Reliability of two measurement indices for gingival enlargement

Abstract

Background and objective: The objective of this study was to analyze the concordance of the vertical gingival overgrowth index (GOi) and the horizontal Miranda & Brunet index (MBi) and to compare their reliability and reproducibility for an early diagnosis of gingival enlargement. A wide range of methods has been employed to determine the severity of drug-induced gingival enlargement (DIGE) that has resulted in uncertainty with regard to the prevalence of this side effect. In recent studies, different indices have been used to grade DIGE. The large variability observed between studies and the differences between vertical and horizontal gingival-enlargement measurements could be the result of the use of nonreliable indices during the measurement process. Some indices involve invasive procedures that require many measurements, or even a data-processing system, while others are less convenient and technically expensive and complex. In previous studies we used two complementary indices - the vertical GOi and the horizontal MBi. The results of these studies found some differences between both indices, with the MBi rendering higher estimates of DIGE prevalence that was attributed to its greater sensitivity for the detection of minimal changes in gingival thickness. To our knowledge, there are no studies comparing different measurement indices for gingival enlargement that are supported by statistical concordance analysis. Material and methods: Twelve plaster casts from patients who had worn orthodontic brackets, and who had different degrees of chronic inflammatory gingival enlargement, were analyzed. Three previously trained examiners registered twice the degree of buccal overgrowth, using the GOi and MBi, in all cast models with a minimum interval of 7 d between the first and the second evaluation. In total, from each cast, measurements from 16 gingival sites were taken using the GOi, and from nine gingival units (mesial and distal sites measurements) using the MBi. Concordance analysis of the registered measurements (intra-examiner and among examiners) for each index and between indices was assessed using the nonweighted Kappa index with a confidence interval of 95%. Results: We obtained 648 values for the GOi and the MBi. The overall score 0 (indicating absence of enlargement) was 32.7% and 19.8% for GOi and MBi, respectively, score 1 (light/moderate) was 39.7% and 48.1%, and score 2 (severe) was 27.6% and 32.1%. Concordance analysis for each index showed intra-examiner Kappa values of 0.820 for the GOi and 0.830 for the MBi. Interexaminer Kappa values were 0.720 for the GOi and 0.770 for the MBi. Concordance between indices showed Kappa values for the same examiner of 0.600, whereas concordance among different examiners was 0.550. Discrepancies between indices indicated a systematic skew, with 79-82.1% of discrepancy associated with a higher value for the MBi compared with the GOi. Conclusion: Both gingival enlargement indices analyzed are reliable, complementary and applicable for measuring gingival overgrowth. However, the MBi shows, with fewer measurements, a greater sensitivity than the GOi for the detection of the early stages of gingival enlargement, being adequate for the screening of large populations at risk.