Abstract
There is a known association of LE as a PNS that often pre-dates the diagnosis of an underlying malignancy. Brierley et al. described three cases of subacute progressive encephalitis, mainly affecting the limbic lobes [1]. One of these three patients had autopsy findings of bronchial lung cancer. Later, Corsellis et al. proposed a link between cancer and LE when they reported three further cases of LE in patients with bronchial lung cancer [2]. Once a diagnosis of LE is made, the treatment is dependent on the underlying cause.
Case report
The patient was a 55-year old school teacher who was brought to the Emergency Department (ED) at Middlemore Hospital, Auckland, New Zealand by ambulance with a 3 day history of headache, fever, confusion and bizarre behaviour noted by the family. The patient had no prior medical history and was taking no medications. He had no prior surgeries. He had allergies to cotrimoxazole and sulfa drugs. Social history was positive for tobacco (2 packs per day). The family reported that he was a heavy alcohol and cannabis user.
In the field the Glasgow coma score (GCS) was 13. At arrival to ED the GCS had decreased to 10 and the patient was combative. On physical exam he was febrile to 38.9°C with a pulse of 96 b.p.m. with a regular rhythm and a blood pressure of 140/60. Exam of eyes indicated no photophobia, and pupils were equal and reactive to light. There was no evidence of lymphadenopathy. Cardiovascular exam indicated dual heart sounds with no murmurs. The respiratory exam indicated a clear chest. Abdominal exam was normal. There was no evidence of meningismus and no petechiae. Blood sugar level was 30 mg dL–1 (1.7 mmol L–1) and increased to 122.5 mg dL–1 after the administration of Intramuscular (IM) glucagon.
Routine blood work was ordered along with non-contrast head computed tomography (CT). The patient was started on i.v. antibiotics for the empiric treatment of meningitis with i.v. ceftriaxone 2 g and i.v. acyclovir 800 mg. CT showed no evidence of an intra-cranial hemorrhage, masses or abnormal enhancement. Cerebrospinal fluid (CSF) spaces were patent with no contra-indication to lumbar puncture.
The CSF from the lumbar puncture at admission was positive for six white blood cells and no red blood cells. There were 12 lymphocytes seen. Glucose level was 177 mg dL−1 10.1 mmol L−1 and protein level was 0.48 g L–1. No organisms were seen on Gram stain or later on culture reported 5 days later. A toxicology screen sent at admission was positive for barbiturates, benzodiazepines and cannabis.
With a working diagnosis of viral encephalitis versus post-leukoencephalopathy, the patient was admitted to the medical intensive care unit for close observation for the first 3 days and later transferred to a regular medical floor for further evaluation.
On the fifth day of admission all cultures from the CSF returned with no growth. Herpes simplex virus, Varicella zoster virus, Epstein–Barr virus, mycobacterium tuberculosis and enterovirus polymerase chain reaction (PCR) were negative. The Borrelia serology was consistent with remote past or recent infection with Borrelia burgdorferi. Acid-fast stains for tuberculosis were negative.
Magnetic resonance imaging (MRI) approximately 1 week into the admission indicated bilateral, symmetrical signal change involving the cortex and subcortical white matter of the medial temporal lobes extending towards but not involving the posterior corpus callosum. No meningeal enhancement was seen. The supratentorial brain showed ischaemic white matter change prominent for age. There were no abnormalities in the brainstem, cerebellum and CSF spaces and no evidence of recent infarct or hemorrhage. The clinical impression was that of either a herpes encephalitis or LE. Given the negative PCR, a work up of LE began.
Blood tests were sent to the UK for antibodies against voltage-gated K+ channels that were later reported as negative. In addition, a testicular ultrasound was negative for any echogenic abnormalities. Chest/abdomen and pelvis CT with i.v. and PO contrast was read as negative.
An HbA1C, ordered on day 18 of admission was 10.5 %, which is consistent with poor long-term glycaemic control, and a new diagnosis of diabetes was made. Given the results of the aforementioned exams, extensive input was provided from physicians specializing in neurology, infectious disease and endocrine. He was also seen by a cardiologist for an episode of paroxysmal atrial fibrillation that resolved with amiodarone.
Confusion persisted during the admission and although formal mental status exams were not recorded, there were numerous notes describing the patient as vague and agitated wandering in the halls. On day 16 of admission an occupational therapist noted that patient was not oriented to month, day, season, year, date or ward. He was oriented only to city, country and to the fact that he was in a hospital. The patient also accused the staff of trying to poison and kill him. This prompted a Mini-Mental Status Examination (MMSE) on which the patient scored 15/30. More specifically, he scored zero for all the orientation aspects and showed significant short-term memory impairment. After a psychiatric evaluation the team initiated a trial of risperidone 4 mg at bedtime and clonazepam 2 mg three times a day.
The patient was started on 500 mg methyl prednisone daily, as an empirical treatment for acute LE 3 days later in the setting of worsening confusion and agitation despite antipsychotic medications. After 3 days the steroids were discontinued and the patient was transferred to the secure ward of Middlemore's acute psychiatric unit due to even further worsening of symptoms especially at night.
On the psychiatric ward the symptoms slowly improved and the clonazepam was tapered. An electroencephalogram (EEG) did not show any diagnostic abnormalities. The background activity was consistent with drowsiness but there were bursts of frontal rhythmic delta activity. The neurologist who interpreted the EEG report remarked that the bursts may represent a mild generalized non-specific abnormality consistent with encephalopathy. There was no evidence of epileptiform or seizure activity or focal or lateralized abnormalities.
A second MMSE conducted 2 weeks after the first showed improvement with a score of 19/30. The patient was able to score 4/10 in orientation. He continued to show signs of short-term memory impairment with a 5 min recall score of zero. Another noted improvement was the ability to perform four serial seven subtractions correctly.
On a final MMSE completed 2 days prior to discharge the patient scored 28/30, and scored 3/3 on 5 min recall. At discharge the patient was taking metformin and intermediate-acting insulin for newly diagnosed diabetes. In terms of psychiatric medications, the patient was discharged on risperidone and clonazepam.
Discussion
LE was diagnosed based on clinical presentation and radiographic findings. After a thorough work-up for infectious and metabolic causes, the final conclusion of idiopathic LE was made. The picture was initially clouded by the positive toxicology screen at presentation. Despite this, the clinician cannot overlook the association between neurological abnormalities and PNS. The most frequently associated cancers are small cell lung carcinoma (SCLC), breast, ovarian and testicular carcinoma.
Neurological PNS are rare disorders and are seen in <0.01% of patients with carcinoma. In 2004 Honnorat and Cartalat-Carel defined PNS as ‘neurological syndromes of unknown cause that often antedate the diagnosis of an underlying, usually not clinically evident, cancer’ [3]. There have been 137 patients with paraneoplastic limbic encephalitis (PLE) identified in the English-language literature, as cited in the Humayun Gultekin et al. study [4].
Increased technological advances have identified onconeuronal antibodies against neurological antigens, which suggests that PNS arises from an immune-mediated response. The immune-mediated response is directed against antigens common to tumor and neuronal cells. PNS can affect any branch of the nervous system. Antibodies have been discovered that seem to be highly predictive of specific neoplasms. Anti-Hu antibodies are associated with SCLC while anti-Ma2 antibodies are associated with testicular and lung cancer.
There are also antibodies against the voltage-gated potassium channels (VGKC), which have been associated with non-paraneoplastic or idiopathic syndromes. The incidence of this type of LE is thought to be at least as common as PLE and is more often seen in men at an average age of 65 years. VGKC antibody titres are usually >400 pmol and are frequently >1000 pmol in idiopathic cases. Normal levels are <100 pmol.
Management
The treatment of LE has two arms. One arm is to treat the underlying cause and the second focuses on immunosuppression. In cases of truly idiopathic LE, plasma exchange or i.v. immunoglobulin therapy in addition to high-dose steroids have been the treatment of choice.
If a malignancy is identified, treatment of the underlying tumor has been shown to be more effective than steroids in cases of PLE. Rees cited Graus's work with 200 patients that showed that treating the underlying tumor was an independent predictor of improvement/stabilization of the LE [5]. In the Humayun Gultekin et al. study of 137 patients followed for 8 months after diagnosis, 44% of patients showed improvement regardless of treatment [4]. The study also noted a greater neurological benefit in patients who received treatment for tumor rather than immunosuppression. The conclusion was that a prompt detection and treatment of underlying malignancy is the most important step in PLE.
Immunosuppression is the other arm of treatment in LE. There are currently no established protocols. In Lambert–Eaton myasthenic syndrome, another neurological PNS, plasma exchange with i.v. immunoglobulin has been an effective treatment. In other neurological PNS T-cell activation is thought to be the cause of symptomatology in which case treatment with tacrolimus and myophenolate mofetil have been attempted. Given that the exact pathogenesis is unknown in many PNS, combination treatment with plasma exchange with or without corticosteroids, cyclophosphamide and tacrolimus have been used. A specific protocol of plasma exchange or i.v. immunoglobulin with high-dose alternate day oral steroids has been suggested by Buckley [6].
Whatever the protocol, the effectiveness of immunosuppressive has not been ideal in many PNS. In fact, Rees stated that use of immunosuppressives can add risk of hastening tumor growth and dissemination [5]. This highlights the importance of early detection and treatment of tumor.
The focus of treatment of LE is an early diagnosis based on clinical manifestations and MRI. The next step is a thorough evaluation for underlying cause whether it is infectious, metabolic, malignant, or idiopathic. It is reported that the neurological symptoms can pre-date the detection of tumor and the need to be diligent in follow up is implied. There are no clear guidelines to date in terms of imaging, laboratory tests or follow up in patients diagnosed with idiopathic LE for possible paraneoplastic aetiology.