Locked‐in syndrome after central pontine myelinolysis, an outstanding outcome of two patients

Abstract

Objective

Central pontine myelinolysis (CPM) is a rare demyelinating disease that affects the pons and which can cause extreme disabilities such as locked‐in syndrome (LIS) in the initial phase. The aim of the study was to describe the evolution over a 12‐month period of two patients with CPM causing an initial LIS.

Method

We retrospectively report the unexpected clinical outcome of these two patients in relation with the anatomical damages documented by brain MRI, associated with diffusion tensor imaging and reconstruction of corticospinal tracts in tractography. The following clinical parameters systematically assessed at 3, 6, 9, and 12 months: muscle testing on 12 key muscles (Medical Research Council), prehension metrics (box and block test and purdue pegboard), and independence for acts of daily living (functional independence measure).

Results

Both patients showed a progressive recovery beginning between 2 and 3 months after the onset of symptoms, leading to almost complete autonomy at 12 months (FIM > 110), with motor strength greater than 4/5 in all joint segments (MRC > 50/60). On brain MRI with tractography, CST appeared partially preserved at pons level.

Interpretation

The possibility of a near‐complete functional recovery at 12 months is important to consider given the ethical issues at stake and the discussions about limiting care that may take place initially. It seems to be the consequence of reversible myelin damage combined with partially preserved neurons. Development of collateral pathways or resolution of conduction block may explain this recovery. MRI comprising DTI and tractography could play a key role in the prognosis of motor recovery.

Introduction

Central pontine myelinolysis (CPM) is a rare demyelinating disease that affects the pons and which can cause extreme disabilities such as locked‐in syndrome (LIS) in the initial phase. ¹ , ² CPM was first used in 1959 by Adams et al. ³ to describe a symmetrical demyelinating lesion of the central pontine bases occurring in a 38‐year‐old patient with chronic alcoholism. The current hypothesis to explain the disease is a too rapid correction of hyponatremia ⁴ , ⁵ leading to intracellular dehydration, endothelial damage, and a blood–brain barrier alteration that is responsible for the exposure of the glial cells to cytokines and complement, leading to myelin degradation and/or oligodendrocyte apoptosis. ⁶ , ⁷ It seems that oligodendrocyte is the most vulnerable type of cell to osmotic stress related to CPM, whereas astroglia and neurons are most resistant. ⁸ This phenomenon, named osmotic demyelination syndrome (ODS), can be centered within the pons (CPM) and/or affect other regions (Extrapontine myelinolysis, EPM). ⁹ Based on isolated cases or case‐series, ¹ , ² , ³ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ its epidemiology remains poorly documented but the incidence of which was estimated to be 2.5% among overall stays in a recent study performed in India over 5 years in an intensive care unit. ¹⁶

Intrinsic factors favoring the appearance of myelinolysis have been identified ² ; these include chronic systemic disease (cancer, liver disease, ¹⁰ liver transplantation, ¹⁷ diabetes, sepsis, severe burns, ¹⁸ electrolyte disorders, etc.), but also chronic alcoholism and malnutrition. ¹⁹ The importance of these factors is highlighted by the description of CPM without any history of hyponatremia or too rapid variation of natremia. ¹⁰ , ¹⁸ , ²⁰ , ²¹ , ²²

The clinical course of CPM is classically biphasic, beginning with encephalopathy secondary to hyponatremia (headache, myalgias, hyporeflexia, etc.) followed by a spontaneous gradual regression, and then, a second phase of neurological degradation is reported to occur in the following days with the appearance of dysarthria, obnubilation, quadriplegia, ataxia, and even LIS or death. ² , ⁸ , ⁹ It is thus a disease with a potentially very poor prognosis although improvements in neuroimaging and treatments have led to improve the morbi‐mortality of CPM ⁸ , ¹¹ , ²³ ; this is illustrated by two case series studies that found that more than half of the patients with ODS regain independent function. ¹² , ²⁴ However, to the best of our knowledge information related to the course of the most severe cases, in particular those presenting initial LIS, is limited to only the diagnosis and hospital discharge. ¹ , ⁴ , ²⁴ More detailed clinical and paraclinical examinations (electrophysiology, neuroimaging) regarding the course of the disease during the hospital stay would be interesting to better understand the mechanisms of recovery and predict the functional prognosis in such severe cases of CPM for which important ethical decisions may be made.

In this context, we report herein two patients with CPM who presented a disorder of consciousness followed by a LIS who were admitted to our hospital. We describe the course of their recovery, over a period of 12 months with regular assessments, but also the results of neurophysiological evaluations and imagery available at the initial phase in order to discuss possible implications in terms of recovery mechanisms and prognostication.

Clinical Presentation and Diagnosis

Patient 1

A 39‐year‐old man with a history of diabetes insipidus treated daily by desmopressin was admitted to the emergency room for general physical deterioration, nausea, dizziness, and insomnia in the context of fasting that had begun 2 weeks earlier. The laboratory workup showed severe hyponatremia (Na⁺ = 102 mmol/L). The state of consciousness then deteriorated (Glasgow coma scale score = 3), in a context of rapid overcorrection of the natremia (Na⁺ = 154 mmol/L) at 72 hours requiring a transfer to intensive care unit. Oro‐tracheal intubation was performed 3 days after admission. The level of consciousness then improved transiently concomitantly with a new episode of hyponatremia (Na⁺ = 120 mmol/L). Brain MRI performed at Day 4 was considered normal. A new episode of alteration of consciousness occurred at Day 7 concomitant with a second episode of hypernatremia (Na⁺ = 151 mmol/L). MRI performed at Day 12 found demyelinating damage of the entire protuberance, extending to the midbrain and associated with an abnormal signal in the basal ganglia corresponding to a diagnosis of CPM with EPM.

In the absence of clear clinical signs of consciousness at Week 4 (Glasgow coma scale score = 8; E3V1M4), a more complete evaluation was performed in a neuro‐intensive care unit at Week 6. Consciousness was assessed twice using the coma recovery scale‐revised ²⁵ and the scores (13 and 15; Au3*‐V5*‐M2/3*‐O1‐C0‐Ar2/3) corresponded to a minimally conscious state. The somatosensory evoked potentials (SEPs) ²⁶ , ²⁷ after right and left median nerve stimulation at the wrist showed preserved N9 peripheral nerve responses at Erb's point and N13 cervical response associated with an abolition of the cervico‐bulbar junction P14 response and of the N20 cortical parietal response. This pattern was compatible with a bilateral impairment of the somesthetic pathways at the brainstem level. The motor evoked potentials (MEPs) performed with transcranial magnetic stimulation and recording sites at the abductor pollicis brevis and anterior tibialis muscles showed an abolition of the responses for the four limbs. This confirmed the severe functional impact of the ponto‐mesencephalic lesion visible on MRI. Notably, despite brainstem auditory evoked potentials showing an increase of the I‐V waves interpeak interval compatible with a functional impairment at the brainstem level, middle latency auditory evoked potentials (AEP's) showed preserved primary auditory cortex responses and auditory event‐related potentials showed preserved cortical P3 cognitive response to the patient's name. ²⁸ At Week 8, a new clinical assessment combined with an eye‐tracker allowed the identification of reliable functional communication (correct responses to closed questions, see videos in the appendix) compatible with a LIS (complete quadriplegia, mutism, dysphagia requiring a gastrostomy).

Patient 2

A 46‐year‐old malnourished woman with chronic alcohol consumption was hospitalized for purulent peritonitis following a perforated peptic ulcer. This was complicated by acute myocarditis and, at Day 10 following admission, by a progressive deterioration of the neurological state with consciousness disorders (hallucinations, confusion, Glasgow coma scale score = 11) in a context of oedemato‐ascitic decompensation. A progressive correction of the natremia from 129 mmol/L to 150 mmol/L occurred over a period of 7 days during this phase. The neurological status evolved toward a LIS. The clinical evaluation performed at Week 8 found almost complete quadriplegia (flicker of contraction of the right elbow), with mutism and dysphagia. The patient was able to open and close her eyes allowing a yes/no code and functional communication using a visual alphabet. The electroencephalogram (EEG), the lumbar puncture, and the CT scan did not reveal any obvious abnormality; brain MRI performed at Day 20 showed a CPM. The persistence of swallowing disorders due to bulbar damage justified a percutaneous tracheotomy and a gastrostomy.

Clinical evaluation during follow‐up

We retrospectively report the course of several clinical parameters systematically assessed at 3, 6, 9, and 12 months. Motor testing of the four limbs was evaluated using the Medical Research Council (MRC) score. ²⁹ This scale rates the strength (between 0 and 5) for three muscle groups for each limb: arm abduction, forearm flexion, wrist extension, thigh flexion, leg extension, and foot dorsal flexion. The total score ranges from 0 to 60.

Prehension abilities were evaluated using the Box and Block Test ³⁰ (gross manual dexterity) and the Purdue Pegboard ³¹ (measurement of coarse arm, wrist and finger movements, and finger dexterity). The results are expressed as absolute values and as a percentage of the norm, for the right hand, which was dominant in both patients. Both tests were performed by occupational therapists.

Autonomy in daily living was measured using the functional independence measure (FIM) score ³² that evaluates the ability of patients to perform several tasks of daily life independently.

Radiological assessment

Anatomical damage was documented by brain MRI, associated with diffusion tensor imaging and reconstruction of corticospinal tracts in tractography. Diffusion‐weighted images were acquired on a 3‐T Ingenia Elition MRI machine (Philips Medical System, Netherlands) with the following parameters: 32 directions, b‐value 0 and 1000 s/mm², TE/TR: 55/5002 ms, in‐plane resolution: 0.875 mm; slice thickness: 2 mm, no slice gap. A distortion correction method was applied for diffusion‐weighted images of each participant, and motion correction was conducted with b‐table rotated. The diffusion data were reconstructed using generalized q‐sampling imaging, with a diffusion sampling length ratio of 1.25. With DSI Studio software, a deterministic fiber tracking algorithm was used with augmented tracking strategies to improve reproducibility. ³³ , ³⁴ A region of interest (ROI) was drawn on pyramids of the medulla oblongata to track the corticospinal tract. The tractography algorithm used the following parameters: adaptative anisotropy threshold, fiber length: 50 to 300 mm, angular threshold: 60°, and step size: randomly selected from 0.5 to 1.5 voxels. A total of 1 000 000 seeds were calculated. Fibers were displayed as 0.10 mm‐diameter tubes. The tensor metrics analyzed were as follows: (i) fractional anisotropy (FA), a summary measure of microstructural integrity, which is lower than in healthy subjects when the white fiber integrity is impaired; (ii) mean diffusivity (MD), an inverse measure of membrane density and cellularity, which is higher than healthy subjects in case of edema or necrosis; (iii) axial diffusivity (AD), ³⁵ , ³⁶ which is classically positively correlated with axonal density; and (iv) radial diffusivity (RD), positively correlated with demyelination. Statistics of tractograms (number of fibers, mean length of fibers per tract in mm, tract volume in mm³, total surface area of the tract in mm²) were then extracted. Then, a ROI volume corresponding to the pontine myelinolysis was drawn manually, and tensors metrics (FA, MD, and RD) of corticospinal tracts (CST)—previously tracked—within the pontine myelinolysis were extracted. These metrics were compared to healthy subjects, previously acquired.

Motor and Functional Evolution

Patient 1

The first signs of recovery were noted in the second month: the patient became able to express himself with words and then sentences, although cognitive difficulties remained. Swallowing gradually improved, enabling the gastrostomy to be weaned after 3 months. In terms of motor function, the patient began to recover limb motor skills starting during the third month after the onset of symptoms. At 12 months, muscle strength was almost normalized according to the MRC score (59/60). Similarly, the other neurological impairments (damage to the cranial nerves and cognitive functions) fully recovered during the year of follow‐up. Neuro‐orthopedic complications occurred in the sub‐acute phase with equino‐varus of the feet and para‐osteo‐arthropathy of the elbows and hips. Corrective orthopedic surgeries were performed at 12 months follow‐up. The patient was able to return home after 15 months of in‐patient hospitalization and was fully autonomous for activities of daily living (FIM: 125/126; Fig. 1, Table 1).

Figure 1.

Graphical representation of motor strength, prehension and autonomy during the 1‐year follow‐up. (A) Motor strength represented by Medical Research Council (MRC) score. (B) Prehension represented by Box and Block test and Purdue Pegboard and expressed as a percentage of the standard. (C) Autonomy represented by functional independence measure (FIM).

Table 1.

Grip scores for both patients at 0, 3, 6, 9, and 12 months, expressed as absolute values.

	6 months		9 months		12 months
	Patient 1	Patient 2	Patient 1	Patient 2	Patient 1	Patient 2
Box and Block (number of blocks moved in 1 minute)	39	42	45	38	55	40
Purdue Pegboard (number of stems)	7.3	7.3	9.7	9.3	10.3	11

Brain MRI associated with diffusion tensor imaging (DTI), performed 5 months after the onset of symptoms, showed CPM. Compared to healthy subjects, tensor metrics showed anomalies on CST axonal integrity (FA was lower), CST myelination (RD was higher), and supposed CST edema (MD was higher; Table 2). It is of note that at the pontine myelinolysis level, FA was lower than in healthy subjects; MD, RD, and AD were higher (Table 3) and the values of which were greater than those found in CST. Right and left CST were symmetrical in tractography. CST crossed the pontine myelinolysis without warped fibers at this level (Figs. 2 and 3).

Table 2.

Diffusion tensor images and tractography metrics of the right and left CST of each patient.

	Patient 1		Patient 2		Healthy subjects (12 CST)
	Right CST	Left CST	Right CST	Left CST
Number of fibers	4592	5706	670	4456	11412
Mean length (mm)	121.6	122.2	116	115	114.7
Volume (mm3)	12143	9426	2458	5212	32203.8
Total surface area (mm2)	25497	21692	6579	9744	86688.2
Mean FA	0.45	0.45	0.41	0.46	0.54
Mean MD	0.89	0.81	0.83	0.81	0.73
Mean RD	0.58	0.61	0.64	0.59	0.48
Mean AD	1.70	1.49	1.40	1.46	1.23

AD, axial diffusivity; CST, corticospinal tract; FA, fractional anisotropy; MD, mean diffusivity; RD, radial diffusivity.

Table 3.

Diffusion tensor images and tractography metrics of the pontine myelinolysis area.

	Patient 1	Patient 2
Pontine myelinolysis volume (mm3)	3102	3405
Mean FA	0.21	0.20
Mean MD	1.52	1.78
Mean RD	1.38	1.61
Mean AD	2.02	2.36

AD, axial diffusivity; CST, corticospinal tract; FA, fractional anisotropy; MD, mean diffusivity; RD, radial diffusivity.

Figure 2.

Brain magnetic resonance imagery (MRI) at month 5 for patient 1. (A) 3D T1 EG sagittal: centropontic hyposignal. (B) 3D Sagittal FLAIR: centropontic hypersignal. (C) Axial T1 EG reconstruction: Centropontic hyposignal, corticospinal tracts partially identified.

Figure 3.

T2 diffusion sequences with tractography of the corticospinal tracts of the brain magnetic resonance imagery (MRI) at month 5 for patient 1. (A) Orientation distribution function (ODF) map focused on the pontine myelinolysis area. Corticospinal tract (CST) is highlighted (in blue due to fibers craniocaudal direction). Right and left CST crossed the pontine myelinolysis. (B) Tractography rendering of left and right CST; in sagittal and coronal views. The yellow volume represents the pontine myelinolysis area (3102 mm³). Left and right CST passed within the pontine myelinolysis without warped fibers at this level.

Patient 2

Neurological evolution was marked by the onset of recovery 3 months after the beginning of symptoms, enabling decanulation and the resumption of oral feeding. Limb recovery was considerable between 3 and 6 months, then more gradual, with an MRC score of 46 out of 60 at 6 months and 50 out of 60 at 12 months (motor strength ≥4/5 in key muscles). The patient was discharged home after 9 months hospitalization and continued her rehabilitation in the day hospital. She recovered almost complete autonomy for activities of daily living (FIM: 115/126 at 12 months, then 126/126 at 15 months; Fig. 1, Table 1). Brain MRI associated with DTI, performed 20 days after the onset of symptoms, showed marked CPM, with partial ADC restriction. On tractography rendering, the right and left corticospinal tracts were asymmetric. The volume of the right CST lower than that of the left CST. The latter was divided into two parts (a ventral part that crossed the pontine myelinolysis and a dorsal part pushed back to it; Figs. 4 and 5). Compared to healthy subjects, tensor metrics showed anomalies on CST axonal integrity (FA was lower), CST myelination (RD was higher), and supposed CST edema (MD was higher; Table 2). It is of note that at the pontine myelinolysis level, FA was lower; MD, RD, and AD were higher (Table 3).

Figure 4.

Brain magnetic resonance imagery (MRI) at month 3 for patient 2. (A) 3D T1 TFE: centropontic hyposignal. (B) 3D FLAIR centropontic hypersignal. (C) Axial T1 TSE reconstruction: centropontic hyposignal.

Figure 5.

T2 diffusion sequences with tractography of the corticospinal tracts of the brain magnetic resonance imagery (MRI) at month 3 for patient 2. (A) Orientation distribution function (ODF) map focused on the pontine myelinolysis area. Left corticospinal tract (CST) fibers cross the pontine myelinolysis (ventral part) and are pushed back (dorsal part) explaining the green color direction (anteroposterior direction). Right CST fibers are only seen dorsally to the pontine myelinolysis. (B) Tractography rendering of left and right CST in sagittal and coronal views. The yellow volume represents the pontine myelinolysis area (3405 mm³). Left and right CST are asymmetric. Pontine myelinolysis dissected the left CST in two parts: the ventral part of the left CST passes within the pontine myelinolysis and fibers are repelled laterally, while the dorsal part of the left CST is pushed back behind the pontine myelinolysis. The right CST has 6 times fewer fibers than the left CST and no fibers are tracked within the pontine myelinolysis: they are only pushed back dorsally.

Discussion

The two cases, with initial very severe neurological impairment, had a near‐complete functional recovery at 12 months, which is important to consider given the ethical issues at stake and the discussions about limiting care that may take place in the intensive care unit. Longitudinal follow‐up with regular assessments over a 12‐month period interestingly showed that motor recovery occurred for both patients between 3 and 6 months after the onset of symptoms. This raises the question of the physiological mechanism underlying motor recovery. Importantly, CPM is histologically characterized by a symmetric loss of the oligodendrocyte myelin and a relative preservation of the underlying axons. ⁶ , ³⁷ The results of tractographic analysis in the two patients presented herein support this hypothesis as corticospinal tracts were identified inside the myelinolysis demonstrating a partial preservation of the corticospinal fibers. It should be noted that the tensor metrics of CST were abnormal; in both patients, FA was lower than in healthy subjects, suggesting an alteration of CST microstructure; RD, AD, and MD were higher, and greater at the pontine myelinolysis level, in favor of a demyelination (RD was higher) and vasogenic edema (MD was higher). ³⁴ , ³⁸

Interestingly, the recovery process in acute motor axonal neuropathy (AMAN) has some similarity with the cases presented herein. For example, motor recovery takes place for many cases during the first year after onset of treatments and has been attributed to the development of collateral pathways or the resolution of conduction block at the nodes of Ranvier. ³⁹ , ⁴⁰ Hence, the suppression of a transient conduction block analogous to neuropraxia in the peripheral nervous system could explain the improvement of the neural traffic within the myelinolysis area. ⁴¹ Furthermore, by analogy with multiple sclerosis, the clear T1 hyposignal on brain MRI observed in both patients would have suggested that the lesion also concerns axons. ⁴² , ⁴³ , ⁴⁴ However, the profile of tensor metrics in both patients (FA was lower than in healthy subjects, and MD, RD, and AD were higher) may also be coherent with the consequences on CST axons of a severe focal edema at the pontine level, whose morphometric correlate was the paradoxical AD increase. Indeed, an AD higher than in healthy subjects has also been attributed to the greater extracellular water content consecutively to the atrophy of the white‐matter fibers and to alterations of axonal water content. ³⁸ The functional correlate of this axonal impairment would be, in this hypothesis, the conduction block observed with the neurophysiological evaluation. It would therefore be interesting to compare, in a future study, tractography images from the patients reported herein to those of a patient with an initial state of LIS secondary to a mechanism other than myelinolysis (such as a stroke or a traumatic lesion) and whose clinical course is different.

The results of the present study also suggest that MRI comprising DTI and tractography could play a key role in the prognosis of motor recovery in the initial phase after myelinolysis, ⁴⁵ in particular when neurophysiological evaluation suggests a very severe and potentially irreversible functional impairment at the central level, related to the severity of the demyelination process. ⁴⁶ The information obtained by DTI could temper the initial pejorative neuro‐prognostication based on standard MRI and evoked potentials by indicating that: (i) the severity of lesion visible on morphological imaging is not a proof of CST interruption; (ii) there are alternative pathological explanations for SEPs and MEPs other than axonal damage (e.g., conduction block related to axonal consequences of focal edema and demyelination assessed by the increased AD); (iii) the profile of tensor metrics (FA was lower compared to heathy subjects and MD, RD, AD higher, and at least partial preservation of the CST tract) suggests that there is sufficient axonal fibers to allow motor recovery. Therefore, the technical issues raised by an incorrect interpretation of tests beyond their initial context of validity (stroke, anoxia, traumatic brain injury in this case) should be taken into account before any ethical discussion.

To conclude, the cases reported herein highlight the possibility of a complete recovery after CPM with initial extreme disability such as disorder of consciousness and LIS that seems to be the consequence of reversible myelin damage combined with partially preserved neurons. The development of collateral pathways or the resolution of conduction block may explain this recovery together with intrinsic and extrinsic favorable factors such as the patient's age and long‐term multi‐disciplinary rehabilitation.

Funding Information

None.

Author Contributions

MC and JL contributed to study design and interpreted the data. MC, JL, CD, JJ, and FG drafted the manuscript. MC, JL, SC, CD, FC, SB, MBG, FG, TJ, MDQ, DM, JJ, and EV were involved in acquisition of data. JJ and NAO performed and interpreted the neurophysiological evaluation. CD and FC performed and analyzed the radiological assessment. All authors reviewed and revised the manuscript.

Conflict of Interest

No conflict of interest has been declared by any of the authors.

Supporting information

Video S1.