Effect of silver and fluoride ions on enamel demineralization: a quantitative study using micro-computed tomography


Professor Edward CM Lo
3/F, Prince Philip Dental Hospital
34 Hospital Road
Hong Kong
Email: edward-lo@hku.hk


Background:  This study aimed to investigate the effect of silver and fluoride ions on demineralization of enamel.

Methods:  The coronal parts of 40 extracted sound premolars were prepared into tooth blocks. An unvarnished occlusal surface window (OW) and a flat buccal/lingual surface window (FW) were created for each tooth by covering all other surfaces with an acid-resistant varnish. These blocks were randomly allocated into four groups of 10 blocks each and immersed in respective solutions for 5 minutes: Group 1 – 2.36 M silver fluoride; Group 2 – 2.36 M potassium fluoride; Group 3 – 2.36 M silver nitrate; and Group 4 – deionized water. After 7-day immersion in a buffered demineralization solution at pH 4.4, micro-CT scans were taken.

Results:  Mean lesion depth in the FW area for tooth blocks in AgF, KF, AgNO3 and control groups were 0 μm, 3.3 ± 10.3 μm, 156.3 ± 30.8 μm, and 173.6 ± 48.6 μm, respectively (p < 0.001). The difference in mean lesion depth between the AgNO3 and control groups was not statistically significant (p > 0.05). Similar OW and FW lesions were observed in tooth blocks in the AgNO3 and control groups.

Conclusions:  Topical application of a 2.36 M fluoride solution can inhibit demineralization of enamel while topical application of silver ions has little effect.

Abbreviations and acronyms:

average lesion depth


cross-section microhardness


flat surface window


micro-computed tomography


occlusal surface window


polarized light microscopy


region of interest


silver diamine fluoride


transverse microradiography


volume of interest


Silver diamine fluoride (SDF) solution is a colourless aqueous solution which contains both ammonia and silver fluoride (AgF). Ammonia helps to stabilize AgF in the solution by forming a complex and stable ion with the chemical formula [Ag (NH3)2]+ through reversible reaction. AgF is the active agent in the solution when it is applied onto the tooth surface.1 Results of clinical studies have shown that SDF is effective in preventing dental caries development.2–5 However, its mechanism of action is unclear. It has been reported that SDF or AgF can significantly enhance the acid resistance of enamel,6,7 but the actions and effects of the different ions have not been specified. It is known that fluoride is effective in preventing dental caries by inhibiting demineralization and promoting remineralization of dental tissues, especially enamel.8–10 Few studies have investigated if there is any specific effect of silver ion on enamel demineralization.

Micro-computed tomography (micro-CT) is an X-ray tomography system which has an X-ray source with a micro focus (focal spot diameter <100 μm). It was developed in the early 1980s,11 and micro-CT devices from different commercial companies are now available. Micro-CT is able to provide high resolution two-dimensional (2-D) and three-dimensional (3-D) digital images through non-destructive scanning. The scan images can be used for qualitative and/or quantitative evaluation and analysis. Micro-CT devices are frequently used in the study of hard tissues, especially bone, in the biomedical field.12 They have also been used for different purposes in dental research.13–16 It has been found that using micro-CT images is a reliable method in making linear measurements compared to direct measurements,17 and that the ability of micro-CT in detecting changes in artificial caries lesions after remineralization is comparable to that of micro-radiography.18

The objective of this study was to evaluate the effect of silver and fluoride ions on enamel demineralization by using micro-CT.

Materials and Methods

The flow of procedures in this laboratory study is shown in Fig. 1. Sound permanent premolars were taken from the tooth collection of a dental hospital. These teeth were from patients who had their teeth extracted for orthodontic reasons. Patients’ informed consent was obtained before tooth collection. Teeth with heavy staining in the occlusal fissures or with signs of early caries were excluded. The collected teeth were placed in 6% sodium hypochlorite solution for 24 hours after extraction to remove all soft tissues, staining and plaque. They were then stored in a 0.5% thymol solution at a temperature of 4 °C. Fissures were cleaned carefully by using a mini-brush mounted on a slow-speed handpiece. A laser-induced fluorescence caries detection device (DIAGNOdent, KaVo, Germany) was used to assess the status of the fissure. Teeth with DIAGNOdent readings equal to or above 15, indicating probable presence of enamel caries,19 were excluded from this study.

Figure 1.

 Flow of the study procedures.

The selected premolars were sectioned transversely at the crown–root junction and the coronal parts were prepared into tooth blocks. Thin sections were cut off from the mesial/distal and buccal/lingual surfaces of the coronal part of each tooth by using a home-made microtome saw. The cusps were ground slightly when necessary to create tooth blocks with natural fissures and with a height, length and width of no more than 6 mm. Stickers measuring 3 × 5 mm were used to cover the fissures in occlusal surfaces while stickers measuring 1 × 2 mm covered the fresh enamel area of buccal/lingual surfaces of each tooth block before a red-coloured acid-resistant nail varnish was painted onto the surfaces. The stickers were later carefully removed when the varnish was dry. The unvarnished surface windows (occlusal surface window – OW and flat enamel surface window – FW) were exposed for intervention and demineralization (Fig. 2).

Figure 2.

 Prepared tooth block showing the unvarnished surface windows.

Forty tooth blocks were prepared. One OW and one FW was created for each tooth block. The tooth blocks were randomly allocated into four study groups with 10 blocks each to be treated by different solutions, namely: Group 1 – 2.36 M (mole/L) silver fluoride (AgF); Group 2 – 2.36 M potassium fluoride (KF); Group 3 – 2.36 M silver nitrate (AgNO3); and Group 4 – control (deionized water). These concentrations of silver and fluoride ions, i.e. 2.36 M, were equivalent to those in a 38% AgF solution which was used in a number of clinical trials on caries control.2,3

According to their group allocation, the tooth blocks were immersed in the assigned solution or water undisturbed for 5 minutes. They were then taken out, dried completely with a blast of compressed air for 1 minute and transferred to four separate cups containing 50 ml of demineralization solution to begin the demineralization process. The demineralization solution used in this study contained 2.2 mM CaCl2, 2.2 mM NaH2PO4, and 50 mM acetic acid; 1 M KOH was used to adjust the pH of the solution to 4.4. Mini-brushes were used to fully soak the pit and fissure in the OW with the demineralization solution at the beginning of the demineralization process. The four cups were kept on a shaker (Orbit 1000, Labnet, USA) during the whole demineralization process (168 hours).

The tooth blocks were taken out of the demineralization solution at the end of the demineralization process, then cleaned and dried. The mesial or the distal flat surface of each tooth block was stuck onto one of the flat ends of a cylindrical resin bar 1 cm in diameter and 3 cm in length. The cylindrical resin bars, together with the tooth blocks on them, were then fixed in container tubes and scanned by micro-CT individually. The micro-CT machine used in this study was SkyScan 1076 (SkyScan, Antwerp, Belgium). It is a cone beam micro-CT with a focus spot size of less than 5 μm and a highest spatial resolution of 9 μm. The X-ray source was operated at a voltage of 100 kV and a current of 80 μA. The highest spatial resolution of 9 μm was used for the scanning. Signal-to-noise ratio was chosen at 5, and a 1 mm aluminium filter was used to cut off the softest X-rays.

Assessment of the micro-CT scans and data analysis

Scanning results of each tooth block were reconstructed using the reconstruction software NRecon (SkyScan Company, Antwerp, Belgium). The reconstructed 3-D images were viewed and processed using the data analysing software CTAn (SkyScan Company, Antwerp, Belgium).

From the reconstructed 3-D image of each tooth block, cross-sectional images showing the whole lesion in the FW area in the tooth block were located. From these lesion images, 20 images were selected by systematic random sampling. The demineralization lesions shown in these 20 images were chosen to represent the whole lesion body of that particular window.

Region of interest (ROI) was drawn for each of the sampled image. The ROI drawn covered the whole body of the demineralized enamel lesion. The external border of the ROI was drawn on the surface margin of the enamel and the other borders of the ROI went beyond the lesion body into sound enamel (Fig. 3). The volume within the ROI was the selected volume of interest (VOI) for one selected slice in this study. Combining the VOIs of the 20 selected slices resulted in the VOI for the particular window of a tooth block.

Figure 3.

 An ROI for taking measurement of average lesion depth.

Greyscale values of the sound enamel (GE) in the image were estimated from the image profile. Image area with a greyscale value of more than 95% of the GE was defined as sound enamel. The volume of sound enamel within the VOI of one particular window was calculated by the software CTAn. Total tissue (any substances) volume within that VOI was also calculated. Total lesion volume within the VOI was then calculated by subtracting the sound enamel volume from the total tissue volume. Width of the demineralization lesion in each of the selected images was measured. The sum of the width measurements of the demineralization lesion from the 20 selected images multiplied by the slice thickness gave a result for the surface area of the lesion of the VOI. Finally, the average lesion depth (ALD) of the demineralization lesion in each window was calculated by dividing the total lesion volume by the surface area of the lesion. Differences in the mean ALD between the four study groups were compared. One-way ANOVA and independent samples t-test were used.


After 7-day demineralization, visual examination found that there were demineralization lesions (white opacity) in both OW and FW areas of the tooth blocks in the control group. Demineralization lesions in the tooth blocks in the KF group were not apparent. The OW and FW areas of the tooth blocks in the AgNO3 and AgF groups had a greyish silver or light brown appearance which hindered direct observation of the demineralization lesions.

Demineralization lesions could be easily seen in both OW and FW areas in the 3-D images of all the tooth blocks in the control and the AgNO3 groups (Fig. 4). The demineralization lesion observed in the OW area was similar to that in the FW area. No tooth block in the AgF group showed demineralization lesion and only one tooth block in the KF group showed a shallow demineralization lesion.

Figure 4.

 Examples of micro-CT images of tooth blocks after demineralization. Image (a) is from one of the tooth blocks in Group 1 (AgF); image (b) is from one of the tooth blocks in Group 2 (KF); image (c) is from one of the tooth blocks in Group 3 (AgNO3); image (d) is from one of the tooth blocks in Group 4 (control).

Mean ALD values in the FW area of tooth blocks in the AgF, KF, AgNO3 and control groups were 0 μm, 3.3 ± 10.3 μm, 156.3 ± 30.8 μm, and 173.6 ± 48.6 μm, respectively. The mean ALD of the KF group was significantly lower than that of the AgNO3 or the control group (p < 0.001). Independent samples t-test showed that there was no statistically significant difference (p > 0.05) between the mean ALD values of the AgNO3 and the control groups.


This randomized controlled study was conducted to compare the effect of silver and fluoride ions on the prevention of enamel demineralization. The ALD values of the tooth blocks that had been immersed in solutions containing 2.36 M fluoride ions (AgF or KF) before the demineralization process were either zero or close to zero. This is in great contrast to the observations in the tooth blocks in the control group where the mean ALD was 173.6 μm. This finding shows that topical application of a high concentration (2.36 M or more than 30 000 ppm) fluoride solution is effective in preventing enamel demineralization. This result is in agreement with previous studies.8,10 Clinical studies have also found a positive correlation between fluoride concentration and its effectiveness in dental caries prevention.20,21

It is important to note that the results were true for both the OW and the FW. This shows that fluoride is effective in preventing enamel demineralization not only on flat surfaces but also in pits and fissures, and with similar effectiveness. Evidence from clinical studies has shown that fluoride is effective in preventing pit and fissure caries in permanent molars.3 A recent review on glass-ionomer cement (GIC) suggests that GIC sealant can release fluoride to prevent dental caries in pits and fissures.22 By using 3-D micro-CT scanning, the results of this study provide the required laboratory evidence to support the findings of the clinical studies.

The present study found that there was no significant difference between the mean ALD values of the AgNO3 and the control groups. Therefore, it may be concluded that silver ion in the solution has little effect in preventing enamel demineralization. This finding should not be taken as a contradiction to those of previous studies which found that silver nitrate was an effective agent in preventing the formation of artificial caries lesions.23,24 This is because the present study focused on investigating silver ion’s effect on preventing the formation of enamel demineralization lesion which differs from the artificial caries lesion produced in a real or artificial mouth model involving bacteria in the other studies. The ability of silver ion to prevent artificial caries lesion development may be due to its strong antibacterial effect.

Clinical development of dental caries involves the co-existing processes of demineralization and remineralization. In this laboratory study, effect of the ions on preventing the development of enamel demineralization lesion induced by acid was investigated but not that of remineralization. This is a limitation of this study in relation to clinical dental caries prevention. Despite this, the present study contributes to understanding the mechanism of enamel demineralization prevention. We have shown that topical application of a high concentration fluoride solution was highly effective in inhibiting enamel demineralization which is part of the process of clinical dental caries development and the crucial process of enamel erosion by acid.

We chose measurements taken of the demineralization lesion in the FW area for two reasons. First, micro-CT scanning results revealed that the appearance and depth of the demineralization lesion in the OW area was similar to that in the FW area. Thus, lesions in either window areas can represent the phenomena. Second, taking measurement of demineralization lesions in the FW area was easier.

Transverse microradiography (TMR), cross-section microhardness (CSMH) testing and polarized light microscopy (PLM) are sensitive evaluation methods which can be used for the assessment of demineralization lesions.25 However, these techniques require destructive, complicated and time-consuming specimen preparation procedures.26 Therefore, non-destructive longitudinal demineralization and remineralization studies of tooth specimens can hardly be conducted by using these techniques. Use of high-resolution micro-CT can overcome these problems. The ALD calculated from the micro-CT scanning results in this study provides a quantitative measure of demineralization by taking advantage of the 3-D information available. This makes it more suitable for making comparisons between different lesions. There is still a technical problem with the method used in this study when assessing lesions in pit and fissure areas because they have irregular surface borders. The difficulty lies in drawing ROI properly and measuring the path length of the irregular lesion surface border accurately. Future improvement in computer software may solve this problem.

Only the effect of fluoride and silver ions on enamel demineralization were investigated in this study. Nonetheless, the findings contribute to understanding the mechanism of action of topical application of AgF in the prevention of dental caries in clinical studies. By using micro-CT scanning which allows for the non-destructive study of tooth blocks, further studies on the effects of fluoride and silver ions on tooth blocks undergoing a pH-cycling challenge or bacterial challenge in an artificial mouth model should be conducted to better understand the specific role which fluoride and silver ions may play in the caries development process.

In conclusion, topical treatment of the enamel surface with a high concentration fluoride solution can inhibit demineralization of enamel in both occlusal and flat tooth surfaces, while topical treatment with silver ions has little effect.


This study was funded by the Research Grants Council of Hong Kong (Ref no. HKU771207M) and was based on a thesis submitted to the Faculty of Dentistry, University of Hong Kong, in partial fulfilment of the requirement for a PhD degree.