BASIC SCIENCE RESEARCH

The Effects of Abutment Taper, Length andCement Type on Resistance to Dislodgementof Cement-Retained, Implant-SupportedRestorationsGuillermo Bernal, DDS, MSD,1 Mitsunobu Okamura, DDS, MSD,2 andCarlos A. Munoz, DDS, MSD3

Purpose: The purpose of this study was to compare the effect of 20 degrees and 30 degrees of totalocclusal convergence (TOC), the occlusocervical dimension, and the type of cement on the tensileresistance to dislodgement of cement-retained, implant-supported restorations.

Materials and Methods: Cylindrical preparations with TOC angles of 20 degrees and 30 degrees andocclusocervical dimensions of 4 mm (S) and 8 mm (L) were made through a machining process. The cylindershad a shoulder finish line of 1.0 mm in depth. Eight impressions were made of each machined cylinder andpoured in type IV dental stone, for a total of 32 dies. Die spacer was applied to each die. A master wax patternwas designed, and the 32 wax patterns were marginated, invested, and cast in type IV gold alloy (n � 8). Thegold crowns were cemented with Fleck’s cement (zinc phosphate cement), Temp-Bond (zinc oxide eugenolcement), Temp-Bond plus Vaseline (30% by weight), and IMProv temporary cement (acrylic/urethanecement) under a 10-kg load and placed in a humidor at 37°C for 1 hour before testing. A uniaxial tensile forcewas applied to the crown using an Instron machine with a crosshead speed of 5 mm/min until cement failureoccurred. Analysis of variance models were fit to determine the effect of TOC, occlusocervical dimension, andcement type of the restorations on the mean tensile strength.

Results: For each type of cement, the mean tensile strengths were significantly higher at 20 degrees ofTOC and 8 mm of occlusocervical dimension compared with the other preparations. At this preparation,IMProv had the highest mean tensile resistance to dislodgement (47.7 � 8.4 kg), followed by Fleck’s(38.2 � 8.8 kg), Temp-Bond (35.9 � 4.4 kg), and Temp-Bond plus Vaseline (8.2 � 2.2 kg). No statisticallysignificant difference was observed between Temp-Bond and Fleck’s zinc phosphate cement when 20degrees of TOC and 8 mm of occlusocervical dimension was used. There was no statistical difference inthe mean tensile resistance to dislodgement for Temp-Bond plus Vaseline with different preparationdesigns (p >0.05) except when 20 degrees of TOC and 8 mm of occlusocervical dimension was used. Themean tensile strength was significantly different between Temp-Bond and Temp-Bond plus Vaseline foreach of the 4 preparation designs (p < 0.05). Among the cements tested, IMProv exhibited higher values,which were statistically different (p < 0.05). Restorations with greater occlusocervical dimension and lessTOC exhibited higher tensile resistance to dislodgement.

Conclusions: Preparations with 20 degrees of TOC and 8 mm of occlusocervical dimension hadsignificantly higher mean retentive values for all of the cements tested. Significant differences inmean tensile strength were observed, with the highest tensile resistance seen with IMProv, followedby Fleck’s cement, and the lowest tensile resistance seen with Temp-Bond plus Vaseline.

J Prosthodont 2003;12:111-115. Copyright © 2003 by The American College of Prosthodontists.

INDEX WORDS: dental cement, screw retention, cementation, implant-supported prosthesis

SCREW-RETAINED, implant-supported pros-theses were developed in response to the need

for retrievability of restorations should removal berequired. As techniques continue to evolve, thesurvival rates of implant-retained restorations areimproving.1 Consequently, the use of cement-re-tained, implant-supported restorations has in-creased, due in part to the ability to optimize oc-clusal interdigitation, enhance esthetics in areasthat would otherwise be the locations of screw

1Professor and Director, Advanced Education Program in Prosthodon-tics, Loma Linda University, Loma Linda, CA.

2Private practice, Fukuoka, Japan.3Professor and Director, Center for Dental Research, Loma Linda

University, Loma Linda, CA.Accepted December 13, 2002Correspondence to: Guillermo Bernal, DDS, MSD, Loma Linda Uni-

versity School of Dentistry, Advanced Education in Prosthodontics, 11092Anderson Street, Loma Linda, CA 92350. E-mail: gbernal@sd.llu.edu

Copyright © 2003 by The American College of Prosthodontists1059-941X/03/1202-0001$30.00/0doi: 10.1016/S1059-941X(03)00006-8

111Journal of Prosthodontics, Vol 12, No 2 ( June), 2003: pp 111-115

access holes, and provide a passive fit, which mayactually improve loading characteristics.2

Screw retention remains the retention mecha-nism of choice for many practitioners, because ofease of retrievability. Simple retrieval is particu-larly important if complications arise, such as abut-ment screw loosening. Conversely, it is more diffi-cult to retrieve cemented restorations withoutpossibly compromising the integrity of the crown,abutment screw, or the implant itself.

Goodacre et al,3 in a review of 17 studies, foundthat abutment screw loosening occurred in 2%–45% of the reported cases. The highest incidencewas noted with single crowns placed in the premo-lar and molar areas.4 One possible cause for failuresin these early designs was the use of titaniumabutment screws, which resulted in a high coeffi-cient of friction between the screw and the abut-ment.5 With the introduction of gold alloy abut-ment screws and new screw designs, the reportedincidence of screw loosening has decreased.6,7

Many of the present-day implant systems haveabutments onto which superstructures can be ce-mented. This design permits the development ofthe desired occlusal interdigitation, improved es-thetics, and correct loading characteristics. Abut-ment preparation designs and cementation tech-niques now mimic conventional fixed prosthodonticprocedures for natural teeth. Moreover, the cementspace that exists between the crown and abutmentcan help compensate for minor discrepancies in thefit of the prosthesis.8-14 However, cementation of asuperstructure makes removal of cemented resto-rations more difficult.8

Several factors influence the retention and resis-tance form of conventional cement-retained resto-rations.10,11 These factors may affect the abutment,the casting, and the luting agent. The abutment isaffected by (1) total occlusal convergence (TOC),(2) surface area and height, (3) surface finish orroughness, and (4) intracoronal auxiliary featuresin the preparation. Factors affecting a casting in-clude (1) relative adaptation of the restoration tothe abutment, (2) texture of the internal surface ofthe casting, (3) splinting of multiple units, and (4)strength properties of the cast metal superstruc-ture. The luting agent is affected by (1) type ofcement, (2) use of venting or internal casting relief,(3) variations in cement viscosity, (4) variations inthe seating forces, and (5) variations in the dura-

tion of the force. The effect of these factors oncement-retained, implant-supported restorationshas yet to be established.

The purpose of this study was to compare theeffect of 20 degrees and 30 degrees of TOC, differ-ences in occlusocervical dimension, and differenttypes of cement on the tensile resistance to dis-lodgement of cement-retained restorations.

Materials and MethodsCylindrical preparations with TOC angles of 20 degreesand 30 degrees and occlusocervical heights of 4.0 mm (S)and 8.0 mm (L) were machined from a titanium alloycommonly used for abutments that can be prepared tocustom shapes (preparable abutments). The cylindershad a 1.0 mm shoulder finish line (Fig 1).

Eight poly(vinyl)siloxane impressions were made ofeach machined cylinder and poured in an AmericanDental Association (ADA)-certified type IV dental stone(Silky Rock; Whip Mix, Louisville, KY), creating a total of32 dies. Two coats of die spacer (Belle de St Claire; KerrLaboratories, Orange, CA) were applied to each die by 1investigator (M.O.). A master wax pattern coping wasdesigned with a loop on the occlusal surface to permit thecastings to engage a special device of the Instron ma-chine. A silicone mold of the master wax pattern wasfabricated, so standardized wax patterns could beformed. The wax was injected around each die, and 32standardized wax patterns were produced.

The 32 wax patterns were marginated and invested ina noncarbon, phosphate-bonded investment (Fast Fire15; Whip Mix). The castings were made using a gold-palladium alloy (Lodestar; Williams-Ivoclar, Amherst,NY), then cleaned and examined under �10 magnifica-tion for nodules before being carefully seated onto theirrespective gypsum die to check passive fit. Eight castingsfor each type of preparation were cemented to the abut-ments using 1 of the following 4 types of commonly usedluting agents in implant prosthodontics: zinc phosphate(Fleck’s; Mizzy, Cherry Hill, NJ), zinc oxide eugenol(Temp-Bond; Kerr Laboratories), acrylic/urethane ce-ment (IMProv; SteriOss, Yorba Linda, CA), and Temp-Bond mixed with 30% petroleum jelly (Vaseline; Chese-brough-Pond, Greenwich, CT) by weight (n � 8).

The crowns were cemented to the simulated abut-ment under a 10-kg load and placed in a humidor at 37°Cfor 1 hour before testing. A uniaxial tensile force wasapplied to the gold crowns using an Instron machine witha crosshead speed of 5 mm/min until cement failureoccurred (Fig 2). After the castings were dislodged fromthe abutments, the abutments were placed into the ul-trasonic unit using a cement-removal solution for 20minutes, then rinsed and placed again in distilled waterfor 5 minutes. A new group of crowns was checked for

112 Resistance to Dislodgement of Restorations ● Bernal, Okamura, and Munoz

internal surface configuration, passivity, and marginaladaptation before cementation. Differences in mean ten-sile strength were analyzed by fitting analysis of variance(ANOVA) models. Two factors were studied: (1) thecombination effect of the TOC angle and occlusocervicaldimension and (2) the cement type. First, a 2-wayANOVA model was fit with main effect terms for the 2factors and a term for their 2-way interaction. The inter-action effect was evaluated to determine whether theeffect of cement type on the mean tensile strength wasthe same for the different convergence angle/dimensioncombinations. In the presence of a significant interactioneffect, 8 separate 1-way ANOVA models were fit tocompare the mean tensile strength between the cementtypes separately for each of the 4 angle/dimension com-binations and to compare the mean tensile strengthbetween the angle/dimension combinations separatelyfor each of the 4 cement types. If differences were de-tected, then multiple comparisons were made usingthe Student-Newman-Keuls multiple comparisons proce-dure. All calculated p values were 2-sided, and p valuesbelow 0.05 were considered statistically significant. Thestatistical analysis was performed using Sigma Stat, ver-sion 2.03 (SPSS, Chicago, IL).

ResultsResults (Table 1) indicate that castings cementedwith IMProv for abutments with a 20-degree TOCangle and 8.0 mm of occlusocervical height had thehighest mean tensile resistance to dislodgement(47.7 � 8.4 kg), followed by Fleck’s zinc phosphate(38.2 � 8.8 kg) and then Temp-Bond (35.9 � 4.4kg). Temp-Bond plus Vaseline had the lowest meantensile resistance (8.2 � 2.2 kg). Figure 3 illustratesthe differences in mean tensile strength betweenthe 4 types of cement for each of the 4 preparationdesigns. At a 20-degree TOC and an 8.0-mm oc-clusocervical height, there was no statistically sig-

nificant difference observed between Temp-Bondand Fleck’s zinc phosphate (p � 0.05); however,mean tensile resistance was significantly differentin these 2 cements for the other 3 preparationdesigns (p � 0.05). Preparations with a 30-degreeTOC and an 8-mm occlusocervical height yieldedstatistically significant differences in mean reten-tion among all of the cements studied, with thehighest mean retention observed with the IMProvcement (p � 0.001). There was no significant dif-ference among crowns cemented with Fleck’s zincphosphate or IMProv when the preparations wereshort (4.0 mm) and had either a 20-degree or30-degree TOC angle (p � 0.05). However, themean tensile resistance was significantly higher forcrowns cemented with IMProv compared with ei-ther of the 2 Temp-Bond cements for each of the 4preparation designs (p � 0.05). The mean tensileresistance was significantly different for Temp-Bond and Temp-Bond plus Vaseline for each of the4 preparation designs (p � 0.05).

Figure 4 illustrates the differences in mean ten-sile strength among the 4 preparation designs foreach of the 4 types of cement. There were nostatistically significant differences (p � 0.05) be-tween the 20-degree TOC angle, 4.0-mm occluso-cervical height preparation and the 30-degree TOCangle, 8.0-mm occlusocervical height preparationwhen the crowns were cemented with Temp-Bondplus Vaseline, Temp-Bond, or Fleck’s zinc phos-phate. There were no statistically significant differ-ences (p � 0.05) between the 20-degree TOC angle,4.0-mm occlusocervical height and the 30-degree

Figure 1. Cylindrical preparations with TOC angle of 20and 30 degrees and occlusocervical dimension of 4.0 mm(S) and 8.0 mm (L).

Figure 2. A crown dislodged after a uniaxial tensileforce was applied to the gold crown using a universaltesting machine.

113June 2003, Volume 12, Number 2

TOC angle, 4-mm occlusocervical height prepara-tions when IMProv was used.

DiscussionCement-retained, implant-supported prostheseshave gained popularity because they allow comple-tion of clinical procedures using conventional fixedprosthodontic techniques. In the absence of occlusalscrew access openings, cemented implant-sup-ported restorations offer enhanced esthetics and anincreased number of occlusal contacts. Cementedrestorations compensate for minor fit discrepanciesthrough use of a luting agent.

One of the major concerns with cemented res-torations is the challenge of retrieval when an abut-ment screw loosens. A cast restoration that fits wellmay be difficult to remove after provisional cemen-tation to a prepared tooth. Likewise, a casting foran implant abutment may be difficult, if not impos-sible, to retrieve without sectioning it.

When selecting a luting agent, it is importantthat the cement be easy to manipulate and removewithout damaging implant components or sur-rounding soft tissue. If residual cement remains, itshould not cause peri-implant complications. In aclinical report, Pauletto et al12 described soft tissue

complications arising due to excess subgingival ce-ment. In each instance, surgery was necessary togain access to remove the remaining cement. Analternative to cementation is the use of a horizontallingual set screw, but this approach is more com-plicated and costly.

Hebel and Gajjar2 reported that Temp-Bondmixed with petroleum jelly provided adequate re-tention for fixed partial dentures when multipleabutments are splinted. However, this studyshowed that with an increased TOC angle and adecreased occlusocervical height, the retentive val-ues for Temp-Bond and petroleum jelly were de-scribed as very low.

According to its manufacturer, IMProv is a eu-genol-free acrylic/urethane cement for temporarycementation. The results from this study showedthat IMProv had a greater resistance to dislodge-ment than the permanent zinc phosphate cementtested. The use of this cement in situations whereimplant component loosening occurs may makeremoval of the prosthesis difficult. If componentloosening is a realistic possibility, then the clinicianmay wish to consider using a cement that allowsprosthesis retrieval without damaging the prosthe-

Figure 4. The effect of TOC angle and occlusocervicaldimension on the tensile resistance to dislodgement ofcement-retained restorations.

Table 1. Mean Tensile Strength for the Four Cements Tested (Kg)

Tempbond V Tempbond Zinc Phosphate IMProv

Mean SD Mean SD Mean SD Mean SD

30° & 4 mm 2.4 1.4 5.7 1.1 10.8 6.4 12.8 3.920° & 4 mm 3.5 1.0 12.7 2.5 23.0 6.3 19.1 6.630° & 8 mm 2.2 1.0 10.4 6.2 20.4 6.8 27.2 4.820° & 8 mm 8.2 2.4 35.9 4.4 38.2 8.8 47.7 8.4

N � 8. Groups connected by vertical lines are not statistically different (p � 0.05).

Figure 3. The effect of cement type on the tensile re-sistance to dislodgement of cement-retained restorations.

114 Resistance to Dislodgement of Restorations ● Bernal, Okamura, and Munoz

sis. Michalakis et al13 evaluated the cement failureof 4 provisional luting agents used for the cemen-tation of fixed partial dentures (FPDs) supported by2 or 4 implants. This study also demonstrated thatIMProv had the highest mean retentive strength 24hours postcementation, and TempBond had thelowest tensile strength. The authors concluded thatNogenol may be more appropriate for the cemen-tation of implant-supported FPDs when retrievabil-ity is important. The current study evaluated theaddition of a lubricant (Vaseline) to a provisionalcement (Temp-Bond) and found that this was theleast retentive of all the cements tested. Usingthese results, the clinician should carefully considerthe choice of cement when the risk of componentloosening is high. In those situations, a weakercement may be clinically effective in allowing pros-thesis retrieval. It appears to be important to eval-uate surface area and TOC of the abutments, aswell as interocclusal space and occlusion, because insome instances when those factors are compro-mised, the risk of the prostheses loosening is high.

TOC angle and occlusocervical dimension influ-ence the amount of retention provided in a cement-retained prosthesis. Jorgensen10 investigated therelationship between retention and TOC angle infull-veneer crowns and found that the greatest re-tention occurred with axial wall parallelism andrapid decline in retention as TOC angle increased.This observation was confirmed in the presentstudy, where results indicated that cylindrical prep-arations with a 20-degree TOC and 8.0-mm oc-clusocervical dimension produced the highest re-tentive values with all the cements tested.

ConclusionsWithin the limitations of this study, we can drawthe following conclusions:

1. At an occlusocervical dimension of 8 mm, a TOCangle of 20 degrees demonstrated a significantincrease in retention for all cements comparedto a TOC angle of 30 degrees.

2. An 8-mm axial wall height causes significantlyhigher removal forces than a 4-mm axial wallheight for all cements except Temp-Bond plusVaseline.

3. IMProv and Fleck’s zinc phosphate cement dem-onstrated significantly greater resistance to re-moval than Temp-Bond or Temp-Bond plus Va-seline.

References1. Adell R, Eriksson B, Lekholm U, et al: A long-term follow-up

study of osseointegrated implants in the treatment of totallyedentulous jaws. Int J Oral Maxillofac Implants 1990;5:347-359

2. Hebel KS, Gajjar RC: Cement-retained versus screw-re-tained implant restorations: Achieving optimal occlusionand esthetics in implant dentistry. J Prosthet Dent 1997;77:28-35

3. Goodacre CJ, Kan JY, Rungcharassaeng K: Clinical compli-cations of osseointegrated implants. J Prosthet Dent 1999;81:537-552

4. Becker W, Becker BE: Replacement of maxillary and man-dibular molar with single endosseous implant restorations:A retrospective study. J Prosthet Dent 1995;74:51-55

5. McGlumphy EA, Mendel DA, Holloway JA: Implant screwmechanics. Dent Clin North Am 1998;42:71-89

6. Eckert SE, Wollan PC: Retrospective review of 1170 endos-seous implants placed in partially edentulous jaws. J Pros-thet Dent 1998;79:415-421

7. Henry PJ, Laney WR, Jemt T, et al: Osseointegrated im-plants for single-tooth replacement: A prospective 5-yearmulticenter study. Int J Oral Maxillofac Implants 1996;11:450-455

8. Guichet DL, Caputo AA, Choi H, et al: Passivity of fit andmarginal opening in screw- or cement-retained implantfixed partial denture designs. Int J Oral Maxillofac Implants2000;15:239-246

9. Chiche GJ, Pinault A: Considerations for fabrication ofimplant-supported posterior restorations. Int J Prosthodont1991;4:37-44

10. Jorgensen KD: The relationship between retention and con-vergence angle in cemented veneer crowns. Acta OdontolScand 1955;13:35-40

11. Kaufman EG, Coelho DH, Colin L: Factors influencing theretention of cemented gold castings. J Prosthet Dent 1961;11:487-502

12. Pauletto N, Lahiffe BJ, Walton JN: Complications associatedwith excess cement around crowns on osseointegrated im-plants: A clinical report. Int J Oral Maxillofac Implants1999;14:865-868

13. Michalakis KX, Pissiotis AL, Hirayama H: Cement failureloads of 4 provisional luting agents used for the cementationof implant-supported fixed partial dentures. Int J Oral Max-illofac Implants 2000;15:545-549

14. Campagni WV, Preston JD, Reisbick MH: Measurements ofpaint-on die spacers used for casting relief. J Prosthet Dent1982;47:606-611

115June 2003, Volume 12, Number 2