Malignancy and Sensory Neuropathy of Unexplained Cause:  A Prospective Study of 51 Patients FREE

PERIPHERAL SENSORY neuropathy of an unexplained cause, sometimes defined as primary¹ or idiopathic,² is a rare disease that has been linked with cancer since its first description.¹ Several series of patients have been described^3- 9 but little is known about the frequency of the association between the neuropathy and cancer. The aim of our prospective study was to obtain further information on this topic.

The Hospital Riuniti in Bergamo, Italy, is a 1205-bed general hospital that serves a population of 896000. On January 1, 1988, we started to examine prospectively patients with peripheral sensory neuropathy with no identified cause. The patients were selected on the basis of the following criteria: (1) symptoms and signs of sensory loss at the extremities; (2) absence of deep tendon reflexes; (3) preservation of strength in the neurologic disability score¹⁰; (4) absence or reduction in amplitude of sensory nerve action potentials, and no motor involvement in the results of the electrophysiologic study, performed as described by Windebank et al⁸; fibrillation potentials were allowed in 2 leg muscles; and the possible diagnosis of chronic inflammatory demyelinating polyneuropathy was ruled out according to the criteria of the Ad Hoc Subcommittee of the American Academy of Neurology¹¹; (5) no history of cancer; (6) no history or evidence of diabetes mellitus, alcoholism, uremia, porphyria, amyloidosis, vasculitis, diseases of the connective tissue, cryoglobulinemia, human immunodeficiency virus infection, chronic viral hepatitis, drug addiction, toxin exposure, B₁₂ or folate hypovitaminosis, pyridoxine abuse, inherited diseases, acute idiopathic polyneuritis, or recent febrile illness; and (7) neither serum anti–myelin-associated glycoprotein antibodies¹² nor antisulfatides antibodies.¹³

We decided to include in the study patients with monoclonal gammopathy of uncertain significance (MGUS) without reactivity to myelin-associated glycoprotein, since no relationship has been established between those gammopathies and neuropathies.¹⁴

The aims of the study and what was involved were clearly explained to both the patients and their relatives. Written consent was obtained.

Patients included in the study underwent cerebrospinal fluid examination; morphologic examination of the sural nerve¹⁵; search for circulating immunoglobulins against neural antigens,^16- 18 called anti-Hu, anti-Yo, and anti-Ri antibodies, detected with previously described methods¹⁹; and a protocol of investigation for tumor detection, including determination of the humoral markers for neoplasms (carcinoembryonic antigen, α-fetoprotein, Ca 19-9, tissue polypeptide antigen, Ca 15-3 in women and prostate-specific antigen in men), search for occult blood in stool specimens, chest radiography, total body computed tomographic scanning, echography of the abdomen, pelvis, and thyroid gland, endoscopic examination of the esophagus and stomach, urologic or gynecological examinations, and rectal examination by a surgeon. The medical treatment was based on symptoms.

The patients were invited to return every 6 months for a checkup that included an electrophysiologic study and a search for cancer. This involved an accurate medical examination, routine blood examinations, investigation for occult blood in the stool specimens, echography of the abdomen, pelvis, and thyroid, urologic or gynecological examination, and chest radiography. If found, any malignancy was histologically confirmed.

The inclusion of patients in the study ended December 31, 1995. The end point of the follow-up was December 31, 1996.

From 1988 through 1995, we observed 363 patients with peripheral sensory neuropathy. Of 363 patients, the causes of the observed peripheral sensory neuropathy included the following: diabetes mellitus (183), alcohol abuse (41), hepatitis C virus–associated cryoglobulinemia (24), uremia (11), recent febrile illness (10), arteritis (6), history of acute idiopathic polyneuritis (4), anti–myelin-associated glycoprotein IgMκ MGUS (4), human immunodeficiency virus infection (3), occupational poisoning (3), inherited disease (2), primary amyloidosis (2), Sjögren disease (1), neurosarcoidosis (1), previous tumor without chemotherapy (6, of whom there were 2 with gastric tumor, 2 with non–Hodgkin lymphoma, 1 with Hodgkin lymphoma, and 1 with uterine tumor), cisplatin treatment (4, of whom there were 2 with colon adenocarcinoma, 1 with lung adenocarcinoma, and 1 with ovarian cancer), vincristine sulfate treatment (4, all of whom had non–Hodgkin lymphoma and were selected for a high-dose protocol of 2.5 mg/m²), and prolonged metronidazole treatment (1). The cause of the neuropathy was not identified in 53 patients, 2 of whom refused further scheduled visits. Thus, we examined and followed up 51 patients, 42 men and 9 women, with a mean age of 64.5 years (range, 19-80 years). Of the 51 patients, 9 patients (17.6%) had an MGUS: 5 had the IgGκ isotype, 2 had the IgMκ isotype, and 2 had the IgAκ isotype. The range between the onset of neuropathy symptoms and clinical diagnosis of peripheral sensory neuropathy was from 3 to 72 months (mean, 13.9 months).

Patients experienced the following physical manifestations of the disease: numbness and/or a tight or swollen sensation in the extremities (44 [86.3%]); tingling or prickling and/or a burning or aching sensation (33 [64.7%]); unsteadiness of gait (24 [47.1%]); fatigue (13 [25.5%]); and cramps in the calves (24 [47.1%]). During the clinical examination, 41 patients (80.4%) had some impairment of superficial sensation and 42 (82.4%) had impairment of deep sensation.

The results of the electrophysiologic study revealed the absence of sensory nerve action potentials in the sural nerve in 34 patients (66.7%) and the ulnar nerve in 22 patients (43.1%) and fibrillation potentials in the leg muscles in 10 patients (19.6%).

Cerebrospinal fluid protein content was increased in 13 patients (25.5%), 6 of whom had an MGUS. Oligoclonal bands were found only in patients with MGUS, and they corresponded to the serum monoclonal proteins. No malignant cells were found in the cerebrospinal fluid.

A morphologic study was made of the sural nerve in 41 patients (80.4%), who all showed signs consistent with a primary axonopathy. Chronic axonopathy was evident in 34 patients, while the remaining 7 patients had severe acute axonopathy with myelin ovoids at different stages of degeneration.

The results of immunological studies demonstrated antineural antibodies in only 1 patient who had anti-Hu antibodies.

The mean follow-up of the patients at the end point was 51.4 months (range, 14-94 months); during that time, the symptoms and signs tended to stabilize or progress slowly. Mild to moderate muscle atrophy developed in 14 patients (27.5%), 8 of whom had a mild reduction in the distal muscle strength. The results of the electrophysiologic studies revealed that 11 patients (21.6%) had a slight reduction in the motor conduction velocity in the peroneal and tibial nerves (never below 10% of the lower limit of normal).

Cancer was diagnosed in 18 patients (35.3%) after a mean interval of 27.9 months (range, 3-72 months) after the onset of symptoms of neuropathy. The patients, 16 men and 2 women, had a mean age of 66.2 years when the clinical diagnosis of neuropathy was made.

Malignancy was discovered in the following: liver in 4 patients (all had a primary hepatoma); lymphatic glands in 3 (all had non–Hodgkin lymphoma); the bladder in 3; the lungs in 2 (both had small cell lung cancer); the prostate gland in 2; the breast in 1; a sublingual gland in 1; the pancreas in 1; and bone in 1 (metastatic sarcoma).

Three patients with MGUS developed primary hepatoma (IgM MGUS), non–Hodgkin lymphoma (IgA MGUS), and metastatic sarcoma (IgG MGUS).

Table 1 shows the characteristics of the 51 patients.

The patient with anti-Hu antibodies was diagnosed as having small cell lung cancer 9 months after the diagnosis of neuropathy and 16 months after the onset of neurologic symptoms.

No paraneoplastic autoantibodies were found in the serum samples of the other patient who developed the same neoplasm, neither when neuropathy was diagnosed nor when the test was repeated after the detection of the cancer.

After cancer treatment, 4 patients showed neurologic improvement. They included the patient with small cell lung cancer but without anti-Hu antibodies, the patient with sublingual gland tumor, 1 patient with primary hepatoma, and 1 patient with bladder cancer.

All of them improved dramatically within a few weeks of treatment with the exception of the patient with bladder cancer, who improved slowly and did not complain of pain or a burning or aching sensation. This patient had a nearly complete recovery of the vibratory sensation after 12 months of tumor eradication.

Of the 4 patients, 3 were alive at the end point (follow-up of tumor at 58, 34, and 22 months), and none complained of neurologic relapses. The fourth patient, who had the small cell lung cancer, was healthy for 6 months after the treatment of the tumor, but her condition declined rapidly because of multiple cerebral metastases.

Of the other patients, 3 survived long enough to be periodically examined. These 3 patients included the patient with breast cancer, 1 patient with non–Hodgkin lymphoma, and 1 patient with bladder cancer. All 3 patients experienced a slow progression of the neurologic disease, but none lost the ability to walk.

The other patients died a few months after the diagnosis of cancer, including the patient with anti-Hu antibodies.

Malignancy was discovered in more than one third of our patients approximately 2 years 4 months after the onset of symptoms of sensory neuropathy.

The high tumor rate in our patients with peripheral sensory neuropathies of an unexplained cause was not subject to bias for 3 reasons. First, our department is not a specialty neuromuscular clinic, and it is located in a general hospital. Second, all but 2 of the patients came from the surrounding area of Bergamo. Third, follow-up examinations were conducted in all but 2 of the initially examined 53 patients. Compared with the most recent studies^8- 9 reported, our tumor rate is higher. In their study of 42 patients, Windebank et al⁸ found malignancy in only 2 patients who died of metastatic cancer 11 and 19 years after the onset of neuropathy, whereas in their study of 75 patients with neuropathy of an uncertain cause, Notermans et al⁹ found no tumor in 20 patients with sensory neuropathy.

However, these studies were somewhat different. The study by Windebank et al⁸ was retrospective and used less restrictive inclusion criteria than our study used. We carried out a prospective study and did not include patients with previous febrile illness, Sjögren disease, vasculitis, or chronic viral hepatitis as they did. Furthermore, due to the design of the study it is possible that those authors did not include patients who had developed a malignancy within 1 or 2 years of the diagnosis of neuropathy. The follow-up in the study by Notermans et al⁹ was shorter than our follow-up, being limited to 6 months after the clinical diagnosis of neuropathy.

Therefore, we suggest that it is reasonable to suspect malignancy in patients with peripheral sensory neuropathy without any identified cause and also any associated disease, and those patients should be observed for an appropriately long time. However, from a practical point of view the results of our study were disappointing.

For several reasons, the discovery of an occult neoplasm was not easy at diagnosis of neuropathy despite our detailed and extensive search. First, it could be that since the time from neuropathy diagnosis to neoplasm discovery is so long, usually longer than a year, the size of the malignancy was below the baseline sensitivity of the investigation methods. A second reason is that cancer had no predominant localization. For example, lung cancer, which is historically the most frequent neoplasm associated with sensory neuropathy,²⁰ was found in only 2 patients. In addition, there was no apparent common embryonic origin of the neoplasms that we discovered.

A third reason is that none of the information resulting from the investigations, such as clinical examinations, electrophysiologic examinations, cerebrospinal fluid analysis, or sural nerve biopsy, could reveal the development of a neoplasm, which is the commonly reported experience.²¹

The finding of anti-Hu antibodies¹⁶ can draw clinical attention to the lungs,²² but results that are negative for anti-Hu antibodies do not rule out small cell lung cancer.²³

The main outcomes of our study are to highlight the need to look for cancer in patients with a peripheral sensory neuropathy of unexplained cause and to note that a malignancy can appear outside the common sites primarily involved in the pathogenesis of paraneoplastic neuropathies. However, the results of our study reveal that the chances of anticipating a diagnosis of an associated malignant neoplasm are low. Furthermore, our data are from a relatively small population and cannot be compared with those from community studies, preventing the translation of our experience into a survey for tumor types. Therefore, there is a need for further data to design an appropriate clinical strategy for patients with peripheral sensory neuropathy of unexplained cause.

Since the manuscript was accepted for publication, another patient from our study developed a high-grade malignant neoplasm. In April 1997, this 76-year-old patient was diagnosed as having a highly undifferentiated adenocarcinoma of the prostate gland, 56 months after the onset of symptoms and 47 months after the diagnosis of neuropathy.

Accepted for publication December 2, 1997.

The immunological studies were conducted with financial support (grant 613) from Telethon-Italy, Milan (Dr Nemni).

The authors also thank the Foundation Francesca Petazzi, Bergamo, Italy, for the statistics regarding the incidence of cancer and the mortality rate in the area of Bergamo.

Reprints: Massimo Camerlingo, MD, Neurology Division, Hospital Riuniti, Largo Barozzi 1, 24128 Bergamo, Italy.