Dr Rahul Ravi, Dr Adnan Kheyroolla, Dr Omkar Shetty, Dr Rubina Tabassum, Dr Gaurang Mistr, Dr Kunal Mehta

P.G Student,

Dean and Professor,



Associate Professor,

Department of Prosthodontics and Implantology,

D.Y Patil University, School of Dentistry, Navi Mumbai, India


Implant therapy offers increased longevity, improved function, bone preservation and quality of life. This article discusses the literature related to the dynamics in effect when an implant is placed in the bone. Osseous and soft tissue changes take place around the implant subsequent to their placement. These changes determine the esthetic outcome of the implant.

Keywords : Implants, osseointegeration, esthetics, implant-abutment connection, platform switching

Citations : Ravi R, Kheyroolla A, Shetty O, Tabassum R, Mistry G, Mehta K. The implant - abutment connection and its relation to crestal bone - a review. J Prosthodont Dent Mater 2020;1(1& 2): 51-59.


Dental implants is a widely accepted device used as a predictable and reliable tool for dental reconstruction, however, it is still necessary to ensure that the height of the peri-implant crestal bone is maintained. Directly after insertion of a dental implant, a cascade of biological events occurs during the bone healing process. The change in bone shape and continuity is a result of this bone healing process, contrary to a possible pathological bone loss. Osseointegration is considered to be the phenomenon of direct apposition of bone on an implant surface, which subsequently undergoes structural adaptation in response to a mechanical load. Over time the shape of crestal bone around the implant changes both horizontally and vertically.

One of the criteria for the success of dental implant treatment is the amount of crestal bone change. Albrektsson et al.8 proposed that a dental implant can be considered successful if peri-implant crestal bone loss is less than 1.5 mm during the first year after implant placement and less than 0.2 mm annually after that.

Numerous theories are put forward to explain the crestal bone loss around implants including, surgical trauma due to excessive heat and pressure generation while preparing the osteotomy for the implant and periosteal flap elevation, or secondary to the attachment of prosthetic components due to occlusal overload and the presence of a micro gap between the implant and the abutment & its positioning in relation to the crestal bone. Although the success of a dental implant is probably multifactorial and dependent on some or all of the above factors, this review specifically addresses the relationship of the implant-abutment connection and its relation to crestal bone changes.

Implant-Abutment Connections and Micro gaps:

In the two-stage implant placement technique, the implant is placed at the bone crest level. After 3-6 months, a prosthetic abutment is installed on the implant to connect the implant to future prosthetic restorations (crowns, bridges or dentures). It is postulated that a potential microscopic space exists at the abutment/implant interface, along the abutment screw threads and at the base of the screw chamber known as micro gaps.

Three main factors are identified as possible causes for the formation of micro gaps: occlusal load during physiological function, manufacturing tolerance and micromotion between the implant–abutment connections. Microorganisms occupy this gap and set up a bacterial reservoir, resulting in an area of inflamed soft tissue facing the fixture abutment junction. This microbial leakage at the implant-abutment interface is a chief challenge for the constructing the two-stage implant systems & is a major contributing factor for peri-implant inflammatory reactions.


Fig 1: Microgaps at the Implant-abutment connection


Many studies have shown that this microscopic space between implant and abutment (micro gap) facilitates the infiltration of fluids and macromolecules from crevicular fluid and saliva, enabling bacterial invasion and proliferation, even in patients with good oral hygiene.

On the other hand, bacterial infiltration may also arise during the first stage or the second stage of implant surgery. Moreover, findings from several studies have documented that bacterial infiltration may occur both from an external source to the inner area of an implant, and in reverse. This migration of bacteria is probably facilitated through the unavoidable presence of micro gaps between the fixture and the abutment components of the assembled system.

The bacterial contamination may be correlated with gap sizes or misfits. Gap dimensions ranging from 20μm to 168μm have been reported. The level of contamination depends not only on the precision of fit but also on the degree of the applied micromovement and torque. The incidence of loads and unscrewing of the prosthetic abutment can increase infiltration, whereas optimal adaptation, minimal micromovement and exceptional prosthetic and occlusal planning are factors which can minimize microleakage, but do not seem to prevent it completely.

The presence of a micro gap in close relation to bone plays a role in developing of peri-implant inflammation and bone loss. Put merely, bacterial invasion of the micro gaps can interfere with the osseointegration of an implant during the healing phase of the surgical intervention and cause peri-implantitis. Pathogenic bacterial microflora may also influence the outcome of guided bone regeneration in the treatment of peri-implantitis.

Broggini et al. demonstrated an increase in inflammatory cells in the peri-implant soft tissues at the level or slightly coronal to the implant–abutment junction due to the bacterial presence, which when combined with osteoclasts formation results in alveolar bone loss. This is in contrast to an implant system with a lack of implant-abutment interface that shows little evidence of the presence of inflammatory cells. This infiltration of inflammatory substances is irrespective of the amount of plaque accumulation.

Comparison of Microleakage in Different Implant-Abutment Connections:

Verdugo CL et al. used external connection implant and conical internal connection (Morse taper) implants. The results of the study showed that less microleakage was shown by Morse taper connection implants than external connection implants. A gap of 10 μm was presented by external connection implants more than Morse taper implants with a gap of 2-3 μm.

Canullo L et al. conducted a five year follow-up study on humans for different implant connections under functional loading. The results showed that microbial contamination was seen in all the connections. Internal Hex and conical connection implants showed less leakage of bacteria at the peri-implant sulcus and inside the connection than external hexagon implants.

Do Nascimento C et al., in their in vitro study, used 43 microbial species, which were very common in the human oral cavity. They evaluated prostheses supported by External Hexagon or Morse Cone implants under dynamic loading conditions. Results revealed that higher microbial count was found in External Hexagon implants than Morse Cone implants. Many microbial species including, peri-implant diseases causing organisms were detected in internal part of External Hex implants. Internal surfaces of Morse Cone implants showed no colonization of microorganisms, as micro gaps present in conical connections were much smaller at the implant-abutment interface.

Baggi L et al. in their study, found that tube-in-tube interface implants were more resistant to colonization than flat to flat interface. Contradictory results were obtained by Al-Jadaa A et al. where they found, implants with a flat-to-flat interface (internal hexagonal) mating surfaces showed the best performance with regard to leakage under both static and dynamic conditions. This study also proved that if implants under static conditions were tight and would provide better sealing ability under dynamic conditions.

Koutouzis T et al. evaluated microleakage of internal Morse-taper connection and found that there was minimal penetration of bacteria down to the implant-abutment interface. Dynamic loading increases the penetration of bacteria as there was micro movement at the implant-abutment interface, which causes a pumping effect and leads to detrimental effects on marginal bone stability.13Contradictory results were obtained by Harder S et al. where conical implant-abutment connections do not prevent microleakage on a molecular level in even unloaded conditions.

From the studies mentioned above, it is possible to draw the inference that internal implant-abutment connections and conical (Morse Taper) implant-abutment connections show lesser microleakage than external implant-abutment connections. The external connections, of which the external hexagonal connection is the most common, provides a short and narrow connection with the abutment and provides for only limited screw engagement and a short fulcrum arm, which together allow for frequent screw loosening. The instability of the external connection leads to open the micro gaps.

Besides, it is often challenging to seat components on the hex easily and with confidence, especially in the posterior parts of the mouth, even for an experienced clinician. Minute rotational changes at a single abutment location can result in the misfit of the superstructure. Increasing the flat-to-flat width and the height of the connection has reduced these complications but not eliminated them. Internal connections, due to their design, have a more stable implant-abutment connection resistant to joint- opening forces.

Lateral forces are distributed deep within the implant, and the long internal wall engagement with the implant shields the abutment screw and buffers vibrations. Internal connections are therefore, less prone to micromovements and opening of the implant-abutment connection. Conical connections, specifically, have the advantage of the smallest inherent micro gaps due to frictional fit of the abutment into the implant, virtually removing any gap between the two. However, no implant-abutment interface can completely prevent microleakage in either loaded or unloaded states.

Position of the Implant-Abutment Connection in Relation to Crestal Bone:

Numerous studies have shown that bone resorption around the implant neck does not start until the implant is uncovered and exposed to the oral cavity. It is hypothesized that bacterial contamination of this micro gap during the second stage surgery results in peri-implant inflammation, leading to bone remodeling. Bone remodeling will progress until the biologic width has been created and stabilized to seal off the connective tissue from the colonized micro gap. Not only does this width progress apically, along the vertical axis, but according to studies conducted by Tarnow et al there is also a horizontal component amounting to 1–1.5 mm.


Fig. 2: Establishment of Biologic Width after second stage implant surgery

So, the position of the implant-abutment connection in relation to the crestal bone determines the amount of bone loss around an implant. A study by Broggini et al revealed that as the apical position of the implant-abutment interface was progressively increased, the total number of peri-implant inflammatory cells was increased in parallel, i.e., the deeper the interface, the greater the magnitude of peri-implant inflammation. Further, the maximum density of neutrophils adjacent to supracrestal implants was significantly less than for crestal and subcrestal implants. Moreover, the peri-implant location with maximum neutrophil density was also dependent upon the depth of the implant-abutment interface. Thus, for supracrestal implants, this location was near the implant-abutment interface (ie, above the original bone crest), whereas for subcrestal implants, this location was immediately coronal to the implant-abutment interface (below the original bone crest).

In parallel with differences in peri-implant inflammatory cell accumulation, the apico-coronal dimension of connective tissue was also progressively expanded as the depth of the implant-abutment interface was increased. This primarily reflected increases in the connective tissue compartment apical to the original alveolar crest (ie, alveolar bone loss). Specifically, there was significantly greater bone loss associated with subcrestal implants as compared to supracrestal position.

Therefore, this study demonstrated that moving the interface supracrestally, effectively changing the location of the inflammatory stimulus, also reduces peri-implant bone loss. Thus, minimal inflammation (and bone change) occurred when the interface was above the original bone crest, whereas the greatest inflammation (and bone loss) occurred when the interface was below the alveolar crest. These clinical observations are highly relevant, since the maintenance of crestal bone height appears to be an important predictor of soft tissue margins in both natural dentition and implants. These findings have several important clinical implications relative to limiting inflammation and bone loss around implants.

First, implant design could be either one-part or transmucosal to eliminate the interface. Second, the interface could be positioned supracrestally. Third, the interface might be made in such a way that excludes microbes, i.e. a more stable interface with smaller micro gaps. In these scenarios, inflammation would not be expected to develop near the alveolar crest, consequently reducing the potential for bony changes. Support for this speculation comes from another animal study in which implants were placed with the interface approximately 3mm above the original alveolar crest. Bone loss around these implants was minimal. Further, in patients with transmucosal implants placed so that the implant interface was approximately 3 mm above the original alveolar crest, minimal bone loss was observed over an eight-year period.


Platform Switching Concept:

Platform switching or platform shifting is a method used to preserve the alveolar bone around dental implants. The platform switching effect was first observed in the mid-1980s. At the time, larger-diameter implants were often restored with narrower abutments (Ankylos Dentsply, Friadent, Germany; Astra-Zeneca, Sweden; Bicon, Boston), as congruent abutments were often still unavailable. As it later turned out, this was a remarkable coincidence. The abutments used with conventional implant types are generally flush with the implant shoulder in the contact zone. This results in the formation of micro gaps between the implant and the abutment.

The bacterial contamination of these micro gaps adversely affects the stability of the peri-implant tissues and leads to a reduction of horizontal and vertical marginal bone levels. If the microcrack is located close to the bone, the creation of the biologic width will occur at the expense of the bone. The platform-switching concept requires that this microcrack be placed away from the implant shoulder and closer toward the axis in order to increase the distance of this microcrack from the bone. This generally implies the use of a reduced-diameter abutment. The inward, horizontal repositioning of the abutment inflammatory cell infiltrate (right) will move the abutment inflammatory cell infiltrate away from the crestal bone and into a more confined area.


There appear to be two results of the horizontal inward repositioning of the implant-abutment interface. First, with the increased surface area created by the exposed implant seating surface, there is a reduction in the amount of crestal bone resorption necessary to expose a minimum amount of implant surface to which the soft tissue can attach. Second, and perhaps more important, by repositioning the implant-abutment interface inward and away from the outer edge of the implant and adjacent bone, the overall effect of the inflammatory cell infiltrate on the surrounding tissue as described by Ericsson et al. and Abrahamsson et al. may be reduced, thus decreasing its resorptive effect on crestal bone. It is further suggested that platform switching repositions the inflammatory infiltrate further away from crestal bone and locates it within an approximate ≤90-degree confined area of exposure instead of a ≤180-degree area of direct exposure to the surrounding hard and soft tissues. As a consequence, the reduced exposure and confinement of the platform-switched abutment inflammatory cell infiltrate may result in a reduced inflammatory effect within the surrounding soft tissue and crestal bone.

It is important to note that to benefit from the platform-switching bone preservation technique, reduced-diameter components, beginning with the healing abutment, must be used from the moment that the implant is exposed to the oral environment, because the process of biologic width formation begins immediately following exposure to the oral environment. Thus, whether an implant is placed using a one- or two-stage surgical procedure, the first component placed on the implant must be of a smaller diameter if a horizontally repositioned biologic width is to be accomplished. This is important because after crestal bone has remodeled to a post-restorative resting position around the top of an implant, it will not return to its presurgical level if platform-switching principles are implemented at a later time.


The long-term predictability of dental implants is now a well-documented fact. Virtually all the major manufacturers can document success rates greater than 90%, and the more refined systems have achieved well above that number for more than ten years. A grey area has been the long-term stability of the abutment and prosthesis. Tremendous progress has been made in this area mainly, due to improved clinical machining tolerances. The transition to internal connections has been gradual but profound. The internal connections available today are far more stable, physically stronger, easier to restore, more amenable to excellent esthetics, and definitely more user-friendly.


1. Terheyden H, Lang NP, Bierbaum S, Stadlinger B. Osseointegration--communication of cells. Clin Oral Implants Res. 2012 Oct;23(10):1127–35.

2. Cochran DL, Nevins M. Biologic width: a physiologically and politically resilient structure. Int J Periodontics Restorative Dent. 2012 Aug;32(4):371–3.

3. Hermann JS, Buser D, Schenk RK, Higginbottom FL, Cochran DL. Biologic width around titanium implants. A physiologically formed and stable dimension over time. Clin Oral Implants Res. 2000 Feb;11(1):1–11.

4. Hermann JS, Cochran DL, Hermann JS, Buser D, Schenk RK, Schoolfield JD. Biologic Width around one- and two-piece titanium implants. A histometric evaluation of unloaded nonsubmerged and submerged implants in the canine mandible. Clin Oral Implants Res. 2001;12(6):559–71.

5. Linkevicius T, Apse P. Influence of abutment material on stability of peri-implant tissues: a systematic review. Int J Oral Maxillofac Implants. 2008 May;23(3):449–56.

6. Ericsson I, Nilner K, Klinge B, Glantz PO. Radiographical and histological characteristics of submerged and nonsubmerged titanium implants. An experimental study in the Labrador dog. Clin Oral Implants Res. 1996 Mar;7(1):20–6.

7. Ericsson I, Persson LG, Berglundh T, Marinello CP, Lindhe J, Klinge B. Different types of inflammatory reactions in peri-implant soft tissues. J Clin Periodontol. 1995 Mar;22(3):255–61.

8. Albrektsson T, Zarb G, Worthington P, Eriksson AR. The long-term efficacy of currently used dental implants: a review and proposed criteria of success. Int J Oral Maxillofac Implants. 1986 Summer;1(1):11–25.

9. Steinebrunner L, Wolfart S, Bössmann K, Kern M. In vitro evaluation of bacterial leakage along the implant-abutment interface of different implant systems. Int J Oral Maxillofac Implants. 2005 Nov;20(6):875–81.

10. Shareef N, Levine D. Effect of manufacturing tolerances on the micromotion at the Morse taper interface in modular hip implants using the finite element technique. Biomaterials. 1996 Mar;17(6):623–30.

11. do Nascimento C, Barbosa RES, Issa JPM, Watanabe E, Ito IY, Albuquerque RF. Bacterial leakage along the implant–abutment interface of premachined or cast components. Int J Oral Maxillofac Surg. 2008;37(2):177–80.

12. Teixeira W, Ribeiro RF, Sato S, Pedrazzi V. Microleakage into and from two-stage implants: an in vitro comparative study. Int J Oral Maxillofac Implants. 2011 Jan;26(1):56–62.

13. Hermann JS, Schoolfield JD, Schenk RK, Buser D, Cochran DL. Influence of the size of the microgap on crestal bone changes around titanium implants. A histometric evaluation of unloaded non-submerged implants in the canine mandible. J Periodontol. 2001 Oct;72(10):1372–83.

14. Byrne D, Houston F, Cleary R, Claffey N. The fit of cast and premachined implant abutments. J Prosthet Dent. 1998 Aug;80(2):184–92.

15. Jansen VK, Conrads G, Richter EJ. Microbial leakage and marginal fit of the implant-abutment interface. Int J Oral Maxillofac Implants. 1997 Jul;12(4):527–40.

16. Ricomini Filho AP, Fernandes FS de F, Straioto FG, da Silva WJ, Del Bel Cury AA. Preload loss and bacterial penetration on different implant-abutment connection systems. Braz Dent J. 2010;21(2):123–9.

17. Rimondini L, Marin C, Brunella F, Fini M. Internal Contamination of a 2-Component Implant System After Occlusal Loading and Provisionally Luted Reconstruction With or Without a Washer Device. J Periodontol. 2001;72(12):1652–7.

18. Dias ECL de C e. M, Bisognin EDC, Harari ND, Machado SJ, da Silva CP, Soares GD de A, et al. Evaluation of implant-abutment microgap and bacterial leakage in five external-hex implant systems: an in vitro study. Int J Oral Maxillofac Implants. 2012 Mar;27(2):346–51.

19. Quirynen M, Alsaadi G, Pauwels M, Haffajee A, van Steenberghe D, Naert I. Microbiological and clinical outcomes and patient satisfaction for two treatment options in the edentulous lower jaw after 10 years of function. Clin Oral Implants Res. 2005 Jun;16(3):277–87.

20. Quirynen M, Bollen CM, Eyssen H, van Steenberghe D. Microbial penetration along the implant components of the Brånemark system. An in vitro study. Clin Oral Implants Res. 1994 Dec;5(4):239–44.

21. Gross M, Abramovich I, Weiss EI. Microleakage at the abutment-implant interface of osseointegrated implants: a comparative study. Int J Oral Maxillofac Implants. 1999 Jan;14(1):94–100.

22. Piattelli A, Scarano A, Paolantonio M, Assenza B, Leghissa GC, Di Bonaventura G, et al. Fluids and microbial penetration in the internal part of cement-retained versus screw-retained implant-abutment connections. J Periodontol. 2001 Sep;72(9):1146–50.

23. Hamilton A, Judge RB, Palamara JE, Evans C. Evaluation of the fit of CAD/CAM abutments. Int J Prosthodont. 2013 Jul;26(4):370–80.

24. Sumi T, Braian M, Shimada A, Shibata N, Takeshita K, Vandeweghe S, et al. Characteristics of implant-CAD/CAM abutment connections of two different internal connection systems. J Oral Rehabil. 2012 May;39(5):391–8.

25. Harder S, Quabius ES, Ossenkop L, Kern M. Assessment of lipopolysaccharide microleakage at conical implant-abutment connections. Clin Oral Investig. 2012 Oct;16(5):1377–84.

26. do Nascimento C, Miani PK, Watanabe E, Pedrazzi V, de Albuqerque RF Jr. In vitro evaluation of bacterial leakage along the implant-abutment interface of an external-hex implant after saliva incubation. Int J Oral Maxillofac Implants. 2011 Jul;26(4):782–7.

27. Esposito M, Hirsch JM, Lekholm U, Thomsen P. Biological factors contributing to failures of osseointegrated oral implants. (II). Etiopathogenesis. Eur J Oral Sci. 1998 Jun;106(3):721–64.

28. Quirynen M, Vogels R, Peeters W, van Steenberghe D, Naert I, Haffajee A. Dynamics of initial subgingival colonization of “pristine” peri-implant pockets. Clin Oral Implants Res. 2005;17(1):25–37.

29. Broggini N. Periimplant Inflammation Defined by the Implant-abutment Interface. 2003. 42 p.

30. Broggini N, McManus LM, Hermann JS, Medina RU, Oates TW, Schenk RK, et al. Persistent acute inflammation at the implant-abutment interface. J Dent Res. 2003 Mar;82(3):232–7.

31. Mawhinney J, Connolly E, Claffey N, Moran G, Polyzois I. An in vivo comparison of internal bacterial colonization in two dental implant systems: Identification of a pathogenic reservoir. Acta Odontol Scand. 2014;73(3):188–94.

32. Larrucea Verdugo C, Jaramillo Núñez G, Acevedo Avila A, Larrucea San Martín C. Microleakage of the prosthetic abutment/implant interface with internal and external connection: in vitro study. Clin Oral Implants Res. 2014 Sep;25(9):1078–83.

33. Canullo L, Penarrocha-Oltra D, Soldini C, Mazzocco F, Penarrocha M, Covani U. Microbiological assessment of the implant-abutment interface in different connections: cross-sectional study after 5 years of functional loading. Clin Oral Implants Res. 2015 Apr;26(4):426–34.

34. do Nascimento C, Ikeda LN, Pita MS, Rafael Cândido Pedroso, Pedrazzi V, de Albuquerque Junior RF, et al. Marginal fit and microbial leakage along the implant-abutment interface of fixed partial prostheses: An in vitro analysis using Checkerboard DNA-DNA hybridization. J Prosthet Dent. 2015;114(6):831–8.

35. Baggi L, Di Girolamo M, Mirisola C, Calcaterra R. Microbiological evaluation of bacterial and mycotic seal in implant systems with different implant-abutment interfaces and closing torque values. Implant Dent. 2013 Aug;22(4):344–50.

36. Al-Jadaa A, Attin T, Peltomäki T, Heumann C, Schmidlin PR. Impact of Dynamic Loading on the Implant-abutment Interface Using a Gas-enhanced Permeation Test In Vitro. Open Dent J. 2015 Mar 31;9:112–9.

37. Koutouzis T, Mesia R, Calderon N, Wong F, Wallet S. The Effect of Dynamic Loading on Bacterial Colonization of the Dental Implant Fixture–Abutment Interface: An In Vitro Study. J Oral Implantol. 2014;40(4):432–7.

38. Tarnow DP, Cho SC, Wallace SS. The effect of inter-implant distance on the height of inter-implant bone crest. J Periodontol. 2000 Apr;71(4):546–9.

39. Broggini N, McManus LM, Hermann JS, Medina R, Schenk RK, Buser D, et al. Peri-implant inflammation defined by the implant-abutment interface. J Dent Res. 2006 May;85(5):473–8.

40. Gargiulo AW, Wentz FM, Orban B. Dimensions and Relations of the Dentogingival Junction in Humans. J Periodontol. 1961;32(3):261–7.

41. Tarnow DP, Magner AW, Fletcher P. The effect of the distance from the contact point to the crest of bone on the presence or absence of the interproximal dental papilla. J Periodontol. 1992 Dec;63(12):995–6.

42. Hermann JS, Buser D, Schenk RK, Schoolfield JD, Cochran DL. Biologic Width around one- and two-piece titanium implants. Clin Oral Implants Res. 2001 Dec;12(6):559–71.

43. Hermann JS, Cochran DL, Nummikoski PV, Buser D. Crestal bone changes around titanium implants. A radiographic evaluation of unloaded nonsubmerged and submerged implants in the canine mandible. J Periodontol. 1997 Nov;68(11):1117–30.

44. Buser D, Mericske-Stern R, Dula K, Lang NP. Clinical experience with one-stage, non-submerged dental implants. Adv Dent Res. 1999 Jun;13:153–61.

45. Hermann F, Lerner H, Palti A. Factors influencing the preservation of the periimplant marginal bone. Implant Dent. 2007 Jun;16(2):165–75.

46. Maynard JG Jr, Wilson RD. Physiologic dimensions of the periodontium significant to the restorative dentist. J Periodontol. 1979 Apr;50(4):170–4.

47. Gardner DM. Platform switching as a means to achieving implant esthetics. N Y State Dent J. 2005 Apr;71(3):34–7.

48. Lazzara RJ, Porter SS. Platform switching: a new concept in implant dentistry for controlling postrestorative crestal bone levels. Int J Periodontics Restorative Dent. 2006 Feb;26(1):9–17.

49. Baumgarten H, Cocchetto R, Testori T, Meltzer A, Porter S. A new implant design for crestal bone preservation: initial observations and case report. Pract Proced Aesthet Dent. 2005 Nov;17(10):735–40.

50. Abrahamsson I, Berglundh T, Lindhe J. Soft tissue response to plaque formation at different implant systems. A comparative study in the dog. Clin Oral Implants Res. 1998 Apr;9(2):73–9.

51. Abrahamsson I, Berglundh T, Lindhe J. The mucosal barrier following abutment dis/reconnection. An experimental study in dogs. J Clin Periodontol. 1997 Aug;24(8):568–72.