We studied the relationship between clinico-electrophysiological variables and metabolite ratios on cervical cord 1H-MRS. NAA/Cr and NAA/m-Ins were significantly decreased, and m-Ins/Cr was significantly increased in ALS patients. NAA/Cr and NAA/m-Ins were correlated with disease severity and progression. Moreover, marked alternations of these metabolite ratios were found in 9 patients with upper limb onset who had ongoing denervation and chronic neurogenic changes in both C2 paraspinal and upper limb muscles.
Recent studies of cervical cord 1H-MRS have shown that several metabolites are useful biomarkers in patients with multiple sclerosis (MS).23–25 Cervical cord 1H-MRS disclosed decreases of NAA concentrations23,24 and NAA/Cr,25 and increases of Cho/Cr and Cho/m-Ins in MS patients.25 The metabolite levels were compared with T2-hyperintensity lesions on brain and spinal cord MRI in MS patients. Increased Cho levels might result from inflammation and demyelination. In the later stage, neuronal dysfunction or loss was thought to cause decreased NAA levels. Elevation of m-Ins could suggest gliosis in MS lesions.25,26 Aside from these findings in MS patients, 2 studies of spinal cord 1H-MRS have been published in ALS patients and their relatives.18,19 Carew et al. reported cervical cord 1H-MRS in ALS patients18 and asymptomatic carriers with SOD1 mutation.19 In the first study, the clinical profile of 14 patients (11 men and three women) showed 8 patients with definite, 1 with probable, and 5 with possible ALS according to the revised El Escorial criteria.18 The mean age (SD) was 54 (12) years, and the mean disease duration was 26.7 months. All patients had limb onset. The mean (SD) of ALSFRS was 39 (10) points, and the FVC was 84 (23)%. Four metabolites were measured in the high cervical cord at the level of the C1–C2 vertebral bodies using a 3.0T scanner. Compared with 16 controls, NAA/Cr, Cho/Cr and NAA/m-Ins were decreased at 40%, 20%, and 38% in ALS patients, respectively. There was no statistical difference of m-Ins/Cr between ALS patients and controls.18 In our study, significant reductions of NAA/Cr and NAA/m-Ins were found at 25% and 35% in ALS patients, respectively. Cho/Cr did not differ statistically between ALS patients and controls. m-Ins/Cr was increased by 37% in ALS patients compared with controls. Therefore, NAA/Cr and NAA/m-Ins in our patients were comparable to those ratios in a previous study of Carew et al.18 However, Cho/Cr and m-Ins/Cr showed different results between the present and previous studies. With respect to the clinical background of ALS patients, only patients with limb onset (12 patients with upper limb onset and 2 with lower limb onset) were recruited in the study of Carew et al.18 Our study included 3 patients with bulbar onset. Our electrophysiological data showed that 4 patients with leg onset and 1 with bulbar onset had no denervation or neurogenic findings in the cervical region. These patients had no significant changes of metabolite ratios compared with controls. Two previous studies by Carew et al. did not describe electromyographic findings in ALS patients.18,19 Different clinico-electrophysiological backgrounds may have led to dissimilar results of Cho/Cr and m-Ins/Cr between the present and previous studies. With respect to the relationship between clinical severity and metabolite ratios, Carew et al.18 suggested significant associations between NAA/Cho, NAA/m-Ins and FVC. However, the ALSFRS was not related to any metabolite ratios. We found that NAA/Cr and NAA/m-Ins were correlated with ALSFRS and FVC at the time of 1H-MRS measurement. Both ratios were inversely linked to the annual declines of ALSFRS and FVC. There was no relationship between these metabolite ratios and bulbar ALSFRS. Previous studies of brain 1H-MRS disclosed that ALSFRS was associated with NAA/Cr or NAA/m-Ins in the motor cortex.9,12 The bulbar subscore of ALSFRS was inversely correlated with glutamate + glutamine/Cr in the medulla oblongata.21 In general, cervical cord 1H-MRS seems to detect damage in both upper and lower motor neurons. Neuronal degeneration and gliosis are thought to cause decreased NAA levels and increased m-Ins levels in ALS patients, respectively. NAA/Cr and NAA/m-Ins were markedly decreased in our 2 patients with possible ALS who had upper limb onset and electromyographic findings in the C2 paraspinal muscles, in addition to definite and probable patients. Therefore, this combined study of C1–C3 cord 1H-MRS and clinico-electrophysiological examination indicated that damage of lower motor neurons innervating the neck, upper limb and respiratory muscles could contribute to reductions of NAA/Cr and NAA/m-Ins.
There were several limitations to this study. The number of ALS patients was relatively small. C1–C3 cord 1H-MRS failed to detect remarkable changes of metabolite ratios in some patients who did not have clinico-electrophysiological findings in neck and upper limb muscles. In the technical limitation of 1H-MRS, single voxel 1H-MRS is not sufficient to differentiate the gray and white matter of the spinal cord. In ours and previous studies of spinal cord 1H-MRS, a VOI was positioned at the high cervical cord.18,19,23–25 The method for evaluating metabolite levels in the low cervical cord or lumbar cord containing the anterior horn cells that innervate the hand and leg muscles is preferred. However, magnetic inhomogeneities and artifacts by respiratory, cardiac and aortic motion, and additional effects of spinal spondylosis cause nonreliable data on 1H-MRS. The major problem with 1H-MRS studies is the physical state of ALS patients. For setting and measurement of 1H-MRS, complete immobility is needed, and patients must be kept in the supine position for 15–20 min. This situation is unsuitable or difficult in ALS patients who require ventilator use, have increased respiratory secretions, or are bedridden. Spinal cord 1H-MRS is currently limited by methodological restrictions and the patient's condition. Technical improvement of 1H-MRS measurement is expected to resolve these limitations.
In conclusion, we found significant changes of NAA/Cr, m-Ins/Cr, and NAA/m-Ins in ALS patients. NAA/Cr and NAA/m-Ins were correlated with ALSFRS and FVC. Both metabolite ratios were inversely linked to the annual declines of ALSFRS and FVC. C1–C3 cord 1H-MRS could be a valuable test for evaluating disease severity and progression in ALS patients. Changes of NAA/Cr and NAA/m-Ins may reflect the damage of the anterior horn cell and gliosis. Further longitudinal studies in patients with early-stage ALS are needed to elucidate whether serial changes of these metabolite ratios are associated with clinico-electrophysiological progression.