S c r e w - a u g m e n t e d F i x a t i o n of A c e t a b u l a r Components A Mechanical Model to Determine Optimal Screw Placement
S t e v e n K. S t r a n n e , M D , J o h n J. C a l l a g h a n , M D , S t e v e n H. E l d e r , BS, R i c h a r d R. G l i s s o n , BS, a n d A n t h o n y V. S e a b e r
Abstract: Sixteen embalmed hemipelves were used to determine the optimal acetabular screw placement to provide maximal screw pull-out ztrength in uniconical and biconical screw fixation. The anterior column, superior ilium, posterior column, ischium, and pubis regions of the pelvis were tested using 6.5mm titanium alloy screws and a hydraulic servo-controlled 1321 Instron testing machine. Force vs displacement data were acquired. Bicortical fixation was stronger than unicortical fixation in the four zones compared. This difference 9was s!gnificant in the superior ilium, posterior column, and ischium. The anterior column could not accept unicortical screws due to inadequate bone depth, whichranged between only 6 mm and 10 mm. Bicortical fixation was significantly greater in the superior ilium, posterior column, and ischium than in the anterior column or pubis. Unicortical fixation was greatest in the superior ilium. This information may aid decisions concerning the positioning of screws to augment acetabnlar component fixation. Key words: total hip anhroplasty, screwaugmented acetabu]ar component
the consideration of optimal screw placement in the acetabular region. With respect to mechanical considerations, Lachiewicz et al. investigated the strength of fixation of various acetabular t:up designs under torsional loading. The screw-augmented u n c e m e n t e d design was delermined to be the most stable u n c e m e n t e d design, with three screws preferable to two. However, specific placement of the screws was not reported (11). Other mechanical studies concerning techniques of acetabular fracture stabilization (4, 14, 16) provide little information to aid in the optimization of screw placement in the hip arthroplasty condition. With respect to anatomic considerations, o u r laboratory has conducted studies to identify neurovas-
In primary and revision u n c e m e n t e d total hip arthroplasty (THA), b o n e screws have been employed to augment the initial fixation of acetabular components to aid in the stability of the implant, which is reported to be essential for bony ingrowth (2). The complex external and internal architecture of the b o n y pelvis and the close proximity of intrapelvic neurovascular structures are important variables in
From the Orthopaedic Research Laboratory, Duke University Medical Cenler, Durham, North Carolina.
This project was funded by a grant from Zimmer USA and the Piedmont Orthopaedic Society. Reprint requests: John J. Callaghan, biD, Department of Orthopaedics, University of Iowa Hospitals, Iowa City, IA 52242.
301
302
The Journal Of Arthroplasty Vol. 6 No. 4 December 1991
cular structures at risk of inadvertent screw penetration and to quantify distances between these structures and the inner pelvic wall. The posterior acetabulum, comprised of the superior ilium, posterior iliac column, and ischium, was determined to be the safest area for screw placement from the standpoint of anatomic structures at risk Of injury. In the vicinity of the anterior column, the lilac vein passes within 5 mm of the inner pelvic wall without soft-tissue protection, while the iliac artery courses within 10 mm of the bony pelvis with some softtissue protection from the iliacus muscle. Near the pubis, the obturator vessels pass within 5 mm of the bone. Neurovascular structures associated with the posterior acetabulum, such as the superior gluteal artery and the sciatic nerve, do not pass as closely to the bony pelvis nor are more easily visualized during THA procedures than those vessels described anteriorly (9). Previous investigators have questioned also the safety of bone screws in the anterior acetabulum (8, 18). The goals of this investigation were to compare the initial mechanical strength of screw implantation between various zones of the acetabulum, to compare the initial mechanical strength between bicortical and unicortical acetabular screw placement, and to compare bone depth available for screw implan-ration in different zones of the acetabulum.
Materials and Methods Sixteen hemipelves, devoid of evidence of bone disease, were obtained for mechanical study from twelve embalmed cadavers. Average age was 70.4 years (range, 37-84 years). All soft tissue and articular cartilage was removed manually, leaving only the bony pelvis, including the subchondral plate, intact. Each acetabulum was divided into five zones, defined to allow accurate interspecimen comparison (Fig: 1). The zones separated each hemipelvis into components consisting of the anterior lilac column, superior ilium, posterior iliac column, ischium, and pubis. Drill holes, 3.2 mm in diameter, were created from the geometric center of each zone. All holes were created by the same surgeon using standard handheld power equipment. Half of the hemipelves were designated for bicortical testing and half for uniconical study. Bicortical holes were oriented so as to maximize bone depth, passing through both the subchondral plate and the far cortex. An attempt was made to orient the holes perpendicular to the sub-
2
Fig. 1. The five zones of the pelvis. Zone 1, the anterior column extending to the anterior superior iliac spine; zone 2, the superior ilium; zone 3, the posterior column; zone 4, the ischium extending to the ischial tuberosity; and zone 5, the pubis.
chondral bone. Uniconical holes were oriented so as to maximize bone depth without penetrating the far cortex. Drilling ceased if resistance was encountered due to contact with the far cortex. No unicortical screw testing was performed in zone 1 due to inadequate bone depth. Thus, five tests were performed on each of the eight bicortical hemipelves, and four tests were performed on each of the eight unicortical hemipelves. Hole depth was quantified using a standard depth gauge. To facilitate mounting of the specimens, each hemipelvis was sectioned by handsaw into three pieces corresponding to the ilium, ischium, and pubis. Care was taken such that drill holes were not within 15 mm of any cut edge (less than two screw diameters). Rigid mounting was achieved by embedding the specimens in a fast-curing polyester resin. Caverns were made in the polymer to allow bicortical screws to pass through the far cortex without contacting polyester. Caverns were created by placing foam spacers over and around each drill-hole exit through the far cortex. The foam was removed after the polyester hardened and before screw implantation. A lack of contact between all screws and the mounting polyester was confirmed by direct visualization. The caverns prevented polyester from contacting the far cortex of the
Screw-augmented Fixation of Acetabular Components
bony specimen within 15 m m of any bicortical screw hole. Identical 6.5-mm titanium alloy screws (Zimmer) were used in all tests and were inserted with a standard hand-held screw driver. In biconical placement, the end of each screw extended past the far cortex until full-sized threads engaged bone throughout the length of s c r e w - b o n e contact. In uniconical placement, each screw was inserted until the tip reached the depth of the uniconical drill hole, as previously determined by depth gauge. Screw-tip design provided 2 m m of distance between the point and the beginning of the threads, thus allowing some margin of error in depth of insertion of unicortical screws without engaging the far cortex. No unicortical screw was inserted past a depth of 30 mm, regardless of location of the far cortex. The testing rig consisted of a universal joint, an attachment to transfer axial tensile force to the undersurface of each screw head, and a specially designed mounting device. The mounting device al-
Fig. 2. A hemipelvis mounted inthe jig with a screw prepared for pull-out testing.
9 Stranne et al.
303
lowed both accurate alignment of each screw axis with the axis of test and also rigid mounting of each of the variously sized and shaped specimens (Fig. 2). Uniaxial pull-out testing was performed using a hydraulic servo-controlled 1321 Instron testing machine at a constant rate of 10 mm/s. Force vs displacement data were acquired each millisecond and stored digitally on a computer. The order of testing was randomized between specimens in zones 1, 2, and 3, which were mounted as one piece. Specimens were maintained uniformly moist throughout testing by both spraying exposed surfaces with normal saline solution periodically and wrapping specimens in plastic during preparation for mechanical tests.
Results In all cases, pull-out testing resulted in extraction of a cylindrical bone plug in the threads of the screw. Throughout the cancellous bone, failure occurred along the s c r e w - b o n e interface at the outer diameter of the screw threads. Gross examination revealed no disruption of the cancellous bone beyond 1 m m of the screw diameter. However, a conical torus of the subchondral plate, approximately IO m m in diameter, was extracted with the proximal end of the bone plug in the majority of tests. The m a x i m u m load carried by the specimen was defined to be the pull-out strength. In no instance did catastrophic failure of the titanium screws occur. In one specimen, cracking of the bone was evident beyond the outer diameter of the screw threads. This specimen was discarded. Mean biconical strengths were greater than m e a n unicortical strengths in the four regions in which both bicortical and uniconical testing were conducted (Fig. 3). Using a repeated-measures analysis. of variance (ANOVA), this difference was statistically significant in zone 2 (the superior ilium, P = .013), zone 3 (the posterior iliac column, P = .0001), and zone 4 (the ischium, P = .0001). Although this trend toward increased bicortical strength occurred in zone 5 (the pubis, P = .34), statistical significance was not achieved. With respect to biconical strength, comparison between t h e five zones revealed mean strengths to be greater in the posterior acetabulum, comprised of the superior ilium, the posterior lilac column, and the ischium, than either the anterior iliac column or the pubis. Statistical comparisons by repeated-measures ANOVA between each of the five zones confirmed zones 2, 3, and 4 to be significantly stronger than zones 1 or 5, with no P value exceeding .0026.
304
The Journal of Arthroplasty Vol. 6 No. 4 December 1991
3000
9
conical placemeht, as determined by repeated-mi~asures ANOVA.
[]
o
2000 o o o U.
1000
Discussion
T =
g o
2
t
3
4
5
Acetabular Zones
Fig. 3. Comparison of biconical and uniconical pullout strengths. Statistically significant differences between bicortical and unicortical values were found in zone 2 (P = .013), zone 3 (P = .0001), and zone4 (P = .0001). Comparison of bicortical values revealed zones 2, 3, and 4 to be significantly stronger than zones l or 5 (P -< .026). Comparison of unicortical values found zone 2 to be significantly stronger than all other zones (P -< .012).
Comparison between the pull-out strengths found in the four zones of the uniconical testing revealed zone 2 (the superior ilium) to be significantly stronger than any of the other three zones, with no P value exceeding .012. Depth of drill holes created for bicortical placement ranged from 6 m m to 37 mm (Fig. 4). Zone 1 (the anterior column) consistently presented the least available depth, ranging between 6 m m and 10 mm. Zone 4 (the ischium) had the greatest range of any one zone, ranging from 10 m m to 37 ram. In unicortical placement, zone 2 and zone 4 screws were all inserted to a depth of 30 m m without contacting the far cortex. Zone 3 screws averaged 22.7 + 1.4 mm, and zone 5 averaged 28.9 _ 4.1 mm. No relationship between depth and pull-out strength was identified in either unicortical or bi-
40
~, 30
II |
-
E
..=
"~
20
!
rr
!1
10
1
2
3
4
5
Acetabular Zone Fig. 4. Range of biconica] screw-depth values in the five zones with mean values displayed.
Using a uniaxial pull-out model of screw failure, the posterior acetabulum, consisting of the superior ilium, the posterior column and tile ischium, was found to provide maximal pull-out strength in biconical screw placement, as did the superior ilium in uniconical screw placement. Moreover, previous evaluation of anatomic relationships suggests a greater safety factor in the posterior regions with respect to inadvertent penetration of neurovascular structures in the posterior regions (8, 9, 18). In all zones, biconical placement was stronger than uniconical placement. This paper represents the employment of biomechanical methods to define optim01 positioning of acetabular screws. Detemaination of mechanical purchase, both in biological and engineering structures, traditionally has been measured by pull-out testing (I, 3, 4, 6, 7, 10, 12, 14, 15, 17, 19). W h e n screw geometry, implantation technique, predrilled hole size, and rate of withdrawal are identical, pull-out strength directly reflects local bone strength and length of contact between screw and bone. The local strength of bone is a function of the anisotropic nature of bone, bone density, and the proportions of conical and cancellous bone engaged in the screw threads (1, 17, 19). Loads experienced by in vivo acetabular screws are complex, cyclic, and of varying amplitude and direction. Although uniaxial pull-out does not recreate all of these in vivo variables, pullout strength quantifies an important parameter indicative of screw purchase and provides a reproducible method of comparing screw placement in various regions at various depths. The exact nature of forces experienced by in vivo acetabular screws throughout the regions of the acetabulum have not been defined; however, additional modes of failure, such as shear forces, may be significant with respect to implant stability. This study does not determine if regions of superior pull-out strength are also superior in resistance to other potential modes of failure. This model provides a measure of tile initial screw purchase. In vivo phenomena, such as structural remodeling along lines of principle stress or fibrou., tissue growth, are not modeled in this study. In addition, in order to provide a reproducible model the acetabulae were not reamed. Any partial reaming of the subchondral plate, as occurs clinically, would alter the magnitudes found in this investigation and actually may accentuate the difference found in this
Screw-augmented Fixation of Acetabular Components
investigation between bicortical and unicortical placement. The e m b a l m e d nature of the specimens in this study m a y alter the absolute magnitudes found in mechanical testing (5, 13); however, relative comparisons of pull-out testing are still valid. The standard deviations found in this investigation are consistent with previous biomechanical screw pull-out investigations (1, 3, 4, 6, 7, I0, 12, 14, 15, 18). Such ranges are expected due to the k n o w n significant intraspecimen variability in bone material properties and geometric structure. Potential sources of error in this study include bit wobble during predrilling, misalignment of the screw axis with that of the hole, misalignment of the screw axis with the axis of testing, and mounting restraints. An attempt was m a d e to minimize the effect of m o u n t i n g by distancing m o u n t i n g polymer from protruding bicortical screws and by distancing cut edges from drill holes. No direct relationship between depth and pull-out strength was established in either unicortical or bicortical placement. The depth of cortical bone was not measured independently from t h e cancellous bone. Such a result is consistent with the conclusion that the cortical b o n e carries significantly greater load, especially with respect to load carried per unit depth, than the cancellous bone. Differences seen between the regions m a y reflect either a differential in depth of cortical b o n e or a differential in cortical bone mechanical properties between the regions. Regardless of which hypothesis is correct, this study identifies the acetabular regions that provide overall superior mechanical screw purchase. The length of screw contact in cancellous bone might b e c o m e significant in cases in which a significant a m o u n t of the subchondral plate is removed, as in revision arthroplasty or complex primary THAs, especially if the far cortex is not engaged. This information, along with the anatomic information obtained in other studies from our laboratory (9), m a y aid decisions concerning the positioning of screws to a u g m e n t acetabular c o m p o n e n t fixation. Bicortical placement was found to provid e superior screw pull-out strength in all zones. In bicortical fixation, the posterior half of the acetabulum provides not only less risk for inadvertent penetration of neurovascular structures but also the greatest pull-out strength.
References 1. AnseIl RH, Scales JT: A study of some factors which affect the strength of screws and their insertion and holding power in bone. J Biomech 1:279, 1968 2. Bobyn JD, Pilliar RM, Cameron llV, Weatherly GC: The optimum pore size for the fixation of porous-sur-
3.
4.
5.
6.
7.
8.
9.
I0.
I I.
12. 13.
14.
15.
16.
17.
18.
19.
9 Stranne et al.
305
faced metal implants by the ingrowth of bone. Clin Orthop 150:263, 1980 Cameron HV, Jacab R, MacNab I, Pilliar RM: Use of polymethylmethacrylate to enhance screw fixation in bone. J Bone Joint Surg 57A:655, 1975 Daum WJ, Tencer AF, Cartwright TJ et al: Pull-out strengths of bone screws at various sites about the pelvis: a preliminary study. J Orthop Trauma 2:229, 1988 Evans FG: Significant differences in the tensile strength of adult human compact bone. p. 391. In Blackwood tlJJ (ed): Proceedings of the First European Bone and Tooth Symposium, Oxford, Pergamon, 1964 Frandsen PA, Christoffersen 1t, Madsen T: Holding power of different screws in the femoral head. Acta Orthop Scand 55:349, 1984 tlughes AN, Jordan BA: The mechanical properties of surgical bone screws and some aspects of insertion practice. Injury 4:25, 1972 Keating EM, Ritter MR, Faris PM: Structures at risk from medially placed acetabular screws. J Bone Joint Surg 72A:509, 1990 Kirpatrick JS, Callaghan J J, Vandemark RM, Goldner RD: The relationship of the intrapelvic vasculature to the acetabulum: implications in screw fixation acetabular components. Clin Orthop 258:183, 1990 Koranyi E, Bowman CE, Knecht CD, Janssen M: ttolding power of orthopedic screws in bone. Clin Orthop 72:283, 1970 Lachiewicz PF, Suh PB, Gilbert JA: In vitro initial fixation of cemented and porous coated acetabular total hip components: a biomechanical comparative study. J Anhroplasty 4:201, 1989 Lyon WF, Cochran JR, Smith L: Actual holding power of various screws in bone. Ann Surg 114:376, 1941 McElhaney J, Fogle J, Byars E, Weaver G: Effect of embalming on the mechanical properties of beef bone. J Appl Physiol 19:1234, 1964 Sawaguchi T, Brown TD, Rubash liE, Mears DC: Stability of acetabular fractures after internal fixation. Acta Orthop Scand 55:601, 1984 Schatzker J, Sanderson R, Murnaghan JP: The holding power of orthopaedic screws in vivo. Clin Orthop 108:115, 1975 Shaw JA, Mino DE, Werner FW, Murry DG: Posterior stabilization of pelvic fractures by use of threaded compression rods. Clin Orthop 192:240, 1985 Standard methods of testing mechanical fasteners in wood. p. D1761. In Book of ASTM Standards, Philadelphia, American Society of Testing and Materials, 1986 Wasielewski RC, Cooperstein LA, Kruger MP, Rubash liE: Acetabular anatomy and transacetabular fixation of screws in total hip anhroplasty. J Bone Joint Surg 72A:501, 1990 Zindrick MR, Wiltse LL, Widell Etl et al: A biomechanical study of intrapendun,dular screw fixation in the lumbosacral spine. Clin Orthop 203:99, 1986
S t e v e n K. S t r a n n e , M D , J o h n J. C a l l a g h a n , M D , S t e v e n H. E l d e r , BS, R i c h a r d R. G l i s s o n , BS, a n d A n t h o n y V. S e a b e r
Abstract: Sixteen embalmed hemipelves were used to determine the optimal acetabular screw placement to provide maximal screw pull-out ztrength in uniconical and biconical screw fixation. The anterior column, superior ilium, posterior column, ischium, and pubis regions of the pelvis were tested using 6.5mm titanium alloy screws and a hydraulic servo-controlled 1321 Instron testing machine. Force vs displacement data were acquired. Bicortical fixation was stronger than unicortical fixation in the four zones compared. This difference 9was s!gnificant in the superior ilium, posterior column, and ischium. The anterior column could not accept unicortical screws due to inadequate bone depth, whichranged between only 6 mm and 10 mm. Bicortical fixation was significantly greater in the superior ilium, posterior column, and ischium than in the anterior column or pubis. Unicortical fixation was greatest in the superior ilium. This information may aid decisions concerning the positioning of screws to augment acetabnlar component fixation. Key words: total hip anhroplasty, screwaugmented acetabu]ar component
the consideration of optimal screw placement in the acetabular region. With respect to mechanical considerations, Lachiewicz et al. investigated the strength of fixation of various acetabular t:up designs under torsional loading. The screw-augmented u n c e m e n t e d design was delermined to be the most stable u n c e m e n t e d design, with three screws preferable to two. However, specific placement of the screws was not reported (11). Other mechanical studies concerning techniques of acetabular fracture stabilization (4, 14, 16) provide little information to aid in the optimization of screw placement in the hip arthroplasty condition. With respect to anatomic considerations, o u r laboratory has conducted studies to identify neurovas-
In primary and revision u n c e m e n t e d total hip arthroplasty (THA), b o n e screws have been employed to augment the initial fixation of acetabular components to aid in the stability of the implant, which is reported to be essential for bony ingrowth (2). The complex external and internal architecture of the b o n y pelvis and the close proximity of intrapelvic neurovascular structures are important variables in
From the Orthopaedic Research Laboratory, Duke University Medical Cenler, Durham, North Carolina.
This project was funded by a grant from Zimmer USA and the Piedmont Orthopaedic Society. Reprint requests: John J. Callaghan, biD, Department of Orthopaedics, University of Iowa Hospitals, Iowa City, IA 52242.
301
302
The Journal Of Arthroplasty Vol. 6 No. 4 December 1991
cular structures at risk of inadvertent screw penetration and to quantify distances between these structures and the inner pelvic wall. The posterior acetabulum, comprised of the superior ilium, posterior iliac column, and ischium, was determined to be the safest area for screw placement from the standpoint of anatomic structures at risk Of injury. In the vicinity of the anterior column, the lilac vein passes within 5 mm of the inner pelvic wall without soft-tissue protection, while the iliac artery courses within 10 mm of the bony pelvis with some softtissue protection from the iliacus muscle. Near the pubis, the obturator vessels pass within 5 mm of the bone. Neurovascular structures associated with the posterior acetabulum, such as the superior gluteal artery and the sciatic nerve, do not pass as closely to the bony pelvis nor are more easily visualized during THA procedures than those vessels described anteriorly (9). Previous investigators have questioned also the safety of bone screws in the anterior acetabulum (8, 18). The goals of this investigation were to compare the initial mechanical strength of screw implantation between various zones of the acetabulum, to compare the initial mechanical strength between bicortical and unicortical acetabular screw placement, and to compare bone depth available for screw implan-ration in different zones of the acetabulum.
Materials and Methods Sixteen hemipelves, devoid of evidence of bone disease, were obtained for mechanical study from twelve embalmed cadavers. Average age was 70.4 years (range, 37-84 years). All soft tissue and articular cartilage was removed manually, leaving only the bony pelvis, including the subchondral plate, intact. Each acetabulum was divided into five zones, defined to allow accurate interspecimen comparison (Fig: 1). The zones separated each hemipelvis into components consisting of the anterior lilac column, superior ilium, posterior iliac column, ischium, and pubis. Drill holes, 3.2 mm in diameter, were created from the geometric center of each zone. All holes were created by the same surgeon using standard handheld power equipment. Half of the hemipelves were designated for bicortical testing and half for uniconical study. Bicortical holes were oriented so as to maximize bone depth, passing through both the subchondral plate and the far cortex. An attempt was made to orient the holes perpendicular to the sub-
2
Fig. 1. The five zones of the pelvis. Zone 1, the anterior column extending to the anterior superior iliac spine; zone 2, the superior ilium; zone 3, the posterior column; zone 4, the ischium extending to the ischial tuberosity; and zone 5, the pubis.
chondral bone. Uniconical holes were oriented so as to maximize bone depth without penetrating the far cortex. Drilling ceased if resistance was encountered due to contact with the far cortex. No unicortical screw testing was performed in zone 1 due to inadequate bone depth. Thus, five tests were performed on each of the eight bicortical hemipelves, and four tests were performed on each of the eight unicortical hemipelves. Hole depth was quantified using a standard depth gauge. To facilitate mounting of the specimens, each hemipelvis was sectioned by handsaw into three pieces corresponding to the ilium, ischium, and pubis. Care was taken such that drill holes were not within 15 mm of any cut edge (less than two screw diameters). Rigid mounting was achieved by embedding the specimens in a fast-curing polyester resin. Caverns were made in the polymer to allow bicortical screws to pass through the far cortex without contacting polyester. Caverns were created by placing foam spacers over and around each drill-hole exit through the far cortex. The foam was removed after the polyester hardened and before screw implantation. A lack of contact between all screws and the mounting polyester was confirmed by direct visualization. The caverns prevented polyester from contacting the far cortex of the
Screw-augmented Fixation of Acetabular Components
bony specimen within 15 m m of any bicortical screw hole. Identical 6.5-mm titanium alloy screws (Zimmer) were used in all tests and were inserted with a standard hand-held screw driver. In biconical placement, the end of each screw extended past the far cortex until full-sized threads engaged bone throughout the length of s c r e w - b o n e contact. In uniconical placement, each screw was inserted until the tip reached the depth of the uniconical drill hole, as previously determined by depth gauge. Screw-tip design provided 2 m m of distance between the point and the beginning of the threads, thus allowing some margin of error in depth of insertion of unicortical screws without engaging the far cortex. No unicortical screw was inserted past a depth of 30 mm, regardless of location of the far cortex. The testing rig consisted of a universal joint, an attachment to transfer axial tensile force to the undersurface of each screw head, and a specially designed mounting device. The mounting device al-
Fig. 2. A hemipelvis mounted inthe jig with a screw prepared for pull-out testing.
9 Stranne et al.
303
lowed both accurate alignment of each screw axis with the axis of test and also rigid mounting of each of the variously sized and shaped specimens (Fig. 2). Uniaxial pull-out testing was performed using a hydraulic servo-controlled 1321 Instron testing machine at a constant rate of 10 mm/s. Force vs displacement data were acquired each millisecond and stored digitally on a computer. The order of testing was randomized between specimens in zones 1, 2, and 3, which were mounted as one piece. Specimens were maintained uniformly moist throughout testing by both spraying exposed surfaces with normal saline solution periodically and wrapping specimens in plastic during preparation for mechanical tests.
Results In all cases, pull-out testing resulted in extraction of a cylindrical bone plug in the threads of the screw. Throughout the cancellous bone, failure occurred along the s c r e w - b o n e interface at the outer diameter of the screw threads. Gross examination revealed no disruption of the cancellous bone beyond 1 m m of the screw diameter. However, a conical torus of the subchondral plate, approximately IO m m in diameter, was extracted with the proximal end of the bone plug in the majority of tests. The m a x i m u m load carried by the specimen was defined to be the pull-out strength. In no instance did catastrophic failure of the titanium screws occur. In one specimen, cracking of the bone was evident beyond the outer diameter of the screw threads. This specimen was discarded. Mean biconical strengths were greater than m e a n unicortical strengths in the four regions in which both bicortical and uniconical testing were conducted (Fig. 3). Using a repeated-measures analysis. of variance (ANOVA), this difference was statistically significant in zone 2 (the superior ilium, P = .013), zone 3 (the posterior iliac column, P = .0001), and zone 4 (the ischium, P = .0001). Although this trend toward increased bicortical strength occurred in zone 5 (the pubis, P = .34), statistical significance was not achieved. With respect to biconical strength, comparison between t h e five zones revealed mean strengths to be greater in the posterior acetabulum, comprised of the superior ilium, the posterior lilac column, and the ischium, than either the anterior iliac column or the pubis. Statistical comparisons by repeated-measures ANOVA between each of the five zones confirmed zones 2, 3, and 4 to be significantly stronger than zones 1 or 5, with no P value exceeding .0026.
304
The Journal of Arthroplasty Vol. 6 No. 4 December 1991
3000
9
conical placemeht, as determined by repeated-mi~asures ANOVA.
[]
o
2000 o o o U.
1000
Discussion
T =
g o
2
t
3
4
5
Acetabular Zones
Fig. 3. Comparison of biconical and uniconical pullout strengths. Statistically significant differences between bicortical and unicortical values were found in zone 2 (P = .013), zone 3 (P = .0001), and zone4 (P = .0001). Comparison of bicortical values revealed zones 2, 3, and 4 to be significantly stronger than zones l or 5 (P -< .026). Comparison of unicortical values found zone 2 to be significantly stronger than all other zones (P -< .012).
Comparison between the pull-out strengths found in the four zones of the uniconical testing revealed zone 2 (the superior ilium) to be significantly stronger than any of the other three zones, with no P value exceeding .012. Depth of drill holes created for bicortical placement ranged from 6 m m to 37 mm (Fig. 4). Zone 1 (the anterior column) consistently presented the least available depth, ranging between 6 m m and 10 mm. Zone 4 (the ischium) had the greatest range of any one zone, ranging from 10 m m to 37 ram. In unicortical placement, zone 2 and zone 4 screws were all inserted to a depth of 30 m m without contacting the far cortex. Zone 3 screws averaged 22.7 + 1.4 mm, and zone 5 averaged 28.9 _ 4.1 mm. No relationship between depth and pull-out strength was identified in either unicortical or bi-
40
~, 30
II |
-
E
..=
"~
20
!
rr
!1
10
1
2
3
4
5
Acetabular Zone Fig. 4. Range of biconica] screw-depth values in the five zones with mean values displayed.
Using a uniaxial pull-out model of screw failure, the posterior acetabulum, consisting of the superior ilium, the posterior column and tile ischium, was found to provide maximal pull-out strength in biconical screw placement, as did the superior ilium in uniconical screw placement. Moreover, previous evaluation of anatomic relationships suggests a greater safety factor in the posterior regions with respect to inadvertent penetration of neurovascular structures in the posterior regions (8, 9, 18). In all zones, biconical placement was stronger than uniconical placement. This paper represents the employment of biomechanical methods to define optim01 positioning of acetabular screws. Detemaination of mechanical purchase, both in biological and engineering structures, traditionally has been measured by pull-out testing (I, 3, 4, 6, 7, 10, 12, 14, 15, 17, 19). W h e n screw geometry, implantation technique, predrilled hole size, and rate of withdrawal are identical, pull-out strength directly reflects local bone strength and length of contact between screw and bone. The local strength of bone is a function of the anisotropic nature of bone, bone density, and the proportions of conical and cancellous bone engaged in the screw threads (1, 17, 19). Loads experienced by in vivo acetabular screws are complex, cyclic, and of varying amplitude and direction. Although uniaxial pull-out does not recreate all of these in vivo variables, pullout strength quantifies an important parameter indicative of screw purchase and provides a reproducible method of comparing screw placement in various regions at various depths. The exact nature of forces experienced by in vivo acetabular screws throughout the regions of the acetabulum have not been defined; however, additional modes of failure, such as shear forces, may be significant with respect to implant stability. This study does not determine if regions of superior pull-out strength are also superior in resistance to other potential modes of failure. This model provides a measure of tile initial screw purchase. In vivo phenomena, such as structural remodeling along lines of principle stress or fibrou., tissue growth, are not modeled in this study. In addition, in order to provide a reproducible model the acetabulae were not reamed. Any partial reaming of the subchondral plate, as occurs clinically, would alter the magnitudes found in this investigation and actually may accentuate the difference found in this
Screw-augmented Fixation of Acetabular Components
investigation between bicortical and unicortical placement. The e m b a l m e d nature of the specimens in this study m a y alter the absolute magnitudes found in mechanical testing (5, 13); however, relative comparisons of pull-out testing are still valid. The standard deviations found in this investigation are consistent with previous biomechanical screw pull-out investigations (1, 3, 4, 6, 7, I0, 12, 14, 15, 18). Such ranges are expected due to the k n o w n significant intraspecimen variability in bone material properties and geometric structure. Potential sources of error in this study include bit wobble during predrilling, misalignment of the screw axis with that of the hole, misalignment of the screw axis with the axis of testing, and mounting restraints. An attempt was m a d e to minimize the effect of m o u n t i n g by distancing m o u n t i n g polymer from protruding bicortical screws and by distancing cut edges from drill holes. No direct relationship between depth and pull-out strength was established in either unicortical or bicortical placement. The depth of cortical bone was not measured independently from t h e cancellous bone. Such a result is consistent with the conclusion that the cortical b o n e carries significantly greater load, especially with respect to load carried per unit depth, than the cancellous bone. Differences seen between the regions m a y reflect either a differential in depth of cortical b o n e or a differential in cortical bone mechanical properties between the regions. Regardless of which hypothesis is correct, this study identifies the acetabular regions that provide overall superior mechanical screw purchase. The length of screw contact in cancellous bone might b e c o m e significant in cases in which a significant a m o u n t of the subchondral plate is removed, as in revision arthroplasty or complex primary THAs, especially if the far cortex is not engaged. This information, along with the anatomic information obtained in other studies from our laboratory (9), m a y aid decisions concerning the positioning of screws to a u g m e n t acetabular c o m p o n e n t fixation. Bicortical placement was found to provid e superior screw pull-out strength in all zones. In bicortical fixation, the posterior half of the acetabulum provides not only less risk for inadvertent penetration of neurovascular structures but also the greatest pull-out strength.
References 1. AnseIl RH, Scales JT: A study of some factors which affect the strength of screws and their insertion and holding power in bone. J Biomech 1:279, 1968 2. Bobyn JD, Pilliar RM, Cameron llV, Weatherly GC: The optimum pore size for the fixation of porous-sur-
3.
4.
5.
6.
7.
8.
9.
I0.
I I.
12. 13.
14.
15.
16.
17.
18.
19.
9 Stranne et al.
305
faced metal implants by the ingrowth of bone. Clin Orthop 150:263, 1980 Cameron HV, Jacab R, MacNab I, Pilliar RM: Use of polymethylmethacrylate to enhance screw fixation in bone. J Bone Joint Surg 57A:655, 1975 Daum WJ, Tencer AF, Cartwright TJ et al: Pull-out strengths of bone screws at various sites about the pelvis: a preliminary study. J Orthop Trauma 2:229, 1988 Evans FG: Significant differences in the tensile strength of adult human compact bone. p. 391. In Blackwood tlJJ (ed): Proceedings of the First European Bone and Tooth Symposium, Oxford, Pergamon, 1964 Frandsen PA, Christoffersen 1t, Madsen T: Holding power of different screws in the femoral head. Acta Orthop Scand 55:349, 1984 tlughes AN, Jordan BA: The mechanical properties of surgical bone screws and some aspects of insertion practice. Injury 4:25, 1972 Keating EM, Ritter MR, Faris PM: Structures at risk from medially placed acetabular screws. J Bone Joint Surg 72A:509, 1990 Kirpatrick JS, Callaghan J J, Vandemark RM, Goldner RD: The relationship of the intrapelvic vasculature to the acetabulum: implications in screw fixation acetabular components. Clin Orthop 258:183, 1990 Koranyi E, Bowman CE, Knecht CD, Janssen M: ttolding power of orthopedic screws in bone. Clin Orthop 72:283, 1970 Lachiewicz PF, Suh PB, Gilbert JA: In vitro initial fixation of cemented and porous coated acetabular total hip components: a biomechanical comparative study. J Anhroplasty 4:201, 1989 Lyon WF, Cochran JR, Smith L: Actual holding power of various screws in bone. Ann Surg 114:376, 1941 McElhaney J, Fogle J, Byars E, Weaver G: Effect of embalming on the mechanical properties of beef bone. J Appl Physiol 19:1234, 1964 Sawaguchi T, Brown TD, Rubash liE, Mears DC: Stability of acetabular fractures after internal fixation. Acta Orthop Scand 55:601, 1984 Schatzker J, Sanderson R, Murnaghan JP: The holding power of orthopaedic screws in vivo. Clin Orthop 108:115, 1975 Shaw JA, Mino DE, Werner FW, Murry DG: Posterior stabilization of pelvic fractures by use of threaded compression rods. Clin Orthop 192:240, 1985 Standard methods of testing mechanical fasteners in wood. p. D1761. In Book of ASTM Standards, Philadelphia, American Society of Testing and Materials, 1986 Wasielewski RC, Cooperstein LA, Kruger MP, Rubash liE: Acetabular anatomy and transacetabular fixation of screws in total hip anhroplasty. J Bone Joint Surg 72A:501, 1990 Zindrick MR, Wiltse LL, Widell Etl et al: A biomechanical study of intrapendun,dular screw fixation in the lumbosacral spine. Clin Orthop 203:99, 1986
