Abstract
Granulocyte-colony-stimulating factor (G-CSF) is a major growth factor in the activation and differentiation of granulocytes. This cytokine has been widely and safely employed in different disease conditions over many years. The administration of the growth factors in spinal cord injury (SCI) has been reported elsewhere; here we have tried to see the effect of SCI severity on the neurological outcomes after neuroprotective treatment for SCI with G-CSF. Seventy-four consecutive patients with SCI of at least 6 months' duration, with stable neurological status in the last 3 months, having informed consent for the treatment were included in the study. All the patients had undergone at least 3 months of standard rehabilitation. Patients were assessed by the American Spinal Injury Association (ASIA) scale, Spinal Cord Independence Measure (SCIM) III, and International Association of Neurorestoratology-Spinal Cord Injury Functional Rating Scale (IANR-SCIFRS) just before intervention and periodically until 6 months after subcutaneous administration of 5 μg/kg per day of G-CSF for 7 consecutive days. Multiple linear regression models were performed for statistical evaluation of lesion completeness and level of injury on changes in ASIA motor, light touch, pinprick, IANR-SCIFRS, and SCIM III scores, as a phase I/II comparative study. The study consisted of 52 motor-complete and 22 motor-incomplete SCI patients. There was no significant difference regarding age and sex, chronicity, and level of SCI between the two groups. Motor-incomplete patients had significantly more improvement in ASIA motor score compared to the motor-complete patients (7.68 scores, p < 0.001); also they had significant improvement in light touch (6.42 scores, p = 0.003) and pinprick sensory scores (4.89 scores, p = 0.011). Therefore, G-CSF administration in motor-incomplete SCIs is associated with significantly higher motor improvement, and also the higher the initial ASIA Impairment Scale (AIS) grade, the less would be the final AIS change, and incomplete cases are more welcome into the future studies. This manuscript is published as part of the International Association of Neurorestoratology (IANR) special issue of Cell Transplantation.
Keywords
Introduction
Traditionally, the clinical course of spinal cord injury (SCI) has been subclassified into hyperacute, acute, subacute, and chronic phases, and various forms of treatment have been speculated to be effective for each period (13). Therapeutic approaches in hyperacute and acute phases, including early resuscitation, steroid application, and decompression/stabilization, have been reported to be associated with better outcomes in incomplete SCIs. Also outcomes of autorecovery and rehabilitation protocols in the subacute phase have been significant compared to chronic cases (2,3). It would be plausible to extrapolate these results for outcomes of cellular and neuroprotective treatments.
Many authors have considered the risk of neurological decline after invasive cell therapy modalities (such as intramedullary injection) in incomplete SCIs and advocate these treatments for patients with complete SCI (14). This conservative approach may deprive those with incomplete SCI (with more potential benefit) from receiving these treatments. This, in part, leads to small effect size reports and hinders the endeavors in this way. Depending on the regional epidemiological conditions, incomplete SCIs form a significant proportion of SCI patients, who may benefit more from these treatments (7).
Also, as recommended, limiting the invasive neuroprotective therapies to complete SCI patients for consideration of the safety sounds logical, while, with upcoming noninvasive neuroprotective treatments, it may be possible to expand the domain of clinical application of these treatments to incomplete cases. In this study, a previously reported safe treatment with granulocyte-colonystimulating factor (G-CSF) with a noninvasive delivery route (subcutaneous) has been administered (10). Neurological changes [sensory, motor American Spinal Injury Association (ASIA) scores] in the two groups that are motor complete [ASIA Impairment Scale (AIS) A or B] and motor incomplete (AIS C or D) are reported, and the outcome subscales are compared.
Materials and Methods
This study was conducted in the Brain and Spinal Injuries Research Center at Tehran University of Medical Sciences and was approved by the local ethical review board. The patients were selected in the outpatient clinic for SCIs. They were aged between 18 and 50 years, all having traumatic SCIs, and had stable neurological status in the last 3 months. They had decompressed spinal cord, stabilized spine, and had been rehabilitated for at least 3 months. They had no history of blood dyscrasia, and an internist consult revealed no hematological abnormality. The patients did not pay for their inclusion or treatment in this study.
Patient Characteristics
Among 98 consecutive cases with traumatic SCI presenting to the SCI clinic, 90 patients volunteered for the study; five patients were excluded due to age >50 years, two because of age <18 years, two because of a complicated pressure sore, one for heterotopic ossification, and three were excluded due to severe neuropathic pain (VAS > 70). Those with major general and/or hematological and allergic problems were excluded. This finding was not present in our patients. Seventy-seven patients with SCI were selected for the study. After that, a written informed consent was taken from the patients. Among them, 74 patients signed the informed consent sheets.
Assessments
Assessments were performed by independent observers composed of neurosurgeons, urologists, and occupational therapists. The employed tools included ASIA scale (9), Spinal Cord Independence Measure (SCIM) III (6), and the International Association of Neurorestoratology-Spinal Cord Injury Functional Rating Scale (IANR-SCIFRS) forms (5). The assessment tools for neurological and functional outcomes were recorded at baseline and after 6 months of intervention. The safety assessments were performed at baseline and monthly intervals until 6 months of follow-up, and any evidence of complications were observed and recorded regularly and reported by the evaluation team.
Treatment
Treatment consisted of 5 μg/kg per day of subcutaneous G-CSF (Filgrastim, Neupogen®; Amgen, Thousand Oaks, CA, USA) for 7 consecutive days. Basal complete blood counts were performed with hematology analyzer Sysmex XT-1800i (Norderstedt, Germany) and were repeated daily for the next 7 days. Also it was performed at 3 months and finally at 6 months.
Statistical Analysis
We performed multiple linear regression models for evaluating the effect of lesion completeness and level of injury on changes in ASIA motor, light touch, pinprick, IANR-SCIFRS, and SCIM III scores in separate regression models. Level of injury was considered as a variable with two levels (cervical and noncervical) and also completeness of lesion was defined as a two-level variable (motor complete and motor incomplete). All variables with values of p < 0.20 in univariate analysis were entered into the multiple regression models; those for the final model were selected by using likelihood ratio tests (LRTs). Pearson's linear correlations were applied for evaluating the linear relationship between changes of ASIA scales, functional assessment measures, and length of lesion on T1-weighted MRI. A chi square test of trend was applied for examining the relationship of initial AIS grade and its changes. Changes in every item of SCIM III and IANR-SCIFRS in motor-complete and motor-incomplete groups were also compared with independent sample t-tests. A value of p < 0.05 was assumed to be statistically significant.
Results
In our study sample, there were 52 motor-complete and 22 motor-incomplete patients. There were no significant differences regarding age, sex, cause of SCI, level of injury, and chronicity in the two groups (Table 1). However, length of the lesion on MRI, as expected, was predictably longer in the motor-complete group (24.9 mm vs. 15.4 mm, p < 0.001).
Clinical Characteristics in Motor-Complete and Motor-Incomplete Patients
Data are presented as mean (SE) or frequency (percentage). SCI, spinal cord injury; MVA, motor vehicle accident; AIS, American Spinal Injury Association (ASIA) impairment scale; WBC, white blood cell; SE, standard error.
A multiple linear regression model adjusted for the level of injury was performed for neurological and functional outcomes. Motor-incomplete patients had significant improvement in ASIA motor score compared to the motor-complete patients (7.68 scores, p < 0.001); also they had significant improvement in light touch (6.42 scores, p = 0.003) and pinprick sensory scores (4.89, p = 0.011). Patients with cervical level of injury had significant improvement in ASIA motor (2.86 scores, p = 0.002), light touch (6.87 scores, p < 0.001), and pinprick scores (4.76 scores, p = 0.007) compared to noncervical lesions, adjusted for the completeness of lesion. Length of lesion had significant negative linear correlation with changes in ASIA motor (r = −0.28, p = 0.016), light touch (r = −0.32, p = 0.006), and pinprick (r = −0.22, p = 0.043) scores, while a nonsignificant negative correlation with FRS (r = −0.17, p = 0.157) and SCIM III (r = −0.13, p = 0.262) changes were observed. ASIA motor score was positively correlated with pinprick (r = 0.49, p < 0.001) and light touch sensory scores (r = 0.33, p = 0.003).
The same analysis was performed for functional outcomes. Despite greater improvement in ASIA scores for motor-incomplete patients, nevertheless they had significantly less improvement in total SCIM III scores compared to motor-complete patients (2.56 scores, p = 0.037). Level of injury did not have any significant effect on functional assessment measures (p > 0.40) (Table 2). Changes in functional evaluation measures were not significantly correlated with ASIA motor and sensory scores. Regarding IANR-SCIFRS item changes, motor-incomplete patients had significant improvement in the walking item (p = 0.022) and standing without brace with marginal significance (p = 0.087) compared to motor-complete patients. Regarding SCIM III item changes, patients in the motor-complete group had significant improvement in bathing upper body (p = 0.018), grooming (p = 0.044), bladder management (p < 0.001), and bed–wheelchair transfer (p = 0.013) compared to the motor-incomplete patients. Also marginal significance was observed in bathing lower body improvement (p = 0.057).
Results of Multiple Linear Regression Models
Ranges for score categories are as follows: motor: minimum 0, maximum 100; light touch: minimum 0, maximum 112; pinprick: minimum 0, maximum 112; IANR-SCIFRS: minimum 0, maximum 48; SCIM III: minimum 0, maximum 100. IANR-SCIFRS, International Association of Neurorestoratology-Spinal Cord Injury Functional Rating Scale; SCIM III, Spinal Cord Independence Measure III; SE, standard error.
Regarding AIS grade changes, 29.0% of AIS A patients changed to B and 3.2% to C, and 67.7% remained in AIS A grade, while 4.8% of AIS B patients changed to C, and 95.2% remained unchanged. In AIS C patients, 7.1% changed to D, and 92.9% were unchanged. None of the AIS grade D patients had AIS changes. A chi square test of trend for AIS change revealed that the less severe the AIS grade, the smaller would be the chance of its change (p < 0.001) (Fig. 1).
Percentage of AIS conversions in final assessment expressed on the basis of pretreatment AIS.
Discussion
Neuroprotective processes taking part after SCI need a well-structured extracellular matrix and neural pathway architecture to pave the way for axonal connections (16). Actually, these infrastructures have inductive effects for determination of axonal sprouting direction. One of the important signs of preservation of tissue microstructure is partial preservation of neurological function, clinically identified as an incomplete lesion. In complete SCI, almost all tissue architecture and axonal connection is lost, and there is no inducing milieu for stem cell differentiation; this may be associated with much less clinical effect.
Historically, in many studies, severity of SCI has been proposed to be one of the most important prognostic factors for neurological recovery after acute phase treatment for SCI. Patients with Frankle A grade have 6–13% chance for neurological recovery, while in Frankle B, the chance for neurological recovery would be in two thirds of the cases, after initial resuscitation, steroids, and decompression (4). It is not very farfetched to extrapolate this idea for estimation of cell therapy outcomes.
According to our results, motor improvement was more significant in patients with motor-incomplete SCIs; other studies also support this finding (8). Regarding neurological level, thoracic patients had less motor improvement compared to cervical patients (irrespective of the severity of the lesion); this finding may be attributed to the fact that thoracic motor segments are not represented in ASIA motor scores.
Pinprick changes were more significantly found in patients with incomplete-motor SCIs; other studies also support this finding after cell therapy (12). In contrast to light touch improvement, pinprick improvement is smaller than light touch score changes. The most noticeable improvements are observed in light touch scores. This finding may be due to multiple neuroanatomical afferent tracts correlated with light touch perception. Pinprick score improvements are more correlated with motor improvement than light touch score improvement; this finding is supported by other studies and may be explained based on the neighborhood of spinothalamic and corticospinal tracts within the spinal cord (11).
Regarding functional changes, the most important predictor has been severity of SCI and neurological level. This means that the Frankle grade changes determine the changes in functional scores (1). Given the fact that Frankle grade change is less common in incomplete-motor SCIs, compared to motor-complete ones, significant SCIM III score change was observed only in motor-complete SCI in our series; however, IANR-SCIFRS scoring subscale change was significant in the motor-incomplete SCIs after treatment.
Gender effect has been notified in some studies, as females show more neurological recovery, while functional changes have been reported to be more significant in complete SCIs of male patients (15). In this study, effect of gender on SCIM III and ASIA outcome was assessed, and there was no association between gender and neurological and functional outcomes.
Age over 65 has been correlated with less neurological improvement. None of our patients were in this age range. We found no significant association between age and neurological and functional outcomes. Test of trend for AIS grade revealed the higher the AIS grade, the less would be the chance for AIS grade spurt. This may explain why SCIM III change was more pronounced in the complete subgroup, considering the fact that AIS change is the major factor for SCIM III change.
Regarding effect of injury level on neurological outcomes in our series, all three modalities were more significantly improved in cervical cases, adjusted for completeness of the lesion, compared to thoracic lesions.
Effect of lesion size (a marker of injury severity) was assessed in our patients. Motor-complete SCIs had longer signal change area on T1-weighted MRI scans. Therefore, lesion length was inversely correlated with neurological outcomes, and other studies also support these findings (15).
It should be noted that the results of the study may be generalized for noninvasive neuroprotective treatments (G-CSF), while in invasive neurorestorative treatments (either cell preparation or cell delivery), it may be recommended to include only AIS A patients, especially in the early phases of the study.
Conclusion
G-CSF administration in motor-incomplete SCIs is associated with significantly greater motor improvement, and also the higher the initial AIS grade, the less would be the final AIS change, and incomplete cases may be more welcome into the future studies.
Footnotes
Acknowledgments
The authors are grateful to Dr. Abbas Norouzi Javidan, Dr. Ahmad Aoude, Miss Asal Derakhshanrad, Rasoul Mansouri Ivrigh, and Mrs. Shole Oryani. We would also like to thank the IANR organizing committee and Dr. Morteza Kheirabadi for their kind support. The authors declare no conflict of interests.