Immune mediated paraneoplastic neurological syndromes often become manifest before the underlying malignancy is detected. As a rule, these syndromes do not improve with antineoplastic treatment.1 We report on a case of a patient with small cell cancer with peripheral neurological syndromes that responded favourably to combination chemotherapy.

At the time of admission the patient, a 66 year old woman, had had a combination of peripheral neurological symptoms for 3 months: (a) muscle weakness and muscle pain of the legs so that she could not walk unattended; (b) a numbness of both legs from the foot to the middle of the thigh; (c) dryness of the eyes and mouth; and (d) severe constipation.

Clinical examination showed a load dependent, proximally accentuated symmetric muscle weakness and hypoaesthesia of the legs. The patient was unable to stand or walk without support. The deep tendon reflexes of the arms were decreased on both sides and leg reflexes could not be elicited. No pathological reflexes were detectable. Analysis of CSF yielded normal values for protein content, cell number, and glucose. Besides a slightly increased erythrocyte sedimentation rate (35 mm in the first hour), standard laboratory values showed no abnormalities. Abdominal auscultation and CT were unrevealing.

Electrophysiological investigation (somatosensory evoked potentials of the tibial and median nerves, EMG, and electroneurography) showed delayed latencies and a reduced amplitude of P40 (tibial SEP), N10, and N13 potentials (median SEP) and polyphasic muscle action potentials. Nerve conductance velocities were on the border of the normal range for the motor peroneal and tibial nerves as well as for the sensory sural nerve. Stimulation of the ulnar nerve with 20 Hz yielded an increment of 60%. Besides a polyneuropathy syndrome, the results of the electrophysiological examinations suggested Lambert-Eaton myasthenic syndrome.

Computed tomography showed a small metastasis of an undetectable primary tumour in a mediastinal para-aortal lymph node, which was immediately resected and histologically identified to be small cell cancer.

As an autoimmune paraneoplastic origin of the neurological symptoms, particularly Lambert-Eaton myasthenic syndrome, was suspected, we measured autoantibodies against presynaptic voltage gated calcium channels by immunoprecipitating ¹²⁵I-conotoxin binding proteins. The antivoltage gated calcium channel titre of 105 pM was positive, although not extremely so. The titre of antinuclear antibodies was 1:960; most of them were identified as anti-Scl-70. High titre (1:6000) antineuronal antibodies were seen on immunofluorescence of fresh frozen sections of rat cerebellum. Anti-Hu appeared as a prominent band in a western blot of recombinant Hu protein.

As treatment with 3,4-diaminopyridine (10 mg every 4 hours) improved the muscle weakness but had no effect on constipation and the sensory disorder, these symptoms were thought to indicate anti-Hu syndrome. Four cycles of chemotherapy (ACO-I scheme), however, improved all neurological syndromes. Muscle pain and leg hypoaesthesia disappeared, and muscle strength was fully restored so that the patient was able to walk unattended for long distances. The features of Lambert-Eaton myasthenic syndrome were also no longer detectable by EMG. Constipation was alleviated. All types of autoantibodies declined considerably. The antinuclear antibodies titre was reduced to 1:32; immunoprecipitation for antivoltage gated calcium channel antibodies yielded 22.7 pM, which is within the normal range. Antineuronal antibodies were barely detected on sections of rat cerebellum at a serum dilution of 1:32. The anti-Hu signal was also clearly reduced on a western blot. The clinical, physiological, and laboratory findings before and after chemotherapy—that is, six months after admission—are summarised in the table. Combination chemotherapy brought about a considerable and unexpected improvement of two peripheral neurological syndromes (Lambert-Eaton myasthenic syndrome and anti-Hu syndrome with gastrointestinal pseudoobstruction, muscle pain, and sensory deficits) and a concomitant decrease in autoantibody titres in a patient with small cell cancer.

The results of treatment for paraneoplastic neurological syndromes are still discouraging, despite widespread use of chemotherapy combining antineoplastic and immunosuppressive activity. The reasons for the frequent failures may be irreversible loss of neurons due to autoimmune attack and the differential susceptibility of autoreactive B and T cells inside and outside the blood-brain barrier to the infusion of immunosuppressive drugs. Autoimmune paraneoplastic syndromes involving the CNS are probably caused by intrathecal autoreactive lymphocytes, which may be less accessible to inhibiting drugs than immune cells outside the blood-brain barrier. This hypothesis is supported by the case presented here and by reports of the differential response of central and peripheral syndromes to treatment.2 3

By contrast with the refractory central syndromes, tumour treatment for Lambert-Eaton myasthenic syndrome has been shown to have beneficial effects in many patients. In at least two cases, a differential response to immunotherapy of peripheral and central syndromes was found in the same patients.2 3 In one case,2 a patient with paraneoplastic Lambert-Eaton myasthenic syndrome and subacute cerebellar degeneration due to an initially undetected small cell lung cancer, chemotherapy had a beneficial effect on Lambert-Eaton myasthenic syndrome, but not on subacute cerebellar degeneration. In the second case,3 the same combination of peripheral syndromes was caused by a non-Hodgkin’s lymphoma. Immunosuppression improved neuromuscular, but not cerebellar symptoms.

Paraneoplastic autoantibodies seem to arise from immunological cross reactions between tumour antigens and normal target proteins. They may take part in the immune response to malignancies, or simply be an epiphenomenon of an immune process triggered by the neoplasm.4 The prognosis of patients who mount a humoral antitumour response is usually better than that of patients without response.5 The immune process characterised by production of paraneoplastic autoantibodies, however, is a two edged sword. The neurological disorders often associated with some paraneoplastic autoantibodies are so severe that they require immediate treatment, even before a tumour is detected. In such cases, immunosuppressive treatments bear the risk of supporting tumour growth by inhibiting the ongoing immune attack against the neoplasm without interfering with the metabolism of the malignant cells. Therefore, all efforts should be directed at identifying the underlying malignancy. As soon as the presence of a tumour is confirmed, chemotherapy should be initiated to control its growth and to suppress autoantibody production, at least outside the CNS.