Increased glutamate/glutamine compounds in the brains of patients with fibromyalgia: A magnetic resonance spectroscopy study




Fibromyalgia (FM) has been defined as a systemic disorder that is clinically characterized by pain, cognitive deficit, and the presence of associated psychopathology, all of which are suggestive of a primary brain dysfunction. This study was undertaken to identify the nature of this cerebral dysfunction by assessing the brain metabolite patterns in patients with FM through magnetic resonance spectroscopy (MRS) techniques.


A cohort of 28 female patients with FM and a control group of 24 healthy women of the same age were studied. MRS techniques were used to study brain metabolites in the amygdala, thalami, and prefrontal cortex of these women.


In comparison with healthy controls, patients with FM showed higher levels of glutamate/glutamine (Glx) compounds (mean ± SD 11.9 ± 1.6 arbitrary units [AU] versus 13.4 ± 1.7 AU in controls and patients, respectively; t = 2.517, 35 df, corrected P = 0.03) and a higher Glx:creatine ratio (mean ± SD 2.1 ± 0.4 versus 2.4 ± 1.4, respectively; t = 2.373, 35 df, corrected P = 0.04) in the right amygdala. In FM patients with increased levels of pain intensity, greater fatigue, and more symptoms of depression, inositol levels in the right amygdala and right thalamus were significantly higher.


The distinctive metabolic features found in the right amygdala of patients with FM suggest the possible existence of a neural dysfunction in emotional processing. The results appear to extend previous findings regarding the dysfunction in pain processing observed in patients with FM.

Fibromyalgia (FM) is a clinical syndrome defined by the presence of chronic widespread musculoskeletal pain and the presence of at least 11 of 18 body tender points that are representative of a possible enhanced sensitivity to painful stimulation (1), and these features are often accompanied by other symptoms such as fatigue, poor sleep quality, loss of memory, and mood disturbances. The lack of clear evidence of peripheral tissue damage to which such symptoms could be attributed (2) and the findings of peculiar abnormalities in both the neural pain modulation pattern (3, 4) and the hypothalamic–pituitary–adrenal axis (5) in these patients have contributed to the current understanding of FM as a central sensitization syndrome of unknown origin. In our conceptual framework, FM symptoms are related to neural and psychological abnormalities in sensory and emotional processing that result from biologic (genetic) predispositions, early adverse experiences (emotional neglect, physical or sexual abuse, family dysfunction), psychological vulnerabilities (neuroticism, negative affect, social learning), and biographically related precipitating factors (physical or psychosocial stress) (6, 7).

In the last decade, the important progress made in neuroimaging techniques has enabled the study of several aspects of cerebral pain processing. Research on patients with FM using functional magnetic resonance (MR) imaging has shown an increased activation in the somatosensory cortex (I and II), prefrontal cortex, parietal lobe, anterior cingulated cortex, insula, basal ganglia, putamen, and cerebellum (8–10) in response to both low and high stimuli. However, although these functional MR findings corroborated the understanding of FM as a sensitization state of the nociceptive system, they did not provide information about the particular brain metabolism in patients with FM. More recently, MR spectroscopy (MRS) studies of patients with FM have demonstrated significant relationships between the levels of brain metabolites in the insula and prefrontal cortex and the severity of clinical and experimental pain (11, 12), while 2 other studies have shown significant differences in the hippocampal levels of metabolites between patients with FM and healthy controls (13, 14). Although these MRS findings in patients with FM are heterogeneous and difficult to integrate in a comprehensive way, they are compatible with the notion of a possible neural dysfunction in sensory and emotional processing.

The present study aimed to evaluate the brain metabolite patterns in patients with FM through spectroscopy techniques, with the starting hypothesis being that dysfunctions in the amygdala and thalamus/prefrontal system might play an important role in the sensory, cognitive, emotional, and affective symptom profiles observed in patients with FM. In fact, an important assumption of the study was that the psychopathologic states and emotional effects of pain and disability are not “impure elements” of FM but, rather, are important clinical manifestations that usually form part of the physiopathology of this syndrome.



A group of 35 women with FM was consecutively recruited from among patients seen at the Fibromyalgia Unit of the Hospital Clinic in Barcelona. These patients were all selected and diagnosed according to the American College of Rheumatology 1990 criteria for FM (1). Patients with a clinical history of brain injuries, seizures, substance abuse, diabetes, hypertension, or systemic disease were excluded. All patients gave written informed consent prior to participation in the study. Because 5 women did not complete the whole protocol, the final FM cohort comprised 30 patients. A control group of healthy women (n = 30), matched to the patients by age and education status, was selected according to the same criteria and assessed by means of the same psychological, neuropsychological, and neuroimaging procedures.

Neither the patients nor the control subjects had taken any drugs that act on brain metabolism in the 4 days prior to being assessed. The study design was approved by the Clinical Research and Ethics Committee of the Hospital Clinic of Barcelona.

Clinical assessment.

Demographic variables (age, marital status, education level, and occupational status) and duration of disease evolution were recorded. All patients were symptomatic and were evaluated with the following scales: To measure pain and fatigue, we used a 10-cm horizontal visual analog scale (VAS), ranging from 0 (no pain/no fatigue) to 10 (severe pain/severe fatigue). In addition, all patients were physically examined for the number of tender points, according to the method of the Tender Point Manual (15). Perceived disability level was measured by the Spanish version of the Health Assessment Questionnaire (HAQ) (16), which is based on questions about the difficulty in performing 20 instrumental activities of daily living, distributed into 8 subscales. High scores denote the most limited function. Furthermore, all patients completed the Spanish version of the Fibromyalgia Impact Questionnaire (FIQ) (17). This is a 10-item self-report questionnaire developed to measure the health status of patients with FM. The first item focuses on the patient's ability to carry out muscular activities. In the next 2 items, patients are asked to circle the number of days in the past week during which they had felt good and how often they had missed work. Finally, the last 7 questions address ratings of job performance ability, pain, fatigue, morning tiredness, stiffness, anxiety, and depression, as measured using 10-cm VAS scores.

Psychopathologic assessment.

The anxiety and depression state was evaluated with the Spanish version of the Hospital Anxiety and Depression Scale (HADS) (18, 19). This is a scale of 14 multiple-choice items designed to assess the presence and extent of anxiety and symptoms of depression in medical outpatients. Patients with scores higher than 11 are classified as a case, and patients with scores between 8 and 10 are considered to be a probable case.

MR imaging and spectroscopy acquisition.

Data were obtained on a 1.5T whole-body MR scanner (General Electric Sigma System). A set of high-resolution T1-weighted images was acquired with fast spoiled gradient-recalled acquisitions (time to recovery [TR]/time to echo [TE] 12 msec/5.2 msec, inversion time 300 msec, number of excitations = 1, field of view 24 × 24 cm, 256 × 256–pixel matrix). Data on the whole brain were acquired in a coronal plane, yielding contiguous slices with a thickness of 1.5 mm.

For MRS, 1H-spectrum images were obtained with a standard quadrature head coil. Proton spectra were obtained from 6 volumes of interest (VOIs) placed on 1) both the right and left amygdala (voxel size 3.37 cm3, 1.5 cm longitudinal relaxation [LR] × 1.5 cm action potential [AP] × 1.5 cm signal intensity [SI]), 2) both thalami (voxel size 2.25 cm3, 1 cm LR × 1.5 cm AP × 1.5 cm SI), and 3) both orbitofrontal cortices (voxel size 3.37 cm3, 1.5 cm LR × 1.5 cm AP × 1.5 cm SI). The voxels were placed in the coronal section, and the procedure was applied in the same manner in all subjects, to ensure standard placement (Figure 1). Care was taken to avoid white matter in the locations of the thalami and amygdalae. In the orbitofrontal cortex location, white matter could not be avoided. Spectra were acquired via a double spin echo point-resolved spectroscopy sequence (PRESS), with a TR of 1,500 msec, TE of 35 msec, 2,048 data points, and automatic shimming and water suppression. PRESS is a good method for a no-loss sequence if false signals can be minimized at short TE (20). With a short TE of 35 msec, metabolites with both short and long T2 relaxation times are observed. N-acetylaspartate (NAA), choline (Cho), creatine (Cr), glutamate (Glu)/glutamine (Gln), and myo-inositol (Ins) were studied in the selected regions of interest.

Figure 1.

Coronal magnetic resonance 1H-spectrum imaging showing the locations (boxed areas) of the orbitofrontal cortex (upper row), thalami (middle row), and amygdala (bottom row) of patients with fibromyalgia.

Absolute metabolite quantification by linear combination model fitting.

For the quantification of absolute concentrations of brain metabolites, expressed in mmoles/kg wet weight, we used the user-independent frequency domain–fitting program LCModel, version 6.1-4A (21, 22), applying an eddy current correction and using internal water signal reference to calculate absolute metabolite concentrations. Some metabolites are quite difficult to resolve from others (22), and the sum of the concentrations of metabolites with similar spectra is more accurate than the individual concentrations. Therefore, apart from the individual analysis of the Cr and Ins compounds, we studied the summed concentrations of the following 3 compound pairs: NAA + N-acetyl-aspartylglutamate, referred to as total NAA, glycerophosphocholine + phosphocreatine, referred to as total Cho, and glutamate + glutamine, referred to as Glx. Absolute metabolite values were only considered when the Kramer-Rao lower bound (22) was below 20%, thus indicating that these metabolites could be reliably estimated (Figure 2).

Figure 2.

Spectra detected by magnetic resonance spectroscopy of both amygdala (boxed areas in 1H-spectrum images) of a representative patient with fibromyalgia, with postprocessing using the user-independent frequency domain–fitting program LCModel, version 6.1-4A. In this patient, the results reveal that Glx concentrations are higher in the right amygdala.

Statistical parametric mapping (SPM) segmentation was performed to correct for the amount of gray matter, white matter, and cerebrospinal fluid (CSF) included in the VOI (SPM% software, running in Matlab 6.5; MathWorks). Metabolite concentrations were adjusted for the amount of CSF contained within the VOI by dividing the metabolite concentration between the volume fraction (percentage of nervous tissue gray matter and white matter contained in the VOI) and multiplying by 100.

Although a total of 30 patients with FMS and 30 healthy controls were originally included in the study, absolute metabolite values could not be analyzed for 2 patients and 6 controls, owing to the poor quality of spectroscopy resonance and because the Kramer-Rao lower bound was higher than 20%. Therefore, since a fitting analysis for some metabolites could not be performed in these subjects, the final sample for the spectrum analysis comprised 28 patients with FM and 24 controls.

Statistical analysis.

Descriptive values are expressed as the mean ± SD, and groups were compared using Student's t-test for continuous variables and the chi-square test for categorical variables. Bonferroni post hoc corrections were calculated for every pair of right and left regions. Pain, fatigue, and depression were treated as continuous variables, and the Pearson's correlation test was used to study the relationship between these variables and the metabolite concentrations determined on spectroscopy. P values less than 0.05 were considered significant. For the pain and fatigue variables, a cutoff of 6 cm on the VAS was used to differentiate light-to-moderate pain or fatigue (VAS scores <6 cm) and moderate-to-severe pain or fatigue (VAS scores ≥6 cm). For the depression variable, a cutoff value of 11 on the HADS test was used to differentiate severe depression. We compared different groups using Student's t-test.


The sociodemographic, psychopathologic, and clinical characteristics of the patients with FM and healthy controls are shown in Table 1. There were no significant differences in age, marital status, or education level between patients and healthy controls, but, as would be expected, patients with FM showed significantly more anxiety and more symptoms of depression than did healthy controls.

Table 1. Sociodemographic, psychopathologic, and clinical characteristics of the patients with fibromyalgia (FM) and healthy controls*
VariablePatients with FM (n = 30)Healthy controls (n = 30)P
  • *

    Except where indicated otherwise, values are the mean ± SD. NS = not significant; HADS = Hospital Anxiety and Depression Scale (Spanish version); VAS = visual analog scale; FIQ = Fibromyalgia Impact Questionnaire; HAQ = Health Assessment Questionnaire (Spanish version).

 Age, years42.62 ± 8.7643.86 ± 10.60NS
 Marital status, no. (%)   
  Married20 (66.6)20 (66.6)NS
  Single5 (16.6)4 (13.3)NS
  Divorced5 (16.6)6 (20)NS
 Education status, no. (%)   
  <8 years11 (36.6)10 (33.3)NS
  8–12 years13 (43.3)11 (36.6)NS
  >12 years6 (20)9 (30)NS
 Anxiety, HADS score12.71 ± 4.707.00 ± 3.160.001
 Depression, HADS score9.97 ± 4.202.81 ± 2.400.001
 Duration of evolution, months151 ± 120
 Number of tender points16.19 ± 2.43
 VAS pain score6.98 ± 1.78
 VAS fatigue score7.35 ± 1.52
 Health status, FIQ total score60.45 ± 13.19
 Perceived disability, HAQ score2.26 ± 3.49

On MRS analysis, differences between patients with FM (n = 28) and healthy controls (n = 24) were found in the right amygdala. Patients with FM showed higher Glx levels in the right amygdala than did controls (mean ± SD 13.4 ± 1.7 arbitrary units [AU] versus 11.9 ± 1.6 AU; t = 2.517, 35 df, P = 0.017, P with Bonferroni correction for multiple comparisons [Pcorr] = 0.03), and had higher Glx:Cr ratios in the right amygdala than did controls (2.4 ± 1.4 versus 2.1 ± 0.4; t = 2.373, 35 df, P = 0.02, Pcorr = 0.04). Among patients with FM, there were no significant differences in Glx levels in the amygdala related to disease duration, levels of pain intensity, fatigue, anxiety, and depression, or degree of disability. In contrast, in the left thalamus of patients with FM, higher Glx levels were related to higher VAS scores for fatigue (r = 0.463, P < 0.05) and to greater pain intensity (patients grouped by VAS pain scores ≥6 [n = 16] versus <6 [n = 12]) (t = 2.347, 25 df, P = 0.02), although there were no significant correlations between overall VAS scores for pain and Glx levels.

Patients with FM and healthy controls had similar levels of Glx in the left amygdala, thalamus, and orbitofrontal cortex (Table 2). No differences in the levels of the other metabolites measured in the thalami, both orbitofrontal cortices, and the amygdalae were found between patients and controls. However, within the group of patients with FM, levels of Ins were significantly higher in the right thalamus of those with more pain (VAS scores ≥6) (t = 2.581, 19 df, P = 0.01) and more tender points (r = 0.502, P < 0.01), and levels of Ins were also significantly higher in the right amygdala of patients with more pain (t = 2.609, 24 df, P = 0.01 and r = 0.410, P < 0.05) and greater fatigue (r = 0.560, P < 0.01) as measured on a linear scale (Table 3).

Table 2. Levels of glutamate/glutamine in all regions of the brain in patients with fibromyalgia (FM) compared with healthy controls*
Brain regionPatients with FM (n = 28)Healthy controls (n = 24)
  • *

    Values are the mean ± SD arbitrary units.

  • P = 0.03 versus controls.

 Right13.4 ± 1.711.9 ± 1.6
 Left11.9 ± 1.911.7 ± 1.7
 Right11.6 ± 2.011.2 ± 1.6
 Left11.6 ± 2.211 ± 2.0
Prefrontal cortex  
 Right10.4 ± 1.410.8 ± 1.3
 Left10.5 ± 2.410.1 ± 1.9

Another interesting observation was that patients with FM who reported having higher levels of disability (higher HAQ scores) or reduced health status (lower FIQ scores) had lower levels of Cr, NAA, or Glx in the amygdala or prefrontal cortex, as shown in Table 3. In addition, as shown in Table 3, significant correlations between brain metabolite levels and other clinical variables (number of tender points and duration of disease) were found in patients with FM.

Table 3. Significant linear correlations between levels of brain metabolites and clinical variables in patients with fibromyalgia (n = 28)*
Brain region, brain metaboliteClinical variablerP
  • *

    Cho = choline; Ins = myo-inositol; VAS = visual analog scale; Cr = creatine; HAQ = Health Assessment Questionnaire (Spanish version); NAA = N-acetylaspartate; Glx = glutamate/glutamine; FIQ = Fibromyalgia Impact Questionnaire (Spanish version).

Right amygdala   
 ChoDuration of disease0.485<0.01
 InsVAS fatigue score0.560<0.01
 InsVAS pain score0.410<0.05
 CrDisability (HAQ score)−0.658<0.001
 NAADisability (HAQ score)−0.5600.004
Left amygdala   
 InsDuration of disease0.435<0.05
 InsDisability (HAQ score)−0.630<0.01
 GlxDisability (HAQ score)0.476<0.05
Right thalamus   
 ChoDuration of disease0.558<0.01
Left thalamus   
 InsNumber of tender points0.555<0.05
 GlxVAS fatigue score0.463<0.05
 CrDepression score−0.408<0.03
Right prefrontal cortex   
 ChoFIQ score−0.432<0.05
 NAA:CrFIQ score−0.682<0.001
Left prefrontal cortex   
 NAA:CrFIQ score−0.710<0.001
 NAADisability (HAQ score)0.454<0.01
 GlxFIQ score−0.695<0.001
 InsFIQ score−0.800<0.01
 ChoVAS fatigue score0.582<0.05
 CrNumber of tender points0.436<0.05


To our knowledge, this is the first study to compare brain metabolism in the amygdala, thalami, and prefrontal cortex of patients with FM and healthy controls in a resting condition. Our main finding was the increased levels of glutamate compounds (Glx) observed in the right amygdala. In 2 recent studies by Harris et al, significantly higher levels of glutamate in the right posterior insula were observed in patients with FM than in healthy controls (23), and significant relationships between changes in glutamate levels in this area and pain threshold or changes in multiple pain domains (pain threshold and clinical pain) after acupuncture treatment were observed (11). Although brain metabolites in the insula were not analyzed in our study, the baseline increase in Glx levels found in the right amygdala of our patients with FM suggests that some type of glutamergic dysfunction is occurring in the limbic areas of these patients. In our patients, the higher levels of glutamate compounds in the amygdala were not related to pain, fatigue, duration of disease, disability level, anxiety, or depression. However, FM patients with more pain and fatigue did show significantly higher levels of Ins in the right amygdala. These results suggest a complex role of the amygdala in the biology of FM, since the increased Glx metabolism was associated with the diagnosis of FM, while increased Ins metabolism was related to the most severe symptoms of the disease.

Glutamate is a major excitatory neurotransmitter within the nervous system and functions in pain neuropathways. Its increased levels in the CSF of patients with FM (24) suggest that it could be responsible for the augmented pain transmission observed (8, 9), and also suggest that it might be used as a biomarker of FM disease severity (11). Thus, the increased Glx levels found in the right amygdala of patients with FM in our study, as well as the increased Glu levels described by Harris et al (11) in the insula of patients with FM, could be considered to be regional manifestations of a more general abnormality in the metabolism of Glu in the brains of these patients.

The higher Glx:Cr ratios in the left thalamus of patients with FM appeared to be related to pain intensity and were correlated with the number of tender points. However, in patients with more pain, Ins levels were significantly higher in the right thalamus, and increased levels of Ins in the right thalamus were also positively correlated with the number of tender points. Abnormalities in the left (and right) thalamus in patients with FM have been found in other neuroimaging studies, using voxel-based morphometry techniques (25) and measures of regional cerebral blood flow (26, 27). In all of these studies, the lower functional thalamic activity in patients with FM was considered to be secondary to a mechanism of tonic inhibition maintained by the persistent excitatory output associated with pain. Ins is a brain metabolite mainly located in astrocytes, where it is the most important osmolyte to regulate cell volume, while Glx is a brain compound considered to be a neurone-astrocyte marker. Therefore, total spectroscopy findings in the right amygdala of patients with FM could suggest a possible role for astrocyte activity in the disease.

Metabolic activity in the prefrontal cortex of patients with FM did not differ from that found in healthy controls. However, in the left prefrontal cortex of patients with FM, positive correlations were found between Cr levels and the number of tender points (r = 0.436, P < 0.05), between Cho levels and fatigue (r = 0.582, P < 0.05), and between NAA levels and disability (r = 0.454, P < 0.01). Although differences between patients with FM and controls in terms of the activity of their prefrontal brain metabolism have not been reported previously, our findings suggest that the prefrontal cortex does play a role in the physiopathology of FM, since FIQ scores (which evaluate the impact of FM on health status) were negatively correlated with the levels of Ins, Glx, and NAA in the left prefrontal cortex, and negatively correlated with the levels of Cho and NAA in the right prefrontal cortex. These metabolic findings are similar to the results of a previous study that showed significantly lower NAA levels in the dorsolateral prefrontal cortex of patients with chronic back pain (28).

More recently, higher Cho levels and lower NAA:Cho and NAA:Cr ratios in both hippocampi have been reported in patients with FM (13), as has a significant negative correlation between the NAA:Cr ratio in the right hippocampus and the intensity and effects of pain (14). The finding of metabolic brain differences between patients with FM and healthy controls in neural structures such as the hippocampus and amygdala (both of which pertain to the limbic system and are involved in fear, avoidance, and emotional responses experienced during pain) is compatible with a possible augmented emotional processing in patients with FM, in line with the augmented pain processing proposed by some authors (8, 27).

Our spectroscopy findings seem to show no correlation of brain metabolite levels with the higher prevalence of psychiatric symptoms among patients with FM. It should be mentioned that anxiety was not found to be related to the other clinical variables or to levels of brain metabolites, and that among patients with depression, only 2 differences were found in the spectroscopy analyses: significantly higher Ins levels in the right amygdala, and lower Cr levels in the left thalamus. This apparent lack of relationship between emotional and affective states and the clinical and spectroscopic findings supports the understanding of FM as a systemic disorder, rather than just a set of somatic symptoms secondary to psychopathology.

In summary, patients with FM showed higher levels of glutamate compounds in the right amygdala than did healthy controls, and pain was related to increased glutamate levels in the left thalamus and higher Ins levels in the right amygdala. The role of the amygdala in emotional responses, along with the differences found on spectroscopy in the hippocampus of patients with FM, suggest the existence of some type of abnormal emotional processing that would partially explain the proneness of patients with FM to show neurotic features, anxiety, and symptoms of depression. These findings also outline the possible nature of FM as a systemic disorder that is mainly expressed through sensorineural dysfunction and abnormal neuroendocrine stress responses.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Collado had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Valdés, Collado, Bargalló, Vázquez, Rami, Gómez, Salamero.

Acquisition of data. Valdés, Collado, Bargalló, Vázquez, Rami, Gómez, Salamero.

Analysis and interpretation of data. Valdés, Collado, Bargalló, Vázquez, Rami, Gómez, Salamero.