Session V - Femur

Anterograde vs. Retrograde Reamed Femoral Nailing: A Prospective Randomized Trial

Paul Tornetta, III, MD; Douglas Tiburzi, MD, Boston Medical Center, Kings County Hospital, and Brookdale Hospital, Boston, MA

Purpose: To compare anterograde and retrograde reamed femoral nailing in a prospective and randomized study.

Methods: A consecutive group of 68 patients with 69 femoral shaft fractures (OTA 32) were entered into a prospective, randomized study of anterograde (AG) vs. retrograde (RG) reamed nailing. Fractures within 5 cm of the lesser trochanter or 5 cm of the knee joint were excluded. There were 7 open, 21 gun shot, and 41 closed fractures. 36 were stable and 33 unstable patterns. AG nailing was performed on a fracture table, and RG nailing was performed on a radiolucent table via a medial paratendinous approach. All nails were titanium, placed after reaming to 1.5 mm over cortical chatter, and statically locked. Operative data was gathered prospectively. CT scanograms were obtained postoperatively for unstable (Winquist 3,4) fractures to accurately determine length and rotational deformity. Intra-operative and postoperative complications were noted prospectively. Union was graded both radiographically and clinically.

Results: There were 38 AG and 31 RG nailings in the study. There was no difference between the groups in age (31 v. 33), ISS score (12.4 v 12.5 range 4 - 42), nail width (11.6 v. 11.4), blood loss (215 v. 253), fracture stability, fracture location (prox, mid, dist), or associated injuries. There was a statistical difference (p < 0.05) in operative time AG = 116 ± 40 and RG = 147 ± 58. When set-up time was included, there was no difference in total time. The postoperative reduction revealed longitudinal angulatory deformity 5° in 1/39 AG and 3/31 RG nails. A rotational deformity of > 10° as seen on CT rotation studies occurred in 3/18 (16 %) of the AG and in 5/15 (33 %) of RG nailings. Shortening occurred after 5/15 unstable RG and none of the unstable AG nailings. The average shortening in the RG group was 1.2 cm (1 - 3). Return to the OR for re-locking was necessary for the patient nailed 3 cm short. Complaints of knee pain were found in 25/31 RG and 13/38 AG patients in the perioperative period. These complaints subsided by the time of union in all but 4 RG and 5 AG patients. The time to union was not different between the groups (AG = 98 days, RG = 92 days). No patient in the RG group developed heterotopic ossification in the knee. There was also no difference in final ROM of the knee or hip. SF 36 scores were collected, but found to be very inconsistent in this patient population even when given the same day; they were therefore not compared.

Discussion: Retrograde femoral nailing has become more popular over the past several years. However, there has been no randomized comparison of this technique to the gold standard of anterograde nailing reported, nor has the effect of the portal on knee function been determined in the long term. This study sought to begin to answer these questions. Part of the differences seen between these two groups is due to the different techniques used for nail placement. A fracture table was used for the AG group and a radiolucent table for the RG group. Nailing without a fracture table makes obtaining proper length and rotation more difficult as evidenced by the superior reductions in the AG group. A distractor can correct for length, but rotation in unstable fractures can be difficult to judge. 33% of the Winquist 3 or 4 fractures in the RG group had > 10° of rotational deformity. Surprisingly, the RG surgery took longer to accomplish than the AG surgery after setup. However, the later half of the study showed shorter operative times for the RG nailings, indicating that at least some of the increase was due to the learning curve of the surgeons and staff. Knee pain was reported commonly by the RG group early, but resolved in most cases as quadriceps strength returned. There was no difference in time to union between the groups.

Conclusion: Although retrograde nailing has advantages and is useful in certain circumstances, it is not superior to anterograde nailing for diaphyseal femur fractures. Difficulty in obtaining the correct length and rotation when nailing without a fracture table contributed to the malreductions seen in this study. A distractor is recommended for fractures that are short and under tension. Greater numbers of patients and longer followup is needed to determine the long-term outcome of the intraarticular knee portal before retrograde nailing should be recommended routinely for femur fractures.