Journal Logo

Original Article

Retrograde Versus Antegrade Nailing of Femoral Shaft Fractures

Ricci, William M*; Bellabarba, Carlo; Evanoff, Bradley; Herscovici, Dolfi§; DiPasquale, Thomas§; Sanders, Roy§

Author Information
Journal of Orthopaedic Trauma: March 2008 - Volume 22 - Issue - p S31-S38
doi: 10.1097/BOT.0b013e31816c0255
  • Free

Abstract

Intramedullary nailing of the femur is accepted as an effective method of treatment for patients with femoral shaft fractures. Reliable outcomes have been reported with antegrade reamed and unreamed nailing techniques.4,30,31 Union rates have been reported to be between 70 and 99 percent, and the rate of malunion has been low.4,30,31 The reported disadvantages of antegrade nailing of the femur include the limited application of this method to ipsilateral femoral neck and femoral shaft fractures, the risk of heterotopic ossification about the hip,5,17,24 the risk of injury to the pudendal nerve.3,12 and the need for a fracture table. Retrograde insertion of femoral nails has become an attractive alternative for some surgeons as a method that can minimize or eliminate some of the shortcomings of an antegrade nailing technique.18-20,22

Initial reports of retrograde nailing of the femur for fractures of the femoral shaft were for the treatment of ipsilateral femoral shaft and femoral neck fractures and used a medial femoral condylar starting point.25 The indications for retrograde nailing of femoral shaft fractures expanded to include patients who were pregnant, who had ipsilateral pelvic or acetabular injuries, multiple fractures, or polytrauma,22 traumatic arthrotomy of the knee, or through-the-knee traumatic amputation.20 Good results for the routine use of retrograde femoral nailing through an intercondylar intraarticular starting point have recently been reported.8,11,19

The goal of the current study was to compare the results and complications of retrograde femoral intramedullary nailing in patients with femoral shaft fractures with those of antegrade femoral intramedullary nailing.

MATERIALS AND METHODS

Inclusion Criteria

Between January 1, 1993 and June 30, 1997, 283 consecutive adult patients with 293 femoral shaft fractures were treated with retrograde or antegrade intramedullary nailing at Tampa General Hospital, Tampa, FL (a Level I trauma center) by the same three surgeons. These patients were identified from a prospectively obtained orthopaedic trauma database. All retrograde nails were inserted through an intraarticular intercondylar starting point, and all antegrade nails were placed through the piriformis fossa. The choice of whether the patient underwent antegrade or retrograde intramedullary nailing was at the discretion of the attending surgeon. Although there was no absolute criteria for selecting the method of nailing, retrograde nailing generally was preferred for patients with fractures of the distal femoral shaft, with additional fractures of the ipsilateral femoral neck, intertrochanteric region, acetabulum, or tibial shaft, or with multiple injuries or obesity. One hundred forty nails were inserted in a retrograde fashion. Six patients were excluded from this total: three who had an early amputation, two paraplegics, and one who died early in the postoperative period. One hundred fifty-three nails were inserted in an antegrade fashion. Six of these patients were excluded: two who had a fracture through abnormal bone owing to metastatic carcinoma, two paraplegics, and two who died early in the postoperative period.

Data Collection

Patients were followed up at regular intervals with clinical evaluations and radiographs. Charts and radiographs were retrospectively reviewed. Fractures were considered healed when patients could bear weight without thigh pain and radiographs showed bridging callus across three cortices. Delayed union was defined by the absence of progressive fracture healing for three consecutive months. Nonunion was defined by the absence of healing after six months. The presence or absence of hip or knee pain at the last clinical follow-up examination was recorded. Measurements of hip and knee range of motion were not consistently documented and therefore were not analyzed. The minimum sufficient followup was defined as clinical and radiographic follow-up to fracture healing or to establishment of a nonunion or delayed union. Attempts were made to contact all patients who, after review of their charts and radiographs, had insufficient follow-up. Multiple methods were used to contact these patients. As suggested by Smith and Watts,23 we used telephone numbers in the hospital and trauma databases, correspondence sent by the U.S. mail, and an Internet search to obtain up-to-date telephone numbers. When these methods failed, the Equifax credit reporting agency (Atlanta, GA, USA.) was used to obtain contact information. Patients who were contacted by these means and returned for examination were also included. Sufficient follow-up was ultimately obtained for 191 patients with 198 fractures of the femoral shaft (67 percent). One hundred four of these femoral shaft fractures were treated with retrograde nails (Group R) and ninety-four were treated with antegrade nails (Group A). Five patients with bilateral fractures were treated with bilateral retrograde femoral nails, and two patients with bilateral fractures were treated with retrograde nails on one side and antegrade nails on the other.

Angular deformity was evaluated on anteroposterior and lateral radiographs at the time of fracture union. Limb length and rotational deformity were evaluated clinically. Malunion was defined as more than 5 degrees of angular deformity in the sagittal or coronal planes, malrotation of at least 10 degrees, or more than one centimeter of limb length discrepancy as compared with the uninjured side. Heterotopic ossification was graded according to the classification of Brooker et al.2

Patient Characteristics

The average follow-up time and age and gender distribution were similar between Group R and Group A (Table 1). The mechanisms of injury for patients in Group R and Group A are shown in Table 2. In Group R, seventy-five (76 percent) patients had 173 associated fractures, and in Group A, forty-four (47 percent) patients had sixty associated fractures (Table 3). In Group R, there were ten patients with a second ipsilateral fracture of the femur (seven intertrochanteric fractures, two distal unicondylar fractures, and one femoral neck fracture). In each of these ten patients, the second fracture was treated separately with appropriate internal fixation. Also in Group R, there was one case of a supracondylar fracture with diaphyseal extension that was treated with retrograde nailing. In Group A, there was one case of an ipsilateral femoral neck fracture that was missed preoperatively and treated with screw fixation around the antegrade nail.

TABLE 1
TABLE 1:
Patient Data
TABLE 2
TABLE 2:
Mechanism of Injury
TABLE 3
TABLE 3:
Associated Fractures

Fracture Characteristics

Open fractures were graded according to the classification of Gustilo et al.,10 and fracture comminution was evaluated according to the system of Winquist et al.30 and the Orthopaedic Trauma Association classification system1 (Table 1). There were similar percentages of open fractures and stable fracture patterns (Winquist Types I and II) in Group R and Group A. There were thirty-one fractures of the proximal one third of the femoral shaft, 111 fractures of the middle one third, and fifty-six fractures of the distal one third. Ten femurs with fractures of the proximal one third (32 percent) were treated with retrograde nails, and twenty-one (68 percent) were treated with antegrade nails. Fifty-two femurs with fractures of the middle one third (47 percent) were treated with retrograde nails, and fifty-nine (53 percent) were treated with antegrade nails. Forty-two (75 percent) of the femurs with fractures of the distal one third were treated with retrograde nails, and fourteen (25 percent) were treated with antegrade nails.

Implant Characteristics

The pattern of nail interlocking was based on the fracture pattern and the surgeon's preference. Static interlocking was used on seventy-five (72 percent) femurs treated in Group R and for seventy-six (81 percent) in Group A. Dynamic interlocking was used on twenty-nine (28 percent) femurs treated in Group R and eighteen (19 percent) in Group A. Eighty-seven nails from Group R were inserted without reaming; eleven were inserted with reaming; and for six, the method was unknown. Fiftytwo nails from Group A were inserted without reaming; thirty-nine were inserted with reaming; and for three, the method was unknown. A titanium femoral nail (Synthes USA, Paoli, PA, USA.) was used in 169 cases (ninetynine retrograde and seventy antegrade), a universal femoral nail (Synthes USA) in nineteen cases (one retrograde and eighteen antegrade), a Russell-Taylor femoral nail (Smith and Nephew Richards Inc., Memphis, TN, USA.) in four cases (four antegrade), a Trimax femoral nail (Smith and Nephew Richards Inc.) in two cases (two antegrade), and a Biomet femoral nail (Biomet Inc., Warsaw, IN, USA.) in three cases (three retrograde). The average nail diameter was 10.3 millimeters for Group R (range 9 to 12 millimeters) and 10.5 for Group A (range 9 to 15 millimeters).

Statistical Analysis

Chi-squared analysis was performed to evaluate differences between Group R and Group A in gender, degree of soft tissue injury (i.e., open or closed), fracture comminution, mode of interlocking (i.e., static or dynamic), reaming, presence of heterotopic bone formation, hip pain, knee pain, reoperation, healing after the index procedure, and malunion. To evaluate differences between the two groups in age, nail diameter, length of hospital stay, surgical time, operating room time, and estimated blood loss, t tests were used. Wilcoxon rank sums were used to evaluate differences in American Society of Anesthesia (ASA) score6 and fracture location. The study had a power of more than 0.80 to detect a difference of 15 percent or more in dichotomous study outcomes.

RESULTS

Patient, Fracture, and Implant Characteristics

There were no statistical differences between the patients in Groups R and A with respect to age, gender, number of open and closed fractures, number of stable (Winquist I and II) and unstable (Winquist III and IV) fractures, mode of interlocking (i.e., static or dynamic), or nail diameter (P > 0.05). There were more fractures of the proximal one third of the femur in Group A and more of the distal one third in Group R (P < 0.001). More nails were inserted with reaming in Group A (P < 0.001). The mean ASA score for patients in Group R was higher (2.3; range 1 to 4) than for patients in Group A (1.9; range 1 to 3) (P < 0.01). The mean length of hospital stay was higher for patients in Group R (fifteen days; range 1 to 86 days) than for patients in Group A (nine days; range 2 to 43 days) (P < 0.001).

Healing

There was no statistically significant difference in healing after the index procedure between the two groups (P > 0.05). Healing after the index procedure occurred in ninety-one (88 percent) of the femurs from Group R and in eighty-four (89 percent) from Group A. Delayed union occurred in seven femurs (7 percent) from Group R. Three of the seven were dynamized (two of these ultimately healed and one patient was lost to follow-up); two healed without further intervention; one autodynamized by fracturing interlocking screws and ultimately healed; and in one patient, an above-the-knee amputation was performed to treat chronic osteomyelitis of the proximal tibia. Delayed union occurred in four femurs (4 percent) from Group A. Two of the four healed without further intervention; one was dynamized and healed; and one healed after treatment with a stimulator. Nonunion occurred in six femurs (6 percent) from Group R. Five of the six were treated with exchange reamed intramedullary nailing (all healed), and one was dynamized (patient lost to follow-up). Nonunion occurred in six femurs (6 percent) from Group A. Five of the six were treated with exchange reamed intramedullary nailing (all healed), and one was lost to follow-up. Healing eventually occurred in 101 femurs (97 percent) from Group R and ninety-three femurs (99 percent) from Group A.

Malunion

There was no statistically significant difference in the incidence of malunion between the two groups (P > 0.05). In Group R, there were eleven malunions (11 percent). Five of the malunited femurs had fractures of the proximal one third of the femoral shaft; three had fractures of the middle one third; and three had fractures of the distal one third. Five of the malunited femurs had varus angulation (range 6 to 12 degrees); one had varus and apex anterior angulation (7 degrees in each plane); one had valgus angulation (6 degrees); one had apex posterior angulation (6 degrees); and three had external rotation (10 degrees in each case). In Group A, there were twelve malunions (13 percent). Six of the malunited femurs had fractures of the proximal one third of the femoral shaft; three had fractures of the middle one third; and three had fractures of the distal one third. Four of the malunited femurs had varus angulation (range 6 to 9 degrees); four had apex anterior angulation (range 7 to 15 degrees); one had varus and apex anterior angulation (7 degrees varus and 9 degrees flexion); one had apex posterior angulation (6 degrees); one had external rotation (10 degrees); and one had internal rotation (20 degrees).

Knee and Hip Pain

At the last clinical follow-up examination, knee pain was present in thirty-four (33 percent) patients from Group R and eleven (12 percent) patients from Group A. To determine more accurately when knee pain was related solely to intramedullary nail placement, a subset of patients who had no known injuries involving the knee joint ipsilateral to their femoral shaft fracture and who did not have ipsilateral placement of a tibial nail were examined. This subset included seventy patients from Group R (average follow-up ninety-six weeks; range 23 to 239 weeks) and seventy-eight from Group A (average follow-up eighty-nine weeks; range 20 to 257 weeks). In these subsets, knee pain was present in a significantly larger number of patients from Group R (twenty-five; 36 percent) than from Group A (seven; 9 percent) (P < 0.001). When patients with known injuries to the hip were excluded, ten (10 percent) of eighty-five patients from Group A and three (4 percent) of eighty-five from Group R had hip pain (P < 0.05).

Complications

There was no difference in the number of reoperations for reasons other than delayed or nonunion between the two groups (P > 0.05). In Group R, there were seventeen (16 percent) reoperations. These included removal of painful interlocking screws in six patients, knee arthroscopy in four patients, knee manipulations under anesthesia in two patients, nail revision in two patients (one for rotational malalignment and one for nail fracture), nail removal after migration into the knee in one patient (Fig. 1), and removal of heterotopic bone from the postpatellar region in one patient. One of the patients who underwent knee arthroscopy had a coronal fracture of the medial femoral condyle treated with open reduction and internal screw fixation. At arthroscopy (seventeen months after injury), these screws were removed. The other three patients who had knee arthroscopy had no other fractures involving the knee. The time intervals from injury to knee arthroscopy were seven, fourteen, and thirty-three months. For each of these three patients, the surgical indication was pain or stiffness; the procedure was performed at another institution; and intraoperative findings included medial and lateral compartment fibrous bands and, in one patient, an attenuated anterior cruciate ligament. The indication for knee manipulation was stiffness in both patients (1.5 and five months after injury, respectively). One of these patients had an ipsilateral patella fracture. The patient with nail migration into the knee joint had Winquist Type IV fracture comminution and was treated with a retrograde nail statically locked using two proximal and two distal interlocking screws. The distal interlocking screws fractured, and the nail migrated into the knee joint and required surgical removal. In Group A, there were sixteen (17 percent) reoperations for reasons other than delayed or nonunion. These included removal of painful interlocking screws in five patients, nail revision for shortening in four patients, nail removal in four patients, exchange of a loose interlocking screw in one patient, revision of a bent nail in one patient, and removal of an exostosis in one patient.

FIGURE. 1
FIGURE. 1:
Migration of a retrograde nail onto the knee joint after fracture of distal interlocking screws.A: Anteroposterior view. B: Lateral view. C: Merchant view.

Placement of one retrograde nail was aborted because of proximal fracture extension that was present but unrecognized preoperatively. Antegrade nailing was subsequently performed. In Group R, there were nine (9 percent) femurs with fractured interlocking screws. Two of these patients had subsequent migration of the nail into the knee joint (Fig. 1). In Group A, there were four (4 percent) patients with fracture of interlocking screws, two (2 percent) patients with pulmonary emboli, one (1 percent) patient with fat embolus syndrome, one (1 percent) patient with a deep venous thrombosis, and one (1 percent) patient with a transient pudendal nerve injury.

Heterotopic ossification about the hip was observed only in Group A patients (P < 0.001). Twenty-four hips (26 percent) from Group A patients had radiographic evidence of heterotopic bone formation. The heterotopic ossification was Grade I in fifteen (16 percent) cases, Grade II in five (5 percent) cases, and Grade III in four (4 percent) cases. Sixteen (27 percent) of sixty-four male patients had heterotopic ossification, and eight (28 percent) of twenty-nine female patients had heterotopic ossification. Four (17 percent) of the twenty-four patients with heterotopic ossification had a closed head injury. None of the patients from Group R had evidence of heterotopic ossification about the hip, but one patient from Group R had heterotopic bone in the postpatellar region that was painful and required operative excision.

Surgical Time, Operating Room Time, and Blood Loss

Many patients from both groups had injuries other than femoral shaft fractures that required surgical procedures at the same time as intramedullary nail placement. To evaluate accurately the surgical time, total operating room time, and estimated blood loss associated with intramedullary nailing, only the cases in which femoral nailing was the sole procedure performed were included in this analysis. Open fractures were also excluded because irrigation and debridement effects the surgical and operating room times and the estimated blood loss. Twenty-three patients from Group R and fifty-nine from Group A met these criteria. These patients were further divided into reamed (Group R-R and Group A-R) and unreamed (Group R-UR and Group A-UR) subgroups because reaming increases operative time and blood loss (Table 4). There was a significantly higher estimated blood loss in Group A-UR (207 milliliters) than in Group R-UR (125 milliliters) (P < 0.01). The average surgical time was similar between the two unreamed subgroups (Group R-UR, eighty-six minutes; Group A-UR, ninetyfour minutes) (P > 0.05). The total operating room time was higher in Group A-UR (162 minutes) than in Group R-UR (146 minutes), but with the number of cases available in these subgroups, this difference was not statistically significant (P > 0.05). Because there were only two patients in the Group R-R subgroup, a comparison between Group R-R and Group A-R with respect to surgical and operating room times and estimated blood loss was not useful.

TABLE 4
TABLE 4:
Surgical Time, Operating Room Time, and Estimated Blood Loss

DISCUSSION

In certain clinical situations, retrograde nailing is an attractive alternative to antegrade nailing for the treatment of patients with femoral shaft fractures. These include patients with ipsilateral femoral neck or intertrochanteric fractures, in which a retrograde approach makes treatment of the proximal fractures easier. Retrograde nailing does not violate the surgical approach for repair of ipsilateral acetabular or pelvic injuries. In obese patients or patients with multiple fractures, retrograde nailing is technically easier, allows easier access to other fractures, and obviates the need for traction on a fracture table. Retrograde nailing also has the potential to reduce some of the shortcomings of antegrade nailing, such as heterotopic ossification about the hip, injury to the pudendal nerve, and hip pain. The existing literature suggests that the results of retrograde nailing of femoral shaft fractures approaches the good results of historical controls treated with antegrade nailing.

Retrograde nailing for the treatment of femoral shaft fractures was first reported by Swiontkowski et al.25 They treated seven ipsilateral femoral neck and femoral shaft fractures with cannulated screws and retrograde intramedullary nails, respectively. A medial condylar starting point was used and all femurs were reamed. Fracture comminution was a contraindication for intramedullary nailing. Three (43 percent) of the seven patients required adjuvant hip spica casts because of varus angulation of more than 5 degrees. One patient required knee manipulation under anesthesia for stiffness. All fractures united after the index procedure. The authors advocated treating fractures of the femoral shaft with retrograde nails when there was an ipsilateral femoral neck fracture. They warned that these fractures must be carefully watched for varus angulation.

Sanders et al22 expanded the indications for retrograde nailing of femoral shaft fractures. In addition to ipsilateral femoral neck and femoral shaft fractures as inclusion criteria, pregnancy, ipsilateral pelvic or acetabular fractures, polytrauma, and polyfracture were criteria for inclusion in their prospective study. Twenty-nine fractures met the inclusion criteria. A medial extraarticular starting point was used for nail insertion. Complications included four malreductions (14 percent): three cases of crack propagation at the insertion point (10 percent) and one aborted nailing (1 percent). Twenty-three (92 percent) of twenty-five fractures healed with the index procedure. The authors concluded that retrograde reamed femoral nailing is a suitable alternative to antegrade nailing for the treatment of femoral shaft fractures when proximal access is neither possible nor desirable.

Patterson et al20 further expanded the relative indications for retrograde femoral nailing of femoral shaft fractures to include associated traumatic arthrotomy of the knee and through-the-knee traumatic amputation. Seventeen retrograde femoral nailings were retrospectively reviewed. Three nailings were performed through an extraarticular medial femoral condyle starting point, and fourteen nailings were through an intraarticular intercondylar notch starting point. There were ten complications in six of fourteen patients with follow-up. These included five nonunions (36 percent), three implant failures (21 percent), one deep infection (7 percent), and one shortening (7 percent). The authors recommended the limited use of retrograde nailing in properly selected cases.

Moed and Watson18 found an apparently higher prevalence of nonunion in their series of twenty-two femoral shaft fractures in patients with multiple injuries treated with retrograde intramedullary nailing without reaming, as compared with historical union rates reported with antegrade nailing. All nails were inserted through an intraarticular intercondylar starting point. Nineteen (86 percent) of the twenty-two fractures united after the index procedure. Three fractures required exchange antegrade femoral nailing and subsequently united. There was one malunion (5 percent). Six (27 percent) patients had persistent knee pain at a minimum of one year of follow-up. The authors concluded that, because of the low union rate, this method of fixation should not be used for the routine treatment of isolated fractures of the femoral shaft.

Herscovici and Whiteman11 reported results of forty-five femoral shaft fractures treated with retrograde unreamed intramedullary femoral nailing. They used an intraarticular intercondylar starting point. Forty-three healed after the index procedure (96 percent). Complications included two nonunions (4 percent), one malreduction (2 percent), and one case of reflex sympathetic dystrophy (2 percent). Thirteen (29 percent) patients were found to have knee pain at an average of two-years of follow-up. They concluded that retrograde femoral nailing is a safe method of fracture management and should be considered in multiple trauma or morbidly obese patients.

Gregory et al8 described twenty patients with twenty-two ipsilateral femoral shaft and tibial fractures treated with retrograde femoral nailing and unreamed tibial nailing. Twelve femoral nails were placed through the medial femoral condyle and ten through the intercondylar notch. Seventeen (85 percent) of twenty femurs healed after the index procedure. There were no femoral malunions. Five additional operative procedures were required. No significant knee problems related to the femoral nailing technique were identified. The authors advocated retrograde femoral nailing and antegrade tibial nailing for the treatment of floating knees.

Moed et al19 reported a second series of thirty-five femoral shaft fractures treated with retrograde intramedullary femoral nailing through an intercondylar starting point. They used a protocol that included planned dynamization for fractures showing minimal healing after six to twelve weeks. There were two nonunions (6 percent), no malunions, and knee pain was described as minimal. They concluded that unreamed retrograde nailing is a safe and beneficial fracture fixation method.

The current study directly compared retrograde with antegrade nailing of femoral shaft fractures in concurrent cohorts. There were no clinically meaningful nor statistically significant differences in healing rates between the two groups despite a greater number of antegrade nails inserted with reaming. After the index procedure, 87 percent of the retrograde and 89 percent of the antegrade nail-treated femurs healed. These results are consistent with the previously reported healing rates for femoral shaft fractures treated with retrograde intramedullary nails. The union rate in the current series for antegrade nailing of femoral shaft fractures is lower than what has been previously reported.4,30,31 The reason for this difference is not clear. A difference in fracture severity may explain the differences in healing between the current study and the report by Winquist et al.30 There were more severe open fractures (9 percent Grade III) and unstable fractures (40 percent) in the current study than in the series of Winquist et al.30 (0.4 percent open Grade III and 17 percent unstable fractures). A lower percentage of reamed nails in this series than in the reports by Wiss et al31 and Brumback et al.4 may explain the observed difference in healing rates.

We found no differences in malunion rates between the retrograde and antegrade nail treated groups. Eleven percent of the retrograde and 13 percent of the antegrade nail-treated femurs were malunited. It is difficult to compare malunion rates accurately between different reports because of variations in the definition of malunion. The two most common definitions of angular malalignment after fracture of the femoral shaft are more than or equal to 5 degrees18,25 and more than 5 degrees.8,14,16,19,22 Other reports define angular deformity as more than or equal to 10 degrees31 or as more than 15 degrees.28 Some reports divide malunion into multiple groups based on the severity of angular deformity.13,15,26,27,30 Other reports do not define malunion.4,7,9,11,20,21,29 There are similar variations in the definition of malunion with regard to length and rotation. The reported limit of unacceptable shortening of the femur after intramedullary nailing ranges from more than one centimeter to more than 2.5 centimeters.8,16,18,19,26,28,30 The reported limit of acceptable rotational deformities ranges from more than 5 degrees to more than 20 degrees.9,13,16,18,19,26,29-31 We used more than 5 degrees of angular deformity, more than one centimeter of length deformity, or more than or equal to 10 degrees of rotational deformity to define malunion of femoral shaft fractures because these criteria seem to be the most commonly cited definitions in the existing literature. We propose that these definitions be used in future reports to make future comparisons easier and more consistent.

In the current series, patients with known knee injuries or ipsilateral tibial nails were excluded from the analysis of knee pain. A significantly higher percentage of patients from Group R had knee pain (36 percent) than did those from Group A (9 percent) (P < 0.001). At the last follow-up examination, 24 percent of these patients had knee pain. These results are consistent with the reports of retrograde nailing of femoral shaft fractures by Moed and Watson18 (27 percent incidence of knee pain) and Herscovici and Whiteman11 (29 percent incidence of knee pain). The cause, severity, and functional significance of this knee pain deserves further investigation.

The reoperation rate was 26 percent in Group R and 24 percent in Group A. Seven of the reoperations in the retrograde group were for reasons that may be specific to the retrograde technique. Four patients had knee arthroscopy (one had an ipsilateral coronal fracture of the femoral condyle), two had knee manipulations (one had an ipsilateral patella fracture), and one patient underwent removal of a nail that migrated into the knee joint after fracture of two distal interlocking screws.

To allow comparison of the surgical and operating room times and estimated blood loss, only cases in which intramedullary nailing was the sole procedure performed were included in this analysis. Because reaming affects the operative time and estimated blood loss, the retrograde and antegrade groups were divided into reamed and unreamed subgroups. Only two retrograde nails in this subgroup were inserted as reamed. Therefore, an accurate comparison of retrograde reamed to antegrade reamed nails cannot be made. The average surgical time, operating room time, and estimated blood loss for the R-UR subgroup were less than those for the A-UR subgroup by 9 percent, 10 percent, and 40 percent, respectively. The lower average estimated blood loss seen with retrograde nailing was statistically significant and likely represents the use of a tourniquet during retrograde nail insertion.

The rate of heterotopic ossification about the hip in this study was less than in previous reports. Marks et al.17 found a 68 percent incidence of heterotopic ossification about the hip after antegrade intramedullary nailing of the femur. Twenty percent of their patients had moderate to severe heterotopic ossification and 5 percent required reoperation for decreased hip range of motion or clinical symptoms. Brumback et al.5 prospectively reviewed 100 consecutive femur fractures treated with antegrade intramedullary nails and found a 60 percent incidence of heterotopic ossification. Moderate heterotopic ossification was found in 15 percent, and severe heterotopic ossification was found in 11 percent. Steinberg and Hubbard24 reviewed 118 cases of antegrade femoral nailing and found a 55 percent incidence of heterotopic ossification among patients with trauma (27 percent was graded severe). Identified risk factors included prolonged intubation, brain injury, male gender, and long delay for surgery. In the current study, there was a 24 percent incidence of heterotopic ossification about the hip after antegrade femoral nailing. Male (26 percent) and female (28 percent) patients had similar rates of heterotopic ossification formation. Closed head injury was present in 16 percent of patients with heterotopic ossification and 20 percent of the entire antegrade group.

One of the limitations of this study is its retrospective design. All patients were prospectively entered into an orthopaedic trauma registry, and all procedures were performed during the same period at a single institution by the same set of senior surgeons. Despite intense efforts, a large number of patients were lost to follow-up. Unfortunately, this is a problem inherent in most intermediate- term follow-up studies involving trauma patients. Many of the methods suggested by Smith and Watts23 were used to contact patients, including an Internet search and the use of the Equifax credit reporting agency. Despite these efforts, 33 percent of patients were lost to follow-up. We think that the effect of this transfer bias was minimized in this study because of its comparative nature. The two groups were remarkably similar in characteristics known to affect fracture healing, including age, gender, the number of open fractures, and the degree of comminution. However, there were several differences between Group R and Group A. Patients in Group R had more associated fractures, a higher average ASA score, and a longer length of hospital stay. Because one of the relative indications for retrograde femoral nailing is the treatment of patients with multiple injuries, these differences were predictable. Also, there were more nails inserted without reaming in the retrograde group. Despite these differences, the union rates between the two groups were similar.

We conclude that retrograde nailing of femoral shaft fractures produces healing rates similar to those of antegrade nailing. The benefits of retrograde nailing technique include decreased blood loss (when inserted without reaming), decreased heterotopic bone formation about the hip, decreased hip pain, and more convenient access to other injuries. These benefits must be weighed against the increased incidence of complications related to the knee seen after retrograde femoral nailing.

REFERENCES

1. Anonymous. Fracture and dislocation compendium. Orthopaedic Trauma Association Committee for Coding and Classification. J Orthop Trauma. 1996;10:1-153.
2. Brooker AF, Bowerman JW, Robinson RA, et al. Ectopic ossification following total hip replacement: incidence and a method of classification. J Bone Joint Surg [Am]. 1973;55:1629-1632.
3. Brumback RJ, Ellison TS, Molligan H, et al. Pudendal nerve palsy complicating intramedullary nailing of the femur. J Bone Joint Surg [Am]. 1992;74:1450-1455.
4. Brumback RJ, Uwagie-Ero S, Lakatos RP, et al. Intramedullary nailing of femoral shaft fractures. Part II: fracture healing with static interlocking fixation. J Bone Joint Surg [Am]. 1988;70:1453-1462.
5. Brumback RJ, Wells JD, Lakatos R, et al. Heterotopic ossification about the hip after intramedullary nailing for fractures of the femur. J Bone Joint Surg [Am]. 1990;72:1067-1073.
6. Dripps RD, LaMont A, Eckerenhoff JE. The role of anesthesia in surgical mortality. JAMA. 1961;178:261-266.
7. Gibbons CL, Gregg-Smith SJ, Carrell TW, et al. Use of the Russell-Taylor reconstruction nail in femoral shaft fractures. Injury. 1995;26:389-392.
8. Gregory P, DiCicco J, Karpik K, et al. Ipsilateral fractures of the femur and tibia: treatment with retrograde femoral nailing and unreamed tibial nailing. J Orthop Trauma. 1996;10:309-316.
9. Grover J, Wiss DA. A prospective study of fractures of the femoral shaft treated with a static, intramedullary, interlocking nail comparing one versus two distal screws. Orthop Clin North Am. 1995;26:139-146.
10. Gustilo RB, Mendoza RM, Williams DN. Problems in the management of type III (severe) open fractures: a new classification of Type III open fractures. J Trauma. 1984;24:742-746.
11. Herscovici D Jr, Whiteman KW. Retrograde nailing of the femur using an intercondylar approach. Clin Orthop. 1996;332:98-104.
12. Kao JT, Burton D, Comstock C, et al. Pudendal nerve palsy after femoral intramedullary nailing. J Orthop Trauma. 1993;7:58-63.
13. Karpos PA, McFerran MA, Johnson KD. Intramedullary nailing of acute femoral shaft fractures using manual traction without a fracture table. J Orthop Trauma. 1995;9:57-62.
14. Kempf I, Grosse A, Beck G. Closed locked intramedullary nailing. J Bone Joint Surg [Am]. 1985;67:709-720.
15. Klemm KW, Borner M. Interlocking nailing of complex fractures of the femur and tibia. Clin Orthop. 1986;212:89-100.
16. Majkowski RS, Baker AS. Interlocking nails for femoral fractures: an initial experience. Injury. 1991;22:93-96.
17. Marks PH, Paley D, Kellam JF. Heterotopic ossification around the hip with intramedullary nailing of the femur. J Trauma. 1988;28:1207-1213.
18. Moed BR, Watson JT. Retrograde intramedullary nailing, without reaming, of fractures of the femoral shaft in multiply injured patients. J Bone Joint Surg [Am]. 1995;77:1520-1527.
19. Moed BR, Watson JT, Cramer KE, et al. Unreamed retrograde intramedullary nailing of fractures of the femoral shaft. J Orthop Trauma. 1998;12:334-342.
20. Patterson BM, Routt ML Jr, Benirschke SK, et al. Retrograde nailing of femoral shaft fractures. J Trauma. 1995;38:38-43.
21. Reynders P, Broos PL. Unreamed intramedullary nailing of femoral shaft fractures using a traction device. Injury. 1998;29:81-84.
22. Sanders R, Koval KJ, DiPasquale T, et al. Retrograde reamed femoral nailing. J Orthop Trauma. 1993;7:293-302.
23. Smith JS, Watts HG. Methods for locating missing patients for the purpose of long-term clinical studies. J Bone Joint Surg [Am]. 1998;80:431-438.
24. Steinberg GG, Hubbard C. Heterotopic ossification after femoral intramedullary rodding. J Orthop Trauma. 1993;7:536-542.
25. Swiontkowski MF, Hansen ST, Kellam J. Ipsilateral fractures of the femoral neck and shaft. J Bone Joint Surg [Am]. 1984;66:260-268.
26. Thoresen BO, Alho A, Ekeland A, et al. Interlocking intramedullary nailing in femoral shaft fractures. A report of forty-eight cases. J Bone Joint Surg [Am]. 1985;67:1313-1320.
27. Tornetta P, Tiburzi D. The treatment of femoral shaft fractures using intramedullary interlocked nails with and without intramedullary reaming: a preliminary report. J Orthop Trauma. 1997;11:89-92.
28. Webb LX, Gristina AG, Fowler HL. Unstable femoral shaft fractures: a comparison of interlocking nailing versus traction and casting methods. J Orthop Trauma. 1988;2:10-12.
29. Wilson-MacDonald J, Owen JW, Lowdon I, et al. Early experience with closed interlocking medullary nailing of the femur. Injury. 1987;18:390-395.
30. Winquist RA, Hansen ST, Clawson DK. Closed intramedullary nailing of femoral fractures. J Bone Joint Surg [Am]. 1984;66:529-539.
31. Wiss DA, Brien WW, Stetson WB. Interlocked nailing for treatment of segmental fractures of the femur. J Bone Joint Surg [Am]. 1990;72:724-728.
Keywords:

Femur fracture; Antegrade; Retrograde; Intramedullary nailing.

© 2008 Lippincott Williams & Wilkins, Inc.