Antibiotic Resistance in Bacteria Isolated from Subgingival Plaque in a Norwegian Population with Refractory Marginal Periodontitis

Abstract

The prevalence of β-lactamase-producing bacteria in subgingival plaque from patients with refractory periodontitis in Norway was assessed by the chromogenic nitrocefin method. β-Lactamase activity was detected in 68% of the patients. Enzyme-producing strains belonged predominantly to the genus Prevotella; others were staphylococci, enteric gram-negative rods, and Bacillus spp.

Microorganisms of dental plaque may cause marginal periodontitis, a disease characterized by loss of attachment around the teeth, which is common in the Western world and even more so outside it (6). Most putative periodontal pathogens are gram-negative anaerobic rods, e.g., Porphyromonas gingivalis, Prevotella intermedia, Bacteroides forsythus, and Treponema denticola (24, 31). However, superinfecting organisms such as enteric gram-negative rods, pseudomonads, staphylococci, and yeasts have been isolated from refractory periodontitis lesions (22), as well as taxa not considered a common part of the oral microbiota (5). There are reports of increasing β-lactamase production in subgingival periodontal pathogens in several countries (7, 12, 17, 27). In Norway, however, this situation has not been studied. The aims of this study were (i) to investigate the extent of β-lactamase production in subgingival bacteria isolated from patients with refractory marginal periodontitis in Norway, (ii) to characterize the β-lactamase-producing bacteria using enzymatic and/or biochemical methods and partial sequencing of the 16S rRNA gene, and (iii) to determine their susceptibility to various antimicrobial agents.

Subgingival plaque samples from 25 consecutive cases of refractory marginal periodontitis received from specialists in periodontics were examined. The patients were aged 39 to 66, and none had responded to conventional periodontal treatment. One patient (no. 14 [Table 1]) had received a combination of an aminopenicillin and metronidazole as an adjunct to conventional treatment within 3 months prior to sampling. The other patients had not received antibiotics in the past 3 months. Bacterial samples were collected from the bottom of the periodontal pocket and were immediately transferred to prereduced anaerobically sterilized Dental Transport Medium (Anaerobic System, Morgan Hill, Calif.). Appropriately diluted samples were plated onto nonselective Trypticase soy agar plates supplemented with 5% defibrinated human blood, hemin (5 mg/ml), and menadione (0.05 mg/ml). In addition, selective agar plates (wolinella medium, CVE medium, clindamycin blood agar, TSBV agar, mitis salivarius agar, mannitol salt agar, MacConkey agar, Trypticase soy agar plates supplemented with 5% defibrinated human blood, hemin (5 mg/liter), and N-acetyl muramic acid (10 mg/liter) and blood agar plates supplemented with 3 μg of amoxicillin/ml were inoculated with 0.1 ml of undiluted sample for recovery of periodontopathogens. Nonselective and selective agar plates were incubated anaerobically (90% N₂, 5% H₂, 5% CO₂) at 37°C for up to 14 days in anaerobic jars (Anoxomat WS9000; Mart, Lichtenvoorde, The Netherlands). All morphotypes of bacteria on plates were subcultured and tested for β-lactamase production.

TABLE 1.

Prevalence and identity of β-lactamase-producing bacterial isolates

Patient no.	β-Lactamase status (+ or −)	Identity
		16S rRNA (% homology)	API (% reliability)
1	+	Prevotella melaninogenica (99)	P. melaninogenica (51)
2	+	Staphylococcus saprophyticus (99)	S. saprophyticus (100)
	+	Prevotella sp. oral strain B31FD (99)	Prevotella loescheii (10)
3	+	Serratia sp. (97)	Citrobacter freundii (unacceptable profile)
4	+	P. melaninogenica (99)	Prevotella oralis (50)
5	+	Prevotella buccae (100)	P. buccae (100)
	+	Prevotella veroralis (98)	Prevotella denticola (81)
	+	Erwinia sp. (99)	Pantoea sp. (56)
	+	Staphylococcus cohnii (99)	Staphylococcus xylosus (88)
6	+	P. melaninogenica (100)	P. melaninogenica (83)
	+	P. buccae (100)	P. buccae (100)
	+	Acinetobacter sp. (99)	Acinetobacter sp. (99)
7	+	C. freundii (99)	C. freundii (100)
8	+	P. denticola (99)	P. melaninogenica (82)
9	−
10	+	P. buccae (99)	P. buccae (100)
11	−
12	+	Bacillus licheniformis (100)	B. licheniformis (100)
13	+	B. licheniformis (99)	B. licheniformis (78)
	+	Escherichia coli (99)	E. coli (100)
14	−
15	+	S. cohnii (99)	S. xylosus (79)
16	+	Prevotella sp. oral strain B31FD (99)	P. loescheii (85)
17	−
18	+	P. intermedia (100)	P. intermedia (100)
19	−
20	+	Prevotella oris (98)	P. buccae (identification not valid)
21	+	Staphylococcus epidermidis (100)	S. epidermidis (97)
22	−
23	+	Unidentified oral bacterium (99)	P. melaninogenica (75)
24	−
25	−

β-Lactamase production was assessed by using chromogenic nitrocefin-impregnated disks (BBLDrySlideNitrocefin; Becton Dickinson) and a chromogenic nitrocefin solution (27). β-Lactamase-positive and -negative strains of Staphylococcus aureus, provided by the Microbiology Laboratory at the National Hospital, Oslo, Norway, were included as controls. Enzymatic and/or biochemical profiling relied on commercial diagnostic kits designed for identification of a number of different microorganisms (Analytab Products [API]; bioMérieux, Marcy-l'Etoile, France). The preparation, incubation, and reading of the kits were carried out according to the manufacturer's recommendations. To confirm species identification with the API system, partial sequencing of the 16S rRNA gene was also performed as previously described (26).

Bacterial susceptibility to ampicillin, amoxicillin, amoxicillin-clavulanic acid, tetracycline, minocycline, doxycycline, clindamycin, metronidazole, cefotaxime, and ceftazidime were assessed by using the Etest (AB Biodisk, Solna, Sweden). Etest extended-spectrum β-lactamases (strip of cefotaxime-cefotaxime plus clavulanic acid and strip of ceftazidime-ceftazidime plus clavulanic acid) were used to confirm the presence of clavulanic inhibitable extended-spectrum β-lactamase enzymes. For anaerobes, brucella agar plates were streaked with a McFarland standard no. 1 inoculum in brucella broth. For facultative organisms, PDM agar plates were streaked with a McFarland standard no. 0.5 inoculum suspended in 0.85% NaCl. The Etest MIC was read according to the recommendations of the manufacturer. Appropriate quality controls were included in each run.

At least one strain with β-lactamase activity was detected in 17 of the 25 (68%) patients (Table 1). No β-lactamase-producing bacteria were found in the subgingival plaque of the patient who had received antibiotic treatment within 3 months prior to sampling (patient no. 14 [Table 1]). The most prominent β-lactamase-producing species belonged to the genus Prevotella (52%). Other enzyme-producing strains were enteric gram-negative rods (five patients, 20%), staphylococci (four patients, 16%), and Bacillus species (two patients, 8%). Five of the β-lactamase-producing isolates could not be identified at the species level, two of them belonged to the genus Prevotella, and the remaining three were facultative gram-negative rods. The two Prevotella strains were closely related to strain B31FD in the GenBank database. They differed from one another by 8 nucleotides in the sequenced region and thus might represent new Prevotella species. All β-lactamase-producing Prevotella species were recovered from the blood agar plates supplemented with amoxicillin. Nonselective agar plates, however, were necessary to detect β-lactamase-producing bacteria other than Prevotella. The MICs of ampicillin and amoxicillin were in the range of 0.50 to 32 μg/ml and 1 to 64 μg/ml, respectively, for the Prevotella species. According to the interpretive categories of the NCCLS (18), all Prevotella isolates were susceptible to amoxicillin-clavulanate (MICs ranged from 0.023 to 0.75 μg/ml) (Table 2). The Prevotella species showed variable resistance to the tetracyclines. One of the isolates (Prevotella sp. oral strain B31FD, patient no. 2) was resistant to cefotaxime and ceftazidime (MICs = 256 μg/ml). All the Prevotella isolates were susceptible to metronidazole. Table 2 also lists the in vitro antimicrobial susceptibility of the staphylococci and the facultative gram-negative rods (19). The results of the screening for extended-spectrum β-lactamases were negative for all bacterial isolates.

TABLE 2.

Susceptibilities to 10 antimicrobial agents, as determined by the Etest

Bacterial strain no.	Identity (16S rRNA)	MIC (μg/ml) of antibioticsa
		Ampicillin	Amoxicillin	Amoxicillin- clavulanic acid	Tetracycline	Minocycline	Doxycycline	Clindamycin	Metronidazole	Cefotaxime	Ceftazidime
1c	Acinetobacter sp.	12/I	16/R	12/I	12/S	0.125/S	1.5/S	16/R	256/R	32/I	3/S
2c	B. licheniformis	256d	256d	24d	0.50d	0.38d	0.38d	256d	256d	32d	256d
3c	E. coli	256/R	256/R	8/S	4/S	4/S	6/I	256/R	256/R	0.19/S	0.19/S
4c	Erwinia sp.	3/S	16/R	16/I	4/S	2/S	1.5/S	16/R	256/R	0.064/S	0.25/S
5	Serratia sp.	6/S	8/S	6/S	3/S	1.5/S	2/S	16/R	256/R	0.38/S	0.125/S
6	C. freundii	8/S	64/R	12/I	8/I	16/R	16/R	256/R	256/R	0.38/S	2/S
7	B. licheniformis	256d	256d	16d	1.5d	3d	0.19d	256d	256d	32d	256d
8	S. saprophyticus	0.19/S	0.38/S	0.38/S	0.50/S	0.125/S	0.25/S	0.094/S	256/R	12/I	16/I
9c	S. cohnii	0.125/S	0.125/S	0.125/S	0.50/S	0.25/S	0.50/S	1.5/I	256/R	2/S	16/I
10	S. cohnii	0.125/S	0.125/S	0.125/S	0.125/S	0.064/S	0.125/S	0.125/R	256/R	3/S	16/I
11	S. epidermidis	0.19/S	0.19/S	0.19/S	0.75/S	0.19/S	1/S	0.125/S	256/R	0.38/S	3/S
12	P. melaninogenica	24/R	32/R	0.75/S	0.094/S	0.016/S	0.047/S	0.016/S	0.125/S	12/S	16d
13	Prevotella sp. strain B31FD	32/R	64/R	0.064/S	32/R	3/S	6/I	0.016/S	0.75/S	256/R	256d
14	P. melaninogenica	1/R	2/R	0.125/S	16/R	2/S	4/S	0.016/S	0.19/S	1/S	0.75d
15c	P. veroralis	16/R	32/R	0.19/S	6/I	1.5/S	3/S	0.016/S	0.125/S	24/I	32d
16c	P. buccae	8/R	32/R	0.125/S	6/I	1.5/S	2/S	0.016/S	0.19/S	4/S	16d
17c	P. melaninogenica	8/R	24/R	0.064/S	12/I	2/S	3/S	0.016/S	0.064/S	8/S	6d
18c	P. buccae	3/R	6/R	0.125/S	0.125/S	0.023/S	0.064/S	0.016/S	0.38/S	1.5/S	8d
19	P. denticola	0.75/R	1/R	0.023/S	0.064/S	0.016/S	0.047/S	0.016/S	0.064/S	1/S	0.75d
20	P. buccae	6/R	8/R	0.094/S	8/I	0.5/S	1.5/S	0.016/S	0.016/S	2/S	8d
21	P. intermedia	4/R	8/R	0.125/S	4/S	2/S	1.5/S	0.016/S	0.38/S	3/S	0.75d
22	P. oris	0.50/R	1/R	0.032/S	0.094/S	0.023/S	0.064/S	0.016/S	0.016/S	0.25/S	0.75d
23	Unidentified oral bacterium	NDb
24	Prevotella sp. strain B31FD	NDb

^aR, resistant; I, intermediate; S, susceptible.

^bND, no data provided; fastidious bacteria.

^cOne patient had strains 4, 9, 15, and 16; a second patient had strains 1, 17, and 18; a third patient had strains 2 and 3.

^dNo available interpretive categories.

A high proportion (68%) of the patients with refractory periodontitis in Norway harbored β-lactamase-producing bacteria in their subgingival plaque, as also found in other countries (16, 28), although higher (12, 27) or lower (7, 14, 25) prevalences have been reported. In the present study API identification correlated with that based on partial sequencing of the 16S rRNA gene in 13 of the 24 bacterial isolates. Especially for the Prevotella species, the correlation was limited. Partial sequencing of the 16S rRNA gene resulted in 97 to 100% homology with partial 16S rRNA gene sequences in the GenBank database. All β-lactamase-producing Prevotella species isolated were resistant to ampicillin and amoxicillin. Penicillin-resistant Prevotella species have been reported in several studies (2, 4, 10, 13, 14, 15, 17). These isolates were all susceptible to amoxicillin-clavulanate. Good activity of amoxicillin-clavulanate against Prevotella species has also been reported (3, 4, 7, 8, 16). In accordance with previous findings (1, 9, 10, 11, 30), 100% susceptibility to metronidazole and to clindamycin was observed in all Prevotella isolates. Two of the Prevotella isolates in the present study were resistant to tetracycline, and four showed intermediate resistance. Because of the frequent use of tetracyclines in periodontal practice (21, 23), patients with refractory periodontitis often present with a history of tetracycline therapy and a microflora that is relatively resistant to this antibiotic (20, 29). According to the latest published NCCLS breakpoints (18), all but one of the Prevotella species isolated in our study were sensitive to the cephalosporin cefotaxime. The MICs were somewhat higher for ceftazidime, but no breakpoints for that cephalosporin have been published for anaerobes. For one Prevotella isolate for which no species identification could be made (B31FD), the MICs of cefotaxime (256 μg/ml) and ceftazidime (256 μg/ml) were high, thus indicating the presence of a different β-lactamase. Most recent investigations assessing β-lactamase-producing species in the oral cavity have focused on anaerobic gram-negative rods. Since the majority of periodontal diseases are mixed infections, the present study attempted to identify the range of oral species that can produce the enzyme. Several of the enteric rods were resistant to penicillin and showed variable resistance to tetracyclines. Three of the isolates showed intermediate resistance towards amoxicillin-clavulanate. These findings underscore the importance of enteric rods in refractory periodontal diseases.

In conclusion, β-lactamase activity in subgingival bacteria from refractory periodontitis in Norway was a common feature. A wide variety of β-lactamase-producing bacteria, both putative pathogenic species and accompanying organisms, could be isolated from the subgingival flora of refractory periodontitis patients.