OBJECTIVES:
To compare union rates and complications of retrograde intramedullary nailing of femoral shaft fractures with those of antegrade intramedullary nailing.

DESIGN:
Retrospective.

SETTING:
Level I trauma center.

PATIENTS:
Two hundred eighty-three consecutive adult patients with 293 fractures of the femoral shaft who underwent stabilization with antegrade or retrograde inserted femoral nails were studied. There were 140 retrograde nails and 153 antegrade nails. Twelve fractures in twelve patients were excluded (three in patients who died early in the postoperative period, three in patients because of early amputation, four in patients who were paraplegic, and two in patients who fractured through abnormal bone owing to metastatic carcinoma), leaving 134 fractures treated with retrograde nails and 147 treated with antegrade nails. One hundred four femurs treated with retrograde nails (Group R) and ninety-four femurs treated with antegrade nails (Group A) had sufficient follow-up and served as the two study groups. The average clinical follow-up was twenty-three months (range 6 to 66 months) for Group R and twenty-three months (range 5 to 64 months) for Group A. Both groups were comparable with regard to age, gender, number of open fractures, degree of comminution, mode of interlocking (i.e., static or dynamic), and nail diameter (p > 0.05).

INTERVENTION:
Retrograde intramedullary nails were inserted through the intercondylar notch of the knee, and antegrade nails were inserted through the pirformis fossa using standard techniques.

MAIN OUTCOME MEASURES:
Union, delayed union, nonunion, malunion, and complication rates.

RESULTS:
After the index procedure there were no significant differences in healing or incidence of malunion between Group R and Group A (p > 0.05). Healing after the index procedure occurred in ninety-one (88 percent) of the femurs in Group R and in eighty-four (89 percent) of the femurs in Group A. In Group R, there were seven delayed unions (7 percent) and six nonunions (6 percent). In Group A, there were four delayed unions (4 percent) and six nonunions (6 percent). Healing ultimately occurred in 100 (96 percent) femurs from Group R and in ninety-three (99 percent) femurs from Group A. In Group R, there were eleven malunions (11 percent), and in Group A, there were twelve malunions (13 percent). When patients with ipsilateral knee injuries were excluded, the incidence of knee pain was significantly greater for Group R patients (36 percent) than for Group A patients (9 percent) (p < 0.001). When patients with ipsilateral hip injuries were excluded, the incidence of hip pain was significantly greater for Group A patients (10 percent) than for Group R patients (4 percent) (p < 0.05).

CONCLUSIONS:
Retrograde and antegrade nailing techniques provided similar results in union and malunion rates. There were more complications related to the knee after retrograde nailing and more complications related to the hip after antegrade nailing.