From January 2006 to December 2010, 565 patients sustained trochanteric fractures (AO/OTA classification 31-A1 to A3) and were treated at two institutions with ITST nails (ITST™, Zimmer®, Warsaw, IN, USA). We selected 116 patients between the ages of 70 and 80 years old to minimize the age related factors such as medullary canal diameter and bone mineral density (BMD). Of these patients, 95 were followed for more than 1 year after surgery or revision surgery.

Surgery was performed with the patient in the supine position on the fracture table under general or spinal anesthesia. When possible, we aligned the medial cortices of fractured fragments in the antero-posterior image and the anterior cortices in the axial radiological image before the incision. When closed reduction failed, a long hemostatic forceps or pointed bone forceps was used to hold the fragments in reduced position after incision. A nail entry site was created on the medial edge of the tip of the greater trochanter. A CMN was inserted into the reamed medullary canal after trochanteric reaming. The diameter of the CMN was determined during preoperative planning according to the diameter of the medullary canal. Specifically, we measured the inner diameter of the proximal femur 2 cm below the distal border of the lesser trochanter in the Picture Archiving and Communication System (PACS) using a magnification bar. We used a nail 1 mm smaller than the measured value. In patients whose medullary canal diameter was over 13 mm, we used a 12-mm diameter CMN, the largest available nail in hospitals. If the medullary canal was too narrow to accept a 10-mm nail (the smallest available nail in the hospitals), we reamed the proximal femur with the medullary reamer, up to 11 mm. Reaming of the proximal femur was performed only if it was necessary to accept the nail. The end of the lag screw was positioned at the center of the femoral head in the antero-posterior and axial images. Distal locking was performed with one or two screws. Rehabilitation started at the first postoperative day. Sitting and continuous passive motion of the knee and hip joints were allowed. The majority of patients were encouraged to partially bear their weight within 1 week after surgery with walking aids depending on the patient's condition.

Fracture subgrouping was performed using AO/OTA classification. BMD was measured on the normal side of the hip joint using dual-energy X-ray absorptiometry (DXA; Discovery™, Hologic®, MA, USA) in 79 patients (83.2%) after surgery. Femoral neck T-scores were selected to represent patient BMD. The clinical absence of pain, presence of a callus shadow that passed through the fracture line in three cortices, and closure of the fracture line were used to determine bone union. Radiological measurements were analyzed using the PACS.

Standard anteroposterior (AP) radiographs of the hip were obtained with both legs positioned to an internal rotation of 15°. Lateral radiographs were taken with the opposite hip flexed and abducted. The canal occupancy ratio (COR) was calculated by dividing the CMN size by the femoral canal diameter. FNS after union was measured using the method described by Weil, et al.,7 at the time of the final follow-up exam. Briefly, the contralateral femur was overlaid on the shortened femur in the final AP radiogram. FNS was calculated using vectorial addition of changes in the x-axis (shortening in the abductor arm) and in the y-axis (shortening in the vertical plane). The sliding distance of the lag screw was measured from the antero-posterior radiological images taken immediately after surgery and at the time of the final follow-up examination, using the method of Watanabe, et al. 10 The tip-apex distance (TAD) was measured from the postoperative radiographs. The reduction was categorized as good, acceptable, or poor using the method of Baumgaertner, et al.11

We used SAS version 9.1 (SAS Institute Inc., Cary, NC, USA) for statistical analyses. Univariate and multiple regression analyses were performed to assess the individual patient effects on FNS. A stepwise multiple linear regression analysis was performed with backward elimination to select an appropriate model. A p-value greater than 0.10 was used for removal. A p-value by two tailed test less than 0.05 was considered significant for all analyses.

The study was approved by the Institutional Review Board (IRB No: 3-2011-0126).

The mean sliding distance was 3.11 mm (range, 0.50 to 16.80 mm); 2.55 mm (range, 0.5 to 8.1 mm) in A1, 3.02 (range, 0.7 to 12.1 mm) in A2, and 11.67 mm (range 8.4 to 16.8 mm) in A3. A3 type fractures demonstrated more screw sliding than A1 or A2 type fractures (p=0.004, p=0.006). Univariate regression analysis was performed to test the association of sliding distance to gender, age at the time of surgery, BMD, medullary canal diameter, COR, TAD, and AO/OTA classification. AO/OTA type (A1 versus A3; R²=0.6241, β=9.121, p<0.001), medullary canal diameter (R²=0.174, β=0.412, p<0.001), and COR (R²= 0.104, β=2.046, p=0.004) were significantly correlated with lag screw sliding distance. Multiple regression analyses (R²= 0.522) illustrated that COR (β=-9.410, p<0.001) and AO/OTA type (A1 versus A3; β=7.803, p<0.001) were more strongly correlated (Table 1).

The mean FNS was 4.49 mm (range, 1.7 to 17.8 mm); 3.90 (range, 1.90 to 9.50 mm) in A1, 4.41 mm (range, 1.70 to 13.50 mm) in A2, and 13.27 mm (range, 10.20 to 17.80 mm) in A3. Univariate regression analysis was performed to test the association of FNS to gender, age at the time of surgery, BMD, medullary canal diameter, COR, TAD, and AO/OTA classification. AO/OTA type (A1 versus A3; R²=0.340, β=9.064, p<0.001), medullary canal diameter (R²=0.207, β=0.447, p<0.001), and COR (R²= 0.212, β= -11.052, p<0.001) (Fig. 1) were significantly correlated with FNS. Multiple regression analyses (R²= 0.523) illustrated that COR (β=-8.918, p<0.001) and AO/OTA type (A1 versus A3; β=8.113, p<0.001) were more strongly correlated (Table 2).