Distal stem design and the torsional stability of cementless femoral stems

The Journal of Arthroplasty Vol. 10 No. 4 1995

Distal S t e m D e s i g n and the Torsional Stability of C e m e n t l e s s Femoral S t e m s J. B a r t o n K e n d r i c k

II, M D , P h i l i p C. N o b l e , M S , a n d H u g h

S. T u l l o s , M D

Abstract: In cementless hip arthroplasty, the fit between the implant and the endosteal cavity is a critical determinant of implant stability. Although cementless implants may be stabilized through proximal fit within the metaphysis, many surgeons rely on diaphyseal fixation to provide the necessary resistance to rotational forces, especially in revision hip arthroplasty. The cross-sectional design of the femoral stem at the level of the femoral isthmus was investigated with respect to its effect on the rotational stability of the bone-stem interface. Four cross-sectional designs--a fluted stem, a finned stem, a porous-coated stem, and a slotted fluted stem--were implanted in 12 cadaveric femurs and loaded in torsion. A knurled stem, cemented into each specimen at the conclusion of testing, acted as a control stem. The torque required to cause 100 t~m of displacement at the bone stem-interface ranged from 13.7 + 0.8 N-m with the porous-coated design to 30.1 _+3.7 N-m with the fluted design (P < .0001). intermediate values of 19.5 + 1.4 and 19.9 _+2.3 N-m were observed with the finned and slotted fluted designs, respectively. In all of the cemented control stems, failure occurred at the bone-cement interface at an average torque of 34.0 ± 3.0 N-m. Statistical analysis demonstrated that the porouscoated, finned, and slotted fluted designs were all significantly weaker in torsion than the cemented control stem; however, there was no significant difference between the torsional resistance of the solid fluted (unslotted) and cemented stems. With the exception of the fluted stem design, it is postulated that the cementless stem configurations evaluated would provide insufficient resistance to torsional forces to stabilize a femoral prosthesis solely through distal fixation within the medullary canal. Consequently, rotational stabilization of the cementless prosthesis necessitates proximal and distal contact between the implant and the femur. K e y words: prosthesis, cementless fixation, biomechanics, torsional stability, total hip arthroplasty.

p r o m i s e d a n d is incapable of stabilizing the b o n e - i m p l a n t interface. ~ This has led to the w i d e s p r e a d use of prostheses that h a v e been designed to achieve distal fixation within the diaphysis; however, it has n o t been conclusively demonstrated whether diaphyseal fixation alone can provide suffident rotational stability to control micromotion and whether rotational stability varies as a function of implant design. This study was undertaken to investigate the contribution of the cross-sectional configuration of the distal stem to the stability of the b o n e - i m p l a n t interface and to examine w h e t h e r any cementless design can provide torsional resistance comparable to that of a cemented stem.

The importance of rigid initial fixation of cementless femoral stems is well r e c o g n i z e d in the orthopaedic literature 1-5 (Harris WH et al., u n p u b lished manuscript); however, opinions differ as to w h e t h e r rigid fixation is best obtained t h r o u g h proximal or distal e n g a g e m e n t of the prosthesis within the femur. In revision arthroplasty, the cancellous bone of the proximal f e m u r is often corn-

From the Department of Orthopaedic Surgery, Baylor College of Medicine, Houston, Texas.

Reprint requests: Hugh S. Tullos, MD, Department of Orthopedic Surgery, Baylor College of Medidne, 6550 Fannin, Suite 2625, Houston, TX 77030.

463

464

The Journal of Arthroplasty Vol. 10 No. 4 August 1995 Materials

and Methods

Cementless stems of four cross-sectional designs w e r e selected for evaluation in this study (Fig. 1). These designs included (1) a cylindrical s t e m augm e n t e d w i t h two longitudinal fins located on opposite sides of the stern cross-section, (2) a cylindrical fluted stern of a design similar to a S a m p s o n rod, (3) a cylindrical p o r o u s - c o a t e d stem w i t h sintered beads of pore diameter ranging f r o m 100 to 300 g m (AML, DePuy, Warsaw, IN), 7 and (4) a cylindrical fluted s t e m w i t h a slot in the coronal p l a n e (S-ROM, Joint Medical Products, Stamford, CT) (Fig. 2).7 Additionally, a fifth design, consisting of a cylindrical, solid stainless-steel stern w i t h a k n u r l e d surface, was c e m e n t e d into each specimen to serve as a control stem. Twelve cadaveric femurs w e r e selected f r o m a large a n a t o m i c collection on the basis of standard a n t e r o p o s t e r i o r (AP) a n d lateral r a d i o g r a p h s . F e m u r s w e r e selected w i t h a canal w i d t h of 10 to 13 m m at the isthmus, as m e a s u r e d o n the AP radiograph. All femurs w i t h a distinctly elliptic i n t r a m e d u l l a r y canal (ie, those in w h i c h the isthm u s was 30% wider on the lateral v i e w t h a n the AP view) w e r e excluded. Each b o n e was sectioned transversely at the level of the lesser t r o c h a n t e r a n d at a level 210 m m m o r e distal. This resulted in a cortical specimen 210 m m long, including the isthmus. Each specimen was potted in dental plaster and t h e n sequentially r e a m e d w i t h standard surgical flexible r e a m e r s to a diameter of 13.0 r a m . Preparation was completed by m o u n t i n g each speci m e n in a drill-press to ensure precise a l i g n m e n t a n d m a c h i n i n g the canal w i t h a n adjustable

Finned Design

Fig. 2. Stem designs prepared for attachment to testing device.

reamer. The diameter of the r e a m e d canal was confirmed to a n accuracy of + 0.05 m m using a bore gage a n d an electronic caliper. The m e d u l l a r y canal was r e a m e d to 14.5 to 15.5 m m to m a t c h the s t e m diameter of each design of the implant (Table 1). With the finned design, the canal was r e a m e d to 15.0 ram, the root d i a m e t e r of the stem. Longitudinal grooves were t h e n m a c h i n e d in the m e d u l l a r y surface using an intramedullary guide a n d long drill bits w i t h a radius of 2.0 m m to a c c o m m o d a t e the fins o n the surface of the s t e m with a n overall interference of 0.3 m m . Prior to i m p l a n t a t i o n of the p o r o u s - c o a t e d stem design, the i n t r a m e d u l l a r y canal was r e a m e d to 14.5 m m to allow for an interterence of 0.3 m m w i t h the outer surface of the cavity. Finally, with the fluted and solid fluted stems, the canal was r e a m e d 15.5 m m to create an interference of 0.5 m m b e t w e e n the o u t e r m o s t surface of the s t e m and the r e a m e d canal. The c e m e n t e d control specimen w e r e pre-

Porous Coated Design

Table 1. Preparation of Intramedullary Canal Based on Implant Design

Fluted Design

Slotted Fluted Design

Fig. 1. Cross-sectional designs of stems.

Stem Design

Reamed IM Canal Size (ram)

Interference (mm)

Pinned Solid fluted Porous-coated Slotted fluted

15.0 15.5 14.5 15.5

O.3 0.5 0.3 0.5

IM, intramedullary.

Distal Stem Design

pared by mixing acrylic cement (Surgical Simplex, Howmedica, Rutherford, NJ) by hand for 90 seconds and then introducing it into each specimen using a cement syringe. The knurled stem was inserted into the cement at a rate of 10 m m / s in a central position within the canal until the cement was polymerized. This resulted in a cement mantle 2 to 3 m m thick within the coronal plane. Following preparation, the potted specimens were rigidly mounted in a stainless-steel holding device in a biaxial mechanical testing machine (Bionix, MTS, Eden Prairie, MN). A 50-ram-long segment of each stem design was attached to a stainless-steel rod, which was rigidly attached to the crosshead of the mechanical testing machine. Under machine control, each stem was implanted into the femoral specimen to a depth of 50 ram. The insertional force was monitored during the implantation procedure. A hole was drilled through the cortex of each specimen to allow attachment of an inductive displacement transducer (Spectral Dynamics, San Diego, CA) to the implant surface via a threaded steel pin. A metal target, calibrated to the transducer, was attached to the outer cortical surface of the bone to allow the relative displacement of the stem with respect to the bone to be measured during loading (Fig. 3).

•

Kendrick et al.

465

Each stem was rotated at a constant rate of 0 . 5 ° per second until torsional failure, as indicated by an abrupt change in the slope of the torque-displacement curve. To facilitate later analysis, the rotation of the stem and the output of the displacement transducer were sampled at a frequency of 25 Hz using a computerized data acquisition system. Following each run, the data were analyzed to determine the torque required to cause 100 gm of relative displacement at the bone-stem interface. Statistical analysis of the data was performed using a one-way analysis of variance and the Newman-Keul test for comparison of multiple groups.

Results Eight experiments were performed using each stem configuration, and six experiments were performed using the cemented control stem. The data collected during each run included the force needed initially to insert each stem into the femur and the torque at failure of the bone-implant interface. The average insertional force for all implants was 2,408 + 136 N (Table 2). The largest average insertion force was recorded with the porous-coated design (2,855 + 157 N), and the lowest, with the slotted fluted design (1,732 ± 134 N) (P = .0139). There was no significant difference in the forces recorded during insertion of the fluted and finned designs (2,678 + 348 N vs 2,367 ± 197 N respectively, P = .4786). Four shaft fractures occurred during rod insertion: two with the porous-coated design, at insertional forces of 3,979 N and greater than 4,450 N (the limit of the test configuration), and two with the finned stem at forces of 326 and 1,202 N. With the finned stem, the fractures occurred through the grooves machined in the medullary canal to allow mechanical locking of the implant in the femur. The torsional resistance of each stem design was measured at 100 ~tm of displacement of the bone-implant interface and averaged 20.8 _+ 1.6

Table 2. Insertional Force Observed With Implantation of Stem Designs Force (N) Stem Design

Fig. 3. Testing configuration.

Finned Solid fluted Porous-coated Slotted fluted

Average 2,367 2,678 2,855 1,732

_+ 196.9 ± 347.9 ± 156.6 ± 134.4

Range 1,630-3,702 1,318-4,221 2,263-3,773 1,166-2,399

466

The Journal of Arthroplasty Vol. 10 No. 4 August 1995

N-m for all the cementless stems tested (Table 3). The average torsional resistance varied according to stem design and ranged from 13.7 + 0.8 N-m for the porous-coated stem to 30.1 + 3.7 N-m for the fluted design (P = .0001). The slotted fluted and finned designs failed at intermediate torques of 19.9 -2-_2.3 and 19.5 ± 1.4 N-m, respectively. The torsional resistance of the c e m e n t e d control stem (34.0 + 3.0 N-m) was significantly higher (P < .05) t h a n that Of all of the cementless stems, except the solid fluted design (Fig. 4). Two fractures occurred during torsional loading, both with the finned stem, at interface displacements exceeding 100 pm. Predictably, both fractures occurred t h r o u g h the grooves that had b e e n cut to accommodate the fins of the stem. In the control group, failure uniformly occurred at the b o n e - c e m e n t interface r a t h e r t h a n at the c e m e n t - i m p l a n t junction. Moreover, three fractures occurred with application of torque to the c e m e n t e d control stems, all at interface displacements greater t h a n 100 gin. As the cortices of the c e m e n t e d specimens ranged from 4.5 to 12 mm, these failures did not appear to be associated with excessive thinning of the cortices of the s p e d m e n due to reaming (Table 4). Two of these fractures were explosive in nature, with the generation of long oblique fractures with butterfly fragments. The effective shear stress at 100 gm displacement of the b o n e - s t e m interface was calculated for each e x p e r i m e n t and ranged from 1.0 + 0.1 MPa in the porous-coated design group to 1.7 + 0.2 MPa in the fluted model group. Intermediate values were calculated for the finned and slotted fluted designs at 1.1 +_0.1 and 1.1 + 0.1 MPa, respectively. The average shear stress for all cementless experiments was 1.3 +_ 0.1 MPa (Table 5). A o n e - w a y analysis of variance indicated a significant difference b e t w e e n the four stem types for insertion force (P = .0139), torque at failure (P = .001), and interface stress (P = .004). The N e w m a n - K e u l range test revealed that the insertion force of the slotted fluted stem was significantly less (P < .05) t h a n that of the porous-coated

Table 3. Torque at 100 ~m of Interface Displacement

40

30

E z a)

20

E

O I--

10

Cemented

Fluted

Slotted

Finned

Porous

Stem Design

Fig. 4. Torque at failure for each distal stem design.

and solid fluted designs. The solid fluted stem also displayed significantly higher torque and shear stress at failure (P < .05) t h a n the other cementless stem designs. A significant correlation (r = .923, P = .001) was observed b e t w e e n the force of insertion and the torque at failure for the fluted stem; however, there was no significant correlation b e t w e e n the insertion force and ultimate torque of the other stem designs.

Discussion Cementless methods of implant fixation hold the promise of increasing the durability and reliability of total hip arthroplasty, particularly in y o u n g e r patients and those undergoing revision of a previous arthroplasty2 Although c e m e n t provides immediate mechanical stability and a perfect fit b e t w e e n even a poorly designed femoral prosthesis and the endosteal cavity, cementless technique demands a precise geometric fit and the elimination of m o t i o n if biologic fixation is to ensue. 6'~.9 In this way, the initial stability of a cementless prosthesis is directly correlated with its ultimate clinical success. N u m e r o u s factors influence the ultimate stability of cementless prostheses, although opinions differ regarding the relative importance of the design of the stem and the m o r p h o l o g y of the implant s u r -

Torque (N-m) Stem D e s i g n Finned Solid fluted Porous-coated Slotted fluted Cemented control*

Average 19,5 30,1 13,7 19.9 34

± 1.3 ± 3.7 ± 0.8 ± 2.3 _+ 3.0

T a b l e 4. I n t e r f a c e S h e a r S t r e s s

Range 15.4-27.2 14.8-45.3 10.9-18.6 11.9-28.8 22.4-46.6

*In all specimens, failure occurred at the b o n e - c e m e n t interface,

Stress (MPa) Stem D e s i g n

Average

Range

Finned Solid fluted Porous-coated Slotted fluted

1.1 1.7 1.0 1.1

0.9-I.6 0.8-2.6 0.8-1.4 0.7-1,6

± 0.1 +_0.2 + 0.1 ± 0.1

Distal Stem Design

•

Kendricket al.

467

Table 5. Cortical Thickness of Specimens Cortical Thickness (ram) Stem Design

Porous-coated ~uted Slottedfluted Fmned Cemented

Age of Donor (years)

Medial

Lateral

Anterior

Posterior

69.8±5.1 74.0±4.3 67.0±4.9 72.0±9.8 63.8±4.2

6.4±0.6 6.5±0.6 5.5±0.5 6.6±0.5 7.0±0.9

6.7±0.5 7.1±0.6 7.6±0.4 6.5±0.6 7.0±1.1

6.1±0.5 7.3±0.3 6.5±0.8 6.3±0.4 7.3±0.9

8.1±0.5 8.8±0.6 9.2±0.8 8.1±0.5 8.9±i.5

face. Several design parameters, including the pore size, pore volume, and coating thickness, influence the rate and extent of bony ingrowth into porous coatings and, presumably, the long-term stability of the porous-coated implants. Surprisingly little attention has been focused on the contribution of the design of the implant itself to its rotational stability, especially in situations where diaphyseal fixation is being sought to minimize micromotion at the bone-stem interface?,'° In this study, we measured the effect of the crosssectional geometry of the distal stem on bone-stem micr o mo tio n using four alternate cementless designs. The first design, a custom-machined, cylindrical, finned stem, had the advantage of abundant surface contact between the implant and the endosteal surface, as wel.l as the presence of geometric features to enhance implant stability under rotational loading. The finned configuration had the disadvantage of requiring substantially more preparation of the intramedullary canal than the other three designs. Moreover, the intramedullary channels prepared within the femur created large stress concentrations, as evidenced by four fractures that were initiated at the grooves, two at relatively small insertional forces and two during torsional loading. The second configuration, a custom-machined fluted design, provided stability through mechanical coupling of the bone and flutes. Surface contact between the stem and the endosteal interface was small relative to other configurations, such as the finned and porous-coated designs; however, the relatively sharp flutes were able to cut into the cortex during insertion, t he r eby increasing the mechanical interlock u n d e r torsional loading. Although, potentially, this could lead to an increased incidence of specimen fracture during stem insertion, this complication was not observed during experimental testing despite the presence of 0.5 mm of interference between the outer surface of the flutes and the reamed canal. The third design, the porous-coated stem, had the advantage of a significant area of contact between the surface of the stem and the endosteal

surface. Conceptually, one might imagine that a beaded stem, implanted with an interference of 0.3 mm, would provide significant torsional stability through the friction generated at the b o n e - s t e m interface. In practice, however, the first few rows of beads entering the bone during implantation tend to erode the reamed surface, thus reducing the effective interference between the canal and the beaded coating." It is interesting to note that in two experiments in which the porous-coated stem was loaded in torsion, an area of coating was actually sheared off the surface of the prosthesis. This observation suggests that the torsional resistance of porous-coated implants m ay be determined, to some degree, by the strength of bonding of the ingrowth surface to the underlying substrate. The fourth design, a fluted stem with a slot in the coronal plane, was routinely implanted with an interference of 0.5 ram. In this design, the flutes engage cortical bone and thus provide rotational stability, much like a Sampson intramedullary nail. 7 The dothespin design reduces the bending stiffness of the stem, thus r e d u d n g the stress concentration generated within the diaphysis around the distal tip of the implant. A disadvantage of the increased flexibility of this design is that the stein compresses during insertion, thereby reducing the mechanical interlock between the flutes and the endosteal surface. Although thisdesign had less surface contact with bone relative to the finned and porous-coated designs, it actually demonstrated a greater overall proximity to the endosteal surface than the solid fluted stem because of a larger minor stem diameter. Like the solid fluted model, no fracture was observed with this design during insertion or torsional loading. Finally, with the cemented knurled stem, the presence of bonding between the cement and the implant was evidenced by the absence of failure at the cement-implant interface under torsional conditions. Although we attempted to pressurize the cement during implantation of the knurled stern into the cylindrical diaphyseal segment, in a complete femur, the cement pressures achieved during

468

The Journal of Arthroplasty Vol. 10 No. 4 August 1995

specimen preparation did not match those routinely generated in the intact femur using the same techniques. For this reason, it is expected that in clinical practice, cemented stems would have sustained higher torques then were measured in this experiment and that there would be a larger difference in performance between the cemented and cementless designs. Although previous work has shown that a close fit of the distal stem to the medullary canal leads to decreased rotational and translational micromotion of cementless stems, as well as reduced axial migration, the magnitude of the interference achieved between a porous-coated stem and the reamed medullary canal does not correlate with torque at failure ~ (Harris et al., unpublished manuscript). The constant torsional resistance of porous-coated implants does, however, increase with the area of contact between the stem and the femur. In practice, the elliptic shape of many medullary canals will limit the area of surface contact, regardless of stem design, just as factors defining the quality of the bone (density, strength, cortical thickness) limit the amount of intramedullary reaming that can be performed to increase the cylindrical region of the medullary canaP 2 (Harris et al., unpublished manuscript). It is postulated that these factors account for much of the variability of the results ot our own experiments despite attempts to standardize the selection of femurs. In this study, a statistical difference was present in the relatively small average insertional force required to implant the slotted fluted stem relative to the porous-coated and solid fluted stems (P = .0095). Theoretically, this might suggest a decreased risk of fracture with implantation for the slotted fluted stem. In fact, no fractures occurred with implantation of the slotted fluted stem or with the solid fluted stem, unlike the porous-coated and finned designs. Moreover, a significant difference was noted in the torque and interface shear stress at failure of the solid fluted stem in comparison to all of the other stem designs. Clearly, the fluted stem provided superior rotational stability relative to all other designs and, in fact, approached 90% of the torsional resistance observed in the cemented control stems. Our results demonstrate that the distal stem designs of commonly used cementless femoral prostheses are not sufficiently rigid to prevent 100 btm of distal motion of the bone-implant interface w h e n loaded with torques in excess of 20 N-re. As torques of up to 22 N-m have been reported in stairdimbing and rising from a chair, it is postulated that most prosthetic devices w o u l d not be able to

provide sufficient torsional stability through distal intramedullary fixation alone. For these devices to function satisfactorily, additional contact is necessary between the bone and the implant to supplement the mechanical interlock of the distal stem within the diaphysis. In primary hip arthroplasty, this can be achieved through proximal contact between the stem and the metaphysis; however, in revision arthroplasty, improved distal fixation is essential as the quality of proximal bone is often poor. In some cases, it may be impossible to achieve a good proximal fit, and allografting may be necessary to compensate for bone loss. In these instances, distal stem fixation is often the only method available to stabilize a cementless prosthesis within the femur. Based on the experiments presented in this study, implants with a solid fluted distal stem are recommended for use in revision arthroplasty to provide rotational stability, even in the absence of significant metaphyseal support for the prosthesis. In these cases, it is recommended that the intramedullary canal be reamed to a cylindrical configuration to the extent that bone quality will allow, and that the stem be implanted with an overall interference of 0.5 m m between the fluted implant and the endosteal surface. Although this approach would usually provide sufficient implant stability for a successful clinical result in the short term, it is not know n whether the proximal bone loss observed with extensively porous-coated prostheses will also be a feature of the other distally finned devices. It is also u n k n o w n whether all distally finned implants will present similar difficulty in removal should this become necessary for the treatment of infection, loosening, or malposition. A possible solution to this dilemma is a return to cemented revision arthroplasty, possibly with the use of compacted morcelized bone. Another cementless alternative is an extensively porous-coated prosthesis of lower bending stiffness through the use of composite materials or structural modifications of commercial metal prostheses. Long-term studies demonstrating the relative efficiency of these approaches are still awaited.

References i. Engh CA: Hip arthroplasty with a Moore prosthesis with porous coating. Clin Orthop 176:52, 1983 2. Engh CA, Bobyn JD, Glassman AH: Porous-coated hip replacement. J Bone Joint Surg 69B:45, 1987 3. Engh CA, Bobyn JD, Goroki JM: Biological fixation. of a modified Moore prosthesis. Orthopaedics 7:285, 1984

Distal Stem Design 4. Pilliar RM, Cameron HU, Welsh RE Binnington AG: Radiographic and mm~hologic studies of load-bearing porous-surfaced structured implants. Clin Orthop 156:249, 1981 5. Spector M: Histological review of porous-coated implants. J Arthroplasty 2:163, 1987 6. Stillwell WT: The art of total hip ~frthroplasty. Grune & Stratton, New York, 1987 7. Shepherd BD, Bruce W, Walter W e t ah Difficult hip replacement surgery: problems and solutions. Scientific Exhibit. Annual Meeting of the American Academy of Orthopaedic Surgeons, Las Vegas, NV, February 1989 8. Cook SD, Thomas KA, Haddad RJ: Histologic analysis of retrieved hmnan porous-coated total joint components. Clin Orthop 234:90, 1988

•

Kendrick et al.

469

9. Noble PC, Alexander JW, Lindahl LJ et al: The anatomic basis of femoral component design. Clin Orthop 235:148, 1988 10. Clemow AJT, Weinstein AM, Klawitter JJ et al: Interface mechanics of porous titanium implants. J Biomed Mater Res 15:73, 1981 11. Kamaric E, Noble PC, Alexander JW, Tullos I-IS: The biomechanics of diaphyseal fixation of a porous coated femoral prosthesis. Presented at the Annual Meeting of the Society of Biomaterials, Charleston, SC, May 1990 12. Miller JE, Burke DL, Stachiewicz J: The fixation of major load-bearing metal prosthesis to bond: an experimental study comparing smooth to porous surfaces in a weight-bearing model. Presented at the Annual Meeting of the Orthopaedic Research Society, 1976