Modified Hardinge approach is an alternative to trochanteric flip osteotomy for Pipkin type IV femoral head fracture

Background Pipkin type IV femoral head (FH) fractures generally have poor prognoses. Although several surgical approaches are used, the optimal procedure is still under debate. The purposes of this study were to compare two approaches, the modified Hardinge approach and trochanteric flip osteotomy, for the treatment of Pipkin type IV FH fractures. Methods This retrospective study included 20 patients who underwent surgical treatment for Pipkin type IV FH fractures between January 2011 and September 2017 at a level 1 trauma center. Thirteen were treated using the modified Hardinge approach (group A) and seven with trochanteric flip osteotomy (group B). All patients completed at least 1 year of follow-up. The clinical outcome of the Merle d’Aubigné-Postel score; and radiological outcomes including the quality of the fracture reduction, the osteonecrosis of the FH (ONFH), the hip joint osteoarthritis (OA), and heterotopic ossification (HO); were compared between the two groups. Conversion to THR was also recorded as an outcome measure, analyzed by Kaplan–Meier curve and log-rank test.

approach results in reduced blood loss, similar clinical and radiological outcomes compared with trochanteric osteotomy. The modified Hardinge approach appears to be an acceptable alternative to trochanteric flip osteotomy.

Background
Fracture of the femoral head (FH) is a relatively uncommon injury, and is mostly caused through high-energy trauma such as sports injuries, traffic accidents, falls from a height, and industrial accidents [1][2][3]. In 1957, Garrett Pipkin proposed a simple and useful classification for FH fractures [4]. Among the four types of fractures described, combined FH and acetabular fractures (Pipkin type IV) are believed to be associated with poorer prognoses [4][5][6]. The optimal surgical approach for combined lesions remains controversial because addressing both fractures through a single surgical approach is difficult.
Use of the modified Gibson incision and modified Hardinge approach to treat Pipkin type IV FH fractures was described in our previous study [17]. After the soft tissues around the hip joint are dissected, the joint can be dislocated and the fractured fragments of the FH managed without trochanteric osteotomy. Meanwhile, the acetabular component can be reduced and repaired simultaneously with this single surgical approach.
The aim of the current study was to compare two surgical approaches, modified Hardinge and trochanteric flip osteotomy, which were used for the surgical treatment of Pipkin type IV FH fractures. The quality of fracture reduction and radiographic and clinical outcomes were reported.

Surgical Procedures
Under general anesthesia, the patient was placed in the lateral decubitus position with the injured hip facing up. We used a modified Gibson skin incision, as described by Moed in 2010 [18].
For group A, the modified Hardinge approach was performed as described in our previous paper [17]. In brief, we released the muscle-tendon junction of the gluteus medius (GMe) at the point of insertion of the greater trochanter (GT). We then cut the gluteus minimus (GMi) 1 cm medially to the edge of the GMe. A T-shaped capsulotomy was performed to release the anterior capsule. The hip joint was then dislocated, and the fractured fragments of the FH were viewed. One to three 3.5-mm interfragmental screws were applied to fix and compress the fracture site. The hip was reduced, and the anterior capsule, GMi, and GMe were repaired layer by layer.
For group B, the trochanteric flip approach of Ganz et al was followed without modification [7]. Trochanteric osteotomy for trochanteric flip with a maximal thickness of about 1.5 cm was conducted using an oscillating saw along the line of osteotomy just anterior to the posterior insertion of GMe. A Z-shaped capsulotomy was performed to avoid injury to the deep branch of the medial femoral circumflex artery. The fractured fragments of the FH were reduced and stabilized in the same way as for the modified Hardinge approach. The trochanteric osteotomy was repaired using two 3.5-mm cortical screws.
Then, the injured acetabular posterior wall (PW) or posterior column (PC) was reduced, and a pre-bent Judet reconstruction plate or pelvic-locking plate (Johnson & Johnson, New Brunswick, New Jersey, USA) was applied.

Follow-up Protocol
After the index surgery, plain radiographs of the pelvic anteroposterior (AP) view and two Judet oblique views were obtained. Each patient was allowed toe-touch ambulation for at least 6 weeks, then full weight bearing ambulation for a further 6 weeks. Three months after surgery, patients were allowed to walk freely without assistance.
Patients were assessed at intervals of 1, 3, 6, and 12 months after discharge. Thereafter, follow-ups were conducted annually. Functional and radiological evaluations were performed and documented at each clinic visit. Functional results were evaluated using the Merle d'Aubigné-Postel score [19].
Prophylaxis was not routinely prescribed for HO. No pharmacological thromboprophylaxis was administrated, but compression stockings were applied to all patients for at least 3 months since the time of injury. Perioperative or postoperative complications including neurovascular injury, loss of reduction, dislocation, hardware failure or irritation, infection, and venous thromboembolism were recorded.

Outcome Measures
Four parameters were used to evaluate the radiological outcomes: the quality of fracture reduction using Matta's grading system [20], osteonecrosis of the FH (ONFH) using the Ficat classification [21], hip joint osteoarthritis (OA) using the Tönnis classification [22], and heterotopic ossification (HO) using the Brooker classification [23]. Conversion to total hip replacement (THR) was also recorded as an outcome measure. No THR was performed during immediate post-trauma surgery. All THRs were implanted during postoperative follow-up when indicated.

Statistical Analysis
Numerical data are presented as means ± standard deviation (SD), while categorical data were expressed as absolute frequencies and percentages. Comparison analysis was conducted between the two groups using either the chi-squared or Fisher's exact test for categorical data, and the two-tailed Student's t-test or nonparametric Mann-Whitney Utest for numerical data. Patient survival, censored for THR, was analyzed by Kaplan-Meier curve and log-rank tests.  Comparison of the Two Approaches Patient characteristics including age at injury, gender, follow-up duration, FH fracture pattern, acetabular fracture type, and injury severity score showed no significant differences between the two groups ( Table 1). The mean operative time was 175.8 ± 62.6 min for group A and 138.9 ± 48.7 min for group B (P = 0.275). The estimated blood loss was 233.1 ± 116.8 ml for group A and 435.7 ± 307.8 ml for group B (P = 0.135). There were no perioperative complications. The length of hospital stay did not significantly differ between the two groups (P = 0.067).
With regards to radiographic outcomes, the quality of fracture reduction by Matta's grading did not differ significantly between the groups (P = 1.000), nor did the occurrence of ONFH (group A vs group B = 30.8% vs 14.3%, P = 0.613), incidence of hip joint OA In terms of post-operative complications; no cases of nerve injury, loss of reduction, hardware failure, infection, or symptomatic venous thromboembolism were detected in either group. One 27-year-old male from group A had recurrent dislocation 11 days after the index surgery. Revision surgery was performed and the dislocation was resolved.
However, post-traumatic OA was noted during follow-up, and THR was implanted 10 months after injury. All trochanteric osteotomies healed uneventfully (mean healing time: 16 weeks, range: 10-26 weeks). One patient complained of discomfort caused by the trochanteric screws and requested removal of the implants.

Replacement
Five patients underwent THR during the follow-up period; three (60% of THRs) within 12 months of the initial surgery, and the procedure was performed beyond the first 12 months in two cases. Figure 1 shows the Kaplan-Meier survival curve without THR for all patients. In group A, three patients with Ficat stage IV ONFH and one patient with Tönnis grade III hip joint OA received THR at an average of 21.3 ± 19.6 months after initial surgery. One patient in group B had stage IV ONFH during follow-up, and THR was implanted 11 months postoperatively. The log-rank test showed no significant difference between the two groups (P = 0.629).

Discussion
Approaches for the management of FH fractures are well represented in the literature.
However, relatively few reports exist concerning combined FH and acetabular fracture (Pipkin type IV) due to the infrequency and complexity of this condition, and there is debate over the optimal surgical approach for Pipkin type IV fractures [14,24,25]. Table   2 summarizes the major case series studies of patients with Pipkin type IV FH fractures.
Nevertheless, some previous investigators included nonoperative populations [26,27], and some surgeons used a single approach [9,11,28]. The rarity of Pipkin type IV fractures makes comparative studies difficult. To the best of our knowledge, the present study is the first to compare different surgical approaches for Pipkin type IV FH fractures.
Several methods of surgical dislocation of the hip have been suggested for the management of FH fractures, of which trochanteric flip osteotomy is one [7,8]. This approach with surgical dislocation of the hip can achieve accurate reduction and fixation of both the FH and acetabular fractures with low risk of compromising the FH blood supply [6, 8-11, 25, 29, 30]. However, complications following trochanteric flip osteotomy are frequently observed; including nonunion, trochanter migration, and trochanteric bursitis caused by the prominent implants [7,9,[14][15][16]. In our previous study, the modified Hardinge approach was proposed to avoid these morbidities [17]. In the present study, we further demonstrated that the modified Hardinge approach and trochanteric flip osteotomy can achieve similar reduction qualities and clinical outcomes assessed by the Merle d'Aubigné-Postel score. We also showed that there was also no significant difference in operative time between the two procedures. Though the estimated blood loss was not statistically different, the blood loss of the osteotomy group was nearly double that of the modified Hardinge group-a difference which is likely due to the bony procedure of osteotomy.
The risk of ONFH following surgical dislocation of the hip is always a concern. Ganz et al reported no incidence of osteonecrosis in 213 hips treated with trochanteric flip osteotomy with surgical dislocation of the hip in non-traumatic patients [7]. This appears to be significantly different in traumatic cases, where the rate of osteonecrosis in acetabular and FH fractures treated with the trochanteric flip osteotomy method has been reported to range from 0%-25% [6, 8-11, 25, 29-31]. The overall incidence of ONFH in our series was 25% (5 out of 20). Even though T-shape or Z-shape capsulotomy was performed delicately and electrocautery was used meticulously in order not to jeopardize the blood supply of the FH, our incidence of ONFH was slightly higher than other studies. Notably, the incidence of ONFH in the modified Hardinge group (30.8%) was higher than that in the osteotomy group (14.3%), although this difference is not significant (P = 0.613). In the modified Hardinge group, one patient had only stage II ONFH, and one patient with stage IV ONFH had presented 6 days after trauma with a persistently dislocated hip.
Compromise of the FH blood supply is related not only to the surgery itself, but also to the surgical timing and injury pattern. The risk of ONFH after modified Hardinge approach or trochanteric flip osteotomy may require further study to clarify, with larger case series and prospective study designs.
Conversion to THR after Pipkin type IV FH fracture is considered a treatment failure by many investigators' definitions [9,15,26,27,32]. Some authors have suggested THR as the primary treatment in acute settings, particularly for lesions involving femoral neck fracture in elderly patients [26,32]. However, insufficient and unstable bone stock make THR extremely difficult. Our series aimed to perform reduction and fixation surgeries in the first attempt. Five of 20 cases (25%) required THR during the follow-up period. This result is in agreement with the literature, as a THR rate of 5-33% after Pipkin type IV FH fractures is generally reported [9,11,26,28,32].
Another undesirable complication of FH and acetabular fracture surgeries is HO. The reported incidence of HO following these surgeries ranges from 6%-38% [9-12, 15, 25, 28]. Indomethacin, bisphosphonates, or radiation have been recommended for the prevention of heterotopic bone formation, although these treatments may have a negative impact on fracture healing [16,[33][34][35]. Prophylaxis was not administrated in our series due to the concern of fracture nonunion, and the results revealed a low incidence of HO (5%, only one out of 20 cases). We postulated that the incidence of HO may be related to racial differences, as most investigations which report a high incidence of HO were conducted in western countries [9][10][11]15], while low incidence has been reported for studies conducted in Asian countries [25,28].
The present investigation has some potential limitations which should be noted. First, due to its retrospective nature, this study may have selection bias, although our inclusion criteria were not markedly different from those of previous studies. Second, this study was single center-and single surgeon-based. Therefore, the reproducibility of procedures may be questionable. Given that our institution is one of the largest trauma centers in Taiwan