To characterize the prodromal phase of adult T-cell leukemia (ATL), a prospective follow-up study was conducted on 50 carriers in a putative pre-ATL state. This state was defined by the presence of molecularly- detectable monoclonal proliferation of human T-lymphotropic virus type I (HTLV-I)-infected T lymphocytes, and the absence of clinical symptoms of leukemia. The median observation time was 50 months. The pre-ATL subjects were divided into two groups according to initial white blood cell (WBC) counts: group A, those with a normal WBC count (9,000/microL) (n = 30), and group B, those with an increased WBC count (9,000 to 15,000) (n = 20). Comparisons were made between the two groups and with a group of 25 patients with chronic ATL (group C) who had WBC counts of more than 15,000. Significant differences in survival rate were found between groups A and B (10-year survival 65.7%) and group C (32.8%) (P < .01), and between group A (10-year survival 90.0%) and group B (52.1%) (P < .05). The incidence of transformation to overt ATL was 10% (3 of 30) in group A and 50% (10 of 20) in group B (P < .01). In six transformed cases (one in A and five in B) we found exactly the same integration sites in pre-ATL and overt ATL phases, confirming the multistep leukemogenesis hypothesized for this disease. However, the pre-ATL subjects could be divided into two distinct prognostic groups based on the initial WBC count; those with good and those with poor prognosis. Although the 10% transformation rate (2.5% annually) in group A seemed to be extremely high compared with that in the general population of HTLV-I carriers (around 0.06% to 0.4% annually), the majority of group A subjects and some in group B showed stable clinical courses without transformation. Further, development of ATL was not observed in four group A subjects with HTLV-I-associated myelopathy (HAM), which is rarely associated with ATL. We propose to call this group of rather benign HTLV-I carriers “HTLV-I carriers with monoclonal proliferation of T lymphocytes (HCMPT).” Thus far we have been unable to identify reliable parameters other than WBC counts that prospectively distinguish HCMPT from the true pre-ATL state, in which there is a high probability of developing ATL. Further clinical and biologic approaches should elucidate the natural history of the HTLV-I carrier state and early events in ATL leukemogenesis.