Aims: To investigate the association of acute parvovirus B19 infection with new onset of acute lymphoblastic and myeloblastic leukaemia. Methods: Cerebrospinal fluid (CSF) samples from patients with acute myelogenous leukaemia (AML) at diagnosis (n = 2) and acute lymphoblastic leukaemia (ALL) at diagnosis (n = 14) were analysed for parvovirus B19 DNA by means of nested polymerase chain reaction. In addition, samples from patients with benign intracranial hypertension (BIH) (n = 10) and hydrocephalus (n = 13) were tested as controls.Results: Four leukaemia cases were positive-common ALL (n = 2), null cell ALL (n = 1), and M7 AML (n = 1)-whereas all controls were negative (Yates corrected x 2 value, 3.97; p = 0.046; odds ratio, 16.92; confidence interval, 1.03 to 77.18). All four patients were significantly anaemic, but none was encephalitic or had evidence of central nervous system leukaemia. In three of these patients, serum tumour necrosis a, interferon c, interleukin 6, granulocyte-macrophage colony stimulating factor (range, 34.93-3800.06 pg/ml), and macrophage chemoattractant protein 1 were detectable. All of these four patients carried at least one of the HLA-DRB1 alleles, which have been associated with symptomatic parvovirus B19 infection. Conclusion: Erythroid suppression and immune cell proliferation are both associated with B19 infection and may also be important in the pathogenesis of acute leukaemia. P arvovirus B19 is the aetiological agent of erythema infectiosum, transient bone marrow aplastic crises, arthritis, non-immune hydrops fetalis, and chronic pure red cell aplasia in immunocompromised persons. More than 20 reports describe an association between acute lymphoblastic leukaemia (ALL) and persistent B19 infection which, until recently, has been assumed to represent an opportunistic infection, the chronicity of which results from immunosuppression. However, several reports describe B19 infection preceding and mimicking ALL.