Clinical Research
Patient and Clinical Characteristics Associated with Primary
Healing of Iatrogenic Perforations after Root Canal
Treatment: Results of a Long-term Italian Study
Fabio G. Gorni, DDS,* Anita Andreano, MD,† Federico Ambrogi, PhD,†
Eugenio Brambilla, DDS,‡ and Massimo Gagliani, MD, DDS*
Abstract
Introduction: There are few data on the long-term ef-
ficacy of mineral trioxide aggregate (MTA) in treating
root canal perforations. We investigated the extent of
A t present, root canal treatment is one of the most commonly performed dental in-
primary healing after perforation repair with MTA and
non-healing as a result of repair. We also investigated
patient/clinical characteristics affecting treatment out-
comes and long-term prognosis. Methods: This was a
prospective cohort study that enrolled consecutive pa-
tients with a single dental perforation treated with
MTA (January 1999–June 2009). Patients were followed
up until December 2012 for a maximum of 13 years after
treatment, with analyses carried out at 8 years. Results:
Of the 110 patients (median age, 36 years; 54.5% male)
eligible for inclusion, 101 were judged to have started to
heal at the first (n = 98, 89%) or second (n = 3, 3%)
annual post-treatment checkup, and 9 (8%, 4 women
and 5 men, aged between 18 and 65 years) did not
show any sign of healing. Patients >50 years had a higher
percentage of non-healing perforations compared with
those #50 years (12% versus 7%). The percentages of
perforations at post-treatment analysis that failed to
heal were 13% (intermediate/middle), 4% (coronal),
and 0% (apical). The percentages of non-healing perfora-
tions according to size were 16% for >3 mm, 6% for 2–
3 mm, and 0% for smaller perforations. Characteristics
associated with probability of progressing after initial
healing were gender, positive probing, size, and site of
perforation. Conclusions: Our results show that having
obtained primary healing with MTA, the likelihood of pro-
gressing is very low. They provide good evidence of the
combined effectiveness of experienced operators and
use of state-of-the-art materials. (J Endod 2016;-:1–5)
Key Words
Iatrogenic, MTA, perforation, root canal treatment,
tooth survival
From the *Department of Biomedical, Surgical and Dental Sciences, ‘‘Giorgio Vogel’’ Dental Clinics; †Department of Clinical Sciences and Community Health, Lab-
oratory of Medical Statistics, Biometry and Epidemiology ‘‘G.A. Maccacaro’’; and ‡Department of Biomedical, Surgical and Dental Sciences, IRCCS Galeazzi Institute,
University of Milan, Milan, Italy.
Address requests for reprints to Prof Massimo Gagliani, Department of Biomedical, Surgical and Dental Sciences, ‘‘Giorgio Vogel’’ Dental Clinics, University of Milan,
Milan, Italy. E-mail address: massimo.gagliani@unimi.it
0099-2399/$ - see front matter
Copyright ª 2016 American Association of Endodontists.
http://dx.doi.org/10.1016/j.joen.2015.11.006
JOE — Volume -, Number -, - 2016
terventions, with 15 million procedures carried out in the United States alone every
year (1). After successful treatment, most teeth can remain as functional units within the
dental arch. However, as the number of root canal interventions has increased exponen-
tially in recent years, so have the nature and number of complications. Root canal de-
viations, fractures of endodontic instruments, and root canal perforations are now
among the most common complications observed in modern dentistry (2). Although
up-to-date reliable data are not available, conservative estimates suggest that perfora-
tions occur in around 20% of endodontically treated teeth (3, 4). Furthermore,
perforations were detected in up to 12% of patients during nonsurgical retreatment
of prior endodontic interventions (5). Root perforations are defined as the communi-
cation between the periodontal apparatus of the tooth and the root canal system.
Although some of them are pathologic, the majority are caused by iatrogenic events.
Regardless of the etiology, a perforation is an invasion into the supporting structures
that causes inflammation and loss of attachment. These in turn compromise the health
of the periradicular tissues and ultimately the prognosis of the tooth. Left untreated, per-
forations result in the loss of integrity of the root and further destruction of the adjacent
periodontal tissues. Root canal perforations can occur before, during, or after an inter-
vention. Clauder and Shin (6) reported that a high percentage (53%) occurs as a result
of prosthodontic treatment, with 47% taking place during routine endodontic treatment.
Effective management of root canal perforations depends on many factors,
including early diagnosis, size, shape, location, and nature of the perforation, chosen
treatment, materials used for the obturation, host response, and importantly, the expe-
rience of the practitioner (7, 8). Fuss and Trope (9) concluded that location is probably
the overriding factor affecting prognosis, with crestal root perforations being the most
susceptible to epithelial migration and rapid pocket formation and thus having the
lowest success rate of repair. Successful treatment depends on accurate diagnosis
and visualization of perforations as well as the use of biocompatible materials effective
in sealing the perforation and preventing bacterial penetration.
Mineral trioxide aggregate (MTA), a calcium silicate–based biocompatible
nonabsorbable material, was developed in the early 1990s. Before then, the choice
of effective, well-tolerated sealant materials was limited. In vitro and in vivo studies
have shown that MTA promotes tissue regeneration without causing inflammation and
has good biocompatibility and nontoxic sealing properties (10–17). Several case
series and retrospective clinical studies have reported healing rates of more than
80% by using MTA (18, 19). However, there are few well-designed prospective studies
Root Perforations: Characteristics Associated with Healing 1
Clinical Research
determining its long-term safety and efficacy in the management of
root canal perforations, and there is an urgent unmet medical need
for evidence-based data.
The primary objectives of this study were to investigate the likeli-
hood of primary healing after perforation repair with MTA and the prob-
ability of progression of the inflammatory process after initial healing.
We also determined patient/clinical characteristics affecting treatment
outcomes and long-term prognosis.
Methods
Study Design
This was a prospective cohort study with enrollment from January
1999 to June 2009 at the Unit of Endodontics, DMCO San Paolo, Depart-
ment of Dentistry, University of Milan, Italy. Follow-up was closed on
December 2012. The study was carried out in accordance with Good
Clinical Practice guidelines and the Declaration of Helsinki 1964 as
currently amended. The protocol was approved by the local ethics com-
mittee, and all patients gave written informed consent to participate in
the study.
Patients
Male and female patients with a single dental perforation were
eligible for enrollment. The recruitment was consecutive starting from
January 1999 until June 2009. A detailed medical and dental history
was obtained from each patient. Exclusion criteria were the following:
1. Age < 18 years
2. Women of child-bearing age not using adequate contraception,
pregnant women, and those lactating
3. Compromised immune status
4. Incomplete pretreatment or intra-treatment records
5. Unwillingness to participate in the study
Diagnosis of Perforations
Perforations were diagnosed by clinical (visualization, periodontal
probing, bleeding spots on paper points) and radiographic examina-
tions. Blood on the side of a paper point was recorded as a strip perfo-
ration. Detailed records of presence/location/intensity of pain and
episodes of swelling/inflammation or abscess were recorded together
with details of prior treatment of the affected tooth. Location and
extent/size of each perforation were determined and recorded on
enrollment and at regular intervals, thereafter using a calibrated peri-
odontal probe. In the few cases where it was not clearly visible, size
was estimated on the basis of the length of the root canal.
Treatment of Perforations
Before sealing the perforation, the area was debrided, cleaned,
disinfected, and dried. If there was bleeding from the site of the perfo-
ration, the area was pre-prepared with an antimicrobial agent such as
calcium hydroxide (Ca(OH)2) powder. MTA (ProRoot MTA; Dentsply
Maillefer, Baillagues, Switzerland; gray and white versions) powder
was mixed with distilled water at a ratio of 3:1 as recommended (15,
20, 21). An MTA gun (Dentsply Maillefer) was used to insert the
material into the perforation, and a smooth humid cotton pellet was
gently pressed over the material to allow setting. After 48–72 hours
when the MTA was completely set, the final filling was performed. All
clinical maneuvers were performed by using magnification, either
loupes (5.5) or surgical microscope (8 or higher) by an expert
operator. After the procedure, patients underwent an x-ray to ensure
that the perforation was adequately sealed.
2 Gorni et al.
Preoperative and Intraoperative Measurements
The following data were recorded:
1. Age and gender
2. Location of the tooth (anterior, premolar, or molar)
3. Perforation site/location:
 Coronal: The upper one third of the root canal;
 Intermediate/middle: Middle one third of the root canal
 Apical: Lower one third of root canal
4. Perforation size (#1 mm, 2–3 mm, >3 mm)
5. Probing: Assessed on a gingival level by using a dichotomous
score:
 Negative probing for probing depth <4 mm
 Positive probing for probing depth $4 mm
 No distinction made whether 1 or more aspects of the tooth ex-
ceeded the 4-mm cutoff
6. Findings of x-rays taken at 3 key time points:
 Preoperative (before perforation repair)
 After repair
 At follow-up (up to a maximum of 13 years)
7. Demographic, clinical, and radiologic data recorded on enrollment
and at annual reviews
Outcome Measures
Perforations were classified as healed when there was/were none
of the following:
1. Clinical signs/symptoms: pain, inflammation, bleeding, no sinus tract
2. Loss of function
3. Periradicular periodontitis
4. Radiolucency near perforation site
5. Evidence of ongoing root resorption
Perforations were judged not to have healed if any of the above was
observed/recorded within the second annual follow-up.
Statistical Analysis
The percentages of teeth healing and not showing signs of healing
within each stratum (eg, male) of the patient (eg, gender) or perfora-
tion characteristics were calculated.
Teeth determined to have healed, according to predefined criteria,
at the second annual follow-up (n = 101) were stratified for gender, age
(#50 years versus >50 years), location, probing result, and site and
size of the perforation. Patients were followed up to assess progressing
inflammation after initial healing, and discrete hazard and survival func-
tions were calculated to describe progression over time (22). Person-
time (combining the number of persons and their time contribution)
was computed from primary healing (either first or second annual
follow-up date). Follow-up was curtailed at 8 years because only
24% of patients had a longer follow-up. Median length of follow-up
was determined by using the reverse Kaplan-Meier method (23). A
discrete time hazard model that uses the complementary log-log link
was used to examine the role of patient and clinical characteristics asso-
ciated with progressing during follow-up (22). Exploratory analysis by
using life tables was carried out initially and then followed by univariate
analysis for all predictors: age (#50 and >50 years), tooth location
(anterior + premolar versus molar), perforation location (coronal
versus apical + intermediate), size (#3 mm versus >3 mm), and prob-
ing. Categorization was done on the basis of previous studies and clin-
ical considerations (4).
In addition, exploratory graphical analyses were performed to
verify the proportional hazard assumption together with the presence
JOE — Volume -, Number -, - 2016

of time-dependent effects. The functional form of baseline risk over time
was also investigated. The final multivariable model was adjusted for sex
and age at treatment and included all available predictors with the
exception of tooth location. Results of the model are presented as haz-
ard ratios (HRs) of progressing after initial healing and their 95% con-
fidence intervals (CIs). In addition, because the number of events was
small relative to the number of predictors, the stability of the final model
was verified by using nonparametric bootstrap (50,000 samples) (24).
For each parameter of the model, we computed a bootstrap 95% CI by
using the 2.5 and 97.5 percentiles of the bootstrap distribution. The HR
was considered significant if the bootstrap 95% CI did not include 1.
On the basis of the final discrete time hazard model, adjusted
probability curves for not progressing were fitted for perforation loca-
tion, size, and probing by using the corrected group prognosis
method; survival curves were first calculated for each patient by using
the level of the covariate of interest (eg, coronal site) and the patient’s
values for the other covariates and then averaged (25, 26). In
addition, 95% CIs were estimated by applying bootstrap resampling
(10,000 samples) and the percentile method (24). Calculations
were performed by using the open-source R statistical computing
environment (27).
Results
Primary Healing
One hundred ten consecutive patients (median age, 36 years;
54.5% male) with a single root perforation after endodontic treatment
were eligible and included in the study. A total of 101 were judged to
have healed at the first (n = 98, 89%) or second (n = 3, 3%) annual
post-treatment checkup, whereas 9 (4 women and 5 men aged between
18 and 65 years, 8%) did not show any signs of healing during the study
period (Table 1). Although the maximum potential follow-up was
13 years, analyses were carried out at 8 years because there were too
few patients with a longer follow-up. There was no loss to follow-up until
the second year.
Clinical Research
The proportion of male and female was the same in non-healed
perforations (8%). Similarly, there was a higher percentage of patients
$50 years with perforations classified as non-healed compared with
those #50 years (12% versus 7%, P = .61) (Table 1). The percentage
of non-healed perforations among those with a positive probing result
was higher than for the negative ones (18% versus 2%, P = .003). Over-
all, the site of perforation was not significantly associated with primary
healing (P = .29); the percentage of non-healed perforations was 13%
intermediate/middle, 4% coronal, and 0% apical. Likewise, there was
no significant association between perforation size and primary healing
(P = .16). The percentage of non-healed teeth was 16% for >3 mm, 6%
for 2 3 mm, and 0% for smaller perforations. Perforations in all ante-
rior teeth healed, whereas perforations in 15% of premolars and 9% of
molars did not (P = .35).
Predictive Factors for Risk of Recurrent Problems
as a Result of Primary Healing with MTA
Figure 1 shows the probability (95% CI) of not progressing as a
result of primary healing with MTA in the 101 teeth healing by the sec-
ond follow-up visit during the 8-year follow-up period (hazard, 3%–
6%, for all years). The median length of follow-up was 4 years (95%
CI, 4–5 years). Overall, the inflammatory process progressed in 18
teeth during the 8-year period, with an estimated probability of pro-
gressing at 5 and 8 years of 18% (95% CI, 9%–27%) and 33% (95%
CI, 16%–47%), respectively.
A number of factors/variables were associated with a greater likeli-
hood of progressing as a result of primary healing with MTA. Women had
a higher risk compared with men (HR, 3.1; P = .03; Table 2), and pa-
tients with a positive probing had a higher risk compared with patients
with a negative probing (HR, 21; P # .001). Patients with an apical or
intermediate perforation were more likely to progress compared with
those with coronal perforations (HR, 3.0; P = .04). The size of the perfo-
ration was also related to the risk of progression; patients with perfora-
tions >3 mm were more likely to progress compared with those with
smaller lesions (HR, 3.6; P = .01). Age was not significantly associated
with a higher risk of progression (P = .16).

TABLE 1. Characteristics Relative to Healing of the Entire Cohort of 110 Teeth

Treated for Perforations from 1999 to 2009 at Endodontics Unit, DMCO San
Paolo, Milan, Italy
Healing, Non-healing, Total, Fisher test
N (%) N (%) N (%) P value
Age (y) .61
#50 87 (93) 7 (7) 94 (85)
>50 14 (88) 2 (12) 16 (15)
Gender 1
Male 55 (92) 5 (8) 60 (54)
Female 46 (92) 4 (8) 50 (46)
Tooth location .29
Anteriors 15 (100) 0 (0) 15 (14)
Premolars 11 (85) 2 (15) 13 (12)
Molars 75 (91) 7 (9) 82 (74)
Probing .003
Negative 65 (98) 1 (2) 66 (60)
Positive 36 (82) 8 (18) 44 (40)
Perforation site .29
Coronal 50 (96) 2 (4) 52 (47)
Intermediate 48 (87) 7 (13) 55 (50)
Apical 3 (100) 0 (0) 3 (3)
Perforation .16
size (mm)
#1 10 (100) 0 (0) 10 (9)
2–3 65 (94) 4 (6) 69 (63)
>3 26 (84) 5 (16) 31 (28) Figure 1. Probability (95% CI) of not progressing as a result of primary heal-
Total 101 (92) 9 (8) 110
ing in 101 teeth.

JOE — Volume -, Number -, - 2016 Root Perforations: Characteristics Associated with Healing 3
Clinical Research
TABLE 2. Discrete Time Survival Analysis of Progression of Inflammation in 101 Teeth after Primary Healing (follow-up was limited to 8 years after treatment)
HR (95% CI)
Bootstrap
Predictors of recurrence N (%) recurring Univariate Multiple P value HR (95% CI)
Age (y)
#50 15/87 (17)
>50 3/14 (21) 1.8 (0.4–5.8) 2.6 (0.6–9.0) .16 NS 1.8 (0.0–8.1)
Gender
Male 6/55 (11)
Female 12/46 (26) 2.8 (1.1–8.0) 3.1 (1.2–9.2) .03 2.9 (1.2–10.9)
Tooth location Not included
Anterior and premolar 2/26 (8)
Molar 16/75 (21) 3.4 (0.9–21.6)
Probing
Negative 2/65 (3)
Positive 16/36 (44) 18.8 (5.3–118.9) 21.2 (5.6–141) <.001 17.1 (4.6–4.7  109)
Perforation site
Coronal 5/50 (10)
Intermediate and apical 13/51 (25) 2.7 (1.0–8.5) 3.0 (1.1–9.8) .04 2.6 (0.9–11.1)
Perforation size (mm)
#3 9/75 (12)
>3 9/26 (35) 3.7 (1.5–9.6) 3.6 (1.3–10.1) .01 2.7 (1.0–8.2)

The stability of the HRs estimated from the final multivariable
model was verified by using nonparametric bootstrap sampling. The
95% CIs of the HRs obtained with bootstrap were always in accordance
with those from the model (ie, when the predictor had a significant P
value in the model, both the model and the bootstrap 95% CI did not
include 1). The only exception was the site of the perforation where
bootstrap 95% CI did include 1 (HR, 2.6; 95% CI, 0.9–11.1), whereas
the model 95% CI did not ( HR, 3.0; 95% CI, 1.1–9.8). The probabil-
ities (95% CI) of not progressing by using our final model in patients
with negative probing at 2 and 5 years, adjusted for the other covari-
ates in the model by using the corrected group prognosis method,
were 99% (99.0%–99.4%) and 97% (95.8%–97.5%), whereas
only 87% (84.2%–89.6%) and 61% (56.1%–66.4%) of patients
with positive probing were problem-free at the same time points
(Fig. 2, left panel) (25, 26). Patients with apical and intermediate
perforations had adjusted probabilities of not progressing at 2 and
5 years of 90% (87.1%–93.6%) and 76% (69.8%–82.0%),
whereas corresponding figures in those with coronal perforations
were 96% (94.6%–97.6%) and 87% (83.0%–91.2%) (Fig. 2, central
panel). Probabilities of not progressing at 2 and 5 years for patients
with a large perforation (>3 mm) were 88% (85.0%–91.7%) and
71% (64.3%–77.6%), compared with 96% (94.8%–97.3%) and
86% (82.1%–90.0%) for patients with a small perforation (Fig. 2,
right panel).
Discussion
Root canal perforations of whatever origin may have serious impli-
cations, and timely and appropriate treatment is imperative to ensure
long-term survival of the affected tooth (3). Effective management of per-
forations represents a challenge even for experienced endodontists. In a
retrospective study investigating healing rates of 70 perforation repairs
performed by 6 endodontic specialists, Pontius et al reported a success
rate of 90% (13, 18, 19). Our study is one of the largest long-term pro-
spective studies to investigate healing of perforations, and as such, our
primary healing rates of 92% (101 of 110) are particularly encouraging.
The results of the present study show that after treatment with MTA and
having obtained primary healing, the risk of progression of the inflam-
matory process was very low, with a 5-year risk of progression of 18%.
The high healing rates obtained in our study were probably due to a
combination of the efficacy of MTA and the expertise of the operators.
The results of our study provide further insights into the prog-
nostic factors influencing healing. Fuss and Trope (9) reported

Figure 2. Eight-year adjusted probability of not progressing (corrected group prognosis method) according to probing levels, perforation site, and perfo-
ration size. Adjusted survival curves from proportional hazards models calculated according to Ghali et al (25).

4 Gorni et al. JOE — Volume -, Number -, - 2016


location and size of perforations as the most important prognostic fac-
tors. Mente et al (19) described how the healing rate for teeth with a
single-root perforation (92%) was higher than that for multirooted
teeth (75%) but was lower in teeth with larger perforations
(>3 mm, 67%) compared with those with smaller (2–3 mm, 88%)
and very small perforations (#1 mm, 90%). None of these differences
were statistically significant, which the authors attribute in part to the
small number of enrolled patients. Results of a larger study with a
longer follow-up by the same group showed differences in healing
rates $15%, depending on location and size of perforation, root-
filling technique, treatment sessions, type of restoration, and place-
ment of a post after treatment (18). As in the previous study, healing
rates were lower (71%) in patients with larger perforations (>3 mm)
than for those with smaller perforations (2–3 mm, 88%) and very
small perforations (#1 mm, 88%). None of these variables were
significantly different on univariate analyses. The authors suggested
that prognostic factors, historically considered to be important in
the pre-MTA era, might no longer be applicable in the post-MTA era
because it provides excellent sealing ability regardless of location
and/or size. The small number of patients in the first study by Mente
et al (19) compared with ours may in part explain the differing results.
Interestingly, in our study women had a higher risk of problems recur-
ring than men, a factor not evidenced by other studies.
Calcium hydroxide powder is widel y used to control bleeding in
pulp-capping procedures because of its excellent antibacterial proper-
ties. Increased bleeding can be indicative of inflammation in the tissues
surrounding the perforation, but in the majority of cases there is only light
bleeding (28). In our patients MTA was applied directly to the prepared
perforation, and in only a few cases was it necessary to first use calcium
hydroxide powder to stop bleeding as described by Yildirim et al (29).
Our results provide further confirmation of the effectiveness of us-
ing MTA as a sealing material in root perforations and in particular its
excellent tolerability on long-term follow-up. They also provide inter-
esting new data on prognostic factors involved in healing. Although
the experience of the operator and the use of magnification were not
taken into account, they merit a mention. Our practice has used magni-
fication for many years, and all clinical operations reported here were
performed by experienced personnel who used magnification, either
loupes (5.5) or surgical microscope (8 or higher). Technology
has revolutionized dentistry in recent years, and magnification is now
considered vital for all dental practitioners regardless of age or visual
acuity. We have seen that successful treatment relies on accurate diag-
nosis and visualization of perforations.
Conclusions
The ultimate goal in the management of root canal perforations is
to effectively seal the area as quickly as possible with an effective
biocompatible agent to prevent inflammation and loss of tissue attach-
ment by preserving healthy tissues at the perforation site. Although this
study was not conducted to compare sealing agents, the high long-term
success rates achieved with MTA confirm that it is an effective and well-
tolerated agent in repairing perforations. In this age of evidence-based,
personalized medicine the results of our study provide valuable data
on how to effectively manage root canal perforations in ‘‘real world’’
clinical practice. Most of previously reported data were based on
case series or small-scale short-term trials and as such offer limited
evidence-based data. Our results provide good evidence of the com-
bined effectiveness of experienced operators and the use of state-of-
the art materials. They also provide information on the factors affecting
prognosis in the post-MTA era, but we are still not in a position to be
able to accurately predict outcome from initial patient and clinical
JOE — Volume -, Number -, - 2016
Clinical Research
characteristics. Further prospective clinical studies on use of MTA to
repair root perforations in routine clinical practice are required.
Acknowledgments
The authors deny any conflicts of interest related to this study.
References
1. Endodontic Facts, American Association of Endodontists. Available at: http://www.
aae.org/about-aae/news-room/endodontic-facts.aspx. Accessed June 1, 2015.
2. Borges
AH, Bandeca MC, Tonetto MR, et al. Portland cement use in dental root per-
forations: a long term followup. Case Rep Dent 2014;2014:637693.
3. Tsesis I, Fuss ZV. Diagnosis and treatment of accidental root perforations. Endodon-
tic Topics 2006;13:95–107.
4. Tsesis I, Rosenberg E, Faivishevsky V, et al. Prevalence and associated periodontal
status of teeth with root perforation: a retrospective study of 2,002 patients’ medical
records. J Endod 2010;36:797–800.
5. Gorni FG, Gagliani MM. The outcome of endodontic retreatment: a 2-yr follow-up.
J Endod 2004;30:1–4.
6. Clauder T, Shin S-J. Repair of perforations with MTA: clinical applications and
mechanisms of action. Endodontic Topics 2006;15:32–55.
7. Kvinnsland I, Oswald RJ, Halse A, Gronningsaeter AG. A clinical and roentgenolog-
ical study of 55 cases of root perforation. Int Endod J 1989;22:75–84.
8. Kaushik A, Talwar S, Yadav S, et al. Management of iatrogenic root perforation with
pulp canal obliteration. Saudi Endod J 2014;4:141–4.
9. Fuss ZV, Trope M. Root perforations: classification and treatment choices based on
prognostic factors. Dent Traumatol 1996;12:255–64.
10. Asgary S, Motazedian HR, Parirokh M, et al. Twenty years of research on mineral
trioxide aggregate: a scientometric report. Iran Endod J 2013;8:1–5.
11. Krupp C, Bargholz C, Br€usehaber M, H€ulsmann M. Treatment outcome after repair
of root perforations with mineral trioxide aggregate: a retrospective evaluation of 90
teeth. J Endod 2013;39:1364–8.
12. Pace R, Giuliani V, Pagavino G. Mineral trioxide aggregate as repair material for fur-
cal perforation: case series. J Endod 2008;34:1130–3.
13. Pontius V, Pontius O, Braun A, et al. Retrospective evaluation of perforation repairs
in 6 private practices. J Endod 2013;39:1346–58.
14. Main C, Mirzayan N, Shabahang S, Torabinejad M. Repair of root perforations using
mineral trioxide aggregate: a long-term study. J Endod 2004;30:80–3.
15. Parirokh M, Torabinejad M. Mineral trioxide aggregate: a comprehensive literature
review—part I: chemical, physical, and antibacterial properties. J Endod 2010;36:
16–27.
16. Torabinejad M, Parirokh M. Mineral trioxide aggregate: a comprehensive literature re-
view—part II: leakage and biocompatibility investigations. J Endod 2010;36:190–202.
17. Parirokh M, Torabinejad M. Mineral trioxide aggregate: a comprehensive literature
review—part III: clinical applications, drawbacks, and mechanism of action.
J Endod 2010;36:400–3.
18. Mente J, Leo M, Panagidis D, et al. Treatment outcome of mineral trioxide aggregate:
repair of root perforations—long-term results. J Endod 2014;40:790–6.
19. Mente J, Hage N, Pfefferle T, et al. Treatment outcome of mineral trioxide aggregate:
repair of root perforations. J Endod 2010;36:208–13.
20. Basturk FB, Nekoofar MH, G€unday M, Dummer PM. The effect of various mixing and
placement techniques on the compressive strength of mineral trioxide aggregate.
J Endod 2013;39:111–4.
21. Roberts HW, Toth JM, Berzins DW, Charlton DG. Mineral trioxide aggregate material
use in endodontic treatment: a review of the literature. Dent Mater 2008;24:149–64.
22. Singer JD, Willett JB. Applied Longitudinal Data Analysis: Modeling Change and
Event Occurrence. Oxford, UK: Oxford University Press; 2003.
23. Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time.
Control Clin Trials 1996;17:343–6.
24. Efron B. Bootstrap methods: another look at the jackknife. Ann Statist 1979;7:1–26.
25. Ghali WA, Quan H, Brant R, et al. Comparison of 2 methods for calculating adjusted
survival curves from proportional hazards models. JAMA 2001;286:1494–7.
26. Nieto FJ, Coresh J. Adjusting survival curves for confounders: a review and a new
method. Am J Epidemiol 1996;143:1059–68.
27. R Development Core Team. R (2013): a Language and Environment for Statisti-
cal Computing—R Foundation for Statistical Computing. Vienna, Austria: R
Foundation for Statistical Computing; 2013. Available at: http://www.R-project.
org. Accessed June 2015.
28. Li Z, Cao L, Fan M, Xu Q. Direct pulp capping with calcium hydroxide or mineral
trioxide aggregate: a meta-analysis. J Endod 2015;41:1412–7.
29. Yildirim G, Dalci K. Treatment of lateral root perforation with mineral trioxide
aggregate: a case report. Oral Surg Oral Med Oral Pathol Oral Radiol Endod
2006;102:55–8.
Root Perforations: Characteristics Associated with Healing 5