Evaluation of Kharma scale as a predictor of lower third molar extraction difficulty

Background The Evaluation of the degree of lower third molar (L3M) extraction difficulty is extremely important for both clinicians and patients. This study aimed to evaluate the validity of a new index (Kharma scale) as a preoperative predictor index of the difficulty of surgical removal of impacted L3M. Material and Methods Extraction difficulty of a series of 49-impacted L3M was predicted preoperatively by Kharma scale, and postoperative difficulty was assessed with a modified Parant scale. Results The sensitivity of Kharma scale, as a predictor of difficulty, was 18.2% and the specificity was 68.4%. Likelihood ratios for the Kharma categories also indicated that the scale is of little value for predicting a difficult extraction. There was no significant association between the Kharma score and duration of operation, but high-modified Parant scores were significantly associated with longer operations. Conclusions The proposed Kharma scale was unreliable as preoperative predictor of the L3M extraction difficulty, and both radiological and clinical information must be taken into account. Key words:Kharma scale, prediction scale, lower third molar, extraction difficulty.


Introduction
Evaluation of the degree of lower third molar (L3M) extraction difficulty is extremely important to design correct treatment strategy and to reduce the risk of complications. Therefore, having optimal scale to predict L3M extraction difficulty continues to challenge clinicians. Several methods have been established for preoperative estimation of difficulty, but they found invalid (1)(2)(3)(4) or of limited clinical use (2,3,5,6). Among these scales, is the Pederson scale, which is widely cited in oral and maxillofacial surgical texts as a useful tool to predict the difficulty of extraction of L3M (1). However, diverse clinicians have questioned its performance (1,3). Recently, Kharma et al. (7) proposed a new scale; which is a modification of Pederson scale; that took into account the anatomical form of the tooth roots. They claimed that this new estimating index is more reliable and accurate than Pederson scale, and significantly correlated with postoperative difficulty rated by modified Parant scale (7). The aim of this study was to evaluate Kharma scale's prediction accuracy. Preoperatively, the surgeon predicted the difficulty of extraction from panoramic radiographs using Kharma scale. After the operation, difficulty was assessed using the modified Parant scale (MPS) ( Table 1). The duration of operation was also recorded by a stopwatch (from start of incision to final suture). By using descriptive statistics of IBM SPSS Statistics 23, sensitivity, specificity, and likelihood ratios were calculated considering the MPS as a reference. In addition, the correlation between the operative time and the difficulty of operation as proposed by both Kharma and modified Parant scales were also assessed by analysis of variance test considering a probability values less than 0.05 as significant.  Table 3. Predictive values (P1, P2) and likelihood ratios of the different Kharma categories for prediction of each modified Parant category.

Results
Likelihood ratios (LR), calculated as P1/P2. Significant likelihood ratios for prediction must be below 0.5 or above 2. P1: Predicted probability of difficulty. P2: Predicted probability of easiness.
which indicated that operations with longer duration was significantly correlated with higher modified Parant scores (P= .007). By contrast, no significant correlation was exist with Kharma scores (P= .716).

Discussion
Clinical and radiographical findings should be considered preoperatively for correct evaluation of L3M extraction. It helps in prediction of operative difficulty for correct treatment plant and improves the level of patients' satisfaction with the treatment received (1,8). Several methods were established for preoperative estimation of difficulty like Pederson scale, which was used by clinicians as a useful tool to predict of L3M extraction difficulty (1). However, The meta-analysis of the current literatures concluded that Pederson scale is not a reliable prediction index in L3M surgery (9). WHARFE scale (10) was also proposed, but is rarely used in practice duo to their complexity (1). Other variables had been considered in MRACBS scale (6) including L3M relation to inferior alveolar and lingual nerves. It is of limited clinical application due to the need to the cone beam computed tomography in classification. Santamaria et al. (11) points to the importance root patterns in determining L3M extraction difficulty. Other researchers (2,3) took into account the curvature, width and number of L3M roots in their difficulty prediction index.
Kharma scale proposed a new difficulty prediction index based on 4 factors: tooth angulation, the depth of the third molar in the mandible, the relationship with the ramus/space available, and root form (7). It is similar to Pederson index in that it measures the same parameters in addition to root forms, and close to Yuasa scale (3) as the former assess the same parameters in addition to tooth angulation. However, Kharma scale, in this study, reported 85.7% false +ve. and 25.7% false -ve. and showed a very low sensitivity (18.2%) and a specificity of 68.4%. In contrast, Yuasa scale in a preliminary study (3) recorded 8.3% false +ve and 15% false -ve. which resulted in high sensitivity (85%) and specificity (92%). The false -ve cases causing problems for both the practitioners and patients (3).
In this study, the MPS was considered as a reference standard index of surgical difficulty as it found reliable and consistent with operative difficulty by researchers (1,7,8,12). The results of this study was in agreement with previous studies where the MPS was found significantly correlated with surgical time (P=0.007). However, Kharma scale fail to correlate with P value of 0 .716.
The results indicate that the Kharma scale has poor sensitivity when over 85% of difficult extractions were not identified. In addition, likelihood ratios for prediction of the different difficulties of the Kharma scale from the categories of MPS also indicate that Kharma scale is of little value in predicting operative difficulty (a significant likelihood ratios for prediction must be below 0.5 or above 2) (13). This is may be owing to the lack of consideration of various relevant factors, such as bone density, periodontal space, flexibility of the cheek, and nerve proximity. In addition, curvature of the root is sometimes an unpredictable factor, as it is often not visible in panoramic radiographs (3).
In conclusion, and depending on the current results, the proposed Kharma scale was unreliable as preoperative predictor of the L3M extraction difficulty, and both radiological and clinical information must be taken into account.