Although it has been recognized previously that several markers are present in the cerebrospinal fluid (CSF), their clinical usefulness of these markers in the diagnosis of malignant lymphoma infiltrating to the CNS has not yet been established. In order to determine their diagnostic usefulness as markers of meningeal infiltration by lymphoma cells in patients with adult T-cell leukemia (ATL), we measured some soluble factors in the CSF of patients with ATL and non-ATL patients. Soluble CD4 (sCD4) was highly elevated in all patients with ATL and meningeal infiltration. The CSF level of the soluble interleukin-2 receptor (sIL-2R; sCD25) was markedly elevated in 13 (72.2%) of 18 patients with ATL and meningeal infiltration. Levels of sCD4 and sCD25 in the CSF of patients with ATL and meningeal infiltration were significantly higher than in non-ATL patients (p < .01 and p < .001, respectively). These findings indicate that levels of sCD4 and sCD25 in the CSF are probably associated with meningeal infiltration by leukemia cells expressing CD4 and CD25 on surface membranes. CSF levels of sCD4 in 14 (60.9%) of 23 ATL patients and sCD25 in 13 (72.2%) of 18 ATL patients without meningeal infiltration were moderately elevated. These findings suggest that a small number of leukemic cells which were not detected by conventional CSF examination may have infiltrated the meninges in these patients. Sequential measurements of sCD4 and sCD25 in CSF obtained from patients with meningeal infiltration by leukemic cells showed that sCD4 and sCD25 levels reflected the activity of leukemic meningitis and correlated with the number of cells in CSF. However, the levels of sCD4 in CSF did not fall below the limit of detection even when the number of cells in CSF became normal. It is thought that the level of sCD4 in CSF is a more sensitive marker for detecting the infiltration of leukemic cells in CSF than the number of cells present in the CSF considering the clinical course of two patients with acute type ATL. Therefore, ATL patients with meningeal infiltration should receive treatments until sCD4 levels become normal and not just until the number of cells become normal. Our results also suggest that measurement of CSF levels of sCD4 and sCD25 is useful for the differential diagnosis of aseptic meningitis and meningeal infiltration by leukemic cells in patients with smoldering ATL. We conclude that measurement of soluble factors in CSF plays an important role in diagnosis, prophylaxis and treatment of meningitis in patients with ATL.