Abstract
We retrospectively studied the results of arthrodesis of the knee with a modular titanium intramedullary nail that couples at the knee. The study group consisted of thirteen patients who had a malignant tumor around the knee, five who had failure of a total knee arthroplasty, and three who had a locally destructive benign tumor about the knee. All of the patients were followed for a minimum of two years.Through a single incision at the knee, one nail was inserted retrograde into the femur and the other, antegrade into the tibia; the two nails were joined at the level of the knee by a conical couple and were secured with locking screws. The diameters of the nails were different, to accommodate the dissimilar sizes of the tibial and femoral intramedullary canals.A solid osseous fusion was achieved in nineteen (90 per cent) of the twenty-one patients (sixteen who had had resection of a tumor and three who had had a failed arthroplasty), at an average of 8.4 months (range, three to nineteen months) after the operation. One patient had a delayed union, but fusion was achieved after additional bone-grafting. Of the sixteen patients who were available for clinical and radiographic evaluation at the time of the study, fifteen were satisfied with the over-all outcome and thirteen had either less pain or the same amount of pain as they had had preoperatively. There were no mechanical failures of the implant and no recurrences of tumor. Complications occurred in eight (38 per cent) of the twenty-one patients: three patients had a stress fracture, three had a peroneal nerve palsy (one of which was transient), one had a superficial wound infection, and one had reflex sympathetic dystrophy.
Arthrodesis of the knee has been attempted with a number of operative techniques, the most common of which has been either external fixation or intramedullary fixation with a Küntscher or Sampson device2,5,6,11,13,14,24,25. However, both of these techniques have inherent limitations.
The use of an external fixator has two potential restrictions: the frequent development of pin-track infections may limit the amount of time that the device can be worn, and the knee cannot be stabilized in a patient who has severe bone loss6,15. A review of the literature shows over-all rates of success ranging from three of fourteen to four of four or 169 (99 per cent) of 1712,5,11,15,19,22.
Intramedullary fixation for arthrodesis has had a higher reported rate of success4,7-10,13,14,18-20,24,25 and has been shown to provide greater stability than external fixation16. However, migration and breakage of the intramedullary device, neurovascular injury, tibial fracture during insertion of the device, and delayed union are not uncommon complications1,6,7,10,13,14,18,25. Also, when the diameters of the femoral and tibial intramedullary canals are dissimilar, optimum fixation with a one-piece, one-diameter intramedullary device is difficult.
The purpose of this investigation was to analyze our results with the use of a modular titanium intramedullary nail system for arthrodesis of the knee. The modularity of this system provides a number of advantages: the nail can be inserted through a single incision at the knee, physiological genu valgum can be preserved, the nail can be inserted in such a way as to adjust for malrotation of the extremity, and the system accommodates dissimilar diameters of the femoral and tibial intramedullary canals.
*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. No funds were received in support of this study.
†Department of Orthopaedics, University of Nebraska Medical Center, 600 South 42nd Street, Omaha, Nebraska 68198-1080.
Twenty-four patients had an arthrodesis of the knee and skeletal fixation with use of a mechanically coupled, modular titanium intramedullary nail system that was inserted retrograde into the femur and antegrade into the tibia. All of the patients were operated on by the two senior ones of us (K. L. G. and J. R. N.). Twenty-one patients were followed for a minimum of two years, and they form the basis of this report. Sixteen (76 per cent) of the twenty-one patients were available for clinical and radiographic evaluation and the completion of a detailed written questionnaire at the time of the study. Three of the remaining five patients had died and two had been lost to follow-up, but all five had been followed for at least two years. The records of these five patients were reviewed to evaluate the results of the operation. The hospital records and radiographs of all patients were reviewed. At the time of the latest examination, the knee was assessed for clinical and radiographic signs of fusion and the effect of the procedure on the patient's ability to carry out activities of daily living was evaluated. Radiographic evidence of fusion was defined as osseous trabeculation across the site of the arthrodesis.
Twelve of the patients were male and nine were female. The average age was forty years (range, twelve to eighty years). The average duration of follow-up was 4.8 years (range, two to ten years). There were fourteen left knees and seven right knees. Thirteen patients had a malignant tumor, three had a benign locally destructive tumor, three had failure of a total knee arthroplasty due to infection, and two had aseptic loosening of a constrained knee prosthesis (Table I). Nine of the thirteen patients who had a malignant tumor received chemotherapy before and after the operation.
A modular prosthetic design provides an attractive alternative to a one-piece custom fabrication. Although the implant was originally conceived as an alternative to a single-diameter, one-piece intramedullary nail for skeletal fixation after resection of a tumor, it also can be used for arthrodesis of the knee after aseptic failure of a knee arthroplasty and for post-traumatic osteoarthrosis. When the device is used after failure of an arthroplasty associated with infection, complete débridement of all infected tissues, removal of the implant, and a six to eight-week course of local and systemic antibiotic therapy are recommended before the arthrodesis is performed.
Nail System
The modular nail system (Neff Femorotibial Nail System; Zimmer, Warsaw, Indiana) consists of two titanium-alloy (Ti-6Al-4V) intramedullary nails with longitudinally machined splines, which are joined at the level of the knee by a press-fit conical couple. The modular design allows the surgeon to use a curved femoral nail (152-centimeter radius of curvature) and a straight tibial nail of different diameters, thereby accommodating the often dissimilar diameters of the femoral and tibial intramedullary canals. The conical couple consists of a 4-degree conical taper with a locking groove machined distally to engage three equilaterally spaced locking screws to prevent rotation or possible decoupling. The locking screws further engage the locking groove in the taper, thus enhancing the press-fit and preventing decoupling by distraction and rotation (Figs. 1-A, 1-B, and 1-C). The biomechanical properties of the conical couple have been described previously3,12.
Operative Procedure
A number of approaches to the knee can be used, depending on the indication for the operation. A standard midline approach was typically used for patients who had had failure of a knee arthroplasty (Fig. 2-A). For patients who were to have resection of a tumor, a triradiate, Y-shaped incision, with the apex at the center of the knee (Mercedes-type incision), including any biopsy tracks, was used to provide at least a wide en bloc resection (Fig. 3-A).
For patients who have a failed implant, the opposing surfaces of the femur and tibia are prepared with use of cutting guides to provide flat parallel surfaces. Next, the intramedullary canals are prepared, with use of flexible reamers, by retrograde reaming of the femur and antegrade reaming of the tibia; the reaming progresses in 0.5-millimeter increments until the diameter is 1.5 millimeters greater than the diameter of the nail to be inserted. The isthmus of the femur and that of the tibia remain the best regions for firm intramedullary fixation.
The modular nail system is commercially available with a diameter of eleven, thirteen, or fifteen millimeters and a length of 340 millimeters (the femoral nail) or 370 millimeters (the tibial nail). For a man of average weight (eighty kilograms), a fifteen-millimeter femoral nail and a thirteen-millimeter tibial nail are most often used and are recommended. For a woman of average weight (sixty kilograms), use of a thirteen-millimeter femoral nail and an eleven-millimeter tibial nail has been suggested. The appropriate diameter can be estimated preoperatively from the radiograph templates that are provided with the system, or the diameter of the intramedullary canals can be calculated by determining the diameter of the reamer that best engages the isthmus over a distance of four to six centimeters.
Each nail is cut to an appropriate length with use of a diamond saw (WH2 bit; Midas Rex, Fort Worth, Texas) or a cutting tool intraoperatively, or the nail may be sent to the biomedical engineering shop, where it is cut and reshaped with a bullet tip and then resterilized for use. The length can best be estimated by reaming to provide for a nail five to six centimeters beyond either isthmus. After the canal has been prepared, the proximal tip of the femoral nail should extend four to six centimeters beyond the femoral isthmus, and the distal tip of the tibial nail should be inserted five to six centimeters beyond the tibial isthmus, without penetrating the ankle joint.
The resected surfaces of the proximal part of the tibia and the distal portion of the femur then are reamed, with use of the custom conical couple reamer (provided with the nail system), to a diameter of twenty-two millimeters to accommodate seating of the coupling device (Fig. 2-B). The depths of reaming necessary for the femoral and tibial conical couples are inscribed on the conical couple reamer. These limits should not be exceeded, as the conical couples will migrate in each canal at the time of coupling, making it difficult to join the nails.
Genu varum or genu valgum may be adjusted to approximate that of the contralateral limb by the appropriate insertion of the curved femoral nail. While preparing for the retrograde insertion of the femoral nail, the arc of rotation of the femoral canal (that is, the plane of the physiological curve) can be estimated by insertion of the sizing wands that are provided. Insertion of the femoral nail with internal rotation, as viewed from the knee, results in slight genu valgum, whereas insertion with the plane in external rotation results in slight genu varum. Neutral rotation is most often recommended. The femoral nail is inserted with use of a slap hammer, which is attached to the nail by threads. Protective sleeves, made of polyfluoroethylene fiber-reinforced polyacetyl (Delrin; E. I. Dupont, Wilmington, Delaware) are used to prevent damage to the male and female tapers during insertion and extraction. Insertion should provide a press-fit, engaging the sharp longitudinal splines in the medullary canal.
The tibial nail should be oriented before it is inserted, to provide access to at least two of the three locking-screw holes. The third hole (oriented posteriorly) is often operatively inaccessible. After both nails have been inserted, the retainer plates (provided with the system) are inserted perpendicularly in the threads of the collar of each couple to prevent recession during the process of joining the nails. The knee is brought to full extension with the flat retainer plates attached, and the nails are joined (Figs. 2-C and 2-D). The retainer plates are removed, and the lower limb is externally rotated to approximate the position of the contralateral limb. The heel then is tapped with the gloved hand to seat the taper and compress the site of the arthrodesis. With careful observation, a space, approximately one millimeter (but no larger), is seen between the collars of the conical couples (Figs. 1-C and 2-D). If decoupling is necessary to change the rotation, a thin wedge-like separator can be positioned in the space between the nails to decouple the mechanical taper. Two of the three locking screws are used to secure the position of the nails and to further seat the taper. Bone graft then is placed around the site of the arthrodesis. For large defects, autogenous rotatory grafts of bone from the tibia and the femur, as well as autogenous bone from the fibula, are used for reconstruction9. When necessary, muscle flaps are used to facilitate closure of the wound after resection of a tumor (Figs. 3-A, 3-B, 3-C, 4-A, 4-B through 4-C).
The patients who did not have a segmental defect were permitted to walk bearing weight as tolerated on the first postoperative day. A clamshell polypropylene-lined knee-ankle-foot orthosis was used to stabilize the site of the arthrodesis postoperatively. The patients who had a segmental defect were allowed to walk with toe-touch weight-bearing, while wearing the orthosis, immediately after the operation. All of the patients wore the orthosis until there were early signs of radiographic healing or they were able to bear full weight without pain (usually in about six weeks), and they used crutches or a walker until they had no pain while walking.
A solid osseous fusion was achieved in nineteen (90 per cent) of the twenty-one knees at an average of 8.4 months (range, 3.2 to 19.3 months) after the primary procedure.
Both patients in whom an osseous fusion was not achieved after the arthrodesis had had severe bone loss after a failed arthroplasty, and a titanium spacer was used to restore the length of the limb. For one of these patients (Case 2), a six-centimeter titanium spacer was used and bone-grafting was not performed at the time of the arthrodesis. By the time of the two-year follow-up examination, an osseous fusion had not been achieved. The second patient (Case 12) had a seven-centimeter spacer and bone-grafting with demineralized bone matrix at the time of the index operation. Repeat bone-grafting was necessary nineteen months postoperatively, and fusion was evident fourteen months later (thirty-three months after the primary attempt at arthrodesis).
Of the sixteen patients who were available for clinical and radiographic evaluation at the time of the study, eleven continued to have some pain in the extremity and five had no pain (Table I). The pain was mild in four of the eleven patients, moderate in six, and severe in one. Three patients rated the over-all pain as being worse than before the operation: two patients (Cases 7 and 19) who had had no pain preoperatively had moderate pain after the arthrodesis, and one patient (Case 1) who had had mild pain preoperatively had severe pain after the operation. Despite the discomfort in the extremity, these patients were satisfied with the outcome when they considered the alternatives for successful resection of the tumor. The five patients who had had a failed total knee arthroplasty had severe pain in the knee before the arthrodesis, but they all had less pain after it. Although none of the patients had severe pain around the knee postoperatively, nine patients had pain in the area of the limb corresponding to the end of a nail: five patients had pain in the ankle and four had pain in the thigh.
Data on the walking ability was available for eighteen of the twenty-one patients. Twelve of the patients were able to walk without the use of an assistive device, five used a cane, and one used a walker. Fourteen patients could walk the same distance as preoperatively or farther, and four patients could not walk as far as they could preoperatively. Four of the fourteen patients had no limitations with regard to walking.
Fifteen of the sixteen patients who returned for the follow-up examination stated that they were satisfied with the over-all outcome. Over-all satisfaction was not evaluated solely on the basis of one variable, such as pain or walking distance. The one patient who was not satisfied (Case 6) thought that the arthrodesis was too disabling for activities of daily living and that it was cosmetically unacceptable. Despite this, he had less pain and remained disease-free.
The patients who had resection of a tumor lost an average of 2199 milliliters (range, 325 to 5300 milliliters) of blood intraoperatively, and throughout the hospital stay an average of seven units (range, one to eighteen units) of blood was given. The average duration of the operation for these patients was 8.5 hours (range, seven to eleven hours). The patients who had an arthrodesis for a failed total knee arthroplasty lost less blood (average loss, 865 milliliters; range, 325 to 1300 milliliters) (p = 0.001), received fewer units of blood (average number transfused, two units; range, zero to five units) (p < 0.01), and had a shorter operation (average duration, 3.4 hours; range, 1.5 to 4.8 hours) (p = 0.001).
Nine (43 per cent) of the twenty-one patients needed a soft-tissue flap for coverage of the wound after resection of a tumor. At the time of the primary procedure, twenty (95 per cent) of the twenty-one patients had grafting with combinations of demineralized bone matrix, fibular strut graft, femoral or tibial rotatory graft, bone from the iliac crest, or allograft (Table I).
There were eight complications, for an over-all rate of 38 per cent. Three patients (Cases 5, 6, and 7) sustained a stress fracture at the site of the arthrodesis, which was diagnosed by the onset of new pain; the pain subsided in all three patients after the limb was immobilized in a cast. Three patients (Cases 1, 14, and 21) had a peroneal nerve palsy, which resolved in only one of them. The palsy in all three patients was thought to be the result of the resection of a tumor involving the proximal part of the tibia. One patient (Case 13) had a superficial wound infection, which responded to antibiotics administered intravenously, and one patient (Case 19) had reflex sympathetic dystrophy, which resolved spontaneously.
At the time of writing, there had been no local recurrence of tumor in any patient; however, three patients had died of metastatic disease. There were no mechanical failures of the implant and no signs of radiographic loosening.
Arthrodesis of the knee is rarely indicated, and it may not be successful even when it has been performed technically well. Sixteen (76 per cent) of the twenty-one patients in the present study had an arthrodesis after resection of a tumor about the knee. Enneking and Shirley described the use of this technique for arthrodesis after resection in twenty patients. Their operative indication was a tumor around the knee that could be resected en bloc in a vigorous patient whose lifestyle required a stable, pain-free extremity. Those authors recommended the use of a custom-made fluted nail with a curved femoral portion and a straight tibial portion. Four of the twenty patients in their study had a non-union that eventually healed after subsequent bone-grafting. In addition, a number of complications were associated with the nail and the grafts, and three patients had a transient peroneal-nerve palsy9.
Our results were very similar to these, except that there were no mechanical failures of the implant in our series, whereas there were three failures in the twenty patients in the study by Enneking and Shirley. There were no local recurrences of tumor in either their series or ours.
Five of our patients had an arthrodesis because of failure of a total knee arthroplasty. The indications for an arthrodesis rather than a revision arthroplasty remain poorly defined6, and the choice must be based on the needs of the individual patient. The indications for an arthrodesis have been reported to be infection of the joint with soft-tissue deficits, evidence of a virulent organism, failure of a hinged prosthesis with severe bone loss, instability of the knee, arthropathy with severe deformity, a deficient extensor mechanism, good general health of the patient, a young age, and the activities of daily living and needs of the patient2,6,14,18,19,22,23.
Of the five patients in our series who had a failed total knee arthroplasty, three had failure due to infection and two had failure of a constrained prosthesis with massive bone loss. The arthrodesis was successful in all three of the patients who had infection. Of the two patients who had failure of a constrained prosthesis, one had a delayed union, with fusion occurring after additional bone-grafting, and the other still had a non-union at the time of writing. It has been reported to be much more difficult to achieve union in a patient who has an arthrodesis because of a failed constrained prosthesis2,15,18-20,22. In both such patients in our series, a large titanium spacer was placed between the nails to compensate for shortening of the limb and bone loss. Rasmussen et al. reported a successful fusion in twelve of thirteen patients with at least six centimeters of bone loss who had been managed with a vascularized fibular rotatory graft. In our series, the patient who had a non-union preoperatively (Case 2) had had poliomyelitis and she was confined to a bed or a chair. She was unable to assist in transfers because of extreme pain and instability of the knee. For this patient, the goal of the operation was not necessarily osseous fusion but rather a stable extremity for transfers. The operation was considered to be successful, and both the patient and the caregivers were satisfied.
The rate of complications (38 per cent) in our series is lower than the rates of 40 to 55 per cent that have been reported in the literature8,15,19,20,25. All of the complications in our patients were considered manageable. There were no failures of hardware, as have occurred in other studies1,6,7,9,10,14. The operative time, blood loss, and number of units of blood transfused in our series were comparable with those in other recent reports7,8,20.
The modular titanium system has several potential advantages compared with an external fixator or a Küntscher nail. Unlike external fixators, an intramedullary nail can bridge a large bone defect and the patient gains functional stability immediately after the operation. The modular design allows for insertion through a single incision at the knee, with retrograde insertion into the femur and antegrade insertion into the tibia. This can be advantageous for an obese patient, in whom the approach to the piriformis fossa is difficult. Also, unlike the situation with a standard intramedullary nail18, the physiological genu valgum of the extremity can be maintained. In contrast to a one-piece intramedullary nail, the modular system has two components with a stable coupler, which allows for a more accurate fit of the nail in the intramedullary canal and for accommodation of the nail to the contour of the femoral shaft, thus reducing the risk of fracture during insertion.
Nevertheless, there are a number of potential drawbacks to this system. Of most concern is that the nails can be difficult to retrieve after the arthrodesis, although retrieval was not necessary in any patient in our series. However, if a solid fusion has been achieved, an anterior cortical window is created, and the nail is cut with high-speed drills, with sequential removal of the cut segments. Another disadvantage inherent in any intramedullary system is that high compression forces cannot be obtained at the site of the arthrodesis. All compression is achieved by vertical loading of the extremity during walking. Finally, for patients who may need a total hip arthroplasty in the future, a shorter femoral nail must be considered in order to accommodate the femoral stem of the prosthesis.
We concluded that the use of a fluted modular titanium intramedullary nail results in high rates of osseous fusion and of patient satisfaction. Although arthrodesis of the knee is rarely indicated, it can be a successful means of reconstruction of the limb after resection of a tumor and as a salvage procedure after failure of a total knee arthroplasty.
Alman, B. A.; De Bari, A.; and |and |Krajbich, J. I.: Massive allografts in the treatment of osteosarcoma and Ewing sarcoma in children and adolescents. J. Bone and Joint Surg.,77-A: 54-64, Jan. 1995.77-A54
1995
Brodersoen, M. P.; Fitzgerald, R. H., Jr.; Peterson, L. F. A.; Coventry, M. B.; and |and |Bryan, R. S.: Arthrodesis of the knee following failed total knee arthroplasty. J. Bone and Joint Surg.,61-A: 181-185, March 1979.61-A181
1979
Chao, E. Y.; Kwak, B. M.; and |and |Kasman, R.: Stress analysis of conical coupling joint in modular prosthetic system design. Trans. Orthop. Res. Soc.,9: 103, 1984.9103
1984
Chapchal, G.: Intramedullary pinning for arthrodesis of the knee joint. J. Bone and Joint Surg.,30-A: 728-734, July 1948.30-A728
1948
Charnley, J., and |and |Lowe, H. G.: A study of the end-results of compression arthrodesis of the knee. J. Bone and Joint Surg.,40-B(4): 633-635, 1958.40-B(4)633
1958
Damron, T. A., and |and |McBeath, A. A.: Arthrodesis following failed total knee arthroplasty: comprehensive review and meta-analysis of recent literature. Orthopedics,18: 361-368, 1995.18361
1995
[PubMed]
Donley, B. G.; Matthews, L. S.; and |and |Kaufer, H.: Arthrodesis of the knee with an intramedullary nail. J. Bone and Joint Surg.,73-A: 907-913, July 1991.73-A907
1991
Ellingsen, D. E., and |and |Rand, J. A.: Intramedullary arthrodesis of the knee after failed total knee arthroplasty. J. Bone and Joint Surg.,76-A: 870-877, June 1994.76-A870
1994
Enneking, W. F., and |and |Shirley, P. D.: Resection-arthrodesis for malignant and potentially malignant lesions about the knee using an intramedullary rod and local bone grafts. J. Bone and Joint Surg.,59-A: 223-236, March 1977.59-A223
1977
Fern, E. D.; Stewart, H. D.; and |and |Newton, G.: Curved Küntscher nail arthrodesis after failure of knee replacement. J. Bone and Joint Surg.,71-B(4): 588-590, 1989.71-B(4)588
1989
Green, D. P.; Parkes, J. C., II; and |and |Stinchfield, F. E.: Arthrodesis of the knee. A follow-up study.. J. Bone and Joint Surg.,49-A: 1065-1078, Sept. 1967.49-A1065
1967
Gruenberg, S.: The structural analysis of the femoro-tibial intramedullary nail using finite element and experiment analysis. Dissertation, Michigan Technological University, Houghton, Michigan, 1987.
Harris, C. M., and |and |Froehlich, J.: Knee fusion with intramedullary rods for failed total knee arthroplasty. Clin. Orthop.,197: 209-216, 1985.197209
1985
[PubMed]
Knutson, K.; Lindstrand, A.; and |and |Lidgren, L.: Arthrodesis for failed knee arthroplasty. A report of 20 cases. J. Bone and Joint Surg.,67-B(1): 47-52, 1985.67-B(1)47
1985
Knutson, K.; Hovelius, L.; Lindstrand, A.; and |and |Lidgren, L.: Arthrodesis after failed knee arthroplasty. A nationwide multicenter investigation of 91 cases. Clin. Orthop.,191: 202-211, 1984.191202
1984
[PubMed]
Mazet, R., Jr., and |and |Urist, M. R.: Arthrodesis of the knee with intramedullary nail fixation. Clin. Orthop.,18: 43-53, 1960.1843
1960
Pritchett, J. W.; Mallin, B. A.; and |and |Matthews, A. C.: Knee arthrodesis with a tension-band plate. J. Bone and Joint Surg.,70-A: 285-288, Feb. 1988.70-A285
1988
Puranen, J.; Kortelainen, P.; and |and |Jalovaara, P.: Arthrodesis of the knee with intramedullary nail fixation. J. Bone and Joint Surg.,72-A: 433-442, March 1990.72-A433
1990
Rand, J. A.: Instructional course lecture. Alternatives to reimplantation for salvage of the total knee arthroplasty complicated by infection. J. Bone and Joint Surg.,75-A: 282-289, Feb. 1993.75-A282
1993
Rasmussen, M. R.; Bishop, A. T.; and |and |Wood, M. B.: Arthrodesis of the knee with a vascularized fibular rotatory graft. J. Bone and Joint Surg.,77-A: 751-759, May 1995.77-A751
1995
Rodrigo, J. J.: Arthrodesis of the knee. In Operative Orthopaedics, edited by M. W. Chapman. Vol. 1, pp. 631-636. Philadelphia, J. B. Lippincott, 1988.
Stulberg, S. D.: Arthrodesis in failed total knee replacements. Orthop. Clin. North America,13: 213-224, 1982.13213
1982
Thornhill, T. S.; Dalziel, R. W.; and |and |Sledge, C. B.: Alternatives to arthrodesis for the failed total knee arthroplasty. Clin. Orthop.,170: 131-140, 1982.170131
1982
[PubMed]
Vander Griend, R.: Arthrodesis of the knee with intramedullary fixation. Clin. Orthop.,181: 146-150, 1983.181146
1983
[PubMed]
Wilde, A. H., and |and |Stearns, K. L.: Intramedullary fixation for arthrodesis of the knee after infected total knee arthroplasty. Clin. Orthop.,248: 87-92, 1989.24887
1989
[PubMed]