Introduction

Schanz osteotomy (SO) is a palliative procedure for patients with neglected congenital dislocation of the hip. This surgery aims to reduce pain and limping and to improve hip function by providing anatomical support [1]. It is a low subtrochanteric valgization osteotomy. The apex of the osteotomy is kept at the level of the ischial tuberosity. Because the length of the flexor and abductor group muscles is not changed, the hip motion strength is maintained in these patients [1,2,3].

Total hip arthroplasty (THA) following SO of the proximal femur is challenging due to altered bony anatomy including medullary canal narrowing and ossification, changes in femoral alignment, and soft tissue contractures. In some cases, hardware needs to be removed at the time of THA and prior surgeries do increase the risk of deep implant infections [2,3,4]. However, few studies have reported the clinical and radiological outcomes of THA performed in patients previously treated with SO; these patients were evaluated only as small subgroups of a whole series [2, 5,6,7,8,9,10,11,12]. In the literature, only two studies have concentrated on THA after SO [3, 4]. Different osteotomy techniques, including trochantheric advancement, are suggested for deformity correction in the proximal femur. Complications related to the osteotomy site, neurological deficits, and dislocations are frequent [3, 4].

In this study, we evaluated the short-term radiological and clinical outcomes of THA after SO. Here, we report the perioperative complications of THA procedures performed for hips previously treated with SO.

Materials and methods

This study was approved by our institutional review board. All participants provided written informed consent. We retrospectively evaluated cementless THA procedures performed in 18 patients [2 male, 16 female; mean age, 55.4 (range, 50–66) years] with severely dislocated hips previously treated with SO of the proximal femur. Previous SO was not performed due to an underlying neuromuscular condition in any of the patients. All hips were classified as Crowe type 4. THA surgeries were performed at our institution between 2008 and 2014. Indications for arthroplasty were severe pain and/or considerable difficulty with walking and performing daily activities, joint contractures, limited hip abduction, and leg length discrepancy.

Medical records were reviewed for the degree of preoperative pain, limping, and daily life limitations. All patients were evaluated clinically and radiographically during the preoperative and follow-up periods.

The Harris hip score [13] was used for clinical evaluation. The presence of the Trendelenburg sign was recorded. We performed the modified Trendelenburg test, as described by Hardcastle and Nade [14], to qualitatively evaluate abductor function during the preoperative visit and the latest follow-up visit.

Radiological evaluation was performed using anteroposterior (AP) radiographs of the pelvis, AP and lateral radiographs of the operated hip, and AP full-length radiographs of the lower extremities (Fig. 1). Acetabular loosening was evaluated according to DeLee and Charnley and femoral loosening was evaluated according to Gruen [15, 16]. Union of the osteotomy site and leg length discrepancy were assessed. Union was defined by the presence of mature bone bridging the osteotomy on at least three of four cortices, as seen on AP and lateral radiographs. Leg length discrepancy was evaluated by measuring the length of the lower limbs using the preoperative and postoperative AP full-length radiographs of the lower extremities. Surgical time, mean intraoperative blood loss, perioperative complications, and postoperative complications were recorded.

Fig. 1
figure 1

Anteroposterior (AP) radiograph of the pelvis and AP full-length radiograph of the lower extremities used for preoperative planning

Full size image

Surgical technique

Patients underwent surgery in the lateral decubitus position under general or spinal anesthesia. A posterior or lateral approach was used. Hypertrophic and elongated joint capsules were resected. The anatomic location of the true acetabulum was identified by palpation of the pubis, ischium, and cotyloid notch; all soft tissues within the acetabulum were removed and the acetabulum was prepared. An appropriately sized, cementless, porous-coated acetabular shell was implanted. Sufficient coverage of the shell was achieved in all cases and superolateral acetabular rim grafting was not required. The shell was fixed with cancellous screws. Then, a ceramic or polyethylene liner was inserted. If present, the previous implants were removed before preparation of the femur. Proximal and distal transverse osteotomies perpendicular to the femoral cortices were performed at the level of prior Schanz osteotomies and a conical bone segment was removed (Fig. 2). Thus, shortening and correction of the deformity were achieved. After rasping of the proximal and distal femurs, a femoral stem with an appropriate size and length was inserted (Fig. 3). If required, the osteotomy site was secured with two, three, or more cable wires tensioned over the halved femoral segment obtained during the osteotomy and used as an autograft (Fig. 4) or with a plate and screws (Fig. 2). In some cases, strut allografts were also used with a plate and screws (Fig. 5). After implantation of the acetabular and femoral components, the joint was reduced and tested. Percutaneous adductor tenotomy was performed just after surgery in the patients who had severely limited hip abduction.

Fig. 2
figure 2

The level of the osteotomy sites determined preoperatively

Full size image
Fig. 3
figure 3

Anteroposterior (AP) radiograph of a patient without additional fixation materials or strut allografts

Full size image
Fig. 4
figure 4

Anteroposterior (AP) pelvis radiograph of a patient during postoperative month 18. The osteotomy site was secured with cable wires tensioned over the halved femoral segment obtained during the osteotomy and used as autograft

Full size image
Fig. 5
figure 5

Anteroposterior (AP) pelvis radiographs of a patient with a plate and screws applied with strut allografts for additional stability at the osteotomy site

Full size image

Two suction drains were used in all patients; they were removed during the first postoperative day. All patients were mobile without weight bearing, using two crutches for 6 weeks. First-generation cephalosporin was administered for 72 h as antibiotic prophylaxis. For venous thromboembolism prophylaxis, low-molecular-weight heparin (4000 Anti-Xa IU/0.4 ml; Clexane®, Sanofi Winthrop Industrie, Maisons-Alfort, France) was administered for 3 weeks postoperatively.

Statistical analysis was performed using SPSS 20.0.0 (SPSS Inc., Chicago, IL, USA). Preoperative and postoperative Harris hip scores were compared using Student’s t test. A p value of 0.05 was considered to be significant.

Results

The mean follow-up period was 30.8 (range, 18–56) months. The posterior approach was used in five (27.7%) patients and the lateral approach was used in 13 (62.3%) patients (Table 1). In 11 of the 18 patients (61%), plates and screws which were implanted during the previous Schanz osteotomy were present. These were removed before THA in the same session in all patients. The mean preoperative leg length discrepancy was 48.2 (range, 30–55) mm. The mean femoral shortening was 3.7 (range, 2–5) cm. The mean postoperative leg length discrepancy was 12.7 mm (range 4–2.4)mm. Final leg length discrepancy was less than 1 cm in ten patients, between 1 and 2 cm in five patients, and more than 2 cm in three patients. The mean diameter of the medullary canal at the osteotomy site was measured as 13.6 (range 12.8–15) mm. The osteotomy site was secured with two or three cable wires tensioned over the halved femoral segment obtained during the osteotomy and used as an autograft in eight cases. A plate and screws were used for fixation in ten patients in whom an additional procedure was required to maintain rotational stability perioperatively. In two cases, strut allografts were used with a plate and screws. Percutaneous adductor tenotomy was performed in eight (44%) patients. An acute improvement in abduction ROM was achieved in all patients. The mean duration of the surgery was 188.3 (range, 130–340) min. The mean perioperative blood loss was 1069.4 (range, 600–2250) ml.

Table 1 Patient demographics and clinical outcomes
Full size table

A variety of commercially available cementless acetabular and femoral components was used. The mean outer diameter of the acetabular shells was 47.1 (range, 40–54) mm. The acetabular shell was fixed with a single screw in 3 patients, with 2 screws in 13 patients, and with 3 screws in 2 patients. A ceramic liner was inserted in 14 patients and a polyethylene liner was used in the remaining 4 patients. The mean length of the femoral stems was 197.2 (range, 167–250) mm. The mean length of the stem bypassing the most distal portion of the osteotomy was 121.4 (range, 80–173) mm. A 28-mm head was used in 8 patients, a 32-mm head was used in 5 patients, and a 36-mm head was used in the remaining 5 patients (Table 2).

Table 2 Data related to the surgical procedure and implants
Full size table

There were complications in five (27%) patients. Nondisplaced fracture of the femur occurred during rasping or insertion of the femoral stem in two patients. Additional cable wires were applied to stabilize these cases. No additional complications were observed in these patients during follow-up. One patient had sciatic palsy that resolved completely during the seventh postoperative month. One patient had a superficial infection that was treated with local debridement and antibiotherapy. One patient had a Brooker [17] type 2 heterotophic ossification that was followed-up with rehabilitation. There was no pseudoarthrosis at the osteotomy sites. No dislocation was observed. No acetabular or femoral loosening was detected.

The Trendelenburg sign was positive in five (27.7%) patients and delayed positive in four (22.2%) patients postoperatively. No patient required crutches. The mean Harris hip score improved from 42.7 (range, 28–55) to 78.7 (range, 55–93) postoperatively (p < 0.05).

Discussion

THA in patients with a high hip dislocation in which a subtrochanteric femoral osteotomy was performed previously may reduce pain and improve hip function. However, it is technically demanding and complications frequently occur [3, 4, 7, 18, 19]. The main reasons for these difficulties are distorted anatomy and alignment of the femur, soft tissue alterations, change in the location of the hip center, leg length discrepancy, narrower femoral canal, adductor insufficiency, and insufficient bone stock in the acetabulum [4, 11, 18, 20]. Thorough preoperative planning is essential for THA in patients with previous SO of the proximal femur [3, 4, 11].

Various types of shortening osteotomies, including transverse, double-chevron, V-shaped, and oblique osteotomies, have been described, along with their advantages and disadvantages, for THA for the treatment of severely dislocated hips [4,5,6, 8, 21, 22]. The main technical difficulty in osteotomies is obtaining rotational stability [5, 10,11,12, 19, 23]. Although it was advocated that the rotational stability was maintained with most osteotomies, except for the transverse osteotomy, the comparative biomechanical study by Muratli et al. [22] demonstrated that there was no statistically significant difference in terms of the resistance against axial, torsional, and lateral bending forces among the four most used osteotomies, including the transverse osteotomy. Eskelinen et al. used a proximal shortening osteotomy with distal advancement of the greater trochanter in hips with a previous proximal SO and segmental shortening with angular correction for hips with a previous, more distal SO [4]. In their series involving 36 hips with previous SO, Sonohata et al. [11] used a subtrochanteric double-chevron osteotomy for ten hips and a subtrochanteric V-shaped osteotomy according to the procedure described by Hotokebuchi [11] for 24 hips. In the remaining two hips, they did not perform a subtrochanteric osteotomy. They reported that higher clinical scores were obtained after V-shaped osteotomy than after double-chevron osteotomy. Transverse osteotomy, which we performed in all our patients, is practical, straightforward, time-saving, and effective for the correction of deformities. It is also advantageous for setting the appropriate anteversion of the femur because it allows derotation of the proximal and distal ends of the osteotomy [24].

In our series, we used longer stems in almost all cases because they were also suggested for increased rotational stability in transverse osteotomy [11, 12, 23, 25]. During preoperative planning, the location of the prior Schanz osteotomy was also predictive of the length of the femoral stem. In the patients with more distally performed Schanz osteotomies, longer revision-type femoral stems were predicted to be used and these implants were found to be appropriate peroperatively for better stability and a longer distance to bypass the osteotomy site. Thus, we think that longer revision-type stems should be estimated for implantation in the patients with distally performed Schanz osteotomies and these implants should be kept available peroperatively. Moreover, in some cases, we also used plate and screw constructs with or without strut grafts to obtain additional stability, as suggested by Baz et al. [26]. Although plates may impair the periosteal blood supply and may have a negative effect on the unification of the osteotomy site, Baz et al. did not report pseudoarthrosis in any of their cases [26]. They suggested that plate and screw constructs could allow early weight bearing and that the rehabilitation program could be started earlier. Similar to their study, we did not find any pseudoarthrosis; moreover, we used limited-contact locking plates to preserve periosteal blood supply. Minimally invasive technique was performed in the distal screws of the plates.

Femoral shortening is performed to facilitate reduction and to avoid nerve palsy, which is a serious complication of THA in severely dislocated hips. Lewalle et al. claimed that the sciatic nerve is at risk in patients with acute leg length discrepancies of 2–4 cm [27]. However, Eggli et al. reported no correlation between the amount of lengthening and nerve damage in 508 hips that underwent surgery because of developmental dysplasia [28]. They suggested that direct or indirect trauma is responsible for most nerve palsy complications. Findings that support this view have been reported by Dunn and Hess [29]. They reported only 1 case of sciatic nerve palsy in their series of 22 cases, with average lengthening of 5 cm and maximum lengthening of 9 cm. They claimed that with lengthening of 4 cm, they encountered problems during the implantation of the femoral stem perioperatively.

In our study, we defined the required amount of appropriate shortening using preoperative full-length radiographs of the lower extremities. However, the amount of the shorthening should be determined also by soft tissue tension. The limb is often short at the beginning and a balance between the soft tissue tension and maximization of the leg length should be maintained to correct the leg length discrepancy. We performed a mean shortening of 3.6 cm (maximum, 5 cm). Postoperative transient sciatic palsy was present in 1 (5.5%) patient. A posterior approach was used in this patient and shortening of 4 cm was performed. We think that nerve palsy complication in this patient were not due to the insufficiency of the shortening. The main reason may be the use of the posterior approach and approximal anatomical location of the sciatic nerve in relation to the retractors used for the approach. Nevertheless, most authors recommend a maximum lengthening of 4 cm [30].

Patients with a highly dislocated hip due to developmental dysplasia usually have a narrow femoral canal and thinner cortices. In these cases, large modular femoral components used in THA combined with a femoral osteotomy may cause fissuring or fracture of the femur [10, 23]. Moreover, a previously performed SO of the femur may lead to harder femoral cortices near the osteotomy site, therefore, causing increased rigidity [4, 9, 20]. Eskelinen et al. reported 5 (7%) nondisplaced fractures of the proximal part of the femur out of 68 hips in which THR was performed after a previous SO [4]. In their series of THA after a previous SO, Sonohata et al. reported intraoperative and postoperative femur fractures in 3 (8%) out of 36 hips [11]. In our series the mean outer diameter of the medullary canal at the osteotomy site was measured as 13.6 (range 12.8–15) mm. Although this narrowest site of the femoral medullary canal was resected during shorthening osteotomy, the use of long revision type femoral stems which are only available in relatively thicker diameters caused nondisplaced fractures of the femoral cortices in 2 (11%) patients in our study. These occurred during rasping or implantation of the femoral stem. Local osteoporosis due to reduced loading may also be another factor for this complication encountered during the implantation process of the femur during THA performed in severely dislocated hips.

Insufficiency of the gluteus muscle is responsible for the persisting Trendelenburg sign in patients who have undergone THA due to developmental hip dysplasia or coxarthrosis due to coxa vara [6, 31,32,, 32]. In our study, the Trendelenburg sign was positive in five (27.7%) patients and delayed positive in four (22%) patients postoperatively. In a patient with delayed positive Trendelenburg sign, the gait is normal during the preliminary evaluation, but when the patient is asked to walk quickly, the Trendelenburg gait becomes obvious due to fatigue and the abductor mechanism insufficiency [14]. Thus, the patients who are suspected of gluteus muscle insufficiency should be evaluated further even if they are found Trendelenburg negative during single leg stance. In one of the patients with a positive Trendelenburg sign, pelvic tilt due to underdiagnosed lumbar scoliosis was present. The high frequency of positive Trendelenburg signs in our series may be due to more posterior and lateral anatomical localization of the greater trochanter and, therefore, shorter adductor muscles in the patients previously treated with SO. After reduction of the joint in the true acetabulum, the greater trochanter and the abductors may be left in a posterior position; however, the abductor lever arm was not restored sufficiently. Therefore, an increase in the tension of the gluteus medius muscle may be responsible for the positive Trendelenburg sign. In addition, the four patients with a delayed Trendelenburg sign had relatively shorter follow-up periods and did not have additional plate and screw constructs for fixation of the subtrochanteric osteotomy sites. We believe that the main cause of the positive Trendelenburg sign in these patients was insufficient rehabilitation.

There were several limitations in our study. The inherent nature of a retrospective review has well-known limitations and biases. The small number of patients was another limiting factor when making definitive statistics. However, this limitation may be present in any single-institutional study on this topic because of the rarity of THA procedures performed in patients previously treated with SO. In addition, the mean follow-up period was relatively short. Patients underwent surgery by multiple surgeons using implants of various types and sizes. In addition, there were variations in the types of treatment because additional fixation techniques were applied in transverse subtrochanteric osteotomy sites in some, but not all patients. Despite these limitations, we think that the current study is valuable because there is no sufficient data regarding the outcomes of THA procedures performed in patients previously treated with SO.

Conclusions

THA in patients previously treated with SO is technically demanding and the complications are frequent. Thorough preoperative planning is mandatory. Surgeons should be cautious and prepared during surgery due to severed anatomy, narrow medullary canal, harder cortices, and local osteoporosis of the femur.