Vol 8, No 2, March/April 2000 75 Most early total knee replacement designs did not include resurfac- ing of the patellofemoral joint. Progression of patellofemoral ar- thritis and anterior knee pain were observed in many patients after these procedures. By the late 1970s condylar total knees with a pa- tellofemoral articulation and patel- lar component became the stan- dard for total knee replacement. Within a few years, however, com- plications related to the patella emerged as a common cause for reoperation. These complications included patellar fracture, patellar subluxation or dislocation, exten- sor mechanism disruption, and component wear, loosening, or dis- sociation. Many of these complications occurred only with, or were much more common with, resurfaced patellae. As many as 50% of total knee revision procedures were attributed to patellofemoral com- plications, 1 especially in the 1980s and early 1990s. 2 During this time period, dissociation or failure of a cementless metal-backed patellar component was the most common device-related complication of total knee replacement reported to the Food and Drug Administration. 3 The patellar component failed in part because of malalignment, mal- rotation, or suboptimal design of the femoral component; therefore, many of these knees would have failed even if the patella had not been resurfaced. As a result, nu- merous changes in surgical tech- nique and component design have occurred in the past decade. These changes have had a significant impact on the issue of patellar resurfacing in knee arthroplasty. Thus, much of the older literature on patellar resurfacing and patellar complications is no longer relevant to current practice. The issue of whether or not to resurface the patella in total knee arthroplasty remains controversial. Clinical practice varies widely Dr. Barrack is Professor of Orthopaedic Surgery and Director, Adult Reconstructive Surgery, Tulane University School of Medicine, New Orleans. Dr. Wolfe is Assis- tant Professor of Orthopaedic Surgery and Director, Musculoskeletal Tumor Surgery, Tulane University School of Medicine. One of the authors or the department with which they are affiliated has received some- thing of value from a commercial or other party related directly or indirectly to the sub- ject of this article. Reprint requests: Dr. Barrack, Department of Orthopaedic Surgery, Tulane University School of Medicine, 1430 Tulane Avenue SL32, New Orleans, LA 70112. Copyright 2000 by the American Academy of Orthopaedic Surgeons. Abstract Whether or not to resurface the patella when performing a primary total knee arthroplasty remains an open question. A number of recent studies have added new information relevant to this controversy. Anatomic studies show that there is normally substantial variability in the anatomy of the trochlear groove. Implanting a femoral component therefore results in a change in the surface topography of the knee in a high percentage of cases. Even though a number of intraoperative techniques have been described in an attempt to accurately repro- duce femoral and tibial component rotation, studies of the application of these techniques reveal that component malpositioning or malrotation of a measur- able degree occurs in 10% to 30% of cases, depending on the surgical technique and landmarks used. There has been substantial change in the design of both femoral and patellar components in recent years. Even with current designs, biomechanical studies indicate that some degree of change in kinematics and contact stresses occurs following total knee arthroplasty. However, the results of clinical studies have been extremely variable, with most showing either no difference or very little difference between resurfaced and nonresurfaced patellae in osteoarthritic knees. The decision to resurface the patella or not must be indi- vidualized on the basis of the surgeonÕs training and experience and an intraop- erative assessment of the patellofemoral articulation. J Am Acad Orthop Surg 2000;8:75-82 Patellar Resurfacing in Total Knee Arthroplasty Robert L. Barrack, MD, and Michael W. Wolfe, MD Perspectives on Modern Orthopaedics throughout the world. Patellar resurfacing is usually performed as part of total knee arthroplasty in North America, but is frequently not performed in Japan and many European countries. The wide variability in practice was con- firmed by a recent survey by the British Orthopaedic Association. Of orthopaedic surgeons respond- ing in the United Kingdom, 32% routinely resurfaced the patella, 19% never resurfaced the patella, and 49% resurfaced the patella sometimes, the most common rea- son being patellofemoral arthritis seen at surgery. 4 Recent clinical and laboratory studies evaluating knee anatomy, surgical technique, component design, and biome- chanics have added to our under- standing of the impact of the pa- tellofemoral joint on total knee arthroplasty. Anatomic Studies Anatomic studies have emphasized the variability of the configuration of the distal femur and trochlear groove. The transepicondylar axis is frequently used as a landmark to orient the femoral component. Feinstein et al 5 performed electronic digitization of three-plane radio- graphs of 15 cadaveric femurs. They found that the patellar groove was on average perpendicular to the transepicondylar axis in the coronal plane, but that the range of orientations varied through an arc of about 16 degrees ( − 8.7 to +7.2 de- grees). Similar ranges were ob- served for the transverse and sagit- tal planes. Eckhoff et al 6 studied the morphology of the sulcus in 85 femoral specimens and found that the average displacement of the femoral sulcus from a plane mid- way between the two femoral epi- condyles was 2.4 mm lateral (Fig. 1), but with a wide range (from 2.5 mm medial to 8.3 mm lateral). Both studies used normal, nonarthritic knees. There is evidence to suggest that arthritic knees are more likely to have anatomic abnormalities of limb alignment and rotation. 7 Ana- tomic studies of normal knees may not be representative of the anato- my of the degenerative knee. Con- versely, arthritic knees may have anatomic abnormalities that the surgeon may not want to reproduce when performing an arthroplasty. A study of arthritic knees was undertaken by Poilvache et al, 8 who correlated their findings with radiographic measurements. They found the transepicondylar axis to be the most consistent in determin- ing femoral component rotation. This axis was externally rotated 3.5 degrees to a line tangential to the posterior condylar surfaces, on average, but there was a standard deviation of 2 degrees. The authors noted substantial anatomic differ- ences between knees in male and female subjects and questioned the necessity of external rotation of the femoral component in all knees. Moreland is often credited as first to emphasize the necessity of externally rotating the femoral component to produce a rectangu- lar flexion gap. 9,10 He showed that this was necessary to match the asymmetric cut of the proximal tibia that resulted from cutting per- pendicular to the long axis of the tibia in the face of an average 3- degree varus slope (Fig. 2). Anatomic studies have consis- tently shown that there is wide variation in the anatomy and orien- tation of the distal femur, and thus of the patellofemoral joint. Sur- geons tend to orient the compo- nents in the same position despite these individual differences. Per- forming knee arthroplasty on the basis of measurements of the aver- age knee will invariably result in a substantial change in orientation in at least 10% to 20% of patients. Surgical Technique The critical surgical factors in patel- lar resurfacing are maintaining the preoperative patellar thickness, per- forming a symmetric bone resec- Patellar Resurfacing Journal of the American Academy of Orthopaedic Surgeons 76 Figure 1 In an anatomic study of 85 knees, Eckhoff et al 6 found that the femoral sulcus was located lateral to the midplane of the femoral condyles. (Adapted with permission from Eckhoff DG, Burke BJ, Dwyer TF, Pring ME, Spitzer VM, Van- Gerwen DP: Sulcus morphology of the dis- tal femur. Clin Orthop 1996;331:23-28.) Figure 2 A perpendicular tibial cut requires femoral cuts made by externally rotating the cutting block with reference to the femur in order to maintain a balanced flexion gap, as represented by the center drawing. Drawing on left demonstrates that perpendicular tibial cut without femoral-cut external rotation results in a trapezoidal flexion gap; drawing on right, that 3-degree-varus tibial cut without femoral external rotation results in a rec- tangular flexion gap. (Reproduced with permission from Moreland JR: Mecha- nisms of failure in total knee arthroplasty. Clin Orthop 1988;226:49-64.) Sulcus Midplane tion, and balancing the extensor mechanism. 11 This can be achieved by using a caliper before and after patellar resection to ensure that the composite will reproduce the prere- section patellar thickness and that equal bone thickness remains in all locations. Underresection can lead to restricted flexion and anterior knee pain (Fig. 3). Overresection can predispose to fracture. Asym- metric resection can lead to patellar instability. A patellar cutting guide can assist in achieving more consis- tent patellar cuts. Trial reduction is necessary to assess tracking. The Òrule of no thumbÓ is often recommended to assess the need for lateral release. 12 Component positioning appears to be critical to optimization of patellar tracking. Different studies have suggested that optimal positioning is medial and superior placement of the patellar component, 13 external rota- tion and lateral placement of the femoral component, and lateral placement of the tibial component. Component orientations that should specifically be avoided include femoral or tibial internal rotation, anterior placement or flexion of the femoral component, and increasing the anteroposterior diameter of the distal femur by oversizing the fe- moral component. Despite adhering to all of these principles, one experienced sur- geon reported an 18% incidence of asymmetric patellar resection depth and a 10% incidence of patellar tilt 2.5 years after 50 total knee replacements performed with pa- tellar resurfacing; changes in patel- lar thickness, joint line, and patellar height were also seen in several patients. 11 In other studies, 14,15 pa- tellar tilt, asymmetry, and subluxa- tion occurred in 7% to 14% of total knee replacements with patellar resurfacing. Bindelglass et al 16 found that only 55% of their resur- faced patellae tracked centrally. However, asymmetry and patellar tilt did not correlate with symptoms in any of these studies. In contrast, Pagnano and Trousdale 17 reported asymmetric resurfacing in 21 of 300 knees (7%). Patellofemoral compli- cations were subsequently noted in 11 of those 21 knees (52%) (Fig. 4). Component design and surgical technique are major factors in the success of knee arthroplasty, and both influence patellar tracking. In the estimation of Bindelglass and Dorr, 18 surgical technique was of greater importance than design fac- tors, such as an asymmetric femoral component, raised lateral flange, or angled trochlear groove. Several studies have shown a wide varia- tion in surgical technique. In the study by Eckhoff et al, 19 evaluation of four different methods of deter- mining tibial rotation yielded a range of rotation from 2 degrees of internal rotation to 19 degrees of external rotation. Likewise, Olcott and Scott 20 evaluated the four most commonly used methods of deter- mining femoral component rota- tion, and determined that flexion gap asymmetry of at least 3 degrees occurred in 10% to 30% of cases, depending on the landmarks used. Acceptable component align- ment is not achieved in at least 10% of cases, as they have a measurable degree of component malposition- ing or malrotation when current techniques and instruments are used. This is clinically significant because component malrotation has been found to closely correlate with the occurrence of patello- femoral complications. 21 The ques- tion is then whether the resurfaced patella or the native patella is bet- ter suited to withstand some de- gree of component malpositioning. Component Design Because of the high incidence of patellar complications, the design of patellar and femoral components has changed considerably over the past 20 years (Fig. 5). In the 1980s, the goal of minimal bone resection, particularly in cruciate-retaining designs, predominated. This re- sulted in thin distal femoral and chamfer cuts and a shallow troch- lear groove. Robert L. Barrack, MD, and Michael W. Wolfe, MD Vol 8, No 2, March/April 2000 77 Figure 3 Underresection of the patella creates a patella-prosthesis composite that is too thick, ÒoverstuffingÓ the anterior compartment. This can be associated with painful, restricted flexion and accelerated polyethylene wear. Figure 4 In this case, an asymmetric patellar cut resulted in painful lateral sub- luxation. Note that the patellar remnant is thicker medially than laterally. A study by Petersilge et al 22 showed that a deeper trochlear groove (requiring more inter- condylar bone resection) reduced mediolateral shear on the patellar component. A statistically signifi- cant improvement in gait and stair-climbing was reported by Andriacchi et al, 23 who used a component with a deeper troch- lear groove, which more closely reproduced the femoral radius of curvature (Fig. 6). In that study there was significant improve- ment in function during stair climbing. Total knee designs with a deeper, more congruent patellofemoral articulation require a separate in- tercondylar bone cut. Hsu and Walker 24 tested a variety of patellar components under load, and found that increased conformity of the patellar component decreased the predicted amount of deformation and wear. McLean et al 25 reported one of the few studies in which the effect of femoral component design on the nonresurfaced patella was examined in cadaveric knees. They found that the five components tested all altered contact character- istics (as measured with the use of pressure-sensitive film) to some extent. During terminal extension (0 to 30 degrees), patella-prosthesis contact was satisfactory. However, contact area was decreased 15% to 25% at 60 degrees of flexion, as compared with a normal knee; this continued to decrease at higher flexion angles. Electrogoniometry at 6 degrees of freedom showed that two of the five femoral compo- nents induced 10 to 12 degrees of abnormal tilt of the nonresurfaced patella. Although some femoral components appeared to be better suited for the nonresurfaced patella than others, none replicated the natural state. In fact, two Òana- tomicÓ designs tended to show im- pingement of the patella on the intercondylar notch at high flexion angles. Biomechanical Studies Biomechanical studies of the patel- lofemoral joint in total knee arthro- plasty have focused on either kine- matics or contact stresses. Chew et al 26 examined three-dimensional tracking of the patella by using a magnetic tracking device. Three current designs were tested, all of which have fairly deep, congruent patellofemoral articulations. Com- pared with a normal knee, there was no statistically significant dif- ference in lateral shift or patellar rotation (clockwise or counterclock- wise coronal-plane motion). How- ever, all three components showed significant lateral tilt of approxi- mately 5 degrees, compared with 0.44 degree for the normal patella (Fig. 7). In another study, similar technology was used to examine the effect of patellar thickness on kine- matics. While neither a thicker nor a thinner composite thickness sig- nificantly affected patellar tracking, it was predicted that both would Patellar Resurfacing Journal of the American Academy of Orthopaedic Surgeons 78 Figure 5 There is a wide variation in the geometry and congruency of the trochlear groove of the femoral components of various designs. Figure 6 Correlation of depth of patellofemoral groove with patellofemoral joint function in the study by Andriacchi et al. 23 Left, Configurations of patellar components used in the two groups of patients. Right, The deeper patellofemoral groove of the component used in group 2 patients was associated with more normal knee moment (pattern 1), which resulted in improved function in stair climbing. (Knee moment is measured as the ratio of body weight to height, expressed as a percentage; Qmax and Qmin represent maximum and minimum flexion moment, respectively.) Patients in group 1 were more likely to have the abnormal pattern 2. (Adapted with permission from Andriacchi TP, Yoder D, Conley A, Rosenberg A, Sum J, Galante JO: Patellofemoral design influences function following total knee arthroplasty. J Arthroplasty 1997;12:243-249.) 0.0 2.5 2.5 5.0 Stance Swing Qmax 1 Qmin 1 Pattern 1 Knee Moment Flexion Extension Qmax 2 Pattern 2 Group 2 Group 1 decrease contact area and thus in- crease contact stress. Virtually all studies of contact areas and stresses have demon- strated a substantial decrease in contact area, and a consequent in- crease in contact stress, whenever knee replacement is performed. Matsuda et al 28 found that resurfac- ing decreased contact area com- pared with not resurfacing the patella, a finding confirmed by Benjamin et al. 29 The magnitude of the contact stress varies with the femoral com- ponent design. Buechel et al 30 tested a variety of components and found contact stresses below 10 MPa only with a mobile bearing design. This finding was contradicted by Mat- suda et al, 28 who found much higher peak contact stresses with both con- forming and dome patellar resurfac- ing compared with the nonresur- faced patella. With the conforming LCS patellar component (Low Contact Stress, DePuy, Warsaw, Ind), they found that the thick prox- imal pole tented the quadriceps ten- don with flexion, thus decreasing contact forces between the tendon and the femoral component and transmitting high peak contact stress to the proximal pole of the patellar component. The stresses measured in virtual- ly all of these studies exceed the yield point of polyethylene, and would predict much higher rates of polyethylene wear and component failure than have been observed clinically. The polyethylene may experience cold flow, which would change the contact stresses over time. McNamara et al 31 tested pa- tellar components in which cold flow had already occurred and found only a minimal decline in the contact pressures. A more likely explanation is soft-tissue adapta- tion, rather than a change in the shape of the polyethylene. A pseu- domeniscus of fibrous tissue often forms around the unloaded portion of a patellar component, which may transfer load to the peripheral over- lying soft tissue and thereby dra- matically change the contact area and the stresses exerted on the com- ponent. Biomechanical studies can- not take these biologic phenomena into account. Tracking studies are also performed passively, on cadav- er specimens, and cannot replicate the exact magnitude, direction, or force of active muscle contractions. Clinical Studies All of the laboratory studies of the patellofemoral articulation are based on experimental models and as- sumptions that may not replicate con- ditions in vivo. In addition, it is diffi- cult to replicate the biologic response that occurs over time. For these rea- sons and others, surgeons tend to rely more heavily on the results of clinical studies in judging issues such as whether or not to resurface the patella. Numerous studies relevant to this topic have appeared in the literature over the years. Most early studies were retrospective and uncontrolled. More recently, there have been pub- lished series with control groups, as well as reports of prospective, ran- domized studies. In series of total knee arthroplas- ties in which patellar resurfacing was routinely performed, implant survival rates of 90% to 95% were reported at 10 years or longer, with a patellar complication rate of less than 5%. Without a control group, it is difficult to determine how these patients would have fared without patellar resurfacing. In ad- dition, most of these studies were by experienced surgeons at total joint centers. The complication rate of patellar resurfacing in the hands of less experienced surgeons is un- known and may be higher. A number of studies have dem- onstrated inferior results with a nonresurfaced patella. Picetti et al 32 performed an uncontrolled series of 100 total condylar knee replace- Robert L. Barrack, MD, and Michael W. Wolfe, MD Vol 8, No 2, March/April 2000 79 intact Genesis II NexGen P.F.C. 1 2 0 − 2 − 4 − 6 − 8 − 10 2345678 Tilt, degrees (medial tilt is positive) Range of flexion (in 15-degree intervals) Figure 7 In a kinematic cadaver study, Chew et al 26 tested three modern components (Genesis II, Smith & Nephew Orthopaedics, Memphis; NexGen, Zimmer, Warsaw, Ind; P.F.C. Sigma, Johnson & Johnson, New Brunswick, NJ) and found a statistically significant lateral tilt of the patella compared with the normal knee. Range of flexion (in 15-degree intervals) is indicated by the values on the horizontal axis (e.g., Ò4Ó indicates 60 degrees of flexion). (Adapted with permission from Chew JTH, Stewart NJ, Hanssen AD, Luo ZP, Rand JA, An KN: Differences in patellar tracking and knee kinematics among three differ- ent total knee designs. Clin Orthop 1997;345:87-98.) ments without patellar resurfacing, after which 29% of patients had postoperative anterior knee pain. As a result of this outcome, criteria for resurfacing were suggested, which were based on height, weight, intraoperative chondromalacia grade, and the presence of preopera- tive anterior knee pain. These crite- ria are very restrictive and would, in fact, result in almost routine re- surfacing. Boyd et al 33 reported a retrospec- tive study of 891 knee replacements in 684 patients. The overall compli- cation rate was much higher with a nonresurfaced patella (12% [58 of 495 knees]) than with a resur- faced patella (4% [16 of 396 knees]). When patients with inflammatory arthritis were excluded, however, the complication rates were compa- rable (6% in the nonresurfaced group versus 4% in the resurfaced group). The implant used in this series, which is no longer available, was characterized as a Òsymmetrical femoral component with a non- anatomic unconstrained patello- femoral groove.Ó This would gen- erally not be consistent with current design principles for the patello- femoral articulation. In addition, the methods of component instru- mentation and positioning have changed substantially since that time, and many of these changes are thought to affect patellar track- ing. For these reasons and others, it is difficult to apply the results of older studies to current practice. Stair climbing, an activity that requires high patellofemoral con- tact forces, has often been used as the benchmark for judging the per- formance of the patellofemoral articulation. For example, Schroeder- Boersch et al 34 reported better scores in stair climbing with patel- lar resurfacing than without resur- facing and therefore recommended routine resurfacing. Studies comparing the same knee implant system with and without patellar resurfacing have demonstrated no difference in clin- ical outcome. 35-37 A number of studies have compared patients undergoing bilateral total knee arthroplasty with one side resur- faced and the other not resurfaced. Keblish et al 38 reported no differ- ence in patient preference or per- formance on stairs at 5 years with bilateral LCS total knees with a mobile-bearing patellar compo- nent. In contrast, Enis et al, 39 using a Synatomic patella and a Townley Knee (DePuy), found that patients more frequently preferred the resurfaced side and had better function with resurfacing as well. The expected result may therefore be related to the component de- sign. Shoji et al 40 reported no dif- ference in function at a minimum 2-year follow-up (average, 2.7 years) in a study of 35 patients with rheumatoid arthritis who received bilateral modified total condylar knees with the patella resurfaced on only one side. However, when Kajino et al 41 evaluated the same group of patients at a minimum follow-up interval of 6 years, they noted a higher incidence of anterior knee pain and difficulty with stairs in those patients with a nonresur- faced patella. These findings may reflect a decline in function and an increase in symptoms for nonresur- faced patellae at longer postopera- tive intervals. In recent years, prospective, ran- domized studies examining the issue of patellar resurfacing have begun to appear. Barrack et al 42 reported no difference in Knee Society scores or patient satisfac- tion between patients with and without patellar resurfacing but a 10% incidence of subsequent patel- lar resurfacing when the MG-II prosthesis (Zimmer, Warsaw, Ind) was used. In a similar study in which the AMK Knee (DePuy) was used, Bourne et al 43 reported less pain and higher flexion torque in patients with nonresurfaced patel- lae, and a revision rate of only 4%. Feller et al 44 reported better stair climbing by patients who under- went implantation of the PCA modular prosthesis (Howmedica, Rutherford, NJ) without resurfac- ing, compared with patients who received the same implant but with resurfacing. There is obviously disparity in the reported results of total knee replacement with and without resurfacing of the patella. The dif- ferences may be related to the im- plants used, the surgical technique, and the methods of evaluation and reporting of results. Summary The three options facing the sur- geon performing total knee arthro- plasty are to resurface the patella routinely, selectively, or not at all. There is consensus in the literature that resurfacing should be favored when inflammatory arthritis is pres- ent, when the patella is severely deformed (as with malunion of a prior patellar fracture), or when patellofemoral joint degeneration is the primary indication for the oper- ation. The literature supports leav- ing the patella nonresurfaced when it is particularly small and thin, especially when osteopenia is also present. 45 Although there is little supporting data, many experienced surgeons believe that patellar re- tention is also appropriate for very young patients, especially when the patellar articular surface ap- pears normal. For the typical osteo- arthritic knee that does not fit into one of these categories, there is support in the literature for both patellar resurfacing and patellar retention. If the patella is resurfaced, care- ful attention must be paid to the surgical technique to minimize com- plications. Potential pitfalls include Patellar Resurfacing Journal of the American Academy of Orthopaedic Surgeons 80 overresection and underresection of the patella, asymmetric patellar resection, femoral component mal- rotation, and extensor-mechanism soft-tissue imbalance. The latter two factors are also important when the patella is not resurfaced. It remains unclear whether the resur- faced or the nonresurfaced patella is better able to tolerate suboptimal component positioning and soft- tissue balancing. A mobile-bearing patella, such as in the LCS total knee prosthesis, could theoretically toler- ate minor femoral or patellar com- ponent malpositioning better than a fixed-bearing device, because its tracking is self-correcting to some degree. Clinical results with this prosthesis have shown a low patel- lofemoral complication rate despite its being metal-backed, a design fea- ture associated with the early poor results reported for fixed-bearing patellar devices. If the decision is made not to resurface the patella, a femoral com- ponent that is compatible with the native patella should be chosen. The optimal design probably in- cludes a deep trochlear groove that extends far enough distally to main- tain support of the patella in flexion and avoids catching of the patella on its edge. 25 Although some fe- moral components have been dem- onstrated in clinical and biomechan- ical studies to be more compatible with the native patella, even so- called ÒanatomicÓ designs fail to replicate normal knee kinematics and contact stresses, and can be associated with patellofemoral com- plications. It is important that the patient understand that the need for subsequent resurfacing is a possible consequence of patellar retention. There does not seem to be a femoral component that has been specifically designed for use with the native pa- tella; therefore, the best possible results with patellar retention may remain unrealized, despite prospec- tive, randomized studies. In addi- tion, most studies of total knee ar- throplasty with a nonresurfaced patella have had only short follow- up. The durability of patellar reten- tion in the long-term has therefore not been well documented. With proper component orienta- tion, limb alignment, and balancing of the knee ligaments and extensor mechanism, an excellent result can be expected in the vast majority of total knee replacements, whether or not the patella is resurfaced. Still, some patients seem to have an- terior knee pain after knee replace- ment regardless of whether the patella is resurfaced. 42 Part of the difficulty in resolving this dilemma is that the exact cause of this pain is unknown. It may represent subtle component malpositioning, in- creased tension or pressure on the soft tissue from altered kinematics, a biologic response such as fibrous band formation, or some combina- tion of factors. With the components, instru- mentation, and techniques currently available, patellofemoral complica- tions should no longer be the most common reason for total knee revi- sion, or even a particularly com- mon reason for reoperation. The patella should be routinely resur- faced for the indications outlined. For the remaining patients, who constitute the majority, this deci- sion must be individualized on the basis of the surgeonÕs training and experience and an intraoperative assessment of the patellofemoral articulation. Robert L. Barrack, MD, and Michael W. Wolfe, MD Vol 8, No 2, March/April 2000 81 References 1. Dennis DA: Patellofemoral complications in total knee arthroplasty: A literature review. Am J Knee Surg 1992;5:156-166. 2. Healy WL, Wasilewski SA, Takei R, Oberlander M: Patellofemoral compli- cations following total knee arthro- plasty: Correlation with implant design and patient risk factors. J Arthroplasty 1995;10:197-201. 3. Castro FP Jr, Chimento G, Munn BG, Levy RS, Timon S, Barrack RL: An analysis of Food and Drug Admini- stration medical device reports relat- ing to total joint components. J Arthroplasty 1997;12:765-771. 4. Phillips AM, Goddard NJ, Tomlinson JE: Current techniques in total knee re- placement: Results of a national survey. Ann R Coll Surg Engl 1996;78:515-520. 5. Feinstein WK, Noble PC, Kamaric E, Tullos HS: Anatomic alignment of the patellar groove. Clin Orthop 1996;331: 64-73. 6. Eckhoff DG, Burke BJ, Dwyer TF, Pring ME, Spitzer VM, VanGerwen DP: Sulcus morphology of the distal femur. Clin Orthop 1996;331:23-28. 7. Eckhoff DG: Effect of limb malrotation on malalignment and osteoarthritis. Orthop Clin North Am 1994;25:405-414. 8. Poilvache PL, Insall JN, Scuderi GR, Font-Rodriguez DE: Rotational land- marks and sizing of the distal femur in total knee arthroplasty. Clin Orthop 1996;331:35-46. 9. Moreland JR: Mechanisms of failure in total knee arthroplasty. Clin Orthop 1988;226:49-64. 10. Moreland JR, Bassett LW, Hanker GJ: Radiographic analysis of the axial alignment of the lower extremity. J Bone Joint Surg Am 1987;69:745-749. 11. Rand JA: Patellar resurfacing in total knee arthroplasty. Clin Orthop 1990; 260:110-117. 12. Ewald FC: Leg-lift technique for simultaneous femoral, tibial, and patellar prosthetic cementing, Òrule of no thumbÓ for patellar tracking, and steel rod rule for ligament tension. Techniques Orthop 1991;6:44-46. 13. Scuderi GR, Insall JN, Scott WN: Patellofemoral pain after total knee arthroplasty. J Am Acad Orthop Surg 1994;2:239-246. 14. Ranawat CS: The patellofemoral joint in total condylar knee arthroplasty: Pros and cons based on five- to ten- year follow-up observations. Clin Orthop 1986;205:93-99. 15. Sneppen O, Gudmundsson GH, BŸnger C: Patellofemoral function in total condylar knee arthroplasty. Int Orthop 1985;9:65-68. 16. Bindelglass DF, Cohen JL, Dorr LD: Patellar tilt and subluxation in total knee arthroplasty: Relationship to pain, fixation, and design. Clin Orthop 1993;286:103-109. 17. Pagnano MW, Trousdale RT: Inadver- tent asymmetric resurfacing of the patella in total knee arthroplasty. Orthop Trans 1998-1999;22:19. 18. Bindelglass DF, Dorr LD: Symmetry versus asymmetry in the design of total knee femoral components: An unresolved controversy. J Arthroplasty 1998;13:939-944. 19. Eckhoff DG, Metzger RG, Vandewalle MV: Malrotation associated with implant alignment technique in total knee arthroplasty. Clin Orthop 1995; 321:28-31. 20. Olcott CW, Scott RD: Femoral compo- nent rotation during total knee arthro- plasty. Clin Orthop 1999;367:39-42. 21. Berger RA, Crossett LS, Jacobs JJ, Rubash HE: Malrotation causing patellofemoral complications after total knee arthroplasty. Clin Orthop 1998;356:144-153. 22. Petersilge WJ, Oishi CS, Kaufman KR, Irby SE, Colwell CW Jr: The effect of trochlear design on patellofemoral shear and compressive forces in total knee arthroplasty. Clin Orthop 1994; 309:124-130. 23. Andriacchi TP, Yoder D, Conley A, Rosenberg A, Sum J, Galante JO: Patellofemoral design influences func- tion following total knee arthroplasty. J Arthroplasty 1997;12:243-249. 24. Hsu HP, Walker PS: Wear and defor- mation of patellar components in total knee arthroplasty. Clin Orthop 1989; 246:260-265. 25. McLean CA, Tanzer M, Laxer E, Casey J, Ahmed AM: The effect of femoral component designs on the contact and tracking characteristics of the unresur- faced patella in TKA. Orthop Trans 1995;18:616-617. 26. Chew JTH, Stewart NJ, Hanssen AD, Luo ZP, Rand JA, An KN: Differences in patellar tracking and knee kinemat- ics among three different total knee designs. Clin Orthop 1997;345:87-98. 27. Hsu HC, Luo ZP, Rand JA, An KN: Influence of patellar thickness on patel- lar tracking and patellofemoral contact characteristics after total knee arthro- plasty. J Arthroplasty 1996;11:69-80. 28. Matsuda S, Ishinishi T, White SE, Whiteside LA: Patellofemoral joint after total knee arthroplasty: Effect on contact area and contact stress. J Arthroplasty 1997;12:790-797. 29. Benjamin JB, Szivek JA, Hammond AS, Kubchandhani Z, Matthews AI Jr: Contact areas and pressures between native patellae and prosthetic femoral components. Orthop Trans 1998-1999; 22:18. 30. Buechel FF, Pappas MJ, Makris G: Evaluation of contact stress in metal- backed patellar replacements: A pre- dictor of survivorship. Clin Orthop 1991;273:190-197. 31. McNamara JL, Collier JP, Mayor MB, Jensen RE: A comparison of contact pressures in tibial and patellar total knee components before and after ser- vice in vivo. Clin Orthop 1994;299: 104-113. 32. Picetti GD III, McGann WA, Welch RB: The patellofemoral joint after total knee arthroplasty without patellar resurfacing. J Bone Joint Surg Am 1991;72:1379-1382. 33. Boyd AD Jr, Ewald FC, Thomas WH, Poss R, Sledge CB: Long-term compli- cations after total knee arthroplasty with or without resurfacing of the patel- la. J Bone Joint Surg Am 1993;75:674-681. 34. Schroeder-Boersch H, Scheller G, Fischer J, Jani L: Advantages of patel- lar resurfacing in total knee arthro- plasty: Two-year results of a prospec- tive randomized study. Arch Orthop Trauma Surg 1998;117:73-78. 35. Ishinishi T, Ogata K, Hara M: Com- parison of cementless and cemented fixation resurfaced and unresurfaced patella on the press fit condylar type total knee arthroplasty [poster]. Presented at the 66th Annual Meeting of the American Academy of Ortho- paedic Surgeons, Anaheim, Calif, February 4-8, 1999. 36. Braakman M, Verburg AD, Bronsema G, van Leeuwen WM, Eeftinck MP: The outcome of three methods of patellar resurfacing in total knee arthroplasty. Int Orthop 1995;19:7-11. 37. Abraham W, Buchanan JR, Daubert H, Greer RB III, Keefer J: Should the pa- tella be resurfaced in total knee arthro- plasty? Efficacy of patellar resurfacing. Clin Orthop 1988;236:128-134. 38. Keblish PA, Varma AK, Greenwald AS: Patellar resurfacing or retention in total knee arthroplasty: A prospec- tive study of patients with bilateral replacements. J Bone Joint Surg Br 1994;76:930-937. 39. Enis JE, Gardner R, Robledo MA, Latta L, Smith R: Comparison of patellar resurfacing versus nonresurfacing in bilateral total knee arthroplasty. Clin Orthop 1990;260:38-42. 40. Shoji H, Yoshino S, Kajino A: Patellar replacement in bilateral total knee arthroplasty: A study of patients who had rheumatoid arthritis and no gross deformity of the patella. J Bone Joint Surg Am 1989;71:853-856. 41. Kajino A, Yoshino S, Kameyama S, Kohda M, Nagashima S: Comparison of the results of bilateral total knee arthroplasty with and without patellar replacement for rheumatoid arthritis: A follow-up note. J Bone Joint Surg Am 1997;79:570-574. 42. Barrack RL, Wolfe MW, Waldman DA, Milicic M, Bertot AJ, Myers L: Resur- facing of the patella in total knee arthroplasty: A prospective, random- ized, double-blind study. J Bone Joint Surg Am 1997;79:1121-1131. 43. Bourne RB, Rorabeck CH, Vaz M, Kramer J, Hardie R, Robertson D: Resurfacing versus not resurfacing the patella during total knee arthroplasty. Clin Orthop 1995;321:156-161. 44. Feller JA, Bartlett RJ, Lang DM: Patel- lar resurfacing versus retention in total knee arthroplasty. J Bone Joint Surg Br 1996;78:226-228. 45. Kim BS, Reitman RD, Schai PA, Scott RD: Selective patellar nonresurfacing in total knee arthroplasty: 10 year results. Clin Orthop 1999;367:81-88. Patellar Resurfacing Journal of the American Academy of Orthopaedic Surgeons 82