Selection of studies
The search identified 178 preliminary references, of which 36 were found in PubMed, 46 in Scopus, 1 in the Cochrane Library, 74 in Web of Science, 20 in EMBASE, 0 in the grey literature and 1 in the manual search of cited references. After excluding 30 duplicates, the remaining 148 were examined. Of these, 123 were excluded after reading the title and abstract, as they were not related to the research question.
After reviewing the full texts of the remaining 25 articles, 11 were excluded because they did not meet the inclusion criteria. Finally, 14 articles met the inclusion criteria and were selected for qualitative analysis. The same articles were included in the qualitative synthesis (meta-analysis). The PRISMA 2020 flowchart (Fig. 1) provides an overview of the article selection process.
Characteristics of the studies
Regarding sample size, the studies by Menon et al.11, with 840 stretches of elastomeric chains; by Kumar et al. 12, with 480 stretches; and by Larrabee et al.8, with 450 stretches showed much higher sample values than those in other studies, such as those by Al-Ani, Sadeghian et al.14 and Pithon et al.7,15, in which the samples barely reached values of 80 stretches of elastomeric chains. Regarding the types of links measured in the studies, all the studies analysed short-link chains. In the study by Omidkhoda et al.16, closed-link chains in addition to short-link chains were analysed.
The numbers of links used to measure the strength of the elastomeric chains were quite different. Behnaz et al.17 used elastomeric chains with 6 links. Authors such as Sadeghian et al.14, Javanmardi and Salehi4, Odmikhoda et al.16 and Pithon et al.7,15 used 5 links to carry out measurements, two of which placed pins in the ends, leaving 3 links in the central part. Oshagh et al.18 and Kumar et al.12 used 4 links. Menon et al.11, Ramachandraiah et al.19 and Larrabee et al.8 used only 3-link sections. In the studies by Al-Ani13 and Nahidh et al.20, the numbers of links used for force measurement were not specified.
Regarding the commercial manufacturers of the elastomeric chains assessed in the studies, many differences were found between studies. Morelli® elastomeric chains were used in the Al-Ani13 and Pithon et al.7,15 studies. 3M® brand chains were used in the Ramachandraiah et al.19 and Kumar et al.12 studies. Rocky Mountain® chains were selected in the study by Larrabee et al.8 Ortho Technology® chains were used by Oshagh et al.18 and Nahidh et al.20. Benhaz et al.17, Mirhashemi et al.5 and Sadeghian et al.14 measured loss of strength in elastomeric chains produced by American Orthodontics®. Omidkhoda et al.16 used stretches of elastomeric chains produced by Dentaurum®, and Menon et al.11 used chains commercially produced by Ortho Organizers® for their study. In general, all of the studies used chains from a single commercial company to carry out their studies, except Ramachadraiah et al.19, who used chains from three different commercial companies: Ortho Plus®, 3M® and Ortho Organizers®. There were also discrepancies between authors regarding the distance between the pins to which the elastomeric chains were attached to measure force during the measurement intervals. Menon et al.11, Pithon et al.7,15 and Larrabee et al.8 used a distance between the holding pins of 23.5 mm. Behnaz et al.17, Sadeghian et al.14, Omidkhoda et al.16, Kumar et al. and Oshagh et al.12,18 used pins with a distance of 25 mm. Oshagh et al.18 applied a distance supported the study by Nattrass et al.6, who stated that 25 mm was the average distance between the canine and the first molar. Al-Ani13 and Nahidh et al.20 used a distance of 29 mm, Javanmardi and Salehi 4 used a distance of 15 mm, and Ramachandraiah et al.19 used distances of 18 and 22 mm for prestretched and unstretched chains, respectively.
The measurement intervals applied to assess the strength degradation of the selected elastomeric chains were quite similar among studies. All the authors measured strength degradation at 0 hours, 24 hours, 7 days, 14 days, 21 days, and 28 days. In only the study by Al-Ani13 was the force not recorded at 14 and 28 days. Of the mouthwashes and agents used in the studies included in this review, the most frequently studied was Listerine® (26.9% alcohol) 11,12,13,14,15,19. Some authors studied other types of Listerine®-brand mouthwashes, such as Listerine® Zero (0% alcohol)15, Listerine® Total Care Zero (NaF 0.02%)17, Listerine® Healthy White (NaF 0.02% + hydrogen peroxide)17, Listerine® Whitening (hydrogen peroxide)15 and Listerine® Green Tea20. Rinses containing 0.2% chlorhexidine, such as Chlorhexidine Plus® 11, Cordosyl® 18 and Cleanform®, and 0.12% chlorhexidine, such as Periogard®, were also studied. Other authors reported the results of sodium fluoride (NaF)-based mouthwashes, such as Orthokin® (NaF + chlorhexidine)4, Sensikin® (NaF + potassium nitrate)4, Oral B® 0.05% (NaF 0.05%)18 and Colgate® Phos-Flur (NaF 0.04%)11. Plax mouthwashes (CPC 0.075% + NaF 0.05%) and Plax Whitening (CPC 0.05% + 1.5% hydrogen peroxide) were also studied7. Finally, Nahid et al.20 studied herbal mouthwashes, such as Tebodont, Aloe-dent and Silca Herb 20, while several authors included Persica-based mouthwash in their studies.4,5,16
It should be noted that various authors focused solely on the study of specific agents, such as alcohol at different concentrations, frequently found in a wide variety of mouthwashes. The concentrations studied were 26.9%8,11, 21.6%19, 14%8 and 8.38% alcohol.19
Agents such as chlorhexidine and sodium fluoride have also been extensively studied independently. Chlorhexidine was studied at a concentration of 0.2% in most relevant studies5,11,16, but it was also studied at a concentration of 0.12% in the study by Pithon et al.15. The included studies that analysed the results obtained by NaF did so at concentrations of 0.4% 11, 0.2% 5 and 0.05% 14,16.
In the control groups, the agents used were distilled water and artificial saliva. Of the 14 articles selected, half used distilled water 7,8,12,13,15,18,20, and the other half used artificial saliva. 4,5,11,14,16,17,19.
All the studies applied the mouthwash for 60 seconds twice a day, except for the studies by Pithon et al.7,15 and Behnaz et al.17, who applied mouthwash for 30 seconds twice a day.
The characteristics of the studies, as well as the data extracted from the articles included in this review, are summarized in Table 3.
Results of individual studies
Type of rinse: Of the authors8,11,12,13,14,15,19 who studied the influence of Listerine® rinse on the degradation of the physical properties of elastomeric chains, Menon et al.11 found that Listerine®, among the mouthwashes studied, caused the largest percentage of disintegration (71.6%) in a period of 28 days. These results were very similar to those obtained by Ramachandraiah et al.19, who reported a strength degradation percentage of 69.25%. Other authors8,12− 14 also found that all groups of alcohol-containing mouthwashes, including Listerine®, showed significant increases in strength degradation compared to the control group (p < 0.01). The authors8 who studied Cepacol® mouthwash observed a loss of 54.2% at 24 hours compared to 53.0% for Listerine®.
Regarding the studies that evaluated mouthwashes with bleaching agents, authors such as Behnaz et al.17 found a decrease in strength from 48.34–53.38%, but without statistical significance (p < 0.05) compared to the control group. These results coincided with those of Pithon et al.7, who found that the presence of a bleaching agent did not significantly influence the degradation of the strength of the elastomeric chains17,7.
Of the authors5,11,15,16 who studied mouthwashes with 0.2% chlorhexidine, Omidkhoda et al.16 found a decrease in strength of 27.24% in the first 24 hours compared to non-chlorhexidine mouthwashes, similar to Mirhashemi et al.5, who observed a significant loss of strength in all groups (p < 0.001). However, in the study by Pithon et al.15, chlorhexidine did not show a significant influence on the pattern of strength degradation.
In mouthwashes such as Orthokin® or Sensikin® that contain NaF, no effect on the pattern of strength degradation was observed4. However, studies11,14,16,18 that analysed mouthwashes with 0.4%, 0.2%, and 0.05% NaF showed 41–50% losses at 4 weeks.
Another mouthwash analysed was Persica. Omidkhoda et al.16 reported a strength degradation value of 18.63%, while Mirhashemi et al.5 reported a loss of 32.4%. Similar results were obtained by Javanmardi and Salehi4, who obtained higher values of degradation for Persica mouthwash compared to other mouthwashes.
Measurement intervals: Most of the studies included in this review found that the maximum percentage of strength loss occurred at 24 hours in all study groups, including the control groups4,5,8,11,13,14,12,15,16,19. In the study by Menon et al.11, this loss was 49.48%, similar to values reported by Al-Ani13 and Behnaz et al.17, who observed decreases in strength of 57.85% and 53.38%, respectively.
In contrast, of the time intervals studied, the lowest disintegration percentage was observed at 28 days, and the force was relatively constant. In the study by Behnaz et al.17, the total degradation value was 86.48% for bleaching agents and 66.3% for control agents. In the study by Pithon et al.15, after 28 days of measurement, the greatest percentage of loss of strength occurred in the control group (distilled water).
Quantitative analysis (meta-analysis)
To evaluate the “effect of mouthwash” on the degradation of strength, the difference with respect to the total control group was calculated for each total mouthwash group, combining all studies. To compare the different mouthwashes, any alcohol with a concentration greater than 20% was considered to be of the same type. Similarly, any mouthwash with 0.04% or 0.05% NaF was also considered to be of the same type. Differences were meta-analysed as SMDs.
Measurement at 24 hours
Regarding the global estimations of the differences between mouthwashes and controls at 24 hours, the results showed that, in general, all the mouthwashes we associated with greater force degradation than the control agents (SMD=-0.74; p = 0.006; 95% CI=-1.26, -0.21) (Fig. 2). The heterogeneity of the studies was very high (I2 = 97%; p < 0.001) because certain authors reported extreme levels of degradation compared to the control group.
However, no mouthwash studied separately achieved a significant difference with respect to the control group (Persica SMD = 0.10; p = 0.496; 95% CI=-0.18, 0.37; I2 = 0%; (Qh) p = 0.789; alcohol 26, 9% SMD= -1.93, p = 0.180, 95% CI-4.75, 0.89, I2 = 99.4%, (Qh) p < 0.001; NaF SMD= -0.40, p = 0.063, 95% CI=-0.82, 0.02, I2 = 65.1%; (Qh) p = 0.040; chlorhexidine 0.2% SMD= -0.25; p = 0.203; 95% CI-0.64, 0.14; I2 = 70.8%; (Qh) p = 0.002)), except Listerine®, which presented a strong tendency, although under a setting of strong heterogeneity (SMD= -1.27; p = 0.059; 95% CI=-1.88, 0.06; I2 = 98.5%; (Qh) p < 0.001) .
Regarding the comparison among mouthwashes, no significant differences in the force measured at 24 hours (p = 0.200) were observed. In fact, the proximity of the grey diamonds, which symbolize the global effect measurement of the different mouthwashes in Fig. 2, illustrates these similarities.
Although there were no overall differences between the 5 most frequently tested products, interesting trends were observed in the direct comparisons; this was the case for mouthwashes containing alcohol, which had a significantly larger degradation percentage than Persica (p = 0.046) and those containing chlorhexidine 0.2% (p = 0.098). Listerine® mouthwash was associated with a larger percentage of degradation than Persica, but statistical significance was not observed (p = 0.085) (Table 4).
Regarding the type of connector (short or closed) of the elastomeric chain, statistically significant differences in strength loss were found among studies at 24 hours (p = 0.169; 95% CI=-0.44, 2.50).
Measurements at 7 days
Regarding the overall estimates of the differences between the mouthwashes and controls at 7 days, there were differences among the 5 most frequently used products. In general, all mouthwashes were associated with greater force degradation than the control agents (SMD=-1.19; p = 0.003; 95% CI=-1.99, -0.39) (Fig. 3). The heterogeneity of the studies was very high (I2 = 98.6%; p < 0.001) because there were imprecise studies that reported greater degradation, while others reported contrasting data.
Regarding comparisons among mouthwashes, significant differences in the force measured at 7 days were found among mouthwashes (p = 0.005). Mouthwashes containing alcohol had a significantly higher percentage of degradation than those containing chlorhexidine 0.2% and NaF as well as Persica. Listerine® had a significantly larger degradation percentage than 0.2% chlorhexidine-containing mouthwashes and Persica, and a significant difference compared to NaF was observed. Alcohol-containing mouthwashes and Listerine® degraded the elastomeric chains in a similar way; the other 3 mouthwash types also showed similar degradation patterns pattern to one other (Table 4).
Regarding the type of connector (short or closed), there were no significant differences in the impacts of the mouthwashes depending on whether they were applied to closed systems or short systems (p = 0.418; 95% CI=-2.65, 1.09).
Measurements at 14 days
Regarding the differences between mouthwashes and controls, significant differences were interpreted at 14 days (SMD=-2.09; p = 0.018; 95% CI=-3.82-0.36). The experimental products were associated with a significant loss of strength compared to the control agents (Fig. 4). Heterogeneity was high (I2 = 99.7%; p < 0.001). After 14 days of measurement, alcohol and NaF showed trends toward loss of strength (the result was significant in the case of NaF) compared to their respective control groups (p = 0.005).
Regarding comparisons among mouthwashes, notable differences were observed, but they were not significant (p = 0.076) (Table 4).
Mouthwashes containing alcohol were associated with a larger percentage of degradation than those containing chlorhexidine 0.2% and NaF as well as Persica. Listerine® mouthwash had a significantly larger degradation percentage than mouthwash containing 0.2% chlorhexidine, and there was a large but nonsignificant difference compared to Persica and NaF.
Regarding the type of connector (short or closed), there were no significant differences in the impacts of mouthwashes applied in closed systems or short systems (p = 0.308; 95% CI=-6.03, 1.90).
Measurements at 21 days
Regarding the difference between the mouthwash and control group, a significant difference was observed (SMD=-2.08; p < 0.001; 95% CI=-3.34, -0.82) (Fig. 5). After 21 days of measurement, Listerine® mouthwash and alcohol-containing mouthwashes were associated with great losses of strength compared to their respective control agents (at the limit of significance). In general, the use of mouthwash was associated with a significant loss in strength compared to use of a control agent (p = 0.001); however, these results were obtained under highly heterogeneous conditions (I2 = 99.3%; p < 0.001).
Regarding comparisons among mouthwashes, there were no differences in the strength measured at 21 days among the different mouthwashes (p = 0.120). In general, alcohol-containing mouthwashes degraded the elastomeric chains more than Persica, and Listerine® degraded the elastomeric chains more than mouthwashes containing 0.2% chlorhexidine and NaF as well as Persica (Table 4).
There were no significant differences in the impact of mouthwashes according to closed or short systems (p = 0.183; 95% CI=-4.69, 0.89).
Measurements at 28 days
At 28 days of measurement, statistically significant differences were found between the mouthwashes and the control groups (SMD=-4.19; p < 0.035; 95% CI=-8.08, -0.31) (Fig. 6); however, these results were obtained under highly heterogeneous conditions (I2 = 99.9%; p < 0.001).
Regarding comparisons among mouthwashes, there were no differences in the strength measured at 28 days between the different mouthwashes (p = 0.778) (Table 4).
There were no significant differences in the impacts of the mouthwashes depending on whether they were applied in closed systems or short systems (p = 0.094; 95% CI=-1.33, 16.9), although there seemed to be a certain tendency towards worse performance in closed systems.
Risk of bias of studies and quality analyses
Figure 7 shows the bias diagram and the parameters used for the publications included in this systematic review. An overall judgement of a high, low, or moderate risk of bias was formed for each study based on whether the authors considered the parameters to be similar, using a protocol adapted to systematic reviews of in vitro studies by David et al.10
Positive evaluations were made for studies which had well defined control groups such as artificial saliva or normally distilled water. Having performed a well statistical analysis was also a factor of high quality. The same happened when the articles reported correctly the type of mouthwash used for their studies. Other factors such as the number of links and the distance between pins were also evaluated, with results being quite variable among the studies (Table 5).
Egger’s test (p value) for publication bias
Using Egger’s test, we were able to graphically represent the regression line between the precision of the studies (independent variable) and the standardized effect (dependent variable) and perform a statistical evaluation of bias.
At 24 hours, the study by Ramachandraiah et al.19, assessing 26.9% alcohol, showed strong publication bias (p < 0.001) because the study was more imprecise than those by Menon et al.11 and Larrabee et al.8 due to its smaller sample size. The studies by Menon et al.11, Mirhashemi et al.5, Oshagh et al.18, Omidkhoda et al.16, and Pithon et al.15, which all assessed 0.2% chlorhexidine, showed clear asymmetry in the funnel plot (p < 0.001). The more imprecise the study was, the more strength degradation was found.
At 7 days, total symmetry was observed among studies assessing some mouthwashes, such as Persica4,5,16, Listerine®8, 11,12,13,14,15,19 and NaF-containing mouthwashes14,15,16,18; however, for studies assessing 0.2% chlorhexidine-containing mouthwashes, clear asymmetry was observed in the Funnel plot (p < 0.001). The lower the accuracy was, the greater the reported force degradation.
In comparisons of studies on mouthwashes, important publication bias was observed since some studies that were very imprecise reported great degradation, while others with the same level of imprecision reported opposite results.
At 14 days, asymmetry was observed; this asymmetry was due to the study by Ramachandraiah et al.19, who studied Listerine®, since the results were radically different from those in the other studies8,11,12,14,15 and the most imprecise.
At 21 days, the studies by Menon et al.11, Pithon et al.15 and Omidkhoda et al.16, who studied 0.2% chlorhexidine, caused asymmetry in the funnel plot (p = 0.089). The observed trend was as follows: the greater the precision of the study was, the clearer the disadvantage of the mouthwash compared to the control group.
At 28 days, the heterogeneity due to the Listerine® 8, 11, 12,15 studies was very high (I2 = 99.9%), and the funnel plot showed that the main cause was the study by Ramachandraiah et al.19 (p < 0.001). No additional data were available for further comparisons.