Dislocation of Rotating Hinge Total Knee Prostheses A BIOMECHANICAL ANALYSIS

Transcription

1 448 COPYRIGHT 2003 BY THE JOURNAL OF BONE AND JOINT SURGERY, INCORPORATED Dislocation of Rotating Hinge Total Knee Prostheses A BIOMECHANICAL ANALYSIS BY WILLIAM G. WARD, MD, DAVID HAIGHT, MD, PAUL RITCHIE, MD, STAN GORDON, BS, AND JEFFREY J. ECKARDT, MD Investigation performed at the Department of Orthopaedic Surgery, Wake Forest University Baptist Medical Center, Winston-Salem, North Carolina Background: Symptomatic instability and implant dislocation are occasionally encountered in patients with a rotating hinge total knee prosthesis. A biomechanical study of rotating hinge total knee implants was performed to determine the association between the design (length and taper) of the central rotational stem and the stability of the implant. Methods: The stem lengths and tapers of knee implants made by seven manufacturers were measured. The tilting laxity of each design was tested by measuring the degree of tilting of the central rotational stem within the tibial housing that occurred with increasing amounts of distraction. The maximum amount of distraction that was possible before the stem dislocated was determined for each design. Results: Implant designs with a short and/or markedly tapered central rotational stem had the greatest tilting, laxity, and instability of that stem. The Howmedica, Techmedica, Intermedics/Sulzer Medica, and Wright Medical Technology/ Dow Corning Wright designs required 39 mm of distraction before they dislocated. The Biomet knee implant required 33 or 44 mm of distraction to dislocate, depending on the thickness of the polyethylene tray that was utilized. The S-ROM knee required only 26 mm of distraction before dislocation occurred. Conclusions: The measurements confirmed that the shorter the stem and the greater its taper, the greater the instability and laxity at any given amount of joint distraction. Clinical Relevance: Rotating hinge knee designs with a short, tapered central rotational stem (in the absence of a mechanical stop to distraction) should be used with caution in patients with bone and soft-tissue compromise that may allow excessive distraction and implant dislocation, especially when the patient has a flexion-extension gap mismatch or a flexion gap laxity. Rotating hinge knee implants were designed to provide a stable knee reconstruction when the intrinsic softtissue stability of the knee had been lost as a result of tumor resection, trauma, multiple knee replacements, deformity, or surgical reconstruction 1-5. We have extensive experience with the instability encountered following tumor resections around the knee joint. Most tumor resections and occasionally revision knee arthroplasties require division and/or resection of the entire capsule in addition to division of all cruciate and collateral ligaments. Following such procedures, there is a greater laxity in flexion than in extension. Distraction laxity in extension is resisted by the hamstring muscletendon units, the neurovascular structures, and the skin, which typically prevent distraction of more than several millimeters. Distraction in flexion is prevented only by tension on the extensor mechanism. Intraoperatively, with the patella everted, which removes the stabilizing effect of the extensor mechanism, these knees can usually be distracted at least 5 cm. This flexion-distraction positioning is the standard maneuver used to insert the rotating stem component of a rotating hinge total knee prosthesis. Given that this flexion gap laxity can occur, an analysis of the stability provided by the various rotating hinge knee designs, under increasing amounts of distraction, can provide useful information for the surgeon. Rotating hinge knee designs have a transverse (horizontally) oriented hinge axis that provides flexion-extension motion and a vertically oriented post-in-channel axis that allows internal and external rotation. Each of the rotating hinge knee designs that have been commercially available in the United States over the past twenty years has had a different design of the central rotational stem (Fig. 1). Only the Link America design has a mechanical block to distraction, to our knowledge. With all other designs, distraction and dislocation of the component is prevented only by the restraint of the remaining soft-tissue envelope. Once any distraction occurs, the degree of tilting (laxity) of the central rotational stem within the

2 449 Fig. 1 This photograph shows (left to right and top to bottom) the central rotational components manufactured by Howmedica, Intermedics, Johnson and Johnson (S-ROM), Techmedica, Wright Medical Technology/Dow Corning Wright, and Biomet. The Link America central rotational stem (not shown) projects upward from the tibial baseplate rather than protruding down into a tibial baseplate. A polyethylene shelf located behind the stem prevents component distraction. channel depends on the length, taper, and tolerance of the stem-in-cylinder design (Fig. 2). Although we have encountered clinical instability in four patients, there are few reports on the clinical stability provided by these rotating hinge knee prostheses and we are not aware of any studies comparing the stability of the different designs 1-5. Accordingly, a biomechanical study was designed to examine seven commercially available rotating hinge knee implants. The specific goals were to document, for each of the various designs, the length and degree of taper of the stem, the degree of tilting of the stem with increasing amounts of distraction, and the amount of distraction required for implant dislocation. Materials and Methods e performed a laboratory evaluation of sample rotating W hinge knee implants from each of the seven manufacturers: Howmedica (Rutherford, New Jersey), Techmedica (Camarillo, California), Wright Medical Technology (Arlington, Tennessee)/Dow Corning Wright (Arlington, Tennessee), S-ROM (Joint Medical Products/Johnson and Johnson, Stamford Connecticut), Link America (Pine Brook, New Jersey), Intermedics/Sulzer Medica (Austin, Texas), and Biomet (Warsaw, Indiana). The results of the testing of Dow Corning Wright and Wright Medical Technology components were considered together because Wright Medical Technology has continued to produce the same rotational hinge knee design begun by its predecessor company, Dow Corning Wright. Two Biomet implants were tested. The effective stem length of the Biomet design varies with the thickness of the tibial polyethylene tray (range, 12 to 22 mm in 2-mm increments). Therefore, a sample from each end of their product spectrum was chosen to determine the characteristics of their thinnest and most commonly utilized size (12 mm) and their thickest and least commonly utilized size (22 mm). All rotational stems were measured to determine their length and diameter at both their tip and their base. The diameter of the stem at its tip was subtracted from the diameter at its base. The arc tangent of the stem length divided by the difference in the stem diameters from top to bottom was used to determine the total stem taper. All measurements were performed with a handheld mechanical micrometer caliper that was accurate to ±0.1 mm. Each of the tibial components was secured to a vise with the superior surface horizontally aligned, as verified with a carpenter s level, and the rotational stem cylinder vertically

3 450 aligned. The matching rotational stem was positioned within the tibial rotational cylinder. A long thin extension pointer was attached to the superior surface of the rotational stem with adhesive. Vertical alignment in the neutral position was verified by casting a shadow on a backboard with a halogen projector lamp. Lateral tilting of the stem within the cylinder was generated by light pressure with a fingertip to displace the stem to each side. The stem angulation relative to the tibial baseplate was recorded by measuring the angle of inclination of the projected extension on the backboard with a goniometer whose horizontal axis alignment was verified with a carpenter s level. Lucite spacers of 5-mm incremental lengths were then placed within the cylinder under the tip of the stem to simulate distraction, and the translational measurements were repeated. These measurements were performed in increments of every 5 mm of distraction up to the increment that allowed component dislocation. The point of dislocation was identified as the point at which any laterally directed force caused the stem to topple over or dislocate from the channel. Measurements then were made at 1-mm increments of distraction from the 5-mm increment immediately before the increment that allowed dislocation to determine, within 1 mm, the amount of distraction required for dislocation. Ten measurements were made at each distraction increment, and the average and the standard deviation were reported as the final value for each increment. An interobserver reliability test was performed on the data recorded by two observers (P.R. and S.G.) from two runs of ten measurements each with the S-ROM knee at 15 mm of distraction, the Techmedica knee at 25 mm of distraction, and the Howmedica knee at 30 mm of distraction. The Pearson correlation coefficient between the measurements recorded by the two observers was , with p = , indicating excellent interobserver agreement. The Pearson correlation coefficient indicating reproducibility between the measurements made by each observer for the first and second runs was and , with p for each, indicating excellent agreement. A laxity curve was generated for each knee implant, with the degrees of angular laxity plotted at each level of component distraction. The slope of the curve reflects the extent to which the angular laxity increases with increases in distraction. The end point of each curve reflects the point at which dislocation occurs with the application of any force with a horizontal vector (lateral, medial, anterior, posterior, or any other horizontally directed force). An analysis of variance was performed to compare the tilting angles measured at 25 mm of distraction, and a Tukey Fig. 2 These line diagrams illustrate the minimal varus-valgus toggle or tilt associated with a long stem with a minimal taper compared with the increased varus-valgus toggle or tilt associated with a short, more tapered stem under similar conditions (25 mm) of joint distraction.

4 451 TABLE I Stem Length, Stem Taper, Minimal Distraction to Dislocate, and Tilting Laxity Manufacturer Stem Length (cm) Total Stem Taper (deg) Minimal Distraction to Dislocate (mm) Total Tilting (Laxity) at 25 mm of Distraction (deg) Howmedica Techmedica Intermedics (Sulzer Medica) Wright Medical Technology Biomet (22-mm polyethylene tray) 3.9* Biomet (12-mm polyethylene tray) 4.9* S-ROM Link America Not available *The Biomet stem itself is 5.2 cm long, but the effective length (the amount that interdigitates with the tibial cylindrical component) varies with the thickness of the tibial polyethylene tray, with less stem protruding as the tray thickness increases. The effective stem length (the amount protruding into the cylindrical channel) is listed. The implant has an anti-subluxation feature to prevent distraction, and it dislocates only if that feature fails. multiple-comparisons procedure was done to determine significant differences between the results. An analysis of variance was performed to assess the difference in the mean angles measured at 5 and 25 mm of distraction (approximately the slope of the laxity curve over this range), and a Tukey multiplecomparisons procedure was done to determine which of the differences were significant. The 5 and 25-mm values were selected because they represent the slope of the instability curves through a range at which the slopes are nearly linear. As one approaches the distraction values where dislocation occurs, the slopes change rapidly and slope-to-slope comparisons are less meaningful. Results ive designs (Howmedica, Techmedica, Intermedics, Wright FMedical Technology, and Biomet [12-mm polyethylene tray design with an effective stem length of 4.9 cm]) have a stem length of 4.8 cm. All required 39 mm of distraction to dislocate. Designs with a stem length of <4.7 cm dislocated with less distraction: the S-ROM 4.6-cm stem dislocated at 26 mm of distraction, the Link America 3.3-cm stem (without the anti-subluxation shelf engaged) dislocated at 20 mm of distraction, and the Biomet 22-mm polyethylene tray design with an effective stem length of 3.9 cm dislocated at 33 mm of distraction (Table I). Only the three designs with a stem taper of >5 (Techmedica, Intermedics, and S-ROM) had >5 of tilting laxity at 25 mm of distraction (Tables I and II). An analysis of variance comparing the laxity angles measured at 25 mm of distraction yielded a p value of <0.0001, indicating significant differences among the implants of the different manufacturers. The Tukey multiple-comparisons procedure showed that all means were significantly different from all others at alpha = 0.05, except for Intermedics (mean, 5.7 ) compared with Techmedica (mean, 5.3 ), Howmedica (mean, 2.1 ) compared with Biomet 22-mm polyethylene tray (mean, 1.9 ), and Wright Medical (mean, 0.5 ) compared with Biomet 12-mm polyethylene tray (mean, 1.1 ). An analysis of variance comparing the differences in the tilting angle measured at 25 mm of distraction and the angle measured at 5 mm of distraction (Table II) (the slopes of the laxity curves) indicated that there was a significant difference among the various designs (p < 0.001). This analysis alone cannot identify which slope comparisons differed by a significant amount. Therefore, a Tukey multiple-comparisons procedure was performed, and a significant difference was found among the means of all slopes (alpha = 0.05) with several exceptions: there was no significant difference between the mean slopes of the Intermedics (4.8 ) and Techmedica (4.2 ) implants or among the mean slopes of the Howmedica (1.3 ), Biomet 22-mm polyethylene tray design (0.8 ), Biomet 12- mm polyethylene tray design (0.5 ), and Wright Medical (0.5 ) implants. For example, the laxity curve for the S-ROM implant was significantly different from those for all other implants, and the curve for the Howmedica implant was significantly different from those for the Intermedics, Techmedica, and S-ROM implants but not from those of the Biomet 12 and 22-mm polyethylene tray design or Wright Medical implants. A ranking of the implants by the slope of their laxity curves over this distraction interval (least to greatest) is shown in Figure 3. The slope of the S-ROM curve was less than those of the other manufacturers implants, a finding that confirmed the greater amount of angular laxity or tilting of short tapered stems. The Intermedics and Techmedica components had intermediate slopes, reflecting intermediate levels of angular laxity at similar levels of distraction due to intermediate stem tapers in association with longer stem lengths. The steepest curves (least laxity) were seen with nontapered stems within cylindrical channels (Wright Medical Technology, Biomet, and Link America) and with the long, minimally tapered Howmedica stem. However, two of the straight stems dislocated

5 452 TABLE II Degrees of Tilt According to the Amount of Distraction Tilt (deg) Distraction (mm) Howmedica Intermedics/ Sulzer Medica Techmedica Wright Medical Technology Biomet (12-mm Polyethylene Tray) Biomet (22-mm Polyethylene Tray) S-ROM with much less distraction ( 33 mm) because of the short stem length. The Link America design was tested without engaging its mechanical stop to distraction (an overhanging polyethylene ledge) because, although the anti-subluxation feature should prevent dislocations, we wished to determine the instability indices that would be encountered should this polyethylene mechanical stop ever fail. T Discussion his biomechanical study demonstrated that the Howmedica Kinematic II, Techmedica, Wright Medical Technology Lacey, and Biomet (12-mm polyethylene tray design) rotating-hinge knee implants all required at least 39 mm of distraction for dislocation to occur. In our clinical practice, of the initial 270 knee reconstructions that we performed with these four rotating hinge knee designs, two (0.7%) were followed by dislocation. Conversely, the Biomet 22-mm polyethylene tray design and S-ROM rotating-hinge knee implants required 33 mm of distraction before dislocation occurred. Two of the five initial reconstructions that we performed with these two rotating-hinge knee designs in clinical practice were followed by dislocation (p = 0.003, Fisher exact test). Both of Fig. 3 Distraction-angular displacement curves for each of the knee designs. The slopes of the curves reflect the changes in angular laxity with the variations in distraction. The end points represent the maximal distraction points beyond which any more distraction causes dislocation.

6 453 the patients with a dislocation of the latter two designs had clinically apparent flexion-gap laxity prior to the dislocation. However, many of our patients treated with tumor resection had this as well. Thus, if a surgeon attempts to address the instability of a flexion gap laxity (arbitrarily and empirically assumed to be a flexion gap laxity of about 2.5 cm) or a severe flexion-extension gap imbalance with a rotating hinge knee implant, it is best to use a prosthetic design that has a long, minimally tapered central rotational stem, which requires greater distraction for dislocation to occur. An implant can become unstable if it allows excessive tilting under conditions of mild joint distraction. This clinical instability was most apparent to our patients when they attempted to lift the lower limb into bed from a seated position on the side of the bed, at which time the hips and knees were both typically flexed 90. As the patient lifted the lower limb, the flexed knee distracted and the resulting laxity became apparent to both the patient and the physician. Thus, the ideal implant for such patients should either provide stability under moderate amounts of distraction or have a safe, effective antisubluxation feature. The length and taper of the central rotational stem should be considered as new rotating hinge total knee designs are developed. The stem length and taper should be especially considered in patients with compromised soft tissues in whom a substantial flexion-gap laxity or a flexion-extension gap mismatch is anticipated. Excessive flexion laxity or flexion-extension gap imbalance should be corrected, if possible, at the time of a knee replacement or revision, as a failure to do this can lead to a poorer result, instability, or dislocation 6,7. William G. Ward, MD David Haight, MD Paul Ritchie, MD Stan Gordon, BS Department of Orthopaedic Surgery, Wake Forest University Baptist Medical Center, Medical Center Boulevard, Winston-Salem, North Carolina address for W.G. Ward: wgward@wfubmc.edu Jeffrey J. Eckardt, MD Department of Orthopaedic Surgery, University of California at Los Angeles School of Medicine, Le Conte Avenue, Los Angeles, CA The authors did not receive grants or outside funding in support of their research or preparation of this manuscript. None of the authors, nor any affiliated agencies, foundations, etc., have received any payment, grant, research support, or other financial support or payment related to this work. However, W.G. Ward received research support for other projects (during the time course of this project) from DePuy, Sulzer, and Howmedica, and he has performed paid consultant work for Howmedica. W.G. Ward was an oncology product design team consultant (designer) for Sulzer Medica. J.J. Eckardt is a consultant for Howmedica and has received grant funding from Howmedica. No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, educational institution, or other charitable or nonprofit organization with which the authors are affiliated or associated. All seven manufacturers supplied sample knee implants for the biomechanical testing. References 1. Otis JC, Lane LM. Nonmodular segmental knee replacements: design and performance. In: Enneking WF, editor. Limb salvage in musculoskeletal oncology. New York: Churchill Livingstone; p Shindell R, Neumann R, Connolly JF, Jardon OM. Evaluation of the Noiles hinged knee prosthesis. A five-year study of seventeen knees. J Bone Joint Surg Am. 1986;68: Rand JA, Chao EY, Stauffer RN. Kinematic rotating-hinge total knee arthroplasty. J Bone Joint Surg Am. 1987;69: Walker PS, Emerson R, Potter T, Scott R, Thomas WH, Turner RH. The kinematic rotating hinge: biomechanics and clinical application. Orthop Clin North Am. 1982;13: Ward WG, Eckardt JJ, Johnston-Jones KS, Eilber FR, Namba R, Dorey FJ, Mirra J, Kabo JM. Five to ten year results of custom endoprosthetic replacements for tumors of the distal femur. In: Brown KLB, editor. Complications of limb salvage: prevention management and outcome. Montreal: International Society of Limb Salvage; p Pagnano MW, Hanssen AD, Lewallen DG, Stuart MJ. Flexion instability after primary posterior cruciate retaining total knee arthroplasty. Clin Orthop. 1998;356: Whiteside LA. Selective ligament release in total knee arthroplasty of the knee in valgus. Clin Orthop. 1999;367: