Aseptic Meningitis Associated with High-Dose Intravenous Immunoglobulin Therapy: Frequency and Risk Factors

Methods

Patients

From 1990 to 1994, clinical trials of high-dose intravenous immunoglobulin (2 g/kg body weight) for patients with various neuromuscular diseases have been done according to treatment protocols approved by the Institutional Review Board of the National Institute of Neurological Disorders and Stroke. All patients were admitted to the National Institutes of Health Clinical Center and gave informed consent for participation in a treatment trial with intravenous immunoglobulin.

Before and after each infusion, all patients had neurologic examinations. The diagnosis of aseptic meningitis was based on the development of all the following symptoms: severe headache with nuchal rigidity, drowsiness or lethargy, fever, photophobia, painful eye movements, and nausea and vomiting. In all patients with these symptoms, cerebrospinal fluid was cultured for bacteria and fungi and was analyzed for cell count, total protein, glucose, IgG, and albumin. The levels of IgG, albumin, and glucose were determined in serum drawn concomitantly with the cerebrospinal fluid. The kinetics of IgG within the cerebrospinal fluid compartment and the percentage of IgG transferred from the serum were calculated according to standard methods [20]. In one patient, clearance of the cerebrospinal fluid IgG was measured by a second cerebrospinal fluid analysis done after 24 hours.

To assess whether aseptic meningitis was related to the rate of infusion of intravenous immunoglobulin, to a specific lot, or to a single commercial preparation, five patients in whom aseptic meningitis developed after the first infusion received another infusion 1 month later with immunoglobulin from a different lot (three patients), a different commercial preparation (two patients), a different infusion rate (three patients), or a different dosing schedule (2 g/kg dose over 3 to 4 days) to further slow the infusion rate (two patients). Six different lots and four commercial products were used in 18 infusions in patients with aseptic meningitis.

Results

Aseptic Meningitis Compared with Simple Headaches

We studied 54 consecutive patients, 24 women and 30 men aged 7 to 79 years old (mean age, 46 years). Their primary neuromuscular diseases were inflammatory myopathy (22 patients), paraproteinemic polyneuropathy (12 patients), amyotrophic lateral sclerosis (10 patients), chronic inflammatory demyelinating neuropathy (5 patients), dystrophy (4 patients), and multifocal motor neuropathy with conduction block (1 patient). Twenty-two patients were receiving maintenance corticosteroid therapy during the infusions, with dosages ranging from 2.5 to 60 mg/d. One hundred fifty-nine infusions were given (average, 3 per patient), as previously reported [7, 9, 12].

In 6 of 54 patients (11%; CI, 4% to 23%), aseptic meningitis developed within 24 hours after completion of the high-dose intravenous immunoglobulin infusions. All patients were treated with narcotic analgesics and antiemetic agents. Two patients had an incomplete response to intramuscular injections of meperidine but responded better to intramuscular injections of ketorolac. In 2 patients, headache and neck stiffness were so acute and severe that an emergency computed tomographic scan was obtained, but neither patient had evidence of a subarachnoid hemorrhage. In all patients, symptoms cleared without sequelae within 3 to 5 days. The demographic characteristics of the patients with aseptic meningitis were not different from those of patients without aseptic meningitis. Further, in both groups, the underlying disease was equally represented. Specifically, meningitis developed in 3 of 22 patients with inflammatory myopathy, 1 of 12 patients with paraproteinemic neuropathy, 1 of 11 patients with motor neuron disease (including 1 with conduction block), and 1 of 4 children with muscular dystrophy.

In contrast to aseptic meningitis, headaches occurred in 27 (56.2%) of the other 48 patients during the first infusion. In 15 patients, the headache recurred with subsequent infusions. The headaches were generally mild, and patients responded to nonsteroidal anti-inflammatory drugs or to a reduction of the infusion rate and did not require additional hospitalization or any diagnostic procedure.

Cerebrospinal Fluid Studies

Lumbar puncture was done within 11 to 46 hours after completion of the intravenous immunoglobulin infusion in five of six patients with aseptic meningitis. The cerebrospinal fluid was slightly cloudy in the three patients with the highest leukocyte count. In one patient (patient 2, Table 1), the lumbar puncture was mildly traumatic. Pleocytosis was noted in four of the five patients who had lumbar puncture (Table 1). The patient (patient 5) without cerebrospinal fluid pleocytosis had a low peripheral leukocyte count (1.4 × 10⁶/L). In three patients, the cerebrospinal fluid showed mild eosinophilia (eosinophils, 3% of the total leukocyte count) without an associated increase in the number of eosinophils in the peripheral blood. The cerebrospinal fluid glucose level was normal, and the cultures for bacteria and fungi were negative.

Elevations in the cerebrospinal fluid IgG ranged from 1.5 to 7 times the upper limit of normal (Table 1). The cerebrospinal fluid IgG indices were normal. In one patient (patient 1), a second cerebrospinal fluid analysis 24 hours later showed a 46% decrease in cerebrospinal fluid IgG, with only an 11% decrease in the concurrently drawn serum IgG.

Migraine

Overall, aseptic meningitis developed in 4 of 8 patients with a history of migraine but in only 2 of 48 patients without such a history (P = 0.003; Fisher exact test). One child with a family history of migraine had his first migraine headache within 24 hours of the intravenous immunoglobulin infusion.

The initial symptoms of aseptic meningitis often mimicked the patients' usual migraine headache, but their symptoms intensified further as fever and meningismus developed. Recurrent meningeal symptoms developed in four of the six patients with each subsequent infusion. In the two patients without a history of migraine, aseptic meningitis did not recur when different lots or commercial intravenous immunoglobulin preparations were used in repeat monthly infusions. Slowing the rate of infusion in three patients or spreading the infusion over more days in two patients did not alter the development or the severity of aseptic meningitis. Corticosteroids did not appear protective; three of the 6 patients with aseptic meningitis and 19 (40%) of the 48 patients without aseptic meningitis were receiving corticosteroids at doses ranging from 0.1 to 1 mg/kg and 0.05 to 1 mg/kg, respectively. The development of aseptic meningitis did not correlate with the age or the type of underlying neuromuscular disease.

Discussion

Headaches are a common side effect of intravenous immunoglobulin therapy. In our series, mild to moderate headaches occurred in 56% of the patients and were usually short-lived, responding to acetaminophen, nonsteroidal anti-inflammatory agents, or the slowing of the infusion rate. Of greater concern, we documented clinically severe aseptic meningitis within 24 hours after completion of the infusion in 11% of the 54 patients receiving treatment with high-dose intravenous immunoglobulin for immune-mediated neuromuscular diseases. The headache in patients with aseptic meningitis was intense and pounding; was associated with meningismus, photophobia, vomiting, fever, and cerebrospinal fluid pleocytosis; and did not respond to the above-mentioned measures. Aseptic meningitis, although short-lived and without sequelae, was especially frequent in patients with migraine and caused severe clinical symptoms that necessitated 2 to 3 days of additional hospitalization and treatment with narcotic analgesics. The findings suggest that the morbidity associated with aseptic meningitis may limit some patients from receiving intravenous immunoglobulin at home or in outpatient clinics.

Whether the frequently encountered mild headache is caused by an incomplete form of subclinical chemical meningeal irritation is uncertain. The various underlying diseases that we treated are not known to predispose the patients to development of aseptic meningitis. Further, aseptic meningitis was noted in two patients with neuropathy who had acellular cerebrospinal fluid before intravenous immunoglobulin therapy and in four patients with myopathy in whom the cerebrospinal fluid was known to be normal.

The cause of aseptic meningitis in patients treated with intravenous immunoglobulin is also unknown. The possibility of an allergic hypersensitivity reaction caused by the direct entry of the preparation into the cerebrospinal fluid compartment is supported by the mild cerebrospinal fluid eosinophilia noted in three of our patients. Other drugs, such as co-trimoxazole [21], azathioprine [22], trimethoprim-sulfamethoxazole [23], ibuprofen [24], and OKT3 monoclonal antibody [25], may cause aseptic meningitis with cerebrospinal fluid eosinophilia several hours after exposure [20, 23].

Serum immunoglobulins, especially IgG, can cross the blood-brain barrier [26] and may be another offending factor. When the blood-nerve barrier is impaired, as in patients with neuropathies or radiculopathies, even the high-molecular-weight IgM can cross into the cerebrospinal fluid [27]. Within this dynamic equilibrium, the turnover of IgG within the cerebrospinal fluid is thought to approximate the cerebrospinal fluid reabsorption rate [26]. In the patient in whom a second cerebrospinal fluid analysis was done, the cerebrospinal fluid IgG decreased by 46% within 24 hours, whereas the serum IgG decreased by 11% during the same period. This suggests that the infused IgG remained within the cerebrospinal fluid compartment for at least 48 hours, even though its clearance rate is faster than that of serum IgG. The infused IgG, derived from a pool of approximately 5000 donors [28], is allogenic and may interact within the cerebrospinal fluid compartment with antigenic determinants on the endothelial cells of the meningeal microvasculature. This interaction can release cytokines and lead to an inflammatory reaction, as supported by the observed cerebrospinal fluid pleocytosis.

The increased susceptibility of patients with migraines to the development of aseptic meningitis implies increased sensitivity of their meningeal vasculature to exogenous IgG. In acute migraine, the intracranial vasculature may react to various circulating factors, such as serotonin, histamine, catecholamines, prostaglandins, complement, circulating immune complexes, and cytokines [29, 30]. Some of these factors may be present in the intravenous immunoglobulin preparations, or their production may be triggered by the intravenous immunoglobulin itself [31]. Although cerebrospinal fluid pleocytosis can rarely develop during acute migraines, the cell count is lymphocytic and almost never exceeds 100 × 10⁶/L [32]. Classic migraines without meningitis were provoked by intravenous immunoglobulin in one recently described patient [33], and one of our patients who had a family history of migraines developed his first migraine attack within 24 hours of intravenous immunoglobulin infusion.

Sensitivity to the product's stabilizing agents, such as polyethylene glycol, maltose, sucrose, and glycine, which may be present in the commercial preparations of intravenous immunoglobulin, could also be responsible for the aseptic meningitis. Because aseptic meningitis developed in the same patients who received different commercial intravenous immunoglobulin preparations, this is less likely. Despite reports that corticosteroids may protect against the side effects of intravenous immunoglobulin [34], this did not appear to be the case in our patients for the maintenance doses they were receiving for this underlying disease.

Why did we document such a relatively high rate of intravenous immunoglobulin-associated aseptic meningitis, especially in patients with migraines? We believe this is because the patients were followed closely in the hospital, whereas most patients are sent home before symptoms develop. Clinicians using intravenous immunoglobulin should anticipate this possibility and warn the patients that a more protracted hospitalization may be required for symptomatic therapy. In two of our patients who did not have a history of migraines, changing the lot or product was, for unknown reasons, effective in preventing the recurrence of aseptic meningitis; thus, a different commercial intravenous immunoglobulin preparation or lot should be considered in such patients if additional infusions are required.