Dental malocclusions are not just about small and weak bones: assessing the morphology of the mandible with cross-section analysis and geometric morphometrics

Dental malocclusions are considered a “disease of civilization” [1]. They are common in current western populations, and markedly less frequent in archaic human populations [2] whose diet was composed mainly of more mechanically challenging items like meat (non-highly processed), seeds and abrasive elements in them [3, 4]. They are also less frequent in modern populations with more ancestral lifestyles whose diet consists of less processed items, and where infections leading to premature tooth loss are less common [1, 5]. Tooth loss, and particularly a reduction in masticatory load, has been linked to the transition between hunter-gatherer and agricultural (also urban and modern) subsistence economies, causing a reduction in jaw size [6‐8]. Traditionally in dentistry and physical anthropology, small jaws have been related to alterations like tooth crowding and an altered maxillomandibular relationship [1, 9‐11].

Experimentally, it has been observed that diminished masticatory forces are capable of altering adult cranial and mandibular shape. In mice, for example, a diet based on soft food results in a shorter narrow rostrum, a short coronoid process, a small articular surface of the condyle and labially inclined incisors [12]. Moreover, inducing masseter muscle paralysis in adult mice leads to an increased activity of the bone resorption marker RANKL, followed by a reduction of bone tissue per area and trabecular thickness of the condyle head [13]. Using medical data from modern populations, it has been observed that muscle force is only weakly related to cranial shape [14], but more so to the shape of the mandible [15].

The developmental origin of dental malocclusions, particularly class II and III maxillomandibular relationships, has been difficult to pinpoint. There would be a genetic predisposition, to which habits and muscle malfunctioning would significantly contribute [16, 17]. Class II and III phenotypes are also very variable and thus difficult to relate to isolated factors. Class II individuals can vary in their vertical component as well as their occlusal relationship [16]. For example, Hong and Yi [18], using a modified Delaire’s cephalometric analysis, were able to identify six types of class III morphologies that relate to differences in the intra-individual arrangement of other cranial structures. The effect of function on mandible shape deformities would explain the differences in size found by Park, Kim, Yu, Yi and Lee [19] in the form of the functional units (i.e. areas affected by the surrounding functional tissues) of the mandible among skeletal classes I, II and III. However, the same study does not find differences in the size of the whole mandibles, and in some cases, the affected mandibles show larger linear dimensions in the functional areas.

One of the features that are common in class II and III mandibles is the vertical elongation of the face, either more anteriorly, posteriorly or both [20, 21]. From a structural perspective, a gracile vertically elongated mandible, reduced in bone mass, is a weak mandible [22] due to its reduced capacity to withstand torsional forces that arise during mastication [23, 24]. However, mandibular robusticity, estimated from external linear measurements and measurements of cortical bone amount and inner distribution, is not always a trait present in individuals exerting large masticatory loads [25].

Are then class II and III mandibles small and weak due to their reduced masticatory load, which is a characteristic of modern populations? Due to the nature of linear measurements, of common use in dentistry, the three-dimensional geometry of class II and III mandibles is not completely known. We aim to perform a detailed characterisation of the shape, size (in relation to shape, i.e. allometry) and inner morphology of mandibles affected by class II and III malocclusions (in the broad sense). We combine 3D shape analysis techniques (geometric morphometrics) and traditional cross-section analyses of the internal amount and distribution of cortical bone. Using a novel approach in dentistry that combines knowledge from the dental and anthropological sciences, we use archaeological and current populations that differ in their subsistence strategy (and thus diet consistency) and maxillomandibular relationship to test the null hypothesis that there are no differences in mandibular shape, amount and distribution of cortical bone among groups. If indeed, class II and III mandibles are morphologically modern mandibles that are particularly affected by the masticatory loads derived from the modern soft diet, we expect that they represent more extreme versions of the modern individuals with a harmonic maxillomandibular relationship. That is, that they show more gracile features together with small size and less amount of cortical bone, more elliptically distributed within the mandible’s cross sections.

Shape and structural analyses were performed using the CT scans of individuals from archaeological and modern samples present in anthropological collections and clinical databases. All individuals were adults, as assessed by the eruption of the third molar and/or the ossification of the spheno-occipital synchondrosis in their corresponding crania. In the case of the archaeological sample, a different number/list of individuals was used for each analysis due to their different status of conservation (e.g. there were individuals with or premature unilateral tooth loss or museum individuals with one side destroyed, not suitable for shape analysis but that could be used in cross-sectional analysis). In total, 118 individuals underwent mandible shape analysis, while 118, 97 and 97 individuals were included in cross-sectional geometry analyses of the symphysis, left and right first molar areas respectively. They correspond to current and past populations of current Chilean and Argentinean territories (Table 1). In the case of individuals in the clinical databases, their CTs were obtained for medical reasons, unrelated to this study. The skulls of the remaining individuals are housed in different museums and universities in Chile, Argentina and Italy (as described in Table 1), and were selected according to their conservation status and their dental status (with at least one molar in occlusion to ensure the presence of masticatory loads). The classification of the sample according to their masticatory load was made following previous anthropological studies [4, 6‐8, 25, 32]. These rely on the changes in subsistence economy from ancient to current populations, and what food items were common in them. The most ancestral type of subsistence economy was based in hunting animals and gathering fruits and/or seeds. As populations became sedentary, agriculture and cooking technologies developed, and vegetables replaced animals as the main resource. Meat is mechanically more challenging to chew, and cooking changes the material properties of food in mouth from hard/tough to hard/soft [33, 34]. Nowadays, food items are extensively processed both mechanically and chemically, resulting in a majority of soft foods. Tough (and less so, hard) food items require comparatively larger, more sustained occlusal forces and more masticatory cycles before swallowing [35, 36]. In the present study, we have chosen three groups of individuals based on the intensity of masticatory loads: intense (composed by hunter-gatherers), medium (agriculturalists) and low (individuals from modern, urban populations). As study cases, we have included individuals with low masticatory loads and altered maxillomandibular relationships, namely skeletal classes II and III. A detailed description of the groups is given in Table 1.

The CTs where acquired in different places and under different acquisition protocols. Since the study of large-scale features is of interest (i.e. whole geometry rather than geometry of small areas) and preliminary exploratory analyses showed no differences among individuals of different origin within load groups, the possible differences in acquisition protocol were not considered as a problem. In the contemporary, sample sex was known, while in the archaeological sample, it was estimated using classic cranial and postcranial (when available) features described in Buikstra and Ubelaker [37] by physical anthropologists that worked in each subsample. In some cases, sex could not be determined and had to be assigned (see below).

Three-dimensional reconstructions of each individual were made via threshold-based and manual segmentation using the software Avizo v.9.1 (Science Visualization Group, Burlington, EE.UU.) and saved as polygon (.ply) files. Using these 3D reconstructions, two types of data were obtained: mandible shape and size data, and cortical bone amount and distribution.

The effect of measurement error on results was neglectable on both mandible shape analyses, as interindividual variation was several times larger than intraindividual variation (F = 18.66, p < 0.01) and cross-section raw data (TA χ² = 0.01, p = 1; CA χ² = 0.83, p = 0.37; I_max χ² = 0.83, p = 0.37; I_min χ² = 0.01, p = 1).

The consideration of the modern (contemporary) human jaw as a small, weak structure when compared with its ancestors due to changes in diet has been long associated with the origin of dental malocclusions [1, 9, 10]. Using combined techniques and individuals from current and archaeological populations that differ in their subsistence economy and maxillomandibular relationships, the aim of this study was to perform a detailed characterisation of the shape, size and inner structure of the mandible in order to identify traits that could be linked to diet consistency and dental occlusion type.

The null hypothesis that there are no differences in mandibular shape, size amount and distribution of cortical bone among groups could not be fully rejected. Although there were some differences among groups, these do not completely match the expectation that modern mandibles are intrinsically small and weak. On the contrary, there is a large variation within groups, leading to extensive overlaps. As expected though, class II mandibles are indeed smaller and show gracile features (they are located in the lower right quadrant in the PC plot in Fig. 2a and towards the left end in the allometry plot in Fig. 2b), while class III are larger and also gracile (Figs. 2 and 3). Gracility is a shared feature among the three low load groups, but the low load mandibles with harmonic occlusion are not small. In fact, they are slightly larger (i.e. they are located more to the right end of the X axis in Fig. 2b) than the medium load group. The larger mandibles belong to the intensive load group, but they have a wide range of sizes that includes those of the medium and low load groups.

Internal morphology shows less differences among groups. In terms of amount of cortical bone, however, the three low load groups have more than the intense and medium load groups (Fig. 3), which is against expectations: there is supposed to be a “need” of more cortical bone in mandibles that are subject to larger masticatory forces. The degree of circularity of cortical bone distribution is very similar among groups.

Previous studies of external bone morphology, particularly from anthropological sciences, assign a rather high importance of masticatory load on bone morphology, with notorious differences between hunter-gatherer groups and agriculturalists [6‐8]. Our results show far less differences among groups. Several factors could account for these differences in results, such as ancestry and sample composition (none of the other studies included modern individuals with altered occlusion). Also, the classification in groups according to masticatory loads presents limitations; the material properties of food items cannot be measured in extinct populations, nor strictly controlled in humans. Independently of the place or culture in which an individual grew, what he or she actually did in life cannot be known with certainty. However, it is worth to highlight that our study included individuals that had at least one molar in occlusion in order to ensure that masticatory loads were exerted in life. This is not trivial, since agriculturalists tend to show more infection-based tooth loss [1, 5], and tooth loss causes form changes in the mandible [55]. The dental status of the individuals included in the aforementioned anthropological studies is not explicit. Our results are more in agreement with Katz, Grote and Weaver [56], who found rather modest effects of subsistence economy on mandibular shape. They attribute more importance to other factors, linked to genetics, on shape changes over many generations, particularly in closely related groups (from a similar ancestry) where genetic diversity and shape variation are linked.

Despite broad similarities of external morphology among groups with different masticatory load intensity, class II and III individuals show more marked differences with the rest of the groups, but less so to the low load group. Thus, the three modern samples together show a quite broad range of shapes, with class II and III mandibles in the extremes of the shape variation. This is linked to the allometric effect present in the sample, i.e. the shape of class II mandibles is dependent of their small size and the shape of class II mandibles, of their large size. Same as genetics, masticatory muscle force has shown to have an effect on mandibular shape in modern individuals [15]. But large shape variation leading to the extreme features shown by individuals with malocclusions could be linked to the effect of other factors, such as nutrition and basal metabolism in modern humans. This possibility has been largely overlooked and, unfortunately, it is far less straightforward to assess although there is some evidence that supports it. In pigs, it has been observed that the nutritional changes associated to weaning cause a decrease in skull growth rates, either when weaning occurs early or late in ontogeny [57]. This is surprising since weaning also changes the pattern of muscle activity, increasing jaw closing function (which should lead to an increased bone growth). Moreover, humans whose development and body size has been affected by diseases and malnutrition can recover later in development if properly nourished [26]. These two studies show that (1) within the life of an individual, the timing at which metabolic alterations occur is key [34], and (2) that nutrition (and hence metabolism) is an extremely relevant factor affecting bone morphology. This effect of metabolism could perhaps be comparable to that of mechanics on bone growth (at least in pigs and modern humans) or even larger, if mechanical constraints are reduced. At a larger, evolutionary time scale, the transition from hunting and gathering to agriculture in human populations has also, but less notoriously, been linked to smaller bones due to malnutrition [5] and hormonal changes related to a low-protein diet [27]. Our medium load group is comprised by agriculturalists, and indeed, they show relatively smaller sizes than the intensive load (comprised by hunter-gatherers) and the low and low-class III groups. These last two, in the case of our Chilean population, were likely exposed to a more balanced diet than agriculturalists. Class II individuals would be the exception to this. Thus, the link between a reduced masticatory loads and dental malocclusions in modern, urban populations is not as straightforward as suggested. Unlike other animals, humans are able to reach adult size without the need of exerting large masticatory loads. It is then likely that under the lack of mechanical (masticatory) developmental constraints, nutrition, developmental timing and of course, genetics, have an important effect on mandible shape variation. The effect of nutrition and basal metabolism offers an interesting perspective on malocclusion aetiology, prevention and treatment, and should be looked in deeply. These results also suggest that the relative importance of factors affecting mandible morphology can differ between past and current populations. Consequently, although masticatory loads have been proposed as an important factor determining secular changes among populations sharing the same geographical location [7, 8], the lack of functional constraints due to the modern diet may have led to the emergence of new, more relevant factors that should be taken into consideration.

It is interesting that the amount of cortical bone, on the other hand, seems to behave the opposite than expected. The three low load groups have a larger relative amount of cortical bone. Moreover, their cross section is not particularly elliptical (and thus less able to withstand the torsional loads derived from high masticatory loads). Thus, despite their external gracility, modern mandibles are not intrinsically weak. These results could be due to sampling, as statistics do not show striking differences (Tables 4, 6 and 8) and the number of individuals with altered occlusion is not large. However, it is worth considering an alternative explanation. Bone structure is dynamic, with cortical and trabecular bone changing in response to load in relatively short periods of time through lifespan [28‐30]. In comparison, external morphology changes slowly and, as explained, under the effect of several factors. It is plausible then that with an overall gracile morphology (although not necessarily size), bone seeks to increase its resistance by increasing the amount of cortical bone. Future work could consider changes in trabecular bone structure (trabecular bone thickness, bone volume fraction, orientation, etc.) in order to understand the extent of the effect of masticatory loads on bone morphology.

From a clinical perspective, our study highlights the importance of considering different factors in the aetiology and perhaps treatment of dental malocclusions in relation to an altered mandible shape, particularly since masticatory loads are not always easy to modify. Beyond the more obvious links of our results to orthodontics and surgery, the results regarding the relative amount of cortical bone could be something to take a further look. Mandible cross-section morphology is key in implantology [31], and since there might be a tendency to compensate the gracile mandible shape by increasing the amount of cortical bone, it could be perhaps a good strategy to promote keeping masticatory loads after tooth loss (whenever possible) in order to avoid cortical bone reduction.

In conclusion, the morphology of the human mandible shows a wide range of variation. Its features do not fully support the idea that it is small and weak, leading to malocclusions. We encourage dental practitioners and academics to not take this idea for granted, but to analyse the scenario for each case. Although class II mandibles are indeed smaller, their internal morphology seems to be able to withstand the forces normally produced during chewing. The large variation in the mandible shape of modern individuals could be related to the effect of other factors that, under comparatively low masticatory loads, gain relevance, such as genetics and the largely overlooked effects of nutrition and basal metabolism.