- Email: [email protected]
Retrograde sealing of accidental root perforations with dentin-bonded composite resin
0099-2399/98/2410-0671 $03.00/0 Printed in U.S.A.
JOURNALOF ENDODONTICS Copyright © 1998 by The American Association of Endodontists
VOL. 24, NO. 10, OCTOBER1998
Retrograde Sealing of Accidental Root Perforations with Dentin-Bonded Composite Resin Jorgen Rud, Dr.odont., Vibe Rud, DDS, and Erik C. Munksgaard, Dr.odont.
Surgical treatment was performed on 100 iatrogen perforations, of which 94 showed radiolucency of bone adjacent to the perforation and 83 presence of an exposed post. The perforations were in all cases sealed with a resin composite (Retroplast) bonded to adjacent root dentin with a dentin bonding agent (Gluma). Cases were examined after - 1 yr (first recall) and, if necessary after 11/= to 11 yr (mean: 4.1 yr) (latest recall). It was observed that the presence or absence of bone on the root between the perforation and cervix at the time of operation had no significant effect on the healing result, and that the radiographic classification "partial" healing with a border of cortical bone or a lamina dura often remained unchanged for many years. At latest recall, the healing result of 65 teeth originally having root perforation elsewhere than at the furcation was: 71% complete, 11% partial, 3% uncertain, and 15% failure. This is significantly different from the healing result of 27 molar teeth originally having perforation at the furcation: 30% complete, 4 1 % partial, 11% uncertain, and 18% failure.
after treatment, respectively. Cementum and a periodontal ligament with Sharpey's fibers in close contact to the Retroplast seal was identified, indicating a good tolerance by the tissue (6, 10). In clinical studies, it was shown that infected root canals with periapical infections treated with Gluma-bonded Retroplast demonstrated a healing frequency of - 9 0 % (11), compared with 72% when amalgam was used as a retrofill material (14), in both cases estimated some years after surgery. Cases showing complete healing 1 yr after surgery with Retroplast remained unchanged when re-examined 8 to 9 yr later, indicating stability of both the retrofill material and the bonding (12). In all of the aforementioned studies, Retroplast was applied not as a filling, but only as a seal at a slightly saucerized root end. This geometry prevents contraction gap formation in contrast to a regular box-like preparation such as that used for amalgam in retrofills (15). The purpose of this study was to examine the healing results of surgical Gluma Retroplast seals of iatrogene root perforations, which were not suited for immediate orthograde treatment.
M A T E R I A L S AND M E T H O D S One hundred thirty-five consecutive patients with iatrogenic root perforation, referred to and treated by the authors (V.R. and J.R.) in their private practice from November 1984 to March 1995, form the material in this prospective examination. Among these, the following perforations were not included in the study: 10 cases with root fractures, 5 cases with total removal of the root, and 1 case with reimplantation. Of seven reoperations, only failure in the primary operation were included. Combined orthograde root filling and retrograde seals of perforations were not included in the material. All patients had pre- and postoperative radiographs taken and were recalled - 1 yr after the operation and eventually later. A total of 100 patients responded; 84% of the 119 potential cases. There were 64% females and 36% males with medium age of 50 yr (range: 19-85 yr). None of the perforations in the series were referred for treatment immediately after the accident. Access to the root perforation was obtained after elevating a facial or palatal gingival flap and by removal of covering bone. The root dentin around the perforation was slightly hollowed and treated with Gluma, and the area was sealed with Retroplast using the method for apical sealing as described previously (6-8, 13). Retroplast is a specially designed
Accidental root perforations are normally treated with orthograde gutta-percha, internal placed amalgam, Cavit, or with calcium hydroxide immediately after the accident (1-5). However, cases with large perforations and with overfill do not respond well to this treatment, and perforations caused by posts that cannot be removed may be impossible to treat orthograde. In addition, orthograde treatment of perforations with radiolucency of the surrounding bone reaching the gingival pocket has been considered to have a poor prognosis (2). In such cases, surgical treatment with retrograde closure might be indicated. The sealing effect of dentin-bonded resin composites applied as apical retrograde root filling has recently been evaluated by laboratory experiments and clinically (6-13). Ambus and Munksgaard (9) found that an apical cover with a resin composite (Retroplast) may be gap-free when bonded with Gluma or other dentin-bonding agents. In other studies, tissue from monkeys and humans adjacent to apical Retroplast seals was examined histologically 1 and 3 yr
671
672
Journal of Endodontics
Rud et al.
resin composite with sufficient radiopacity, thixotropic effect, and 2 min working time colored to enable precise application (! 1). In the anterior and premolar regions, access to lingual or palatal perforations was often done from the buccal side by removing interradicular bone and sometimes root substance. Missing root substance was replaced by Retroplast. Access to intraradicular perforations of maxillary molars was obtained after substantial resection or, in some cases, by removal of one of the buccal roots. When treating cases with perforating posts, the post-end was resected until slightly below the root surface and the surrounding dentin was slightly hollowed. Loose posts were cemented before sealing with Retroplast. The location of the perforation and occurrence of bone loss on the root was noted. The healing was classified on the basis of radiographs and clinical symptoms into one of the following categories, which were based on criteria proposed by Rud et al. (16) and adapted to root perforations: 1. Complete healing--(a) Radiographic bone regeneration next to proximal perforations with or without a recognizable periodontal ligament space up to twice as large as normal (Fig. 1); radiographs eventually taken in different projections. (b) Disappearance of formerly recognized preoperative radiolucency adjacent to buccal or palatal/lingual perforations, with unrecognizable periodontal ligament space (Fig. 2). 2. Partial healing--Bone regeneration up to the level of the seal at which location a border of compact bone or lamina dura may have developed (Fig. 3). 3. Uncertain healing--Cases where healing cannot yet be determined or cases undergoing phases of healing; a decreased radiolucency compared with the postoperative or a previous follow-up radiograph, but without a compact bone border or lamina dura. 4. Failure--Unchanged or enlarged rarefaction originating from the perforation; loose sealing material seen on radiograph; presence of clinical signs of inflammation originating from the perforation area such as a fistula or an abscess.
Evaluation The authors (J.R. and V.R.) evaluated the radiographs separately and reached a common verdict in the 18 cases (18%) where primary differences occurred. The results were evaluated by X2-test (M × N, two-tailed or a Fischer's exact test) (17).
FIG 1. Perforation of a post mesiocervical in central incisor. Fortysix-year-old female. (A) Postoperatively. The root has been ground to just below the root surface and covered with Retroplast. A minute bridge of cervical bone was preserved. (B) Nine years and 2 months after operation. Complete bone healing of radiolucency and the cervical area. Reformation of periodontal ligament space and lamina dura is seen.
RESULTS The distribution of the 100 teeth in the present study according to type is shown in Fig. 4. Maxillary teeth were represented in 60% and mandibular teeth in 40% of the treated teeth. Perforations in lower incisors and canines had low representation. Besides 31 perforations at molar furcation areas, perforations were located mesially (52%), buccally (21%), and distally (23%). Lingual/palatal perforations were rare (4%). Midroot perforations were more common, compared with cervical and apical perforations. Premolar furcation perforations occurred in 6% of the cases. Among the 100 cases, 83 had posts in the perforated roots. Most of these posts were exposed in the perforation while perforated through the root. Other filling materials were extruded through the perforation in 13% of the cases. A preoperative-radiolucency of the
adjacent bone was seen in 94% of the cases. Of the six cases without radiolucency, three had a fistula and one an abscess. Of the 100 patients examined after the operation, 89 were seen after 6 to 18 months (mean: 1.0 yr) (first recall). Eleven cases not examined at first recall were seen later. Fifty-four patients were seen between 18 months and 11 yr (mean: 4.1 yr) (latest recall). Eight cases not examined at latest recall were partial or uncertain healings at first recall. The interrelations between healing results of the first and latest recall were as follow: eight cases with complete healing at both the first and latest recall were found. This agrees with previous findings using Retroplast that complete apical healing is unchanged and stable when examined later, unless interference of new factors
Retrograde Sealing of Root Perforations
Vol. 24, No. 10, October 1998
673
TABLE 2. Origin of 92 cases in Table 1 showing various healing types at latest recall Healing at Latest Recall
From healing II first recall From healing III first recall Cases not examined at first recall Found at both first and latest recall Healings transferred from first recall* Sum of latest recall
I
II
III
IV
2 4 6 8 34 54
-6 3 9 -18
---5 -5
-6 2 3 4 15
I, complete healing; II, partial healing; III, uncertain healing; IV, failures. * As explained in the text, first recall healings I and IV remain unchanged by time.
TABLE 3. Healing frequencies at first and latest recall for patients treated for root perforation at molar furcation Healing Category Complete Partial Uncertain Failure Subtotal
First Recall (Mean: 1 yr)
Latest Recall (Mean: 5.4 yr)
No.
%
No.
%
4 8 11 3
15 31 42 12
8 11 3 5
30 41 11 18
26
100
27
100
Not examined at first recall Not examined at latest recall FIG 2. Buccal perforation of an upper first premolar. Forty-eightyear-old female. (A) Postoperatively. A radiolucency is seen on each side of the root. (B) Complete bone healing at 7 months follow-up.
with no connection to the method (e.g. root fracture) occur (12, 13). This means that cases with complete healing at first recall may be transferred to the latest recall group without further examination. This is of cause true also for failures that, almost by definition, will not change with time. Efforts were therefore concentrated on recalling patients with partial or uncertain healing at first recall, which might change healing group at later recall. The healing results in the total material are shown in Table 1. To show the TABLE 1. Healing frequencies at first and latest recall for patients treated for root perforation First Recall Healing Category Complete Partial Uncertain Failure Subtotal Not examined Total
No.
%
Latest Recall No.
%
42 14 26 7
47 16 29 8
54 18 5 15
59 20 5 16
89
100
92
100
11
8
100
100
The first recall occurred between 0.5 and 1.5 yr (mean: 1.0 yr). The numbers showing complete healings and failures at first recall were transferred to latest healing. Apart from the transferred 1-yr results, the latest recall took place between f.5 and 11 yr (mean: 4.1 yr).
Total
31
31
relation between the healing groups, an account was made of the origin of the latest recall, partly based on the results from the first recall (Table 2). Healing frequencies for 31 perforations in molar furcations, observed after - 1 and 5.4 yr, are shown in Table 3. Twenty-five of the perforations in furcations were in mandibular molars and six in maxillary (Fig. 4). Three of the upper perforations were in palatal roots (Fig. 5). Of the 25 mandibular perforations in furcation, 18 (72%) were in the distal root and 2 were between the roots. Twenty-two (88%) of the perforated mandibular roots contained posts. Examples of furcation healings are the following: complete healing (Fig. 6), partial healing with lamina dura (Fig. 7), and partial healing with compact bone (Figs. 3 and 5). Five of the 31 furcations were not examined at first recall, but only at latest recall. Four complete healings and three failures were transferred from first to latest recall. The results in Table 4 are the healing frequencies of cases with perforation elsewhere on the root than at the furcation. They are obtained by deducting the numbers in Table 3 (furcations) from those in Table 1 (total material). Absence of bone on the root extending from the cervix and around the perforation was recorded in 83 cases during the operation. This deficiency derived from the perforation and not as a result of marginal periodontal disease. Table 5 shows healing categories in cases with and without such bone on the root.
674
Rud et al.
Journal of Endodontics
FIG 3. Perforations of posts in furcation area and mesial and distal roots. Twenty-nine-year-old female. (A) Postoperatively. Retroplast covers the distal post and furcation. (B) Almost the whole length of root is perforated by the post. Just covering the post with Retroplast, it need not be resected. (C) After 11/2yr cortical bone regeneration is seen up to the level of Retroplast seal. Tooth was extracted soon after because of a root fracture. See crack in distal and mesial root. TABLE 4. Healing frequencies at first and latest recall for patients treated for root perforation elsewhere on the root than at the furcation
NUMBER OF TEETH P~ 2 2011~ DIO
P2
I--~
Healing Category
M1 M2
__¢
_Q ii -010 0
~ 20
Cases with perforation at the furcation
FIG4. Location of 100 perforations according to tooth type and arch.
A condition for re-establishment of a periodontal ligament and lamina dura adjacent to a root perforation is that a tight and permanent seal can prevent bacteria and their toxins in the root canal from entering the surrounding periodontal tissues (18, 19). It seems in the majority of cases that the cover of resin composite (Gluma and Retroplast) provided such a seal (Table 1). After sealing of root perforations, bone may regenerate until it reaches the root surface and forms part of a periodontal ligament space with a lamina dura (Figs. 1, 6, and 8). In some cases the bone may, however, just regenerate until it reaches the border of the Retroplast seal where further growth stops. The bone surface may then be bordered with compact bone or a lamina dura (Figs. 3, 5, 7, and 9). In the present study, this has been called partial healing. This seems to be different from incomplete apical healing caused by scar tissue (16). The bone regenerated in partial healings may be compared with the compact bone border, which may be seen after a successful periodontal treatment. Apical scar tissue healing
Latest Recall
%
No.
%
Complete Partial Uncertain Failure
38 6 15 4
60 10 24 6
46 7 2 10
71 11 3 15
Total
63
100
65
100
The number stated in the various heaiing categories are those from Table 1 (total material) minus those from Table 3 (furcations).
TABLE 5. Presence or absence ( + / - ) of bone on root between perforation and cervix, among 83 cases with various healing types (Table 1) at latest recall
Healing Category DISCUSSION
First Recall No.
+ Bone
- Bone
Total
Complete Partial Uncertain Failure
25 (62%) 7 (18%) 1 (2%) 7 (18%)
20 (47%) 11 (26%) 5 (12%) 7 (16%)
45 18 6 14
Total
40 (100%)
43 (101%)
83
In 9 cases (8 complete healings and 1 failure), the + / - bone was not registered.
will eventually become complete healing (16, 20). Complete healing with reformation of a periodontal membrane space can, however, not always be expected to occur adjacent to sealed root perforations, even if the seal is tight and the sealing material biocompatible. A communication may exist from the bone defect around the perforation to the gingival pocket. This may lead to infection adjacent to the Retroplast seal. Adequate gingival hygiene, eventually with gingivectomy, may, however, result in partial healing and in some cases complete healing. The cervicular
Voh 24, No. 10, October 1998
Retrograde Sealing of Root Perforations
675
FIG6. Perforation of post in distal root to furcation of mandibular first molar. Forty-seven-year-old female. (,,4)Preoperatively. (/3) One year after sealing perforation with Gluma and Retroplast. Complete bone healing. Bone preserved buccal to distal root.
FIG 5. Divergence of root filling and post in palatal root of first maxillar molar indicate perforation. Sixty-one-year-old female. (A) Preoperatively. (B) Postoperatively. Distobuccal root removed to get access to furcation and palatal root. Post cut below root surface and covered with Gluma and Retroplast. (C) After 11 yr partial healing with bone defect from a distobuccal gingival pocket.
bone loss may, however, be so extensive, that bone healing will occur only to a certain distance on the perforated root (Fig. 9). The presence or absence of bone on the root between the perforation and the cervix was noticed during the operations. Absence of bone, especially in this area, could be expected to influence bone healing around and adjacent to the Retroplast seal. The numbers in Table 5 might indicate that complete healing occurs more often with bone on the root, whereas partial and uncertain healings are found more often in cases without bone. The
apparent differences were, however, not statistical significant (X2 = 4.01, p = 0.26). The findings showed that, even with no bone on the root, a limited bone healing may take place. This has also been found in a follow-up study of apicoectomized teeth with total buccal bone loss, showing 37% successful apical healings (21). However, if just a small bridge of crestal bone exists, it should by all means be preserved. In cases of treated molar furcation perforations (Table 3), complete healings increased from 15 % to 30% from first to latest recall. At latest recall, the proportion of cases showing complete healing was significant less (p = 0.0006) than that observed with restored perforations elsewhere on the root than at the furcation (Table 4), and the overall healing results of these two groups differ significantly (X2 = 16.6, p = 0.0009). These numbers emphasize the general observation that treatment of furcation perforations have a poorer prognosis than nonfurcation cases (1-5). The healing frequencies at first recall of restored perforations elsewhere on the root than at molar furcations (Table 4) were significantly different 0( z = 13.3, p = 0.004) than at latest recall. In addition, complete healing in 60% and 71% of the cases, respectively, is less than that seen for apical healing (11, 13). The proportion of uncertain healings decreased from 24% at first recall to 3% at latest recall, a decrease that is significant (p = 0.0009). This shows that uncertain healings in time change to one of the other healing groups.
676
Rud et al.
Journal of Endodontics
FtG 8. Perforation of the root caused by a too thick post in a lower canine. Forty-three-year-old female. (A) Preoperatively. (B) Complete healing. Reformation of periodontal ligament at 5-yr follow-up.
FIG 7. Perforation of post and cement to furcation in mandibular first molar. Seventy-eight-year-old female. (A) Preoperatively. (B) Postoperatively after Retroplast seal. (C) One-year follow-up. Partial healing. Bone demarcated with lamina dura.
The compact bone border seen in partial healing seems to be stable. Further bone deposits may occur, but slowly. It may take several years before a complete bone healing has been established. On the other hand, no cases of partial healing became uncertain healing or failure at the latest recall (Table 2). That partial healings are almost unchanged is also evident by the findings: 10% at the first recall and 11% at the latest recall in Table 4. Partial healing is the category most often found in furcation cases (Table 3). The proportion of partial healings of the total amount of restored furcations (Table 3), compared with that of the restored nonfurca-
tions at latest recall (Table 4), was statistically significantly different (p = 0.004). Partial healings in furcations may also be stable; in the present series, some were observed up to 11 yr. An analysis of the 15failures presented in Table 1 showed that they were mainly caused by absence of a tight Retroplast seal, as was the case for failures with apical retrograde root seals (13). This explains the development of cases with acute infection, as well as increased radiolucency. A limited number of reoperations was performed, with acceptable healing in approximately half the cases. A comparison of the present results with those of other authors is difficult. As Gutmann and Harrison (1) point out, "surgical repair of root perforations has received sporadic attention in the dental literature and has been supported primarily by case reports or limited studies." Therefore, statistically based analyses of differences are not possible. Despite the results in the present study, nonsurgical methods will and should probably be continued as a first choice. This may not include cases with large perforations, especially with overfill;
Vol. 24, No. 10, October 1998
Retrograde Sealing of Root Perforations
677
osseous defect associated with the perforation. Conventional treatment of the perforation should therefore be conducted shortly after the accident and the patients recalled not later than 1-yr later to assess the outcome. If the result is unsuccessful, surgery with Gluma-Retroplast may be undertaken as early as possible. This study was supported in part by the Danish Dental Association, Research Foundation. We thank Dr. Jens O. Andreasen, Dental Clinic, University Hospital (Copenhagen, Denmark) for his kind review of the manuscript. Drs. J. Rud and V. Bud have a private practice in oral surgery in Copenhagen, Denmark. Dr. E. C. Munksgaard is assistant professor, Department of Dental Materials, School of Dentistry, Faculty of Health Sciences, University of Copenhagen. Address requests for reprints to Dr. Jorgen Rud, Nikolaj Plads 7, DK-1067 Copenhagen, Denmark.
References
FIG 9. Distal perforation o f maxillar lateral incisor with d i s l o d g e m e n t of filling material in 19-yr-old female. (A) Preoperatively. (B) P o s t o p eratively. Perforation sealed with Retroplast. No cervical bone. (C) Follow-up after 6 yr. Partial b o n e regeneration b o r d e r e d with a lamina dura. Tooth slightly loosened.
cases with large radiolucencies near the osseous crest; and cases where posts are involved. Benenati et al. (2) point out that the prognosis for surgical treatment is better if it is completed before a communication is established between the oral cavity and the
1. Gutman JL, Harrison JW. Surgical endodontics. Boston: Blackwell, 1991:411,413. 2. Benenati FW, Roane JB, Biggs JT, Simon JH. Recall evaluation of iatrogenic root perforations repaired with amalgam and gutta-percha. J Endodon 1986;12:161-5. 3. Str6mberg T, Hasselgren G, Bergstedt H. Endodontic treatment of traumatic root perforations in man. Swed Dent J 1972;65:457-66. 4. Kvinnsland R, Oswald B J, Halse A, Gronningsaeter AG. A clinical and roentgenological study of 55 cases of root perforation. Int Endod J 1989;22: 75-84. 5. Alhadainy HA. Root perforations. Oral Surg Oral Med Oral Patho11994; 78:368-74. 6. Rud J, Munksgaard EC, Andreasen JO, Rud V, Asmussen E. Root filling with composite and a dentin bonding agent. I-Vi. Danish Dent J 1989:93, 156-60, 195-7, 223-9, 267-73, 343-5, 401-5. 7. Rud J, Munksgaard EC, Andreasen JO, Rud V, Asmussen E. Retrograde root filling with composite and dentin-bonding agent. 1. Endod Dent Traumatol 1991 ;7:118-25. 8. Rud J, Munksgaard EC, Andreasen JO, Rud V. Retrograde root filling with composite and dentin-bonding agent. 2. Endod Dent Traumato11991 ;7: 126-31. 9. Ambus C, Munksgaard EC. Dentin bonding agents and composite retrograde root filling. Am J Dent 1~)93;6:35-8. 10. Andreasen JO, Munksgaard EC, Fredebo L, Bud J. Periodontal tissue regeneration including cementogenesis adjacent to dentin-bonded retrograde composite fillings in humans. J Endodon 1993;19:151-3. 11. Rud J, Rud V, Munksgaard EC. Retrograde root filling with dentinbonded modified resin composite. J Endodon 1996;22:477-80. 12. Rud J, Rud V, Munksgaard EC. Long-term evaluation of retrograde root filling with dentin-bonded resin composite. J Endodon 1996;22:90-3. 13. Rud J, Rud V, Munksgaard EC. Effect of root canal contents on healing of teeth with dentin-bonded resin composite seal. J Endodon 1997;23:53541. 14. Bud J, Andreasen JO, Moiler Jensen JE. A follow-up study of 1,000 cases treated by endodontic surgery. Int J Oral Surg 1972;1:215-28. 15. Hansen EK, Asmussen E. Cavity preparation for restorative resins used with dentin adhesives. Scand J Dent Res 1995;93:474-9. 16. Rud J, Andreasen JO, Moiler Jensen JE. Radiographic criteria for the assessment of healing after endodontic surgery. Int J Oral Surg 1972;1:195214. 17. Wulff HB, Schlichting P. Medstat. Copenhagen: Astra, 1987. 18. Andreasen JO, Rud J. Correlation between histology and radiography in the assessment of healing after endodontic surgery, int J Oral Surg 1972; 1:161-73. 19. Bud J, Andreasen JO, Meller Jensen JE. A multivariate analysis of the influence of various factors upon healing after endodontic surgery. Int J Oral Surg 1972;1:258-71. 20. Molven O, Halse A, Grung B. Incomplete healing (scar tissue) after periapical surgery--radiographic findings 8-12 years after treatment. J Endodon 1996;22:264-8. 21. Skoglund A, Pearson G. A follow-up study of apicoectomized teeth with total loss of the buccal bone plate. Oral Surg Oral Med Oral Path 1985;59:78-81.
JOURNALOF ENDODONTICS Copyright © 1998 by The American Association of Endodontists
VOL. 24, NO. 10, OCTOBER1998
Retrograde Sealing of Accidental Root Perforations with Dentin-Bonded Composite Resin Jorgen Rud, Dr.odont., Vibe Rud, DDS, and Erik C. Munksgaard, Dr.odont.
Surgical treatment was performed on 100 iatrogen perforations, of which 94 showed radiolucency of bone adjacent to the perforation and 83 presence of an exposed post. The perforations were in all cases sealed with a resin composite (Retroplast) bonded to adjacent root dentin with a dentin bonding agent (Gluma). Cases were examined after - 1 yr (first recall) and, if necessary after 11/= to 11 yr (mean: 4.1 yr) (latest recall). It was observed that the presence or absence of bone on the root between the perforation and cervix at the time of operation had no significant effect on the healing result, and that the radiographic classification "partial" healing with a border of cortical bone or a lamina dura often remained unchanged for many years. At latest recall, the healing result of 65 teeth originally having root perforation elsewhere than at the furcation was: 71% complete, 11% partial, 3% uncertain, and 15% failure. This is significantly different from the healing result of 27 molar teeth originally having perforation at the furcation: 30% complete, 4 1 % partial, 11% uncertain, and 18% failure.
after treatment, respectively. Cementum and a periodontal ligament with Sharpey's fibers in close contact to the Retroplast seal was identified, indicating a good tolerance by the tissue (6, 10). In clinical studies, it was shown that infected root canals with periapical infections treated with Gluma-bonded Retroplast demonstrated a healing frequency of - 9 0 % (11), compared with 72% when amalgam was used as a retrofill material (14), in both cases estimated some years after surgery. Cases showing complete healing 1 yr after surgery with Retroplast remained unchanged when re-examined 8 to 9 yr later, indicating stability of both the retrofill material and the bonding (12). In all of the aforementioned studies, Retroplast was applied not as a filling, but only as a seal at a slightly saucerized root end. This geometry prevents contraction gap formation in contrast to a regular box-like preparation such as that used for amalgam in retrofills (15). The purpose of this study was to examine the healing results of surgical Gluma Retroplast seals of iatrogene root perforations, which were not suited for immediate orthograde treatment.
M A T E R I A L S AND M E T H O D S One hundred thirty-five consecutive patients with iatrogenic root perforation, referred to and treated by the authors (V.R. and J.R.) in their private practice from November 1984 to March 1995, form the material in this prospective examination. Among these, the following perforations were not included in the study: 10 cases with root fractures, 5 cases with total removal of the root, and 1 case with reimplantation. Of seven reoperations, only failure in the primary operation were included. Combined orthograde root filling and retrograde seals of perforations were not included in the material. All patients had pre- and postoperative radiographs taken and were recalled - 1 yr after the operation and eventually later. A total of 100 patients responded; 84% of the 119 potential cases. There were 64% females and 36% males with medium age of 50 yr (range: 19-85 yr). None of the perforations in the series were referred for treatment immediately after the accident. Access to the root perforation was obtained after elevating a facial or palatal gingival flap and by removal of covering bone. The root dentin around the perforation was slightly hollowed and treated with Gluma, and the area was sealed with Retroplast using the method for apical sealing as described previously (6-8, 13). Retroplast is a specially designed
Accidental root perforations are normally treated with orthograde gutta-percha, internal placed amalgam, Cavit, or with calcium hydroxide immediately after the accident (1-5). However, cases with large perforations and with overfill do not respond well to this treatment, and perforations caused by posts that cannot be removed may be impossible to treat orthograde. In addition, orthograde treatment of perforations with radiolucency of the surrounding bone reaching the gingival pocket has been considered to have a poor prognosis (2). In such cases, surgical treatment with retrograde closure might be indicated. The sealing effect of dentin-bonded resin composites applied as apical retrograde root filling has recently been evaluated by laboratory experiments and clinically (6-13). Ambus and Munksgaard (9) found that an apical cover with a resin composite (Retroplast) may be gap-free when bonded with Gluma or other dentin-bonding agents. In other studies, tissue from monkeys and humans adjacent to apical Retroplast seals was examined histologically 1 and 3 yr
671
672
Journal of Endodontics
Rud et al.
resin composite with sufficient radiopacity, thixotropic effect, and 2 min working time colored to enable precise application (! 1). In the anterior and premolar regions, access to lingual or palatal perforations was often done from the buccal side by removing interradicular bone and sometimes root substance. Missing root substance was replaced by Retroplast. Access to intraradicular perforations of maxillary molars was obtained after substantial resection or, in some cases, by removal of one of the buccal roots. When treating cases with perforating posts, the post-end was resected until slightly below the root surface and the surrounding dentin was slightly hollowed. Loose posts were cemented before sealing with Retroplast. The location of the perforation and occurrence of bone loss on the root was noted. The healing was classified on the basis of radiographs and clinical symptoms into one of the following categories, which were based on criteria proposed by Rud et al. (16) and adapted to root perforations: 1. Complete healing--(a) Radiographic bone regeneration next to proximal perforations with or without a recognizable periodontal ligament space up to twice as large as normal (Fig. 1); radiographs eventually taken in different projections. (b) Disappearance of formerly recognized preoperative radiolucency adjacent to buccal or palatal/lingual perforations, with unrecognizable periodontal ligament space (Fig. 2). 2. Partial healing--Bone regeneration up to the level of the seal at which location a border of compact bone or lamina dura may have developed (Fig. 3). 3. Uncertain healing--Cases where healing cannot yet be determined or cases undergoing phases of healing; a decreased radiolucency compared with the postoperative or a previous follow-up radiograph, but without a compact bone border or lamina dura. 4. Failure--Unchanged or enlarged rarefaction originating from the perforation; loose sealing material seen on radiograph; presence of clinical signs of inflammation originating from the perforation area such as a fistula or an abscess.
Evaluation The authors (J.R. and V.R.) evaluated the radiographs separately and reached a common verdict in the 18 cases (18%) where primary differences occurred. The results were evaluated by X2-test (M × N, two-tailed or a Fischer's exact test) (17).
FIG 1. Perforation of a post mesiocervical in central incisor. Fortysix-year-old female. (A) Postoperatively. The root has been ground to just below the root surface and covered with Retroplast. A minute bridge of cervical bone was preserved. (B) Nine years and 2 months after operation. Complete bone healing of radiolucency and the cervical area. Reformation of periodontal ligament space and lamina dura is seen.
RESULTS The distribution of the 100 teeth in the present study according to type is shown in Fig. 4. Maxillary teeth were represented in 60% and mandibular teeth in 40% of the treated teeth. Perforations in lower incisors and canines had low representation. Besides 31 perforations at molar furcation areas, perforations were located mesially (52%), buccally (21%), and distally (23%). Lingual/palatal perforations were rare (4%). Midroot perforations were more common, compared with cervical and apical perforations. Premolar furcation perforations occurred in 6% of the cases. Among the 100 cases, 83 had posts in the perforated roots. Most of these posts were exposed in the perforation while perforated through the root. Other filling materials were extruded through the perforation in 13% of the cases. A preoperative-radiolucency of the
adjacent bone was seen in 94% of the cases. Of the six cases without radiolucency, three had a fistula and one an abscess. Of the 100 patients examined after the operation, 89 were seen after 6 to 18 months (mean: 1.0 yr) (first recall). Eleven cases not examined at first recall were seen later. Fifty-four patients were seen between 18 months and 11 yr (mean: 4.1 yr) (latest recall). Eight cases not examined at latest recall were partial or uncertain healings at first recall. The interrelations between healing results of the first and latest recall were as follow: eight cases with complete healing at both the first and latest recall were found. This agrees with previous findings using Retroplast that complete apical healing is unchanged and stable when examined later, unless interference of new factors
Retrograde Sealing of Root Perforations
Vol. 24, No. 10, October 1998
673
TABLE 2. Origin of 92 cases in Table 1 showing various healing types at latest recall Healing at Latest Recall
From healing II first recall From healing III first recall Cases not examined at first recall Found at both first and latest recall Healings transferred from first recall* Sum of latest recall
I
II
III
IV
2 4 6 8 34 54
-6 3 9 -18
---5 -5
-6 2 3 4 15
I, complete healing; II, partial healing; III, uncertain healing; IV, failures. * As explained in the text, first recall healings I and IV remain unchanged by time.
TABLE 3. Healing frequencies at first and latest recall for patients treated for root perforation at molar furcation Healing Category Complete Partial Uncertain Failure Subtotal
First Recall (Mean: 1 yr)
Latest Recall (Mean: 5.4 yr)
No.
%
No.
%
4 8 11 3
15 31 42 12
8 11 3 5
30 41 11 18
26
100
27
100
Not examined at first recall Not examined at latest recall FIG 2. Buccal perforation of an upper first premolar. Forty-eightyear-old female. (A) Postoperatively. A radiolucency is seen on each side of the root. (B) Complete bone healing at 7 months follow-up.
with no connection to the method (e.g. root fracture) occur (12, 13). This means that cases with complete healing at first recall may be transferred to the latest recall group without further examination. This is of cause true also for failures that, almost by definition, will not change with time. Efforts were therefore concentrated on recalling patients with partial or uncertain healing at first recall, which might change healing group at later recall. The healing results in the total material are shown in Table 1. To show the TABLE 1. Healing frequencies at first and latest recall for patients treated for root perforation First Recall Healing Category Complete Partial Uncertain Failure Subtotal Not examined Total
No.
%
Latest Recall No.
%
42 14 26 7
47 16 29 8
54 18 5 15
59 20 5 16
89
100
92
100
11
8
100
100
The first recall occurred between 0.5 and 1.5 yr (mean: 1.0 yr). The numbers showing complete healings and failures at first recall were transferred to latest healing. Apart from the transferred 1-yr results, the latest recall took place between f.5 and 11 yr (mean: 4.1 yr).
Total
31
31
relation between the healing groups, an account was made of the origin of the latest recall, partly based on the results from the first recall (Table 2). Healing frequencies for 31 perforations in molar furcations, observed after - 1 and 5.4 yr, are shown in Table 3. Twenty-five of the perforations in furcations were in mandibular molars and six in maxillary (Fig. 4). Three of the upper perforations were in palatal roots (Fig. 5). Of the 25 mandibular perforations in furcation, 18 (72%) were in the distal root and 2 were between the roots. Twenty-two (88%) of the perforated mandibular roots contained posts. Examples of furcation healings are the following: complete healing (Fig. 6), partial healing with lamina dura (Fig. 7), and partial healing with compact bone (Figs. 3 and 5). Five of the 31 furcations were not examined at first recall, but only at latest recall. Four complete healings and three failures were transferred from first to latest recall. The results in Table 4 are the healing frequencies of cases with perforation elsewhere on the root than at the furcation. They are obtained by deducting the numbers in Table 3 (furcations) from those in Table 1 (total material). Absence of bone on the root extending from the cervix and around the perforation was recorded in 83 cases during the operation. This deficiency derived from the perforation and not as a result of marginal periodontal disease. Table 5 shows healing categories in cases with and without such bone on the root.
674
Rud et al.
Journal of Endodontics
FIG 3. Perforations of posts in furcation area and mesial and distal roots. Twenty-nine-year-old female. (A) Postoperatively. Retroplast covers the distal post and furcation. (B) Almost the whole length of root is perforated by the post. Just covering the post with Retroplast, it need not be resected. (C) After 11/2yr cortical bone regeneration is seen up to the level of Retroplast seal. Tooth was extracted soon after because of a root fracture. See crack in distal and mesial root. TABLE 4. Healing frequencies at first and latest recall for patients treated for root perforation elsewhere on the root than at the furcation
NUMBER OF TEETH P~ 2 2011~ DIO
P2
I--~
Healing Category
M1 M2
__¢
_Q ii -010 0
~ 20
Cases with perforation at the furcation
FIG4. Location of 100 perforations according to tooth type and arch.
A condition for re-establishment of a periodontal ligament and lamina dura adjacent to a root perforation is that a tight and permanent seal can prevent bacteria and their toxins in the root canal from entering the surrounding periodontal tissues (18, 19). It seems in the majority of cases that the cover of resin composite (Gluma and Retroplast) provided such a seal (Table 1). After sealing of root perforations, bone may regenerate until it reaches the root surface and forms part of a periodontal ligament space with a lamina dura (Figs. 1, 6, and 8). In some cases the bone may, however, just regenerate until it reaches the border of the Retroplast seal where further growth stops. The bone surface may then be bordered with compact bone or a lamina dura (Figs. 3, 5, 7, and 9). In the present study, this has been called partial healing. This seems to be different from incomplete apical healing caused by scar tissue (16). The bone regenerated in partial healings may be compared with the compact bone border, which may be seen after a successful periodontal treatment. Apical scar tissue healing
Latest Recall
%
No.
%
Complete Partial Uncertain Failure
38 6 15 4
60 10 24 6
46 7 2 10
71 11 3 15
Total
63
100
65
100
The number stated in the various heaiing categories are those from Table 1 (total material) minus those from Table 3 (furcations).
TABLE 5. Presence or absence ( + / - ) of bone on root between perforation and cervix, among 83 cases with various healing types (Table 1) at latest recall
Healing Category DISCUSSION
First Recall No.
+ Bone
- Bone
Total
Complete Partial Uncertain Failure
25 (62%) 7 (18%) 1 (2%) 7 (18%)
20 (47%) 11 (26%) 5 (12%) 7 (16%)
45 18 6 14
Total
40 (100%)
43 (101%)
83
In 9 cases (8 complete healings and 1 failure), the + / - bone was not registered.
will eventually become complete healing (16, 20). Complete healing with reformation of a periodontal membrane space can, however, not always be expected to occur adjacent to sealed root perforations, even if the seal is tight and the sealing material biocompatible. A communication may exist from the bone defect around the perforation to the gingival pocket. This may lead to infection adjacent to the Retroplast seal. Adequate gingival hygiene, eventually with gingivectomy, may, however, result in partial healing and in some cases complete healing. The cervicular
Voh 24, No. 10, October 1998
Retrograde Sealing of Root Perforations
675
FIG6. Perforation of post in distal root to furcation of mandibular first molar. Forty-seven-year-old female. (,,4)Preoperatively. (/3) One year after sealing perforation with Gluma and Retroplast. Complete bone healing. Bone preserved buccal to distal root.
FIG 5. Divergence of root filling and post in palatal root of first maxillar molar indicate perforation. Sixty-one-year-old female. (A) Preoperatively. (B) Postoperatively. Distobuccal root removed to get access to furcation and palatal root. Post cut below root surface and covered with Gluma and Retroplast. (C) After 11 yr partial healing with bone defect from a distobuccal gingival pocket.
bone loss may, however, be so extensive, that bone healing will occur only to a certain distance on the perforated root (Fig. 9). The presence or absence of bone on the root between the perforation and the cervix was noticed during the operations. Absence of bone, especially in this area, could be expected to influence bone healing around and adjacent to the Retroplast seal. The numbers in Table 5 might indicate that complete healing occurs more often with bone on the root, whereas partial and uncertain healings are found more often in cases without bone. The
apparent differences were, however, not statistical significant (X2 = 4.01, p = 0.26). The findings showed that, even with no bone on the root, a limited bone healing may take place. This has also been found in a follow-up study of apicoectomized teeth with total buccal bone loss, showing 37% successful apical healings (21). However, if just a small bridge of crestal bone exists, it should by all means be preserved. In cases of treated molar furcation perforations (Table 3), complete healings increased from 15 % to 30% from first to latest recall. At latest recall, the proportion of cases showing complete healing was significant less (p = 0.0006) than that observed with restored perforations elsewhere on the root than at the furcation (Table 4), and the overall healing results of these two groups differ significantly (X2 = 16.6, p = 0.0009). These numbers emphasize the general observation that treatment of furcation perforations have a poorer prognosis than nonfurcation cases (1-5). The healing frequencies at first recall of restored perforations elsewhere on the root than at molar furcations (Table 4) were significantly different 0( z = 13.3, p = 0.004) than at latest recall. In addition, complete healing in 60% and 71% of the cases, respectively, is less than that seen for apical healing (11, 13). The proportion of uncertain healings decreased from 24% at first recall to 3% at latest recall, a decrease that is significant (p = 0.0009). This shows that uncertain healings in time change to one of the other healing groups.
676
Rud et al.
Journal of Endodontics
FtG 8. Perforation of the root caused by a too thick post in a lower canine. Forty-three-year-old female. (A) Preoperatively. (B) Complete healing. Reformation of periodontal ligament at 5-yr follow-up.
FIG 7. Perforation of post and cement to furcation in mandibular first molar. Seventy-eight-year-old female. (A) Preoperatively. (B) Postoperatively after Retroplast seal. (C) One-year follow-up. Partial healing. Bone demarcated with lamina dura.
The compact bone border seen in partial healing seems to be stable. Further bone deposits may occur, but slowly. It may take several years before a complete bone healing has been established. On the other hand, no cases of partial healing became uncertain healing or failure at the latest recall (Table 2). That partial healings are almost unchanged is also evident by the findings: 10% at the first recall and 11% at the latest recall in Table 4. Partial healing is the category most often found in furcation cases (Table 3). The proportion of partial healings of the total amount of restored furcations (Table 3), compared with that of the restored nonfurca-
tions at latest recall (Table 4), was statistically significantly different (p = 0.004). Partial healings in furcations may also be stable; in the present series, some were observed up to 11 yr. An analysis of the 15failures presented in Table 1 showed that they were mainly caused by absence of a tight Retroplast seal, as was the case for failures with apical retrograde root seals (13). This explains the development of cases with acute infection, as well as increased radiolucency. A limited number of reoperations was performed, with acceptable healing in approximately half the cases. A comparison of the present results with those of other authors is difficult. As Gutmann and Harrison (1) point out, "surgical repair of root perforations has received sporadic attention in the dental literature and has been supported primarily by case reports or limited studies." Therefore, statistically based analyses of differences are not possible. Despite the results in the present study, nonsurgical methods will and should probably be continued as a first choice. This may not include cases with large perforations, especially with overfill;
Vol. 24, No. 10, October 1998
Retrograde Sealing of Root Perforations
677
osseous defect associated with the perforation. Conventional treatment of the perforation should therefore be conducted shortly after the accident and the patients recalled not later than 1-yr later to assess the outcome. If the result is unsuccessful, surgery with Gluma-Retroplast may be undertaken as early as possible. This study was supported in part by the Danish Dental Association, Research Foundation. We thank Dr. Jens O. Andreasen, Dental Clinic, University Hospital (Copenhagen, Denmark) for his kind review of the manuscript. Drs. J. Rud and V. Bud have a private practice in oral surgery in Copenhagen, Denmark. Dr. E. C. Munksgaard is assistant professor, Department of Dental Materials, School of Dentistry, Faculty of Health Sciences, University of Copenhagen. Address requests for reprints to Dr. Jorgen Rud, Nikolaj Plads 7, DK-1067 Copenhagen, Denmark.
References
FIG 9. Distal perforation o f maxillar lateral incisor with d i s l o d g e m e n t of filling material in 19-yr-old female. (A) Preoperatively. (B) P o s t o p eratively. Perforation sealed with Retroplast. No cervical bone. (C) Follow-up after 6 yr. Partial b o n e regeneration b o r d e r e d with a lamina dura. Tooth slightly loosened.
cases with large radiolucencies near the osseous crest; and cases where posts are involved. Benenati et al. (2) point out that the prognosis for surgical treatment is better if it is completed before a communication is established between the oral cavity and the
1. Gutman JL, Harrison JW. Surgical endodontics. Boston: Blackwell, 1991:411,413. 2. Benenati FW, Roane JB, Biggs JT, Simon JH. Recall evaluation of iatrogenic root perforations repaired with amalgam and gutta-percha. J Endodon 1986;12:161-5. 3. Str6mberg T, Hasselgren G, Bergstedt H. Endodontic treatment of traumatic root perforations in man. Swed Dent J 1972;65:457-66. 4. Kvinnsland R, Oswald B J, Halse A, Gronningsaeter AG. A clinical and roentgenological study of 55 cases of root perforation. Int Endod J 1989;22: 75-84. 5. Alhadainy HA. Root perforations. Oral Surg Oral Med Oral Patho11994; 78:368-74. 6. Rud J, Munksgaard EC, Andreasen JO, Rud V, Asmussen E. Root filling with composite and a dentin bonding agent. I-Vi. Danish Dent J 1989:93, 156-60, 195-7, 223-9, 267-73, 343-5, 401-5. 7. Rud J, Munksgaard EC, Andreasen JO, Rud V, Asmussen E. Retrograde root filling with composite and dentin-bonding agent. 1. Endod Dent Traumatol 1991 ;7:118-25. 8. Rud J, Munksgaard EC, Andreasen JO, Rud V. Retrograde root filling with composite and dentin-bonding agent. 2. Endod Dent Traumato11991 ;7: 126-31. 9. Ambus C, Munksgaard EC. Dentin bonding agents and composite retrograde root filling. Am J Dent 1~)93;6:35-8. 10. Andreasen JO, Munksgaard EC, Fredebo L, Bud J. Periodontal tissue regeneration including cementogenesis adjacent to dentin-bonded retrograde composite fillings in humans. J Endodon 1993;19:151-3. 11. Rud J, Rud V, Munksgaard EC. Retrograde root filling with dentinbonded modified resin composite. J Endodon 1996;22:477-80. 12. Rud J, Rud V, Munksgaard EC. Long-term evaluation of retrograde root filling with dentin-bonded resin composite. J Endodon 1996;22:90-3. 13. Rud J, Rud V, Munksgaard EC. Effect of root canal contents on healing of teeth with dentin-bonded resin composite seal. J Endodon 1997;23:53541. 14. Bud J, Andreasen JO, Moiler Jensen JE. A follow-up study of 1,000 cases treated by endodontic surgery. Int J Oral Surg 1972;1:215-28. 15. Hansen EK, Asmussen E. Cavity preparation for restorative resins used with dentin adhesives. Scand J Dent Res 1995;93:474-9. 16. Rud J, Andreasen JO, Moiler Jensen JE. Radiographic criteria for the assessment of healing after endodontic surgery. Int J Oral Surg 1972;1:195214. 17. Wulff HB, Schlichting P. Medstat. Copenhagen: Astra, 1987. 18. Andreasen JO, Rud J. Correlation between histology and radiography in the assessment of healing after endodontic surgery, int J Oral Surg 1972; 1:161-73. 19. Bud J, Andreasen JO, Meller Jensen JE. A multivariate analysis of the influence of various factors upon healing after endodontic surgery. Int J Oral Surg 1972;1:258-71. 20. Molven O, Halse A, Grung B. Incomplete healing (scar tissue) after periapical surgery--radiographic findings 8-12 years after treatment. J Endodon 1996;22:264-8. 21. Skoglund A, Pearson G. A follow-up study of apicoectomized teeth with total loss of the buccal bone plate. Oral Surg Oral Med Oral Path 1985;59:78-81.