EARLY AND LATE STAGE EEG FINDINGS IN WERNICKE-KORSAKOFF SYNDROME DUE TO LONG-STANDING STARVATION: CORRELATION WITH THE CLINICAL AND MRI FINDINGS

Demet Kınay, Betül Baykan, Ayşen Gökyiğit, Emel Gökmen, Hüseyin Şahin, Candan Gürses, Rezzan Tuncay, Gencay Gürsoy, Hakan Gürvit
University of İstanbul, İstanbul Faculty of Medicine Department of Neurology

Keywords: Wernicke Encephalopathy, Wernicke-Korsakoff syndrome, EEG, starving, hunger strike, starvation, death fast

* This work was presented in part as a poster in the European Neurological Society Meeting held in 5-9 June 1999 in Milan-Italy

ABSTRACT:

Early and late stage EEG findings in Wernicke-Korsakoff Syndrome (WKS) due to long standing starvation were investigated. After a hunger strike performed by political prisoners, 18 patients (15M, 3 F) with a mean age of 30 years were managed in our clinic. Early stage EEG’s were performed between 20 and 40 days after the termination of the strike and following the treatment. EEG examinations were repeated in 17 patients one year later.

In the early stage, abnormal EEG findings were recorded in 5 patients (27.8%). Main abnormality was theta activity in frontal and frontotemporal regions bilaterally. Paroxysmal abnormalities were not observed. The abnormal EEGs had no relationship with the clinical findings at the acute stage and the follow-up findings after one year. There was no significant correlation between the distribution of abnormal EEG findings and the subtypes of clinical condition as WE or WKS. At the end of the first year, abnormal EEGs returned to normal in 3 cases, and showed some improvement in the other 2 patients. Increased intensity in thalamus and grey matter around the aquaducts were determined in MRI at the acute stage in 3 of 13 patients whose EEG findings were normal and 4 of 5 patients whose EEGs were pathological. There was a positive correlation between the EEG and acute stage MRI findings.

INTRODUCTION:

Wernicke’s encephalopathy (WE) is characterised by nystagmus, gaze palsies, ataxia of the gait, and confusion. These manifestations may occur singly or in various combinations and may follow a slowly evolving pattern or conversely  they are characterised by an abrupt onset. Korsakoff’s syndrome refers to an abnormality of mental activities in which recent memory is affected out of the proportion to other cognitive functions; the patient is otherwise alert and responsive. Most patients who present with the manifestations of WE survive the acute illness and are left with an amnestic state in which case the term Wernicke-Korsakoff syndrome (WKS) is more appropriate. WKS is due to a deficiency of thiamine (vitamin B1). If thiamine therapy is not initiated the patient may become comatose and die. On the other hand, when proper therapy is instituted symptoms resolve rapidly (1-3). Although this syndrome occurs most frequently in association with chronic alcoholism, recent case reports  have emphasised that it can occur in non-alcoholics as well (2,3). In WKS the lesions usually occur symmetrically in corpora mamillare, medial thalami, the areas adjacent to the third ventricle, aquaduct, and fourth ventricle.

There are only a few reports about EEG findings in patients with WE and WKS. In this paper, we aimed to investigate early and late EEG findings, and their relations with the MRI findings, clinical stages and clinical progress of WE/WKS which was due to long term starvation.

MATERIAL AND METHODS

After the termination of the hunger strike performed by political prisoners, 18 patients were managed in our clinic. There were 15 male and 3 female patients whose mean age was 29.94 ± 6.38. Their ages ranged from 23 to 36 years with the exception of one patient who was fifty years old. Duration of hunger was 67 to 69 days for 16 patients (three started 2 days later than the rest), a total of 54 days for one (temporarily ends the strike on the day 44 and resumes on the day 59) and 39 days for the last patient (becomes comatose and is started to be fed parenterally by his comrades). All patients were exposed to similar environmental conditions. Fifteen of them were from the major prison in Istanbul, which hosted the largest number of strikers. The remaining three were from other prisons in and close to Istanbul. Dietary intake was restricted to some amount of salt, sugar and liquids as water, lemonade and linden tea during the strike. Severe lassitude and vomiting starting around day 60 further restricted, often times prevented this daily allowance. Their mean weight loss was 21.8 kg (range from 11 to 31) and mean Body Mass Index (BMI) loss was 16.5 (range from 11.8 to 18.4) during this hunger strike.

Eight patients were admitted on the termination day and 10 others were referred within 20 days. All patients showed neurological signs compatible with Wernicke’s encephalopathy, 10 of whom developed Korsakoff’s syndrome.

Sixteen of 18 admitted patients returned to prison, after 13 to 40 day-hospital stay. One patient had already been released. One patient was transferred to the psychiatry department for further management (Patient no;1).  We followed up all patients for a period of one year.

Clinical features on admission and at the end of the first year were summarised in Table 1. In order to follow the course and determine the prognosis in one year period, the patients were evaluated with a severity scale developed by one of us (HG) (scale 1). The patients were divided to mild (stage1a and 1b), moderate (stage 2) and severe ( stage 3 and 4) stages.

EEG Examinations:

In 17 patients, early stage EEG’s were performed between 20-40 days after termination of the hunger strike and following the thiamine treatment. In one patient, early stage EEG was performed 2 months after termination of the hunger strike and after one month of thiamine therapy. At the time of the first EEG; all blood biochemistry parameters were normal except low albumin levels.

At the time of the first EEG examinations, neurological examination revealed a global confusional state in 6 patients. State of alertness was somewhat comprised in two, and the other four were alert, yet confused, with evident problems in attentional and mnestic tasks in neuropsychological testing.

EEG examinations were repeated in 17 patients one year later.

All EEG’s had been read and interpreted by two experienced electroencephalographers (AG, BB) who were blind to the clinical data.

Recordings were performed with scalp electrodes placed according to the International 10-20 system with both bipolar and referential montages . Samples, each lasting 30 min, were obtained while the subject was awake and relaxed, with eyes closed. Standard activating procedures was performed in all patients. Interictal background activity was studied and selected by visual inspection in artefact-free EEG samples. Slowing of the background activity was classified as delta (0.5-3.5 Hz) and theta (4-7 Hz).

MRI Examinations:

MRI scans were obtained within the first week in 6 patients and within the second week in 3 patients. Nine patients were studied  after the second week .

MRI scans were performed with an 1.0 T unit (Siemens, Magnetom impact). Spin-echo (SE) sequences, 2620/17 (TR/TE) and 2900/85 (TR/TE) were obtained with 5-mm thickness (axial) for the whole brain and a 3-mm thickness (coronal) though the periaquaductal region and third and fourth ventricles. Sagittal and axial SE 600/15 (TR/TE) scans with a 5-mm thickness were also obtained. In all patients, sequences included sagittal spin-echo (TR/TE:600/15) with 5-mm thickness and coronal spin-echo (TR/TE: 600/15) with 3-mm thickness performed after intravenous administration of Gd-DTPA. Lesions of paraventricular regions of the thalamus and hypothalamus and periaquaductal regions of the midbrain were evaluated in all cases by 2 independent observers (RT, GG), as well as mamillary body, cerebral, and cerebellar shrinkage.

RESULTS:

Early phase:

In the early phase of the illness, abnormal EEG findings were recorded in 5 patients (%27.8).  Abnormal EEG findings consisted of mild regular slow activity in theta range which involved frontal regions in 2 patients (no;2,3) and frontotemporal regions in 2 patients (no;7,13). In the last patient, EEG revealed a moderate, regular generalised slow activity with the greatest amplitude over the frontal regions; this was accentuated during hyperventilation (no;1). Paroxysmal abnormalities were not observed.

Two patients with clouded consciousness (no;1,2) and one of the 4 alert but confusional patients (no;3) had pathological changes of the  EEG. Two patients who had abnormal EEG findings were not confusional and showed only motoric signs of WE (no;7,13). There was no significant correlation between abnormal EEG findings and mental state of the patients  (Fisher’s Exact Test; P=0.2682).

Two of the 8 patients who were classified as WE (no;7,13) and 3 of the 10 patients who were classified as WKS (no;1,2,3) had pathological EEG changes. There was no significant relation between the distribution of abnormal EEG findings and the subtypes of clinical condition as WE or WKS (Fisher’s Exact Test; P=1.0000).

Nine of the 13 patients whose EEGs were normal had been rated being in  severer clinical stages (stages 4 and 3) and 4 of them were in moderate clinical stage (stage 2) according to our scaling system. On the other hand, four of the 5 patients who had abnormal EEG findings, were in those severer clinical stages (stages 4 and 3) and the last one was in moderate stage (stage 2). No significant correlation was found between the EEG findings and the severity of clinical findings at this early stage (Fisher’s Exact Test; P=1.0000).

Late Phase:

EEG examinations repeated in 17 patients showed abnormalities in 3 patients one year later. In 2 patients, abnormal EEG findings included slight theta activity over one frontotemporal (no;15) and temporal region (no;13) and these were observed only during hyperventilation. The 3^rd  patient had a slight slow activity over the frontotemporal regions bilaterally in the routine EEG. This abnormality was accentuated during hyperventilation (no;1).

The abnormal EEG findings performed in the early period and at the end of the first year were summarised in Table 2. Two of the 3 patients who had abnormal EEG findings at the end of the first year had also abnormal EEG findings in the early period (no1,13), but the 3^rd patient had a normal EEG in the early period. (no;15).

All of the abnormal EEGs in the early phase of the illness found to improve to a certain degree (more or less) at the end of the first year. In the patient whose EEG had been grossly abnormal, generalised slow activity diminished and localised to frontotemporal regions (no;1). The slow-wave activity were observed only during hyperventilation in one patient (no;13).

Prognosis of the patients after one year according to the clinical stages was seen in table 3. There was no significant correlation between abnormal EEG and the prognosis after one year follow-up (Fisher’ Exact Test; P=0.3260).

Clinical course was also not different in the patients who still have abnormal EEG findings compared with the patients whose EEGs returned to normal.   The exceptional patient who had abnormal EEG at the end of the first year, although his first EEG was normal, did not show any deterioration in the clinical course (no,15).

MRI findings of the patients are listed in Table 4. In the early phase of the illness, symmetrical signal abnormalities on T2-weighted images in the periaquaductal areas and in the medial thalami were found in 4 of 5 patients (Figure 1) whose EEGs were abnormal  (no;1,2,3,7) and 3 of 13 patients whose EEGs were normal (no;4,5,9). We found a significant correlation between the results of  the EEG and the symmetrical signal abnormalities on T2-weighted MRI images in the periaquaductal areas and in the medial thalami (Spearman correlation coefficient r= 0.53, p=0.023). There was also a weak correlation between EEG findings and cortical atrophy on MRI (Spearman correlation coefficient r= 0.47, p=0.045). No significant correlation could be shown between EEG findings in the early period and MRI findings such as mamillary atrophy, third ventricular and aquaductal dilatation and cerebellar and vermian atrophy ( Spearman correlation coefficient r=0.01, p=0.95; r=0.19, p=0.44; r=0.026, p=0.918; r=-0.02, p=0.919; respectively) .

DISCUSSION

The true prevalence of Wernicke’s encephalopathy is clearly higher than what is diagnosed clinically. The adult autopsy incidence of up to 2.2% (4,5) emphasises the underdiagnosis of this syndrome, which is recorded clinically in only 0.04% to 0.13 % of all hospital admissions (4,6) It may be difficult to formulate a clinical diagnosis in patients who have an  atypical history and the incomplete or attenuated forms of the syndrome, so that better tools for the early and definite identification of this disease (and the prediction of the prognosis) are needed in clinical practice (2). Most studies have reported that imaging modalities such as MRI may contribute greatly by providing an early diagnosis for this life-threatening disease (7-9). There are only a few reports about EEG findings in patients with WE and they are usually anectodal case reports.

In this study, EEG findings were examined in a homogenous group of WE/WKS cases who have suffered from long term and severe lack of food for similar periods. The electroencephalograms were performed after the institution of thiamine therapy and slight improvement was observed in clinical findings of all our patients. We followed up our patients in the prison for a period of one year and their EEGs were reevaluated at the end of this period. In the early phase of the illness, the EEG abnormalities were recorded in 27.8% of patients in our study. Victor et al. reported that about half of the patients with Wernicke-Korsakoff disease show EEG abnormalities from a mild to moderate diffuse slow activity (4). In their monograph, almost all patients were alcohol addicts. Kelley and Reilly have reported varying degrees of background slowing with increased amounts of theta and delta activity in WE (10).  EEG abnormalities in reported cases with WE/WKS showed diffuse slow wave activity or frontal/ frontotemporal dominance of the frequent slow waves (9, 11-15). Martinez-Barros et al. emphasised that EEG showed non-specific changes which can appear with many kinds of diffuse encephalopathy regardless of etiology (15).

It is noteworthy that pathologic EEG findings in our group were consisted of mild slowing observed predominantly in frontal, frontotemporal regions. Fratzen has also reported three cases with WE whose EEGs showed a moderate to severe degree of slowing with the greatest amplitude over the fronto-temporal regions (11). The cause of WE in one of his cases was hunger strike as in our cases. Fournet and Lanternier reported diffuse sharp and slow wave complexes in severely ill patients (16). Paroxysmal abnormalities were not observed in any of our patients.

No statistically significant difference could be demonstrated in means of pathological EEG findings between the WE and WKS subgroups in our patients. This is an original finding not reported before.

There was no relationship between the pathologic EEGs and severity of the neurological clinical findings. EEG examinations were normal in most of the cases with severe clinical findings of WE/WKS. Victor et al. have also reported that in some patients with the full-blown acute syndrome, who later proved to have extensive lesions in the diencephalon in postmortem examinations, the EEG was entirely normal (4).   

Total cerebral blood flow and cerebral oxygen and glucose consumption may be greatly reduced in the acute stages of the disease, and these defects may still be present after several weeks of the treatment (3). These observations indicate that profound reductions in brain metabolism need not be reflected in EEG abnormalities or in depression of the state of consciousness (3).

We would like to emphasise some improvement in the EEG findings after treatment. In our cases, all of the EEG abnormalities detected in the early phases of the illness improved to a lesser or more  extend at the end of the first year. The improvement of the EEG abnormalities paralleled clinical improvement. In only one of our cases who had the most severe abnormalities in the early phases, EEG abnormalities continued one year later, although slight improvement was observed. Similar to ours, in three cases reported by Fratzen, EEG abnormalities detected in the early phases became less within several months after the thiamine therapy and returned to normal one year later (11). In a case report with WK caused by psychogenic food refusal, Doraiswamy et al. have reported that the diffuse slowing detected in the first EEG was replaced by mild background (7 Hz) and bitemporal (4 to 5 Hz) slowing seven months later high dose thiamine treatment (14).  In another case report with WE, Yokote et al have reported that the EEG showed basic rhythms of 6-7 Hz theta waves without laterality at the beginning. Two weeks after the administration of thiamine and the patient regained consciousness, the EEG demonstrated improvement with basic rhythms of 8-9 Hz slow alpha waves with frontal-dominant frequent theta waves (9). Dreyfus and Victor found that EEG changes could range from mild to severe and in general, paralleled the severity of neurological deficits (17).

We demonstrated that there was no relation between the initial EEG findings and clinical status at the end of the first year. We can, therefore, assume that EEG abnormalities in WE/WKS are not useful  in the prediction of the prognosis.

We found an interesting significant correlation between the results of EEG and the symmetrical signal abnormalities on T2-weighted MRI images in the periaquaductal areas and bilateral medial thalami. Follow-up MRI scans in some reports have revealed atrophy of mamillary bodies, cortical atrophy and third ventricular and aquaductal dilatation (7-9, 14). Early thiamine replacement may result in complete resolution of thalamic and midbrain MR signal abnormalites (7,8,14). Galluci et al have reported that the high signals on T2-weighted images represent pathologically non-specific findings such as edema, demyelination and gliosis (8). Comparison of the MRI with the pathological findings showed that the high signals seen on the T2-weighted images corresponded to edematous change induced by spongy disintegration of the neurophil (18). In some reports, the mamillary bodies have been enhanced on MRI in acute WE (19,20). Mamillary body atrophy has been reported as the sign of chronic WE (21). The exact mechanisms underlying the pathogenesis of the lesions observed in WE are not understood completely. The role of thiamine in the pathophysiology of WE may be related to its involvement in the function of excitable membranes and neurotransmitter production (8). Antunez et al have that MRI is useful in confirming the diagnosis of acute WE, but the absence of abnormalities on MRI does not exclude the diagnosis and the sensitivity of MR imaging in revealing evidence of this disease was 53% and the specificity, 93%   (22).

In conclusion; abnormal EEGs were found in 27.8% of the cases with WE/WKS, consisting of slow wave activity in frontal and frontotemporal regions which improved markedly after one year. The abnormal EEGs had no relationship with the neurological findings and the subtypes of clinical condition as WE versus WKS. There was a significant correlation between the pathologic results of  the EEG and the T2 hyperintensity on MRI in the periaquaductal areas and in the medial thalami.

REFERENCES

1.   Victor M, Adams RD, Collis GH. The Wernicke-Korsakoff syndrome and related neurologic disorders due to alcoholism and malnutrition, 2nd ed. Philadelphia; Davis, 1989:1-117.

2.   Victor M. MR in the diagnosis of Wernicke-Korsakoff syndrome. AJNR 1990;11:895-896.

3.   Adams RD, Victor M, Ropper AH. Principles of neurology. New York:McGraw-Hill,1997.

4.   Victor M, Adams RD, Collins GH. The Wernicke-Korsakoff Syndrome: A Clinical and Pathological Study of 245 Patients, 82 with Post-Mortem examinations. Philadelphia, Davis Co, 1971;1-174.

5.   Harper C. Wernicke’s encephalopathy: a more common disease than realised. A neuropathological study of 51 cases. J Neurol Neurosurg Psychiatry 1978;42:226-231.

6.   Harper CG. The incidence of Wernicke’s encephalopathy in Australi: a neuropathological study of 131 cases. J Neurol Neurosurg Psychiatry 1983;46:593-598.

7.   Donnal JF, Heinz ER, Burger PC. MR of reversible thalamic lesions in Wernicke syndrome. AJNR 1990;11:893-894.

8.   Galluci M, Bazzao A, Splendiani A, Masciocchi C, Passariello R. Wernicke encephalopathy: MR findings in five patiens. AJNR 1990;11:887-892.

9.   Yokote K, Miyagi K, Kuzuhara S, Yamanouchi H, Yamada H. Wernicke encephalopathy: Follow-up study by CT and MR. Journal of Computer Assisted Tomography 1991;15:835-838.

10.Kelley JT and Reilly EL. EEG, alcohol and alcoholism. In: (1983) EEG and Evoked Potentials in Psychiatry and Behavioral Neurology, edited by JR Hugnes and WP Wilson. Butterworths, Boston.

11. Frantzen E. Wernicke’s encephalopathy. Acta Neurol Scand 1966; 42:426-441.

12.Kitaguchi T, Kobayashi T, Tobimatsu S, Goto I, Kuroiwa Y. Computed tomography and magnetic resonance imaging in a young patient with Wernicke’s encephalopathy. J Neurol 1987; 284:449-450.

13.Roche SW, Lane RJM, Wade JPH. Thalamic hemorrhages in Wernicke-Korsakoff syndrome demonstrated by computed tomography. Ann Neurol1988;23:312-313.

14.Doraiswamy PM, Masse EW, Enright K, Palese VJ, Lamonica D, Boyko O. Wernicke-Korsakoff syndrome caused by psychogenic food refusal: MR findings. AJNR 1994;15:594-596.

15.Martinez-Barros, Ramos-Peek, Escobar-Izqulerdo, Figuera JR. Clinical, electrophysiologic and neuroimaging findings in Wernicke’s syndrome. Clinical Electroencephalography 1994; 25;148-152.

16. Fournet A, Lanternier M. Consultation electroenceplographique dans 17 cas d’encephalopathie de Gayet-Wernicke. Rev Neurol (Paris) 1956;94:644-645.

17. Dreyfus PM, Victor M. Effects of thiamine deficiency on the cerebral nervous system. Am J Clin Nutr 1961;9:414-425.

18. Suzuki S, Ichijo M, Fujii H,  Matsuoka Y, Odawa. Acute Wernicke’s encephalopathy: Comparison of magnetic resonance images and autopsy findings. Internal Medicine 1996;35:831-834.

19. D’Aprile P, Gentile MA, Carella A. Enhanced MR in the acut phase of Wernicke encephalopathy. AJNR 1993;15:591-593.

20. Shogry MEC and Curnes JT. Mammillary body enhancement on MR as the only sign of acute Wernicke encephalopathy. AJNR 1994;15:172-174.

21. Charness ME, DeLaPaz RL. Mammillary body atrohy in Wernicke’s encephalopathy: antemortem identification using magnetic resonance imaging. Ann Neurol 1987;22:595-600.

22. Antunez E, Estruch R, Cardenal C, Nicolas JM, Fernandez-Sola J, Urbano-Marquez A. Usefuness of CT and MR imaging in the diagnosis of acute Wernicke’s encephalopathy. AJR 1998;171:1131-1137.

Scale 1:             Activities of Daily Living Scale for Hunger Strikers

*developed by one of us (HG)

Table 1: Neurological findings at the admission and after one year 

·        Numbers represent individual patients

Table 2.  The pathological EEG findings in the early period and one year later

*This patient had the most severe EEG abnormality in the early period and also after one year.

Table 3 . Prognosis of the patients after 1 year

*the numbers indicates individual patients

L: disclosed the patients whose EEG were abnormal in the early stage (no; 1,2,3,7,13)

** Stage 4 and 3=severe, stage 2= moderate, stage 1b and 1a= mild.