Article Text


Delayed onset of a second paraneoplastic neurological syndrome in eight patients
  1. François Ducray1,2,
  2. Francesc Graus3,
  3. Maria Claudia Vigliani4,
  4. Jean-Christophe Antoine2,5,
  5. Veronique Rogemond1,2,
  6. Albert Saiz3,
  7. Jérôme Honnorat1,2
  1. 1Hospices Civils de Lyon, Hôpital Neurologique, Centre de Référence Maladie Rare ‘Syndromes Neurologiques Paranéoplasiques’, Lyon, France
  2. 2INSERM, U842, Lyon, Université Lyon 1, Lyon, France
  3. 3Department of Neurology, Hospital Clinic, Universitat de Barcelona and Institut d'Investigació Biomèdica August Pi i Sunyer (IDIBAPS), Barcelona, Spain
  4. 4Dipartimento di Neuroscienze, Università di Torino, Torino, Italy
  5. 5Hôpital Bellevue, Neurologie, Saint-Etienne, France
  1. Correspondence to Professor J Honnorat, Neurologie B, Hôpital Neurologique de Lyon, 59 boulevard Pinel, 69677 Bron Cedex, France; jerome.honnorat{at}


Background The occurrence of a second delayed paraneoplastic neurological syndrome (PNS), different from the first, in the same patient, is not well known. However, recognition of this possibility is important in the management of the patient with PNS.

Objective To describe eight patients who presented with two different PNS with a time delay of months to years.

Design, setting and patients Retrospective analysis of eight patients, included in the European PNS database, with two different PNS that occurred months to years apart.

Results The median delay between the two different PNS was 15 months. The antibody repertoire did not change between the two episodes. Anti-Hu was detected in five patients, and anti-CV2 and anti-Ri in one each. One patient did not present onconeural antibodies. Lung cancer was diagnosed in six patients. In five patients, the second PNS revealed a cancer relapse (n=4) or a second cancer (n=1) and in three it occurred without evidence of tumour recurrence. In these three patients, all with anti-Hu antibodies, the second episode consisted of a lower motor neuron disease. Median survival was not reached after a median follow-up of 5 years.

Conclusions A second PNS can reveal a cancer relapse but can also arise in its absence. The long survival of patients with a second PNS suggests that the immune response might be more effective in controlling the cancer.

Statistics from


Patients with paraneoplastic neurological syndrome (PNS) usually display a subacute aggressive evolution and many patients die a few months after the diagnosis as a result of the combined effects of the neurological disability and cancer progression. In the most common PNS, paraneoplastic encephalomyelitis/sensory neuropathy associated with Hu antibodies, median survival time is 11 months and the percentage of survivors beyond 3 years is less than 20%.1 2 Information on the clinical evolution of PNS in those patients who survive more than 3 years is scarce. A recent study suggests that patients who are long term survivors of PNS associated with anti-Hu antibodies slowly progress over the years despite the fact that the cancer is cured, suggesting that the immune response may become independent of the original tumour trigger and remain active for many years.3 In line with this observation, we found that anti-Hu antibody levels remain unchanged several years after remission of the tumour. In this study, two of the 35 patients included presented with a second, different PNS months or years after the diagnosis of the first PNS.4 In the present study, we analysed the clinical and immunological features of a series of eight patients who developed two different types of PNS over several years to ascertain their clinical profile and draw attention to this possibility in patients followed with PNS.


Data from the PNS EURONETWORK Database ( collected between January 2000 and December 2008 (n=979) were retrospectively reviewed.5 Patients with a second PNS different from the first one were identified in four centres (Barcelona (n=3), Lyon (n=2), Torino (n=2) and Saint-Etienne (n=1)). Patients' sera were tested for the presence of Hu, Ri, CV2/CRMP5, Yo, Ma2 and amphiphysin antibodies by immunohistochemistry and western blotting, as previously described.6


Patient characteristics at the time of the first PNS are summarised in table 1 (detailed case reports are provided as supplementary data, available online only). Median age was 54 years (range 42–71 years). The first PNS episodes presented as a subacute sensory neuronopathy (n=2), a sensorimotor neuropathy (n=3), an opsoclonus–myoclonus (n=2) and a brainstem encephalitis (n=1). Serum screening for onconeural antibodies was positive in six patients: anti-Hu (n=4), anti-CV2/CRMP5 (n=1) and anti-Ri (n=1). This distribution was not different from the distribution in the 979 patients of the PNS EURONETWORK Database which contains 38.8% of cases with anti-Hu (n=380), 6% with anti-CV2/CRMP5 (n=59) and 5.1% with anti-Ri (n=50) antibodies, and 28.2% (n=276) without classical onconeural antibodies.5 Associated cancers were small cell lung cancer (SCLC) (n=4), carcinoid anaplastic lung tumour (n=1), prostatic adenocarcinoma (n=1), undifferentiated bronchial carcinoma (n=1) and breast cancer (n=1). After treatment of the underlying tumour in association with immunoactive drugs, the neurological condition improved in five patients and remained stable in three.

Table 1

Patient characteristics at the time of the first paraneoplastic neurological syndrome

Patient characteristics at the time of the second PNS are summarised in table 2. In all patients, the second PNS was different from the first one and occurred without recurrence of the first one except in patient No 6 in whom a mild worsening of the sensory neuronopathy was initially observed. Among the three patients in whom the first and second PNS consisted of a neuropathy (patient Nos 2, 3 and 6), the second PNS (lower motor neuron disease in patient Nos 2 and 6, gastrointestinal pseudo-obstruction in patient No 3) could not be considered as worsening of the initial sensory neuropathy because none of these three patients had suffered these symptoms at the first PNS and the pathological substrate of the new symptoms represented an immune attack of another area of the nervous system (the spinal motorneurons or the myentheric plexuses). Median delay between the first and second PNS was 15 months (range 6–42 months). The diagnoses of the second PNS were encephalomyelitis (n=1), chorea (n=1), stiff person syndrome (n=1), dermatomyositis (n=1), dysautonomia with gastrointestinal pseudo-obstruction (n=1) and lower motor neuron disease (n=3). In patients with CNS involvement, MRI imaging and CSF studies ruled out metastatic disease. These second PNS episodes antedated a cancer relapse in four patients and a second cancer in one. In three patients, the second PNS occurred without evidence of cancer relapse despite repeated whole body positron emission tomography scan imaging. Among the six patients who had an onconeural antibody, the same onconeural antibody was always found at the time of the second PNS. The titre of onconeural antibodies at both episodes was available in four patients: it remained stable (n=2) or decreased, and eventually became undetectable, before increasing again (n=2). In one patient, Hu antibodies were only detected at the time of the second PNS. Cancer was treated in all patients in whom the second PNS was associated with a cancer relapse or a second cancer. One patient (patient No 6) was treated with chemotherapy even though no relapse of the SCLC was evident. Six patients received immunoactive treatments, which consisted of steroids (n=4), intravenous immunoglobulins (IVIg) (n=4), cyclophosphamide (n=1) and ciclosporin (n=1). Three patients improved after treatment, two patients remained stable and three patients experienced progression of the neurological deficit. After a median follow-up of 4.5 years after the first PNS, median survival has not been reached. Five patients are alive without relapse of the cancer and four of these did not experience progression of the neurological symptoms. Three patients died, one from cancer relapse and two from evolution of the second PNS.

Table 2

Patient characteristics at the time of the second paraneoplastic neurological syndrome

The second PNS consisted of a lower motor neuron disease in the three patients without evidence of cancer relapse. These patients had anti-Hu antibodies. In patient No 6, the lower motor neuron disease first involved the lower limbs before extending to the upper limbs, abdomen and trunk. Anti-Hu antibodies became undetectable after the first PNS and reappeared at the onset of the lower motor neuron disease. CSF examination demonstrated only an elevated protein level (0.75 g/l). Treatment included chemotherapy, despite no evidence of cancer relapse, followed by steroids, IVIg and ciclosporin. However, the patient died of respiratory insufficiency after 5 years of evolution. No necropsy was performed. In patient Nos 2 and 7, the lower motor neuron disease was localised. In patient No 2, it was limited to the upper limbs, the trunk and the abdomen. CSF showed mild pleiocytosis (4 lymphocytes/mm3), elevated protein level (0.6 g/l) and oligoclonal bands. After treatment with IVIg and cyclophosphamide, no progression of the lower motor neuron was observed. The patient is alive and stable 8 years after the onset of the lower motor neuron disease without evidence of cancer relapse. The lower motor neuron disease was limited to the legs in patient No 7. CSF showed only an elevated protein concentration (1.5 g/l). The patient was treated with steroids and is alive and stable 4.5 years after the onset of the lower motor neuron disease without evidence of cancer relapse.


The simultaneous involvement of different areas of the nervous system by the paraneoplastic process is not unusual and the term paraneoplastic encephalomyelitis was introduced to define this syndrome that usually occurs in the setting of anti-Hu antibodies and lung cancer.1 Paraneoplastic encephalomyelitis is reported in 11% of patients with PNS.1 In contrast, the simultaneous occurrence of two different classical PNS is less frequent. Well known examples are the coincidence of paraneoplastic cerebellar degeneration and Lambert–Eaton myasthenic syndrome associated with antibodies to calcium channels and lung cancer or the simultaneous development of sensory neuronopathy and gastrointestinal pseudo-obstruction in patients with lung cancer and anti-Hu antibodies.1 7 The delayed onset of a second PNS seems to be much rarer as we identified only eight cases with this condition (0.8%) among the 979 patients of the PNS EURONETWORK database and found only one similar case in the literature.8

Anti-Hu was the most commonly found onconeural antibody in patients with delayed onset of a second PNS. Although the second PNS was different from the first in all cases, the antibody repertoire did not change between the two episodes. In fact, in two patients (patient Nos 1 and 3), the second PNS could be considered to be a different manifestation of the clinical syndrome usually associated with the underlying antibody. Indeed, polyneuropathy and chorea are two typical manifestations of the anti-CV2/CRMP5 syndrome,2 9 10 and sensory neuronopathy and gastrointestinal pseudo-obstruction are two typical manifestations of the anti-Hu syndrome.1 In five patients (62.5%), the second PNS was clearly associated with relapse of the initial cancer or of a second cancer, suggesting that cancer triggered the immune reaction responsible for the second PNS. However, in three cases, no cancer relapse was observed. Surprisingly, all three patients had anti-Hu antibodies and similar lower motor neuron disease. About 15%–20% of patients with anti-Hu syndrome show evidence of lower motor neuron involvement but it is rarely the only manifestation.1 11 12 Because of the particular features of the lower motor neuron disease in these three patients, it is very unlikely that it corresponded to a degenerative motor neuron disease.13 It rather suggests that the immune response had become independent of the cancer. Spontaneous complete remissions have been observed in patients with SCLC and anti-Hu antibodies; thus, an alternative hypothesis would be that cancer relapse was not found because of effective antitumour immunity.14 In any event, this similar lower motor neuron syndrome in patients without tumour relapse suggests a common mechanism of immunisation against motoneurons, probably specific for Hu patients.

Finally, another interesting characteristic of these patients was their prolonged survival. Median survival in the five patients with anti-Hu antibodies was not reached at 5 years. This result contrasts dramatically with the usual median survival reported in anti-Hu patients (11.8–18 months).1 2 This suggests that in patients who develop a second PNS, the immune response is more aggressive but at the same time more effective in eradicating the cancer.


View Abstract


  • Competing interests None.

  • Patient consent Obtained.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.