The PAI calibration process was concluded with inter and intraobserver Kappa scores of 0.941 and 0.985, respectively, indicating good agreement with the ‘gold standard’ and minimal variability in the PAI scoring.
Of the selected study population of 610 subjects, 328 (53.8%) were female and 282 (46.2%) were male. The sex distribution did not differ significantly between the Toronto and Saskatoon populations (P > 0.7).
Descriptive and analytic statistics regarding missing teeth, untreated teeth with apical periodontitis and root-filled teeth are summarized in Table 2. Overall, 309 subjects had at least one missing tooth (excluding third molars), 172 subjects had at least one untreated tooth with radiographic signs of apical periodontitis and 209 subjects had at least one root-filled tooth. Comparing the two study populations, the subjects in Saskatoon revealed significantly more complete, periapically healthy and endodontically untreated dentitions than the subjects in Toronto. Of the potential 17 080 teeth in the total population (28 _610), 932 teeth were missing. Thus, the total number of teeth inspected in the radiographs of all subjects was 16148, of which 10 474 were from the Toronto population and 5674 from the Saskatoon one. Of the total, 330 teeth (2.0%) were endodontically untreated and had apical periodontitis, whilst 411 teeth (2.5%) were root-filled. Because the Saskatoon population had only panoramic radiographs available, the 54 subjects with 97 root-filled teeth identified in the panoramic radiographs were invited to have periapical radiographs exposed in order to enable an appropriate determination of PAI scores for those teeth. Only 36 subjects with 69 (71.1%) root-filled teeth complied; therefore, in the following analyses only these teeth are included.
The distribution of PAI scores for the assessed 383 root-filled teeth is presented in Table 3. The most frequently assigned PAI score was ‘3’ (36.0%), whilst the most infrequent was ‘5’ (2.9%). Dichotomizing the PAI scores revealed that a total of 209 teeth were categorized as ‘healthy’, and the remaining 174 teeth were categorized as ‘diseased’, or having apical periodontitis. The difference in the proportion of ‘diseased’ root-filled teeth in the two populations was not statistically significant (chi-square, P > 0.3).
Of the total of 209 subjects with at least one root-filled tooth,119 subjects responded to the recruitment invitation and enrolled in the study, 83 in Toronto (53.5%) and 36 in Saskatoon (67.9%). The socio-demographic characteristics of the enrolled subjects are summarized in Table 4. The study populations in both sites were suitably matched, except for the significantly higher proportion of immigrants in the Toronto population.

Table 2. Numbers, mean values and prevalence of missing teeth, untreated teeth with apical periodontitis and root-filled teeth.



Table 3. Distribution of PAI scores for root-filled teeth.



Table 4. Socio-demographic characteristics of the subjects enrolled in the clinical examination and interview part of the study.



Table 5. Associations of periapical health (PAI scores1and 2) with antecedent factors in root-filled teeth. Analyses are based on all 383 teeth reviewed radiographically, unless otherwise specified.



Table 6. Odds Ratios (OR) and 95% confidence intervals (CI) estimating the risk of apical periodontitis in root-filled teeth related to treatment quality factors as assessed radiographically.

The root-filled teeth, those reviewed only radiographically (n = 383) and those examined both radiographically and clinically (n = 236), are characterized in Table 5, and the variables related to periapical health (PAI scores 1 and 2). Molars were the most prevalent amongst the root-filled teeth (42.8%), whilst the distribution amongst anteriors, premolars and molars did not differ significantly between the two populations (data not shown in table, P > 0.7). Of the 236 teeth examined clinically and addressed in patient interviews, 82.6% had been treated by generalists,87.3% had been treated over 2 years before the present examination, 16.1% were symptomatic at the time of the examination,36.4% had clinically detectable defective restorations and 19.5% had defects in coronal tooth structure. Of the total 383 teeth reviewed radiographically, the density and length of the root filling were adequate in 60.1 and 58.0%, respectively, whilst the coronal restoration was apparently adequate in only 42.3%. In 9.8% of the sample (n = 224), the restoration was either temporary or missing. The differences in periapical health related to tooth type and dental arch were borderline significant. Differences related to root filling density and length, as well as those related to the radiographic quality and type of the restoration, were all statistically significant. Differences in periapical health related to all other characteristics of the root-filled teeth were not statistically significant, including the difference between teeth treated by generalists and endodontists, and between generalists in Toronto and in Saskatoon (data not shown in table - 53.3 and 48.5%, respectively; P > 0.5).
Analysis of the agreement between the clinical and radiographic quality assessment of the coronal restorations in the 236 clinically examined teeth revealed a Kappa score of 0.235, demonstrating poor agreement between the clinical and radiographic assessments.
The likelihood of an outcome of apical periodontitis is related to radiographically assessed treatment quality factors in Table 6. The odds of developing apical periodontitis were 2.7 times higher when the root filling density was inadequate than when it was adequate, and 2.5 times higher when the root filling length was inadequate (short or long) than when it was adequate. Specifically, the root filling being short (OR = 2.4) or long (OR = 2.8) increased the odds of developing apical periodontitis, with no statistically significant difference between long and short root fillings. With regards to the quality of the coronal restoration, the odds of developing apical periodontitis were 1.7 times higher when it was inadequate than when it was adequate.
The two radiographic quality parameters of the root filling (density and length) found to be significantly associated with periapical health were analysed in various combinations (Table 7). The difference in prevalence of ‘healthy’ scores amongst the four root filling quality combinations was statistically significant (P < 0.001). The odds of apical periodontitis were 3 times higher when the root filling was inadequate (density or length) than when it was adequate. When both the root filling density and length were inadequate, the odds of apical periodontitis were 3.3 times higher than with any other root filling quality combination.
To assess the relative impacts of the apparent quality of the root filling and the restoration on periapical health, different combinations of the quality parameters were analysed (Table 8). For this analysis, the quality of the root filling was noted as inadequate if either the density or the length of the root filling was assessed as inadequate. The difference in prevalence of ‘healthy’ scores amongst the four combinations of root filling and restoration quality was statistically significant (P < 0.001). The odds of apical periodontitis were 4.6 times higher when the root filling or restoration were inadequate than when both were adequate. When both the root filling and restoration were inadequate, the odds of apical periodontitis were 2 times higher than with any other combination of root filling and restoration quality. Comparing teeth with adequate root fillings, the proportion of ‘healthy’ scores was significantly higher for those with adequate restorations than for those with inadequate restorations (P < 0.01). The odds of apical periodontitis were 2.8 times higher when an adequate root filling was coupled with an inadequate restoration than with an adequate one. In contrast, comparing teeth with inadequate root fillings, the difference between those with adequate and inadequate restorations was not statistically significant. When the root filling was inadequate, the odds of apical periodontitis were only 1.2 times higher with an inadequate restoration than with an adequate one.
Quality characteristics of the endodontic and overall treatment are related to providers of treatment in Table 9. Overall, adequate treatment on all accounts was more frequent amongst teeth treated by endodontists than teeth treated by generalists; however, the difference was statistically significant only for the length of the root fillings. Amongst teeth treated by generalists, the frequency of adequate quality of both the root filling and the restoration was significantly higher in the Saskatoon population.
Interviews of 119 subjects to determine the reasons for extraction of the 261 missing teeth revealed that 189 teeth (72.4%) were extracted because of caries and inability or refusal to receive endodontic care. A further five teeth (1.9%) were extracted because of persistence of apical periodontitis or fracture of the tooth after endodontic treatment, and eight teeth (3.1%) were extracted due to advanced periodontal disease. The remaining 59 teeth (22.6%) were extracted for a variety of other reasons, including orthodontic purposes.

Table 7. Combinations of quality of root fillings, as assessed radiographically, and their relation to periapical health (PAI scores 1 and 2) and risk of disease (N = 383 teeth).



Table 8. Combinations of quality of root fillings and restorations, as assessed radiographically, and their relation to periapical health (PAI scores1and 2) and risk for disease (N = 383 teeth).



Table 9. Quality characteristics of the treatment provided by generalists and endodontists.