Measuring enamel erosion: A comparative study of contact profilometry, non-contact profilometry and confocal laser scanning microscopy

Abstract

Objectives

To compare three instruments for their ability to quantify enamel loss after acid erosion.

Methods

6 randomized parallel groups of bovine enamel samples were subjected to citric acid (higher acidity) or orange juice (lower acidity) erosion and remineralisation in a cycling model. Two protected shoulders were created on each of the samples using tape, to serve as reference for analysis. The time of exposure to each acid was varied, along with presence or absence of agitation. After treatment, samples were measured on 3 instruments capable of measuring step height: a contact profilometer (CP); a non-contact profilometer (NCP); and a confocal laser scanning microscope (CLSM) by three different examiners. Additionally, 3D (volume) step height was also measured using the CLSM.

Results

Increasing acid concentration and exposure time resulted in greater erosion, as did agitation of samples while in acid solution. All instruments/methods identified the same statistically significant ( p < 0.05) pair-wise differences between the treatments groups. Further, all four methods exhibited strong agreement (Intra-class correlation ≥ 0.96) in erosion level and were highly correlated, with correlations of 0.99 or higher in all cases.

Significance

All instruments/methods used in this study produced very similar conclusions with regard to ranking of enamel loss, with data showing very high agreement between instruments. All instruments were found to be equally suited to the measurement of enamel erosion.

Introduction

Research in the area of tooth wear has been growing over the last decade or so. There are a number of reports detailing this as a growing problem among Western populations . Tooth wear is widely considered to consist of 3 main aspects, two of which are physical wear (abrasion, attrition) the third which is chemical dissolution (erosion) . It is challenging to separate the two components in clinical models, so in vitro models are often used to help understand each component separately, although combined models which include both aspects are also important .

The methods used for measuring damage caused by tooth wear are many and varied, and discussed at length by Attin . Quantitative assessment of tooth wear has most often been reported using surface profilometry. This has the advantage of being reasonably straightforward to conduct, and simple to understand, as it allows a step measurement (in microns) of enamel lost after exposure to acid, compared to a protected/undamaged (control) portion of the sample. However, surprisingly little research has been reported comparing the different types of instruments for surface profile measurement . A notable comparison the authors could find was conducted by Heurich (2010) , where two contact profilometers (CP) were compared to a confocal laser scanning microscope (CLSM) and an Atomic Force Microscope (AFM). All instruments performed well and the same experimental conclusion would have been reached with each instrument.

Two main types of profilometer are available, contact and non-contact. Contact profilometers use a stylus moved across the surface to record the surface profile. While relatively simple, this traditional method has the potential risk of affecting the reading or even damaging the sample as a consequence of the contact . Non-contact profilometers generally use some type of laser to scan the surface to create the profile. In addition, non-contact profilometers usually generate a surface plane rather than just simple line profiles, which allows volumetric loss analysis . However, while removing the risk of surface damage due to contact, the type of laser scanning system needs careful selection, as reflective and/or translucent surfaces (such as enamel) can cause inconsistencies when profiled. A recent alternative/variation on the non-contact profilometer is the confocal laser scanning microscope (CLSM). This combines the laser scan with capture of a traditional visible light microscope image, producing a detailed 3D image of the surface. Traditionally these instruments did not provide quantitative data (i.e. scans/images calibrated to the μm level), but more modern instruments are now able to do this . CLSM is used extensively elsewhere in science and engineering research .

The null hypothesis for the study was that differing instruments measuring the same samples provided data with poor agreement. The approach was to utilize a laboratory model for erosion to compare the data output from surface profiling instruments with differing analytical protocols and located at different institutions in the UK.

Materials and methods

Study design and procedure

Summary outline

Each of 48 bovine enamel samples (Therametric Technologies Inc., Noblesville, IN, USA) was randomly assigned into one of 6 parallel groups (A1–A6), 8 samples per group. Samples were visually examined under a magnifier prior to inclusion in the study to check for significant defects and discarded as appropriate. A cycling model was employed to induce erosive damage on the samples. Each cycle commenced with 30 min immersion in remineralizing solution (remin solution described by Eisenburger et al. ), at mouth temperature with gentle agitation (Nickel Electro Clifton NE528D Shaking Waterbath with horizontal linear agitation at 24 cycles/min.), followed by deionized water rinse. Samples were then immersed in 100 mL of acid, either citric acid or commercial orange juice, at room temperature (RT, 20 ± 1 °C), followed by a final rinse in deionized (DI) water. The cycle was repeated 24 times. The 0.05 mol/L citric acid, (CA), was freshly made in the laboratory on a daily basis from deionized (DI) water and monohydrate citric acid and had a pH 2.26, with a titratable acidity of 18.0 . The orange juice (Waitrose Essentials Orange Juice, referred to as OJ) had a pH of 3.8 and a titratable acidity of 10.5. The central portion of each sample was exposed to acid demineralization, the remainder of the sample masked to prevent acid contact. Groups A1, A2, A3 and A5 were treated with OJ, while groups A4 and A6 were treated with CA. Groups A1–A4 had a total acid exposure time of 180 min (7.5 min/cycle), while the other two groups had a total acid exposure time of 300 min (12.5 min/cycle). Further, while in acid, group A2 was gently agitated to induce increased erosion (Stuart See-Saw Rocker SSL4, set at 24 cycles/min). Group A3 samples were suspended upside-down in the acid (enamel surface facing down). See Table 1 for a summary. All acid exposure was done at room temperature (20 ± 1 °C).

Table 1
Study design groups.
Group Acid type a Agitation? Suspended? Acid immersion time/cycle (min) Total acid exposure (min)
A1 OJ None No 7.5 180
A2 OJ Rocker No 7.5 180
A3 OJ None Yes 7.5 180
A4 CA None No 7.5 180
A5 OJ None No 12.5 300
A6 CA None No 12.5 300

a OJ = Orange Juice, CA = 0.05 M citric acid.

After 24 cycles had been completed, the samples were allowed to air-dry, before the tape was removed. A clear demineralized line was obvious on all samples. The samples were stored dry in an air-tight container before scanning. Use of the erosion model enabled creation of a diverse set of data to fully test the instruments and their associated measuring protocols.

Measurement of enamel loss

Measurement of enamel loss was conducted on 3 different instruments, all within a 6 week period. A trained and blinded operator experienced in the use of each instrument conducted each set of measurements.

The samples were first analyzed on a Surftest SV-200 (Mitutoyo, UK), a contact profilometer (CP) with diamond-tip stylus, available at Bristol Dental Hospital, UK. Contact profilometry drags the stylus across the eroded gap to output a 2D line profile (step height measurement). Prior to use, the instrument was calibrated using a reference block with a step height range of 2.70–3.30 μm. The maximum vertical resolution of the Surftest SV-200 is 0.01 μm. For each enamel sample, two readings were taken across the eroded sample, analyzed with the software Surpak (Mitutoyo, UK) and values averaged to give the step height of enamel loss in microns. Although perfect repositioning accuracy is impossible at the micron level, the sample was roughly in the same position for every measurement.

The samples were then analyzed on a XYRIS 2000 (Taicaan, UK) available at King’s College London Dental Institute, UK. The profilometry measurement protocol was validated to an accuracy of 0.042 μm for measuring step height enamel loss. The accuracy of measurement was determined using a calibrated 2.47 μm step height reference standard (DSS1 step standard, National Physical Laboratory, Teddington, UK) following metrological good-practice guidelines . This laser microscope scans a group of data points then links them to get a 3D height map. The data were then analyzed on Boddies v1.92 (Taicaan, UK). Two step measurements were taken for each sample and averaged to give the height of enamel loss of the sample, in microns.

Finally, the samples were analyzed on a LEXT OLS3100 (Olympus, UK), available at the National Physical Laboratory, Teddington, UK. This instrument has the option of taking both a real color image using optical light, as well as a surface scan, using a laser with a wavelength of 408 nm. This system has the capability of taking images of both highly reflective and transparent materials. Confocal microscopy consists of capturing a sequence of images, with each image containing an x y dataset within a single focal plane. A 3D dataset is then constructed from the individual in focus image planes. The 3D dataset can then be used to produce a 3D image as well as allowing measurements to be made in 3 dimensions, e.g. profiles, depths, angles, volumes, etc. The instrument was calibrated by the manufacturer. A single image was captured of the surface of each sample using the ×20 objective, making the measure area 0.63 mm × 0.47 mm ( x and y ), with a maximum depth resolution of 0.128 μm.

After capture, the datasets were processed with the LEXT OLS4000 software. Step height was calculated using two methods:

  • (i)

    a simulation of a simple line trace (2D) method. 3 ‘trace lines’ were digitally placed across the eroded surface from one shoulder to the other, with each being 50 μm wide. For each trace line, two points were selected, one from each shoulder, to the lowest point in the eroded enamel area. The software calculated the height difference to each shoulder, and then the two values were averaged to account for any differences in shoulder height (e.g. in case the sample was not completely flat).

  • (ii)

    a calculation of 3D step height , based on the total volume of enamel lost within the CLSM field of view (0.63 × 0.47 mm portion in the center of the sample). The operator outlined the visibly eroded portion on the 2D image (overhead view) and the software was then used to calculate the area, and also the volume of enamel loss in this area up to the height of the reference shoulders. Volume (μm 3 ) was divided by area (μm 2 ) to get an average step height (μm) across the measured eroded portion. A similar approach has been reported previously .

As the CLSM was used last, it was possible to use this instrument to inspect the surface for marks/scratches left by the CP. On some (not all) samples, the trace line from the contact profilometer could be identified, but even where marks were visible, it was only on the eroded portion of the chip. Magnification was increased to the ×50 objective (which allows greater sensitivity) and measurement of the depth of the trace line was attempted. However, it was not possible to reliably measure the depth of CP scratches, indicating the depth was considerably less than 1 μm. The region around the contact profilometer trace line was obviously avoided for measurement by the other instruments, though it is unlikely it would have affected results significantly had it been within the measured area .

Statistical methods

For each method, comparisons of means between groups used analysis of variance assuming unequal variances. Pairwise p -values used the Tukey Honest Significant Difference (HSD) adjustment for multiple comparisons to control the experiment-wise error rate to be a two-sided 5% significance level. For each group, the mean, standard deviation (SD), and coefficient of variation % (CV = 100% × SD/mean) were calculated. The CV measures the percentage of noise relative to the signal of the erosion level of a group. Within each group, repeated measures analysis was used to compare means between methods assuming unequal variances. The inter-method reliability (also known as the intra-class correlation or ICC) was calculated to measure agreement between methods . For the inter-method reliability of agreement, a value near 0 indicates low agreement while a value close to 1 indicates a high level of agreement between methods. In this case, agreement occurs when both methods produce a numerically similar value of erosion. The Pearson correlation coefficient was also calculated to measure whether there was a linear relationship between methods. Note that it is possible to have a strong correlation and still have poor agreement between methods if a systematic shift occurs in the scores of one method relative to another.

Materials and methods

Study design and procedure

Summary outline

Each of 48 bovine enamel samples (Therametric Technologies Inc., Noblesville, IN, USA) was randomly assigned into one of 6 parallel groups (A1–A6), 8 samples per group. Samples were visually examined under a magnifier prior to inclusion in the study to check for significant defects and discarded as appropriate. A cycling model was employed to induce erosive damage on the samples. Each cycle commenced with 30 min immersion in remineralizing solution (remin solution described by Eisenburger et al. ), at mouth temperature with gentle agitation (Nickel Electro Clifton NE528D Shaking Waterbath with horizontal linear agitation at 24 cycles/min.), followed by deionized water rinse. Samples were then immersed in 100 mL of acid, either citric acid or commercial orange juice, at room temperature (RT, 20 ± 1 °C), followed by a final rinse in deionized (DI) water. The cycle was repeated 24 times. The 0.05 mol/L citric acid, (CA), was freshly made in the laboratory on a daily basis from deionized (DI) water and monohydrate citric acid and had a pH 2.26, with a titratable acidity of 18.0 . The orange juice (Waitrose Essentials Orange Juice, referred to as OJ) had a pH of 3.8 and a titratable acidity of 10.5. The central portion of each sample was exposed to acid demineralization, the remainder of the sample masked to prevent acid contact. Groups A1, A2, A3 and A5 were treated with OJ, while groups A4 and A6 were treated with CA. Groups A1–A4 had a total acid exposure time of 180 min (7.5 min/cycle), while the other two groups had a total acid exposure time of 300 min (12.5 min/cycle). Further, while in acid, group A2 was gently agitated to induce increased erosion (Stuart See-Saw Rocker SSL4, set at 24 cycles/min). Group A3 samples were suspended upside-down in the acid (enamel surface facing down). See Table 1 for a summary. All acid exposure was done at room temperature (20 ± 1 °C).

Nov 25, 2017 | Posted by in Dental Materials | Comments Off on Measuring enamel erosion: A comparative study of contact profilometry, non-contact profilometry and confocal laser scanning microscopy

VIDEdental - Online dental courses

Get VIDEdental app for watching clinical videos