We used three-dimensional finite element analysis and Taguchi parameter design techniques to study the effects of design variables -- femoral neck length (45 or 62 mm), proximal bonding (bonded or unbonded), and distal stem geometry (Charnley-type flatsided or cylindrical) -- on the stresses within the mantle and at the interfaces to find the best combination of design variables to prevent failure. The robustness of the best design was determined with respect to uncontrollable, environmental variables: patient weight (750 N and 850 N) and activity (walking, stair ascent, and descent). In addition, we also used these techniques to determine the combination of neck length, proximal bonding, and stem geometry that would be most likely to produce failure. We compared these results with those from clinical follow-up studies.

The best combination of design variables was a proximally bonded with flat-sided distal geometry. Changes in stresses due to neck length were small compared to those for proximal bonding and distal stem geometry. Proximal bonding decreased proximal cement mantle stresses by 35 to 50% to levels below the fatigue strength of cement and decreased compressive stresses in the distal mantle by more than half. For the bonded case, the flatsided implant reduced shear stresses at the cement-prosthesis interface by 23%, but increased the distal cement stresses. For the unbonded cases, the flat-sided implant decreased shear stress by 73% and reduced the maximum principal stress in the proximal mantle by 30% compared to the round geometry. Therefore, the combination of proximal bonding with a rounded distal stem was worse and was exacerbated when proximal bonding is lost. However, the contributions of the environmental variables (patient weight and activity) to increases in cement stress were relatively large, consistent with clinical findings. Therefore, design, fixation and environmental variables must be considered when explaining clinical outcomes.

Aseptic loosening of cemented femoral components continues to be an important long-term problem in cemented total hip arthroplasty. Clinical studies have demonstrated the relationship of cement technique to clinical outcome. Previous autopsy studies performed by us have demonstrated that cement mantle fracture was associated with debonding at the metal cement interface, thin cement mantles, mantle defects, voids in the cement and sharp corners of the prostheses. In order to more thoroughly analyze the relationship of mantle thickness to failure of the cement mantle, a more detailed analysis of autopsy specimens was performed.

Eight femurs containing cemented femoral components from patients with successful primary total hip arthroplasties were retrieved at autopsy at an average of 8.5 years after the index operation. Biomechanical studies were performed to quantify implant stability. Following this, the implants were sectioned at 5 mm increments and the cross sections examined using both light and scanning electron microscopy. A thin mantle was defined as a mantle of less than 1 mm in thickness. Thin cement mantles occupied approximately 9% of the entire cement mantle in these eight specimens. An additional 1.5% of the surface area was occupied by mantle defects where there was no cement between metal and bone.

In these eight specimens, 101 fractures in the cement mantle were detected. Although only 9% of the cement mantle was classified as having a thin cement, 92 of the 101 cracks occurred in areas that were less than 1 mm in thickness. In assessing the remainder of the cement mantle, many interfacial pores, voids, and remnants of bone and marrow were detected in the cement mantle. Seven-percent of the cement fractures were associated with cement voids which ranged in size from 0.15 to 0.5 mm in diameter. Two-percent of the cement fractures were found in association with remnants of bone within the cement mantle. An additional 4% of the cement fractures were associated with the high porosity noted at the interface between stem and cement.

This study documents the relationship between thin cement mantles and fractures in the cement. Approximately 90% of the fractures seen in the cement were associated with a thin mantle. From a clinical standpoint, this supports the concept of stem centralization, avoidance of cement mantles less than 1 mm in thickness, and reduction in the number of voids in the cement.

Advances in the design and long-term performance of cemented hip prostheses have occurred through detailed attention to factors causing mechanical failure of the cement mantle and its interfaces. During weight-bearing, the cement mantle surrounding the femoral stem is exposed to large stresses which vary with the geometry of the prosthesis, the thickness of the cement, and the rigidity of the cement/bone interface. To ensure that these stresses remain below the fatigue strength of the cement itself, the mantle must have a minimum thickness of 2 mm distally and 3-6 mm proximally, and must be supported by a strong interface with the surrounding bone. In practice, these conditions are achieved by pressurizing cement within the canal and then centralizing the femoral stem during insertion into the implantation site.

The key steps for achieving cement/bone interlock are (a) exposure of a strong, porous bony surface through removal of weak cancellous bone, (b) removal of blood, fat, and bony debris, and (c) pressurization of the liquid cement for sufficient time to generate 3-5 mm of bony penetration. In the medullary canal, liquid cement is most easily pressurized with a cement gun, which must be capable of delivering 2-3 units of cement to ensure adequate interdigitation at the cement/bone interface. The canal can only be adequately pressurized if the implantation site is converted into a closed space, usually with an intramedullary plug, distally, and a flexible cannulated seal, proximally. Both of these devices must be capable of withstanding 50 psi of pressure without migration or leakage.

Centralization of the stem within the canal is essential to preserve the integrity of the mantle and to maintain a minimum thickness of cement. Areas of direct stem/bone impingement are common sites of osteolysis and serve as nuclei for mantle fractures. Centralization of the femoral stem is most easily achieved using pre-formed PMMA spacers mounted both proximally and distally on the body of the prosthesis. These devices fit closely within the cavity formed by broaching the femur and maintain a minimum separation between the surface of the implant and the wall of the bony cavity. The distal centralizer must be modular to accommodate different canal diameters, which typically vary by 5-6 mm in femora with the same size metaphysis. Proximally, the prosthesis is often centralized with a slip-on ring of PMMA that matches the contour of the prosthesis and the canal at the osteotomy level.

Radiographic studies have shown that first-generation centralizers reduce the incidence of inadequate cement mantles from 28% to less than 10% proximally, and 37% to 8% distally. However, some complications have been reported, including fracture of modular devices and accumulation of voids around the fins of distal centralizers. These complications may compromise of cement fixation in the long-term, but have not proved detrimental in radiographic follow-up of up to eight years post-implantation. The occurrence of these complications has led to the development of new designs of centralizing devices, including integral molding of proximal centralizing rings directly onto the surface of the prosthesis. Centralizers of the pre-molded design do not fracture during stem insertion and provide additional pressurization of the cement during implantation into the femur. Alternate methods of centralizing the implant within the medullary canal have also been developed that avoid the use of implanted devices within the canal.

Although long-term clinical follow-up is still awaited, mid-term data indicate that modem techniques of cement pressurization and centralization lead to genuine improvements in the reproducibility and longevity of cemented hip replacement.

Mechanical failure of cemented femoral reconstructions continues to be a potential mechanism for premature failure of fixation. The cement mantle around a cemented femoral prosthesis is subjected to significant levels of stress under normal physiological loads. After a cemented total hip arthroplasty, the loads across the hip during ambulation must travel from the implant to the femur through the cement mantle. If the stresses that the cement mantle experiences locally exceed the endurance limit of cement, the cement will ultimately undergo mechanical failure under cyclical loading. The area near the tip of a cemented femoral component has been identified as an area of particularly high cement strains. Cement fractures at this location have been demonstrated in autopsy retrieved specimens as well as in cases prior to revision surgery.

The most important factors influencing the level of cement strain seen near the tip of the implant include the thickness of the cement mantle, the position of the stem within the cement mantle, and the bending stiffness of the distal portion of the implant. The diameters of the medullary canal and the tip of the implant influence the thickness of the cement mantle while stem malposition can lead to locally thin cement mantles near the tip of the implant. The bending stiffness of the distal portion of the implant is directly influenced by the geometry and material properties of the stem.

A variety of stem designs were investigated to determine the optimal geometry that would reduce the bending stiffness of the distal portion of the stem, allow for an adequate cement mantle, and provide a means of centralization that would achieve a neutral stem position at the time of insertion. A stem that narrowed at the tip to accept an externally applied centralizer was found to have a distinct advantage over a stem with a hole drilled in the tip to accept a centralizer. This design was more likely to result in cement strains near the tip of the implant that remain below the endurance limit of cement under normal physiologic loads.

The suggestion that porosity reduction in acrylic cement is necessary for optimization of results in cemented hip arthroplasty derives from the concept that fatigue fracture of cement is a major factor in the late failure of cemented THA’s, and that, in turn, the fatigue strength of cement is improved by porosity reduction.

It is, however, by no means certain that fatigue fracture of cement is a major factor in the late failure of cemented THA’s. But there is no doubt that porosity reduction does significantly increase the fatigue strength of cement in conventional materials fatigue testing. If fatigue fracture of cement is a major factor in the failure of THA, it should be easy to demonstrate a clear difference in outcome between hips in which pore-free cement had been used, by comparison with those in which it had not. No such comparative series have yet been published to demonstrate an improved outcome with pore-free cement, and by contrast, the difference that has emerged in the most recent report (1996) from the Swedish Hip Registry, has, paradoxically, revealed a 30% increase in the risk of failure when vacuum-mixed cement was used. The cause of this apparent anomaly remains uncertain. It may be associated with the use of defective mixing systems, but it may also be due to some other aspect of the behavior of vacuum-mixed cement, for example, its increase shrinkage at polymerization.

The findings from the Swedish Hip Registry in relation to matters of operative technique that do improve the outcome of cemented total hip arthroplasty are clear, and their application in Sweden has led to a progressive reduction in the failure rate of the operation to very low levels. This means that it would be very difficult indeed to prove that an extra measure, such as porosity reduction, had any beneficial effect. This assumes that the apparently anomalous results from the Swedish Registry, mentioned above, are due to defective mixing systems, and not to some other adverse change in the behavior of vacuum-mixed cement. For these reasons, and taking in to account the now considerable evidence that porosities in cement do not have an adverse influence on total hip arthroplasty outcome, it is not possible to state categorically that porosity reduction is essential for cemented total hip arthroplasty. It might even be harmful. There is no evidence that it is necessary.

The curing time of bone cement can be reduced by several minutes by preheating the femoral stem, decreasing operative time and the potential risk of accidental loss of position. Dall et al (1) in an in-vitro model found a 33% reduction in curing time when the stem was heated to 50 degrees Celsius with a decrease in compressive yield strength from 98 Mpa to 90 Mpa.

Bishop et al (2) found an 80 % reduction in surface porosity at the stem-cement interface with stem heated to 44° C in comparison with stems inserted at room temperature. They proposed that with stems at room temperature the polymerization starts at the warmer bone-cement interface and proceeds towards the stem. The stem-cement interface cures last with accompanying pores and flaws which could act as stress risers and propagate leading to mechanical failure. With preheated stems the polymerization should progress centrifugally, with significantly fewer pores and flaws at the stem-cement interface.

Altering the polymerization rate could affect the biomechanical properties of the bone-cement. We investigated in a laboratory model the effect of a heated stem on the biomechanical properties of four common brands of bone cement (Simplex P, Howmedica, Inc., Rutherford, NJ, Zimmer Osteobond and Zimmer Regular, Zimmer, Inc., Warsaw, IN and Palacos R, Smith and Nephew Orthopaedics, Memphis, TN). ASTM "dog bone" cement specimens made at four test temperatures, 0° C, 23° C, 37° C and 50° C, were tested for modulus, ultimate tensile strength, fracture toughness and percent porosity. Fatigue strength was tested in Simplex P at 23° C and 37° C.

No adverse changes in the biomechanical properties were found for any brand of cement at any temperature. Fracture toughness and fatigue strength were unaffected by increased temperature. Ultimate tensile strength increased in Simplex P and Zimmer Osteobond but was unaffected for Zimmer Regular and Palacos. Percent porosity was inversely related to temperature in all brands studied, except for Palacos R which demonstrated consistently low porosity, unaffected by temperature variation. Fatigue strength of Simplex P was unaffected by heating to 37° C. Based on this laboratory model, heating of the femoral stem could provide time savings, without sacrificing the mechanical properties of bone cement.

In centers performing a large volume of cemented total joint replacements, a reduction of several minutes in curing time would provide significant annual operative time savings, without compromising implant performance. At The Hospital for Special Surgery where approximately three thousand total joint replacements are performed per year, a reduction of 5 minutes of surgical time could represent 250 hours less of operative time and a savings of approximately $250,000 per year, if we consider the cost of operating room time to be close to $1,000 per hour. Likewise, a surgeon performing 300 cemented joint replacements per year could reduce the time spent in the operating room by approximately 25 hours.

We only recommend preheating the femoral stem to those surgeons who perceive the duration of polymerization as too long. Perhaps, in the early attempts the stem should be preheated to 30° C, increasing progressively to 37° C as the confidence of the surgical team is established. Surgeons should be experienced in joint replacement surgery and have a skilled surgical team, as the risk of premature polymerization, prior to complete seating of the stem should be recognized, particularly if the temperature of the operating room is warm.

1. Dall, DM, Miles, AW and Juby, GJ: Accelerated Polymerization of Acrylic Bone Cement Using Preheated Implants. Clin Orthop. 211:148-150, 1986

2. Bishop, NE, Ferguson, S. Tepic S: Porosity Reduction in Bone Cement at the Cement-Stem Interface. J Bone Joint Surg. 78B:349-356, 1996

Prospective cohort studies and retrieval analysis were used to compare outcomes of cemented and cementless femoral fixation for primary and revision total hip arthroplasty (THA). The clinical studies used patient derived outcomes which were independent of the doctor.

The early results of a randomized clinical trial of cemented (Exeter) versus cementless primary THA (PCA) indicated that the early clinical outcomes were similar between the two groups and that immediate weight-bearing did not adversely affect cementless fixation. The mid-term results indicated that while the clinical results were comparable between the two groups, there were some mid-term failures of matte surfaced cemented stems. Polished surfaced cemented stems appear to behave differently. Cementless proximally porous coated stem stems had good mid term survival.

In comparison to primary THA, differences in the early outcomes were seen between cemented and cementless revisions of cemented THA. Measurements of the extent of femoral defects were used to allow comparison of the results. Better pain relief, function, and radiographic stability was obtained with cemented mid-length stem revision compared to proximally porous coated mid and long stem cementless revision when used to treat similar segmental deficiencies.

The relationship of stem loosening to cemented femoral component design and surface finish was explored by examining a larger series of matte versus polished tapered cemented femoral stems and by analysis of a large number of retrieved cemented stems. In addition, assessment of the sites of loosening was used to provide better understanding of the mechanisms of failure.

In summary, matte surfaced tapered cemented stems of standard length were associated with osteolysis and loosening of primary THA. Longer length cemented stems for revision gave good results. Cemented stem wear was related to surface finish, material, design, and type of fixation of the stem in cement. These factors influence the likelihood of osteolysis and early loosening.

The rational design of femoral total hip components synergistically incorporates the mechanics of implant geometry, cement utilization, and the cement/metal interface. Over the past 30 years, cemented total hip femoral components have been produced with a variety of surface finish conditions utilizing manufacturing methods including polishing, blasting (with a variety of media), machining, chemical milling and deposited metal coatings. As a result, the implant surfaces have greatly varied in detailed morphology, visual appearance, and surface mechanics. Surface roughness measurement of historical and current hip femoral components from many manufacturers vary by more than two orders of magnitude and cover the range from several micro inches to several hundred micro inches. With increased metal surface roughness, there is generally greater cement interface shear strength. While this increased strength may reduce the potential for cement/metal interface motion, if motion occurs, the rougher implant surfaces have an increased potential to generate particulate debris. Smoother implant surfaces demonstrate lower cement shear strength and lower cement abrasion potential with interface motion.

There is somewhat of a paradoxical choice in selecting implant surface finishes. Rougher cemented implant surfaces have a higher probability of interface fixation and, at the same time, a higher debris generation consequence of interface motion. In contrast, smoother cemented implant surfaces have a higher probability of interface motion and a lower debris generation consequence of that motion. In context with other implant design features, choices of surface finish can be made to either resist cement/metal interface motion or to accept interface motion.

Stem cement debonding is one of the most common forms of fixation failure and is thought to be a prelude to gross loosening of a total hip reconstruction. However, the immediate consequences of debonding remains a matter of controversy. The dynamic effects of stem cement debonding in total hip reconstruction were analyzed using 3-dimensional finite element techniques. Stem cement interface conditions were assumed as completely bonded or unbonded, with or without friction. The dynamic effects were accounted for, as presented by the stance and swing phases of the gait cycle.

It was found that both cyclic micromotions at the stem cement interface and stresses in the cement mantle were effectively reduced by friction. The friction cases produced failure probabilities of the cement mantle that were relatively close to the one generated by the bonded stem. The probability of mechanical failure of the cement bone interface decreased after debonding and decreased more with reduced stem cement friction.

These results show that, although a firm and lasting bond between stem and cement may be desirable for preventing cement failure, the mechanical effects of a debonded stem are less detrimental than were assumed earlier. For straight tapered stem shapes subjected to the loading conditions described, a polished stem may be desirable for the cement bone interface mechanics.

There is controversy in the literature as to the role of femoral stem surface roughness in loosening of the stem of total hip replacement (THR). The roughness of the surface with or without other design modifications such as a collar might affect the cement-bone bond and hence the longevity of THR.

We studied the effect of femoral stem surface roughness and a collar in loosening of the stem of cemented hip replacement. 35 sheep were randomly allocated to receive a hip hemi-arthroplasty with either a polished, matte or matte collared double tapered cemented femoral stem. Stem subsidence was measured with the aid of marker balls inserted at the time of surgery. Nine months later the sheep underwent hemi-arthroplasty of the contralateral hip using the same design of stem. The sheep were then killed to allow comparison of this stem immediately after implantation with that at nine months after implantation. Micromotion was recorded between the prosthesis and bone and the prosthesis and cement.

There was no measurable subsidence of any stem within the cement mantle. At nine months after implantation there was no important difference in axial prosthesis-bone micromotion between a polished and a matte surfaced stem of identical taper (p = 0.6). Axial prosthesis-bone micromotion immediately after implantation was 37mm and at nine months 23mm, suggesting that fixation in cement improved over time(p<0.001). There were no important differences in axial, medio-lateral and antero-posterior prosthesis-bone or prosthesis-cement micromotion between the three stem types.

The lack of differences in prosthesis-bone and prosthesis-cement micromotion between stems suggests fixation of a double tapered femoral stem in cement is not improved by a matte surface or a collar in the sheep model.

Polished stems were introduced into clinical practice in Exeter in November 1970. Experience with such stems, therefore, now exceeds 27 years. Early subsidence of these stems within the cement mantle was initially a cause for considerable concern but, with the passage of time, it has become clear that this subsidence is important in accommodating the visco-elastic properties of acrylic bone cement and that the polished surface is an integral part of the successful function of a collarless tapered stem.

The first 433 polished stems have been reviewed at regular intervals over 25 years. Revision for aseptic stem loosening remains at under 3%. The outcome of only 2% of the hips is unknown. Of the 49 survivors, at an average follow-up of 21.2 years, no stem is clinically or radiologically loose. Only 1 stem (2%) has endosteal bone lysis extending through greater than 1 Gruen zone. 80% of the hips have lost <2 mm of calcar height.

In 1976, with the benefits of the polished stem not yet fully appreciated, the number of stem sizes was increased and surface finish was changed to matte. Stem shape remained identical. This was a retrograde step. A review in 1990, of 180 cases operated on in 1980 with an average follow up of 10 years, revealed a 9% revision rate for aseptic stem loosening.

The polished stem was re-introduced in 1986 in monoblock form. A personal review of 130 hips operated on at that time, with an average follow up of 9 years, has confirmed the benefits of the polished stem, with no revisions and no osteolysis in the polished stem group versus a 4.2% revision rate and a 16.7% osteolysis rate in the matte stem group.

The universal polished Exeter stem was introduced in 1988. The first 160 cases have been reviewed at 5-8 years. There have been no revisions for aseptic loosening and there are no cases of osteolysis in this group. A study of a consecutive series of 85 THA's in 70 patients under the age of 50 followed for an average of 6 years, has revealed no revisions for aseptic loosening. Our experience with the Exeter collarless tapered stem has clearly demonstrated significant advantages of the polished over the matte stem.

In the author's 27-year experience cementing femoral components, stems with smooth and rough surfaces were implanted. Prospective data has been kept on all patients allowing extensive long-term follow-up. The Charnley stem, which had a 22-mm head and a stainless steel polished surface, was implanted in 168 hips from 1971 to 1975. To date only eight stems (4.8%) have required revision, four of these due to a stem fracture.

The experience with the T-28 and TR-28 allows comparison of similar stems but with different surface finish. From 1972 to 1977, 209 polished stainless steel T-28 were inserted and to date 14 (6.7%) have required revision. None had associated bone lysis. Ten (71%) of the 14 revisions were due to stem fracture and all were in the first one hundred hips, before the cross sectional area was increased. From 1977 to 1982, 225 second generation T-28 matte surface, forged cobalt-chrome stems were implanted. Five (2.2%) required revision. Radiographic review of both series revealed minimal lysis in four (1.9%) polished stems and in 11 (4.9%) matte stems. Three of these 11 had significant osteolysis despite the shorter length of follow-up.

From 1980 to 1993, 1,149 matte surface cobalt-chrome Iowa stems were implanted. In 1986 proximal precoat was added, which required further roughening of the surface. In 1995, Mohler et al. reported the early results of the author, combined with those of the developer of this prosthesis, Dr. Richard C. Johnston. Early loosening in 1.5 % was associated with significant bone lysis, complicating revision. By 1997, 32 (2.8%) stems have required revision, all associated with significant bone loss. Nine required early revision, between two and five years after insertion. These early failures prompted Dr. Johnston and the author to return to a polished stem with a geometry almost identical to the Charnley.

A series of early femoral component failures prompted a detailed retrospective clinical and radiographic review of 176 hybrid cemented total hip arthroplasties (THA) using a PMMA coated femoral prosthesis (Precoat, Zimmer, Warsaw IN). All surgeries were performed utilizing third generation cement techniques. Average length of follow-up was 6.3 years (range 4-11 yrs). Twenty-one patients died, and 1 was revised due to sepsis. Of the remaining 154 THA’s, twenty-three (15%) of the femoral components mechanically failed (21 revised, 2 definitely loose). The average time to revision was 3.9 years. None of the acetabular components failed.

Comparison between the failure and non-failure groups revealed that poor cement mantles (grades C/D) were statistically significant predictors of femoral failure. The presence of cement mantle defects was the primary reason for a poor cement mantle grade. The most common mechanism of failure was progressive, circumferential cement-bone interface osteolysis with subsequent implant migration and mechanical failure of the cement mantle. Debonding of the cement column from the prosthesis was a late finding and occurred in only 45% of failed cases.

The implementation of modifications to improve the quality of the cement column, particularly centralization and centrifugation, led to a significant improvement in clinical results. Strengthening of the cement-prosthesis interface may magnify the deleterious effects of a poor cement mantle and predisposes the cement-bone interface to failure.

Follow-up data of total hip replacements covering the first two decades of in vivo service have shown excellent long-term results with matte finished surface cemented femoral components. Similarly, long-term results have been achieved with collared femoral stems. Data supporting matte finished cemented femoral components have come from many different surgeons and different countries. These data establish the success of a collared cemented matte finished femoral component at all intervals spanning the first two decades.

Precoating the metal stem with a thin layer of methyl methacrylate strengthens the interface with bone cement.

Our ten year data using a precoated matte finished femoral stem are also excellent.

Recent reports have indicated a low incidence of failure during the first decade of some matte surface cemented femoral components and some precoated matte surface cemented femoral components. Analysis of these early failures show that the causes are multifactoral, including compromised cementing techniques, increased torsional moments, decreased torsional resistance and other factors of implant design, in addition to issues of patient activity, the role of precoating, and the influence of different degrees of surface roughness. An integrated synthesis of the data involving these factors will be presented relative to both the success and the failure of matte finished cemented femoral stems.

Dislocation following total hip replacement surgery (THR) remains a significant clinical problem. Malposition of the acetabular component increases the occurrence of impingement, reduces the "safe" range of motion and increases the risk of dislocation. Not fully understanding the interaction between pelvic orientation and final acetabular cup alignment may be one of the main contributing factors in the continued significant incidence of dislocations following total hip replacement.

There has been little clinical research to examine the effects of patient positioning and pelvic motion on the alignment of the acetabular implant during total hip replacement surgery. Until now, no tools were capable of accurately measuring these variables during the actual procedure. As part of a broader program in medical robotics and computer assisted surgery, we have developed several enabling technologies that provide surgeons with a new class of image guided measurement tools and assist devices. These surgical navigation tools provide position and alignment information never before available intraoperatively. Our Hip Navigation system (HipNav) continuously and precisely measures and tracks pelvic location and relative implant alignment. HipNav technology is used to gauge current clinical practice and provide intraoperative feedback to surgeons in order to improve the precision and accuracy of acetabular alignment during THR. These tools were successfully introduced into the clinical practice of surgery with results showing that: a) there exist unpredictable and large variations of the initial position of patients' pelvii on the OR table as well as significant pelvic movement during surgery and during intraoperative range of motion testing; b) current mechanical acetabular alignment guides do not account for these variations, and result in variable and in some cases unacceptable acetabular alignment; and c) press fitting oversized acetabular components influences the final cup orientation.

Keywords

Total hip replacement, dislocation, acetabular implant alignment, computer-assisted surgery, image-guided surgery, and navigational guidance.

The aim of this study was to assess the efficiency of a new cementing technique developed to prevent the risk of intraoperative pulmonary embolism. 70 patients with coxarthrosis entered into a prospective randomized clinical trial. In the control group of 35 cases the total hip arthroplasty was cemented conventionally. In the second group a proximal drainage placed along the Linea aspera and a distal drainage placed in the diaphysis created a vacuum in the medullary cavity of the femur during the insertion of the stem. The operation was performed under blood gas analysis, hemodynamic and transesophageal echocardiography monitoring.

Embolic events were observed in 94% of the cases of the control group and in 14% of the cases of the vacuum group (p=0.005). Embolism occurred during the insertion of the femoral component and continued after reduction of the hip joint. A significant decrease of oxygen saturation (p=0.001) and increase of the pulmonary shunt values (p=0.01) was observed in the cases operated conventionally.

The rise of intramedullary pressure in the femur is the most decisive pathogenic factor of pulmonary embolism during total hip arthroplasty. The logical prophylactic measure to prevent intravasation of fat and bone marrow is to create sufficient drainage. The cohorted investigation demonstrated the value of the vacuum cementing technique for a substantial reduction of intraoperative embolism and pulmonary impairment.

The purpose was to develop an animal model for the study of heterotopic ossification (HO) and to analyze the effects of perioperative radiation. In Phase One, NZW rabbits (n=18) underwent surgery either with or without muscle injury on each hip to establish the most reliable model in which to study HO. In Phase Two, rabbits (n=36) underwent either 400, 800, or 1200 centigray radiation to one hip 24 hours after bilateral hip surgery to establish a dose-response relationship for postoperative radiation therapy (RT). In Phase Three, rabbits (n=24) underwent preoperative RT (800 centigray) at 4, 16 or 24 hours preoperatively to investigate the mechanism of action and efficacy of preoperative RT. Monthly radiographs were graded by blinded observers for severity of HO. Mean grade, intraobserver and interobserver variability, and statistical significance were evaluated.

In Phase One, 17 of 18 rabbits generated HO in both hips, and the mean grade of HO was always greater on the operative side with intentional muscle injury. Variability in the grading was considered "excellent' (Kappa >0.80). Phase Two revealed that 800 centigray was the minimal effective dose (p=0.027) Contrary to hypothesis, Phase Three revealed an increasing grade of HO coinciding with a decreasing preoperative time interval, with the difference in HO grade with 24 hour versus 4 hour preoperative radiation being significant (p=0.0064).

The rabbit model is reliable and reproducible and closely resembles the human clinical situation following hip surgery. Both preoperative and postoperative radiation effectively prevented HO formation. The results support the use of preoperative radiation and establish a need for further investigation regarding the mechanism of action and timing of preoperative radiation therapy.

Dislocation is the second most common complication of total hip replacement following loosening. This is a mentally crippling complication for a patient who requires immediate treatment and counseling by the surgeon. The cause of dislocation is 1) poor patient position early postoperatively, most probably caused by impingement without capsular healing; 2) component malposition; and 3) soft tissue imbalance of the hip. Two-thirds of dislocations are recurrent. One of four dislocations occur after the first year; including late dislocations the surgical approach does not matter, but 22 mm head size is more frequent. The incidence of dislocation is increased with alcoholism, neurologic and muscle imbalance, prior hip surgery, hip fractures, and surgical inexperience.

Treatment of dislocations should be surgical if more than two occur. Four categories of treatment exist: 1) closed reduction without any further dislocation; 2) surgical treatment without further dislocation; 3) surgical treatment with subequent dislocations successfully treated with closed reduction and 4) surgical treatment with recurrent dislocations and recurrent surgical treatment.

Seventy-four hips with dislocation were reviewed. Fifty-one dislocated two or more times. The primary cause was impingement in 21, component malposition in 30, and soft tissue imbalance in 43 with 22 having 2 or more causes. Categories of treatment were 20 in 1; 22 in 2; 8 in 3; and 24 in 4 with average revisions of these being ±3.

A previously physically validated three-dimensional finite element model was used to study how several total hip component design and surgical placement variables contribute to resisting the propensity for posterior dislocation accompanying an erectly seated leg crossing motion. The computational formulation incorporated treatments of both polyethylene material nonlinearity and large-displacement sliding contact. The primary outcome measures were the peak intrinsic moment developed to resist dislocation, and the ranges of motion prior to neck-on-lip impingement and prior to frank dislocation.

Modifications of the acetabular liner design (chamfer bevel angle, lip breadth, head center inset) involved trading off improved peak resisting moment for compromised range of motion, and vice versa. Increases of head size led to substantial improvements in peak resisting moment, but if the head-to-neck diameter ratio was held constant, had almost no influence on the component range of motion. For the leg-crossing event studied, increased component anteversion, and even more so increased tilt (i.e., less net abduction), achieved improvements in both range of motion and in peak resisting moment, but these changes imply diminished resistance to anterior dislocation from extension/adduction motion inputs.

Three clinical studies, two retrospective and one prospective randomized, were performed to evaluate the relationship of femoral head size and acetabular component outer diameter to the prevalence of dislocation of modular total hip arthroplasty.

GROUP I: Between January 1994 and October 1995, 69 primary total hip arthroplasties in 63 patients were performed through a posterior approach by one surgeon using either a 22 mm or 28 mm diameter femoral head on a nonrandomized basis. The size of the femoral head was selected to allow a polyethylene liner thickness of more than 8 mm. The mean follow-up time was 2 years (range, 1-4 years). Four of 69 hips (5.7%) dislocated. Two of 13 hips (15%) with a 22 mm head dislocated, compared to 2 of 56 hips (3.5%) with a 28 mm head. This difference in dislocation rate was not statistically significant.

GROUP II: Between October 1995 and April 1996, the two senior authors performed a prospective, randomized study of 31 primary total hip arthroplasties in 30 patients. Each hip was randomized to surgical approach and to a femoral head diameter of 22 or 28 mm, for both acetabular components with an outer diameter of 56 mm or larger and 54 mm or smaller. With a minimum follow-up time of one year, 5 of 31 hips (16%) dislocated. Five of 14 hips (36%) with

22 mm femoral heads dislocated, compared to none of 17 with 28 mm heads (p=.012). All dislocations occurred in hips with 22 mm femoral heads and acetabular diameters of greater than 56 mm. Five of 16 hips (31%) with acetabular outer diameters 56 mm or larger dislocated, compared to none of 15 with outer diameters 54 mm or smaller (p=.043). Three hips had recurrent dislocations and two required revision. The surgical approach and acetabular component orientation did not significantly influence the rates of dislocation. Patient gender, age, preoperative diagnosis and obesity did not influence the rate of dislocation.

GROUP III: From December 1984 to January 1994, 308 primary total hip arthroplasties were performed through a posterior approach by one surgeon. A 28 mm modular femoral head was utilized with the Harris-Galante porous-coated acetabular component and a variety of cemented and uncemented femoral components. The rate of dislocation for components with an outer diameter of

60 mm and larger was significantly increased compared to those with an outer diameter of 58 mm and smaller (p=.035).

Based on these studies, hips with a 22 mm modular femoral head had an increased rate of dislocation compared to a 28 mm head, and this was statistically significant in the prospective, randomized study. The rate of dislocation for each head size was related to outer acetabular component size, with increased rates for component diameters 56 mm and greater with the 22 mm head and 60 mm and greater with the 28 mm head.

Although an acetabular component with an elevated rim is thought to improve the postoperative stability of a total hip prosthesis, the actual clinical value has not yet been demonstrated. To address this question, we reviewed the results of 5,167 total hip arthroplasties that had been performed at our institution from April 1, 1985, through December 31, 1991. The prostheses included 2,469

acetabular components with an elevated-rim liner (10 degrees of elevation) and 2,698 with a standard liner. The cumulative probability of dislocation was estimated as a function of time since the operation with use of the Kaplan-Meier survivorship method. Forty-eight of the 2,469 hips that had the elevated-rim acetabular liner dislocated within two years, compared with 101 of the 2,698 hips that had the standard acetabular liner. The two-year probability of dislocation was 2.19 per cent for the hips with the elevated-rim liner and 3.85 per cent for those with the standard liner (p=-0.001). Although these data demonstrae improved stability after total hip arthroplasty when an elevated liner is used, particularly in hips that are at greater risk for dislocation of the prosthesis, the long-term effect of this elevated liner on wear and loosening is a concern.

For this reason the cumulative probability of revision of these 5,167 because of implant loosening was estimated as a function of time since operation using Kaplan-Meier survivorship method. The 5 year Kaplan-Meier probability of survival was 98.8 per cent (95 per cent CI: 97.9 to 99.6) for the elevated group and 98.3 (95 per cent CI: 97.7 to 99.0) per cent for the standard group (p=.87). The effect of the elevated liner on the Kaplan-Meier probability of failure due to acetabular or femoral loosening was analyzed for the following subgroups: cemented components, uncemented components, males, females, primary total hip arthroplasty and revision total hip arthroplasty. There were no statistically significant differences found in the Kaplan-Meier probability of acetabular or femoral component survival for any of the various subgroups.

The two senior authors (PMP, RP) independently began using an identical enhanced posterior soft tissue repair following total hip replacement through a posterior approach. In the first author's experience, a dislocation rate of four percent in 395 patients prior to utilizing the enhanced closure was reduced to zero percent in 395 patients in whom the enhanced closure was performed. In the second author's experience, 160 THR's had a dislocation rate of 6.2 percent prior to the enhanced closure while 124 THR's had a dislocation rate of 0.8 percent following the enhanced closure. These results are highly statistically significant.

Dislocation is the second most common complication following total hip arthroplasty (THA) with reported incidence from less than one per cent to almost ten per cent. Multiple causes of dislocation have been identified including component malposition, muscle and soft tissue imbalance or absence, impingement, and various patient factors. Nonoperative treatment of a single dislocation is often effective, however one per cent of patient’s undergoing THA will require revision surgery for hip instability. Unfortunately the overall success rate of the various operative options is only about sixty per cent. The purpose of this study was to describe our clinical experience with a constrained acetabular insert. (Omnifit® Constrained Acetabular Insert, Osteonics Corporation, Allendale, NJ).

Between April 1988 and February 1993, 101 constraining acetabular components were implanted into 98 patients at one of two centers (Iowa Methodist Medical Center or Indiana University Hospital). Two-thirds of the cases were female; average age at time of surgery was 71 years (range, 31 to 92). Indications for use of the constraining component were recurrent dislocation (average six, range two to 20) in 56 cases, intraoperative instability in 38 cases, and neurologic impairment in seven cases. Average number of prior ipsilateral hip procedures was three. At latest follow-up, 74 patients (77 hips) were living, 23 patients (23 hips) had died, and one patient (one hip) was lost to follow-up. Living patients had an average clinical follow-up of 61 months (range, 24 to 97) while deceased patients had an average clinical follow-up of 29 months (range, one to 81).

At latest clinical follow-up 88 per cent had no or mild pain only and 72 per cent walked with no or a mild limp. There were four cases of recurrent dislocation following placement of the constraining component. In two of these cases component failure was a contributing factor. Fifty-four of the 56 cases (96 per cent) reoperated for recurrent dislocation have had no further dislocations.

In summary, this new constraining acetabular component was successful in preventing instability in over 97 per cent of living cases at an average five-year follow-up. This short-term clinical success rate appears to be superior to any previously reported treatment for total hip instability. However, it should be noted that the two surgeons (WNC, RCJ) who performed these cases were experienced in techniques of revision THA and have complex arthroplasty practices. The authors recommend judicious use of this component as a salvage measure for desperate cases of total hip instability because of potential long-term concerns for increased polyethylene wear and interface stresses.

We now have between one and two decades of experience using cementless femoral components in the total hip arthroplasty. Initially, with the development of this technology and procedure, there were manufacturing concerns (i.e. stem breakage and bead shedding) and technical concerns (i.e. learning curve, femoral fracture). The need for torsional stability at the time of implantation and the need for avoidance of non-circumferentially porous coatings was realized.

Four strategies have evolved with second generation stems: extensive porous coatings, proximal porous coatings with increased torsional stability, taper stem geometries, and the use of enhancement coatings, i.e. hydroxyapatite coatings. Excellent clinical and radiographic results have been reported with all of these components.

In the future, to decide if all of these strategies provide the same clinical satisfaction and the same long-term fixation, better assessments will be necessary. Prospective studies (preferably randomized trials) using both health care provider assessments and patient administered questionnaires along with the use of newer pain assessments (including pain diagrams) may give insight into any differences in clinical results obtained with various femoral implant designs.

This paper addresses the concerns that have been raised regarding the use of extensively porous-coated femoral components. These concerns include failure of fixation requiring revision, end-of-stem (thigh) pain, pronounced proximal bone resorption due to stress shielding, and osteolysis. We reviewed 509 consecutive, unselected cases (482 patients) treated with an extensively porous-coated stem of one design and a porous-coated cup. Fifty-six had died and 30 were lost prior to their five-year follow-up evaluation. There were thirteen reoperations that involved the femoral side of the arthroplasty (7 stem revisions and 6 cases with femoral osteolysis) for an overall femoral reoperation rate of 2.6%.

Of the 382 cases with clinical follow-up at a mean of 9 years, only 15 (4%) reported thigh pain severe enough to limit their activities. Only one of these 15 cases had fibrous tissue fixation; the other 14 were bone ingrown. Despite having notable thigh pain, 13 of 15 stated they were satisfied with the outcome of their operation.

Of the 445 cases with two-year radiographs, 113 (25%) had pronounced stress shielding (proximal bone resorption extending to the level of the lesser trochanter or below). Cases with pronounced stress shielding were older, weighed less, were predominantly female, had proportionally more large diameter stems (15.0 mm or greater), and had fewer reoperations. There was no difference in function or satisfaction at latest follow-up between patients with and those without stress shielding. Femoral osteolysis (cyst 1.5 cm² or larger) was observed in 11 (3.1%) of 351 unreoperated hips at a mean of 8.7 years radiographic follow-up. All of the lesions were confined to Gruen zones 1 and 7: five hips had a lesion in zone 1, three had a lesion in zone 7, and three had a lesion in zones 1 and 7.

Including the six cases reoperated because of femoral osteolysis, the overall incidence of femoral osteolysis in this patient group was 4.8 % (17 of 35). On the basis of these results and our clinical experience, we continue to use an extensively porous-coated femoral component in nearly all of our primary and revision cases.

Biological fixation of cementless femoral stems requires adequate cortical contact to resist out of plane rotational forces and prevent axial migration. Cortical contact can be achieved via metaphyseal or diaphyseal fit with extensive porous coating. Cementless components that rely on proximal cortical contact for fixation are as follows:

1. Proximal porous coating, non-anatomic designs.
2. Proximal porous coating, anatomic.
3. HA coated anatomic and non-anatomic.
4. Tapered geometry stems with proximal coatings.
5. Modular proximally coated.

Proximally coated, non-tapered stems with non-circumferential coating and smooth diaphysis had less bone ingrowth, unacceptable thigh pain and higher rates of femoral osteolysis. Third body wear generated from HA coated stems has been documented and particle generation from modular junctions is also a concern.

Achieving cortical contact is more difficult with all types of stems that rely on proximal fixation due to greater anatomic variance of the proximal femur. Fixation in a "champagne flute" type femur is more consistent but use of proximal devices in "stove pipe" femurs is less reliable. Extensively coated devices can be used in all femoral geometries since only isthmus fixation is needed.

Adjustments in leg length and offset by changing implant height are limited because optimal fit and fill is provided at one level of neck resection for any given implant size in anatomic and tapered designs. Increasing offset to prevent dislocation can only be accomplished by lengthening the leg. Using a smaller metaphyseal filling implant at a lower neck resection compromises stability. Large proximal filling prostheses also limit the ability to adjust component version. In contrast, extensively coated, non-anatomic stems can be adjusted up or down without affecting stem fit since fixation relies on an isthmus fit, thus adapting to a wider variety of internal femoral geometries.

Higher fracture rates occur with proximal cortical contact stems that have diaphyseal engagement from flutes, which direct the bulky proximal stem asymmetrically. Modular stems have overcome some of these problems. However, unlike extensively coated stems, distal biologic fixation does not occur, resulting in more micromotion that can lead to thigh pain in stovepipe femurs.

If there is inadequate proximal stability or proximal fracture altering fixation and the diaphysis has been reamed, there will be removal of cancellous bone. Conversion to a cemented device is not desirable since a hostile intramedullary environment for cement exists. This is an ideal situation for an extensively coated stem since the fracture can be bypassed.

Finally, removal of bulky ingrown anatomic or tapered devices is more destructive to the proximal femur than the non-proximal filling extensively coated device. The distal portion of the extensively coated device is easily removed with a hollow trephine after it is sectioned in-situ with a metal cutting burr in conjunction with an extended trochanteric osteotomy.

Ten years of clinical experience has given us insight into the biologic and mechanical properties of HA coated implants. The most frequently asked questions regarding HA are: does HA enhance bone ongrowth or ingrowth?; will HA lead to increased polyethylene wear or an increased incidence of osteolysis?; will HA disappear, and if so, what will be left to maintain fixation?

An overwhelming body of evidence from both animal studies and clinical experience has demonstrated an enhancement of bony ongrowth or ingrowth to an HA coated titanium implant when compared to smooth, grit blasted, or porous coated titanium implants. This enhanced bone apposition has been demonstrated to occur even in the presence of osteoporosis, gaps between the implant and bone, and early micromotion following implantation in contrast to the negative effects these factors have on non-HA coated press-fit cementless implants. In addition micromotion and subsidence of stable femoral stems has been shown to be less for HA coated implants compared to porous coated implants without HA of the same design. The six to ten year clinical results utilizing HA coated titanium femoral stems and HA threaded sockets have been excellent and comparable to or better than published results for porous coated implants for the same follow-up period.

Anecdotal reports have implicated HA particles to osteolysis. These reports are derived from failed and/or loose implants where histological sections of membranes have identified particles of metal, polyethylene, as well as HA. Autopsy and revision retrievals of HA implants utilized in our patient population have not revealed any evidence for accelerated polyethylene wear, scratching of the femoral head, or inflammatory reaction secondary to the HA particles. Animals studies have clearly shown that HA particles do not incite an inflammatory response and are resorbed through normal biologic processes as opposed to PMMA and polyethylene particles. Furthermore PMMA and metal particles when implanted into animal models result in scratching of chrome cobalt femoral heads whereas HA particles do not. Finally, ten years of clinical experience has shown that HA seals the femur against polyethylene migration and not a single case of endosteal lysis along an HA coated femoral stem has been identified.

When pure and highly crystalline HA is properly applied to a grit blasted titanium femoral stem in a thin dense layer it is both mechanically and biologically stable. Retrievals have demonstrated physiologic remodeling of the HA coating with replacement of bone directly adjacent to the titanium surface in areas of high stress. Clinical experience has demonstrated early and extensive bone growth and remodeling around HA coated femoral stems and DEXA studies have demonstrated increased bone accretion around proximally HA coated femoral implants. While early and lasting fixation has been seen with HA coated femoral stems and HA threaded acetabular components, the results using a smooth press-fit HA coated acetabular implant has not met expectations. These press-fit implants were well-fixed through four years, but we found unacceptable rates of aseptic loosening between four and six years. These results we believe demonstrate the need for an interlock of bone particularly in zone three to maintain long termfixation. HA threaded sockets as well as porous coated sockets achieve bony and/or soft tissue interlock to resist the force loads seen by the acetabular shell. This is in contrast to the femur where the combination of axial, bending, and rotational forces wedge a tapered component into the femur resulting in an accretion of bone and a secondary stabilization of the implant.

In conclusion, experimental and clinical data has provided convincing evidence that HA enhances bony attachment to titanium through osteoconductive mechanisms even in the presence of bone gaps, osteoporosis, and early micromotion. HA particles are not osteolytic in their own right and there is no scientific evidence that they lead to increased polyethylene wear. Circumferential coatings of HA on femoral components seal the femur against polyethylene particle migration and endosteal lysis has not been seen in the first decade of use. While HA clearly enhances bone ongrowth, the underlying design of the implant is crucial for long term success. We have shown that an interlock of bone is not necessary with a well designed HA coated femoral stem but does appear to be necessary on the acetabular component for predictable long term fixation.

As a class, tapered cementless total hip replacements have proven effective in relieving pain and restoring function to patients crippled with arthritis of the hip. The evidence favoring cementless tapers is overwhelming, including our randomized clinical trial in which we found cementless tapered implants to be equal to cemented total hip replacements in terms of every disease specific, patient specific, global health, functional capacity and cost/utility outcome measure which we examined. The prevalence of thigh pain was no different between a tapered cementless and cemented implant. Many others have substantiated these findings. In addition, tapered cementless devices have dealt with the issue of what is left after a failed total hip replacement in that stress shielding and osteolysis do not appear to be significant problems.

Our experience with more than 1,000 cementless tapered total hip replacements (CLS, Mallory Head, Synergy) has made this our implant of choice in 75% of patients. In the remaining 25% who have poorer bone stock because of age (over 75 years), secondary osteoporosis (inflammatory arthritis) or pre-existing bone disease (Paget's Disease or previous irradiation), we prefer to cement the femoral stem. The rationale for using alternative femoral stems in these cases is based on our desire to avoid the use of excessively stiff stems which stress shield the femur.

In summary, tapered cementless total hip replacements have proven effective, durable and easy to insert. Our algorithm for cementless tapered stems is as follows:

The efficacy, durability and limits of cementless femoral fixation have yet to be clearly delineated. This study outlines the experience of the author's group using a flat tapered wedge during the past 5 to 8 years in 2,616 patients.

The efficacy in terms of pain relief and restoration of function was studied in a typical patient population using, first, a matched pair analysis comparing patients with cement versus cementless fixation. A second group of patients was studied who had cemented fixation on one side and cementless fixation on the contralateral side. In both of these study groups equivalent pain relief and restoration of function was achieved using cementless fixation compared to contemporary cement fixation.

Durability was studied examining the first 100 patients undergoing cementless femoral fixation, with a revision rate under 1%. Radiographic parameters indicated dependable fixation and a low rate of lysis.

Of particular interest, two subgroups with compromised bone stock were examined to determine their appropriateness for cementless fixation - the elderly and the rheumatoid. Thirty-nine octogenarians were followed for 2 to 5 years with cementless femoral fixation and compared to a control group of 70 patients with cemented femoral fixation followed 2 to 9 years. Surprisingly, both groups achieved a Harris Hip Score of 85 and no revisions were noted in either group. Thigh pain was present in 10 cementless and 5 cemented patients. Radiographic data indicated good fixation in 94% of the cementless group with no measurable subsidence in either group.

Fifty-two rheumatoid hips were treated with cementless femoral fixation and followed 2 to 7 years (mean 3.6). They achieved excellent outcomes with an average Charnley Pain Score of 5.7. There was a 6% incidence of thigh pain and no fractures. There were no cases of radiographic loosening and 98% showed evidence of spot welds. No revisions were performed.

The authors' experience indicates dependable clinical and radiographic outcomes with cementless femoral fixation in an 8-year time frame. This is true not only in typical osteoarthritic patients, but also in octogenarians and rheumatoid patients as well. With good early outcomes, it is therefore possible that cementless fixation will allow long term advantages to become manifest in these patients.

Proximally coated femoral stem prostheses have been extensively used for nearly 15 years. Problems associated with first generation implants included inadequate instrumentation, insufficient size range, and discontinuities in the proximal coating. In correcting these deficits, 2nd generation implants have achieved a high degree of clinical and radiological success as reported at mid-term (5-7 years).

Proximally coated implants depend on achieving mechanical stability, in both the short and long-term, principally in the metaphysis, with only supplemental fixation in the diaphysis through the uncoated portion of the stem. In addition, permanent stability relies on biological fixation only within metaphyseal bone. Based on the authors' experience and the literature, there are a few conditions that do not lend themselves to proximally coated prosthesis. These include; metabolic bone disease (e.g. Gaucher's disease, sickle cell disease, osteomalacia), significant anatomical distortion (e.g. CDH with anteversion, prior osteotomy, intertrochanteric fracture), and Dorr C bone. Most revision situations do not lend themselves to proximal fixation and require alternative fixation.

The authors' extensive experience with proximally coated femoral stems leads them to the conclusion that excellent clinical and radiological results can be achieved in the majority of patients presenting for primary arthroplasty and in some revision situations. If the clinical results of proximally coated prostheses and extensively coated prostheses are equivalent, and the literature strongly suggests that they are, the principal advantages include less proximal bone stress shielding and easier removal, should that become necessary.