Introduction

Total hip arthroplasty (THA) for the treatment of severe dislocation related to developmental dysplasia of the hip has been one of the most technically challenging procedures. Several previous reports mentioned contraindication of THA in these patients or a high incidence of complications and failure of THA because of severe anatomical abnormalities, including deficiency of acetabular bone stock, abnormal location of the hip center, abnormal neurovascular structures, leg length discrepancy, muscular contractures, increased anteversion of the femoral neck, abductor muscle insufficiency, and a narrow femoral intramedullary canal [1–4]. Acetabular components placed high on the thin iliac bone with cement have been shown to loosen and, in addition, the abductor muscle insufficiency remains and leg length discrepancy is not corrected. To overcome these problems, placement of the acetabular component at the center of the true acetabulum has been emphasized [5–7].

Femoral shortening osteotomy in THA for severely dislocated hips permits the reduction of the femoral head into the acetabular component at the level of the true acetabulum. The original method of shortening is resection of the proximal femur coupled with trochanteric osteotomy and cemented fixation of the femoral component [4, 5]. This procedure does have the advantages reducing the femoral head, improving leg length discrepancy, restoring abductor muscle function, and avoiding complications such as femoral and sciatic nerve palsy. However, the incidence of trochanteric nonunion is high [8], and excision through the intertrochanteric region sometimes results in massive loss of the proximal femoral bone stock. Recently, techniques of subtrochanteric femoral shortening osteotomy in THA have been developed to preserve proximal femoral bone stock, in order to reduce the need for trochanteric osteotomy and to achieve reduction of the femoral head into the acetabular component without neurovascular injury. There are many reports of surgical procedures for subtrochanteric femoral shortening osteotomy, including step-cut, oblique-cut, double-chevron, and V-shaped osteotomy [9–12]. Interestingly, all of them are based on the cementless THA technique. There have been only a few reports of subtrochanteric femoral shortening osteotomy coupled with cemented THA [13, 14].

In this study, we retrospectively evaluated the postoperative results of cemented THA with subtrochanteric femoral shortening transverse osteotomy in patients with severely dislocated hips, especially in terms of bone union at the site of femoral osteotomy.

Patients and methods

Between May 1999 and June 2005, cemented THAs with subtrochanteric femoral shortening transverse osteotomy in 15 severely dislocated hips (11 patients) were performed. Under general anesthesia, all of the patients were operated on in the lateral decubitus position using an anterolateral approach with trochanteric osteotomy as previously reported by Dall [15]. The acetabular component was placed in the true acetabulum by PMMA cement. The resected femoral head was applied to the defective acetabulum as an autogenous bone graft if necessary. A femoral shortening transverse osteotomy was performed by resecting the femur below the lesser trochanter in accordance with preoperative planning (Fig. 1a). If necessary, the femur was derotated. While holding the osteotomy site tightly with bone clamps (Fig. 1b), cement fixation of the femoral component was performed. The resected cylindric bone segment was cut longitudinally and these two pieces were used as grafts to prevent rotational instability and to accelerate bone union of the osteotomy site, as previously described (Fig. 1c) [16]. Morselized cancellous bone autografts from the resected femoral heads augmented around the osteotomy site. Two weeks after the operation, partial weight-bearing with crutches was allowed and progressively increased. Ambulation with a single cane continued until there was radiological evidence of bone union at the osteotomy site.

Fig. 1
figure 1

Operative technique for subtrochanteric femoral shortening transverse osteotomy. a Subtrochanteric transverse osteotomy is completed with an osteotome. The arrows indicate the osteotomy sites. b Femoral shortening and derotation are performed. The docking site is held with bone clamps. The arrowhead indicates the docking site. c After cement fixation of the stem, the osteotomy site is stabilized by bone grafting with removed segments of the femur

Full size image

Patients were evaluated preoperatively and postoperatively with the Merle d’Aubigné and Postel hip score [17]. A routine radiographic examination was performed. A line drawn through the teardrops was referred to in order to evaluate the level of the femoral osteotomy site and the radiographic leg lengthening. The amount of radiographic leg lengthening was defined as the length obtained by subtracting the amount of intraoperative femoral resection from the distance between the top of the greater trochanter preoperatively and postoperatively on radiographs, as previously described [18]. Before this operation, we obtained informed consent from the patients and their families.

Statistical analysis

Analysis of correlation between the time to bone union and several factors was performed via Spearman’s correlation coefficient by rank. p ≤ 0.05 was considered significant.

Institutional review board approval was obtained for publication of the study. The patients and their family were informed that data from the cases would be submitted for publication, and gave their consent.

Results

Fifteen hips (11 patients) received cemented THA with subtrochanteric femoral shortening transverse osteotomy. There were one man and ten women. The average age of the patients was 58.9 years (range 42–77 years). The average weight and height were 41.5 kg (range 33–53.6 kg) and 147.0 cm (range 136.5–160 cm), respectively. All of the patients had congenital dislocation of the hip, and one patient had a complication of rheumatoid arthritis. According to Crowe’s classification [4], the joint dysplasia was graded as grade III in one hip and grade IV in 14 hips. Unilateral dislocated hips with normal contralateral hips were observed in 2 patients, unilateral dislocated hips with osteoarthritis of contralateral hips in 5 patients, and bilateral dislocated hips in 4 patients. Preoperative radiographs showed that the mean Sharp angle of the acetabulum was 52.3° (range 40–70°). Clinical evaluation revealed that the mean preoperative leg length discrepancy was 28.5 mm (range 0–67 mm), and that the Merle d’Aubigné and Postel hip score was 8.1 ± 2.5.

The average operation time was 3.3 h (range 2.8–4 h), and the mean intraoperative blood loss was 578 ml (range 190–870 ml). The femoral head was applied as an autogenous graft in all but one of the cases to augment the superolateral wall of the acetabulum. The average size of the acetabular components was 40 mm (range 38–44 mm). The femoral components included 6 PHS type 6 straight (JMM, Kyoto, Japan), 1 PHS type 7 regular (JMM, Kyoto, Japan), 3 PHS wide straight (JMM, Kyoto, Japan), 1 PHS standard (JMM, Kyoto, Japan), and 4 HS3 (JMM, Kyoto, Japan), The mean length of the stem bypassing in the most distal portion of the osteotomy was 95.2 mm (range 70–118 mm). The mean proximal level of the transverse osteotomy was 10.7 mm (range 0–30 mm), which was below the inferior border of the lesser trochanter. The mean length of intraoperative femoral resection was 38.3 mm (range 10–76 mm). The highest point of the greater trochanter postoperatively was lower by a mean of 67 mm (range 38–90 mm) than its preoperative point on the radiograph. The calculated measurement of radiographic leg lengthening was a mean of 27.1 mm (range 10–55 mm). The mean postoperative leg length discrepancy was 3.3 mm (range 0–14 mm).

Two patients experienced postoperative dislocations, but dislocation did not become recurrent. No clinical evidence of postoperative neurologic and vascular injuries was observed. Follow-up periods averaged 75.7 months (range 33–125 months). Twelve osteotomy sites achieved bone union, including one patient who had low-intensity pulsed ultrasound (LIPUS) accelerated bone union. The average time to bone union at the osteotomy sites was 9.2 months (range 6–15 months). No statistically significant correlation between the time to bone union and several factors was found (Table 1). However, the length of the resected femur exhibited a weak correlation with the time to bone union. Nonunion occurred in three osteotomy sites of 2 patients and revision surgeries were performed. At the latest follow-up, no radiological loosening was noted in the cemented acetabular and femoral components, and the Merle d’Aubigné and Postel hip score had improved to 15.1 ± 1.3.

Table 1 Correlation of several factors with the time to bone union
Full size table

Case reports

Case 1

A 42-year-old woman with Crowe IV dislocated bilateral hips was treated using PHS type 6 straight stems (Fig. 2a, b). Six months after operation, bone union at the osteotomy sites was complete (Fig. 2c).

Fig. 2
figure 2

Radiographs of a 42-year-old woman with Crowe IV dislocated bilateral hips. a Before the operation, b at THA using PHS type 6 straight stems, c six months after the operation; bone union by bridging callus had occurred

Full size image

Case 2

A 57-year-old man with Crowe IV dislocated bilateral hips was treated using PHS standard and wide straight stems (Fig. 3a, b). He was also treated for rheumatoid arthritis. Two weeks after the operation, the THA was dislocated at the left hip. Seventeen months after the operation, a left femoral shaft fracture occurred (Fig. 3c) and an open reposition and internal fixation (ORIF) was performed using an AO plate (Fig. 3d). Four months after the ORIF, the AO plate was dislodged (Fig. 3e), and then revision of the stem was performed using the Huckstep hip system (Fig. 3f). Six years after THA of the right hip, a right femoral shaft fracture occurred (Fig. 3g) and revision of the stem was performed using the Huckstep hip system (Fig. 3h). Intraoperative findings revealed fibrous nonunion at the osteotomy sites in bilateral femora. At the time of final follow-up, bone union was not obvious at either osteotomy site (Fig. 3i, j).

Fig. 3
figure 3

Radiographs of a 57-year-old man with Crowe IV dislocated bilateral hips. a Before the operation, b at THAs using PHS standard and wide straight stems, c thirteen months after the operation (a left femoral shaft fracture had occurred), d at ORIF, e 4 months after ORIF (the plate was dislodged), f at revision, g 6 years after THA (a right femoral shaft fracture had occurred), h at revision, i seven years after revision (bone union of the left femoral osteotomy site is still not obvious), j three years after revision (bone union of the right femoral osteotomy site is not obvious)

Full size image

Case 3

A 55-year-old woman with a Crowe IV dislocated left hip was treated using HS3 stem (Fig. 4a, b). Five months after the operation, radiography showed an increased gap at the osteotomy site. Because bone union had not occurred by 3 years after the operation (Fig. 4c), reoperation was performed using strut bone allografts (Fig. 4d). Intraoperative findings revealed disruption of the cement mantle at the osteotomy site and subsequent rotational instability between the stem and the cement mantle. Three months after the operation, LIPUS was applied by an SAFHS2000 Sonic Accelerated Fracture Healing System (Smith & Nephew Inc., Memphis, TN, USA, provided by Teijin Pharma Ltd., Japan) at the upper and lower ends of the allografts for 20 min everyday. Bone fusion between the femur and the lower end of a strut allograft currently observed (Fig. 4e).

Fig. 4
figure 4

Radiographs of a 55-year-old woman with a Crowe IV dislocated left hip. a Before the operation, b at THA using HS3 stem, c 3 years after the operation (the docking site had a gap, as shown by the arrowheads, and no callus formation was observed), d at revision of the stem with the strut allografts, e 1 year after revision (the distal end of the strut allograft has achieved bone union; see the arrow)

Full size image

Case 4

A 58-year-old woman with a Crowe III dislocated left hip was treated using PHS type 7 stem (Fig. 5a, b). Six months after the operation, she complained of left thigh pain. Radiography showed resorption of grafted autogenous bones and incomplete callus formation around the osteotomy site (Fig. 5c). Patient treatment began with LIPUS (applied by an SAFHS2000) for 20 min everyday, and 9 months after LIPUS application bone union was observed at the osteotomy site (Fig. 5d, e). At present, the patient has no thigh pain and ambulates without a cane.

Fig. 5
figure 5

Radiographs of a 58-year-old woman with a Crowe III dislocated left hip. a Before the operation, b at THA using PHS type 7 stem, c 6 months after the operation (the patient had thigh pain). d Fifteen months after the operation (bone union was complete). e A boxed region in d is shown at a high magnification. Bridging callus formation is obvious. The arrowheads denote cement invasion into the docking site

Full size image

Discussion

Our retrospective study indicates that a fair clinical outcome of cemented THA with subtrochanteric femoral shortening transverse osteotomy is achieved eventually in patients with severely dislocated hips. The major reasons for the great restorations observed in clinical findings are both relief from pain and improvement in ambulant activity. In all patients, we placed the acetabular components on the true acetabulae to increase the tension of the gluteus medius muscle, resulting in the disappearance of Trendelenburg limp. Furthermore, postoperative leg length discrepancy was less than 1.4 cm in all patients. Because of the small diameter of the acetabulum, we preferred cement fixation of the acetabular component with a structural autogenous bone graft. The first reason to use this is to obtain greater polyethylene thickness of the acetabular component in cement fixation than that in the cementless component with the same diameter. Second, we have reported the long-term results in cemented THA with acetabular autogenous bone grafting in our hospital, and we have shown that Kaplan–Meier survivorship analysis predicted a rate of survival of the acetabular components at 15 years of 96% [19]. Thus, our findings indicate longer-term success in autogenous acetabular bone grafting compared with previous reports [20, 21]. Indeed, none of the acetabular components of cemented THA in severely dislocated hips in this study exhibited radiological loosening for the follow-up periods.

The procedure of subtrochanteric femoral shortening transverse osteotomy has the advantage of easy preoperative anteversion correction. In other osteotomies, including step-cut and V-shaped using cementless stems, the techniques used for derotation of the femur are not simple if specific components with modular necks are not used [21, 22]. In addition, it is difficult to keep the osteotomy site stable using transverse osteotomy with cementless stems; additional fixation or complete press-fit of the stem at the osteotomy site is therefore required. In contrast, in cemented fixation of the stem, the intramedullary cement mantle hampers rotational instability of the osteotomy site. However, in case 3, rotational instability and subsequent nonunion occurred, because the columnar shape of the distal portion of the HS3 stem produced rotational instability at the stem–cement interface. Thus, distal-tapered and rectangular-sectioned stems should be used in subtrochanteric femoral shortening transverse osteotomy with a cemented stem to ensure that the cement mantle is sufficiently thick and to inhibit rotational instability. Otherwise, the transverse osteotomy site should be stabilized, by a unicortical plate for instance.

The major concern about this procedure is the apparent delayed union of the osteotomy site compared with that seen with cementless stems. Although some nonunions have been reported [22–24], the average time to bone union is generally less than 6 months when cementless stems are used [18, 25]. In cemented THA, the presence of the cement in the intramedullary cavity results in a loss of bone marrow cells and degeneration of the endosteum, both of which have the potential to form new bone at the osteotomy site. Furthermore, postoperative radiography shows the invasion of a small amount of cement between the docking sites, which precludes bridging callus formation in the cortex. Moreover, the preparation of osteotomy sites on the femur by removing periosteum circumferentially from the outer surface of the cortex reduces the potential activity towards bone formation of the periosteum. Indeed, the patient in case 2 has not formed callus at the osteotomy site due to a complication, rheumatoid arthritis, which is believed to suppress bone formation activity. Analysis by Spearman’s correlation coefficient by rank of the correlations between the time to bone union and several factors such as the length of femoral resection, the lowering of the greater trochanter, final leg lengthening, the level of the osteotomy site from the lesser trochanter, and the length of the stem below the osteotomy site indicated that there was no statistically significant correlation among these. However, the length of resected femur exhibited a weak correlation with the time to bone union (p value 0.081), meaning that the grafting of intercalary cortical bone fragments can be considered critical to bone union. To overcome disadvantage of this procedure, we applied noninvasive LIPUS to the osteotomy site in case 4. Accumulating evidence shows that LIPUS accelerates fracture healing by stimulating bone growth [26–28]. In this patient, the intercalary bone fragments and the autogenous bone chips placed around the osteotomy site were resorbed, and the bridging callus formation outside the osteotomy side was incomplete, possibly resulting in the complaint of thigh pain 6 months after the operation. The application of LIPUS for 9 months induced bridging callus formation, suggesting that LIPUS is a valuable tool for facilitating bone union at the osteotomy site for this procedure.

In conclusion, cemented THA with subtrochanteric femoral shortening transverse osteotomy is one of the possible procedures for the treatment of severely dislocated hips. However, the high incidences of nonunion and delayed union associated with this technique may lead to high loosening rates of the femoral stem, and we need to evaluate the long-term results of this procedure. We also need to improve the technique in order to avoid the invasion of cement into the osteotomy site, to keep the periosteum intact as much as possible, to graft the intercalary cortical bone fragments at the osteotomy site for as long as possible, and to prevent rotational instability at the transverse osteotomy site with internal fixation.