Introduction

Over the years, the erosive potential of soft and fruit drinks on the dentition has been the focus of many investigations (Dugmore and Rock 2004; Tahmassebi et al. 2006; Blacker et al. 2011; Blacker and Chadwick 2013). The acid content of these drinks lead to erosion and the sugar content, metabolised by plaque microorganisms to generate organic acids will bring about dental caries. While the natural biological process in the mouth can neutralise a single acid attack, frequent use will lead to remarkable detrimental effects (Lussi et al. 2004).

In the UK, consumption of shop-purchased smoothies has risen dramatically from 6 million litres in 2001 (Mercer 2007) to 51 million litres in 2010. It is said that the true level of smoothie consumption is not known, as these figures exclude consumption of homemade drinks, as well as, those bought from cafes, coffee shops and juice bars (Blacker et al. 2011). The promotion of healthy eating has been one of the driving forces behind the continuing exceptional rate of growth of this industry.

Smoothies are made mainly from pureed fruits, and they are considered healthy owing to their high level of antioxidants, fibres, and vitamins; however, the consumption of smoothies is also viewed as being potentially detrimental to health owing to the high sugar and acid content of such drinks. Considering the drink from a dental perspective, demineralisation may occur as a direct result of consumption, therefore leading to dental erosion and dental caries.

In recent years, much of the dietary advices have been encouraging an increase in the intake of eating fresh fruit and vegetables, and apparently drinking smoothies is a practical and an easy method in gaining the benefits of fruits. The 5-a-day campaign has also helped to encourage intake of this drink. The literature published on the subject provides only very little data concerning the impacts of smoothies consumption on dental erosion. More research needs to be conducted to produce scientific proof of smoothie drinks to guide professionals in giving advice and educating the public about the consequences of consuming these drinks.

Materials and methods

The present study was carried out as a pilot study and therefore power calculations were not performed. Sample number was chosen based on studies that had used a similar method but with a different sample of drinks (Abdullah 2009). There were 50 enamel slabs included in the study, with 10 in each subgroup.

Enamel slabs

Enamel samples were from sound premolar teeth extracted for orthodontic reasons only. Tissue Bank approval was sought to obtain the teeth. Teeth were stored in distilled water and 0.1 % thymol (Sigma Aldrich). Crowns were sectioned using a water cooled, Diamond Wire Saw, cutting machine (Well®Walter EBNER, CH-2400 Le Loche). The buccal and palatal surfaces of each crown were removed, and the slabs were prepared according to the relevant standard operating procedure (about 2 mm wide, 2 mm length, and 2 mm depth).

The enamel slabs were mounted in circular resin blocks of 3 mm thickness. To ensure flatness of the blocks and to make sure no resin covered the enamel surface, a grinding machine was used. Fine grit abrasive papers were used, from 600 grade, to 1,200 grade and 2,500 grade (wet or dry paper, 3 M). Minimal enamel was removed. This was easily achieved because the resin blocks were held in rectangular steel blocks which were slightly less than 3 mm deep due to wear. The blocks were then cleaned with methanol to remove any remnants of abrasive paper. Surfaces were polished with 5 and 1 μm alumina paste. Finally, the enamel blocks were rinsed with DD water. At all times, the slabs were kept moist in DD water in micro-centrifuge tubes and left at room temperature. This was to prevent dehydration of the enamel. Profiles of the resin blocks were assessed using a surface profilometer (Scantron ProScan 2000).

To achieve flatness of the enamel samples, it was ensured that average height to average depth (Rz) range was within 1.0 μm. Enamel slabs that were confirmed to be flat were tested with a surface microhardness test. A computer-aided Duramin Indenter Machine (Struers A/S, DK 26-10, Denmark) was used. Indentations were made using a Knoop diamond under 100 g load for 30 s. This was performed with care, and the middle area was covered with tape in order not to touch it. The length of indenter penetration was measured using an image analysis system. Three indentations were performed and measured, and then the mean was determined.

Slabs with appropriate microhardness (range of 60–70 μm) were accepted to be included in the study. They were then covered with nail varnish except for the middle area, where a window of 1 mm × 2 mm was left exposed. Those that were out of range were rejected from the study. Enamel slabs were put in a bag and randomly picked, then placed in a tray with 10 holes. They were secured in position using adhesive wax. The trays were then put in a bag and randomly picked out and labelled into five groups, ready to go through the cycling process. It was not possible to achieve blindness for this study as all the drink samples could easily be differentiated.

Materials

  1. (a)

    Strawberries and bananas smoothie (Innocent® pure fruit smoothie, London)

  2. (b)

    Mangoes and passion fruit smoothie (Innocent® pure fruit smoothie, London)

  3. (c)

    Diet Coke® (Coca Cola Company, USA)

  4. (d)

    Distilled de-ionised water (negative control)

  5. (e)

    Citric acid 0.3 % (positive control)

Testing the drinks for pH and titratable acidity

Using a pH meter (VWR International Orion, Orion research, UK) the pH of each of the sample drinks (100 ml) were checked immediately on opening. Tests were carried out at room temperature with a magnetic stirrer placed in the beaker to mix the sample drinks well at 875 rpm. Between use, the electrode was rinsed thoroughly to avoid cross-contamination. Every day, pH was calibrated using standard buffers of pH 4.0 and 7.0.

To test the titratable acidity of the sample material, 100 ml of each drink was put in a beaker with a magnetic stirrer continuously moving at the speed on 875 rpm throughout the test. pH value was noted and then 0.1 μmol sodium hydroxide (NaOH) solution was gradually pipetted until the pH of the sample drink reached 7.0. The measurement was performed in triplicate and an average value was calculated. Temperature of the drinks was at about 21 °C.

Protocol/regime for pH cycling

The trays containing the 50 enamel slabs were randomly divided into one of the 5 intervention groups, 10 enamel slabs per each group. Slabs were immersed under static conditions for 2 min each time, five times daily in fresh 50-ml aliquots of sample drinks for 21 days (Fig. 1). A gap of 60 min was left between immersions, where slabs were placed in daytime saliva. Overnight, slabs were kept in night time saliva. All this was carried out at room temperature. The two smoothies and the Diet Coke used during the immersion were used chilled (around 15 °C) and the DD water and the temperature for 0.3 % citric acid as per previous studies on erosion was kept at room temperature (21 °C); whereas as the smoothie drinks are always served chilled and therefore they were kept at a temperature of around 15 °C.

Fig. 1
figure 1

Flowchart representing 21-day regimen

Full size image

Before and after immersion with the sample drinks, the slabs were rinsed with DD water. Fresh artificial saliva was changed after each immersion: 1 h before starting the pH cycling regime and 1 h after the final immersion.

Results

pH and titratable acidity

Sample materials were analysed for their pH and titratable acidity. Table 1 shows that the smoothies were acidic in nature. Diet Coke had the lowest pH level at 2.95 followed by SB smoothie at 3.59 and MP smoothie at 3.73. Both smoothies had a high titratable acidity. The amount of sodium hydroxide (NaOH) needed to neutralise the smoothies were 3.5–4 times more than the amount needed for Diet Coke and citric acid 0.3 %.

Table 1 Properties of sample materials
Full size table

Pearson’s Correlation coefficient was determined to observe associations between surface loss with pH and titratable acidity. As shown in Table 2 there was a positive correlation r = 0.59 with titratable acidity which was statistically significant (p < 0.05). However, a value of 0.59 is considered only a fair/medium association.

Table 2 Result of Pearson’s correlation between surface loss and pH as well titratable acidity
Full size table

Test of normality of the data

Shapiro–Wilk test was used to check the normality of the data. This test showed that all groups had p > 0.05, thereby the measurements from this study were of normal distribution. Therefore, parametric tests could be used for all groups (Table 3).

Table 3 Shapiro Wilk test result for all groups
Full size table

Figure 2, shows the distribution of the enamel surface loss by separate sample materials. MP smoothie had caused the largest change in surface loss, followed by Diet Coke and SB smoothie. DD water which was the negative control for this study showed very minimal volume loss.

Fig. 2
figure 2

Distribution of surface loss of all sample groups

Full size image

Differences between sample groups

The differences between each group were compared. Multiple comparisons were conducted with Bonferroni correction to identify where the differences lay. Table 4 shows the result of independent samples test of SB smoothie or MP smoothie against other sample materials.

Table 4 Results comparing mean differences of surface loss experienced by control and test groups using independent samples t tests
Full size table

Comparing smoothies against other sample materials

Comparing SB smoothie against DD water, p value was lower than 0.01 which was the level of significance after Bonferroni correction. Therefore, the difference was statistically significant. Confidence intervals was from 1.82 to 3.14 and did not include zero. In comparison with water there was significantly more surface loss by SB smoothie.

When comparing Diet Coke and SB smoothie, the p value was 0.29, indicating there was no significant difference between the two groups at 5 % level. The mean difference in surface loss was 0.57 μm with a 95 % confidence interval of −1.66 to 0.53. As this confidence interval includes zero, we could confidently state that there was no difference between SB smoothie and Diet coke. No significant difference was seen when comparing SB smoothie with citric acid as well, which was the positive control.

On comparing MP smoothie to DD, MP smoothies caused significantly more surface loss after the erosive challenge. There were also significant differences when comparing MP smoothie with SB smoothie and citric acid 0.3 %. However no statistical significant difference was found when comparing MP to Diet coke, the p value was >0.01 and therefore the enamel loss by MP smoothie was not significantly different to the surface loss attained by erosive challenges with Diet Coke.

Reproducibility

A sample of 10 enamel slabs were randomly selected and re-measured. The differences between the two readings were plotted to the average of the two readings in the Bland–Altman Plot. Mean of the differences or the bias was 0.15 μm. Bias was considered quite small, indicating good agreement between the readings. The 95 % limits of agreement (−0.09, 0.398) included 0, interpreting that the measurements were likely to agree on certain enamel slabs.

Discussion

Smoothies are a relatively new type of soft drinks compared with fruit juices, however, their sales have soared in recent years. A survey has shown that 37 % of consumers occasionally or regularly drink smoothies (YouGov 2008). In the UK, “Innocent” is the market leader in smoothies where they claimed their smoothies contained 100 % fruits and did not contain sugar, water or preservatives. The Innocent smoothies contain 10–18 % fibres and count as two of the 5-a-day portion (Ruxton 2008). Innocent’s top selling smoothies are the ‘strawberries and bananas’ and the ‘mangoes and passion fruit’ which were the test materials used for this research.

Pure fruit and vegetable juice offer similar health benefit as having whole fruits and vegetables (Ruxton et al. 2006). Undoubtedly, their content of antioxidants, polyphenol and fibres are good for health. On the other hand, health concerns from drinking smoothies include the high energy and sugar content as well as the effect of caries risk from non-milk extrinsic (NME) sugars present in the juices/smoothies. Both concerns were shown to have no higher than the impact and risks of having two portions of whole fruits (Ruxton 2008; Hussein et al. 1996; Beighton et al. 2004). Smoothies contain acids from fruits, and theoretically this may lead to dental erosion. There are only limited studies in the literature relating smoothies to erosion (Blacker and Chadwick 2013).

Carbonated drinks dominate the sales of soft drinks in the UK. In 2009, there was consumption of 98.3 l per person. Coca Cola® is a widely known example of carbonated soft drink produced by the Coca Cola Company, and the Diet Coke® is a popular variation of it. The reason Diet Coke was included in this study was because the consumption of this drink has increased among adolescents who have become more weight conscious (Mintel 2008). The association between erosion and Coke is so well established, that it has been used to create artificial erosive lesions in many studies (Amaechi et al. 1999; Hooper et al. 2007).

In the current study, an in vitro model was used as it has several advantages over in vivo studies. The in vitro model allows for several experimental variables to be controlled as well as flexibility of the study design (Koulourides and Chien 1992). In vitro studies have low operation costs (White 1992) and are highly controlled allowing single experimental parameters to be investigated (ten Cate 1994). In addition, highly sensitive detection and analytic techniques, as used in the present study, can be applied to measure the TSL.

On the other hand, the in vitro model does have some limitations. Dental tissue is non-vital and thus cannot reproduce biological responses to erosion. Moreover, the specific interactions of dental hard tissue and saliva, as well as the relationship between sample surface area and fluctuations in the volume and composition of oral fluids cannot be adequately simulated in vitro (ten Cate 1990). Also other factors such as rate of consumption and manner of swallowing, which would have influence on enamel erosion cannot be duplicated in the in vitro model.

Different studies have immersed teeth in various types of acidic challenges and using different time durations, usually at prolonged periods of time. Therefore, their results were exaggerated as there was no modifying influence of saliva. Amaechi et al. (1999) introduced a model to produce dental erosion lesions using a simple in vitro technique. Lesions were produced by cycling teeth in aliquots of 20 ml of orange juice, at regular intervals six times per day (5 min every time). The six times of exposure was to simulate drinking at breakfast, midday, lunch, late afternoon, dinner and bedtime. Immersion carried out done at room temperature and the bovine teeth were kept in artificial saliva in between immersion and at night. The experiment was carried out in such way for 24 days, giving a total of 12 h exposure (Amaechi et al. 1999). The 5 min exposure was chosen because there was an observation that pH of saliva and its calcium phosphate saturation returned back to baseline after a 5 min rinse with citric acid rinse (Bashir and Lagerlof 1996). This study concluded that the technique used was suitable to mimic the conditions in vivo and could be used in looking at various parameters on dental erosion.

The method used for this study was from a protocol developed at the University of Leeds as used in a study by Abdullah (2009). The protocol was slightly modified from the method used by Amaechi et al. (1999). The six times dipping of 5 min emersion was thought to be an overestimation of the real life situation. Therefore, a protocol consisting of five times of dipping with 2 min of immersion each time was used. Two types of artificial saliva were used. The daytime saliva was used for remineralisation and the night time artificial saliva was used for maintaining the minerals. The slabs were immersed in 50-ml fruit solution each time. This amount was chosen as it is believed that most consumers will drink about one bottle of smoothie per day (Mintel 2008). So, five exposures were thought to represent a bottle of 250 ml of smoothie every day.

Tooth wear measurement in this study was represented by difference in height, using unexposed areas as a reference point. The three point height difference was performed using the laser profilometry software. Profilometry is method of choice to measure loss of surface. This technique is simple, quick and allows measurement of surface loss of a large area with a high precision. This method is also widely used in many studies, allowing comparisons.

Fruit juices contain weak acids which exist in solution as either undissociated molecules or in dissociated form; pH measures the hydrogen ion (H+) concentration and titratable acidity (TA) measures concentration of all H+ ions, both free and bound to undissociated acids and anions. Generally, the smoothies have a higher pH than the Diet Coke (2.95) and Citric acid (3.60). However, their TA was 3.5–4 times more than both Diet Coke and Citric Acid. For water, theoretically the pH of pure water (H2O) is 7.0 at 25 °C, but when exposed to the carbon dioxide in the atmosphere this equilibrium results in water becoming slightly acidic with time. Water also does not have buffering capacity; the pH reading shoots up even with very small amount of NaOH. Hence, no reading for pH and titratable acidity for DD water was performed.

There have been debates on whether pH or TA is a better predictor of a drink’s erosive potential. pH measures H+ concentration while TA determines the actual hydrogen ion availability for interaction with tooth surface (Boulton 1980). Some studies have shown that pH was a better predictor and some believed that TA was more important (Edwards et al. 1999; Cairns et al. 2002). There have been suggestions that pH is a good predictor for the first few minutes of erosion challenge, whereas the TA better characterises the erosive potential during longer exposure times (Jensdottir et al. 2006). This was agreed by Hannig who believed that when there is short-term exposure of enamel to acidic environment; pH and type of acid impose a bigger effect on erosion rather than titratable acidity (Hannig et al. 2005). In this present study, there was a positive correlation to TA (r = 0.59) but the value shows fair/medium association and should be treated with suspicion that there are other variables influencing the association. Generally, it is agreed that pH and TA alone do not readily explain the erosive potential of food and drinks.

It was noted that smoothies are viscous drinks, containing a lot of fibres compared to the other sample drinks. Mangoes and passion fruit smoothie had 18 % fibre and the strawberry and banana smoothie had 15 % of its content consisting of fibres. During the process of titrating, a magnetic stirrer had to be used to the speed of 875 rpm to mix the NaOH well. It also took the pH meter a longer time to come to a reading, as it needed a stable reading before giving a measurement.

When comparing smoothies with Diet Coke, it was found that erosion from both smoothies were not significantly different to the erosion caused by the Diet Coke. In other words, drinking smoothies will cause almost the same amount of erosion as when slabs were exposed to Diet Coke.

Conclusions

This study showed that Innocent Strawberries and Bananas smoothie and Innocent Mangoes and Passion fruit smoothie are acidic and had a high titratable acidity. Both smoothies caused significant surface loss/erosion to enamel after 21-day pH cycling regimen. Innocent Mangoes and Passion fruit smoothie caused significantly more surface loss compared with Innocent Strawberries and Bananas smoothie. Both smoothie caused surface loss/erosion to enamel similar to the effect of Diet Coke after the 21-day pH cycling regimen.