar, the suspected colonieswere subcultured on a 5% horse blood agar (SSI) toexamine for b-haemolysis.

Detection and differentiation of F. necrophorum

F. necrophorum was detected based on the colonymorphology as shown in Fig. 1A and b-haemolysis asdemonstrated in Fig. 1B. The smell of butyric acidand fluorescence of a chartreuse colour in ultravioletlight Fig. 1C were also suggestive of F. necrophorum(6). Furthermore, the identification of F. necrophorumrequired detection of Gram-negative pleomorphicrods as shown in Fig. 2 and susceptibility to kanamy-cin (1000 lg; Oxoid), metronidazole (5 lg; Oxoid),colistin (10 lg; Oxoid) and resistance to vancomycin(5 lg; Oxoid) as described by Tage Justesen (http://tagejustesen.dk/anaerobebakterier/). F. necrophorumwas differentiated from other Fusobacterium spp. byb-haemolysis on horse blood agar and the typical col-ony morphology, and F. necrophorum subsp. funduli-forme was differentiated from F. necrophorum subsp.necrophorum by colony morphology and characteris-tic pleomorphic morphology observed under themicroscope (6).

Table 1. Ingredients in the three anaerobic agarmedia

Water1 900 mLPeptone 13.4 gNaCl2 4.5 gCitric acid 0.14 gK2HPO4 1.6 gStarch 1.8 gYeast extract 0.9 gAgar 6.8 gLiver autolysate 16 mLDefibrinated horse blood2 62 mLCysteine HCl 0.45 gVitamin K 0.0009 gNalidixin3 5.0 mgVancomycin3 2.5 mg

SSI, Statens Serum Institute; AVN, anaerobic vanco-mycin and nalidixin.1The pH was adjusted to 7.4 0.1.2The SSI anaerobe agar and AVN agar were heat-treated, which made them unusable for detection ofb-haemolysis.3Nalidixin and vancomycin were added to the AVNagar and Fusobacterium necrophorum selective agar.

A SELECTIVE AGAR FOR F. NECROPHORUM

2010 The Authors APMIS 2010 APMIS 995

Antibiotic susceptibility

All the F. necrophorum strains were tested on SSIanaerobic agar for susceptibility to penicillin (5 lg;NeoSensitabs, Rosco Diagnostica A S, Taastrup,Denmark) and metronidazole (5 lg, disc; Oxoid).

Real-time PCR and DNA extraction

Before DNA extraction the throat swabs were vor-texed for a few seconds and 700 lL of the saline was

purified on Arrow (NorDiag, Oslo, Norway) accord-ing to the manufacturers instructions, eluted in50 lL elution buffer, and stored at 4 C until use.Real-time polymerase chain reaction (PCR) was

performed as described by Jensen et al. (3). A 10-foldserial dilution of known concentrations (104107

CFU mL) of F. necrophorum subsp. funduliformeATCC 51357 was used to quantify the concentrationon the swabs.

Clinical information

Written clinical data were provided by most of thepractitioners or the hospital and registered in theMADS database (http://www.madsonline.dk), a sys-tem for standardized handling of samples and admin-istration data from the laboratory.

Statistical analysis

The chi-squared test was used to analyse difference inthe sensitivity of F. necrophorum on the agar media.Statistical analysis was performed using SPSS forWindows version 16.0 (SPSS Inc., Chicago, IL,USA).

RESULTS

Culture and real-time PCR results

A total of 139 throat swabs were screened forF. necrophorum. The swabs were from patientsaged 057 years, with a median of 20 years.Ninety-four per cent of the patients were

A B C

Fig. 1. Fusobacterium necrophorum subsp. funduliforme: colony morphology, b-haemolysis and fluorescence.Colony morphology, b-haemolysis and fluorescence from a pure culture of F. necrophorum subsp. funduliforme asseen on the F. necrophorum selective agar. (A) The colonies appear circular, slightly elevated with a smooth edgeand a white-buff, sometimes light yellow, often with a waxy appearance. The colonies can sometimes slide on theagar when touched. (B) A clear zone around the F. necrophorum colonies is seen as a result of b-haemolysis of thehorse blood in the agar. (C) When the colonies are exposed to ultraviolet light at 365 nm (Woods light), a char-treuse fluorescence colour can be seen.

Fig. 2. Fusobacterium necrophorum subsp. funduli-forme: Gram staining. Gram staining of a clinicalsample containing F. necrophorum subsp. funduli-forme. F. necrophorum is Gram negative with longand short rods sometimes with curling and tanglingrods, as shown.

BANK et al.

996 2010 The Authors APMIS 2010 APMIS

between 10 and 40 years of age. Clinical infor-mation was extracted from the MADS databaseand indicated that at least 81 of the patients hada sore throat. Sparse or no information was pro-vided for the remaining 58 patients and only afew swabs were taken for control examination.F. necrophorum was identified from 45 swabs

by real-time PCR and by culture from 32, 29and 11 of the swabs on the F. necrophorum selec-tive agar, AVN agar and SSI anaerobe agar,respectively. Data are demonstrated in Table 2.Ninety-one per cent (29 32) of F. necrophorum,identified on the F. necrophorum selective agar,was observed as characteristic curling and tan-gling rods on microscopy as shown in Fig. 2.F. necrophorum was identified by real-time

PCR and on the F. necrophorum selective agarfrom 25 (31%) and 18 (22%) of the 81 patientswith sore throat, respectively. A concentrationof 106108 CFU swab was obtained in 16(20%) of the 81 swabs.The F. necrophorum selective agar was very

reliable for detecting F. necrophorum at mediumand high concentrations, but the sensitivity stea-dily declined at lower concentrations as illus-trated in Table 2. The SSI anaerobe agarcommonly used in Denmark had a significantlylower sensitivity than the F. necrophorum selec-tive agar (p = 0.0008). F. necrophorum wasidentified almost three times as often on the F.necrophorum selective agar, 75% (24 32) ofthem at high concentrations, compared with theSSI anaerobe agar as seen in Table 2.Confluent growth in the first streak was

observed on the F. necrophorum selective agarwhen the concentration was above 9.1 105CFU swab, 10 or less colonies were observedon the first streak when the concentration was1058.6 105 CFU swab and 5 or less colonieswere observed when the concentration was lessthan 105 CFU swab. When there was confluentgrowth in the first streak, there was also alwaysgrowth in the second and third streaks.From one swab, F. necrophorum was detected

on the F. necrophorum selective agar and AVNagar, but was negative by real-time PCR.

Antibiotic susceptibility

All the F. necrophorum strains were susceptibleto penicillin and metronidazole. Susceptibilitywas usually read within 24 h. T

able

2.Sensitivityandspecificity

forthedetectionofFusobacterium

necrophorum

onthethreeanaerobicagarmedia

F.necrophorum

selectiveagar

AVN

agar

SSIanaerobeagar

CFUth

roatsw

ab