IADR Abstract Archives
Periodontal Profile Classes Predict Periodontitis Progression and Tooth Loss
Objectives: Previous clinical disease classifications have had limited utility for predicting either disease progression or tooth loss. Our goal was to develop data analytical tools that enable the agnostic identification, and definition of distinct periodontal and tooth profile classes (PPC/TPC) that incorporate missing teeth patterns and better predict disease progression and tooth loss than traditional disease definitions and summary scores.
Methods: Full-mouth clinical periodontal measurements (7-indices) from 6,793 subjects from the Dental Atherosclerosis Risk in Communities Study (DARIC) were used to identify PPC/TPC. A custom Latent Class Analysis (LCA) procedure was developed to identify seven distinct PPCs that defined person-level disease patterns and severity and 7 TPC that created 7 categories of tooth-level disease. LCA procedures assigned each subject into one of 7 PPCs and composite scores were computed for each subject that combined for each subject the scoring of existing teeth by TPC that was weighted by 10-year risk for tooth loss using the DARIC data for model development. The NHANES (2009-2010/2011-2012) and the Piedmont 65+ study samples were used for validation using a total of 7,785 subjects.
Results: LCA analyses grouped study participants into seven distinct periodontal profile classes (PPC A-G). PPC-D (moniker) and PPC-G (moniker) showed the highest risk for tooth loss Relative Risk RR=3.62 (2.61-5.01) and RR=3.83 (2.87-5.12), respectively, which were higher than the CDC/AAP severe classification RR=2.82 (1.97-4.04). PPC-G also showed the highest risk for periodontitis progression RR=7.82 (2.96-20.7). Class assignment was robust with small misclassification error in the presence of missing data.
Conclusions: These findings suggest that periodontal/tooth profile classes derived by LCA can provide not only robust periodontal clinical definitions that reflect disease patterns in the population, but this precise patient stratification can provide accurate estimates of person- and tooth type-specific periodontitis progression as well as 10-year tooth loss.
Methods: Full-mouth clinical periodontal measurements (7-indices) from 6,793 subjects from the Dental Atherosclerosis Risk in Communities Study (DARIC) were used to identify PPC/TPC. A custom Latent Class Analysis (LCA) procedure was developed to identify seven distinct PPCs that defined person-level disease patterns and severity and 7 TPC that created 7 categories of tooth-level disease. LCA procedures assigned each subject into one of 7 PPCs and composite scores were computed for each subject that combined for each subject the scoring of existing teeth by TPC that was weighted by 10-year risk for tooth loss using the DARIC data for model development. The NHANES (2009-2010/2011-2012) and the Piedmont 65+ study samples were used for validation using a total of 7,785 subjects.
Results: LCA analyses grouped study participants into seven distinct periodontal profile classes (PPC A-G). PPC-D (moniker) and PPC-G (moniker) showed the highest risk for tooth loss Relative Risk RR=3.62 (2.61-5.01) and RR=3.83 (2.87-5.12), respectively, which were higher than the CDC/AAP severe classification RR=2.82 (1.97-4.04). PPC-G also showed the highest risk for periodontitis progression RR=7.82 (2.96-20.7). Class assignment was robust with small misclassification error in the presence of missing data.
Conclusions: These findings suggest that periodontal/tooth profile classes derived by LCA can provide not only robust periodontal clinical definitions that reflect disease patterns in the population, but this precise patient stratification can provide accurate estimates of person- and tooth type-specific periodontitis progression as well as 10-year tooth loss.