Secondary Logo

Journal Logo

SECTION I: SYMPOSIUM: Papers Presented at the 2006 Meeting of the Knee Society

Extensor Mechanism Allograft Reconstruction in TKA at a Mean of 56 Months

Burnett, R, S. J; Butler, R, A; Barrack, R, L

Section Editor(s): Laskin, Richard S MD, Guest Editor

Author Information
Clinical Orthopaedics and Related Research: November 2006 - Volume 452 - Issue - p 159-165
doi: 10.1097/01.blo.0000238818.25530.2b


Disruption of the extensor mechanism after or during total knee arthroplasty (TKA) is an infrequent complication with an incidence ranging from 0.10% to 2.5%.8,10,18,22 A disruption of the extensor mechanism may occur at different anatomic locations and have several etiologies.21 Quadriceps tendon rupture, patellar fracture, patellar tendon rupture, and avulsion of the patellar tendon are the most common causes of extensor mechanism failure in association with TKA.5 Primary repair of the disrupted extensor mechanism in these cases is frequently unsuccessful.11,18,19,22 The use of local autogenous tissue to augment a primary repair has been described,1,8,14,16,22,24 but these autogenous tissues may be compromised in patients who have undergone multiple previous knee procedures.

The use of allograft tissues to reconstruct the deficient extensor mechanism initially showed promise, but was associated with failures.12,17 Modification of allograft surgical techniques5,20 have evolved, and currently the use of allograft knee extensor mechanism (quadriceps tendonpatella-patellar tendon-tibial bone block)2,4,5,13,20 and Achilles tendon allografts (Achilles tendon-calcaneal bone block) have shown promise2,9 in clinical studies, and with histologic evidence of allograft incorporation into host tissues and bone.6 One of us (RLB) previously reported2 a short-term followup of 14 patients treated with allograft reconstruction with encouraging results, but with less impressive outcomes than recently published series using similar techniques.4,5,9,20

We describe the results of extensor mechanism allograft reconstruction at midterm followup for extensor lag, loss of flexion, functional ambulatory capacity, and patient satisfaction, and to compare these results to reported studies4,5,9,12,13,17,20,21 of similar patient cohorts.


We retrospectively reviewed 19 consecutive patients (19 knees) with chronic extensor mechanism disruption after TKA who underwent allograft reconstruction. Since the previous report of 14 patients,2 five additional patients have been added to the cohort and the followup is longer. There were 13 women and six men with a mean age of 66 years (range, 51-81 years) at the time of allograft reconstruction. Before surgery, three patients had been wheelchair bound and the remaining 16 patients required some form of ambulatory assist device (crutches, walker, or cane) to ambulate and were classified as household ambulators. Minimum followup was 2 years (mean 56 months; range, 24-96 months). This study was approved by the institutional review board.

Fourteen extensor mechanism disruptions occurred after a primary TKA and four disruptions after revision TKA. One reconstruction was performed at a knee-fusion conversion to TKA. In 13 knees, the patellar tendon was ruptured or avulsed chronically. In the remaining six patients, four had a comminuted patellar fracture that had failed an attempt at internal fixation or a prior patellectomy resulting in complete extensor lag, one patient underwent takedown or conversion of a prior knee arthrodesis to TKA, and one patient had an inferior pole avulsion fracture of the patella associated with a chronic extensor lag in association with a TKA. The average time from extensor mechanism disruption to allograft reconstruction was 6.6 months (range, 1-24 months). None of these patients had an acute intraoperative avulsion or laceration of the patellar tendon. In five patients, complete revision of the TKA components occurred at the time of the extensor mechanism reconstruction, whereas in the remaining 14 patients the femoral and tibial components were retained. In nine of these 14 patients, the tibial insert was exchanged because of polyethylene damage and wear, and in the remaining patients a thicker, more constrained compatible modular insert was used to improve stability.

A Knee Society clinical rating score15 was obtained preoperatively and then at a minimum of 2 years postoperatively. This score consists of two categories: one for pain, range of motion, and stability, and another for functional status. There is a maximum score of 100 points for each category. The degree of extensor lag was measured preoperatively and at most recent followup using a goniometer. Preoperatively, 17 of the 19 patients had a complete extensor lag of 90°, one had a lag of 45°, and one had a fused knee. We recorded the patients' ambulatory status and use of assistive devices preoperatively and at most recent followup.

We used two different allograft reconstruction techniques: an Achilles tendon allograft with a calcaneal bone block (ATA), or a complete extensor mechanism composite allograft consisting of quadriceps tendon-patella-patellar tendon-tibial tubercle (EMA). The choice of graft type depended on the location and status of the remaining patella, mobility of the quadriceps, and the status of the bone stock at the tibial tubercle. This algorithm attempts to maintain a viable portion of the host extensor mechanism when appropriate. If the patella was retracted excessively and could not be mobilized to a position placing the patella within 2 to 3 cm of the joint line with the knee in extension, we used a total EMA. If the patella and patellar component were intact and could be mobilized within 2 to 3 cm of the joint line in extension (Fig 1), we retained the patella and used an Achilles tendon allograft with a calcaneal bone block (attached and fixed into the tibial tubercle) sewn to the quadriceps mechanism with the knee in full extension after the quadriceps was mobilized and placed under tension. If there was massive osteolysis, including the tibial tubercle, then an Achilles tendon with calcaneal bone block was used, placing the calcaneal bone block distal to the area of the lytic lesion. If the patella was absent or insufficient, then we used an EMA to help restore the lever arm of the patella. Ten patients had reconstruction with an ATA (Fig 2) and an additional nine patients had a total EMA. All surgeries were performed by the senior author (RLB).

Fig 1A
Fig 1A:
D. This figure shows a reconstruction using an Achilles tendon allograft (ATA). (A) A preoperative lateral radiograph shows a primary TKA with a ruptured patellar tendon. There is a patella alta, and the previously resurfaced patella is intact. The primary TKA components are in satisfactory position. This patient has a complete extensor lag. (B) A postoperative lateral radiograph of TKA following ATA reconstruction is shown. (C) A postoperative anteroposterior radiograph of TKA after ATA reconstruction is shown. (D) A clinical photograph was taken 24 months after ATA reconstruction. This patient has full, active extension with no extensor lag.
Fig 2A
Fig 2A:
F. This figure shows the technique of Achilles tendon allograft reconstruction in TKA. (A) Achilles tendon fresh-frozen allograft is harvested, with an attached calcaneal bone block. A generous bone block is later cut to size for the tibial insertion. (B) The host tibial trough is prepared to accept the allograft bone block. The trough is typically 1.5 cm wide × 2.5 cm long × 1.5 cm deep, and is just medial to the host tibial tubercle. The trough is created smaller than the allograft to allow for a press fit of the slightly oversized allograft bone block. (C) Three wires or cables are passed beneath the floor of the host tibial trough for allograft fixation. (D) The allograft is tamped into place and the wires are tightened. (E) The allograft tendon is passed through a slit in the lateral retinaculum (posterior and lateral) to host patellar tendon remnant, then pulled proximally anterior to the host patella. (F) The allograft is tensioned and sewn into the host quadriceps mechanism with Ethibond suture.

In all cases, a rectangular bone block of approximately 1.5 × 2.5 cm was fashioned and press-fit into a trough medial to the tibial tubercle. Three 18-gauge wires were placed at the base of this trough and twisted over the press-fit bone block (Fig 3A). The quadriceps mechanism was mobilized and placed under tension. We positioned the total extensor mechanism allografts in the midline and used the remaining host retinaculum to oversew or cover the allograft with host tissue whenever possible. This has been described previously.4,5 The Achilles tendon allografts were tensioned with the knee in full extension while the extensor mechanism composite grafts were tensioned in 30° of flexion (to avoid creating a severe patella alta). After the bone block of the Achilles tendon allograft was secured, the distal allograft tendon was passed posterior and lateral to the remnant of the host patellar tendon (Fig 3B). We passed the graft through an opening in the lateral retinaculum and then pulled proximally through the anterolateral retinacular opening and anterolaterally over the patella. This allows the allograft to be passed under host tissue at the level of the joint line. The allograft tendon is then fanned out (widened to maximal allograft surface area) proximally and sewn into the host quadriceps and retinaculum with an interrupted #2 Ethibond suture (Ethicon, Johnson & Johnson, Somerville, NJ). During this procedure, the host extensor mechanism was pulled distally while the allograft was tensioned proximally to bring the native patella to a height of approximately 2 cm above the joint line with the knee in full extension.

Fig 3A
Fig 3A:
B. Schematic drawings of Achilles tendon allograft reconstruction show (A) bone trough preparation with cable passage. The scar tissue lateral to the disrupted patellar tendon is retained. (B) A slit incision is created in the remnant of scar tissue or remaining host patellar tendon, and the allograft Achilles tendon is brought through this slit and then over the host patella and sewn into the proximal quadriceps extensor mechanism.

We standardized the physical therapy protocol for all 19 patients. The knees of all patients were placed in a knee immobilizer in extension for 4 weeks and the patients were mobilized to partial weightbearing with a walker. At 4 weeks, isometric quad- riceps strengthening and straight leg raising exercises were started. Active ROM allowing flexion to 45° also was started and maintained for an additional 4 weeks (total 8 weeks). During the next 8 weeks (postoperative weeks 8 to 16), active ROM in flexion was permitted for up to 90° of flexion, and the patients continued to ambulate with a knee immobilizer in extension. The immobilizer was removed for supervised therapy.

After surgery, patients were evaluated at intervals of 2 weeks, 4 weeks, 12 weeks, 6 months, 1 year, 2 years, and annually thereafter. The degree of patient satisfaction with their function and the outcome of their surgery also was rated at the most recent followup. We rated the success of the procedure by the ability of patients to return to normal daily activities of daily living without reports of knee instability. Patient satisfaction after the procedure was measured using a four-point satisfaction score as extremely (four points), very (three points), moderately (two points), slightly satisfied (one point), or unsatisfied (0 points) (a non-validated previously reported7 scale). Success with the procedure was defined as a return to a community ambulatory status with or without the use of an assistive device, an extensor lag of less than 30°5,17and without symptoms of knee instability.

Statistical analysis was performed using SPSS Version 5.0 software (Chicago, IL). The Wilcoxon signed ranks test was performed on preoperative and postoperative data. Statistical significance was defined as p < 0.05.


The mean Knee Society score improved (p < 0.001) from 46 points preoperatively to 133 points postoperatively. (A supplemental data table is available via the Article Plus feature at You may locate this article then click on the Article Plus link on the right.) The mean total knee score improved (p < 0.001) from 27 points preoperatively (range, 15-51 points) to 76 points (range, 61-87 points) postoperatively. The mean total function score improved (p < 0.001) from 19 points preoperatively (range, 11-34 points) to 57 points (range, 22-78 points) postoperatively.

The mean postoperative extensor lag improved (p < 0.001) to 13.9° (range, 0°- 90°). At most recent followup, one woman with a total EMA had an extensor lag of 45° (slightly satisfied; community ambulatory with a walker), and one woman had a lag of 90° with a failure of the Achilles tendon allograft after a fall and refused additional surgery (community ambulatory with a walker; unsatisfied). One patient had an extensor lag of 30°, ambulated with a cane, and was moderately satisfied. One patient had a 15° extensor lag and the remaining patients had extensor lags of 10° or less (three patients, no lag; seven patients, 5° lag; five patients, 10° lag).

No patients experienced loss of knee flexion after the allograft procedure. A mean preoperative flexion of 99° (range, 0°-120°) was unchanged at 98° postoperatively (range, 80°-110°).

Postoperatively, all 19 patients were community ambulators. Six patients used a cane, six patients used a walker, and seven patients ambulated without assistive devices. At most recent followup, 13 patients required a lesser level of ambulatory assistive device (cane/crutches to nothing; walker to cane; wheelchair to walker, etc), two patients required two levels less of an assistive device (walker to independent, etc), and two had no change. Two patients regressed one level of assistive device.

Eighteen of 19 patients thought the procedure improved their functional status and did not report knee instability at most recent followup. One patient thought the surgery was unsuccessful in terms of returning to participation in normal activities. This patient had a disruption of the allograft reconstruction, reported instability, and refused further surgery.

Thirteen patients (68%) were extremely or very satisfied with the surgery, four were satisfied, one was slightly satisfied and one was unsatisfied. The mean satisfaction score was 2.9 points out of a maximum of four points (non-validated scale). All patients said they would choose to have the surgery again, including the two patients with extensor lags greater than 45°. When the results of extremely, very, and moderately satisfied are combined, the overall satisfaction rate is 17 of 19 patients. In both of these patients, the initial surgery was successful, but was complicated by later failure. One patient fell 6 months after surgery and had displacement of the tibial tubercle bone block. Before falling, the patient achieved full active extension. The bone block was replaced in the trough and rewired, and an Achilles tendon allograft was placed distally and oversewn to the quadriceps tendon to reinforce the repair. This same patient fell again 2 years later, at which time her extensor lag increased from 10° to 45°. This patient is able to ambulate well with a walker and declined any additional surgical intervention. The second patient fell and ruptured the allograft at 36 months, and has refused additional surgery at 90 months after the procedure.


The results of extensor mechanism allograft reconstruction are encouraging. At a mean followup of 55 months, our study provides the longest followup reported with the use of an extensor mechanism allograft used to reconstruct a failed extensor mechanism in TKA. The improvement in preoperative extensor lag, maintenance of knee flexion, and low failure rate of the procedure are similar, but less successful than recent reports4,5,13,20,23 using these techniques. A successful return to activities of daily living without instability of the knee was achieved, and patients reported being satisfied with their surgical outcome. Our results show the use of allograft tissue can reliably restore active knee extension and improve functional status in patients who have had extensor mechanism disruption after TKA.

This study is limited by small numbers that make the results difficult to generalize. To control for potentially confounding variables (including type of graft), larger numbers of extensor mechanism allograft procedures would be necessary to obtain adequate power, although this is an uncommon diagnosis. Regardless of potentially confounding variables, the procedures yielded reasonable results. The procedures are demanding in surgical technique, require the availability of different forms of allo- graft tissue, and a rehabilitation protocol that may be closely monitored. Therefore, the results reported may not be uniformly reproducible.

This study differs from previously published series5,9,12,13,17,20 of allograft reconstruction in TKA because we used a total extensor mechanism allograft and an Achilles tendon allograft based on the status of the host extensor mechanism. We believe this more accurately represents the extensor mechanism failure treatment spectrum the surgeon will encounter. It is useful for the surgeon to be familiar with both treatment techniques. In our series, if the host patella and quadriceps mechanism were intact and able to be mobilized in proximity to the joint line, we used an Achilles tendon allograft for the reconstruction. This allows for preservation of the host tissues and still allows for a successful allograft reconstruction technique. Similarly, we believe if the patella (±patellar component) is intact and may be appropriately mobilized distally, there is no need to sacrifice these structures to implant a total EMA as described by others.4,5

Previous series on of the use of an Achilles tendon allograft have reported encouraging short-term results.6,9,23 The technique recommended for success when using a total extensor mechanism allograft has been reported5,20 and described in detail.3,4 However, the specific technique and surgical procedure have not been well described, to our knowledge, with the use of an Achilles tendon allograft. Crossett et al9 described the successful use of an Achilles allograft in TKA in nine patients. In a retrieval study,6 the use of an Achilles tendon allograft was described using a similar technique to ours, with a modification of the proximal Achilles tendon allograft being split into two arms and then woven into the host extensor mechanism reinforced with nonabsorbable suture. We have found the technique of bringing the Achilles allograft posterior then lateral to the host patellar tendon remnant, and then advancing the allograft through an opening in the retinaculum, with a proximal anastomosis, an effective construct. This allows the allograft to be partially covered by host tissue distally. This has been previously recommended with the use of an EMA.4,5

Early reports of failures using allograft reconstruction in TKA have allowed surgeons to learn from this experience,4,5 and a focus on the techniques of this form of reconstruction has led to improved clinical results. In our series, the Achilles tendon allograft and extensor mechanism composite graft were used depending on the specific indication. If the patella and patellar component were intact and could be mobilized within 2 to 3 cm of the joint line in extension, then an Achilles tendon allograft was used. If the quadriceps mechanism was disrupted chronically and retracted too far proximally, then an Achilles tendon allograft also was used because an extensor mechanism composite graft would not allow sufficient length to restore continuity to the extensor mechanism. In cases where the patella was absent or insufficient, the extensor mechanism composite graft was used as described previously.6,20,25

The results of the current study are in agreement with several recently published series of EMA reconstruction. Similar to the current study, Nazarian and Booth20 reported a postoperative extensor lag of 13° occurred, but was tolerated well and patients generally had a successful clinical outcome, despite a 20% allograft rerupture rate. In a recent report on EMA reconstruction,4,5 the importance of the technique is described. Minimal extensor lag and no loss of knee flexion were reported. Leopold et al17 and others12,13 have reported the high failure rate associated with this procedure if careful attention is not given to technique. The results of the current study are in agreement with recently published series5,9,20; restoration of knee function by reducing extensor lag, maintaining knee flexion, reducing symptoms of instability, and returning patients to activities of daily living may be achieved with the use of an EMA.

We tensioned the EMA tightly before incorporating the host soft tissues. In contrast to a recent series (in which the total extensor mechanism allograft was tensioned with the knee in full extension),5 the senior author (RLB) of the current series prefers to tension the total EMA in slight flexion to avoid a severe patella alta. The Achilles EMAs in this series were tensioned tightly with the knee in full extension. We agree with the principle of tensioning allografts tightly in full extension as recommended by Burnett et al.5

The emphasis on a graduated physical therapy protocol with initial immobilization followed by a protected return to flexion has been previously highlighted by one of the coauthors of the current series (RSB).5 In the current series, the knee was immobilized in full extension for 4 weeks, and during the next 4 weeks flexion was permitted up to 45°. At 8 weeks, patients were permitted to flex up to 90°. This differs from a recently published report by one of the coauthors (RSB) in which no flexion was allowed for 8 weeks, with flexion to 90° only permitted at 12 weeks. The extensor lag in the current series measured 14°, and with the more conservative flexion protocol4,5 the lag measured only 4°. The results of the current study are thus inferior with respect to extensor lag, and we have modified our initial period of immobilization from 4 to 8 weeks, with a more graduated return to flexion as described by Burnett et al.5

Certain technical points probably are crucial for success. It is important to mobilize the extensor mechanism and to free adhesions above and below the quadriceps muscle group. Fresh frozen allograft is preferable to freeze-dried graft, and the graft should be nonirradiated. None of the patellas in EMA used in this series were resurfaced. We prefer tibial fixation with wires instead of screws because we have observed failures through screw holes used for allograft reconstruction. It is important to tension the graft appropriately. We prefer to tension extensor mechanism composite grafts tightly at 30° flexion so the repair is under tension at 45° passive flexion. We maintain the knee extension for at least 4 weeks, after which active-assisted flexion and isometric quadriceps strengthening are started under the supervision of a therapist. Despite this and other studies on this subject being performed at a tertiary care orthopaedic center, we believe this procedure can be performed safely by any orthopaedic surgeon who treats extensor mechanism disruption with a TKA. A careful review of history, physical examination, radiographs, and the location of the native patella (if still viable) should allow the surgeon to predict which type of allograft is needed for the reconstruction.

Extensor mechanism disruption is a serious complication after TKA. Prevention is crucial; it is important to recognize patient and surgical factors possibly predisposing to extensor mechanism disruption. The results of the current study are in agreement with recently published series,5,9,20 and restoration of knee function by reducing extensor lag, maintaining knee flexion, reducing symptoms of instability, and returning patients to activities of daily living may be achieved with the use of an EMA.


1. Abril JC, Alvarez L, Vallejo JC. Patellar tendon avulsion after total knee arthroplasty: a new technique. J Arthroplasty. 1995;10:275-279.
2. Barrack RL, Stanley T, Butler AR. Treating extensor mechanism disruption after total knee arthroplasty. Clin Orthop Relat Res. 2003;416:98-104.
3. Bozic K, Busfield B, Ries M. Patella fractures and total knee arthroplasty. Techniques in Knee Surgery. 2005;4:55-61.
4. Burnett RS, Berger RA, Della Valle CJ, Sporer SM, Jacobs JJ, Paprosky WG, Rosenberg AG. Extensor mechanism allograft reconstruction after total knee arthroplasty. J Bone Joint Surg Am. 2005;87(Suppl 1):175-194.
5. Burnett RS, Berger RA, Paprosky WG, Della Valle CJ, Jacobs JJ, Rosenberg AG. Extensor mechanism allograft reconstruction after total knee arthroplasty: a comparison of two techniques. J Bone Joint Surg Am. 2004;86:2694-2699.
6. Burnett RS, Fornasier VL, Haydon CM, Wehrli BM, Whitewood CN, Bourne RB. Retrieval of a well-functioning extensor mechanism allograft from a total knee arthroplasty: clinical and histological findings. J Bone Joint Surg Br. 2004;86:986-990.
7. Burnett RS, Haydon CM, Rorabeck CH, Bourne RB. Patella resur- facing versus nonresurfacing in total knee arthroplasty: results of a randomized controlled clinical trial at a minimum of 10 years' followup. Clin Orthop Relat Res. 2004;428:12-25.
8. Cadambi A, Engh GA. Use of a semitendinosus tendon autogenous graft for rupture of the patellar ligament after total knee arthroplasty: a report of seven cases. J Bone Joint Surg Am. 1992;74:974-979.
9. Crossett LS, Sinha RK, Sechriest VF, Rubash HE. Reconstruction of a ruptured patellar tendon with achilles tendon allograft following total knee arthroplasty. J Bone Joint Surg Am. 2002;84:1354-1361.
10. Dobbs RE, Hanssen AD, Lewallen DG, Pagnano MW. Quadriceps tendon rupture after total knee arthroplasty: prevalence, complications, and outcomes. J Bone Joint Surg Am. 2005;87:37-45.
11. Doolittle KH 2nd, Turner RH. Patellofemoral problems following total knee arthroplasty. Orthop Rev. 1988;17:696-702.
12. Emerson RH Jr, Head WC, Malinin TI. Reconstruction of patellar tendon rupture after total knee arthroplasty with an extensor mechanism allograft. Clin Orthop Relat Res. 1990;260:154-161.
13. Emerson RH Jr, Head WC, Malinin TI. Extensor mechanism reconstruction with an allograft after total knee arthroplasty. Clin Orthop Relat Res. 1994;303:79-85.
14. Gitomirski ML, Finn HA. Medial gastrocnemius flap for reconstruction of knee extensor mechanism disruption after total knee replacement (TKR). Surg Technol Int. 2004;12:221-228.
15. Insall JN, Dorr LD, Scott RD, Scott WN. Rationale of the Knee Society clinical rating system. Clin Orthop Relat Res. 1989;248:13-14.
16. Jaureguito JW, Dubois CM, Smith SR, Gottlieb LJ, Finn HA. Me- dial gastrocnemius transposition flap for the treatment of disruption of the extensor mechanism after total knee arthroplasty. J Bone Joint Surg Am. 1997;79:866-873.
17. Leopold SS, Greidanus N, Paprosky WG, Berger RA, Rosenberg AG. High rate of failure of allograft reconstruction of the extensor mechanism after total knee arthroplasty. J Bone Joint Surg Am. 1999;81:1574-1579.
18. Lynch AF, Rorabeck CH, Bourne RB. Extensor mechanism complications following total knee arthroplasty. J Arthroplasty. 1987; 2:135-140.
19. MacCollum MS 3rd, Karpman RR. Complications of the PCA anatomic patella. Orthopedics. 1989;12:1423-1428.
20. Nazarian DG, Booth RE Jr. Extensor mechanism allografts in total knee arthroplasty. Clin Orthop Relat Res. 1999;367:123-129.
21. Parker DA, Dunbar MJ, Rorabeck CH. Extensor mechanism failure associated with total knee arthroplasty: prevention and management. J Am Acad Orthop Surg. 2003;11:238-247.
22. Rand JA, Morrey BF, Bryan RS. Patellar tendon rupture after total knee arthroplasty. Clin Orthop Relat Res. 1989;244:233-238.
23. Venkataramanan N, Sinha R, Crossett L. Extensor mechanism reconstruction after total knee arthroplasty with achilles tendon allo- graft. Techniques in Knee Surgery. 2004;3:117-121.
24. Wilson FC, Venters GC. Results of knee replacement with the Walldius prosthesis: an interim report. Clin Orthop Relat Res. 1976; 120:39-46.
25. Zanotti RM, Freiberg AA, Matthews LS. Use of patellar allograft to reconstruct a patellar tendon-deficient knee after total joint arthroplasty. J Arthroplasty. 1995;10:271-274.
© 2006 Lippincott Williams & Wilkins, Inc.