Complete Spinal Cord Injury Treatment Using Autologous Bone Marrow Cell Transplantation and Bone Marrow Stimulation with Granulocyte Macrophage-Colony Stimulating Factor: Phase I/II Clinical Trial


  • Seung Hwan Yoon,

    1. Inha Neural Repair Center, Inha University College of Medicine, Incheon, Korea
    2. Departments of Neurosurgery, Inha University College of Medicine, Incheon, Korea
    Search for more papers by this author
  • Yu Shik Shim,

    1. Inha Neural Repair Center, Inha University College of Medicine, Incheon, Korea
    2. Departments of Neurosurgery, Inha University College of Medicine, Incheon, Korea
    Search for more papers by this author
  • Yong Hoon Park,

    1. Inha Neural Repair Center, Inha University College of Medicine, Incheon, Korea
    2. Inha Research Institute for Medical Sciences
    Search for more papers by this author
  • Jong Kwon Chung,

    1. Inha Neural Repair Center, Inha University College of Medicine, Incheon, Korea
    2. Anesthesiology, Inha University College of Medicine, Incheon, Korea
    Search for more papers by this author
  • Jung Hyun Nam,

    1. Diagnostic Pathology, Inha University College of Medicine, Incheon, Korea
    Search for more papers by this author
  • Myung Ok Kim,

    1. Rehabilitation Medicine, Inha University College of Medicine, Incheon, Korea
    Search for more papers by this author
  • Hyung Chun Park,

    1. Inha Neural Repair Center, Inha University College of Medicine, Incheon, Korea
    2. Departments of Neurosurgery, Inha University College of Medicine, Incheon, Korea
    Search for more papers by this author
  • So Ra Park,

    1. Inha Neural Repair Center, Inha University College of Medicine, Incheon, Korea
    2. Physiology, Inha University College of Medicine, Incheon, Korea
    3. Center for Advanced Medical Education by BK21 Project, Inha University College of Medicine, Incheon, Korea
    Search for more papers by this author
  • Byoung-Hyun Min,

    1. Department of Orthopedics, Molecular Science, and Technology, Ajou University Hospital, Suwon, Korea
    Search for more papers by this author
  • Eun Young Kim,

    1. Departments of Neurosurgery, Inha University College of Medicine, Incheon, Korea
    Search for more papers by this author
  • Byung Hyune Choi,

    1. Inha Neural Repair Center, Inha University College of Medicine, Incheon, Korea
    2. Inha Research Institute for Medical Sciences
    Search for more papers by this author
  • Hyeonseon Park,

    1. Inha Neural Repair Center, Inha University College of Medicine, Incheon, Korea
    2. Departments of Neurosurgery, Inha University College of Medicine, Incheon, Korea
    Search for more papers by this author
  • Yoon Ha M.D., Ph.D.

    Corresponding author
    1. Inha Neural Repair Center, Inha University College of Medicine, Incheon, Korea
    2. Departments of Neurosurgery, Inha University College of Medicine, Incheon, Korea
    3. Center for Advanced Medical Education by BK21 Project, Inha University College of Medicine, Incheon, Korea
    • Department of Neurosurgery, Inha University College of Medicine, 7-206, Sinheung-dong 3-ga, Jung-gu, Incheon 400-712, Korea. Telephone: +82-32-890-3449; Fax: +82-32-890-3099
    Search for more papers by this author


To assess the safety and therapeutic efficacy of autologous human bone marrow cell (BMC) transplantation and the administration of granulocyte macrophage-colony stimulating factor (GM-CSF), a phase I/II open-label and nonrandomized study was conducted on 35 complete spinal cord injury patients. The BMCs were transplanted by injection into the surrounding area of the spinal cord injury site within 14 injury days (n = 17), between 14 days and 8 weeks (n = 6), and at more than 8 weeks (n = 12) after injury. In the control group, all patients (n = 13) were treated only with conventional decompression and fusion surgery without BMC transplantation. The patients underwent preoperative and follow-up neurological assessment using the American Spinal Injury Association Impairment Scale (AIS), electrophysiological monitoring, and magnetic resonance imaging (MRI). The mean follow-up period was 10.4 months after injury. At 4 months, the MRI analysis showed the enlargement of spinal cords and the small enhancement of the cell implantation sites, which were not any adverse lesions such as malignant transformation, hemorrhage, new cysts, or infections. Furthermore, the BMC transplantation and GM-CSF administration were not associated with any serious adverse clinical events increasing morbidities. The AIS grade increased in 30.4% of the acute and subacute treated patients (AIS A to B or C), whereas no significant improvement was observed in the chronic treatment group. Increasing neuropathic pain during the treatment and tumor formation at the site of transplantation are still remaining to be investigated. Long-term and large scale multicenter clinical study is required to determine its precise therapeutic effect.

Disclosure of potential conflicts of interest is found at the end of this article.


It has long been believed that intrinsic repair is quite restricted after spinal cord injury (SCI) because neurogenesis rarely occurs in the central nervous system (CNS). As a result, cell transplantation has become a promising therapeutic option for SCI patients. Many experimental studies have suggested that the transplantation of bone marrow cells (BMCs), neural progenitor cells, or olfactory ensheathing cells could promote functional improvements after SCI [1, [2], [3]–4]. In recent years, some have found that BMCs differentiate into mature neurons or glial cells under specific experimental conditions [5, [6]–7]. These findings imply the therapeutic potential of BMCs in patients with neurological diseases and can obviate ethical problems.

The grafting of BMCs into the SCI models has been actively studied. Transplanted BMCs were found to improve neurological deficits in the CNS injury models by generating neural cells or myelin producing cells [8, 9]. Furthermore, BMCs can produce neuroprotective cytokines, which rescue the neurons with impending cell death after injury [10, 11]. Several clinical trials have explored the hypothesis that cell transplantation may enhance the recovery of neurologic functions after SCI. However, it was reported that significant functional recovery after cell transplantation was rarely achieved in the human clinical trials [12, 13]. These reports raise a concern that treatment protocols using only cell transplantation are not sufficient to achieve the required therapeutic goals in SCI.

Recently, it has been reported that hematopoietic cytokines including granulocyte macrophage-colony stimulating factor (GM-CSF), granulocyte-colony stimulating factor (G-CSF), or erythropoietin had neuroprotective effects and improved neurologic functions after CNS injury [14, [15], [16], [17]–18]. These findings suggest that GM-CSF, which is popularly used and considered safe for hematologic disease, could be used for SCI treatment.

In our previous study, GM-CSF decreased neuronal apoptosis and improved the functional outcome in SCI animal models [19]. Preliminary results of our study demonstrated that BMC transplantation and GM-CSF treatment didn't increase serious complications [14]. Furthermore, it has been published that GM-CSF stimulates microglial cells to increase brain-derived neurotrophic factor (BDNF) synthesis [18].

As a result, it was hypothesized that administration of GM-CSF to SCI patients could be an adjunct measure to improve the therapeutic effects of BMC transplantation (Fig. 1). In this study, autologous BMCs were transplanted, and the bone marrow was stimulated with GM-CSF in SCI patients to study the therapeutic and adverse effects of the treatment.

Figure Figure 1..

Therapeutic strategy of BMC transplantation and GM-CSF. GM-CSF stimulates and mobilizes the bone marrow stem cells (direct pathway). In addition, GM-CSF can have intrinsic spinal cord repair mechanisms (indirect pathway), including neuroprotection from apoptosis, endogenous stem cell activation, inhibition of glial scar formation, and microglial cell activation. Abbreviations: BMCs, bone marrow cells; GM-CSF, granulocyte macrophage-colony stimulating factor.

Materials and Methods

Patient Selection

This phase I/II nonrandomized study was approved by the Institutional Review Board at Inha University Hospital, and all procedures were performed after obtaining written informed consent. Detailed protocols for patient selection and for BMC transplantation and the GM-CSF administration have been reported previously [14]. In brief, patients were eligible when they were admitted to Inha University Hospital with complete SCI (American Spinal Injury Association [ASIA] Impairment Scale [AIS] grade A). Patients first underwent successful spinal cord decompression and stabilization with anterior cervical fusion using an autologous iliac bone graft and demonstrated persistent complete paralysis below the level of injury. Exclusion criteria were the presence of anatomical transection of the cord, visualized by spinal magnetic resonance imaging (MRI), fever (above 39°C), ventilator-assisted breathing, and serious pre-existing medical diseases. According to the time window between the injury and transplantation, eligible patients were divided into three groups: acute (<2 weeks), subacute (2–8 weeks), and chronic (>8 weeks). In the control group, all 14 patients were managed with decompression surgery of the spinal canal and then referred to the rehabilitation clinic for active rehabilitation. Patient groups are summarized in Figure 2A and Table 1. Patients in the control group were not treated with BMC transplantation or GM-CSF administration. All 14 patients in the control group were selected randomly from complete SCI patients (AIS A) who were admitted to Inha University Hospital from January 2001 to May 2003. The patients in the treatment groups were selected from the patients admitted to the hospital after June 2003. Demographic data is summarized in Table 1. This study was designed as a nonrandom and open-label safety trial with observer-blind neurologic evaluations of patients.

Figure Figure 2..

Patient groups and treatment schedule. (A): Patient compliance, selection criteria, and follow-up from the beginning of the study to the 10-month follow-up check. (B): Bone marrow cell transplantation and GM-CSF administration protocol. In the acute stage, cell transplantation was done within 14 injury days. Daily administration of GM-CSF was performed for the first 5 days of each month over a 5-month period. Diagnostic studies, including serial neurologic exams, EMG, spine MRI, functional MRI, and EP, were performed both at the beginning of and during the treatment. Abbreviations: ASIA A, American Spinal Injury Association Impairment Scale grade A; ASIA B, American Spinal Injury Association Impairment Scale grade B; d, days; EMG, electromyography; EP, evoked potentials; fMRI, functional magnetic resonance imaging; GM-CSF, granulocyte macrophage-colony stimulating factor; m, months; MRI, magnetic resonance imaging.

Table Table 1.. Summary of patient demographics, spinal cord injury level, and duration of the MRI follow-up
original image

Separation of Human Bone Marrow Cells

Bone marrow blood (100–150 ml) was aspirated from the iliac bone and diluted in Hanks' balanced salt solution (HBSS) at a ratio of 1:1. After the samples were centrifuged (1,000g for 30 minutes) through a density gradient (Ficoll-Paque Plus, 1.077 g/l; Amersham Biosciences, Piscataway, NJ,, the mononuclear cell layer was recovered from the gradient interface and washed with HBSS. The cells were centrifuged at 900g for 15 minutes and resuspended in 1.8 ml of phosphate-buffered saline at a density of 1.1 × 108 cells per milliliter. The concentrated BMCs were then transferred to the operation room. All the procedures were carried out in a sterile environment.


In acute SCI patients, transplantation was done within 14 days after admission. Neurologic examination was performed immediately before operation to confirm complete SCI. Complete laminectomy was performed from one vertebra above to one below in order to provide sufficient access to the transplantation site. The dura mater was then incised, sparing the arachnoid, which was subsequently opened separately with microscissors. The dorsal surface of the contusion site was located using high-power microscopic magnification. After exposure of sufficient surface at the contusion site, 300-μl aliquots of cell paste (total volume of 1.8 ml) were injected into six separate positions surrounding the lesion site with the injection depth of 5 mm from the dorsal surface and 5 mm lateral from the midline. To avoid mechanical injury during injection, 2 × 108 cells were injected at a rate of 300 μl/minute using a 21-gauge needle attached to a 1-ml syringe. To prevent cell leakage through the injection track, the injection needle was left in position for 5 minutes more after completing the injection. The dura mater and arachnoid were then closed. The muscle and skin were closed layer by layer.

GM-CSF Injection Schedule

After surgery, a total of five cycles (daily for the first 5 days of each month over 5 months) of GM-CSF (Leucogen; LG Life Science, Seoul, Korea, were injected subcutaneously (250 g/m2 of body surface area).

Follow-Up Studies

The neurological status of the patients was determined in terms of AIS. Using AIS, a five-scale subdivision was used: A, a complete loss of motor and sensory function; B, sensory but not motor function is preserved below the neurological level and includes the sacral segments S4–S5; C, motor function is preserved below the neurological level, and more than half of the key muscles below the neurological level have a muscle grade of less than III; D, motor function is preserved below the neurological level, and at least half of the key muscles below the neurological level have a muscle grade of III or more; E, motor and sensory function are normal. The mean duration for follow-up was 10 months postoperatively. During the GM-CSF administration, vital signs were checked and complete blood analysis was taken daily.

We performed electrophysiological studies such as motor evoked potentials, somatosensory evoked potentials, and electromyography to differentiate voluntary muscle contraction from reflex or involuntary spontaneous limb movement. The spinal MRI was also carried out to examine any changes in the spinal cord and surrounding tissues. Changes in activity patterns in the cortical sensorimotor networks were measured using functional MRI during the proprioceptive stimulation with repetitive passive toe movement. All preoperative and postoperative follow-up studies and the treatment schedules for the acute group are summarized in Figure 2B.

Statistical Analysis

Statistical analyses included analysis of variance or Mann-Whitney U for the nonparametric variables (age, follow-up duration, and peripheral blood leukocyte number); χ2 test or Fisher's exact test were performed to analyze the nominal or ordinal variables (sex, injury level, adverse events, neuropathic pain, AIS improvement, and MRI findings). All tests were considered significant at p values less than .05. Statistical comparisons were analyzed with SPSS 12.0 (SPSS, Chicago,

Role of the Funding Sources

The sponsors of the study had no role in the study design, datum collection, datum analysis, datum interpretation, or the writing of this report. The corresponding author had full access to all data in the study and had final responsibility for the decision to submit for publication.


Among the 53 patients treated, 48 (90.5%) were included in the 10-month follow-up (Fig. 1). Five patients (9.5%) were excluded because two patients did not want BMC transplantation, and three patients were ASIA B in preoperative neurologic status. The demographic characteristics of the study population were similar across the three treatment groups and the control group. The mean follow-up period varied from 10.1–11.3 months.


No patients experienced serious complications requiring open surgery such as sepsis or wound infection. Table 2 outlines the adverse events that occurred during the treatment. A higher degree of febrile reaction was noted in the GM-CSF treatment group. However, it didn't increase the meaningful mortality or morbidity. Serious wound infection or gastrointestinal bleeding was not observed in this study. Theoretically, mechanical injury to the spinal cord neurons may occur during cell transplantation directly into the spinal cord, resulting in the risk of worsening the neurologic symptoms postoperatively. In this study, one patient in the chronic stage group showed a transient deterioration of neurologic symptoms, including a mild loss in the grasping power of both hands (Table 2). This negative effect was, however, transient and minimal, with the patient recovering his previous neurologic status within 1 month. Three patients showed increased muscle rigidity in the posterior neck muscles after surgery. One of these patients showed severe spastic contraction of the posterior neck muscles with muscle pain requiring antispasmodics and pain management. Spasticity resolved at 2 months postoperatively.

Table Table 2.. Summary of adverse events
original image

Neuropathic Pain

It has been reported that neuropathic pain is one of the serious complications after SCI. Furthermore, some studies have reported that cell transplantation strategy increased the risk of neuropathic pain postoperatively [20, 21]. To identify the potential risk of neuropathic pain under the current protocol, the incidence of neuropathic pain in each group was investigated (Table 3). It was found that some of the patients in the treatment group and control group developed neuropathic pain requiring treatment. Overall, 20% of the patients in the treatment group developed neuropathic pain, whereas only 1 of the 13 (7.7%) patients in the control group complained of it during the follow-up period. Interestingly, the neuropathic pain occurred mainly in the subacute or chronic treatment group (33.3% each). In the acute treatment group, however, only 1 of 17 patients (5.9%) experienced it. These results imply that BMC transplantation and GM-CSF administration may be associated with a greater risk of developing neuropathic pain, particularly in the subacute and chronic patients.

Table Table 3.. Incidence of neuropathic pain
original image

Aberrant regeneration of axons in damaged spinal cord is one suggested mechanism of increasing neuropathic pain after cell transplantation [20, 21]. Currently, we don't have any method to assess aberrant regeneration in the patients directly. However, we could visualize an example of aberrant regeneration using functional MRI (fMRI) study (Fig. 3C). The predominant activation of the temporal lobe and occipital lobe rather than the contralateral sensorimotor cortex was identified in postoperative fMRI at 4 months.

Figure Figure 3..

Neurologic outcome and magnetic resonance imaging (MRI) findings. (A): Change in the neurologic status and spinal MRI findings produced by the effect of bone marrow cell (BMC) transplantation and granulocyte macrophage-colony stimulating factor administration. Neurologic improvements in the acute and subacute treatment groups were significantly higher than the control group (7.7%) and the natural recovery of spinal cord injury reported previously (12.5%) [39]. (B): Follow-up MRI findings. Sagittal T2W (Ba), T1W (Bb), and postcontrast MRI (Bc) performed at 4 months postoperatively showing a small patch enhancement in the BMC transplantation area (arrows) and slight cord expansion. (C): Functional MRI (fMRI) findings. Preoperative initial (Ca) and postoperative follow-up (Cb) fMRI at 4 months. The fMRI was performed following the right great toe proprioceptive stimulation. Initial fMRI did not reveal any cortical activities, suggesting complete disruption of sensory pathways. The 4-month postoperative fMRI demonstrated increased signal activities in bilateral temporal and ipsilateral occipital lobes indicating aberrant regeneration (white circle in [Cb]). Abbreviation: wks, weeks.

Neurologic Improvements

Figure 3A provides histograms of the estimated mean change in the percentage of patients improving on the AIS scale. Overall, 29.5% of patients in the acute treatment group showed neurologic improvement from AIS A to B or C. In the subacute treatment group, 33.3% of patients improved to AIS B or C during the follow-up period. However, the patients in the chronic treatment group didn't show any changes in the neurologic status.

GM-CSF Effect on the Leukocytosis and Neurologic Outcome

To investigate the systemic effect of GM-CSF on the outcome of SCI, the number of white blood cells in the peripheral blood was compared with the neurologic outcomes (Fig. 4). The administration of GM-CSF has been used extensively to trigger peripheral leukocytosis and to induce bone marrow hematopoietic stem cell mobilization. The total number of recruited white blood cells in the peripheral blood was elevated after GM-CSF administration. The number of white blood cells in patients showing improved neurologic function was significantly higher than that in the patients without neurologic improvement.

Figure Figure 4..

Change in the peripheral leukocyte number produced by GM-CSF administration. The mean number of peripheral leukocytes significantly increased in patients with improving American Spinal Injury Association Impairment Scale grades while showing no significant increase in patients without improvement (* p < .05). Abbreviations: Conc., concentration; GM-CSF, granulocyte macrophage-colony stimulating factor; WBC, white blood cells.

Spinal MRI Findings

Changes in the MRI findings for the patients with treatment are given in Table 4 and Figure 3B. Overall, 42.9% of patients in the treatment group showed an increase in the diameter of the spinal cord at the cell transplantation site. However, 33.3% of the patients showed atrophic changes (a decrease in diameter) at the transplantation site. Six patients (28.6%) showed evidence of spinal cord enhancement. Other findings, including spinal cord edema demonstrated by increased T2 weight images, cystic degeneration, or spinal cord atrophy distal to the lesion, were observed. But no other significant changes such as tumor formation of the transplanted BMCs were found.

Table Table 4.. Changes in spinal magnetic resonance imaging findings
original image


The present study has demonstrated that patients at acute stage of complete SCI did not show any serious adverse response after the combined therapy of BMC transplantation and GM-CSF administration. The use of autologous BMCs for stem cell therapy in SCI patients has more advantages. First, one can avoid all problems associated with the immunological rejection or graft-versus-host reactions [22], which are frequently caused in allogenic cell transplantation. Second, autologous BMC therapy is considered safe by not being associated with carcinogenesis [23]. Third, extensive scientific data on BMCs have been accumulated from previous experiences in BMC transplantation for hematological diseases. These advantages have made cell therapy using BMCs widely applicable and investigated clinically in various neurologic diseases. However, it also has some disadvantages. First, the procedure requires open surgery to approach the injury area, which results in the increase in the potential adverse effects. Second, the sorting of BMCs needs to be done in vitro, thus increasing the risk of contamination. Third, it is still unclear whether some BMC components may have a deleterious effect on the functional improvement. In this study, approximately 100–150 ml of bone marrow was aspirated from the iliac bone, and mononuclear cells were sorted and concentrated to a final volume of approximately 1.8 ml. These BMCs were then transplanted by injection into the surrounding area of the injury site. This procedure required approximately 3–4 hours to complete. Therefore, some technical modifications are needed to reduce the operation time and enhance the safety and therapeutic effects.

BMCs are a mixed cell population including hematopoietic stem cells, mesenchymal stem cells, endothelial progenitor cells, macrophage, and lymphocytes. It has been shown in other studies that hematopoietic or mesenchymal stem cells had a neuroprotective effect and increased neurite outgrowth [11, 24]. Furthermore, activated macrophages are currently being investigated as a therapeutic target for SCI patients [25, 26]. In contrast, inflammatory cells in BMCs could have adverse effects, such as increasing the inflammatory response, thus inhibiting the regeneration process [27, 28]. As a result, further studies comparing the effects of whole BMCs with those of a subpopulation with inflammatory cells removed are required.

Recombinant human GM-CSF has long been used safely in patients with bone marrow suppression. It is well known that the classic role of GM-CSF is hematopoiesis by inducing the growth of several different hematopoietic cell lineages. It also enhances the functional activities of mature effector cells involved in antigen presentation and cell-mediated immunity, including neutrophils, monocytes, macrophages, and dendritic cells. Recently, some studies have reported that GM-CSF prevented apoptotic cell death not only in hematologic cells but also in neuronal cells [18, 29, 30]. From these findings, it was speculated that administration of GM-CSF to SCI patients would result in the improvement of neurologic functions without any significant complication. In this study, a hypothesis was developed that GM-CSF would not only activate patient bone marrow for stem cell mobilization but would also have a direct effect on the transplanted BMCs by enhancing their survival in the spinal cord and activating them to excrete neurotrophic cytokines (Fig. 1). Recently, some investigators found that GM-CSF stimulated microglial cells to produce neurotrophic cytokines such as BDNF [18]. Furthermore, other hematopoietic cytokines including G-CSF and erythropoietins were also shown to have neuroprotective effects and promote neurologic outcomes in CNS injury models [15, [16]–17]. Since the safety of most hematopoietic cytokines is already verified in the clinical fields, these experimental studies could be rapidly turned into clinical investigation.

Complications such as cerebral and myocardial infarction after hematopoietic cytokine administration have been reported, and the primary mechanism involved was presumed to be an acute arterial occlusion by thrombus [31]. The thrombogenic properties of the peripheral blood result from increased viscosity due to the mobilization of hematopoietic cells. In this study, no serious complications by GM-CSF were observed that caused an increase in the morbidity or mortality. Most patients suffered from a fever, rash, or leukocytosis. These findings could be considered not only as adverse effects but also as indicators for the systemic effect of GM-CSF. Methylprednisolone therapy has been considered as a standard therapy for acute SCI. However, recent publications noticed that the methylprednisolone therapy significantly increased the morbidities and mortalities [32, 33]. It increased serious complications such as wound infection, gastrointestinal bleeding, and urinary tract infection. In order to avoid the detrimental effects of methylprednisolone, the safety of cell therapy protocols comparing that of methyl prednisolone therapy needs to be investigated.

The AIS grading is determined by the clinical neurologic findings and not by pathologic severity. As a result, AIS grading does not completely reflect pathologic severity. It is possible that one AIS grade could include a wide spectrum of severity. Less severely injured patients with an AIS grade of A could, therefore, have a greater potential for improvement. In this study, approximately 30% of the patients in AIS A improved to grade B or C (Fig. 3A). It is believed that this partial effect was caused by an additive or alleviating role inherent in the protocol rather than the innate ability of natural recovery in the less severely injured patients. Interestingly, a relationship between leukocytosis after GM-CSF administration and neurological improvement was identified. The mean number of peripheral leukocytes significantly increased in patients with improving AIS grades while showing no significant increase in patients without improvement. These findings imply that patients who are more responsive to GM-CSF might have a greater capacity for improvement. As a result, interpersonal inherited variability in response to GM-CSF is one possible explanation for the observation that the protocol appeared to have a therapeutic effect on only a small population of the patient group. Furthermore, these findings also support the idea that the mobilizing role of GM-CSF could be one important mechanism of functional improvement. However, it is not clear at this point that the higher level of leukocytosis is a prerequisite for the GM-CSF effect and/or the recovery from SCI.

Many experimental studies have been completed to find the optimal period for cell transplantation. To avoid the destruction of transplanted cells by the inflammatory process in the acute phase (less than 1 week), many researchers consider the subacute phase between 10 and 14 days as an optimal period for cell transplantation. Results from animal studies show that significant gliosis is a major obstacle inhibiting axonal regeneration during the chronic stage [34, 35]. Interestingly, few patients in this study showed neurologic improvement even if BMC transplantation was performed between 2 and 8 weeks after injury. These results indicate that the optimal period of BMC transplantation in SCI patients should not be restricted to patients less than 2 weeks postinjury.

However, our patients in the subacute group were relatively younger (mean 29.4 years) than the other groups (mean >40 years). Therefore, it should be considered that their relatively younger age could have influenced their ability to recover better than the other groups.

Few studies have been published that identify cell transplantation therapy as an ideal option for improving the neurologic functions in patients at the chronic stage [36, [37]–38]. This study also did not show any therapeutic benefits for chronic patients. Even though a few patients experienced the improvement in the sensory or motor functions (3–5 dermatome) immediately after surgery, these improvements disappeared in 3–4 months. Because of the sudden and immediate nature of these improvements, it cannot be suggested that these short-term improvements were due to the effect of the engrafted stem cells, as this process needs an incubation period to expand cell numbers. Furthermore, cytokines released from the engrafted stem cells also require some time for signal transduction to show final effects. It is known that restoration of cerebrospinal fluid flow or the microtrauma caused by the opening of the dura and injection of BMCs could exert transient changes in the environment around the SCI lesion site such as the disruption of the glial scar barrier or the transient excitation of quiescent neurons. In the chronic stage treatment group, some patients demonstrated new or changing clinical features such as transient increasing myelopathy, ascending neurological level, or pain. In this study, increase in the spinal cord dimension at the site of transplantation was found. However, no tumor infiltrating the surrounding spinal cord tissue was identified. These findings imply that autologous BMC transplantation doesn't increase the risk of malignant transformation, which is still a field under investigation in adult stem cell transplantation.

Approximately 20% of patients who were treated with BMCs and GM-CSF complained about pain during the follow-up period. Furthermore, one patient showed the extension of the spinal cord edema associated with an increase in neuropathic pain (data not shown). These findings reinforce the suggestion that cell transplantation, especially in chronic stage, increases adverse effects like neuropathic pain. Therefore, we should be informed that the stem cell therapy for chronic patients doesn't guarantee promising results and should be cautiously performed in selected cases. Recently, it has been reported that stem cell transplantation increases neuropathic pain in animal study. Neuropathic pain might be related to the aberrant regeneration of damaged axons. Our preliminary results of fMRI studies also support this idea. Further studies using fMRI to understand the correlation between aberrant regeneration and neuropathic pain would provide us with the objective pain monitoring method as well as the information to understand the nature of neuropathic pain in SCI patients.

There are several limitations in this study. First, complete double-blind case control studies were not possible. Although the examiners were well trained and qualified by the controlled education system, potential bias produced between examiners could not be removed completely. Second, neurologic improvements were described as advancing AIS grades. Therefore, small improvements without changes in AIS grade could not be identified. In order to overcome these limitations, it is hoped this study can be repeated using a random sample multicenter double-blind protocol with specially designed assessment methods. Finally, due to the limited number of complete SCI patients, we could not understand how each of the BMC transplantation and GM-CSF treatments affected the recovery from SCI. Further large group studies are required to clarify the effects of either therapy.

This study suggests that, for the functional recovery of damaged spinal cord, human BMCs can be transplanted into the spinal cord of patients with complete SCI. The BMC transplantation and GM-CSF administration caused no new deficit in patients and appeared to be safe in short- and longer-term assessments. In addition to protocol safety, preliminary evidence showing the effect of cell therapy in patients with complete SCI was also provided. This study was performed in a limited number of patients. It is recommended that the therapeutic effects should be established in a more comprehensive multicenter study.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.


This study was supported by the Grant of the Korea Health 21 R&D Project, Ministry of Health & Welfare, Republic of Korea (A050082).