The effect of diameter and length of hydroxylapatite-coated dental implants on ultimate pullout force in dog alveolar bone

J Oral Maxillofac Surg 48: 174·178. 1990

The Effect of Diameter and Length of Hydroxylapatite-Coated Dental Implants on Ultimate Pullout Force in Dog Alveolar Bone MICHAEL S. BLOCK, DMD,* ARMANDO DELGADO, DMD,t AND MARK G. FONTENOT, DDS, MENG:t: This study compared the effect of diameter and length on the pullout force required to extract hydroxylapatite-coated implants from dog alveolar bone. After 15 weeks for integration, implants of 3.0, 3.3, and 4.0 mm diameter and 4, 8, and 15 mm length were pulled. The results showed that the ultimate pullout force correlated strongly to implant length, but not diameter. Cortical bone contributed more to implant retention than cancellous bone.

In a recent study by Gerner et aI, tricalcium phosphate- (TCP) coated and uncoated cylindrical implants with a slight taper were placed in the canine iliac crest.f At 14 weeks, the average ultimate shear strength of the TCP-coated and uncoated implants were 1.60 MPa and 0.36 MPa, respectively. In this study, the forces had both shear and tensile strength components and were actually fixation strengths rather than shear strengths. De Groot et ae and Geesink et al" in 1987 reported on the ultimate shear strength of their HAcoated implants in the dog femur. They found the tensile bonding strength of the HA-titanium-bone interface to be 85 MPa. Uncoated titanium implants were reported to have an ultimate interfacial shear strength of 0.6 MPa. Thomas et al 5 examined the mechanical characteristics of cylindrical implants that differed in macrostructure from previous studies in that they had semicircular annular grooves approximately 750 urn in depth. The implants were tested with and without an HA coating in the dog femur. Pushout testing on these implants with mechanical retentive designs resulted in an interfacial shear strength of 17 MPa. Cook et al reported on the shear strength of HAcoated, porous titanium and uncoated, porous titanium implants." The results indicated that the use of HA coating on porous titanium increased attachment strength only in the early periods. None of the previously described studies considered the mechanical consequences of variations in cortical and

The anatomy of the alveolar bone often limits the size of an endosseous implant for reconstruction of the edentulous patient. Thus, implants of different diameters and lengths are required for alveolar ridges that vary in width and height. The purpose of this study was to compare the effect of diameter and length on the amount of pullout force required to extract hydroxylapatite- (HA) coated dental implants from dog alveolar bone. Cook et al reported on the ultimate shear strength of HA-coated and uncoated grit-blasted cylindrical titanium implants in the dog femur. I Implants were mechanically tested by a simple pushout test so that the bone-implant interface was placed in shear. At 32 weeks, the HA-coated implants had an ultimate shear strength of 6 to 7 MPa (megapascal), compared with less than 1.2 MPa for the uncoated implants. Received from Ihe LSU School of Dentistry, New Orleans, LA. * Associate Professor, Department of Oral and Maxillofacial Surgery. t Graduate Student, Department of Prosthodontics. :j: DEng Candidate and Resident, Department of Oral and Maxillofacial Surgery. Supported by Basic Research Support Grant, LSU School of Dentistry, and Calcitek, Inc, Carlsbad, CA. Address correspondence and reprint requests to Dr Block: Department of Oral and Maxillofacial Surgery, LSU School of Dentistry, 1100 Florida Ave, New Orleans, LA 70119. © 1990 American Association of Oral and Maxillofacial Surgeons 0278-2391/90/4802-0009$3.00/0

174

175

BLOCK,DELGADO,ANDFONTENOT

cancellous bone contact to the implant surface, or the influence of differences of implant length and diameter. This study analyzed implants of different sizes which were evaluated in posterior mandibular alveolar bone in order to provide both cortical and canceIlous bone contact to the implant. Materials and Methods IMPLANTS TESTED

Nine implant sizes were evaluated. All implants (Calcitek, Inc, Carlsbad, CA) were HA coated with a 75-lJ.m thick layer." The implants were either 3.0, 3.3, or 4.0 mm in diameter and either 4, 8, or 15 mm in length (Fig 1). The 8- and IS-mm-Iong implants had apical holes. The 4-mm-long implants had no apical flutes or other mechanical retentive features. All implants had internal threads for placement of an adapter for mechanical testing. ANIMAL MODEL

Six mongrel dogs (30 to 40 kg) were used for this study. Under general anesthesia, their mandibular

Table 1. Mechanical Testing Results Length

Diameter

No.

15 15 15 8 8 8 4 4 4

4 3.3 3 4 3.3 3 4 3.3 3

9 7 7 9 7 7 II 6 5

Pullout Force 38.06 ± 35.72 ± 34.60::!: 33.02 ± 25.78 ± 25.08 ± 17.50 ± 18.16 ± 14.94 ±

8.1 3.83 7.24 4.69 3.81 9.34 5.50 4.90 2.86

Force per Unit Area 130.36 148.23 157.97 211.65 200.43 214.52 224.61 282.45 255.76

± 28.12 ::!: 15.89 ::!: 33.06 ± 30.05 ::!: 29.64 ::!: 80.15 ::!: 70.67 ::!: 76.17 ::!: 48.89

Pullout force and force per unit area values are given as mean ± SD.

four premolars and their first and second molars were extracted. After 12weeks of healing, a cortical bone surface was present along the superior surface of the edentulated alveolar ridge. Using standardized techniques with graduating-size drills and internal irrigation, six to seven randomly sized implants (Table 1) were placed into both sides of the mandible, approximately 1 mm below the alveolar crest (Fig 2). The incisions were closed with 4-0 polyglactin sutures. AIl animals were given procaine penicillin for S days. The animals received a mush diet for the subsequent 15 weeks. After 15 weeks, the dogs were killed with an overdose of pentobarbital, and their mandibles were immediately retrieved for mechanical testing. MECHANICAL TESTING PROTOCOL

Immediately upon death, the mandibles were stripped of soft tissue, and lateral radiographs were taken (Fig 3). The bone that covered the implant heads was then removed with a curette (Fig 4). Hemimandibles were embedded in acrylic in a jig fabricated for this study. After the mandibles were places into the jig, bolts were placed through the acrylic to prevent movement of the mandible due to shrinkage of the acrylic. The jig was then placed into a pressure pot of warm water, and the acrylic

FIGURE I.

The nine implants evaluated.

FIGURE 2.

Implants placed into the mandible.

176

HA IMPLANT PULLOUT STUDIES

FIGURE 3. Lateral radiograph showing implant placement.

was allowed to set. After the acrylic hardened, the jig was placed on the platen of a Mechanical Testing System (MTS 810, Minneapolis, MN). The test apparatus allowed movement of the jig in three dimensions (Fig 5) to permit all of the implants to be independently aligned to their vertical axis during the pullout testing. An adapter was screwed into the implant body, a 24-inch straight rod was attached to the adapter (Fig 6), and a universaljoint attachment connected the rod to the load cell of the MTS 810. The universal joint allowed for correction of small nonaxial alignments. The shaft of the rod was aligned with the axis of the implant, and the pullout test was run using a displacement of 2 mm/min (Fig 7). The resultant force-displacement curve was plotted on a chart recorder; the maximum force recorded at the failure point was used for statistical analysis. The mechanical testing was completed within 6 hours of death. Following pullout, the implants and bone were placed in 95% ethanol for later evaluation. The ANOVA of the SAS statistical package (SAS Institute, Cary, NC) was used to analyze pullout force (measured in pounds) and force per unit area (pounds per square inch or megapascals). Force per unit area was calculated by dividing the pullout

FIGURE 4. Hemimandibles stripped of soft tissue before being put into the testing apparatus.

FIGURE 5. Universal testing apparatus for three-dimensional adjustment to ensure parallelism.

force.by the surface area of the implant. The surface area was approximated as a cylinder (3.1416 x diameter x implant length). This approximation did not include surface roughness or the bottom of the implant. However, because the areas of all implants were approximated in a similar fashion, force per unit area calculations could be compared within the study. These two force-dependent variables were analyzed using a two-way ANOVA. The two independent variables were diameter (three values) and length (three values). In addition to these two main effects, an interaction was also tested. If diameter or length was found to be significantly different, Duncan's multiple-range test was used to identify which of the three measurements were different from one another. Results

All of the implants were integrated, and bone completely covered to implants. This overlying bone was removed with a curette to expose the healing screw. Less than 1 mm of crestal bone loss was present on three 8-mm-Iong and two 15mm-long implants; this was less than 12% of the

FIGURE 6. Hemimandible in testing apparatus.

177

BLOCK, DELGADO, AND FONTENOT

FIGURE 7. surface.

Implant pulled from bone. Note bone on implant

implant length. All remaining implants had no crestal bone loss. The forces recorded at the failure point of the pullout curve were analyzed statistically. First, the relationship of diameter and length to pullout force was analyzed. The ANOVA showed no significant difference in the pullout force due to variation in diameter (F(2,67) == 1.27, P > .2876), but did show a significant difference due to length (F(4,67) == 58.62, P < .001). There was no significant interaction (F(4,67) == 0.95, P > .4397). This is graphically depicted in Fig 8. This figure and Duncan's multiple range test indicated that the 4-mm length implants

.-

~(

.. c . ... . r

c:::J 1) It

• tt.r n e [m) ' l'Ih n

FIGURE 8. Bar graph depicting pullout strength for each implant tested.

were extracted 'with significantly smaller forces than the longer 8-mm implants, and the 8-mm implants were again extracted with significantly smaller forces than the l5-mm implants. The two-way ANOV A analysis of the force per unit area calculations showed no significant difference in the force per unit area due to variation in diameter (F(2,67) == 0.66, P > .52), but did show a significant difference due to length (F(2,67) == 24.46, P < .001. There was no significant interaction (F(2,67) == 0.95, P > .4434). This is graphically depicted in Fig 9. This and Duncan's multiple-range test indicated that the 4-mm length implants had a greater force per unit area calculated value than the longer 8-mm implants, and the 8-mm implants had a greater force per unit area calculated value than the longer 15-mm implants. When the implants were extracted from the alveolar bone, several consistent observations were made. The implant separated from its crypt along both the HA-bone and HA-metal substrate interfaces. The HA-substrate interface separated along the superior (or crestal) aspect of the implant and also along the apical (or bottom) portion of the implant. The HA-bone interface separated along the middle or cancellous portion of the implant site. Discussion

Anatomic constraints for implant placement include limitations in bone width and height. Ideally, an implant should be completely surrounded by bone in order to facilitate optimal hard and soft tissue maintenance. When confronted with a narrow ridge, a thinner implant is desirable in order to have bone surround the implant body. Our pullout data indicate that diameter changes from 4 mm to 3.3 mm (17.5% decrease in diameter) were less significant than length changes of larger magnitude (15 mm to 8 mm == 46.7% decrease). If surface area is linearly related to pullout force, then one would expect that the pullout extraction force to be 46.7% less for the 8-mm implants than for the longer 15mm implants. However, the decrease in pullout extraction force for the 4- and 8-mm-Iong implants was not linearly related to the decrease in implant length when compared with the I5-mm-Iong implants. Thus, a linear relationship of ultimate pullout force to surface area was not observed, implying that other geometric and physical factors involving cortical and cancellous bone structure contributed to differences in pullout force between the different length implants. The force put unit area calculation represents an expression of the bone-implant shear and tensile strengths, and was not linearly related to surface area. The difference in the calculated force per unit

178

HA IMPLANT PULLOUT STUDIES

....

MPA 2.07 1.93 1. 79 1.65 1.52 1. 38 1.24 1.10 0.96

c..

0.83

'" .... Q)

FIGURE 9. Bar graph depicting force per unit area of each implant tested.

-

<:

c:

::::l

'" Q)

u

.... a

Ll..

0.69 0.55 0.41

PSI 300.--------------------+-----, 280 260 240 220 200 180 160 140 120 100 80 60

0.28 40 0.14 20

7 8

6 4

O~.........,.:L-,....__=~------"'..:uu..>o.1..._._~'--">L----''''-''-'-~_r--"''u:u......=

~

Length of Implants in mm

4 mm diameter

area in these implants may represent an unused or mechanically less significant portion of the bone implant interface, which would then result in a decrease in force per unit area since the pullout force was expressed per unit surface area. In the radiographs (Fig 3) denser bone can be seen in the crestal region, with portions of the 8-mm and I5-mm-Iong implants being located in a cancellous bone region. One may hypothesize that a less mechanically significant part of the implant exists within the cancellous bone. The force per unit area values were less than those reported from pushout studies in femur cortical bone. t •2 The mechanical difference between a dense femur cortex and alveolar bone was not examined in this study, but would be expected to be significantly different. Statistically this study indicates that the thinner implants had similar pullout strengths to wider implants of the same length. In addition, this study indicates that the mechanical effects of cortical and cancellous bone may not be equal. Clinical ramifications of the relatively small mechanical contribution of the cancellous bone in the posterior mandible of the dog may indicate that a larger number of implants may be required to support a posterior prosthesis when compared with the anterior mandible, where longer implants can be placed into two cortical bone layers

c:::::J 3.3 mm diameter

f;;:;:;::;::;::~

3 mm diameter

without disrupting the inferior alveolar neurovascular bundle.

Acknowledgment We would like to thank Don Hendricks for fabricating the universal testing apparatus used in this study.

References I. Cook SD, Kay JF, Thomas KA, et al: Interface mechanics

2.

3. 4. 5.

6. 7.

and histology of titanium and hydroxylapatite coated titanium for dental implant applications. Int J Oral Maxillofac Implants 2:15, 1987 Gerner BT, Barth E, Albrektsson T, et al: Comparison of bone reactions to coated tricalcium phosphate and pure titanium dental implants in the canine iliac crest. Scand J Dent Res 96:143, 1988 deGroot K, Geesink RGT, Klein CPAT, et al: Plasma sprayed coatings of hydroxylapatite. J Biomed Mater Res 21:1375, 1987 Geesink RGT, deGroot K, Klien CPAT: Chemical implant fixation using hydroxyl coatings. Clin Orthop 225:147, 1987 Thomas KA, Kay JF, Cook SD, et al: The effect of surface macrotexture and hydroxylapatite coating on the mechanical strength and histologic profile of titanium implant materials. J Biomed Mater Res 21:1395, 1987 Cook SD, Kay JF, Thomas KA, et al: Hydroxylapatite coated porous titanium for use as an orthopedic histologic attachment system. Clin Orthop 230:303, 1988 Block MS, Kent IN, Kay JF: Evaluation of hydroxylapatitecoated titanium dental implants in dogs. J Oral Maxillofac Surg 45:601, 1987