Three different toothpastes were investigated with regard to their remineralization properties of artificially demineralized human enamel:

Karex (Kurt Wolff, Bielefeld, Germany)

Elmex (GABA, Hamburg, Germany)

Ajona (Liebe, Leinfelden-Echterdingen, Germany)

Karex is a fluoride-free toothpaste with the active ingredient HAP. Elmex is a fluoride-containing toothpaste with amine fluoride. Ajona is a toothpaste that contains neither fluoride nor HAP and is therefore considered a control toothpaste (see Table 1).

Preparation of the samples

For the present in vitro study, 19 intact human wisdom teeth were used, which were surgically removed before tooth eruption. Surgical removal of all teeth was performed at the Department of Oral and Maxillofacial Surgery at Münster University Hospital (Germany). All patients were of legal age, were fully informed about the planned study, and gave written informed consent to participate. All human specimens were handled strictly according to the “Declaration of Helsinki” (Local ethics protocol code 2021‐608‐f‐N).

Immediately after surgery, teeth were stored for one hour in a 0.05% thymol solution for disinfection, rinsed thoroughly with distilled water, and stored in distilled water at 5 °C in a refrigerator until further processing.

To define exactly reproducible enamel surfaces on the extracted teeth, 3 vestibular and 4 oral grooves were made in the mesio-distal direction with a diamond cut-off wheel (946; shank 104-HP, size 220, Brasseler, Lemgo, Germany). Subsequently, a further groove was created vestibular-cervically with a surgical round-headed rotating instrument (H71; socket 104-HP, size 008, Brasseler). This defined 4 enamel surfaces next to each other on each tooth, which could be used for demineralization and remineralization and will be referred to as analysis surfaces in the following. The analysis surfaces had a size of 2 mm in width and 8 mm in length. In addition, there were two control surfaces, each of which was located between the analysis surfaces. The control surfaces had a size of 1 mm in width and 8 mm in length.

The two control surfaces and the outer marginal occlusal and cervical areas were covered with clear nail varnish (Superstay Forever Strong 7 Days crystal clear 25; Maybelline, New York, USA) and dried for 60 s. The surfaces covered with nail varnish (control surfaces) were thus protected from contact with acid and toothpaste (chemically and mechanically) during the further course of the test. (Fig. 1) Accordingly, neither de- nor remineralization processes took place on these surfaces. This was verified in preliminary tests.

Fig. 1

Systematic drawing of how the respective enamel surfaces on a tooth were demineralized and remineralized with the corresponding toothpastes in order to obtain the individual samples for SEM evaluation. (large image: cross-section with surfaces S0—S4, small image: schematic view from occlusal with vertical cuts to obtain the SEM specimens.)

Demineralization of the samples

All teeth were then placed in 90% lactic acid (pH 3; Pharmacy of the University Hospital, Münster, Germany). After 8 h, the teeth were removed from the lactic acid, rinsed with distilled water (Pharmacy of the University Hospital) and dried.

The following analysis surfaces (S) were defined on the teeth in each case.

S0: control surface

S1: demineralization with lactic acid only

S2: demineralization with lactic acid + remineralization with Karex

S3: demineralization with lactic acid + remineralization with Elmex

S4: demineralization with lactic acid + remineralization with Ajona

Remineralization of the samples

For the examination of Karex, the first analysis surface S1 was covered with clear nail varnish after acid treatment. In contrast, the analysis surfaces S3 and S4 were reliably protected from access of toothpaste, and thus mechanical and/or chemical influences, with liquid rubber dam (Easydam; DeltaMed, Friedberg, Germany). Subsequently, Karex was applied to the analysis surface S2 and brushed for 15 s with a manual toothbrush (Sensitive Super Soft; Dontodent, Karlsruhe, Germany) at a brushing load of 200 g. The brushing load was checked in a preliminary test using a scale. Afterwards, Karex remained on the tooth surface for further 105 s, resulting in a total contact time of 2 min. After thorough rinsing of the tooth with distilled water and appropriate drying, the analysis surface S2 was covered with clear nail varnish and thus fixed.

Easydam was then carefully removed from the analysis surface S3 using a scalpel. In analogy to the treatment with Karex, S3 was brushed with Elmex and a manual toothbrush for 15 s with circular movements at a brushing load of 200 g. The toothpaste remained on the tooth for further 105 s. After rinsing with distilled water and drying, the analysis surface S3 was also covered with clear nail varnish and fixed. Easydam was then carefully removed from analysis surface S4 using a scalpel. In analogy to the treatment with Karex and Elmex, S4 was brushed with Ajona and a manual toothbrush for 15 s with circular movements at a brushing load of 200 g. The toothpaste remained on the tooth for further 105 s. After rinsing with distilled water and drying, the analysis surface S4 was covered with clear nail varnish and fixed.

Scanning electron microscope evaluation

For further scanning electron microscopy (SEM) analysis of the surfaces, the 19 teeth were cut using a saw (WOCO 50; Uniprec, Clausthal-Zellerfeld, Germany). For this purpose, the clinical tooth crown was first separated from the root by a horizontal cut at the enamel-cement junction. Vertical cuts in the vestibular-oral and occlusal-cervical directions, respectively, were used to obtain three specimens from one tooth, resulting in a total of 57 specimens. The cuts were made perpendicular to the demineralized surfaces in identical distances. These 57 specimens were degreased with acetone (density 0.792 g/cm3; Meffert, Bremgarten, Switzerland) and dried for 24 h in a drying oven (T5028; Heraeus, Hannover, Germany). This was followed by embedding in acrylic (TransOptic Compression Mounting Compound; Buehler, Lake Bluff, USA). The resulting specimens were smoothed with diamond-coated abrasive paper (18 μm grit; Waterproof Silicon Carbide Paper, Struers, Willich, Germany) in a grinder (MetaServ 250 Grinder-Polisher; Buehler, Lake Bluff, USA), sprayed with diamond spray (DP-Spray P; Struers, Willich, Germany), and polished to a mirror finish in a polishing machine (PM5; Logitech, Glasgow, United Kingdom). The finished samples were then coated with carbon 25 μm thick in a carbon coater (Carbon coater 208 carbon, Cressington, Dortmund, Germany).

The four analysis surfaces were photographed at different magnifications (150 × , 300 × , 500 × , and 1000 ×) using a scanning electron microscope (JSM-6510LV, JEOL, Zurich, Switzerland). In the next step, the 300 × magnification images were transferred to imaging software (ImageJ; Wayne Rasband, NIH, Bethesda, USA) and subjected to pixel analysis. The images of 300 × magnification proved to be particularly suitable in preliminary tests, since here the largest possible analysis area was given with sufficient visibility of the fusion prisms within the image section. Each image was inserted individually into ImageJ and first calibrated using uniform scale at a height of 50 μm in the image. The image was then rotated so that the enamel surface was aligned parallel or perpendicular to the image frame. A “region of interest” was determined by placing a rectangle along the enamel surface. The depth of the rectangle was fixed at 100 μm. The length of the rectangle was chosen to be as large as possible to define the largest representative area. Finally, the Crop function was used to crop out the “region of interest” from the original image. Using the Threshold function, the percentage of demineralization was calculated in the precisely defined, reproducible image section. For this purpose, a black and white image was generated in which the grey pixels represent the intact enamel (enamel prisms) and the black pixels represent the defective enamel (corresponding to demineralization by the lactic acid). The higher the percentage grey, the better the remineralization and the lower the demineralization. The higher the percentage black, the more demineralization and less remineralization. (Figs. 2 and  3).

Fig. 2

SEM images of a representative specimen before and after demineralization of 8 h. Grey areas represent the intact enamel (enamel prism) and the black areas the defective enamel (corresponding to demineralization by the lactic acid). (a = control group without treatment; b = surface after 8 h demineralization with lactic acid)

Fig. 3

SEM images of a representative specimen in the different toothpaste groups after demineralization of 8 h and remineralization (brushing with the appropriate toothpastes). Grey areas represent the intact enamel (enamel prism) and the black areas the defective enamel (corresponding to demineralization by the lactic acid). (a = control group without treatment; b = Karex; c = Elmex; d = Ajona)

Further demineralization of the samples

In order to check whether the treatment of the enamel specimens with the respective toothpaste protects against a new acid attack, the tooth specimens that have been already evaluated in the SEM were bedded out of the acrylic with the aid of acetone and thoroughly rinsed with distilled water. Afterwards, all tooth specimens were again immersed in 90% lactic acid (pH 3) for 2 h, thoroughly rinsed with distilled water, re-embedded in acrylic and prepared for SEM analysis in analogy to the first series of experiments. An identical analysis to the first series of experiments was performed using SEM. (Fig. 4).

Fig. 4

SEM images of a representative specimen in the different toothpaste groups after demineralization of 8 h, remineralization (brushing with the appropriate toothpastes) and new demineralization of 2 h. Grey areas represent the intact enamel (enamel prism) and the black areas the defective enamel (corresponding to demineralization by the lactic acid). (a = control group without treatment; b = Karex; c = Elmex; d = Ajona)

A total of 513 images were analyzed. Of these, 5 different surfaces were analyzed in the 1st test series with 57 images each and 4 different surfaces (minus the control surface) with 57 images each in the 2nd test series. All images generated by ImageJ (n = 513) were evaluated independently by two expert observers (LG, AR). Only minor differences were found. In case of discrepancies (n = 15 ≙ 2.9%), the affected images were evaluated by another independent expert observer (TD) in a third round. All participants were previously calibrated for evaluation. All samples were blinded, i.e. the evaluators did not know how the samples had been treated beforehand.

Statistical evaluation

Data were analysed using SPSS software (IBM SPSS Statistics 27, Armonk, USA). Distribution of the data was evaluated using the Kolmogorov–Smirnov test. Normally distributed data (demineralization within the previously defined image section of the respective analysis areas S0 to S4) were further analysed using ANOVA (Analysis of Variance) with post-hoc Scheffé test. For analysis of non-normally distributed data, the Kruskal–Wallis test was applied. Pairwise comparisons were performed using the Mann–Whitney U-test. The level of significance was set at p < 0.05.