OBJECTIVE To determine whether subjects coinfected with HTLV-I and HIV have a higher frequency of myelopathy than subjects singly infected with HIV. METHODS A prospective, nested case-control study of HTLV-I and HIV coinfected (cases) and HIV singly infected adults (controls) participating in a prospective HIV cohort study at a university hospital outpatient HIV clinic in Rio de Janeiro, Brazil. METHODS Subjects were evaluated for evidence of myelopathy by a neurologist unaware of their HTLV serologic status. Patients with at least two pyramidal signs, such as paresis, hypertonicity or spasticity, hyperreflexia, clonus, diminished or absent superficial reflexes, or the presence of pathologic reflexes (e.g., Babinski or Hoffmann), were defined as having myelopathy. Myelopathy severity was quantified using the Kurtzke Functional Disability Scale (FDS); patients with FDS scores > or = 4 were considered to have significant myelopathy. Selected patients with myelopathy underwent lumbar puncture for the evaluation of intrathecal synthesis of HTLV-I antibodies. RESULTS Of 15 coinfected subjects, 11 (73%) had evidence of myelopathy versus 10 of 62 subjects (16%) with HIV single infection (adjusted odds ratio [OR] = 13.0, p = 0.00002). When only myelopathy patients with FDS scores of > or = 2 or > or = 4 were included, the association between coinfection and the presence of myelopathy remained (OR = 7.3, p = 0.0003 for scores > or = 2; and OR = 8.9 for scores > or = 4, p = 0.04). In addition, a higher proportion of coinfected subjects had peripheral neuropathy (40%) than controls (16%) (OR = 3.5, p = 0.07). CONCLUSIONS Coinfection with HTLV-I was strongly associated with myelopathy among subjects infected with HIV. The relative contribution of HTLV-I versus HIV in the pathogenesis of coinfection-associated myelopathy is not known. Coinfection may also be associated with peripheral neuropathy. Further studies are needed to elucidate the mechanisms of coinfection-associated neurologic conditions.