Introduction.
Bacteria that are found in an infected root canal with apical periodontitis form a selected group of the total oral microflora. When the pulp becomes necrotic there is an increase of strict anaerobic Gram-negative and Gram-positive species such as Fusobacterium spp., Prevotella spp., Porphyromonas spp., Peptostreptococcus spp.,Eubacterium spp., and Actinomyces spp., which occurs at the expense of facultative anaerobic species (Fabricius et al. 1982). Amongst the important ecological factors in this dynamic process are nutrition, pH, reduction-oxidation potential, temperature, host resistance and microbial interactions.
In the apical part of the root canal (where interaction with the periapical tissues takes place), proteolytic bacteria constitute the main proportion of the microflora (Slots & Taubman 1992). The close contact with vital periapical tissues may be presumed also to give a higher oxygen tension inhibiting their growth. However, the small proportion of facultative anaerobic species present in the apical region may consume the available oxygen resulting in a low reduction-oxidation potential favoring colonization of strict anaerobic species (Slots & Taubman1992).
The interactions between different species can be beneficial to one or more microorganisms, whilst with other microbes it can be antagonistic. Positive interactions, e.g. co adherence and nutrient supply have been described to occur in plaque (Hillman et al.1985, Grenier & Mayrand 1986, Marsh 1989), the periodontal pocket (van Winkelhoff et al. 1987, Dzink et al. 1988, Socransky et al. 1988, Ashimoto et al. 1996) and in the root canal (Fabricius et al.1982, Sundqvist1992, Gomes et al.1994).
Combinations of bacteria may be related to specific signs and symptoms of oral diseases. Socransky et al. (1988) related periodontal attachment loss and deepest pockets to specific clusters of bacteria. Gomes et al. (1996b) related pain of endodontic origin with combinations of Peptostreptococcus spp. and Prevotella spp., whilst Haapasalo (1989) found Porphyromonas endodontalis and P. gingivalis to be related to acute symptoms. Fabricius et al. (1982) showed that specific combinations of bacteria had the greatest potential to induce apical periodontitis.
The first step in the investigation of microbial interactions in the root canal of teeth with apical periodontitis is to study the prevalence and the simultaneous isolation of bacterial species. The aim of this study was to investigate positive and negative associations of bacteria found in root-canal infections of teeth with periapical bone destruction without clinical signs and symptoms.

Materials and methods.

Patient material and microbiological sampling.
The content of 58 root canals in 50 patients was sampled for microbial analysis. All selected teeth were asymptomatic, did not respond to sensitivity testing, had not received previous endodontic treatment and showed radiographic evidence of periapical bone loss. All patients were systemically healthy and had not used antibiotics 6 months prior to the study.
After cleaning of the tooth with pumice and isolation with rubber dam, the crown and surrounding rubber dam were disinfected with 80% alcohol for 2 min. An access cavity was made with sterile high-speed diamond burs under irrigation with sterile saline. Before entering the pulp chamber, the access cavity was disinfected again for 2 min with 80% alcohol. A swab sample was taken from the surface and streaked on blood agar plates to check disinfection. The pulp chamber was accessed with sterile burs and rinsed with reduced transport Јuid (RTF, Syed & Loesche 1972) and suctioned with sterile tips. RTF was then introduced to the root canal by a syringe with a sterile 27-gauge needle. Care was taken not to overfi ll the canal.With sterileHedstro« mfi les the canal was enlarged to a size 20 fi le to the estimated working length as calculated from the preoperative radiograph. Five sterile paper points were consecutively placed in the canal and left for10 s and thenplaced in sterile tubes containing 1mL RTF and transferred to the laboratory within15 min for microbiological processing.

Microbiological procedures.
Ten-fold serial dilutions of the samples were prepared and100 mL of each dilution was inoculated on blood agar plates supplemented with 5% horse blood, 5 mg L  hemin and 1mg L   menadione. Plates were anaerobically (80% N2, 10% H2,10% CO2) incubated in an anaerobic box at 378C for 7 days. After incubation, the total colony forming units (CFU) and the different colony morphotypes were counted with the use of a stereomicroscope at 16x magnification (Zeiss, Oberkochen, Germany).
All colony morphotypes were streaked to purity, incubated in air and 5% CO2 (BBL Gaspak CO2 systems, Becton Dickinson and Company, Cockeysville, MD, USA) anaerobically to determine strict anaerobic and facultative anaerobic growth. Identification was made on the basis of Gramstain, catalase activity and a commercially available identification kit ATB rapid ID32A (Biomerieux SA, Lyon, France), for strict anaerobes and ATB rapid ID32Strep for facultative anaerobic cocci (Biomerieux SA).
The blood agar plates with the total samples were kept incubated under the anaerobic conditions for up to 14 days to allows low growing species to develop. Newly emerging colonies were also streaked to purity and identified.

Statistics.
Differences in prevalence of different species in the 58 root canal samples were tested with Cochran test for nominal nonparametric data (present vs. not present).
The Fisher’s Exact test was applied to test the null hypothesis that there was no relationship between any of the species (P > 0.05).When there was a significant difference, indicating that a relationship between species was present, the Odds ratio was calculated. Positive associations were those with an Odds ratio >2 (the odds of detecting one organism was twice as likely with another organism present as with it absent) and negative associations were those with an Odds ratio <0.5.

All swab samples were negative, indicating a disinfected operation field.
All 58 samples contained cultivable microorganisms. The median number of CFU mLx1 was 8 x104 (range 50-3 x107 CFU mLx1) per sample. The number of species per canal varied between 4 and 7 (mean 5.1). Of the isolated species, 87% were strictly anaerobic. The most prevalent bacterial species were P. intermedia, P. micros and A. odontolyticus, present in, respectively, 33, 29 and 19% of the cultured canals (Table 1, Fig.1). P. intermedia and P. micros were present in significantly more samples than any of the other species (P < 0.05). A. odontolyticus was present in significantly more samples than most species except for F. nucleatum, P. intermedia, Capnocytophaga spp., Eubacterium lentum, Propionobacterium acnes and P. micros.
A total of 81 combinations of bacterial species were tested with the Fisher’s exact test. For four microbial combinations, the null hypothesis had to be rejected and therefore indicated a positive association between two species. Table 2 shows all four combinations that had an Odds ratio >2, indicating a positive relationship. Negative relationships were not found.

Table 1. Prevalence of bacterial species in 58 root canals.


* Prevotella intermedia and Peptostreptococcus micros were significantly more present than all other species (P < 0.05).
** Actinomyces odontolyticus was significantly more present than all except Fusobacterium nucleatum, Prevotella intermedia, Capnocytophaga spp., Eubacteriumlentum, Peptostreptococcusmicros and Propionibacteriumacnes (P< 0.05).

Figure 1. Percentages of isolates from 58 root canals.



Table 2.

OR, Odds ratio.
CI, confidence interval.

Discussion.
Anaerobic species were dominant in all root canals and constituted 87% of the isolates. Sundqvist (1992) found 90% and Gomes et al. (1994) 64% strict anaerobes. The most dominant species in our study were P. intermedia, P. micros and A. odontolyticus. These species are commonly found in root-canal infections in most studies where comparable patient material and microbial techniques have been used (Sundqvist 1992, Brauner & Conrads 1995, Gomes et al. 1996a). The Odds ratio calculation is a standard method to measure associations between microorganisms in clinical samples. However, the Odds ratio can be high, suggesting a relationship between species, whilst the sample size is very small. In these cases a wide 95% confidence interval can be present and, therefore, it is necessary to be very cautious about the interpretation of the finding. The Fisher’s exact test offers a calculation to find a relationship between two samples with very small expected frequencies. In this study higher Odds ratios were found for many combinations of bacteria. The Fisher’s exact test, however, contradicted the existence of such relationships inmost instances. When comparing the results of similar studies these factors have to be taken into account (Altman 1991).
Positive correlations between P. intermedia and P. micros and Eubacteriumwere previously found in periodontal pockets by Lewis et al. (1988) and Socransky et al. (1988) as well as in endodontic infections of the root canal (Sundqvist 1992, Gomes et al. 1994). Brook et al. (1991) found a consistent pattern of a combination of Bacteriodes spp. and Peptostreptococcus spp. in periapical abscesses. These findings are in accordance with our study where clear positive relations were found between P. intermedia and P. micros. P. intermedia and P. micros have been described as a pathogenic combination of bacterial species in experimental anaerobic infections in mice (Brook &Walker1983).
Gomes et al. (1994) and Drucker et al. (1992) found that P. oralis was associated with P. melaninogenica whilst in our study P. oralis was positively associated with P. intermedia. The confidence interval in our study, however, was very wide whilst Gomes et al. (1994) did not state the 95% confidence interval associated with the Odds ratio of12.Drucker et al. (1992) did not calculate the Odds ratio but showed a significant relationship between the species (P < 0.05).The positive relation between P.micros and A. odontolyticus was described previously by Sundqvist (1992), who found an Odds ratio of 3 between P. micros and Actinomyces spp. In our study a positive association was also found between Veillonella spp. And Bifidobacterium spp. Both isolates are not considered significant human pathogens (Slot & Taubman 1992). Veillonella parvula was described a shaving no or negative relations to other bacteria (Sundqvist1992) and is associated with less active periodontal sites (Dzink et al. 1988, Socransky et al. 1988, ). This discrepancy between results may be due to the small sample size and very wide confidence interval that was found in the present study.
Specific combinations of bacteria are found in the root canal and they may contribute to population shifts of the flora by different mechanisms of interaction. For example, one species or a combination of species might favour another by providing growth factors or by changing the physical-chemical environment. The newly developed combinations may contribute to the development of clinical signs and symptoms or to apical bone destruction.

References.

Altman DG (1991) Practical statistics for medical research.London, UK: Chapman & Hall Publishers.
Ashimoto A, Chen C, Bakker I, Slots J (1996) Polymerase chain reaction detection   of 8 putative periodontal pathogens in subgingival plaque of gingivitis and   advanced periodontitis lesions.Oral Microbiology and Immunology 11, 266-73.
Brauner AW, Conrads G (1995) Studies into the microbial spectrum of apical   periodontitis.International Endodontic Journal 28, 244-8.
Brook I, Frazier EH, Gher ME (1991) Aerobic and anaerobic microbiology of   periapical abscess.Oral Microbiology and Immunology 6, 123-5.
Brook I, Walker RI (1983) Infectivity of organisms recovered from polymicrobial   abscesses.Infection and Immunity 42, 986-9.
Drucker DB, Lilley JD, Tucker D, Gibbs AC (1992) The endodontic microflora   revisited.Microbios 71, 225-34.
Dzink JL, Socransky SS, Haffajee AD (1988) The predominant cultivable microbiota   of active and inactive lesions of destructive periodontal diseases.Journal   of Clinical Periodontology 15, 316-23.
Fabricius L, Dahle¤ n G, Holm SE, Moller AJ (1982) Influence of combinations   of oral bacteria on periapical tissues of monkeys.Scandinavian Journal   of Dental Research 90, 200-6.
Gomes BP, Drucker DB, Lilley JD (1994) Associations of specific bacteria   with some endodontic signs and symptoms.International Endodontic Journal   27, 291-8.
Gomes BP, Drucker DB, Lilley JD (1996a) Positive and negative associations   between bacterial species in dental root canals.Microbios 80, 231-43.
Gomes BP, Lilley JD, Drucker DB (1996b) Associations of endodontic symptoms   and signs with particular combinations of specific bacteria.International   Endodontic Journal 29, 69-75.
Grenier D, Mayrand D (1986) Nutritional relationships between oral bacteria.Infection and Immunity 53,616-20.
Haapasalo M (1989) Bacteriodes spp. in dental root canal infections.Endodontics and Dental Traumatology 5, 1-10.
Hillman JD, Socransky SS, Shivers M (1985) The relationships between streptococcal   species and periodontopathic bacteria in human dental plaque.Archives of   Oral Biology 30, 791-5.
Lewis MAO, MacFarlane TW, McGowan DA, MacDonald DG (1988) Assessment of the   pathogenicity of bacterial species isolated from acute dentoalveolar abscesses.Journal of Medical Microbiology 27, 109-16.
Marsh PD (1989) Host defenses and microbial homeostasis: Role of microbial   interactions.Journal of Dental Research 68(special issue), 1567-75.
Slots J, Taubman MA, eds (1992) Contemporary Oral Microbiology and Immunology.St. Louis,Missouri: MosbyYear Book Inc.
Socransky SS, Haffajee AD, Dzink JL, Hillman JD (1988) Associations between   microbial species in subgingival plaque samples.Oral Microbiology and Immunology   3, 1-7.
Sundqvist G (1992) Associations between microbial species in dental root   canal infections.Oral Microbiology and Immunology 7, 257-62.
Syed SA, LoescheWJ (1972) Survivalof dental plaque flora in various transport   media.Applied Microbiology 24,638-44.
Winkelhoff vanAJ, Kippuw N, DeGraaff J (1987) Cross-inhibition between black-pigmented Bacteroides species.Journal of Dental Research 66, 1663-7.