A Toddler With New Seizures, Progressive White Matter Lesions, and Multifocal Microhemorrhages

The patient is a 16-month-old male with a diagnosis of acute disseminated encephalomyelitis (ADEM) referred to our PICU for further diagnostic evaluation and management in the setting of progressive neurologic deterioration leading to severe cerebral edema and diffuse brain injury.

The onset of symptoms occurred 3 months before the referral, when the patient presented to a local hospital with a 1 day history of cough, emesis, and new onset seizures. Review of systems was negative for fevers, recent illnesses, witnessed falls, and ingestions. Family history was significant for juvenile idiopathic arthritis (JIA) (father), neurofibromatosis type 1 (paternal uncle), and epilepsy (paternal cousins). Neurologic exam at presentation was significant for encephalopathy without focal neurologic deficits.

His initial laboratory evaluation at the referring hospital was significant for enterovirus/rhinovirus respiratory infection, mild thrombocytopenia, and transaminitis (Table 1). A right upper-quadrant ultrasound showed hepatomegaly. Bilirubin, albumin, prothrombin time, international normalized ratio, and partial thromboplastin time were normal. He had a normal EEG. However, brain MRI revealed numerous scattered hyperintense T2 foci throughout the white matter in both cerebral hemispheres and cerebellum, as well as in the deep brain nuclei (Fig 1A). There were microhemorrhages and patchy enhancing abnormality in the subcortical and cerebellar white matter (Fig 1A). A lumbar puncture demonstrated elevated white blood cells (93% lymphocytes) and protein, absent oligoclonal bands, and negative infectious studies (Table 1). On the basis of these findings, he was diagnosed with ADEM.

His neurologic symptoms were initially well controlled with high-dose steroids. However, with the tapering of steroids, he developed progressive decreased use of the left upper and lower extremities, preferential slumping to the left side, drowsiness, and eventual seizure recurrence. He was readmitted for ADEM relapse and received pulse dose steroids and intravenous immunoglobulin. The laboratory investigation on readmission was unremarkable (Table 2). However, repeat MRI was notable for interval progression of multifocal hyperintense T2 signal, with more confluent appearance and increased patchy enhancement abnormality in the subcortical, periventricular, and deep white matter of the cerebral hemispheres, as well as worsening of thalamic/basal ganglia involvement (Fig 1B). Scattered new foci of microhemorrhages in the cerebral and cerebellar hemispheres were also observed. A repeat lumbar puncture was performed, revealing 17 white blood cells (81% lymphocytes), no red blood cells, elevated protein (58 mg/dL), and negative Gram stain and culture. Despite immune modulating therapies, his neurologic exam continued to decline, with onset of new high-grade fevers (39.6°C), cerebral salt wasting, and respiratory failure necessitating transfer to the local PICU. Repeat head computed tomography was significant for diffuse cerebral edema, with follow-up MRI demonstrating diffuse brain injury (Fig 2). He was subsequently transferred to our center for further evaluation and management.

Given the chronicity, previous interventions, progression, and severity of the symptoms, our priority on his arrival to the PICU is to distinguish clinical signs and symptoms associated with the primary illness from those associated with interventions and potential illness-related complications. Are the initial presentation, MRI, cerebrospinal fluid (CSF) findings, and seizures consistent with ADEM?

ADEM is a pediatric demyelinating condition with estimated incidence of 0.23 to 0.4 per 100 000 children.¹  Average age of onset is 3.6 to 7 years; most have reported a preceding viral illness.¹  Common clinical symptoms include encephalopathy, pyramidal signs, cerebellar signs, and cranial nerve deficits. The initial presentation of encephalopathy and seizures, along with lymphocytic pleocytosis in the CSF and predominantly white matter lesions on the MRI, are consistent with ADEM.¹  Although a preceding viral infection is not required for a diagnosis of ADEM (reported in 55%–86% of cases¹ ), a positive finding for enteroviral infection suggests the presence of a viral trigger, which has been reported to be associated with ADEM, as well as other etiologies. However, several features are atypical in this child and warrant mention. Although ADEM occurs predominantly in early childhood, other etiologies in the differential are more frequent in children aged <2 years.²  The presence of transaminitis, hepatomegaly, and thrombocytopenia is uncommon in children with ADEM and suggests a more systemic underlying process. Rhinovirus is not known to be associated with transaminitis or thrombocytopenia. Enterovirus can cause hepatitis but is not associated with thrombocytopenia in an uncomplicated disease.³  Therefore, the presence of transaminitis and thrombocytopenia in this child cannot be attributed simply to these concurrent infections. In addition to ADEM, the differential diagnosis should include infectious meningoencephalitis, other neuroinflammatory conditions such as neuromyelitis optica spectrum disorder, antimyelin oligodendrocyte glycoprotein (MOG) antibody-associated disorder, central nervous system (CNS) vasculitis, and hemophagocytic lymphohistiocytosis (HLH) with CNS manifestations.⁴  The presence of anti-MOG antibody could have further supported ADEM as the primary diagnosis in the presence of atypical signs and symptoms because a significant number of patients with ADEM, especially severe, recurrent, or relapsing ADEM, are positive for anti-MOG antibody.^5,6 

I agree that the initial MRI suggests ADEM as a possibility. However, contrast enhancement is somewhat atypical for ADEM. On the basis of the MRI findings, additional considerations include CNS vasculitis and chronic lymphocytic inflammation with pontine perivascular enhancement responsive to steroids (CLIPPERS).^7–9  Multiple sclerosis is unlikely because of abnormal patchy enhancement in the cerebellum and subcortical supratentorial white matter.⁷  Dr Risen mentioned HLH with CNS manifestations as a potential diagnosis. MRI findings associated with HLH are nonspecific and include diffuse parenchymal volume loss (widening sulci, subarachnoid space, or ventriculomegaly), as well as brain parenchyma tissue involvement ranging from discrete patchy to diffuse confluent lesions.^10–12  Contrast enhancement is common in HLH.^10,11,13,14 

The child initially improved with steroids, which is the expected treatment response for ADEM. Concurrently with the steroid taper, his clinical condition worsened with the development of new neurologic symptoms and fever. His subsequent MRI also exhibits progression of the abnormalities culminating in diffuse cerebral edema and diffuse brain injury. Is relapsing ADEM common? Do the serial neuroradiological findings at this point further support or exclude any of the candidate diagnoses? Have cerebral edema and diffuse brain injury been described as complications of ADEM or other neurologic conditions in the differential diagnosis?

The clinical course of ADEM is most often monophasic, but can be multiphasic, with complete or partial recovery.¹⁵  The International Pediatric Multiple Sclerosis Study Group updated the diagnostic criteria for monophasic and multiphasic ADEM in 2013. Criteria for a diagnosis of monophasic ADEM are the following (Table 3):

1. a first polysymptomatic clinical event, with presumed inflammatory cause;

2. encephalopathy is present;

3. typical brain MRI includes diffuse, poorly demarcated, large (≥1–2 cm) predominantly in the cerebral white matter though deep gray matter lesions (thalamus, basal ganglia) can be present; T1 hypointense white matter lesions are rare; and

4. new or recurrent symptoms or MRI changes within 3 months of the initial presentation are considered a part of the initial monophasic ADEM diagnosis.

Under these new diagnostic criteria, recurrent ADEM is classified under multiphasic ADEM, which is defined as a new event of ADEM 3 or more months after the initial event that can be associated with new or reemergence of previous clinical and MRI findings. Among children with ADEM and positive anti-MOG antibody, monophasic illness is associated with a transient positive anti-MOG serology, whereas relapsing ADEM is associated with the persistence of anti-MOG antibody.^16,17  Although it is currently not a standard laboratory study, given this emerging literature supporting the role of anti-MOG antibody in atypical, severe, or recurrent ADEM, providers should strongly consider serological monitoring for anti-MOG antibody as part of the initial workup in ADEM diagnostic investigation. Reemergence of neurologic symptoms and changes in MRI findings in this child could be consistent with monophasic ADEM because it recurs within 3 months of the initial presentation, though still with multiple atypical features. Relapse of ADEM, particularly during steroid taper, has been reported with worsening symptoms and neuroimaging; however, the progression of neuroimaging with hemorrhagic lesions and significant neurologic deterioration is notable. A rare, hyperacute variant of ADEM (acute hemorrhagic leukoencephalitis [AHL]) can lead to severe neurologic findings, including cerebral edema and death.^18–21  However, AHL is often described as the initial ADEM presentation and not as a recurrence.

The age of onset, the lack of preceding viral illness and other neurologic deficits at presentation, slightly atypical initial MRI findings, unusual clinical course, and the presence of other systemic manifestations should motivate us to explore alternative diagnoses. Interestingly, Dr Tran mentioned CLIPPERS as a potential diagnosis on the basis of the initial MRI findings. A recent report described a case of CLIPPERS as CNS-restricted HLH in an 18-month-old female with defects in UNC13D, which is 1 of the genetic causes of HLH (familial HLH).²²  Perhaps the constellation of findings in this patient represents HLH with CNS manifestations? Do the MRI findings support HLH with CNS involvement as the leading diagnosis?

The follow-up MRI demonstrating new microhemorrhages is not specific for either AHL or HLH. However, AHL tends to have more necrotizing and aggressive hemorrhage than the minimal microhemorrhage in this case. Cerebral edema and diffuse brain injuries have been reported in AHL and HLH with CNS involvement.^{18–21,23,24}  There are radiologic features that are more commonly observed in HLH as compared with ADEM.¹¹  They are the centrally located and often symmetric white matter lesions, the presence of microhemorrhages, and contrast enhancement (Table 4), all of which are much more prominent in the follow-up MRI and more suggestive of HLH. Therefore, we need to reevaluate the diagnosis of ADEM and investigate HLH with CNS involvement fully as the alternative diagnosis.

AHL and HLH with CNS involvement are severe manifestations of 2 uncommon conditions (ADEM and HLH, respectively). At the initial presentation, the patient did not undergo diagnostic evaluation for HLH. How common are CNS manifestations as the presenting symptoms of HLH? Currently, he does not meet 5 of 8 criteria that are required for the clinical diagnosis of HLH (Table 5).²⁵  However, he had been treated with a prolonged course of steroid, which would have altered the laboratory findings. Can we still use the HLH–2004 clinical criteria for diagnosis? If we cannot rely on the clinical diagnostic criteria, do we have high enough index of suspicion to perform specialized testing now?

HLH is an inflammatory syndrome characterized by clinical signs and symptoms of pathologic and extreme inflammation.²⁶  HLH presents with a wide constellation of clinical and laboratory findings, many of which are also features of other inflammatory conditions.^26,27  This makes the establishment of a diagnosis challenging and often leads to delays in the initiation of treatment, with subsequent poor outcomes. The pathologic hyperactive inflammation can arise from many causes that include mutations associated with impaired cytotoxic function of the natural killer (NK) cells and T-cells (PRF1, UNC13D, Rab27a, STX11, SH2D1A, BIRC4, and others), infection, malignancy, rheumatologic conditions, and many other rare etiologies.²⁶  Clinical criteria have been established to identify patients who are likely to have HLH (Table 5).²⁵  In patients meeting the clinical criteria, additional specialized diagnostics should be used to confirm the diagnosis. They include genetic testing, NK cell function assessment, CD 107a mobilization, serum levels of soluble IL-2 receptor, CXCL9, sCD163, IL-18, and perforin/granzyme expression.

In addition to systemic symptoms, CNS manifestations such as seizures, meningismus, cranial nerve palsy, and impaired consciousness occur in a significant number of the patients.^{4,10,11,23,24}  Although once considered to be rare, there is increasing recognition that patients with familial (genetic) HLH can present with predominant neuroinflammatory symptoms with minimal systemic manifestations (CNS-restricted HLH).^22,28,29 

At the time of initial presentation, it was unlikely that he would have met HLH–2004 clinical diagnostic criteria²⁵  (Table 5), most notably because he did not have persistent fevers or cytopenia. On readmission, in the context of an acutely deteriorating clinical condition, laboratory investigation that includes fibrinogen, triglyceride, and ferritin, in addition to complete blood count and liver function test, might have been useful in raising suspicion toward HLH workup. However, as Dr Lai mentioned, the child has been treated with steroids chronically. Therefore, these systemic HLH markers may have limited utility in aiding the diagnosis. Another notable challenging factor in this situation is the family history of JIA in the patient’s father. Macrophage activation syndrome (MAS) is another life-threatening condition and can be described as an acquired form of HLH that occurs in association with rheumatologic disorders.³⁰  Although the patient himself has no previous symptoms of JIA, MAS could be the initial presentation of systemic onset JIA.²⁵  It is challenging to distinguish between familial HLH and MAS initially given the significant overlap between these 2 conditions.

I agree with Dr Tlais that this patient most likely would not have met the clinical criteria for HLH at the initial presentation. However, the patient now has 3 of 8 clinical criteria of HLH on arrival to our hospital (cytopenias, fevers, and splenomegaly). This highlights the necessity of continued reevaluation of clinical symptoms and findings, and adjusting the working diagnosis and management as necessary. At present, the traditional clinical and laboratory findings will be difficult to interpret in the setting of a prolonged course of steroids, given that steroids constitute a cornerstone therapy in HLH. Based on the imaging findings, his clinical progression, systemic symptoms, and first-degree relative with an autoinflammatory disease, the index of suspicion for HLH with CNS manifestations is sufficiently high that we should embark on the specialized diagnostics, even though this child does not meet the clinical criteria.

Further workup in this patient revealed absent NK-cell activity, bone marrow hemophagocytosis (Fig 3), and elevated sIL-2 receptor levels, confirming our suspicion for HLH on hospital day 5. The patient was then started on HLH-directed therapy before the results of the confirmatory genetic testing with etoposide, dexamethasone, and intrathecal methotrexate. He was ultimately found to have compound heterozygous PRF1 mutations on genetic testing, consistent with the diagnosis of familial HLH presenting as CNS-restricted HLH. The confirmatory genetic testing sent on admission to our hospital resulted almost 4 weeks later, which coincided with 4 months since initial presentation to medical care. Concurrent with HLH-directed therapy, the patient underwent HLA typing for possible allogeneic hematopoietic stem-cell transplantation (HSCT). However, he developed worsening neurologic status while awaiting HSCT, and, because of the irreversible nature of his neurologic deficits (seizure disorder, static encephalopathy, inability to swallow with G-tube dependence, and difficulty controlling oral secretions), and the risks associated with HSCT, the patient’s family opted to delay transplantation indefinitely. The patient has since required multiple hospitalizations for management of acute illnesses, including septic shock secondary to presumed bacterial pneumonia, increased seizure frequency, respiratory failure, and systemic HLH exacerbations (manifested by elevated liver enzymes, elevated ferritin, and pancytopenia).

It is well recognized that neurologic symptoms such as seizures and acute encephalopathy can occur in patients with systemic HLH manifestations.^{4,10,11,23,24}  Henter and Nennesmo described in 1997 classic neuropathologic findings of lymphocytic and histiocytic/macrophage infiltrate into the leptomeninges, perivascular space, brain parenchyma, and, ultimately, tissue necrosis, reflecting increasing disease severity.²³  Similarly, there is a wide spectrum of imaging findings that include normal, diffuse parenchymal vol loss, intraparenchymal lesions, and cerebral edema.^10–14  Although MRI findings are nonspecific, certain patterns in location and appearance may be more commonly observed in HLH as compared with ADEM (Table 4),¹¹  which is illustrated in this case.

Once considered a rare presentation, there are increasing reports of familial HLH with CNS-restricted manifestations that had been initially classified as neuroinflammatory conditions including ADEM and CLIPPERS.^{22,28,29,31–33}  We were presented with 2 rare etiological possibilities with significant clinical overlap, and that can both result in devastating neurologic sequelae. Our patient’s initial presentation and laboratory findings, as well as the subsequent clinical progression, were unusual for ADEM. On the other hand, the HLH–2004 clinical diagnostic criteria focus primarily on evidence of the systemic manifestation, which limits its utility in aiding the diagnosis in CNS-restricted HLH such as this case. The current time frame for genetic testing (4–6 weeks³⁴ ) limits its utility in the acute setting where the rapid initiation of disease-directed therapy is crucial. Ultimately, molecular diagnostics were the key element to arrive at the diagnosis in this case. Interestingly, mutations in the PRF1 gene comprise the majority of the reported cases of CNS-restricted HLH.²⁸  Whether this simply reflects the prevalence of PRF1 mutation in familial HLH or pathologic mechanisms that are unique to PRF1 mutation remains to be explored. Timely diagnosis is especially crucial to improve outcomes and quality of life because prompt treatment with CNS-directed HLH therapy followed by hematopoietic stem cell transplantation is necessary to prevent neurologic decline and symptom progression.²⁹  Our case, along with previously published reports, highlights the need to maintain a high index of suspicion and broaden the differential diagnosis in patients whose initial symptoms and subsequent clinical course do not conform to the expected neuroinflammatory conditions. Therefore, early molecular diagnostic testing in critically ill children may be indicated to ensure timely diagnosis and prompt treatment initiation.

We thank Dr Kenneth L. McClain for sharing his expertise on the diagnosis and management of HLH and for the editorial review. We also thank Dr Monira Haque for sharing her hematopathology expertise in the diagnosis of HLH.

ADEM

acute disseminated encephalomyelitis

AHL

acute hemorrhagic leukoencephalitis

CLIPPERS

chronic lymphocytic inflammation with pontine perivascular enhancement responsive to steroids

CNS

central nervous system

CSF

cerebrospinal fluid

HLH

hemophagocytic lymphohistiocytosis

HSCT

hematopoietic stem-cell transplantation

JIA

juvenile idiopathic arthritis

MAS

macrophage activating syndrome

MOG

myelin oligodendrocyte glycoprotein

NK

natural killer