Sexual dimorphism applying the mandibular canine index in a Brazilian sample: a pilot study

This research study aimed at applying the morphometric quantification of the canine index for sexual dimorphism in a Brazilian sample. This was an observational study and the convenience sample consisted of 90 adult participants (45 male subjects and 45 female subjects), aging from 18 to 35 years. With the aid of a digital caliper, the intraoral mesiodistal measurement of the permanent mandibular right canine (MD43) and the intercanine distance (IC) were taken by three examiners. The measurements were applied to the mandibular canine index formula to estimate sex based on the morphometric features of human canines. The applicability of this approach for sexual dimorphism was assessed based on the mandibular canine index (MCI) calculated by the formula. The MCI was higher in male than in female subjects. In the total sample, the MCI overall mean accuracy rate for sexual dimorphism was 52.22%. In male subjects, the MCI was able to properly differentiate sex in 82.22% of the sample, while in female subjects the accuracy rate decreased to 22.22%. These results call the attention to the careful use of MCI especially for Forensic Anthropology. In particular, the accuracy of the method was close to the random of a sample that contained both sexes. Thus, the MCI should not be used as the only tool for sexual dimorphism.

For optimal outcome, the MCI approach was tested for validation in several populations (Muller, Lupi-Pegurier, Quatrehomme, & Bolla, 2001;Acharya & Mainali, 2009;Paramkusam et al., 2014;Rajarathnam et al., 2016;Silva et al., 2016;Dixit, Sreedevi, Nagarajappa, & Bhashin, 2017;Gandhi et al., 2017;Kumawat et al., 2017;Mohsenpour, Gangadhar, & Samehsalari, 2017;Patel et al., 2017;Singh, Garg, & Singh, 2017) and different results were verified. For example, Kumawat et al. (2017) and Singh et al. (2017) have tested the MCI predicative value for sexual dimorphism in the Indian population and have verified accuracy values of 79.66 and 83.50% respectively. On the other hand, Mohsenpour et al. (2017) have tested the MCI in the Iranian population and found accuracy values of 53.00%, a result close to the one found by Silva et al. (2016), who have verified accuracy values of 54.20% in the Portuguese population. In view of these differences, it is important to test the applicability of the MCI in several populations, as cases presenting a lack or fragmentation of the pelvic bones would make it impossible to apply other methods for sexual estimation. In these situations, dental features can be assessed due to the hardness of teeth.
Some studies have investigated the morphometric information of human canines for sexual dimorphism in Brazilian samples, however, none of them was designed to specifically test the MCI approach (Costa et al., 2012;Fernandes, Sathler, Natalício, Henriques, & Pinzan, 2013;Sabóia et al., 2013;Silva et al.,2016). Based on the above discussion, this study aimed at testing the MCI approach, considering the hypothesis that sexual dimorphism is feasible and accurate, having as a sample mandibular canines of Brazilian male and female subjects.

Material and methods
This observational study was approved by the Human Research Ethics Committee (CAAE 67803717.1.0000.5419). Participants were prospectively selected and signed an informed consent form. The inclusion criteria consisted of Brazilian adults who, through clinical evaluation, presented the permanent mandibular right (Fédération Dentaire Internationale: #43) and left (Fédération Dentaire Internationale: #33) canines with no caries, cavities, fractures, restorations or prosthetic crowns. Participants with agenesis/absence or mandibular anterior crowding, supernumerary teeth, wear on the proximal surfaces of teeth, diastemas and severe parafunctional habits, such as bruxism and dental clenching, were excluded from this sample, as well as participants with syndromes.
Following the MCI approach, the MD and the IC measurements of tooth #43 were intraorally taken by three trained examiners, in each participant, three times. Thus, 18 measurements (9 MD and 9 IC) and 9 indexes (MCI) were obtained for each participant, since each examiner performed the analysis three times. The measurements were taken in millimeters using a caliper (WesternPro DC-6 150 mm, 0.3 mm/0.01, Western, China). The MD measurement was taken between the most prominent part of the mesial and distal surfaces of the crown ( Figure 1A). The IC was established as the distance between the tips of the mandibular canine teeth ( Figure 1B). The measurements were taken with the caliper display facing downwards and the values were assessed and registered by an external examiner, ensuring the blindness of the examiners.
Intra-and inter-examiner reproducibility tests were performed by using the Intraclass Correlation Coefficient (ICC). The ICC values for intra-and inter-examiner reproducibility were ≥ 0.75, confirming the correct training and calibration of the examiners. The MCI was calculated applying the following formula: The obtained threshold was used to classify male and female subjects. Correct and incorrect (accuracy) classifications were compared to the original sex of participants. Descriptive statistics was applied to quantify absolute (n) and relative (%) accuracy rates. The software package R (R foundation, Vienna, Austria) was used in the study.

Results and discussion
In Anthropology, the morphology of teeth and bones provides information about sex, age, height and ancestry of a deceased body. Sex figures as the first and, possibly, the most impacting anthropological feature as it places potential victims in two groups -male and female subjects. Consequently, sexual dimorphism is an important tool for the identification of skeletal remains. This pilot study has tested the MCI approach proposed by Rao et al. (1989) for sexual dimorphism in a Brazilian sample, as it is a simple and practical method of evaluation. Thus, the convenience sample comprised 90 participants (45 male subjects and 45 female subjects), aging from 18 to 35 years.
Although the MD distance of both mandibular canine teeth used in isolation presented a higher predictive value than the MCI in the research by Azevedo, Pereira, Gouveia, Tavares and Caldas (2019), the hypothesis of the current study firstly predicted the possibility of obtaining sexual dimorphism applying the original formula of MCI in a Brazilian sample. This hypothesis was confirmed by the MCI values from male subjects (0.255 ± 0.01) and female subjects (0.248 ± 0.01) ( Table 1).
In particular, the scientific literature corroborates the fact that morphometric differences in canine size may occur in millimeter levels, in which male subjects present larger teeth than female subjects (Sabóia et al., 2013;Mitsea, Moraitis, Leon, Nicopoulou-Karayianni, & Spiliopoulou, 2014;Sai Kiran et al., 2014;Shireen et al., 2014;Aggarwal et al., 2016;Pandey & Ma, 2016;Silva et al., 2016). However, it is important to consider that the opposite may occur. Female subjects were previously found with larger teeth (incisors, canines and premolars) than male subjects in specific populations such as in Nepal, Dharwad, Orissa and South India  In this research study, it was decided to use the MD distance of tooth #43 to calculate the MCI formula since this was the measure of the original method applied by Rao et al. (1989), in addition to subsequent studies comparing contralateral teeth, founding no statistically significant differences (Ayoub et al., 2014;Costa et al., 2012;Vijayan et al., 2019). The mean MD distance, in male and in female subjects and, the mean IC distance, in male and in female subjects, are shown in Table 1.
The MCI standard applied as a threshold for classifying males and females was 0.24. According to this threshold, the sample was correctly classified in 52.22% of the cases. Specific classification between male and female subjects is shown in Table 2. As shown in Table 2, the results of the present research study were not as satisfactory as the results found by Rao et al. (1989) who have verified accuracy rates reaching 84.3% in male subjects, 87.5% in female subjects, and 85.9% in general for India. Additional studies conducted in the country have also presented good accuracy rates, as Dixit et al. (2017), who have found overall accuracy rate of 71.00%; Gandhi et al. (2017), who have found accuracy rate of 76.66%; Kumawat et al. (2017), who have found accuracy rate of 79.66%; Patel et al. (2017), who have found 78.75%; and Singh et al. (2017), who have found an overall accuracy of 83.50%.
However, studies with samples from Iran (Mohsenpour et al., 2017) and Portugal (Silva et al., 2016) have reached overall lower accuracy rates of 53.00 and 54.20% respectively, and considered heterogeneity to the international scenario for the MCI approach. Thus, specific characteristics of the population for the canines may be the most acceptable explanation for the differences between the studies. This is the reason why validation studies are important prior to the application of a method in practice (Atreya et al., 2019).
In the literature, there are no previous studies considering the MCI in the Brazilian population. Therefore, the results from this study may not be representative for the entire Brazilian population -which is heterogeneous and with miscegenation from several European and African countries (Ferreira, Mendes, Wiezel, Luizon, & Simões, 2006;Prabhu & Acharya, 2009;Atreya et al., 2019). Thus, the miscegenation of the Brazilian population may explain the lower accuracy rate found in this research (Table 2).
Research studies applying the MCI in samples of the Brazilian population, considering different regions of the country and classified based on genetic analyzes of ancestry, could provide a better understanding of the effect of miscegenation on the odontometric measures in the Brazilian population (Ferreira et al., 2006;Prabhu & Acharya, 2009;Valle-Silva et al., 2019). However, not, disconsidering reverse dimorphism or monomorphism as a result of human evolution (Boaz & Gupta, 2009;Prabhu & Acharya, 2009).
A clear example of miscegenation in the Brazilian population can be observed comparing our results ( Table  1) to the studies of Sabóia et al. (2013) and Costa et al. (2012). In the former, the authors have observed a mean MD width of 6.9 mm for female subjects from Southeast Brazil, while the latter observed a similar value (6.99 mm) for male subjects from Northeast Brazil. These findings show that similar morphometric information of canines can be found between male subjects and female subjects from different regions in Brazil. In practice, it could cause uncertainty and, even, wrong identification. Therefore, the results of the present study contribute to practice by calling the attention to the application of the MCI. More specifically, it shows that there are morphometric differences between the permanent mandibular canines of male and female subjects, but the proposed approach does not have power to accurately classify men and women based on teeth, as demonstrated by Dony, Reddy and Kakodkar (2018).
Additional studies should be conducted with increased sample size, comparing populations from different countries, proposing specific statistical corrections of the population to best fit the MCI to each country, and even include samples of participants with crowding and diastemas in order to challenge the MCI approach with dental conditions that are commonly found by dentists and anthropologists. Thus, the development of adapted odontometric sexual dimorphism formulae considering specific populations should be considered.

Conclusion
The accuracy of the MCI approach ranged slightly above 50% when differentiating male from female subjects, having as basis the morphometric features of the permanent mandibular canines. This result suggests that, in practice, the MCI should not be considered as the only reliable method for sexual dimorphism.