Percutaneous autologous impaction bone graft for advanced femoral head osteonecrosis: a retrospective observational study of unsatisfactory short-term outcomes

Background Half of osteonecrosis of the femoral head (ONFH) patients suffer femoral head collapse at initial diagnosis, and more than half are bilaterally affected. This study developed a percutaneous autologous impaction bone graft (IBG) technique as a modification of core decompression (CD). We also summarized the short-term results and treatment efficacy of percutaneous autologous IBG in advanced ONFH. Methods Twenty patients (12 males, 8 females) with nontraumatic, postcollapse ONFH except one case underwent CD (10-mm core diameter) and reverse IBG. Radiological changes of the ONFH stage and type were analyzed. Survival analysis using Kaplan–Meier estimates was performed with conversion to total hip arthroplasty (THA) as the endpoint. In addition, the Harris hip score (HHS) and University of California, Los Angeles (UCLA) activity rating scale were evaluated. Results Percutaneous autologous IBG was performed successfully, with an average operation time of < 1 h and small blood loss, and 7 patients (35%) needed conversion to THA at an average of 17 months postoperatively. We observed radiological progressive change in 60% of the patients during a mean observation period of 3 years. The mean clinical scores, except data recorded, after THA significantly improved (before vs. after 3 years: UCLA activity score, 3.7 vs. 5.2 [P = 0.014]; HHS, 57.6 vs. 76.5 points [P = 0.005]). In addition, 6 patients showed radiological progression but no clinical deterioration. Conclusions Percutaneous autologous IBG was technically simple and minimally invasive, but short-term results were unsatisfactory for advanced ONFH. Indications for this procedure should be carefully examined to improve it in order to enable bone formation.


Background
Osteonecrosis of the femoral head (ONFH) is a multifactorial disease with unclear pathogenesis. It occurs at3 0 years of age and is the main reason for total hip arthroplasty (THA) in young patients. Patients undergoing systemic corticosteroid pulse therapy and people who abuse alcohol are at risk of ONFH [1][2][3][4][5]. The use of magnetic resonance imaging (MRI) helps in diagnosing ONFH at an early stage [1-3, 5, 6]. In the USA, the number of surgical procedures for ONFH in 2008 was double that in 1992 [7]. Because the long-term outcomes of THA for ONFH have improved [2][3][4]8], THA is selected as the first, not the final, surgical option [7]. However, ONFH patients are young, and THA is not an easy choice because of the risks of nondurability, infection, and dislocation of artificial implants [2]. Historically, in Europe and the USA, bone grafts and core decompression (CD) were performed to preserve joints [1][2][3][4][5][6][7], while in Japan and the rest of Asia, osteotomy was used [1][2][3][4][5][6]9]. In the 1940s, Phemister et al. [10] performed tibial bone graft from the lateral femur to the femoral head via a core, which has evolved over time [1][2][3][4][5][6]. Since the 1960s, CD has also become popular [1][2][3][4][5][6][7]11]. With increasing use of THA, the development of regenerative medicine combining CD with cell transplantation, bone substitute, and growth factors has shown promise, especially for early to stage 2 ONFH, prior to femoral head collapse [1,2,[4][5][6], with THA being the final surgical option for ONFH. However, half of advanced (stage 3 or greater) ONFH patients already have femoral head collapse at initial diagnosis [12], and more than half are bilaterally affected [1,5,12]. Therefore, it is necessary to establish joint-preserving surgery for advanced ONFH. This study developed a percutaneous autologous impaction bone graft (IBG) technique as a modification of CD and the Phemister method. The femoral bone obtained from CD was itself used as a reverse autologous bone graft. We also summarized the short-term (3-year) results and treatment efficacy of percutaneous autologous IBG in advanced ONFH.

Ethics
This retrospective observational study was approved by the institutional review board of Kyoto University Graduate School and Faculty of Medicine. The study was conducted in accordance with the World Medical Association Declaration of Helsinki. All patients provided written informed consent for the procedure and publication of study results.

Patients
We enrolled 20 patients (20 hip joints; 12 men and 8 women; mean age, 41 years [range 18-70 years]) with advanced ONFH except one case in this study. The inclusion criteria was nontraumatic ONFH patients (stage 2 and stage 3) who requested and accepted for the jointpreserving surgery. The key exclusion criteria were traumatic ONFH and patients with stage 4 diagnosed by magnetic resonance imaging (MRI) in accordance with the 2001 Japanese Investigation Committee (JIC) guidelines [13]. Stage 3 shows radiological femoral head collapse but no acetabular deformity. Of the 20 patients, 12 were associated with corticosteroid use, 5 were associated alcohol intake, 2 were associated corticosteroid use and alcohol intake, and 1 was idiopathic. As for JIC classification, 1 had stage 2, 17 had stage 3A, and 2 had stage 3B ONFH; 2 patients had type C1 and 18 had type C2 lesions; and 10 patients had bilateral morbidity and 3 had already undergone THA. Table 1 shows patient demographics at the baseline, along with key results 3 years postoperatively.

Surgical procedure
We performed percutaneous autologous IBG on the 20 patients from December 2016 to January 2018. Each patient was anesthetized and placed in the supine position on a fracture operating table equipped with a C-arm image intensifier. A 3-5-cm longitudinal skin incision was made distal to the lateral aspect of the femur near the level of the lesser trochanter, and the femoral fascia and vastus lateralis muscle were split to reach the lateral part of the femoral cortex, the CD entry site. The approach was very similar to inserting a lag screw in γ-nail surgery for femoral neck fracture. On the basis of preoperative planning, using C-arm images, CD was carefully directed toward the center of the ONFH area. Next, using a soft-tissue protector, a 10-mm-diameter trephine was inserted from the femoral neck to the ONFH target site while checking two directions of fluoroscopy. Then, the trephine was pulled out, and the cylindrical bone inside it removed. IBG was performed using this cylindrical autologous bone in the opposite direction so that the healthy part was implanted in the ONFH area; fluoroscopy was performed to reconstruct the femoral head collapse (Fig. 1). For bone defects post-CD, we used two types of cylindrical artificial bones: 9-mm cylindrical hard hydroxyapatite (Neobone®, Aimedic MMT Co., Ltd., Tokyo, Japan) and 10-mm cylindrical unidirectional porous β-tricalcium phosphate (Affinos®, Kuraray Co., Ltd., Tokyo, Japan). The former artificial bone was used for the CD entry site, and the latter artificial bone was used for the proximal part of the femur.
In all 20 patients, surgery was performed as scheduled, with no problematic complications such as intraoperative fractures or postoperative infection. In 2 patients (cases 5 and 7), simultaneous bilateral surgeries were performed.

Rehabilitation
Postoperative rehabilitation was performed under the guidance of a physical therapist with non-weight-bearing gait training from day 1 to 3 weeks postoperatively. Postoperative gait training included one-third partial weight bearing (PWB) at 3 weeks, half PWB at 4 weeks, twothirds PWB at 5 weeks, and full weight bearing at 6 weeks postoperatively.

Clinical assessment
We monitored perioperative complications related to surgery and assessed two clinical scores at the baseline and every 6 months postoperatively: the University of California, Los Angeles (UCLA) activity rating scale (1 [no physical activity] to 10 [regular participation in impact sports]) [14] and the Harris hip score (HHS) (0-100.025 points, with higher scores indicating better Both ONFH sides were operated simultaneously. THA was performed first, followed by bone grafting on the opposite side using the collapsed femoral head removed during THA c The total volume loss including THA on the opposite side outcomes) [15]. The physical therapist assessed the patients' walking capability, including passive hip range of motion as an HHS subcategory. We also performed survival analysis using the Kaplan-Meier method with conversion to THA as the endpoint.

Radiographic assessment
We performed radiological evaluation based on the JIC ONFH classification [13]. At the time of the patients' visit and every 6 months postoperatively, we took X-rays in two directions of the bilateral hip joint at the baseline. Computed tomography (CT) and MRI were also performed at the baseline. In addition, we evaluated a change in the JIC stage (1, 2, 3A, 3B, 4: early to secondary osteoarthritis via femoral head collapse) and JIC type (A, B, C1, C2: small to large) by X-ray.

Statistical analysis
All statistical analyses were performed using JMP Pro 14 (SAS Institute, Cary, NC, USA). Clinical scores were analyzed using a repeated-measures linear mixed effects model [16]. The cumulative survival rate with THA as the endpoint was evaluated using the Kaplan-Meier method. Scores measured after THA conversion were excluded from clinical score analysis. P < 0.05 was considered statistically significant.

Results
The mean observation period was 36 months (range 29-42 months). The average surgery time was 54.5±21.6 min and tended to be longer in earlier stages of ONFH. The average blood loss was 42.7±83.0 g. At follow-up, 7 patients (35%) were converted to THA; the conversion was as trouble-free as normal primary THA. The survival rate with conversion to THA as the endpoint was 65% at 3 years postoperatively. Figure 2 shows results of Kaplan-Meier survival analysis with conversion to THA as an endpoint. The average time to THA was 17 months postoperatively (range 12-29 months). Radiological progression was observed in 12 patients (60%), including the 7 patients converted to THA, while 8 patients showed no stage change since surgery. Two patients (cases 5 and 7) underwent simultaneous surgery with unilateral THA and unilateral IBG surgery. One Fig. 1 Surgical procedure, as shown in pre-and postoperative radiographs and intraoperative photographs. a Representative intraoperative fluoroscopic image at CD. b Photograph of a cylindrical bone taken by CD. The bloody zone is healthy, and the yellowish-white part on the right is the osteonecrotic part. Reverse IBG was performed. Postoperatively, the gap was repaired. Core defects in the healthy zone were supplemented with artificial bone avoiding fracture. c, c' Changes on AP X-ray. d, d' Changes on LAT X-ray. Downward yellow arrows indicate the gap made by femoral head collapse, while the upward yellow arrow indicates repair of the femoral head sphericity by IBG. CD, core decompression; IBG, impaction bone graft; AP, anteroposterior; LAT, lateral patient (case 11) showed femoral head collapse and osteosclerosis in the weight-bearing area, but the clinical course was good (Fig. 3). In contrast, case 20 showed bone resorption and bone cysts within the femoral head, leading to collapse and needing conversion to THA (Fig. 4) Distinguishing the preoperative stage with or without femoral head collapse, Table 3 compares the results of this study to other surgical methods. The joint preservation rates of different CD-based surgical treatment options range from 13.0 to 92.8% in postcollapse ONFH [18][19][20][21][22][23][24]. The short-term results of this study do not compare favorably with other treatments. There are several possible reasons. First, the quality of bones collected by CD was variable. Although not assessed the volume of the collected bone intraoperatively, the amount of bone volume might be small for IBG. Cancellous bone quality is unpredictable preoperatively, and additional autogenous bone collection from the iliac bone should be considered. Surgery with only artificial bone grafting has poor outcomes. Liu et al. [24] implanted betatricalcium phosphate (β-TCP) in 23 advanced ONFH patients, of whom 20 (87%) were converted to THA at 5 years postoperatively. Allograft might still be better than artificial bone for ONFH, although its efficacy should be tested.
Second, only autologous bone grafting might be insufficient. In the past 20 years, regenerative medicine has been developed for early-stage ONFH. The use of cells and growth factors has shown good results. Regenerative medicine can be applied to advanced cases, too. Wang et al. performed bone grafting by the light-bulb method combined with demineralized bone matrix with recombinant human bone morphogenetic protein (rhBMP) in advanced ONFH and reported an 88.7% (63/71) survival rate at 2 years postoperatively [19]. We have also reported good results of regenerative therapy for earlystage ONFH with recombinant human fibroblast growth factor-2 (rhFGF-2) [25]. Studies have reported low efficacy of cell therapy in advanced ONFH patients (Table 3). Tomaru et al. [23] reported a 43.2% (16/37) survival rate at 12 years postoperatively using bone marrow mononuclear cell (BMMNC) transplantation for advanced ONFH [23]. Wang et al. also reported a 66.6% (6/9) survival rate with BMMNC therapy at 2.3 years postoperatively [19]. This may be because glucocorticoid-induced ONFH patients have decreased angiogenic and increased apoptotic activity [26,27]. Therefore, for the patients with glucocorticoid-induced ONFH, adjunct growth factors such as rhBMP might be more useful than the BMMNC therapy.  Third, the timing of weight bearing in postoperative rehabilitation is an issue. One-third PWB gait training started at 3 weeks postoperatively might be too early. Other techniques have postoperative weight-bearing restrictions for 1 or 2 months. It is necessary to establish a system to determine when gait training can begin.
Fourth, bilateral ONFH is difficult to treat. The bilateral ONFH incidence is as high as 75% [28]. A multicenter study reported that of 310 ONFH patients (505 hips), 62.9% (195/310) had bilateral ONFH [11]. While there is no definite surgical plan, ONFH progresses and eventually both sides might need THA. There is no consensus on management for different stages of bilateral ONFH.
Fifth is the mode of femoral head collapse related to the postoperative clinical score. The relationship between the mode of femoral head collapse and symptoms should be considered. Ito et al. [29] reported that sclerotic changes around ONFH show trabecular hypertrophy, slow bone resorption progression, and have a good clinical course. It is necessary to establish a Data of the upper row are expressed as mean ± SD; middle row, 95% CI; and bottom row, median (range). The P-value was calculated for the effect of time in a repeated-measures linear mixed-effects model. The HHS ranged from 0 to 100.0, with a high positive value indicating a more functional hip joint. The UCLA activity rating scores ranged from 1 to 10, with 10 indicating "regularly participate in impact sports" and 1 indicating "wholly inactive, cannot leave residence" HHS Harris hip score, UCLA University of California, Los Angeles, SD standard deviation, CI confidence interval  [7]. This increase is accounted for by THA and hip resurfacing, with > 90% of surgeries being performed with artificial prostheses. Long-term THA outcomes for ONFH have been improving [7,8]. Therefore, THA is readily selected as the final surgical option for advanced ONFH. However, if THA is selected as the final therapeutic solution in ONFH patients who underwent osteotomies with metal plates [9] or CD with tantalum rods [30], conversion to THA is difficult [1][2][3][4] and therefore on the decline [2,6,7]. We believe that for jointpreserving surgery, conversion to THA should be performed without difficulty.
This study has some limitations that must be noted. This was a retrospective study, with a small number of patients. The postoperative radiological assessment was used by X-ray for evaluating ONFH stage. But the results of the postoperative CT or MRI analysis are needed to evaluate bone regeneration on a radiological base. Further well-designed, high quality studies are required for future research.

Conclusions
We developed and performed percutaneous IBG surgery for advanced ONFH. Although the surgical technique was simple and acceptable, the short-term results were not satisfactory. The clinical indications for this method will require caution. There are issues to be remedied such as the use of additional autologous bone, adjunctive regenerative therapy, and improvement of the rehabilitation therapy.