Norbert EnklingPeter [obienVictoria KlimbergStefan BayerRegina Mericske-SternSeren Jepsen

Authors' affiliations:Norbert Enkling, Victoria Klimberg, Regina Mericske-Stem, Department of Prosthodontics, University ofBerne, Berne, SwitzerlandPeter lohien, Department of Oral Surgery, University ofWitten/Herdecke, Witten, GermanyStefan Bayer, Department of Prosthetic Dentistry,Propaedeutics and Material Science, University ofBonn, Bonn, GermanySeren Jepsen, Department of Periodontology, Operativeand Preventive Dentistry, University of Bonn, Bonn,Germany

Correspondingauthor:Norbert EnklingDepartment of ProsthodonticsSchool of Dental MedicineUniversity of BernFreiburgstrasse 73010 BernSwitzerlandTel.: +41316328705Fax: +41316324933e-mail: norbert.enkling@zmk.unibe.ch

Date:Accepted 15 September 2010

To cite this article:Enkling N, Iohren P, Klirnberg V, Bayer S, Mericske-StemR, Jepsen S. Effect of platform switching on peri-implantbone levels: a randomized clinical trial.CIin. Oral Impl. Res. 22, 20I1; 1185-II92.

doi: 10.1 I I r/j.I6oo·050I.20IO.02090.X

© 2011 John Wiley & Sons A/S

Effect of platform switching 01~

peri-implant bone levels: a randomizedclinical trial

Key words: bacteria, bone level, bone loss, dental implants, implant design, platform switching

AbstractObjective: The concept of platform switching has been introduced to implant dentistry based on

observations of reduced peri-implant bone loss. However, randomized clinical trials are still lacking.This study aimed to test the hypothesis that platform switching has a positive impact on crestalbone-level changes.Material and methods: Two implants with diameters of 4 mm were inserted epicrestally into one sideof the posterior mandibles of 25 subjects. After 3 months of submerged healing, the reentry surgerywas performed. On the randomly placed test implant, an abutment 3.3 mm in diameter was mounted,resulting in a horizontal circular step of 0.35 mm (platform switching). The control implant wasstraight, with an abutment 4 mm in diameter. Single-tooth crowns were cemented provisionally.

All patients were monitored at short intervals over the course of 1year. Standardized radiographs andmicrobiological samples from the implants' inner spaceswere obtained at baseline (implant surgery),and after 3, 4, and 12 months.Results: After 1 year, the mean radiographic vertical bone loss at the test implants was0.53 ± 0.35 mm and at the control implants, it was 0.58 ± 0.55 mm. The mean intraindividualdifference was 0.05 ± 0.56 mm, which is significantly <0.35 mm (P= 0.0093, post hoc power79.9%). The crestal bone-level changes depended on time (P<O.OOl), but not on platform switching(P= 0.4). The implants' internal spaceswere contaminated by bacteria, with no significant differencesin the total counts between the test and the control at any time point (P= 0.98).Conclusions: The present randomized clinical trial could not confirm the hypothesis of a reducedperi-implant bone loss at implants restored according to the concept of platform switching.

Marginal crestal bone loss at implants is oftenattributed to a microbial impact (Linclhe& Meyle2008). It appears that with two-piece implants,an abutment-associated inflammatory cell infil-trate (ICT) forms around the bacterially contami-nated micro-gap between the implant and theabutment and extends from the micro-gap cor-onally, apically, and laterally for 0.5-0.6 mm(Ericsson et al. 1995, 1996). In animal studies,it was shown that by positioning the micro-gapmore apically, i.e., below the crestal bone, theabutment ICT is established subcrestally follow-ing bone resorption (Hermann et al. 2000).

The concept of platform switching suggests anabutment or a suprastructure with a diameter atthe implant-platform level that is smaller thanthe implant diameter. This configuration resultsin a circular horizontal step, which enables ahorizontal extension of the biological width.The rationale for platform switching is to locatethe micro-gap of the implant-abutment connec-

tion away from the vertical bone-to-implantcontact area. Compared with the conventionalrestorative procedure using an identical size im-plant and suprastructure diameter, platformswitching is suggested to prevent or reduce crestalbone loss (Gardner 2005; Broggini et al. 2006;

Lazzara & Porter 2006; Prosper et al. 2009).So far, the results on platform switching are

controversial, but most clinical studies havereported a positive impact of platform switchingon crestal bone stability. The reduction in boneloss appears to be correlated with the size of thecircular step. In a prospective clinical studyinvolving 69 implants in 31 patients, Canulloand colleagues found a bone loss of 1.49 mm atimplants with matching abutments, 0.99 mm atimplants with a 0.2 mm circular step, 0.82 mmwith a 0.5 mm step, and 0.56 mm with a0.85 mm step. Thus, the mean positive impacton bone resorption 33 months after implantoperation was larger in size than the circular

II8S

Enklinget al. PlatformswitchingRCT

step of platform switching (Canullo et al. 2010).This is in agreement with the results of Hurzeleret al. (2007) and Vela-Nebot et al. (2006), whoreported that by circular steps of 0-45 and0.5 mm, respectively, the implants with platformswitching exhibited 1.8mm less bone loss thanthe implants without platform switching within12 and 6 months, respectively, after implantoperation. Also, Capiello et al. (2008) studied131 implants in 45 patients and found a bone-protective effect of 0.72mm that was larger insize than the circular step of 0.4 mm. In contrastto these results, Vigolo et al. (2009) found asignificantly different bone resorption pattern ina clinical study with 182 implants in 144 pa-tients with a 60-month follow-up after crownmounting, with platform-switching implantsshowing 0.3 mm less crestal bone loss than im-plants with matching abutments; however, thecircular step was 0.45 mm and, therefore, largerthan the difference in bone remodeling. Also,Trammel et al. (2009) reported a statisticallysignificant bone-protective effect of platformswitching of 0.2 mm, which was smaller thanthe circular step, for 25 implants in IO patientsafter a 24-month follow-up. Crespi and collea-gues, however, did not find any differences increstal bone loss when comparing platformswitching and nonplatform switching. Theyused different implant systems and performedimmediate implant placement and immediateloading (Crespi et al. 2009). This was in agree-ment with two animal studies that did not showSignificantly different bone resorption patterns atimplants with platform switching compared withimplants and matching abutments (Weng et al.2008; Becker et a1. 2009). To date, data fromRCTs comparing marginal bone-level alterationson two adjacent implants restored with and with-out platform switching, exhibiting an identicaldiameter (4mm) and the same design character-istics, have been missing.

Therefore, it was the aim of this randomizedtrial to test the hypothesis that platform switch-ing would significantly reduce peri-implant boneloss during the first year after implant placement.

Material and methods

Experimental designA prospective, single-blind, controlled clinicaltrial was designed. Twenty-five patients ran-domly received two implants in the posteriormandible, which were restored with (test) orwithout (control) platfonn switching after a sub-

merged healing phase of 3 months. Patients wereexamined and data were collected at five timepoints, with the time of surgery as the baseline.Standardized radiographs and microbiological

I r86 ) Clin. Oral Impl. Res. :U, 20II / II85-II92

samples of the implants' internal spaces wereobtained at each time point, except month 8.

SubjectsTwenty-five patients (10 females, IS males, age5I ± ro·5 years) were recruited at the DentalClinic Bochum/University of Witten-Herdecke,Germany, for the study. Prerequisites for inclu-sion in the study were good general health andabsence of infectious disease, diabetes, and osteo-pathy. Other requirements included no activeperiodontitis, no drugs influencing bone metabo-lism, no lactating or pregnant women, an eden-tulous gap in the posterior mandible for theplacement of two implants, sufficient boneheight above the alveolar nerve, bone width forthe placement of implants 9.5mm in length and4 mm in diameter, 4 mm of keratinized mucosain the prospective implant position in the bucco-lingual direction, and a medium or a thick soft-tissue biotype.

The patients were informed in detail about thepossible risks and benefits, and all signed aninformed consent. The study was performed incompliance with Good Clinical Practice and theDeclaration of Helsinki last revised Edinburgh2000; the study protocol was reviewed and ap-proved by the Clinical Trials Committee of theUniversity of Witten/Herdecke, Witten, Ger-many.

HVD:-:, ,, ,, ,-_ ....•..••...._ ......•.......•--_., I, ,

•••••-~I-o •••.••...••.:,

ImplantsThe implant system used in the present studywas the SIC Ace'" implant (SIC-Invent, Basel,Switzerland). It has an internal hex connectionwith an interlocking clearance fit (Zipprich et al.2007) and a medium-rough sand-blasted, acid-etched surface. All implants used in the studyhad a length of 9.5mm and a diameter of 4mm.The standard SICAceeo implants are designed forplatform switching, with a 45° beveled platformswitching shoulder, resulting in a circular step of0.35 mm (Fig. r a]. The corresponding abutmenthas a diameter of 3.3mm. The control implantwas manufactured without the bevel, but with anidentical internal configuration for abutmentconnection. The abutment with a 4mm dia-meter could be used (Fig. rb],

Radiographic examination and evaluationThe radiographic assessments were based on fourstandardized panoramic images taken immedi-ately after implant surgery (baseline), immedi-ately before reentry (3 months post-op],immediately after crown insertion (4 monthspost-op], and at the second recall (12 monthspost-op). For standardization, the patients' mand-ibles were fixated by a customized bite splint,and the panorex unit was readjusted individuallyto the respective patient position. The digitalorthopantomographs were obtained with the

Fig. I. (a) Study results after I year: bone-level alterations at the implant with platform switching (Test). (b) Study results afterI year: bone-level alterations at the implant without platform switching (Control). The red area demonstrates the bone-levelalteration since baseline (implant insertion operation); measured distances from reference points: IBL, vertical bone level at theimplant, reference point: micro-gap HVD, horizontal aspect of the vertical bony defect at the implant, reference point: implantsurface GBL, general horizontal bone level; reference point: micro-gap.

© 20II John Wiley & Sons A/S

Promax RPX 232574 (Planmeca, Helsinki, Fin-land) and evaluated and measured using theDimaxis Software 4.3.1 [Planmeca] with a mea-suring precision of 0.01 mm. The regions ofinterest on the radiographic images were magni-fied using the software tools to the highestpossible level ( x 20), and the bone height mea-surements were always calibrated at the respec-tive implant length of 9.5 mm.

The marginal bone levels at the mesial anddistal implant surfaces were assessed by measur-ing the distance between a reference point at theimplant and the bone level: the reference pointfor the vertical implant bone level (IBL)and forthe general horizontal bone level (GBL)was themicro-gap between implant body and abutment,and for the horizontal extent of the vertical defect(HVD), the implant shoulder adjacent tothe crestal edge of the vertical bony defect. Ateach implant, six measurements were performed(Fig. ra and b): IBL:A, mesial, B, distal; HVD: C,mesial, D, distal; GBL:E, mesial, F, distal. HVDand VVD are horizontal and vertical measuringunits to describe angular bony defects, wherebyGBL is the horizontal level at which the angulardefect begins. Three calibrated dentists experi-enced in oral radiology performed the radio-graphic evaluation independently, resulting in atotal of 3600 measurements. If the differences inmeasurements among the three examiners were0.1 mm or less, the mean of the three measure-ments was used. If the differences were>o. I mm, the three examiners re-analyzed thespecific implant together, and consensus wassought.

Clinical procedure. randomization. andallocation concealmentIn January 2007, after a promotion of the study inthe local press, more than 200 persons werescreened as potential study subjects (Fig. 2).Twenty-five patients were selected according tothe inclusion and exclusion criteria; the subjectsgave their informed consent and were enrolled inthe study. In April 2007, all implant surgerieswere performed within I week at the DentalClinic Bochum/University of Witten-Herdecke,Germany. After a terminal injection with localanesthetics (Ultracain UDS Forte, epinephrineI: 100,000, Sanofi-Aventis, Paris, France), acrestal incision was made, a full-thickness flapwas raised, the surgical site was exposed, and theimplants were inserted at the bone level. Theminimal distance between the two implants was3mm and between the tooth and the implant, it

o 3 4recall 2 (treatment)surgery reentry I crowns

Fig. 2. Time schedule of the study.

© 20II John Wiley & Sons A/S

was 2 mm. Surgical bone recontouring of thealveolar bone was not allowed. For each patient,the locations of the test and control implantswere randomized according to a computer-gener-ated list. The implants healed in a submergedposition. In July 2007, 3 months after the im-plant operation, the implants were exposed afterelevating a mini full-thickness flap. After im-plant insertion and after second-stage surgery, thetest persons rinsed with a 0.2 % chlorhexidinegluconate mouthwash (Meridol Perio, GABA,Therwil, Switzerland) twice daily for I min forI week until removal of the sutures. After 2

weeks, impressions were taken; after anotherweek, a try-in of the casted crown frameworkswas performed. After 7 days (4 months after theimplant operation), fully ceramic-veneered castedsingle crowns were produced, SIC Ace® standardtitanium abutments (SICNo. 936163) were tigh-tened with 25 N em on the implants, and thecrowns were mounted on the abutments withprovisional cement. Follow-up appointmentswere scheduled at 8 and 12 months post-implantinsertion for the assessment of oral and implanthealth and hygiene based on the full-mouthsulcular bleeding index [Muhlemann & Son1971), full-mouth plaque index (Silness & Loe1964), peri-implant probing depth, and sulcularflow rate. Moreover, at the IS appointments ofthe study, the oral hygiene of the test subjectswas reinforced by hygiene reinstruction or, ifnecessary, by professional plaque control.

Microbiological analysisMicrobiological samples were harvested from theimplant intemal cavity at surgery (baseline), atsecond-stage surgery (3 months), at placement ofthe suprastructure (4 months), and at the secondrecall (I2 months). At the second recall, thecrowns and abutments had to be removed. Afterharvesting the microbiological samples from theimplants' inner cavities, the abutments weretightened, and the crowns were again mountedwith provisional cement. All microbiologicalsamples were collected in a standardized way bythe same investigator. Samples were taken usingthree consecutively inserted sterile paper points(ISO #90, Roeko, Langenau, Germany), whichwere inserted into the internal cavity of theimplant for 20 s and then immediately transferredinto a sterile transport tube for evaluation byreal-time PCR [Carpegen" Perio Diagnostics,Carpegen GmbH, Munster, Germany) for deter-mination of the total bacterial counts (TBC)(Jervoe-Storm et al. 2005). Before sampling,

8 12 (months)

recall 1

Enklinget al PlatformswitchingRCT

selected implants and their adjacent regionswere isolated with cotton rolls; great care wastaken to avoid any kind of contamination of theimplants during removal of the internal screw.

Statistical methodsAccording to the sample size calculation, 25

subjects were included in the study. The primaryend-point of the study was the change in theIIffiL./I

The primary hypothesis to be tested was asfollows:

[i] I year after implant insertion, the test im-plant shows at least 0.35 mm less bone lossthan the control implant.

The secondary hypotheses to be tested were asfollows:

[i] The IBL values change within the studyperiod.

[ii] This change depends on the implant type.[iii] There is a difference between the TBC of

the test and the control implants over thestudy period.

Differences between IBL at the test and thecontrol implants were chosen as the primaryoutcome parameter. Sample-size calculationswere performed using C'Power 3 for matchedpairs (Faul et al. 2009). Based on data fromprevious studies on implants in the lower jawand in accordance with power calculations ofother studies (Hildebolt et al. 1998; Astrandet al. 1999), it was considered possible to detecta true difference of at least 0.35 mm with an SDof 0.65 between the test and the control implantsin this randomized split-mouth study designwith80% power and 23 patients. This estimate wasbased on a one-tailed test of matched pairs con-ducted at the 5% level of significance. To com-pensate for possible dropouts in this prospective5-year clinical trial, the sample size was adjustedto 25 patients.

The primary hypothesis [i] was that I yearafter implant placement, the control implantwould lose at least 0.35 mm more bone thanthe test implant, meaning that the intraindivi-dual difference of the IBLvalues between test andcontrol would be 0.35 mm or more. The primaryhypothesis was tested using the Wilcoxon test.The global test of the dependence of the IBLvalueon implant type (hypothesis iii] and time (hy-pothesis ii] was tested using the nonparametricmodel of Brunner et al. (2002). The intraindivi-dual difference between the TBC of the test andcontrol implants over the study period (hypoth-esis iv] was also tested globally using the non-parametric model of Brunner and Langer. Thesoftware SAS 9.2 (SAS Institute, Heidelberg,

1187 I elin. Oral Impl. Res. 22, 20II / II8S-1192

Enkling et al- Platform switching RCT

Germany) was used for the Brunner-Langermodel. The Wilcoxon tests were performed usingthe software R (R Foundation for StatisticalComputing, Vienna, Austria), and for the posthoc power analysis, the software PASS (NCSS,Kaysville, UT, USA) was used. The graphs weredesigned using the software Prisma (GraphPadSoftwareFinna, La Jolla, CA, USA).

Results

All patients were available for all follow-upexaminations. After I year, the implant survivalrate was 100%. Patients presented with good oralhygiene (Table I) and healthy peri-implant con-ditions: the maximum probing depth around theimplants was 3 mm, and no bleeding was ob-served on probing.

On the radiographs, two types of bone-levelchanges were found: (I) a general horizontal boneloss (GBL) and (2) a defect with a vertical (VVD)and a horizontal (HVD) component. The verticalIBLwas the sum of GBL and VVD. At the I-yearfollow-up, compared with the baseline, the bone-level changes (L'.) were as follows: L'.IBL(meanof the distances A, B)was -0.53 ± o.j y nun forthe test implants and - 0.58 ± 0.55 mm forthe control implants; L'.GBL (mean of thedistances E, F) was - 0.34 ± 0.35 mm for the

Table 1. Full-mouth plaque index (PI, 5ilness &Loe 1964) and sulcular bleeding index (581, Muh-lemann & 50n 1971), categorized in time (n = 25)

Index Time Mean 5D Minimum Maximum

PI Before 0.55 0.32 0.18 1.383 months 0.52 0.33 0.14 1.384 months 0.45 0.29 0.07 0.978 months 0.28 0.21 0 0.7312 months 0.36 0.23 0 0.86

5BI Before 0.34 0.27 0 0.93 months 0.52 0.4 0 14 months 0.32 0.24 0 0.98 months 0.21 0.19 0 0.812 months 0.25 0.24 0 0.83

test implants and -0.19 ± OA7mm for thecontrol implants; and HVD (mean of the dis-tances C, D) was - 0.21 ± 0.38 mm for the testimplants and - 0.35 ± 0.44 mm for the controlimplants (Table 2). In general, VVD and HVDwere similar; the vertical defect measured fromthe new general bone level had about the samedepth and width (Fig. ra and b).

After I year, the intraindividual difference wassmall: 0.05 ± 0.56mm for IBL and 0.14 ±0.61 mm for HVD. The box plot for the intrain-dividual differences of IBL between test andcontrol shows that the values vary around zero(Fig. 3); in some patients, bone levels were bettermaintained at implants with platform switching,whereas in other patients, bone levels were bettermaintained at implants without platform switch-ing (Fig. 4a and b). The nonparametric 95%lower confidence bound for the median of theintraindividual difference of IBL between controland test was - o. I 8 mm and hence smaller than0.35 mm. Therefore, hypothesis [i] had to berejected (P = 0.9907). Because the P-value wasclose to I and the nonparametric 95 % upperconfidence bound was 0.32 mm and thereforesmaller than 0.35mm, the opposite hypothesiswas tested exploratively. The difference in IBLbetween test and control was significantly smal-ler than 0.35mm (P=0.0093, post hoc power79·9%)·

The Brunner-Langer model was able to de-monstrate that the changes in IBL were signifi-cantly dependent on time (P<O.OOI), but not onthe implant type (P = 0A) or an interaction oftime and implant type (P=0.57I). The mainbone remodeling occurred for all implants withinthe first 4 months after implant insertion, withapproximately 0.15 mm vertical bone loss permonth. Between the crown placement and thesecond recall, only a minimal additional verticalbone loss of approximately - o. I ± 0.35 mmwas observed. The progression of bone loss overtime was similar for the test and the control

implants (Fig. 5a and b), with the mean intrain-dividual differences between the two treatmentmodalities always <0.1 mm.

The Brunner-Langer model showed that theTBC was significantly dependent on time(P<O.OOI), but not on implant type (P=0.982)or an interaction of time and implant type(P =OA5 3). TBC were high in both implant typesat baseline, and slightly higher counts were foundin the test group (difference control minus test:- 3.7 x 106,95 % CI [ - 10.5 X 106; 1.2 X 106]).

TBC decreased from baseline to the 12-monthevaluation in both implant types to almost thesame level (Table 3).

Discussion

The results of the present study indicate that theprimary hypothesis (i) must be rejected. Themean intraindividual difference in the L'.IBLva-lues after I year was 0.05 ± 0.56 mm and, there-fore, statistically Significant smaller than theexpected 0.35 mrn, the extent of the platformshifting (P=0.0093, post hoc power 79.9%). Insome patients, the implants with platformswitching lost more bone, whereas in oilierpatients, the implants without platform switch-ing lost more bone (Fig. 4a and b). A significantperi-implant bone-level alteration was observed

c:~ 1.6E 1.22lCii 0.8OJ 0.4>~~E 0.0.,Ec:~

-0.40.ciii -0.8u

-1.2'E~ -1.6

Difference Control vs. Test

Fig. 3. Box plot showing the intraindividual difference in thevertical bone-level alterations of implants with [Test] andwithout [Control] platform switching after 12 months.

Table 2. Peri-implant bone-level alterations compared with the baseline (0 months) for time and implants with (Test) and without (Control) platformswitching: mean of the mesial and distal measurements on the radiographs (mm) (Fig. 1a and b)

Time Measured Control Test Difference (Test- Control)distance

Mean 5D Median 95% CIfor Mean 5D Median 95% CIfor Mean 5D Median 95% CIforthe median the median the median

3 months II.IBL -0.38 0.43 -0.3 -0.45; -0.1 -0.33 0.52 -0.05 - 0.41; 0 0.05 0.51 0.04 - 0.02; 0.25II.HVD -0.2 0.41 0 0; 0 -0.15 0.33 0 0; 0 0.05 0.47 0 0;0II.GBL -0.26 0.33 -0.17 - 0.38; 0 -0.1 0.21 0 -0.05; 0 0.16 0.36 0.04 0; 0.23

4 months II.IBL -0.46 0.55 -0.55 -0.7; -0.2 -0.44 0.42 -0.39 -0.61; -0.21 0.01 0.57 0.14 - 0.34; 0.28II.HVD -0.25 0.39 0 - 0.37; 0 -0.22 0.36 0 - 0.29; 0 0.03 0.47 0 0; 0.13II.GBL -0.2 0.46 -0.11 -0.27; 0 -0.22 0.25 -0.21 - 0.29; 0 -0.02 0.42 0 - 0.28; 0.27

12 months. AIBL -0.58 0.55 -0.63 - 0.92: - 0.43 -0.53 0.35 -0.61 - 0.7; - 0.28 0.05 0.56 0.05 -0.18;0.32II.HVD -0.35 0.44 - 0.32 -0.57; 0 -0.21 0.38 0 0; 0 0.14 0.61 0 0; 0.4II.GBL -0.19 0.47 -0.18 - 0.37; 0 -0.34 0.35 -0.3 -0.5; 0 -0.15 0.46 -0.04 - 0.3; 0

II.IBL,vertical bone-level alteration at the implant compared with the baseline (0 months); II.HVD,horizontal component of the vertical bony defect at the implant,alteration compared with the baseline (0 months); ~GBL,general horizontal bone-level alteration compared with the baseline (0 months).

1188 I Clin. Oral Impl. Res. 22, 20II / rr85-II92 © 20II [ohn Wiley s: SonsA/S

(b)

Fig. 4. Radiographs at implant placement and at I2 months post-op. [a] The test implant [left] underwent less peri-implantcrestal bone loss than the control implant [right]. (bl The test implant [left) underwent more peri-implant crestal bone lossthan the control implant [right].

over time (P<o.ooII, and this mainly took placeduring the first 4 months; i.e., between implantoperation, reopening, impression, framework try-in, and crown mounting (the period when theperi-implant mucosa was frequently manipu-lated) [Abrahamsson et al. 19971. Between bothtreatment modalities, no difference in theperi-implant bone resorption pattern could bemeasured at any time (P =OA): the bone lossdevelopedwith a similar magnitude and velocity.From the 4- to the I2-month follow-ups, i.e.,during the loaded period, only a minimal changeof 0.1 ± 0.35 mm occurred. This time-relatedcrestal bone-level change is typical at implants,and has also been reported for other implantsystems (Bragger et al. 1998; Astrand et al.2004; Broggini et al. 2006).

Compared with the baseline, a ilGBLof - 0.19± oA7 mm at control and - 0·34 ± 0·35mm attest was measured I year after implant place-ment. Raising a flap at implant placement and atsecond-stage surgery could influence and accel-erate bone resorption. Thus, ilGBL might occurregardless of implant placement. In addition to

© 20II JohnWiley&. SonsA/S

GBL, vertical defects were detected. The meanhorizontal extent of this HVD was 0.I4mmlarger for the control than for the test implantsafter 12 months. Nevertheless, the vertical de-fects were throughout similar in width and depth,which is in accordance with the results of Ro-driguez-Ciurana et al.(2009). The intraindividualdifference of the HVD values between the testand the control implants was related to the factthat the ilGBLwas generally more pronounced atthe region of the test than of the control implants;therefore, the vertical defects around the testimplants had, on average, merged more into thegeneral horizontal bone loss than was the casearound the control implants. The fine verticaldefects around the implants may have beeninduced by surgical trauma during implant sitepreparation and implant insertion. Stresses buildup in the cortical bone during surgery, which canlead to local necrosis permanently manifested bythe formation of small vertical defects (Erikssonet al. I984).

In the present study, measurements of thecrestal bone level were performed with standar-

Enklinget al. PlatformswitchingRCT

dized panoramic images. Radiographs are reliableand serve as an altemative to the histologicalevaluation (Albandar I989; Schulze et al. 2000;Hermann et al. 200Ia; Deppe et al. 2004). Pa-noramic images have been used in numerousclinical implant studies, although some authorsrate the quality of the panoramic images inferiorto that of the intraoral images (Benn I990).Nevertheless, in vitro studies could show thatthe panoramic image of the posterior mandibularregion offers a quality that is comparable tointraoral films (De Smet et al. 2002; Rockenbachet al. 2003; Deserno et al. 2009). Peri-implantbony defects measured with panoramic radio-graphs may be overestimated by 52% comparedwith the histological measurements (De Smet etal. 20021, but because both implants were im-mediately adjacent to each other on the sameradiograph, this was similar for both test andcontrol.

In the present study, the bone-level alterationsat implants with platform switching and at im-plants with matching abutments were similar tothe data from the literature for implants withplatform switching [Vela-Nebot et al. 2006).Therefore, the results for the control group werebetter than the data from implants with matchingabutments in comparative studies [Vela-Nebotet al. 2006) or the reference criterion for implantsuccess with 1.5mm bone resorption after I year[Albrektsson et al. 1986; Albrektsson & Isidor1994), which has also been used for comparingthe data of implants with platform switching(Cochran et al. 2009). Our data, which showedno relevant bone-protective effect of platformswitching, are in agreement with results fromanimal studies, in which no differences in verti-cal bone-level alterations between platformswitching and nonplatfonn switching implantscould be demonstrated (Wenget al. 2008; Beckeret al. 2009). The reason for the differences be-tween the findings of the present study andclinical data in the literature with regard to plat-fonn switching may be that, in comparativestudies, mostly implants with a smaller diameterwith matching abutments were compared withimplants of a wider diameter and platformswitching (Hiirzeler et al. 2007; Cappiello et al.2008; Crespi et al. 2009; Prosper et al. 2009;Trammell et al. 2009; Canullo et al. 2010).Therefore, implant diameter may be biased withrespect to bone-level alteration. In the presentstudy, to the best of our knowledge, identicalimplants, different only at the implant shoulder,inserted into similar sites were intraindividuallycompared for the nrst time. Platform switchingwas first defined as IIaltering the horizontalposition of the micro-gap" by Gardner (2005).Before the concept of platform switching wasadopted from implant companies, this technique

I I 89 I elm. Oral Impl. Res. 22, 20II / II8S-II92

Enkling et al- Platform switching RCT

E 0.75 a§.~ 0.50.[~ 0.25<:~ 0.00 +----1f--.---,--t--t-,---.-----.--r--,.----,---r---,----,

~<Ii -0.25>~ -0.50~o\) -0.75~co"is.-1.00E'6d: -1.25<:.& -1.50-I!!u:: -1.75

Test

K ........r--------.J

135 6 7 8 9 10 11

Observation time [months]

E 0.75§.~ 0.50.[•• 0.25\)e~ 0.00 +---:l:---7-----:2!:--+-t----=5-~6:---:7;----;8!:---:9!:---:1:':0----:1.;:.1-1t2~13

l!<Ii -0.25>11-0.50c:oo -0.75c:co"is.-1.00.5~ -1.25CD<:.& -1.50

i!u:: -1.75

b

Observation time [months]

Control

~,1-----1

Fig. 5. [a and b]: Means ± SO of the vertical implant bone level, i.e., first marginal bone-to-implant contact in relation to thereference point [implant-aburment micro-gap I at the test and control implants [data presentation adapted from Astrand et al.20041. The bone level was measured immediately after implant surgery [baseline= a months I, immediately before reentry 13months post-op], immediately after crown insertion 14 months post-op], and at the second recall (12 months post-op]. A

negative value at the Y-axis indicates that the most coronal bone-to-implant contact was more apical than the reference point,and vice versa.

Table 3. Total bacterial counts (TBe) in samples from the internal cavities of test and control implantsanalyzed with real-time PCR

Time TBC control TBC test TBC difference

Mean 95% CI Mean 95% CI Mean 95% CI

o months 18.8 10.7; 24.7 22.5 16.8; 28.5 -3.7 -10.5; 1.23 months 7.4 4.0; 7.6 6.7 4; 6.5 0.7 -1.1; 2.84 months 3 1.9; 3,9 3 1.8; 4.1 0.03 -0.9; 112 months 3.6 2.1; 4.3 3.6 1.7; 3.8 0.01 -0.1; 1.4

The Brunner-Langer model showed that the TBCwas significantly dependent on time (P<O.OOl), but not onimplant type (P=O.982) or an interaction of time and implant type (P=0.453).Data presented in 106 units.

could only be realized by using an abutment witha diameter smaller than that of the implantshoulder. However, this horizontal offset is asso-ciated with a sharp edge at the implant shoulder.As the platform switching concept was integratedinto implant designs, most implant companiestried to avoid this sharp edge. Therefore, the

implant shoulders were slightly rounded or bev-eled like the implant used in the present study(SIC Ace"', SIC-Invent). Because of the differencein the outer shape of the implant shoulder be-tween the test and the control implant, thereference point was the micro-gap between theimplant and the abutment. Hence, the micro-gap

I I90 I Clin. Oral Impl. Res. :U, 20II / II8S-II92

was positioned epicrestally (Fig. ra and b). All 50implants studied had the same length (9.5 mm)and diameter (4mm). Both implant groups had anidentically sized internal space, and the config-uration of the implant-abutment connection hadidentical internal connections with clearance fitsthat allowed a certain pumping effect betweenthe implant inner cavity and the peri-implantenvironment close to the micro-gap (Zipprichet al. 2007). If a bone-protective effect of platformswitching is to be proven, an internal connectionwith a clearance fit is ideal so that the effect ofplatform switching is assumed to be correlatedwith the micro-gap and its abutment ICT,whereas for the negative impact of the micro-gap on the bone, a certain mobility at the tin-plant-abutment connection must be present(Hermann et al. 200Ib). Capiello and colleaguescompared an implant with an internal conicalconnection and a diameter of 4.8 mm with animplant of 4. I mm diameter and a clearance fitconnection. The advantage of the platformswitching implants of 0.72 mm is that less boneresorption might be attributed to the differentconfigurations of the internal connections andadditionally to the difference in implant diameter(Cappiello et al. 2008). In the study of Vigolo andGivani with a y-year follow-up, only implantswith a 5 mm diameter were investigated to en-sure no possible bias of implants' diameters;however, that study'S definition of baseline wasdenture mounting. Therefore, bone-level altera-tions from implant operation to baseline were notmonitored. The advantage of 0.3 mm less boneresorption at implants with platform switchingmight be attributed to the remodeling of thenonmeasured bone between the implant opera-tion and the baseline (Vigolo & Givani 2009). Ifdata from different studies are to be compared,the time point designated for baseline measure-ments is crucial. Some investigators used thesurgical operation as the baseline (Vela-Nebot etal. 2006; Hurzeler et al. 2007; Crespi et al. 2009;Prosper et al. 2009; Canullo et al. 20IO; Coc-chetto et al. zoro], whereas others started theirmeasurements when the prostheses were deliv-ered [Hurzeler et al. 2007; Cappiello et al. 2008;Vigolo & Givani 2009). In the present study, asin other prospective implant studies, the mainbone-level alteration occurred between implantoperation and denture mounting (Astrand et al.2004; Hurzeler et al. 2007). Therefore, the base-line definition of the present study was "implantoperation," the entire bone-level alteration afterimplant insertion was analyzed, and additionally,the date of "crown mounting" was documentedradiographically.

Bacteria are considered to play an importantrole (Lindhe & Meyle 2008) in crestal bone lossat implants. The bacterial colonization can occur

© 201 I John Wiley & SonsA/S

as a consequence of leakage at the implant-abutment interface or as a result of contamina-tion during implant placement (van Winkelhoffet al. 2000). In the present study, the TBCharvested from the inner cavities of both implanttypes were similar. The high counts of bacteria atbaseline were possibly caused by unintentionalcontamination during surgery. Nevertheless,TBC decreased to almost the same level withinthe rz-rnonth study time, with no significantdifferences between both implant types after 3, 4,and 12 months (P=0.982). Thus, a possibleinfluence of differences in the bacterial coloniza-tion between test and control implants on thebone resorption, which could be a possible con-founder, does not appear to have been present.With the standardized operation procedure atcomparable surgical sites in the lower posteriorarch and with constant oral hygiene control,other possible confounding factors should havebeen minimized.

In summary, the measured mean MBL of< 0.6 mm after I year is small and is comparableto the results of implants with platform switch-

References

Abrahamsson, 1., Berglundh, T. & Lindhe, ). (1997)The mucosal barrier following abutment dis/recon-nection. An experimental study in dogs. Journal ofClinical Periodontology 24: 568-572.

Albandar, ).M. (1989) Validity and reliability of alveolarbone level measurements made on dry skulls. Journalof Clinical Periodontology 16: 575-579·

Albrektsson, T. & Isidor, F. (1994) Consensus reportof session iv. In: Lang, N.P. & Karring, T.,eds. Proceedings of the 1st European Workshop onPeriodontology, 365-369. London: QuintessencePublishing Co ..

Albrektsson, T., Zarb, G., Worthington, P. & Eriksson,A.R. (1986) The long-term efficacy of currently useddental implants: a review and proposed criteria ofsuccess. The International Journal of Oral eJ Max-illofacial Implants I: 11-25.

Astrand, P., Engquist, B., Dahlgren, S., Engquist, E.,Feldmann, H. & Grondahl, K. (1999) Astra techand branemark system implants: a prospective 5-year comparative study. Results after one year.Clinical Implant Dentistry eJ Related Research I:

17-26.Astrand, P., Engquist, B., Dahlgren, S., Grondahl, K.,

Engquist, E. & Feldmann, H. (2004) Astra tech andbranemark system implants: a y-year prospectivestudy of marginal bone reactions. Clinical Oral Im-plants Research 15: 413-420.

Becker, )., Ferrari, D., Mihatovic, 1., Sahm, N., Schaer,A. & Schwarz, F. (2009) Stability of crestal bonelevelat platform-switched non-submerged titanium im-plants: a histomorphometrical study in dogs. Journalof Clinical Periodontology 36: 532-539.

Benn, D.K. (1990) A review of the reliability ofradiographic measurements in estimating alveolarbone changes. Journal of Clinical Periodontology IT14-2 I.

© 20II John Wiley & Sons AjS

ing and a tight inner cone connection [Astrand etal. 2004; Norton 2006). These results wereachieved with and without platform switchingusing an interlocking clearance fit implant-abut-ment connection, i.e., no inner-cone connection.Therefore, the prosthetic restoration of the im-plant used in the present trial, i.e., treatmentwith or without platform switching, can bechosen according to the personal preferences ofthe clinician. At the annual examinations of thisprospective y-year clinical study, it has to bedetermined whether the bone level will stabilizeat the I-year level, as suggested by other pro-spective implant studies [Astrand et al. 2004), orwhether, in the long term, an advantage of theplatform-switching concept concerning the peri-implant bone-level alteration will be seen (Vigolo& Givani 2009).

Acknowledgements: The authorswould like to acknowledge the supportprovided by Dr Claudia Weiss (KKSKoln,Coordinating Centre for Clinical Trials,

Bragger, U., Hafeli, U., Huber, B., Hammerle, C.H.& Lang, N.P. (1998) Evaluation of postsurgicalcrestal bone levels adjacent to non-submergeddental implants. Clinical Oral Implants Research 9:218-224.

Broggini, N., McManus, L.M., Hermann, ),S., Medina,R., Schenk, R.K., Buser, D. & Cochran, D.L. (2006)Peri-implant inflammation defined by the implant-abutment interface. Journal of Dental Research 8S:473-47S.

Brunner, E., Domhof, S. & Langer, F. (2002) Nonpara-metric Analysis of Longitudinal Data in FactorialExperiments. New York: Wiley-Interscience.

Canullo, L., Fedele, G.R., Iannello, G. & Jepsen, S.(2010) Platform switching and marginal bone-levelalterations: the results of a randomized-controlledtrial. Clinical Oral Implants Research 21: 1I5-12I.

Cappiello, M., Luongo, R., Di Iorio, D., Bugea, c.,Cocchetto, R. & Celletti, R. (2008) Evaluation ofperi-implant bone loss around platform-switched im-plants. The International Journal of Periodontics eJRestorative Dentistry 28: 347-355.

Cocchetto, R., Traini, T., Caddeo, F. s: Celletri, R.(2010) Evaluation of hard tissue response aroundwider platform -switched implants. The InternationalJournal of Periodontics eJ Restorative Dentistry 30:

163-17I.Cochran, D.L., Bosshardt, D.o., Crize, L., Higginbot-

tom, F.L., Jones, A.A., lung, R.E., Wieland, M. &

Dard, M. (2009) Bone response to loaded implantswith non-matching implant-abutment diameters inthe canine mandible. Journal of Periodontology 80:

609-617·Crespi, R., Cappare, P. & Gherlone, E. (2009) Radio-

graphic evaluation of marginal bone levels aroundplatform-switched and non-platform-switched im-plants used in an immediate loading protocol. The

Enkling et al· Platform switching RCT

Cologne, Germany) for the statisticalevaluation of the study data and the supportprovided by Prof. Dr rer. nat. Thomas M.Deserno [ne Lehmann) (Department ofMedical Informatics, Aachen University ofTechnology [RWTH], Aachen, Germany) forthe evaluation of the radiographs. Statisticalanalysis using the Brunner and Langermodel and the post hoc power analysis forresults of the primary hypothesis wereperformed with the support of Mr Yves Bartelsand Mr Niki Zumbrunnen (Institute ofMathematical Statistics and Actuarial Science[IMSV],University of Bern, Switzerland).Additionally, the authors would like to thankDr med. dent. Torben Klimberg, Mrs MartinaPopovic, and Mrs Sabine Sieverding(Dental Clinic Bochum/University ofWitten-Herdecke, Germany) for theirexcellent support in organizing the studyappointments. Conflict of interest andsource of funding statement: The authorsdeclare that they have no conflict ofinterest.

Intemational Journal of Oral eJ Maxillofacial Im-plants 24: 920-926.

Deppe, H., Wagenpfeil, S. & Donath, K. (2004) Com-parative value of attachment measurements in im-plant dentistry. The International Journal of Oral eJMaxillofacial Implants 19: 208-215.

Deserno, T.M., Rangarajan, ).R., Hoffmann, )., Bragger,u., Mericske-Stern, R. & Enkling, N. (2009) Aposteriori registration and subtraction of panoramiccompared with intraoral radiography. Oral Surgery,Oral Medicine, Oral Pathology. Oral Radiology andEndodontology IOS: e39-e45.

De Smet, E., Jacobs, R., Gijbels, F. & Naert, 1. (2002)The accuracy and reliability of radiographic methodsfor the assessment of marginal bone level aroundoral implants. Dentomaxillofacial Radiology 31:176-18r.

Ericsson, I., Nilner, K., Klinge, B. & Glantz, P.O. (1996)Radiographical and histological characteristics of sub-merged and nonsubmerged titanium implants. Anexperimental study in the Labrador dog. ClinicalOral Implants Research T 20-26.

Ericsson, 1., Persson, L.G., Berglundh, T., Marinello,c.P., Lindhe, ). & Klinge, B. (1995) Different typesof inflammatory reactions in peri-implant softtissues. Journal of Clinical Periodontology 22: 255-26I.

Eriksson, R.A., Albrektsson, T. & Magnusson, B.(1984) Assessment of bone viability after heat trauma.A histological, histochemical and vital microscopicstudy in the rabbit. Scandinavian Journal of Plasticand Reconstructive Surgery and Hand Surgery 18:261-268.

Faul, F., Erdfelder, E., Buchner, A. & Lang, A.G. (2009)Statistical power analyses using g*power 3.1: tests forcorrelation and regression analyses. Behavior Re-search Methods 41: II49-II60.

I 191 I Clin. Oral Impl. Res. 22, 20II / IIS5-II92

Enkling et al. Platform switching RCT

Gardner, D.M. (20051 Platform switching as a means toachieving implant esthetics. The New York StateDental [oumal. 71: 34-37·

Hermann, J.S., Buser, D., Schenk, R.K. & Cochran,D.L. (20001 Crestal bone changes around titaniumimplants. A histometric evaluation of unloaded non-submerged and submerged implants in the caninemandible. [oumal of Periodontology 71: 1412-1424.

Hermann, J.S., Schoolfield, J.D., Nummikoski, P.V.,Buser, D., Schenk, R.K. & Cochran, D.L. (2001alCrestal bone changes around titanium implants: amethodologic study comparing linear radiographicwith histometric measurements. The Intemational[oumal. of Oral etJMaxillofacial Implants 16: 475-485.

Hermann, J.S., Schoolfield, J.D., Schenk, R.K., Buser,D. & Cochran, D.L. (2001bl Influence of the size ofthe microgap on crestal bone changes around titaniumimplants. A histometric evaluation of unloaded non-submerged implants in the canine mandible. [ouinalof Periodontology 72: 1372-1383.

Hildebolt, e.F., Pilgram, T.K., Yokoyama-Crothers, N.,Fletcher, G., Helbig, J.L., Bartlett, T.Q., Gravier, M.,Vannier, M.W. & Shrout, M.K. (19981 Reliabilityof linear alveolar bone loss measurements ofmandibular posterior teeth from digitized bitewingradiographs. [oumal of Clinical Periodontology 25:850-856.

Hurzeler, M., Fickl, S., Zuhr, O. & Wachtel, n.c.(20071 Peri-implant bone level around implants withplatform-switched abutments: preliminary data froma prospective study. [otunal of Oral and MaxillofacialSurgery 65: 33-39·

[ervoe-Storm, P.M., Koltzscher, M., Falk, W., Dorfler,A. & Jepsen, S. (20051 Comparison of culture andreal-time per for detection and quantification of fiveputative periodontopathogenic bacteria in subgingivalplaque samples. [oumol of Clinical Periodontology32: 778-783.

Lazzara, R.J. & Porter, S.S. (20061Platform switching: anew concept in implant dentistry for controllingpostrestorative crestal bone levels. The Intemational[ouinal of Periodontics etJRestorative Dentistry 26:

9-17·

I 192 I Clin. Oral Impl. Res. 22, 20II j II85-II92

Lindhe, J. & Meyle, J. (20081 Peri-implant diseases:consensus report of the sixth European workshop onperiodontology. joumal of Clinical Periodontology35: 282-285.

Muhlemann, H.R. & Son, S. (197I1 Gingival sulcusbleeding - a leading symptom in initial gingivitis.Helvetica Odontologica Acta IS: I07-II3·

Norton, M.R. (20061 Multiple single-tooth implantrestorations in the posterior jaws: maintenanceof marginal bone levels with reference to theimplant-abutment microgap. The International[oumal of Oral etJ Maxillofacial Implants 21:777-784.

Prosper, L., Redaelli, S., Pasi, M., Zarone, F., Radaelli,G. & Gherlone, E.F. (20091A randomized prospectivemulticenter trial evaluating the platform -switchingtechnique for the prevention of postrestorative crestalbone loss. The Intemational [oumal of Oral etJMaxillofacial Implants 24: 299-308.

Rockenbach, M.l., Sampaio, M.e., Costa, L.J. & Costa,N.P. (20031 Evaluation of mandibular implant sites:correlation between panoramic and linear tomogra-phy. Brazilian Dental [oumal 14: 209-213.

Rodriguez-Ciurana, X., Vela-Nebot, X., Segala-Torres,M., Calvo-Guirado, J.L., Cambra, J., Mendez-Blanco,V. & Tarnow, D.P. (20091 The effect of interimplantdistance on the height of the interimplant bone crestwhen using platform-switched implants. The Inter-national [oumal of Periodontics etJRestorative Den-tistry 29: 141-151.

Schulze, R., Krummenauer, F., Schalldach, F. & d'Hoedt,B. (20001 Precision and accuracy of measurements indigital panoramic radiography. DentomaxillofacialRadiology 29: P-56.

Silness, J. & Loe, H. (19641 Periodontal disease inpregnancy. Ii. Correlation between oral hygiene andperiodontal condition. Acta Odontologica Scandina-vica 22: 121-135.

Trammell, K., Geurs, N.e., O'Neal, S.J., Liu, P.R.,Haigh, S.T.,McNeal, S., Kenealy, J.N. & Reddy, M.S.(2009) A prospective, randomized, controlled compar-ison of platform-switched and matched-abutmentimplants in short-span partial denture situations.

The Intemational [oumal of Periodontics etJRestora-tive Dentistry 29: 599-605.

van Winkelhoff, A.T., Goene, R.J., Benschop, e. &

Folmer, T. (20001 Early colonization of dental im-plants by putative periodontal pathogens in partiallyedentulous patients. Clinical Oral Implants ResearchII: 5II-pO.

Vela-Nebot, X., Rodriguez-Ciurana, X., Rodado-Alonso,e. & Segala-Torres, M. (20061Benefits of an implantplatform modification technique to reduce crestalbone resorption. Implant Dentistry IS: 313-320.

Vigolo, P. & Givani, A. (20091 Platform-switched re-storations on wide-diameter implants: a y-year clin-ical prospective study. The International [ournal ofOral etJMaxillofacial Implants 24: 103-I09.

Weng, D., Nagata, M.J., Bell, M., Bosco, A.F., de Melo,L.G. & Richter, E.J. (20081 Influence of microgaplocation and configuration on the periimplant bonemorphology in submerged implants. An experimentalstudy in dogs. Clinical Oral Implants Research 19:1I41-II47·

Zipprich, H., Weigl, P., Lange, B. & Lauer, H.e. (20071Micromovements at the implant-abutment interface:

measurement, causes, and consequences. Implanto-logie IS: 31-46.

Supporting Information

Additional supporting information may be foundin the online version of this article:

Table SI. Supporting information in accor-dance with the CONSORT Statement 2001

checklist used in reporting randomized trials.

Please note: Wiley-Blackwell is not responsiblefor the content or functionality of any supportingmaterials supplied by the authors. Any queries(other than missing material) should be directedto the corresponding author for the article.

© 20II John Wiley & Sons AjS