Refers to Erratum

Journal of Endodontics, Volume 49, Issue 10, October 2023, Pages 1376-1377

View PDF

Journal of Endodontics, Volume 49, Issue 6, June 2023, Pages 664-674

Yoshi Terauchi, Mahmoud Torabinejad, Kingsley Wong, George Bogen

View PDF



No clinical studies have examined the effect of mineral trioxide aggregate (MTA) obturation levels on the outcome of endodontic retreatment. This retrospective study examined treatment outcomes in three cohorts that compared overfilling, flush filling, and underfilling after orthograde retreatment using MTA.


Two hundred fifty patients with 264 teeth diagnosed with previously treated root canals and apical periodontitis retreated in a private endodontic practice were included. All teeth received MTA obturation and the extent of the final filling level was measured in relation to the major apical foramen. After 6-month follow-ups, all nonhealing cases were treated surgically. After 24- to 72-month reviews, the effect of preoperative lesion size and the degree of MTA obturation level were assessed. Multiple linear regression and time-to-event analysis using Stata 17 software (StataCorp LLC, College Station, TX) were used to evaluate the data.


Within the three cohorts, 99 out of 108 overfilled teeth (91.7%), 90 out of 103 flush fills (87.4%), and 10 out of 53 underfilled teeth (18.9%) healed and were successfully retreated without surgery at 48-months. When surgical outcomes were included, the combined healed proportion was 93.2%. Preoperative lesion size was found to be an important predictor for retreatment nonhealing. A 1-mm increase in lesion size at baseline resulted in an estimated 11% (95% CI 1.04, 1.18)–38% (95% CI 1.22, 1.58) increase in the risk of surgery. Compared to overfilling and flush filling, underfilling was associated with an approximately three-fold increase in requiring surgery and characterized by delayed healing.


MTA obturation is a viable retreatment option for teeth with nonhealing endodontic treatment. MTA overfills or flush fillings do not adversely affect healing outcomes. However, MTA underfilling increases the chances for nonhealing and surgical intervention.

Key Words

MTA cement



periapical lesion



The outcomes of this retrospective cohort study suggest that MTA obturation is an effective option for retreatment of previously root canal–treated teeth with periapical lesions. MTA extrusion does not reduce the rate of periapical healing.

Obturation of the root canal system is an integral component of nonsurgical endodontic treatment. Ideally, the procedure should ensure that the entire root canal system is obturated without extrusion of the root-filling material beyond the apical constriction1. This prevents irritation of the periradicular tissue2, inhibits microbial recolonization, and denies residual microorganisms access to nutrients and potentially further propagation3. The quality of root filling and obturation length have been shown by meta-analysis to be a significant prognostic factor for nonsurgical retreatment4,5.

Thermoplastic and cold-compaction obturation techniques have both been shown to provide acceptable obturation quality, but show limited resistance to microleakage6,7. The two commonly-used obturation techniques have also been shown to have similar treatment outcomes7. Importantly, an adequate coronal restoration has been shown to be a significant predictor of success8,9. Moreover, the ability of gutta-percha and sealer to prevent microleakage is questionable in modern endodontic treatment6,10. Consequently, bioceramic sealers have been developed for the purpose of improving the seal-ability, antimicrobial effect, and biocompatibility of conventional obturation methods11,12. Although absent in well-controlled clinical trials, bioceramic sealers combined with a well-fitted single gutta-percha cone have shown promising success rates in a retrospective cohort study13. Despite these materials demonstrating advanced sealing properties, they may not have characteristics preferable to unmodified mineral trioxide aggregate (MTA) cement. Improved antibacterial properties, generation of an interfacial layer with dentin and consistent reformation of the periodontium may be expected when MTA cement is applied in retreament cases14, 15, 16, 17.

Obturation using unmodified MTA or other calcium silicate cements (CSCs) is also associated with some drawbacks that may not be crucial in teeth requiring retreatment. Staining teeth under full coverage restorations and slower setting-times do not necessarily adversely affect esthetic or biological outcomes15. Additionally, removal and retreatment of set MTA may not be a critical factor in healing, as nonhealing in the majority of these cases often simplifies indicated surgical treatment16,18. This can be advantageous where access to surgical sites is difficult and placement of retrograde fillings is challenging16. Moreover, efficient obturation of canals using MTA or other CSCs is expected when operators become familiar with the handling properties of these materials.

Root canal retreatment presents a unique clinical challenge that is complicated by the need to remove prior obturation materials, overcome procedural errors, reduce bacterial contamination, and promote apical healing when pathosis is evident. Failure of root canal treatment can be attributed to both intraradicular and extraradicular infections involving biofilms, cysts, foreign body reactions, and cholesterol clefts19,20. Frequently, teeth with long-standing apical periodontitis can also exhibit large lesions and inflammatory apical root resorption21,22. Obturation using CSCs such as MTA does not predictably ensure successful treatment but has shown promise for cases with root resorption and open apices23,24. However, it is often difficult to prevent unwanted extrusion of root filling material in cases with open apices without providing apical barriers25. Furthermore, filling material extrusion has consistently been shown to adversely affect treatment outcomes1,5,26.

In cases refractory to primary orthograde treatment, retreatment procedures should focus on methods that promote the highest probability of periapical healing and limit secondary interventions27. Although gutta-percha materials and bioceramic sealers are the generally accepted treatment modality, clinicians may want an alternative retreatment method for the more challenging nonhealing cases. A search of literature shows an absence of clinical studies that have examined the effect of the level of MTA root canal obturation on the healing rates of root canal retreatment in teeth exhibiting periapical disease. The aim of this retrospective study was to investigate the effect of MTA obturation levels on the outcome of endodontic retreatment.

Materials and Methods

Study Design and Participants

The study examined a patient pool comprised of cases treated between 2013 and 2022 by an endodontic specialist at a private practice in Tokyo, Japan. Informed consents were obtained from all adult patients and parents of all minors after discussing risks, benefits and outcomes of the treatment. Patients showing periapical radiolucencies 3 mm or larger in diameter on preoperative cone-beam computed tomography (CBCT) imaging (FineCube; Yoshida Dental Mfg. Co., Ltd., Tokyo, Japan), received MTA obturation during endodontic retreatment. Patients included in the study were recalled a period of atleast 24–72 months for both orthograde and retrograde treatment outcomes.

A total of 250 patients (out of 3327) met the inclusion criteria. Patients with unremarkable health histories, previous orthograde treatments, and adequate restorations were included. Furthermore, the possible influence of these factors was distributed between all three groups. Case selection also included patients showing periapical lesions measuring at least 3 mm or greater in diameter and diagnosed with symptomatic or asymptomatic apical periodontitis. Teeth with detectable preoperative vertical root fractures (VRF), advanced periodontal disease, or deemed unrestorable were excluded. Data included tooth number, tooth type, gender, the pre/postoperative maximum lesion diameter, maximum extent (amount) of the MTA overfill, underfill or flush fill, whether periapical surgery was required, and the time course to radiographic healing. Assessments were based on CBCT imaging with a maximum quantified image resolution of 0.3 mm. MTA obturations 0.3 mm or more beyond the major apical diameter were categorized as overfills (n = 108). Obturations even or less than 0.3 mm above or below the major apical foramen were considered flush fills (n = 103) and MTA fills 0.3 mm or greater below the major apical diameter designated as underfills (n = 53). All measurements were completed using FineCube software. Teeth that failed to demonstrate a measurable reduction in lesion size from CBCT review at 6 months after MTA retreatment were surgically treated using microsurgical techniques. Asymptomatic cases that showed reduction of the lesion size to 0.3 mm or smaller in diameter on CBCT imaging at least 6 months or longer after retreatment or surgery were classified as healed. After 24 to 72-month reviews, the effect of pre-operative lesion size and the degree of MTA obturation level were assessed.

Nonsurgical Retreatment and MTA Obturation

All endodontic treatments were performed under dental dam isolation using the dental operating microscope. After disassembly of restorations or post and cores, root canal filling materials were removed initially with Terauchi File Retrieval Kit-S ultrasonic tips (Guilin Woodpecker Medical Instrument Co., Ltd, Guangxi, China) and an XP Shaper/Finisher (FKG Dentaire SA, La Chaux-de-Fonds, Switzerland). When 3 mm or less of gutta-percha remained apically, an XP Shaper/Finisher rotating at 1000–2500 rpm in combination of a GPR hand instrument (Kohdent Roland Kohler Medizintechnik GmbH & Co. KG, Germany) was used with chloroform and removal confirmed radiographically. If separated instruments were present, they were removed prior to final obturation. The canals were then preflared with no. 2 or 3 Gates Glidden drills prior to negotiation of the apical third. The working length was established with an electronic apex locator (EAL) (Root ZX II; J Morita Corp, Kyoto, Japan) or Ai-Pex; (Guilin Woodpecker Medical Instrument Co.,Ltd, Guangxi, China) and confirmed radiographically. Definitive canal preparation was performed with either Vortex Blue/ProTaper Universal files (Dentsply Tulsa Dental Specialties, Johnson City, TN), EdgeTaper Platinum files (EdgeEndo, Albuquerque, NM), HyFlex EDM/CM files (Coltene Whaledent, Altstatten, Switzerland), or XP Shaper/Finisher files using QMix 2in1 chelation (Dentsply Tulsa Dental Specialties, Johnson City, TN). Root canal irrigation consisted of using 5.25% sodium hypochlorite solution (NaOCl) followed by 3% hydrogen peroxide solution (H2O2) throughout the procedure. All the canals were instrumented to a minimum size of ISO 35.

Further removal of obturation material was accomplished with the use of the no. 25 XP Finisher rotating at 1000–2500 rpm. This was followed by irrigation with 5.25% NaOCl for 30 seconds and Qmix 2in1(Dentsply Tulsa Dental Specialties, Tulsa, OK) for 30 seconds. Each root canal was then ultrasonically irrigated with 5.25% NaOCl using with the ProUltra PiezoFlow (Dentsply Tulsa Dental Specialties, Johnson City, TN) for sixty seconds. Finally, the root canals were saturated with 5.25% NaOCl for a further 10–20 minutes, replenishing every 5 minutes before drying the canals with sterile paper points or a Stropko irrigator (DCI International, Newberg, OR).

The apical size was gauged with no. 35 to 60 NiTi K-file (Dentsply Tulsa Dental Specialties, Johnson City, TN). A shaping file 1 or two sizes smaller than the apical size was used to condense MTA during canal obturation16,18. Gray or White ProRoot MTA (Dentsply Tulsa Dental Specialties, Johnson City, TN) was mixed with phosphate buffered saline on a glass slab and carried into the canal with a Buchanan Hand Plugger 0 (Kerr Corporation, Brea, CA). MTA was compacted with the selected rotary file in a few millimeter push/pull motion while rotating counterclockwise either manually or at 100 rpm using an endodontic motor. The rotary file was initially connected to an electronic apex locator that allowed increased accuracy during initial MTA placement. After apical closure was established, the file was moved coronally as the MTA plug was completed using various sized ultrasonic tips, pluggers, and rotary files. MTA compaction was advanced coronally from the apex to the level of the expected post insertion or core placement depth. No apical barrier membranes were used. Wet cotton pellets were placed over the MTA and the access cavity was temporally sealed with Cavit G (3M ESPE, St Paul, MN). The patients returned a week later for core or post and core placement and then recalled for an initial 3-month follow-up. Patients were further evaluated at 6 to 72-month observation periods.

Statistical Analysis

Descriptive statistics were used to summarize the characteristics of the study cohorts (overfilling, flush filling, and underfilling). For each cohort, generalized linear regression models with Poisson distribution, log-link function, and robust standard error were used to estimate the unadjusted and adjusted risk ratios (RR) and their 95% confidence intervals (CI) of requiring surgery after endodontic treatment had failed. The median healing time was estimated from the time-to-healing functions generated using the Kaplan-Meier method, and the log-rank test was used to test the equality of the functions across the three cohorts (Fig. 1). The data were also fitted using the Cox proportional hazards regression models to estimate the unadjusted and adjusted hazard ratios (HR) of healing. The assumption of proportional hazards was tested using Schoenfeld residuals. In the regression models, the exposure variables of interest were MTA amount and lesion size at baseline and the covariates were age, gender and tooth type. A univariable analysis was first performed to determine the unadjusted RR and HR, followed by a multivariable analysis using all variables to obtain the adjusted estimates which were subsequently reported.



  1. Download : Download high-res image (197KB)
  2. Download : Download full-size image

Figure 1. Estimated Kaplan-Meier time to healing curve showing the probability of healing at follow-up appointments in patients retreated with MTA overfilling, flush filling, and underfilling (n=205). excluded canine teeth (n = 2). MTA, mineral trioxide aggregate.

After combining the three cohorts into 1, the same analysis for an individual cohort was repeated with MTA obturation level as an additional exposure variable of interest. All analyses were carried out using Stata (version 17.0, StataCorp LLC, College Station, TX), and the Kaplan-Meier curve was plotted using R (version 4.2.2, R Foundation for Statistical Computing, Vienna, Austria) as well as the package “survminer” (version 0.4.9,


The baseline characteristics of the 264 teeth included in the study are shown in Table 1. The majority of the teeth were from female patients and the average age at the time of retreatment was 47.1 years (SD = 13.1), 48.3 years (SD = 12.6), and 50.1 years (SD = 11.6) in the overfilling, flush filling, and underfilling cohort, respectively. Molars were the predominant tooth type in the nosurgery and surgery subgroups across the three cohorts. Seven teeth in the surgery subgroup and 1 in the nonsurgery subgroup were extracted. Five teeth were extracted because of VRF, and three teeth were extracted due to periodontal conditions.

Table 1. Baseline Characteristics of 264 Teeth Included by MTA Filling Level and Retreatment Group

Characteristics Overfilling (n = 108) Flush filling (n = 103) Underfilling (n = 53)
No surgery Surgery No surgery Surgery No surgery Surgery
Number of teeth 99 (91.7) 9 (8.3) 91 (88.4) 12 (11.7) 17 (32.1) 36 (67.9)
Age, years 47.2 (13.6) 46.6 (3.8) 48.0 (13.1) 50.5 (6.6) 48 (9.6) 51.0 (12.4)
 Male 20 (20.2) 4 (44.4) 24 (26.4) 5 (41.7) 4 (23.5) 13 (36.1)
 Female 79 (79.8) 5 (55.6) 67 (73.6) 7 (58.3) 13 (76.5) 23 (63.9)
Tooth type
 Incisor 11 (11.1) 1 (11.1) 6 (6.6) 2 (16.7) 2 (11.8) 1 (2.8)
 Canine 2 (2.2)
 Premolar 7 (7.1) 2 (22.2) 8 (8.8) 3 (25.0) 3 (17.7) 8 (22.2)
 Molar 81 (81.8) 6 (66.7) 75 (82.4) 7 (58.3) 12 (70.6) 27 (75.0)
Lesion size at baseline, mm 6.6 (5.3,8.1) 9.8 (5.7,13.1) 6 (4.7,8.3) 9.5 (8.4,11.2) 5.1 (4.6,7.8) 8.5 (6.7,10.0)
MTA amount, mm 0.88 (0.52) 0.93 (0.57) 0 0 −1.42 (0.75) −1.80 (0.99)


n, number of teeth; MTA, Mineral Trioxide Aggregate; SD, standard deviation; Q1, first quartile; Q3, third quartile; mm, millimeter

The values are count and proportion unless otherwise stated.


Presented as mean (standard deviation).
Presented as median (1st quartile, 3rd quartile).

Overfilling Cohort

At the 6-month follow-up appointment, close to 2/3 of these teeth (n = 65 out of 99, 65.7%) were classified as healed (Table 2). Ninety-nine of 108 teeth in this cohort (91.7%) were successfully retreated with MTA overfilling at the 48-month recall period (overfill amount: 0.88 mm, SD: 0.52) (Table 2), (Fig. 2). The median healing time was 6 months (95% CI 6, 6) (Fig. 1). In the nosurgery subgroup, the mean MTA overfill amount at baseline was 0.88 mm (SD: 0.52) (Table 1). The median lesion size at baseline was 6.6 mm (first quartile [Q1] 5.3 and third quartile [Q3] 8.1) (Table 1). In comparison, for teeth that required surgical intervention (n = 9 out of 108, 8.3%), the median lesion size was 9.8 mm (Q1 7.7, Q3 13.1). Two teeth in the surgery subgroup were eventually extracted due to VRF.

Table 2. Proportions of 264 Teeth at Recall Periods Classified as Healed by MTA Filling Level and Retreatment Group

Overfilling Flush filling Underfilling Overall
NS S Total NS S Total NS S Total NS S Total
n 99 9 108 91 12§ 103 17 36 53 207# 57∗∗ 264
Recall period
 6 months 65 (65.7) (60.7) 5 (55.6) (4.6) 61 (67.0) (59.2) 6 (50.0) (5.8) 6 (35.3) (11.3) 25 (69.4) (47.2) 132 (63.8) (50.0) 36 (63.2) (13.6)
 12 months 91 (91.9) (84.3) 7 (77.8) (6.5) 86 (94.5) (83.5) 10 (83.3) (9.7) 8 (47.1) (15.1) 30 (83.3) (56.6) 185 (89.4) (70.1) 47 (82.5) (17.8)
 24 months 98 (99.0) (90.7) 88 (96.7) (85.4) 10 (58.8) (18.9) 196 (94.7) (74.2)
 36 months 99 (100) (91.7) 89 (97.8) (86.4) 198 (95.7) (75.0)
 48 months 90 (98.9) (87.4) 199 (96.1) (75.4)
Overall, N (% of total) 106 (98.1) 100 (97.1) 40 (75.5) 246 (93.2)


n, number of teeth; MTA, Mineral Trioxide Aggregate; NS, no surgery; S, surgery


At the first recall period that showed the tooth had healed.
Includes extracted or non-healing teeth: n = 2.
n = 1.
n = 2.
n = 7.
n = 6.
n = 8.
n = 10.


  1. Download : Download high-res image (974KB)
  2. Download : Download full-size image

Figure 2. MTA retreatment overfill. (A) Preoperative radiograph shows periapical lesions associated with maxillary right second molar (arrows). (B, C) CBCT sagittal and coronal views show large periapical lesion surrounding palatal root with radiolucent material extrusion (arrows). (D, E) CBCT sagittal and coronal images showing small periapical lesions on both mesiobuccal (MB) and distobuccal roots (arrows). (F) CBCT axial view of large periapical lesion on palatal root extending to the MB root (arrow). (G) Microscope photograph shows fissure extending from the palatal root to the MB root (arrow). (H) Photograph of the groove filled with white MTA (arrow). (I) Postoperative radiograph shows MTA obturation in all canals with significant MTA overfill in the palatal root (arrow). (J) Six-month postoperative radiograph shows advancing remineralization of pre-existing periapical lesions (arrows). (K) Twenty-four-month postoperative follow-up radiograph shows normal periapical tissues with a permanent full coverage restoration. (L, M) Twenty-four-month postoperative CBCT sagittal and coronal images revealing complete resolution of periapical lesions and circumferential bone reformation around palatal root (arrow). (N–O) Twenty-four-month postoperative CBCT sagittal and coronal images displaying complete resolution of the periapical lesions on both buccal roots (arrows). (P) Twenty-four-month postoperative CBCT axial view shows remineralization of supporting bone between palatal and mesiobuccal roots (arrow). CBCT, cone-beam computed tomography; MTA, mineral trioxide aggregate.

Lesion size at baseline appeared to be associated with surgical intervention. With each millimeter increase in size resulting in an estimated 16% (adjusted RR 1.16, 95% CI 0.98, 1.36) increase in risk (Table 3). Among the tooth samples successfully retreated, increasing lesion size at baseline negatively impacted the healing time by 11% (adjusted HR 0.89, 95% CI 0.81, 0.98) for each millimeter increase (Table 4).

Table 3. Estimated adjusted Relative Risks of Requiring Surgery at 24–72 month Review as Predicted by Lesion Size at Baseline and MTA Filling Level (n = 262)

Variables Overfilling (n = 108) Flush filling (n = 101) Underfilling (n = 53) Combined (n = 262)
aRR (95% CI), P value aRR (95% CI), P value aRR (95% CI), P value aRR (95% CI), P value
MTA amount, mm 1.19 (.47, 3.03), .719 n/a .83 (.69, .99), .040 .76 (.62, .93), .009
Lesion size at baseline, mm 1.16 (.98, 1.36), .080 1.38 (1.22, 1.58), <.001 1.11 (1.04, 1.18), .002 1.18 (1.10, 1.27), <.001
MTA filling level
 Over n/a n/a n/a .86 (.38, 1.95), .724 ref
 Flush n/a n/a n/a ref 1.16 (0.51, 2.62), 0.724
 Under n/a n/a n/a 3.08 (1.58, 6.01), .001 3.56 (1.50, 8.49), 0.004


n, number of teeth; MTA, Mineral Trioxide Aggregate; aRR, adjusted risk ratio; ref, reference category; n/a, not applicable.


The risk ratios were adjusted for other reported variables and covariates including age, sex and tooth type.
excluded canine teeth (n = 2).

Table 4. Estimated adjusted Hazard Ratios for Time Required for Retreatment Case Classification as Healed When Predicted by Lesion Size at Baseline and MTA Filling Level (n = 205)

Variables Overfilling (n = 99) Flush filling (n = 89) Underfilling (n = 17) Combined (n = 205)
aHR (95% CI), P value aHR (95% CI), P value aHR (95% CI), P value aHR (95% CI), P value
MTA amount, mm 1.06 (.73, 1.54), .768 n/a 1.27 (.39, 4.12), .689 .95 (.67, 1.36), .789
Lesion size at baseline, mm .89 (.81, .98), .013 .91 (.82, 1.00), .060 1.32 (.85, 2.06), .215
MTA filling level
 Over n/a n/a n/a 1.06 (.69, 1.62), .794 ref
 Flush n/a n/a n/a ref 0.95 (0.62, 1.44), 0.794
 Under n/a n/a n/a .36 (.15, .84), .018 0.34 (0.12,0.97), 0.043


n, number of teeth; MTA, Mineral Trioxide Aggregate; aHR, adjusted hazard ratio; ref, reference category; n/a, not applicable.


The hazard ratios were adjusted for other reported variables and covariates including age, sex and tooth type.
excluded canine teeth (n = 2).
Lesion size at baseline violated the proportional hazards assumption. The covariate was categorized and included in the model as strata, and hence its results were not shown.

Flush Filling Cohort

The proportion of teeth reported to be healed at the 6-month follow-up appointment after nonsurgical retreatment was 67.0% (n = 61 out of 91). The healed proportion at the 6-month recall after surgery was 50.0% (n = 6 out of 12) (Table 2). The number of teeth successfully healed in the nosurgery subgroup increased to 90 out of 103 teeth (87.4%) after the 48-month follow-up (Table 2, Fig. 3). In this subgroup, the median healing time was six months (95% CI 6, 6) (Fig. 1) and the median lesion size at baseline (6 mm, Q1 4.7, Q3 8.3) was smaller than that of the surgery subgroup (9.5 mm, Q1 8.4, Q3 11.2) (Table 1). One tooth in the nosurgery subgroup with a VRF was extracted and two in the surgery subgroup were deemed refractory to treatment at 36 and 48 months. Similar to the previous cohort, increasing lesion size at baseline was associated with a heightened risk of surgery after retreatment (adjusted RR 1.38, 95% CI 1.22, 1.58) and delayed healing (adjusted HR 0.91, 95% CI 0.82, 1.00) (Tables 3 and 4).


  1. Download : Download high-res image (836KB)
  2. Download : Download full-size image

Figure 3. MTA retreatment flush fill. (A) Preoperative radiograph of previously treated maxillary right lateral incisor displaying a large cast post and full coverage restoration with periapical lesion and apical root resorption. (B–D) CBCT sagittal, coronal and axial images showing extent of lesion and perforation of the palatal bone (arrows). (E) Microscope photograph of previous filling materials in a ledged labial space after post and core removal. (F) Microscope photograph of MTA obturated. (G) Immediate postoperative radiograph confirming flush MTA obturation. (H–J) Recall radiographs at 3, 6, and 24 months showing bone reformation of previous periapical defect. (K) CBCT coronal view at 24 months showing osseous repair of the palatal plate (arrow). (L, M) CBCT coronal and axial views confirming remineralization of lesion and reformation of palatal bone at 24 months (arrows). CBCT, cone-beam computed tomography; MTA, mineral trioxide aggregate.

Underfilling Cohort

Contrary to the other two cohorts, the majority of teeth in this cohort were treated surgically (n = 36 out of 53, 67.9%) (Table 2), Furthermore, most of the teeth (n = 25 out of 36, 69.4%) in the surgery subgroup were healed at the 6-month follow-up appointment after surgery. However, only approximately 1/3 (n = 6 out of 17, 35.3%) of the teeth in the no surgery subgroup were observed to be healed at the same time point. Ten of the 53 teeth (18.9%) retreated with MTA underfilling healed without requiring surgery at the 24-month review (Table 2, Fig. 4), and the median healing time was 24 months (95% CI 6) (Fig. 1). In the nosurgery subgroup, the average amount of MTA underfilling at baseline was -1.42 mm (SD: 0.75), slightly less than that in the surgery subgroup (-1.80 mm, SD: 0.99) (Table 1). The median lesion size at baseline of the nosurgery and surgery subgroup in this cohort was 5.1 mm (Q1 4.6, Q3 7.8) and 8.5 mm (Q1 6.7, Q3 10.0), respectively. Four out of 36 (11.1%) surgical cases were extracted at the time of surgery due to VRF and 1 due to nonhealing. There were more nonhealing teeth in the nosurgery subgroup (n = 7 out of 17, 41.2%) than in the surgery subgroup (n = 1 out of 38, 2.8%).


  1. Download : Download high-res image (1MB)
  2. Download : Download full-size image

Figure 4. MTA retreatment underfill. (A) Preoperative radiograph shows previously treated maxillary left first and second molars (no. 14, 15) with periapical pathosis associated with mesiobuccal (MB) root no. 15 (arrow). (B) CBCT sagittal view reveals mucositis of the sinus floor (arrows) and the apical foramen location of the curved MB root (arrow). (C) Sagittal CBCT view shows extent of periapical lesion measuring 8mm in length (arrow). (D) Coronal CBCT view reveals periapical lesions associated with both MB and palatal roots no. 15. (E) CBCT axial image demonstrates loss of supporting bone and cortical plate of buccal roots no. 15 (arrows). (F) Periapical radiograph shows negotiation file blocked by previous ledge formation (arrow). (G) Periapical radiograph showing MTA retreatment obturation of both molars with bonded composite cores (no. 14, 15). Note obturation short of apical foramen no. 15. (H) Twenty-four-month postoperative radiographic displays recently placed full coverage restorations and remineralization of MB periapical lesion no. 15. (I) Twenty-four-month CBCT sagittal image reveals normal mucous membrane (arrows). Note transported MTA obturation at perforation site of ledged MB canal and osseous repair of periapical area (arrow). (J, K) Twenty-four-month sagittal and coronal CBCT recall images confirming successful repair of periapical bone (arrow). (L) Twenty-four-month axial CBCT follow-up shows reformation of periradicular supporting bone. CBCT, cone-beam computed tomography; MTA, mineral trioxide aggregate.

For each millimeter of decreased MTA underfilling (ie, less underfilling), the surgery risk after retreatment was reduced by 17% (adjusted RR 0.83, 95% CI 0.69, 0.99) (Table 3). As observed in other two cohorts, lesion size at baseline was a risk factor for surgery, with each millimeter increase in size resulting in an estimated 11% (adjusted RR 1.11, 95% CI 1.04, 1.18) increase in risk. There were no notable associations between healing and MTA level or lesion size at baseline (Table 4).

Combined Cohort

The combined cohort’s success fraction for nonsurgical retreatment in this retrospective study was 75.4% (n = 199 out of 264) (Table 2). When only the overfilling and flush filling cohorts were considered, the proportion increased to 89.6% (n = 189 out of 211). The success fraction of non-surgical retreatment that healed using the MTA obturation technique was 91.7% (n = 99 out of 108) for overfilling, 87.4% (n = 90 out of 103) for flush filling, and 18.9% (n = 10 out 53) for the underfilling cohorts. After long-term review, the healed proportion of MTA retreated teeth combined with nonhealing teeth requiring surgery was 93.2% (n = 246 out of 264) (Table 2).

MTA underfilling, as a group, was associated with at least a three-fold increase in requiring surgical intervention after retreatment when compared to flush filling (adjusted RR 3.08, 95% CI 1.58, 6.01) and overfilling (adjusted RR 3.56, 95% CI 1.50, 8.49) (Table 3). However, there was no discernible difference in the risk between the overfilling and the flush filling cohorts (adjusted RR 1.16, 95% CI 0.51, 2.62). Notably, an incremental increase in MTA level lessened the likelihood of having surgery by 24% (adjusted RR 0.76, 95% CI 0.62, 0.93) (Table 3, Fig. 5). Healing was delayed close to three-fold in the underfilling cohort compared to the flush filling (adjusted HR 0.36, 95% CI 0.15, 0.84) and overfilling cohorts (adjusted HR 0.34, 95% CI 0.12, 0.97). Nevertheless, the healing time was not affected by incremental change in the extent of MTA obturation level (adjusted HR 0.95, 95% CI 0.67, 1.36) (Table 4).


  1. Download : Download high-res image (758KB)
  2. Download : Download full-size image

Figure 5. MTA retreatment overfill. (A) Preoperative radiograph shows large osseous lesion associated with the maxillary right first (no. 3) and second molars (no. 2) (arrow). (B, C) CBCT sagittal and coronal views show large periapical lesions between the palatal roots of the molars extending to the crestal bone (arrows). (D) CBCT coronal and sagittal views reveal lesions elevating the sinus floor (arrow). (E, F) CBCT axial and coronal images showing perforation of the palatal plate and loss of supporting bone (arrows). (G) Microscope photograph showing hemorrhaging upon access in the MB canal no. 3 (arrow). (H) Postoperative radiograph shows MTA overfills in each canal of the first molar with new composite core. (I) Three-month postoperative radiograph shows MTA extrusion in no. 2 and advancing remineralization of osseous defects. (J) Forty-eight-month postoperative radiograph shows resolution of the pre-existing periapical lesions with a new coronal restoration. (K, L) Postoperative 48-month CBCT sagittal and coronal views shows complete reformation of the supporting bone no. 2, 3 (arrows). (M) Forty eight-month postoperative CBCT sagittal view of no. 2 showing resolution of lesions with a normal sinus floor. (N, O) Axial and coronal CBCT recall images demonstrating reformation of the cortical bone and repair of osseous defects (arrows). CBCT, cone-beam computed tomography; MTA, mineral trioxide aggregate.


Nonsurgical root canal retreatment is a challenging procedure as the presence of previous iatrogenic errors, root canal filling materials, missed main or accessory canals and residual bacteria/biofilms in inaccessible areas create unique problems not typically present in the initial treatments4,5,20. Excluding the underfill cohort, the outcome for MTA overfilling and flush filling in this retrospective study compares favorably to the 69%–81% success rates in retreatment reported in previous studies4,5. However, the underfilling cohort showed the lowest healing rate, requiring surgical intervention in the majority of cases at 6 months (69.4%). Importantly, this study retreated many compromised teeth not included in previous published studies. A large proportion of MTA obturated and underfilled teeth were challenging cases primarily attributed to blocked, ledged, or calcified canal systems (Fig. 4). Although direct comparison with previous outcome studies is difficult due to differences in methodology and criteria, the strict definition of success used in this study should limit any potential bias of the outcome rate.

The influence of preoperative lesion size on the outcome of nonsurgical endodontic treatment is not universally agreed upon. The results of this study found that preoperative lesion size was a significant predictor for requiring surgical intervention. For every millimeter increase in lesion size, the risk for requiring surgical treatment increased by an estimated 16% in overfilling, 38% in flush filling, and 11 % in the underfilling cohort. Correspondingly, the nosurgery subgroup in the flush filling cohort demonstrated smaller median lesion sizes at baseline than those of the surgical subgroup. Furthermore, increasing lesion size at baseline was also associated with delayed healing in the overfilling and flush filling cohorts. There was inconclusive evidence that the risk of surgery differed by gender, age, and tooth type in the three cohorts.

Our findings related to the size of the lesions are in agreement with the results reported in a systematic meta-analysis, where orthograde retreatments with larger preoperative lesions were associated with significantly lower healing rates than those with smaller periapical lesions9. The data also appear to support the concept that larger preoperative lesions may be more critical to healing outcomes than the actual MTA obturation level. In contrast, Sjogren et al., reported that the diameter of the preoperative lesion was not a significant predictor for treatment outcome1. A possible explanation for reduced healing rates in teeth with large periapical lesions may be the presence of more established intraradicular or extraradicular biofilms. These can be more difficult to eradicate through conventional nonsurgical treatment20. The improvement in success rates in the two cohorts over those in recent studies could be attributed to stringent irrigation protocol and/or the antibacterial and bioactive properties of MTA/CSC materials28,29.

The underfilling cohort, although smaller in size, demonstrated slower healing in the nosurgery subgroup compared to the other two cohorts. Moreover, the majority of teeth (67.9%) in this sample required surgical intervention to resolve nonhealing lesions. Interestingly, the surgical cases in this cohort healed more rapidly than both the other cohorts at the six-month recall period. It might be surmised that nonnegotiable canals encountered during MTA retreatment that were blocked and could not be obturated flush or beyond the major apical foramen, had higher risk for surgical treatment. Nonnegotiable canals were most likely the result of complex anatomy, calcifications, missed apical bifurcations, previously transported canals, or ledged apical preparations (Fig. 4). In clinical practice, this can be an important consideration in the management and informed consent of patients undergoing retreatment that receive underfillings after MTA obturation.

The retreatment outcomes of different MTA obturation levels in this study suggest that healing rates may improve using unmodified MTA in cases where the level of filling is flush or beyond the major apical foramen (Figs. 2, 3 and 4). Additionally, after inclusion of all three cohort nonhealing cases treated surgically, the combined healing rate was noteworthy. This favorable outcome may be attributed to the many physicochemical properties of MTA. They include the formation of an interfacial hybrid apatite diffusion zone, generation of an alkaline pH, and the phenomenon of intratubular mineralization14, 24, 29. During the setting process, leaflet-like crystals form along the length of the dentinal and intracanalicular tubules, subsequently neutralizing persistent microbial pathogens29. However, this process may be ineffective in reducing extraradicular biofilms if the obturation material cannot come in direct contact with microorganisms. Bacteria residing outside the canal system protected by biofilms can prevent healing and increase indications for surgical alternatives20,27.

The biocompatibility and bioactivity properties of MTA are further supported in this study as the degree of MTA extrusion was infrequently associated with nonhealing (Fig. 5). This result differs to that reported previously, where extrusion of root filling material was associated with compromised outcome and slower healing1, 5, 26, 30, 31. MTA cement has been established to be well-tolerated when in contact with periradicular tissues in studies examining apexification and perforation repairs23, 24, 25. This can be a distinct advantage in cases with advanced apical root resorption and long-standing periapical pathosis.

Despite its relatively large sample size and long follow-up periods, there are some shortcomings in this study that deserve consideration. The retrospective nature of this investigation is its main drawback. Ideally, it would have been preferred that this study had been carried out in a randomized prospective format with adequate controls. However, setting up randomized clinical prospective studies with various cohorts such as level of obturation and appropriate controls that could include other obturation materials has its own difficulties and ethical issues. The only manner to set up randomized prospective clinical investigations with adequate controls is to carry them out in experimental animals. Notwithstanding, these studies also have their own disadvantages.

Although the number of teeth in the underfilling cohort group was only approximately half of the two other cohorts, it is plausible that the inclusion of additional underfilling cases may have produced different data. However, it is also important to note that many nonhealing MTA retreament cases can be successfully managed by surgical intervention, supporting the ultimate goal of continued tooth retention. Hence, data outcome from surgical treatment was included in this study. Another limitation of this investigation is the fact that all of the cases were performed by one experienced operator (Y.T.). The results might have been different if this investigation was completed in multiple locations by various operators with different training and experience using MTA obturation.

All teeth that received retreatment and demonstrated no measurable reduction in lesion size after CBCT review at 6-months were surgically treated. There still remains the probability that some of these MTA obturated cases could have healed allowing longer observation periods without surgical intervention32. Studies have shown delayed healing can be expected in a large portion of teeth that have undergone conventional retreatment30,31. However, long-term follow-up periods do not ensure that healing will emerge after retreatment33. Accordingly, the slower healing dynamics and large number of nonhealing cases reflected in the underfilling cohort may have been the consequence of untreated apical canal spaces33. The advantages of surgical intervention in these cases included expedited healing periods and reduction of possible late post-treatment complications34.

Our findings show that approximately 2/3 of teeth in the nosurgery overfill and flush fill subgroups showed definitive signs of healing at 6 months. This is higher than Orstavik’s study where only 50% of successfully treated teeth showed signs of healing at 6 months35. Other studies have demonstrated delayed healing in retreated teeth by up to 17–27 years, especially in cases with extruded root filling materials32, 33. It appears that the time required for complete remineralization of apical defects may be shortened when unmodified MTA is used as an obturation material. The influence of MTA on the speed of healing needs to be further explored in future randomized clinical trials or case-controlled studies.


MTA obturation is a viable alternative to the conventional filling materials currently used during retreatment procedures. In addition, extrusion of MTA does not adversely influence the outcomes of nonsurgical retreatment. However, indications for surgical treatment increase after retreatment when MTA obturation cannot reach the apical foramen.

CRediT authorship contribution statement

Yoshi Terauchi: Conceptualization, Resources, Investigation, Data curation, Writing – review & editing. Kingsley Wong: Software, Formal analysis, Data presentation, Writing – review & editing. Mahmoud Torabinejad: Methodology, Validation, Writing – review & editing. George Bogen: Conceptualization, Writing – original draft, Visualization, Writing – review & editing.


Dr. Terauchi is the named inventor of the Terauchi File Retrieval Kit and has patent licensing arrangements and is the recipient of royalties from the promotion and distribution of TRFK products through Guilin Woodpecker Medical Instrument Co., Ltd, Guangxi, China. Other authors deny any conflicts of interest related to this study.

There were errors in the origiinally pubished article. The article has been corrected and is reprinted here. For citation purposes, please use the original publication details: J Endod 2023;49(6):664-674.

View Abstract© 2023 American Association of Endodontists.

Postgraduate Certificate

Year 1 of the MSc in Endodontology. An additional 6 days of study, perfecting everything you have learned from the 10 day programme to OSCE standards.

Postgraduate Diploma

Year 2 of the MSc in Endodontology. 12 days of studying advanced topics such endodontic microscopy, innovations in endodontics, research methods and statistics.

MSc in Endodontology

Year 3 of the MSc in Endodontology and the final component for the Masters Degree. You’re given approximately 12 months to submit a dissertation based on an endodontic topic of your choosing. You’ll be appointed a personal supervisor to who will provide 1:1 coaching.