MR. STEPHEN KEBKE (Orcid ID : 0000-0002-4526-2165)
DR. FERNANDO JOSÉ MOTA DE ALMEIDA (Orcid ID : 0000-0003-3388-9451)
Article type : Original Scientific Article
Tooth survival following root canal treatment by general dental practitioners in a Swedish county – a 10-year follow-up study of a historical cohort
S. Kebke1, H. Fransson2,3, M. Brundin4 & F.J. Mota de Almeida5
1Piteå Hospital, Norrbotten Public Dental Service, Piteå, 2Department of Endodontics, Faculty of Odontology, Malmö University, Malmö, 3Department of Endodontology, Institute of Odontology, The Sahlgrenska Academy, University of Gothenburg, Gothenburg, 4Department of Odontology/Endodontics, Faculty of Medicine, Umeå University, Umeå, 5Tandvårdens Kompetenscentrum, Norrbotten Public Dental Service, Luleå, Sweden
Running title: 10-year tooth survival
Key words: Root Canal Therapy, Analysis, Survival, General Practice, Dental, Prognosis, Endodontics, Kaplan-Meier Estimate, Public Health, Tooth Extraction, Treatment Outcome
Corresponding author: Fernando José Mota de Almeida
Tandvårdens Kompetenscentrum, Vuxenspecialiteter, Skeppsbrogatan 22, Box 922, SE-971 28 Luleå, Sweden
Email: fernando.mota-de-almeida@norrbotten.se Telephone: 0046-(0)920-71926
ABSTRACT
Aim To evaluate the 10-year survival rate of root filled teeth treated by general dental practitioners (GDPs), and to identify possible prognostic factors.
Methodology In 2006, 3 676 individuals had at least one tooth root filled by a GDP within the Norrbotten Public Dental Service, Sweden. Over the next 10 years, 331 individuals died and were excluded. A random sample of 302 of the remaining individuals were included in the study, of whom 280 (n=280 teeth) were included in the analysis. Dental records were reviewed retrospectively by a calibrated researcher to collect predetermined data regarding individual, preoperative, intra-operative, and postoperative factors. The outcome measure was tooth extraction over time, and cases with no events were censored, regardless of apical status or symptoms, until last known date of tooth survival. In case of missing data, individuals were recalled for a control visit. Kaplan-Meier survival tables, and Cox regression models were used for
analysis. P<0.05 was considered statistically significant.
Results The cumulative 10-year survival was 81.7% (standard error: 2.6%), and the mean incidence of tooth extraction during the 10 years was 1.8% per year. The univariate analysis identified three possible prognostic factors (p<0.05) that were associated with extraction: molars, two or more emergency inter-appointment visits during the treatment, and root canal treatments consisting of five or more separate sessions. A multivariate regression analysis revealed no significant relationships for the variables gender, tooth type, number of contacts, any emergency visits during endodontic treatment, number of sessions to complete endodontic treatment, pulp diagnosis, or type of permanent restoration and extraction.
Conclusions The mean incidence of tooth loss over the first 10 years after completion of root canal treatment performed by a GDP was approximately 2% per year. No prognostic factors could be identified.
INTRODUCTION
Strindberg (1956) suggested that a successful root canal treatment is established by the absence of apical periodontitis, where clinical and radiographic normalcy is maintained or re-established. A systematic review of the literature reported that the success rate of primary root canal treatment completed at least one year previously ranged from 68% to 85% when this strict criterion was applied, figures that have not improved in recent decades (Ng et al. 2007). However, less strict criteria for the outcome of root canal treatment have been proposed, where success is defined by the fact that the apical radiolucency is reduced or unchanged in size and the patient has no symptoms (Bender et al. 1966). Thus, a successful treatment could also be defined by tooth survival; that is, an asymptomatic and functional tooth regardless of its radiologic appearance (Friedman & Mor 2004). Tooth survival instead of absence of the disease may be regarded as an insufficient biological goal, but for the individual this may be an acceptable result of the treatment, as the tooth is functioning normally in daily life. Root filled teeth may be lost for various reasons with prosthetic, caries, and periodontal reasons being reported as common reasons for extraction (Vire 1991, Toure et al. 2011, Tzimpoulas et al. 2012). Normally, a root filled tooth is extensively restored, and the root canal treatment may further weaken the tooth. For this reason, the results of studies surveying tooth survival may be of interest for prognostic purposes and therapy planning.
In a systematic review of the literature, almost all published in the 2000s, the pooled probability of tooth survival for 2–10 years ranged from 86% to 93% after primary root canal treatment (Ng et al. 2010). Since then, additional studies from several geographical areas have been published with survival rates after root canal treatment ranging from 79% to 95% with follow-up times of 3–20 years (Chen et al. 2008, Lumley et al. 2008, Ng et al. 2011, Bernstein et al. 2012, Lin et al. 2014, Landys Boren et al. 2015, Raedel et al. 2015, Fransson et al. 2016, Pratt et al. 2016, Skupien et al. 2016, Fernandez et al. 2017, Khalighinejad et al. 2017, Prati et al. 2018, Kwak et al. 2019, Vahdati et al. 2019).
Few studies focus on tooth survival following root canal treatment performed by general dental practitioners (GDPs), even though these are the dentists who perform the vast majority of root canal treatments (Chen et al. 2008, Lumley et al. 2008, Bernstein et al. 2012, Raedel et al. 2015, Fransson et al. 2016). Moreover, those studies that do exist usually use data from health insurance databases, where information about potentially interesting variables that could predict treatment outcome is more likely to be absent, and where drop-outs, discontinuers, and reliability of data are not well controlled. This suggests a need to study the survival of root canal treated teeth performed by GDPs and possibly to identify prognostic factors not necessarily accessible from insurance data.
The primary aim of this study was to evaluate the long term (10-year) survival rate of root filled teeth treated by GDPs working in the Public Dental Service of the county of Norrbotten (also known as North Bothnia), Sweden. The secondary aim was to identify possible prognostic factors that affected that outcome.
Accepted Article
MATERIAL AND METHODS
A historic cohort of individuals who had undergone completed root canal treatment by GDPs in Norrbotten was randomly chosen and followed for a period of 10 years.
Context
When the study began in 2006, Norrbotten, the northernmost county in Sweden, had a total population of 251 886 (www.statistikdatabasen.scb.se) and 29 general dental clinics within the Public Dental Service. These clinics were located throughout the entire county, with at least one clinic in each municipality, and served approximately 70% of the population in Norrbotten including children and adolescents.
In 2006, 149 GDPs in the county reported that they completed at least one root canal treatment (range: 1‒256 treatments). The GDPs had varying degrees of experience. Most of them had their undergraduate training in Sweden, but some had studied in Belgium, Germany, or Portugal.
Cohort
All individuals with at least one root canal treatment performed by GDPs in the Norrbotten Public Dental Service between 1 January and 31 December 2006 were identified via the county’s electronic
administrative dental records. There were 3 676 such individuals, with a total of 3 925 root filled teeth (range: 1–5 teeth). Of these, 331 individuals died during the following 10 years and were therefore excluded. The remaining 3 345 dental records were arranged in random order using Microsoft Excel (version 14, 32-bit option, 2010, Redmond, WA, USA). All dental records had a corresponding personal identity number to the treated patient. These numbers were entered in a column in a Microsoft Excel sheet. In a new column, the programme´s RAND function was used for all patients that randomly attributed a 10 digit number between 0 and 1 to each entry. The columns were rearranged in numerical order of the generated random values, thus creating a randomised list of patients. The first 302 individuals on the randomised list were contacted by letter in order to obtain their passive consent to participate in the study. A total of 280 individuals treated by 102 GDPs in 28 clinics were included. Reasons for exclusion along with demographic data (gender and age) are given in Figure 1. If an individual had several teeth treated in 2006, one of these teeth was randomly chosen for inclusion in the analysis by randomising a list of their teeth in the similar way as described above for randomising patients, and then choosing the first tooth on the list.
Data collection
Analogue and digital dental records including dental radiographs were retrieved from the clinics; 110 of the analogue records could not be fully retrieved, which prevented the collection of some variables. The collection period lasted from September 2016 to May 2018.
Initially, the dental records were screened for survival of the root filled tooth to 2016. This was either assessed by presence or absence of tooth in existing radiographs, or from the dental record with a specific
Accepted Article
treatment or diagnosis of the tooth of interest. The date of any data indicative of the root filled tooth being present or extracted was recorded. At this stage it was possible to collect data on 150 of 280 root filled teeth as of 2016. For the remaining 130 individuals there were no available data indicating either survival or loss of the root filled tooth. These 130 individuals were contacted by telephone or mail for a voluntary basic dental examination at the nearest public dental clinic in order to establish whether or not the tooth was retained in the mouth. The examination was performed either by one of the authors (SK) or by a caregiver at the nearest dental clinic. If the root filled tooth was missing without a record of extraction in the dental records, due to extraction at a private clinic or outside the county, the individuals were asked to give a written consent to retrieve the records from that particular caregiver. If the exact time of the extraction was impossible to retrieve, an estimation of the timing of the extraction was calculated as the mean between the last date of the tooth being retained and the first date of the tooth being absent in the dental records. For individuals not attending the clinical follow-up, the last known date of the retained tooth was used. Five individuals were asked over the telephone whether the tooth was retained or not. Exclusions and drop-outs were recorded (Figure 1), and dental records were reviewed thoroughly for relevant variables (Table 1) such as pulp and periapical status. The type of final coronal restoration was also registered by referring to the records and categorised as following: any type of restoration in the orifice, placement of a post upon root-filling completion, and the extent of the final restoration (partial or full coverage). The elapsed time between the completion of the root filling and final restoration was recorded.
The outcome measure was tooth extraction over time. The recorded date of the event of interest (that is, tooth extraction) was used to calculate the elapsed time in days since the date of completion of root canal treatment. If there were no data suggesting that tooth extraction had taken place, the last known date for observation of presence of the tooth, disregarding the apical status or symptoms, was used for (right) censoring. The independent variables are shown in Table 1. Of these, the most important possible confounders (that is, variables for which it is difficult to separate their effects on the response variable) were tooth type, number of contact points, and terminal arch location in one group; presence of sinus tract and apical status in another group; acute problems and number of sessions in another group; and finally pulp bleeding and pulp necrosis.
Assessments and calibration
The first 12 dental records used for data collection were reviewed by two authors (SK, FMA) for calibration purposes. All variables were assessed for this purpose except age and gender (which were automatically registered), root filling quality, and apical status. The next 20 dental records were assessed independently by the same authors to assess the inter-examiner agreement. SK reviewed the rest of the records alone. A second inter-examiner agreement (SK, FMA) was performed at the end of the collection process, with 15 new patient records. If there was disagreement, discussion took place until consensus was
achieved. SK reviewed an additional 10 dental records after a wash-out period of 6 months, to assess intra-observer agreement.
Statistical analysis, reporting, and ethics
No sample size calculation was performed, but in order to improve comparability with another similar and relevant study which had 330 participants (Landys Boren et al. 2015) the initial aim was to include 302 individuals in the present study. Missing data were left as such without any imputation procedures. Agreement was calculated with descriptive statistics. Survival tables were used to calculate the
accumulated survival rates at 1 825 and 3 650 days (equivalent to 5 and 10 years). A Kaplan-Meier survival curve was plotted for the overall survival, including censoring which denotes that the event of interest was not observed during the observation period. Univariate and multivariate Cox regression models were used to analyse relationships of possible prognostic factors, with p < 0.05 considered statistically significant. The variables selected were subsequently used in a multivariate Cox regression model which also included the variables that had a p-value of 0.2 or lower in the univariate models. Version 26 of IBM SPSS Statistics (SPSS Inc., Chicago, IL, USA) was used for the statistical analysis.
The reporting of the present study complies with the STROBE statement (von Elm et al. 2014), and the study was approved by the Regional Ethical Board at Umeå University, Sweden (ref: 2016-141-31M). RESULTS
The first inter-observer agreement was 90%, and the second was 91%. The intra-observer agreement was 95%. During the observation period (range: 0‒4 433 days), 47 teeth (16.7%) were known to have been extracted. The mean observation time for the non-extracted teeth was 3 164 days (standard deviation: 1 307, median: 3 774). Information on whether the tooth had survived or had been extracted at 3 467 days (9.5 years) after completion of root canal treatment was available for 199 patients (71.1%); the observation period was more than 365 days (1 year) in all except a small number of patients (n=22, 7.9%). The reasons for extraction were fractures (n=16), caries (n=11), endodontic reasons (n=10), prosthetic reasons (n=3), periodontic reasons (n=1), and unknown reasons (n=6). The calculated cumulative survival rate was 91.6% (standard error: 1.7%) after 1 825 days (5 years) and 81.7% (standard error: 2.6%) after 3 650 days (10 years). The survival curve is plotted in Figure 2.
Three prognostic factors reached the p<0.05 threshold in the univariate Cox regression analysis: tooth type (higher hazard for molars), acute problems (higher hazard for presenting twice or more with acute problems at start and during treatment), and number of sessions (higher hazard for treatments lasting five or more sessions) (Table 1).
The multivariate analysis was performed with these three variables (p<0.05) in conjunction with four others that had potential prognostic value (0.05<p<0.20): gender, number of contact points, pulp necrosis, and extent of permanent first restoration. No variable reached the p<0.05 threshold in the multivariate analysis.
Accepted Article
DISCUSSION
This study, based on a historical cohort, found a 10-year survival rate of 81.7% for teeth following root canal treatment performed by GDPs in a Swedish county.
Discussion of methods
It is not easy to conduct long-term randomized clinical studies of the outcome of root canal treatments. A non-interventional study following a specified cohort, such as the present one, can provide useful data, as the randomised sample is from a real-world population in a real-world setting. However, inevitably there are several drawbacks with this study design; some more general, others more particular.
Selection bias is common in retrospective cohort studies, but it must be stressed that the cohort studied here was a randomised sample derived from all patients having completed a root canal treatment during one full year within the Public Dental Service in Norrbotten. All dentists are obliged to report the completion of a root canal treatment, as it is connected to the payment for the treatment. One hundred and two GDPs contributed at least one root canal treatment to the final analysis. The present study is not a true
retrospective cohort study, but could be considered to be a historical prospective cohort study (Kirkevang & Vaeth 2019). This implies that it is retrospective with regard to baseline exposure, but prospective in terms of determining the incidence of extractions. If the aim was to investigate the same outcome in a general population, the collection of data would have been very cumbersome due to needing to access different dental record systems, and would probably have resulted in a greater number of individuals with incomplete data.
Recall bias is a common systematic error in observational case-control studies. This is a type of information bias stemming from the fact that cases are more likely than controls to recall a prior exposure (Sedgwick 2014). The information used for the present study was primarily drawn from dental records, and so the only risk of recall inaccuracy will have arisen from the few individuals who were asked by telephone if the root canal treated tooth was still retained or had been extracted, and if so when the extraction had been
performed. However, it is unlikely that any inaccuracy relating to this will have been due to a systematic error; that is, a bias. In 2006, the county had a double journal system, with both analogue and digital dental records being used (sometimes alternately), and some analogue records could not be fully retrieved. However, it is not expected that this to have led to a systematic error. The results of the study are further dependent on the quality of the notations in the dental records, which were heterogeneous and not
calibrated; some caregivers were more detailed than others when recording information. It is possible that the dentists wrote more careful notes on patients whose teeth were for some reason perceived as
problematic, and so similar factors might not have been noted in all dental records. Moreover, dentists who were more interested and possibly more skilled in endodontics could have been more prone to properly register the cases, leading to a potential risk of bias in the interpretation of the variables. On the other hand,
Accepted Article
this study design also has advantages, even though the data were collected retrospectively, as the design enabled recording of exposure to risk factors before the occurrence of the outcome (extraction), which strengthens the robustness of the results (Sedgwick 2014). Nevertheless, for some variables there was no recordable information, diminishing the statistical power for many secondary outcomes or leading to less accurate interpretations of the true situation. The selection of collected variables was limited to the data in the dental records, so there was no information on risk factors such as remaining tooth structure (Al-Nuaimi et al. 2020). This problem relates mainly to the gathering of the secondary variables, and not to the
reporting of either the completed root canal treatment or any tooth extractions, both of which were mandatory to report in a digital system. Thus, the primary outcome has more robust data and should be regarded as more reliable.
The collection of data from the dental records was calibrated, and the reliability was estimated to be high apart from two variables: root filling quality and apical status. It would have strengthened the study if all dental records could have been reviewed by more than one researcher, but this was impractical under the circumstances, and could be considered to be of less importance since the reliability of the source itself (that is, the dental records) was questionable in certain cases.
The choice of including one tooth per individual was to avoid the possibility that different teeth in the same individual might not have been independent in terms of outcome, and consequently the approach facilitated the statistical analysis by avoiding the need to take clustering into account (Hannigan & Lynch 2013). The outcome of root canal treatment is dependent on time. For the classical outcome of radiographic normalcy, a longer follow-up should be expected to show a larger proportion of teeth with apical
radiolucency. In the systematic review by Ng et al. (2007), the pooled success rates increased with longer follow-up periods when the strict criterion of healing was used. Conversely, tooth survival should only decrease with time, as extraction is cumulative for a population. In either case, it is important to relate the findings to a given time frame. Survival analysis is appropriate for studying tooth survival as an outcome. Besides recording the elapsed time until extraction, the elapsed time to the most recent data on the presence of the tooth was also recorded, preferably through information in the most recent dental records or
radiographs. Censoring is the term used in survival analysis when the event of interest (tooth extraction in this case) has not been observed during the observation period (Hannigan & Lynch 2013).
Discussion of results
The 5-year survival in this study is comparable to that found in a registry study which covered the entire Swedish adult population including the present county (Fransson et al. 2016), though it should be noted that the present study population also included children and young adults (<20 years). The results seem to indicate that over the first 10 years after root-canal treatment performed by GDPs, the mean incidence of tooth loss is just under 2% per year; however, in the 9th year the incidence was almost 5%. The mean incidence of tooth loss is equivalent to the incidence that can be calculated from other recent studies,
Accepted Article
ranging from 0.8% to 3.6% (Chen et al. 2008, Lumley et al. 2008, Ng et al. 2011, Bernstein et al. 2012, Lin et al. 2014, Landys Boren et al. 2015, Raedel et al. 2015, Fransson et al. 2016, Pratt et al. 2016, Skupien et al. 2016, Fernandez et al. 2017, Khalighinejad et al. 2017, Prati et al. 2018, Kwak et al. 2019, Vahdati et al. 2019). These studies are, however, quite heterogeneous in relation to inclusion criteria and how drop-outs during follow-up were handled, and comparisons should therefore be made cautiously.
The frequencies of the different variables (especially those with few missing cases) should represent the endodontic treatment panorama that GDPs faced in Norrbotten in 2006, and could be extrapolated to some degree to the present whole population of Sweden. One aspect worth noting is the very high proportion of primary treatments and very low proportion of secondary treatments. These proportions are similar to those reported in another Swedish study based on GDPs working in the Public Dental Service (Wigsten et al. 2019), but different from those in a Swedish study based on a specialist clinic where a large proportion of the included cases were retreatments (Landys Boren et al. 2015). This might reflect differences in case selection between patients attending GDPs and specialist clinics (that is, the degree of difficulty of such cases), or perhaps some propensity of GDPs to avoid retreatment cases, instead referring to specialists or choosing other treatment options.
Very few potential prognostic factors were found in the univariate analysis, possibly due to the limited numbers of cases presenting specific variables (for example, previous root filling, marginal status,
procedural accidents, and postoperative pain). These variables were not frequent in the material, making it difficult to obtain a powerful statistical analysis. One example might be the use of rubber dam, which in the present material was used in 85% of the treatments. Rubber dam use has previously been linked to an improvement of 1.5–2.2% in tooth survival 3.4–5 years after completed root canal treatment in an Asian population where dentists reported a low frequency of such use (11–15%) (Lin et al. 2014, Kwak et al. 2019). The reported survival tables after 10 years were still not sufficiently divergent to allow detection of such a small difference with the existing material.
The type and quality of coronal restorations of teeth with completed root canal treatments have previously received attention as potentially being predictive of outcome (Dawson 2020). However, in the present material, despite a pattern of better outcome for teeth with full cusp coverage (see variable “Extent of permanent first restoration after completed treatment” in Table 1), no statistically significant difference could be found. Nevertheless, this could be interesting to explore in future studies with larger samples. The three factors that were identified as statistically significant in the univariate analysis for tooth survival have been described before. Molars have been reported to have worse survival than other teeth in
populations treated by GDPs (Lin et al. 2014, Fransson et al. 2016, Kwak et al. 2019). The other two factors might be intimately related to each other, as individuals presenting several times with acute problems might need their treatment to be spread over several sessions. Cheung (2002) reported shorter
survival of root filled teeth treated in more than two sessions. However, none of the aforementioned variables showed an independent association with the outcome in the multivariate analysis.
CONCLUSION
This historical cohort study revealed that the mean incidence of tooth loss after completion of root canal treatment performed by GDPs over 10 years was approximately 2% per year. No independent prognostic factors could be identified.
ACKNOWLEDGMENTS
The authors would like to thank Robert Lundqvist for assistance in the statistical analysis, and the caregivers involved in the collection of data. This work was funded by the Norrbotten Public Dental Service, Sweden.
Conflict of Interest statement
The authors have stated explicitly that there are no conflicts of interest in connection with this article.
REFERENCES
Al-Nuaimi N, Ciapryna S, Chia M, Patel S, Mannocci F (2020) A prospective study on the effect of coronal tooth structure loss on the 4-year clinical survival of root canal retreated teeth, and retrospective validation of the Dental Practicality Index. International Endodontic Journal doi:10.1111/iej.13322
Bender IB, Seltzer S, Soltanoff W (1966) Endodontic success–a reappraisal of criteria. Part II. Oral Surgery, Oral Medicine, and Oral Pathology 22, 790-802.
Bernstein SD, Horowitz AJ, Man M, et al. (2012) Outcomes of endodontic therapy in general practice: a study by the Practitioners Engaged in Applied Research and Learning Network. Journal of the American Dental Association 143, 478-87.
Chen SC, Chueh LH, Hsiao CK, Wu HP, Chiang CP (2008) First untoward events and reasons for tooth extraction after nonsurgical endodontic treatment in Taiwan. Journal of Endodontics 34, 671-4.
Cheung GS (2002) Survival of first-time nonsurgical root canal treatment performed in a dental teaching hospital. Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and Endodontics 93, 596-604. Dawson V (2020) Coronal restoration in root-filled and non root-filled teeth: studies on periapical status, tooth survival, subsequent treatments and treatment decisions (PhD Thesis). Malmö, Sweden: Malmö University.
Fernandez R, Cardona JA, Cadavid D, Alvarez LG, Restrepo FA (2017) Survival of endodontically treated roots/teeth based on periapical health and retention: a 10-year retrospective cohort study. Journal of Endodontics 43, 2001-8.
Fransson H, Dawson VS, Frisk F, Bjorndal L, EndoReCo, Kvist T (2016) Survival of root-filled teeth in the Swedish adult population. Journal of Endodontics 42, 216-20.
Friedman S, Mor C (2004) The success of endodontic therapy — healing and functionality. Journal of the California Dental Association 32, 493-503.
Hannigan A, Lynch CD (2013) Statistical methodology in oral and dental research: pitfalls and recommendations. Journal of Dentistry 41, 385-92
Khalighinejad N, Aminoshariae A, Kulild JC, Wang J, Mickel A (2017) The influence of periodontal status on endodontically treated teeth: 9-year survival analysis. Journal of Endodontics 43, 1781-5.
Kirkevang L, Vaeth M (2019) Epidemiology, treatment outcome, and risk factors for apical periodontitis. In: Ørstavik D, ed. Essential Endodontology: Prevention and Treatment of Apical Periodontitis, 3rd edn; pp. 143-177. Hoboken, NJ: Wiley-Blackwell.
Kwak Y, Choi J, Kim K, Shin SJ, Kim S, Kim E (2019) The 5-year survival rate of nonsurgical endodontic treatment: a population-based cohort study in Korea. Journal of Endodontics 45, 1192-9.
Landys Boren D, Jonasson P, Kvist T (2015) Long-term survival of endodontically treated teeth at a public dental specialist clinic. Journal of Endodontics 41, 176-81.
Lin PY, Huang SH, Chang HJ, Chi LY (2014) The effect of rubber dam usage on the survival rate of teeth receiving initial root canal treatment: a nationwide population-based study. Journal of Endodontics 40, 1733-7.
Lumley PJ, Lucarotti PS, Burke FJ (2008) Ten-year outcome of root fillings in the General Dental Services in England and Wales. International Endodontic Journal 41, 577-85.
Ng YL, Mann V, Gulabivala K (2010) Tooth survival following non-surgical root canal treatment: a systematic review of the literature. International Endodontic Journal 43, 171-89.
Ng YL, Mann V, Gulabivala K (2011) A prospective study of the factors affecting outcomes of non-surgical root canal treatment: part 2: tooth survival. International Endodontic Journal 44, 610-25. Ng YL, Mann V, Rahbaran S, Lewsey J, Gulabivala K (2007) Outcome of primary root canal treatment: systematic review of the literature – part 1. Effects of study characteristics on probability of success. International Endodontic Journal 40, 921-39.
Prati C, Pirani C, Zamparini F, Gatto MR, Gandolfi MG (2018) A 20-year historical prospective cohort study of root canal treatments. A multilevel analysis. International Endodontic Journal 51, 955-68. Pratt I, Aminoshariae A, Montagnese TA, Williams KA, Khalighinejad N, Mickel A (2016) Eight-year retrospective study of the critical time lapse between root canal completion and crown placement: its influence on the survival of endodontically treated teeth. Journal of Endodontics 42, 1598-603.
Raedel M, Hartmann A, Bohm S, Walter MH (2015) Three-year outcomes of root canal treatment: mining an insurance database. Journal of Dentistry 43, 412-7.
Skupien JA, Opdam NJ, Winnen R, et al. (2016) Survival of restored endodontically treated teeth in relation to periodontal status. Brazilian Dental Journal 27, 37-40.
Sedgwick P (2014) Retrospective cohort studies: advantages and disadvantages. British Medical Journal 348, g1072.
Strindberg LZ (1956) The dependence of the results of pulp therapy on certain factors: an analytic study based on radiographic and clinical follow-up examinations (PhD Thesis). Stockholm, Sweden.
Toure B, Faye B, Kane AW, Lo CM, Niang B, Boucher Y (2011) Analysis of reasons for extraction of endodontically treated teeth: a prospective study. Journal of Endodontics 37, 1512-5.
Tzimpoulas NE, Alisafis MG, Tzanetakis GN, Kontakiotis EG (2012) A prospective study of the extraction and retention incidence of endodontically treated teeth with uncertain prognosis after endodontic referral. Journal of Endodontics 38, 1326-9.
Vahdati SA, Torabinejad M, Handysides R, Lozada J (2019) A retrospective comparison of outcome in patients who received both nonsurgical root canal treatment and single-tooth implants. Journal of Endodontics 45, 99-103.
Vire DE (1991) Failure of endodontically treated teeth: classification and evaluation. Journal of Endodontics 17, 338-42.
Accepted Article
von Elm E, Altman DG, Egger M, et al. (2014) The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: guidelines for reporting observational studies. International Journal of Surgery 12, 1495-9.
Wigsten E, Jonasson P, EndoReCo, Kvist T (2019) Indications for root canal treatment in a Swedish county dental service: patient- and tooth-specific characteristics. International Endodontic Journal 52, 158-68.
Table 1 Frequencies of various factors in the material derived from dental records and radiographs, numbers of extraction events (NOE) in relation to each factor, and hazard ratios regarding possible prognostic factors related to tooth survival after completed root canal treatment performed by general dental practitioners.
NOE Cox regression
univariate analysis
Variable Frequencies, and
means when appropriate (n, means when appropriate) n Hazards ratio (95% CI) P value Individual-related factors
Age (reported in 2006) years mean 45.24 (range
10-80, SD 15.16) n=280 47 0.997 (0.980-1.016) for each additional year 0.784 Gender * male (0) female (1) 140 140 28 19 0.672 (0.375-1.203) 0.181
Disease (reported; not healthy = diabetes, immunosuppressive medications or conditions, heart conditions) healthy (0) not healthy (1) 242 23 40 5 1.340 (0.526-3.412) 0.539 Preoperative factors
Tooth type ˠ non-molars (0)
molars (1) 135 145 16 31 1.899 (1.038-3.474) 0.037
Accepted Article
Number of contact points * (assessed by radiographs; at baseline, or after if assessment could reflect the situation at baseline)
0 (0) more than 0 (1) 7 237 3 36 0.406 (0.125-1.319) 0.134
Terminal arch location (assessed by radiographs) no (0) yes (1) 228 27 37 2 0.417 (0.100-1.729) 0.228
Presence of sinus tract (reported or radiographic signs of tracing sinus tract) no (0) yes (1) 199 23 30 3 0.900 (0.274-2.952) 0.862
Apical status (assessed by radiographs, according to Strindberg’s criteria [Strindberg 1956]) normal status (0) abnormal status (1) 61 69 3 10 1.779 (0.646-4.90) 0.265
Previous root filling (reported or assessed by radiographs; signs of presence of any filling material in pulp space before treatment)
no (0) yes (1) 222 8 33 2 1.798 (0.431-7.499) 0.420
Marginal status (assessed by radiographs; estimation of bone level in relation to root length) cervical 1/3 (0) apical 2/3 (1) 230 11 21 3 2.033 (0.621-6.653) 0.241 Intra-operative factors
Antibiotics (reported during treatment) no (0)
yes (1) 173 44 28 6 0.733 (0.303-1.774) 0.492
Intra-pulpal anaesthesia (reported upon access) no (0) 208 34 0.048 0.525
yes (1) 5 0 (0.000-571.504) Acute problems ˠ (presenting with acute problems at start and during treatment) once or less (0)
twice or more (1) 203 15 26 7 3.879 (1.682-8.943) 0.001
Caregivers (number of caregivers involved from access to completion of root canal treatment) one (0) two or more (1) 159 73 25 11 0.917 (0.458-1.837) 0.808
Number of sessions ˠ (to complete treatment) less than five (0) more than five (1)
202 31 25 11 2.930 (1.441-5.955) 0.003
Elapsed time between access preparation and root filling Days mean 118 (range 0‒1697, SD 176) n=216 33 1.001 (0.999-1.003) for each additional day 0.233
Rubber dam (evidence of use of rubber dam; i.e., clamp seen in radiograph or notation in journal) no (0) yes (1) 19 109 2 18 1.040 (0.241-4.497) 0.958
Pulp bleeding (reporting of bleeding pulp in all canals upon access) no (0) yes (1) 177 33 27 6 1.239 (0.511-3.004) 0.635
Pulp necrosis * (reporting pulp necrosis in at least one canal upon access) no (0) yes (1) 130 80 24 9 0.539 (0.250-1.160) 0.114
Accepted Article
Apical size (reporting maximal apical size preparation in smallest canal) (ISO) 208 (range #10 ‒ #80) 33 0.993 (0.969-1-018)
0.582
Type of root canal filling (reporting gutta percha in combination with) sealer (0) rosin-chloroform (1) 152 77 25 0 0.767 (0.368-1.597) 0.479
Root-filling quality (assessed on radiograph: good if length approximately 0–2mm from apex and no voids)
adequate (0) not adequate (1) 116 69 17 10 0.997 (0.447-2.134) 0.953
Procedural accidents (assessed on radiographs, or reporting of file fractures or iatrogenic perforations) no (0) yes (1) 170 4 24 1 1.834 (0.247-13.588) 0.553 Postoperative factors
Postoperative pain (reporting of care-seeking for postoperative pain up to 4 weeks after completed treatment) no (0) yes (1) 222 8 36 1 0.858 (0.117-6.282) 0.880
Extent of first permanent restoration after completed treatment * (partial coverage with direct filling and full coverage – direct or indirect, new or old)
partial (0) full (1) 171 77 33 7 0.468 (0.207-1.058) 0.068 Elapsed time between completed treatment and first restoration Days mean 32.81 (range
0‒581, SD 74.51) n=277 46 0.999 (0.995-1.004) for each additional day 0.737
Accepted Article
Use of plug in the orifice of the root canal coronal to the root-filling no (0) yes (1) 164 105 29 17 0.916 (0.503-1.667) 0.774
Post (placement of post after completed treatment) no (0) yes (1) 211 52 35 7 0.759 (0.337-1.709) 0.505
Abutment (tooth used as an abutment for fixed [n=2] or removable prosthesis [n=4] after completed treatment) no (0) yes (1) 253 6 40 1 1.45 (0.199-10.586) 0.714
95% CI – 95% confidence interval; NOE – number of events (i.e. extractions); SD ‒ standard deviation ˠ variables with p<0.05; * variables with 0.05>p<0.20
FIGURE LEGENDS
Figure 1 Flow chart of included and excluded individuals from the total cohort of patients receiving a non-surgical root canal treatment performed by GDPs in Norrbotten during 2006.
Figure 2 Kaplan-Meier survival curve for teeth with completed non-surgical root canal treatments performed by GDPs Norrbotten during 2006.
