Toxic leukoencephalopathy versus delayed post-hypoxic leukoencephalopathy after oral morphine sulphate overdose

Abstract

Toxic leukoencephalopathy (TLE) is a rare pathology caused by various substances including opioids (notably heroin), immunosuppressants, chemotherapy agents, cocaine, alcohol and carbon monoxide. However, although heroin is metabolised by the body into morphine, there is a striking paucity in cases of primary oral morphine-induced TLE, especially in the adult population. We present the case of a man in his 40s admitted to hospital in respiratory depression with a Glasgow Coma Scale (GCS) score of 6 after taking an overdose of oral morphine sulphate. Following a complete recovery to baseline, he was then readmitted with an acute deterioration in his neurobehavioural condition. Initial investigations returned normal but MRI showed changes characteristic for TLE.

In cases of opioid toxicity such as ours, TLE is difficult to differentiate from delayed post-hypoxic leukoencephalopathy, due to their similar clinical presentation, disease progression and radiological manifestation. We explore how clinicians can approach this diagnostic uncertainty.

Background

Toxic leukoencephalopathy (TLE) is a demyelinating disease principally of the periventricular white matter (PVWM) occurring in response to exposure to a toxin.¹ The exact pathophysiology remains uncertain, especially given the associated hypoxia often involved with these cases and the damage this can cause independently of the toxin itself. Observed changes include the destruction of oligodendrocytes, microvascular damage and resultant demyelination of the white matter.² Substances may be directly toxic to oligodendrocytes, as has been demonstrated for the metabolite of heroin (morphine-3-glucuronide), enhancing the permeability of the blood-brain barrier and causing cerebral oedema and focal ischaemia.^{3 4}

TLE may present immediately following toxic insult with characteristic neurological and radiological symptoms or, more rarely, after a 1–3 week period of lucidity.⁵

TLE is a rare condition and diagnosis is based on a triad of features:

• Exposure to a toxin such as chemotherapy agents (most notably fludarabine and methotrexate), heroin (inhaled or injected), other opioid medications such as oxycodone, immunosuppressants (including ciclosporin and tacrolimus), cocaine (snorted or inhaled), alcohol and carbon monoxide.¹

• Characteristic neurobehavioural changes such as altered mental state, psychomotor changes, depression, anxiety, visuospatial deficit and coma.⁶

• Hallmark radiological findings on CT and MRI.

The principal MRI findings for TLE are bilateral, symmetric and confluent PVWM changes, representing cytotoxic oedema. These may be visible as hypodense lesions on CT images but are most apparent as hyperintense lesions on MRI. Findings of herniation and cerebellar oedema are indicators for poorer outcomes.⁵ Heroin-inhalation (often referred to as ‘chasing the dragon’) induced spongiform TLE is well reported in the literature and can be seen as a distinct entity.

Specific changes seen consist of T2 hyperintensities in various deeper white matter structures including posterior limbs of the internal capsule (with sparing of the anterior limb), the cerebellar peduncles and splenium of the corpus callosum. There may also be symmetric involvement of the corticospinal tract, medial lemniscus and tractus solitarus. Diffusion-weighted imaging (DWI) may show restricted diffusion in the posterior limb of the internal capsule, temporal lobes and globus pallidus.^{2 7 8} Additional toxin-specific changes described in the literature, however, are based only on case report level of evidence and should be viewed in this context. Observed changes include sparing of the cortical U-fibres and involvement of both cerebellar grey matter, basal ganglia, hippocampi and central tegmental tracts in methadone-induced TLE.¹ Oxycodone-induced TLE involves the cerebellum and globus pallidus, while fentanyl affects the deep white matter in both cerebral and cerebellar regions.¹ It should be noted that the number of oxycodone and fentanyl-related cases of TLE described in the literature is smaller than for methadone and heroin, and findings should be viewed in this context.¹ Although ingested morphine sulphate-induced TLE has been described in children,^{9 10} we were unable to find any examples of it in the adult population described in the literature.

Following resuscitation, emergency management aims to minimise risk factors for secondary brain injury. This includes reducing raised intracranial pressure, maintaining cerebral perfusion and managing additional risk factors, such as alcohol excess and thiamine deficiency. Patients may require respiratory support or use of vasopressors in intensive care and suboccipital craniectomy may be required for cerebellar oedema with herniation or hydrocephalus. The use of steroids has also been reported but they are not established as a routine part of management. Patients may require extensive neurorehabilitation, with huge variability in prognosis, varying from complete return to prior neurological function to severe brain damage and death.⁶

Case presentation

A man in his 40s presented to the emergency department of a small District General hospital with a 3-day history of increasing confusion and agitation. That morning, he was found by his partner urinating and pouring sugar on the kitchen countertops and aggressively slamming doors in their home. He also complained of a severe headache and tinnitus in the left ear. Due to his agitation, only a limited examination was possible, but he was found to have globally increased tone, brisk reflexes and clonus.

He had a 15-year history of chronic back pain due to various spinal pathologies, including intervertebral disc prolapse, cervical myelopathy and spinal stenosis, for which he had undergone three operations. His General Practioner (GP) had been prescribing tramadol to manage the pain for 15 years, but this was escalated to modified-release morphine sulphate with morphine for breakthrough pain around 6 months before his admission.

The patient had been discharged from hospital 3 weeks prior to his readmission following an intentional morphine overdose. Paramedics had found 60 empty packs of modified release morphine and two empty bottles of 10 mg/5 mL morphine solution (200 mL in total). His oxygen saturation was 70% (increasing to 98% with 15 L oxygen through a non-rebreathe mask), respiratory rate 16 rpm, blood pressure 127/86 mmHg, heart rate 118 bpm and he was apyrexial. They gave him six doses of 0.4 mg of naloxone with minimal response. He was admitted to intensive care for three nights for intubation, ventilation and vasopressor support. He scored 6 on the Glasgow Coma Scale (GCS) and there was evidence of acute kidney injury, toxic liver insult and possible hypoxic brain injury. A CT-head was attempted but unfortunately, due to the patient being agitated, the images were significantly distorted by movement artefact and cannot be included in this report. Over the next 3 days, he recovered sufficiently to be extubated and moved to a ward for further treatment. He was discharged when back to his baseline 5 days later.

Investigations

The cause of his second admission was investigated extensively. Blood tests including liver function tests, full blood count, renal profile, C reactive protein, vitamin B12, ammonia, thyroid function tests, magnesium, coagulation profile, glucose and arterial blood gases were all returned as satisfactory. A urine dipstick returned completely normal while a CT-head found no evidence of acute intracranial abnormality. A lumbar puncture yielded clear and colourless cerebrospinal fluid (CSF) with glucose and protein levels within range. CSF meningitis and encephalitis panels were negative and no oligoclonal bands were seen.

Finally, an MRI-head was performed, 27 days after the initial overdose. Symmetrical white matter hyperintensities were seen bilaterally, involving both cortical and subcortical PVWM of both cerebral hemispheres (figure 1). Areas of restriction were seen on the DWI sequence and apparent diffusion coefficient (ADC) map (figure 2). High signal intensity was seen in the external capsule across all sequences (figures 1 and 2). These findings combined with the clinical presentation are highly suspicious for TLE.

Figure 1.

First MRI brain (27 days after overdose). T2 sequence with blade (above), fluid attenuated inversion recovery (FLAIR) sequence (below). (A) Symmetrical bilateral subcortical hyperintensities; (B) high intensity within the external capsule and (C) symmetrical bilateral cortical hyperintensities.

Figure 2.

First MRI brain (27 days after overdose); DWI sequence (above), apparent diffusion coefficient (ADC) mapping (below); (A) restricted diffusion in the external capsule; (B) example areas of restricted diffusion. DWI, diffusion-weighted imaging.

Treatment

He was managed conservatively while an inpatient with intravenous fluids, IV Pabrinex (high potency solution containing vitamins B1,2,6, nicotinamide, vitamin C and glucose), analgesics and anti-anxiolytics. He was seen by the mental health team who suggested he be started on quetiapine, titrated up to 50 mg ON. A specialist neurological opinion was sought, which agreed with the diagnosis of TLE and suggested a prescription of coenzyme Q10 along with vitamins C and E.

After 1 week, the patient’s wife began noticing improvement in his condition as he was able to recognise her again. A week later, he was reviewed again by a neurologist who noted he had improved but that he was still encephalopathic. His tone and clonus had improved but reflexes remained brisk.

Outcome and follow-up

A repeat MRI-head performed 6 weeks after the first found improvement in the focal T2 and fluid attenuated inversion recovery (FLAIR) white matter hyperintensities but also found increased diffuse high signal within the white matter throughout the supratentorial brain compared with previously (figures 3 and 4). New areas of high signal intensity were seen in the globus pallidus bilaterally (figure 3).

Figure 3.

Second MRI brain (2 months and 4 days after overdose); T2 sequence with BLADE (above), fluid attenuated inversion recovery (FLAIR) sequence (below); (A) new high signal intensity at the globus pallidus bilaterally; (B) example areas of reduced focal points of high signal intensity but increased diffuse high signal intensity.

Figure 4.

Second MRI brain (2 months and 4 days after overdose). DWI sequence (above), apparent diffusion coefficient (ADC) mapping (below); (A) diffuse areas of restricted diffusion. DWI, diffusion-weighted imaging.

Concurrently, he was showing signs of good cognitive improvement, including becoming oriented to time and place, and aware of his diagnosis and disability. However, he still showed diurnal variation in his ability to manage his mood and respond to environmental challenges. He remained unaware of the events leading up to his suicide attempt. It was decided that he would benefit from a transfer to a neurobehavioural unit under neuropsychiatric care given the risk of further suicide attempt and his significant ongoing needs. Once there, he underwent talking therapy, compassion-focused therapy and the use of mindfulness strategies, improving his mood and enabling him to combat residual feelings of guilt following his overdose. He underwent psychometric assessment, which evidenced weakness in non-verbal memory, novel problem-solving skills and reduced processing speed. However, it was felt that these ongoing deficits would not impact severely on his daily life, and he has since been able to return home and to work.

Discussion

Various diseases mimic TLE clinically and/or radiologically and this diagnostic uncertainty is highlighted by our case. The plethora of normal laboratory results and lack of any signs of infection or other systemic disease means many of these can be discounted, including:

• Septic encephalopathy—no clinical signs of sepsis.¹¹

• Acute disseminated encephalomyelitis—no prodromal flu-like symptoms and normal inflammatory markers.^{12 13}

• Acute haemorrhagic leukoencephalitis—no prodromal infection symptoms or recent vaccination and normal inflammatory markers.¹⁴

• Viral encephalitis—no signs of infection and normal CSF analysis.

• Progressive multifocal leukoencephalopathy—no immunocompromise.¹⁵

• Acute hepatic leukoencephalopathy—no history of liver disease or alcohol misuse, normal liver function tests.¹

• Extrapontine myelinolysis—normal electrolyte levels.¹⁶

Posterior-reversible encephalopathy (PRES) shares a similar clinical picture to TLE; however, it can be distinguished easily on MRI. In TLE, the most noticeable changes are seen in the PVWM whereas in PRES they are seen in the cortical/subcortical white matter (most commonly in the parietal and occipital distribution). Also, abnormalities on DWI are seen in TLE (figures 2 and 4) but not usually in PRES.¹

Given the history of our patient presenting with a reduced level of consciousness (GCS 6) and respiratory depression following his morphine overdose, there was likely a degree of cerebral hypoxia. A delayed post-hypoxic leukoencephalopathy (DPHL) is therefore a reasonable differential diagnosis. DPHL is a rare complication following hypoxic brain injury, most commonly seen after carbon monoxide poisoning¹⁷ but also in the context of respiratory failure caused by opioid overdose.^18–21 The disease progression is characterised by an acute hypoxic episode followed by a ‘lucid interval’—a period of relative clinical stability, until a representation with new neurological symptoms 1–4 weeks after the initial event.²² Patients typically display one of two groups of symptoms: those presenting with parkinsonism (with associated symptoms of agitation, hallucinations, dystonic posturing and apathy) or akinetic-mutism (with associated symptoms of apathy, bowel and bladder incontinence, minimal primitive responses to pain and pathological laughter or crying).²³ On MRI, the predominant finding is extensive bilateral T2 and FLAIR hyperintensity in both the cortical and subcortical PVWM and centrum semiovale (dorsal frontal and parietal lobes), representing demyelination with axonal preservation.²² The exact pathophysiological mechanism underpinning DPHL is not understood. It may be that the toxin causes direct insult to myelin, due to deficiency in the enzymes controlling myelin turnover (notably arylsulfatase-A) or alternatively due to delayed post-hypoxic apoptosis of oligodendrocytes (responsible for myelin production) in the affected white matter.²³ However, DPHL has also been reported as a unilateral variant following mechanical recanalisation therapy for ischaemic stroke.²⁴ This lends greater credence to the hypothesis of delayed apoptosis given the absence of a toxin in this case. Prognosis is inversely related to the age of the patient, with most patients improving within a period of 3–6 months. Some cases have fatal outcomes but mostly patients recover with varying amounts of lasting neurological impairment.²³

The challenge of differentiating TLE from a hypoxic brain injury has been discussed in the literature, notably in a case of oxycodone TLE mistaken for an acute global hypoxic brain injury in a patient presenting with a GCS of 5 after an overdose.²⁵ Our case, however, is distinct given the patient’s period of lucidity and delayed representation, raising the comparison with DPHL.

Typically, TLE presents within a few hours of exposure, termed acute TLE which is certainly not applicable to our case given the delayed presentation. However, there are some cases of methadone-induced TLE occurring 1–3 weeks after an initial recovery and period of lucidity following the incident without any hypoxic involvement.⁵ This lends weight to the idea that delayed onset TLE is a separate entity from DPHL, although it should be emphasised that delayed presentation is more widely associated with DPHL. It has also been posited that a new term—delayed toxic-hypoxic leukoencephalopathy should be used for cases with both opioid overdose and significant hypoxia, as recognition that there may be two separate disease mechanisms present,^{26 27} although it is not yet in general use.

In our case, the clinical presentation could represent either or both conditions. The principal initial MR findings of symmetrical bilateral cortical and subcortical PVWM changes could again be the result of either pathology and the same is true of the high signal intensity seen in the globus pallidus on the second MR.^{5 28} One difference that can point to TLE being a more likely diagnosis is the lack of high intensity seen in the overlying cortical grey matter in our case,^{1 29} however, this is based purely on case-series level evidence. We do not believe that, on the balance of evidence, it is possible for us to definitively agree with either diagnosis in this case. Further research is required to be able to confidently differentiate between the two pathologies and understand how they may overlap. Specifically, research is needed to isolate the pathological processes of both conditions, as well as identify radiological differences to enable diagnosis.

Learning points.

• Toxic leukoencephalopathy (TLE) is a rare condition caused by a variety of substances including opioids, most commonly inhaled heroin.

• TLE can be difficult to differentiate from delayed post-hypoxic leukoencephalopathy (DPHL) due to its similar clinical and radiological presentation, especially in opioid-induced cases due to respiratory depression. The two pathologies may also coexist, further confusing the picture.

• More research is required to help differentiate of TLE and DPHL and how they overlap, specifically looking at the pathological processes and radiological differences.

Ethics statements

Patient consent for publication

Consent obtained directly from patient(s).