HIV-Associated Vacuolar Encephalomyelopathy

Abstract

We report a case of HIV-associated vacuolar encephalomyelopathy with progressive central nervous system dysfunction and corresponding vacuolar degeneration of the spinal cord, cranial nerves, and brain, the anatomic extent of which has not previously been described. Vacuolar myelopathy classically presents as a spinal syndrome with progressive, painless gait disturbance in the setting of advanced HIV and AIDS. Vacuolar involvement of the brain and cranial nerves, as illustrated in this case report, is a newly described variant of this condition that we term vacuolar encephalomyelopathy.

CASE REPORT

A 33-year-old man, recently diagnosed with HIV, was hospitalized first in June 2018 with dizziness, hearing loss, and tinnitus. He was diagnosed with vestibulocochlear neuritis that was managed symptomatically with clonazepam. His initial CD4 cell count and HIV viral load were 34 cells/mm³ and 14 000 copies, respectively. He was started on antiretroviral therapy (ART) with dolutegravir (DTG) and emtricitabine/tenofovir alafenamide (FTC/TAF), as well as trimethoprim/sulfamethoxazole and azithromycin for opportunistic infection prophylaxis. He was hospitalized 3 weeks later with multiple progressive neurologic symptoms including lower extremity weakness, ataxia, paresthesias, dysphagia, and decreased visual acuity.

The patient was from Virginia and had no recent travel. Medical history was notable only for hypertension managed with lifestyle modifications. He did not use tobacco products, alcohol, or illicit drugs and was prescribed no medications other than ART. Family history was notable for his mother having multiple myeloma. The patient was first diagnosed with HIV at a free clinic in May 2018 and disclosed prior sexual encounters with men.

Cerebrospinal fluid (CSF) analyses demonstrated 0 white blood cells/mm³, glucose 72 mg/dL, and protein 140 mg/dL. CSF viral polymerase chain reaction (PCR) testing was negative, including herpes simplex virus (HSV), cytomegalovirus (CMV), Epstein-Barr virus (EBV), varicella zoster virus (VZV), and JC virus. Serum and CSF cryptococcal antigen testing and Bartonella serologies were negative. CSF antibodies to aquaporin-4 and oligoclonal immunoglobulin bands were not observed. Serum and CSF Lyme and syphilis serologies were negative. B12 and folate levels were normal. After 1 month of ART, repeat HIV viral load was undetectable and CD4 count was 26. Magnetic resonance imaging (MRI) of the central neuraxis demonstrated prominent lateral cervical and thoracic cord signal abnormalities extending from C2-C7 and T1-T10. There was also a single punctate focus of signal abnormality noted in the right thalamus.

Inflammatory myelopathy such as neuromyelitis optica, vs a vacuolar myelopathy, was suspected, along with concomitant immune reconstitution inflammatory syndrome (IRIS). The patient was managed with steroids, multiple plasma exchanges, and intensification of ART with the addition of maraviroc for putative immunomodulatory effect. Despite these interventions, the patient’s neurologic deficits progressed rapidly over the next 2 weeks, with the development of quadriplegia, dysarthria, and blindness.

MRI repeated 12 days later showed new more conspicuous hyperintensities involving the bilateral posterior thalami (Figure 1), superior cerebellar peduncles/vermis, dorsal white matter tracts of the brainstem, and possible diffuse enhancement of the optic nerve sheaths. Increased signal abnormalities were also seen in the cervical and thoracic cord, with dorsal nerve root enhancement. The patient later developed fever and hypoxemia after a suspected aspiration event, at which time aggressive care was withdrawn. The patient died on hospital day 28.

Figure 1.

Compared with previous magnetic resonance imaging of the brain (A) 12 days prior, there were new (B) symmetric, punctate foci of diffusion restriction (circled) within the bilateral posterolateral thalami, with corresponding fluid-attenuated inversion recovery hyperintensity (shown) without associated edema.

A brain and body autopsy was performed. The brain and spinal cord were grossly unremarkable other than demonstrating mild cerebral edema. Microscopic examination of the brain and spine showed discrete and confluent white matter intramyelinic vacuoles accompanied by macrophages loaded with myelin debris (Figure 2). The vacuolar changes were most severe in the spinal cord, particularly the upper/mid thoracic and cervical levels, where the lateral and ventral corticospinal tracts were involved; there was relative sparing of the posterior columns, particularly the gracile fasciculus. The brain stem, corticospinal tract, and cerebellar peduncles were also involved. The optic chiasm and optic pathways showed similar vacuolar changes without inflammatory infiltrates. In the cerebrum, vacuolar changes were prominent in the bilateral thalami, with no other significant areas of involvement. A significant observation of the lesions was the lack of lymphoplasmacytic infiltrates, microglial nodules, and/or multinucleated giant cells, ruling out the possibility of other pathologic processes including HIV-associated leukoencephalopathy, acute demyelinating process like neuromyelitis optic, or infectious processes. Immunohistochemical stains for HIV p24 antigen performed in several areas of the brain and spinal cord were negative. The remainder of the autopsy was remarkable for oral thrush, esophageal ulcers due to HSV1 infection, and extensive pseudomembranous colitis with Gram-positive bacilli consistent with Clostridium difficile infection. The lungs showed bilateral bronchopneumonia with abscess formation in the right lower lobe, without demonstrable microorganisms.

Figure 2.

Representative sections of the brain and spinal cord. The thoracic levels of the spinal cord were the most affected by the vacuolar changes, with extensive involvement of lateral corticospinal tract (A) and portions of the posterior column (A). The vacuolar changes, highlighted by myelin stain (Luxol Fast Blue [LFB]) in (B), were accompanied by an accumulation of macrophages containing myelin (C), as demonstrated by CD68 immunostain. Similar vacuolar changes were present in the brain stem (superior cerebellar peduncle) (D) and the bilateral thalami (E), with macrophages containing myelin debris (LFB) (F). The extension of macrophagic infiltration can be appreciated by CD68 immunostain in this thoracic level of the spinal cord (G), with the entire posterior, lateral, and ventral tracts involved. The optic chiasm and optic tracts showed similar vacuolar changes (H).

DISCUSSION

Vacuolar myelopathy is described as a syndrome that primarily affects the spinal cord. Weakness and sensory deficits are typically described in the lower extremities, usually without an associated sensory level [1]. Unfortunately, there is no reliable way to confirm vacuolar myelopathy premortem, and it remains a diagnosis of exclusion.

Although the exact etiology of vacuolar myelopathy is unknown, the association between HIV infection and vacuole formation is thought to be indirect, perhaps mediated by a metabolic derangement. As pathologic findings resemble those of subacute combined degeneration of the spinal cord (typically associated with B12 deficiency), the B12-dependent transmethylation pathway is thought to be important, possibly mediated by macrophage activation leading to a local deficit in methyl group donors [2]. Vacuoles develop from areas of focal swelling of the myelin sheath, and on histologic examination of the spinal cord, 10–100-micron-sized vacuoles are seen containing cellular debris or macrophages. The posterior and/or lateral columns are principally affected [1].

CSF findings may be normal; however, protein elevation and/or lymphocytic pleocytosis may also be seen [1]. Measurement of the CSF viral load would have been reasonable to check for CSF “escape” despite a rapid virologic response to antiretroviral therapy. Unfortunately, we did not have a large enough CSF sample to send this test. Although CSF viral replication has been associated with vacuolar myelopathy, its pathogenic role in the condition has not been characterized [3].

MRI of the spinal cord is usually normal; in some cases, nonspecific hyperintensity of the spinal cord can be seen [4]. CMV polyradiculomyelitis may mimic the clinical presentation of vacuolar myelopathy. CMV polyradiculomyelitis is usually distinguishable by the absence of spasticity on exam and MRI findings of diffuse, multisegmental signal abnormalities that are seen in both gray and white matter, with thickening/enhancement of nerve roots [5].

Optic neuropathy was reported in association with vacuolar myelopathy in an immunosuppressed HIV-negative liver transplant recipient [6]. However, HIV/AIDS-related vacuolar myelopathy has not been reported to be associated with optic neuropathy, and in this case it prompted initial diagnostic testing for a neuromyelitis optica spectrum disorder and a trial of plasma exchange. To our knowledge, this is the first case of vacuolar encephalomyelopathy associated with HIV.

This case describes paradoxical worsening of the neurological condition after initiation of antiretroviral therapy. IRIS has been described in the context of vacuolar myelopathy [7].

Given the widespread use of ART without CD4 count restriction, later-stage manifestations of AIDS such vacuolar myelopathy have become less common but are all the more necessary to describe their variable presentations. Myelopathy is thought to be under-recognized in AIDS patients, often ascribed to debilitation or wasting, and is present on autopsy in 23%–47% of patients with AIDS [8–10].

A large case series of 215 AIDS autopsies identified vacuolar myelopathy in 100 (47%) patients but did not identify brain involvement in any of those cases apart from multinucleated giant cells [9]. Although the pathologic prevalence of vacuolar myelopathy in AIDS appears to be quite common, the clinical prevalence of vacuolar myelopathy is probably much lower. For example, among patients with vacuolar myelopathy on pathology, Pan et al. found symptoms of myelopathy in only 2 out of 56 patients with documented neurologic evaluations. We postulate that clinical vacuolar myelopathy cases are likely underrepresented in the literature, and the literature may not have captured the less common occurrence of vacuolar encephalomyelopathy.

There are no proven effective treatment options for vacuolar myelopathy apart from initiation of ART in antiretroviral-naïve patients. Neither scheduled L-methionine supplementation as tested in a clinical trial design nor administration of IVIG has been shown to have benefit [11, 12]. Plasma exchange has not been systematically studied in the context of vacuolar myelopathy.