Prospective randomized study of reamed versus unreamed femoral intramedullary nailing: an assessment of procedures.

OBJECTIVE

To determine whether the procedure of unreamed femoral nailing is simpler, faster, and safer than reamed femoral intramedullary nailing.

DESIGN

Prospective randomized.

SETTING/PARTICIPANTS: One hundred femoral shaft fractures without significant concomitant injuries admitted to an academic Level 1 urban trauma center.

INTERVENTION

Stabilization of the femoral shaft fracture using a reamed or unreamed technique.

OUTCOME MEASUREMENTS

The surgical time, estimated blood loss, fluoroscopy time, and perioperative complications were prospectively recorded.

RESULTS

One hundred patients with 100 femoral shaft fractures were correctly prospectively randomized to the study. Thirty-seven patients received reamed and sixty-three patients received unreamed nails. All nails were interlocked proximally and distally. The average surgical time for the reamed nail group was 138 minutes and for unreamed nail group was 108 minutes (p = 0.012). The estimated blood loss for the reamed nail group was 278 milliliters and for the unreamed nail group 186 milliliters (p = 0.034). Reamed intramedullary nailing required an average of 4.72 minutes, whereas unreamed nailing required 4.29 minutes of fluoroscopy time. Seven perioperative complications occurred in the reamed nail group and eighteen in the unreamed nail group. Two patients in the unreamed group required an early secondary procedure. Iatrogenic comminution of the fracture site occurred during three reamed and six unreamed intramedullary nailings. Reaming of the canal was required before the successful placement of three nails in the unreamed group because of canal/nail diameter mismatch.

CONCLUSIONS

Unreamed femoral intramedullary nailing involves fewer steps and is significantly faster with less intraoperative blood loss than reamed intramedullary nailing. The unreamed technique, however, was associated with a higher incidence of perioperative complications, although the difference was not statistically significant (p = 0.5).