Article Text


Immunoglobulin and lymphocyte decrease concurrent with adverse reactions induced by methotrexate for RA
  1. Department of Allergy and Immunological Diseases
  2. Tokyo Metropolitan Komagome Hospital
  3. Tokyo, Japan
  1. Dr Shigeko Inokuma, Tokyo Metropolitan Komagome Hospital, 3–18–22 Honkomagome, Bunkyoku, Tokyo, 113–8677, Japan

Statistics from

The limiting factor in low dose pulse methotrexate treatment for rheumatoid arthritis (RA) has been its toxicity.1 2We recently treated a female patient with RA, in whom pneumonitis and granulocytopenia developed during methotrexate treatment; her white blood cell count was 1.10×109/l and Pao 2 was 37 mm Hg. Before treatment, at the time of development of adverse reactions, and after recovery after methotrexate was withdrawn, her IgG levels were 17.99, 10.15, 16.75 g/l; IgA 5.14, 3.69, 4.33 g/l; IgM 1.73, 1.06, 1.36 g/l; and lymphocyte count 1.96, 0.42, 1.56×109/l, respectively. We then investigated whether immunoglobulin levels and lymphocyte count decrease when adverse reactions to methotrexate treatment develop.

One hundred consecutive patients with RA (80 women and 20 men, mean (SD) age 57.5 (9.2) years) receiving between 2.5 and 15 mg of methotrexate weekly in Tokyo Metropolitan Komagome Hospital were followed up from 1991 to 1998. When the patients did not respond and had no adverse reactions, the dose was increased by 1.25 to 2.5 mg/week. Response to treatment, assessed by the patient's impression of improvement, a decrease in swelling and pain of more than two joints, a decrease of >20 mg/l in the C reactive protein (CRP) level, adverse reactions, lymphocyte and eosinophil counts, serum concentrations of immunoglobulins, γ fraction, rheumatoid factor, and albumin were studied.

Sixteen adverse reactions occurred in 15 patients; the reaction affected the liver (six patients), the lung (three), the skin (three), the bone marrow (three), and the oral mucosa (one). They recovered after methotrexate was discontinued or reduced, without steroid treatment. Thirteen of these 15 patients showed a mean (SD) decrease in CRP from 63 (36) to 32 (55) mg/l, whereas all 22 non-responders who had no adverse events showed a decrease in CRP from 46 (39) to 41 (34) mg/l. A significant relation was found between a good response to treatment and the appearance of adverse reactions. The patients with adverse reactions had a higher creatinine level, and more frequent use of steroid at high dose (7.9 (7.8) v 2.8 (3.5) mg/day prednisolone). The patients with the higher creatinine level were older.

The albumin level increased more in responders. The rheumatoid factor titre decreased in responders and in patients without adverse reactions. The eosinophil count did not correlate either with response or adverse reactions.

After treatment the levels of IgG, IgA, and IgM, γ fractions, and lymphocyte count in the 15 patients who had adverse reactions were significantly reduced compared with the values before treatment. The reductions and reduction ratios compared with pretreatment values were significantly greater in patients with adverse reactions than in those without. Table 1 gives the results obtained and the threshold values that could differentiate between patients with and without adverse reactions. When the patients were grouped according to therapeutic response, significant reductions were seen only in the levels of immunoglobulins and γ fractions, but no reduction was seen in lymphocyte counts, in 78 responders. The greater decreases in responders than in non-responders were seen only in IgG and IgA levels. The reductions and reduction ratios of immunoglobulins were greater in patients with adverse reactions, grouped according to toxicity (30–35%, table 1), than in those with therapeutic response grouped according to efficacy (13–14%, data not shown).

Table 1

Pretreatment value, decrease, decrease ratio, and threshold value of immunoglobulin levels and lymphocyte count in patients used to differentiate between patients with and without adverse reactions. Values are shown as mean (SD)

Our study shows that when a patient's immunoglobulin levels and lymphocyte count decrease globally by as much as 25% or more from the pretreatment level, this decrease is suggestive of toxicity. Conversely, in patients without adverse reactions, the decreases were less than 20%. The clinical improvement contributed only partially to the reductions; steroid treatment was not likely to have been the cause either, as they had been given for a long time without a significant change in the dose.

Recently, we reported that the immunoglobulin level decreases with adverse reactions, during a disease modifying antirheumatic drug, bucillamine, treatment.3-5 A reduction in interleukin 6 level was reported to parallel an improvement during methotrexate treatment.6 7 The reduction in lymphocyte numbers is controversial.8 9 Immunomodulation might relate mainly to adverse reactions, whereas the effect might appear owing to anti-inflammatory mechanisms.10

It can only be speculated whether consumption or leakage of immunoglobulin plays a part in the previously supposed mechanism of acute hypersensitivity or cytotoxicity, or in an independent epi-phenomenon. There is the encouraging possibility that monitoring the immunoglobulin level and the lymphocyte count might disclose life threatening reactions and enable the doctor to know when to reduce the dosage or to stop the drug entirely.


The authors thank Dr Victoria Elegant and Ms Keiko Miyahara for their help.


View Abstract


  • Email: inokuma-k{at}

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.