Bilateral pallidal hemorrhage in toxoplasmosis update of acute symmetric lesions of deep nuclei

Abstract

As acute symmetric lesions of deep gray nuclei are often associated with an impaired level of consciousness and neuroimaging by itself cannot distinguish between etiologies, diagnosis may be problematic. Appreciation of the cause of the various neuroimaging patterns in conjunction with the history, examination and laboratory investigations allows for accurate diagnosis in the vast majority of cases. Given the metabolic vulnerability of deep gray nuclei, other than bi-thalamic infarction, it follows that toxic-metabolic and hypoxic-ischemic events account for the majority of cases. Nevertheless, the differential diagnosis is broad and diverse. We here describe two cases of bilateral pallidal hemorrhage in AIDS-associated toxoplasmosis, and review conditions recently described with acute symmetric deep gray nuclei lesions on neuroimaging.

Introduction

The neuroimaging patterns of symmetric lesions of deep gray nuclei (DGN) although not specific are highly suggestive of the various etiologic conditions.^1,2 As such, familiarity with the causes associated with the different patterns is helpful in determining diagnosis. In the context of two patients with bilateral pallidal hemorrhage in association with AIDS-related toxoplasmosis, we highlight the neuroimaging features of recent reports helpful in the diagnosis of acute symmetric lesions of DGN.

Case histories

Case 1

A 29-year-old man from Guatemala, in the US for 7 years, presented with a 5 day history of progressive frontal headache and hiccoughs, with occasional vomiting, and a 2 day history of veering to one side on walking. He denied any illicit drug use or high risk sexual activity. Neurologic examination showed the patient to be lethargic but arousable to answer questions and was otherwise non-focal. Computed tomography (CT) and magnetic resonance imaging (MRI) showed bilateral globus pallidus lesions (Figure 1(a) to (d)), and a lesion in the left parieto-occipital area. The attenuation coefficient of the CT lesion measured 50 HU. Other than mild anemia, routine laboratory studies were unremarkable. Screening HIV antigen/antibody assay was positive and RNA viral load was 148,045 (N < 40), CD4 cell count 9 (N = 535–1451) and toxoplasmosis antibody IgG was greater than 900 (N < 0.9). Cerebral toxoplasmosis in the setting of HIV/AIDS was diagnosed, and treatment with sulfadiazine 1500 mg four times daily, pyrimethamine 75 mg daily and leucovorine 25 mg daily was begun. The patient’s condition gradually and progressively improved on the above medication regimen over the next 10 days, and he was discharged home to be followed in the outpatient clinic. Repeat CT imaging after 4 months showed complete resolution of the increased pallidal attenuation and decreased on MRI gradient echo (Figure 1(e) and (f)).

Figure 1.

Computed tomography with increased attenuation in globus pallidus bilaterally (arrows) and hypodensity of left parieto-occipital region (a); magnetic resonance (MR) fluid-attenuated inversion recovery with increased signal of the pallidum and left parieto-occipital region (b); MR gradient echo with hemosiderin deposition in pallidum (arrows) (c); MR T1-weighted gadolinium enhanced with target sign on left (arrow) (d); CT 4 months post-treatment shows resolution of increased attenuation (e); MR gradient echo sequence shows reduced but persistent signal (f).

Case 2

A 31-year-old man with a history of intravenous drug use and a history of HIV infection for 3 years, non-compliant with medications, presented with a one week history of change in mental status. His mother noted that he had complained of frontal headache for past 3 days with vomiting for one day prior to admission. Neurologic examination showed the patient to be lethargic and slow to respond to questions. There was a left facial weakness, 3/5 left-sided motor weakness with brisk deep tendon reflexes on the left, with a positive Romberg sign. CT and MRI showed bilateral globus pallidus lesions (Figure 2(a) to (d)). The attenuation coefficient of the CT lesion measured 55 HU. The patient was known to be hepatitis C positive with a CD4 cell count of 18 and a viral load of 340,000. A brain thallium single-photon emission computed tomography scan did not show any significant uptake consistent with the clinical and imaging suspicion of toxoplasmosis. He was treated with sulfadiazine but was switched to clindamycin because of renal failure. In addition, pyrimethamine and leucovorin were added. He showed continuous clinical improvement and repeat MRI after 3 weeks showed a decrease in lesion size and signal intensity on fluid-attenuated inversion recovery (FLAIR) with increased hemorrhage on gradient echo sequence (Figure 2(e) and (f)). After a one month hospitalization the patient was discharged home to be cared for by his family.

Figure 2.

Computed tomography with increased attenuation in globus pallidus bilaterally (arrows) (a); magnetic resonance (MR) fluid-attenuated inversion recovery (FLAIR) with increased signal of the pallidum (b); MR gradient echo with hemosiderin deposition in pallidum (arrows) (c); MR T1-weighted gadolinium enhanced with target sign on left (arrow) (d); Gradient echo sequence 3 weeks post treatment shows reduced signal and progression of hemorrhage (f).

Discussion

The development of symmetric DGN lesions most often relates to the vulnerability of these structures, notably to their high glucose and oxygen utilization seen on functional neuroimaging. As seen with the two patients presented here, etiologies other than hypoxic-ischemic or toxic-metabolic causes may uncommonly manifest with acute symmetric lesions of DGN (Figure 3).

Figure 3.

Acute Deep Gray Nuclei Lesions.

Distinct from unilateral hemorrhage, bilateral hemorrhage of DGN is rare and the etiology varies with the nuclei involved. The putamen and the thalamus are most commonly related to hypertension,^3,4 with isolated case reports of ‘basal ganglia’ hemorrhage due to mucormycosis⁵ and of uncertain etiology for the cuadate nucleus.⁶ Bilateral pallidal hemorrhage on the other hand is almost always caused by hypoxic-ischemia and/or toxic-metabolic causes related to intoxication with carbon monoxide, cyanide, methanol or drug overdose, uremia, hypoglycemia or primary cardiac event with hemorrhage into damaged tissue.^1,2 Similar risk factors may also be associated with hemorrhage into other nuclei as seen with putamenal hemorrhage with methanol ingestion.⁷

Hemorrhage with cerebral toxoplasmosis is rare, may involve the cortical gray, basal ganglia, brainstem or cerebellum and usually occurs after the onset of treatment.^8,9 Symmetric hemorrhage of the DGN in toxoplasmosis with AIDS is limited to a recent neuropathological case report.¹⁰ The neuroimaging correlate of this finding is here described in two patients, one of whom heralded the disease. The occurrence of bilateral basal ganglia lesions in the absence of known associated risk factors is rare. Nevertheless, the neuroimaging findings of our patients are best appreciated considering toxoplasmosis is the most common cause of a mass lesion in AIDS patients, has a propensity for the basal ganglia and has been associated with lesion hemorrhage. Diffuse lymphocytic infiltrate that accompanies toxoplasma infection may compromise vessel wall integrity and predispose to hemorrhage into the infected area.¹⁰

Other causes of hemorrhage to consider in an AIDS patient include varicella-zoster encephalitis, herpes simplex encephalitis, cytomegalovirus infection, cerebral infarction with non-bacterial endocarditis, vasculitis, mycotic aneurysm, hemophilia, thrombocytopenia, hemorrhagic metastatic disease (Kaposi’s sarcoma) and cerebral lymphoma.^8,9,11 In addition to there being no evidence of associated predisposing conditions, both our patients were seropositive for toxoplasmosis, manifest characteristic imaging features including target sign on gadolinium enhanced MRI, diagnostics of the disease¹² (Figures 1(d) and 2(d)) and responded clinically to anti-toxoplasma therapy. Calcification may be seen in congenital toxoplasmosis and may also be seen following treatment of the acquired form with the calcification exhibiting a dot-like or dystrophic pattern in this setting. Both of our patients initially presented with areas of high attenuation on CT within the globus pallidus prior to the initiation of anti-toxoplasma therapy strongly favoring symmetric hemorrhage based on their clinical history. Furthermore, the degree of edema surrounding the areas of increased attenuation, the CT Hounsfield unit values, the degree of MRI blooming artifact on the gradient recalled echo series, and the evolution of these findings over 3 week’s time (case 2) after initiating therapy all support the presence of evolving hemorrhage.^8,9 Most convincing was complete resolution of the increased pallidal attenuation on CT imaging with reduced but persistent signal on gradient echo (case 1) after 4 months. A patient with bilateral symmetric cerebellar hemorrhage on CT is the only other imaging report of a symmetric intracranial hemorrhage in a similar clinical setting.¹¹ Bhagavati et al.⁹ reported a 50% frequency of hemorrhagic lesions at initial toxoplasmosis presentation, suggesting that hemorrhage may be more common than previously reported. This conclusion awaits confirmation from other studies given the absence of CT evidence of hemorrhage in four of the seven reported cases, the lack of pathologic correlation, and the reliance on T1-weighted imaging to demonstrate hemorrhage without a gradient series; calcium and subacute methemoglobin may exhibit increased signal on T1-weighted imaging. In future cases susceptibility weighted imaging can be used to distinguish between the presence of blood and calcium. Susceptibility weighted imaging is a MRI sequence that is sensitive to compounds that locally distort the magnetic field. Paramagnetic blood products and diamagnetic calcium distort the magnetic field in opposite ways, allowing them to be distinguished on post-processed phase maps and more accurately determine the true incidence of hemorrhage with cerebral toxoplasmosis.¹³

Symmetric pallidal hemorrhage in the absence of other risk factors should raise concerns for toxoplasmosis given its predilection for the basal ganglia and association with lesion hemorrhage, and initiate a search for an underlying immunodeficiency, to include HIV.

In addition to toxoplasmosis, recent reports of conditions involving DGN include eastern equine encephalitis, acute hepatic encephalopathy with hyperammonemia and diabetes-dialysis syndrome.^14–16 The first two mainly show restricted diffusion of the thalami, while the third involves the lenticular nuclei with predominantly FLAIR MRI changes.

In the appropriate clinical setting and geographic location restricted diffusion of the thalami can be a helpful in the diagnosis of encephalitis, particularly eastern equine encephalitis.¹² Other acute viral encephalitides in addition to eastern equine encephalitis can affect the DGN and include Epstein–Barr virus, influenza, West Nile virus and Murray Valley virus¹⁴ (Figures 3 and 4). Restricted diffusion changes have been described with acute hepatic encephalopathy associated with elevated ammonia that characteristically involves the insular and cingulate cortices and thalami bilaterally.¹⁵ Also, diabetes-dialysis syndrome, that presents clinically with headache, dysarthria and gait and movement disorder, and occurs predominantly in patients of Asian descent in association with metabolic acidosis, manifests with reversible FLAIR changes of the lenticular nuclei.¹⁶ Other infectious, toxic-metabolic and hypoxic-ischemic conditions including malaria, ethylene glycol, heat stroke and tripanosomiasis may show acute change of DGN (Figure 3).

Figure 4.

Restricted-Diffusion of Deep Gray with/and without Cortical Involvement.

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Conflict of interest

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.