| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Radiology (E.R.D., C.T.), Memory Disorders Research Group (T.V.E., J.D., G.W., M.H.), Medical Endocrinology Rigshospitalet (Å.K.R., U.F.-R.), and Department of Endocrinology Bispebjerg Hospital (H.P.), University of Copenhagen, DK-2100 Copenhagen, Denmark
Address all correspondence and requests for reprints to: E. R. Danielsen, Ph.D., Department of Radiology, Rigshospitalet, 9 Blegdamsvej, DK-2100 Copenhagen, Denmark. E-mail: else.rubaek.danielsen{at}rh.regionh.dk.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Objective: Our objective was to evaluate cerebral biochemistry in acute and treated Graves’ disease.
Design: We conducted a prospective study, investigating volunteers once and patients before and 1 yr after treatment.
Setting: The study was performed at a radiology department, a memory disorder clinic, and two endocrinology clinics.
Patients and Other Participants: Of 53 consecutively referred, newly diagnosed, and untreated patients with Graves’ thyrotoxicosis, 27 patients (34 ± 8 yr) and 33 matched volunteers were included.
Interventions: Patients were treated with thionamide.
Main Outcome Measures: We assessed brain metabolite concentrations.
Methods: Proton magnetic resonance spectroscopy of the brain and a battery of biochemical, affective, and cognitive tests were used.
Results: Previously reported findings of reduced choline and myo-inositol in acute Graves’ disease were confirmed and reversibility was demonstrated. Parieto-occipital white matter glutamine was and remained significantly reduced (P < 0.01). Acute phase parieto-occipital white matter total choline correlated significantly (r = –0.57; P < 0.01) with impaired thyroid function. Pretreatment total T3 predicted posttreatment occipital gray matter glutamine (r = –0.52; P < 0.01). Occipital gray matter total choline (r = –0.53; P < 0.01) and parietooccipital white matter glutamate (r = –0.54; P < 0.01) correlated with initial values of selected attention and concentration cognitive scores and predicted them at follow-up.
Conclusions: The persistent reduction of glutamine in white matter, the decreasing glutamate in occipital gray matter, and the correlation with severity of the initial disease as well as with attention and concentration cognitive scores indicated that there was a persistent and possibly progressive disturbance of the glutamate glutamine cycling in Graves’ disease.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The diagnosis of Graves’ disease was based on increased peripheral thyroid hormones, suppressed TSH, positive TSH-receptor antibodies, and diffuse uptake on a Tc-99m pertechnetate scintigraphy of the thyroid gland. Patients with a history of previous thyroid disease or psychiatric disorders were excluded from the study. Likewise, patients with a history of previous neurological disorders known to influence neuropsychiatric functions were excluded from the study. Patients unable to read Danish were also excluded. Twenty-seven of these patients (26 females, one male; age 34 ± 8 yr, mean ± SD) were included in the study. The causes of exclusion were previous psychiatric disease (n = 5), ongoing drug abuse (n = 1), foreign primary language (n = 6), cerebral metastases (n = 1), multiple sclerosis (n = 1), familial relation to the investigator (n = 1), vacation of the investigator (n = 1), no informed consent (n = 5), claustrophobia (n = 2), and temporary closure of the MRS unit at the time of enrollment (n = 2), and in one patient, Graves’ disease was never confirmed because the patient never showed up for further investigation (n = 1). The patients received methimazole therapy, but three were changed to propylthiouracil due to adverse effects (itching). Twenty-five patients with normalized thyroid hormones completed the follow-up examination 379 ± 24 d (mean ± SD) after the initial examination, and the two remaining were excluded due to pregnancy; all patients were still treated with antithyroid medication at this time. For comparison, 33 age-, sex-, and education-matched healthy volunteers were included, recruited from advertising. Similar exclusion criteria were applied to the group of healthy volunteers with addition of exclusion upon family history of thyroid disease. The demographic characteristics, the thyroid status, and affective and cognitive data for the subjects have previously been reported in Vogel et al. (4) from a slightly larger group of subjects because four patient and one volunteer who could not have the MR examination due to claustrophobia or temporary closure of the MR scanner were included in Vogel et al.’s material. The new data in this manuscript are confined to MRS data and its relations to thyroid status and affective and cognitive measures.
Serum concentrations of TSH, total T4, free T4, total T3, and TSH-receptor antibodies were measured. Other laboratory blood tests included liver function and electrolytes, for the purpose of excluding systemic diseases known to affect the concentration of some metabolites measured in the brain (10). The laboratory variables were measured by standard methods. Affective rating scales included the Hamilton Depression Rating Scale (11, 12), the Hamilton Anxiety Scale (13, 14), and the Bech-Rafaelsen Mania Scale (15). The neuropsychological test batteries to evaluated cognitive parameters were described in detail previously (4).
MRS examinations were performed before and 1 yr after treatment initiation of the Graves’ disease patients. Standard T1- and T2-weighted MR images were acquired from all participants to exclude cerebral pathology. Positioning and repositioning of the volumes of interest were based on the MR images. Short echo time spectra (6) were acquired from three volumes of interest: a parietooccipital location primarily containing white matter (poWM), a midoccipital location primarily containing gray matter (oGM), and a midfrontal location primarily containing gray matter (fGM). Metabolite quantification involved fully automated and user-independent spectral evaluation by LCMODEL (16). Objective criteria for rejecting MR spectra of poor technical quality were based on objective output parameters from the LCMODEL analysis: sufficient spectral resolution (full width at half-minimum 
0.07 ppm) and sufficient information (signal to noise ratio 
7). The concentrations of N-acetylaspartate + N-acetylaspartylglutamate, myo-inositol, total creatine, total choline, and glutamate + glutamine were calculated as described in detail in Elberling et al. (6). In addition, the present analysis included individual concentrations of glutamine and glutamate for the purpose of group comparisons.
The desired sample sizes were chosen to 30 to get the statistical power 80% for detecting differences between patients and controls if the concentrations deviated by more than 1 SD from each other, based on the first 18 volunteers (6), which corresponded to 7% (N-acetylaspartate + N-acetylaspartylglutamate), 13% (myo-inositol), 7.0% (total creatine), 10% (total choline), and 15% (glutamate + glutamine) detection limits.
Schirmer and Auer (17) have shown that the intra-individual coefficients of variance for the concentrations measured by MRS are about half compared with the interindividual coefficients of variance, the day-to-day variation of MRS is thus expected to be between 3% (N-acetylaspartate + N-acetylaspartylglutamate) and 8% (glutamate + glutamine).
Statistical analyses were done using paired two-tailed Student’s t test for comparing results before and after treatment and using unpaired two-tailed Student’s t tests, equal variance not assumed, for comparing patients with healthy volunteers.
Correlation analyses were performed to test whether the initial thyroid status (total T4, free T4, total T3, and TSH-receptor antibodies) predicted the MRS findings in the acute phase and/or at follow-up and whether the initial MRS findings predicted affective rating scales and/or cognitive rating scales during the acute phase or at follow-up. The affective scales Hamilton Depression Rating Scale, Hamilton Anxiety Scale, and Bech-Rafaelsen Mania Scale were used. A subset of the cognitive rating scales representing three of the main groups attention and concentration (Median sentence repetition, Digit span forwards, Digit span reverse orders, Stroop test congruent, Stroop test incongruent), psychomotor speed and executive functions (Trail making A, Trail making B, and symbol digit modalities test), and memory and recognition (Forced choice face recognition and Forced choice word recognition) were used in the correlation analyses.
The MRS results from fGM were excluded from the correlation analyses due to the low number of spectra that passed the objective spectral quality test describer above. Spearman’s rank order correlation (SPSS Inc., Chicago, IL) was used.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
.
|
). The quality of the spectra between the groups was not significantly different as judged by full width of half-minimum and by the signal to noise ratio.
|
Total choline was consistently reduced in acute Graves’ disease compared with healthy volunteers in all three volumes of interest (P < 0.00001 in poWM; P < 0.000001 in oGM; P < 0.01 in fGM). Myo-inositol was reduced in acute Graves’ disease compared with healthy volunteers in all three volumes of interest as well but significantly so only in oGM and poWM, (P < 0.0001 in poWM; P < 0.05 in oGM; P < 0.075 in fGM). These abnormalities were reversible because both myo-inositol and total choline were in the normal range at follow-up. The paired analysis of MRS in the acute phase and at follow-up showed that both myo-inositol and total choline changed statistically significantly in all three volumes of interest (Table 2
).
Total T3 correlated with total choline in poWM of untreated Graves’ disease patients (r = –0.57; P < 0.01). Less significant correlations were found for free T4 and TSH-receptor antibodies (see Table 3). After treatment, neither total choline nor myo-inositol correlated with any of the thyroid parameters in the acute phase.
|
). Initial total choline in oGM showed a weak correlation with Median sentence repetition (r = –0.38; P < 0.05) and predicted that this persisted at follow-up (r = –0.53; P < 0.01) (see Table 5). Initial total choline in poWM predicted Hamilton Depression Rating scale and Digit span forwards at follow-up, but the correlations were weak (see Tables 4 and 5). Initial myo-inositol in poWM correlated with Forced choice face recognition at follow-up (r = 0.42; P < 0.05).
|
|
The combined glutamate + glutamine was significantly reduced both in the acute phase of Graves’ disease and at follow-up (P < 0.01), but only in poWM. Individual analyses of glutamine and glutamate revealed that glutamine was reduced in poWM, significantly, however, only at follow-up (P < 0.01 follow-up vs. healthy volunteer; P < 0.06 acute Graves’ disease vs. healthy volunteer). Individual analyses of glutamine and glutamate furthermore showed that glutamate decreased in Graves’ disease from the acute phase to the follow-up in oGM (P < 0.05).
The initial thyroid function parameters did not correlate with the corresponding measures of glutamate and glutamine, but initial total T4 (r = –0.42; P < 0.05) and total T3 (r = –0.52; P < 0.01) correlated with glutamine in oGM of the treated patients (see Table 3).
None of the glutamate and glutamine measures correlated with any of the affective scales (see Table 4), but correlations were found for three of the cognitive tests. Initial glutamate + glutamine in oGM correlated with initial Trail making A (r = –0.40; P < 0.05). Initial glutamate + glutamine in poWM correlated with the initial Forced choice word recognition (r = –0.45; P < 0.05). Initial glutamate in poWM correlated with the initial (r = –0.54; P < 0.01) and predicted the follow-up (r = –0.48; P < 0.05) Stroop test incongruent, the glutamate + glutamine showed similar correlations (see details in Table 5).
N-Acetylaspartate + N-acetylaspartylglutamate and total creatine
There was a small but significant reductions of N-acetylaspartate + N-acetylaspartylglutamate and total creatine in poWM (P < 0.05). These abnormalities were reversible, but only total creatine changed significantly in the paired analysis. Finally, opposite N-acetylaspartate + N-acetylaspartylglutamate changes were found in fGM and oGM (P < 0.05).
No laboratory tests correlated with the initial measures of N-acetylaspartate + N-acetylaspartylglutamate or total creatine.
Weak correlations were found for initial N-acetylaspartate + N-acetylaspartylglutamate or total creatine and Hamilton Depression Rating Scale and Hamilton Anxiety Scale at follow-up as well as initial Forced choice face recognition (see details in Tables 4 and 5).
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Total choline measured by MRS represents primarily phosphocholine and glycerophosphocholine. Elevated total choline is mainly associated with membrane breakdown as seen in brain injuries, with tumor growth or hypernatremia. Decreased total choline is found in patients with hyponatremia or liver disease (10).
Myo-inositol is almost exclusively located in astrocytes, where it is believed to be the most important osmolyte. Myo-inositol is elevated in conditions where gliosis has occurred and in hypernatremia. Myo-inositol is decreased in hyponatremia and liver disease.
This study confirmed the previously reported MRS abnormalities in Graves’ disease and showed that the reduced myo-inositol and reduced total choline concentrations were reversible after successful antithyroid treatment. The mechanisms of these reversible biochemical abnormalities may be disputed, although several explanations have been given previously (5, 6). Myo-inositol and total choline are known to be reduced in hyponatremia because myo-inositol and total choline act as organic osmolytes, and their intracellular concentrations, as measured by MRS, are reduced to counterbalance the extracellular electrolyte reduction (18). Myo-inositol and total choline are also known to be reduced in hepatic liver disease (19), where one hypothesis is that the abnormalities are also due to an osmolyte effect. Both these conditions were excluded in the examined subjects, but a hypothesis of a related mechanism might explain both abnormalities. A possible explanation could thus be that organic osmolytes are shifted, because the intracellular electrolyte composition is altered in Graves’ disease due to abnormal transport through the cell membrane. Thyroid hormones are known to modulate electrolyte transport activities in rat renal and intestinal brush border membranes (20, 21, 22), but whether similar mechanisms are present in the brain is an open question. Another explanation of the reduced myo-inositol and total choline measured in Graves’ disease could be a reduced nuclear magnetic resonance visibility because larger pools of the compounds are bound in the cell membranes. As support for this hypothesis, Prasad et al. (21) has demonstrated an increased membrane pool of phosphatidyl inositol in renal brush border membranes of hyperthyroid rats.
The abnormalities in poWM total choline correlated significantly (r = –0.57; P < 0.01) with impaired thyroid function, which indicated that thyroid hormones modulated the choline pools of the cellular membranes in the cerebral white matter.
Initial total choline in oGM predicted Median sentence repetition at follow-up (r = –0.53; P < 0.01) and shows a weaker correlation with the initial Median sentence repetition (r = –0.38; P < 0.05), these correlations are negative, an indication that lower total choline as seen in Graves’ disease corresponds to higher scores; the correlations are therefore unlikely to be related to Graves’ disease.
Initial total choline in oGM showed a weak correlation with scores on the Hamilton Anxiety scale. These correlations were positive and are therefore unlikely to be related to the mechanisms leading to decreased total choline in acute Graves’ disease.
Initial myo-inositol in poWM showed a weak correlation with the Forced choice face recognition test score at follow-up; it may, however, be a coincidental finding due to the large number of correlations.
Glutamate and glutamine
The amino acid glutamate is the brain’s most abundant excitatory neurotransmitter. In excessive amounts, it becomes neurotoxic, and it is an import metabolite in the mitochondrial redox cycle. Another amino acid, glutamine, is an inhibitory neurotransmitter. Glutamine is viewed as an astrocyte marker, whereas glutamate is also associated with neuronal tissue. Glutamate-glutamine cycling is a major metabolic pathway in the brain. Glutamate is released from the presynaptic neuron and taken up by the astrocyte, where it is synthesized into glutamine and shuttled back into the excitatory nerve terminals (10).
Glutamate and glutamine are usually reported as glutamate + glutamine, because the resonances partially overlap. It is difficult to separate the two using 1.5-T MR systems, unless one of them is severely elevated. But separate analyses of the two are possible in group studies as ours. Many studies only report the combined metabolites, so the glutamate + glutamine was included in our results mainly for the purpose of comparison with other studies.
The persistent reduction of glutamine in poWM, the decreasing glutamate in oGM, and the relation between abnormal initial thyroid parameters and the follow-up oGM glutamine as well as the significant correlations between poWM glutamate (also glutamate + glutamine) and the Stroop test incongruent indicated that there was a persistent and possibly progressive disturbance of the glutamate-glutamine cycling in Graves’ disease. The abnormalities were objective measures that may be related to the long-term cognitive and affective complaints reported in otherwise well treated patients with Graves’ disease. A similar pattern of reduced glutamate in gray matter has previously been reported in depression (23). One explanation could be that persistent disturbance of the glutamate-glutamine cycling may lead to cognitive and affective symptoms in Graves’ disease patients. Future MRS studies of treated patients with Graves’ disease with and without long-term cognitive and affective complaints are needed to strengthen this hypothesis.
The antithyroid therapy that the patients still received at follow-up could have caused the persistent reduced glutamine in poWM, falsification of such a theory needs additional studies.
N-Acetylaspartate + N-acetylaspartylglutamate and total creatine
N-Acetylaspartate and N-acetylaspartylglutamate overlap; hence, the term N-acetylaspartate + N-acetylaspartylglutamate was used. The major component is N-acetylaspartate, a metobolite almost exclusively found in neuroaxonal tissue (10). Total creatine represents the combined signals of phosphocreatine and creatine. These compounds are involved in energy metabolism via the creatine kinase reaction generating ATP (10).
The reduced but reversible N-acetylaspartate + N-acetylaspartylglutamate and total creatine in poWM may be related to a transient toxic effects on the white matter similar to transient white matter abnormalities sometimes seen in toxic encephalopathies.
The clinical significance of the weak correlations with follow-up affective scales scores is uncertain because they are positive (see Table 4). One would have expected that reduced N-acetylaspartate + N-acetylaspartylglutamate and total creatine if at all would lead to increased affective scales.
Conclusions
We have demonstrated significant MRS abnormalities in the acute phase of Graves’ disease. The abnormalities linked to glutamate-glutamine cycling persisted after treatment. The abnormalities have been demonstrated between groups of patients and healthy volunteers. Some of the abnormalities such as the decreased total choline in oGM were detectable in individual patients. Other abnormal findings such as the decreased neurotransmitter glutamate were visible as differences between groups. Improved methods for example using 13C MRS designed to access the glutamate-glutamine cycling are required before these abnormalities can be detected in individual patients and hereby become relevant in clinical practice.
Multiple correlations were analyzed so the findings may not all be significant, but the most significant relate to total choline, glutamate, or glutamine. The biochemical parameter that showed the most significant (P < 0.01) correlations with MRS quantities (poWM total choline and oGM glutamine) was total T3. Worse total T3 lead to worse MRS.
All correlations between MRS and affective scales were weak (0.01 < P < 0.05). Among the correlations between MRS and the cognitive scales, the most significant (P < 0.01) were related to oGM total choline or poWM glutamate and one of the attention and concentration scores (initial as well as at follow-up), the Stroop test incongruent or the Median sentence repetition. The data thus indicate that worse MRS data (oGM total choline or poWM glutamate) correspond to worse initial attention and concentration scores and predicts worse attention and concentration scores at follow-up.
Overall, the MRS findings for myo-inositol and total choline from the three volumes of interest were similar, indicating that the abnormalities represented diffuse and global neurochemical disturbances. Differing MRS findings were mainly found in poWM compared with fGM and oGM, which is likely to be due to the white matter vs. gray matter content rather than due to the regional locations in the brain.
The MRS abnormalities may be caused solely by the hyperthyroidism or by other factors related to autoimmune disease. Future studies of patients with nonimmune causes of hyperthyroidism are needed to address this issue.
|
Footnotes |
|---|
Disclosure Information: All authors have nothing to declare.
First Published Online May 20, 2008
Abbreviations: fGM, Midfrontal location primarily containing gray matter; MRS, magnetic resonance spectroscopy; oGM, midoccipital location primarily containing gray matter; poWM, parietooccipital location primarily containing white matter.
Received September 25, 2007.
Accepted May 14, 2008.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
