Background The relationship between periodontitis (PD) and rheumatoid arthritis (RA) is well established but there is a lack of (large observational studies including) validated measurements of PD.
Objectives To evaluate a patient reported PD-module that can be used to investigate the PD-status in large epidemiological studies.
Methods For 353 patients from an early arthritis cohort (<6 months symptom duration) patient-reported PD and dentists' assessments of PD status were available for analysis. 232 patients had further independent evaluation of their PD status using radiographs obtained from their dentists. Patient reports included 12 items in total. Assessments by the patients' dentists and an independent evaluation of radiographs were treated as gold standard with possible outcomes: no, mild, moderate or severe PD. Relevant patient-reported items were chosen using factor analysis, age was included as a covariate. Sensitivity, specificity and AUC were calculated for all 3 possible allocations “no” vs. “mild, moderate or severe” PD, “no or mild” vs. “moderate or severe” PD and “no, mild or moderate” vs. “severe” PD.
Results Patients had a mean age of 56 years and a mean number of teeth of 19. According to dentist assessment, 30% had no, 34% mild, 26% moderate and 10% severe PD. Patient-reported items included in the PD-score were number of teeth, receding of jaw bone, presence of gum pockets/loose teeth/receding gums and extraction of teeth because of gum pockets. The factor analysis resulted in the same sample of items no matter which gold standard was used. Choosing the cutoff that optimized sensitivity and specificity simultaneously, the score was used to determine patient's PD-status. With dentist's assessment as gold standard, severe PD could be identified very well while considering the evaluation using radiographs patients with no PD could be identified best. For both assessments of PD, number of teeth alone was also a good predictor with AUCs only slightly lower than those for the six item score.
Conclusions With good sensitivity and specificity for the detection of any vs. no PD the patient reported score is a helpful instrument. We recommend using it to further investigate the relationship between PD and clinical parameters of RA.
Acknowledgements CAPEA was supported by an unconditional grant from Pfizer. The periodontitis project was supported by the German League of Rheumatology.
Disclosure of Interest : None declared