A pilot study of regenerative therapy using controlled release of recombinant human fibroblast growth factor for patients with pre-collapse osteonecrosis of the femoral head

The treatment protocol was approved by the Ethics Committee of Kyoto University Graduate School and Faculty of Medicine. Before enrollment, written informed consent was obtained from each participant included in the study. The study was registered with the public Japanese clinical trials registry, the University Hospital Medical Information Network Clinical Trials Registry (UMIN000009250).

We designed this phase II, prospective, open-label clinical trial to evaluate the safety and clinical outcomes of a single local administration of rhFGF-2-impregnated gelatin hydrogel as a treatment for the precollapse stage of ONFH. The study was conducted at a single institution over 12 months, with patients enrolled and treated between March 1, 2013, and March 27, 2013, with follow-up ending on June 19, 2014.

The eligibility criteria were age between 20 and 80 years and presence of ONFH at precollapse stage 1 or 2 according to the classification system for ONFH staging developed by the Japanese Investigation Committee of Health and Welfare (JICHW) [10]. Factors underlying ONFH were determined as associated with steroid use, associated with alcohol intake, or idiopathic.

Patients were placed in the supine position on a fracture operation table with a C-arm image intensifier. A 1-cm skin incision was created over the lateral aspect of the femur near the level of the lesser trochanter, and 800 μg of rhFGF-2-impregnated gelatin hydrogel was embedded percutaneously. In brief, under fluoroscopic control, a 2.4-mm Kirschner guide wire was inserted in the direction of the ONFH lesion. The first drilling was carefully directed to the center of the ONFH region using a C-arm image based on preoperative planning. With a soft-tissue protector in place, core decompression was performed via the guide wire with a 4.8-mm diameter trephine. A 3.5-mm cylinder was inserted into the core decompression site, and the hydrogel was inserted in the ONFH region using a special pusher (Iso Medical Systems, Tokyo, Japan).

Weight bearing was prohibited on the day of surgery; however, walking was encouraged under the guidance of physical therapists on day one post-operatively. Before treatment, the physical therapist assessed walking capability, including passive hip motion, in all participants.

Recombinant hFGF-2 was provided by Kaken Pharmaceutical Co. (Tokyo, Japan). For controlled-release, biodegradable gelatin hydrogel was prepared using glutaraldehyde cross-linking of acidic gelatin purified from natural bovine bone, as reported previously [11, 12].

Safety was monitored by the Department of Clinical Trial Design and Management Translational Center according to the occurrence and severity of adverse events (AEs). AEs were defined as any sign of worsening of the patients’ condition after treatment, and they were classified as either serious or non-serious. AEs requiring hospitalization were classified as serious and were reported and assessed by an external data monitoring committee, with the severity classified as mild, moderate, or severe. Causal relationships between the clinical trial and resolution of each AE were judged and evaluated.

All participants underwent preoperative clinical assessment, and they were prospectively followed-up clinically and radiographically. Patients were evaluated at day one, three, six, and 12 months post-operatively. The primary endpoint was safety assessment, and the secondary endpoints were changes in the clinical scores, radiographic stage of ONFH, and bone regeneration as evaluated by radiography, computed tomography (CT), and MRI.

The following clinical score tools were used at baseline (pre-operative) and at three, six, and 12 months (post-operatively). Patients were evaluated using the visual analog scale (VAS) for pain, the University of California, Los Angeles (UCLA) activity score [13], and the Harris hip score (HHS) [14]. VAS for pain is a self-assessment scale for patients, which ranges from 0 mm (no pain) to 100 mm (maximum pain). The UCLA activity score uses a simple scale ranging from 1 to 10. A rating of 1 is defined as “no physical activity, dependent on others” and 10 is defined as “regular participation in impact sports.” HHS ranges from 0 to 100.025 points, with higher scores indicating better outcomes, and has the following subcategories: pain (44 points), limp (11 points), support (11 points), distance walked (11 points), sitting (5 points), stairs (4 points), putting on shoes and socks (4 points), absence of deformity (4 points), and range of motion (5.025 points). HHS for range of motion was assessed by a physical therapist. For patients with bilateral hip involvement, both hips were examined.

Serial anteroposterior and frog lateral radiographs of the affected hip were used to examine radiographic progression of the femoral head collapse in hips (before surgery to one year follow-up) according to each JICHW classification stage. The JICHW stage and type classifications were used to describe ONFH and localization of the necrotic lesion [10]. During pre-operative planning, the drilling route to the necrotic lesion was planned using navigation software (Orthomap 3D, Stryker, Kalamazoo, MI, USA; Fig. 1). This software was used to evaluate bone regeneration at the implanted area of the femoral head. The fraction of osteonecrotic volume to the whole femoral head was calculated using integrated data from CT and MRI scans. To clarify the location of bone regeneration in CT images, further radiological analysis (not included in the protocol of the clinical trial) was performed. The coronal CT images obtained pre-operatively were compared with those obtained at the one year post-operative follow-up after converting into a binary data format using Image J software (U.S. National Institutes of Health, Bethesda, MD, USA). For evaluating binary format data, thresholds for each specimen were determined at the same condition used to clarify if and where the bone regenerated. The difference in the two binary format images was defined as the regenerated bone region, as described previously [9].

All statistical analyses were performed on an intention-to-treat basis. Safety and efficacy analyses were conducted on all patients who received treatment. The incidence of AEs and baseline characteristics were described for each patient. The changes over time in the VAS for pain scores, UCLA activity scores, and HHSs were summarized and investigated using a repeated-measures linear mixed-effect model containing terms for time with a compound symmetry covariance structure [15]. The effect of time was analyzed with a one-tailed t-test. Advances in the disease stage and the incidence of femoral head collapse were evaluated for each patient by radiographic evaluation. Statistical analyses were performed using SAS version 9.3 software (SAS Institute, Inc., Cary, NC, USA). All data are reported as mean ± standard deviation. The linear mixed-effect model was used for statistical analyses, unless indicated otherwise. All P values <0.05 were considered significant.