T. V. Castrisos, J. E. A. Palamara & P. V. Abbott
School of Dental Science, University of Melbourne, Melbourne, Victoria, Australia.

Introduction.
Several techniques are available for the removal of posts from roots and these include the use of ultrasonic vibration and various post removal devices. Most of the literature on post removal discusses the various methods used in the form of clinical case presentations, although some studies have investigated the efficiency of post removal techniques. Buoncristiani et al. (1994) compared the efficiency of ultrasonic and sonic vibration to remove parallel-sided preformed posts and concluded that ultrasonic vibration could remove them in less than 10 min. Berbert et al. (1995) and Johnson et al. (1996) have also examined the influence of ultrasonic vibration on the forces required to remove posts. Berbert et al. (1995) used cast posts/cores and concluded that the force required to remove them after application of ultrasonics for 2 to 5 min was 30% less than the control group with no ultrasonic vibration. Johnson et al. (1996) concluded that 16 min of ultrasonic vibration could significantly reduce the amount of force required to remove parallel-sided preformed posts in extracted human premolars. Yoshida et al. (1997) examined the effect of using one or two ultrasonic tips simultaneously placed on a cast post and core and they reported that the use of two ultrasonic tips was significantly quicker. The use of the Eggler post remover (Automaton-Ventriebs-Gesellschaft, Germany) and the Gonon (Thomas) post remover (Thomas Extracteur De Pivots; FFDM, Bourges, France) have been described by Bando et al. (1985), Stamos & Gutmann (1991) and Machtou et al. (1989), who all described them as highly efficient methods for removing posts.
A survey of the methods used by American endodontists to remove intraradicular posts found that post removal devices such as the Gonon and Eggler post removers were not commonly used (Stamos & Gutmann 1993). The reasons given by the respondents to this survey for not using these devices were that they were too dangerous, could not be universally used, or did not work. In contrast, a recent survey of endodontists in Australia and New Zealand found that the Eggler post remover was the most frequently used post removal device, with 42% of the respondents using it regularly (Castrisos & Abbott 2001). Stamos & Gutmann (1991) stated that the advantages of using these types of post removers were:

1. conservation of remaining tooth structure
2. reduced risk of root fracture
3. reduced risk of root perforation, and
4. reduced risk of torque forces being placed on the root.

Although several articles have presented case reports about the efficiency of various devices, the safety of these devices has not been established. Altshul et al. (1997) compared the prevalence of dentinal cracks in teeth with cast posts that had been removed by either ultrasonic vibration or the Gonon post removal system. There was a statistically significant difference between the number of dentinal cracks present following ultrasonic removal of posts compared with teeth where the post was not removed. The dentinal cracks were seen more frequently at the cervical level following the use of ultrasonics and the time taken to remove the posts with ultrasonic vibration was significantly longer than with the Gonon system.
As a result of the design and size of the Eggler post remover (Fig. 1), it is usually used to remove posts in anterior teeth. It is considered by most endodontists who use it to be of greater benefit for the removal of cast posts/ cores than ultrasonics (Castrisos & Abbott 2001), where the time required to remove the post is greater.

Figure 1. The Eggler Post Remover.

To date, no studies have reported the amount or nature of forces applied to teeth during post removal. However, several methods have been used in various studies to determine the amount of force applied to teeth with various loads during other dental procedures. These methods include two-dimensional and three-dimensional photoelastic stress analysis (Henry 1977, Mentink et al. 1998), finite element stress analysis (Huysmans et al. 1993) and strain gauge measurements. The latter have been used to assess forces applied when obturating root canals using lateral condensation (Saw & Messer 1995, Lertchirakarn et al. 1999), when cementing posts (Ross et al. 1991, Obermayr et al. 1991) and in measuring the reduction in stiffness of endodontically treated teeth (Reeh et al. 1989). The strain gauges used in these studies were applied to the external surface of the root or crown to measure the strain on the tooth.
The aim of this study was to measure the surface strains caused by the application of an Eggler post remover parallel to the long axis of the tooth and to compare them with strains produced by loads applied at a 10 angle to the long axis.

Materials and methods.
Twenty extracted single-rooted human teeth were obtained from the clinics of the Royal Dental Hospital of Melbourne. At all times prior to and during the experiment, the teeth were kept moist by either storing them in phosphate buffered saline or by covering them with gauze moistened in phosphate buffered saline. Radiographs were taken from the buccal and mesial directions to ensure that only a singleroot canal was present and all the teeth were examined under a microscope ( 16 magnification) to ensure they were free of fractures and cracks. The crowns of the teeth were removed with a diamond bur at a level 1 mm coronal to the labial cemento-enamel junction and the teeth were then measured. The mesio-distal dimensions of the teeth at the cemento-enamel junction ranged from 4.2 mm to 5.9 mm and each tooth was checked to ensure that the repeller arms of the Eggler post remover could be rested on the root surface.
The root canals of each tooth were negotiated and instrumented up to a size 40 Hedström file at a length that was 1 mm short of the apical foramen. The apical third of each canal was flared with the step-back technique and the coronal third was flared with sizes 2 and 3 Gates Glidden burs. Whilst being instrumented, the canals were irrigated with 1% sodium hypochlorite between each file size. They were then filled with guttapercha and AH26 (Dentsply DeTrey GmbH, Konstanz, Germany) using lateral condensation; a post space was provided in each tooth to a length of 10 mm by removing gutta-percha with a heated plugger. A minimum of 4 mm of gutta-percha remained in each tooth. Post spaces were further prepared with Parapost drills (Coltene Whaledent, Konstanz, Germany), ensuring that at least 1 mm of dentine remained on both the mesial and distal sides of the root canals at the apical extent of the post holes. The dimensions of the post holes were verified radiographically.
The teeth were divided into two groups, with approximately equal distribution of root lengths and diameters based on visual inspection. The post hole preparations were then finalized: in group 1, the post holes were flared with Gates Glidden burs to leave a thickness of 1 mm of dentine coronally on the buccal, lingual, mesial and distal sides; in group 2, post holes were prepared so as to maintain 2 mm of dentine on each side of the root face. The final shape of each post was parallel-sided for the apical half with a tapered coronal half to accommodate the thickness of coronal dentine required for each group. Direct patterns for cast posts/cores were formed using Parapost burn out points and GC Pattern Resin (GC Corporation, Tokyo, Japan). The core was cuboid-shaped with dimensions of 2 mm 2 mm and a height of 3 mm to allow placement of the Eggler post remover. The post/ cores were cast in a non-precious alloy and cemented with zinc phosphate cement (SS White, Gloucester, UK) which was allowed to set for 1 week prior to removal of the posts with the Eggler post remover.
Strain gauges (EA-06–125BT-120, Micro-Measurements Group Inc., Raleigh, NC, USA) were trimmed to fit the root surface. Four strain gauges were placed vertically 1.5 mm below the coronal root surface after each tooth had been lightly scraped on the buccal, lingual, mesial and distal surface to facilitate their placement. Pilot studies indicated that this was the most appropriate site for placement of the strain gauges so that some extra dentine could be removed for the second stage of the experiment when the posts were removed at a 10 angle and also to allow clearance between the point of loading and the wires of the strain gauge. The root surface was etched with 37% orthophosphoric acid for 30 s, washed with water for 15 s and dried with air. The strain gauges were cleaned with chloroform, primed with a catalyst and attached to the root surface with cyanoacrylate adhesive (M-Bond 200, Micro-Measurements Group Inc., Raleigh, NC, USA). Excess cyanoacrylate adhesive was removed by applying pressure to the surface of the strain gauges until adhesion occurred, leaving a thin layer of cyanoacrylate that did not interfere with the measurement of strain. The strain gauges, solder contacts and roots were covered with silicone (Dow Corning 3140 RTV Coating, Dow Corning Corp., Midland, MI, USA) to protect the strain gauges from moisture. The strain gauges were then connected to a data acquisition board (AT-MID-16E-2, National Instruments Corporation, Austin, TX, USA) and the data was stored on a computer using process control software (LabVIEW 4.0, National Instruments Corporation, Austin, TX, USA). Four separate channels were used to allow simultaneous measurements from each of the strain gauges and the measurements were made continuously, whilst removing the posts with the Eggler post remover. Throughout the experiment, one tooth was used as the experimental tooth, and a second tooth, which had been prepared in the same manner, served as a compensator for apparent strain that may arise from temperature fluctuations and electrical heating.

Figure 2. Application of the Eggler Post Remover to the core section of a cast post/core with strain gauges attached to the external root surfaces



Figure 3. The Eggler Post Remover in use. Note the repeller arms have splayed as they are pushed against the mesial and distal surfaces of the root face which has caused the left repeller arm to almost slide off the tooth.

In order to remove the post, the forceps of the Eggler device were placed over the buccal and lingual surfaces of the core and tightened with the large inner wheel (Fig. 1). The repeller arms were then lowered by tightening the small thumbscrew at the top of the device. As the repeller arms were lowered, they contacted the mesial and distal tooth structure and the forceps moved away from the tooth as the post was removed (Figs 2,3).
Strain measurements were initially recorded when the cast post/cores were removed from the teeth in a direction that was parallel to the long axis of each tooth and then again when removed at an angle of 10 to the long axis. After the initial removal of the posts, any retained luting agent was cleaned from the posts and from within the post spaces by scraping it off with a probe and an excavator. The teeth were examined under magnification (16 ) to ensure complete removal of cement and for any evidence of fractures or cracks. The posts were recemented after a minimum period of 24 h and the teeth were kept moist in phosphate buffered saline during this time. The height of the mesial dentine of each tooth was reduced so that when the Eggler post remover was applied to the cores, it was at a 10 angle to the long axis of each tooth. The 10 angle was measured with a protractor following application of the Eggler post remover. Then the posts/cores were removed from the teeth again and strain gauge measurements were recorded. The teeth were examined once again under magnification to determine whether there were any fractures or cracks.

Statistical analysis.
Strain measurements were recorded continuously on the buccal, lingual, mesial and distal surfaces of the tooth whilst each post was being removed. The maximum strains recorded when removing the post along the long axis and at a 10 angle to the long axis within the same group were compared using the Wilcoxon Signed Rank test. Comparisons were made between group 1 and group 2 for removal of the post along the long axis and at the 10 angle to the long axis using the Mann–Whitney U -test. All comparisons were made with a 95% level of confidence to determine statistical significance and data were analysed using the SPSS statistical software program ( SPSS for Windows 6.1.31995, SPSS Inc, Chicago, IL, USA).

All posts were removed in less than 2 min with the Eggler post remover. Strain measurements could not be recorded in two teeth in group 1 when removing the post at 10 to the long axis of the tooth because the repeller arms did not contact the root surface when attempting to remove the post.
Strain is the change in length divided by the original length of the object and it has no unit of measurement. In this study, the amount of deformation was reported as ‘ strain’ (where 10 000 strain = 0.01 strain = 1% deformation). A positive value indicated tension, whilst a negative value indicated compression. Tables 1 and 2 list the ranges, means and standard deviations for the greatest strains during post removal, as well as the character of the strain (tension or compression). The greatest strain on any surface for each tooth was identified and the strains on the other three surfaces were measured at that same point in time.

Root strain during post removal parallel to the long axis.
The magnitude of strains measured on the buccal and lingual root surfaces was lower than those measured on the mesial and distal surfaces where the repeller arms of the Eggler post remover contacted the tooth. When posts were removed along the long axis of the teeth, the strain measurements were similar on the mesial and distal surfaces for most teeth, although there were large discrepancies between strain measurements in three teeth from group 1 and four teeth from group 2. All strain measurements in group 1 indicated compressive stresses on the mesial and distal surfaces, whilst seven out of 10 teeth had tensile stresses on the buccal and lingual surfaces. Compressive stresses were present in nine out of the 10 teeth on the mesial and distal surfaces in group 2. It was noted that the post was difficult to remove from the tooth that had tensile stresses and there was some rotation of the Eggler post remover during use on this tooth. Tensile stresses were present on the buccal or lingual surfaces of two of the 10 teeth in group 2. The Mann–Whitney U -test indicated that the strain measurements of group 1 were not statistically different to those of group 2 (Table 3) at the 95% level of confidence.

Table 1. Strain measurements for group 1 (1 mm thickness of dentine).


Strain measurements do not have a unit of measurement but are expressed as ' strain' where 1  strain represents a change in dimension of 1 part in 1 million. Positive values indicate tension, whilst negative values indicate compression.

Table 2. Strain measurements for group 2 (2 mm thickness of dentine).


Strain measurements do not have a unit of measurement but are expressed as ' strain' where 1  strain represents a change in dimension of 1 part in 1 million. Positive values indicate tension, whilst negative values indicate compression.

Table 3. Statistical analysis: Comparisons between groups 1 and 2 using the Mann-Whitney U-test at the 95% level of confidence.

Root strain during post removal at 10 to the long axis.
The magnitude of the strains measured on the buccal and lingual surfaces were lower than those measured on the mesial and distal surfaces and the highest strains were recorded on the mesial surface. These strains were higher in seven out of eight teeth in group 1 and in seven out of 10 teeth in group 2. All teeth in both groups had compressive stresses on the mesial and distal surfaces. The Mann–Whitney U -test indicated that there was no significant difference in the strain measurements between groups 1 and 2 (Table 3) at the 95% level of confidence.

Comparison of strains within each group.
Comparisons were made within the same group using the Wilcoxon Signed Rank test when removing the post parallel to the long axis and at 10 to the long axis of the root. There were no significant differences between the strain measurements on the buccal, lingual, mesial and distal surfaces in both groups (Table 4) with a 95% level of confidence.

Table 4. Statistical analysis: comparisons within group 1 and within group 2 between removal of the posts parallel to the long axis and at a 10° angle to the long axis using the Wilcoxon Signed Rank Test at the 95% level of confidence.



Table 5. Number of root fractures that occurred in each group during post removal.



Figure 4. Buccal surface of a tooth with an oblique root fracture following removal of a cast post/core with the Eggler Post Remover applied at a 10  angle to the long axis of the root. This specimen was from group 1 with only 1 mm thickness of dentine remaining in the coronal portion of the tooth.

Root fractures during post removal.
Fractures of the root occurred in four teeth – three of these were from group 1 and one from group 2. All of these fractures occurred when the posts were being removed at an angle of 10 to the long axis of the root (Table 5). The fractures in three of these teeth (two from group 1; one from group 2) were small slivers of dentine which resulted in loss of 1–2 mm of dentine on the outer surface of the root and all of these fractures occurred at the point where the repeller arms of the Eggler post remover contacted the tooth. The other tooth from group 1 had an oblique root fracture on the buccal surface which extended at least 6 mm in a disto-apical direction (Fig. 4).

Discussion.
The greatest strains measured were on the mesial and distal root surfaces where the repeller arms of the Eggler post remover contacted the root. There was considerable variation in strain from a tensile value of 214 strain to a compressive strain of 5406 strain. The strains measured at the root surface were below the slightly greater than 1% strain required to exceed the compressive elastic limit where fracture may occur (Waters 1980).
Posts were cemented and then removed twice for each tooth – first along the long axis of the root (0 ) and then at a 10 angle to the long axis. Post removal was performed in this order because there were concerns that root fractures may occur when the posts were removed at 10 to the long axis of the tooth and this was confirmed with four teeth during the second stage of the experiment. The teeth were examined under a microscope ( 16) for evidence of any fractures or cracks prior to recementing the post, but none were seen. However, it is recognized that multiple cementation and removal of each post, and the order in which the posts were removed, could possibly have had deleterious effects on the tooth roots but this could not be determined in the current study. The same teeth were used at 0 and then at 10 in an attempt to provide a more accurate comparison of the strains produced by loading the roots at the two different angles.
Four teeth (20% of cases) suffered damage when the Eggler post puller was not used correctly – that is, at an angle of 10 to the long axis of the tooth root. Only one of these was a root fracture that clinically would have rendered the tooth unrestorable and necessitated extraction. The other three teeth had small (1–2 mm) sliver fractures of dentine where the repeller arms of the Eggler post remover contacted the teeth. These findings emphasize the need to be careful to ensure that the Eggler post remover is applied parallel to the long axis of the tooth before applying any force.
There were several factors that may have contributed to these fractures. First, there was an uneven application of load by the repeller arms during post removal, as demonstrated by the strain measurements in the teeth that fractured where there was a difference between the distal and mesial surfaces with compressive strain values ranging from 490 up to 3481. Secondly, it was observed that the distance between the two repeller arms increased as the load applied to the root surface was increased (Fig. 3). This action introduces lateral shear forces on the mesial and distal root surfaces, which may increase the risk of having small fractures as was seen at the site of application of the repeller arms. Fractures may occur since dentine is weaker under shear or tensile loading than under compression (Waters 1980). In the region where the repeller arms contacted the cut root faces, it is possible that point, tensile or shear stresses may have initiated cracks or exceeded the ultimate strength of dentine to cause these fractures.
One tooth had an oblique fracture on the buccal root surface, which is typical of most vertical root fractures. The reason for this has not been established, although it could be related to the stress patterns that developed within the tooth. The tooth that fractured exhibited tensile stresses on the buccal surface and the fracture line progressed in a distal direction as it moved apically, which was opposite to the direction in which the post was being pulled out of the canal. The fracture also may have been related to the amount of remaining dentine thickness, as shown by Tjan & Whang (1985), who measured the forces required to fracture roots when removing cast posts/cores with different thickness of buccal dentine. They reported that there was no difference in the load required for fracture, but the type of failure that occurred varied according to the thickness of remaining dentine. Tjan & Whang (1985) concluded that roots with less than 1 mm of dentine remaining were more prone to fracture than those with 2 mm or more of buccal dentine. In the current study, the tooth that had an oblique fracture had only 1 mm of dentine remaining coronally.
The Eggler post remover is designed so that the gripping forceps engage the core portion of the post/core in order to pull the post out of the root canal. The repeller arms partially surround the gripping forceps and push against the root face to provide the load to remove the post. The distance between the repeller arms is 3.2 mm and hence the Eggler device can not be used on teeth with a mesio-distal distance between the adjacent teeth that is less than this. In addition, the mesio-distal dimension of the tooth root will determine how the repeller arms contact the surface of the root. In narrow teeth the repeller arms may only contact the outer portion of the cut root face, whilst in wider teeth the entire cross sectional area of the repeller arm could be in contact with the cut root face. In narrow teeth, the surface area through which the load is applied will be decreased and most of the load will be applied through the outer dentine of the root face (Fig. 4), which may increase the risk of root fracture occurring.
When the posts were removed along the long axis of the teeth, the mesial and distal strains were equal in the majority of cases, but in seven teeth there was a large discrepancy. This may have been due to incorrect alignment of the post remover along the long axis of the tooth in these specimens. Since the alignment of the post remover is determined by the positioning of the forceps on the core and if it is placed at an angle to the long axis of the tooth, then one of the repeller arms will contact the tooth before the other, which results in a greater load being applied through that arm. Another possibility is that some excess localized strain may develop under one of the repeller arms. When the posts were removed at a 10 angle to the long axis of the tooth, the mesial strain was higher than the distal strain in seven teeth in group 1 and eight teeth in group 2, indicating that there was uneven loading through the repeller arms.
Whilst using the Eggler device in this study, it was noted that there is a discrepancy between the width of the cross bar that connects the two repeller arms (2.7 mm) and the dimensions of the gripping forceps (3.2 mm) that the repeller arms partially surround (Fig. 2). Hence, as the repeller arms are lowered to contact the tooth, they splay out and will contact the tooth at divergent angles. This introduces a lateral component to the load applied to the tooth which may have contributed to the occurrence of the dentinal ‘sliver’ fractures seen in this study. This problem is more likely to occur in narrow teeth where only the outer dentine is contacted by the repeller arms. The splaying of the Eggler’s repeller arms during use could be rectified with some simple modifications to the design of the device. Tensile forces were found on the buccal or lingual surfaces of seven teeth in group 1 and two teeth in group 2 when the posts were removed parallel to the long axis of the tooth. The vertical tensile strain on the buccal or lingual surfaces was most likely associated with principal compressional strains in the horizontal direction. Horizontal compressional strains on the buccal and lingual surfaces are generated when the mesial and distal surfaces are pushed out. This may have occurred when the repeller arms of the Eggler post puller splayed apart during use.

References.

Altshul JH, Marshall G, Morgan LA, Baumgartner JC (1997) Comparison   of dentinal crack incidence and of post removal time resulting from post removal   by ultrasonic or mechanical force. Journal of Endodontics 23 , 683-6.
  Bando E, Kawashima T, Tui IT, Kubo Y, Nakano M (1985) Removing dowels in   difficult teeth. Journal of Prosthetic Dentistry 54 , 34-6.
  Berbert A, Filho MT, Ueno AH, Bramante CM, Ishiikiriama A (1995) The influence   of ultrasound in removing intraradicular posts. International Endodontic   Journal 28 , 100-2.
  Buoncristiani J, Seto BG, Caputo AA (1994) Evaluation of ultrasonic and sonic   instruments for intraradicular post removal. Journal of Endodontics 20 ,   486-9.
  Castrisos TV, Abbott PV (2002) A survey of methods used for post removal   in specialist endodontic practice. International Endodontic Journal 35 ,   172-80.
  Henry PJ (1977) Photoelastic analysis of post core restorations. Australian   Dental Journal 22 , 157-9.
  Huysmans MC, Peters MC, Van der Varst PG, Plasschaert AJ (1993) Failure behaviours   of fatigue tested post and cores. International Endodontic Journal 26 ,   294-300.
  Johnson WT, Leary JM, Boyer DB (1996) Effect of ultrasonic vibration on post   removal in extracted human premolar teeth. Journal of Endodontics 22 , 487-8.  
  Lertchirakarn V, Palamara JEA, Messer HH (1999) Load and strain during lateral   condensation and vertical root fracture. Journal of Endodontics 25 , 99-104.  
  Machtou P, Safarti P, Cohen AG (1989) Post removal prior to retreatment.   Journal of Endodontics 15 , 552-4.
  Mentink AG, Creugers NH, Hoffenbrouwers PM, Meeuwissen R (1998) Qualitative   assessment of stress distribution during insertion of endodontic posts in photoelastic   material. Journal of Dentistry 26 , 125-31.
  Obermayr G, Walton RE, Leary JM, Krell KV (1991) Vertical root fracture and   relative deformation during obturation and post cementation. Journal of   Prosthetic Dentistry 66 , 181-7.
  Reeh ES, Douglas WH, Messer HH (1989) Stiffness of endodontically treated   teeth related to restoration technique. Journal of Dental Research 68 ,   1540-4.
  Ross RS, Nicholls JI, Harrington GW (1991) A comparison of strains generated   during placement of five endodontic posts. Journal of Endodontics 17 , 450-6.  
  Saw LH, Messer HH (1995) Root strains associated with different obturation   techniques. Journal of Endodontics 21 , 314-20.
  Stamos DE, Gutmann JL (1991) Revisiting the post puller. Journal of Endodontics   17 , 466-8.
  Stamos DE, Gutmann JL (1993) Survey of endodontic retreatment methods used   to remove intraradicular posts. Journal of Endodontics 19 , 366-9.
  Tjan AHL, Whang SB (1985) Resistance to root fracture of dowel channels with   various thickness of buccal dentine walls. Journal of Prosthetic Dentistry   53 , 496-500.
  Waters NE (1980) Some mechanical and physical properties of teeth. In:   Vincent D, ed. The Mechanical Properties of Biological Materials . Cambridge,   UK: Cambridge University Press, 63- 134.
  Yoshida T, Gomyo S, Itoh T, Shibata T, Sekine I (1997) An experimental study   of the removal of cemented dowel retained cast posts/cores by ultrasonic vibration.   Journal of Endodontics 23 , 239-41.