Between December 2003 and January 2007, fifty-five consecutive high subtrochanteric fractures were treated with clamp-assisted reduction and intramedullary nail fixation at two level-I trauma centers (Tampa General Hospital and the University of New Jersey Medical Center). We defined a "high" subtrochanteric fracture as a fracture within two cortical diameters of the lesser trochanter that exhibited the typical deformity of flexion, external rotation, and abduction of the proximal fragment. Pathologic fractures due to neoplasms were excluded. We also excluded more distal subtrochanteric fractures (more than two femoral diameters from the lesser trochanter) as these fractures do not exhibit the classic deformities seen with short proximal fragments. These fractures presented no unique challenges and therefore were treated with traditional antegrade nailing techniques, without the need for adjunctive reduction incisions or clamping. Fractures with an Orthopaedic Trauma Association31 classification of 32 A1.1, A2.1, and A3.1 as well as 32 B1.1, B2.1, and B3.1 were included. Institutional review board approval and patient consent were obtained for our retrospective review. Patients were followed until clinical and radiographic union had occurred or for a minimum of six months, and the mean duration of follow-up was five months (range, four to twenty months). Two patients died, and nine were lost to follow-up. Of the remaining forty-four patients, twenty-seven were male and seventeen were female, and the mean age was fifty-five years (range, twelve to seventy-six years). Forty-three fractures were closed, and one was open. Nineteen fractures resulted from falls and twenty-five, from high-energy mechanisms (motor-vehicle accidents).
The surgery was performed with the patient supine on a fracture table, after administration of a muscle relaxant and a general anesthetic. Prophylactic antibiotics, typically a first-generation cephalosporin (Ancef [cefazolin]), were given. Gentle traction was placed on the limb, and the fracture was visualized in two planes fluoroscopically. If the reduction of the fracture was satisfactory without the need for adjunctive techniques, standard antegrade nailing was performed without adjunctive reduction techniques. These fractures were excluded from the study since they were essentially fractures of the proximal part of the femoral diaphysis. If the fracture exhibited the typical deformities of abduction, flexion, and external rotation of the proximal fragment despite gentle traction, then no attempt was made to achieve a starting point in the proximal fragment in its displaced position. A small incision, about 3 to 5 cm in length, was made in the lateral aspect of the thigh just distal to the lateral prominence of the greater trochanter and was deepened through the fascia lata. A longer incision may be necessary for obese patients. If the fibers of the vastus lateralis were not already disrupted, they were bluntly split in the direction of their fibers with a palpating finger. The clamp was then inserted through this interval in the muscle fibers, not below them (as in a traditional subvastus approach). No attempt was made to visualize the fracture directly.
Oblique and spiral fractures were reduced anatomically and were clamped with a large reduction clamp (large bone forceps; Synthes, Paoli, Pennsylvania). For transverse fractures, the clamp was placed on the proximal fragment, and the deforming forces were reversed—i.e., the fragment was adducted, extended, and internally rotated (Figs. 1 and 2). An assistant held the clamp in the appropriate position as the primary surgeon performed the nailing. The reduction was checked fluoroscopically, and, if it was satisfactory, a guidepin was placed into the proximal part of the femur, either in the piriformis fossa (thirteen patients) or the greater trochanter (thirty-one patients), according to the surgeon's preference, and the nailing was completed (Figs. 3, 4, and 5). It should be noted that the fracture site was not visualized, and the fracture hematoma was not evacuated with suction, although inevitably the hematoma was decompressed by the exposure.
The clamp was placed percutaneously, and the reduction was checked fluoroscopically and with a palpating finger. No deep retractors were used. All nails were inserted with a reaming technique, with the surgeon typically choosing a nail diameter that was 1 to 2 mm less than the size of the reamer that first caused endosteal chatter.
All nails were statically locked with two distal locking screws. The proximal locking configuration was based on the fracture pattern. When the fracture was associated with an intact lesser trochanter and good bone quality (four [9%] of the forty-four fractures), a standard locking screw was placed into the lesser trochanter. When the fracture involved the lesser trochanter or bone of poor quality (forty fractures [91%]), fixation into the femoral head (so-called cephalomedullary fixation) was performed. When the patient had a long spiral or oblique fracture, special attention was directed to observing the behavior of the fracture fluoroscopically after clamp removal with a palpating finger. If there was a tendency for the fracture to displace after the clamp was removed, then a single cerclage cable was placed through the same incision in the same location at which the clamp had been previously positioned. The goal was to allow osseous contact and avoid displacement. This step was performed in nine (20%) of the forty-four patients. It should be stressed that no broad-based medial retractors were used, and only a single cable was placed. The best position for the cable is the best position for the clamp, so additional dissection was not required.
Patients were typically allowed to bear weight as tolerated with two-arm support for the first six weeks; they then progressed to using a cane as tolerated. Radiographs were made immediately postoperatively, at six weeks, at three months, and as needed afterward if union had not occurred.
Clinical union was defined as a painless fracture site while the patient walked without the use of gait aids. Radiographic union was defined as bridging callus across three of four cortices as seen on orthogonal radiographs. Nonunion was defined as hardware breakage or cutout or a lack of radiographic union by six months postoperatively. Alignment was measured immediately postoperatively and at the time of the last follow-up specifically to identify any varus or flexion deformity of the proximal fragment. The central axis of the femoral shaft distally was used as the reference line with which to measure fracture alignment. The cortical contour, laterally on the anteroposterior radiograph and anteriorly on the lateral radiograph, was also used to confirm any malalignment. In addition, the relationship between the center of the femoral head and the tip of the greater trochanter was measured, in order to identify varus malreduction. A deviation of >5 mm relative to the contralateral, unaffected hip on an anteroposterior radiograph (the center of the femoral head being more distal than the tip of the trochanter) was considered to represent varus angulation of the proximal fragment. Rotation was evaluated clinically by comparing internal and external rotation of the hip with internal and external rotation of the contralateral hip and observing the foot progression angle with gait. Limb lengths were measured from the medial malleolus clinically.
Source of Funding
No external funding was received for this study.
Forty-three (98%) of the forty-four fractures united. There were no infections or wound complications. All nine fractures that were treated with supplementary cerclage cable fixation healed. One eighty-one-year-old patient had a nonunion. She had been treated with an antegrade nail with a greater-to-lesser trochanter screw configuration and without use of a cerclage cable, and the alignment was anatomic. The nonunion was initially treated with external bone stimulation followed by dynamization at twenty months after the fracture repair, but it failed to heal.
All reductions were within 5° of anatomic alignment in both planes. Thirty-eight (86%) of the forty-four fractures had an anatomic reduction. Six (14%) of the forty-four fractures had minor varus alignment, of between 2° and 5°, but none of these fractures failed to unite. Importantly, no sagittal plane malunions were noted. No patient had a symptomatic limb-length discrepancy or rotational malunion at the time of the last follow-up.
By virtue of its load-sharing characteristics, the shorter lever on the proximal fixation, and its biologically friendly implantation techniques, intramedullary nail fixation of subtrochanteric fractures has resulted in high union rates1. Although biomechanically and biologically superior to plate fixation, closed nailing techniques can be technically difficult as a result of the position of the proximal fragment1-30,32-43. Awkward starting trajectories have often resulted in eccentric nail channels in the proximal fragment. Starting points were often too lateral, resulting in varus alignment of the proximal fragment. Several limited incision-reduction tools, including Schanz pins and ball-spiked pushers, have been recommended. As a result of our frustration with these "tricks," we began to routinely use a clamp, which we applied percutaneously through a small lateral incision, to reduce these displaced fractures prior to nailing.
We found that overcorrecting the deformity by hyperadducting the proximal fragment greatly facilitates creation of the best starting-point trajectory for the nail, especially in larger patients. We believe that our union rate of 98% is better than or equivalent to any reported in the literature for these problematic fractures, despite our routine use of a clamp to facilitate the nailing. It should be noted, however, that careful placement of the percutaneous clamp and use of a cerclage cable only in selected cases that exhibited instability after nailing probably contributed to our high union rate. We do not advocate wide exposures, periosteal stripping, the use of broad-based retractors, or indiscriminate use of multiple cerclage cables for these fractures42.
Although many have discouraged the use of cerclage cable fixation because of a fear of fracture devascularization1,33,42, the results in our series demonstrate that, when applied in a biologically friendly manner, cerclage fixation can facilitate the achievement of excellent reductions without adversely affecting the union rate44. It is plausible that the excellent osseous apposition provided by the cerclage cable may outweigh the risks of the additional soft-tissue trauma caused by its passage.
The high rate of anatomic reductions in this series is noteworthy but not surprising given the fact that the fractures were reduced first and then nailed. Also, there were no limb-length discrepancies or rotational malunions. This is not surprising, since the use of a reduction clamp, fluoroscopy, and, if necessary, a palpating finger probably allowed relatively accurate assessments of length and rotation as well.
The weaknesses of this study include its retrospective nature, the short duration of follow-up, and the fact that it was a case series without a control group. Additionally, detailed information on operative time and blood loss was not available in sufficient detail to allow comparison of those variables between our procedure and closed nailing techniques.
In conclusion, we found that clamp-assisted reduction and intramedullary nail fixation with selective use of a cerclage cable provided a high rate of fracture union and excellent reduction quality, with no apparent increase in complications, in a large series of "high" subtrochanteric fractures of the femur. Careful attention to the minimization of additional soft-tissue trauma is recommended when surgeons are performing this technique. 