Detection of in vitro demineralization adjacent to restorations using quantitative light induced fluorescence (QLF)

Dental Materials 19 (2003) 368–374 www.elsevier.com/locate/dental

Detection of in vitro demineralization adjacent to restorations using quantitative light induced fluorescence (QLF) I.A. Prettya,*, P.W. Smithb, W.M. Edgara, S.M. Highama a

Department of Clinical Dental Sciences, The University of Liverpool, Edwards Building, Daulby Street, Liverpool L69 3GN, UK b Unit of Prosthodontics, Turner Dental School, The University of Manchester, Manchester M13 9PL, UK Received 13 June 2001; revised 4 March 2002; accepted 12 June 2002

Abstract Aim. Quantitative light-induced fluorescence (QLF) is a technique for the detection, quantification, and longitudinal monitoring of early carious lesions. The technique is non-destructive and can be used in vivo. Using the natural fluorescence of teeth, and the loss of such fluorescence in demineralized enamel, QLF is a repeatable and valid optical caries monitor. Previously used in smooth and occlusal surfaces, the purpose of this pilot study was to determine if QLF could detect, and longitudinally monitor, demineralization adjacent to a range of restorative materials. Method. Fifteen previously extracted lower third molars were selected based upon the lack of any visible demineralization. A single burr hole was placed on the buccal surface and the cavity restored with amalgam, composite, compomer, glass ionomer or a temporary filling material. The buccal surface was then coated in an acid resistant nail varnish leaving an exposed area around the restoration and also a similar sized control region. The teeth had QLF images taken at baseline and were then subjected to a demineralizing buffer, further QLF images were subsequently taken at 72 and 144 h. Transverse microradiography was used to confirm the presence of early, subsurface lesions at the completion of the cycle (144 h). QLF images were analyzed by a single blinded examiner and values for change in radiance fluorescence were computed. These values were recorded as loss of radiance fluorescence loss integrated over area of lesion and expressed as DQ. Results. The appearance of each material under QLF and the change in fluorescence is described. Amalgam, glass ionomer and the temporary material all exhibited reduced fluorescence, while composite and compomer showed increased fluorescence, when compared with surrounding enamel. There was no change in fluorescence of the materials when subjected to experimental demineralizing conditions. Readings at 72 and 144 h demonstrated demineralization adjacent to the restorations and at the exposed control. Significant differences were detected between baseline, 72 and 144 h using ANOVA on all restorations with the exception of compomer where significance was noted between baseline and 144 h, p . 0.05. Conclusion. This pilot study has demonstrated the ability for QLF to detect and monitor secondary caries. Analysis techniques should be based upon the subtraction of baseline DQ scores from subsequent images. Further research is required to assess the ability of QLF to detect secondary lesions in vivo. q 2003 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. Keywords: Secondary caries; Restoration; QLF; Enamel; Detection

1. Introduction 1.1. Quantitative light-induced fluorescence Cariologists and clinicians are currently concerned with the detection of early carious lesions of the kind that can be reversed following fluoride or similar interventions. The diagnosis of such lesions is a major research focus [1]. * Corresponding author. Tel.: þ44-151-706-5288; fax: þ 44-151-7065809. E-mail address: [email protected] (I.A. Pretty).

Diagnostic methods to detect early demineralization (decalcification) in vivo include visual inspection [2], tactile examination with dental probe [2], radiography, electronic monitoring (ECM) [3], fibre-optic transillumination (FOTI) [4], and laser fluorescence (DIAGNODent) [5]. All of these methods have limitations affecting either their diagnostic ability or their practicality in a clinical setting. Several of the methods (i.e. radiography) cannot detect the carious process until it is significantly advanced [6] while others produce arbitrary data that cannot be correlated to actual mineral loss, an important parameter in caries research. Methods have been developed that permit this quantifiable

0109-5641/03/$ 30.00 þ 0.00 q 2003 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved. PII: S 0 1 0 9 - 5 6 4 1 ( 0 2 ) 0 0 0 7 9 - 9

I.A. Pretty et al. / Dental Materials 19 (2003) 368–374

369

Fig. 1. A diagrammatic representation of the QLF system. Courtesy of Inspektor Research Systems BV, NL.

detection (i.e. transverse microradiography, TMR) of very early lesions but are destructive and cannot be used in vivo [7]. The development of quantitative light-induced fluorescence (QLF, Inspektor Research Systems BV, NL) presents a solution to the problem by combining the sensitivity and specificity of the destructive in vitro techniques with the clinical usefulness of those employed in vivo, [8]. Initially, the technique was developed using lasers and was demonstrated by Bjelkhagen et al. [9]. With concerns existing over the intra-oral use of lasers, de Josselin de Jong et al developed a system using filtered visible light, QLF [10]. The principle behind the technique is that enamel will auto-fluoresce under certain lighting conditions [11]. Demineralized enamel will fluoresce less and this loss of fluorescence can be detected, quantified, and longitudinally monitored using QLF [12]. The technique is non-destructive, easy to master and offers researchers a truly quantifiable, and archivable research tool. The contrast between sound and demineralized enamel is increased by a factor of 10 and the lack of specular reflections facilitates reliability and reproducibility. 1.2. The QLF method The teeth are illuminated from an arc lamp using a liquid light guide with a peak intensity of 370 nm. A yellow highpass filter (520 nm) is placed in front of a CCD microcamera which captures the tooth image. A live image of the tooth under examination is displayed on a PC screen, and when a quality image is obtained the examiner depresses a foot pedal to save the image to disk. The CCD camera and light

guide are formed into a hand held unit similar to a dental handpiece that also includes a mirror for capturing lingual surfaces, Fig. 1. The proprietary analysis software detects the darker areas of the image and simulates the fluorescence radiance of sound enamel at the lesion site via a reconstruction algorithm. This is performed by a twodimensional linear interpolation of sound enamel values adjacent to the lesion. The absolute decrease in fluorescence is determined by calculating the percentage loss between actual and reconstructed fluorescence and is expressed in the value DF. The program also calculates the area of the lesion, in mm2 and from this can calculate DQ. This value is defined as the fluorescence radiance loss integrated over the lesion area, and is the reportable value for in vivo studies. DQ is comparable to the total mineral loss from the lesions (DZ ) as measured by TMR, the current gold standard for in vitro studies [8]. The QLF device has been validated by comparison with TMR [13] and the analysis method has shown to be reliable and reproducible [14]. 1.3. Secondary caries The clinical diagnosis of secondary caries is by far the commonest reason given for the replacement of restorations [15 –17]. Secondary caries is thought to be a localized lesion occurring around restorations that is identical in aetiology and histology to primary caries [15]. There is agreement in the literature that secondary caries can be difficult to detect clinically [15,18,19] and measures that would allow early detection and institution of preventive strategies have great potential to reduce the need for replacement of restorations [20].

Manufacturer

DENTSPLY Caulk

DENTSPLY Caulk DENTSPLY Caulk

Colte`ne Whaledent 3M ESPE

0007000346 Expiry date 06/2002

001228 Expiry date 12/2005 0012000329 Expiry date 11/2003

KJ488 Expiry 2004 FW0069028 Expiry 12/2003

I.A. Pretty et al. / Dental Materials 19 (2003) 368–374

Batch number

370

QLF offers the potential for early diagnosis and quantification of mineral loss in primary carious lesions in vivo [12,13]. In addition, should it also demonstrate the ability to detect and quantify mineral loss adjacent to existing restorations, then this would be highly relevant to the clinician, who would not only have the ability to detect secondary caries early, but also monitor the efficacy of any preventive strategies that might be employed [21,22]. The purpose of this pilot study is twofold. Firstly to investigate the appearance of common restorative materials under QLF conditions and determine how this might affect the detection of lesions adjacent to them. Secondly to determine the ability of QLF to detect artificial secondary carious lesions adjacent to such restorative materials and the analysis techniques involved.

Temporary Glass ionomer

Urethane dimethacrylate resin, polymerizable trimethacrylate resin, strontium aluminium fluorosilicate glass, strontium flouride (2.5 mg/m3) Alloy of silver, tin, copper,and mercury Urethane modified bis-GMA dimethacrylate, titanium dioxide, silica fume (as dust) Eugenol-free temporary material Capsulated polymaleinate glass ionomer Compomer

Encaspulated spherical, amalgam Composite

Constituents Type

2.1. Tooth preparation Fifteen previously extracted lower third molars were selected for use in this study. Selection was based upon the absence of any clinically visible enamel demineralization. Each tooth was subsequently examined using QLF in order to detect any subclinical lesions or defects. The roots were removed by diamond disc leaving the anatomical crown. Each crown was gently pumiced (SS White, UK) and abraded with wet-and-dry paper. Each tooth was then randomly allocated to one of five restorative materials (three teeth per material) which were placed by a single operator. A single burr hole (Hi Di 644, medium grit) was drilled on the buccal surface and then restored using the appropriate material following the manufacturer’s instructions. The cavity extended through enamel and into dentin. Restorative materials used were compomer (Dyract, A3 Dentsply, Caulk), amalgam (Megalloy, Dentsply, Caulk), temporary material (Coltosol, Colte`ne, Whaledent), composite (Spectrum, A3 Dentsply, Caulk), and glass ionomer (Ketac Fil, ESPE), Fig. 2 (Table 1). The teeth were then coated in a transparent acid resistant nail varnish (MaxFactor, Procter and Gamble, UK) leaving two areas exposed: (a) the restoration and adjacent area, region x and (b) a control area of similar size, region y. Region z was coated, Fig. 2. The control region was required to ensure that demineralization of the tooth was possible and to control for restorative material effect on adjacent enamel, i.e. fluoride containing and the F2 history of the tooth. The teeth were mounted in to the black plastic caps of 50 ml bottles (Jensons, UK) using dental compound (Kent Dental, UK).

Coltosol Ketac Fil

Megalloy Spectrum

Dyract

2.2. Lesion production Material

Table 1 Details of the restorative materials employed

2. Materials and methods

Baseline QLF images and white-light digital photographs were taken of each of the teeth. The caps were then

I.A. Pretty et al. / Dental Materials 19 (2003) 368–374

371

72 h and left to air dry for 30 min. QLF images and whitelight high-resolution digital photographs (3.3 Megapixels, Sony, Japan) were taken. Images were taken using a method that standardized the position of the teeth in relation to the mounted QLF camera. This was achieved by the placement of a cross hair on the caps containing the samples that was subsequently aligned to a similar cross hair on the QLF platform [11]. The images were taken in Class 1 ASA darkroom conditions [11]. The teeth were then returned to the demineralizing solution for a further 72 h (144 h total) and the examinations repeated. QLF images were stored on the PC for later analysis at the completion of the study. 2.3. QLF analysis Fig. 2. (a) Cross-section through tooth showing the depth of the cavity that is filled with wax for easy visualization. The cavity extends through enamel into enamel. (b) A study tooth mounted in black plastic cap with amalgam restoration. (c) The same tooth as (b) under QLF conditions. Note the lack of reflections in comparison with (b) and the appearance of the amalgam. (d) Diagrammatic representation of the regions of the tooth. Region x is free of varnish and includes the restoration, region y is the control area and the remainder of the tooth z is coated with acid resistant nail varnish.

attached to the bottles containing 30 ml of a demineralizing solution based upon the techniques described by ten Cate [23] (acetic acid, pH 4.5, F 0.05 ppm) and placed into an orbital incubator at 37 8C to provide gentle agitation. Fluoride was included within the demineralizing solution to prevent surface softening and erosion of the enamel. The samples were removed from the demineralizing solution at

Analysis of the lesions was conducted by a single blinded examiner using QLF version 2.00 (Inspektor Research Systems BV, NL). Each image file was duplicated to permit separate analysis of the control and experimental regions of the tooth. All images were analyzed in a random manner to remove bias for duration in solution and restorative material. A standard analysis technique was employed to reduce the subjectivity of the method [12,14], Fig. 3 presents details of the analysis method. Data recorded were the DF and DQ values for both control and experimental regions. In contrast to previous experiments where the software was used to detect and exclude the restoration from the DQ value the restorations were included in this case. The DQ value from the baseline image was subtracted from the two subsequent images

Fig. 3. Example of a QLF analysis of a restoration (amalgam) prior to demineralization. (a) The analysis patch is placed around the area of interest ensuring that the patch line falls on sound enamel. (b) The reconstructed image with the restoration subtracted. (c) The restoration or lesion is represented in a variety of colors relating to % fluorescent loss. (d) The DQ and DF values are shown, including the ‘lesion’ area at various fluorescent thresholds.

372

I.A. Pretty et al. / Dental Materials 19 (2003) 368–374 Table 2 Baseline QLF values for each material Restorative material

DQ (SD) region x

DQ (SD) region y

Compomer (Dyract) Amalgam (Megalloy) Temporary (Coltosol) Composite (Spectrum) Glass ionomer (Ketac-Fil)

20.033 (0.058) 2211.467 (36.839) 2115.533 (41.128) 20.033 (0.057) 228.867 (7.569)

20.066 (0.115) 20.033 (0.057) 20 (0) 20.066 (0.057) 20.066 (0.057)

Ebner, Switzerland). Sections were taken through all three experimental zones, x, y, and z. These sections were mounted onto custom brass anvils with nail varnish and polished using a diamond disc to give planoparallel specimens of 80 mm thickness. The fine sections were then mounted onto a microradiographic plate-holder with an aluminium stepwedge (25 mm steps). Kodak high-resolution plates (type 1A) were employed with a 15 min exposure using a Cu(Ka) X-ray source (Philips B.V., The Netherlands) operating at 25 kV and 10 mA at a focus-specimen distance of 30 cm. Plates were developed using Kodak brand materials following manufacturers’ instructions. Following developing the microradiographs were analyzed using a Leica DMRB microscope (Leica, Germany) with image capture via a CCD video camera (Sony, Japan) connected to a PC. The presence or absence of a visible subsurface lesion was recorded. Images of the sections were retained.

3. Results 3.1. Appearance of restorative materials under QLF conditions

Fig. 4. Appearance of each restorative material under both white-light and QLF conditions prior to demineralisation: (a) compomer (Dyract), (b) amalgam (Megalloy), (c) temporary (Coltosol), (d) composite (Spectrum) and (e) glass ionomer (Ketac Fil).

(72 and 144 h) to represent any increase in mineral loss. This avoids any mineral loss potentially being excluded with the restoration in error. Following completion of the analysis, the data were decoded and entered into Excel and then exported to SPSS for statistical analysis. 2.4. Transverse microradiography Following completion of the QLF analysis stage of the study, each of the teeth was sectioned (approximately 250 mm) using a water-cooled diamond saw (Well, Walter

An example of each of the materials under white-light and QLF conditions is shown in Fig. 4. Baseline DQ values for each material are shown in Table 2. The 5% DQ threshold value was used in all cases. The control areas all showed that no demineralization was present at baseline. Amalgam, temporary material and glass ionomer all caused a reduction in fluorescence, however, the compomer and composite restorations demonstrated a relative increase in fluorescence with respect to the controls. Fig. 4 demonstrates that these materials appear lighter under QLF conditions. The values obtained at baseline were subtracted from subsequent reading to indicate the overall decrease in fluorescence (if any) associated with mineral loss. Table 3 shows the change (from baseline) of DQ values for each material at 72 and 144 h combined with control and TMR data. In all cases except composite and compomer (at 72 h) the control areas exhibited higher (less loss of fluorescence) DQ values than the restored areas. Following TMR analysis and sectioning all exposed areas (zones x, y ) demonstrated enamel subsurface lesions (Fig. 5). All covered areas (zone y) exhibited no subsurface lesions,

373

0,0,0 0,0,0 0,0,0 0,0,0 0,0,1

Fig. 5. (a) QLF image of tooth (restored with amalgam) at 72 h showing darker areas in regions x and y that were not clinically visible and which relate to early enamel demineralization. (b) White-light photographed tooth of the same tooth at 144 h showing clinically visible demineralization in regions x and y.

although there was evidence of surface softening on tooth three restored with glass ionomer, Fig. 6. Statistically, significant differences were detected between baseline and 72 and 144 h using ANOVA followed by t-tests on all restoratives except compomer, where significance was only noted between baseline and 144 h, p . 0.05.

4. Discussion and conclusions

Data for all three teeth in each group are shown. 0 ¼ no subsurface lesion detected, 1 ¼ subsurface lesion detected.

The described study presents an initial investigation to determine the ability for QLF to detect enamel demineralization adjacent to a variety of dental restorative materials. The study also shows the appearance of such restorative materials under QLF lighting conditions. In the study, all teeth demonstrated demineralization in the control and restorative zones. The control zones demineralized less than

a

1,1,1 1,1,1 1,1,1 1,1,1 1,1,1 23.7 (1.706) 210.733 (2.542) 24.5 (2.364) 213.533 (3.245) 27.566 (a4.909) 21.467 (0.586) 24.933 (3.647) 22.9 (2.163) 27.766 (5.552) 26.866 (4.366) 20.333 (0.416) 267.567 (30.886) 244.167 (59.092) 24.367 [7.563) 222.1 (0.796) Compomer (Dyract) Amalgam (Megalloy) Temporary (Coltosol) Composite (Spectrum) Glass ionom er (Ketac-Fil)

23.9 (1.480) 278.567 (142.984) 271.267 (65.999) 29.133 (8.237) 234.7 (6.907)

1,1,1 1,1,1 1,1,1 1,1,1 1,1,1

z y x

144 h region y DQ increase (SD) 144 h region x DQ increase (SD) 72 h region y DQ increase (SD) 72 h region x DQ increase (SD) Restorative material

Table 3 Data from the QLF and TMR investigations

TMRa examination zone

I.A. Pretty et al. / Dental Materials 19 (2003) 368–374

Fig. 6. (a) Transverse microradiograph of amalgam (amal) and enamel (enam) at 5 £ 0.11 magnification. (b) Transverse microradiograph of enamel adjacent to amalgam restoration showing a subsurface lesion at 20 £ 0.46 magnification: (I) buccal surface, (II) subsurface zone of demineralization, (III) sound enamel. (c) Transverse microradiograph of composite (comp) and enamel (enam) at 5 £ 0.11 magnification. (d) Transverse microradiograph of enamel adjacent to composite restoration showing a subsurface lesion at 20 £ 0.46 magnification: (I) buccal surface, (II) subsurface zone of demineralization, (III) sound enamel.

374

I.A. Pretty et al. / Dental Materials 19 (2003) 368–374

the restored areas as measured by decrease in DQ. This can be explained by the entirely intact enamel surface in the control zones and the occurrence of microleakage adjacent to the restoration. In this study a longitudinal increase in demineralization was observed from 72 to 144 h in both control and restorative areas. The amalgam restoration demonstrated the most demineralization followed by the temporary material, glass ionomer, composite, and finally compomer. Indeed, it is surprising that a temporary filling material exhibited less demineralization than amalgam [24,25]. However, statistical tests to determine differences between these materials were not performed as the sample size is small, and the study is at a ‘proof-of-concept’ stage. Caution in interpreting these results as material resistance to secondary decay must be shown. Despite the small sample size, this pilot study has demonstrated that QLF can detect demineralization longitudinally adjacent to restorative materials. The use of an initial baseline image and DQ value should be used to subtract from further images to determine the increase in mineral loss. The use of a threshold DQ to eliminate a restoration has been suggested. This could only be used in those restorative materials that exhibit a loss in fluorescence greater than that seen in demineralized enamel, e.g. amalgam. This would exclude the composite materials (which demonstrate no loss in fluorescence) and others such as glass ionomers that exhibit fluorescence loss similar to that of demineralized enamel. By using a baseline value and subtraction technique an analysis method common to all restorative materials can be employed. Further research is required to determine if QLF can detect differences in demineralization resistance of restorative materials, although a larger sample size will be required. As a predominately in vivo device, studies using the device in a clinical setting with longitudinal monitoring will develop the technique further. Difficulties in optimizing conditions for QLF in vivo exist, such as ambient light levels, dryness of teeth, stain, and plaque which must all be considered [14]. Researchers accept that current methods of in vivo caries detection, particularly adjacent to restorations, are insufficiently sensitive [19]. The development of QLF to detect early, non-visible lesions will increase the detection of such lesions. This early detection will allow the implementation of preventative regimes that can reverse such demineralization and reduce the need to replace restorations.

References [1] Angmar-Mansson B, al-Khateeb S, Tranaeus S. Monitoring the caries process. Optical methods for clinical diagnosis and quantification of enamel caries. Eur J Oral Sci 1996;104(4):480 –5. Pt 2. [2] Angmar-Mansson B, ten Bosch JJ. Advances in methods for diagnosing coronal caries—a review. Adv Dent Res 1993;7(2):70– 9. [3] Ashley PF, Blinkhorn AS, Davies RM. Occlusal caries diagnosis: an in vitro histological validation of the electronic caries monitor [ECM] and other methods. J Dent 1998;26(2):83–8.

[4] Hintze H, Wenzel A, Danielsen B, Nyvad B. Reliability of visual examination, fibre-optic transillumination, bite-wing radiography, and reproducibility of direct visual examination following tooth separation for the identification of cavitated carious lesions in contacting approximal surfaces. Caries Res 1998;32(3):204 –9. [5] Ross G. Caries diagnosis with the DIAGNOdent laser: a user’s product evaluation. Ont Dent 1999;76(2):21–4. [6] King NM, Shaw L. Value of bitewing radiographs in detection of occlusal caries. Community Dent Oral Epidemiol 1979;7(4): 218 –21. [7] De Josselin de Jong E, van der Linden AH, ten Bosch JJ. Longitudinal microradiography: a non-destructive automated quantitative method to follow mineral changes in mineralised tissue slices. Phys Med Biol 1987;32(10):1209–29. [8] AlKhateeb S, ten Cate JM, Angmar-Mansson B, de Josselin de Jong E, Sundstrom G, Exterkate RA, et al. Adv Dent Res 1997;11(4):502–6. [9] Bjelkhagen H, Sundstrom F, Angmar-Mansson B, Ryden H. Early detection of enamel caries by the luminescence excited by visible laser light. Swed Dent J 1982;6(1):1– 7. [10] De Josselin de Jong E, Sundstrom F, Westerling H, Tranaeus S, ten Bosch JJ. A new method for in vivo quantification of changes in initial enamel caries with laser fluorescence. Caries Res 1995;29(1):2 –7. [11] Pretty IA, Edgar WM, Higham SM. The effect of ambient light on QLF analysis. J Oral Rehabil 2002;29(4):369–73. [12] Van der Veen MH, de Josselin de Jong E. Application of quantitative light-induced fluorescence for assessing early caries lesions. Monogr Oral Sci 2000;17:144 –62. [13] Shi XQ, Tranaeus S, Angmar-Mansson B. Comparison of QLF and DIAGNOdent for quantification of smooth surface caries. Caries Res 2001;35(1):21– 6. [14] Pretty IA, Hall AF, Smith PW, Edgar WM, Higham SM. The intraand inter-examiner reliability of quantitative light-induced fluorescence (QLF) analyses. Br Dent J 2002;193(2):105 –9. [15] Mjor IA, Toffenetti F. Secondary caries: a literature review with case reports. Quintessence Int 2000;31(3):165–79. [16] Burke FJ, Cheung SW, Mjor IA, Wilson NH. Reasons for the placement and replacement of restorations in vocational training practices. Prim Dent Care 1999;6(1):17 –20. [17] Wilson NH, Burke FJ, Mjor IA. Reasons for placement and replacement of restorations of direct restorative materials by a selected group of practitioners in the United Kingdom. Quintessence Int 1997;28(4):245 –8. [18] Fontana M, Gonzalez-Cabezas C. Secondary caries and restoration replacement: an unresolved problem. Compend Contin Educ Dent 2000;21(1):15– 18. [19] Goldberg AJ. Deterioration of restorative materials and the risk for secondary caries. Adv Dent Res 1990;4:14–18. [20] Dionysopoulos P, Kotsanos N, Papadogiannis Y. Lesions in vitro associated with a Fl-containing amalgam and a stannous fluoride solution. Oper Dent 1990;15(5):178–85. [21] Blackman D, van der Veen M, Lagerweij M, Ando M, Stookey G. Laser fluorescence diagnosis of demineralised enamel adjacent to resin restorations. J Dent Res 1997;76:131. Abstract 943. [22] Tranaeus S, de Josselin de Jong E, Lussi A, Angmar-Mansson B. Quantitative light-induced fluorescence for assessment of enamel caries around fillings: a pilot study. Caries Res 1997;31:324. Abstract 132. [23] ten Cate JM, Duijsters PP. Alternating demineralization and remineralization of artificial enamel lesions. Caries Res 1982;16(3):201–10. [24] Derand T, Birkhed D, Edwardsson S. Secondary caries related to various marginal gaps around amalgam restorations in vitro. Swed Dent J 1991;15(3):133–8. [25] Dijkman GE, Arends J. Secondary caries in situ around fluoridereleasing light-curing composites: a quantitative model investigation on four materials with a fluoride content between 0 and 26 vol%. Caries Res 1992;26(5):351–7.