Wernicke’s encephalopathy: the role of cranioencephalic magnetic resonance in a difficult case

Description

Thiamine pyrophosphate is an essential coenzyme in several biochemical pathways in the brain, including intermediate energy production and axonal conduction. Early detection of thiamine deficiency is difficult as symptoms may be vague.¹ Wernicke’s encephalopathy (WE) is a severe life-threatening illness resulting from thiamine deficiency, with an estimated prevalence of 0.4–2.8%.² It can be observed in patients with a history of bariatric surgery and, less frequently, after placement of an intragastric balloon (IGB).^3–5

A morbid obese woman in her 40s, with no history of alcoholic habits, is brought to the emergency department (ED) due to frequent vomiting combined with lethargy and apathy. The patient had a history of an IGB placement 2 months before, followed by symptoms that took her to the ED on three occasions, initially with vomiting, followed by symptoms of vertigo, gradual onset of weakness and paresthesia affecting the distal lower limbs, causing gait imbalance and disorientation and culminating in a state of apathy and prostration observed in this episode. Two days before, the patient was diagnosed by her general practitioner with a urinary tract infection caused by Candida norvegensis; however, she did not take the medication due to persistent vomiting. On admission to the ED, the patient presented with a state of prostration and fever but was capable of simple communications. The physical examination showed a state of superficial stupor, with a Glasgow Coma Scale (GCS) score of 14 (E3, V5, M6), bilateral horizontal nystagmus and apparent ataxia of lower limbs, without clear meningeal signs. Laboratory findings showed an elevation of inflammatory parameters and an acute kidney injury: leucocytes: 13.0×10⁹/L; C reactive protein (CRP): 115 mg/L; haemoglobin: 117 g/L with an mean corpuscular volume (MCV) of 107 fL; blood urea nitrogen: 28.56 mmol/L; creatinine: 221 µmol/L; magnesium: 0.26 mmol/L; phosphate: 0.19 mmol/L; potassium: 3.0 mmol/L and creatine kinase: 6600 U/L. The toxicological study result was negative. A cranioencephalic-CT (CE-CT) scan was unremarkable. A central nervous system infection was ruled out by a normal cerebrospinal fluid (CSF) analysis. It was assumed the diagnosis of urinary sepsis, having started treatment with voriconazole, fluid therapy and correction of the electrolyte disturbances. Despite the treatment, she maintained persistent emesis and progressive worsening of her neurological condition with a GCS score of 8 (E2, V2, M4) after 3 days of in-hospital stay. A new CE-CT and lumbar punction were performed, showing normal results. A cranioencephalic-MRI (CE-MRI) was impossible to perform due to the patient’s weight at the time (body mass index of 56 kg/m²). An electroencephalogram (EEG) was compatible with metabolic encephalopathy. The patient was intubated and admitted into an intensive care unit (ICU). An abdominal CT scan revealed a complete occlusion of the gastric lumen by the IGB (figure 1), causing hyperemesis, so the balloon was removed. It was raised the hypothesis of vitamin deficiency due to hyperemesis, malnutrition and the concomitant infectious disease that increased metabolic rate. All this contributed to an increased metabolic demand leading to WE.

Figure 1.

Abdominal CT examination showing the presence of intragastric balloon occupying all gastric cavities.

Supplementation of thiamine (500 mg three times a day intravenously for 2 days, then 200 mg per day intravenously for 5 days, followed by 100 mg per day), folic acid (400 μg per day for 10 days) and cyanocobalamin 1 mg/week intramuscularly (maintained for 2 weeks) was started on admission to the ICU, only 5 days after the admission of the patient to the ED. Laboratory analysis found low vitamin B^₁₂ levels (79 pg/mL) but normal thiamine levels (92 pg/mL). Given the dehydration, with evidence of rhabdomyolysis with elevated creatinine and urea levels, a renal replacement technique was started in the ICU and maintained for 2 days to cover uraemic encephalopathy as the cause of the neurological deterioration. After 5 days, despite improvement in renal function and good control of the focus of infection, when sedative medications were stopped, there was no improvement in the patient’s neurological status with a GCS score of 4 (E1, V1T, M2) and bilateral horizontal nystagmus evoked by horizontal gaze to both sides. As the coma persisted, the CE-CT and EEG were repeated during hospitalisation with normal results. CSF study (syphilis, Borrelia infection, toxoplasmosis, Mycoplasma infection, Herpes simplex virus infection, Cytomegalovirus infection, Epstein-Barr virus infection and bacteriological culture) and vasculitis study were negative. On the 26th day of hospitalisation, after the loss of weight during length of hospitalization, a CE-MRI was performed and reported as normal, with no lesions that would justify the neurological condition. However, a new reassessment of the MRI was requested to a different neuroradiology expert who pointed to the presence of small hyperintensities in fluid-attenuated inversion recover/T2 in the periaqueductal topographies (figure 2), mammillary bodies (figure 3) and quadrigeminal plate (figure 4), all findings compatible with WE. Her neurological condition improved slowly, from a GCS score of 3 to a GCS score of 12 (E4, M6, V3), and she was discharged to the internal medicine unit after 45 days.

Figure 2.

Three-dimensional axial fluid-attenuated inversion recover (T2-weighted images) images showing hyperintense lesions involving periaqueductal grey matter (solid arrow) and mamillary bodies (dotted arrow).

Figure 3.

Three-dimensional coronal fluid-attenuated inversion recover (T2-weighted images) MRI showing mamillary bodies hyper-signal (dotted arrow).

Figure 4.

Three-dimensional sagittal fluid-attenuated inversion recover (T2-weighted images) with hyperintense lesions involving quadrigeminal plate (dotted arrow).

On the date of discharge, the patient was conscious and able to follow simple orders, without nystagmus, but with severe myopathy. Motor rehabilitation continued, with increased interaction with caregivers, maintaining, however, a marked functional limitation. A new MRI was scheduled after the patient’s discharge from ICU to internal medicine unit; however, the patient died before the examination due to nosocomial pneumonia with septicaemia.

WE diagnosis is difficult and is often delayed since the clinical triad (encephalopathy, ophthalmoplegia and ataxia) is present in only 15–30% of cases and even lesser in non-alcoholic patients.^{6 7} WE should be suspected and treated in patients with two of the following four Caine criteria: dietary deficiency, oculomotor abnormalities, cerebellar dysfunction and either altered mental status or mild memory impairment.⁸ There is no analytical study that can confirm the diagnosis since the serum or plasma thiamine testing has poor sensitivity and specificity, and levels do not necessarily reflect intracerebral thiamine levels.^{7 8} Therefore, a normal quantification of thiamine should not guide the initiation or suspension of the appropriate supplementation.^{7 8} It should be noted that magnesium plays an important role as a cofactor in the conversion of thiamine/vitamin B₁ into its biologically active form, thiamine pyrophosphate.¹ Low levels of magnesium can contribute to WE itself but also lead to a refractory response to thiamine treatment, so hypomagnesaemia, as presented in this case, should be evaluated and corrected promptly.¹

CE-CT has low sensitivity in diagnosing this condition.⁹ MRI can point us to the diagnosis by the presence of subtle findings that should be sought in these situations, such as small foci of periaqueductal grey (figure 2), mamillary bodies (figure 3 and figure 4) and quadrigeminal plate hyper-signal (figure 4), the latter being described in a lower percentage of the reported studies.^{1 9–12} In this clinical case, all four Caine criteria were met, with the diagnosis of WE being supported by a recent bariatric surgery and the appearance of specific neurological symptoms/signs after a few days of repeated vomiting. Knowing that numerous conditions can give rise to those same symptoms (figure 5), with WE also being triggered by some of these differential diagnosis conditions, in the face of a suggestive history, the appropriate treatment should not be delayed by the diagnostic testing and must be made while excluding other diagnoses, as intravenous administration of thiamine is safe, simple and inexpensive and could prevent the progress of WE to coma and death.⁸ Brain MRI remains the most valuable method to confirm a diagnosis of WE since it can show altered signal intensities in selective diencephalic and brainstem areas, at least 2–3 weeks after the occurrence of a definite thiamine deficiency.¹

Figure 5.

Differential diagnosis of WE made in this clinical case. WE, Wernicke’s encephalopathy.

Learning points.

• The diagnosis of Wernicke’s encephalopathy (WE) should be considered in patients with a history of intragastric balloon placement and the presence of specific neurological signs/symptoms.

• No laboratory findings or cranioencephalic-CT scan can make the diagnosis. MRI is the most valuable examination to confirm the diagnosis of WE, and we must seek some signals at first may go unnoticed.

• Clinicians should have a lower threshold to initiate vitamin supplementation and hypomagnesaemia correction, if needed, being aware that the clinical improvement after WE coma may take a long period of time.

Ethics statements

Patient consent for publication

Consent obtained from next of kin.