Brain Advance Access originally published online on October 22, 2007
Brain 2007 130(12):3102-3110; doi:10.1093/brain/awm165
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Association between cortical metabolite levels and clinical manifestations of migrainous aura: an MR-spectroscopy study
1Stroke Prevention Research Unit, University Department of Clinical Neurology, Radcliffe Infirmary, Woodstock Road, Oxford OX2 6HE, 2MRC Biochemical and Clinical Magnetic Resonance Unit, Department of Biochemistry, University of Oxford, John Radcliffe Hospital, Headington,Oxford OX3 9DU and 3Department of Neurology, Northampton General Hospital NHS Trust, Cliftonville, Northampton NN1 5BD, UK
Correspondence to: Ursula Schulz, University Department of Clinical Neurology, Radcliffe Infirmary, Woodstock Road, Oxford OX2 6HE, UK E-mail: ursula.schulz{at}doctors.net.uk
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionConclusionReferences |
|---|
Previous studies suggest an abnormal cerebral cortical energy metabolism in migraineurs. If causally related to the pathophysiology of migraine, these abnormalities might show a dose–response relationship with the duration and severity of aura symptoms. While such a trend has been suggested in phosphorus spectroscopy (31P-MRS) studies, it has not been considered in proton spectroscopy (1H-MRS) studies and it has not been studied in cerebral white matter. We aimed to determine whether for any of the metabolites measured by 31P-MRS or 1H-MRS there was a dose-response relationship with aura duration and severity, and whether such an association was also present in cerebral white matter. We studied patients with migraine with aura and healthy controls with 31P-MRS and with 1H-MRS. We measured metabolite ratios in grey and in white matter and in the patients, we related metabolite levels to the clinical characteristics and duration of the aura. In patients, the phosphocreatine/phosphate (PCr/Pi) ratio decreased significantly with increasing aura duration and was significantly lower in patients with hemiplegic migraine than in patients with non-motor aura. Overall the metabolite ratios did not differ significantly between patients and controls, but compared with controls the PCr/Pi ratio in patients with hemiplegic migraine and in patients with persistent aura >7 days was significantly lower. These changes were only present in grey matter. Results for 1H-MRS did not differ significantly between patients and controls, and they showed no association with duration or severity of symptoms. In this study, metabolite ratios differed significantly between patients with different aura phenotypes and with increasing aura duration. In addition, only in some patient subgroups were metabolite ratios significantly different from controls. These findings support the concept that migraine with aura is a heterogeneous disorder with distinct pathophysiological subtypes. They further suggest that rather than determining the susceptibility to developing a migraine attack, changes in cortical energy metabolism may determine the clinical manifestations of the migrainous aura once an attack has started.
Key Words: MR-spectroscopy; migraine with aura; pathophysiology; migraine pathogenesis
Abbreviations: ATP, adenosine tri-phosphate; Cre, creatinine; Cho, choline; 1 DSI, one dimensional spectroscopic imaging; MRI, magnetic resonance imaging; NAA, N-acetyl aspartate; PCr/Pi, phosphocreatinine/phosphate; SNR, signal to noise ratio
Received March 8, 2007. Revised May 26, 2007. Accepted June 28, 2007.
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionConclusionReferences |
|---|
The pathophysiology of migrainous aura is incompletely understood. The currently most widely accepted explanation is the concept of the spreading depression of Leão (Leão, 1944
). This proposes that migrainous aura is caused by a short burst of intense neuronal activity, followed by a slowly propagating wave of neuronal and glial depolarization, which spreads across the cortex at a speed of 3–5 mm/min (Milner, 1958). The changes in neuronal activity are associated with a brief increase in cerebral blood-flow, followed by a longer lasting decrease in regional cerebral perfusion. Although various mechanisms have been proposed, the mechanisms causing the cortical depression are not known (Sanchez-del-Rio and Reuter, 2004). It has been suggested that changes in cortical excitability may increase the susceptibility to developing migrainous aura, and that the level of cortical excitability is associated with abnormalities in cerebral energy metabolism (Welch et al., 1990). Such changes have been reported by several phosphorus magnetic resonance spectroscopy (31P-MRS) studies, which found reduced energy reserves in the brains of migraineurs compared to controls (Welch et al., 1989; Montagna, 1995; Lodi et al., 2001), although more recent results have been conflicting (Boska et al., 2002). There have also been some small studies of proton spectroscopy (1H-MRS) in migraine patients, but these showed inconsistent results (Watanabe et al., 1996; Macri et al., 2003; Dichgans et al., 2005b; Sarchielli et al., 2005).
Recent studies suggest that migraine with aura is a condition consisting of distinct subtypes, and that the pathophysiological mechanisms underlying different migraine subtypes may differ (Eriksen et al., 2006). Our hypothesis was that if metabolic abnormalities are causally linked to the pathophysiology of migrainous aura, they might differ between different migraine subtypes, and they might show a dose-response relationship: the metabolic deficit should increase with increasing severity and/or duration of the migraine aura. Such trends have already been described for 31P-MRS, especially between migraine patients with and without aura (Montagna, 1995; Lodi et al., 2001; Boska et al., 2002). While two studies reported a trend for a higher metabolic deficit in patients with prolonged aura (Lodi et al., 2001) or with complicated migraine (Montagna, 1995) compared to patients with ‘migraine with aura’, these studies did not consider the association between the metabolic deficit, the duration and the clinical phenotype of the aura. Moreover, none of the 1H-studies considered symptom duration or severity. In addition, all previous MR-spectroscopy studies concentrated on metabolic abnormalities in cortical tissue, and did not consider potential white matter changes in detail. However, given that white matter lesions on MRI-scanning are more frequent in migraineurs than in controls (Swartz and Kern, 2004), one might also expect to find metabolic abnormalities in cerebral white matter. Finally, recent data suggest that not only the metabolites measured in 31P-MRS reflect mitochondrial function, but also the levels of N-acetyl-aspartate (NAA), which is measured with 1H-MRS (Clark, 1998). Changes in 31P-MRS may therefore be associated with concomitant changes in 1H-MRS. However, there are no studies of migraine patients with concurrent 31P-MRS and 1H-MRS.
We studied patients with a diagnosis of migraine with aura with 31P-MRS and with 1H-MRS to determine whether any metabolic deficit was related to duration or the clinical phenotype of the aura. We determined whether metabolic changes differed between grey matter and white matter, and whether there was any association between the findings of 31P-MRS and of 1H-MRS.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionConclusionReferences |
|---|
Patients
Patients with a history of migraine with aura who attended general neurology clinics in a teaching hospital and affiliated district general hospitals were selected for the study. Patients were eligible if the attacks fulfilled the international headache classification criteria (ICHD-2, Headache Classification Committee of the International Headache Society, 2004
) for migraine with aura, and if the type of symptoms and the duration of the attacks were similar on each occasion. Each patient underwent a detailed standardized interview. To enable us to study metabolite levels according to duration of symptoms, we recruited patients according to pre-defined categories of aura duration: 
1 h; >1–24 h; >1–7 days, >7 days. We chose this subdivision according to the ICHD-2, which defines ‘typical aura’ as including visual, sensory and/or speech disturbance, with each aura symptom lasting <1 h. The aura in hemiplegic migraine includes hemiplegia and any of the above symptoms, and each aura symptom may last up to 24 h. The ICHD-2 also categorizes ‘persistent aura without infarction’ as aura symptoms lasting for more than 7 days. Attacks with an aura duration between these cut-offs are not specifically categorized. In addition to the typical duration of the attacks, we also recorded the type of aura symptoms (non-motor aura versus hemiplegic migraine), the frequency of attacks and whether patients had a family history of migraine. We defined the symptomatic hemisphere as the hemisphere giving rise to the aura symptoms. Patients who consented to take part in the study and who had no contra-indications to magnetic resonance imaging (MRI) were scanned in the interictal period, at least 2 weeks after resolution of the most recent migraine attack. Control subjects were age and sex matched and healthy, with no history of migraine. The study was approved by the local ethics committee.
Imaging and spectroscopy
Imaging and multi-nuclear MRS were performed using a 2T whole-body Bruker Avance spectrometer (Bruker Biospin GmbH, Ettlingen, Germany). Images and proton spectra were acquired using a circularly polarized birdcage head coil, while phosphorus spectra were acquired using an 8 cm diametre circular surface coil. Patients were positioned in a purpose made head holder and a 2 cm diametre water-filled microsphere was placed adjacent to the temporal bone to allow for subsequent positioning of the phosphorus surface coil (see subsequently). A series of T2 weighted (Fast Spin Echo, TR = 3 s, TE = 80 ms) and T1 weighted (TR = 0.5 s, TE = 10 ms) images were collected to guide placement of spectroscopy voxels. A 3 x 3 x 2 cm3 proton spectroscopy voxel (PRESS) (Bottomley, 1987) was defined within the left hemisphere at the level of the basal ganglia, to include as much white matter as possible and carefully minimize contributions from CSF. The voxel was then shimmed to a line-width of <7 Hz. CHESS water suppression (Haase et al., 1985) was optimized within the voxel and metabolite spectra were collected (PRESS, TR = 1.5 s, TE = 135 ms, 128 averages, 2500 Hz spectral width, 1024 data points). A second voxel placed symmetrically in the right hemisphere was then examined using the same protocol. With this imaging protocol, we achieved the following signal to noise ratios (SNR) for the quantified peaks in the proton spectra: NAA
50, creatine30 and choline35.
In preparation for the phosphorus data collection, field homogeneity was adjusted over a 5 x 5 x 5 cm3 PRESS localized voxel, centred on each proton voxel and shim parameters stored. The patient bed was then withdrawn from the magnet while maintaining the patient position and the phosphorus surface coil was placed lateral to the left hemisphere, centred over the location of the proton spectroscopy voxel (using measurements from the images and the microsphere as the external landmark). Shim parameters were recalled for this hemisphere and Phosphorus spectra were acquired using a one dimensional spectroscopic imaging (1DSI) sequence (adiabatic 90° excitation, TR = 2.5 s, 2000 Hz spectral width, 512 data points and 32 encode steps) to encode spectra from a stack of 1-cm thick sagittal slices. Localization in the anterior–posterior and inferior–superior directions was limited by the surface coil receiver profile only. The patient bed was once again withdrawn and the surface coil relocated over the right hemisphere. Shim parameters for the right hemisphere were recalled and a second 1DSI phosphorus data set was collected. As in this imaging protocol the 31P signal was collected with a surface coil, SNR varied with depth. Typical SNR of phosphocreatine (PCr) in the grey matter and white matter regions was 30 and 15–20, respectively.
Data processing
Phosphorus 1DSI data sets were Fourier transformed in the spatial dimension only to produce pseudo-FIDs corresponding to each spatial location. These FIDs were then processed using time-domain fitting software (MRUI) (van den Boogaart et al., 1996) with the AMARES algorithm (Vanhamme et al., 1997). Peak integrals for PCr, adenosine tri-phosphate (ATP) and inorganic phosphate (Pi) were determined. Tissue pH was calculated from the difference in chemical shift of PCr and Pi. (Moon and Richards, 1973) The spatial localization of each spectrum in the set was defined according to the MR Images and spectra were assigned to originate predominantly from white matter, grey matter and mixed tissue. A typical 31P-spectrum and placement of the stacked 1-cm slices is shown in Fig. 1A.
|
Proton spectra were analysed using WIN-NMR software (Bruker Biospin GmbH, Germany). Free induction decays were Lorentz-Gauss converted, Fourier transformed, zero-order phase-corrected and baseline corrected between 1 and 4 ppm using a fifth order automatic polynomial function. Gaussian lineshapes were then fitted to the peaks from N-acetyl-aspartate (NAA), creatine (Cre) and choline (Cho) and peak integrals determined. A typical 1H-spectrum and the placement of the 1H-MRS voxel is shown in Fig. 1B.
Statistical analysis
Patients were categorized according to type of symptoms (migraine with non-motor aura—including visual and sensory symptoms and dysphasia; hemiplegic migraine—hemiparesis in addition to any of the above symptoms) and to symptom duration (1 h, >1–24 h, >24 h–7 days, >7 days). In 31P-MRS, we determined the ratios of PCr/Pi, PCr/ATP, Pi/ATP and the pH in grey matter, in white matter and in the total hemisphere (white matter + grey matter + mixed tissue) for each patient. We used analysis of variance (ANOVA) to determine whether the mean ratios differed between patients with different aura symptoms and with different aura duration, and analysed whether there were any systematic changes with increasing symptom duration. To determine the consistency of our findings, we performed the analyses for the symptomatic and for the contralateral hemisphere.
In 1H-MRS, we determined the ratios of NAA/Cho, NAA/Cre and Cho/Cre. We also assessed whether a lactate peak was present. We compared mean ratios between patients and controls and related the ratios to symptom type and duration as described earlier. We correlated the results of 1H-MRS and 31P-MRS to determine whether there was any association between the findings obtained with both methods. All statistical analyses were done with SPSS Version 12.01®.
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionConclusionReferences |
|---|
Twenty-one patients [nine men, mean (SD) age 42 (11) years] and 16 age- and sex-matched healthy controls with no history of migraine [eight men, mean (SD) age 39 (15) years] were recruited. The clinical details for the patients are shown in Table 1. Findings on neurological examination and structural imaging were normal in all study subjects, and at the time of scanning no patient was on any migraine prophylactic treatment. Ten patients had migraine with non-motor aura. All 10 experienced visual symptoms. Three also had sensory symptoms and two had dysphasia. Eleven patients typically experienced hemiplegic migraine. Of these, four also had sensory symptoms and two had had episodes of migrainous coma. The typical duration of the aura was
1 h in six patients, between 1 and 24 h in two patients, from >24 h to 7 days in four patients and >7 days in nine patients. There was a strong association between symptom type and duration of aura, in that all patients with hemiplegic migraine had an aura duration of >24 h, whereas in 6 of the 10 patients with non-motor aura the usual aura duration was 1 h. All patients had unilateral symptoms which allowed classification of the symptomatic hemisphere.
|
Phosphorus spectroscopy 31P-MRS
All 21 patients underwent 31P-MRS. In 15 patients, both hemispheres were scanned, whereas the first six patients only underwent scanning of the symptomatic hemisphere. Table 2 shows the results for 31P-MRS in relation to the phenotype of the aura and in relation to symptom duration for the symptomatic hemisphere. The PCr/Pi ratio was significantly lower in patients with hemiplegic migraine compared to patients with non-motor aura, and it decreased with increasing symptom duration. These changes were highly significant in the grey matter, but not present in the white matter. The individual ratios of PCr/ATP and Pi/ATP suggest that the reduction in the PCr/Pi ratio was mainly due to an increase in Pi in patients with hemiplegic migraine and with increasing symptom duration. To investigate whether phenotype and duration were independently associated with metabolite ratios, we studied the association between metabolite levels and aura duration separately in patients with non-motor aura and in patients with hemiplegic migraine. Of the 10 patients with non-motor aura, six had a typical aura duration of
1 h, two patients had a typical aura duration of >1 to 24 h, and one patient each of >24 to 7 days and of >7 days. Despite patient numbers in each category being very small, there was significant heterogeneity between the groups (Phet = 0.03) with a clear trend for the PCr/Pi ratio to decrease with increasing aura duration (Plin = 0.008). In all 11 patients with hemiplegic migraine the typical aura duration was >1 day. However, there was still a significant difference in the cortical PCr/Pi ratio between the three patients with an aura duration of 7 days [mean (SD) = 1.89 (0.22)] and the eight patients with an aura duration of >7 days [1.59 (0.14), P = 0.023]. We also tried to determine whether the association between aura phenotype and metabolite ratios was independent of aura duration. However, as there was very little overlap in duration between the two aura phenotypes, it was not possible to perform a meaningful analysis.
|
Table 2 shows the metabolite ratios which we obtained in controls. We found no significant difference between controls and the cohort of migraine patients overall. However, when we compared each individual patient subgroup with the controls [mean (SD) PCr/Pi ratio = 1.99 (0.33)], we found that the PCr/Pi ratio in the cortex of the symptomatic hemisphere was significantly lower in patients with with aura duration >7 days [1.59 (0.13); P = 0.002], and in patients with hemiplegic migraine [1.68 (0.21), P = 0.01]. There were no significant differences in metabolite ratios between any of the other patient groups and controls. Metabolite ratios did not differ significantly between patients with and without a family history of migraine, and we found no association between the frequency of migraine attacks and metabolite ratios. There was some inconsistent heterogeneity in pH between the different categories of aura duration, but the pH showed no consistent change with increasing aura duration, and it did not vary between aura phenotypes. All of the above results apply to measurements obtained in the symptomatic hemisphere (Table 2). The results obtained in the asymptomatic hemisphere (Table 3) confirmed that the PCr/Pi ratio was lower in patients with hemiplegic migraine compared to patients with non-motor aura, and that it decreased with increasing aura duration. However, the results did not always achieve statistical significance.
|
A finding which was equally present in patients and in controls was that the metabolite ratios differed significantly between grey and white matter: the PCr/Pi ratio was higher in the white matter than in the grey matter [mean (SD) 2.32 (0.47) versus 1.80 (0.28), P < 0.0001], whereas Pi/ATP and pH were both higher in the grey matter [0.63 (0.13) versus 0.50 (0.13), P < 0.0001 and 7.10 (0.05) versus 7.05 (0.05), P < 0.0001]. The PCr/ATP ratio did not differ significantly between grey and white matter.
Proton spectroscopy (1H-MRS)
Thirteen patients and 13 controls underwent 1H-MRS in addition to 31P-MRS. In all of these, both hemispheres were scanned. The results for 1H-spectroscopy are shown in Table 4. We found no association between any of the metabolite ratios (NAA/Cre, Cho/Cre and NAA/Cho) and type or duration of aura symptoms. We found no lactate peaks in any of the subjects. To see whether there was an association between a reduced PCr/Pi ratio and a reduction in NAA, we correlated PCr/Pi with NAA/Cre and NAA/Cho in both the symptomatic and in the contralateral hemisphere. We did not find any association between these variables.
|
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionConclusionReferences |
|---|
In this study, we found that patients with hemiplegic migraine had significantly lower PCr/Pi ratios than patients with non-motor aura. We also found that the PCr/Pi ratio decreased with increasing aura duration. These changes were only present in cortical tissue. Concomitant 1H-MRS showed no abnormalities in migraine patients, and it showed no association with the results of 31P-MRS.
Phosphorus spectroscopy (31P-MRS)
Two previous groups have studied migraine patients with 31P-MRS (Welch et al., 1989; Barbiroli et al., 1992). The Bologna group (Barbiroli et al., 1992; Montagna, 1995; Lodi et al., 2001) found that compared with controls energy reserves were significantly reduced in all migraine patients, regardless of whether they had migraine with or without aura. The authors found a trend for patients with ‘complicated migraine’ (Montagna, 1995) or patients with ‘prolonged aura’ (Lodi et al., 2001) to have a more severe metabolic deficit than patients with ‘aura’. In these studies, patients with hemiplegic migraine were not considered as a separate group, although a further study by the Bologna group found abnormal brain energy metabolism in a family with hemiplegic migraine (Uncini et al., 1995). In contrast to the Bologna group, the Detroit group more recently found no differences in energy reserves between migraine patients and controls. However, they found a non-significant trend for the PCr/Pi ratio to be lower in patients with migraine with aura than in migraine without aura. This difference was most marked in hemiplegic migraine (Boska et al., 2002). Our aim was to compare the energy metabolism between different subtypes of migraine with aura in detail. Like the Detroit group, we also found no overall difference between patients and controls. However, in patients with hemiplegic migraine and in patients with persistent aura >7 days, the PCr/Pi ratio was significantly lower than in controls. Among migraine patients, hemiplegic migraine and increasing aura duration was associated with a decreasing PCr/Pi ratio. Although the currently available 31P-MRS studies do not show consistent differences between migraineurs and controls, they do indicate an altered energy metabolism in patients with migraine with aura. The findings of our study further suggest that, because not all patients differed from controls, rather than influencing the susceptibility to developing a migraine attack, changes in energy metabolism may determine the clinical manifestation of an attack once it has started.
The ICHD-2 classifies migraine with aura and hemiplegic migraine as two separate entities. In addition to differences in the clinical presentation and the presence of genetic variation, our finding of lower PCr/Pi ratios in patients with hemiplegic migraine further supports the concept of different pathomechanisms underlying these two migraine subtypes. The identification of specific mutated genes in familial hemiplegic migraine has pointed to disturbances in ion transport mechanisms playing a crucial role in the pathogenesis of migraine aura (Goadsby, 2007). In familial hemiplegic migraine, three mutated genes have so far been identified (Ophoff et al., 1996; de Fusco et al., 2003; Dichgans et al., 2005a). These affect ion homoeostasis, eventually leading to cortical hyperexcitability by increasing intracellular calcium influx and synaptic glutamate levels (Sanchez-del-Rio and Reuter, 2004; Goadsby, 2007). Such changes in ion homoeostasis and neuronal excitability may result in increased cellular energy demands and reduced energy reserves. In non-hemiplegic migraine, no specific mutations have yet been identified, which suggests different pathomechanisms with different effects on neuronal energy metabolism in this migraine subtype.
Our study showed a close association between increasing aura duration and decreasing energy reserves. The duration of the aura reflects the time it takes neurones to regain function after having been affected by the cortical spreading depression. Lower energy reserves may prolong this recovery time. In keeping with previous studies (Thomsen et al., 2003; Kirchmann, 2006), we also found a close association between aura duration and aura phenotype. One possible explanation is that an underlying energy deficit determines both the number of neurones affected by the spreading depression and the extent of dysfunction in individual neurones after they have been affected by the spreading depression. Alternatively, a common underlying aetiology, for example changes in ion homoeostasis, may affect the aura phenotype and aura duration by different mechanisms. The aura phenotype is determined by the area of cortex affected by the spreading depression. Changes in ion homoeostasis may increase cortical excitability, leading to more extensive cortical depression and causing a more severe clinical deficit. Altered ion homoeostasis may also increase cellular energy demands and lead to a decrease in energy reserves, prolonging the time required by neurones to recover from the effects of the cortical spreading depression. The close association of aura phenotype and aura duration may be explained by a higher level of neuronal excitability also leading to an increase in cellular energy demands.
Migraine patients have white matter lesions on MRI scanning more frequently than controls, (Swartz and Kern, 2004) which raises the possibility that metabolic abnormalities may also involve the white matter. Furthermore, metabolic abnormalities have been found in muscle tissue in migraineurs, (Barbiroli et al., 1992) suggesting a generalized metabolic deficit which should also involve cerebral white matter. However, in this study we did not find any reduction in energy reserves in the white matter. This suggests that white matter lesions in migraine patients are not related to an underlying energy deficit. If the deficit in energy metabolism is generalized, it may nevertheless be restricted to metabolically very active tissues, such as muscle or neurones.
Proton spectroscopy (1H-MRS)
We are only aware of four other small studies which used 1H-MRS in migraine (Watanabe et al., 1996; Macri et al., 2003; Dichgans et al., 2005b; Sarchielli et al., 2005) and which have reported conflicting results: one study reported an increase in lactate in the interictal period, (Watanabe et al., 1996) and one study showed a reduction in choline levels in the interictal period, (Macri et al., 2003) but these results were not reproduced by the other studies. A reduction in NAA was reported by two studies (Dichgans et al., 2005b; Sarchielli et al., 2005), whereas the other two studies found no changes in NAA (Watanabe et al., 1996; Macri et al., 2003). Two studies had placed their areas of interest into the cerebellum (Macri et al., 2003; Dichgans et al., 2005b), whereas the other two studies (Watanabe et al., 1996; Sarchielli et al., 2005) examined occipital lobe areas. All studies had placed the proton voxels into grey matter. Our study differed from the previous ones in that we studied 1H-MRS in cerebral white matter. We were mainly interested in the NAA-levels, because while NAA has generally been regarded as a marker of neuronal integrity, it has more recently also been shown to be present in oligodendrocytes, and its main role is believed to be in lipid and myelin formation. (Clark, 1998) Changes in NAA levels may therefore also reflect damage in cerebral white matter (Garnett et al., 2000). Since migraine patients have an increased prevalence of white matter lesions compared to controls (Swartz and Kern, 2004), one might expect NAA levels in migraine patients to be lower than in controls. However, we were unable to detect such a difference.
In addition to its role in myelination, it has also been suggested that NAA may be a marker of mitochondrial function, because it is synthesized and located in neuronal mitochondria (Clark, 1998). One might therefore expect a positive association between PCr/Pi and NAA levels. However, we did not find such an association. This may be due to the way we placed the proton voxel—it was located in the subcortical white matter, where we did not find any deficit in mitochondrial function in 31P-MRS. Furthermore, even if NAA is related to mitochondrial function, this may involve metabolic pathways different from the ones measured with 31P-MRS.
Potential limitations of the study
We feel that our study provides important new insights into the pathophysiology of migraine with aura. However, it also has some potential limitations. First, the variability in our data reflects a combination or true variation between study subjects and experimental variation, which mainly arises from the SNR of the acquired spectra. It is possible that the SNR, in particular in 31P-MRS with the inherent lower sensitivity of the 31P-nucleus, did not allow us to detect significant differences between controls and some patient sub-groups. However, the SNRs in our study compare favourably with those of other studies, and the variability in our data is similar to the other published studies in migraine. We therefore feel that the data quality in our study was as high as was technically possible and that it is unlikely to be the cause of any differences between our findings and those of others. Second, the areas of brain we investigated differed from those in some of the previous studies, which to a large majority had focused on the occipital cortex. It is possible that differences in the investigated brain areas may account for differences in the results between our and previous studies. However, according to previous studies the changes in energy metabolism in migraineurs affect the entire body and, for example, are also present in muscle tissue (Barbiroli et al., 1992). Any metabolic abnormalities should therefore be present in the entire cortex, and it is unlikely that voxel location should have influenced our results. Furthermore, one of our aims was to study metabolic changes in the white matter, and the voxel and slice locations we chose allowed us to do this within the approved time frame for scanning. Finally, it is possible that differences in patient characteristics between ours and other studies may account for some of the differences in study results. This is particularly true for several of the 1H-MRS studies, which included mixed populations of migraine patients, whereas our population consisted solely of patients with migraine with aura. Other factors, such as time since last migraine attack, attack frequency or duration may also account for differences in study results. These details were not recorded in all of the previous studies, and it is therefore not possible to determine the extent to which they may have affected the results.
|
Conclusion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionConclusionReferences |
|---|
In this study, abnormal findings were restricted to 31P-MRS of cerebral grey matter. The PCr/Pi ratio was lower in patients with hemiplegic migraine than in patients with non-motor aura, and it decreased with increasing aura duration. Compared to controls, energy metabolism was abnormal in patients with hemiplegic migraine and in patients with persistent aura >7 days, but not in the other patients. Our findings support the concept of migraine with aura being a heterogeneous disorder with distinct pathophysiological subtypes. Disturbances in energy metabolism may not determine the susceptibility to developing a migraine attack, but they may determine the clinical manifestations of the aura once an attack has started.
|
Acknowledgements |
|---|
The study was funded by the UK Medical Research Council, the Wellcome Trust and by MSD Pharma, who provided an educational grant which was used to pay a part of the salary for Dr U.G.S. in her first year as a Clinical Research Fellow.
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionConclusionReferences |
|---|
Barbiroli B, Montagna P, Cortelli P, Funicello R, Iotti S, Monari L, et al. Abnormal brain and muscle energy metabolism shown by 31P magnetic resonance spectroscopy in patients affected by migraine with aura. Neurology (1992) 42:1209–14.
Boska MD, Welch KMA, Barker PB, Nelson JA, Schultz L. Contrasts in cortical magnesium, phospholipid and energy metabolism between migraine syndromes. Neurology (2002) 58:1227–33.
Bottomley PA. Spatial localization in NMR spectroscopy in vivo. Ann NY Acad Sci (1987) 508:333–48.[Web of Science][Medline]
Clark JB. N-acetyl-aspartate: a marker for neuronal loss or mitochondrial dysfunction. Dev Neurosci (1998) 20:271–6.[CrossRef][Web of Science][Medline]
De Fusco M, Marconi R, Silvestri L, Atorino L, Rampoldi L, Morgante L, et al. Haploinsufficiency of ATP1A2 encoding the Na+/K+ pump a2 subunit associated with familial hemiplegic migraine type 2. Nat Genet (2003) 33:192–6.[CrossRef][Web of Science][Medline]
Dichgans M, Freilinger T, Eckstein G, Babini E, Lorenz-Depiereux B, Biskup S, et al. Mutation in the neuronal voltage-gated sodium channel SCN1A causes familial hemiplegic migraine. Lancet (2005a) 366:371–7.[CrossRef][Web of Science][Medline]
Dichgans M, Herzog J, Freilinger T, Wilke M, Auer DP. 1H-MRS alterations in the cerebellum of patients with familial hemiplegic migraine type 1. Neurology (2005b) 64:608–13.
Eriksen MK, Thomsen LL, Olesen J. Implications of clinical subtypes of migraine with aura. Headache (2006) 46:286–97.[CrossRef][Web of Science][Medline]
Garnett MR, Blamire AM, Rajagopalan B, Styles P, Cadoux-Hudson TAD. Evidence for cellular damage in normal-appearing white matter correlates with injury severity in patients following traumatic brain injury. A magnetic resonance spectroscopy study. Brain (2000) 123:1403–9.
Goadsby PJ. Recent advances in understanding migraine mechanisms, molecules and therapeutics. Trends Mol Med (2007) 13:39–44.[CrossRef][Web of Science][Medline]
Haase A, Frahm J, Hanicke W, Matthaei D. 1H NMR chemical shift selective (CHESS) imaging. Phys Med Biol (1985) 30:341–4.[CrossRef][Web of Science][Medline]
Headache Classification Committee of the International Headache Society. The International Classification of Headache Disorders. Cephalalgia (2004) 24:1–160.[CrossRef][Web of Science]
Kirchmann M. Migraine with aura: new understanding from clinical epidemiologic studies. Curr Opin Neurol (2006) 19:286–93.[Web of Science][Medline]
Leão AAP. Spreading depression of activity in cerebral cortex. J Neurophysiol (1944) 7:379–90.
Lodi R, Iotti S, Cortelli P, Pierangeli G, Cevoli S, Clementi V, et al. Deficient energy metabolism is associated with low free magnesium in the brains of patients with migraine and cluster headache. Brain Res Bull (2001) 54:437–41.[CrossRef][Web of Science][Medline]
Macri MA, Garreffa G, Giove F, Ambrosini A, Guardati M, Pierelli F, et al. Cerebellar metabolite alterations detected in vivo by proton MR spectroscopy. Magn Reson Imaging (2003) 21:1201–6.[CrossRef][Web of Science][Medline]
Milner PM. Note on a possible correspondence between the scotomas of migraine and spreading depression of Leão. Electroencephalogr Clin Neurophysiol (1958) 10:705.[CrossRef][Medline]
Montagna P. Magnetic resonance spectroscopy in migraine. Cephalalgia (1995) 15:323–7.[CrossRef][Web of Science][Medline]
Moon RB, Richards JH. Determination of intracellular pH by 31P NMR. J Biol Chem (1973) 248:7276–8.
Ophoff RA, Terwindt GM, Vergouwe MN, van Eijk R, Oefner PJ, Hoffman SM, et al. Familial hemiplegic migraine and episodic ataxia type-2 are caused by mutations in the Ca2+ channel gene CACNL1A4. Cell (1996) 87:543–52.[CrossRef][Web of Science][Medline]
Sanchez-del-Rio M, Reuter U. Migraine aura: new information on underlying mechanisms. Curr Opin Neurol (2004) 17:289–93.[CrossRef][Web of Science][Medline]
Sarchielli P, Tarducci R, Presciutti O, Gobbi G, Pelliccioli GP, Stipa G, et al. Functional 1H-MRS findings in migraine patients with and without aura assessed interictally. Neuroimage (2005) 24:1025–31.[CrossRef][Web of Science][Medline]
Swartz RH, Kern RZ. Migraine is associated with magnetic resonance imaging white matter abnormalities: a meta-analysis. Arch Neurol (2004) 61:1366–8.
Thomsen LL, Ostergaard E, Jes Olesen J, Russell MB. Evidence for a separate type of migraine with aura: sporadic hemiplegic migraine. Neurology (2003) 60:595–601.
Uncini A, Lodi R, Di Muzio A, Silvestri G, Servidei S, Lugaresi A, et al. Abnormal brain and muscle energy metabolism shown by 31P-MRS in familial hemiplegic migraine. J Neurol Sci (1995) 129:214–22.[CrossRef][Web of Science][Medline]
van den Boogaart A, Van Hecke P, Van Huffel S, Graveron-Demilly D, van Ormondt D, de Beer R. MRUI: a graphical user interface for accurate routine MRS data analysis. (1996) In: Proceedings of the ESMRMB 13th Annual Meeting: Prague. 318.
Vanhamme L, van den Boogaart A, Van Huffel S. Improved method for accurate and efficient quantification of MRS data with use of prior knowledge. J Magn Reson (1997) 129:35–43.[CrossRef][Web of Science][Medline]
Watanabe H, Kuwabara T, Ohkubo M, Tsuji S, Yuasa T. Elevation of cerebral lactate detected by localized 1H-magnetic resonance spectroscopy in migraine during the interictal period. Neurology (1996) 47:1093–5.
Welch KMA, Levine SR, D’Andrea G, Schultz LR, Helpern JA. Preliminary observations on brain energy metabolism in migraine studied by in vivo phophorus 31 NMR spectroscopy. Neurology (1989) 39:538–41.
Welch KMA, Tepley N, Barkley G, Ramadan NM. The concept of migraine as a state of central neuronal hyperexcitability. Neurol Clin (1990) 8:817–28.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  U. G. Schulz, A. M. Blamire, P. Davies, P. Styles, and P. M. Rothwell Normal Cortical Energy Metabolism in Migrainous Stroke: A 31P-MR Spectroscopy Study Stroke, December 1, 2009; 40(12): 3740 - 3744. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||