The long-term success of total knee arthroplasty is dependent, in part, on proper restoration of the mechanical axis and soft-tissue balancing. While most arthritic knees have some degree of associated deformity and secondary imbalance of the soft-tissue envelope, both conditions are usually periarticular and easily addressed with adjustment of the resurfacing osteotomies and soft-tissue releases on the concavity of the deformity. However, in the presence of extra-articular deformity of 10 degrees or more in the coronal plane or 20 degrees or more in the sagittal plane, complex imbalance of the collateral ligaments may result when the deformity is addressed solely with modified intra-articular bone resection and soft-tissue releases5,10. Such deformities may be found after fracture malunion or periarticular osteotomy or they may be associated with metabolic diseases such as osteomalacia, or rickets, and Paget disease.
To limit such ligament imbalance and to reduce the need for constrained implants, these patients can be treated with simultaneous or staged osteotomy and resurfacing knee arthroplasty. We report our experience with the use of simultaneous corrective osteotomy and total knee arthroplasty for the treatment of eleven patients with osteoarthritis of the knee and ipsilateral extra-articular angular deformity of the femur.
*One or more of the authors has received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article. In addition, benefits have been or will be directed to a research fund, foundation, educational institution, or other nonprofit organization with which one or more of the authors is associated. No funds were received in support of this study.
†Department of Orthopaedic Surgery, Hospital of the University of Pennsylvania, 3400 Spruce Street, 2 Silverstein, Philadelphia, Pennsylvania 19104. E-mail address for J. H. Lonner: jlonner@mail.med. upenn.edu.
‡Department of Orthopaedic Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 02114.
Between 1990 and 1996, we performed simultaneous femoral osteotomy and total knee arthroplasty in eleven patients with extra-articular femoral deformity associated with advanced bicompartmental or tricompartmental osteoarthritis of the ipsilateral knee. There were five primarily uniplanar deformities (four varus deformities and one antecurvatum deformity), five biplanar (varus and antecurvatum) deformities, and one triplanar (varus, antecurvatum, and internal rotation) deformity (Table I). The femoral deformity resulted from fracture malunion in ten patients and hypophosphatemic rickets in one. One patient had a fibrous union of a fracture of the medial femoral condyle in addition to a supracondylar malunion. For all patients, use of a template on preoperative radiographs led to a concern that complex ligamentous imbalance would result if the deformity was addressed by standard distal femoral resection perpendicular to the mechanical axis of the femur; therefore, corrective osteotomy was performed to restore femoral alignment prior to, and during the same operative session as, the total knee arthroplasty (Figs. 1-A, 1-B, 1-C and 1-D).
There were five women and six men, with a mean age of sixty-three years (range, forty to seventy-four years) at the time of the index operation. All of the operations were unilateral. For the patients who had a fracture malunion, the interval between the treatment of the fracture and the index operation averaged forty-two months (range, eight months to eleven years). One of the original fractures was open, and none were infected. The patient with rickets had had multiple femoral and tibial osteotomies, followed by healing, as a child.
Clinical and functional evaluations were carried out by comparing preoperative and postoperative objective scores according to the 200-point system of the Knee Society3. This scale includes a 100-point knee score for the evaluation of such categories as pain, range of motion, and stability, with deductions for flexion contracture, extension lag, instability, and malalignment. It also assigns a maximum of 100 points for functional capacity, which includes such parameters as the ability to walk on level ground and on stairs as well as the necessity for assistive devices. Postoperative knee scores were based on the most recent follow-up data. The angular deformity and the mechanical axis of the femur were determined preoperatively on weight-bearing full-length radiographs that incorporated the ipsilateral hip and ankle. A theoretical intra-articular resection of femoral bone was determined on the same radiographs, with a plan for a 90-degree osteotomy relative to the mechanical axis. A corrective osteotomy was considered necessary if, on the full-length weight-bearing radiograph, it appeared that distal femoral bone resection at a right angle to the mechanical axis would compromise the integrity of the insertion of either the medial or the lateral collateral ligament.
The postoperative anatomical and mechanical axes were also determined on full-length weight-bearing radiographs incorporating the hip and ankle. The alignments of the femoral and tibial components were measured on the initial postoperative anteroposterior and lateral radiographs. In addition, serial radiographs were reviewed to determine the presence of interface radiolucencies and to document any progression.
Seven procedures were performed through an extensile anterolateral subvastus approach to the knee, with osteotomy of the tibial tubercle, to facilitate exposure, hardware removal, and subsequent femoral osteotomy and fixation. Hardware was removed from three patients; the other patients either had had hardware removed previously or had not had hardware.
The corrective osteotomy was performed in the distal femoral metaphysis, distal to the deformity, in three patients, and it was performed through the fracture malunion in eight. In limbs with multiplanar deformities, corresponding multiplanar osteotomies were performed. The site of the femoral osteotomy was secured with blade-plate fixation in seven patients, with a spanning noncemented long-stemmed femoral component in two, and with a retrograde femoral nail in two. A standard medial parapatellar arthrotomy was used in the four patients in whom a retrograde nail or a long-stemmed femoral component was employed to secure the osteotomy site. In the two patients who were treated with a retrograde nail, separate small lateral incisions were made proximally, at the site of the deformity, to allow corrective osteotomy at the apex of the deformity. When the anterolateral subvastus approach is used, exposure can be optimized by displacing the osteotomized tibial tubercle medially, leaving it attached to the medial sleeve of soft tissue. All of the sites of the tibial tubercle osteotomies were fixed with lag screws. Scar tissue was excised if necessary.
The corrective osteotomy was performed, prior to the total knee arthroplasty, with use of fluoroscopic assistance to ensure its proper location and amount of correction. An extramedullary femoral alignment system (Johnson and Johnson, Warsaw, Indiana, or Osteonics Insight, Osteonics, Allendale, New Jersey) was used in eight patients. The extramedullary alignment system allowed stable fixation of the femoral osteotomy site before the total knee arthroplasty was performed. The remaining patients were treated with an intramedullary alignment technique in which provisional fixation of the femoral osteotomy site was provided by small wires and an intramedullary alignment rod was passed across the osteotomy site into the isthmus of the femur.
A Press-Fit Condylar implant (Johnson and Johnson) was used in eight patients; an Osteonics implant, in two; and a Legacy implant (Zimmer, Warsaw, Indiana), in one. Two patellae were not resurfaced because they exhibited little degeneration and they articulated congruently with the femoral trochlea. All of the prostheses were posterior-cruciate-substituting designs, and all were cemented into place.
The duration of follow-up averaged forty-six months (range, twenty-six to eighty-eight months). The mean postoperative knee score was 87 points (range, 61 to 98 points), representing a mean improvement of 77 points compared with the preoperative score. The flexion arc increased a mean of 33 degrees, to a mean of 89 degrees (range, 65 to 115 degrees) at the time of follow-up, and the flexion contractures were reduced to a mean of 2 degrees. The postoperative function score averaged 81 points (range, 30 to 100 points), representing a mean improvement of 59 points (Table II).
The postoperative limb alignment, based on measurement of the mechanical axis in the coronal plane on a weight-bearing full-length anteroposterior radiograph of the lower extremity, was restored to within 2 degrees of normal in each patient. Two knees had nonprogressive radiolucent lines of less than two millimeters in width in tibial zones2 1 and 4, and two patients had nonprogressive radiolucent lines of less than two millimeters in width in femoral zone2 1.
Ten of the eleven femoral osteotomies and all seven of the osteotomies of the tibial tubercle were followed by healing within sixteen weeks. One femoral osteotomy site, in a patient treated with a press-fit long-stemmed component (Case 6), did not heal despite bone-grafting. Two patients required manipulation under anesthesia to improve motion within six weeks after the surgery; one of these knees (Case 8) was manipulated into extension and was treated with a dynamic extension splint to maintain reasonable extension. One patient had an acute pulmonary embolism seven days postoperatively and was treated with intravenous heparin therapy. One patient required removal of the tibial tubercle screws, which were prominent and causing tenderness, at ten months after the surgery. There were no postoperative infections.
A successful outcome after total knee arthroplasty requires, among other things, restoration of the mechanical axis, balancing of periarticular soft tissues, and accurate orientation of the individual prosthetic components4. An extra-articular, extraligamentous deformity may not be amenable to routine techniques of intra-articular bone resection and ligamentous balancing. In such situations, the treatment alternatives include adjustment of resurfacing cuts with ligamentous release or advancement, unicondylar wedges, or corrective extra-articular osteotomy in conjunction with, or staged with, total knee arthroplasty.
Mann et al. recently reported the results of the preferred technique of Insall after it had been used to treat a series of deformities, eleven of which were femoral, that were associated with arthrosis of the knee5. Coronal plane deformities averaged 14 degrees (range, 5 to 22 degrees), and sagittal plane deformities averaged 12 degrees (range, 0 to 38 degrees). The authors stated that such deformities may be adequately addressed with modified intra-articular bone resection and ligament balancing, obviating the need for additional osteotomy. The authors reported marked improvement in scores, according to the system of The Hospital for Special Surgery, after a minimum of two years of follow-up; however, there were radiolucent lines in three knees, although none were considered progressive and only one knee required a revision because of failure of a metal-backed patellar component. This technique may be reasonable if the function and balance of the collateral ligaments are not compromised. Also, while it appeared to be effective at two years, this technique may alter joint-line position and the femoral component may not always rest flush on the cut bone surfaces; rather, oblique gaps left between the prosthesis and the bone, on the concavity of the deformity, may need to be filled with polymethylmethacrylate, structural bone graft, or wedges5. The theoretical concerns in such cases are that structural support may be insufficient and oblique shear forces may make the bone-cement interfaces vulnerable to early demarcation and perhaps lead to long-term failure1.
When a patient has a femoral diaphyseal or metaphyseal deformity of 10 degrees or more in the coronal plane (that is, a varus or valgus deformity) or 20 degrees or more in the sagittal plane, cutting the distal part of the femur with proper orientation to the mechanical axis will result in a resection that is oblique to the epicondylar axis in extension, and the consequence will be ligamentous imbalance in extension6. The posterior aspects of the femoral condyles, however, maintain their appropriate relationships with the tibia in flexion. This resultant scenario is difficult to manage. Ligament advancement, to address an asymmetrical soft-tissue envelope, has yielded mixed and unpredictable results, in part because patients do not have a comparable imbalance in both extension and flexion8.
It has been our experience, and that of others6,10, that if the origin of a collateral ligament is compromised by bone resection, or if the resection creates large asymmetrical gaps that make soft-tissue balancing difficult, corrective osteotomy to normalize long-bone anatomy prior to total knee arthroplasty is prudent. Another concern, pointed out by Wolff et al., is that, in the presence of notable femoral deformity, restoration of the normal mechanical axis with intra-articular resection may normalize orientation of the knee but hip adduction or abduction is still necessary to keep the knee and ankle parallel to the ground in the stance phase of gait10. This may result in localized areas of stress concentration in the hip and ultimately contribute to accelerated coxarthrosis.
The degree of corrective osteotomy depends not only on the extent of the deformity but also on the location of the osteotomy in relation to the deformity. In order to restore alignment of the knee, an osteotomy performed between the apex of the deformity and the knee may be smaller than the deformity because intra-articular resurfacing cuts may be modified and because of the ability of the hip and ankle to compensate10. Using trigonometric formulas, Wolff et al. calculated the angular malalignment of the knee in the presence of extra-articular deformity at a variety of distances from the joint10. Functional ligamentous lengthening was then determined for each scenario. As the authors discussed, the surgeon can use this information in preoperative planning to determine whether significant instability would result without extra-articular corrective osteotomy.
Performing corrective extra-articular osteotomy and knee arthroplasty during the same operative session is technically difficult but effective. A laterally based incision is frequently necessary, particularly when internal fixation has been used for initial fracture treatment or if the surgeon prefers to fix the osteotomy site with a plate rather than with an intramedullary nail. Through this approach, hardware can be removed, osteotomy can be performed, and stable internal fixation can be achieved. It has been our experience that osteotomy of the tibial tubercle in this setting facilitates exposure for both the extra-articular osteotomy and the arthroplasty. The technique described by Whiteside and Ohl is effective, with a high rate of union9. It may not be possible to fix the femoral osteotomy site with a modular press-fit long-stemmed condylar resurfacing implant when acute angular malalignment at the osteotomy site has been created iatrogenically after restoration of the mechanical axis6,7. In our experience, a precontoured blade-plate or a locked intramedullary nail has provided stable fixation of the osteotomy site, allowing intensive postoperative physiotherapy and uniform rates of union. The femoral osteotomy site was secured with a press-fit long femoral stem in two patients in the present series, and one of the sites did not heal. Currently, we prefer to secure the osteotomy site with either a locked intramedullary nail or a blade-plate and then to perform the distal femoral resection with an extramedullary femoral alignment system.
Consistent with the reports by Mann et al.5 and by Rand and Franco6, we found that the procedure was successful in improving function, reducing pain, and maintaining stability in our relatively small series of complex cases. The flexion arc may be improved considerably with mobilization of the quadriceps, adequate periarticular soft-tissue releases, restoration of mechanical and anatomical alignment, and stable internal fixation of the sites of both the femoral osteotomy and the osteotomy of the tibial tubercle, which allows early and intensive physiotherapy.
In the setting of supracondylar metaphyseal or diaphyseal malunion or metabolic deformity, operative options should be based on a preoperative determination of the relationship between the mechanical axis and the planned intra-articular resection on full-length radiographs. Simultaneous corrective femoral osteotomy, to correct severe mechanical malalignment prior to total knee arthroplasty, is advisable to ensure maintenance of structural support of the prosthesis and to avoid complex ligamentous imbalance. Ultimately, this approach may enhance implant function and survival in these complex cases.