Changes in Metabolites after Treatment with Memantine in Fibromyalgia. A Double‐Blind Randomized Controlled Trial with Magnetic Resonance Spectroscopy with a 6‐month Follow‐up

Summary

Aim

To evaluate the efficacy of memantine on metabolite levels in different areas of the brain and to determine whether changes in metabolite levels correlate with clinical variables in Fibromyalgia (FM) patients.

Methods

Doubled‐blind parallel randomized controlled trial. Twenty‐five patients diagnosed with FM were enrolled in the study. Patients were administered questionnaires on pain, anxiety, depression, quality of life, and cognitive impairment, and single‐voxel MRS of the brain was performed. All assessments were performed at baseline and after 6 months of treatment with memantine or placebo.

Results

Patients treated with memantine exhibited a significant increase in the glutamate (P = 0.010), glutamate/creatine ratio (P = 0.013), combined glutamate + glutamine (P = 0.016) and total N‐acetyl‐aspartate (NAA+NAAG) (P = 0.034) in the posterior cingulate cortex compared with those on placebo. Furthermore, the memantine group exhibited increases in creatine (P = 0.013) and choline (Cho) (P = 0.025) in the right posterior insula and also a correlation between choline and the Fibromyalgia Impact Questionnaire (FIQ) in the posterior insula (P = 0.050) was observed.

Conclusion

Memantine treatment resulted in an increase in cerebral metabolism in FM patients, suggesting its utility for the treatment of the illness.

Introduction

Fibromyalgia (FM) is a chronic rheumatic disease of currently unknown etiology that is characterized by the presence of diffuse musculoskeletal pain and painful sensitivity to touch in at least 11 of 18 defined trigger points 1. The prevalence of this syndrome in Europe is calculated at approximately 2.9% (95% CI: 2.4–3.4) 2. In Spain, the prevalence of FM in rheumatology consultations is 12% (2.2% in men and 15.5% in women) 3. Fibromyalgia is one of the main health problems presently affecting western countries owing to its high prevalence, clinical impact on patient disability, diminished quality of life, and significant healthcare costs that it incurs. FM treatments have limited efficacy, with an effect size of approximately 0.5 4.

The cause of hyperalgesia in FM is unknown, although it is suspected that there is a functional alteration in the central nervous system structure. In recent years, the neurophysiology of the pain process has led to increased interest in identifying the brain structures activated when patients experience pain. Different neuroimaging methods have been used to identify these structures, including a positron emission tomography (PET) study in FM revealing abnormalities of regional cerebral blood flow in the thalamus 5 and presynaptic dopaminergic activity in several brain regions in which dopamine plays a critical role in modulating nociceptive processes 6. Single photon emission computed tomography (SPECT) 7, 8 demonstrates a reduction in the cerebral blood flow of FM patients, and diffusion tensor and volumetry 9, 10 reveals white and gray matter abnormalities. Previous functional neuroimaging studies on FM confirm that patients with FM exhibit augmented neuronal responses to both innocuous and painful stimuli 11. These studies are consistent with the allodynia and hyperalgesia observed in this condition 12, 13.

A substantial body of literature, using Proton magnetic resonance spectroscopy (¹H‐MRS), has demonstrated that there are abnormalities in brain metabolites observed in individuals with FM. The most frequent findings are an increase in glutamate (Glu) and glutamate + glutamine (Glx) levels in the posterior insula 14 increase in Glx levels and an increase in the Glx/creatine (Glx/Cr) ratio in the amygdale 15 as well as the posterior cingulate cortex 16, 17 and a significant N‐acetyl‐aspartate (NAA) reduction in the hippocampus in patients with FM, suggesting a neuronal abnormality 18.

Memantine acts as a neuroprotectant by decreasing glutamate excitotoxicity and has been known to increase levels of brain‐derived neurotrophic factor (BDNF), thus influencing synaptic plasticity in rats and reducing β‐amyloid‐induced apoptotic death and neuroinflammation in the hippocampus 19. In the Morris water maze, memantine improved acquisition performance and spatial accuracy and increased the durability of synaptic plasticity 20. The NMDA receptor antagonist memantine has been used to treat Parkinson's disease, spasticity, convulsions, vascular dementia, and Alzheimer's disease and has an excellent clinical safety record spanning more than 20 years. Over the past decade, memantine drug trial research has been conducted in several neurological disorders with cognitive disabilities, including Down Syndrome, Fragile X Syndrome, and Rett Syndrome 21. Memantine exhibits a very low incidence of side effects in human clinical trials 22, 23, and a recent extension of trials has demonstrated the drug's clinical tolerability even with prolonged use 24. Recent research has highlighted the efficacy of memantine for the treatment of complex regional pain syndrome 25 and phantom limb pain 26.

The purpose of this study is to evaluate the efficacy of memantine in modifying the levels of brain metabolites in FM patients. The secondary aim is to assess the efficacy of memantine for pain treatment and other key FM symptoms such as cognitive dysfunction, depression, anxiety, and quality of life in a pre–post study. We also expected that the differences in metabolite ratios would correlate with the degree of clinical improvement.

Materials and Methods

This study is a randomized double‐blind placebo‐controlled clinical trial performed on patients suffering from FM with a 6‐month follow‐up with no commercial interest. The patients were randomized in two parallel groups: a treatment group, which was given 20 mg of memantine daily after a titration period of 1 month, and a control group, which received a placebo. There was a 6‐month follow‐up period (including the dose adjustment period of 1 month).

The study was conducted in accordance with the standards of good clinical practice and was performed according to the Initiative on Methods, Measurement and Pain Assessment in Clinical Trials (IMMPACT) 27, which recommends the inclusion of a set of core outcome domains in the clinical trials of pain treatments. The study also followed the recommendations established by the Consolidated Standards of Reporting Trials (CONSORT) statement 28, 29 for randomized, controlled trials. The protocol of this study has been published previously 30.

Patients diagnosed with FM were recruited for inclusion in the study from primary healthcare centers in Zaragoza, Spain. The patients were selected based on the following criteria: age 18–65, Spanish comprehension skills and an FM diagnosis from an American College of Rheumatology‐certified rheumatologist 6. Additional inclusion and exclusion criteria have been published in the trial 30.

Of the 33 patients initially screened for inclusion, eight patients were withdrawn from the trial because they revoked informed consent, for safety or efficacy reasons (as determined by the researcher), when it was in the best interest of the patient and when the patient did not comply with the treatment for more than seven consecutive days.

Intervention

The intervention consisted of administering 20 mg of memantine to 13 subjects. The treatment duration was 6 months, including 4 weeks of dose adjustment in which patients started with 5 mg of memantine in the first week and increased the dosage by 5 mg each week until the full dose was reached in the fourth week. The 20 mg daily dosage continued for an additional 5 months. Medications permitted during the trial included those listed on the summary of product characteristics and the patient information leaflet for memantine. No other treatment for chronic pain was permitted.

The number of tablets in each dispensed container was monitored at each evaluation, and researchers kept track of the number of pills the patient should have taken and how many should have been remaining upon completion of the treatment 30. Film‐coated tablets were administered orally. One qualified person prepared, conditioned, labeled and released the study drug according to the principles of good manufacturing practice under the responsibility of H. Lundbeck A/S. Upon completion of the trial, patients continued with the standard FM treatment according to clinical practice guidelines.

Sociodemographic and Clinical Variables. Sociodemographic Data

Gender, age, marital status, education, and occupation were collected. The variables were measured and assessed according to the following tests. The pain threshold was measured by a sphygmomanometer 31. Perceived pain was measured using a visual analog scale (PVAS) 32. Cognitive state was measured using the Cognition Mini‐Exam (MEC) 33. FM health status was measured with the Fibromyalgia Impact Questionnaire (FIQ) 34; the validated Spanish‐language version of this questionnaire was used 35. Anxiety and depression levels were measured with the Hospital Anxiety Depression Scale (HADS) 36. Quality of life was measured with the EuroQol 5D (EQ5D) questionnaire 37. Illness severity was measured with the CGI‐Severity scale 38. The psychological variables were evaluated in the first month at the start of the trial (baseline evaluation) and at 6 months.

We then performed ¹H‐MRS of the brain to explore metabolite levels in five cortical areas: the posterior cingulate cortex, right anterior and posterior insula, and both hippocampi (Figure 1). The following exploration areas were chosen: (1) areas in which the authors have observed increased glutamate levels powered by MRS 14, 15, 16, 17, (2) brain structures that are activated during painful conditions in healthy controls 39 and in FM patients 40, and (3) regions (mentioned in prior reports) that have been implicated in cognitive impairment 41, 42. The examinations were performed on a 1.5 T Signa HD clinical scanner (GE Healthcare Diagnostic Imaging, Milwaukee, WI, USA). All of the images were acquired using an eight‐channel phased‐array head coil in both the transmit and the receive mode (NVHEAD A). Localized ¹H‐MRS was performed using a short echo time (TE) of 35 ms, a repetition time (TR) of 2000 ms and 128 accumulations using a single voxel with a spin echo technique that uses selective excitation with gradient spoiling for water suppression. The mode of spectral acquisition was the proton brain spectroscopy‐point resolved spectroscopy (PROBE‐p) technique. To quantitate the absolute concentrations of brain metabolites and the ratios of the metabolites relative to creatine (which is considered the most stable metabolite, as an internal reference) 43 expressed in mmoles/kg wet weight, we used the user‐independent frequency domain‐fitting program LCModel, version 6.2‐0 (Stephen Provencher, Oakville, ON, Canada) 44. The voxel positions were not significantly different between the two exams. We used the three planes of space. Concentration values were not corrected for CSF contributions.

Figure 1.

Voxelsplacement in the right hippocampus (A), left hippocampus (B), anterior insula (C), posterior insula (D), posterior cingulate area (E) and spectrum acquired with the LCModel software (F).

Sample Size and Power

The patient sample (n = 25) was chosen, for analyzing the objectives of the study. Similar sample sizes have been used to identify significant differences in the levels of glutamate and NAA in different brain regions of FM patients and control groups 14, 15, 16, 17.

Statistical Analysis

The mean and standard deviation were used to describe social and demographic variables by treatment group. Baseline variables were compared by treatment group using the Mann–Whitney U‐test. Baseline and post‐treatments scoring of the main outcome (metabolites and clinical variables) was compared using the Wilcoxon test for paired samples. In addition, we used the nonparametric Spearman′s rho correlation in the memantine group and in the whole sample to study the relationship between brain metabolites for which the levels were significantly different and for which the clinical variables were studied. The statistical tests were two‐tailed, and calculations were made using the Statistical Package for the Social Sciences (SPSS) for Windows, version 15 (IBM España S.A. Madrid, España).

Ethics and Consent

Informed consent was obtained from the participants before they were aware of to which group they were assigned. The study follows Helsinki Convention norms with posterior modifications as well as the Declaration of Madrid of the World Psychiatric Association. The study protocol was approved by the Clinical Research Ethics Committee of Aragón (June 2012) and the Medicines and Health Products Agency of Spain (EUDRACT 2011‐006244‐73).

Results

Sample Recruitment and Sociodemographic, Clinical and Psychological Variables

Thirty‐three patients were invited to participate in the study. Eight of the patients were ruled out, six patients did not comply with the treatment for more than 7 consecutive days, one received a diagnosis from a general practitioner rather than a rheumatologist, and the remaining patient did not provide informed consent. A total of 25 patients with FM completed the trial, and the patients were randomized into two groups: (A) treatment group (n = 13; 12 women and one man), which was administered 20 mg of memantine daily; and a (B) control group (n = 12; 11 women and one man), which was administered a placebo. The study patient sample exhibited the expected characteristics of sociodemographic variables in FM patients, including a female predominance (92%) and middle‐aged patients in the treatment group and placebo group (mean: 48.1 [SD = 8.70] vs. 48.5 [SD = 8.18]; P = 0.814). In the memantine group, the mean duration of the disorder was 11.90 [SD = 11.16] years, whereas in the placebo group, the mean duration was 13.06 [SD = 9.40]. The psychological profiles revealed the usual psychological characteristics of FM patients: high scores in anxiety (61.29% vs. 59.38%) and depression (64.52% vs. 50%) assessed with the HADS. The results of the assessments using the different neuropsychological scales before and after treatment with memantine and placebo are presented in Table 1. At 6‐month follow‐up, fibromyalgia patients treated with memantine exhibited improved the mean CGI‐severity scale scores (4.50 [SD = 0.67] vs. 4.00 [SD = 0.82]; P = 0.025), FIQ scores (62.56 [SD = 17.21] vs. 49.36 [SD = 11.59]; P = 0.009) and quality of life scores (assessed by the visual analog scale of the EQ5D; 42.08 [SD = 17.77] vs. 55.00 [SD = 14.34]; P = 0.021). Patients treated with placebo exhibited decreased scores with respect to Pain sphygmo (mean: 108.46 [SD = 44.32] vs. 77.27 [SD = 17.37]; P = 0.044) and EQ5D (mean: 53.77 [SD = 20.55] vs. 45 [SD = 15.81]; P = 0.039).

Table 1.

Scores (mean [SD]) on neurophysiological scales before and after 6 months of treatment with placebo and memantine in 25 patients with fibromyalgia

Scale	Memantine (n = 13)					Placebo (n = 12)
	Baseline		Post‐treatment		P‐value	Baseline		Post‐treatment		P‐value
	Mean	SD	Mean	SD		Mean	SD	Mean	SD
CGI	4.5	0.7	4.0	0.8	0.02	4.3	0.9	4.9	0.5	0.14
PAIN sphygmo	97.9	15.9	114	27.6	0.08	108	44.3	77.3	17.4	0.04
PVAS	6.9	1.8	5.0	1.9	0.06	5.7	2.7	6.9	1.4	0.09
MEC	34.2	1.1	34.9	0.3	0.13	33.5	2.7	34.1	0.8	0.81
HADS anxiety	11.8	5.5	10.9	4.7	0.88	11.3	3.2	12.3	4.7	0.92
HADS Dep	9.0	4.9	7.2	3.5	0.16	8.9	4.3	11.2	3.8	0.06
HADS total	20.0	6.6	20.4	9.6	0.69	20.9	9.0	20.6	8.8	0.79
FIQ total	62.6	17.2	49.4	11.6	0.01	63.8	16.6	62.1	17.8	0.53
EQ5D	42.1	17.8	55.0	14.3	0.02	53.8	20.5	45.0	15.8	0.04

CGI, Scale that measures illness severity (CGIS), global improvement or change (CGIC) and therapeutic response; PAIN sphygmo, Pain threshold measured by a sphygmomanometer; PVAS, Perceived pain was measured using a visual analog scale; MEC, Spanish version of the Mini‐Mental State Examination; HADS, Hospital Anxiety Depression Scale; FIQ, Fibromyalgia Impact Questionnaire; EQ5D, Spanish‐language version of the Quality of life questionnaire.

Neuroimaging Variables. Magnetic Resonance Imaging

In all subjects, the conventional MRI brain parenchyma images were normal. Magnetic Resonance Spectroscopy: In Table 2, the mean values of the metabolite ratios, the absolute values and the P‐values are reported for every exploration area. Significant changes between the placebo and post‐treatment metabolite levels were observed in the patients. After 6 months of memantine treatment, a significant increase in the glutamate (6.76 vs. 7.66; P = 0.010), glutamate/creatine ratio (1.22 vs. 1.32; P = 0.013), NAA+NAAG (8.42 vs. 8.75; P = 0.034) and glutamate + glutamine (8.83 vs. 9.90; P = 0.016) parameters in the posterior cingulate cortex was observed in memantine‐treated patients (see Figure 2). A significant increase in the creatine (5.43 vs. 5.81; P = 0.013) and choline (Cho) (1.32 vs. 1.43; P = 0.025) levels in the right posterior insula was observed in the memantine‐treated patients (see Figure 3). In the memantine group, a correlation between choline and FIQ (P = 0.050) was observed in the right posterior insula cortices (Table 3).

Table 2.

Absolute values and ratios to creatine (mean [SD]) of various metabolites in the posterior cingulate cortex and posterior insula in 25 patients with fibromyalgia after 6 months of treatment with placebo and memantine

Location, metabolite	Placebo (n = 12)		Memantine (n = 13)		P‐value
	Mean	SD	Mean	SD
Posterior cingulate cortex
Glu	6.8	0.9	7.7	0.9	0.01
Glucr	1.2	0.1	1.3	0.1	0.01
NAA+NAAG	8.4	0.4	8.7	0.5	0.03
Glu+Gln	8.8	1.2	9.9	1.0	0.02
Posterior insula
Cr	5.4	0.4	5.8	0.4	0.01
Cho	1.3	0.2	1.4	0.2	0.02

Glu, glutamate; Glucr, Glutamate/creatine ratio; NAA+NAAG, N‐acetyl‐aspartate+N‐acetyl‐aspartyl‐glutamate; Glu+Gln, glutamate + glutamine (Glx); Cr, creatine; Cho, choline.

Figure 2.

Box plots representing glutamate (A) ratio of glutamate/creatine (B), glutamate + glutamine (C), N‐acetyl‐aspartate + N‐acetyl‐aspartyl‐glutamate (D) in the posterior cingulate area in patients with fibromyalgia syndrome who were treated with placebo or memantine for 6 months. Glu represents baseline values, while Glu2 represents values after treatment.

Figure 3.

Box plots representing the Creatine (A) and Choline (B) in the posterior insula in patients with fibromyalgia syndrome who were treated with placebo or memantine for 6 months. ‘cr’ and ‘cho’ represents baseline values, while ‘cr2’ and ‘cho2’ represents values after treatment.

Table 3.

Correlation among brain metabolites and clinical variables in memantine group

Memantine group
Changes in end–initial metabolites	Change in end–initial clinical variables
	CGI	FIQ	EQ5D
Posterior cingulate cortex
Glu	−0.4 (0.28)	0.3 (0.33)	0.05 (0.89)
Glucr	−0.1 (0.78)	0.4 (0.26)	0.11 (0.75)
NAA+NAAG	0.2 (0.6)	0.05 (0.88)	0.50 (0.11)
Glu+Gln	−0.2 (0.5)	0.32 (0.36)	0.19 (0.59)
Posterior insula
Cr	0.1 (0.77)	−0.09 (0.80)	−0.27 (0.44)
Cho	0.3 (0.38)	0.62 (0.05)	−0.04 (0.89)

CGI, Scale that measures illness severity (CGIS), global improvement or change (CGIC) and therapeutic response; FIQ, Fibromyalgia Impact Questionnaire; EQ5D, Spanish‐language version of the Quality of life questionnaire. Glu, glutamate; Glucr, Glutamate/creatine ratio; NAA+NAAG, N‐acetyl‐aspartate+N‐acetyl‐aspartyl‐glutamate; Glu+Gln, glutamate + glutamine (Glx); Cr, creatine; Cho, choline.

Spearman′s correlation coefficient (P‐value)

In the posterior cingulate, the standard deviation of the metabolites was as follows: 3–5% in Cr; 3–5% in tNAA; 7–15% in Glu; and 5–15% in Glx. The signal/noise ratio was 16–22, and the line width was between 0.038 and 0.048. In the posterior insula, the standard deviation was 4–5% in Cr and 4–7% in Cho. The signal/noise ratio was 11–17, and the line width between 0.038 and 0.067. No other studied brain areas exhibited changes in metabolites.

Discussion

This is the first randomized, controlled study of memantine for the treatment of fibromyalgia. Memantine treatment, compared with the placebo, was effective in patients with fibromyalgia with respect to clinical global impression (measured by CGI), global function (assessed by FIQ), and quality of life (assessed by the visual analog scale of the EQ5D). Memantine might be expected to be useful for the treatment of FM based on its pharmacological effects 25, 26, 45. Therefore, longer follow‐ups are necessary to confirm the time stability of the improvement produced with memantine.

Because there is no definitive biomarker to detect and monitor progression in FM, neuroradiological techniques are helpful for this purpose 46. MRS was able to reflect the inter‐individual variations in the clinical response that occur in clinical trials. These results point to a potential value of MRS to monitor response to treatment in FM, which might be of help in investigating the effect of new drugs.

Magnetic resonance spectroscopy is a promising technique to monitor the response to drugs in FM because it is noninvasive and reproducible in nature. MRS studies in fibromyalgia have reported a strong correlation between high levels of glutamate and pain 14, 15, 16, 17; therefore, the use of an agent inhibiting the NMDA receptor aimed at decreasing brain glutamate levels could be expected to improve pain. Indeed, some authors previously suggested the use of memantine in FM 14, 47. A trial including patients with fibromyalgia and normal controls treated with the N‐methyl‐D‐aspartate receptor (NMDAR) dextromethorphan concluded that other mechanisms need to be considered for the widespread pain of FM patients. These mechanisms might include tonic nociceptive inputs from peripheral tissues and/or enhanced descending facilitation 48. The benefits of memantine in FM treatment are thus expected to be 3‐fold: (1) neuroprotection via antagonism of NMDARs, (2) analgesia through the normalization of dysregulated pronociceptive and antinociceptive pathways, and (3) enhanced analgesia and prevention of opioid tolerance in a combinatorial analgesic approach 49.

In this trial, we observed a significant increase in the glutamate (Glu), glutamate/creatine ratio (Glu/Cr), glutamate + glutamine (Glx), and total NAA (N‐acetyl‐aspartate + N‐acetyl‐aspartyl‐glutamate [NAA+NAAG]) in the posterior cingulate cortex in patients treated with memantine compared with placebo. These amino acids are of the greatest importance in neurological function and in many other metabolic processes. Their detection and monitoring in disease is one of the most promising areas of clinical MRS. Given the extremely high concentration of glutamate in brain tissue paired with its excitotoxic potential, tight physiological regulation of extracellular glutamate levels and receptor signaling is required to assure optimal excitatory neurotransmission and limit excitotoxic damage. The tight control of glutamatergic neurotransmission is an energy costly process, requiring multiple regulatory processes and high levels of glucose and oxygen consumption 50. Therefore, if glutamate is involved in the pathophysiology of FM, brain energy metabolism may be an underlying influence in FM and its therapeutic remediation. It is important to note that in addition to its role as a neurotransmitter, glutamate also serves as a metabolic precursor to GABA and as a component of various amino acid‐based derivatives, for example, the antioxidant glutathione. Consistent with the key role of glutamate in multiple aspects of brain physiology, metabolic studies have determined that virtually all of the glucose that enters the CNS is eventually converted to glutamate 50.

Mitochondria provide the necessary energy for cell life and cause cell death by necrosis and/or apoptosis in cases of noxious stimuli. The effects of calcium concentrations on mitochondrial bioenergetics and oxidative mechanisms may provide valuable clarifications in the responses of isolated mitochondria. Furthermore, it may be beneficial to design therapeutic strategies on the basis of mitochondrial dysfunction parameters, as mitochondrial dysfunction is prominent in common pathologies following traumatic brain injury, spinal cord injuries and other neurodegenerative conditions 51.

¹³C‐MRS studies suggest that the majority of energy related to Glu metabolism is used to support events associated with Glu neurotransmission 52 and provide a quantitative measure of synaptic glutamate release 53. The efficacy of memantine in reducing glutamate receptor site concentrations is due to the low affinity for and rapid voltage‐dependent interaction with the NMDA receptor. Studies demonstrate that memantine, which belongs to the N‐methyl‐D‐aspartate (NMDA) family of receptor antagonists, could reduce the harmful effects of excessively high levels of brain glutamate that are found in Alzheimer′s disease 54, which may be consistent with its anti‐excitotoxic properties 55.

The mechanism of action underlying the anti‐excitotoxic properties of memantine is not a reduction in the levels of glutamate or its release; instead, the neurotoxic effect of glutamine is reduced, preventing the entry of excess calcium as it blocks the (NMDA) receptor 48. A recent study 56 suggested that administration of the NMDA receptor antagonist memantine for 21 days significantly reduced glutamate + glutamine (Glx) concentrations, and this reduction was associated with a reduction in brain activation in the prefrontal cortex. Glutamate (Glu) and glutamine (Gln) are strongly compartmentalized (in neurons for Glu and in astrocytes for Gln) and are directly connected to energy metabolism and neurotransmission 57. Mechanisms that might increase the combined Glx (Glu+Gln+GABA) concentration include greater presynaptic vesicular release of Glu 58, faster breakdown of NAAG into NAA and Glu 59, slower conversion of Gln to GABA 60 and net production rather than consumption of Glu by the Krebs cycle in neurons and astrocytes 61, 62. The physiological responses observed on functional MRIs and PET scans and that arise from increased energy demand with neuronal activation have been proposed to be directly related to this glutamate/glutamine (Glx) cycle 6.

N‐acetyl‐aspartate is a marker of density neuronal and axonal viability and mitochondrial metabolism 63. The function of NAA within axons in the white matter is unknown, but one of its roles may involve the synthesis of neurotransmitters 64. N‐acetyl‐aspartyl‐glutamate (NAAG) is colocalized with NAA in neurons and releases NAA and glutamate when it is cleaved by N‐acetylated alpha‐linked dipeptidase 65. glutamate and possibly NAAG are excitatory amino acids. Within physiological concentrations, glutamate can be neurotoxic. Data suggest that NAAG may be the form in which the neuron stores glutamate to protect the cell from the excitatory and potentially neurotoxic action of glutamate. For example, the lower hippocampal NAA levels suggest a neuronal or an axonal metabolic dysfunction or some combination of these processes 66. Recent studies have indicated that NAA reflects functional rather than structural neuronal characteristics, suggesting that NAA is most informative in the investigation of functional abnormality 67. Beyond being osmolytes, NAA and NAAG also exhibit specific roles as “molecular water pumps” and are active in transporting very large quantities of water outside the neuron in proportion to their molar concentration 68. As much of this water originates from the cell energy chain, essentially all of the canonical metabolites can be linked to energy through water regulation and in other ways.

The default mode network 69 comprises a set of brain regions that are coactivated during passive task states, demonstrates an intrinsic functional relationship, and is connected via direct and indirect anatomic projections. The medial temporal lobe and the medial prefrontal subsystems converge on important nodes of integration, including the posterior cingulate cortex. The general increase of NAA+NAAG, Glu and Glx values in default mode regions may indicate that the posterior cingulate cortex plays an important role in manipulating the functioning of the default mode network. Nonetheless, given that most of the modifications were “increased”, one could assume that the function of the default mode should also “increase”. We believe that the abnormality in our study reflects the metabolic sensitivity of the default mode regions. The exact mechanism of the effect of memantine in FM patients is poorly understood. Importantly, the default mode network (DMN), which plays a key role in attention, is hypoactive in Alzheimer's disease and is under glutamatergic control. Some studies have measured a positive effect of memantine treatment for 6 months for moderate to severe Alzheimer's disease by functional magnetic resonance imaging (fMRI), resulting in an increased resting DMN activity 70.

In this study, we observed an elevation in creatine and choline in the right posterior insula in patients treated with memantine compared with placebo. Creatine and creatine phosphate are measured as a single peak with ¹H MRS. Creatine phosphate serves as a reserve for high‐energy phosphates in the cytosol of muscle and neurons and buffers cellular ATP/ADP reservoirs. Tissues such as muscle and brain, where the largest changes in energy metabolism occur, have the highest concentrations of creatine kinase. Creatine is used as an internal reference value, as it is the most stable cerebral metabolite. Creatine has a role in the energetic system of the brain and in osmoregulation 43.

Choline and glycerophosphocholine (GPC) are the precursors for acetylcholine (ACho) and phosphatidylcholine (PtdCho). The synthesis of ACho occurs only within cholinergic neurons, whereas all cells use choline to synthesize PtdCho, which is a major constituent of the cell membrane 71. Choline is a marker of the phospholipid metabolism and cellular membrane turnover, reflecting cellular proliferation. Membrane synthesis, however, is an active process, suggesting that Cho levels may also reflect cell energetics 72. Another remarkable fact is the significant positive correlation between choline levels in the right posterior insula with FIQ scale scores. The posterior insula is known to play a prominent role in pain and interoceptive sensory processing 14. A previous study by our group confirms a significant reduction in choline in the FM and somatization disorder groups compared with controls 16.

Conclusion

Despite the limitations of the small sample size, it can be concluded from our study that memantine may induce some short‐term recovery of neuronal function in the posterior cingulate cortex and in the posterior insula of patients with chronic FM. On the basis of our results, the posterior cingulate cortex and posterior insula should be included in the areas to be explored in FM patients. This is a positive trial in which memantine demonstrated remarkable spectroscopic effects, and we conclude that memantine increases brain metabolism in patients with FM. Given the significant correlation found in the right posterior insula between metabolite values and clinical scales, MRS may be useful to monitor progression and response to treatment in FM.

Conflict of Interest

The authors declare no conflict of interest. Appropriate approval and procedures were used concerning human subjects.