OTA 1997 Posters - Femur Fractures

Reamed Intramedullary Nailing of the Femur: A Review of 551 Cases and Radiolucent Table versus Fracture Table for Reamed Intramedullary Nailing of the Femur: Effect on Fracture Alignment

Philip R. Wolinsky, MD, Eric McCarty, MD, Yu Shyr, PhD, Kenneth Johnson, MD

Vanderbilt University Medical Center, Nashville, Tennessee, USA

Purpose: To determine the results of reamed intramedullary nailing of the femur (RIMNF) performed using closed techniques and interlocking nails, and to determine if the alignment of fractures of the femoral shaft treated with a reamed intramedullary nail inserted with the leg draped free on a radiolucent table differs from the alignment obtained when nails are inserted with the use of a fracture table.

Materials and Methods: Fractures of the femoral shaft treated with a first generation intramedullary nail at our institution from 6/1986 to 3/1996 were identified. The following inclusion criteria had to be met: male patients had to be 13 years of age or older at the time of injury, and females had to be 12 years of age or older. Follow-up of at least 6 months was required unless healing was documented to have occurred if the follow up period was less than 6 months. A review of the patient's medical record and radiographs was then performed. Univariate statistical analysis was carried out using either the analysis of variance (ANOVA), or the Chi-square test for continuous or categorical variables respectively. Logistic regression models with 95% profile likelihood confidence limits for parameter estimates were carried out to address the association between fracture malalignment greater than 5° and other possible covariates. Differences were considered to be statistically significant when the p value was <0.05.

Results: 822 fractures were identified. 551 fractures in 515 patients met the inclusion criteria. 87% of the fractures occurred as the result of either a motor vehicle or motorcycle accident. The average age was 27.2 years (range 12-87), and the male to female ratio was 345:170. 461 of the fractures were closed. Of the 90 open fractures, 37 were grade 1, 38 grade 2, 9 grade 3A, 3 grade 3B, and 3 were grade 3C. 93% of all the fractures and 94% of the open fractures were stabilized within 24 hours of injury.

Five hundred forty-five fractures healed for a union rate of 98.9%. 26 fractures needed a second procedure to heal: 11 underwent a nail exchange, 15 required dynamization by removal of locking bolts, and 3 needed nail exchange and dynamization. One nail and 13 (2.4%) of the locking bolts broke. 6 (1.1%) fractures developed an infection. Three of the infections were deep and 3 were superficial. All the infections developed in closed fractures, none of the open fractures developed an infection. 211 or 38% of the fractures required hardware removal. Eighty locking bolts and 131 of the nails were removed. 73% of the bolts and 58% of the nails were removed because of patient complaints of pain.

Adequate follow-up films were available to evaluate alignment of the femoral shaft for 418 fractures. No fractures healed with more than 10° of varus/valgus (v/v), or procurvatum/recurvatum (p/r). 374 (89%) of the fractures healed with less than 5° of angulation in either plane. 40 (9.6%) healed with more than 5° in one plane, but not the other, and 4 (1%) healed with greater than 5° of angulation in both planes. Univariate and multivariate statistical analysis revealed an association between angulation of >5° and a fracture of the distal 1/3 of the femur. We do not routinely measure rotation. However, no patient required a derotational osteotomy for a clinically significant rotational deformity. We also do not routinely measure leg lengths. 2 patients who had the nails inserted using a fracture table were lengthened approximately 2 cm. Two patients nailed without the use of a fracture table were shortened 2 cm, and one patient shortened after removal of locking bolts.

90 open fractures were treated with RIMNF. 94% of these fractures were stabilized within 24 hours of injury. All the open fractures healed, and none developed an infection.

255 fractures were stabilized without the use of the fracture table and 296 with the use of a fracture table. Adequate follow up radiographs were available for 229 of the fractures stabilized without the use of a fracture table and for 189 of those with the use of a fracture table. The difference in the proportion of films available forthe two groups is due to the fact that our radiology department purges the files of inactive patients. Since the fractures stabilized with the use of a fracture table were performed at an earlier point in our study, a higher percentage of these files were purged. Statistical analysis revealed no differences in the sex, age, side of fracture, proportion of open stabilized injuries, fracture location, pattern, or Winquist classification between the two groups. No fracture healed with more than 10° of varus/valgus (v/v), or procurvatum/ recurvatum (p/r). 202 of the fractures stabilized without a fracture table healed with less than 5° degrees in either plane as did 172 of the fractures treated with a fracture table. 27 fractures treated without, and 17 of the fractures treated with a fracture table healed with more than 5° but less 10° of deformity in either plane. Statistical analysis revealed no association between choice of operating room table and malalignment. When fracture location was analyzed, a distal 1/3 fracture location was found to be associated with a greater chance of malalignment (p=0.0001).

Multivariate analysis was performed using age, sex, fracture location, and use of the fracture table as covariates. Only a distal 1/3 fracture location was found to be associated with a higher incidence of malalignment (p=0.0001). The choice of the operating room table was not associated with malalignment (p=0.5418).

Discussion and Conclusion: A review of 551 fractures treated with a first generation intramedullary interlocking nail revealed a high union rate, low infection rate, low rate of malunion, and low incidence of hardware failure. No open fracture developed an infection. However, 38% of all fractures required removal of hardware, most commonly because of complaints of hardware related pain. Angulation of more than 5 degrees in either plane was found to correlate with a distal 1/3 fracture, perhaps because of the flare of the canal in the metaphysis.

RIMNF is usually performed with the use of a fracture table to gain mechanical advantage to regain length and alignment. Since 1991 we have shifted to performing RIMNF with the leg draped free on a radiolucent table. Analysis of the alignment of fractures treated with both methods reveals that the choice of operating room tables has no effect on angular alignment of the femoral shaft. Rather, fractures of the distal 1/3 have a higher incidence of malalignment regardless of the table used. No patient required a derotational osteotomy. Two fractures treated using the fracture table required reoperation for excessive lengthening and two treated using the radiolucent table required operative treatment for excessive shortening.

The alignment obtained treating femoral shaft fractures with the nail inserted without the use of the fracture table does not differ from fractures treated with a nail inserted using a fracture table.