Abstract
The extracellular matrix has been identified as an essential determinant of virulence of dental biofilms. Therefore, enzymes that degrade the matrix are a promising preventive approach for biofilm control. Laboratory studies have consistently reported significant effects of enzyme treatment on biofilm formation, but information from randomized clinical trials (RCTs) is limited. The present triple-blind RCT investigated the effect of lozenges containing 2 different concentrations of the matrix-degrading enzymes mutanase, beta-glucanase, and DNase on de novo plaque formation, gingivitis development, and the plaque microbiome in healthy adults. Eighty subjects were enrolled with random allocation to either placebo, active 1, or active 2 (3-fold enzyme concentration). All subjects completed the study without major protocol deviations. Plaque formation was assessed after 1, 7, and 14 d of intervention without self-performed oral hygiene, using the Turesky modification of the Quigley and Hein plaque index (TM-QHPI) and the planimetric plaque index and the thickness index. Gingival index (GI) scores were registered on day 14. Moreover, the microbial plaque composition was analyzed by 16S rRNA gene sequencing, and the amount of autofluorescent plaque and plaque removal with an Airfloss device was quantified. Plaque formation was significantly lower for day 7, active 1 compared with placebo and for the pooled active groups (post hoc analysis) as well as for the pooled recordings across all time points (post hoc analysis) and for autofluorescent plaque on day 7, active 2. No significant differences in plaque formation were observed for the other time points, GI scores, or plaque removal. A nonsignificant trend toward reduced species richness was found in both active groups compared with placebo. In conclusion, multienzyme treatment may promote oral health by slightly delaying plaque formation and maturation and might serve as a supplement to mechanical plaque removal. The study was registered at Clinicaltrials.gov (NCT05082103).
Keywords
Introduction
The most prevalent biofilm-induced oral diseases—dental caries and periodontitis—are caused by a breakdown of the symbiotic relationship between host and oral microbiome (Marsh 2018). Self-performed oral hygiene is the mainstay in preventing the uncontrolled buildup of dental biofilm and thereby oral disease, but complete plaque removal is difficult to achieve, especially in areas that are hard to reach by mechanical means (Claydon 2008). In addition to fluoride, oral care products therefore typically contain antimicrobial agents that seek to kill harmful bacteria in dental biofilms. Their effect, however, is unspecific and more prone to eliminate commensal microorganisms on mucosal surfaces than bacteria in thick plaque deposits (Adams and Addy 1994; James et al. 2017; Bescos et al. 2020).
Bacteria in dental biofilms are embedded in a highly diverse extracellular polymeric matrix that comprises polysaccharides, glycoconjugates, extracellular DNA (eDNA), proteins, and lipids (Jakubovics et al. 2021). The extracellular matrix provides protection against environmental stresses such as desiccation or antimicrobials, and it especially confers mechanical stability to dental biofilms. Therefore, matrix-degrading enzymes have been identified as a promising therapeutic approach to destabilize dental biofilm and facilitate its removal (Pleszczyńska et al. 2017). In vitro studies have consistently demonstrated convincing treatment effects for single enzymes such as mutanase, dextranase, or DNase and likewise for combinations of different enzymes applied in concert (Hayacibara et al. 2004; Wiater et al. 2013; Otsuka et al. 2015; Singh et al. 2021; Del Rey et al. 2024; Nielsen et al. 2024). However, compared with laboratory models, in vivo–grown dental plaque harbors a more complex biofilm matrix, and it may therefore be more resilient to enzymatic removal (Dige et al. 2022). As of today, only a few clinical trials have assessed the effect of enzymatic treatment on plaque accumulation in the mouth, and most of them did not use combinations of matrix-degrading enzymes (Kelstrup et al. 1978; Midda and Cooksey 1986; Daly et al. 2019; Paqué et al. 2021; Hoffstedt et al. 2023; Schlafer et al. 2024).
A recent laboratory study investigated the effect of combined treatment with mutanase, beta-glucanase, and DNase on a saliva-derived biofilm model and reported considerable removal of established biofilms and almost complete prevention of de novo biofilm formation (Rikvold et al. 2024). Mutanase and beta-glucanase cleave alpha-1,3- and beta-1,3 and 1,4-glycosidic bonds, respectively, whereas DNase hydrolyzes phosphodiester linkages in the backbone of eDNA. Thereby, the 3 enzymes target key components of the dental biofilm matrix (Jakubovics and Burgess 2015; Pleszczyńska et al. 2017; Jakubovics et al. 2021). Inspired by the in vitro results, 2 clinical pilot trials were conducted. One assessed the effect of a mouthwash containing the 3 enzymes on habitual plaque formation in orthodontic patients and found reduced plaque levels after 8 d of intervention (Hoffstedt et al. 2023). The other trial applied the enzymes in a lozenge format for 7 d and investigated their effect on de novo plaque formation and the oral microbiome in healthy individuals (Schlafer et al. 2024). The results provided evidence of retarded plaque formation and maturation in the treatment group, as indicated by a lower plaque index and species diversity, and laid the foundation for the present randomized controlled trial (RCT).
The aim of this RCT was to systematically investigate the effect of enzyme treatment with lozenges containing 2 different concentrations of mutanase, beta-glucanase, and DNase on plaque formation in healthy individuals. Specifically, the trial assessed the amount of de novo plaque accumulation, the development of gingivitis, and the microbial plaque composition in the absence of self-performed oral hygiene. In addition, the amount of de novo formed red autofluorescent plaque and the amount of plaque removed by accelerated micro-droplets of air and water (Airfloss) were quantified. The null hypothesis was that enzymatic treatment does not affect (a) dental plaque accumulation, (b) the development of gingivitis, (c) the microbial plaque composition, (d) the amount of red autofluorescent plaque, and (e) plaque removal with an Airfloss device, compared with placebo treatment. Dental plaque accumulation was assessed after disclosure and determined using the Turesky modification of the Quigley and Hein plaque index (TM-QHPI), the planimetric plaque index (PPI), and the thickness index (TI) (Turesky et al. 1970; Del Rey et al. 2023; Rikvold, Del Rey, et al. 2023). The microbial plaque composition was analyzed by next-generation 16S rRNA sequencing. The amount of red autofluorescent plaque and plaque removal were quantified using the PPI.
Materials and Methods
Ethical Approval and Study Registration
The study protocol was approved by the Ethical Committee of Region Midtjylland (1-10-72-259-21) and registered at ClinicalTrials.gov (NCT05082103) and at the internal database for research projects at Aarhus University (2021-0214784). The study was conducted in accordance with the Helsinki Declaration and its amendments, and the article was written according to the CONSORT reporting guidelines (Schulz et al. 2010).
Trial Design and Test Product
The study was a triple-blind RCT with 3 parallel arms and equal allocation. Investigators, participants, and outcome assessors were blinded to intervention assignments. Participants were allocated to either control treatment with a placebo lozenge or active treatment with an enzyme-containing lozenge (active 1 or 2). The active lozenges contained a combination of the enzymes DNase, mutanase, and beta-glucanase (BioFresh® Clean S, Novozymes A/S), applied in 2 different concentrations. Active 2 lozenges had a 3-fold higher activity than active 1 lozenges did. Except for the enzymes, the placebo and active lozenges had the same ingredients and were identical in appearance, taste, and texture (Appendix Table 1). The study was monitored by the sponsor and conducted according to the guidelines of good clinical practice (International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use 2016).
Sample Size, Eligibility Criteria, and Recruitment
Based on the results of a pilot trial (Schlafer et al. 2024), a sample size of 38 participants was calculated for each group. An interim analysis was performed after the first 60 participants. Healthy participants with at least 20 natural teeth and no active carious lesions, signs of periodontitis, or soft-tissue pathology were enrolled in the study. For details on sample size calculation, eligibility criteria, and recruitment process, see the Appendix.
Outcomes and Statistical Analysis
The primary outcome of the study was (1) the plaque accumulation after 1 d of intervention, determined by the TM-QHPI. Secondary outcomes were (2) plaque accumulation after 7 and 14 d of intervention, determined by the TM-QHPI; (3) plaque accumulation after 1, 7, and 14 d of intervention, determined by the PPI; and (4) the development of gingivitis after 14 d of intervention, assessed by the gingival index (GI) of Löe and Silness (Löe et al. 1965). Exploratory outcomes were (5) the composition of the plaque microbiome after 14 d of intervention, assessed by 16S rRNA gene sequencing; (6) the plaque thickness after 1, 7, and 14 d, determined by the TI; (7) the amount of red autofluorescent plaque after 1, 7, and 14 d; and (8) the amount of plaque removed by an Airfloss device after 14 d, determined by the PPI. Data were reviewed and statistical analyses performed before unblinding according to a predefined plan (Appendix). For hypothesis-generating post hoc analyses (not part of the predefined plan), the active groups and data from different time points were pooled (Appendix). For secondary and exploratory outcomes, no correction for multiplicity was applied.
Study Intervention
The intervention period lasted 14 d, during which the participants had 4 physical appointments: a baseline visit (day 0) and visits on day 1, 7, and 14 (see Fig. 1A). During all visits, a thorough clinical examination of the oral mucosa was performed to identify abnormalities or potential adverse events (AEs) of the investigational product.
Study intervention and participant flow diagram. (
At baseline, the GI was recorded and the participants received a professional tooth cleaning. They were instructed to refrain from all self-performed oral hygiene procedures and to not use mouthwash or chewing gum during the intervention period. Instead, participants were instructed to take 1 lozenge 3 times/d, 30 min after meals. Instructions were also given not to chew the lozenges but to place them on the dorsum of the tongue and let them dissolve while moving the saliva around in the mouth. The GI recording was repeated after 14 d. Plaque samples were collected with toothpicks from 3 approximal spaces and stored at −80 °C until analysis.
On day 1, 7, and 14 of the intervention period, the TM-QHPI, PPI, and TI were assessed on fluorescence camera images after plaque disclosure. After the first set of fluorescence images was acquired on day 14, a defined mechanical treatment with an Airfloss device was performed to assess if the enzyme-treated plaque was destabilized. In addition to the assessment of disclosed plaque, undisclosed plaque was quantified planimetrically on day 1, 7, and 14, to investigate the amount of red autofluorescent plaque. These assessments were performed before the disclosing solution was applied. At the end of each visit, the tooth surfaces that had been disclosed were cleaned professionally. Details on examiner calibration; assessment of TM-QHPI, PPI, GI, and TI; and the subsequent digital image analysis are provided in the Appendix (Methods and Appendix Table 2).
16S rRNA Gene Sequencing
The microbial plaque composition was determined by 16S rRNA gene amplicon sequencing. For details on DNA extraction, amplification, data processing, and statistical analysis, see the Appendix.
Results
Interim Analysis and Study Abortion
The interim analysis tested for efficacy and futility of the intervention, based on the primary outcome, the TM-QHPI on day 1. The average TM-QHPI was slightly, but not significantly, lower in both active groups than in the placebo group (Appendix Table 3; placebo vs. active 1: P = 0.22; placebo vs. active 2: P = 0.26). Conditional powers for the maximum predetermined sample size of 60 participants per group were 43.2% (placebo vs. active 1) and 21.5% (placebo vs. active 2), respectively. The study was therefore aborted, as the desired conditional power of 80% could not be reached within the predetermined limits of budget and sample size. By the time the first 60 participants completed the study, another 20 participants had been enrolled and started the intervention. These participants completed the study while the interim analysis was performed, and the final data analysis therefore comprised 80 participants.
Participant Flow, Baseline Data, and AEs
An overview of the participant flow is shown in Fig. 1B (for details, see Appendix). Baseline demographic data and GI recordings, as well as the number of lozenges taken, are shown in the Table. Seven AEs were reported (Appendix), but none of them precluded the participants from completing the study (Appendix Table 4).
Baseline Demographics, Gingival Index (GI) at Baseline, and Compliance Data (Total Number of Lozenges Taken in the Intervention Period).
Primary and Secondary Study Outcomes (TM-QHPI, PPI, and GI), Post Hoc Analyses
After 1 d of intervention, the average TM-QHPI did not differ significantly between groups (Fig. 2 and Appendix Table 5). After 7 d of intervention, TM-QHPI was significantly higher in the placebo group compared with active 1 (P = 0.012) but not active 2 (P = 0.23). No significant differences in average TM-QHPI were observed between the groups after 14 d of intervention. Less plaque formed in the lower than in the upper jaw, resulting in generally lower scores on day 14.
Effect of enzymatic treatment on de novo plaque accumulation, assessed by the Turesky modification of the Quigley and Hein plaque index (TM-QHPI). Average TM-QHPI (SD) did not differ significantly between groups after 1 and 14 d of intervention. After 7 d of intervention, the average TM-QHPI was significantly lower in the active 1 group compared with placebo (*P = 0.012) but not in the active 2 group (P = 0.226). Boxplots and whiskers show minimum (bottom whisker), first quartile (lower boundary of the box), median (horizontal line), mean (cross symbol), third quartile (upper boundary of the box), and maximum (top whisker) values. SD, standard deviation of the mean. Note that plaque levels were higher on day 7 than on day 14, because assessment on day 7 was performed in the upper jaw and on day 14 in the lower jaw, where plaque levels are lower in general.
Average PPI did not differ significantly between groups on day 1 and 14 (Fig. 3 and Appendix Table 5). On day 7, average PPI was significantly higher for placebo than for active 1 (P = 0.02) but not active 2 (P = 0.06). Both TM-QHPI and PPI were significantly lower for the pooled active groups on day 7 (Appendix Table 11) but not on day 1 or 14. PPI was significantly lower for active 1 (−4.28%, 95% confidence interval [CI] [−8.25%, −0.31%]; P = 0.035) and active 2 (−4.43%, 95% CI [−8.44%, −0.42%]; P = 0.03) compared with placebo when data were pooled for all 3 time points, while the TM-QHPI was significantly lower for active 1 (−0.24, 95% CI [−0.40, −0.07]; P = 0.005) but not active 2 (−0.14, 95% CI [−0.31, 0.03]; P = 0.1). Null hypothesis (a) that enzymatic treatment does not affect plaque accumulation was partially rejected.
Effect of enzymatic treatment on de novo plaque accumulation, assessed by the planimetric plaque index (PPI). The PPI was calculated semiautomatically on fluorescence images acquired after plaque disclosure. (
The GI values increased during the intervention period compared with baseline, but no significant difference was observed between the groups (Appendix Table 6; placebo vs. active 1: P = 0.29; placebo vs. active 2: P = 0.59; active 1 vs. active 2: P = 0.61). Null hypothesis (b) was accepted.
Exploratory Outcomes (TI, Plaque Removal, Red Autofluorescent Plaque. and Microbial Composition)
No significant TI differences were observed between groups after 1, 7, and 14 d of intervention (Appendix Table 7). Airfloss treatment removed little plaque, irrespective of the intervention group (Appendix Fig. 1), and no statistically significant differences were found between groups (Appendix Table 8). Significantly less autofluorescent plaque had formed after 7 d in the active 2 group compared with both active 1 and placebo (Appendix Table 9). In general, only small amounts of red autofluorescent plaque had formed compared with the total disclosed plaque. Null hypothesis (d) was partially rejected, while null hypothesis (e) was accepted.
Plaque samples harbored a diverse bacterial community representative of mature dental biofilms. Samples were dominated by Streptococcus spp. (12.43% ± 7.31 SD), Leptotrichia spp. (9.54% ± 5.87 SD), and Actinomyces spp. (8.66% ± 5.05 SD; Fig. 4A and Appendix Fig. 2). Enzyme-treated samples showed a nonsignificant trend toward reduced species richness (Fig. 4B). Seven zero-radius operational taxonomic units (OTUs) were significantly less abundant in the active groups (active 1: n = 2; active 2: n = 7), while 1 OTU was more abundant in the placebo group (Fig. 4C). The bacterial composition was mostly influenced by the individual participant, rather than by the intervention group, as evidenced by principal component analysis (Fig. 4D). Null hypothesis (c) was partially rejected.
Microbial composition of dental plaque after 14 d of intervention. (
Discussion
The present trial performed an in-depth analysis of the effect of lozenges containing 2 different concentrations of 3 matrix-degrading enzymes (mutanase, beta-glucanase, and DNase) on dental plaque formation and gingivitis development in healthy adults. A trend toward lower or delayed plaque buildup in the active groups compared with placebo was observed, but most investigated variables, including the primary outcome, were not significantly different between groups.
These findings contrast in vitro experiments on the same enzymes, which showed more than 60% biofilm removal and almost complete prevention of de novo biofilm formation in a saliva-based model (Rikvold et al. 2024). This difference between laboratory and clinical data illustrates that dental biofilms grown intraorally have a more diverse matrix composition than in vitro models do and are not as easily disrupted (Dige et al. 2022; Rikvold, Kambourakis Johnsen, et al. 2023; Rikvold et al. 2024).
Results from previous RCTs assessing the effect of enzymes on plaque formation are ambiguous. Paqué et al. (2021) found no significant effect of toothpaste slurries containing the antibacterial enzyme glucose oxidase on 4-d-old undisturbed plaque, whereas Daly et al. (2019) observed significantly less plaque formation after 13 wk. Likewise, Hoffstedt et al. (2023) reported lower plaque scores in orthodontic patients after 8 d with a mouthwash containing the same enzymes as the actives in this RCT.
The results from the present trial align with a previous pilot RCT that found significantly reduced plaque formation after 1 d but not after 7 d of intervention (Schlafer et al. 2024). Compared with the pilot trial, the present RCT is based on a larger sample size, it tested the application of 2 different active concentrations, and it used a more elaborate methodology for plaque quantification. Specifically, plaque levels were assessed after disclosure, and in addition to the TM-QHPI, the PPI was quantified semiautomatically, using a method that reduces subjectivity (Del Rey et al. 2023; Rikvold, Del Rey, et al. 2023). Moreover, the levels of autofluorescent plaque were measured, and the intervention period was prolonged to 14 d to allow for a more reliable assessment of the development of gingivitis. Other strengths of the present study were a high participant compliance and no dropouts. The present trial has a more robust design compared with the pilot RCT and may explain differences in the outcomes.
The observed trend of slightly reduced biofilm formation for both active groups was consistent across all investigated plaque-related outcomes. After 7 d of intervention, the TM-QHPI and PPI were significantly lower for the active 1 group compared with placebo. However, it remains to be determined whether a reduction of 0.37 TM-QHPI units or a PPI difference of 8.53% results in clinical benefits for the individual. Some previous RCTs have argued that TM-QHPI reductions as small as 0.1 units are clinically relevant (Pabel et al. 2018; Daly et al. 2019), while other trials have aimed for larger effect sizes (van Palenstein Helderman et al. 2006; Adam et al. 2020). Daly et al. (2019) found that TM-QHPI reductions of 0.11 units were associated with significant reductions in the modified GI of 0.33 units. These results, however, were observed after a 13-wk intervention period during which participants performed regular oral hygiene. In the present study, GI scores increased during the intervention period, but no significant differences between groups were observed, which may speak against a clinical relevance of the reductions in plaque scores or else indicate that the intervention period was too short.
The exploratory outcomes assessed in the present RCT confirm a slightly delayed plaque formation in the active groups. The amounts of autofluorescent plaque, which have been associated with both biofilm maturity and virulence (Lennon et al. 2006; van der Veen et al. 2006), were lower in the active groups throughout the intervention period but only significantly reduced for active 2 after 7 d. Compared with placebo, the active groups showed a lower abundance of the late colonizers Capnocytophaga, Fusobacterium, and Tannerella and a trend toward reduced species richness, which may indicate a delayed plaque maturation caused by the enzymes (Kistler et al. 2013). These findings are in agreement with Adams et al. (2017), who observed a shift toward health-associated species in plaque samples after 3 mo of enzyme treatment. After 14 d of plaque formation, Airfloss treatment was administered based on the hypothesis that plaque subjected to enzymatic treatment was destabilized and could be more easily removed mechanically. However, hardly any biofilm was removed in any of the groups, which indicates either that the plaque was not destabilized or that the Airfloss device has limited ability to remove plaque. The latter is consistent with a previous report on interproximal plaque removal (Worthington et al. 2019).
Active 1 and 2 performed similarly across the investigated parameters. It is therefore not likely that increased plaque removal can be achieved with higher enzyme concentrations. Study participants were not allowed to perform regular oral hygiene during the trial, which led to the build up of thick plaque deposits. Possibly, the enzymes were not able to penetrate the biofilm and reach their targets, or else they did not specifically target all matrix components that are important for biofilm stability. In addition to glucans and eDNA, the dental plaque matrix harbors a wide range of other polymers that contribute to mechanical stability (Tawakoli et al. 2017; Jakubovics et al. 2021; Dige et al. 2022), and a 3-enzyme combination may not be sufficient to degrade this stabilizing scaffold. A recent large-scale screening study identified several promising enzyme combinations capable of removing 48-h old in situ–grown dental biofilms (Nielsen et al. 2024). The study achieved in situ biofilm removal with 3 enzyme formulations that contained 5 to 6 different enzymes. In addition to a cocktail of carbohydrases such as mutanase, dextranase, and glucanase, these effective formulations all contained DNase, although it was previously shown to be inefficient at disrupting intraoral biofilms older than 7.5 h (Schlafer et al. 2017). In addition, 2 of the 3 formulations contained a lipase, which had a modest effect in vitro. Successful plaque removal may therefore require expanding the formulation to a more diverse cocktail of enzymes that cater to the diverse matrix composition in dental plaque. Future research should continue mapping the principal components of the dental biofilm matrix and explore the clinical potential of a multitargeted enzymatic treatment approach.
The present RCT had an intervention period of 14 d, which is in the high end for a de novo plaque formation study that prohibits self-performed oral hygiene. Study designs in which participants maintain their regular hygiene allow for the monitoring of the effect of therapeutic adjuncts over longer periods, and other clinical outcomes, such as caries incidence, can be investigated. In studies that use enzymatic treatment as an adjunct to mechanical cleaning, however, the participants’ toothbrushing habits and skills are a considerable confounder, as the primary intended treatment effect is reduced plaque formation.
In conclusion, enzyme-containing lozenges were well tolerated and may serve as a supplement to regular mechanical plaque removal. While plaque accumulation after 1 d of intervention was not significantly affected, the enzymes showed the potential to delay dental plaque accumulation over time. The treatment effect may be improved by expanding the diversity of enzymes in the formulation.
Author Contributions
P.D. Rikvold, S. Schlafer, contributed to conception, design, data acquisition, analysis, and interpretation, drafted and critically revised the manuscript; K.K. Johnsen, Y. Chokyu Del Rey, contributed to conception, design, data acquisition, analysis, and interpretation, critically revised the manuscript; L.B.S. Hansen, contributed to analysis and interpretation, critically revised the manuscript; I. Knap, C. Holz, R.L. Meyer, contributed to conception, critically revised the manuscript; M.R. Jørgensen, contributed to conception, design, data interpretation, critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work.
Supplemental Material
sj-docx-1-jdr-10.1177_00220345251347959 – Supplemental material for The Effect of Enzymes on Dental Plaque: A Randomized Controlled Trial
Supplemental material, sj-docx-1-jdr-10.1177_00220345251347959 for The Effect of Enzymes on Dental Plaque: A Randomized Controlled Trial by P.D. Rikvold, K.K. Johnsen, Y. C. Del Rey, L.B.S. Hansen, I. Knap, C. Holz, R.L. Meyer, M.R. Jørgensen and S. Schlafer in Journal of Dental Research
Footnotes
Acknowledgements
The authors would like to give special thanks to Jouni Junnila and Jonathan Jaeger for performing statistical analyses. Lene Grønkjær, Anette Aakjær Thomsen, Javier E. Garcia, Charlotte K. Vindbjerg, Sussi B. Eriksen, Matthias Beck, and Dirk Leonhardt are acknowledged for their excellent technical assistance. The authors would also like to thank Manish K. Tiwari, Thomas Durhuus, and Delphine Saulnier for their support.
Declaration of Conflicting Interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Pernille D. Rikvold is an industrial PhD student at Aarhus University, co-funded by the Innovation Fund Denmark (9065-00244B) and Novozymes A/S. Lea B. Hansen, Caterina Holz, and Inge Knap are employed at Novozymes A/S (part of the Novonesis group), and Mette R. Jørgensen was also employed here during the study period. A patent application (WO 2023/110900) filed by Novozymes A/S is related to this work. 16S rRNA gene sequencing analysis was performed by Lea B. Hansen, and all other statistical analyses were performed by Jouni Junnila and Jonathan Jaeger, external consultants hired by Novozymes A/S.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by Novozymes A/S and by the Innovation Fund Denmark (9065-00244B). The funders had no influence on data collection or the decision to publish.
A supplemental appendix to this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
