Secondary Logo

Journal Logo

Total Knee Arthroplasty After High Tibial Osteotomy

Meding, John, B.; Keating, E., Michael; Ritter, Merrill, A.; Faris, Philip, M.

Clinical Orthopaedics and Related Research: June 2000 - Volume 375 - Issue - p 175-184
Section II: Original Articles: Knee
Free
SDC

Between 1980 and 1995, 95 consecutive total knee replacements were performed at an average of 10 years 4 months after high tibial osteotomy. The average age of the 82 patients was 66 years, with a preoperative diagnosis of osteoarthritis in 94 knees. One patient died 6 months after surgery. The followup of the remaining 81 patients (94 knees) averaged 8.6 years (range, 2-17 years). Knee Society knee score at final followup improved to an average of 87.6 points from a preoperative average of 38.1 points. No pain was present in 86.2% of knees, and 12.8% of knees had only mild or occasional pain. Tibial radiolucencies were identified in 12 (12.8%) knees at final followup, and in only four knees were radiolucent lines found about the lateral zones. Only one tibial component required revision 3 years after surgery. Although no preoperative factor was identified that predisposed to an inferior knee score, function score, or pain score, the severity of the preoperative flexion contracture and the number of previous surgeries did relate to diminished postoperative motion. However, an increased number of patellar radiolucencies were seen in the knees in which the lateral joint line was raised (referenced from the fibular head) a greater degree. The clinical results of total knee replacement after high tibial osteotomy appeared similar to those of primary total knee replacement. The previous high tibial osteotomy had no adverse effect on the eventual results of a cemented posterior cruciate retaining total knee replacement.

From the Center for Hip and Knee Surgery, Orthopaedics Indianapolis, St Francis Hospital Mooresville, Mooresville, IN.

Reprint requests to John B. Meding, MD, The Center for Hip and Knee Surgery, 1199 Hadley Road, Mooresville, IN 46158.

Received: April 23, 1999.

Revised: November 22, 1999.

Accepted: January 5, 2000.

Because total knee replacement generally is contraindicated in younger and more active patients, those with unicompartmental osteoarthritis of the knee may be considered candidates for high tibial osteotomy. The short-term results of high tibial osteotomy have been very successful,1,4,11 even to the point where the need for total knee replacement is delayed if not eliminated.21 However, permanent pain relief with high tibial osteotomy is unlikely,5,7,9 with overall survival approximately only 6 years.17 Thus, high tibial osteotomy generally buys time, but many of these patients eventually will need a total knee replacement.9

Several authors have studied the result of total knee replacement after high tibial osteotomy with conflicting results.3,5,12,15,16,19,20,21 Several reports have described similar results in patients undergoing total knee replacement with or without a previous high tibial osteotomy.2,3,16,19,20 However, several other reports have described inferior results with total knee replacement in knees that have undergone a previous high tibial osteotomy.12,15,21 Unfortunately, many of these reviews (even those patient matched studies) have evaluated rather heterogeneous patient populations, grouping together cemented and uncemented total knee replacement and posterior cruciate retaining and posterior stabilized total knee replacement.2,12,15,16,19,20 There have been some reviews in which posterior cruciate substituting and posterior cruciate retaining groups were not mixed.3,21 In addition to the conflicting results, the diversity of these patient populations does little to clarify these results. Thus, although high tibial osteotomy does not preclude total knee replacement, the effect of a previous high tibial osteotomy on the eventual outcome of total knee replacement remains a matter of controversy.

The purpose of this study was to evaluate total knee replacement after high tibial osteotomy in a more homogenous patient population and to identify any potential subset of these patients that may portend to an inferior result. The current authors hypothesized that high tibial osteotomy should have no adverse effect on the eventual results of a cemented posterior cruciate retaining total knee replacement.

Back to Top | Article Outline

MATERIALS AND METHODS

Five thousand fifty-three primary total knee replacements were performed in 3506 patients at the authors' institution between 1980 and 1995. From this patient base 95 consecutive total knee replacements in 82 patients were performed at an average of 10 years 4 months after high tibial osteotomy (range, 24-289 months; standard deviation, 66.2 months). The average age of the patients was 66 years. There were 61 men and 21 women. Thirteen patients underwent bilateral total knee replacement after bilateral high tibial osteotomy, all of which were performed simultaneously. Forty patients underwent bilateral total knee replacement after unilateral high tibial osteotomy, with 32 of these performed simultaneously. The diagnosis was primarily osteoarthritis (94 knees). Table 1 shows the remaining demographic data.

TABLE 1

TABLE 1

The previous high tibial osteotomy was a dome type in 10 (10.5%) cases and a closing wedge type in 85 (89.5%) cases. In 56 (58.9%) knees, internal fixation was used at the time of the osteotomy (staples, 49 cases; plate and screws, six cases; screws and wire, one case). Thus, in 39 (41.1%) knees, no internal fixation was used at the time of high tibial osteotomy. Only one known complication was associated with the previous high tibial osteotomy (nonunion). There were no known cases of avascular necrosis of the tibia, infection, or peroneal nerve palsy. Twenty-five knees had previous surgery in addition to the high tibial osteotomy. Nineteen knees underwent medial meniscectomy (six open, four arthroscopic). One knee had repeat high tibial osteotomy. Six knees underwent at least two other surgeries (five open medial meniscectomy and subsequent hardware removal after high tibial osteotomy, and one knee had open medial meniscectomy and arthroscopic meniscectomy).

The Knee Society clinical8 and radiographic6 evaluation systems were used to evaluate knees before surgery and at each followup. Knee alignment, the Insall and Salvati ratio,10 and the location of the lateral joint line (measured from the fibular head) were recorded on the preoperative and initial postoperative radiographs. Table 2 shows these preoperative data. Knee scores averaged 38.1 with an overall alignment of approximately 4° valgus. Although the average Insall and Salvati ratio was 1.1, 18 (18.9%) knees had a ratio of less than 0.8 (patella infera). The average distance from the lateral joint line to the osteotomy site was 23.4 mm (range, 10-65 mm; standard deviation, 9.4 mm). In addition, the preoperative position of the lateral joint line ranged from 5 mm distal to the fibular head to 31 mm proximal to the fibular head.

TABLE 2

TABLE 2

All knees were approached through the standard medial parapatellar incision. In 93 (97.9%) knees, an anatomic graduated component prosthesis (AGC, Biomet, Warsaw, IN) was used. In two (2.1%) knees, a Posterior Cruciate Condylar prosthesis (Howmedica, East Rutherford, NJ) was used. Thus, a cruciate retaining prosthesis was used in every case. In addition, all tibial components, all patellar components, and 87 (92.6%) femoral components were cemented. In nine (9.5%) knees, an all polyethylene AGC tibial component was used, and in six (6.3%) knees the patella was metal backed. Although significant soft tissue scarring between the osteotomy site and the surrounding soft tissues, including the medial collateral ligament and patellar ligament, was observed, in no case was a lateral release required before exposure. Tibial tubercle osteotomy or quadriceps snip were not required. Hardware was removed at the time of total knee replacement in 17 knees and before total knee replacement in 16 knees. In 23 knees, hardware removal was not required before or during total knee replacement. An attempt was made to resect a minimal amount of lateral tibial bone in each case. In addition, intraoperative ligament releases were recorded at the time of surgery.

In all cases ambulation was started on postoperative Day 1 with range of motion (ROM) initiated on postoperative Day 2. The initial postoperative radiograph was obtained 8 weeks after surgery with patients observed again at 6 months and 1, 3, 5, 7, 10, 12, 15, and 17 years. Clinical and radiographic evaluations were performed at each followup. All radiographs were evaluated for tibial, femoral, and patellar interface radiolucencies. One patient died 6 months after surgery of an acute myocardial infarction. The followup in the remaining 81 patients (94 knees) averaged 8.6 years (range, 2-17 years). Two patients died during this followup interval, and two other patients were lost to followup after their 3- and 5-year evaluations, respectively.

Statistical analysis was performed using the Statistical Analysis System (SAS, Cary, NC).

Back to Top | Article Outline

RESULTS

The average knee score at final followup (8.6 years, 94 knees) improved to 87.6 points, from an average of 38.1 points before surgery. No pain was present in 86.2% of knees, and 12.8% of knees had only mild or occasional pain. Similarly, function scores improved from an average of 45.1 points before surgery to 83.7 points after surgery. In addition, ROM improved to an average of 113.6° flexion to negative 1° extension (contracture). The average alignment on the initial postoperative radiograph was 6.6° valgus (range, 3° varus to 17° valgus). Anteroposterior femoral alignment averaged 8.2° valgus at final followup (range, 0°-15°), with an average tibial alignment on the anteroposterior radiograph of 2.1° varus (range, 8° varus to 5° valgus). The average Insall and Salvati ratio (1.11; range, 0.67-1.49) was unchanged from the preoperative average.

Intraoperatively, lateral retinacular release was required in only 43 (45.3%) knees. In no case was a lateral retinacular release required for knee exposure. Table 3 shows the number of medial, posterior, and lateral ligamentous releases required at the time of surgery. Only 90% of knees required release of the deep medial collateral ligament, and almost 90% of knees required no release of the iliotibial band, lateral capsule, or popliteus tendon. Only five (5.3%) knees required release of the iliotibial band, and another five knees required release of the lateral joint capsule with or without release of the popliteus tendon.

TABLE 3

TABLE 3

An average of 4 mm of lateral tibial bone was resected at the time of surgery (Table 4). Compared with the preoperative radiographs, the distance between the lateral joint line and the proximal fibula increased an average of 6.6 mm (range, 0-30 mm). In no case was less than a 10-mm thick tibial component used. Thus, with minimal lateral tibial bone resection and the use of a relatively thicker tibial component, the preosteotomy lateral joint line position (anatomic) was reestablished (Fig 1).

TABLE 4

TABLE 4

Fig 1A

Fig 1A

Table 5 shows the distribution of bone-cement interface radiolucencies. No cement-prosthetic interface radiolucencies were observed. Tibial radiolucencies were identified in 12 (12.8%) knees at final followup. Only four knees had radiolucent lines found about the lateral zones (Zones 3 and 4). Only one revision of the tibial component was required. This was an all polyethylene AGC tibial component that was revised 3 years after surgery. At the time of the current review, this patient is pain free 5 years after revision surgery. Other reoperations included one case of fibrous tissue removal in the suprapatellar pouch, one debridement of a superficial wound infection, three cases of revision of loose metal backed patellar components, and one deep infection requiring excisional arthroplasty 1 month after surgery (Staphylococcus aureus). Five (5.3%) knees required postoperative manipulation. All but one of these knees achieved motion within 5° of full extension and more than 90° flexion. One knee retained an 11° flexion contracture and only 75° flexion. Other complications included five patella fractures, all of which were treated nonoperatively. One patellar tendon avulsion occurred 3 months after surgery and resulted in a temporary 30° extensor lag. Treatment was nonoperative using a hinged knee brace initially locked in full extension for 3 weeks, then allowing progressive knee flexion. Four other cases of superficial wound infection were treated with oral antibiotics and all healed. Only one case of knee instability was observed; this knee had 10 mm of medial instability and was treated nonoperatively.

TABLE 5

TABLE 5

Table 6 compares the most recent postoperative clinical scores, ROM, and interface radiolucencies, with potential factors that have been associated with a poor outcome of total knee replacement after high tibial osteotomy, including preoperative flexion contraction, the number of previous surgeries before total knee replacement, retained hardware, patellar infera, and preoperative alignment.14,15 In addition, the degree of raising the lateral joint line referenced from the fibular head also was evaluated as a potential factor affecting total knee replacement after high tibial osteotomy. Postoperative pain, knee score, and function score were independent of the degree of flexion contracture, the number of previous surgeries, the presence or absence of retained hardware at the time of total knee replacement, patella infera, preoperative alignment, or the degree to which the lateral joint line was raised. Diminished motion after surgery was more common in the knees with greater degrees of flexion contracture and a greater number of previous surgeries. However, ROM was unrelated to the other factors. Four of the six knees that underwent at least two other surgeries before total knee replacement had femoral radiolucencies (p = 0.043). In addition, the Insall and Salvati ratio in each of these four knees was less than 0.8 (p = 0.043). The presence of retained hardware after total knee replacement (23 knees) related to an increased number of tibial radiolucencies (p = 0.008). Finally, the greater degree to which the lateral joint line was raised (increased distance from fibular head) related to an increased number of patellar radiolucencies (p = 0.033).

TABLE 6

TABLE 6

The number of soft tissue releases required at the time of surgery also was compared with the preoperative alignment and the Insall and Salvati ratio. Preoperative alignment had no effect on the number of lateral retinacular releases (p = 0.753), medial releases (p = 0.637), lateral knee capsular releases (p = 0.587), or posterior capsular releases (p = 0.778). Although patella infera was not related to the number of lateral patellar retinacular releases (p = 0.771), lateral knee capsular releases (p = 0.352), or posterior capsular releases (p = 0.614), a diminished Insall and Salvati ratio (patella infera) was associated with an increased number of medial ligamentous releases (p = 0.028). Nine knees requiring at least release of the superficial medial collateral ligament with or without semimembranosus release had a preoperative Insall and Salvati ratio of less than 0.8.

Back to Top | Article Outline

DISCUSSION

The issue of whether a previous high tibial osteotomy has any potential adverse effect on a subsequent total knee replacement remains a matter of controversy. Some authors report inferior results in patients who have undergone previous high tibial osteotomy,12,15,21 and other authors report similar results after total knee replacement with or without a previous high tibial osteotomy.2,3,16,20,19 Two reports have reviewed total knee replacement after high tibial osteotomy without patient matched or comparison studies. Windsor et al21 reviewed patients with 45 cemented total condylar total knee replacements with an average followup of 55 months. Although they had good or excellent results, in 80% of the patients in the study, the authors concluded that total knee replacement after high tibial osteotomy had results similar to those after revision surgery, with inferior results compared with primary total knee replacement. However, Staeheli et al19 reviewed patients with 35 cemented total condylar posterior stabilized and posterior cruciate retaining total knee replacements with an average followup of 44 months and found good to excellent results in 89% of the knees and thought the previous high tibial osteotomy had no deleterious effect on a subsequent total knee replacement.

Four other reports have reviewed total knee replacement after high tibial osteotomy comparing patients with primary total knee replacement matched for age, gender, prosthesis, and followup with or without matching weight, Charnley class, and deformity. Katz et al12 reviewed patients with 21 cemented and uncemented posterior cruciate retaining knees with an average followup of 35 months and found more poor results (81% good to excellent versus 100% good to excellent) in the high tibial osteotomy group than in the patient matched group. In a followup from the same institution, Mont et al15 found greater fair and poor results in the high tibial osteotomy group reviewing patients with 73 cemented and uncemented posterior cruciate retaining knees with an average followup of 73 months. Only 64% good to excellent results were seen in the high tibial osteotomy group, compared with 89% good to excellent results in the primary total knee replacement group. Conversely, Amendola et al2 reviewed patients with 42 cemented and uncemented posterior cruciate retaining knees with an average followup of 37 months and found no clinical difference between the two groups, although they did observe diminished ROM in the previous high tibial osteotomy group. Similarly, Toksvig-Larsen et al20 reviewed 40 cemented and uncemented posterior cruciate retaining and posterior stabilized total knee replacements and saw no clinical difference with respect to overall knee scores, pain, ROM, migration, or component loosening.

Bergenudd et al3 reviewed 99 hybrid posterior cruciate retaining total knee replacements, 14 of which were after high tibial osteotomy, and found no clinical radiographic differences between these two groups. The followup of the total knee replacement group was longer (96 months) than that for the primary total knee replacement group (72 months). Finally, Nizard et al16 compared patients with 63 cemented posterior cruciate retaining and posterior stabilized knees with a patient matched group with an average followup of 55 months and found no significant differences in the clinical scores or ROM. Because of the technically demanding nature of total knee replacement after high tibial osteotomy these authors still suggested that the indications of high tibial osteotomy should be reevaluated for older and less active patients.16

Ritter et al18 reviewed the survival of the posterior cruciate ligament retaining AGC total knee prosthesis. At the 7-year followup (327 knees), the knee score averaged 75.3 points and the function score averaged 80.1 points. Although 1-year, 3-year, and 5-year knee scores were 77, 77, and 78 points, respectively, 6-month knee scores averaged 75.3 points. Ten-year survival was greater than 98%.

In the current study, with an average followup of 8.6 years, knee and function scores averaged 87.6 and 83.7 points, respectively. Excluding the one excisional arthroplasty secondary to infection and the three knees revised for loose metal backed patellar components, at the time of this review, only one knee has been revised secondary to component loosening (an all polyethylene tibial component). Thus, 98.8% of knee prostheses remain in place at the most recent followup. If the two patients who were lost to followup are considered as having failed surgery, 96% of the knees prostheses still are functioning. Both of the posterior cruciate retaining total condylar total knee replacements performed are functioning well with no evidence of loosening at 15 and 17 years' followup. Based on this review, the clinical results of total knee replacement after high tibial osteotomy are similar to results of primary total knee replacement.

Factors that have potential deleterious effects of total knee replacement after high tibial osteotomy include the presence of multiple surgical incisions, diminished ROM (flexion contracture after high tibial osteotomy), peroneal nerve palsy after high tibial osteotomy, reflex sympathetic dystrophy, collateral ligament laxity, multiple previous surgeries, patella infera, nonunion after high tibial osteotomy, retained hardware after high tibial osteotomy, a truncated lateral tibial metaphysis commonly found after high tibial osteotomy, and the valgus deformity that commonly is present in a patient after high tibial osteotomy.14,15

Because no patient reported in the current review had a peroneal nerve palsy or reflex sympathetic dystrophy after high tibial osteotomy, these two factors were not studied. In addition, because only one patient reported in the current review had nonunion after high tibial osteotomy, this factor was not evaluated specifically. Difficulty attributable to a truncated lateral tibial metaphysis was not encountered in seating the keel of the prosthesis in any knee. The absence of this potential problem in the current study may be attributable to the few severely overcorrected knees after high tibial osteotomy (average preoperative alignment, 3.9° valgus) in the current study and that no stem extensions were used with the knee prostheses. In no case was the tibial component medialized or downsized because of abutment of the keel against the truncated lateral metaphysis.

Of the remaining factors, only two adversely affected the clinical results of total knee replacement after high tibial osteotomy and then only with respect to ROM. Although an increased flexion contracture and increased number of previous surgeries correlated with diminished postoperative motion, the severity of these factors was not associated with postoperative pain, knee score, or function score. Although not clinically evident, an increased number of femoral radiolucent lines were observed in the knees with an increased number of previous surgeries and with patella infera. In addition, the presence of retained hardware after high tibial osteotomy did relate to an increased number of tibial radiolucencies. Finally, the greater degree of which the lateral tibial joint line was raised, referenced from fibular head, did correlate to an increased number of patellar radiolucent lines. Although the reason for this correlation is uncertain, it may be attributable to the relative change in the patellar position with respect to the actual joint line. However, in theory the position of the lateral joint line after total knee replacement would be that of the position of the lateral joint line before high tibial osteotomy with respect to the tibial metaphysis.

Although the patient population in this study was more homogeneous than that of previous reports, differences did exist before surgery. All patients had a previous high tibial osteotomy, yet the previous surgery varied. Dome and wedge osteotomies were included. Other differences included the number of previous surgeries and the presence or absence of hardware. The latter two had an effect on the overall result of total knee replacement after high tibial osteotomy (Table 6). The results of total knee replacement according to the osteotomy type were not compared. Patients with a dome osteotomy were included in the current study because the lateral tibial metaphysis remains truncated relative to the fibular head. However, regardless of the type of osteotomy, the position of lateral joint line relative to the fibular head varied tremendously (preoperative range, 5 mm distal to the fibular head and 31 mm proximal to the fibular head). Raising the lateral joint line appeared to have no effect on postoperative pain, knee score, function score, or ROM. For example, the postoperative knee score in the patient whose preoperative joint line was positioned 31 mm proximal to the fibular head was 90 points, with a function score of 80 points with no pain.

Major technical considerations in performing total knee replacement after high tibial osteotomy include exposure, proximal tibial bone resection, and ligamentous balancing. Mont et al14 and Windsor et al21 stressed the potential need for lateral patellar release to facilitate surgical exposure. Although such a lateral release was performed in 45% of knees reported in the current review, in no case was it required for initial exposure. No lateral release, quadriceps snip, or tibial osteotomy was required to gain exposure in patients reported in the current study, which may be a reflection of the ROM that was present before surgery (Table 2) and that facilitated exposure, and not a consequence of surgical technique.

The patients in the current study were not specifically compared with patients in a primary series with respect to blood loss, transfusion rate, surgical time, hospital stay, or rehabilitation time. Although these variables represent potential differences between total knee replacement with or without a previous high tibial osteotomy, the end result appears to be as good as an index arthroplasty.

To reestablish the anatomic preosteotomy position of the lateral joint line, minimal lateral bone resection is required. At least a 10-mm thick tibial component was used in every case, despite an average of only 4 mm being resected from the lateral proximal tibial metaphysis. In addition, although several authors have recognized the possible need for lateral ligament as balancing, especially in the over-corrected valgus knee,13,14,21 almost 90% of the cases in this review required no such lateral ligament as balancing. Windsor et al21 reported that a lateral fibular collateral ligament release was needed to restore axial alignment in 10 of 45 knees. In the current study, only five knees required release of the lateral capsule with or without popliteus tendon release and release of the iliotibial band. This result may be attributable to only 13 of the 94 knees in this study having a valgus alignment greater than 10° before total knee replacement

At an average of 8.6 years' followup, the clinical results of total knee replacement after high tibial osteotomy appeared similar to those of primary total knee replacement, with average knee scores of 87.6 points with 90% good or excellent results and 99% of patients having no or mild pain. Although no factor leading to an inferior knee, function, or pain score was identified, the severity of preoperative flexion contracture and the number of previous surgeries did relate to diminished postoperative motion.

Back to Top | Article Outline

References

1. Aglietti P, Rinonapoli E, Stringa G, Taviani A: Tibial osteotomy for the varus osteoarthritic knee. Clin Orthop 176:239-251, 1983.
2. Amendola A, Rorabeck CH, Bourne RB, Apyan M: Total knee arthroplasty following high tibial osteotomy for osteoarthritis. J Arthroplasty 4(Suppl):511-517, 1989.
3. Bergenudd H, Sahlström A, Sanzén L: Total knee arthroplasty after failed proximal tibial valgus osteotomy. J Arthroplasty 12:635-638, 1997.
4. Coventry MB, Bowman PW: Long-term results of upper tibial osteotomy for degenerative arthritis of the knee. Acta Orthop Belg 48:139-156, 1979.
5. Coventry MB, Ilstrup DM, Wallrichs SL: Proximal tibial osteotomy: A critical long-term study of eighty-seven cases. J Bone Joint Surg 75A:196-201, 1993.
6. Ewald FC: The Knee Society total knee arthroplasty roentgenographic evaluation and scoring system. Clin Orthop 248:9-12, 1989.
7. Holden DL, James SL, Larson RL, Slocum DB: Proximal tibial osteotomy in patients who are fifty years old or less: A long-term follow up study. J Bone Joint Surg 70A:977-982, 1988.
8. Insall JN, Dorr LD, Scott RD, Scott WN: Rationale of the Knee Society clinical rating system. Clin Orthop 248:13-14, 1989.
9. Insall JN, Joseph DM, Misika C: High tibial osteotomy for varus gonarthrosis: A long-term followup study. J Bone Joint Surg 66A:1040-1048, 1984.
10. Insall JN, Salvati E: Patella positioning in the normal knee joint. Radiology 101:101-104, 1971.
11. Insall JN, Shoji H, Mayer V: High tibial osteotomy: A five year evaluation. J Bone Joint Surg 56A:1397-1405, 1974.
12. Katz MM, Hungerford DS, Krackow KA, Lennox DW: Results of total knee arthroplasty after failed proximal tibial osteotomy for osteoarthritis. J Bone Joint Surg 69A:225-232, 1987.
13. Krackow KA, Holtgrewe JL: Experience with a new technique for managing severely overcorrected valgus high tibial osteotomy at total knee arthroplasty. Clin Orthop 258:213-224, 1990.
14. Mont MA, Alexander N, Krackow KA, Hungerford DS: Total knee arthroplasty after failed high tibial osteotomy. Orthop Clin North Am 25:515-525, 1994.
15. Mont MA, Antonaides S, Krackow KA, Hungerford DS: Total knee arthroplasty after failed high tibial osteotomy: A comparison with a matched group. Clin Orthop 299:125-130, 1994.
16. Nizard RS, Cardinne L, Bizot P, Witvoet J: Total knee replacement after failed tibial osteotomy: Results of a matched-pair study. J Arthroplasty 13:847-853, 1998.
17. Ritter MA, Fechtman RA: Proximal tibial osteotomy: A survivorship analysis. J Arthroplasty 3:309-311, 1988.
18. Ritter MA, Worland R, Saliski J, et al: Flat-on-flat nonconstrained, compression molded polyethylene total knee replacement. Clin Orthop 321:79-85, 1995.
19. Staeheli JW, Cass JR, Morrey BF: Condylar total knee arthroplasty after failed proximal tibial osteotomy. J Bone Joint Surg 69A:28-31, 1987.
20. Toksvig-Larsen ST, Magyar G, Önsten LR, Lindstrand A: Fixation of the tibial component of total knee arthroplasty after high tibial osteotomy: A matched radiostereometric study. J Bone Joint Surg 80B:295-297, 1998.
21. Windsor RE, Insall JN, Vince KG: Technical considerations of total knee arthroplasty after proximal tibial osteotomy. J Bone Joint Surg 70A:547-555, 1988.
© 2000 Lippincott Williams & Wilkins, Inc.