A woman with ataxiaBMJ 2009; 338 doi: https://doi.org/10.1136/bmj.b180 (Published 29 January 2009) Cite this as: BMJ 2009;338:b180
A 41 year old woman was admitted to hospital with a one week history of transient episodes of blurred vision and progressive dysarthria, incoordination and ataxia, such that she could no longer stand.
Three months earlier she had experienced the gradual onset of a mild occipital headache and episodes of transient blurred vision. Her speech had felt slower and slurred and her left hand was clumsy. These symptoms had been preceded by a severe outbreak of cold sores, which had been treated with oral acyclovir. On this occasion, her neurological symptoms gradually resolved over a month. She had no family history or other history of note, and she was not taking any drugs except the oral contraceptive pill.
On examination there were cold sores around her mouth and she had bilateral cerebellar signs, more prominent on the left. She had negative viral serology and a normal magnetic resonance imaging of her head and venous sinuses, and electroencephalogram. Lumbar puncture showed an acellular cerebrospinal fluid with mildly raised protein (0.52 g/l) and unmatched oligoclonal banding. Results of viral polymerase chain reaction testing were negative. Antineuronal antibodies were sent and returned a positive anti-Yo antibody.
(1) What is the significance of this result?
(2) What is your management plan?
(3) What is the likely prognosis?
1 The presence of antineuronal antibodies indicates that the woman has a paraneoplastic cerebellar syndrome.
2 Any patient with positive antineuronal antibodies on testing needs full investigation to exclude an underlying malignancy. Although the paraneoplastic phenomenon is thought to occur only once metastasis or micrometastases have occurred, the underlying malignancy can be difficult to find. Investigations may include ultrasonography, computed tomography, and positron emission tomography scanning.
3 The prognosis depends on the underlying malignancy. The neurological improvement once the underlying cancer has been treated depends upon the type of nerve involved, but it is usually limited.
Last test result
The presence of anti-Yo antibodies indicates that this woman has a paraneoplastic cerebellar degeneration syndrome.
Paraneoplastic neurological syndromes occur in about one in 10 000 patients with cancer. Up to a third of them may be cerebellar syndromes.1 With increasing age it is more likely that a subacute cerebellar syndrome is caused by an underlying neoplasia—for example, in women more than 50 years old, about two thirds of which are caused by underlying malignancy.2
Cerebellar degeneration syndromes were first described in association with breast or ovarian cancers in 1938. In the mid-1980s antibodies to neurones in cancer patients were discovered.3 Since then nine different types of antibodies have been described with paraneoplastic cerebellar degeneration1 and specific antibodies seem to be associated with specific types of neoplasm. For example, anti-Yo antibodies (directed against Yo proteins) are predominantly associated with gynaecological cancers in women.4 They occur in about 2.3% of ovarian cancers and 1.6% of breast cancers.5 Anti-Hu antibodies are associated with squamous cell carcinomas, most often in the lungs, in both men and women. Anti-Yo is not 100% specific for gynaecological cancers, and it has also been associated with gastrointestinal cancers.6 It is, however, thought to be close to 100% specific to an underlying malignancy1 4 so a positive result should prompt thorough investigation. Although the specificity of antineuronal antibodies in paraneoplastic syndromes is high, their sensitivity is low at about 50%. Therefore, as in the case in question where initial tests were negative, response to empirical treatments was poor, and an obvious diagnosis was not apparent, an underlying malignancy must be sought. Antineuronal antibodies are sometimes quantified as a marker of the severity and response to treatment of the underlying malignancy. These antibodies, however, are affected by many other factors such as the strength of the host’s immune system, and they do not correlate well with the progression of the tumour. They are, therefore, not quantified in many UK laboratories.
Tumours in patients with anti-Yo antibodies express the Yo antigen.7 The damage in paraneoplastic syndromes is thought to be caused by a cytotoxic T cells immune response directed at the tumour, which cross reacts with the neurones.8 9 10 This may be a response aimed at protecting the host against the tumour. All cases so far have reported metastases or micrometastases at presentation.11 In about 60% of patients the cerebellar syndrome precedes the diagnosis of cancer1 4 11 12 and antineuronal antibodies can be detected months or years before an underlying tumour manifests. A gap of 10 years between the development of the neurological syndrome and detection of the underlying cancer has been reported.7 In some cases, however, the clinical syndrome has occurred after the underlying cancer has presented or even during or after treatment,11 which indicates that it is not simply the presence of metastatic disease that triggers the immune response.
It is unclear whether the presence of this response slows the progression of the underlying neoplasia. No evidence indicates that patients with a clinically worse cerebellar syndrome have better outcomes. It seems that the overall outcome may be similar irrespective of the presence or severity of the cerebellar syndrome.13
Clinically, paraneoplastic cerebellar degeneration usually presents as a progressive gait ataxia followed by truncal and limb ataxia with dysarthria and nystagmus.4 It is usually symmetrical and progresses over days to weeks,4 but it can rarely occur over hours to days. The woman’s illness occurred in a biphasic manner, and it seemed that the first episode was responsive to acyclovir given by the general practitioner. Her varicella zoster virus test, however, was negative for IgM antibodies in both episodes, indicating an unusual biphasic presentation of a paraneoplastic syndrome rather than a viral aetiology. In about 20% of cases mild memory and cognitive deficits exist,4 and they can be associated with brainstem symptoms, as in this case. This may be because Yo proteins are found in the large neurones of the brainstem as well as Purkinje cells.
From this woman’s initial presentation saggital sinus thrombosis was considered because of her history of generalised insidious headache, visual disturbances, and use of the oral contraceptive pill. With the recurrence and worsening of her symptoms, multiple sclerosis or post-infective acute disseminated encephalomyelitis were possibilities. The raised protein and unmatched oligoclonal bands in the cerebrospinal fluid were consistent with this (acute disseminated encephalomyelitis may or may not be associated with cerebrospinal oligoclonal banding). However, normal magnetic resonance imaging and magnetic resonance venography preclude these diagnoses.
Viral infections can cause cerebellar syndromes. Herpes simplex virus encephalitis can rarely present with a predominantly cerebellar syndrome, rather than the classic limbic encephalitis.14 The gradual presentation and normal cerebrospinal fluid and magnetic resonance image made this diagnosis unlikely. If herpes simplex virus is suspected, acyclovir, which has a relatively safe side effect profile, should be considered.
In a paraneoplastic syndrome, blood tests are usually normal. Cerebrospinal fluid shows mild inflammatory changes and oligoclonal bands in about 60% of cases. Protein can also be raised, and therefore the presence of these does not necessarily support a diagnosis of an underlying inflammatory process. Magnetic resonance images are usually normal in early stages, but may show mild cerebellar atrophy in later stages.2 15 16
2 Management plan
Cerebellar syndrome has no effective treatment. Immunoglobulins, plasmaphoresis and cyclophosphamide have all been tried but the results are disappointing. The underlying malignancy is treated. In this case, there were no localising clues on history, examination, or on the chest radiograph so the next stage was for the woman to have a mammogram, a breast ultrasound scan, and transvaginal ultrasonography—all of which were normal. A computed tomogram of her chest, abdomen, and pelvis was also normal. She then went on to have a positron emission tomography scan, which showed increased uptake in the left axilla, and she had a left axillary lymph node biopsy after this. The biopsy showed metastatic cancer, but the sample was too small for further analysis. She therefore had a formal lymph node excision. This showed adenocarcinoma staining for human epidermal growth factor receptor 2, which with the presence of anti-Yo antibodies indicated a primary cancer of the breast. Magnetic resonance imaging was used to image the breasts and showed a 6 mm area in the lower inner quadrant of the left breast. Targeted ultrasonography then showed a further three or four small hypoecogenic nodules in the inner lower quadrant of the left breast. She went on to have a left radical mastectomy and axillary node clearance that showed metastases in three out of eight lymph nodes. She was treated with adjuvant carboplatin, herceptin, and docetaxal chemotherapy.
The neurological symptoms usually stabilise once the underlying cancer is treated17 but do not tend to improve despite attempts to modify the immunological response with corticosteroids, immunoglobulins, plasmapheresis, or cyclophosphamide either at presentation or after treatment of the underlying neoplasm.3 Cases have been reported, however, where improvements have occurred, and it may sometimes be worth trying these treatments. The patient here made little improvement with intravenous immunoglobulins and plasma exchange. Her syndrome stabilised after surgery. The overall prognosis depends on the nature and stage of the underlying neoplasm at diagnosis. A year later she remains in remission.
Differential diagnosis of a subacute cerebellitis include inflammatory, vascular, neoplastic, and paraneoplastic causes. Blood samples should be tested for antineuronal antibodies.
The presence of antineuronal antibodies strongly supports a paraneoplastic diagnosis, but their absence does not exclude it. Have a high degree of suspicion if no other explanation for the clinical picture can be found
The cerebrospinal fluid may show inflammatory changes, a raised protein, and positive oligoclonal bands
Brain magnetic resonance imaging is usually normal in early stages, but may show cerebellar atrophy later
Treatment of the neurological syndrome is usually ineffective
Treatment is directed at the underlying neoplasm
Cure of the cancer does not usually result in a remission in the neurological syndrome.
Cite this as: BMJ 2009;338:b180
Competing interests: None declared.
Patient consent obtained.