Comparison of the efficacy of periodontal prognostic systems in predicting tooth loss

Association between asthma and periodontitis: A case-control analysis of risk factors, related medications, and allergic responses

AIMS

This study aimed to investigate the association between asthma, related allergies and medication use, and the presence and severity of periodontitis among individuals at the University of Michigan School of Dentistry.

METHODS

Employing a case-control design, the study analyzed data from 892 patients, half with asthma and half without asthma. Data collection included demographics, asthma history, medication use, allergies, and periodontal examination outcomes, including probing pocket depth (PPD), mobility, furcation involvement, and radiographic bone loss (RBL). Logistic regression models assessed the relationship between asthma and periodontitis, adjusting for confounders.

RESULTS

Asthmatic patients exhibited significantly lower odds of periodontitis (OR = 0.10, p < .001) and were less likely to present with advanced stages (OR = 0.23, p < .001) and grades of the disease (OR = 0.31, p < .001) compared to non-asthmatic patients. The study also found a higher proportion of females in the asthmatic group (67% vs. 51.8%, p < .001). Smoking was identified as a significant factor associated with periodontitis in patients with asthma, with former smokers at more than double the odds (OR = 2.28, p = .035) and current smokers at a slightly lower yet significant odds (OR = 1.87, p = .050). Additionally, asthmatic patients on adrenergic inhalers had an increased likelihood of developing periodontitis (OR = 1.76, p = .045). Allergies to codeine and latex were associated with higher odds of periodontitis, with ORs of 3.41 and 6.09, respectively.

CONCLUSIONS

Asthma was found to be associated with lower odds of periodontitis. However, this association appears to be modified by smoking habits and the use of certain asthma medications, which are related to an increased likelihood of periodontitis among asthmatic patients.

Systematic Review of Prognosis Models in Predicting Tooth Loss in Periodontitis

This study reviews and appraises the methodological and reporting quality of prediction models for tooth loss in periodontitis patients, including the use of regression and machine learning models. Studies involving prediction modeling for tooth loss in periodontitis patients were screened. A search was performed in MEDLINE via PubMed, Embase, and CENTRAL up to 12 February 2022, with citation chasing. Studies exploring model development or external validation studies for models assessing tooth loss in periodontitis patients for clinical use at any time point, with all prediction horizons in English, were considered. Studies were excluded if models were not developed for use in periodontitis patients, were not developed or validated on any data set, predicted outcomes other than tooth loss, or were prognostic factor studies. The CHARMS checklist was used for data extraction, TRIPOD to assess reporting quality, and PROBAST to assess the risk of bias. In total, 4,661 records were screened, and 45 studies were included. Only 26 studies reported any kind of performance measure. The median C-statistic reported was 0.671 (range, 0.57-0.97). All studies were at a high risk of bias due to inappropriate handling of missing data (96%), inappropriate evaluation of model performance (92%), and lack of accounting for model overfitting in evaluating model performance (68%). Many models predicting tooth loss in periodontitis are available, but studies evaluating these models are at a high risk of bias. Model performance measures are likely to be overly optimistic and might not be replicated in clinical use. While this review is unable to recommend any model for clinical practice, it has collated the existing models and their model performance at external validation and their associated sample sizes, which would be helpful to identify promising models for future external validation studies.

Influence of residual pockets on periodontal tooth loss: A retrospective analysis

BACKGROUND

Individuals enrolled in supportive periodontal therapy (SPT) can still present with tooth loss due to periodontitis (TLP). There is limited evidence on the influence of residual pockets (RPc) and a defined "threshold" at which a patient's profile is set to be at high risk for TLP in the literature. Therefore, this study aimed to assess the influence of RPc on TLP and determine the prognostic performance of RPc compared to the staging and grading of periodontitis on TLP risk.

METHODS

Clinical data from 168 patients (3869 teeth) treated for periodontitis and receiving SPT for at least 10 years were evaluated in this retrospective study. TLP and the percentage of sites with RPc ≥ 5 mm or ≥6 mm per patient were collected. The prognostic performance of RPc was compared to the staging and grading of the disease on TLP using a multilevel Cox proportional hazard regression model.

RESULTS

Over a median follow-up of 25 years, 13.7% of teeth were lost, 4.6% of which were due to periodontitis. Most patients with TLP had ≥1 site with RPc ≥5 mm (90.8%) or ≥6 mm (77.6%). Multivariate multilevel Cox regression revealed that patients with >15% of sites with RPc ≥5 mm had a hazard ratio of 2.34, and grade C had a hazard ratio of 4.6 for TLP compared to RPc ≤4 mm/grade A. Grading exhibited the best discrimination and model fit.

CONCLUSION

Patients with RPc ≥5 mm at >15% of the sites are at risk for tooth loss. Grading and RPc ≥5 mm displayed very good predictive capability of TLP.

Tooth loss during supportive periodontal care: A prospective study

AIM

To assess periodontal stability and the association between tooth- and patient-related factors and tooth loss during supportive periodontal care (SPC).

MATERIALS AND METHODS

A prospective observational study was carried out on previously treated periodontitis patients followed up for 5 years in SPC. The risk profile (low, moderate, high) of each patient based on periodontal risk assessment (PRA) scoring at baseline was evaluated, and tooth loss rates were analysed.

RESULTS

Two hundred patients were included in the study, and 143 had 5-year follow-up data available for analysis. The overall annual tooth loss per patient was 0.07 ± 0.14 teeth/patient/year. Older age, smoking, staging and grading were associated with increased tooth loss rates. Most patients whose teeth were extracted belonged to the PRA high-risk group. Both PRA and a tooth prognosis system used at baseline showed high negative predictive value but low positive predictive value for tooth loss during SPC.

CONCLUSIONS

Overall, the tooth loss rate of periodontitis patients in this prospective cohort study under SPC in private practice was low. Both tooth-based and patient-based prognostic systems can identify high-risk cases, but their positive predictive value should be improved.

Probing pocket depth reduction after non-surgical periodontal therapy: Tooth-related factors

BACKGROUND

To investigate tooth-related factors that influence the reduction of probing pocket depths (PPD) after non-surgical periodontal therapy (NST).

METHODS

Seven hundred forty-six patients with a total of 16,825 teeth were included and retrospectively analyzed. PPD reduction after NST was correlated with the tooth-related factors; tooth type, number of roots, furcation involvement, vitality, mobility, and type of restoration; using logistic multilevel regression for statistical analysis.

RESULTS

NST was able to reduce probing depth overall stratified probing depths (1.20 ± 1.51 mm, p ≤ 0.001). The reduction was significantly higher at teeth with higher probing depths at baseline. At pockets with PPD ≥ 6 mm, PPD remains high after NST. Tooth type, number of roots, furcation involvement, vitality, mobility, and type of restoration are significantly and independently associated with the rate of pocket closure.

CONCLUSIONS

The tooth-related factors: tooth type, number of roots, furcation involvement, vitality, mobility, and type of restoration had a significant and clinically relevant influence on phase I and II therapy. Considering these factors in advance may enhance the prediction of sites not responding adequately and the potential need for additional treatment, such as re-instrumentation or periodontal surgery, to ultimately achieve the therapy end points.

Prevalence of stable and successfully treated periodontitis subjects and incidence of subsequent tooth loss within supportive periodontal care: A systematic review with meta-analyses

AIM

To identify (i) the prevalence of meeting the endpoints of 'stable periodontitis' (probing pocket depth [PPD] ≤ 4 mm, bleeding on probing [BoP] < 10%, no BoP at 4 mm sites), 'endpoints of therapy' (no PPD > 4 mm with BoP, no PPD ≥ 6 mm), 'controlled periodontitis' (≤4 sites with PPD ≥ 5 mm), 'PPD < 5 mm' and 'PPD < 6 mm' at the start of supportive periodontal care [SPC]) and (ii) the incidence of tooth loss in relation to not meeting these endpoints within a minimum of 5 years of SPC.

MATERIALS AND METHODS

Systematic electronic and manual searches were conducted to identify studies where subjects, upon completion of active periodontal therapy, entered into SPC. Duplicate screening was performed to find relevant articles. Corresponding authors were contacted to confirm inclusion and retrieve required clinical data for further analyses to assess the prevalence of reaching endpoints and incidence of subsequent tooth loss, if available, within at least 5 years of SPC. Meta-analyses were carried out to evaluate risk ratios for tooth loss in relation to not reaching the various endpoints.

RESULTS

Fifteen studies including 12,884 patients and 323,111 teeth were retrieved. Achievement of endpoints at baseline SPC was rare (1.35%, 11.00% and 34.62%, respectively, for 'stable periodontitis', 'endpoints of therapy' and 'controlled periodontitis'). Less than a third of the 1190 subjects with 5 years of SPC data lost teeth-a total of 3.14% of all teeth were lost. Statistically significant associations with tooth loss, at the subject-level, were found for not achieving 'controlled periodontitis' (relative risk [RR] = 2.57), PPD < 5 mm (RR = 1.59) and PPD < 6 mm (RR = 1.98).

CONCLUSIONS

An overwhelming majority of subjects and teeth do not achieve the proposed endpoints for periodontal stability, yet most periodontal patients preserve most of their teeth during an average of 10-13 years in SPC.

Prognostic value of a composite outcome measure for periodontal stability following periodontal regenerative treatment: A retrospective analysis at 4 years

BACKGROUND

Recently, a composite outcome measure (COM) was proposed to describe the short-term results of periodontal regenerative treatment. The present retrospective study aimed at evaluating the prognostic value of COM on clinical attachment level (CAL) change over a 4-year period of supportive periodontal care (SPC).

METHODS

Seventy-four intraosseous defects in 59 patients were evaluated at 6 months and 4 years following regenerative treatment. Based on 6-month CAL change and probing depth (PD), defects were classified as: COM1 (CAL gain ≥3 mm, PD ≤4 mm); COM2 (CAL gain <3 mm, PD ≤4 mm); COM3 (CAL gain ≥3 mm, PD >4 mm); or COM4 (CAL gain <3 mm, PD >4 mm). COM groups were compared for "stability" (i.e., CAL gain, no change in CAL or CAL loss <1 mm) at 4 years. Also, groups were compared for mean change in PD and CAL, need for surgical retreatment, and tooth survival.

RESULTS

At 4 years, the proportion of stable defects in COM1, COM2, COM3, and COM4 group was 69.2%, 75%, 50%, and 28.6%, respectively, with a substantially higher probability for a defect to show stability for COM1, COM2, and COM3 compared with COM4 (odds ratio 4.6, 9.1, and 2.4, respectively). Although higher prevalence of surgical reinterventions and lower tooth survival were observed in COM4, no significant differences were detected among COM groups.

CONCLUSIONS

COM may be of value in predicting CAL change at sites undergoing SPC following periodontal regenerative surgery. Studies on larger cohorts, however, are needed to substantiate the present findings.

The prognostic prediction of periodontal non-surgery therapy in periodontitis patients based on surface-enhanced Raman measurements of pre-treatment saliva

Periodontitis is one of the most prevalent dental diseases, and the patients with periodontitis often suffer from refractory periodontitis or recurrence of disease due to improper or inadequate treatment. In clinical practice, the early and accurate assessment of post-treatment prognosis in periodontitis patients is always very important in order to implement timely interventions. In this study, a pre-treatment saliva SERS based prognostic protocol was explored to predict the prognosis of periodontal non-surgery therapy in periodontitis patients. According to the biomolecular analysis, significant differences in the levels of ascorbic acid, uric acid and glutathione are observed between good prognosis group and poor prognosis group, which are expected to serve as potential prognostic markers. Furthermore, high accuracy, sensitivity and specificity can also be achieved by using the proposed prognostic model. The excellent performance of the proposed method has demonstrated its potential for fast, accurate, and non-invasive prognostic prediction of periodontal non-surgery therapy in periodontitis patients, even at the time before implementing treatment, thus is expected to benefit timely and rational guidance on clinical interventions.

Development and international validation of logistic regression and machine-learning models for the prediction of 10-year molar loss

AIM

To develop and validate models based on logistic regression and artificial intelligence for prognostic prediction of molar survival in periodontally affected patients.

MATERIALS AND METHODS

Clinical and radiographic data from four different centres across four continents (two in Europe, one in the United States, and one in China) including 515 patients and 3157 molars were collected and used to train and test different types of machine-learning algorithms for their prognostic ability of molar loss over 10 years. The following models were trained: logistic regression, support vector machine, K-nearest neighbours, decision tree, random forest, artificial neural network, gradient boosting, and naive Bayes. In addition, different models were aggregated by means of the ensembled stacking method. The primary outcome of the study was related to the prediction of overall molar loss (MLO) in patients after active periodontal treatment.

RESULTS

The general performance in the external validation settings (aggregating three cohorts) revealed that the ensembled model, which combined neural network and logistic regression, showed the best performance among the different models for the prediction of MLO with an area under the curve (AUC) = 0.726. The neural network model showed the best AUC of 0.724 for the prediction of periodontitis-related molar loss. In addition, the ensembled model showed the best calibration performance.

CONCLUSIONS

Through a multi-centre collaboration, both prognostic models for the prediction of molar loss were developed and externally validated. The ensembled model showed the best performance in terms of both discrimination and validation, and it is made freely available to clinicians for widespread use in clinical practice.

Machine Learning in Predicting Tooth Loss: A Systematic Review and Risk of Bias Assessment

Predicting tooth loss is a persistent clinical challenge in the 21st century. While an emerging field in dentistry, computational solutions that employ machine learning are promising for enhancing clinical outcomes, including the chairside prognostication of tooth loss. We aimed to evaluate the risk of bias in prognostic prediction models of tooth loss that use machine learning. To do this, literature was searched in two electronic databases (MEDLINE via PubMed; Google Scholar) for studies that reported the accuracy or area under the curve (AUC) of prediction models. AUC measures the entire two-dimensional area underneath the entire receiver operating characteristic (ROC) curves. AUC provides an aggregate measure of performance across all possible classification thresholds. Although both development and validation were included in this review, studies that did not assess the accuracy or validation of boosting models (AdaBoosting, Gradient-boosting decision tree, XGBoost, LightGBM, CatBoost) were excluded. Five studies met criteria for inclusion and revealed high accuracy; however, models displayed a high risk of bias. Importantly, patient-level assessments combined with socioeconomic predictors performed better than clinical predictors alone. While there are current limitations, machine-learning-assisted models for tooth loss may enhance prognostication accuracy in combination with clinical and patient metadata in the future.