Anaerobic microflora under Class I and Class IIcomposite and amalgam restorationsChristian Spiieth, PD Dr Med Dent HabilVOIaf Bernhardt, Dr Metí DentVAnnegret Heinrich, Dr Med Dent̂ /Hannelore Bernhardt, Prof Dr Rer Nat HabiP/Georg Meyer, Prol Dr Med Dent Hábil''

Objectives: The microflora around and beneath restorations is an important factor of restoration failure.The aim of this pilot study was to determine and compare the microbiai spectrum under composite andamaigam restorations with speciai attention to the anaerobic flora. Method and materials: Ten compositeand five amalgam restorations scheduled for replacement were clinically evaluated tor marginal gaps, frac-tures, and secondary caries. After their remcvai and carles diagnosis, a dentin sampie just below therestoration was taken under steriie conditions, stored in a prereduced transport medium for anaerobic bac-teria, and immediateiy transferred to a fabcratcry for microbiai diagnosis Results; The ciinical parametersshowing mostly moderate marginal imperfections and the ratios of aercbic tc anaerobic flora were compara-ble for composite and amalgam restorations (11,4%:88,6% and 15,4%:a4,5%, respectively). The microbiaivariety under composite restorations was much greater compared to amalgam, and it was similar to that ofinfected root canals including anaerobic gram-negative rods, such as Fusobacterium species or Porphy-romonas species Beneath amalgam, the microbiai flora was similar to the one found in carious dentin andplaque, with anaerobic and facultatively anaerobic gram-positive rods dominating. Quantitatively, there wereup to eight times more microorganisms under composite restorations. The number of bacterial strains corre-lated with the caries activity and the fiiiing materiai, the number of anaerobic rods correlated highly withcaries activity and localization. In a linear regression, caries activity and the tilling material had statisticailysignificant influence on the bacterial load. Conclusion: Although caries activity and location had the great-est influence on the micrcbial flora under the restorations, the kind of restoration material seemed tc havean additional effect on the composition ct the microtlora. This pilot study indicates that inadequate compos-ite restorations may promote the growth of cariogenic, as weil as obligate anaerobic and potentially pulpo-pathogenic bacteria, which shculd be confirmed by further studies, (Quintessence !n¡ 2003.34:497-503)

Key words: amalgam, anaerobic, composite, microfiora, restoration

CLINICAL RELEVANCE: The bacterial flora growing ad-jacenf to a restoration is of special interest for itslong-term survivai, especially if it is influenced by therestorative material itself.

'Assistant Professor, Department cf Restorative Dentistry, Periodontoiogyand Pediatrro Dentistry, University of Greifswaid, Greifswald, Germany.

'Clinical Instructor, Department oí Restcrative Denlistry Periodontciogy,and Pédiatrie Dentistry University of Greifswald, Greifswald, Germany,

^Professor, Department of Ciinicai Microbiology in the Clinic of internaifuledicine A, University of Greifswaid, Greifswald, Germany.

"Prcfessor, Department cf Restcrative Dentistry, Periodcntology, andPédiatrie Dentistry, University of Greifswald, Greifswaid, Germany.

Reprint requests: Dr Christian Spiietfi, Sciiool of Dentistry, Universityof Greifswaid, Rotgerberstrasse 3, 17467 Greifswald, Germany E-mail:splieth@uni-g reifswai d, de

During the past two decades, much research hasbeen done to determine bacterial growth on cari-

ous and restored teeth,'-^ The bacterial flora growingadjacent fo a restoration is one of the most relevant fac-tors for its long-term survivai, as bacterial plaque candemineralize the tooth structure around the restorationand result in secondary caries. Secondary caries is themost common cause of restoration failure,-"-' and it ischaracterized by Mjör artd Toffenetti* as "primary carieslocated at the margin of a restoration," Adhesiverestorations such as composites offer the possibility ofavoiding a marginal gap and thus, avoiding the problemof recurrent caries or activation of remaining cariesthrough the marginal gap. Many studies, eg, Gaengler etal,̂ mostly performed under ideal conditions and bymotivated researchers, have shown that compositerestorations can reach high-quality levels and achievelong-lasting clinical success. On the other hand, epi-demiologic data indicate that at least in some countries,the routine use of composite materials does not reach

Quintessence Internationai 497

•Spliethetal

Fig 1 Examined composite restcnalion witn mar-ginai imperfections and iow clinical caries activity.

sufficient longevity. In Germany, 53O/o of the examinedcomposite restorations needed replacement, and inSweden, about one quarter of 2- to 4-year-old anteriorcomposites were clinically unacceptable."*" In insuffi-cient composite restorations, the microflora in the mar-ginal gap and beneath the restoration seems to be amajor problem, especially as the diagnosis of secondarycaries is difficult.'' It is very lüíely that material proper-ties of dental restorations influence plaque accumula-tion and the development of secondary or remainingcaries.'- Although a few studies could not find signifi-cant differences in the microbial flora around differentdental material, other studies detected specific differ-ences, proved an antibacterial effect of metal ions, orfound a correlation between the roughness of dentalmaterials and bacterial accumulation.''-''Svanberg etai"" detected much higher total colony-forming unit(CFU) counts for mutans streptococci at margins ofcomposite restorations than of comparable amalgamrestorations. Regarding microbial flora, it is likely thatbacteria present in dental plaque which are involved inthe etiology of primary caries also play a major role inthe formation of secondary earies." This was confirmedby Gonzalez-Cabezas et al'^ who found Streptococcusmutans, Actinomyces naeslundii, and Lactobacilluscasei widely present in secondary carious lesionsaround amalgam restorations. Adhesive compositerestorative materials, which have been increasingly usedfor premolars and molars since the 19y0s, offer the op-portunity of creating a bacteria-tight seal of the restora-tion and eliminating a marginal gap between the tooth

and tbe restoration. On the other hand, polymerizationshrinkage and the load of chewing pressure often resultin cracking and microleakage of the composites.'' Thismarginal gap could be an écologie niche for microor-ganisms,̂ " especially as composites do not have the an-tibacterial effects of, eg, Hg ions of amalgam.̂ ' Matasa"even described a "microbial attack" on composites usedas bonding adhesives for orthodontic appfications.

Qvist,-^ who examined the correlation betweenmarginal adaptation of resin composite restorationsand bacterial growth in cavities, concluded that bacte-ria in a cavity can be used as an indicator of marginalleakage along the restoration, in amalgam restora-tions, Kidd et aP-* found that marginal gaps over 0.4mm were colonized by bacteria to a significantlygreater extent than were smaller gaps. Studies inwhich a technique was used that allowed for furtherclassification of the collected samples showed that thebacterial spectrum around composite restorations wassimilar to that of dental plaque with predominantlyStreptococci and Actinomyces species."'^^ As moststudies concentrate on bacterial growth around themarginal gap on the axial wall of restorations, very lit-tle is known about the bacterial flora under restora-tions. Thus, the aim of this study was to determine andcompare the mierobial spectrum under composite andamalgam restorations, paying special attention to theanaerobic flora, as anaerobic bacteria dominate indental plaque and carious lesions.^' As the cultivationof anaerobic bacteria from the oral cavity is very la-borintensive, this pilot study was performed to evalu-ate whether there are indications to justify the expen-diture of greater resources for further research.

METHOD AND MATERIALS

The sample of this pilot study comprised 10 compositeand 5 amalgam restorations taken from 15 patientsv^hich needed removal due to clinical parameters(tractures, discoloration) or the patient's wish Therestorations had been placed in different private prac-tices. None of the patients suffered from toothachesand all teeth were vital, apart from one composite-re-stored tooth that had already been endodonticallytreated The patients were asked for the approximateage of the restoration and whether rubber dam wasused dunng application.

Ciinicai procedure

The restorations (Fig 1) were evaluated for mareinaladaptation and bulk fractures by one experienced detist using a dentai probe and magnifying lenses !2xoriginal magnification) according to modified criteria of

498 Voiume 34, Number 7, 3QQ3

• Spiieth el ai

TABLE 1 Results of clinical examination for the distribution of marginaladaptation, bulk fractures, and secondary caries in composite andamalgam restorations (percentages and absolute numbers)

Restoration

CompositePetcentagesAbsoiute no.

AmalgamPetcentagesAbsolute no.

Matginai gaps

Nogap

0%0

0%0

Inenamei

70%7

80%4

indentin

30%3

20%1

Ftactures

Yes

60%6

80%4

No

40%4

20%1

Caries iocalization

Marginal

70%7

80%4

Pulpalwali

30%3

20%1

Caries activity

Low

80%8

80%4

Higii

20%2

20%1

the California Dental Association.'^"'^s After placingrubber dam, the restorative material was removed firstusing rotating diamond instruments, then using sterilehand instruments. The presence of caries was examinedvistially and tactilely with a sterile explorer. If carieswas present, tbe location was recorded as marginal orpulpal, and caries activity as low (probe leaves a scratcbmark in dentin like in wax) or bigb (soft dentin). Thesubsequent treatment also was registered. Finally, asample of dentin from tbe pulpai wall of tbe cavity justbelow the restoration was taken under sterile condi-tions with a second band instrument, stored in 1 mLprereduced transport medium for anaerobes (as modi-fied by Cary Blair's^'), and immediately transferred totbe laboratory for cultivation in an anaerobic chamber.Sampling and transport took approximately 5 minutes.

Microbiologie procedures

Tbe samples were dissoluted for 10 seconds witb tbehelp of a vortex mixer. Fifty microliters of the sampleswere plated on Wilkens Chalgren blood agar plates(Difco, Voigt Global) and a selective WilkensChalgren blood agar plate (Oxoid) for gram-negativeand nonspore- forming rods. In tbe case of highly cari-ous dentin, tbe sample was diluted 10 times beforeplating. Tbe cultures were incubated in a glove box(Forma Scientific) at 37*'C for at least 4 days. Colonieswere selected and counted under a stereomicroscopeat 12x original magnificatiion.

The quantities of different colony types were esti-mated. Preliminary identification for levei I and II wasdone by Gram's staining method, checks for aerotoler-ance, sensitivity to antibiotics, fermentation of carbo-hydrates, and production of indole, nitrate, andothers,'^ For a further classification of the anaerobicgram-negative bacteria, the computer-guidedANI-Identification Card System (Bio Merieux Vitek)was used. Most of the selected anaerobic bacteriacould be identified down to tbe species level. It waspossible to identify the organisms only to tbe genus

level in a few cases of low anaerobic bacterial growtband for the aerobic bacteria.

Statistical analysis

The number of bacterial strains was counted and tbebacterial load (total number of bacteria) was calcu-lated for eacb restoration. Differences between bacter-ial colonization under composite- and amalgam-re-stored teetb were analyzed using tbe Mann-Whitney Utest (SPSS 10.0). In addition, the statistical signifi-cance of correlations between restorative materials,different subgroups of bacteria, and caries activity, aswell as location was examined with tbe Spearman testFinally, collinearity between the variables was testedusing multivariate analysis (linear regression).

RESULTS

Most patients stated that tbeir restorations were olderthan 2 years {n = 13). Only two composite restora-tions were placed 1 to 2 years ago according to the pa-fients. Ali patients denied tbe use of rubber dam forposterior restorations.

Clinical observations

In the clinical examination, none of tbe restorationshad perfect marginal adaptation and most restorationshad bulk fractures (Table 1). Tbe distribufion of mar-ginal adaptation and fractures witbin tbe subgroups ofcomposite and amalgam restorations were comparable.Although most of the removed restorations shewedsome degree of caries, it was mostly limited to tbe mar-ginai walls and was of low activity within the sub-groups of composite and amalgam restorations (seeTable 1). In most teetb, restorations could be placedwithout prior endodontic treatment. In one tooth only,which had a composite restoration, a pulp treatment(direct capping with calcium hydroxide) was necessary.

Quintessenoe international 499

Spiiethelai

TABLE 2 Distribution of bacteriai species in coiony-forming units (xio ') |oi composite-restored teeth

Bacterial strains '

Anaerobio cocciS saccharoiyticusS inlermediusS consteliatusS sangiusPeptostreptococcus speciesVeilloneila species

Anaerobic rodsL acidophiiusL case/L catenaformeL ¡enseniiA israeliiA naesiunöiiA odontoiyticusA viscosus 0.4Bifiáobacterium speciesEubaclerien speciesE aerofaciensBacteroides/Prevotelia speciesP biviaPbuccaeP corporisP disiensP intermediaP ioesci^eii/üenticolaP melaninogenioaP oraiisPorisF necrophorumF nucleatumCapnocytophaga species

Aerobic coccianhemoiytic streptococciStaphyiococcus speciesE faeciumNeisseria species 0,2

Aerobio rodsCoryneforme rodsS mailophiia

Total no. of coiony-forming units 0.6

Total no, ot strains 2

'Caries located at the pulpat wait rather tha

2t

20

10

40

20

40

2

3

1

2

2

2

0.2

2

2

0,2

4

4

1

10

0,2

0,2

2

10

2

2

132,0134 0 29,4

6 9 8

n the marginal gap.tlndicates high caries activity ratlier tlian low cari s activity.

P

5

4

tient

6 ' t

2

4200

5

0,2

8

10

27,25

206,03

T

6

2

8

20

36,04

8

10

330

4

4

41

10

0.2

5

3

19 10-

2 3

20

10

20 0,2

10

6

2

0,04

82.2 50,0 23.212 6 4

Microbial observations

Altogether, a total of 83 different bacterial species wasisolated, of which 70 were found in composite-restoredteeth and 13 in amalgam-restored teeth. The ratios ofaerohic to anaerobic flora were comparable; under com-

posite 11,40/0:88,6%; and under amalgam 15,4%:84,5o/o,Anaerohic species dotninated. The microbial varietyunder composite restorations was much greater com-pared to amalgam: 34 strains of strictly anaerobic non-spore-forming gram-negative rods, 17 strains of anaero-bic or facultatively anaerobic nonspore-forming

500Volume 34, Number 7 2003

• SpNeth et al

TABLE 3 Distribution of bacterial species in coiony-forming units (xio^)of amalgam-reslored teeth

Bacterial strains

Anaerobic cocciS intermediusPeplostreptocoocus species

Anaerobic rodsL caseiL ¡enseniiActinomyces speciesA odontoiyticusBifidobacterium speciesEubacterien speciesBacteroides/Prevoteila species

Aerobic cocciStapfiylococcus species

Totai no. ol colony-forming unrtsTotal no. of strains

'Canes located a\ the pulpai waii rathe•"Indícales iiigh caries activity rattier ttia

11

20

6

1

27.03

12

2

20

0.22

1

25.5

5

ttian ttie marginal gapn low ca es activity.

Patient

13 14

0 6

0.6 0.02 0

15

2

Q.2

2.2

2

gram-positive rods, nine strains of anaerobic gram-posi-tive cocci, and two strains of anaerobic gram-negativecocci (Tabie 2). Beneatii amalgam, one strain of strictlyanaerobic nonspore-forming gram-negative rods, sevenstrains of anaerobic or facultatively anaerobicnonspore-forming gram-positive rods, and tbree strainsof anaerobic gram-positive cocci were found (Tabie 3).More species and bigber quantities of lactobacilli wereisolated from composite restorations.

Quantitatively, tbere were up to eigbt times moremicroorganisms under composite restorations (seeTables 2 and 3), Differences between bacterial colo-nization under composite and amaigam were statisti-cally significant for anaerobic rods (P < .05) but notfor aerobic rods or anaerobic and aerobic cocci, dueto tbeir lower numbers.

Correlations between parameters

The number of bacterial strains correlated moderatelywith the restorative materiai and tbe caries activity(r = 0.6; P = .02), wbile the correlation between thebacterial load and the restorative material was on tbeverge of statistical significance (r = 0.51; P = .053). Astbe anaerobic rods comprised tbe largest proportion ofthe bacterial flora, tbeir bigh correlations witb cariesactivity and localization were statistically bigblysignificant (r = 0.8, r = 0.7, respectively; P < .01), tbecorrelation witb the kind of restorative material wasjust moderate {r = 0,6, P = -02}. In tbe subgroup ofamalgam restorations, caries activity and localization

correlated highly witb the number of bacteriai strainsand anaerobic rods [r = 0.9, P < .01). wbile in compos-ite restorations, caries localization was independent (r< 0.44, P > .2) of all bacterial parameters, and oniycaries activity and bacteriai load/anaerobic rods corre-lated higbly (both r = 0.8, P ^ .005). In tbe linear re-gression, wbieb yielded a statisticaiiy significant modeiin spite of tbe low numbers (P < .001), caries activityand tbe restorative material proved to be tbe most im-portant factors for tbe bacterial load (P = ,0001 and P= .0499, respectively), wbile the caries localization andmarginal integrity were excluded in tbe modei [P > .3).No collinearity was detected in tbis model.

DISCUSSION

Tbe teetb selected for the present study were placed indifferent private practices, several years ago. As thedistribution of marginal gaps and fractures as well astbe degree and iocation of caries within the subgroupsof composite and amaigam restorations were similar,the teetb selected for eacb group were comparable,Tbe majority of tbe samples that were taken from thepulpal walls of tbe cavities were clinically not carious,as secondary caries was mostly limited to tbe marginalwalls and was of low activity. Still, tbe ratios of anaer-obic and aerobic bacteria in the present study weresimilar to tbe ratios found in dental plaque or cariousdentin,^' Tbe activity of caries determines tbe amountand variety of the microbial flora in all cavities, as

Quintessence I nier national 501

• Splietti et al

documented In a study exatnining the microbiai com-position in carious lesions,̂ ^ This finding indicates thatthe clinical caries diagnosis after removal of therestoration was a valid parameter in Ihe present study.Bacterial colonization under amalgam was similar tothe flora of carious denfin and cariotts plaque,^"" withanaerobic and facultatively anaerobic gram-positiverods dominating. This distribution was also widelypresent in secondary carious lesions around amalgamrestorations,•* In contrast to this, a considerableamount of Bacteroides and Prevotella species was de-tected under many composite restorations. This spec-trum was similar to the microflora of infected rootcanals with potentially pulpopathogenic microbes, '̂-'"'Hoshino et aP^ found these obligate anaerobes to bethe predominant invading flora in nonexposed dentalpulps of 19 freshly extracted teeth with deep cariouslesions.

In contrast to the present resuits, Mejare et al-^-^detected a bacterial spectrum under compositerestorations similar to dental plaque with Streptococciand Actinomyces species dominating. This bacterialcomposition could be due to the short persistence ofthe restorations in the oral cavity in their study.Microorganisms which penetrate the marginal gaporiginated from the dental plaque, and require a cer-tain length of time to change their composition. Therestorations evaluated in the present study were ex-posed to the oral cavity for several years. Thus, ageand the high degree of imperfection of the compositerestorations detected by the clinical examinationcould have led to the differences in bacterial spectrumthe current authors found compared to that of dentalplaque. In addition, the composition ofthe compositemateriai itself might have contributed fo the shift froma dental plaquelike flora to an anaerobic tnicrobialspectntm, because amalgam restorations of compara-ble age and quaiity showed a quite different spectrum.

Furthermore, the localization of caries was impor-tant for the amount and variety of bacteria detected inthe dentin satnple under the amalgam restoration inthe present study. Marginal caries of lower activity didnot result in high bacterial counts at the pulpal wall.In contrast to this, the localization of caries and themicrobiai results were independent in teeth with eom-posite restorations. Penetration of bacteria along thecavity walls Is likely. The moderate correlafion be-tween the restorative material and the number of bac-teria indicates an influence of the restorative materialon the bacterial growth, independent of the cariousstatus of the cavity. Composite restorations with lowcaries activity showed a greater microbiai spectrumand more obligate anaerobes than the correspondingamalgam restorations. This might be due to the anti-microbial effect of metal ions in amalgam,̂ " the high

degree of imperfection of the restorations detected tnthe clinical examination, and/or the lack of rubberdam use during restoration placement, as differentprocedures infiuence the marginal adaptation and thebacterial growth,̂ '''3t it is likely that better marginaladaptation and bonding or tbe use of cavity liners willresult in less bacteriai penetration and growth aroundcomposite restorations.̂ '̂̂ '̂̂ ^

CONCLUSION

Although caries activity and location had the greatestinfluence on the microbiai flora under the restora-tions, the kind of restorative material used seemed tohave an additional effect on the composition of themicroflora. This pilot study indicates that inadequatecomposite restorations may promote the growth ofcariogenic, as well as obligate anaerobic and poten-tially pulpopathogenic bacteria, which should be con-firmed by further studies.

REFERENCES

1, Bergenholtz G, CON CF, Loesche WJ, Syed SA. Bacterialleakage around dental restorations: Its efïect on the dentalpulp. J Oral Pathol 1982;ll:439-450,

2, Brännström M, Nyborg H, The presence of bacteria in cavi-ties filled with silicate cement and composite resin materi-als. Swed Dent J 1971;64:149-155,

3, Lundin SA, Noren )G, Warfvinge J. Marginal bacterial leak-age and pulp reactions in Class II composite resin restora-tions in vivo, Scand Dent) 1990;14:185-192.

4, Mjor IA, Frequency of seeondary caries at various anatomi-cal locations. Oper Dent 1985;10(3):88-92,

5, Friedl KH, Hiller KA, Schmalz G. Placement and replace-ment of composite restorations in Germany, Oper Dent

6, Klausner LH, Green TG, Charbeneau GT Placement andreplacement of amalgam restorations: A challenge for theprofession, Opcr Dent 1987;12:105-112.

7, Mjör IA, Moorhead ¡E, Dahl JE. Reasons for replacement ofrestorations in permanent teeth in general dental practiceInt Dent | 2000;50;361-366.

8, Mjor lA, Toffenetti F, Secondary caries: A literature reviewwith ease reports. Quintessence Int 2000;31 165-179

9, Gaengler P, Hoyer I, Montag R, Clinical evaluation of pos-terior composite restorations: The 10-year report I AdhesDent2001;3:185-194, repon, j AQnes

10. Pieper I^ Mausberg R, Curdt R, Uhde V Clinical evaluationof the quahty of amalgam and po!yn,er filling materials aftervarious periods of function, A study of servicemen of theGerman armed forces. Dtsch Zahnarztl Z 1989 44 707 710

11. Allander L, Birkhed D, Bratthall D, Quality evaluation ofanterior restorations in private practice SWPH n . !1989;13(41:141-150, ' ^^'^^ '

502 Volume 34, Number 7, 2003

• Splieth et al

12. Svanberg M. Class It amalgam restorations, glass-ionomertunnel restorations, and caries development on adjacenttooth surfaces: A 3-year clinical study. Caries Res 1992;26:515-318.

13. Brauner A. Die bakterielle Besiedelung von Zahn ärztlich anWerkstoffen: Amalgame. Zahnarztl Prax 1988;5:67-170.

14. Hahn R, Weiger R, Netusehil L, Brück M. Microbial accu-mulation and vitality on different restorative materials. DentMater 1993,9:312-316.

15. Summanen P, Baron E|. Wadsworth Anaerobic BacteriolggyManual, ed 5. Belmont, CA: Star Publishing, 1993.

16. Svanberg M, Mjor IA, Orstaiik D. Mutans streptococci inplaque from margins of amalgam, composite, andglass-ionomer restorations. ] Dent Res 1990;69(3):S61-864.

17. Kidd EA, Joyston-Bcchal S, Beighton D. Microbiologicalvalidation of assessments ot caries activity during cavitypreparation. Caries Res 1993;27:402-408.

18. Gonzalez-Cabezas C, Li Y, Gregory RL, Stoolcey GK.Distribution of three cariogenic bacteria in secondary cari-ous lesions around amalgam restorations. Caries Res1999;33:357-365.

19. Choi KK, Condon JR, Ferracanc JL. The effects of adhesivethickness on polymerization contraetion stress of compos-ite. J Dent Res 2Q0O;79812-817.

20. Klimm W, Buchmann G, Reissig D. Schneider H.Vergleichende lichtmikroskopische sowie transmissions-und rasterelektronenmikroskopische Untersuchungen zurVerifikation von Mikroorganismen im Dentin gefüllterZähne. Stomatologie DDR 198S;38:702-707

21. Grossman ES, Matejka JM. Amalgam restoration and invitro caries formation. J Prosthet Dent 1995:73:199-209.

22. Matasa CG. Microbial attack of orthodontic adhesives. AmJ Orthod Dentofacial Orthop 1995;10S:132-141

23. Qvist V. Correlation between marginal adaptation of com-posite resin restorations and bacterial gronih in cavities.Scand J Dent Res 1980;88:296-300.

24. Kidd EA, Joyston-Bechal S, Beighton D. Marginai ditchingand staining as a predictor of secondary earies aroundamalgam restorations: A clinical and microbiological study.J Dent Res 1995;74:1206-1211.

25. Mejare B, Mejare I, Edwardsson S. Acid etching and com-posite resin restorations. A culturing and histologie study onbacterial penetration. Endod Dent Traumatol 1987;3:1-5.

26. Mejare B, Mejare 1, Edwardsson S. Bacteria beneath com-posite restorations. A culturing and histobacteriologicalstudy. Acta Odontoi Scand 1979;37:267-275.

27 Loesche WJ, Syed SA. The predominant cultivable flora ofcarious plaque and carious dentine. Caries Res 1973;7:201-216.

28. Ryge G. Clinical criteria. Int Dent J 1980;30:347-358.29. Björndai L, Larsen T. Changes in the cultivable flora in deep

carious lesions following a stepwise excavation procedure.Caries Res 2000;34:502-508.

30. Kidd EA, Joyston-Bechal S, Beighton D. The use of a cariesdetector dye during cavity preparation: A microbiologicalassessment. Br Dent J 1993;174:245-248.

31. Drucker DB, Lilley JD, Tucker D, Gibbs ACC. The en-dodontic microñora revisited. Microbios 1992;71:225-234.

32. Hoshino E, Ando N, Sato M, Kota K. Bacterial invasion ofnon-exposed dental pulp. Int Endod J 1992:25{suppi l):2-5.

33. Stiebe B, Poethe I, Bernhardt H. Die Baiíteriologie der nor-malen Wundheilung nach Zahnextraktion unter besondererBerücksichtigung der Anaerobierdiagnostik. Zahn-Mund-Kieferheilkunde 1990:78:247-251.

34. Qvist V, Qvist |. Marginal leakage along Concise in relationto filling procedure. Scand J Dent Res 1977;85:305-312.

35. Eriksen H. Pears G In vitro caries related to marginal leak-age around composite resin restorations. J Oral Rehabil1978:5:15-20.

Quintessence International 503