Use of autologous bone marrow stem cell implantation for osteonecrosis of the knee in sickle cell disease: a preliminary report

Osteonecrosis of the knee (OK) secondary to sickle cell disease (SCD) is a debilitating disorder that may progress to subchondral collapse and end-stage arthritis of the knee at an early age. Intravascular erythrocyte sickling and vaso-occlusion of microcirculation lead to ischemic death and necrosis of the constituents of the bone and marrow. In its early stage, the overlaying articular cartilage is intact, however, the changes in the subchondral bone may rapidly progress to structural collapse, chronic pain and joint destruction [1‐3]. Advanced OK-SCD eventually requires knee arthroplasty, which is associated with increased risk of adverse outcome and higher revision rates in sickle cell disease patients [4]. For these reason, the early treatment is very important.

OK-SCD is usually part of a multifocal disease, and the knee is a particularly vulnerable joint in sickle cell disease. There have been a limited number of OK-SCD case reports and a high occurrence of osteonecrosis of the knee in asymptomatic patients [2, 5], therefore the exact occurrence of OK in sickle cell disease patients remains unclear. Flouzat-Lachaniete et al. [6] reported in a large study that 34% of SCD patients diagnosed with multifocal osteonecrosis had involvement of the knee. They also observed an annual incidence rate of 3.6 cases per 100 patients, which makes joint-preserving techniques important to initially attempt.

Treatment options for osteonecrosis of the knee in SCD patients are limited. For symptomatic patients at an early stage, many conservative procedures are used to enable knee preservation instead of replacing it [7]. However, the efficacy of these joint-preserving techniques, including core decompression or debridement of the osteonecrotic lesion, has been controversial in SCD and many patients still progress to advanced osteonecrosis [8‐10]. Therefore, even if positive results are obtained, the treatment of the OK-SCD continues to be a daunting problem for orthopedic surgeons.

Cell therapy with autologous implantation of concentrated bone-marrow mononuclear cells (BMC) has been used successfully for treatment of hip osteonecrosis [11, 12]. Although no reports exist on the implantation of BMC in the knee in SCD patients, several authors described that BMC implantation can significantly decrease the pain and other joint symptoms caused by osteonecrosis and delay, or avoid the progress of this disease toward higher osteonecrotic stages [13, 14]. The rationale for the use of cell therapy for osteonecrosis in sickle cell disease patients relies on the presence of bone marrow-derived mesenchymal stromal cells (MSCs) and endothelial progenitors cells (EPCs) in BMC, which may enhance the mechanism of blood vessel formation and improve tissue repair [12, 15]. Other nucleated cells (macrophages, T cells, B cells, dendritic cells, natural killer cells and neutrophils) are also present in BMC and once introduced into the osteonecrotic lesion may help improve pain and function [13‐15]. After encouraging results were obtained in the hip, the same technique has been transferred to the knee joint in sickle cell disease patients.

In this prospective pilot study we aimed to evaluate safety, feasibility, clinical and radiological results after minimally invasive implantation of autologous bone marrow mononuclear cell (BMC) in OK lesions secondary to sickle cell disease.

At presentation, 21 patients (63.6%) had unilateral osteonecrosis (the right knee was affected in 9 cases and the left knee in 12 cases) and 12 patients (36.4%) had bilateral osteonecrosis. OK was stage I in 16 patients (21 knees) and stage II in 17 patients (24 knees), according to classification system (Table 1). Osteonecrotic lesions were predominantly located in the medial femoral condyle (19 knees, i.e. 42.3%) followed by the lateral femoral condyle (Table 2). Examination of preoperative MRI consistently revealed subchondral areas of early-stage osteonecrosis and diffuse bone marrow edema (Figs. 1 and 2).

Overall, the interventional procedure was well tolerated by SCD patients, with no complications or serious adverse event (deaths, hematoma, nerve injury and others). Only four episodes of small adverse events (site infection following cell injection, and transient local pain) were noted. No serious adverse events or failure led to treatment discontinuation or study termination. The mean observation follow up period was 27.3 months (range: 12–60).

All SCD patients suffered from symptomatic knee joint and limitation of knee motion at the time of enrollment. Pre-operative KSS was poor in 8/33 patients (24.2%) and fair in 17/33 patients (51.6%) whereas KFS was poor in 31/33 patients (94%) and fair in 2/33 patients (6%) (Table 3). After the percutaneous BMC implantation improvements were observed in both KSS and KFS rating systems compared to the preoperative state. After surgery, 97% (32 out of 33 patients) had excellent outcome in KSS score whereas the KFS score was good in three patients (9%), and excellent in the remaining thirty patients (91%). The mean KSS (and standard deviation) was 64.3 ± 9.7 (range: 45–80) before surgery and 92.2 ± 4.1 (range: 84–100) at a mean of 27.3 ± 13.8 months (p < 0.001), while the KFS was 44.5 ± 8.0 (range: 30–55) before surgery and 91.6 ± 5.8 (range: 80–100) at a mean of 27.3 ± 13.8 months (p < 0.001) (Fig. 3a, b) (Table 3). KSS and KFS evaluation preoperatively and at the time of follow-up showed clinical improvement was maintained over time (Fig. 4a, b). However, no correlation was seen between change in clinical scores (changes in KSS or KFS) and gender (p = 0.02; 0.008), stage of the osteonecrosis (p = 0.26; 0.11), genotype of the sickle cell disease (p = 0.35; 0.12), and presence of bilateral knee osteonecrosis (0.015; 0.10), respectively.

At beginning, all 33 patients had moderate or severe pain usually localized, occurring both at motion, weight-bearing and at rest, causing significant functional impairment in all cases. After BMC implantation, pain decreased was observed in 87.8% (29/33 patients) and remained unchanged only in 4 (12.2%) patients. The mean NRS pain score decreased from 6.7 ± 1.2 (mode: 7; range: 4–9) to 4.5 ± 0.8 (mode: 4; range: 2–5) after BMC implantation (p < 0.001) (Fig. 3c) (Table 3). Among 33 patients, 29 (87%) were highly satisfied with the surgical outcome, three (13%) were moderately satisfied. The mean NRS satisfaction score obtained was 9.2 at the most recent follow-up.

The total amount of BMC-nucleated cell ranged from 10 to 37 × 10⁶ cell/ml (range: 4.5–16.7 × 10⁸ nucleated cell/patient), including < 1.2% CD34+ CD45^dim positive cells, which are precursors of hematopoietic cells, and 2.1 +/− 1.4 × 10⁴ cells of CFU-F as an indicator of mesenchymal stromal cell activity. These results show that up to 19.9 × 10⁶ CD34+ CD45^dim progenitor cells were injected through and distributed into the ostenecrotic zones.

In contrast to the improved clinical outcome, radiographic signs did not change appreciably after BMC implantation. We could not perceive any sign of subchondral fracture or collapse of the femoral condyles or the tibial plateau at the follow-up times after intervention. At the latest radiographic evaluation, any significant subchondral bone changes of radiographic signs, including presence of crescent sign, increased thickness, density, and irregularity were noted. Postoperative MRI scans demonstrated no subchondral bone erosion or fracture, normal contour of the epiphyseal head and preservation of the joint space. Most lesions that appeared stable on MRI were clinically also stable or improved. These results indicate that the joint space was preserved and the subchondral bone was stable after BMC implantation during follow-up period. Figs. 5 and 6 shows representative cases from our series.

The most important finding of the present study is that early stages of osteonecrosis of the knee in SCD patients achieved clinically meaningful and significant improvement after the implantation of BMC, compared to the preoperative situation. The results also demonstrated that BMC could be safely delivered in a concentrated manner to osteonecrotic lesions of the knee and no radiographic evidence of subchondral collapse at midterm follow-up (average: 27.3 months) in SCD patients.

Osteonecrosis of the knee in sickle cell disease is most commonly seen in patients who are younger than 45 years of age, with typical bilateral knee involvement. The medial and lateral femoral condyles are relatively frequent sites for its occurrence [6, 7]. As with osteonecrotic hip in sickle cell disease, variable mechanisms may be implicated in the knee, including interruption of the vascular blood supply to the subchondral bone. If not treated, early stages of OK-SCD have a high likelihood of progression to symptomatic disease and subchondral collapse, as demonstrated for hip, shoulder and ankle [2, 23, 24]. In these situations, conservative treatment is generally not successful, leading to secondary arthritis and total joint replacement in approximately 80% of cases after 2–5 years in SCD patients [2, 13].

Joint preservation therapy is the immediate goal of treatment considering the prevalence of OK-SCD among young adult patients in our series. Instead, joint replacement is a controversial option considering the risk and reduced durability of total knee arthroplasty in SCD patients. Recently, Perfetti et al. [4] have shown that sickle cell disease patients admitted for total knee arthroplasty have a 137% higher risk for perioperative complication than that of non-SCD patients. Moreover, joint replacement was associated with several complications in SCD patients: intra-operative bleeding, infections, and loosening or early loss of the prosthesis [25]. For a small and preliminary study as ours, in which the main focus was safety and feasibility, the results confirmed that our knee joint-preserving procedure was promising and had a limited rate of procedure-associated adverse effects. Therefore, we suggest that BMC implantation might be a valuable therapeutic joint preserving option for osteonecrosis of the knee in SCD patients.

Conventional treatments are unable to reverse or halt early stage osteonecrosis of the knee [26, 27] and its efficacy remains controversial for SCD patients [8, 9]. MRIs and morphological analysis have shown that no significant repair process in the osteonecrotic lesion could be achieved with core decompression alone [14], calling for additional alternative treatment options, usually with some form of bone grafting. In our series, we included SCD patients with OK in early pre-collapse stages, since its recognition is important for the long-term result of this procedure [11, 12].

Here, we treat osteonecrosis by implantation of autologous progenitor stem cells into the necrotic lesion. In sickle cell disease, implantation of BMC has been described for over twenty years with successful results for patients suffering with early stage osteonecrosis of the humeral and femoral head. [13]. This procedure has been used in the senior author’s practice for the past ten years for osteonecrosis of the femoral head, and the same technique has been transferred to the knee joint in sickle cell disease patients. The percutaneous minimally invasive technique in our study uses small-diameter marrow-needle as a modification of core decompression, proper patient immobilization and knee positioning, with minimal morbidity being associated with the procedure. Furthermore, surgery was performed carefully, with minimal trauma to avoid unexpected drilling through the articular cartilage surface or supracondylar femur fracture in OK-SCD patients.

Autologous concentrated bone marrow-based approaches for knee osteonecrosis have been limited to case reports or to osteonecrosis associated with other risk factors rather then sickle cell disease [27]. Recently, Goodman et al. [28] have shown that 12 young patients (fourteen knees) with OK associated to corticosteroid and submitted to local debridement before osteoprogenitor cell grafting had excellent pain relief and function, integration of the graft and no radiographic evidence of subchondral collapse at up to 5 years postoperatively. Marulanda et al. [26] reported a series of 61 knees (38 patients) with OK treated by a technique using multiple small percutaneous 3 mm drillings. A total of 56 (92%) had a successful clinical outcome, with 5 (8%) knees progressing to further surgery. Buda and colleagues [29] evaluated the midterm (average follow-up: 29 months) results of 30 patients with post-traumatic osteochondral lesions of the knee treated predominantly by arthroscopic transplantation of concentrated bone-marrow aspirate with a scaffold into the lesion site. They reported good to excellent results in all the patients based on the IKDC and KOOS scores, and growth of bone and cartilage, nearly complete defect filling and satisfactory integration of the graft at follow-up in 80% of cases [29]. Gobbi et al. [30] reported a long-term and sustained benefit of concentrated bone-marrow aspirate compared to microfracture for the treatment of osteochondral lesions of the knees. Zellner et al. [31] described promising clinical and radiological results, achieving successful bone regeneration of large and deep osteochondral defects of the knee joint after the treatment with concentrated bone-marrow aspirate combined with bone augmentation. In our study, arthroplasty was avoided in all the patients during the follow-up period. There was also significant improvement in the KSS, KFS and NRS pain score at the time of the latest follow up. Only 4 knees (out of 33) had persistent painful symptoms after treatment, which required pharmacological management. These outcomes are promising in that knee osteonecrosis can be safely and effectively treated with BMC implantation. Although these initial results seem quite encouraging, longer-term follow up is clearly needed.

Consistent with the previously published results from our group and others, significant improvement of pain was observed in knee joints following injection with BMC [12, 32]. Pain in the knee is often referred in patients with hip disorders [33]. However, patients were selected based on the knee pain, and the distribution of knee pain was not significantly different when patients had previous diagnosis of hip osteonecrosis. We suggest that SCD patients with knee osteonecrosis often present with pain because the main lesion is on the knee, especially in the medial femoral condyle (load transmitting area). Treatment of knee osteonecrosis with BMC showed only local improvement and no effect on the other joints, which is consistent with previous reports [30‐32].

Cell therapy efficacy for osteonecrotic lesions depends on the quantity of implanted viable BM nucleated cells and stem cells progenitors [13]. Hematopoietic/endothelial progenitors (CD34CD45^low), mesenchymal stromal cells (CFU-F) and nucleated cell into the osteonecrotic lesion may help by replenishing damaged joint structures and providing modulation of the immune response, thus alleviating the symptoms and progression of the disease [13, 14, 30]. However, the correlation between BMC implantation and clinical success is in open investigation. Prior studies investigating a relationship between the cellular dose contained within the bone marrow concentrate and efficacy of the treatment for knee osteoarthritis have shown that patients receiving a higher concentration of cells (< 10⁸ nucleated cell/patient) reported a better pain outcome in comparison with the lower dose group [32]. In our study, the cellular doses administered were adequately high above the BMC-nucleated cell indicated. Moreover, the frequency of hematopoietic/endothelial progenitors (CD34CD45^low) and mesenchymal stromal cells (CFU-F) were consistent with those in the literature, and the number of CFU-F that was injected was in accordance with previous clinical studies (between 25,000 and 200,000 CFU-Fs) [13, 14]. Therefore, despite a low number of sample sizes tested, these results suggest that there is no deficiency in the quantity of grafted SCD patient’s marrow stromal cell populations.

We are conscious that our experimental design has several limitations. The presented data collection is not a controlled study. Sickle cell disease is a systemic life-threatening hemoglobinopathy associated with several comorbidities, and there is no current standard treatment for osteonecrosis in these patients. Also, the inclusion of a control group (placebo group) was not in accordance with ethical recommendations of the national authorities that have analyzed and provided the financial support for this study. Quite clearly, prospective randomized studies including a comparison with an active control group would be nevertheless desirable, in order to determine more comprehensively the efficacy of the treatment described in this study.

Our study has several others limitations: 1) The number of patients is small and the follow-up short to intermediate term; 2) We selected only patients who had associated osteonecrosis of the hip. Because there may be patients with knee osteonecrosis who do not have hip osteonecrosis, this study population might not represent all sickle cell patients with osteonecrosis of the knee; 3) We cannot be sure that implanted BMCs remained at the site of implantation during the follow up period; 4) Another limitation is consequent to the fact that only a few patients underwent MRI during the follow-up because of the costs associated with imaging asymptomatic patients with sickle cell disease. Monitoring bone formation in the necrotic area or examine how BMC implantation may have affected outcomes should be incorporated in future studies.

Despite these limitations, implantation of BMC in young sickle cell disease patients with secondary osteonecrosis of the knee was relatively simple, promoted pain relief and functional improvement at midterm follow-up. This technique does not preclude further reconstructive operations, if necessary. This prospective pilot study suggest that the treatment strategy of BMC is a promising strategy and may be useful to treat early stage knee osteonecrosis in sickle cell disease patients without subchondral bone collapse and large degenerative changes of the cartilage.