Effect of gingival fluid collection on subgingival plaque sampling

fa

J. Dent. 1994; 22: 223-228

Effect of gingival fluid collection subgingival plaque sampling B. Mullally,

L. Wolff*,

The Queen’s USA

University

N. Hardie*, of Belfast,

D. Aeppli*

on

and B. Pihlstrom*

UK and *Clinical

Periodontal

Research

Center,

University

of Minnesota,

ABSTRACT The purpose of this study was to determine whether gingival crevicular fluid (GCF) sampling by paper strip removes sufficient bacteria to affect subsequent subgingival plaque sampling using a curette. In 25 subjects, one healthy, gingivitis and periodontitis site was sampled for GCF using a strip followed by subgingival plaque sampling with a curette. Bacterial assays indicated that GCF strips removed significant numbers of bacteria when placed intracrevicularly for 5 s. A greater proportion of total bacteria was removed with strip sampling at healthy rather than gingivitis or periodontitis sites. Qualitative assessment of presence or absence of spirochaetes and dark-pigmented species indicated potential for significant interference of curette sampling by the strip at gingivitis and healthy sites. We concluded that paper strip GCF sampling may significantly affect curette sampling at the same sites. The magnitude of this impact depended on the clinical classification of specific sites and the assay performed. KEY WORDS:

Periodontal diseases,

J. Dent. 1994; December

22:

223-228

Dental plaque, Sampling,

(Received

16 August

1993;

Gingival crevicular fluid reviewed

1 November

Correspondence University

1993;

accepted

9

1993)

should be addressed to: Mr B. Mullally, Department of Restorative of Belfast, School of Clinical Dentistry, Grosvenor Road, Belfast 8T12 68P.

INTRODUCTION Clinical investigations of risk assessment in periodontology may require subgingival sampling of both gingival crevicular fluid (GCF) and plaque’. Moreover, clinical tests for diagnostic evaluation of GCF and subgingival plaque are likely to become more common in clinical practice*. If these diagnostic procedures are applied to the same sites during a single patient visit, there is potential for the first procedure to interfere with the second. Indeed, it has been reported that the plaque sampling technique contributes more to variance in bacterial recovery than laboratory processing3. It is obvious that subgingival plaque sampling using a curette would have a major effect on subsequent GCF sampling at the same site since curette plaque sampling often causes bleeding. However, the effect of paper strip GCF sampling on subsequent plaque sampling with a curette is less obvious. The purpose of this study was to determine whether paper strip sampling of GCF removes a sufficient quantity ofbacteria to affect subsequent subgingival plaque sampling using a curette. It is important to document the magnitude of this o 1994 Butterworth-Heinemann 0300-57 12/94/040223-06

Ltd.

Dentistry,

Queens

effect to allow accurate interpretation diagnostic tests obtained subsequent

MATERIALS Clinical

of bacteriological to GCF sampling.

AND METHODS

methods

Twenty-five adults (16 males and nine females; mean age f s.d. = 47.1 f 13.1 years) with periodontitis were recruited for this study. All gave written informed consent and none had a dental prophylaxis, antibiotic or non-steroidal antiinflammatory medication in the 6 months prior to examination. Individuals with systemic conditions requiring antibiotic prophylaxis, haematological disorders or diabetes were not included. Three sites from each participant were screened for investigation based on visual inspection, previously existing clinical records and radiographs. In each subject, one healthy, gingivitis and periodontitis site was selected. Clinical screening criteria for health included: (1) Gingival Index (GI)4 = 0; (2) probing depth < 3 mm without

224

J. Dent. 1994;

22: No. 4

bleeding on probing according to existing periodontal charting; and (3) absence of crestal bone loss as determined from radiographic records. Criteria for gingivitis included: (1) GI > 1; (2) probing depth < 3 mm with prior evidence of bleeding on probing: and (3) absence of crestal bone loss. Criteria for periodontitis included probing depth > 5 mm and the presence of crestal bone loss. Following selection of the three sites for each subject. a Plaque Index5 and GI were scored at each site. After gentle air drying and isolation with cotton rolls, a paper strip (Periopaper, IDE Interstate, Amityville, NY. USA) was inserted into the gingival crevice until resistance was felt and left in place for 5 s6. A subgingival plaque sample was then obtained using a Gracey curette machined to a blade width of0.5 mm to facilitate access to the subgingival area. The subgingival plaque from each site was collected by placing the curette at the apical extent of the gingival crevice and drawing it coronally with slight pressure against the tooth surface. A minimum of three passes was made for each site and in all cases sampling was continued until the investigator felt confident that all subgingival plaque had been removed. The paper strip GCF and curette plaque samples from each site were placed in separate vials containing 250 yl reduced transport fluid’ and were immediately transferred to an anaerobic chamber (Coy Mfg. Co., Ann Arbor. MI, USA) for laboratory processing. To avoid disturbing the subgingival area prior to sampling, no probing measurements were made until after plaque sampling had been completed. Following plaque sampling. calculus was scored using the calculus component of the Oral Hygiene Indexx. Clinical measurements of probing depth and attachment level were obtained to confirm site classifications into the clinical diagnostic categories of health, gingivitis and periodontitis. All clinical measurements were made by a single calibrated investigator (B.M.). Pocket depth was measured from the gingival margin to the base of the pocket in whole millimetres using a North Carolina periodontal probe (Hu-Friedy Mfg. Co., Chicago, IL. USA) and clinical attachment level was measured from the cement-enamel junction with the same probe using the method first described by Ramfjord9.

Frequency distributions of spirochaete pigmented CFUs/ml tabulated within category were also used for qualitative bacterial recovery from paper strip and

Laboratory

Results

All laboratory

methods

procedureslO were carried out under sterile conditions in an anaerobic chamber. Following transfer to the anaerobic chamber, samples were dispersed by passing the sample S-10 times through a 25 gauge tuberculin syringe. A 50 ul aliquot was then removed for darkfield microscopy. The remaining solution was made up to 2 ml and sonicated for 20 s using a Kontes Cell Disrupter (Kontes. Vineland, NJ, USA). Serial dilutions were plated onto supplemented blood agar media using a Spiral Plater (Spiral Systems Inc., Cincinnati, OH, USA). Agar plates were incubated for 10 days in the anaerobic chamber at 35°C. Total colony forming units/ml (CFUs/

ml) and dark pigmented bacteria (CFUs/ml) were counted and no attempt was made to speciate bacteria in the plaque samples. Darktield microscopic counts were obtained using a Petroff-Hausser chamber at a magnification of 10001o. Total bacteria which included spirochaetes and spirochaetes alone were enumerated and, where possible, approximately 100 organisms were counted for each sample. In some samples from healthy sites, fewer than 100 organisms were available for counting.

Statistical

procedures

Sites not confirmed by clinical examination after plaque sampling as fulfilling criteria for the three diagnostic categories were excluded from all analyses. For descriptive purposes. mean probing depth. attachment level and median GI, PI and calculus scores were calculated across subjects for each diagnostic category (healthy, gingivitis. and periodontitis). Because of the non-normal distribution of microbiological data, non-parametric methods were used for all statistical analyses. Medians and interquartile ranks were calculated for each bacterial assay and disease classification. A Wilcoxon signed rank test was used to test for statistically significant differences between paper strip and curette samples within each disease category. Percentage calculations of bacteria removed by the paper strip were based on the assumption that all bacteria had been removed by the combined strip and curette sampling methods (% bacteria removed by strip = 100 X strip bacterial count/(strip plus curette bacterial count)). In order to protect against spurious statistical significance due to multiple comparisons, a Friedman two-way ANOVA was performed to determine if there were any differences between disease categories with respect to proportions of bacteria removed by the paper strip for each bacterial assay. For assays having significant findings, a Wilcoxon signed rank test was then used for pairwise comparisons of disease categories.

counts and dark each diagnostic comparisons of curette samples.

Clinical Mean + s.d. (mm) depth (PD) and attachment level (AL) for each diagnostic category were as follows: (1) health: PD = 1.9 f 0.7, AL = 0.8 + 1.5: (2) gingivitis: PD = 3.0 + 0, AL = 1.1 i- 1.8; (3) periodontitis: PD = 6.8 rt 1.6. AL = 6.1 + 2.5. Median PI and GI scores for healthy sites were both 0, while the median PI was 0.5 for gingivitis and 1.0 for periodontitis sites. The median GI was 1.0 for gingivitis and periodontitis sites. The median calculus score was 0 for healthy and gingivitis sites and 2.0 for sites with periodontitis.

Mullally

et al.: GCF sampling

effect on subgingival

plaque

225

Table 1. Median counts and number of subjects (n) for each diagnostic category and bacterial assessment DPB CFWml

(X 104)

Spirochaere darkfield

Disease classificarion

Strip

Curette

n

Healthy Gingivitis Periodontitis

0 0 20.0

0 0.1 30.0

24 24 25

0 0* 4.0*

0 2.0 25.0

21 23 23

0.02

73

0*

1.6

67

All sites

0

*Stattstically significant

differences,

coun;u@JeW Strip

P < 0.005. DPB = dark-pigmented

n

Total bacteria CFWml (X 104) Strip Curette n 0.15 50:::: 7.0*

0.15

Total darkfield microscopic coun;ugt:,04) Strip

n

2:00::00

24 24 25

40.0 35.0* lOO.O*

35.0 100.0 350.0

21 23 23

1000.0

73

44.0*

100.0

67

bacteria.

Table II. Median percentage of total counts removed by paper strip sampling method (100 X no. strip/ (no. strip + no. curette)) Disease classification

______

Healthy

DPB CFU/m/

Spirochaete darkfield counts

92%

19%

(5)

(2

Total bacteria CFU/m/

Total darkfield microscopic counrs

n Gingivitis (o%;io%) n Periodontitis n

27% (1 l%, 58%) 23

(o%“;o%) i0 15% (10%. 32%) 22

(12%. 35%) 25

(14%.

34%) 23

Values for first and third quartiles are in parentheses. Brackets indicate statistically significant difference at P < 0.05. n = Number of subjects, DPB = dark-pigmented bacteria.

Bacteriological Median bacterial counts per sample for each clinical diagnostic category and sampling method are given in Table 1. There were no statistically significant differences between sampling methods at healthy sites for any method of bacterial assessment and there were no statistically significant differences in dark-pigmented bacteria between strip and curette samples for any disease classification. For both gingivitis and periodontitis sites, curette samples had statistically significantly (P < 0.005) more bacteria than the strip samples for total CFUs/ml and darkfield counts of spirochaetes and total bacteria. The median percentages of total counts and values for interquartile bacterial percentages of bacteria removed by the paper strip sampling method are given in Table ZZ. There were no statistically significant differences in the median proportion of dark-pigmented CFUs/ml or spirochaete counts removed by the paper strip between any disease category. However, for both total CFUs/ml and total darkfield microscopic counts, statistically significantly (P < 0.05) greater proportions of bacteria were removed by the paper strip from healthy sites compared to sites having either gingivitis or periodontitis. The interquartile ranks in Table ZZ illustrate that a significant proportion of bacteria may be removed by the strips within any given disease classification. For example, the total darkfield microscopic count interquartiles for healthy sites were 46% and 76%. This means that in the third quartile or in 25% of the healthy sites, the proportion

of total darkfield microscopic counts removed by the paper strip ivas 76% or more. For this same bacterial assay (first quartile), 75% of the subjects had at least 46% of the microorganisms removed by the paper strip. A 2 X 2 frequency distribution of detection or nondetection of spirochaetes and dark-pigmented bacteria by each sampling method is given in Table III. Substantial disagreements in bacterial detection occurred between the two sampling methods which were dependent on disease classification. For the majority of healthy sites, spirochaetes were not detectable (17 of 21) using either sampling method. However, of two sites positive with the strip, only one was positive with the subsequent curette sample. At healthy sites, overall agreement for the presence and absence of spirochaetes between the two methods was 86% (18/21). For sites with gingivitis, there was agreement between sampling methods for detection (six sites) and non-detection (three sites) of spirochaetes in only nine of 23 sites (39%). Of the nine strip sites positive for spirochaetes at sites with gingivitis, only six were positive for the curette sample indicating the strip may have interfered with plaque sampling in about 30% of the curette samples. At sites with periodontitis, there was good agreement between curette and paper strip sampling for spirochaete detection (19 sites) and non-detection (one site) in 20 of 23 sites (87% agreement). For dark-pigmented bacteria (Table III), overall agreement for the two sampling methods (21 of 24) at healthy sites was 88%. Six healthy sites were positive for

226

J. Dent. 1994; 22: No. 4

Table 111.Subject frequency distributions of qualitative assessment of detection (Yes) and nondetection (No) of bacteria for each sampling method within each clinical diagnostic category Healthy

Gingivitis

Periodontitis

Curette

Curette

Curette

Spirochaete darkfield counts Strip

DPB CFU/ml Strip

DPB = dark-pigmented

No

18

0

9

Yes

3

3

2

bacteria.

the strip method while only three of these were positive with subsequent curette sampling. Moreover, there were no sites that were positive for curette and negative for the strip. Agreement was somewhat less (15 of 24; 63%) for sites with gingivitis. Dark-pigmented bacteria were found in eight of the sites with gingivitis sampled by the paper strip and 13 sampled by curette. Overall agreement between sampling methods was 76% for sites with periodontitis (19 of 25).

DISCUSSION In addition to clinical measurements, attempts to identify specific biochemical and microbiological markers are frequently included in clinical studies. This may involve collection of GCF and plaque’. Although the effect of plaque sampling on the subgingival microbiota has been addressed in other studies”, little attention has been paid to the potential interference of GCF collection using paper strips on subsequent subgingival plaque sampling from the same site. It has been estimated that the number of organisms in a periodontal pocket available for sampling can vary from lo2 to IO9 depending on the status of the periodontal tissues Il. In the present study, it was postulated that insertion ofthe paper strip would remove a substantial number of these microorganisms from the pocket. By comparing the numbers and proportions of microorganisms removed by the paper strip and the curette, an estimate of the effect of one sampling method on the other may be obtained. For periodontitis and gingivitis sites. statistically significantly greater microscopic counts of total bacteria. spirochaetes and total CFUs/ml were found in curette samples compared to paper strip samples. For healthy sites and for dark-pigmented CFUs/ml, there was no significant difference between sampling methods. This indicates that GCF sampling with a paper strip may have more effect on recovery of bacteria from subsequent curette samples at healthy sites than at sites with either gingivitis or periodontitis. This would not be surprising considering that more severely diseased sites contain a higher number of microorganisms. The expectation that

the paper strip method would have more effect at healthy sites with fewer organisms than at diseased sites with greater numbers of bacteria was confirmed by the finding that the paper strip sampling method resulted in a statistically greater percentage of the total CFUs/ml and darkfield microscopic counts from healthy as compared to gingivitis and periodontitis sites. Although there is no previous information in the literature regarding the removal of bacteria from periodontal pockets by paper strips. an analogous situation is the use of paper points for subgingival plaque sampling. Different methods of sampling subgingival plaque have been reviewed and it was concluded that scaler samples remove a higher proportion of microorganisms from pockets than any other method in use”. Kiel and Lang’? reported that a scaler removed 61-91% of the pocket microflora compared to 7-48% removed by paper points. However. Moore et al.13 reported similar total bacterial counts for single paper point samples and subsequent curette samples obtained at the same sites. They also reported no differences between sampling methods for the proportion of sites or subjects positive for the presence of A. actinomycefemcomitans. The present study indicated no significant difference in the numbers of dark-pigmented CFUs/ml between the two sampling methods in sites with periodontitis (Table I). This indicates that the paper strip may remove a large proportion of dark-pigmenting bacteria in such pockets prior to curette sampling. This is confirmed by data in Table ZZ. indicating that in 50% of the samples from sites with periodontitis. at least 27% of the dark-pigmented CFUs/ml were removed by the paper strip. Paper points have been shown to selectively remove the loosely adherent microorganisms. Kiel and Lang’* reported a higher percentage of black-pigmented Bacteroides species isolated by paper points compared to scaler sampling. This may be because gram-negative organisms. which increase with increasing severity of periodontal disease’“. 14,are frequently found in the loosely adherent subgingival plaqueL5. Renvert et al.16 found that significantly higher proportions of CFUs/ml and spirochaetes were removed by sampling with three paper points for 15 s as compared to a

Mullally

et al.: GCF sampling

single Morse scaler sample. If medians of the multiple pass curette sampling data from Table I for periodontitis sites are compared to means of baseline data reported by Renvert et al. l6 for single pass scaler samples after paper point sampling at similar sites, more total CFUs/ml were recovered in the present study (200.0 X 105) compared to (14.0 and 6.0 X 105) as reported by Renvert’s groupi6. The numbers of spirochaetes removed from similar pockets were greater with multiple passes using a curette in the present study (250.0 X 103) compared to a single pass using a Morse scaler (46.0 and 113.0 X 103) in the study by Renvert et al.‘“. However, similar numbers of total CFUs/ ml (50.0 X 105) were obtained in the present study using a 5 s insertion of a paper strip as was reported by Renvert er a1.i6 (75.0 and 84.0 X 105) who sampled similar sites using three paper points inserted into pockets for 15 s. Therefore, while comparisons between studies may be hazardous, it appears that multiple passes using a Gracey curette can remove substantially more subgingival bacteria and spirochaetes than a single pass with a Morse scaler. It also appears that sampling for 5 s with a paper strip may result in similar recovery of bacteria as sampling with three paper points for 15 s. If the object of a diagnostic test is to simply determine the presence or absence of spirochaetes or dark-pigmented bacteria, it is unlikely that GCF sampling prior to curette plaque sampling will have a major effect at periodontitis sites. Overall agreement between the two sampling methods varied between 76% and 87% at sites with periodontitis depending on the bacterial assay. On the other hand, qualitative assessment of both these bacterial types may be severely compromised at sites with gingivitis because of the relatively low overall agreement between the methods for spirochaetes (38%) and dark-pigmented bacteria (63%). This would be expected because relatively large numbers of both these organisms were present in periodontitis and relatively fewer numbers were present in health. In periodontitis, sufficient numbers existed so that prior paper strip sampling did not appreciably interfere with detection by subsequent curette sampling. For healthy sites. only a few organisms of these types were present and overall agreement between methods would be expected. However, data in Table III also indicates that there is potential for interference of strip sampling on subsequent curette sampling in the small proportion of healthy sites where these organisms were present. The results of this study demonstrated that GCF strips removed a significant number of bacteria when placed to the base of the gingival crevice for 5 s. Furthermore. a greater proportion of total bacteria assessed either by cultural or microscopic methods were removed with paper strip sampling at healthy sites than at sites having gingivitis or periodontitis. For qualitative assessment of presence or absence, there was potential for significant interference of GCF sampling with subsequent curette plaque sampling at gingivitis and healthy sites for spirochaetes and dark-pigmented bacteria.

effect on subgingival

plaque

227

Some investigators have described a method of GCF sampling using periopaper placed at the orifice of the crevice’. This would be expected to result in less disturbance of the subgingival microflora. However, this technique may not be very effective in removing a sufficient volume of GCF for assay. Paper strips are frequently used to remove GCF for the purpose of sampling biochemical markers of disease activity. This study has established that subgingival paper strip GCF sampling may have a significant impact on subsequent curette sampling at the same sites. The magnitude of this impact was dependent on the clinical diagnostic classitication of specific sites and the bacterial assay which was performed.

Acknowledgements Brian Mullally is grateful to the Royal College of Physicians and Surgeons (Glasgow) for their support in awarding him the T.C. White Travelling Fellowship to facilitate his travel to the University of Minnesota where his work was carried out. In addition we are also grateful to Mr Gerard Linden who was instrumental in setting up this visit. This study was supported by USA NIDR/NIH PSO-DE 08489 and the Erwin M. Schaffer Periodontal Research Chair.

References I. Wolff L, Smith Q, Snyder W et al. Relationship between lactate dehydrogenase and myeloperoxidase levels in human gingival crevicular fluid and clinical and microbial measurements. J Clin Periodontal 1988; 5: 110-l 15. 2. Fine D. Incorporating new technologies in periodontal diagnosis into training programs and patient care: a critical assessment and a plan for the future. J Periodon tol 1992; 63: 383-393. 3. Wikstriim M, Renvert S, Dahlen G eta]. Variance in recovery of periodontitis-associated bacteria caused by sampling technique and laboratory processing. Oral Microbiol h~munol 1991; 6: 102-106. 4. Ltie H and Silness J. Periodontal disease in pregnancy. 1. Prevalence and severity. Acta Odontol Stand 1963: 21: 533-551. 5. Silness J and Lde H. Periodontal disease in pregnancy 11. Correlation between oral hygiene and periodontai condition. Acta Odontol Stand 1964; 22: 121-135. 6. Smith QT. Au GS, Freese PM et al. Five parameters of gingival crevicular fluid from eight surfaces in periodontal health and disease. J Periodont Res 1992; 27: 466-475. 7. Holdeman L, Cato E and Moore W. Anaerobic Laboratory Manual, 4th edn. Blacksburg: Virginia Polytechnic Institute and State University, 1977; pp 141-142. and uses. J 8. Greene J. The oral hygiene index-development Periodontol 1967; 38: 625-635. 9. Ramfjord SP. Indices for prevalence and incidence of periodontal disease. J Periodonrol 1959; 30: 51-59. 10. Wolff L, Liljemark W. Bloomquist C et al. The distribution of Actinobacillus actinomycetemcomitans in human plaque. J Periodont Res 1985; 20: 237-250.

228

J. Dent 1994; 22: No. 4

II. Tanner A and Goodson J. Sampling of microorganisms associated with periodontal disease. Oral Microbial Immlmol 1986: 1: 15-20. 12. Kiel R and Lang N. Effect of subgingival sampling techniques on periodontal microbiological culturingl Dent Res 1983: 62: 247 (Abstr). 13. Moore W. Holdeman L. Cato E et al. Comparative bacteriology of juvenile periodontitis. Infect Imrnmo/ 1985: 48: 507-519.

International

14. Slots J and Genco RJ. Black pigmented Bactcroidrs species. Capnocyrophaga species. and Acrir~ohacillus actinomycetemcomitans in human periodontal disease. J Dent Re,\ 19X4; 63: 412-421. 15. Listgarten M. Structure ofthe microbial flora associated with periodontal health and disease in man. A light and electron 1976: 47: l-18. microscopic study. J Pcriodonfol 16. Renvert S. Wikstrijm M. Helmersson M et al. Comparative study of subgingival microbiological sampling techniques. .\ Periodon to/ 1992: 63: 797-801.

Abstracts

Study on the method of three-dimensional cephalometrics and clinical application. Zhou Jie-ming et a/. Beijing J Stomato/ 1993;

1: 1-9.

Some craniomaxillary landmarks are selected individually on the lateral and posteroanterior craniomaxillary radiographs. With the assistance of a computer image processing system and ‘C’ language software, the threedimensional figure of craniomaxilla is rebuilt. Then some new measuring index of the three-dimensional craniomaxillary figure is inducted. This technique is accurate and reliable, with a systematic accuracy of 0.45 mm. This technique has provided a new system for craniomaxillary radiograph examination and improves the diagnosis and treatment of malocclusion. The technique has been used in craniomaxillary measuring of children with normal occlusion in Beijing. We obtained the normal measuring range of normal occlusion in three periods: deciduous teeth, mixed dentition, permanent teeth. (7 references) Liao Fang-Gang and Yu Liu-Ning

Improvement in surface grazing agent (Bell Feel Brightener) for resin composite restorations. Hosoda H, Yamada T, Onoe N, Morigami M. Jap J Conserv Dent 1993; 36: 1595-l 608. In order to improve colour characteristics of the conventional surface glazing agent Bell Feel Brightener for resin composite restorations, an experimental clear surface glazing agent Bell Feel Brightener II without camphorquinone has been newly developed. Several properties of the agent such as the width of the airinhibition layer, tooth brushing abrasion resistance, surface gloss, etc. were compared with the conventional agent in vitro. A clinical study on the agent was also performed. As a result, the new agent was superior to the conventional in vitro and in vivo. The new agent was useful as a varnish and glazing agent for glasspolyalkenoate cement restorations and as a glazing agent for resin composite inlay restorations. (4 references) H. Hisamitsu

Clinical evaluation of a bonding system All-bond 2 using BPDM. Masutani S, Hinoura K, Hosaka M, Hirai Y, Onose H, lshikawa T. Jap J Conserv Dent 1993; 36: 1233-l 249. A number of new generation dentine bonding agents are now available to the dental profession. Laboratory studies on the bonding strength of resin composites to dentine were undertaken to evaluate these products. However, the evaluation of the adhesive to dentine was made from the clinical point of view. Therefore, this study was designed to evaluate the clinical performance of a newly developed dental adhesive All-bond 2 system. The 6-month clinical investigation of this system in combination with Bis-fil M and P resin composite restorations was conducted on 75 patients who had signed informed consent forms at two different dental university hospitals. All the restorations were evaluated by using the modified USPHS criteria at baseline (1 -week) and I-, 3and 6-month recalls. Assessments were made of pulpal response, gingival condition, retention of restoration, colour match, surface staining, marginal discolouration, marginal adaptation, anatomical form (wear), surface texture, marginal fractures, restoration fractures, secondary caries, replacement and retreatment. One clinical case displayed discomfort due to cold water. However, the serious pulpal reaction and harmful effect on the gingival tissue were not evident. The optimum marginal integrity was maintained up to 6 months. As for bulk fractures, no case of secondary caries and no dislodgement of the restorations to treat were observed. Therefore, this bonding system was determined to be safe and useful for clinical application. (46 references) H. Hisamitsu