The influence of adipose-derived stromal vascular fraction cells on the treatment of knee osteoarthritis

This clinical study was designed to evaluate the safety, feasibility, and efficacy of autologous SVF cells in patients with knee OA. The grade of knee OA was evaluated by the Kellgren-Lawrence (KL) classification, and all patients with grades I to IV OA participated in this study. The study protocol conformed to the Declaration of Helsinki and was approved by the appropriate ethics committees. All patients provided informed consent prior to participation.

The inclusion criteria were (a) patients diagnosed with knee OA at any age, (b) exhibiting substantial pain and loss of function, (c) ineffectiveness of conservative treatment including rehabilitation, medication, and intra-articular injection of hyaluronic acid or steroids, and (d) written informed consent. The exclusion criteria were (a) severe bony defect seen on preoperative radiographs, (b) previous knee injury requiring operation, (c) active or previous knee joint infection, and (d) a serious past history, such as systemic inflammatory diseases and vascular changes.

Patients were asked to perform daily home exercises by themselves according to a standardized rehabilitation protocol of the hospital after treatment, in addition to rehabilitation by a physical therapist through regular hospital visits.

The Celution® 800/CRS system (Cytori Therapeutics Inc., San Diego, CA) was used to extract SVF cells from the patient’s abdominal or breech subcutaneous fat. This system consists of two parts: one for tissue washing and digestion, and the other for cell concentration. All subjects underwent a liposuction procedure to obtain 100–360 mL of adipose tissue under general anesthesia; the extracted tissue was then processed using the Celution® 800/CRS System according to manufacturer instructions. Briefly, the tissue was washed to remove blood and debris. Celase® GMP, which was a mixture of highly purified collagenase and neutral protease enzymes, was then added and incubated at ~ 37 °C for 20 min with continuous mixing to digest the aspirated adipose tissue. After digestion, the SVF cells were concentrated using centrifugation and washed to remove the Celase® reagent. SVF cells were then extracted from the system and counted to prepare the specified dose in 5 mL of lactated Ringer’s solution. The whole system can be operated aseptically using clinical-grade solutions such as saline and lactated Ringer’s, and single-use Celution™ consumable sets. The SVF cell count and viability were determined at each investigational site using the NC-100™ NucleoCounter® Automated Cell Counting System (Chemometec, Allerod, Denmark). This system is an image cytometer based on fluorescence from the fluorescent dye, propidium iodide. When a sample is mixed with Reagent A100 and Reagent B, lysis of the viable cell membrane occurs, rendering all the cell nuclei susceptible to staining with propidium iodide facillitating in a total cell count. However, non-viable cells are permeable without treatment and are therefore stained directly with propidium iodide, resulting in a non-viable cell count. Thus, the cell viability of a sample is determined using the total cell count and the count of non-viable cells.

We administered an intra-articular injection of 2.5 × 10⁷ SVF cells to each patient according to the number of purified SVF cells and guidelines previously stated in a similar report [22]. Cell transplantation into the knee joint was performed without anesthetic and under echo guidance after aspiration if the joint fluid level was excessive.

The primary endpoint of this study was patient improvement based on clinical evaluations and scores. Clinical evaluations included knee range of motion (ROM) and muscle force of knee extension and flexion using a hand-held dynamometer. Clinical scores included Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), visual analog scale (VAS) for pain (0–100), Japanese Knee Osteoarthritis Measure (JKOM), and Knee injury and Osteoarthritis Outcome Score (KOOS). For measurement of muscle force of knee extension and flexion, patients were tested in a prone position with their knee at 45 degrees flexion. The hand-held dynamometer was placed at the center of their lower leg. The examiner asked subjects to bend their knee and hold for 3 s to measure hamstrings strength, and to straighten their knee and hold for 3 s to measure quadriceps strength. The examiner added resistance to maintain the knee at 45° and measured the displayed value as muscle strength. These tests were performed three times and the average value was recorded.

As a secondary endpoint, imaging evaluations, which included the hip-knee-ankle (HKA) angle assessed via radiography, and T2 mapping value using a 1.5-T magnetic resonance imaging (MRI) unit (Sigma Exite HDx; GE Healthcare, Waukesha, Wis) [23, 24] were also assessed. The method of calculating the T2 mapping value is as follows; we selected a central slice that passed through the center of the weight-bearing cartilage surrounded by the anterior and posterior margins of the meniscus on a sagittal slice of T1-weighted fast-field echo images. In addition to the central slice, we added two slices neighboring the central slice anteriorly and posteriorly (Fig. 2a). The region of interest (ROI) was then set at the weight-bearing full-thickness cartilage of the medial and lateral femoral condyle and medial and lateral tibial plateau on the central slice of the coronal image (Fig. 2b). The ROI was also set using the same method on both the anterior and posterior slice. Overall, the T2 mapping values of 12 ROIs were measured. According to this analysis, the lower the T2 mapping value, the lower the degree of articular cartilage degeneration.

Both clinical and imaging evaluations were performed preoperatively and at 1, 3, 6 and 12 months postoperatively after intra-articular injection of SVF cells. Clinical evaluations were performed by an independent experienced physiotherapist. Image analyses were performed by an independent orthopedic surgeon with 15 years of experience in MRI analysis of knee OA features. For safety evaluations, incidence, severity, and outcome of all adverse events were recorded.

All values were expressed as mean ± standard deviation. Results were analyzed using a statistical software package (Statview 5.0; Abacus Concepts, Inc., Berkeley, CA, USA). Clinical and imaging evaluations were compared between the five time periods using repeated measures analysis of variance. Furthermore, we evaluated the clinical scores preoperatively and at 12 months postoperatively, and investigated the improvement rate of clinical scores from preoperatively to 12 months postoperatively among the KL classification by using repeated measures analysis of variance. P < 0.05 was considered statistically significant. A statistical power analysis was performed prior to the study, which was expected to require a power of 0.8, based on a prespecified significance level of α < 0.05 and assuming a medium effect size (effect size = 0.30) using G power 3 [25]. The estimated sample size was 45 patients.

The mean ROM improved from a baseline of − 6.0°–131.3° to − 4.8°–133.9° at 1 month, − 4.3°–134.3° at 3 months, − 3.7°–134.5° at 6 months, and − 3.5°–132.6° at 12 months postoperatively. The improvement in the mean extension angle from baseline to 6 and 12 months was statistically significant. Muscle force of knee extension and flexion improved from a baseline of 202.5 Nm and 99.5 Nm, respectively, to 198.9 Nm and 108.2 Nm at 1 month, 219.0 Nm and 116.6 Nm at 3 months, 235.4 Nm and 124.2 Nm at 6 months, and 261.9 Nm and 126.8 Nm at 12 months postoperatively, respectively. The mean muscle force of knee extension and flexion was significantly better at 12 months postoperatively than preoperatively.

The improvement in the mean total WOMAC scores from baseline to 12 months postoperatively was from 33.4 to 22.6 points, which showed a statistically and clinically significant difference (Table 2). The improvement in VAS scores from baseline to 12 months postoperatively was from 46.5 to 32.8 points, which showed a statistically and clinically significant difference (Table 2). An improvement was also evident in the mean total JKOM scores from baseline to 12 months postoperatively (from 34.9 to 26.8 points), which showed a statistically significant difference (Table 2). Improvements in the KOOS score were also evident, including an average score of 5 subscales and all subscale scores 12 months postoperatively (Table 2, Additional file 1). The improvements from baseline in the mean scores of pain, symptoms, activities of daily living, sports, and quality of life subscales were from 53.1 to 66.3 points, 57.8 to 67.0 points, 70.0 to 77.5 points, 27.6 to 37.5 points, and 33.6 to 44.6 points with statistical significance, respectively. The improvements in the mean scores of pain, symptoms, sports, and quality of life subscales also showed clinically significant differences (Additional file 1).

According to the KL classification, there were no grade I patients in this study. There was no significant difference in preoperative clinical scores of WOMAC, VAS, JKOM, and KOOS among the KL classifications. Clinical scores of WOMAC, JKOM and KOOS at 12 months postoperatively were significantly better for grade II than for grade III. Furthermore, clinical scores of WOMAC, VAS, and JKOM at 12 months postoperatively were also significantly better for grade II than for grade IV. The improvement rate of WOMAC from baseline to 12 months was significantly better for grade II than for grade III. There was no significant difference in improvement rates of VAS, JKOM, and KOOS among the KL classifications (Table 3).

The mean HKA angles changed from a baseline of 6.7° to 7.1° at 1 month, 6.7° at 3 months, 6.6° at 6 months, and 6.8° at 12 months postoperatively. However, this change was not statistically significant at any time. In contrast, the mean T2 mapping values of the lateral femur and tibia in the anterior areas at 12 months postoperatively were significantly lower than those preoperatively. The mean T2 mapping value of the medial tibia in the central area at 6 and 12 months postoperatively was also significantly better than that preoperatively. Furthermore, the mean T2 mapping value of the lateral femur in the posterior area at 12 months postoperatively was significantly better than that preoperatively (Table 4, Fig. 4).

Neither deaths nor life-threatening adverse events were observed during the 12-month follow-up after cell therapy. Furthermore, there was no mild to moderate adverse event such as swelling, local heat of the knee, or infection during follow-up.