Anti-Caries Agents and Dental Caries Among High-Risk Adults
Anti-Caries Agents and Dental Caries Among High-Risk Adults
This study was one of the few to evaluate caries outcomes among high-risk adults following delivery of non-operative preventives, and to our knowledge, the first study to examine outcomes according to the frequency with which preventives were received. Delivery of non-operative anti-caries agents at multiple visits was associated with a 19 % reduction in DFT increment but no statistically significant difference in the incidence of new decay (DFT increment >0). This indicates that in high-risk populations, anti-caries agents might be most useful for reducing disease severity among caries affected individuals, as opposed to preventing caries entirely. In a recently reported randomized controlled trial of high caries risk adults, a combined antibacterial and fluoride therapy intervention resulted in no difference in the probability of having no incremental lesions in a zero-inflated Poisson model but yielded a statistically significant 24 % decrease in DMFS increment for the count portion of the model. Similarly, a community randomized trial featuring fluoride varnish applications for Aboriginal children in Australia reported a prevented fraction for incremental lesions of approximately 30 % but no statistically significant change in the prevalence of children with ≥1 affected teeth.
In a similar retrospective study of administrative data from two large US dental health plans, in one plan, a formal recommendation for at-home fluoride treatment for high-risk patients was associated with a non-statistically significant 11 % reduction in caries increment. However, in the second dental plan, high-risk patients who received in-office topical fluoride demonstrated higher caries increment than high-risk patients given no therapy. The authors speculated that dentists in this plan "further stratified" patients within the high risk category, choosing to deliver preventive therapy more often to a subset of high-risk patients that they deemed even more likely to experience future decay. In the present study, we attempted to account for such confounding by indication by adjusting for the number of decayed teeth at baseline, and in the adjusted analysis, we did observe a statistically significant difference in DFT increment according to therapy received.
While receipt of anti-caries agents at multiple visits was associated with lower caries incidence, equivalent to slightly more than one decayed/restored tooth prevented over 18 months for every three patients treated, more than half of these high-risk patients did not receive any form of anti-caries agent. Whether less than universal use of non-operative therapies reflects reticence on the part of providers, patients, or both, our results suggest that greater prevention could be achieved if non-operative therapies were more widely utilized.
Notably, in the exploratory subgroup analysis, having dental benefits through a public program was associated with the greatest reduction in DFT increment with repeated delivery of anti-caries agents. A special arrangement between the dental school and the administrator of the state Medicaid dental program made it possible for the university clinic to collect reimbursement for risk-based preventive treatments and to provide them at no charge to the patients who agreed to accept them. Therefore, it is plausible that this reimbursement mechanism eased the way for more intensive preventive therapy, both in terms of the frequency of delivery and the number of different types of products provided, potentially leading to a larger impact among these patients.
We had no measure of patient adherence to recommended regimens for home-use anti-caries agents. We speculate that poor patient adherence accounts for the lack of anti-caries effectiveness associated with one-time therapy. In contrast, we hypothesize that patients who were dispensed agents on multiple occasions reflect adherence patterns consistent with continuing home-use and return for agent replenishment. The vast majority of patients who received anti-caries agents on multiple occasions were given more than one type of agent (fluoride, xylitol, or chlorhexidine), which follows the documented protocol emphasized in this clinic. Thus, it was not possible to determine if any one agent was most effective. The CAMBRA approach, which aims both to decrease pathological factors (antibacterial therapy) and simultaneously to enhance preventive or reparative therapy (e.g., via high concentration fluoride product), likely operates through multiple mechanistic pathways.
Harnessing routinely collected data from electronic health records for clinical research presents challenges but also promises to expand clinical research capacity. Data analyzed in this study were not collected specifically for use in research. For example, student providers did not undergo a specific calibration exercise in caries detection, and there was no rigid methodology applied to treatment planning or caries risk assessment, although all providers were part of the same educational program, which teaches and emphasizes CAMBRA, and at the time of this study, allowed for radiographic, visual, and tactile methods to be used in caries detection. Also, it is possible that our calculation of the DFT increment included some restorations that were placed for reasons other than dental caries, leading to an overestimation of caries occurrence in all comparison groups. Such limitations were partly balanced by access to a large analytic sample that reflects realistic treatment decisions made outside the context of a formal intervention study.
Further research is required to determine whether the results observed in this study can be generalized beyond this educational clinic, which predominantly serves lower-income patients at high caries risk and in which dental students are primarily responsible for diagnostic and preventive care. Additionally, most patients observed at baseline did not return to the clinic for a follow-up examination, which could also decrease the generalizeability of our findings. However, we implemented inverse probability censoring weighting to account for differences in measured characteristics between the baseline and follow-up samples. We did not evaluate therapy outcomes among low, moderate, or extreme risk patients due to the smaller number of patients in these categories, particularly in the extreme risk group: a category marked by severe hypo-salivation and for which guidelines suggest intensive preventive care in multiple forms, which may surpass the level of prevention provided to high-risk patients in this study. Furthermore, consistent with CAMBRA guidelines, relatively few low- and moderate-risk patients received anti-caries agents.
We obtained adjusted outcome values through implementation of a doubly-robust version of the g-computation estimator, a technique rarely applied in oral health research, despite increasingly common use in epidemiology, generally. An attractive aspect of this approach is the ease of interpretation: estimates take the form of expected caries outcomes associated with each category of interest under equal covariate distributions. However, as with all observational studies, analyses must account for confounding variables, and it is possible that unmeasured factors could have affected the results.
Discussion
This study was one of the few to evaluate caries outcomes among high-risk adults following delivery of non-operative preventives, and to our knowledge, the first study to examine outcomes according to the frequency with which preventives were received. Delivery of non-operative anti-caries agents at multiple visits was associated with a 19 % reduction in DFT increment but no statistically significant difference in the incidence of new decay (DFT increment >0). This indicates that in high-risk populations, anti-caries agents might be most useful for reducing disease severity among caries affected individuals, as opposed to preventing caries entirely. In a recently reported randomized controlled trial of high caries risk adults, a combined antibacterial and fluoride therapy intervention resulted in no difference in the probability of having no incremental lesions in a zero-inflated Poisson model but yielded a statistically significant 24 % decrease in DMFS increment for the count portion of the model. Similarly, a community randomized trial featuring fluoride varnish applications for Aboriginal children in Australia reported a prevented fraction for incremental lesions of approximately 30 % but no statistically significant change in the prevalence of children with ≥1 affected teeth.
In a similar retrospective study of administrative data from two large US dental health plans, in one plan, a formal recommendation for at-home fluoride treatment for high-risk patients was associated with a non-statistically significant 11 % reduction in caries increment. However, in the second dental plan, high-risk patients who received in-office topical fluoride demonstrated higher caries increment than high-risk patients given no therapy. The authors speculated that dentists in this plan "further stratified" patients within the high risk category, choosing to deliver preventive therapy more often to a subset of high-risk patients that they deemed even more likely to experience future decay. In the present study, we attempted to account for such confounding by indication by adjusting for the number of decayed teeth at baseline, and in the adjusted analysis, we did observe a statistically significant difference in DFT increment according to therapy received.
While receipt of anti-caries agents at multiple visits was associated with lower caries incidence, equivalent to slightly more than one decayed/restored tooth prevented over 18 months for every three patients treated, more than half of these high-risk patients did not receive any form of anti-caries agent. Whether less than universal use of non-operative therapies reflects reticence on the part of providers, patients, or both, our results suggest that greater prevention could be achieved if non-operative therapies were more widely utilized.
Notably, in the exploratory subgroup analysis, having dental benefits through a public program was associated with the greatest reduction in DFT increment with repeated delivery of anti-caries agents. A special arrangement between the dental school and the administrator of the state Medicaid dental program made it possible for the university clinic to collect reimbursement for risk-based preventive treatments and to provide them at no charge to the patients who agreed to accept them. Therefore, it is plausible that this reimbursement mechanism eased the way for more intensive preventive therapy, both in terms of the frequency of delivery and the number of different types of products provided, potentially leading to a larger impact among these patients.
We had no measure of patient adherence to recommended regimens for home-use anti-caries agents. We speculate that poor patient adherence accounts for the lack of anti-caries effectiveness associated with one-time therapy. In contrast, we hypothesize that patients who were dispensed agents on multiple occasions reflect adherence patterns consistent with continuing home-use and return for agent replenishment. The vast majority of patients who received anti-caries agents on multiple occasions were given more than one type of agent (fluoride, xylitol, or chlorhexidine), which follows the documented protocol emphasized in this clinic. Thus, it was not possible to determine if any one agent was most effective. The CAMBRA approach, which aims both to decrease pathological factors (antibacterial therapy) and simultaneously to enhance preventive or reparative therapy (e.g., via high concentration fluoride product), likely operates through multiple mechanistic pathways.
Harnessing routinely collected data from electronic health records for clinical research presents challenges but also promises to expand clinical research capacity. Data analyzed in this study were not collected specifically for use in research. For example, student providers did not undergo a specific calibration exercise in caries detection, and there was no rigid methodology applied to treatment planning or caries risk assessment, although all providers were part of the same educational program, which teaches and emphasizes CAMBRA, and at the time of this study, allowed for radiographic, visual, and tactile methods to be used in caries detection. Also, it is possible that our calculation of the DFT increment included some restorations that were placed for reasons other than dental caries, leading to an overestimation of caries occurrence in all comparison groups. Such limitations were partly balanced by access to a large analytic sample that reflects realistic treatment decisions made outside the context of a formal intervention study.
Further research is required to determine whether the results observed in this study can be generalized beyond this educational clinic, which predominantly serves lower-income patients at high caries risk and in which dental students are primarily responsible for diagnostic and preventive care. Additionally, most patients observed at baseline did not return to the clinic for a follow-up examination, which could also decrease the generalizeability of our findings. However, we implemented inverse probability censoring weighting to account for differences in measured characteristics between the baseline and follow-up samples. We did not evaluate therapy outcomes among low, moderate, or extreme risk patients due to the smaller number of patients in these categories, particularly in the extreme risk group: a category marked by severe hypo-salivation and for which guidelines suggest intensive preventive care in multiple forms, which may surpass the level of prevention provided to high-risk patients in this study. Furthermore, consistent with CAMBRA guidelines, relatively few low- and moderate-risk patients received anti-caries agents.
We obtained adjusted outcome values through implementation of a doubly-robust version of the g-computation estimator, a technique rarely applied in oral health research, despite increasingly common use in epidemiology, generally. An attractive aspect of this approach is the ease of interpretation: estimates take the form of expected caries outcomes associated with each category of interest under equal covariate distributions. However, as with all observational studies, analyses must account for confounding variables, and it is possible that unmeasured factors could have affected the results.
Source...