The Classic: Modular Total Knee-Replacement Arthroplasty. A Review of Eighty-nine Patients

The basic technique was described by Marmor. Certain technical points warrant emphasis. In nearly all of our patients we used a single, relatively straight medial parapatellar incision carried directly down to the joint without developing skin flaps. Proximally, the incision separated the interval between the rectus femoris and the vastus medialis muscles. Distally, 20 per cent of the insertion of the patellar tendon into the tibial tubercle was elevated sharply. This permitted easy lateral displacement of the patella and wide exposure of the knee joint. With the knee hyperflexed, a wedge of the popliteal surface of each femoral condyle five millimeters thick can be removed to give wide exposure to the entire posterior aspect of the tibial plateaus. The Hall II Pneumatic Instrument, fitted with a contra-angle handpiece and a large, sharp cutting burr, facilitates the preparation of the tibial plateau. The anterior portion of the lateral femoral implant is countersunk into the femoral condyle to prevent possible patellar impingement during flexion. Anchoring holes three centimeters deep are made in the proximal surface of the tibia and are directed downward and anteriorly. After all prosthetic components have been implanted and the methylmethacrylate has set, the tourniquet is released and hemostasis is obtained. The patella should be replaced in the trochlear groove and free motion verified during passive flexion and extension of the knee. In patients with a preoperative valgus deformity, it often is necessary to release the lateral patellar retinaculum and imbricate the medial retinaculum to restore patellofemoral alignment. The capsule of the knee is closed with the knee in flexion, while the remaining layers are closed with the knee fully extended. Closed Hemovac drainage is maintained for forty-eight hours. At that point the postoperative Robert Jones dressing and the drains are removed, and a gentle range of motion is begun. The patients are allowed to bear weight on the third postoperative day and walk with either a walker or crutches as tolerated. Subsequently a cane is used for a minimum of four weeks. Quadriceps-setting isometric exercises are begun preoperatively and continued through the entire postoperative period. All patients are started on intravenous cephalosporin two hours prior to surgery, and this is maintained intravenously for forty-eight hours after surgery. In the initial portion of the study, no specific antithromboembolic medications were used. For the past year, however, all patients were started on aspirin, 600 milligrams orally twice a day, beginning the day before surgery and continuing until the day of discharge from the hospital.

Modular components fabricated by the Richards Manufacturing Company were implanted in the first thirty patients. All subsequent procedures were performed using the Zimmer compartmental type of modular knee prosthesis.

During the study period, eleven patients underwent unicompartmental tibiofemoral replacement for “apparently” localized degenerative arthritis involving either the medial or the lateral compartment. Since the biomechanical implications and, indeed, the rationale for unicompartmental replacement differ from those for bicompartmental tibiofemoral replacement, these eleven patients were not included in the series. We also eliminated from the series two osteoarthritic and five rheumatoid patients who underwent bilateral modular knee replacement, because of the difficulty of comparing functional results in patients with bilateral disease with results in those who have undergone unilateral total knee replacement. At times, functional results are difficult to assess in the presence of disabling problems in other weight-bearing joints of the same or the contralateral extremity. This problem existed to some extent in the osteoarthritic patients, but was more manifest in the rheumatoid patients. In the latter group, functional ability was correlated as well as possible with the involved knee only. Relief of pain was relatively simple to correlate with the affected joint; however, the need for support and the ability to walk distances were greatly influenced by disease of adjacent joints.

The series under discussion comprises eighty-nine patients, fifty-eight with osteoarthritis and thirty-one with rheumatoid arthritis. All rheumatoid patients were seropositive, and in those with a proliferative invasive synovium an anterior synovectomy was performed at the time of joint replacement. All patients who had been taking systemic corticosteroids at any time during the twelve months prior to surgery were begun on steroid medication preoperatively and maintained on it during the operation and for seventy-two hours thereafter. Subsequent steroid requirements as well as general medical management of the patients were variable and no standard regimen was followed.

There were thirty-one male and twenty-seven female osteoarthritic patients, 79 per cent of whom were over the age of sixty. Only two of the thirty-one rheumatoid patients were men, and the greatest number of patients (seventeen) were between fifty and sixty years old.

In this series, the major purpose of total knee replacement was to relieve pain. Eighty-six per cent of the osteoarthritic patients and 78 per cent of the rheumatoid patients had pain which was rated as moderate, severe, or disabling. In each group there was one patient who did not complain of any pain but who was essentially bedridden, with severe limitation of joint motion. The purpose of surgery in these two patients was to regain motion for adequate nursing care and chair-to-bed transfer. There were seven rheumatoid patients (23 per cent) whose major complaint was recurrent swelling associated with pain and limitation of motion. At surgery, large areas of cartilage erosion were visualized, and despite the presence of minimum roentgenographic changes I thought that synovectomy alone would he inadequate. Therefore. bicompartmental replacement was also performed in these knees.

The patients were evaluated and rated preoperatively and at three, six, twelve, eighteen, and twenty-four months postoperatively. It was noted early in the study that the great majority of patients achieved a plateau in terms of functional and numerical rating by the third postoperative month, and that further improvement beyond this point was minimum. No deterioration of results was encountered during follow-up.

All patients were operated on by myself and my two associates (Dr. M. A. Gruber and Dr. A. J. Zimmerman) or by senior resident orthopaedic surgeons under our direction. I personally performed all evaluations of the patients. Questions regarding pain and function were asked in a standardized manner. The evaluation of gems varum or valgum was made on the basis of weight-bearing anteroposterior roentgenograms made with the knee in maximum extension. The degree of flexion contracture or recurvatum, when present, was measured on lateral roentgenograms with the patient recumbent and the leg suspended by the ankle so that maximum knee extension was achieved.

A rating system similar to that described by Ranawat and Shine was used. Out of a total of 100 points, thirty were allocated to pain, thirty-five to function, and ten to muscle strength. Static passive range of motion (five points), flexion deformity (ten points), instability (ten points), and extensor lag (up to five negative points) were less important in the rating, although these parameters obviously influenced both the patient’s function and the initial choice as to whether a resurfacing implant would be the best prosthesis for the reconstruction.

After evaluation, we did not accept for operation any patient in whom varus or valgus instability exceeded 20 degrees, or in whom the sum of the degrees of instability and flexion contracture or recurvatum exceeded 40 degrees. Recurvatum of over 15 degrees was thought to indicate absence of the posterior cruciate ligament and was a contraindication for resurfacing arthroplasty. A history of prior intro-articular infection also was an absolute contraindication.

No patient had clinical signs of thrombophlebitis [5] or of pulmonary embolus, although some patients were maintained on salicylates and others were not. No patient who received salicylates in our series had gastrointestinal bleeding. There were no cases of fat embolism. One patient had a deep infection (Staphylococcus aureus) which required debridement and closed suction-irrigation. This patient had rheumatoid arthritis and had been taking steroids for fourteen years prior to surgery. At final follow-up she had no drainage, although she continued to have disabling pain with walking. Roentgenograms revealed three millimeters of settling. Three other patients had superficial infections: culture of fluid from one revealed Pseudomonas aeruginosa: from one, Staphylococcus epidermidis: and one culture was sterile. The drainage in these patients ceased spontaneously and did not recur. All three of these patients also had rheumatoid arthritis and had been taking steroid medication for a long time. In four other rheumatoid patients, erythema developed around the operative wound during the first postoperative week; these lesions subsided without drainage. None of these patients was febrile, nor did they manifest a leukocytosis. One rheumatoid patient had slight skin necrosis which healed spontaneously. There was one case of capsule rupture that occurred approximately two months after surgery. This patient was osteoarthritic and had previously undergone two unsuccessful tibial osteotomies for genu valgum. Two months after the arthroplasty, she had massive swelling of the knee and an obvious rupture of the sutured capsule. When the knee was re-explored, an exuberant amount of synovial hypertrophy was found covering the articular surfaces and the implants. She underwent synovectomy and reapproximation of the capsule. Microscopic examination of the synovium revealed a nonspecific synovitis without lymphoid nodules. There was no evidence of foreign-body reaction to the cement or of particulate matter in the tissue. The synovial fluid and the synovium were cultured for aerobic, anaerobic, fungal, and mycobacterial organisms, and all cultures were sterile. The exact cause of the complications in this patient could not be determined. She died of renal disease before the eighteen-month evaluation, and therefore is not included in Table 1. Had she survived, she probably would have had a poor result, considering both function and relief of pain.

No cases of frank loosening of the prosthesis occurred in this series. One of our patients with a unicompartmental replacement did have complete loosening of the tibial unit at the cement-bone interface.

Marked dysesthesia developed on the sole of the foot in one osteoarthritic patient immediately postoperatively. This decreased somewhat but did not completely disappear. Despite the fact that this patient had no pain in the knee and had an arc of motion of 110 degrees, the result was classified as poor because of the disabling foot pain. There were no other neurological problems in the series.

Although the modular arthroplasty is completely unconstrained (there are no tracks in the polyethylene of the tibial components), and although some latitude in orientation of the implants is permissible, stresses applied to the implant by improper placement may lead to early failure. One patient in the series had a valgus deformity of the knee preoperatively, associated with a moderate flexion contracture. In this patient the operative exposure was extremely difficult, despite the use of a second incision. Postoperative roentgenograms revealed that both tibial implants were severely angulated in the coronal plane. Although there was initial relief of pain, there was recurrence of pain and limitation of motion in the knee by the sixth postoperative week, and the prosthesis showed settling of up to three centimeters. The patient had no systemic or local signs of infection, but by the third month she had become so disabled that revision of the arthroplasty was required. We did a bone resection and implanted a total condylar knee prosthesis, and after nine months she was completely free of pain.