Pathophysiological Evaluation in a Case of Wernicke's Encephalopathy by Multimodal MRI
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.266252
Source of Support: None, Conflict of Interest: None
To report a patient with Wernicke's encephalopathy (WE) using multimodal magnetic resonance imaging (MRI) including conventional MRI, diffusion-weighted MRI (DWI), arterial spin labeling (ASL), and proton MR spectroscopy (MRS). A 50-year-old woman of WE with a history of cholecystectomy and acute pancreatitis was given MRI scans including DWI, MRS, and ASL pre- and post-thiamine treatment. Two weeks after admission, the patient's condition rapidly improved. The typical MRI findings and lesions in the frontal cortex at baseline disappeared or resolved partially. The reduced apparent diffusion coefficient value in part of the thalamus lesion, the elevated cerebral blood flow in the frontal cortex, the lactate doublet peak in the right thalamus lesion, and in cerebral spinal fluid, all resolved after treatment. The combination of conventional MRI with DWI, proton MRS, and ASL, offers a powerful diagnostic tool and a better understanding of the pathophysiological and hemodynamic mechanisms.
Keywords: Diffusion-weighted imaging, magnetic resonance imaging, magnetic resonance spectroscopy, pseudo-continuous arterial spin labeling, Wernicke's encephalopathy
Wernicke's encephalopathy (WE) is an uncommon but severe neurological syndrome, caused by thiamine (vitamin B1) deficiency. It is characterized by a sudden onset of altered consciousness, ophthalmoplegia, and ataxia. Conventional magnetic resonance imaging (MRI) showing typical (thalami, mammillary bodies, tectal plate, and periaqueductal area) and atypical (cerebellum, cranial nerve nuclei and cerebral cortex) signal intensity (SI) alterations, is an essential tool to get the right diagnosis, especially when clinical presentation is incomplete.
However, there are few reports of perfusion and proton MR spectroscopy (MRS) in WE.,,,, Proton MRS has been used as a non-invasive tool for evaluating brain metabolites in vivo. Pseudo-continuous arterial spin labeling (PCASL) using a three-dimensional fast spin echo sequence is entirely non-invasive to measure brain perfusion with an excellent intra- and inter-scanner reliability and reproducibility.
We herein report a patient with clinical and biochemical features consistent with WE using multimodal MRI including conventional MRI, diffusion-weighted MRI (DWI), PCASL, and proton MRS.
A 50-year-old woman with acute pancreatitis a week ago was admitted to our center in August 2017, because of persistent aphasia and walking instability for 2 days. The symptoms aggravated with unconsciousness and convulsion when the patient was hospitalized. Three weeks earlier, she had undergone an endoscopic cholecystectomy. After the cholecystectomy, she had malnutrition with taking in little porridge every day. At admission, a neurologic examination showed mild inattention, horizontal nystagmus, and bilateral ataxia. Laboratory studies showed elevated lactic acid of 3.8 mmol/L (normal range, 0.7–2.5 mmol/L) in blood, elevated serum amylase of 171 U/L (normal range, 35–135 U/L), and keto-bodies of 5 mg/dl in urine. Her brain computed tomography (CT) scan showed normal, so she was given an MRI scan to exclude ischemic stroke. Based on the clinical and imaging features, a possible WE diagnosis was made. Then intravenous administration of thiamine was performed. Two weeks later, a repeated MRI scan was performed.
MRI scans were performed on the night of admission and 2 weeks after thiamine treatment, respectively, on a 3.0 Tesla MR scanner (Discovery 750, GE Healthcare, Milwaukee, WI, USA) with an 8-channel phase-array head coil. The MRI scans included conventional MRI (axial T1WI, T2WI, axial, or coronal FLAIR), axial DWI, proton MRS, and PCASL. For proton MRS, single voxel of point-resolved spectroscopy (PRESS) technique was utilized. The voxel was placed in the dorsomedial thalamus and specifically in the anterior horns of lateral ventricles, as they did in Feng et al.'s study. The cerebral blood flow (CBF) map was generated using the GE AW 4.6 workstation.
This report was approved by the Institutional Review Board of our hospital and informed consent was obtained from the patient's family.
In the initial MRI scan, the patient showed symmetrical mildly to brightly high SI on T2WI/FLAIR in bilateral thalami, mammillary body, and periaqueductal region [Figure 1]f and [Figure 2]e. Besides, there were also lesions in the bilateral perirolandic cortex and frontal cortex, which is atypical finding of WE. DWI showed similar lesions of high signal in different extent [Figure 1]a and [Figure 1]c. Apparent diffusion coefficient (ADC) value in part of the thalamus lesion was reduced [Figure 1]b while not in the other lesions. Proton MRS with voxel positioning in the right thalamus lesion demonstrated a normal value of N-acetylaspartate (NAA)/phosphocreatine and creatine (Cr) ratio (1.26), choline-containing compounds (Cho)/Cr ratio (0.6), and positive lactate doublet peak at 1.33 ppm [Figure 2]a and [Figure 2]b, while voxel positioned at the lateral ventricles to measure the concentration of metabolites in the cerebral spinal fluid (CSF) showed marked lactate doublet peak at l. 33 ppm (NAA/Cr ratio = 1.04, Cho/Cr ratio = 0.95, respectively) [Figure 2]c and [Figure 2]d. PCASL measured symmetrically elevated CBF in the bilateral frontal cortex including the cortical lesions [Figure 1]d and [Figure 1]e.
Two weeks later, the patient's condition rapidly improved. For the second MRI scan after vitamin B1 treatment, the lesions in the bilateral thalamus partially resolved on DWI images [Figure 1]g with no obvious decrease of ADC values [Figure 1]h and lesions of the mammillary body and periaqueductal region disappeared [Figure 2]j. The cortical lesions disappeared on DWI images [Figure 1]i only with a focal high signal in the left perirolandic cortex on FLAIR [Figure 1]l. Single-voxel spectra of proton MRS of the thalamus lesion measured metabolites of normal ratios of NAA/Cr (1.28), Cho/Cr (0.71), and no lactate doublet peak at 1.33 ppm with short TE (35 ms) [Figure 2]f and [Figure 2]g, and no lactate doublet peak was found in spectra of voxel put in the lateral ventricles, either (NAA/Cr ratio = 1.04, Cho/Cr ratio = 0.95, respectively) [Figure 2]f and [Figure 2]g. PCASL investigated no abnormal CBF alterations in cortical regions [Figure 1]j and [Figure 1]k.
The present study reported an evaluation of WE by combining conventional MRI with advanced functional MRI techniques including DWI, single voxel proton MRS, and PCASL, which can render a better understanding of pathophysiology and hemodynamic changes of WE.
Thiamine is stored in body tissues, predominantly as thiamine diphosphate (TDP). TDP plays a significant role in the conversion of glucose into energy, acting as an essential cofactor for several enzymes in the Krebs cycle and in the pentose phosphate pathway. In the case of thiamine deficiency, intracellular TDP is depleted leading to a series of metabolic alterations in the central nervous system. The decreased activities of the pyruvate dehydrogenate complex, the alpha-ketoglutarate dehydrogenase complex, and the transketolase resulting in a reduction of Krebs cycle and pentose phosphate pathway deficiency, induce a cellular energy deficit due to reduced passage of pyruvate into the citric acid cycle, consequent intracellular accumulation of toxic intermediates such as lactate and alanine, and cerebral lactic acidosis, consistent with which were our patient's elevated serous lactic acid and urine keto-bodies. All these biochemical changes promote cell death by necrosis and apoptosis, leading to the histologic lesions widely described in previous reports, including intra- and extracellular edema and proliferation of pleomorphic microglia as the earliest modification.
The typical MRI findings of WE show increased T2/FLAIR or DWI SI involving symmetrical thalami, mammillary bodies, tectal plate, and periaqueductal area. And atypically, regions of the cerebellum, cranial nerve nuclei, and cerebral cortex can also be involved. In our patient above, similar manifestations (dorsomedial thalami, mammillary body, periaqueductal region, perirolandic cortex, and frontal cortex) were shown on conventional MRI. The clinical evolution and the symmetrical hyperintensities disclosed by T2WI/FLAIR and DWI suggest an acute stage of the disease in this patient, considering that this hyperintense SI is correlated with edema, especially when it is seen on DWI. The ADC of the thalamus lesion in our patient was reduced, which is a finding that has been associated with cytotoxic edema. However, the ADC in the other lesions showing no decrease was attributed to vasogenic edema. After thiamine treatment, neurologic symptoms in our patient and lesions in the thalamus regressed partially, and the cortical lesions disappeared almost completely, which are probably due to the osmotic dysregulation induced by decreased cellular energy levels.
Proton MRS is a powerful tool to biochemically characterizes metabolic brain diseases in vivo. To our knowledge, there have been very few reports of proton MRS in cases of WE.,,, Murata et al. and Mascalchi et al. both reported decreased NAA/Cr ratio in the lesion of thalamus and increased NAA/Cr ratio after thiamine treatment speculating it was presumably due to edema., Nevertheless, our patient showed normal NAA/Cr ratio in the thalami before and after thiamine treatment, similarly with what Rugilo and his collaborators reported. The acquired SI in the voxel of the right thalamus lesion probably averaged contributions from the lesion and surrounding normal brain tissues. Consistently with previous reports, the lactate doublet peak detected in the thalamus of our patient before thiamine treatment disappeared after treatment, which is probably a result of the increased anaerobic oxidation of carbohydrates due to thiamine deficiency. Besides, we positioned a voxel in the anterior horns of lateral ventricles to measure the metabolic alterations of CSF  and disclosed a remarkable increase in lactate without decreased NAA/Cr ratio, which is, we hypothesize, because lactate produced in lesion of the brain diffuses to CSF, where lactate is not normally detected. We speculate that the few normal brain tissues included in the voxel account for the NAA signal detected in this voxel. However, other studies are still required to further assess the utility of MRS as a potential diagnostic tool in WE.
As far as we know, only two documentary reports were involved with brain perfusion of WE in the literature., The two studies using xenon contrast CT and single-photon emission computed tomography (SPECT), respectively, both demonstrated reduced CBF in bilateral gray matter of WE patients. By using PCASL, a recently developed non-invasive tool to measure brain perfusion, we detected conversely elevated symmetrical elevated CBF in the bilateral frontal cortex including the cortical lesions. We hypothesize that in the acute phase of WE, accumulation of toxic intermediates such as lactate and cerebral lactic acidosis might lead to a change of local vascular features, such as changes in increased vascular permeability or auto-regulatory mechanisms by vasodilation and breakdown of blood–brain barrier that induces the local edema and hyper-perfused state. Yet it needs future studies to fully clarify this. In a word, the hemodynamic change during the disease onset is at least complex and much needs to be done to completely understand the hemodynamic process of WE.
In summary, the combination of conventional MRI with some of its new developments such as DWI, proton MRS, and PCASL offers not only a powerful diagnostic tool but also a better understanding of the pathophysiological and hemodynamic mechanisms.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2]