Time-dependent effect of intense capsule-coffee and bleaching on the color of resin-infiltrated enamel white spot lesions: an in vitro study

Sample preparation

A total of 20 surfaces from ten extracted sound human third molars were included in this study. All teeth were sterilized according to the guidelines proposed by the Center for Disease Control and Prevention (CDC) (Kohn et al., 2003). Teeth were cleaned and inspected for defects and visible stains that would interfere with the study procedures. The root section of each tooth was embedded in a cylindrical polyvinyl silicon base (Speedex Putty; Coltene Whaledent, Altstätten, Switzerland) for handling purposes. The crown part of the embedded teeth was covered with an acid-resistant nail varnish except for two small (4 × 4 mm) square windows—one on the buccal surface (test (n = 10)) and the other on the lingual surface (control (n = 10)). All teeth were stored in deionized water at 37 °C.

Subsurface lesion formation

The coronal portions of the embedded teeth were immersed in a demineralization solution {(2.2 mM CaCl₂, 10 mM NaH₂PO₄, 50 mM acetic acid, 100 mM NaCl, 1 ppm NaF, 0.02% NaN₃ ; pH 4.5) (Bakhsh et al., 2018; Bakry et al., 2018), 16 mL demineralization solution per exposed enamel surface (32 mL_solution/tooth)} for 6 days. The total volume of the demineralization solution used was determined using the following formula: ${\displaystyle {\mathbf{V\; olume}}_{\mathbf{total}}=2{\mathrm{mL}}_{\mathrm{solution}}/1{\mathrm{mm}}_{\mathrm{enamel area}}^2\text{(Ferreira et al, 2007)}.}$ Using this volume of similar demineralization solutions results in an almost 43 µm deep subsurface lesion (Magalhães et al., 2009). The establishment of sub-enamel demineralized lesions on each tooth’s exposed buccal and lingual surfaces was validated using micro-computed tomography (SkyScan 1172; Kontich, Belgium) (Oliveira et al., 2020). Figure 1 shows a micro-CT image of the artificial WSLs created.

Resin infiltration

Infiltration resin (ICON; DMG, Hamburg, Germany) was applied to the enamel surface of the buccal windows of teeth following the manufacturer’s instructions. After drying the teeth, Icon-Etch was applied for 2 min and then suctioned off and rinsed with deionized water for 30 s. Subsequently, the buccal surfaces were rinsed with 100% ethanol for 30 s and then air-dried. Icon-resin infiltration gel was applied after that for 3 min using a smooth surface ICON applicator with continuous massaging. The infiltrated buccal surfaces were then cured using an LED curing unit (3M Elipar™ LED curing light; 3M ESPE, Saint Paul, MN, USA) for 40 s with 1,200 mW/cm² irradiance. A second layer of ICON-infiltration resin was applied and light cured. The exposed lingual enamel surfaces of teeth were not infiltrated and served as a positive control. All teeth were then stored in deionized water at 37 °C for 24 h; then, baseline color measurements were recorded.

Baseline color measurement

Color measurement was performed against a white background under the same light source (daylight LED bulb) during all stages of the study and by one investigator (H.Y.) to avoid variability in color measurements. The tip of the spectrophotometer was positioned vertically, in a marked location, resting on the surface to be measured. To standardize color measurement throughout the study, an average of three readings on each surface were recorded for each color parameter.

Baseline spectrophotometric color measurements were determined for the test (buccal), and control (lingual) surfaces using the CIE (Commission International de L’Eclairage; Luo, 2014) L*a*b* parameters using a handheld spectrophotometer (VITA Easy shade advance, VITA Zahnfabrik, Bad Säckingen, Germany), where L* indicates lightness (a value of 100 indicates white and zero indicates black); a* represent red (positive) and green colors (negative); b* represents yellow (positive) and blue colors (negative). $L_v^{\ast }$ was recorded for test and control surfaces. L_v represents the tooth structure’s lightness difference in color space between the measured tooth shade and the closest matching VITA classical (A1-D4 shades), where higher values are positive and lower values are negative.

Coffee immersion procedures and color change

Following resin infiltration, all the samples were immersed in capsule-coffee staining solution (Starbucks Espresso Roast coffee capsules by Nespresso, Nestlé S.A., Vaud, Switzerland) for a total of 8 days. In a study by Ertas et al. (2006), it was reported that an average of 3.2 cups of coffee is consumed daily by average coffee drinkers and the average duration per cup was around 15 min (Ertas et al., 2006). Therefore, 24 h of storage in coffee would correspond to almost one month of coffee drinking (Ertas et al., 2006; Gamal, Safwat & Abdou, 2022; Hasanain, 2022). Consequently, eight days of 24 hours/day storage would correspond to 8 months of coffee drinking. The coffee was prepared using the espresso brewing setting in a 19-bar high-pressure pump coffee capsule machine (Nespresso lattissima touch; Nespresso by De’Longhi, Treviso, Italy). The coffee solution was allowed to cool down to 37 °C before immersing the teeth. The immersed teeth were kept in the coffee solution and stored in an incubator at 37 °C (Memmert, Schwabach, Germany). The coffee staining solution was replaced daily with a freshly brewed coffee solution.

After every two days of immersion in the staining solution, the samples were removed and rinsed gently with deionized water and dried with absorbent paper. The color of exposed surfaces (both test and control surfaces) was recorded using the CIE L*a*b* parameters in the spectrophotometer (VITA Easy shade advance; VITA Zahnfabrik, Bad Säckingen, Germany), as described above in the Baseline Color Measurement Section.

The color change due to immersion (ΔE_x) in the capsule-coffee staining solution was determined using the CIE 1976 L* a* b* (Luo, 2014) color system according to the following formula: ${\displaystyle \Delta {\mathbf{E}}_{\mathit{x}}=\sqrt{{\left({\mathit{L}}_{\mathit{staining}\left(\mathit{xdays}\right)}-{\mathit{L}}_{\mathit{baseline}}\right)}^2+{\left({\mathit{a}}_{\mathit{staining}\left(\mathit{xdays}\right)}-{\mathit{a}}_{\mathit{baseline}}\right)}^2+{\left({\mathit{b}}_{\mathit{staining}\left(\mathit{xdays}\right)}-{\mathit{b}}_{\mathit{baseline}}\right)}^2}}$ ${\displaystyle =\sqrt{\Delta {\mathit{L}}_{\mathit{x}}^2+\Delta {\mathit{a}}_{\mathit{x}}^2+\Delta {\mathit{b}}_{\mathit{x}}^2}}$ where _x indicates the number of days of immersion in the staining solution. ΔE₂, ΔE₄, ΔE₆, and ΔE₈ were determined as representing color changes after 2, 4, 6, and 8 days of immersion, respectively. Average ΔE_x values from the test and control surfaces were recorded at each time point.

Change in lightness $L_v^{\ast }$ due to capsule-coffee immersion (ΔL${}_{\mathrm{vx}}^{\mathrm{\ast }}$) was determined according to the following formula: ${\displaystyle \Delta {\mathit{L}}_{\mathit{vx}}={\mathit{L}}_{\mathit{v staining}\left(\mathbf{x days}\right)}-{\mathit{L}}_{\mathit{v baseline}}}$ where _x indicates the number of days of immersion in the capsule-coffee staining solution. ΔL${}_{v2}^{\ast }$, ΔL${}_{v4}^{\ast }$, ΔL${}_{v6}^{\ast }$, and ΔL${}_{v8}^{\ast }$ were determined and represent color changes after 2, 4, 6, and 8 days of immersion, respectively.

Bleaching and color change

After 8 days of staining, teeth were rinsed with deionized water and dried using air jets. An approximately one mm thick layer of 35% hydrogen peroxide in-office bleaching gel (Whiteness HP Blue; FGM, Joinville, Brazil) was applied to both test and control surfaces following the manufacturer’s instructions. The gel was allowed to remain in contact with the enamel surface of teeth for 40 min, after which the gel was suctioned off, and the enamel surfaces were rinsed thoroughly with deionized water.

Color change due to bleaching was calculated using the following formula: ${\displaystyle \Delta E_{\mathrm{bx}}=\sqrt{{\left({\mathit{L}}_{\mathit{bx}}-{\mathit{L}}_{\mathit{staining}\left(8\mathit{days}\right)}\right)}^2+{\left({\mathit{a}}_{\mathit{bx}}-{\mathit{a}}_{\mathit{staining}\left(8\mathit{days}\right)}\right)}^2+{\left({\mathit{b}}_{\mathit{bx}}-{\mathit{b}}_{\mathit{staining}\left(8\mathit{days}\right)}\right)}^2}=\sqrt{\Delta {\mathit{L}}_{\mathit{bx}}^2+\Delta {\mathit{a}}_{\mathit{bx}}^2+{\mathit{b}}_{\mathit{bx}}^2}}$ where b indicates post bleaching color measurements, and x indicates the bleaching round. The CIE L*a*b* parameters and $L_v^{\ast }$ at 8 days of staining were used as the baseline measurement to determine color change after bleaching. Mean $L_{vb1}^{\ast }$ and ΔE_b1 values for the test and control surfaces were recorded. Teeth were stored in deionized water at 37 °C. Both test and control surfaces received a second round of bleaching using 35% hydrogen peroxide in-office bleaching gel (Whiteness HP Blue, FGM, Joinville, Brazil), as described above. CIE L*a*b* color parameters, $L_{vb2}^{\ast }$, and ΔE_b2, were recorded.

Change in lightness $L_{vb}^{\ast }$ due to bleaching (ΔL${}_{\mathrm{vbx}}^{\ast }$) was determined according to the following formula: ${\displaystyle \Delta {\mathit{L}}_{\mathit{vbx}}={\mathit{L}}_{\mathbf{v}\mathit{bx}}-{\mathit{L}}_{\mathit{v staining}\left(8\mathit{days}\right)}}$ ${\displaystyle \Delta \Delta {\mathit{L}}_{\mathit{vb}}={\mathit{L}}_{\mathbf{v}\mathit{b}2}-{\mathit{L}}_{\mathit{vb}1}}$ where _b indicates post-bleaching color measurements and _x indicates the number of bleaching rounds.

Mean values of ΔL_vbx ^∗, Δ ΔL_vb ^∗, and ΔE_bx were statistically analyzed.

Change in the change of color (ΔΔE_b) between first and second bleaching rounds was determined according to the following formulas: ${\displaystyle \Delta \Delta {\mathit{E}}_{\mathit{b}}=\Delta {\mathit{E}}_{\mathit{b}2}-\Delta {\mathit{E}}_{\mathit{b}1}}$ where _b indicates post-bleaching color measurements and _x indicates the number of bleaching rounds. Mean Δ ΔE_b of test and control surfaces were statistically analyzed.

Statistical analysis

The color parameter ΔL${}_v^{\mathrm{\ast }}$ and color change ΔE were statistically analyzed. All the color-related data were collected, tabulated, and analyzed statistically. Statistical analysis was performed using SPSS (version 20). Data organization and graphical representation were conducted using Microsoft Office Excel. The following quantitative variables were described; mean, standard deviation (SD), range (minimum–maximum values), standard error (SE), 95% confidence interval of the mean, and coefficient of variation. A Shapiro–Wilk normality test was used to test all quantitative variables for hypothesis normality to choose the appropriate parametric and non-parametric tests. All tested variables were normally distributed; therefore, parametric tests were used. A general linear mixed model (GLMM) repeated measure ANOVA was conducted for statistical analysis of ΔL ${}_v^{\ast }$ and ΔE changes due to staining over time and changes due to bleaching rounds with a significance level of p <  0.05 (S), while p < 0.001 was considered a highly significant (HS) difference. Multiple pairwise comparisons were conducted using the Bonferroni method. Two-tailed tests were utilized for the analysis of all statistical tests. Independent sample t-test was used for the comparison between every two groups.

Assessment of staining effect on enamel surface lightness

Table 1 shows the $L_v^{\ast }$ color lightness parameter mean values and standard deviation (SD) of the test (received ICON infiltration treatment) and control (no ICON infiltration done) surfaces at baseline and after 2, 4, 6, and 8 days of immersion in the capsule-coffee staining solution.

The gradual decrease in the mean values of the $L_v^{\ast }$ color parameter over time for both test and control enamel surfaces in the study is demonstrated in Fig. 2. Analysis of time as a main effect on the $L_v^{\ast }$ values of surfaces was statistically highly significant (p < 0.001). Both the interaction of time with surfaces under investigation and the type of surface (test vs. control) as a main effect were non-significant (p = 0.47 and p = 0.35, respectively).

Multiple comparisons of the changes in surface lightness (ΔL${}_v^{\ast }$) between the various staining times regardless of the surface type by the Bonferroni method showed highly significant changes when ($L_v^{\ast }$) was compared to baseline values (p < 0.001) and statistically significant differences between 2 days staining and the 8 days staining (p = 0.028). Even though there was a slight decrease in the surface lightness in the test group over the control group, as shown in Fig. 2. However, no statistically significant difference between the surface lightness of test and control surfaces was detected using the Bonferroni method (p = 0.35).

There were no statistically significant differences (p >0.05) in the change in lightness (ΔL_{v staining(xdays)}) relative to lightness at baseline (ΔL_v.baseline) between the two surfaces (ΔL_vx) when analyzed using the independent sample t-test at all time points. The changes in lightness at different time points relative to baseline L_v values are presented in Fig. 3.

Assessment of staining effect on enamel surface change of color

Table 2 shows the change of color (ΔE_x) mean values and standard deviation (SD) of the test (received ICON infiltration treatment) and control (no ICON infiltration done) surfaces after 2, 4, 6, and 8 days of immersion in the capsule-coffee staining solution.

The changes in the mean (ΔE_x) values over time for both test and control enamel surfaces in the study are demonstrated in Fig. 4. Mauchly’s test of sphericity was non-significant, indicating non-violation of the sphericity assumption. Analysis of time as a main effect on the (ΔE_x) values of surfaces was statistically highly significant (p < 0.001). In contrast, time interaction with surfaces under investigation was non-significant (p = 0.7). However, the color change (ΔE_x) when analyzing the type of surface as a main effect was statistically highly significantly higher in the test surfaces than in the control surfaces (p <  0.001). Pairwise comparison between test and control surfaces using the Bonferroni method revealed a highly significant difference between the surfaces in response to the staining solution (p < 0.001).

Multiple comparisons of color change between the time points of immersion in the capsule-coffee solution, regardless of the surface type, were determined with the Bonferroni correction method at a significance level of p < 0.05. There was a statistically significant increase in ΔE_x between the eighth and second days and between the fourth and eighth days of staining (p = 0.03 and p = 0.01, respectively).

Assessment of bleaching effect on lightness of capsule-coffee-stained enamel surfaces

Table 3 shows the mean lightness and change of color values for the test and control surfaces with each bleaching round.

General linear model repeated measure mixed design ANOVA was used to analyze the effect of bleaching rounds on the lightness of test and control surfaces. The gradual increase in the mean values of the $L_{vbx}^{\ast }$ color parameter with bleaching round for both test and control enamel surfaces is demonstrated in Fig. 5. Mauchly’s test of sphericity was significant, indicating a violation of sphericity assumption. Therefore, Greenhouse-Geisser correction was applied. Bleaching round number as a main effect on the color parameter $L_v^{\ast }$ was statistically highly significant (p < 0.001). The mean color parameter $L_{vb2}^{\ast }$ was significantly higher compared to the mean $L_{vstaining\left(8days\right)}^{\ast }$ value and the mean $L_{vb1}^{\ast }$ value in both test and control surfaces (p < 0.001). In contrast, the interaction of bleaching round and type of surfaces was statistically non-significant (p = 0.69).

Multiple comparisons with Bonferroni revealed a statistically highly significant higher $L_{vb1}^{\ast }$ and $L_{vb2}^{\ast }$ than L_{vstaining(8days)∗} (p < 0.001). Mean $L_{vb1}^{\ast }$ was statistically highly significantly higher than mean $L_{vb2}^{\ast }$ (p < 0.001). However, the difference between $L_{vb1}^{\ast }$ and $L_{vb2}^{\ast }$ mean values was not statistically significantly different between the test and control surfaces (p = 0.4). The mean lightness $L_{vstaining\left(8\right)days}^{\ast }$) was used as a baseline to compare the change in lightness between the test and control surfaces. Independent samples t-test for comparing the change in the lightness ΔΔL${}_{vb}^{\mathrm{\ast }}$, ΔL_vb1*, and ΔL${}_{vb2}^{\ast }$ between test and control surfaces revealed no statistically significant differences, as demonstrated in Fig. 6.

Assessment of bleaching effect on capsule-coffee-stained enamel surfaces change of color

Figure 7 shows the increase in change of color (ΔE_bx) mean values of the test (received ICON infiltration treatment) and control (no ICON infiltration done) surfaces after the first and second bleaching rounds.

All variables were normally distributed, and parametric tests were used. General linear model repeated measure mixed design ANOVA was used to analyze the effect of bleaching rounds on the color change of test and control surfaces. Analysis of bleaching round as a main effect on the (ΔE_b) values of surfaces was statistically highly significant (p <0.001). In contrast, the interaction of bleaching round with surfaces under investigation was non-significant (p = 0.49). However, the color change (ΔE_b) when analyzing the type of surface as a main effect was statistically significantly higher in the test surfaces than in the control surfaces (p = 0.01). Pairwise comparison by Bonferroni between the first and second bleaching rounds revealed a highly significant increase in the change of color of tooth structure (p < 0.001). Pairwise comparison by Bonferroni revealed a statistically significant higher change of color in test surfaces compared to control surfaces (p = 0.01). Independent samples t-test revealed a statistically significantly higher (ΔΔE_b) in test surfaces compared to control surfaces (p = 0.02), as detailed in Table 4.