Multivariate Regression Models for Predicting the Prognosis of Luxation Injuries of Permanent Teeth: Periodontal and Pulp Analyses

Abstract

Background

This study aimed to analyse the relationships between injury variables, treatment variables and the status of the periodontium or pulp of luxated teeth.

Materials and Methods

The electronic medical records and imaging data of patients who underwent treatment for luxation injuries of permanent teeth in the paediatric dentistry and dental emergency departments of the Stomatology Hospital, Zhejiang University School of Medicine between 1 January 2012 and 1 January 2022 were retrospectively analysed. First, the patients' records were reviewed to collect background and follow-up imaging data; then, all early clinical variables were reviewed and used to construct models to predict the periodontal and pulp status of the luxated teeth. The models included the significant factors identified in the univariate analysis, and multivariate analysis was performed to evaluate the relationships between the variables and the final status of the periodontium or dental pulp. The chi-square test and Fisher's exact test were employed to analyse the differences in the frequency of each variable. The variables were systematically screened based on their odds ratios, and significance levels were determined by Pearson's chi-square test.

Results

A total of 188 patients were initially identified and ultimately included. The age range of the patients was 7–56 years. Logistic regression models for periodontal and pulp prognosis of the luxated teeth were established. The models were refined by combining the results of feature selection, parameter testing and goodness-of-fit testing. The final model included four variables and accurately predicted the periodontal outcome in 65.79% of the cases. The prognostic model for the dental pulp included three significant factors and had an overall accuracy of 94.59%.

Conclusion

The prognostic models developed to predict the influence of various factors on the status of the periodontium and dental pulp of luxated teeth demonstrated notable accuracy and practical utility. Therefore, these models are potentially valuable tools for long-term prognostic assessments.

Trial Registration

Approval no. ChiCTR2100044897