Search

Objective: The aim of this study was to describe specificities of patients with NMDA receptor antibody (NMDAR-Ab) encephalitis associated with a malignant tumor., Methods: Retrospective observational study of 252 patients with NMDAR-Ab encephalitis of the French Paraneoplastic Neurological Syndrome Reference Center. Patients were classified in three groups: (1) non-malignant ovarian teratomas, (2) malignant ovarian teratomas (immature), and (3) other malignant tumors., Results: Sixty patients (23.8%) had an associated tumor and 15 (6%) were malignant. No particular neurological symptom was observed in these patients. Ovarian teratomas were the most frequent (51 cases) with 6 of them immature (11.8% of teratomas). Nine patients (3.6%) developed other malignant tumors (3 small cell lung carcinomas, 1 uterine adenocarcinoma, 1 prostate adenocarcinoma, 1 Hodgkin lymphoma, 1 pineal dysgerminoma, 1 neuroblastoma and 1 pancreatic neuroendocrine tumor). Among patients with a cancer other than teratoma, 6/9 were elderly patients (median age 65 years, representing 30% of elderly patients with such encephalitis) compared to a median age of 26 years in adult patients included herein. The clinical course was similar in the three groups, other than a higher death rate among patients with malignant tumors (86 versus 2%; p < 0.001) mainly due to tumor progression (5/7 deaths)., Conclusion: Immature ovarian teratomas represent 11.8% of all teratomas in patients with NDMAR-Ab encephalitis. The other malignant tumors are mainly observed in elderly patients. The presence of a malignant tumor does not impact the neurological presentation but is directly associated with a higher risk of death.

Sera from 71 patients with primary Sjögren's syndrome (PSS) and from 102 patients with systemic lupus erythematosus (SLE) were tested by immuno-dot blotting against HuD, Ri, Yo and amphiphysin recombinant proteins. For Ri, Yo and amphiphysin antigens, no immunoreactivity was found in the 173 sera tested. One PSS patient with a clinical picture of subacute sensory neuronopathy had high titers of anti-Hu antibodies. An extensive search for an underlying tumor was initially negative but a small cell lung cancer was eventually discovered three years later. Another patient with SLE and a clinical picture of demyelinating polyradiculoneuropathy had anti-Hu antibodies. Repeated search for an underlying tumor remains negative after five years follow-up in this young non-smoking patient. In addition, the neuropathy progressively improved and the anti-Hu antibodies titer slowly decreased from 1:8000 to 1:2000, making the diagnosis of paraneoplastic syndrome unlikely in this patient. This study indicates that the detection of anti-Hu antibodies in patients with known symptomatic systemic autoimmune diseases such as PSS or SLE should induce the same work-up than the detection of these antibodies in the absence of other immune diseases, i.e. repeated search for occult cancer during several years. As illustrated by our first patient, this strategy may be fruitful. Nevertheless, the clinician should know that anti-Hu antibodies may exceptionally (0.6% in this series) occur in systemic autoimmune disorders with neurological complications, in the absence of an underlying neoplastic disease.

Objectives: To evaluate the effect of a combination of immunoglobulins (IVIg), cyclophosphamide (CTX), and methylprednisolone (MP) on the clinical course of patients with paraneoplastic neurological syndrome (PNS) and antineuronal antibodies (Abs)., Methods: Seventeen patients with paraneoplastic encephalomyelitis/sensory neuropathy (PEM/SN) with anti-Hu Abs (n = 10) or cerebellar degeneration (PCD) with anti-Yo Abs (n = 7) received one to nine cycles (mean 3.5) of a combination of IVIg (0.5 g/kg/day from days 1 to 5), CTX (600 mg/m2 at day 1) and MP (1g/day from day 1 to 3). The Rankin scale (RS) was used to evaluate the response. A positive response was considered as either improvement or stabilisation in patients who were still ambulatory (RS< or =3) at the onset of treatment, whereas only improvement, and not stabilisation, was considered a therapeutic benefit in bedridden patients (RS> or =4)., Results: Tolerance was good and no patient experienced grade 3/4 toxicity (World Health Organisation). Sixteen patients were evaluable for response. Of the seven patients with RS> or =4, none improved. Of the nine patients with RS< or =3, none improved but three (two SN and one PCD) stabilised for 4, 35, and 16 months., Conclusions: This study suggests that vigorous immunosuppressive treatment is not useful in severely disabled PNS patients with antineuronal Abs. In a minority of patients (mainly with SN) who are not severely disabled at the onset of treatment, a transient stabilisation is possible and deserves further evaluation.

Objective: To evaluate 1) the effect of the tumor treatment on the clinical course of paraneoplastic encephalomyelitis (PEM) with anti-Hu antibodies, 2) the impact of immunotherapy on the tumor evolution, and 3) the outcome of the small cell lung cancer (SCLC) of PEM patients compared with that of patients without PEM., Methods: The authors retrospectively analyzed 51 PEM patients (42 with SCLC, 9 with other tumors) who received antineoplastic treatment with (25 patients) or without (26) concomitant immunotherapy. Tumor response was assessed at the end of the antineoplastic treatment. Progression of PEM was defined as a change of at least 1 point in the Rankin scale measured at the onset and at the end of the tumor treatment. To evaluate the outcome of SCLC, 27 PEM patients with SCLC were matched one-to-one with SCLC patients without PEM for age, performance status, tumor stage, and type of antineoplastic treatment., Results: Thirty-six (70%) patients were neurologically stable at the end of the tumor treatment. In a logistic regression analysis, tumor complete response was the only predictor of PEM stabilization (OR 7.07; 95% CI 1.68 to 29.76; p = 0.006). Immunotherapy did not modify the outcome of the tumor and PEM. Median survival was similar in SCLC patients with and without PEM, but the probability of survival at 30 months was higher in PEM patients with SCLC (OR 5.26; 95% CI 1.0004 to 27.6902; p = 0.03)., Conclusions: Complete response of the tumor seems to have a favorable influence on the course of paraneoplastic encephalomyelitis (PEM). Concomitant immunotherapy does not adversely affect the tumor outcome. The small cell lung cancer of PEM patients may have a slightly better evolution than that of patients without PEM.

Paraneoplastic encephalomyelitis or subacute sensory neuronopathy associated with small-cell lung cancer (SCLC) and high titers of anti-HuD antibodies, also called the "anti-Hu syndrome," is believed to result from an immune response triggered by tumor antigens and misdirected to the neurons. To further assess the issue of cell-mediated immunity in this disease, the peripheral blood lymphocyte surface phenotype was studied in 15 patients suffering from the anti-Hu syndrome (seropositive group) and in two control groups consisting of 12 seronegative SCLC patients without neurological syndrome and 15 healthy volunteers. In addition, the recombinant human HuD protein was used to stimulate in vitro peripheral blood mononuclear cells of 10 seropositive patients and of 10 patients from each control group. Phenotypic analysis of the peripheral blood lymphocytes revealed a significant increase of the memory helper (CD45RO+CD4+) T cells in the seropositive group in comparison with the two control groups. Antigen-specific proliferation of peripheral blood mononuclear cells, measured by [3H]thymidine uptake after HuD antigen stimulation, was much higher in the seropositive group than in the two control groups, and phenotypic analysis of proliferating cells revealed a significant expansion of the CD45RO subpopulation of T cells in the seropositive group. Furthermore, after HuD stimulation, a significant increase of the interferon-gamma/interleukin-4 ratio was found in culture supernatants of the seropositive group compared with seronegative SCLC patients and normal controls. Taken together, these results indicate that HuD protein is an antigenic target for autoreactive CD4+ T cells, presumably of the Th1 subtype, which could therefore be directly involved in cell-mediated injury of the nervous system as well as in antitumoral immunity.

Patients with stiff man syndrome and breast cancer develop anti-amphiphysin I antibodies that primarily recognise the C terminus of the protein. Anti-amphiphysin I antibodies have also been identified in a few patients with paraneoplastic neurological disorders (PND) and small cell lung cancer (SCLC). The frequency of anti-amphiphysin I antibodies in patients with SCLC and PND was analysed and the epitope specificity of these antibodies was characterised. Anti-amphiphysin I antibodies were evaluated by immunohistochemistry on human and rat cerebellum and immunoblots of rat brain homogenates. Serum samples included 134 patients with PND and anti-Hu antibodies (83% had SCLC), 44 with SCLC and PND without anti-Hu-antibodies, 63 with PND and either Yo, Ri, or Tr antibodies, 146 with SCLC without PND, and 104 with non-PND. Positive serum samples were confirmed with immunoblots of recombinant human amphiphysin I and immunoreacted with five overlapping peptide fragments covering the full length of the molecule. Serum samples positive for anti-amphiphysin I antibodies included those from seven (2.9%) patients with PND and two (1.4%) with SCLC without PND. Six of the seven anti-amphiphysin I antibody positive patients with PND had SCLC (three with Hu-antibodies), and one had anti-Hu-antibodies but no detectable tumour. The PND included encephalomyelitis/sensory neuropathy (five patients), cerebellar degeneration (one), and opsoclonus (one). All anti-amphiphysin I antibodies reacted with the C terminus of amphiphysin I, but seven also recognised other fragments of the molecule. In conclusion, anti-amphiphysin I antibodies are present at low frequency in patients with SCLC irrespective of the presence of an associated PND. All anti-amphiphysin I antibody positive serum samples have in common reactivity with the C terminus of the protein.

Paraneoplastic neurological diseases are a group of neurological disorders associated with neoplastic tumors but not due to tumoral extension, metabolic, infectious, vascular or toxic complications of these tumors or their treatment. In the majority of paraneoplastic neurological disorders, circulating autoantibodies directed against neurons have been found in the serum and/or the CSF suggesting, and in some cases implicating, autoimmunity in the pathophysiology of these diseases. The finding of autoimmune phenomena during the course of paraneoplastic neurological disorders is of importance: from a practical point of view, since the detection of anti-neuronal autoantibodies is of great diagnostic help and should lead to the thorough search of the associated tumor often at an early stage of its development; from a theoretical point of view, these disorders represent a peculiar type of molecular mimicry. Tumoral neontigens having structural homology or identity with neuronal autoantigens elicit autoreactivity. The immunological effector mechanisms involved in the pathophysiology of paraneoplastic syndromes appear to differ according to the disease: autoantibodies are pathogenic in Lambert-Eaton syndrome whereas, in paraneoplastic cerebellar degeneration and in the Hu syndrome, the cellular immune response might play a greater role.

We compared the human leukocyte antigen alleles found in a group of 17 patients with small-cell lung cancer (SCLC), paraneoplastic neurologic syndromes (PNS), and high titers of anti-Hu autoantibodies with those in 30 patients with SCLC but no PNS and no anti-Hu antibodies (control group). There was no difference between the two groups, suggesting that specific haplotypes are not required for the development of the "anti-Hu syndrome."

Several cancers, especially lung, ovarian and breast, can cause paraneoplastic cerebellar degeneration. The presence of different antineuronal antibodies associated with different cancers and paraneoplastic cerebellar degeneration suggests that several immunological mechanisms may result in the same neurological disorder. In patients with small-cell lung cancer, paraneoplastic cerebellar degeneration may occur with or without Hu antineuronal antibodies (HuAb), indicating that patients with the same tumour can develop paraneoplastic cerebellar degeneration by different immunological mechanisms. Furthermore, paraneoplastic cerebellar degeneration sometimes occurs in association with the Lambert-Eaton myasthenic syndrome. In order to try to understand the clinical implication of antineuronal antibodies in patients with small-cell lung cancer, we examined the serum of 57 patients with presenting symptoms of paraneoplastic cerebellar degeneration for the presence of HuAb and P/Q- and N-type voltage-gated calcium channel antibodies. Patients with paraneoplastic cerebellar degeneration who were HuAb positive were compared with HuAb negative patients with respect to neurological symptoms, course of the neurological disorder, response to treatment, tumour prognosis, pathological findings, and cause of death. The tumour outcome and serological findings of these patients were also compared with those of 109 small-cell lung cancer patients without paraneoplastic syndromes of the CNS. Titres of HuAb were classified as 'high' (immunoblot titre > 1:10,000) or 'low' (< 1:10,000), the latter similar to the antibody titres detected in some small-cell lung cancer patients without paraneoplastic symptoms. Twenty-five patients with paraneoplastic cerebellar degeneration (44%) had high titres of HuAb, four (7%) had low titres of HuAb, and 28 (49%) were HuAb negative; for clinical comparisons with the patients with high titres of HuAb, the four patients with low antibody titres were included in the HuAb negative cohort. None of the 109 small-cell lung cancer patients without paraneoplastic symptoms had high titres of HuAb. The presence of high titres of HuAb defined a subset of patients who differed from the HuAb negative paraneoplastic cerebellar degeneration cohort, HuAb positive patients were more likely to be female (P < 0.01), to have multifocal neurological disease (brainstem encephalopathy and sensory neuropathy being common extracerebellar manifestations) (P < 0.002), and be severely disabled (P < 0.005). A total of nine patients (16%) from both paraneoplastic cerebellar degeneration groups developed electrophysiologically confirmed Lambert-Eaton myasthenic syndrome. Seven of these nine patients had serum available for P/Q-type voltage-gated calcium channel antibody testing and all seven were positive. In addition, 20% of HuAb negative paraneoplastic cerebellar degeneration patients without clinically identified Lambert-Eaton myasthenic syndrome had P/Q-type voltage-gated calcium channel antibodies, while only 2% of small-cell lung cancer patients without paraneoplastic symptoms had these antibodies. Treatment of the tumour and/or immunomodulation did not alter the course of paraneoplastic cerebellar degeneration, but improved Lambert-Eaton myasthenic syndrome symptoms. At the time of death, in 60% of HuAb positive and 20% of HuAb negative paraneoplastic cerebellar degeneration patients, the tumour was either not evident or localized to the chest (P < 0.007); neurological disease was the cause of death of 65% HuAb positive paraneoplastic cerebellar degeneration and 10% HuAb negative paraneoplastic cerebellar degeneration patients (P < 0.001). (ABSTRACT TRUNCATED)

Paraneoplastic limbic encephalitis (LE) is considered a particular manifestation of paraneoplastic encephalomyelitis (PEM), a remote effect of cancer almost always associated with anti-neuronal antibodies (anti-Hu; also called ANNA 1) and small cell lung carcinoma (SCLC). In order to define the frequency of anti-Hu antibodies in LE with SCLC and to analyse possible clinical differences between patients with and without anti-Hu antibodies, the charts of 16 patients with LE and SCLC were reviewed. Eight patients (50%) had anti-Hu antibodies (anti-Hu+) whereas eight patients (50%) had no detectable anti-neuronal antibodies (anti-Hu-). The clinical and laboratory features of LE and time to diagnosis of SCLC were similar in the anti-Hu+ and anti-Hu- groups. Involvement of other areas of the nervous system compatible with the diagnosis of PEM was observed in seven (87.5%) patients of the anti-Hu+ group but in only one (12.5%) of the anti-Hu- group (P = 0.012). Five patients, including four of the anti-Hu- group, had a partial improvement of the LE after treatment of the SCLC. Another anti-Hu- patient improved spontaneously. Six patients of the anti-Hu+ group died from the neurological disorder, whereas in the anti-Hu- group the cause of death was progression of the SCLC in the three patients who died. The results of this study indicate that the absence of anti-Hu antibodies does not rule out the presence of an underlying SCLC in patients with a clinical diagnosis of LE. Patients with LE and SCLC who are without anti-Hu antibodies are less likely to develop PEM and seem to improve more often after treatment of the cancer than those who present anti-Hu antibodies.

Paraneoplasic neurological syndromes are defined as an association between neurological disease and cancer in the absence of metastatic, infectious or metabolic complications. These syndromes occur in less than 1% of patients with cancer. With the exception of the Lambert-Eaton syndrome in which an autoimmune mechanism has been proven, the pathogenesis of paraneoplasic neurological syndromes remains unknown despite a frequent association with anti-Hu, anti-Yo, and anti-Ri antibodies. Neither passive transfer of these antibodies nor immunization with recombinant proteins has been able to produce paraneoplasic syndromes in animal models. It is most likely that cell immunity or other antigenic mechanisms are also implicated and that the synthesis of anti-Hu, anti-Yo or anti-Ri is simply one manifestation of a more complex immune reaction. Nevertheless, the discovery of these antibodies can be useful in the diagnosis of cancer, especially in the early stages. Other antibodies, specific for paraneoplasic neurological syndromes, will undoubtedly be discovered in the future, increasing the capacity of immunological diagnosis. A better understanding of immune cross reactions between the tumor and the nervous system may help further our understanding of the immune system and lead to progress in specific treatment for paraneoplasic neurological syndromes.

Objective: To evaluate the effect of intravenous high dose human immunoglobulin (IVIg) therapy on the clinical course and autoantibody titres of patients with neurological paraneoplastic syndromes., Methods: Twenty two patients with paraneoplastic encephalomyelitis and sensory neuronopathy syndrome associated with anti-Hu antibodies (18) or paraneoplastic cerebellar degeneration (PCD) with anti-Yo antibodies (four), were treated with 1-26 (mean 5.8) cycles of IVIg. The Rankin scale was used to evaluate the response., Results: The only serious toxicity was one case of haemolytic anaemia. Twenty one patients were evaluable for therapeutic response. One patient, with subacute sensory neuronopathy (SSN), improved for at least 15 months, 10 remained stable (eight with anti-Hu and two with anti-Yo antibodies), and 10 deteriorated (eight with anti-Hu and two with anti-Yo antibodies). In seven of the 10 patients who stabilised, the syndrome had already made a plateau when the treatment was started but three patients (one with anti-Hu and two with anti-Yo antibodies) who had still been progressing stabilised for six, eight, and more than 48 months, including one patient with SSN who achieved stabilisation when the neurological dysfunction was only moderate (Rankin scale = 3). Another patient with SSN and initial stable response worsened when IVIg was reduced and improved when it was increased. No significant predictive factors of outcome could be identified but improvement or stabilisation was more frequent in patients with isolated involvement of the peripheral nervous system (62%) than in patients with evidence of CNS damage (37%) at the onset of treatment. Stabilisation in patients with CNS involvement was only achieved when the neurological dysfunction was already severe (Rankin scale > 3). The titres of autoantibodies did not change significantly., Conclusion: Treatment with IVIg at the doses given in the present protocol was not effective in paraneoplastic CNS syndromes associated with antineuronal antibodies. The role of this regime in the treatment of SSN should be further evaluated.

Background: Patients with neuroendocrine-related tumors and paraneoplastic encephalomyelitis (PEM) or paraneoplastic sensory neuronopathy (PSN) develop high titers of antibodies, called anti-Hu, against neuronal proteins expressed in their tumors, usually small cell lung cancer (SCLC). These tumors appear to be more indolent than those not associated with anti-Hu antibodies. The aims of this study were to determine 1) if patients with neuroblastoma (NB) also have anti-Hu antibodies, 2) the correlation between antibody titer and survival, and 3) if coexpression of Hu antigens and major histocompatibility proteins (MHC) by the tumor correlates with the development of anti-Hu associated PEM/PSN:, Methods: Using immunohistochemistry and Western blot analysis, the sera of 109 patients with NB whose neurologic condition was concealed at the time of the study were examined for the presence of anti-Hu antibodies. The expression of Hu antigens and MHC proteins in 50 nonselected NB and 26 SCLC (16 known to be from seropositive and 10 from seronegative patients) was examined using immunohistochemistry., Results: Four Stage 4 NB patients were seropositive and had longer survival (median 86 months) than 71 seronegative patients in the same age group and with the same tumor stage (median survival, 28.5 months). Seventy-eight percent of NB and all SCLC expressed Hu antigens. Overall, 17 of 20 tumors from seropositive patients expressed both Hu and MHC Class I proteins, but only 4 of 30 tumors from seronegative patients expressed both proteins (P < 0.0001)., Conclusions: 1) Some patients with NB develop anti-Hu antibodies; a search for that type of tumor is indicated in seropositive children, 2) most NBs and SCLCs express Hu antigens but only a few are associated with anti-Hu antibodies, and 3) Class I MHC expressed by some Hu antigen-bearing tumors may play a role in the development of anti-Hu associated PEM/PSN:

Paraneoplastic sensory neuropathy (PSN) usually runs a subacute progressive course, leaving the patient with severe sensory dysfunction in weeks to months. We describe five patients with PSN, high titers of anti-Hu antibodies (type 1 antineuronal nuclear autoantibodies), and an indolent clinical course. The patients had a median age of 55 years (range, 41 to 72). Four had small-cell (3) or undifferentiated large-cell (1) lung cancer. Patients presented with mild, asymmetric sensory symptoms; in two, the neuropathy was predominant in the arms. Two patients also had a visceral neuropathy causing gastrointestinal dysfunction. The PSN was stable or progressed very slowly without treatment for a median of 18 months (range, 5 to 32) and remained so after treatment with immunoglobulins (1 patient), chemotherapy (3), or both therapies (1). All patients were ambulatory, leading an independent life up until the time of the last visit or until death from the tumor (2 patients). The median follow-up was 36 months (range, 22 to 52). A paraneoplastic origin should be considered in patients with mild, very slowly progressive sensory neuropathies.

We studied the intrathecal synthesis of the anti-Hu antibody (also called type 1 antineuronal nuclear autoantibody) in 14 patients with isolated paraneoplastic subacute sensory neuronopathy (SSN) and 16 with paraneoplastic encephalomyelitis (PEM). Patients with PEM had higher anti-Hu titers in the CSF (p = 0.003) but not in the serum than those with SSN. Only one patient (7%) with SSN had a positive intrathecal anti-Hu antibody synthesis whereas this was present in 14 (88%) of the 16 patients with PEM (p < 0.0001). The correlation between the intrathecal production of anti-Hu antibodies and PEM supports the role of autoimmune mechanisms in the pathogenesis of PEM. The absent intrathecal synthesis of anti-Hu antibodies in patients with SSN suggests easier accessibility of the systemic immune reaction to the sensory neurons probably due to the partial absence of blood-nerve barrier in the dorsal root ganglia.

Paraneoplastic cerebellar degeneration (PCD) is a rare complication of systemic cancer. PCD may present as a "pure", severe pan-cerebellar syndrome of subacute progression or be only one clinical feature in the setting of extensive CNS disease. The most characteristic form of "pure" PCD is associated with the presence of an anti-Purkinje cell antibody (AB), called anti-Yo, in patients with breast or ovarian cancer. The primary tumor is very often unknown when the cerebellar signs occur, and extensive investigations, including laparotomy or prolonged follow-up may be required to demonstrate its presence. More rarely, others AB than anti-Yo are discovered during PCD. Almost 50% of patients with "pure" PCD do not have circulating anti-neuronal AB. In the cases, the primary cancer is more often known and the clinical course of the cerebellar syndrome may be slower. Cerebellar degeneration may also occur during paraneoplastic encephalomyelitis. In this setting, the cerebellar signs which may be isolated at the onset, become associated with other signs of neuraxis involvement (limbic encephalitis, brainstem encephalitis, myelitis and particularly, subacute sensory neuronopathy) during the course of the disease. When a paraneoplastic encephalomyelitis is associated with a small cell lung cancer, an antineuronal AB called anti-Hu is frequently found. Finally PCD may be associated with the opsoclonus-myoclonus syndrome with the Lambert-Eaton syndrome.

Background: Twenty-four patients with neurologic involvement and anti-Hu antibodies were studied with the aim of defining the type of tumor associated, evaluating whether the clinico-pathologic picture agreed with the concept of paraneoplastic encephalomyelitis (PEM) and evaluating the treatments used., Methods: The study was retrospective with the clinical histories being reviewed to define the neurologic syndromes, their evolution and response to the different treatments, time of appearance and type of tumor as well as the neuropathologic changes in the patients undergoing autopsy., Results: In 18 patients a neoplasm was diagnosed as small cell pulmonary carcinoma (SCPC) in 89% of the cases. The neurologic picture preceded the tumor by an average of five months. The clinical pictures included: sensitive neuropathy (20 patients), cerebellous and truncus encephalicus involvement (8 patients), motor neuropathy (6 patients), cortical involvement (5 patients) and neurovegetative dysfunction (4 patients). In 55% of the patients more than one area was altered. Post mortem studies carried out on 5 patients demonstrated inflammatory infiltrates and neuronal loss in multiple areas of the nervous system. None of the patients improved with treatment. The 9 patients who only received immunodepressants evolved in a way similar to those who were not treated. In 7 of the 11 patients who received antitumoral therapy, the neurologic syndrome stabilized for at least 6 months., Conclusions: The clinicopathological picture and the associated tumor seen in patients with anti-Hu antibodies are identical to those seen in PEM. Antitumoral treatment seems to be more effective than immunodepressant treatment.

We retrospectively evaluated the effect of plasmapheresis (PE) in seven patients with paraneoplastic encephalomyelitis (PEM), small-cell lung carcinoma, and anti-Hu antibodies, and four patients with paraneoplastic cerebellar degeneration (PCD), ovarian or breast cancer, and anti-Yo antibodies. In addition to PE, patients received prednisone (nine), cyclophosphamide (eight), or treatment of the tumor (five). All but one patient were severely disabled by the time PE began. The clinical outcome was compared with that of five patients (PEM, four; PCD, one) who only had treatment of the tumor. Only one of these five patients had a severe neurologic deficit at the onset of the antineoplastic treatment. No patient improved. Two patients treated with PE and antineoplastic therapy and three who only received treatment of the tumor remained stable for at least 6 months. Four of the five patients with a stable course started the treatment when the neurologic deficit was not severe. We conclude that the efficacy of PE with other immunosuppressive therapies in the stabilization of the neurologic deficit is uncertain.

Paraneoplastic syndromes affecting the nervous system are rare and their diagnosis is often difficult when the original cancer is unknown. Recently, high levels of antineuronal antibodies (AB) have been found in serum and CSF of some patients with paraneoplastic syndromes. The anti-Yo AB recognizes 2 proteins of 34 and 62 kd in the cytoplasm of Purkinje cells and in malignant cells of patients suffering from paraneoplastic cerebellar degeneration associated with ovarian and breast cancer. The anti-Hu AB recognizes a 37-40 kd protein in nuclei of neurons and in tumor cells of patients suffering from subacute sensory neuronopathy and encephalomyelitis associated with small cell lung cancer. Other antineuronal AB have been more rarely identified. The presence of high titer of one of these AB in a patient with suspected paraneoplastic syndrome is of great practical interest since it confirms the neurological diagnosis and strongly suggests the location of the primary tumor when the malignancy is unknown. The pathogenetic role of the antineuronal AB is unknown but it is likely that some paraneoplastic syndromes affecting the nervous system are due to an immune reaction against antigens shared by the tumor and the nervous system. To date, no efficient treatment has been found.