A Reappraisal of the Accuracy of the Tactile Method for the Detection of the Subgingival Cementoenamel Junction: An In Vivo Study.

Objectives:  This article reappraises the accuracy and factors associated with the detection of the cementoenamel junction (CEJ) using the tactile method.

Materials and methods:  A total of 111 tooth sites of 7 patients scheduled for flap surgery were selected for the study. The CEJ was detected in a blind manner using the conventional tactile method with a standard periodontal probe by a single, trained examiner. A custom-made stent was prepared to standardize the measurements and the distance from a fixed reference point on the stent to the CEJ was measured before (apparent CEJ) and after (real CEJ) opening a gingival flap. To evaluate the effect of local anesthesia (LA) on the measurement error, assessment with and without LA given prior to the measurement was also evaluated. The bone crest-CEJ distance at each site was also recorded in all sites.

Statistical analysis:  The measurement error of apparent versus real distance, if any, was compared using Cohen's weighted kappa coefficient (WKC) (± 1 mm).

Results:  A weak WKC (WKC = 0.539) was found between the apparent and real CEJ distance. Higher WKCs were noted at posterior and proximal sites than the anterior and buccal/lingual sites, respectively (0.840 and 0.545 vs. 0.475 and 0.488). A higher confluence of the agreements was noted when CEJ distance was measured in anesthetized sites (WKC = 0.703). Sites without bone loss showed more coronal deviation of CEJ detection, as opposed to apical deviation seen at sites with bone loss.

Conclusion:  The conventional CEJ detection using the tactile method was relatively imprecise depending on the anatomical location of the tooth and the bone loss at the site of measurement. However, the detection accuracy improved when the sites were anesthetized. In clinical terms, our data, reported here for the first time imply that, in the absence of visual cues, posterior tooth site measurements of periodontal attachment loss were more reliable in comparison to the other sites. The bone crest level also impacted the measurement deviation to some extent, implying that, possible overestimate of clinical attachment loss may occur at sites without bone loss.