Heidenhain variant sporadic Creutzfeldt-Jakob disease diagnosed as an autoimmune encephalitis due to a false-positive GAD autoantibody

Abstract

A 64-year-old man presented with a subacute history progressive visual field defects, illusions and misperceptions. An initial MRI brain revealed a right occipital signal abnormality on diffusion-weighted imaging (DWI) with serum glutamic acid decarboxylase (GAD) autoantibodies markedly elevated. A diagnosis of autoimmune encephalitis was made, with the patient being treated with intravenous immunoglobulin. One month after discharge, the patient represented with worsening frank and well-formed visual hallucinations, ataxia and progressive cognitive impairment. Progress MRI displayed characteristic T2 ribboning on diffusion weighted imaging (DWI) and fluid-attenuated inversion recovery (FLAIR) sequences, along with periodic sharp wave complexes on electroencephalogram (EEG) and a raised CSF protein 14-3-3. Repeat serum, as well as cerebrospinal fluid (CSF), GAD antibodies were again markedly elevated as measured by ELISA (RSR, Cardiff, UK), although archival CSF from the original presentation as well as CSF from the second presentation had undetectable GAD autoantibodies as measured via radioimmunoassay (DIAsource, Ottignies-Louvain-la-Neuve, Belgium). Creutzfeldt-Jakob disease was confirmed at autopsy.

Background

Creutzfeldt-Jakob disease (CJD) is a neurodegenerative disease caused by the prion protein. The disease can be sporadic, transmitted or hereditary. While the diagnosis can only be confirmed at autopsy, typical MRI, electroencephalogram (EEG) and CSF studies are found in a large number of sporadic cases, and when present, offer an overall reasonable positive likelihood ratio for diagnosis.¹ The Heidenhain variant of CJD is characterised by predominant involvement of the occipital lobes that results in florid, well-formed visual hallucinations.² There is no treatment with any proven efficacy, and no patient with CJD has survived.^{1 2}

The rapidly evolving field neuroimmunological disorders has seen an increase in autoantibody testing, in an attempt to diagnose a range of immune-mediated conditions, that largely, are highly amenable to treatment. Case series and reports have shown that some patient who have been diagnosed with CJD, in fact have a treatable autoimmune disorder,^3–5 although a consequence of ‘screening’ for all possible autoantibodies in atypical neurological presentations includes being misled by false-positive results, which may deliver unrealistic treatment expectations.^{6 7}

Here, we describe a cautionary case of Heidenhain variant CJD that was previously diagnosed as ‘autoimmune’ in aetiology, but proven to be prion spongiform encephalopathy at autopsy.

Case presentation

A 64-year-old man was evaluated for a 3-week history of progressive visual field disturbance that included left-sided visual field illusions (‘wavy’ vision) and misperceptions (ill-defined shadows). A thorough neurological examination failed to show a reproducible visual field defect. Motor, sensory and cerebellar examinations were largely normal, except for some difficulty performing tandem steps. An external MRI of his brain at that time was reported as having an acute cortical stroke in the right occipital lobe, although the imaging features were atypical for stroke. Given the progressive nature of his symptoms, the initial neurological opinion was that of diagnosis of autoimmune encephalitis; a lumbar puncture at that time revealed an acellular CSF, with protein of 0.7 g/L (reference interval 0.15–0.45) and glucose of 3.5 mmol/L (reference interval 2.5–5.6). Serum GAD antibody levels were markedly elevated (>2000 kU/L reference <5) with other neurological autoantibodies usually associated with autoimmune encephalitis being negative. A dual positron emission tomography/CT (PET-CT) was ordered as part of a paraneoplastic evaluation, but this failed to show a malignancy.

He was commenced on an induction dose of intravenous immunoglobulin (IVIg), and then discharged with plans for outpatient follow-up.

The patient represented within 1 month of his previous assessment for a second evaluation and opinion. In that time, his visual field disturbances had progressed to frank—and remarkably well-formed—hallucinations (eg, he aptly described a ‘Japanese wedding dessert cart’ with great attention to the finer details of the construction) as well as progressive visual illusions, particularly of dynamic macropsia and dyschromatopsia of healthcare workers’ facial features. His neurological examination had progressed and now included overt confusion, a gross left visual field defect, complex multidirectional nystagmus, along with gross imbalance (positive Romberg’s assessment) and ataxia with an unsafe gait.

Investigations

A progress MRI of his brain showed evolution of the diffusion restriction, with cortical FLAIR ribboning in the right parieto-occipital region (figure 1). An EEG displayed characteristic periodic sharp wave complexes (PSWC) that were most pronounced over the occipital regions (figure 2). CSF protein 14-3-3 was positive.

Figure 1.

Axial brain diffusion-weighted imaging MRI showing posteriorly predominant cortical T2 ribboning.

Figure 2.

EEG with occipital predominant periodic sharp wave complexes. Performed as a 16-channel longitudinal bipolar (double banana) montage, sensitivity set to 10 µV/mm, notch filter off, with low-pass and high-pass filters set at 70 and 0.5 Hz, respectively. EEG, electroencephalogram.

Repeat CSF and serum autoimmune studies (NMDAR, CASPR2/LGi1, GABAbR, AMPAR, DPPX, IgLON5 surface neuronal receptor antibodies, as well as traditional paraneoplastic antibody targets, including Hu, Ri, Yo, PCA2 Ma/Ta2, CRMP-5, DNER, amphiphysin and SOX-1) were again negative, although repeat serum and CSF GAD antibodies were positive at >2000 and 181 kU/L as performed via ELISA (RSR, Cardiff, UK), respectively.

A false-positive GAD antibody result was suspected, and samples were re-evaluated for GAD antibodies via radioimmunoassay (RIA, DIAsource, Ottignies-Louvain-la-Neuve, Belgium); archival CSF from the original lumbar puncture 1 month previously, as well as CSF from the second collection were both negative (<0.7 U/mL), while serum from the second collection was positive at 115.5 U/mL (reference <1 U/mL) (table 1).

Table 1.

Comparison of GAD antibody results via two different assays (ELISA and RIA) on samples collected on both presentations

Sample type	Assay method	Firstpresentation	Second presentation
CSF	ELISA (kU/L)	NT	181
	RIA (U/mL)	Negative (<0.7)	Negative (<0.7)
Serum	ELISA (kU/L)	>2000	>2000
	RIA (U/mL)	NT	115.5 (post IVIg)

IVIg, intravenous immunoglobulin; NT, not tested; RIA, radioimmunoassay.

CSF GAD antibodies were negative via RIA on both occasions, while the ELISA was markedly positive on the second collection. Serum GAD antibody results were strongly positive on both collections, while only markedly positive via RIA on the second collection. The second serum collection was post IVIg which may have contributed to a false-positive RIA result; not tested.

Differential diagnosis

Heidenhain variant CJD.

Outcome and follow-up

During his second hospital inpatient admission, his condition had exponentially deteriorated. He was initially trialled on a course of intravenous steroids pending the results of the EEG, protein 14-3-3 and GAD antibody assay via RIA. The patient became progressively confused, then delirious and agitated. His ataxia had also progressed with him suffering a fall. The medical team—in discussion with the family—then arranged for palliative care input.

He passed away exactly 6 weeks from his original hospital presentation.

A limited postmortem brain autopsy was consented to by his family. Histopathology confirmed CJD (figure 3).

Figure 3.

High power microscopy of H&E stained right temporal lobe showing widespread microvacuolar spongiform change with occasional larger coalescent vacuoles and some associated mild gliosis (A). Immunohistochemical staining of frontal cortex with anti-12F10 (prion protein) shows a fine synaptic pattern with occasional larger coarse aggregates (B). Final histopathological diagnosis was sporadic Creutzfeldt-Jakob disease (sCJD) MM1+2c.

Discussion

Heidenhain variant CJD is a progressive and devastating neurodegenerative illness without cure. Treatment is largely directed by the need for supportive measures, with adequate pre-emptive palliation a cornerstone of management.

The field of neuroimmunology has advanced exponentially in recent times, although in the wake of this expansion there exists a great temptation to label all unusual neurological presentations and associated imaging as autoimmune in origin. Furthermore, the field of neuroimmunological serology diagnostics is still largely in its infancy, with many disorders only being able to be diagnosed when the presence of specific pathognomonic autoantibodies are detected (such as anti-NMDAR, anti-voltage-gated potassium channel (VGKC) complex or even GAD antibody-associated encephalitis); radiological imaging such as MRI, while invaluable, does not consistently give the same patterns of pathology, even within case series of the same disorder. MRI findings can even be completely normal, despite gross clinical pathology.⁸

Unlike other serological testing modalities that employ a two-step testing process (eg, a highly sensitive screening assay followed by a highly specific confirmatory assay), most testing for antibodies in immune-mediated neurological disorders, particularly the autoimmune encephalitides, are limited to one assay.⁸ These assays, like all assays, are not infallible; all tests intrinsically will generate false-positive results occasionally. Additionally, the level of positivity does not correlate with the likelihood of the result being actually true; for example, a low-level false positive and a high-level false positive are both simply that—false.⁹ Without confirmatory assays being generally available, and with imaging such as MRI being inconsistently positive in autoimmune encephalitis, a false-positive result has the potential to mislead clinicians and cause undue distress and harm from inappropriate interventions or treatment expectations.

Here, we present a case of Heidenhain variant CJD with classical cortical ribboning sign on MRI that subsequently was investigated at a much later time point and found to have PSWC on EEG and a positive CSF protein 14-3-3. The diagnosis was eventually confirmed at autopsy. This patient, in retrospect, had a ‘full-hand’ clinicoradiographic presentation consistent with CJD, with congruent CSF protein 14-3-3 results. In an otherwise ‘incomplete’ presentation, other antemortem testing, such as the novel CSF real-time quaking-induced conversion (RT-QuIC) assay may have been insightful when the patient was initially evaluated.¹⁰

Our patient did not carry a diagnosis of type 1 or 2 diabetes mellitus. The original and repeat CSF studies on this patient failed to show GAD antibodies via RIA, while the second CSF sample was positive for GAD antibodies via ELISA. Both serum samples were strongly positive via ELISA, with the second sample being moderately positive via RIA (there was no serum remaining from the original collection to assess GAD antibodies via RIA). While the presence of GAD antibodies may have been genuine epiphenomena secondary to neurodegeneration or even true anti-GAD associated neurological autoimmunity,¹¹ the wildly discordant results are more likely to be due to a false positivity, given the recognised higher sensitivity of the RIA,^{12 13} with CSF results consistently negative by this method. The positive serum RIA result is difficult to interpret, although it is well established that IVIg preceding sample collection may cause a falsely elevated result, likely to exogenous GAD antibodies from normal blood donors.¹⁴

Not only was the correct diagnosis delayed by an erroneous GAD-antibody result, but inefficacious treatment with IVIg and steroids (the latter which had the potential to contribute to morbidity from delirium) were given, as well as establishing incorrect outcome expectations from family members and healthcare staff. Furthermore, the accelerated deterioration while on the inpatient ward (pending the above tests) impeded the ability to adequately arrange palliative discharge planning into more comfortable surrounds.

Patient’s perspective.

Patient’s wife described that during the night his ensuite in the hospital had turned into a beautiful vertical interior garden and went on to explain the type of plants in great detail.

The week before his second admission was the worst; I would find him lost in various places in the house during the night. He did not know what he was doing and I would have to yell at him to try and get him out of his trance-like state. He would yell out for me sometimes that he did not know where he was. A very upsetting sad time for us both. Hopefully, this case report can help other doctors to pick up on Creutzfeldt-Jakob disease earlier and stop unnecessary treatment.

Earlier diagnosis would have allowed for a more structured palliation.

Learning points.

• Creutzfeldt-Jakob disease is a progressive neurodegenerative disorder mediated by a pathogenic prion protein, and is universally fatal.

• Immune-mediated neurological diseases, particularly antibody-associated autoimmune encephalitis, are largely defined by the presence of a particular autoantibody and thus diagnosis often relies solely on one serological test.

• All tests—including assays to detect autoantibodies found in autoimmune encephalitis—have the potential to be fallible and should always be interpreted with a degree of caution, particularly in the face of other diagnoses.

• The degree of positivity of an assay does not necessarily correlate with the degree of true positivity.

• False-positive autoantibody serological tests have the potential to lead clinicians to an incorrect diagnosis that in turn may lead to inefficacious treatment, or expectations of treatment.