A twenty-four-year-old man sustained a closed fracture of the left femur, just distal to the isthmus, in a high-velocity motor-vehicle accident. The fracture was classified as B3 according to the AO system11 and as C2 according to the system of Tscherne and Oestern (Fig. 1-A). Other injuries included cerebral contusions, fractures of the right femoral condyle and the right tibial plateau, and dislocations of the right elbow and the right carpometacarpal joint. After initial resuscitation, the left femoral fracture was stabilized with a femoral nail, 420 by nine millimeters, made of titanium alloy (Ti-6Al-7Nb; Synthes, Paoli, Pennsylvania). The nail was inserted without reaming and was locked proximally and distally with two locking screws, each with a core diameter of 4.3 millimeters and an outer diameter of 4.9 millimeters7 (Fig. 1-B).
At the initial operation, the dislocations of the right upper extremity were reduced and stabilized. The fracture of the distal part of the right femur and the fracture of the right tibial plateau were treated temporarily with external fixation and then, on the fourth postoperative day, were treated definitively with internal fixation. Because the patient had bilateral injuries, he was kept in bed for forty-two days, during which time a physical therapy program was carried out for both lower extremities. Six weeks after the operation, the patient was allowed to walk, with use of crutches, with full weight-bearing on the left lower limb and partial weight-bearing on the right lower limb. Radiographs of the left femur, made at six weeks, showed no displacement of the fracture but little callus formation.
At nine weeks, the patient had a sensation of instability at the site of the left femoral fracture. There was no history of any additional injury or fall. Radiographs of the femur showed that the nail had broken through the more proximal of the two distal screw-holes (Fig. 1-C). Three days after the broken nail had been detected, it was removed with use of a push-out technique and renailing was performed (Figs. 2-A, 2-B, 2-C, 2-D, 2-E, 2-F, 2-G, 2-H through 2-I).
The patient was placed in the supine position on a radiolucent table. Image-intensifier control was used throughout the procedure. The proximal part of the nail, the end cap, and all of the locking screws were removed with use of standard technique (Fig. 2-A). To remove the distal segment, a four-centimeter longitudinal incision was made over the lateral femoral condyle and was deepened through the iliotibial band onto the lateral condyle. Under image-intensifier control, a guide-wire was placed in the lateral femoral cortex, approximately two centimeters proximal to the lateral articular surface (anterior to the lateral collateral ligament), and was passed obliquely upward through the cancellous bone to the distalmost screw-hole. An opening instrument (UTN; Synthes), such as is usually used with a tibial nail to be inserted without reaming6, then was passed over the guide-wire and a core cylinder of corticocancellous bone was removed, creating a so-called lateral working channel (Figs. 2-B and 2-C).
The bone cylinder was set aside for reimplantation. The working channel then was enlarged proximally, with use of a sharp curet, until the nail was visible (Fig. 2-D). A spike was placed in the distalmost screw-hole, and the nail fragment was worked proximally until a curved narrow Hohmann retractor could be placed underneath its tip (Figs. 2-E and 2-F). Another nine-millimeter femoral nail, designed to be inserted without reaming, then was inserted through a standard opening in the proximal part of the femur and passed distally to the nail segment (Fig. 2-G). The Hohmann retractor was used like a shoehorn to guide the nail through the lateral working channel, and the nail fragment was pushed out. After the fragment had been removed, the corticocancellous bone cylinder was replaced. Another Ti-6Al-7Nb femoral nail (Synthes), 420 by nine millimeters, was inserted without reaming, with care being taken to note the position of the previous distal screw-holes (Figs. 2-H and 2-I). Malalignment in 20 degrees of internal rotation was also corrected. The duration of the operation was 110 minutes, and the duration of fluoroscopy was two minutes and thirty-nine seconds.
Postoperatively, the patient walked with partial weight-bearing (as much as twenty kilograms) for six weeks and then was permitted to walk bearing half of his body weight for three weeks before full weight-bearing was allowed. Union of the fracture was demonstrated radiographically four months after the renailing. At the sixteen-month follow-up examination, the patient was free of pain and had stability and a normal range of motion of the knee (Fig. 3).
An intramedullary nail provides excellent stability for a fractured long bone as well as a good biomechanical environment for the fracture to heal, and it allows early rehabilitation of the limb and the patient. However, the fracture may not heal before fatigue failure of the implant occurs due to cyclical loading8. Several studies have shown a high prevalence of implant failure associated with femoral fractures located distal to the isthmus3,4. The failure is due to stress concentration within the nail, focused around the more proximal of the two distal screw-holes, which predisposes locked nails to break at this level, especially when the fracture is distal to the isthmus5.
We attributed the early fatigue failure of the nail in our patient to the unstable fracture pattern, the distal location of the fracture, and the small diameter of the nail. Early weight-bearing and interlocking nailing in slight distraction also could be important factors. Although they have not yet been proved by well controlled clinical studies, there are several theoretical advantages to the use of a solid femoral nail without reaming; these include avoidance of the damage caused by reaming, less blood loss, and decreased risks of pulmonary complications and infection7,9,12. As the use of femoral nails without reaming increases, the prevalence of breakage of such nails will also increase and orthopaedic surgeons will be faced with the practical problem of how to remove a distal fragment of a solid nail. Although several methods to remove hollow intramedullary nails have been described1,4,8,10,15, there have been no reports, to our knowledge, on techniques to remove large fragments of solid femoral nails. Franklin et al. recommended a grasping device, such as a biopsy forceps, for the removal of a loose fragment of a small-diameter nail. However, those authors stated that they "did not have to contend with any large-diameter solid fragments, which could have been extremely difficult to remove."
The push-out technique described here has several advantages: minimum soft-tissue dissection is needed, reaming of the medullary cavity is not necessary, and extensive intra-articular exposure of the knee is not needed to remove the segment of the nail. A major limitation of this technique is that it cannot be used to remove long fragments of a broken nail. For example, if a nail breaks at the level of the proximal screw-holes or at the level of the fracture, the push-out technique must be modified. With use of the same distal approach, the long fragment is pushed out in a reverse direction (from distal to proximal). The fragment also could be removed with use of an antegrade transarticular intercondylar push-out technique.
Because the procedure is performed in the cancellous part of the distal end of the femur, nails with a larger diameter also can be removed with this technique. When the tip of the nail is located in the infra-isthmal portion of the femur (because the nail was too short), removal is easier than we have described, as there is no need to work the distal end of the nail in a proximal direction. However, a lateral channel that is made too proximal on the femur may cause a potential stress-riser, with a subsequent fracture.
NOTE: The authors thank Jeremy C. Kaye, F.R.C.S., Royal Liverpool University Hospital, Liverpool, Great Britain, and Pierre Guy, F.R.C.S.(C), McLaughlin Fellow, Orthopaedic Division, McGill University, Montreal, Canada, for their revision of the manuscript.