Hyposalivation is mostly associated with various syndromes, diabetes, vitamin deficiency, menopause, salivary gland hypofunction due to inflammation, infection, the use of various drugs, or radiotherapy. Since incidence of head-and-neck cancer  and amount of drug revenue increases with age, primarily older patients suffer from hyposalivation . Furthermore, periodontal diseases occur predominantly in elderly patients and often lead to gingival recessions . Additionally, progressive attrition/abrasion during prolonged utilization of the teeth causes dentin exposure . Since dentin is not as resistant as enamel to acid exposure, earlier and more severe demineralization can be expected . Thus, carious lesions located at the cervix of the teeth develop easily during radiation therapy [40, 41]. Therefore, in the present study, both enamel and dentin were assessed.
The teeth used in the present study were obtained from newly slaughtered cattle. One of the major disadvantages of human teeth compared to bovine teeth is the occurrence of defects such as initial carious lesions. It can be assumed that bovine teeth are rarely influenced by external factors (e.g., acids, fluorides), and, thus, the biological spread of bovine teeth is relatively small [25, 42]. In contrast, human teeth usually exhibit an inconsistent age and source, which might result in a variable composition that leads to larger variations in the test response. Due to their similar chemical composition, general availability and large size, bovine teeth are a suitable substitute for human dental hard tissues in in situ examinations, and a more uniform reaction can be expected using these specimens . Moreover, lesion formation in bovine enamel is very similar to the demineralization process observed in human enamel . However, bovine enamel is generally considered a more porous material compared to human enamel, and, thus, bovine enamel may be susceptible to accelerated demineralization . For in situ studies into caries prevention, sterilization of tooth specimens is essential. Sterilization of enamel and dentin using ethylene oxide has not been considered to induce relevant effects with in situ studies of de- and remineralization .
As a precursor of dental caries, initial subsurface lesions appear in many patients . These lesions were simulated in the present study by the artificial demineralization (i.e., subsurface lesion) of half of the enamel specimens before in situ incorporation. Furthermore, an investigation of the remineralization processes requires the utilization of artificially demineralized enamel specimens that are adjacent to sound specimens. Due to the possibility of disintegration during the in situ exposure, the dentin specimens were not artificially demineralized. Four specimens that were obtained from one bovine tooth were embedded into the removable denture of one patient during each experimental phase (Figures 3a and 4). This procedure ensured that the specimens from one tooth were only used for one patient and that one bovine individual-to-one patient regimen was maintained. Furthermore, the enamel and dentin specimens correlated with respect to their source during each test phase, and this permitted a comparison of the in situ effects of demineralization and remineralization between dentin and enamel.
To investigate the effects of the study medication on the specimen surfaces after a wearing period of five and seven weeks, the surfaces were divided as described (Figure 3). By covering the respective surface areas with acid-resistant nail varnish and a flowable composite, the sound and artificially demineralized regions were maintained during the entire in situ investigation for the microradiographic analyses. Additionally, the coverage of the effect fields (E1; Figure 3) after five weeks ensured that this surface areas were not affected by additional de- and remineralizing effects. This method allowed the determination of distinguishable test areas.
Specimens attached on natural teeth could interfere with mastication and articulation, and, therefore, specimens were fixed in removable dentures. However, it has to be taken into account that wearing a dental prosthesis (full or partial) can significantly affect the composition of the resident oral microflora, and, following, the results might be influenced by this factor . Therefore, this factor should be kept in mind when interpreting the upcoming data. Unlike former in vitro studies , the coverage areas were additionally fixed with flowable composite to ensure the ability to withstand mechanical forces during the study period. Up to now no visible gaps between the specimen surfaces and the coverings have been recognized. Moreover, the additional test field coverage with composite resin seems to be advantageous for protection of surface areas. So far no coverage losses could be observed.
The intention of the chosen in situ model was to mimic the natural caries process and provide clinically relevant information in a relatively short period of time without causing irreversible tissue changes in the panelists' dentition. Generally, clinical caries trials are limited to investigations by a dental explorer, and to using radiographs to identify and validate demineralizations at sites usually not visible directly. Thus, the caries process is determined at a relatively late stage and patients are exposed to X-rays resulting in an increasing radiation exposure. In contrast, the in situ model presented here offers the integration of transversal microradiography (TMR) as a basic science analytical technique. TMR offers high sensitivity and ensures waiving radiation exposure for panelists . Generally, various experimental methods are available for analyzing subsurface lesions of bovine enamel and dentin specimens. These include transversal microradiography , polarized microscopy , microhardness testing [50, 51], electric caries monitoring , transversal wavelength-independent microradiography , optical coherence tomography , and scanning electron microscopy [55, 56]. With the exception of transversal microradiography (TMR), all mentioned technologies reveal some shortcomings with regard to accuracy when specimens are analyzed according to the mentioned parameters mineral loss and lesion depth. In contrast, TMR allows a direct measurement of the longitudinal mineral distribution as a profile in a subsurface lesion and has long been established and recognized as a gold standard for analyzing mineral content changes over time . Consequently, TMR is considered a surrogate outcome measure that is directly impacted by the intervention. Thus, microradiographic outcome was selected as a surrogate endpoint in the present study.
Since fluorides are commonly found in foodstuff such as salt, fish, and mineral water, the study implementation required a renunciation of several foods in terms of a fluoride-free diet. This adjustment was only established for the (short) in situ period, and thus, might represent a lower personal burden for the patients compared with long-term clinical trials . Altogether, an in situ model might better conform with the patients' requirements, and these always have to be weighed against the background of protracted cancer-related therapies.
Saliva substitutes such as the widespread Glandosane, which contains acids and a relatively low amount of calcium and phosphates (and, therefore, is undersaturated with respect to calcium phosphates), have demonstrated demineralizing properties in vitro [1, 25, 26, 33]. From the perspective of dental medicine, neutral or even remineralizing effects of artificial salivas would seem preferable . Glandosane has a pH value of 5.3, whereas the pH value of the modified SN is approximately 5.98. Due to the strict permission criteria for the modification of existing saliva substitutes (German Act of Medical Devices), a higher pH value could not be adjusted for SN. Even solutions with low pH values or a large amount of titratable acids do not cause demineralization in dental hard tissues, if an appropriate saturation is used with respect to relevant calcium phosphates via a reduction of the apatite solubility . In previous in vitro studies, supersaturated solutions such as modified SN was observed to remineralize enamel and dentin over a period of five weeks, and this has been attributed to the addition of calcium, phosphates and fluorides . Human saliva can be considered a supersaturated solution with respect to hydroxyapatite (HA), and has a mean pH value of 6.7 under physiological conditions . In addition to flushing purposes, saliva also acts as a buffer. Bicarbonate and phosphate buffers in particular allow for neutralization of organic acids via diffusion through the dental plaque. The saliva of patients who suffer from hyposalivation after radiotherapy reveals a decreased pH value (i.e., acidic pH value), which promotes the demineralizing processes of dental hard tissues.
The primary component of enamel is HA. The solubility of a substance is characterized by its solubility product (K). For a solution such as saliva, the ion product (IP) is based mainly on the calcium, phosphate, and hydroxyl concentrations. The thermodynamic driving force of enamel de- or remineralization is a function of the degree of saturation with respect to HA (DSHA) [59, 60], and the degree of saturation of a solution (DS) can be defined by dividing the product of the ion activities (IP) by K, and exposing this result to the number of ions in the formula unit . If DS is greater than 1, the solution is supersaturated with respect to the calculated mineral phase; if DS is less than 1, the solution is undersaturated; in case of DS equals to 1, the solution is saturated, and no net dissolution or precipitation occurs. For aqueous solutions, the degree of saturation with respect to apatites [HA, octacalcium phosphate (OCP), dicalcium phosphate dihydrate (DCPD), and calcium fluoride (CaF2)] can be calculated, if the pH and the concentrations of certain ions are known . In a previous study  that determined the composition of the SN used in the present one, IONPRODUCT, which was developed by Peter Shellis , was used. IONPRODUCT is a computer program that calculates the ionic activity of products and the degree of solution saturation with respect to biologically relevant calcium phosphates. The input values are the concentrations of the constituent ions, pH, temperature, and atmospheric pressure. A database contains the required solubility products of each of the minerals and the relevant dissociation constants. The program can be used to determine the DS at different pH values for a given calcium concentration.
Because patient acceptance of slightly acidic saliva substitutes is higher compared to products with neutral pH values, the current pH values of SN as well as of G might support patient compliance [27, 33]. In the present study, an experimental period of seven weeks was selected. The present test field design refers to former in vitro studies exhibiting effects after an investigation of five weeks [33, 48, 63]. Since artificial saliva was clinically administered ad libitum, no maximum daily dose could be assumed. These procedures, employed in previous in vitro studies, simulated an extremely intense contact between saliva substitutes and specimens during an experimental time of up to five weeks, and, hitherto, these conditions could not be replicated with clinical settings and conditions. Therefore, in comparison to former in vitro studies a longer examination time of seven weeks was selected.
It is important to note that unlike the conditions observed in vitro, a total absence of saliva is rarely observed in clinical situations. Due to a variable presence of saliva, compositional changes , and the decreased pH values detected in the present in situ study, slightly altered results should be considered. Despite the omission of factors such as nutrition, previous studies have noted the development of erosions on the specimen surface even under strictly controlled in vitro conditions. Variability in patient diets, especially the consumption of sour drinks/food, might lead to the development of erosion. To reduce the influence of diet on mineralization, the patients were provided with nutritional counseling. In the present study, the patients were asked to restrain from the consumption of any sour foodstuffs during the test period to avoid any bias. Furthermore, panelists' were asked to abstain from any additional intake of fluorides from food or any form of fluoride-containing oral hygiene products. Therefore, advisory information and nutrition counseling regarding fluoride uptake was conducted for all patients to illustrate the need of a fluoride-free diet. In particular, the panelists' were required to abstain from fluoride-containing toothpaste; instead, fluoride-free toothpaste was administered.
To receive the patients' personal assessment concerning their well-being and usage of salivary substitutes, questionnaires were handed out as described before (Figure 2). The simplicity of the questionnaires which was enabled by referring to the German school mark scale (1-6; 1 = very good, 6 = poor), being familiar to the patients, allowed for straightforward answers to the questions and provided a fast and simple evaluation by the investigators. The evaluated personal well-being and the self-assessed intensity of dry mouth before and after the use of each saliva substitute can be compared with each other and with the collected data containing measured values of the amount of natural salivation and saliva substitutes consumption during each trial. In addition, the questionnaires provide important information on the general acceptance of the use of saliva substitutes in terms of taste and duration.
Optimal oral hygiene and regular oral screening is necessary in individuals with hyposalivation due to an increased risk for caries and periodontal diseases. Saliva substitutes should not only relieve the symptoms of oral dryness, but should also provide protection against demineralization. In addition to other in situ studies investigating caries formation and the protective effect of fluorides, the present study is the first clinical trial to evaluate the effects of saliva substitutes on dental hard tissues in situ. It is difficult to control human behavior, and, therefore, the data which will be obtained herein might differ from those generated in previous in vitro studies. Despite these considerations, the present study has the potential to provide clinical data that can improve the effects of salivary substitutes and their acceptance among patients. Thus, the results of the present study should be important for future treatment guidelines for patients suffering from hyposalivation.