14.3. The Impact of Ante-Mortem (AMTL) and Post-Mortem Damage (Including Anti-Mortem and Post-Mortem Tooth Loss) in Estimating the Level of Disease in Ancient Populations
Studying periodontal disease in an ancient population with dried skulls, however, can be challenging. For example, both ante-mortem (AMTL) and post-mortem damage (including anti-mortem and post-mortem tooth loss) was the most common problem faced in this study that led to an underestimation of the various dental pathologies identified in the skull collection including periodontal disease, caries, and calculus. For example, the skeletal remains from the original collection (n = 516) were excluded for the following reasons: (1) 228 missing skulls, (2) 43 adults were unclassified or under 18 years of age, (3) 84 adults were either edentulous or had post-mortem damage, and (4) 56 adult skulls were unexamined due to time constraints. Some of the teeth were present but could not be placed in the correct position by RA due to several factors such as extensive tooth wear, extensive caries, or a damaged tooth socket. In the present study the prevalence of AMTL was higher in older individuals with average of 8.8 teeth lost per individual. Both the maxilla and the mandible showed similar numbers of AMTL (average 3.5 and 2.8), the reason for tooth loss in this sample may have been due to trauma, caries, extensive tooth wear, periapical pathology such as an abscess, or periodontal disease. Individuals who were edentulous were excluded from this study. In retrospect, this decision may have resulted in the underestimation of the total AMTL in the population, especially in the older age group.
Although most of the skulls were in good condition, the alveolar margins were commonly damaged, during burial, excavation, and processing, which may have been a major cause of post-mortem tooth loss [
13,
17]. Single rooted teeth were mostly affected, which could be due to the root anatomy and loss of the soft tissue including the periodontal ligament [
18]. A common problem identified in the present study was damage to the buccal bone including fractures, which led to the exclusion of the teeth and as such led to an underestimation of the prevalence of bone loss due to periodontal disease. Furcation involvement was also common in the St Bride’s sample, and periapical lesions (bone defect) were observed in half of the sample with different shapes and severity ranging from a small defect to a large defect involving more than one tooth, which was a similar observation in the Romano-British sample [
3].
14.4. Attempts to Overcome the Problem of Inaccurate Linear Measurement of the Distance between the CEJ and AC When Examining Dried Skulls
One of the previous issues in assessing the prevalence of periodontal disease (periodontitis) in ancient populations was the accuracy of the linear measurement of the distance between the CEJ and AC in the skull collections. This measurement, however, ignored the continuous eruption of the teeth during the individual’s life and consequently, overestimated the prevalence of periodontitis [
3]. To overcome this possibility, the Goncalves et al. [
9] methodology was used in the present study for the CEJ-AC measurements. According to Goncalves et al. [
9], subtracting 2 mm corresponded to the CEJ-AC distance which was considered as normal in healthy young individuals based on the radiographic imaging of the teeth [
19]. For healthy individuals over 45 years several investigators have recommended that an average of 3 mm should be considered normal [
20,
21]. The rationale for these recommended measurements was that measurements over this average value would be considered as alveolar bone loss due to periodontal disease [
9]. Except for the youngest age group, teeth with a higher level of tooth wear presented a similar mean level of CEJ-AC distance when compared to teeth with a low tooth wear grading. In the youngest age group, there was a large difference between the two samples (wear 0/1 and wear 2/3), which may have occurred due to the small sample size. Furthermore, one of the individuals in this age group had a large amount of ante-mortem tooth loss and extensive tooth wear, which was unusual for an individual of that age. According to the general pathology data, this individual also had trauma with an oblique non-union fracture of the right femur together with evidence of bone pathology which may have influenced the bone remodelling process. The results from the St Bride’s Lower Churchyard sample agreed with the study of a UK late-medieval population study [
9] in that, regardless of the wear level, the CEJ-AC was always similar, and any measurements obtained were more likely to be bone loss due to persistent inflammation because of the periodontal disease progression. It should however be acknowledged that the index used in the Goncalves et al. [
9] study was the same as that used in the Smith and Knight study [
22], whereas in the present study the index used was the same as that employed in the Raitapuro-Murray et al. [
3] study. Larsen [
23] also indicated that, in general, there appeared to be an increase in periodontal disease (over time) with a transition from traditional to western-style processed diets. For example, past populations consuming large amounts of plant carbohydrates or processed foods have higher rates of periodontitis compared to foragers with substantial amounts of animal protein in their diets.
14.5. Problems When Comparing Disease Prevalence in Ancient and Modern Populations
Variation in the methodology, case definition, and the statistical interpretation of the data from previous studies made comparison of the results difficult, particularly regarding the use of a case definition to assess periodontitis. In the present study the methodology of Raitapuro-Murray et al. [
3] was used, and for example, in case definition I there were no individuals affected in the 36–45 years age group, whereas nine individuals were considered as affected using case definition II. The reason for the difference was that for case definition I, four quadrants were required to be considered as periodontally affected and in this age group several of the quadrants were edentulous. The other reason was that in some individuals only one jaw was present, therefore, the individual was excluded from the affected category. It should be noted, however, that the measurements recorded in the present study represented an estimation of bone loss and therefore cannot be viewed as an accurate representation of periodontal disease progression.
Comparison of the prevalence rates from the various studies proved problematic due in part to differences in methodology and case definitions when recording the prevalence of periodontal disease from dried skulls (with the absence of soft tissue including periodontal pockets) in a historical collection of human remains compared to living individuals within a modern-day population. For example, as previously mentioned, the prevalence of moderate to severe periodontitis in the 18–19th century skull sample was 21–24%, which was higher than in the Romano-British sample (5.6%) [
3], whereas the prevalence of periodontitis in adults over 30 years (as classified by the Centers for Disease Control and Prevention (CDC)/American Academy of Periodontology (AAP) case definitions) was 7.8% for severe periodontitis and 34.4% for non-severe (mild to moderate) periodontitis respectively [
24]. Similar prevalence rates were also recorded in the Adult Dental Health Survey in the United Kingdom (excluding Scotland) with 9% of dental adults with severe periodontitis and 37% with pocketing of 4–6 mm respectively [
25].
In the present study, there was a low number of teeth presenting with vertical bone loss. Cases of localized vertical bone loss of pulpal origin were extremely severe and extensive and were usually associated with deep carious lesions and pulpal exposure due to caries, extensive tooth wear, or tooth fracture. Other reasons for localized vertical bone loss may include food or foreign object impaction, trauma, or root fracture [
3]. Most of the St Bride’s sample suffered from attrition with tooth wear ranging from mild to extensive wear. There were no differences between the age groups regarding the severity of tooth wear. In several individuals the tooth wear was so severe that it led to a pulpal exposure, which was also observed in the Romano-British sample [
3]. The index used for recording tooth wear in the present study however was not sensitive in detecting any differences between the age groups since most of the population, even those in the younger age group, had exposed dentine which in some cases exposed the pulp chamber. Attrition is the normal result of tooth use, either for mastication or as a tool. The primary cause of tooth wear has been previously assumed to be dietary in nature with the fibrous or abrasive nature of food [
3,
26] although a secondary cause of tooth wear, historically, was the use of teeth as a tool.
Evidence of heavy smoking was observed in 14% of the sample. Clay pipes were used in the 18–19th century which left a clear circular wear pattern on the tooth surface. According to Albandar et al. [
27], pipe smoking may have similar adverse effects on periodontal health and tooth loss as cigarette smoking. Furthermore, smoking increased both the prevalence and severity of periodontal disease in susceptible individuals and exposure was associated with a 2- to 3-fold increase in the odds ratio of developing periodontitis [
28].
A high percentage of skulls in the present study also suffered from significant and widespread oral pathology including dental caries (89.5%; 4.8 mean per tooth) and periapical lesions (40%; 0.7 mean per tooth affected) (
Table 7). Dental caries varied between the skulls and the most affected teeth were the first molars followed by the second molars (involving mainly the occlusal and root surfaces). The St Bride’s sample was previously studied for the prevalence of caries by Mant and Roberts [
29]. The prevalence of dental caries in the total population was 78.9%, and of those skulls identified with dental caries, 26.5% of all teeth present were affected. These investigators compared the data to other populations and concluded that there were no significant differences in the rate(s) of caries in individuals either by gender or social status [
29]. More recently, Smith [
30] compared post-medieval and modern-day caries exposure of adults living in the East London borough of Tower Hamlets, UK, using recorded data from human remains excavated from New Churchyard located under a major London railway station (Liverpool Street) and data from the East London Oral Health Inequality study (ELOHI) [
31]. From the results of this comparative data, it was evident that there were significantly lower rates of dental caries (decay) in a post-medieval London sample than in a modern-day population. One of the problems of comparing data from historical skull collections and modern populations is the method of data collection, which has changed. For example, in the St Bride’s [
29], Raitapuro-Murray et al. [
3], and the New Churchyard [
30] studies caries was not collected for analysis using the Decayed, Missing, Filled Teeth index (DMFT), whereas in the modern population (ELOHI) study (cited by Smith [
30,
31]), examiners were trained and calibrated in the use of the index resulting in a good level of examiner reproducibility. It should be noted, however, as Smith [
30] indicated in his paper, that he was unable to compare the raw data from the ELOHI study when comparing the New Churchyard study and a representative modern-day sample (ELOHI). Nevertheless, the recorded caries rate in the Romano-British sample (43.9% to 75.4%; mean teeth affected 1.5–2.3 depending on the age cohort) and post-medieval samples (St Bride’s 89.5% (
n = 94), 4.8 mean per tooth and New Churchyard 27.9%, F: average 4.5 decayed teeth and 6 missing teeth, M: average 4.3 decayed teeth and three missing teeth) was lower than that of a more modern comparable sample (ELOHI (“White British” M/F) DMFT mean (95% CI): 13.47 (12.52–14.42)) despite the differences in data collection [
3,
31,
32].
The progression from a relatively low caries rate as shown in the Romano-British [
3] and the post-medieval samples (St Bride’s and New Churchyard) culminating in a higher caries rate as evidenced in a more modern population [
31] can be associated with poor oral hygiene and a cariogenic diet. For example, the use of sugar throughout the 18th century was a key factor in the development of caries, and its increase in use is directly linked to an increase in tea drinking throughout the UK, resulting in a carbohydrate and soft diet that allowed plaque accumulation around the teeth. This trend has continued in a modern-day population with access to a more cariogenic diet [
30]. It should also be acknowledged that the frequency of the caries rate in the present study could be underestimated due to both ante-mortem and post-mortem tooth loss as well as the short life spans of the individuals within the sample. Furthermore, it is evident from recent studies that the frequency and distribution of caries and ante-mortem tooth loss increased in ancient populations with a longer life span [
17,
29,
30,
31,
33]. In the present study there was no evidence of ‘routine’ dental treatment other than extraction, which may be indicative of the low socio-economic status of the individuals in the sample, an observation that appears to be supported by Hillam [
34] who reported that there was limited data of any conservative dental treatment practiced in both Britain and mainland Europe up to the end of the 18th century.
Although the study of ancient populations may contribute to an understanding of both the epidemiology and natural history of a disease there is a need for investigators to agree on guidelines using the same methodologies (including training and calibration) and case definitions when examining periodontal disease and other oral pathologies in dried skull collections.