crestal bone changes around titanium implants. a ... · methods: a total of 59 titanium implants...

Volume 71 • Number 9

Crestal Bone Changes Around TitaniumImplants. A Histometric Evaluation ofUnloaded Non-Submerged and SubmergedImplants in the Canine MandibleJoachim S. Hermann,*† Daniel Buser,‡ Robert K. Schenk,‡§ and David L. Cochran*

1412

Background: Today, implants are placed using both non-submerged and submerged approaches, and in 1-and 2-piece configurations. Previous work has demonstrated that peri-implant crestal bone reactions differ radi-ographically under such conditions and are dependent on a rough/smooth implant border in 1-piece implantsand on the location of the interface (microgap) between the implant and abutment/restoration in 2-piece con-figurations. The purpose of this investigation was to examine histometrically crestal bone changes aroundunloaded non-submerged and submerged 1- and 2-piece titanium implants in a side-by-side comparison.

Methods: A total of 59 titanium implants were randomly placed in edentulous mandibular areas of 5 fox-hounds, forming 6 different implant subgroups (types A-F). In general, all implants had a relatively smooth,machined coronal portion as well as a rough, sandblasted and acid-etched (SLA) apical portion. Implant typesA-C were placed in a non-submerged approach, while types D-F were inserted in a submerged fashion. TypeA and B implants were 1-piece implants with the rough/smooth border (r/s) at the alveolar crest (type A) or1.0 mm below (type B). Type C implants had an abutment placed at the time of surgery with the interfacelocated at the bone crest level. In the submerged group, types D-F, the interface was located either at the bonecrest level (type D), 1 mm above (type E), or 1 mm below (type F). Three months after implant placement,abutment connection was performed in the submerged implant groups. At 6 months, all animals were sacri-ficed. Non-decalcified histology was analyzed by evaluating peri-implant crestal bone levels.

Results: For types A and B, mean crestal bone levels were located adjacent (within 0.20 mm) to therough/smooth border (r/s). For type C implants, the mean distance (± standard deviation) between the inter-face and the crestal bone level was 1.68 mm (± 0.19 mm) with an r/s border to first bone-to-implant contact(fBIC) of 0.39 mm (± 0.23 mm); for type D, 1.57 mm (± 0.22 mm) with an r/s border to fBIC of 0.28 mm(± 0.21 mm); for type E, 2.64 mm (± 0.24 mm) with an r/s border to fBIC of 0.06 mm (± 0.27 mm); and fortype F, 1.25 mm (± 0.40 mm) with an r/s border to fBIC of 0.89 mm (± 0.41 mm).

Conclusions: The location of a rough/smooth border on the surface of non-submerged 1-piece implantsplaced at the bone crest level or 1 mm below, respectively, determines the level of the fBIC. In all 2-pieceimplants, however, the location of the interface (microgap), when located at or below the alveolar crest, deter-mines the amount of crestal bone resorption. If the same interface is located 1 mm coronal to the alveolarcrest, the fBIC is located at the r/s border. These findings, as evaluated by non-decalcified histology underunloaded conditions, demonstrate that crestal bone changes occur during the early phase of healing afterimplant placement. Furthermore, these changes are dependent on the surface characteristics of the implantand the presence/absence as well as the location of an interface (microgap). Crestal bone changes were notdependent on the surgical technique (submerged or non-submerged). J Periodontol 2000;71:1412-1424.

KEY WORDSAlveolar bone/anatomy and histology; dental implants; bone resorption/etiology; titanium.

* Department of Periodontics, Dental School, University of Texas Health Science Center at San Antonio, TX.† Department of Prosthodontics, University of Basel Dental School, Basel, Switzerland.‡ Department of Oral Surgery, University of Bern, School of Dental Medicine, Bern, Switzerland.§ University of Bern, Institute of Pathophysiology.

9204_IPC_AAP_553300 9/11/00 1:06 PM 412

J Periodontol • September 2000 Hermann, Buser, Schenk, Cochran

The placement of endosseous dental implants hasevolved using 2 surgical approaches.1,2 One ofthese approaches uses a submerged technique

where the implant is placed into the bone and the topof the implant is placed at or below the alveolar crest.3

The soft tissues are closed over the bone and implantthereby submerging the implant. The other surgicalapproach involves an implant that extends coronallybeyond the alveolar crest where the soft tissue flap isplaced around the coronal portion of the implant body.4

The placement of additional restorative componentswith the submerged implant creates an interface orgap (often termed a “microgap”) at or below the alve-olar crest. In the non-submerged approach, whenrestorative components are placed, the interface islocated 2 to 3 mm coronal to the alveolar crest, whichis even with or slightly below the gingival level. Addi-tionally, such 1-piece implants generally exhibit an“apical” or endosseous portion with a roughenedimplant surface and a “coronal” or transgingival por-tion with a relatively smooth surface.

Longitudinal descriptions of the use of both sub-merged and non-submerged techniques in humanshave demonstrated that the implants can be clinicallysuccessful in a large percentage of cases over extendedperiods of time.5-8 In the clinical descriptions ofendosseous implant use in humans, one criterion ofsuccess has typically included radiographic alveolarcrestal bone changes. This criterion is a mean crestalbone loss of less than 1.5 mm in the first year afterabutment connection and less than 0.2 mm in subse-quent years.9,10 This criterion was established usingthe submerged approach and a machined screwimplant where an interface was created at or belowthe alveolar crest at the uncovering (second-stage)surgery. For 1-piece implants placed in a non-sub-merged approach, several long-term studies indicatethat the level of crestal bone remodels to approxi-mately the border between the rough and smoothimplant surfaces.11-13 A detailed radiographic evalu-ation of approximately 100 such implants revealedthat more implants gained bone over an 8-year periodcompared to the number of implants that lost crestalbone.14 The authors concluded that mean bone levelchanges are not as informative as is a distributionanalysis of bone changes.

One inherent advantage of a non-submerged ap-proach is the fact that only one surgical procedure hasto be performed. Therefore, some investigators havebegun to attach secondary components on submergedimplants at the time of implant placement.15-21 Thisrepresents a non-submerged approach; however, aninterface (microgap) remains at or below the alveolarcrest, similar to the traditional submerged approach.In this situation, the interface is created earlier in theprocedure at implant placement rather than after a

submerged healing period for the implant. Radiographicanalyses of such 2-piece non-submerged implants,however, reveal crestal bone loss changes similar to thetraditional submerged approach, except that thechanges are observed earlier.22

The purpose of the present investigation was to eval-uate crestal bone changes by histological analysesaround submerged and non-submerged implantsplaced in a side-by-side comparison. Additionally, 1-and 2-piece implant/abutment systems were studiedto determine the influence of the location of the inter-face (microgap) in 2-piece systems or the rough/smooth border on 1-piece implant systems in relationto the first bone-to-implant contact.

MATERIALS AND METHODSImplant SurfacesA full-body screw design was chosen for all 6 differ-ent cylindrical titanium implants (A-F) which weremade from cold-worked, grade IV commercially puretitanium� (Fig. 1). The total length was 9 mm, and the

1413

� Institut Straumann AG, Waldenburg/BL, Switzerland.

Figure 1A.Schematic (true to scale) of implants type A-C at time of implantplacement in relation to the crestal bone level.The solid black linedelineates the border between rough and smooth implant surfaces,whereas the broken line shows the location of the interface (microgap)where appropriate.

Figure 1B.Schematic (true to scale) of implants type D-F at time of implantplacement in relation to the crestal bone level.

9204_IPC_AAP_553300 9/11/00 1:06 PM Page 1413

Crestal Bone Around Non-Submerged and Submerged Implants Volume 71 • Number 9

olar crest. For type B implants, the rough part (SLA)measured 5.0 mm, with the rough/smooth border placed1.0 mm below the alveolar crest. For all other implants(types C-F), the rough implant surface (SLA) was 4.5mm in vertical dimension, with the rough/smoothimplant border located 1.5 mm below the crest. TypesA and B were 1-piece implants, meaning there wasno interface (microgap) present; implant types C-Fconsisted of 2 pieces, with an interface or microgapof about 50 µm in size between the implant and thesecondary component, the abutment. The location ofthe interface was defined to be at the bone crest levelfor types C and D; for types E and F, the interface waslocated 1 mm above and 1 mm below the crest,respectively. Implant types A-C were placed accord-ing to a non-submerged approach and types D-F wereplaced using a submerged technique.

Power AnalysisOne of each of the 6 types of implants could be placedon each side of the dog mandible. To ensure that posi-tion in the arch did not influence the results, thesequence of placement was randomized for the leftand right sides of the mandible in each dog. Becausethe goal of the study was to detect implant group dif-ferences of 1 mm or more, a sample size of 60implants was confirmed by power analysis to be suf-ficient to identify significant implant group differencesusing analysis of variance at the 0.01 level with powerof 80%. Five dogs, each with 12 implants placed, wereused in the study. Ten different sequences of implanttype placement were selected and randomly assignedto the left and right sides of the mandible of each dog.

AnimalsThe protocol was approved by the Institutional Ani-mal Care and Use Committee of the University of TexasHealth Science Center at San Antonio. Five lab-bred,male American foxhounds were used. The dogs wereapproximately 2 years of age at the beginning of thestudy and had a body weight of about 30 to 35 kg. Noheartworms were found and all dogs were quarantined.

Surgeries Extraction. Extractions were performed in an oper-ating room under intravenous (iv) general anesthe-sia and sterile conditions utilizing 4% thiopenthal-Nasolution (0.4 ml/kg bw) for premedication purposes(Fig. 4).¶ The animals were put on a heating pad, intu-bated and inhaled with 1.5 to 2% isoflurane,# and mon-itored with an electrocardiogram during the surgery.First, the surgical site was disinfected with 10% povi-done-iodine solution/1% titratable iodine.** Subse-

1414

Figure 2.Scanning electron micrograph (SEM) of a relatively smooth, machinedsurface (bar = 20 µm; original magnification ×2,100).

Figure 3.Scanning electron micrograph (SEM) of a sandblasted (large-grit) and acid-etched surface (SLA) (bar = 20 µm; original magnification×2,100).

¶ Pentothal, Abbott Laboratories, North Chicago, IL.# Ærane, Ohmeda Carbide Inc., Liberty Corner, NJ.** Clinidine, Clinipad Co., Guilford, CT.

body diameter 3.5 mm. The threads added 0.6 mm inthe apical end of the implant so, with threads, the outerdiameter measured 4.1 mm. The relatively smooth,coronal portion of each implant consisted of amachined titanium surface (Fig. 2) with an outer diam-eter of 3.5 mm. Each abutment with a machined sur-face was also 3.5 mm in diameter, which matched thediameter of the body of the implant without threads.The rough, apical part of each implant had a sand-blasted (large-grit) and HCl/H2SO4 acid-attacked sur-face (SLA) exhibiting 2 levels of roughness, one at 20to 40 µm peak to peak, and a superimposed secondlevel at 2 to 4 µm peak to peak (Fig. 3). For type Aimplants, the SLA surface was 6.0 mm in length,revealing the rough/smooth implant border at the alve-

9204_IPC_AAP_553300 9/11/00 1:06 PM Page 1414


quently, 2% lidocaine HCl with epinephrine 1:100,000was given as local anesthetic,†† and all 4 premolars(P1-P4) and the first molar (M1) were extracted care-fully. Before extraction, all teeth were scaled andcleaned, and P2-M1 were sectioned to help preventtooth fracture. Finally, interrupted sutures were usedallowing for an appropriate approximation of the woundmargins.

On the day of surgery, the animals were given 20mg of the analgesic nalbuphine subcutaneously (sc)bid (10 mg/ml).‡‡ In addition, after a period of 7 to 10days, the dogs received 3 ml of the antibiotic benzathinepenicillin (150,000 I.U.) combined with procaine peni-cillin G (150,000 I.U.) sc sid every 48 hours.§§ The ani-mals were then briefly anesthetized with a combination(1.1 ml/15 kg bw) of xylazine (7.1 mg/ml),�� acepro-mazine (2.1 mg/ml),¶¶ atropine (0.1 mg/ml),¶¶ and ket-amine (50.0 mg/ml) iv.## After anesthesia, sutures wereremoved following disinfection of the wound site with a0.12% chlorhexidine digluconate-soaked gauze.***

Implant placement. After a healing period of 6months (Fig. 4), non-submerged and submergedimplants were placed under the same surgical condi-tions as when tooth extraction had been performed(operating room, anesthesia, sterility). A crestal inci-sion was performed to maximize keratinized gingiva oneach side of the incision. Consequently, mucoperiostealflaps were reflected on the lingual and buccal aspect.Foramina mentalia were dissected and exposed. Usingan acrylic bur, the edentulous osseous ridge was care-fully flattened and combined with copious irrigationwith chilled sterile physiologic saline. A Boley gaugewas used to help distribute 6 test implants on eachside of the mandible. Implant site preparations werecarried out with torque reduction rotary instrumentsat 500 rpm using chilled saline. Implant types A-C

were inserted according to a non-submerged approach; i.e., for typeC, implants and abutments werescrewed together at the time offirst-stage surgery. Implant typesD-F were placed according to asubmerged technique. Finally, oneof each kind of test implant wasinserted per side in a randomizedfashion. Thus, no implant typehad a biased position in the arch.

If indicated, periosteal relievingand contouring incisions were car-ried out on the buccal and lingualaspects of each implant to obtaintension-free adaptation of thewound margins for close adapta-tion of the gingiva to the transgin-gival portion of the nonsubmergedimplants (types A and B) and the

abutment of type C. Wound closure over the submergedimplant types D-F was achieved using horizontal mat-tress and interrupted sutures. The animals received 20mg of the analgesic nalbuphine sc bid (10 mg/ml) onthe day of surgery. Three ml of the antibiotic benza-thine penicillin (150,000 I.U.) with procaine penicillinG (150,000 I.U.) was given sc sid every 48 hours for14 days. On day 1, 100 mg of the antibiotic genta-micin was given sc bid, and the same amount sid fromday 2 through 10 (50 mg/ml).††† The animals received2 ml of the anti-inflammatory dexamethasone inter-muscularly (im) sid day 1 and day 4 (2 mg/ml).‡‡‡

Suture removal was performed after 7 to 10 days asdescribed above. To minimize loading, the dogs werefed a softened diet. Three times per week, mechanicaland chemical plaque control were carried out utilizinga soft toothbrush and a soft sponge in combination witha 0.2% chlorhexidine gel.§§§

Abutment connection. Three months after implantplacement, second-stage surgery was carried out andabutments were connected for submerged implanttypes D-F. Surgical conditions were the same asdescribed above. First, the surgical sites were disin-fected and local anesthesia administered. Over the topof these implants, a midcrestal incision was performed,combined with a small vertical relieving incision at thebuccal and lingual aspect. Implants were uncoveredafter the elevation of a full-thickness flap. In the case

1415

†† Henry Schein Inc., Port Washington, NY.‡‡ Nubain, Astra Pharmaceutical Products Inc., Westborough, MA.§§ Pen-B, Pfizer Inc., Lee’s Summit, MO.�� Miles Inc., Shawnee Mission, KS.¶¶ Burns Veterinary Supply, Oakland, CA.## Fort Dodge Laboratories Inc., Fort Dodge, IA.*** Peridex, Procter & Gamble Co., Cincinnati, OH.††† Gentocin, Schering-Plough Animal Health Corp., Kenilworth, NJ.‡‡‡ Dexaject, Burns Veterinary Supply.§§§ PlakOut Gel, Hawe-Neos AG, Bioggio/TI, Switzerland.

Figure 4.Study design.

9204_IPC_AAP_553300 9/11/00 1:06 PM Page 1415


of implants partially covered with bone, mostly in typeF implants, a minor osteotomy was performed usinghand instruments (chisel, mallet). Consequently, flat-head cover screws could be removed in the submergedimplant group. Abutments of individual lengths wereconnected specifically for each implant type, meaningthat all implants emerged to the same level after abut-ment placement. Interrupted sutures combined withsmall V-shaped gingivectomies were used for woundclosure around the abutments. Postoperative care andsuture removal were performed the same way as aftertooth extractions. Abutments (types C-F) were dis-connected and immediately tightened afterwards at 4,8, and 10 weeks after second-stage surgery to imitatethe placement of another healing abutment, impressiontaking, as well as the placement of the final prostheticcomponent.

Three months after abutment connection, all dogswere sacrificed. Euthanasia was carried out with anoverdose of pentobarbital sodium iv (0.2 ml = 65mg/kg bw).� � � The block-resection of the mandibleswas performed using an oscillating autopsy saw.¶¶¶

The recovered segments with the implants wereimmersed in a solution of 4% formaldehyde combinedwith 1% CaCl2 for histologic preparation and analysis.

Non-Decalcified Histologic AnalysisPreparation. Each implant with surrounding tissueswas prepared for non-decalcified histology.23 Speci-mens were embedded in methyl methacrylate andstained with basic fuchsin. One to 3 mesio-distal andup to 5 orofacial sections at a thickness of approxi-mately 80 µm were obtained to maximize the availabledata per implant.

Histometry. Histometric quantification was carriedout using a light microscope at different magnifica-tions to best locate anatomical reference points.###

The microscope was connected to a high-resolutionvideo camera**** and interfaced to a monitor†††† aswell as a personal computer.‡‡‡‡ This optical systemwas associated with a digitizing pad and a bone his-tometry software package with image capturing capa-bilities.§§§§ All implants had a rough/smooth border(r/s) at a particular level in relation to the crest of thebone. Due to the manufacturing process, however,there was a very small zone where only acid attackoccurred just apical to the machined surface. About 0.1mm more apical, the sandblasted surface combinedwith the acid-etched surface could be detected. Thislevel, where both sandblasting and acid attackoccurred, was defined as the rough/smooth border forthe histologic evaluation. Finally, the following mea-surements were performed at each implant site (Fig. 5):

1. Distance between the top of the implant (top)and the first bone-to-implant contact (fBIC) for implanttypes A and B.

2. Distance between the interface (microgap) of theimplant (IF) and the first bone-to-implant contact(fBIC) for implant types C through F.

3. Distance between the top of the implant (top)and the rough/smooth border (r/s) for implant typesA and B.

4. Distance between the interface (microgap) of theimplant (IF) and the rough/smooth border (r/s) forimplant types C through F.

Statistical AnalysisOf particular interest in this study was the determina-tion of the position of the first bone-to-implant contact(fBIC) relative to the location of the rough/smooth (r/s)border. This was done by initially obtaining histomet-ric measurements of the distance from the top of theimplant (types A and B), or the interface (types C-F),respectively, to the fBIC (Tables 1 and 2). Subse-quently, histometric measurements from the top orinterface to the r/s border were made on each implantsite individually. Each implant had 4 or 5 sections with2 sites for which histometric data were collected. Inorder to verify that the IF:fBIC values obtained from thehistometric evaluation were not influenced by exam-

1416

�� Euthanasia-5 Solution, Henry Schein Inc.¶¶¶ Stryker Co., Kalamazoo, MI.### Vanox-T, Olympus, Tokyo, Japan.**** CCD-Iris color video camera, Sony Corp., Fujisawa, Japan.†††† Multisync XV17+, NEC, Itasca, IL.‡‡‡‡ Vectra VL, Hewlett Packard, Palo Alto, CA.§§§§ Image-Pro Plus, Media Cybernetics, Silver Spring, MD.

Figure 5.Composite schematic (not true to scale) of histometric evaluation withthe following measurements: distance between the top of the implant(types A and B) or the interface (IF; types C-F), respectively, to the first bone-to-implant contact (fBIC) as well as the rough/smoothborder (r/s).

9204_IPC_AAP_553300 9/11/00 1:06 PM Page 1416


also large in a histometric sense, as the buccal sites haddistances that exceeded mesial sites by an average of0.28 mm per implant and both lingual and distal sitesby an average of 0.38 mm per implant. Furthermore,these differences were definitely related to the positionof the bone for buccal sites, as evidenced by the factthat the measured distances from the top or interfaceof the implant to the r/s border were not significantlydifferent for the 4 types of sites (P >0.5), so that dis-tortion among sections and sites was quite small. Theseresults indicate that only the lingual, mesial, and distalsites should be used to calculate mean values for eachimplant from the top or interface to the fBIC in thisstudy. For the purposes of consistency, buccal siteswere also excluded in the calculation of mean valuesfor each implant from the top or interface to the r/sborder. In addition, the difference between these 2 mea-sures was also calculated for each site (buccal sitesexcluded), and the mean difference for each implantwas determined. If either measure was unreadable fora site, then that site was excluded from the analysis. Themean implant measures were then analyzed using F-tests to verify that the effects of dog and position in thearch for the implant were not significant (P >0.05). One-sample Student t tests were performed for each implanttype to determine if the implant type mean (obtainedfrom the mean implant measures) of the distance fromthe r/s border to the fBIC was significantly different from0 mm. F-tests were then performed to check for sig-nificant (P <0.05) differences across implant types. Fordimensions in which the F-test was significant, Bon-ferroni-adjusted unpaired Student t tests were carriedout to identify group differences.

RESULTSClinical DataFifty-nine out of the possible 60 implants could beplaced. One implant could not be placed since thebone at this site was too soft and, therefore, primarystability could not be achieved. All the implants placedin this study achieved clinical stability and no com-plications occurred during healing or during the fol-low-up period. A previous publication of the same dataset based on monthly radiographic evaluations demon-strated that no peri-implant radiolucencies were foundaround any of the implants and that only crestal boneloss was noted around specific implant types.22 Thus,these implants achieved excellent hard tissue integra-tion by clinical and radiographic analyses. Althoughboth meticulous mechanical and chemical plaque con-trol were carried out 3 times per week, different degreesof peri-implant inflammation could be detected. TypeA and B implants (1-piece, non-submerged) presentedwith only minimal clinical signs of peri-implant inflam-mation. However, type C implants (2-piece, non-sub-merged) as well as type D-F implants (2-piece, sub-

1417

Table 1.

Histometric Data for Non-SubmergedImplants

Type A Type B Type CVariables n = 9 (42) n = 10 (47) n = 10 (50)

IF: fBIC 2.98 ± 0.27 * 3.88 ± 0.42 † 1.68 ± 0.19 ‡

IF: r/s 2.79 ± 0.12 3.87 ± 0.22 1.29 ± 0.06

Mean values ± standard deviation (mm); n = number of implants analyzed;( ) = number of measured implant sites; note that for groups A and B, thereference point (IF) is the top of the implant.* P <0.02.† P >0.9 (not significant).‡ P <0.001.

Table 2.

Histometric Data for Submerged Implants

Type D Type E Type FVariables n = 10 (47) n = 9 (48) n = 10 (49)

IF: fBIC 1.57 ± 0.22 * 2.64 ± 0.24 † 1.25 ± 0.40 ‡

IF: r/s 1.29 ± 0.10 2.58 ± 0.15 0.36 ± 0.06

Mean values ± standard deviation (mm); n = number of implants analyzed;( ) = number of measured implant sites.* P <0.01.† P >0.4 (not significant).‡ P <0.001.

iner bias, the primary examiner carried out 2 mea-surements, as did a second examiner, for 6 implants.Each of the 6 implant types and each of the 5 dogswere represented in the examiner verification analysis.The 6 implants that were evaluated had a total of 51sites for which IF:fBIC could be measured. For the pri-mary examiner, the 2 readings differed by less than0.10 mm for 44 (86.3%) of the 51 sites, and the max-imum difference was 0.21 mm. For the other exam-iner, the 2 readings differed by less than 0.10 mm for49 (96.1%) of the 51 sites, and the maximum differ-ence was 0.13 mm. When all 4 readings were com-pared collectively, they differed by less than 0.10 mmfor 36 (70.6%) of the 51 sites, and 49 (96.1%) differedby less than 0.20 mm, and the maximum differencewas 0.39 mm.

After excluding sites that were unreadable, a meanhistometric dimensional value was calculated for eachimplant. However, the fBIC for buccal sites tended to belower than that for the lingual, mesial, or distal sites. Infact, the mean distance between the top or interface ofthe implant and the fBIC was significantly greater forbuccal sites compared to any of the other three sitesfor all implants combined (P <0.001 for Bonferroni-adjusted paired Student t test). These differences were

9204_IPC_AAP_553300 9/11/00 1:06 PM Page 1417


merged) showed moderate to severe clinical signs ofperi-implantitis. These particular results will bedescribed in a subsequent publication.

Histometric AnalysisLight microscopic evaluation of the bone-to-implantcontact indicates that hard tissue integration of theimplants was achieved as well (Figs. 6 through 11).In all implant types, intimate contact of bone was founddirectly adjacent to the sandblasted and acid-etched(SLA) surface. As expected, dense cortical bone hadlarge areas of bone-to-implant contact compared tocancellous bone areas where more marrow space wasfound. It can be noted, however, that in more cancel-lous bone, osseous tissue was found along the SLAsurface demonstrating and confirming the excellentosteoconductive nature of this surface.24-37

Nine 1-piece, non-submerged implants (type A)were analyzed in 34 histological sections allowing 42implant site measurements (Fig. 1A). The mean dis-tance from the top of the implant to the fBIC was 2.98mm (Table 1, Fig. 12A). The mean distance betweenthe rough/smooth border (r/s) to the fBIC was 0.19mm, which was marginally statistically different (P<0.02). This indicated that the fBIC was found approx-imately 0.2 mm below the r/s border which had beenplaced clinically as close to the alveolar crest as pos-sible.

For the 1-piece, non-submerged type B implantswith the r/s border located approximately 1 mm belowthe crest (Fig. 1A), the mean distance between thetop of the implant and the fBIC was 3.88 mm, whilethe distance of the r/s border to the fBIC was 0.01

1418

Figure 6.Mesio-distal section of a type A implant (1-piece, non-submerged).Basic fuchsin stain; original magnification ×2.5; original inner/outerimplant diameter = 3.5 mm/4.1 mm.The original cover screw wasremoved and replaced with a small diameter screw to facilitatehistologic processing.

Figure 7.Mesio-distal section of a type B implant (1-piece, non-submerged).Basic fuchsin stain; original magnification ×2.5; original inner/outerimplant diameter = 3.5 mm/4.1 mm.The original cover screw wasremoved and replaced with a small diameter screw to facilitatehistologic processing.

9204_IPC_AAP_553300 9/11/00 1:06 PM Page 1418


mm (Table 1, Fig. 12A). The fBIC was not significantlydifferent from the location of the r/s border, indicatingthat in the case of type B implants, the bone waslocated at the r/s border. In this group, 10 implantswith 36 sections (47 readable sites) could be analyzed.

The third type of implant (type C) placed in a non-submerged approach had the r/s border located clin-ically approximately 1.5 mm apical to the crest ofbone and differed from type B by having an interface(microgap) located clinically at the alveolar crest(Fig. 1A). The mean distance based on 50 mea-surements from the interface between the implant andabutment to the fBIC was 1.68 mm (Table 1, Fig. 12A).This was based on analyzing 10 implants with 41 sec-tions. The r/s border to fBIC was 0.39 mm, which wasstatistically significantly different (P <0.001).

Type D implants were exactly like type C implants

(2-piece, level of r/s border, interface placed clinicallyat the alveolar crest); however, they were placed in asubmerged approach (Fig. 1B). Thus, the implant wasplaced such that the top of the implant was located atthe alveolar crest, and the soft tissue flap covered theimplant. After 3 months of healing, the implant wasuncovered and an abutment was placed such thatimplant types C and D were now identical, with theexception being the time of exposure to the oral cavity(nonsubmerged versus submerged approach). At thetime of sacrifice for type D implants, the mean distancefrom the interface to the fBIC was 1.57 mm. The meandistance of the r/s border to the fBIC was 0.28 mm,which indicated that the location of the fBIC was sig-nificantly different (P <0.01) from the r/s border (Table2, Fig. 12B). These measurements were based on 10implants with 42 sections and 52 measurable sites.

1419

Figure 8.Mesio-distal section of a type C implant (2-piece, non-submerged).Basic fuchsin stain; original magnification ×2.5; original inner/outerimplant diameter = 3.5 mm/4.1 mm.The abutment has beenremoved and replaced with a small diameter screw to facilitatehistologic processing.

Figure 9.Mesio-distal section of a type D implant (2-piece, submerged). Basicfuchsin stain; original magnification ×2.5; original inner/outer implantdiameter = 3.5 mm/4.1 mm.The abutment has been removed andreplaced with a small diameter screw to facilitate histologic processing.

9204_IPC_AAP_553300 9/11/00 1:06 PM Page 1419


Type E implants were 2-piece implants placed in asubmerged technique, but had an extension of theendosseous portion projecting clinically, approximately1.0 mm above the alveolar crest. This placed the inter-face (microgap) above the bone crest level. The r/s bor-der, however, was similar to types B, C, D, and F in thatit was located apical to the alveolar crest (Figs. 1A and1B). For one of the type E implants, the fBIC could notbe observed on any of the 5 sections taken. The remain-ing 9 implants with 39 sections and 48 measurable siteswere analyzed. The mean distance from the interface tothe fBIC was 2.64 mm with an r/s border to fBIC dis-tance of 0.06 mm (Table 2, Fig. 12B). The r/s borderand fBIC measurements were not significantly different(P >0.5) indicating that, similar to type B implants, thebone was located at the r/s border (Table 2).

Type F implants were placed with a submerged tech-

1420

Figure 10.Mesio-distal section of a type E implant (2-piece, submerged). Basicfuchsin stain; original magnification ×2.5; original inner/outer implantdiameter = 3.5 mm/4.1 mm.The abutment has been removed andreplaced with a small diameter screw to facilitate histologic processing.

Figure 11.Mesio-distal section of a type F implant (2-piece, submerged). Basicfuchsin stain; original magnification ×2.5; original inner/outer implantdiameter = 3.5 mm/4.1 mm.The abutment has been removed andreplaced with a small diameter screw to facilitate histologic processing.

nique also, but the top of the implant was placed clin-ically approximately 1.0 mm below the alveolar crestand the r/s border about 0.5 mm apical to the inter-face/microgap (Fig. 1B). Ten implants with 40 sec-tions and 52 measurable sites were analyzed. Themean distance from the interface to the fBIC was 1.25mm, with a distance of 0.89 mm between the r/s bor-der and the fBIC (Table 2, Fig. 12B). This latter distancewas significantly (P <0.001) apical to the location of ther/s border, indicating that crestal bone loss occurredpast the interface clinically located 1.0 mm below thealveolar crest and approximately 1.0 mm apical to ther/s border (located apical to the interface/microgap).

A comparison of the amount of crestal bone lossamong the 6 implant types revealed that the greatestbone loss (2.25 mm) was observed around type Fimplants (Tables 1 and 2, Figs. 12A and 12B), which

9204_IPC_AAP_553300 9/11/00 1:07 PM Page 1420


the change occurs. An important aspect of this studyis that all 6 implant types were experimental implantswith no differences in shape and that the implantswere randomized in a side-by-side comparison. Cre-stal bone changes that occurred correlated with eachspecific implant type irrespective of its position withinthe arch, left or right side, or the individual dog. Thecrestal bone loss that occurred around implant type Ccompared to type B was due only to the difference inthe implants, the existence of an interface (microgap).

Many studies have demonstrated that bone tissuefavors rough implant surfaces compared to relativelysmooth titanium surfaces.24-37 The advantage of therough surface reported in these studies has recentlybeen summarized.38 Based on prior studies, it wasexpected that the first bone-to-implant contact wouldoccur at the rough/smooth border on the type A as wellas type B implants. However, for type A implants, avery small difference, statistically only marginally sig-nificant, existed between the r/s border and the fBIC.This may be due to the fact that type A implants hadbeen placed with their r/s border even with the crestof the bone, meaning that the r/s border could just stillbe observed. The histometric results (at time of sac-rifice) revealed that the crestal bone did not movecoronally about two-tenths of a millimeter. These find-ings are further supported by the fact that the fBICoccurred exactly at the r/s border in type B implants.In addition, this small difference was about within thelimits of the precision of the histometric method applied(intra/interexaminer evaluation).

Type B and type C implants were virtually identicalin regards to both shape and location of therough/smooth border. The main difference betweenthese 2 implant types was the presence of an interface(microgap) which was located at the alveolar crest.Thus, the significant bone loss that occurred beyondthe rough/smooth border around type C implants(while none was associated with type B implantsbeyond this border) was due to the presence of theinterface (microgap). Both of these implants wereplaced in a non-submerged technique, indicating thatthe bone loss was related to the presence of the inter-face (microgap) and not due to a surgical technique(nonsubmerged versus submerged). This finding takeson clinical significance as investigators are reportingthe use of 2-piece implant systems (typically sub-merged implants plus an abutment) placed in a non-submerged approach.15-21 These findings demonstratethat submerged implants placed with an abutment atthe time of implant placement will also experience sig-nificant crestal bone resorption.

Type D implants were identical in all respects(shape, location of the r/s border, and location of theinterface/microgap) to type C implants except that thetype D implant was placed with a submerged surgical

1421

Figure 12A.Schematic (true to scale) of hard tissues around non-submergedimplants type A-C at time of sacrifice in relation to the rough/smoothborder (solid black line) as well as the location of the interface(microgap; broken black line) where appropriate.

Figure 12B.Schematic (true to scale) of hard tissues around submerged implantstype D-F at time of sacrifice in relation to the rough/smooth border(solid black line) as well as the location of the interface (microgap;broken black line).

was significantly different (P <0.01) from the other 5implant types. F-tests comparing implant positionwithin the arch (P = 0.17) and among dogs (P = 0.09)were not significant.

DISCUSSIONThe results from these histologic analyses demon-strate that the first bone-to-implant contact (fBIC)around a 2-piece endosseous dental implant is depen-dent on the location of an interface (microgap)between the implant and the abutment/restoration(types C-F). In the absence of an interface (a 1-pieceendosseous dental implant; types A and B), the firstbone-to-implant contact is dependent on the borderbetween a rough and smooth implant surface (r/s bor-der). The fact that the implant is placed in a sub-merged or a non-submerged approach does not alterthe magnitude of the change, only the time in which

9204_IPC_AAP_553300 9/11/00 1:07 PM Page 1421


technique where osseous healing could occur prior toan uncovering surgery and the placement of an abut-ment (and creation of an interface/microgap). Thus,the only difference between type C and D implantswas in the timing of exposure to the oral cavity. Sig-nificant crestal bone loss occurred around type Dimplants, similar in amount to type C implants. Thispattern of crestal bone loss is common for 2-piecesubmerged implants, and therefore has been wellaccepted as one major criterion of success utilizingthis approach.8-9 However, contradictory reports fromexperimental studies have been published in the 1990swhere different 2-piece submerged implants have beeninvestigated in animal studies not experiencing thesame amount of crestal bone loss as observed clini-cally in many previous studies.39,40 An important detailseems to be the fact that, according to the protocol inthe present study, abutments on all 2-piece implants(types C-F) were disconnected and immediately tight-ened, trying to imitate clinically relevant steps like theexchange of a healing abutment, impression taking,as well as the insertion of the superstructure. Thesesteps were not carried out in the 2 above mentionedstudies;39,40 however, once the same research grouprepeated their experiments, including the disconnec-tion/tightening of abutments, the same amount of cre-stal bone loss could be observed and, thus, the resultsof this study were confirmed.41 Overall, these resultssuggest that differences occurring in crestal bone lossdo not depend on whether the implant is placed in asubmerged or non-submerged technique, but on thedifferences in timing of implant exposure (and inter-face/microgap creation) to the oral cavity.

This experiment was designed to determine if thepresence of an interface (microgap) had an influenceon crestal bone resorption. It was reasoned that if suchan effect was in fact due to the presence of an interface,then, if the interface was moved coronally away fromthe alveolar crest, less bone loss would take place; sim-ilarly, if the interface was moved apical to the alveolarcrest, greater amounts of bone loss would occur. Implanttypes E and F represent these conditions, respectively.Interestingly, moving the interface (microgap) coro-nally 1 mm with these implants, combined with therough/smooth border located 1.5 mm apical to the alve-olar crest, resulted in the first bone-to-implant contactat the rough/smooth border. This was identical to whatoccurred around type B implants. Thus, this combina-tion of interface (microgap) and rough/smooth borderlocation resulted in crestal bone changes around this 2-piece implant system similar to a 1-piece system. Theseresults reinforce the lack of effect of the specific surgi-cal technique (type B was placed in a non-submergedapproach, type E in a submerged approach) and thesignificance of the location of the interface (microgap)and r/s border.

The largest amount of bone loss was associatedwith type F implants, which was significantly greaterthan that found around any of the other 5 implanttypes. This finding was expected, based on the premisethat if bone loss was associated with an interface(microgap), then placement of that interface in a loca-tion apical to the alveolar crest would result in thegreatest amount of bone loss. This finding further sup-ports the fact that bone loss is associated with an inter-face (microgap) and has significant clinical implica-tions, particularly in two indications. Some clinicianshave advocated the placement of submerged (2-piece)implants well below (approximately 2 to 3 mm) thealveolar crest in the maxillary anterior sextant in orderto have an esthetic implant-borne restoration.42-45 Theresults of the present study demonstrate that signifi-cant crestal bone loss can occur around these typesof restorations. Secondly, in the posterior sextants inthe mouth, limited bone height may be available forimplant placement due to the location of the mandibu-lar canal/nerve and the maxillary sinus. Often times,in fact, bone grafting procedures are performed inthese areas to increase the amount of bone availablefor implant placement. In these cases, placement of a2-piece endosseous implant at or below the alveolarcrest would result in a significant amount of crestalbone loss, and thus counteract the bone augmenta-tion procedure and/or increase the crown-to-implantratio in an area of the mouth where the greatest amountof occlusal force is found. Thus, in these clinical situ-ations, based upon the 6 implant types used in thisstudy, a one-piece implant designed with a rough/smoothborder at the alveolar crest should be used to ensurethe least amount of crestal bone loss.

In conclusion, this experimental study in the caninemandible compared 6 implant designs placed in 2surgical approaches: submerged and non-submerged.Both 1-piece and 2-piece implants (with an inter-face/microgap) were included with variations in thelocation of the interface (microgap) and r/s border.The results demonstrate that significant amounts ofcrestal bone loss occur around 2-piece (typically sub-merged) implant designs depending on the locationof the interface. Secondly, the location of a rough/smooth border on a 1-piece (typically non-submerged)implant has an influence on the first bone-to-implantcontact. Lastly, the surgical technique of submergingor not submerging the implant has no influence onthe amount of crestal bone loss that occurs. More sig-nificantly, the creation as well as the location of aninterface (microgap) influences crestal bone loss andthe first bone-to-implant contact.

ACKNOWLEDGMENTSThe authors would like to thank Patricia Navarro fortyping the manuscript and Sonja A. Bustamante, H.T.

1422

9204_IPC_AAP_553300 9/11/00 1:07 PM Page 1422


(ASCP), for her assistance throughout the study. Wegreatly appreciate the valuable help of Dr. Richard J.Haines, clinical veterinarian, and his team for excellentcare of the animals. Drs. James P. Simpson and DanielSnétivy, Institut Straumann AG, are acknowledged formanufacturing the test implants and for providing thescanning electron micrographs (SEMs). In addition,we would like to express our gratitude to John D.Schoolfield for helping with the statistical analysis. Thisstudy has been supported by grant 02-94/061 from theITI-Foundation for the Promotion of Oral Implantology,Waldenburg/BL, Switzerland; the Swiss Society of Peri-odontology, Bern; the Swiss National Science Foun-dation, Basel; and the Swiss Foundation for Medicaland Biological Stipends, Bern.

REFERENCES1. Cochran DL, Mahn DH. Dental implants and regenera-

tion. Part I. Overview and biological considerations(chapter 59). In: Hardin J, ed. Clark’s Clinical Dentistry,vol. 5. Philadelphia: JB Lippincott Company; 1992:1-7.

2. Mahn DH, Cochran DL. Dental implants and regenera-tion. Part II. Clinical considerations (chapter 59A). In:Hardin J, ed. Clark’s Clinical Dentistry, vol. 5. Philadel-phia: JB Lippincott Company; 1992:1-11.

3. Brånemark PI, Adell R, Breine U, Hansson BO, Lind-strom J, Ohlsson A. Intra-osseous anchorage of dentalprostheses. I. Experimental studies. Scand J Plast Recon-str Surg 1969;3:81-100.

4. Schroeder A, van der Zypen E, Stich H, Sutter F. Thereactions of bone, connective tissue, and epithelium toendosteal implants with titanium-sprayed surfaces. JMaxillofac Surg 1981;9:15-25.

5. Cochran DL. Implant therapy. I. Ann Periodontol 1996;1:707-790.

6. Buser D, Mericske-Stern R, Bernard JP, et al. Long-termevaluation of nonsubmerged ITI implants. Part 1: 8-yearlife table analysis of a prospective multi-center studywith 2,359 implants. Clin Oral Implants Res 1997;8:161-172.

7. Adell R, Lekholm U, Rockler B, Brånemark PI. A 15-year study of osseointegrated implants in the treatmentof the edentulous jaw. Int J Oral Surg 1981;10:387-416.

8. Worthington P, Bolender CL, Taylor TD. The Swedishsystem of osseointegrated implants: Problems and com-plications encountered during a 4-year trial period. IntJ Oral Maxillofac Implants 1987;2:77-84.

9. Albrektsson T, Zarb G, Worthington P, Eriksson AR. Thelong-term efficacy of currently used dental implants: Areview and proposed criteria of success. Int J Oral Max-illofac Implants 1986;1:11-25.

10. Smith DE, Zarb GA. Criteria for success of osseointe-grated endosseous implants. J Prosthet Dent 1989;62:567-572.

11. Buser D, Weber HP, Brägger U, Balsiger C. Tissue inte-gration of one-stage ITI implants: 3-year results of a lon-gitudinal study with hollow-cylinder and hollow-screwimplants. Int J Oral Maxillofac Implants 1991;6:405-412.

12. Weber HP, Buser D, Fiorellini JP, Williams RC. Radi-ographic evaluation of crestal bone levels adjacent tononsubmerged titanium implants. Clin Oral Implants Res1992;3:181-188.

13. Pham AN, Fiorellini JP, Paquette D, Williams RC, WeberHP. Longitudinal radiographic study of crestal bone lev-

els adjacent to nonsubmerged dental implants. J OralImplantol 1994;20:26-34.

14. Buser D, Mericske-Stern R, Dula K, Lang NP. Clinicalexperience with one-stage, nonsubmerged titanium den-tal implants. Adv Dent Res 1999;13:153-161.

15. Ericsson I, Randow K, Glantz PO, Lindhe J, Nilner K.Clinical and radiographical features of submerged andnonsubmerged titanium implants. Clin Oral Implants Res1994;5:185-189.

16. Ericsson I, Randow K, Nilner K, Petersson A. Some clin-ical and radiographical features of submerged and non-submerged titanium implants. A 5-year follow-up study.Clin Oral Implants Res 1997;8:422-426.

17. Collaert B, De Bruyn H. Comparison of Brånemark fix-ture integration and short-term survival using one-stageor two-stage surgery in completely and partially eden-tulous mandibles. Clin Oral Implants Res 1998;9:131-135.

18. Bernard JP, Belser UC, Martinet JP, Borgis SA. Osseoin-tegration of Brånemark fixtures using a single-step oper-ating technique. A preliminary prospective one-yearstudy in the edentulous mandible. Clin Oral ImplantsRes 1995;6:122-129.

19. Ericsson I, Nilner K, Klinge B, Glantz PO. Radiographi-cal and histological characteristics of submerged andnonsubmerged titanium implants. An experimental studyin the Labrador dog. Clin Oral Implants Res 1996;7:20-26.

20. Levy D, Deporter DA, Pilliar RM, Watson PA, ValiquetteN. Initial healing in the dog of submerged versus non-submerged porous-coated endosseous dental implants.Clin Oral Implants Res 1996;7:101-110.

21. Becker W, Becker BE, Israelson H, et al. One-step sur-gical placement of Brånemark implants: A prospectivemulticenter clinical study. Int J Oral Maxillofac Implants1997;12:454-462.

22. Hermann JS, Cochran DL, Nummikoski PV, Buser D.Crestal bone changes around titanium implants. A radi-ographic evaluation of unloaded nonsubmerged and sub-merged implants in the canine mandible. J Periodontol1997;68:1117-1130.

23. Schenk RK, Olah AJ, Herrmann W. Preparation of cal-cified tissues for light microscopy. In: Dickson GR, ed.Methods of Calcified Tissue Preparation, vol. 1. Amster-dam: Elsevier Science Publishers BV; 1984:1-56.

24. Claes L, Hutzschenreuter P, Pohler O. Relationshipbetween removal torque values of leg screws, time ofimplant placement, and surface texture (in German).Arch Orthop Unfall-Chir 1976;85:155-159.

25. Steinemann SG, Eulenberger J, Mäusli PA, SchroederA. Adhesion of bone to titanium. In: Christel P, MeunierA, Lee AJC, eds. Advances in Biomaterials - Biologicaland Biomechanical Performance of Biomaterials, vol. 6.Amsterdam: Elsevier Science Publishers BV; 1986:409-414.

26. Wilke HJ, Claes L, Steinemann S. The influence of var-ious titanium surfaces on the interface shear strengthbetween implants and bone. In: Heimke G, Soltész U, LeeAJC, eds. Advances in Biomaterials-Clinical ImplantMaterials, vol. 9. Amsterdam: Elsevier Science Publish-ers BV; 1990:309-314.

27. Buser D, Schenk RK, Steinemann S, Fiorellini JP, FoxCH, Stich H. Influence of surface characteristics on boneintegration of titanium implants. A histomorphometricstudy in miniature pigs. J Biomed Mater Res 1991;25:889-902.

28. Cochran DL, Simpson JP, Weber H-P, Buser D. Attach-

1423

9204_IPC_AAP_553300 9/11/00 1:07 PM Page 1423


ment and growth of periodontal cells on smooth andrough titanium. Int J Oral Maxillofac Implants 1994;9:289-297.

29. Wong M, Eulenberger J, Schenk R, Hunziker E. Effectof surface topology on the osseointegration of implantmaterials in trabecular bone. J Biomed Mater Res 1995;29:1567-1575.

30. Cochran DL, Nummikoski PV, Higginbottom FL, Her-mann JS, Makins SR, Buser D. Evaluation of anendosseous titanium implant with a sandblasted andacid-etched surface in the canine mandible: Radiographicresults. Clin Oral Implants Res 1996;7:240-252.

31. Kieswetter K, Schwartz Z, Hummert TW, et al. Surfaceroughness modulates the local production of growth fac-tors and cytokines by osteoblast-like MG-63 cells. J Bio-med Mater Res 1996;32:55-63.

32. Martin JY, Dean DD, Cochran DL, Simpson J, Boyan BD,Schwartz Z. Proliferation, differentiation, and protein syn-thesis of human osteoblast-like cells (MG63) culturedon previously used titanium surfaces. Clin Oral ImplantsRes 1996;7:27-37.

33. Schwartz Z, Martin JY, Dean DD, Simpson J, CochranDL, Boyan BD. Effect of titanium surface roughness onchondrocyte proliferation, matrix production, and dif-ferentiation depends on the state of cell maturation. JBiomed Mater Res 1996;30:145-155.

34. Buser D, Nydegger T, Hirt HP, Cochran DL, Nolte LP.Removal torque values of titanium implants in the max-illa of miniature pigs. Int J Oral Maxillofac Implants1998;13:611-619.

35. Cochran DL, Schenk RK, Lussi A, Higginbottom FL,Buser D. Bone response to unloaded and loaded titaniumimplants with a sandblasted and acid-etched surface: Ahistometric study in the canine mandible. J Biomed MaterRes 1998;40:1-11.

36. Buser D, Nydegger T, Oxland T, et al. Interface shearstrength of titanium implants with a sandblasted andacid-etched surface: A biomechanical study in the max-illa of miniature pigs. J Biomed Mater Res 1999;45:75-83.

37. Cochran DL, Schenk R, Buser D, Wozney JM, JonesAA. Recombinant human bone morphogenetic protein-2 stimulation of bone formation around endosseous den-tal implants. J Periodontol 1999;70:139-150.

38. Cochran DL. Endosseous dental implant surfaces inhuman clinical trials. A comparison using meta-analy-sis. J Periodontol 1999;70:1523-1539.

39. Berglundh T, Lindhe J, Ericsson I, Marinello CP, Liljen-berg B, Thomsen P. The soft tissue barrier at implantsand teeth. Clin Oral Implants Res 1991;2:81-90.

40. Abrahamsson I, Berglundh T, Wennström J, Lindhe J.The peri-implant hard and soft tissues at different implantsystems. A comparative study in the dog. Clin OralImplants Res 1996;7:212-219.

41. Abrahamsson I, Berglundh T, Lindhe J. The mucosalbarrier following abutment dis/reconnection. An exper-imental study in dogs. J Clin Periodontol 1997;24:568-572.

42. Saadoun AP, Sullivan DY, Krischek M, Le Gall M. Sin-gle tooth implant - management for success. Pract Peri-odontics Aesthetic Dent 1994;6:73-80.

43. Palacci P. Implant placement. In: Palacci P, Ericsson I,Engstrand P, Rangert B, eds. Optimal Implant Position-ing & Soft Tissue Management for the Brånemark Sys-tem, vol. 1. Berlin: Quintessence Publishing Co., Inc.;1995:35-39.

44. Spiekermann H. Single-tooth implants. In: RateitschakKH, Wolf HF, eds. Color Atlas of Dental Medicine - Implan-tology, vol. 1. New York: Thieme Medical Publishers,Inc.; 1995:267-298.

45. Nevins M, Stein JM. The placement of maxillary ante-rior implants. In: Nevins M, Mellonig JT, eds. ImplantTherapy - Clinical Approaches and Evidence of Success,vol. 2. Chicago: Quintessence Publishing Co., Inc.;1998:111-127.

Send reprint requests to: Dr. David L. Cochran, Universityof Texas Health Science Center at San Antonio, DentalSchool, Department of Periodontics, 7703 Floyd Curl Drive,San Antonio, TX 78284-7894. Fax: 210/567-6299; e-mail:[email protected]

Accepted for publication February 22, 2000.

1424

9204_IPC_AAP_553300 9/11/00 1:07 PM Page 1424

crestal bone changes around titanium implants. a ... · methods: a total of 59 titanium implants...

Documents